Mitochondrial ascorbate synthesis acts as a pro-oxidant pathway and down-regulate energy supply in plants

Attempts to improve the ascorbate (AsA) content of plants are still dealing with the limited understanding of why exists a wide variability of this powerful anti-oxidant molecule in different plant sources, species and environmental situations. In plant mitochondria, the last step of AsA synthesis is catalyzed by the enzyme L-galactone-1,4-lactone dehydrogenase (L-GalLDH). By using GalLDH-RNAi silencing plant lines, biochemical and proteomic approaches, we here discovered that, in addition to accumulate this antioxidant, mitochondria synthesize AsA to down-regulate the respiratory activity and the cellular energy provision. The work reveals that the AsA synthesis pathway within mitochondria is a branched electron transfer process that channels electrons towards the alternative oxidase, interfering with conventional electron transport. It was unexpectedly found that significant hydrogen peroxide is generated during AsA synthesis, which affects the AsA level. The induced AsA synthesis shows proteomic alterations of mitochondrial and extra-mitochondrial proteins related to oxidative and energetic metabolism. The most identified proteins were known components of plant responses to high light acclimation, programmed cell death, oxidative stress, senescence, cell expansion, iron and phosphorus starvation, different abiotic stress/pathogen attack responses and others. We propose that changing the electron flux associated with AsA synthesis might be part of a new mechanism by which the L-GalLDH enzyme would adapt plant mitochondria to fluctuating energy demands and redox status occurring under different physiological contexts.


INTRODUCTION 51
In plants, the mitochondrial electron transport chain (mETC) consists of a series of electron 52 transporters that function to oxidize reducing equivalents, NADP(H) and FADH 2 (Schertl and Braun, 53 2014). A widely accepted model about the electron transfer is that the electrons normally enter 54 via complex I (NADH:ubiquinone oxidoreductase) or through a diversity of "alternative" NAD(P)H 55 dehydrogenases using flavin mononucleotide (FMN) as electron acceptor (Pineau et al., 2005). 56 Alternatively, complex II (succinate:ubiquinone oxidoreductase) and other dehydrogenases such as 57 glyceraldehyde 3-phosphate dehydrogenase (G3-PDH), the "electron transfer flavoprotein-58 ubiquinone oxidoreductase" (ETFQ-OR) and the proline dehydrogenase (ProDH) supply electrons 59

RESULTS 121
The alternative respiration is modulated by mitochondrial ascorbate synthesis 122 To answer the question how the L-GalLDH enzyme affects the mitochondrial electron transport 123 chain (mETC), we examined the effects of respiratory inhibitors on mitochondrial respiration of 124 RNAi-plant lines harboring silenced L-GalLDH activity. As expected, the mixture of AOX and COX 125 inhibitors (5 mM SHAM and 3 mM NaN 3 ) decreased the oxygen uptake rate of leaf mitochondria 126 purified from wild type plants and L-GalLDH-RNAi plant lines ( Figure 1A). Residual oxygen uptakes 127 were observed in presence of both inhibitors. When leaf mitochondria were pre-treated with the 128 L-GalLDH substrate (5 mM L-GalL), absolute respiration was greatly reduced and was not sensitive 129 to the mixture of both inhibitors in wild type mitochondria. Nonetheless, a significant blockage of 130 respiration occurred in the L-GalL-treated leaf mitochondria from L-GalLDH-RNAi plant lines 131 ( Figure 1A), which suggest that part of the AOX and COX pathways is active. The western blot 132 analysis showed lower levels of L-GalLDH protein abundance in both L-GalLDH-RNAi plant lines 133 (~21% and ~63% of Wt for 8-14 and 5-13 plant lines, respectively), which resulted in decreased L-134 GalLDH activity ( Figure 1B). Notably, the level of L-GalLDH suppression but not the enzyme activity 135 was more marked in the L-GalLDH-RNAi line 8-14 as compared to 5-13 line. 136 To further explore the causes of the differences in respiratory rates, we analyzed the flow of 137 electrons through the AOX pathway in the presence of L-GalLDH substrate. Clearly, the Figure 1C  138 shows that leaf mitochondria of the L-GalLDH-RNAi plants had significant alternative respiration ~9 139 nmol O 2 /mg protein was resistant to NaN 3 and sensitive to SHAM in presence of L-GalL. However, 140 it was not detected in wild type mitochondria ( Figure 1C Figure 1D). Taken together, these data suggest that the AOX molecules in wild type 154 mitochondria are significantly inhibited by L-GalL due to the L-GalLDH activity. In the plant lines, 155 the suppression of L-GalLDH enzyme could, in turn, prevent AOX inhibition by L-GalL. 156 Interestingly, the 8-14 plant line, which has the higher L-GalLDH suppression ( Figure 1B) showed 157 the lower level of oxidized AOX ( Figure 1D). 158 We examined the respiratory capacity of mitochondria purified from heterotrophic tissues (fruits) 159 of other plant species and using other respiratory substrates and inhibitors. The mitochondrial 160 preparations of fruit purified with a Percoll density-gradient method had intact mitochondria 161 (≥80% integrity). The content of mitochondria (based on mitochondrial protein) ranged from 1 to 162 3.4 mg protein. The oxygen uptake of papaya, strawberry and tomato fruit mitochondria was 163 blocked by respiratory inhibitors (supplementary data IIA). However, respiration was insensitive to 164 inhibitors in the presence of L-GalL (supplementary data IIA). Moreover, when energizing 165 mitochondria with other substrates that enter electrons through complexes I (malate, glutamate) 166 or II (succinate), the alternative respiration was blocked by L-GalL (supplementary data IIB). 167 Clearly, the insensitivity of oxygen uptake to inhibitors (supplementary data IIA) and the loss of 168 alternative respiration in presence of L-GalL (supplementary data IIB) were responses in fruit 169 mitochondria that resembled to those found in wild type tomato leaf mitochondria treated with L-170 GalL ( Figure 1A and 1C). It supports the hypothesis of that the L-GalLDH activity down-regulates 171 mitochondrial electron flux by inhibiting the alternative oxidase pathway. Clearly, this is a general 172 effect in both autotrophic and heterotrophic plant tissues. 173 To get further insights about the mechanism inactivating AOX pathway, we adopt papaya fruit 174 mitochondria as model because their ability to synthesize AsA and the significant bulk of active 175 mitochondria with high AOX capacity that can be easily obtained our results here and (Oliveira et 176 al., 2015). Respiration and ascorbate production of papaya mitochondria were stimulated by 177 increasing L-GalL concentrations up to about 5mM. Higher concentrations were progressively 178 inhibitory, being the respiratory activity more sensitive to the inhibition by the substrate 179 concentration (supplementary data III). 180 The alternative oxidase but not the Cytc oxidase is critical for AsA biosynthesis 181 By using inhibitors that target specific points in the mETC, we analyzed the possible role of 182 terminal oxidases during mitochondrial AsA synthesis. The current Bartoli's model explaining 183 mitochondrial AsA synthesis implies that Cytc and Cytc oxidase are absolute requirements for AsA 184 production ( Bartoli et al., 2000). As Cytc oxidase re-oxidizes Cytc quickly, the L-GalLDH activity, 185 which was measured as rate of Cytc reduction, is assayed in presence of Cytc oxidase inhibitor. It 186 was confirmed that the treatment of mitochondria with the inhibitor of Cytc oxidase (NaN 3, azyde) 187 led to over-accumulation of reduced Cytc (~6 µmol cytc.min -1 mg protein -1 , Figure IIB), consistent 188 with a lower Cytc re-oxidation by this terminal oxidase. However, azyde-treated mitochondria still 189 maintained a little capacity to synthesize ascorbate (~0.35 µg AsA mg protein -1 , Figure IIA). This 190 suggested that part of AsA synthesis could be independent of Cytc oxidase. On the other hand, the 191 addition of the inhibitor of AOX pathway (SHAM) affected drastically the Cytc reduction by L-GalL 192 (<1 µmol cytc.min -1 mg protein -1 , Figure IIB), and provoked a very low level of AsA content ( Figure  193 IIA), it suggests that SHAM limits electron flux through Cytc. 194 As AOX gene expression and capacity increase during papaya fruit ripening (Oliveira et al.,195 2015),we comparatively analyzed the L-GalLDH activity between green-mature and fully ripe 196 papaya fruit. The mitochondria from ripe fruit showed lower L-GalLDH activity but had increased 197 AsA synthesis capacity ( Figure IIC). 198 In addition, other inhibitors also showed significant effects during AsA synthesis. Mitochondria 199 treated with antimycin A, an inhibitor of complex III, showed the higher value of Cytc reduction (~7 200 µmol cytc.min -1 mg protein -1 ) ( Figure  Cytc oxidase is a main factor affecting FAD recycling during AsA synthesis 205 We followed changes in fluorescence of exogenously supplemented FAD in the presence of L-GalL 206 and then assessed the effect of inhibitors on such changes. It was recorded a variation of FAD 207 fluorescence (about 2500 Units) following incubation with L-GalL ( Figure 2D). It indicates that FAD 208 redox state changes during AsA synthesis. All inhibitors tested in this study decreased the effect of 209 L-GalL in FAD fluorescence, having the Cytc oxidase inhibitor, azyde, the highest effect ( Figure 2D). 210 These data may suggest that the FAD redox state during AsA synthesis is basically controlled by 211 Cytc oxidase, but other respiratory components could be also involved, albeit indirectly. 212 Mitochondrial uncoupling and ROS over-production are associated with low AOX capacity during 213

AsA synthesis 214
Given that alterations of mETC and AOX pathway may affect the pumping of H + and the 215 mitochondrial coupling (Millar et al., 2011), we explored if mitochondrial oxidative 216 phosphorylation is also affected during AsA synthesis. As expected, there was a 19% of membrane 217 depolarization (based on the respiratory increase induced by the uncoupling agent, CCCP) in 218 NADH-respiring mitochondria. However, mitochondrial respiration was insensitive to CCCP in 219 presence of L-GalL (Table I), suggesting that the generation of the proton gradient is affected. 220 Moreover, both the phosphorylation efficiency, measured as ADP:O ratio and the mitochondrial 221 coupling efficiency, determined as RCR, decreased in presence of L-GalL (Table I). Intriguingly, 222 these L-GalL-dependent alterations correlated with higher H + -ATPase activity of complex V and an 223 unexpected higher mitochondrial capacity to reduce NAD + into NADH (Table I). As the ubiquinone 224 redox state and the mitochondrial energy production are regulated by AOX pathway 225 (Vanlerberghe, 2013), we analyzed possible changes in ubiquinone redox state and the 226 mitochondrial ability to synthesize ATP. It was noted that, in the presence of L-GalL, the 227 mitochondrial capacity to maintain UQ in its reduced state (UQH 2 ) enhanced (about three times 228 more reduced ubiquinone in the L-GalL treatment than in control). Besides, the mitochondrial ATP 229 synthesis capacity was inhibited by L-GalL (Table I). These results suggest that AsA synthesis could 230 cause the over-reduction of mETC and UQ pool, resulting in a decrease (~20%) in ATP synthesis 231 capacity. 232 As AOX is an important ROS scavenger, we hypothesized that the inactivation of AOX pathway 233 during AsA synthesis would enhance ROS. By measuring the H 2 O 2 level, using the Amplex Red 234 method, an increased H 2 O 2 formation was detected within 5-15 min following incubation of 235 mitochondria with L-GalL, having maximal H 2 O 2 increases between 5-20 mM L-GalL whereas 236 response was extremely low or non-detected at concentrations below 5 mM L-GalL 237 (supplementary data IV). 238 We explored the possible sources of mitochondrial ROS during AsA synthesis. Figure  We also demonstrate the lower production of H 2 O 2 in fruit mitochondria from L-GalLDH-RNAi 256 plant lines, which was consistent with an increased AOX respiration during AsA synthesis 257 (supplementary data V). However, despite these mitochondria showed decreased L-GalLDH 258 activity (lower Cytc reduction rate), their abilities to produce AsA and alter FAD redox status were 259 similar to that of wild type fruit mitochondria (supplementary data V). It suggests that AOX 260 pathway may sustain AsA synthesis in mitochondria with low L-GalLDH activity by reducing H 2 O 2 261 level. 262 To further examine the role of alternative respiration during AsA synthesis, we performed an 263 opposite experiment in which the AOX is previously activated before the treatment with L-GalL. To 264 this, mitochondria were firstly treated with pyruvate, a known allosteric AOX activator, and 265 subsequently L-GalL was added to inhibit alternative respiration. Surprisingly, the AOX respiration 266 (7.6 nmol O 2 mg min -1 protein -1 ) was not inhibited in the presence of pyruvate (supplementary data 267 VI). Moreover, this lack of inhibitory effect was not related to lower L-GalLDH activity given that its 268 capacity to reduce Cytc remains in presence of pyruvate. However, the higher AOX capacity 269 correlates with lower H 2 O 2 production and enhanced AsA synthesis (supplementary data VI). 270 The possibility of that AsA synthesis leads to shifts in the overall functional status of cell was 271 tested by performing a comparative proteomic analysis between untreated and L-GalL-treated 272 papaya fruit tissue. Of the set of 53 proteins identified, 24 (45%) and 29 (55%) were up-regulated 273 and down-regulated by L-GalL, respectively (Table II). Possible roles of these proteins will be 274 discussed later with regards to an involvement of AsA synthesis in ROS and energy metabolism as 275 well as in the plant responses to abiotic and biotic stresses. 276

Regulation of seedling emergence by L-GalLDH is associated with altered ATP content 277
To get further insights about the physiological role of L-GalLDH on seedling establishment (an 278 energy-demanding process) , we evaluated AsA synthesis and ATP content in both L-GalLDH-RNAi 279 lines and wild type in germinating seeds and in seedlings that reach the autotrophy capacity. Wild 280 type and L-GalLDH-RNAi seedlings showed similar ability to synthesize AsA, but, wild type ones 281 contain less ATP in dark ( Figure 4A). Interestingly, wild type germinating seeds also had a little less 282 ATP content (data not shown). Treatment of seeds with L-GalL inhibited wild type seedling 283 emergence and consequently, they showed shorter seedlings one-week post germination ( Figure  284 4B). However, this inhibitory effect of L-GalL was not evident in L-GalLDH-RNAi lines and these 285 seedlings elongated faster ( Figure 4B). Under the growth conditions used in the experiment, these 286 differences in size between wild type and plant lines disappeared when seedlings became larger. 287 In fact, at 30-days-old stage, wild type seedlings had higher size ( Figure 4B). Nonetheless, carbon 288 dioxide fixation and biomass were comparable between wild type and L-GalLDH-RNAi plant lines 289 at 30 days after germination ( Figure 4C). 290

DISCUSSION 291
Over many years, it has been believed that plants synthesize AsA basically to produce this 292 powerful antioxidant molecule, which has multiple functions (Smirnoff, 2018). The localization of 293 the L-GalLDH enzyme within mitochondria has supported the obvious paradigm that AsA synthesis 294 exists for producing AsA, which promotes ROS scavenging. We unexpectedly found that AsA 295 synthesis triggers ROS content and down-regulates energy supply. Most specifically, the electron 296 flux derived from L-GalLDH activity is poorly used for the generation of proton gradient and Similarly to the demonstrated role of AsA synthesis, the AOX pathway can also decline ATP 318 production; however, the later pathway does not reduce the mitochondrial capacity to utilize 319 reducing equivalents (Schertl and Braun, 2014). This difference may be strongly linked with the 320 close inter-relationship between AOX and L-GalLDH expressions (Bartoli et al., 2006). Likely, when 321 plants need to decrease mitochondrial ATP generation, the L-GalLDH enzyme could be expressed 322 to inhibit the respiratory electron flux unlike the AOX pathway, which maintains a significant part 323 of such flux intact, contributing to energy loss. Therefore, the AsA synthesis may be activated to 324 avoid the loss of energy when the need for mitochondrial ATP is low. 325 Our study also shows that the generation of hydrogen peroxide (H 2 O 2 ) during AsA synthesis is 326 crucial for determining AsA level. Paradoxically, AsA is needed for H 2 O 2 elimination (Smirnoff,  AsA synthesis capacity was expressed as total AsA per fresh weight (mg). 457

Measurement of cytochrome c reduction capacity 458
The L-GalL-induced Cytc reduction capacity was assessed following procedure described in (Ôba,

Assessment of NADH production capacity by isolated mitochondria 495
The NADH production was based on the difference in the absorption spectra of NAD + and NADH. 496 The NADH shows absorption maxima at 340 nm but NAD + does not absorb light at 340 nm 497 (Renault et al., 1982). In brief, mitochondria were osmotically-broken by incubating into 10 mM 498 MOPS (without mannitol) pH 7.2, 0.1 mM EDTA during 10 min. Then, a reaction mixture was 499 prepared consisting of 5 mM L-GalL, 20 mM NAD + , 3 mM azyde and 1 mM SHAM, allowed to react 500 during 30 min at 25°C and NADH absorbance (extinction coefficient NADH = 6.220 M −1 cm −1 ) was 501 read with UV/Vis spectrophotometer. Control mixture had the same components, except L-GalL. 502 The relative NADH production was determined as the difference in absorbance at 340 nm 503 following incubation for 30 min. 504

Measurement of H + -ATPase pumping activity 505
The H + -ATPase activity was measured using sub-mitochondrial particles obtained from intact 506 mitochondria treated or not with L-GalL. Particles were prepared by sonication, as described in 507 (Ragan et al., 1987), with some modifications. Briefly, intact mitochondria (200 µg) were incubated 508 in 10 mM MOPS containing 0.35 M mannitol with or without 5 mM L-GalL over 60 min at 25°C. 509 Following treatment, mitochondria were sonicated by 6-10 s pulses with 30 s intervals and 510 supernatant ultra-centrifuged and the resulting pellet (sub-mitochondrial particles) were re-511 suspended into the same buffer. Then, sonicated-disrupted mitochondria solution was incubated 512 with 1 mM ATP and its capacity to hydrolyze ATP was monitored by measuring the release of 513 inorganic phosphate (P i ) colorimetrically at 720 nm, as previously reported in (Subbarow, 1925). 514 H + -ATPase activity was expressed as the amount of released Pi during 1 min into the reaction 515 medium. 516

Measurement of ATP level 517
To determine the mitochondrial capacity to synthesize ATP during ascorbate biosynthesis, 50 µg of To determine ATP content in plants, 600-900 mg of aboveground plant tissue were collected at 531 night-time (two hours before lighting) and were quickly incubated at 100°C for 15 min in 1 mL of 532 boiled water, as described in (Yamamoto et al., 2002). Tissues were homogenized at 4°C and then 533 centrifuged at 9000g for 15 min at 4°C. Supernatant (200 µL) was used for ATP quantification, as 534 described above. 535

Western blot analysis 536
Mitochondrial proteins were reduced using 2.5% (v/v) 2-mercaptoethanol into sample buffer, 537 loaded onto one-dimensional SDS/PAGE gels and run following standard procedures. For the 538 detection of oxidized AOX, mitochondrial proteins were prepared in absence of 2-539 mercaptoethanol. Molecular weight markers (24-102 kDa, GE Healthcare) were used and equal 540 loading of gels (25 µg protein) was checked by Ponceau staining. The proteins were transferred to 541 a nitrocellulose membrane (Hybond ECL, Amersham/GE Healthcare); the membrane was blocked 542 in nonfat milk 5% overnight at 4°C and antibodies against L-GalLDH and AOX (commercially 543 provided by Agrisera) were used in dilutions 1:500. Then, the membranes were washed three 544 times in PBS buffer with milk 5%, incubated with goat anti-rabbit secondary antibody for 2h, and 545 subsequently washed in PBS buffer three times. Results were visualized by chemiluminiscence 546 with a ECL Western Blotting Detection System (Amersham/GE Healthcare) and quantified using 547 ImageJ densitometric software (https://imagej.nih.gov/ij/). 548

Experimental design for comparative proteomic analysis 549
Papaya fruit mesorcarp discs (500 mg fresh weight) were treated with 5 mM of L-GalL, 50 mM 550 MOPS buffer pH 7.0 for two hours at 25°C. A control sample without L-GalL was also incubated 551 under the same conditions. Then, three independent samples of proteins were extracted from 552 each treatment. Procedures for protein extraction, digestion and mass spectrometry analysis are 553 performed following previous works described in (Heringer et al., 2017). 554

Bioinformatic analysis 555
Progenesis QI for Proteomics Software V.2.0 (Nonlinear Dynamics, Newcastle, UK) were used to 556 spectra processing and database searching conditions. The analysis were performed using 557 following parameters: Apex3D of 150 counts for low energy threshold, 50 counts for elevated 558 energy threshold, and 750 counts for intensity threshold; one missed cleavage, minimum fragment 559 ion per peptide equal to two, minimum fragment ion per protein equal to five, minimum peptide 560 per protein equal to two, fixed modifications of carbamidomethyl (C) and variable modifications of 561 oxidation (M) and phosphoryl (STY), and a default false discovery rate (FDR) value at a 4% 562 maximum, peptide score greater than four, and maximum mass errors of 10 ppm. The analysis 563 used the Carica papaya v. 0.4 protein databank from Phytozome (https://phytozome.jgi.doe.gov/). 564 Label-free relative quantitative analyses were performed based on the ratio of protein ion counts 565 among contrasting samples. After data processing and to ensure the quality of results, only 566 proteins present in 3 of 3 runs were accepted. Furthermore, differentially abundant proteins were 567 selected based on a fold change of at least 1.5 and ANOVA (P ≤0.05). Functional annotation was 568 performed using Blast2Go software v. 3.4 (Conesa et al., 2005). 569

Seed treatment, plant growth and photosynthesis 570
To determine growth of tomato plants with induced AsA synthesis, seeds of wild type and L-571 GalLDH-RNAi plant lines were subjected to imbibition treatment with 20 mM GalL for 6 hours. In 572 parallel, control seeds were treated in absence of L-GalL. Imbibited seeds were sown on soil pots 573 filled with commercial substrate and irrigated with Hoagland solution. Then, seedlings (one per 574 pot) were grown for four weeks at a growth chamber at 25°C with a 16-h photoperiod at a light 575 intensity of 500 μmol m 2 s -1 . Height and biomass of plants were determined at given time points. 576 Dry weights were determined by drying aboveground tissue in an air circulation oven at 80°C for 577 one week. Height was determined by measuring the distance from the ground to the top of 578 canopy. At four-weeks after sowing, instantaneous gas exchange measurements were done on six 579 recently fully expanded leaves in the upper part of the wild type plant and L-GalLDH-RNAi plant 580 lines. Measurements were taken between 2-4 hours after the start of light period using a gas 581 exchange system (LiCOR, Biosciences, Lincoln, NE, USA). Determinations of CO 2 assimilation were 582 performed at light intensity 500 µmolm -2 s -1 , 400 ppm CO 2 and temperature 24-26°C. 583

Statistical analysis 584
Data from at least three independent biochemical experiments were averaged and subjected to 585 ANOVA and, when needed, means were analyzed following Tukey test at P ≤ 0.05.       Figure 1A. NADH-driven respiration of leaf mitochondria purified from 30-days-old wild type, 8-14 and 5-13 transgenic lines. Mitochondrial preparations were pre-treated or not with 5 mM L-GalL and the oxygen uptake rates were determined following the addition of 10 mM malate. Then, the respiration was blocked by a mixture of respiratory inhibitors 5 mM SHAM and 3 mM NaN3. 1B. Immunoblot of L-GalLDH in leaf mitochondria from 30-days-old wild type, 8-14 and 5-13 transgenic lines detected by Western Blot using anti-L-GalLDH (Agrisera). Relative abundance, expressed as % of wild type signal, was obtained by densitometry. L-GalLDH activity (measured as rate of Cytc reduction) was determined in the mitochondria from 30-days-old wild type, 8-14 and 5-13 transgenic lines. Values represent means ± standard error and asterisks represent significant differences between inhibited and non-inhibited reactions analyzed by one-way ANOVA following by Tukey test (P < 0.05). Measurements from three independent mitochondrial preparations (n=3). 1C. Rates of oxygen uptake SHAM-sensitive and NaN3-resistant (alternative respiration) determined in mitochondria from wild type, 8-14 and 5-13 transgenic lines. Asterisks represent significant differences of alternative respiration of transgenic leaf mitochondria when compared to wild type by one-way ANOVA following by Tukey test (P < 0.05). 1D. Immunoblot of AOX detected when mitochondrial proteins are loaded without the reducing agent of free sulphydryl residues (2-mercaptoethanol) into sample buffer. A 66 kDa protein was detected by Western Blot using anti-AOX antibodies (Agrisera) in leaf mitochondria purified from 30-days-old wild type, 8-14 and 5-13 plants. Immunoblot of AOX (molecular weight of about 33 kDa) obtained when the reducing agent was added into sample buffer and subsequent detection with anti-AOX antibodies. Equal loading of gels was checked by Ponceau staining. Relative abundances were assessed by quantification of signals through densitometry and expressed as % of wild type level.  Figure 2A. Ascorbate production capacity measured in pure mitochondria treated or not with inhibitors (3 mM azyde, NaN3, 1mM SHAM, 2 µM antimycin A, 20 µM rotenone, and 5 mM DPI). Ascorbate synthesis was initiated in presence of 5 mM L-GalL. 2B. Activity of L-GalLDH enzyme (assessed as capacity to reduce Cytc) determined in purified mitochondria incubated with 1mM SHAM, 2 µM antimycin A, 20 µM rotenone, and 5 mM DPI). Cytc reduction was started by adding 5mM L-GalL in presence of the Cytc oxidase inhibitor, azyde. 2C. Ascorbate production and L-GalLDH activity determined in mitochondria from green-mature and full-ripe papaya fruit. 2D. Ascorbate synthesis-dependent changes in flavin adenine dinucleotide (FAD) fluorescence measured in pure mitochondria incubated with the inhibitor compounds indicated above in figure 2A. AsA synthesis was started with 5mM L-GalL. Bars represent means ± standard error from at least three independent experiments. Figure 3A. Induction of hydrogen peroxide (H2O2) production by 5 mM L-GalL in pure mitochondria treated with 1 mM SHAM, 3 mM azyde, 20 µM rotenone, 2 µM antimycin A, and 5 mM DPI. H2O2 content was quantified using Amplex Red/horseradish peroxidase (HRP) assay and relative H2O2 level for all treatments was normalized to their corresponding controls without L-GalL. Bars are means ± standard error (n=3). 3B.
Representative confocal images of co-localization of ROS and mitochondria stains in fruit mesocarp tissue incubated with 5 mM L-GalL and the same inhibitor compounds as above. Mitochondria were localized with Mito-Tracker Red and ROS detection was performed with (2,7dichlorodihydrofluorescein diacetate, DCF-DA). Scale bar 10 µm.  Figure 4A. Ascorbate synthesis capacity and ATP content of leaf tissue (in dark) determined in leaf tissue from four weeks-days-old wild type, 8-14 and 5-13 transgenic lines. 4B. Effect of seed treatment with 20mM L-GalL on seedling growth from wild type, 8-14 and 5-13 transgenic lines measured one-week and four-weeks following the chemical treatment. 4C. Net photosynthesis of fully expanded leaves (Dotted line) and dry weight of aboveground tissue (Grey line) determined in wild type, 8-14 and 5-13 transgenic lines at the four-weeks growth stage. All bars are means ± standard error from three independent experiments.