Protective Responses at the Biochemical and Molecular Level Differ between a Coffea arabica L. Hybrid and Its Parental Genotypes to Supra-Optimal Temperatures and Elevated Air [CO2]

Climate changes with global warming associated with rising atmospheric [CO2] can strongly impact crop performance, including coffee, which is one of the most world’s traded agricultural commodities. Therefore, it is of utmost importance to understand the mechanisms of heat tolerance and the potential role of elevated air CO2 (eCO2) in the coffee plant response, particularly regarding the antioxidant and other protective mechanisms, which are crucial for coffee plant acclimation. For that, plants of Coffea arabica cv. Geisha 3, cv. Marsellesa and their hybrid (Geisha 3 × Marsellesa) were grown for 2 years at 25/20 °C (day/night), under 400 (ambient CO2, aCO2) or 700 µL (elevated CO2, eCO2) CO2 L−1, and then gradually submitted to a temperature increase up to 42/30 °C, followed by recovery periods of 4 (Rec4) and 14 days (Rec14). Heat (37/28 °C and/or 42/30 °C) was the major driver of the response of the studied protective molecules and associated genes in all genotypes. That was the case for carotenoids (mostly neoxanthin and lutein), but the maximal (α + β) carotenes pool was found at 37/28 °C only in Marsellesa. All genes (except VDE) encoding for antioxidative enzymes (catalase, CAT; superoxide dismutases, CuSODs; ascorbate peroxidases, APX) or other protective proteins (HSP70, ELIP, Chape20, Chape60) were strongly up-regulated at 37/28 °C, and, especially, at 42/30 °C, in all genotypes, but with maximal transcription in Hybrid plants. Accordingly, heat greatly stimulated the activity of APX and CAT (all genotypes) and glutathione reductase (Geisha3, Hybrid) but not of SOD. Notably, CAT activity increased even at 42/30 °C, concomitantly with a strongly declined APX activity. Therefore, increased thermotolerance might arise through the reinforcement of some ROS-scavenging enzymes and other protective molecules (HSP70, ELIP, Chape20, Chape60). Plants showed low responsiveness to single eCO2 under unstressed conditions, while heat promoted changes in aCO2 plants. Only eCO2 Marsellesa plants showed greater contents of lutein, the pool of the xanthophyll cycle components (V + A + Z), and β-carotene, compared to aCO2 plants at 42/30 °C. This, together with a lower CAT activity, suggests a lower presence of H2O2, likely also associated with the higher photochemical use of energy under eCO2. An incomplete heat stress recovery seemed evident, especially in aCO2 plants, as judged by the maintenance of the greater expression of all genes in all genotypes and increased levels of zeaxanthin (Marsellesa and Hybrid) relative to their initial controls. Altogether, heat was the main response driver of the addressed protective molecules and genes, whereas eCO2 usually attenuated the heat response and promoted a better recovery. Hybrid plants showed stronger gene expression responses, especially at the highest temperature, when compared to their parental genotypes, but altogether, Marsellesa showed a greater acclimation potential. The reinforcement of antioxidative and other protective molecules are, therefore, useful biomarkers to be included in breeding and selection programs to obtain coffee genotypes to thrive under global warming conditions, thus contributing to improved crop sustainability.


Introduction
Climate changes associated with global warming are expected to endanger ecosystems and food security [1]. Climate changes are believed to be closely driven by greenhouse gas (GHG) emissions into the atmosphere. Among them, air [CO 2 ] already exceeds 400 µL L −1 and continues to increase at a rate close to 2 µL CO 2 L −1 per year [2], with estimates pointing to unprecedented values between 730 and 1000 µL L −1 by 2100, depending on the measures to control future GHGs emissions. This air [CO 2 ] increase is believed to trigger a temperature rise between 1.0-1.8 • C (best scenario) and 3.3-5.7 • C (worst scenario without additional efforts to limit the emissions), as compared to 1850-1900 [3,4]. Climate changes are already affecting the frequency and severity of extreme events, such as heat waves, longer and harsher droughts, unpredictable rainfalls, etc. [4]. This will have severe impacts on agricultural ecosystems, with the consequent decline of crop yields, quality [3], and suitable areas, under increasing pressure for feed and food availability to fulfill the demands of a growing world population [5] that is expected to approach 10,000 million people by 2050 [6][7][8].
High temperature is one of the major abiotic stresses that pose growing and serious challenges to plant growth and development [9]. Heat stress affects several physiological processes [10], e.g., it could alter membrane permeability and fluidity, influencing cellular homeostasis [11,12], and cause denaturation and aggregation of proteins [13], cell damage and ion leakage, interfering with important processes such as respiration and photosynthesis [14]. Indeed, the photosynthetic apparatus is highly sensitive to high temperatures [15], namely at the PSII level [9], associated with the dissociation of the D1 protein, but the electron transport chain (ETC) and the oxygen-evolving complex (OEC), and photosynthetic enzymes, such as RuBisCO activase, could also be inactivated [10,16,17]. Additionally, temperature rise stimulates photorespiration and respiration more than photosynthesis due to decreases in the affinity of RuBisCO for CO 2 and the solubility of CO 2 , both relative to O 2 , thus reducing the relative rate of carboxylation to oxygenation and C-assimilation [18][19][20].
Unfavorable environmental conditions that inhibit energy use through photochemistry may promote the over-reduction of the ETC, and the accumulation of molecules in the excited state [9,21], such as singlet oxygen ( 1 O 2 ), the singlet chlorophyll ( 1 Chl) and triplet state of chlorophyll ( 3 Chl), leading to the production of superoxide radical (O 2 * − ), hydrogen peroxide (H 2 O 2 ) and hydroxyl radical (OH •− ) [22]. The reactive oxygen species (ROS), mainly produced in chloroplasts, mitochondria, and peroxisomes [23], are harmful oxidants able to cause lipid peroxidation, enzyme inactivation, and degradation of pigments, proteins, and DNA [24] leading ultimately to cell apoptosis [25]. The control of highly reactive molecules of Chl and O 2 is achieved by promoting energy dissipation mechanisms that prevent its formation (e.g., photoprotective pigments) and the expression of enzymatic and non-enzymatic antioxidants that scavenge the ROS already produced [13,24]. Therefore, the reinforcement of mechanisms dedicated to preventing ROS overproduction and/or its efficient scavenging is usually crucial to plant tolerance to a wide number of environmental Table 1. Leaf carotenoids concentration (mg g −1 dry weight, DW) in C. arabica cv. Geisha 3, cv. Marsellesa and their Hybrid (Geisha 3 × Marsellesa) plants grown under 400 or 700 µLCO 2 L −1 at the control temperature (25/20 • C, day/night), submitted to supra-optimal temperatures (31/25 • C, 37/28 • C, 42/30 • C), and after 4 (Rec4) and 14 (Rec14) days of recovery. For each parameter, the mean values ± SE (n = 4-6 plants) followed by different letters express significant differences between CO 2 treatments for each temperature (a, b) or between temperatures for the same CO 2 treatment (A, B, C, D), always separately for each genotype, where a > b and A > B > C > D.

Pigment
Genotype Under aCO 2 , the single impact of temperature rising to 31/25 • C and then further to 37/28 • C led to a general increase of most carotenoids in the three genotypes, although significantly only in some cases, especially in Geisha 3 and Hybrid plants. That was reflected in maximal increases in total carotenoid concentration in Geisha 3 (76%), Hybrid (43%), both at 31/25 • C, and Marsellesa (40%) at 37/28 • C (the latter was not significant), always when compared to their control values.
As the temperature reached the maximum value (42/30 • C), the carotenoid concentration declined in aCO 2 plants from their maxima to values that usually did not differ from their controls at 25/20 • C, except for lutein in Geisha 3 plants that maintained an increased value, or violaxanthin that showed reduced values (all genotypes). Still, DEPS tended to greater values than at 25/20 • C, with rises in Geisha 3 (45%), Marsellesa (315%), and Hybrid leaves (98%). A strong decline was observed in carotenes pools, particularly in α-carotene, implicating a clear decline in the (α/β)-carotene ratio of all genotypes.
The plants grown under eCO 2 submitted to temperature rise showed notable stability of the studied carotenoids, without significant modifications from 25/20 • C up to 37/28 • C (lutein) or even 42/30 • C (neoxanthin, zeaxanthin, V + A + Z, carotenes, total carotenoid) for the three genotypes. That also implicated the stability of DEPS values, even at 42/30 • C, although the (α/β)-carotene ratio tended to lower values in Geisha 3 and Marsellesa at that temperature.
Only a few significant differences were depicted between the two [CO 2 ] conditions regarding the leaf xanthophyll concentration at maximal temperature. Noteworthy were the observations that, at 42/30 • C, the eCO 2 Marsellesa plants showed greater concentrations of V + A + Z and, especially, lutein than their aCO 2 counterparts, but such eCO 2 superimposition reduced DEPS in all genotypes (significantly only in Hybrid) as compared with aCO 2 plants. Additionally, at 42/30 • C, the eCO 2 plants tended to have greater values of α-carotene (significant in Hybrid), β-carotene, (α + β) carotene (both significant in Marsellesa), and (α/β) carotene (significant in Hybrid).
Overall, during the recovery period (at 25/20 • C), most carotenoid concentrations returned to values that did not differ from the initial values, usually regardless of CO 2 condition and genotype. That was the case from Rec4 onwards of neoxanthin (all genotypes), zeaxanthin (only by Rec14 for Hybrid plants), V + A + Z, α-carotene, β-carotene, and (α + β) carotene (except the Hybrid plants under aCO 2 for the latter). In contrast, lutein and total carotenoids maintained increased levels by Rec4 in all genotypes under aCO 2 , and still by Rec14 in Geisha 3 and Hybrid plants, whereas α-carotene had reduced concentrations even by Rec14 in the three genotypes and both [CO 2 ]. DEPS showed greater values than the control at 25/20 • C for Marsellesa and Hybrid plants both in Rec4 and Rec14.

Antioxidant Enzyme Activity
At control temperature, the single eCO 2 exposure altered the activity of antioxidative enzymes, depending on the genotype and enzyme (Figures 1 and 2). In Marsellesa, the potential activities of SOD ( Figure 1) and GR ( Figure 2) declined 49% and 58%, respectively, relative to their aCO 2 plants, whereas APX and CAT were not affected. Geisha 3 plants showed an approximately doubled APX activity and ca. half of CAT, without changes in SOD and GR. Notably, all four enzymes were mostly insensitive to eCO 2 in the Hybrid plants.
At control temperature, the single eCO2 exposure altered the activity of antioxidative enzymes, depending on the genotype and enzyme (Figures 1 and 2). In Marsellesa, the potential activities of SOD ( Figure 1) and GR ( Figure 2) declined 49% and 58%, respectively, relative to their aCO2 plants, whereas APX and CAT were not affected. Geisha 3 plants showed an approximately doubled APX activity and ca. half of CAT, without changes in SOD and GR. Notably, all four enzymes were mostly insensitive to eCO2 in the Hybrid plants. the mean values ± SE (n = 3 plants) followed by different letters express significant differences between CO 2 treatments for each temperature, separately for each genotype (a, b), or between temperatures for the same CO 2 treatment (A, B, C), always separately for each genotype, where a > b and A > B > C.
With further temperature increase to 42/30 • C, APX activity was severely depressed (Figure 1), whereas CAT showed values even greater (Geisha 3 and Hybrid) or unaltered (Marsellesa) as compared with those at 37/28 • C (Figure 2). At this harshest condition (42/30 • C), SOD activity continued to decline (except in the Hybrid), whereas GR values were mostly maintained in all genotypes when compared to the values under 37/28 • C. plants grown under 400 or 700 µLCO2 L −1 at control (25/20 °C, day/night), submitted to supraoptimal temperatures (37/28 °C, 42/30 °C), and after 14 days of recovery (Rec14). For each enzyme, the mean values ± SE (n = 3 plants) followed by different letters express significant differences between CO2 treatments for each temperature, separately for each genotype (a, b), or between temperatures for the same CO2 treatment (A, B, C), always separately for each genotype, where a > b and A > B > C.

Figure 2.
Changes in leaf cellular activities of glutathione reductase (GR) and catalase (CAT) enzymes in C. arabica cv. Geisha 3, cv. Marsellesa and their Hybrid (Geisha 3 × Marsellesa) plants grown under 400 or 700 µLCO2 L −1 at control (25/20 °C, day/night), submitted to supra-optimal temperatures (37/28 °C, 42/30 °C) and after 14 days of recovery (Rec14). For each enzyme, the mean values ± SE (n = 3 plants) followed by different letters express significant differences between CO2 treatments for each temperature, separately for each genotype (a, b), or between temperatures for the same CO2 treatment (A, B, C), always separately for each genotype, where a > b and A > B > C.
With further temperature increase to 42/30 °C, APX activity was severely depressed (Figure 1), whereas CAT showed values even greater (Geisha 3 and Hybrid) or unaltered (Marsellesa) as compared with those at 37/28 °C (Figure 2). At this harshest condition (42/30 °C), SOD activity continued to decline (except in the Hybrid), whereas GR values were mostly maintained in all genotypes when compared to the values under 37/28 °C. . For each enzyme, the mean values ± SE (n = 3 plants) followed by different letters express significant differences between CO 2 treatments for each temperature, separately for each genotype (a, b), or between temperatures for the same CO 2 treatment (A, B, C), always separately for each genotype, where a > b and A > B > C.
The superimposition of eCO 2 at 37/28 • C altered, in a few cases, the enzyme activities observed under aCO 2 . A clearer case was observed with CAT, whose activity decreased under eCO 2 both at 37/28 • C and 42/30 • C in all genotypes. Additionally, APX also declined in Geisha 3 (42/30 • C), Marsellhesa and Hybrid (both at 37/28 • C). In contrast, SOD activity increased in Geisha 3 (by 63% at 42/30 • C), Marsellhesa and Hybrid (by 64% and 35%, respectively, both at 37/28 • C), although not showing differences to aCO 2 at 42/30 • C for the last two genotypes. Marsellesa showed the lowest values among the three genotypes. Finally, GR activity was unchanged between CO 2 conditions at 37/28 • C and 42/30 • C for the three genotypes, except for a decline at the harshest temperature in Geisha 3.
After 14 days of recovery (Rec14) at 25/20 • C, the aCO 2 plants tended to approach the value observed at the beginning of the experiment (also at 25/20 • C). That was the case for all enzymes in Geisha 3 and Marsellesa. In the Hybrid plants, there were reduced activities of SOD and APX and greater values of CAT.
Still, by Rec14, the eCO 2 plants showed a distinct pattern from their aCO 2 counterparts. All genotypes showed lower SOD activity under eCO 2 than at aCO 2 by the end of the experiment (Figure 1), but for the other enzymes, a genotype-dependent situation was observed. In eCO 2 plants, Geisha 3 showed a declined APX activity, in opposition to the rise in CAT. Marsellesa kept close APX activities between [CO 2 ] conditions and a reduction of CAT under eCO 2 (but always similar to the initial controls), whereas the Hybrid plants showed greater (130%) and lower (89%) values for APX and CAT, respectively. Finally, GR was mostly irresponsive to [CO 2 ] by Rec14 in the studied genotypes, showing values close to those at the beginning of the trial (Figure 2).

Expression of Selected Genes Associated with Protective Roles
The single exposure to eCO 2 at 25/20 • C did not significantly alter the expression of any of the studied genes, irrespective of genotype, in comparison to their aCO 2 plant counterparts ( Table 2).
Regarding the single temperature exposure, a moderate down-regulation impact was observed only for VDE2 at 37/28 and/or 42/30 • C, together with a total recovery under Rec14 for the three genotypes compared with the initial expression values at 25/20 • C. For all the other studied genes, considering both those encoding for antioxidative enzymes (CAT, the two CuSODs, and the three APX isoforms) and other protective proteins (HSP70, ELIP, Chape20, Chape60), the high temperature was the major driver of gene expression changes, promoting a strong up-regulation at 37/28 • C and a further rise at 42/30 • C, when the highest expression values were usually attained in the three genotypes. Notably, maximal gene expression was always found in Hybrid plants (aCO 2 , 42/30 • C), except for CAT (Geisha 3) and APX Cyt (Marsellesa).
In general, the abundance of gene transcripts associated with antioxidative enzymes was the highest up-regulated in the three genotypes. For instance, the expression of genes encoding for Cu, Zn-SOD (CuSOD1, CuSOD2), APX (cytosolic enzyme APX, APX Cyt ; chloroplast APX, APX Chl ; stromatic APX, APX t+s ), and CAT presented a common pattern of marked up-regulation at 42/30 • C. Among SOD, APX and CAT genes, APX Cyt was the most strongly up-regulated one in Geisha 3 (70-fold) and, especially, in Marsellesa (104-fold), whereas in Hybrid plants, the greatest up-regulation was observed for CuSOD2 (55-fold; Table 2). CAT was always significantly up-regulated under heat (between 2-fold in Marsellesa and 6-fold in Geisha 3) but was the less up-regulated of this group of genes encoding for antioxidative enzymes. By Rec14, the transcripts abundance of all studied genes (but VDE) declined in comparison with 42/30 • C, but significantly higher values than the control were still found in all cases, mainly in the antioxidative enzymes (Table 2).
Under heat, eCO 2 strongly attenuated the observed up-regulation under aCO 2 , for all studied genes and in the three genotypes. That eCO 2 impact occurred at 37/28 • C (except for HSP70, ELIP, APX Chl , APX t+s, and VDE in Geisha 3, and APX Chl and VDE in Hybrid) and 42/30 • C. However, a significant up-regulation was still usually observed in eCO 2 plants at supra-optimal temperatures when compared with their values at 25/20 • C, meaning that under eCO 2 , gene expression reinforcement still occurred, although to a lesser extent than at aCO 2 . Among the largest gene expression attenuations observed, CuSOD1, CuSOD2, and APX Cyt stood out in the three genotypes, both under 37/28 • C and/or 42/30 • C. Such declines were frequently to values close to or lower than half (Table 2), with striking reductions of 88% (CuSOD2, Hybrid, 37/28 • C), 78% (APX Cyt , Marsellesa, 37/28 • C), and 75% (APX Cyt , Geisha 3, 42/30 • C) when compared with their respective aCO 2 counterparts at the same temperature. Additionally, in Rec14, eCO 2 plants still maintained lower expression values than the aCO 2 plants but were closer to their controls, suggesting a better recovery to the initial status. Table 2. Real-time PCR expression studies relative to the expression value observed under control conditions of temperature and CO 2 (25/20 • C, 400 µL CO 2 L −1 ) from leaves of C. arabica cv. Geisha3, cv. Marsellesa and their Hybrid (Geisha 3 × Marsellesa) plants grown under 400 or 700 µL CO 2 L −1 at control (25/20 • C, day/night), submitted to supra-optimal temperatures (37/28 • C, 42/30 • C), and after 14 days of recovery (Rec14). For each gene transcript, the mean values (n = 3 plants) followed by different letters express significant differences between [CO 2 ] levels for each temperature treatment (a, b) or between temperature treatments for the same CO 2 treatment (A, B, C, D), always separately for each genotype, where a > b and A > B > C > D.

Plant Material, Growth Conditions and Experimental Design
For these experiments were used plants from Coffea arabica L. cv. Geisha 3, cv. Marsellesa and their hybrid (Geisha 3 × Marsellesa), which result from breeding efforts to find new cultivars (and in this case also their hybrid) to be used under shaded environments, and to better cope with the new climate conditions estimated to occur along this century. The applied experimental design was similar to that described in [34], with minor modifications. Following a completely randomized design, potted plants (6 to 8 per treatment) were grown in 20 L pots from the seedling stage until ca. two years of age in walk-in growth chambers (EHHF 10000; ARALAB, Albarraque, Portugal) under controlled environmental conditions of temperature (25/20 • C, day/night), irradiance (ca. 700-800 µmol m −2 s −1 ), air humidity RH (70%), and photoperiod (12 h), and either ambient (400 µL CO 2 L −1 , aCO 2 ) or elevated (700 µL CO 2 L −1 , eCO 2 ). For the entire experiment, the plants were kept under well-watered conditions (predawn water potential higher than −0.3 MPa), adequate mineral nutrient supply (provided as in [59]), and without restrictions as regards space for root growth, as judged by visual examination at the end of the experiment after removing the plants from pots.

Temperature Rise Implementation
Plants were submitted to a gradual temperature increase to allow them to express their potential acclimation capability. The temperature was raised from 25/20 • C (day/night) up to 42/30 • C, at a rate of 0.5 • C day −1 (of the diurnal temperature), with 5-7 days of stabilization at 31/25 • C, 37/28 • C, and 42/30 • C to allow for programmed plant material collection. Subsequently, the temperature was readjusted to 25/20 • C and plants were monitored over a recovery period of 4 (Rec4) and 14 (Rec14) days. The control conditions refer to the plants grown at 25/20 • C and aCO 2 .
All measurements were performed on newly matured leaves from the upper third part of the plant canopies, which were flash-frozen in liquid nitrogen and stored at −80 • C until analyses.
The carotenoid concentration and the enzyme activities were given per dry weight units (DW) since it is a more stable basis of expression due to the eventual change of leaf hydration status under high-temperature conditions. In this way, the relation between fresh weight (FW) and DW was obtained from the same leaves used for biochemical measurement, similarly to what is established for leaf relative water content (RWC) determinations in Coffea spp. [26]. For that, eight foliar discs of 0.5 cm 2 each were punched from the leaves and FW was immediately determined, whereas DW was obtained after drying the discs at 80 • C for 48 h.

Carotenoids Evaluation
Carotenoids were assessed from three leaf discs (each of 0.5 cm 2 ) from 4 to 6 plants per treatment, cut after 1.5-2 h of leaf exposure to diurnal illumination, flash frozen in liquid nitrogen, and stored at −80 • C until analysis. The leaf tissue homogenization and subsequent reversed-phase HPLC analysis were performed as described in [60] using an end-capped C 18 5 µm Spherisorb ODS-2 column (250 × 4.6 mm, Waters, Milford, MA, USA). Carotenoid detection was performed at 440 nm in an HPLC system (Waters Alliance e2695, Milford, MA, USA) coupled with a diode-array detector (Waters 2996, Milford, MA, USA). Identification and quantification of each carotenoid were performed using specific standards. The de-epoxidation state, involving xanthophyll cycle components, was calculated as [DEPS = (zeaxanthin + 0.5 antheraxantin)/(violaxanthin + antheraxantin + zeaxanthin)].

Activity of Antioxidative Enzymes
Maximal cellular enzyme activities were assayed using 100 mg fresh weight (FW) of leaf tissue for the most important treatments (the enzyme activity assays were not performed for the 31/25 • C and Rec4 treatments), taken from 3 plants per treatment. All the procedures, from homogenization to enzyme activity measurements, were performed following [26], with some modifications. Briefly, 1 mL of buffer containing 200 mM Tris-HCl (pH 8), 10 mM MgCl 2 6H 2 O, 30 mM β-mercaptoethanol, 4 mM DTT, 2% Triton X-100, "Complete cocktail EDTA" (2 pills) and 10% glycerol was used, with addition of 1% (1 mL) of polyvinylpolypyrrolidone (PVPP) to each sample in the homogenization phase. The samples were centrifuged (13,000× g, 20 min, 4 • C), and the supernatant was used to evaluate enzymatic activities.
Superoxide dismutase (SOD, EC 1.15.1.1) reaction mixture contained 20 mM adrenaline in 50 mM phosphate buffer (pH 7.8) and sodium carbonate buffer (pH 10.4) with EDTA 0.125 mM and 50 µM, and 50 µL of the enzyme extract in a final volume of 1 mL. The activity was spectrophotometrically assessed at 480 nm.
Ascorbate peroxidase (APX, EC 1.11.1.11) reaction mixture contained 20 mM ascorbate and 0.1 mM H 2 O 2 in 50 mM phosphate buffer (pH 7.8) and 10 µL of the enzyme extract in a total volume of 1 mL. The sample reaction was assessed through the H 2 O 2 -dependent oxidation of ascorbate at 290 nm, using an extinction coefficient of 2.8 mM −1 cm −1 for calculations.
Catalase (CAT, EC 1.11.1.6) reaction mixture contained 40 mM H 2 O 2 in 50 mM phosphate buffer (pH 7.8) and 10 µL of the enzyme extract in a total volume of 1 mL; activity was evaluated through the rate of H 2 O 2 consumption at 240 nm, using an extinction coefficient of 3.94 mM −1 cm −1 .

Expression of Genes Associated with Antioxidant and Protective Molecules
Genes encoding antioxidant and other protective proteins were selected based on [26,33] for real-time qPCR studies, using malate dehydrogenase (MDH) and Ubiquitin-conjugating enzyme E2 (UBQ), which were found to be among the most stable pair of genes to be used as reference genes for the studied conditions of temperature and [CO 2 ] [39]. All primer sequences are presented in Table 3.
Total RNA was isolated from 100 mg of frozen material taken from 3 plants per treatment and processed as described in [35], using the innuPREP Plant RNA Kit (Analytik Jena, Jena, Germany) following the manufacturer's protocol. The intactness of the extracted RNA was verified by electrophoresis on a 1.5% agarose gel by evaluating the integrity of the 28S and 18S ribosomal RNA bands and the absence of smears. cDNA was synthesized from 1 µg total RNA using the SensiFAST TM cDNA Synthesis kit (Meridian BioScience, Cincinnati, OH, USA), according to the manufacturer's recommendations. The presence of a single amplification product of the expected gene size was verified by electrophoresis on a 1.5% agarose gel. RT-qPCR reactions were prepared using the SensiFAST TM SYBR No-ROX kit (Meridian BioScience, Cincinnati, OH, USA) according to the manufacturer's protocol. One negative was included for each primer pair, in which cDNA was replaced by water. Reactions were carried out in 96-well plates using a qTOWER 2.2 Thermal Cycler (Analytik Jena, Jena, Germany) using the following parameters: hot start activation of the Taq DNA polymerase at 95 • C for 10 min, followed by 40 cycles of denaturation at 95 • C for 15 s, annealing at 60 • C for 30 s, elongation at 72 • C for 30 s. A melting curve analysis was performed at the end of the PCR run by a continuous fluorescence measurement from 55 • C to 95 • C with sequential steps of 0.5 • C for 15 s (single peaks were obtained). Three technical replicates were used for each biological replicate.

Statistical Analysis
Data were analyzed using two-way ANOVA (p < 0.05) to evaluate the differences between the two atmospheric [CO 2 ] (aCO 2 or eCO 2 ) or between the different temperature and recovery treatments (25/20 • C, 31/25 • C, 37/28 • C, 42/30 • C, Rec4, Rec14), and their interaction, followed by Tukey's HSD test for mean comparisons, except when otherwise stated. The ANOVA for each parameter was performed independently for each of the studied genotypes. For gene expression analysis, the relative expression ratio of each target gene was quantified based on its real-time PCR efficiencies and the crossing point (CP) difference of the unknown sample versus the control (25/20 • C, 400 µL CO 2 L −1 air) within each genotype, as described in [33,39], followed by the same statistical procedure described above. Data were analyzed using Statistica, v8 (StatSoft, Tulsa, OK, USA).

Xanthophylls
Under the control temperature, eCO 2 by itself did not significantly alter the level of most xanthophylls in Geisha 3 and Hybrid plants but promoted greater concentrations of several xanthophylls in Marsellesa (neoxanthin, lutein, the pool of V + A + Z involved in the xanthophyll cycle; Table 1), as also reported in other coffee genotypes [34], revealing its potential photoprotective capability in Marsellesa leaves, even in the absence of stressful conditions. By contrast, under aCO 2, the rise of temperature to 37/28 • C promoted an increase in the concentration of several carotenoids (particularly neoxanthin and lutein) in all genotypes, although especially in Geisha 3 and Hybrid plants, as also reflected in a global reinforcement of total carotenoids. Still, both zeaxanthin and DEPS values were not significantly altered in all genotypes, which would be linked to the maintenance of the use of energy through photosynthesis at this temperature [12,34]. This was in line with the moderate rise of zeaxanthin and DEPS under 42/30 • C when the photochemical use of energy (and net photosynthesis) was strongly depressed (data not shown). Such rise in zeaxanthin (and DEPS) were also observed in C. arabica cv. Icatu and cv. IPR108 at 42/34 • C, irrespective of [CO 2 ], when the photochemical use of energy was compromised [12,34], and plant acclimation (namely of photosystems, electron carriers, and chloroplast membranes) strongly depended on photoprotective thermal dissipation and antioxidative mechanisms [33]. The presence of adequate levels of xanthophylls, and especially zeaxanthin and lutein (but also of β-carotene), in the light-harvesting complexes (LHC), are associated with a higher capability of reducing excess excitation energy through thermal dissipation, thus preventing the formation of highly reactive molecules of oxygen and Chl [61,62].
Notably, the eCO 2 plants showed mostly stable carotenoid values up to 37/28 • C or even 42/30 • C. This suggests a lower need for thermal dissipation when compared with aCO 2 counterparts, likely associated with the maintenance of a greater photosynthetic functioning due to a greater CO 2 availability at the chloroplast level and the associated lower energy excess. Furthermore, at 42/30 • C, the Marsellesa plants showed a better potential photoprotective capability due to greater concentrations of lutein and the pool of V + A + Z. Still, this was accompanied by a lowered DEPS value (as also in the other two genotypes) as compared with their aCO 2 plants. Zeaxanthin was reported to rise (as well as DEPS) in coffee plants exposed to conditions that strongly depress the photochemical use of energy [26,32] since it is associated with a higher need for thermal dissipation and prevention of lipid peroxidation by removing the epoxy groups from the oxidized double bonds of thylakoid fatty acids [63,64]. Therefore, our findings suggest that under the imposed conditions, zeaxanthin was not needed in these plants, although we cannot discard that they were unable to convert violaxanthin into zeaxanthin through VDE action, both of which were suggested by the maintenance or down-regulation of VDE (Table 2). Notably, for Hybrid (regardless of [CO 2 ]) and Marsellesa (aCO 2 ) plants, the need for energy dissipation persisted by Rec4 since zeaxanthin concentration (and DEPS) increased in comparison with both 25/20 • C and 42/30 • C values. Furthermore, Geisha 3 plants showed a higher DEPS value at 42/30 • C, although associated with the maintenance of zeaxanthin and a declined V + A + Z pool, thus supporting the view of an incomplete ability of these genotypes to cope with heat at the xanthophyll cycle level.
Neoxanthin and lutein are found in the periphery of the LHC of photosystems, and they play important functions as they maintain the correct assembly and stability of antenna proteins [65]. In addition, the lutein-epoxide cycle and neoxanthin (as well as β-carotene) are quenchers of 3 Chl* and 1 O 2 , thus, scavenging these important lipoperoxidation initiators [66][67][68][69][70]. In fact, besides the structural contribution to the antenna complexes, neoxanthin can influence energy harvest, transfer, and dissipation in the photosystem, with implications for photoprotection, namely by increasing the efficiency of 3 Chl* quenching by lutein, thus contributing to preventing ROS formation and the consequent photoinhibition [71]. Still, neoxanthin did not significantly differ between [CO 2 ] treatments and was mostly stable within each [CO 2 ] treatment except in Hybrid (increase by 37/28 • C under aCO 2 ) and Geisha 3 plants (rise at 31/25 • C and 37/28 • C). By contrast, only the eCO 2 plants of Marsellesa at 42/30 • C displayed a greater concentration of lutein (than their aCO 2 counterparts) which likely strengthened the thermal dissipation protection of photosystems against photooxidation in those plants. In fact, the eCO 2 positive impact on the performance of the photosynthetic apparatus and stress defenses was reported at physiological and molecular levels in studies involving other C. arabica and C. canephora genotypes, where a significant up-regulation of photosynthetic, antioxidant, and lipid metabolism genes and/or proteins was found under eCO 2 [28,29,35]. That ultimately mitigated the harsh effects of drought [27][28][29] and heat [34,36], supporting the maintenance of higher photosynthetic performance under eCO 2 in the studied coffee genotypes.

Carotenes
The α-and β-carotenes are accessory pigments found in the reaction centers and core antennae of PSI and PSII [72]. Besides the function of absorbing light (especially blue light), carotenes, and especially β-carotene, have important photoprotective functions. The latter has the ability to protect lipid components of membranes, and chlorophyll a from oxidation, quenching 1 O 2 and 3 Chl* by forming triplet carotenes that dissipate energy through heat [31,73,74]. Furthermore, β-carotene protects the cytochrome b 6 f complex from photobleaching promoted by 1 O 2 [75], and decreased levels of this pigment in the reaction centers have been associated with higher vulnerability to photodamage [74]. The carotenoid biosynthesis pathway has been previously found to be significantly enriched in C. canephora cv. CL153 plants grown under eCO 2 [35], although without a corresponding carotenoid rise [33]. Such lack of impact of the single exposure to eCO 2 was also found in the present work, except in Marsellesa, which was the only genotype to present a significant rise (55%) in total carotenoids, together with rising tendencies of α-and β-carotenes (and a 53% rise for (α + β) carotene). Therefore, Marsellesa plants would have a better potential photoprotection capability (in addition to the already mentioned lutein significant rise), namely against 3 Chl* and 1 O 2 [66][67][68]70].
Regarding the single temperature implementation, α-and/or β-carotenes tended to have greater concentrations at 31/25 • C and 37/28 • C, similarly to other pigments. That was accompanied by a tendency to higher (α/β) carotene ratios in all genotypes. However, Marsellesa plants were the only ones to show maximal concentrations of (α + β) carotenes at 37/28 • C, thus supporting a greater protective function at this moderately high temperature when the other two genotypes showed already a tendency to decline. However, at 42/30 • C, the α-and β-carotenes strongly declined (as compared to the values at 37/28 • C), especially the first, which was reflected in strong reductions in the (α/β) carotene ratios of all genotypes, as also found in other genotypes submitted to high temperatures [33]. Although the strong concomitant decline of β-carotene suggested that the photosynthetic apparatus was largely affected, the decrease in (α/β) carotene ratio was interpreted as reflecting a protective mechanism against the energy in excess under cold conditions in Coffea sp. [31]. This view agrees with the maintenance of lowered (α/β) carotene ratio values, even if β-carotene and (α + β) carotene values tended to recover (e.g., in Marsellesa) by Rec14, irrespective of genotype and [CO 2 ]. Contrasting with aCO 2 plants, there were remarkable stability of α-, β-and (α + β) carotenes in the eCO 2 plants from 25/20 • C up to 42/30 • C, thus suggesting the maintenance of photosynthetic structures to which these pigments are associated, e.g., the photosystems. The decline in (α/β) carotene in Marsellesa at 42/30 • C could suggest a higher thermotolerance capability at the harshest temperature, in line with the finding of a greater potential functioning of the photosynthetic apparatus (evaluated through photosynthetic capacity, A max -data not shown). However, in the present case, such (α/β) carotene decline resulted from a tendency to α-reduction and not from an increase in β-carotene (which was maintained), contrary to what was found in C. arabica cv. Icatu under heat stress, where β-carotene showed its maximal values under 42/34 • C as compared to their initial values at 25/20 • C [33,34]. Nevertheless, as compared with their aCO 2 counterparts, under 42/30 • C, only the eCO 2 plants of Marsellesa displayed simultaneously greater concentrations of β-carotene, lutein, and (V + A + Z), thus denoting a greater potential for photoprotection of the photosynthetic apparatus acting in a complementary way [33].

Antioxidant Enzyme Responses and the Associated Gene Expression
The studied enzymes play important roles in oxidative stress control since they react with several ROS, such as O 2 •− (SOD) and H 2 O 2 (APX, CAT). The ROS control is greatly accomplished, namely, through the ascorbate-glutathione cycle, with the participation of enzymes such as SOD, APX, and GR, complemented with extra-chloroplast scavenging systems, such as CAT [25,67,76].
Although the single effect of eCO 2 on the antioxidant enzymes was genotype-and enzyme-dependent, the studied enzyme activities were always maintained or reduced, as compared with the aCO 2 plant counterparts, with the only exception of APX in Geisha 3, which showed a significant rise. In fact, in the Hybrid plants, no significant changes were observed for all these enzymes, whereas in Marsellesa, SOD and GR activity values declined (Figures 1 and 2). This low responsiveness to eCO 2 was in accordance with the absence of significant expression changes in SODs, APXs, and CAT ( Table 2), showing that under non-stressed conditions, the eCO 2 had a low, if any, impact on the expression of these genes. This was also in line with the minor changes detected in the primary metabolite profile of C. canephora cv. CL153 and C. arabica cv. Icatu genotypes are grown under eCO 2 and control (well-watered and 25 • C) conditions [77]. Our present results also revealed a genotype-dependent response to eCO 2 among C. arabica genotypes but contrasted with previous findings in the above-mentioned genotypes where it was highlighted that eCO 2 promoted a significant up-regulation of a considerable number of genes related to photosynthetic, antioxidant, and lipid metabolism. These supported the maintenance of increased photosynthetic potential promoted by eCO 2 and the absence of photosynthesis down-regulation [35]. In fact, single eCO 2 promoted moderate responsiveness regarding the antioxidative response, which was much lower than the one felt under drought [78] or heat constraints [28,29,36] in those same C. canephora and C. arabica genotypes.
Contrasting with the single eCO 2 impact, the single exposure to supra-optimal temperatures was a strong driver of response since it greatly promoted changes in the activity of APX, CAT, and GR (although not of SOD; Figures 1 and 2) and up-regulated genes associated with the antioxidant enzymes (CAT, CuSODs, APXs; Table 2) at 37/28 • C. CAT and APX activities increased in the three genotypes under aCO 2 , reinforcing the potential for H 2 O 2 control and reflecting a clear response toward the acclimation of the three genotypes. This agrees with the increases in the activity of APX, GR, and CAT under aCO 2 in C. canephora cv. CL153, under eCO 2 of GR and CAT in C. arabica cv. Icatu, as well as of Cu, Zn-SOD and CAT in C. arabica cv. IPR108 up to 37 • C [33]. In fact, thermotolerance can be improved by up-regulating gene expression and protein levels of ROS-scavenging enzymes [79,80]. Furthermore, the key protective role of the antioxidant defense system in coffee plants was previously reported to increase under several environmental stress conditions. For instance, Cu, Zn-SOD, and APX activities were enhanced in some coffee genotypes subjected to a gradual cold treatment [24] since an efficient ROS control is crucial to the acclimation response of C. arabica and C. canephora cultivars exposed to single and combined drought and cold conditions [26]. As stated above, our findings are consistent with the strong up-regulation of genes coding for CAT and APX enzymes at 37/28 • C, which was greater under aCO 2 than under eCO 2 , both for the enzyme activity and for gene expression. On the other hand, SOD activity tended to decline (significantly in Hybrid plants at 37/28 • C), although the expression of SOD genes was markedly up-regulated. In fact, although they were often coherent, transcript level changes of genes coding for antioxidant enzymes were not always fully in line with the enzyme's activity pattern. The notion that gene expression does not always perfectly follow biochemical patterns has been previously mentioned in other studies [26,28,33]. For example, at the two highest temperatures, CuSOD1 and CuSOD2 were overexpressed in the three genotypes and, to a much greater extent, under aCO 2 than under eCO 2 . This contrasted with the absence of rises (or even decline) of enzyme activity in most cases, and even with the greater activity at 37/28 • C (Marsellesa and Hybrid) and 42/30 • C (Geisha 3) under eCO 2 (Figure 1), what can be justified by post-transcriptional and post-translational mechanisms regulating protein synthesis and enzyme functional conformation. This can greatly alter the relation between transcription levels and the biochemical results (e.g., of enzyme activities), thus reinforcing the need for data integration of complementary transcript and other molecular profiling, physiological and biochemical studies to have a clear picture of the real plant response to stress [36].
The difference between enzyme activity and gene expression was particularly striking at 42/30 • C, as also found for chloroplast APX activity and APX Chl expression in C. canephora cv. Conilon [33]. Gene expression associated with antioxidant enzymes showed their maximal absolute values (CAT, CuSODs, APXs) at the highest temperature when the enzyme activities showed quite different and variable impacts. In fact, the activities of GR and CAT were mostly maintained, and the one of APX declined in all genotypes, whereas SOD showed different variations in the genotypes (declined in Hybrid and Marsellesa; maintained in Hybrid), compared to the values at 37/28 • C. Noteworthy is the common response regarding all genotypes and both [CO 2 ] (although greater in aCO 2 ) at this extreme temperature regarding the activity of CAT and its gene expression. The complementary ability to control the oxidative stress promoted by H 2 O 2 through CAT reinforcement could be of particular importance due to the concomitant severe decline of cellular APX activity in all genotypes in both [CO 2 ] conditions. These findings confirm the high heat sensitivity of APX found in Coffea spp. since it was the greatest negatively affected enzyme under 42 • C [33]. Furthermore, it points to a change of H 2 O 2 control from chloroplast APX [33] or other cellular APXs ( Figure 1) until 37/28 • C to an extra-chloroplast control through CAT (which is predominantly located in mitochondria, peroxisomes, and glyoxissomes) at higher temperatures (42 • C). This is of recognized great importance since H 2 O 2 is capable of diffusing passively across membranes, turning the extra chloroplastic scavenging systems into important H 2 O 2 detoxification pathways [24,67,81,82].
Regarding GR, the higher (Geisha 3 and Hybrid) or stable (Marsellesa) activity at the two highest temperatures will contribute to regenerating GSH and indirectly ascorbate in the ascorbate-glutathione cycle [67,83,84]. In most cases, a significant up-regulation both at 37/28 • C and 42/30 • C was observed in eCO 2 plants, as compared with their values at 25/20 • C. However, under such supra-optimal temperatures, the eCO 2 greatly reduced the transcript abundance of CAT, CuSODs, and APXs, frequently below 50%, with striking attenuations in some cases (e.g., to 12% in CuSOD2, Hybrid, 37/28 • C), compared with their aCO 2 counterparts. Notably, among the studied enzymes, the activity of CAT was the best example of an eCO 2 -induced down-regulation in the three genotypes ( Figure 2). Similarly, the activity of APX also declined in Geisha 3 (42/30 • C), Marsellhesa, and Hybrid (both at 37/28 • C). Higher gene expression and enzyme activity under aCO 2 (than at eCO 2 ) suggests a greater presence of ROS and, therefore, a higher need to control them, namely of H 2 O 2 , which is a known signaling molecule that also triggers the expression of genes encoding its scavenging enzymes [82,85]. The relaxation of part of the antioxidant system (considering both the activity of some enzymes and the gene expression of this type of enzymes) under eCO 2 can be interpreted as reflecting a lower need for a robust antioxidant system [33,86]. That is a consequence of the presence of higher C-assimilation associated with greater photochemical use of energy (data not shown [34]) and minor photorespiration, the latter directly decreasing H 2 O 2 production [87,88]. This ultimately reduces the energy and H 2 O 2 pressure on the photosynthetic apparatus, always when compared with aCO 2 plants. Additionally, Geisha 3 plants under eCO 2 showed a greater SOD activity than that of aCO 2 plants, although with a lower gene expression. The overexpression of genes coding for Cu, Zn-SOD, and greater SOD and APX activities has been associated with a greater tolerance to oxidative stress in transgenic tobacco (Nicotiana tabacum L.) [89] since a greater SOD activity indicates a stronger control of superoxide radicals, although with an increase in H 2 O 2 production that could be controlled through APX and CAT action. The antioxidant response under eCO 2 and its potential mitigating effect under stress conditions have been addressed in other species. The wide variety of obtained results showed not only a strong species-dependency but also varying as well between different cultivars, as reported in this study. For instance, when submitted to heat and drought stress, Arabidopsis thaliana plants showed high levels of ascorbate and CAT activity under eCO 2 [90], whereas in soybean (Glycine max (L.) Merr.) declines in SOD, CAT, APX, and GR activities were observed [91], and in alfalfa (Medicago sativa L. cv. Aragón) reductions of antioxidant molecules and CAT activity was also reported [86].
After the recovery period (Rec14), with very few exceptions (e.g., CAT in Hybrid under aCO 2 and in Geisha 3 under eCO 2 ), the activity of the antioxidative enzymes declined in comparison with the values observed at 42/30 • C (or were maintained when such values were already similar to controls) regardless of [CO 2 ] conditions. That was fully in line with strong declines in transcript abundance of CAT, SODs, and APXs, thus approaching the initial expression levels under control. However, under aCO 2, a higher expression of these genes was maintained when compared with eCO 2 . This pattern was particularly evident for CuSOD2 in Marsellesa and Hybrid and the APX genes in the three genotypes, which points out that a greater need for ROS scavenging is still present in aCO 2 plants 2 weeks after the end of stress exposure.

Expression of Genes Associated with Other Protective Molecules
Besides the antioxidant enzymes and photoprotective pigments, other protective molecules were assessed through gene expression studies. These molecules are reported to have crucial roles in the maintenance of cellular homeostasis under several environmental stresses, including heat stress [92,93]. Transcript levels of protective molecules, other than the antioxidative enzymes discussed above, i.e., chaperonins 20 and 60 (Chape20 and Chape60), early light-inducible protein (ELIP), and 70 kD Heat Shock Proteins (HSP70s), showed a similar pattern of variation. These confirmed (1) an absence of response to the single exposure to eCO 2 in the expression of these genes in the three genotypes (similarly to the findings of [33] Regarding the specific roles of the proteins encoded by these genes, chaperonins ensure the correct folding of new proteins, especially plastid proteins (e.g., RuBisCO), thus playing an important role in heat stress tolerance [94][95][96]. ELIPs are found in thylakoid membranes and protect plants under different environmental stresses since they can participate in the antioxidative stress response by dissipating excess energy and preventing the formation of radicals [97][98][99]. As for HSP70s, although in this study, the maximum transcript accumulation was found at 42/30 • C, the greatest abundance of the protein was previously found at 37 • C [33]. These proteins are considered one of the most important protective molecules in plant responses to stress [95,100,101]. Furthermore, previous studies in Coffea spp. showed that the HSP70 protein synthesis is among the earlier responses to high temperatures [33], but also to moderate and severe drought, its presence further amplified under the superimposition with eCO 2 [29]. In the present study, a lower HSP70 expression under eCO 2 might also point to a lower need for stress protection and enhanced thermotolerance in these coffee plants since the up-regulation of HSP70 genes and the greater presence of this protein are usually found under stress conditions [36,95]. Furthermore, HSP70s are involved, among other roles, in PSII repair, and a positive correlation between the expression of the gene encoding for APX and heat shock transcription factors (HSFs) was reported in transgenic plants of Arabidopsis under heat stress [102]. As terminal components in the signal transduction chain triggered by heat stress, HSFs bind to the heat shock elements (HSEs) involved in downstream heat-inducible genes, playing a central role in the heat response stress [103]. Thus, a greater presence of HSP70 could also contribute to protecting coffee plants from oxidative stress. Overall, the lower gene expression of these protective molecules (HSP70, ELIP, Chape20; Chape60) under eCO 2 plants when compared with their aCO 2 counterparts reinforces the suggestion of a lower need for protecting photosynthetic components from photoinhibition due to increased use of energy through photochemical processes (and lower photorespiration rate), as reflected in the observed photosynthetic performance under eCO 2 [34,53], which is the better photoprotective mechanism against the build-up of energy overpressure in the chloroplast structures [34].
Our findings highlight that, although, with similar patterns of response, a genotypedependent response to heat and/or eCO 2 was clear, both among the C. arabica genotypes studied here or regarding previously studied ones where a stronger response relative to protective molecules (e.g., HSP70) was observed [27,36]. Such relevant differences were also found between C. arabica and C. canephora genotypes [28] and strongly highlight the need to search for thermotolerance biomarkers to be used in breeding programs [54]. Also, some "hybrid vigor" might occur in the studied Hybrid plants when compared to their parental genotypes. This was supported by the greatest responsiveness of all of the studied genes associated with proteins linked to protective roles. Still, that was clear only in aCO 2 plants at 42/30 • C, which might also suggest a greater need for protective mechanisms, thus needing further studies to understand if this wide up-regulation configures a greater vigor or, by opposition, a stronger sensitivity to the imposed conditions. Both issues regarding the genotype-dependent response to heat and/or eCO 2 and the "hybrid vigor" potential strongly advise the implementation of accurate breeding and selection programs to get new coffee genotypes. These, together with the use of adequate crop management practices, such as agroforestry [40,104], will be decisive in guaranteeing the environmental and economic sustainability of this crop.

Conclusions
Overall, eCO 2 alone barely altered most protective components as compared with aCO 2 . Most xanthophylls were maintained in Geisha 3 and Hybrid plants, but Marsellesa tended to rise the concentration of neoxanthin, lutein, the pool of V + A + Z, (α + β) carotene, and total carotenoids, improving its photoprotective ability. The enzyme activities did not increase upon the single effect of eCO 2 (with the exception of APX in Geisha 3), in line with the absence of significant modifications of genes associated with antioxidant enzymes and other protective proteins, always relative to aCO 2 plant counterparts.
Temperature was the main driver of plant response at 37/28 • C and 42/30 • C., At 37/28 • C several carotenoids (particularly neoxanthin and lutein) increased in all genotypes, whereas at 42/30 • C a moderate increase of zeaxanthin and DEPS was found, likely contributing to preventing the formation of highly reactive molecules of O 2 and Chl when the photochemical use of energy would be quite depressed. Marsellesa showed greater additional carotenoids photoprotection at 37/28 • C than the other genotypes, including maximal (α + β) carotene concentration, although the latter declined in all genotypes at 42/30 • C.
With the exception of VDE (down-regulated with heat), all genes encoding for antioxidative enzymes (CAT, CuSODs, APX) or other protective proteins (HSP70, ELIP, Chape20, Chape60), exhibited a high responsiveness to single heat in all genotypes, with a strong up-regulation at 37/28 • C. This was usually even greater at 42/30 • C, with the genes associated with antioxidative enzymes showing the greatest transcripts abundance. Accordingly, 37/28 • C greatly promoted the activity of APX and CAT in all genotypes (likely controlling H 2 O 2 presence), and of GR (except in Marsellesa), but not of SOD (that tended to decline despite the large increase of SODs transcripts). CAT activity deserves a special mention since it increased even at 42/30 • C, thus compensating for the strong APX activity decline at this temperature in all genotypes. An increased potential thermotolerance could arise also through the reinforcement of HSP70, ELIP, Chape20, and Chape60 molecules, as supported by a large up-regulation of their associated genes. Hybrid plants showed the greatest gene up-regulation of most genes under 42/30 • C.
In contrast with the low responsiveness to single eCO 2 exposure, the superimposition with high temperatures revealed an interaction between heat and eCO 2 . The eCO 2 plants showed mostly stable carotenoid values up to 37/28 • C or even 42/30 • C, but only Marsellesa plants denoted greater photoprotective capabilities at 42/30 • C, through greater concentrations of lutein, V + A + Z, and β-carotene, as compared their aCO 2 plants. The eCO 2 attenuated the gene up-regulation observed in aCO 2 plants under heat and lower lowered the antioxidative system (CAT activity is the best example), pointing to a lesser need for ROS control (e.g., H 2 O 2 ) supported by the persistence of photochemical use of energy and low photorespiration in eCO 2 plants.
An incomplete recovery by Rec14 was suggested in the Hybrid and Marsellesa (regardless of [CO 2 ]) due to greater zeaxanthin (and DEPS) values than in control, reflecting the persistence of a thermal dissipation need. Although most antioxidant enzyme activities declined towards their initial values, the partial recovery in all genotypes was further pointed out by the maintenance of up-regulation of all genes, usually much greater under aCO 2 than in eCO 2 counterparts, what denoted a better/faster recovery under eCO 2 . Still, higher levels of antioxidant components (molecules and gene expression) in aCO 2 plants can also be seen as a protection strategy, allowing the plants to better endure new stress events, whereas, in eCO 2 counterparts, such a role could be performed by the greater photochemical use of energy.
Altogether, C. arabica plants responded to heat and/or eCO 2 in a genotype-dependent manner, with a greater acclimation potential in Marsellesa. Heat was the main response driver for the addressed protective molecules and genes, whereas eCO 2 alone did not greatly altered plant status but usually attenuated the heat response, likely supported by a greater use of energy through photochemistry. The importance of the acclimation process and their responsiveness turn these molecules/genes useful biomarkers to breeding and selection programs, which should also explore the heterosis advantages that might arise from select hybrid crosses. Together with adequate management practices, these can help this crop to thrive under global warming conditions.