Involvement of a Plastid Terminal Oxidase in Plastoquinone Oxidation as Evidenced by Expression of the Arabidopsis thaliana Enzyme in Tobacco*

Chlororespiration has been defined as a respiratory electron transport chain in interaction with photosynthetic electron transport involving both non-photochemical reduction and oxidation of plastoquinones. Different enzymatic activities, including a plastid-encoded NADH dehydrogenase complex, have been reported to be involved in the non-photochemical reduction of plastoquinones. However, the enzyme responsible for plasquinol oxidation has not yet been clearly identified. In order to determine whether the newly discovered plastid oxidase (PTOX) involved in carotenoid biosynthesis acts as a plastoquinol oxidase in higher plant chloroplasts, theArabidopsis thaliana PTOX gene (At-PTOX) was expressed in tobacco under the control of a strong constitutive promoter. We showed that At-PTOX is functional in tobacco chloroplasts and strongly accelerates the non-photochemical reoxidation of plastoquinols; this effect was inhibited by propyl gallate, a known inhibitor of PTOX. During the dark to light induction phase of photosynthesis at low irradiances, At-PTOX drives significant electron flow to O2, thus avoiding over-reduction of plastoquinones, when photo- synthetic CO2 assimilation was not fully induced. We proposed that PTOX, by modulating the redox state of intersystem electron carriers, may participate in the regulation of cyclic electron flow around photosystem I.

In photosynthetic organisms like photosynthetic bacteria or cyanobacteria, photosynthesis and respiration operate in close interaction within the same membranes where they share some electron transport components such as the plastoquinone (PQ) 1 pool (1). In chloroplasts, the existence of a respiratory electron transport chain (chlororespiration) in interaction with photosynthesis has been suggested (2,3), and this activity has been proposed to originate from the cyanobacterial ancestor of chloroplasts (1). Chlororespiration would involve non-photochemical reduction of the PQ pool and subsequent oxidation by a plastoquinol terminal oxidase. Non-photochemical reduction of PQs is a well established phenomenon that likely occurs during cyclic electron transfer reactions around photosystem I (PS I). A plastid-encoded NADH dehydrogenase (Ndh) complex showing homologies with bacterial complex I has been characterized in thylakoid membranes from higher plants (4 -7). Inactivation of some ndh genes using plastid transformation of tobacco showed the involvement of the Ndh complex in nonphotochemical reduction of PQs (8,9). It was proposed that the Ndh complex participates in both chlororespiration and cyclic electron transfer around PS I (8 -11). A role of the Ndh complex in cyclic electron flow around PS I was recently confirmed by photoacoustic measurements performed in tobacco ndh mutants (12). It should be noted that alternate activities, such as a putative ferredoxin-PQ reductase (FQR) (13,14) or a Ndh-2 type activity may also be involved in these processes (15,16).
Nevertheless, the involvement of a plastid terminal oxidase in chlororespiration has been the subject of controversy during the last decade (2,3,16,17). Initially, the existence of chlororespiration was based mainly on the effect of respiratory inhibitors such as cyanide and CO on the redox state of the PQ pool (2). However, such effects can be alternatively explained by an inhibition of mitochondrial respiration and the existence of redox interactions between chloroplasts and mitochondria (16 -20). Recently, the study of an Arabidopsis mutant (immutans) showing a variegated phenotype led to the identification of a protein involved in carotenoid biosynthesis (21,22). Based on sequence homology with mitochondrial alternative oxidases, this protein was suggested to act as a plastid terminal oxidase (PTOX) (21,23). Expression of Arabidopsis PTOX (At-PTOX) in Escherichia coli conferred a cyanide-resistant O 2 uptake sensitive to propyl gallate, a known inhibitor of alternative oxidases (24,25). In PS I-less mutants of the green algae Chlamydomonas reinhardtii, a limited but significant electron flow from photosystem II (PS II) to molecular O 2 was measured. Based on the effects of inhibitors (insensitivity to KCN and CO and sensitivity to propyl gallate) on this process and on the detection of a thylakoid protein that cross-reacted with an antibody raised against PTOX, it was proposed that the chlororespiratory O 2 uptake is because of a Chlamydomonas homologue of PTOX (24). In higher plants, an involvement of PTOX in PQ oxidation has not been experimentally evidenced. Based on the study of a reconstituted system, Casano et al. (6) proposed that a peroxidase using hydrogen peroxide as an electron acceptor may be involved in chlororespiration.
To get further insight into the function of PTOX and in particular to determine whether this protein can achieve quinol oxidation in chloroplasts, tobacco plants constitutively expressing At-PTOX have been generated. We show that At-PTOX facilitates the oxidation of reduced PQs using O 2 as a terminal acceptor.

EXPERIMENTAL PROCEDURES
Plant Material-Tobacco plants (Nicotiana tabacum var. petit havana) were grown on compost in a phytotron (25°C day/20°C night; 12-h photoperiod) under an irradiance of 300 mol photons⅐m Ϫ2 ⅐s Ϫ1 supplied by quartz halogen lamps (HQI-T 400W/DV, Osram, Germany). Plants were watered with a half-diluted Hoagland's nutritive solution.
Production of Transgenic At-PTOX Tobacco Plants Overexpressing At-PTOX-The Arabidopsis thaliana PTOX cDNA (GenBank TM accession number AJ004881) was used as a template for PCR amplification using the primers 5Ј-CCGCTCGAGCCTGACGGAGATGGCGGCGATT-TCAGG-3Ј and 5Ј-CCCGAGCTCTTATTAACTTGTAATGGATTTCTTG-AGGC-3Ј, respectively, containing an XhoI and an SacI restriction site at the 5Ј and 3Ј end. The amplified fragment started 9 bp upstream to the coding sequence of the At-PTOX cDNA and contained two stop codons (the start codon and two stop codons are underlined). After digestion, the amplified fragment was introduced in a sense orientation into a plant binary vector (pKYLX71). Expression of At-PTOX was driven by a double sequence of the cauliflower mosaic virus 35 S-labeled constitutive promoter (26). The recombinant plasmid was introduced by electroporation into Agrobacterium tumefaciens (strain C58), which was used for tobacco transformation employing the standard leaf disc transformation method (27). Two independent transformation experiments were carried out, and six transformants were recovered on a kanamycin-selective medium (100 mg⅐liter Ϫ1 ). Two independent transgenic lines (PTOX 1 ϩ and PTOX 2 ϩ ), overexpressing high amounts of PTOX, were selected and were self-pollinated. The T1 generation was used for further experiments.
Preparation of Osmotically Lysed Chloroplasts for O 2 Exchange and Chlorophyll Fluorescence Measurements-Leaves were harvested at the end of the night period, and intact chloroplasts were isolated at 4°C on a Percoll gradient according to a modification of the method described by Mills and Joy (51). Approximately 30 g of leaves were ground in a blender for 2 s in 100 ml of medium A containing 330 mM sorbitol, 50 mM Tricine-NaOH, pH 7.8, 2 mM EDTA, 1 mM MgCl 2 , 2 mM ascorbic acid, and 5 mM dithiothreitol. After filtration through 250-and 60-m nylon net, followed by centrifugation (2000 ϫ g, 3 min), the crude extract was resuspended in medium A (dithiothreitol-free) and layered onto a Percoll step gradient formed with two layers of medium A containing 90 and 40% (v/v) Percoll, respectively. After centrifugation in a swing out rotor at 3,500 ϫ g for 15 min, intact chloroplasts were recovered from the 40:90% Percoll interphase, washed with 60 ml of medium A, pelleted at 2,000 ϫ g for 3 min, and osmotically lysed by resuspension in 10 mM MgCl 2 and 1 mM phenylmethylsulfonyl fluoride for 30 min. Lysed chloroplasts were diluted at a final concentration of 200 g of chlorophyll⅐ml Ϫ1 in 30 mM Hepes-KOH buffer, pH 7.5, containing 0.3 M sorbitol, 5 mM NaCl, 10 mM MgCl 2 , 2.5 mM NaHPO 4 , 50% (v/v) glycerol, 1 mM phenylmethylsulfonyl fluoride. Aliquots of the chloroplast preparation were stored at Ϫ20°C. For O 2 exchange and chlorophyll fluorescence measurements, aliquots were resuspended in 30 mM Hepes-KOH buffer, pH 7.5, containing 0.3 M sorbitol, 5 mM NaCl, and 10 mM MgCl 2 . (F m Ϫ F 0 )/F m measured on chloroplasts samples was 0.7 (Ϯ0.02, 6 experiments). Anaerobiosis was achieved by addition of glucose (20 mM) and glucose oxidase (2 mg⅐ml Ϫ1 ) to the chloroplast suspension. Reactive oxygen species generated by the glucose oxidase activity were scavenged by adding superoxide dismutase (500 units⅐ml Ϫ1 ) and catalase (1,000 units⅐ml Ϫ1 ).
Mass Spectrometric O 2 Exchange Measurements-For mass spectrometric measurements of O 2 exchange, osmotically lysed chloroplasts (20 g of chlorophyll⅐ml Ϫ1 ) were placed in the measuring chamber (1.5-ml reaction volume). The sample was sparged with N 2 to remove 16 O 2 , and 18 O 2 (95% 18 O isotope content, Euriso-Top, Les Ulys, France) was then introduced to reach an O 2 concentration close to the equilibrium with air. Dissolved gases were introduced into the ion source of the mass spectrometer (model MM 14 -80, VG Instruments, Cheshire, UK) through a Teflon membrane. Light was supplied by a fiber optic illuminator (Schott, Main, Germany) supplying a light intensity of 150 mol photons⅐m Ϫ2 ⅐s Ϫ1 . All gas exchange measurements were performed at 25°C. The use of 18 O 2 allowed the in vivo determination of O 2 evolution by PS II (originating from the photolysis of water which is not enriched) in the presence of O 2 consuming processes.

Chlorophyll Fluorescence Measurements in Chloroplasts and
Leaves-Chlorophyll fluorescence was measured at 25°C using pulsemodulated fluorimeters (PAM 101-103 and PAM 2000, Walz, Effeltrich, Germany for chloroplasts and leaves, respectively). The maximal chlorophyll fluorescence level (F m ) was measured under a 0.8 -1-s saturating pulse (about 8,000 -10,000 mol photons⅐m Ϫ2 ⅐s Ϫ1 ) in dark-adapted leaves, on which the basal fluorescence (F 0 ) was recorded before the pulse. The maximal photochemical yield of PS II was determined as (F m Ϫ F 0 )/F m . Fluorescence levels F m , F s (fluorescence in the light), F mЈ (maximal fluorescence in the light, using a saturating pulse), and F 0Ј (basal fluorescence of light-adapted leaves, recorded after rapid reoxidation of the PQ pool using far-red light) were used to calculate PS II photochemical yield (F mЈ Ϫ F s /F mЈ ), non-photochemical quenching (qN ϭ 1 Ϫ F mЈ /F m ), and photochemical quenching (qP ϭ F mЈ Ϫ F s /F mЈ Ϫ F 0Ј ) under different irradiances (28). Apparent photosynthetic electron transport rates (mol electrons⅐m Ϫ2 .s Ϫ1 ) were estimated as (F mЈ Ϫ F s )/F mЈ ϫ PFFD i ϫ LA ϫ 0.5, where PFFD i is the incident photosynthetic photon flux density; LA is the leaf absorbance (0.84), and 0.5 the factor accounting for the light partition between the two photosystems.
For chlorophyll fluorescence measurements in stripped leaf discs, leaf samples were placed on a wet paper filter at 25°C in ambient air. Chlorophyll fluorescence measurements on attached leaves were performed using the gas exchange cuvette of a Licor gas exchange system (LI-6400, Li-Cor Inc, Lincoln, NE) to control leaf temperature (25°C) and gas atmosphere. Illumination was provided by a homemade red (663 nm) LEDs source.
Inhibitor Treatment of Leaf Discs-Leaf discs were sampled from 5to 8 week-old plants. After stripping the lower epidermis, leaf discs were soaked in water for 60 min. 2,5-Dibromo-3-methyl-6-isopropyl-pbenzoquinone (DBMIB, 50 M final concentration) or propyl gallate (1 mM final concentration) was added diluted in methanol (maximal final methanol concentration was 0.5%). Control leaf discs were soaked in water containing methanol. It should be noted that the DBMIB concentration (50 M) used for leaf discs was much higher than the concentration generally used to obtain specific inhibition of the cytochrome b 6 /f complex on isolated thylakoids (1 M). Despite this relatively high concentration necessary to obtain an effect in leaf discs, we checked that DBMIB did not act as an electron acceptor.
Photosynthetic CO 2 Fixation Measurements on Attached Leaves-Net CO 2 exchange measurements were performed on attached leaves using a portable gas exchange system (LI-6400, Li-Cor Inc, Lincoln, NE) and a homemade red (663 nm) LEDs source. Leaf temperature was maintained at 25°C, and leaf vapor pressure deficit was maintained around 0.8 kPa. Various O 2 and N 2 concentrations were provided by mixing pure gases. O 2 concentration was measured using a paramagnetic O 2 analyzer (MAIHAK, Hamburg, Germany). This mixing system was also used for fluorescence measurements in attached leaves. Quantum yield of CO 2 fixation in air and under non-photorespiratory conditions (O 2 1.5% (v/v); CO 2 750 l⅐liter Ϫ1 ) were calculated from the slope of the linear portion of the light response curve (5 measurements at irradiances between 40 and 80 mol photons⅐m Ϫ2 ⅐s Ϫ1 and 10 measurements at irradiances between 10 and 100 mol photons⅐m Ϫ2 ⅐s Ϫ1 for air and non-photorespiratory conditions, respectively).
Electrophoresis and Western Analysis on Chloroplast Fractions-Intact chloroplasts were isolated and purified from leaves using discontinuous Percoll (Amersham Biosciences) gradients as described previously (29). Chloroplasts were osmotically lysed in a solution containing 20 mM MES, pH 6.0, 15 mM NaCl, and 5 mM MgCl 2 and centrifuged for 20 min at 35,000 ϫ g. Stroma lamellae and grana membranes were separated following a stacking step carried out as described previously (4).
To prepare total insoluble proteins, tobacco leaves (1 g fresh weight) were frozen in liquid nitrogen and ground to a fine powder with a chilled pestle and mortar. The powder was resuspended in a 5-ml extraction buffer (50 mM Tris-HCl, pH 8.0) containing 50 mM ␤-mercaptoethanol and 1 mM phenylmethylsulfonyl fluoride. After 30 min of stirring (4°C) and centrifugation (40,000 ϫ g for 20 min), the pellet was resuspended in the same buffer containing 1% SDS. After 30 min of stirring (4°C) and centrifugation (40,000 ϫ g for 20 min), proteins contained in the supernatant were precipitated with acetone (80% final concentration).
Transgene Transcript Analysis-RT-PCR analysis of At-PTOX transcripts was carried out as described previously (25)  Protein and Chlorophyll Determination-Protein content was determined using a modified Lowry method (Sigma). Chlorophyll content was measured according to the method of Lichtenthaler and Wellburn (31).

RESULTS
Expression of At-PTOX in Tobacco-Transgenic tobacco plants expressing the At-PTOX cDNA sequence under the control of the doubled constitutive 35 S promoter of the cauliflower mosaic virus were generated by Agrobacterium-mediated transformation. Two lines, PTOX 1 ϩ and PTOX 2 ϩ , showing a particularly strong expression of the transgene were selected among six transformant lines and were further studied (Fig.  1A). Note that although no signal was observed in WT tobacco (Fig. 1A, upper panel), a faint band was detected after reamplification ( Fig. 1A, lower panel). We checked that amplified RT-PCR fragments, including the faint band amplified in WT tobacco (Fig. 1A), cross-hybridized with the At-PTOX probe by Southern analysis (data not shown). Antibodies raised against At-PTOX were used to characterize At-PTOX expression in tobacco transgenic lines using Western analysis. Both transformant lines showed large amounts of a 41-kDa band corresponding to At-PTOX in total insoluble leaf proteins (25), whereas no signal was observed in wild type (Fig. 1B). In both lines, At-PTOX was targeted to the chloroplasts, thanks to the presence of an N-terminal transit peptide (22), and was found to be associated with thylakoid membranes, essentially stroma lamellae, with only small amounts being found in grana (Fig.  2). Subsequent experiments were performed on both PTOX 1 ϩ and PTOX 2 ϩ lines and yielded similar results. Because PTOX has been reported previously (21,22) to be involved in carotenoid biosynthesis, the pigment content of transgenic plants was analyzed. High pressure liquid chromatography measurements did not reveal any significant difference in chlorophyll or carotenoid content in WT and PTOX ϩ extracts (data not shown). In addition, after transfer to high light conditions, similar amounts of xanthophyll cycle carotenoids (violaxanthin, zeaxanthin, and antheraxanthin) were found in both plants. PTOX 1 ϩ and PTOX 2 ϩ plants did not show any particular phenotype, and growth was comparable with WT plants when cultivated under normal conditions (not shown).
Expression of At-PTOX Suppresses the Post-illumination F 0 Fluorescence Increase-When intact WT leaves were illuminated for a few minutes and then placed in the dark, a transient increase in the F 0 chlorophyll fluorescence level occurred (Fig. 3A) (see Refs. 32 and 33). The post-illumination fluorescence transient was absent in PTOX ϩ leaves, and the F 0 fluorescence level rapidly decreased after switching off the light (Fig. 3B). As reported previously, the fluorescence increase was absent in Ndh-less mutants (Fig. 3C), but interestingly the fluorescence signal decreased more slowly that in PTOX ϩ . The absence of a post-illumination chlorophyll fluorescence increase in Ndh-less mutants was interpreted as the involvement of the Ndh complex in the re-reduction of the PQ pool occurring in the dark after a period of illumination (8,9,34). This experiment suggests that like the Ndh complex At-PTOX was able to modulate the redox state of PQ in the dark, most likely by oxidizing reduced plastoquinones. In agreement with this interpretation, when leaf discs were treated with propyl gallate, a potent inhibitor of PTOX (24), a reversal of the loss of the F 0 fluorescence rise was observed (data not shown). Subsequent experiments were designed to characterize the role of At-PTOX in PQ oxidation.
Involvement of At-PTOX in the Dark Oxidation of the PQ Pool-In the experiment described in Fig. 4, chlorophyll fluorescence changes were measured in dark-adapted leaves in response to a saturating light pulse. During a pulse, PS II primary electron acceptors were fully reduced, and chlorophyll fluorescence rapidly reached a maximum level (F m ). After the light pulse, the chlorophyll fluorescence level decreased in the dark, and this decay was related to the reoxidation of PS II primary acceptors (Q A ) in redox equilibrium with the PQ pool. The fluorescence decay was clearly biphasic. The fast phase was similar in WT and PTOX ϩ . On the other hand, the slowly decreasing phase was much faster in PTOX ϩ than in WT, indicating that PQs were more efficiently reoxidized in transgenic plants. Addition of propyl gallate severely slowed down the fluorescence decay, which came close to that observed in WT leaves (Fig. 4B). On the other hand, cyanide (KCN 1 mM) had no significant effect on the fluorescence decay measured in PTOX ϩ (data not shown). In order to check that PS II acceptors were more reduced in WT than in PTOX ϩ during the fluorescence decay shown on Fig. 4A, a control experiment was performed by flashing a second light pulse 4 s after the first pulse (Fig. 5). Under such conditions, because no non-photochemical quenching of F m occurred, the upper area delimited by the fluorescence induction curve reflected the relative pool size of electron acceptors of PS II, mainly the PQ pool (2,35). Fig. 5A shows that in WT leaves, 4 s after the first pulse, PS II acceptors are more reduced than in dark-adapted leaves. In contrast, the redox state of PS II acceptors measured in PTOX ϩ leaves 4 s after a pulse illumination was close to that measured in dark-adapted leaves (Fig. 5B). We concluded from these experiments that At-PTOX was functional in transgenic tobacco leaves and was able to oxidize efficiently reduced PQs following their reduction by a saturating light pulse. We found that propyl gallate slightly (but in a reproducible manner) affected the slow phase of the fluorescence decay measured in WT leaves (Fig. 4A), possibly indicating the contribution of a putative tobacco PTOX in PQ oxidation.
At-PTOX Is Active in Thylakoids and Used Molecular O 2 as a Substrate-The activity of PTOX on PQ oxidation was then investigated in chloroplast preparations. Addition of exogenous NADH to osmotically lysed chloroplasts isolated from WT leaves increased the apparent F 0 chlorophyll fluorescence level measured under low non-actinic light, indicating an increase in the redox state of the PQ pool (Fig. 6A). Note that in chloroplast preparations, NADH-induced PQ reduction was not mediated by the Ndh complex (Ndh-1), which likely inactivated during the extraction procedure, but rather by an alternative (Ndh-2 like) activity (15,16,34). Under aerobic conditions, the NADHinduced fluorescence increase was significantly slower in PTOX ϩ than in WT chloroplasts (Fig. 6B). Addition of propyl gallate increased the chlorophyll fluorescence level in PTOX ϩ chloroplasts, whereas no significant effect could be detected in WT chloroplasts. Removing O 2 from the sample strongly increased the chlorophyll fluorescence signal in a similar manner in both WT and PTOX ϩ chloroplasts, and the F m level corresponded to a full reduction of PQs being rapidly reached (Fig. 6,  A and B). This experiment showed that At-PTOX was functional in isolated tobacco chloroplasts and that the redox state of the PQ pool resulted from a competition between reduction by NADH and oxidation by PTOX.
Mass spectrometric measurements of O 2 exchange were then performed on chloroplast preparations using 18 O 2 , to determine which electron acceptor was used during PTOX-mediated PQ oxidation (Fig. 7). In the absence of either cytochrome b 6 /f or PS I, an electron flow from PS II to O 2 involving PTOX occurred in Chlamydomonas cells (36). When tobacco chloroplasts were treated with DBMIB (a potent inhibitor of the cytochrome b 6 /f complex) and illuminated, simultaneous O 2 production by PS II and O 2 uptake were observed using 18 O 2 and mass spectrometry. PS II activity was higher in PTOX ϩ than in WT (Fig. 7). Addition of 25 M of the PS II inhibitor 3-(3,4-dichlorophenyl)-1,1-dimethylurea fully suppressed the electron transfer activity, showing the involvement of the PQ pool in both cases. Treatment by propyl gallate largely suppressed the difference in electron transfer activity observed between WT and PTOX ϩ , showing that this difference was likely to be due to the activity of the oxidase. In these experiments, the PS II-mediated electron flow was balanced by a simultaneous increase in the O 2 uptake rate, thus supporting the view that, as in the case of Chlamydomonas PTOX, At-PTOX is a true quinol oxidase, using O 2 as an electron acceptor and releasing H 2 O as a final product (24).

Measurement of a PS II-mediated Electron Flow to At-PTOX in
Leaves-Similar fluorescence experiments were performed in stripped leaf discs treated with DBMIB, measuring chlorophyll fluorescence in order to probe PS II activity (Fig. 8). In the absence of inhibitor, similar electron flow rates were observed at low irradiances both in WT and transgenic leaf discs. In WT leaf discs, DBMIB strongly inhibited linear electron flow (90% inhibition at 75 mol photons⅐m Ϫ2 ⅐s Ϫ1 ), whereas in transgenic leaf discs inhibition by DBMIB was much less pronounced (60% inhibition at 75 mol photons⅐m Ϫ2 ⅐s Ϫ1 ). The DBMIB-insensitive electron flow observed in PTOX ϩ was inhibited by propyl gallate (Fig. 8) and reached basal rates measured in WT discs treated with propyl gallate. This experiment showed that in leaves placed under low light intensity, when linear electron flow to PS I was inhibited, a significant part of PS II-driven electron flow (about 35% of the maximal electron flow to PS I) can be directed toward PTOX and O 2 .
Involvement of PTOX during Photosynthesis-We were then interested to determine whether the activity of At-PTOX, which can be evidenced either in the dark (Figs. 3 and 4) or in the light in the absence of functional electron transfer to PS I (Fig. 8), could be observed in the light during normal conditions of photosynthesis. During a dark to light induction of photosynthesis, typical variations in chlorophyll fluorescence were observed (37). Under low light intensity, these variations reflected changes in the electron transfer rate occurring during the activation of photosynthesis. The transient increase in fluorescence commonly observed in WT during the induction phase reflected the transient accumulation of plastoquinols due to the initial absence of PS I electron acceptors. In fact, an activation of the PS I acceptor side and of Calvin cycle enzymes was generally required to initiate CO 2 assimilation and further reoxidize NADPH. This transient was almost completely abolished in PTOX ϩ at the lowest irradiance (8 mol photons⅐m Ϫ2 ⅐s Ϫ1 , Fig. 9A), indicating a highly efficient plastoquinol oxidation before the activation of PS I. After a few minutes of illumination, both F s and F mЈ values were identical in WT and PTOX ϩ . At this low irradiance, F mЈ was close to F m , showing the absence of non-photochemical quenching. At higher light intensity (50 mol photons⅐m Ϫ2 ⅐s Ϫ1 ), a difference between WT and PTOX ϩ was also observed, F s values remained lower in PTOX ϩ than in WT during the first 3 min of illumination (Fig. 9B). When illumination was prolonged, the decrease in F s was more pronounced in WT, and after 10 min reached a lower level than in PTOX ϩ . It should be noted that variations in F s values were accompanied by concomitant changes in F mЈ (Fig. 9B). As a consequence, both non-photochemical (qN) and photochemical (qP) quenching parameters were lower in PTOX ϩ than in WT after 10 min of illumination. This effect on qN and qP was also observed at higher irradi- ances (for example at 750 mol photons⅐m Ϫ2 ⅐s Ϫ1 ; Table I). However, at all irradiances, after 1 h of illumination when a steady state was reached, qN and qP of WT and PTOX ϩ became identical (Table I). Despite these fluctuations in fluorescence quenchings, no significant differences in PS II photochemical yields, measured either at 10 or 60 min, could be evidenced between WT and PTOX ϩ (Table I). Moreover, measurements of net CO 2 gas exchange at steady state showed no significant difference in quantum yield of CO 2 fixation in air as well as under non-photorespiratory conditions (Table II). In WT and PTOX ϩ , rates of CO 2 fixation measured at saturating irradiance were also similar (Table II). DISCUSSION We have shown in this paper that when expressed in tobacco, At-PTOX is targeted to the chloroplasts and functions as a PQ oxidase. The activity of At-PTOX could be evidenced in intact leaves, following either photochemical or non-photochemical reduction of PQs and also in thylakoids, when PQs were reduced by exogenous NADH. Based on chlorophyll fluorescence and mass spectrometric measurements performed on thylakoids, we propose that At-PTOX drives PQ oxidation using molecular O 2 as a terminal electron acceptor. This agrees with previous conclusions reached from mass spectrometric measurements on Chlamydomonas mutants deficient in PS I (24). Because the Ndh complex is involved in the non-photochemical reduction of the PQ pool (8,9,34) (see Fig. 3C) and At-PTOX is involved in its non-photochemical oxidation, we conclude that a chlororespiratory electron transfer involving the plastid Ndh complex, the PQ pool, and At-PTOX occurs from NAD(P)H to O 2 in chloroplasts of transgenic tobacco expressing At-PTOX. In the dark, the redox status of PQs therefore depends on an equilibrium between its reduction by the Ndh complex and oxidation by PTOX.
In thylakoid membranes, PS I reaction centers and ATPase complexes are essentially located in stroma lamellae, whereas PS II are restricted to grana, cytochrome b 6 /f complexes being found in both types of membranes. Like the Ndh complex (11,38,39), At-PTOX was found mainly in stroma lamellae, indicating that chlororespiration is restricted to stroma lamellae and is absent in granal thylakoids. Previously, the involvement of a propyl gallate-sensitive PQ oxidase in chlororespiration had been evidenced in Chlamydomonas cells (24). It was proposed that an At-PTOX homologue was functional in Chlamydomonas thylakoid membranes (24), but the corresponding gene has not yet been identified (16). In higher plants, first evidence for the existence of chlororespiration was based on the effect of respiratory inhibitors such as cyanide (40) or CO (41). Such effects cannot be explained by the inhibition of PTOX, because this protein was reported to be insensitive to these compounds (24,25). This was confirmed in this study by the insensitivity to cyanide of the slow phase of the chlorophyll fluorescence decay. Therefore, the effects of respiratory inhibitors such as cyanide or CO more likely result from the inhibition of mitochondrial respiration that has been reported to affect the redox state of the PQ pool due to the existence of redox interactions between chloroplasts and mitochondria (16,42). Such effects may alternatively reflect the existence of an alternative PQ oxidation pathway sensitive to cyanide and CO. In this respect, Fig. 7 indicates the existence in chloroplasts of a propyl gallateinsensitive mechanism for PQ oxidation. Recently, Casano et al. (6), studying a reconstituted system containing the Ndh complex and a plastidial hydroquinone peroxidase, proposed the existence of a PQ oxidation pathway using hydrogen peroxide as a terminal acceptor.
In addition to an involvement in dark reactions, we have  shown that At-PTOX may interact with photosynthetic electron transport reactions in illuminated leaves. In WT plants, a transient over-reduction of photosynthetic electron carriers occurs during the induction phase of photosynthesis. This is due to the fact that the photosynthetic carbon reduction cycle is not operative in the dark, because some of the enzymes of the cycle require light-induced activation by reduced thioredoxins (43).
In transgenic tobacco plants expressing At-PTOX, the transient over-reduction of photosynthetic electron carriers is greatly decreased, indicating that electrons are diverted to O 2 via PTOX. This suggests that PTOX can potentially prevent over-reduction of PQs in the light. In plant mitochondria, alternative oxidase has been suggested to function as an "energy overflow," its activity being increased when the cytochrome pathway is saturated with electrons (44). Overexpression of alternative oxidase in this organelle has been shown to limit the generation of reactive oxygen species by preventing overreduction of electron carriers (45). It should be noted, however, that expression of At-PTOX did not result in increased resistance of transgenic lines to photoinhibition (data not shown). Differences in qN and qP values between WT and PTOX ϩ were transitorily observed during the 10 -30-min period of illumination, whereas the photochemical yield of PS II remained identical in both WT and PTOX ϩ . The fact that both photochemical yield of PS II and rate of CO 2 fixation are identical suggests that at the end of the transitory induction period of photosynthesis, the oxidase function of PTOX does not contribute to drive significant electron flow compared with photosynthetic carbon reduction and oxidation cycles. On the other hand, lower qN and qP in PTOX ϩ between the initial induction period and steady state suggests that the pH gradient is lower and that PS II acceptors (Q A ) are more reduced compared with the WT. A lower pH gradient could indicate that cyclic electron reactions around PS I are down-regulated in PTOX ϩ . Cyclic electron reactions around PS I have been reported to be controlled by the redox poise of some electron carriers; this effect was possibly mediated by molecular O 2 (46). Overexpression of PTOX, by modifying the redox poise of intersystem electron carriers, may perturb the establishment of cyclic electron transfer reactions. Interestingly enough, a role of chlororespiration in the control of cyclic electron flow around PS I was recently deduced from photoacoustic measurements performed in leaves under low O 2 concentration (12). The fact that chlororespiration and cyclic electron transfer reactions around PS I operate within the same membranes (stroma lamellae, see Ref. 16) further strengthens the hypothesis of a functional link between these two activities.
At steady state, qP and qN values were similar in WT and PTOX ϩ , indicating that both the redox state of Q A and the pH gradient reached similar levels. This may indicate that at steady state the contribution of cyclic electron flow around PS I is decreased compared with its high activity during the induction phase when terminal electron acceptors are not fully available. Alternatively, this effect might reflect the involvement of regulatory mechanisms that could be turned on under these conditions. For instance, the expression of some nuclear genes, like cab genes encoding light harvesting complex apoprotein, has been shown to be controlled by the redox state of PQs (47). The higher reduction of the PQ pool observed in PTOX ϩ during the induction of photosynthesis may trigger such long term adaptation processes and explain why similar pH gradients and Q A redox state are finally reached at steady state in both types of plants. Analysis of gene expression in PTOX ϩ plants should inform us of the possible existence of such adaptive mechanisms.
If the role of At-PTOX in PQ oxidation could be demonstrated in transgenic tobacco, the involvement of a functional PTOX in WT tobacco appears more difficult to establish. A faint band, specific to the native ptox transcripts, was amplified in WT tobacco by RT-PCR (Fig. 1A). However, by using an antibody raised against At-PTOX, no signal corresponding to native PTOX was detected in insoluble proteins prepared from WT tobacco leaves (Fig. 1). This may be due to the fact that either the antibody raised against the Arabidopsis enzyme does not cross-react with the tobacco enzyme or that the native enzyme is present in too small an amount to be detected. The latter hypothesis is the most probable, because this antibody crossreacts with chromoplast preparations from pepper, another Solanacae species (25), and also with chloroplast preparations from C. reinhardtii (24). In this respect, a doublet that may correspond to the native tobacco PTOX was detected in purified stroma lamellae preparations probed with the Arabidopsis antibody (data not shown). It should be noticed that the plastid Ndh complex, the other probable component of chlororespiration, has been reported to be present in leaves in very low amounts (4,39). The slight effect of propyl gallate on the slow phase of the chlorophyll fluorescence decay measured in WT leaves (Fig. 3), may reflect a contribution of the native tobacco PTOX to the oxidation of PQs. In agreement with this interpretation, it has been reported recently (12) that in tobacco leaves the re-reduction rate of the oxidized primary electron donor in PS I (P 700 ϩ ) is increased by propyl gallate. This effect was interpreted as the re-routing of electrons toward PS I when the putative tobacco plastid terminal oxidase is inhibited. Since PTOX most likely represents a minor component of thylakoid membranes, at least when plants are grown under normal conditions, a regulatory role (for instance in the control of cyclic electron flow) seems more probable than a direct bioenergetic role. However, more work remains to be done to determine clearly the involvement of native PTOX in leaves.
Different lines of evidence suggest that PTOX might become more abundant at particular developmental stages or under particular growth (or stress) conditions. In higher plant chloroplasts, the role of PTOX in carotenoid biosynthesis has been demonstrated from the analysis of Arabidopsis and tomato mutants (21,22,25). The variegated phenotype of the Arabidopsis mutant immutans was explained by an involvement of PTOX in phytoene desaturation, an important step in carotenoid biosynthesis occurring during the early stage of the greening process (21,22). As suggested previously (16,19), native PTOX might be more abundant in non-green plastid under conditions where the photosynthetic apparatus is not functional. High amounts of PTOX were reported in achlorophyllous membranes prepared from chromoplasts of red pepper fruits, where carotenoid biosynthesis is particularly active (25). Overexpression of At-PTOX did not influence the leaf carotenoid content, thus indicating that the PTOX level is not a limiting factor regulating carotenoid biosynthesis. Interestingly, the IM (or PTOX) promoter was shown to be active, and IM mRNAs were expressed ubiquitously in Arabidopsis tissues and organs throughout development, arguing in favor of a more global role for this protein in plastid metabolism (48).
In C 4 plants, subunits of the Ndh complex have been reported to be much more strongly expressed in bundle sheath chloroplasts than in mesophyll chloroplasts (49). In bundle sheath chloroplasts, only low levels of PS II are detected. In these cells, ATP required for CO 2 fixation is generated by cyclic electron transport around PS I. Interestingly, it was recently reported that bundle sheath cells from C 3 leaves have photosynthesis features close to those of C 4 leaves (50). It will be interesting to determine whether ndh and ptox genes are more strongly expressed in bundle sheath chloroplasts of C 3 plants than in mesophyll cells and participate in the regulation of cyclic electron transfer reactions around PS I.