Future Climate CO2 Levels Mitigate Stress Impact on Plants: Increased Defense or Decreased Challenge?

Elevated atmospheric CO2 can stimulate plant growth by providing additional C (fertilization effect), and is observed to mitigate abiotic stress impact. Although, the mechanisms underlying the stress mitigating effect are not yet clear, increased antioxidant defenses, have been held primarily responsible (antioxidant hypothesis). A systematic literature analysis, including “all” papers [Web of Science (WoS)-cited], addressing elevated CO2 effects on abiotic stress responses and antioxidants (105 papers), confirms the frequent occurrence of the stress mitigation effect. However, it also demonstrates that, in stress conditions, elevated CO2 is reported to increase antioxidants, only in about 22% of the observations (e.g., for polyphenols, peroxidases, superoxide dismutase, monodehydroascorbate reductase). In most observations, under stress and elevated CO2 the levels of key antioxidants and antioxidant enzymes are reported to remain unchanged (50%, e.g., ascorbate peroxidase, catalase, ascorbate), or even decreased (28%, e.g., glutathione peroxidase). Moreover, increases in antioxidants are not specific for a species group, growth facility, or stress type. It seems therefore unlikely that increased antioxidant defense is the major mechanism underlying CO2-mediated stress impact mitigation. Alternative processes, probably decreasing the oxidative challenge by reducing ROS production (e.g., photorespiration), are therefore likely to play important roles in elevated CO2 (relaxation hypothesis). Such parameters are however rarely investigated in connection with abiotic stress relief. Understanding the effect of elevated CO2 on plant growth and stress responses is imperative to understand the impact of climate changes on plant productivity.


INTRODUCTION
The changing earth's atmosphere includes a gradual increase in CO 2 to possibly double the current concentration (IPCC, 2012). Such increase in primary carbon (C) source, will affect plant metabolism, growth, and development (fertilizing effect), especially under favorable water and nutrient conditions. Such effect may be transient, and differ among plant groups, particularly between C3-and C4-type photosynthesis. This subject is extensively reviewed (e.g., Long et al., 2006;Ainsworth et al., 2008;Albert et al., 2011;Dieleman et al., 2012;Xu et al., 2013Xu et al., , 2015Huang and Xu, 2015;Pandey et al., 2015;Kimball, 2016), and is commonly covered in plant physiology text books.
Less documented are the interactions of elevated CO 2 with plant responses to the environment, such as in stress conditions. Nevertheless, the importance of understanding such interactions, given the high-CO 2 future climate scenario's, is increasingly recognized, (e.g., Mittler and Blumwald, 2010;Feng et al., 2014;Xu et al., 2015), also in text books (e.g., see "stress matrix" in Taiz et al., 2015). One effect of elevated CO 2 on plant responses, is the reduction of stress impact. This is demonstrated at the plant growth level, but also at the level of cellular oxidative damage (e.g., lipid peroxidation, protein oxidation), and at the level of stress-generated reactive oxygen species (ROS; Geissler et al., 2010;Mishra et al., 2013;Zinta et al., 2014;AbdElgawad et al., 2015).
A reasonable number of papers has reported the stressreducing effect of elevated CO 2 . Nevertheless, until recently, this topic was rarely covered in reviews. Now, recent overviews of CO 2 effects in plants, start dedicating attention to this effect (Feng et al., 2014;Misra and Chen, 2015;Xu et al., 2015). It is generally recognized that CO 2 effects on abiotic stress impact vary considerably, and, that the underlying mechanisms remain elusive. It is clear that, in addition to providing extra C, elevated CO 2 induces stomatal closing. This improves water use, protecting against drought stress, and helps to explain reduced impact of ozone stress (reduced uptake). However, reduced oxidative damage and ROS levels under elevated CO 2 , probably involves so called non-stomatal factors (Ghannoum, 2009), including metabolic changes. More specifically, increased C availability, possibly resulting in increased supply of defense (antioxidant) molecules, is often held primarily responsible for improved protection against oxidative damage in elevated CO 2 (antioxidant hypothesis).
This conclusion is indeed supported by studies, showing increased antioxidant levels and/or antioxidant enzyme activities (Lin and Wang, 2002;Geissler et al., 2010;Pintó-Marijuan et al., 2013;Zinta et al., 2014). However, there is a considerable number of reports in which elevated CO 2 had little or no effect on antioxidants, or even decreased their levels (Erice et al., 2007;Pérez-López et al., 2010;Farfan-Vignolo and Asard, 2012;Mishra et al., 2013). This indicates that the stress-mitigating effect of elevated CO 2 cannot be universally attributed to increased antioxidant defenses. A key alternative process probably involved in the effect of elevated CO 2 on oxidative stress, is photorespiration. Elevated CO 2 promotes carboxylation over oxygenation at rubisco, reducing reactive oxygen species (ROS) formation (relaxation hypothesis; Long and Drake, 1991;Booker et al., 1997;Ainsworth et al., 2008;Zinta et al., 2014;AbdElgawad et al., 2015). Moreover, when measured simultaneously, reduced photorespiration correlates well with the decrease in H 2 O 2 and lower oxidative damage levels under high CO 2 in some studies (Aranjuelo et al., 2008;Mishra et al., 2013). Therefore, whether elevated CO 2 reduces stress impact through "increased defense or decreased challenge, " remains unaddressed (e.g., Tausz

A SYSTEMATIC LITERATURE ANALYSIS
To gain insight in this issue, we performed a systematic literature analysis, using an "as complete as possible" collection of studies (Web of Science-indexed), meeting the following criteria; (1) analyzing oxidative stress markers (electrolyte leakage, protein oxidation, lipid peroxidation) and antioxidants (molecules and enzymes), (2) in plant shoots grown under ambient and elevated CO 2 , in presence and absence of abiotic stress. The selection of papers was based on a search using the keyword combination "elevated CO 2 " AND antioxidants OR oxidative stress AND plants, resulting in 238 hits (December 2015). From the combined lists, double references were removed; manuscripts were removed because measurements were not performed on shoots; because they were reviews; or because one of the keywords (elevated CO 2 , antioxidants, oxidative stress) was in fact not studied in detail but occurred only in the extended keyword list. Eventually, 105 papers were analyzed (details in Supplementary Table 1 and Datasheet 1).
To perform the analysis in a quantitative manner, we scored the number of occurrences (observations) in which a given parameter (e.g., lipid peroxidation, ascorbate level, APX activity,...), significantly increased (+), remained unchanged (=), or decreased (−). This number of observations (+ or = or −), is expressed relative, as a fraction (%) of the total number (sum = n) of observations on that parameter. As not all 105 studies report on all parameters, the number of observations is often lower than 105. On the other hand, often papers report on changes in one particular parameter, in multiple measurements and/or conditions, or development stages, with and without elevated CO 2 . Therefore, the number of observations can be considerably higher than 105 (see figures).

ELEVATED CO 2 MITIGATES OXIDATIVE STRESS IN PLANTS
First we addressed the question whether changes in "overall oxidative damage, " induced by C, S, and CS, correlate with observations of "overall changes in antioxidants." Such correlation could point to a causal link, e.g., increased antioxidant defenses being responsible for reduced stress impact in elevated CO 2 (antioxidant hypothesis). Obviously, estimating overall levels is a rather crude approach, and does not eliminate the possibility of elevated CO 2 specifically affecting one or more antioxidant components (analyzed below).
However, changes induced by C, S, and CS in "overall antioxidants" (Figure 1A) are different from changes in oxidative damage and ROS. In particular, stress induces increases in antioxidants less frequent (43%, "+") than increases in oxidative damage (55% "+") or ROS (67% "+"). And, antioxidants were more often reported to not change (50% "="), then this is the case for oxidative damage (36% "=") and ROS (33% "=") changes. Therefore, it appears questionable whether increases in antioxidants can be considered a likely general cause of lowered ROS and oxidative damage in elevated CO 2 .

EFFECT OF ELEVATED CO 2 ON INDIVIDUAL ANTIOXIDANT COMPONENTS
It is conceivable that elevated CO 2 specifically increases particular antioxidants. Such information might be lost in the somewhat "crude" summation of all observations (as done above). We therefore zoomed-in on individual molecules and enzymes from various antioxidant defense pathways ( Figure 1B). This analysis confirms reported increases of the activities of many antioxidant components under abiotic stress (lanes S). Increases were particularly frequently observed for PHEN (54% "+"), DHAR (61% "+"), and TOC (59% "+"). However, for all antioxidant components there is a considerable number of observations indicating no changes (from 22 to 62% "=").
In elevated atmospheric CO 2 during stress exposure, the levels and activities of nearly all molecular and enzymatic antioxidants (except GPX) have been reported to increase, but with variable and often low (10-40%) frequencies. Increases were reported particularly frequently for PHEN (38%), POX (26%), SOD (29%), and MDHAR (40%). None of all antioxidant components accounts solely, throughout all studies, for increased antioxidant defenses in high CO 2 . The frequently observed increases of DHAR and MDHAR, together with ASC and GSH levels reported to remain largely unchanged (56 and 54% "="), may indicate that the ASC/GSH-cycle is often involved in the antioxidant response in elevated CO 2 . However, MDHAR/DHAR activities are reported in only 12 of the 105 papers, and further verification is therefore necessary.

ARE ANTIOXIDANTS SPECIFICALLY RESPONSIVE TO ELEVATED CO 2 IN PARTICULAR SPECIES OR GROWTH CONDITIONS?
We next addressed the question whether increases in antioxidant capacity in elevated CO 2 were possibly specific for metabolism type (C3 vs. C4), species-group, stress type or growth facilities (details in Supplementary Table 1). For example, one could hypothesize that in C4 plants, extra CO 2 is more likely to alleviate stress impact through increased antioxidant capacity, then through suppression of photorespiration. Our dataset contains 54 species with C3, and 8 with C4 metabolism. Heat maps summarizing the observations on oxidative damage and antioxidant changes (Figure 1C), show considerable reduction of stress impact by elevated CO 2 and increases in antioxidants for C3 as well as C4 plants.
A large number (95%) of all species in which abiotic stress, elevated CO 2 and antioxidant changes are studied, are either legumes (10), grasses (18), or trees (22) (Supplementary Table 1). Oxidative damage as a result of abiotic stress, was considerably more frequently observed in the legumes (58%) and trees (68%), compared to grasses (17%) (Figure 1C). Also the reduction in oxidative damage by elevated CO 2 , was more frequently reported in legumes (47%), than in grasses (16%) and trees (11%). It is not immediately clear what the basis is for this difference in responsiveness. Increases in antioxidants are reported equally frequent in each species group.
Another factor that could explain some inconsistency in reports on antioxidant increases in elevated CO 2 , is the variation in growth facility. Elevated CO 2 effects on growth seem less pronounced in FACE experiments, than in growth cabinetexperiments (Ainsworth et al., 2008). On the other hand, in field conditions (e.g., OTC, FACE) stress is often more severe  and prolonged (Mittler and Blumwald, 2010). The literature analysis shows that reduction of stress impact on cell damage, and increases in antioxidants, are reported in all facilities (Figure 1C, insufficient data for FACE).
Finally, we also sorted the observations by stress type (Figure 1C). It is apparent that the stress mitigating effect on oxidative damage occurs for all abiotic stresses, with the notable exception of heavy metal stress. However, the effect of elevated CO 2 on heavy metal stress is only reported in 4/105 papers, and needs further confirmation.
Increases in antioxidants are reported for all stresses, and are therefore not stress-type specific. Interestingly, reported increases in antioxidants for drought and ozone stress, suggest that in addition to stomatal closure by elevated CO 2 , other processes contribute to the reduction of drought and ozone impact. The data sorted by stress type, also illustrate that in the majority of the reports, that whereas antioxidant activities generally increase in response to the stress (S), they do not increase further or even decrease when the stress is combined with elevated CO 2 (CS).

CONCLUSIONS
The question whether mitigation of stress impact by elevated CO 2 occurs through "increased defense or decreased challenge" is probably best answered by "both." Clearly increased defenses have been demonstrated, but only in a minority of the reports, and, this effect is not specific for any particular antioxidant, C3, or C4 metabolism, for a particular species group, growth facilitytype or stress type. This suggests that decreased challenge also plays an important role in stress mitigation in elevated CO 2 . The primary candidate process for this, is reduced hydrogen peroxide production by elevated CO 2 in photorespiration. Effects of elevated CO 2 on photorespiration in stress conditions, have also been reported, and point to effects on photosynthesis (Booker et al., 1997;Aranjuelo et al., 2008;Pérez-López et al., 2012;AbdElgawad et al., 2015). However, despite the possibly important role of photorespiration in stress responses under future climate conditions, these aspects are rarely investigated simultaneously (only 4 of 105 papers investigated photorespiration changes). It is therefore of great interest to further unravel the role of antioxidants and photorespiration in elevated CO 2 effects. Moreover, the causal role of antioxidant increases to reduce oxidative damage and ROS under elevated CO 2 , is almost exclusively inferred from "correlative changes." This conclusion is not, yet, supported by studies performing plant manipulations, e.g., use of mutants, overexpressor lines, or pharmacological treatments. Engineered plants with elevated antioxidant enzymes, sometimes show increased stress resistance (Eltayeb et al., 2007;Lee et al., 2007;Avramova et al., 2015). However, these lines have not been tested under elevated CO 2 . Recently, a mutant screening assay was developed employing the effect of elevated CO 2 on hydrogen peroxide production (Queval et al., 2012), which underlines the relevance of understanding photorespiration in altered CO 2 conditions. However, also in C4 plants, in which photorespiration is not very active, elevated CO 2 reduces ROS levels and oxidative damage, without changes in antioxidants. This suggests that other non-stomatal processes, apart from antioxidant defenses and photorespiration, contribute to stress mitigation. Stress induces ROS production at various cellular sites (e.g., Foyer et al., 1994;Schwanz et al., 1996;Gill and Tuteja, 2010;Miller et al., 2010;Das and Roychoudhury, 2014), which each can be affected by elevated CO 2 (Figure 2). Apart from changes in antioxidants and photorespiration (described above), some evidence indicates that elevated CO 2 reduces mitochondrial and chloroplast ROS formation, and NADPH oxidase activity, but, little is known at the level of β-oxidation or other cellular oxidases (Gonzalez-Meler et al., 1996;Booker et al., 1997;Gonzàlez-Meler and Siedow, 1999;Lin and Wang, 2002). Understanding the effect of elevated CO 2 in plant stress responses is further complicated by the fact that additional C affects multiple metabolic processes, as demonstrated by non-targeted (omics) approaches (e.g., Sicher and Barnaby, 2012;Xu et al., 2014;Zinta et al., 2014;AbdElgawad et al., 2015;Misra and Chen, 2015).
An explanation for the relatively large variety in antioxidant responses to elevated CO 2 , is that these responses possibly occur relatively far downstream from the CO 2 primary targets. As a result, changes in antioxidants may not directly correlate to the stress and CO 2 treatment only, but are an integrated response of changes in various metabolic processes. The primary targets for CO 2 , i.e., where cellular "perception" of altered CO 2 levels first occur, are probably changes in stomatal opening, suppression of photorespiration and increased levels of carbohydrates through increased C fixation. Changes in antioxidants are probably an integrated downstream overall result from changes in these processes. It therefore appears that an important topic, as the response of plants to adverse growth conditions, in future climate-levels CO 2 , is underexplored, and leaves many unanswered questions (also Xu et al., 2015).
Finally, with regard to textbook knowledge, our analysis indicates that positive interactions of CO 2 have been shown for salinity and heat stress. We therefore suggest to update the very instructive Mittler and Blumwald "stress matrix" (Mittler and Blumwald, 2010;Taiz et al., 2015) accordingly.

AUTHOR CONTRIBUTIONS
HAb: Literature analysis and review writing. GZ: Literature analysis and review writing. GB: Review writing. IJ: Review writing. HA: Literature analysis and review writing.