Coordination of Morpho-Physiological and Metabolic Traits of Cistus incanus L. to Overcome Heatwave-Associated Summer Drought: A Two-Year On-Site Field Study

The shrub Cistus incanus L. is well-adapted to Mediterranean conditions thanks to its morpho-anatomical, physiological and biochemical traits. However, its distribution and survival in coastal dunes will be likely threatened by ongoing runaway climate change. We investigated how the harsh climatic conditions generated by the 2015 summer heat wave triggered specific anatomical, physiological, and biochemical responses of this species in its natural environment. These adjustments were compared to those measured in summer 2014. The drier and hotter conditions of summer 2015 determined an increment in leaf lamina thickness, due to a greater palisade parenchyma, thus leading to an increase in the whole leaf mass per area. These morphoanatomical adjustments enhanced leaf resistance against dehydration, optimized carbon assimilation, and delayed leaf senescence. In addition, the higher amount of secondary metabolites detected for 2015, in particular tannins and monoterpenes, contributed to the maintenance of physiological performances even under hotter and drier conditions, preventing oxidative stress through the consumption of excess reducing power. In conclusion, our study offers new evidence on the integration of morphophysiological and metabolic adjustments of this species growing in its natural habitat to cope with ongoing climate change.


INTRODUCTION
Mediterranean climate is characterized by mild and humid winters and hot and dry summers, with most of the rainfalls concentrated in the cold season (Lionello et al., 2006). Summer drought is considered the primary constraint affecting both productivity and distribution of Mediterranean vegetation due to a combination of high irradiance, high temperatures and water scarcity (Gratani et al., 2013;Flexas et al., 2014).
Mediterranean maquis vegetation mostly consists of evergreen sclerophylls with a deep root system and semi-deciduous shrubs, which reduce transpiring surface during stressful periods (Galmés et al., 2007;Medrano et al., 2009). Shrub species, in particular, are a widespread life form in the coastal ecosystem and their potential resilience to environmental change deserves special attentions Galle et al., 2011;Puglielli et al., 2017a). Mediterranean shrubs display a vast array of strategies to cope with severe climatic constraints (Lefi et al., 2004;Sardans and Peñuelas, 2013), which may in part correlate with the large biological diversity encountered in Mediterranean areas suffering from most severe environmental pressures (Van Der Plas et al., 2016). Evergreen species usually display a conservative resource use strategy (David et al., 2007), characterized by low net photosynthesis and relative growth rate, and the utilization of the fresh assimilated carbon for leaf construction (Sardans and Peñuelas, 2013). By contrast, semi-deciduous species display a higher physiological plasticity compared to the evergreen counterparts, with a decline in photosynthetic performances during central hours in summer season, when high solar irradiance occurs in concomitance with severe reduction in soil water availability and excessive air temperature (Di Ferdinando et al., 2014).
Among these metabolic adjustments, the increase in polyphenol contents in stressed plants has been linked to improve photoinhibition tolerance and, in general, to the protection of photosynthetic organs from oxidative damages . Mediterranean shrubs also emit large amounts of biogenic volatile organic compounds (BVOCs), among which mono-and sesqui-terpenes are the most abundant. These compounds play prominent functions against biotic and abiotic stresses as well as in mediating ecological interactions (Loreto and Schnitzler, 2010;Niinemets et al., 2013;Loreto et al., 2014;Mu et al., 2018). Specifically, monoterpenes have been shown to protect leaves against oxidative and thermal damages, decreasing the rate of formation of reactive oxygen species (ROS) that can damage the photosynthetic apparatus (Loreto et al., 1998(Loreto et al., , 2014. It is known that water stress may result in increased monoterpene emission (Ormeño et al., 2007) or in a significant change in the blend composition of stored monoterpenes (Brilli et al., 2013).
Cistus L. is a genus of dicotyledonous plants colonizing open areas of stony and infertile soils (Papaefthimiou et al., 2014). Among Cistus spp., Cistus × incanus L. (pink rock-rose, hereby C. incanus) belongs to the autochthonous Tertiary Mediterranean flora (Quézel, 1985). This species is classified as a summer drought semi-deciduous species, shedding a high percentage of leaves and branches and increasing the mass investments per leaf area unit during drought (LMA, leaf mass per area) (Catoni et al., 2012;De Dato et al., 2013;Grant et al., 2015). The semi-deciduousness allows this species to be well-adapted to summer drought (Correia and Ascensao, 2016;Puglielli, 2019). In addition, also seasonal dimorphism, expressed both in the leaf morphoanatomical traits as well as in the wood properties, may be considered another important adaptation of this species to the Mediterranean harsh climate (Aronne and De Micco, 2001;De Micco and Aronne, 2009;Arena et al., 2013). Indeed, summer leaf anatomy is characterized by a crimped lamina partially rolled to form crypts in the lower surface and a thick palisade parenchyma in the mesophyll tissue (Aronne and De Micco, 2001;Rotondi et al., 2003).
Despite being well-adapted to multiple stress factors, C. incanus distribution and survival in coastal dune areas will be likely negatively affected by the ongoing climatic change conditions (Nogués et al., 2015;Gori et al., 2019). In this respect, very few field investigations have been conducted on the effects of drought and heat stress on this Mediterranean shrub (Gratani et al., 2008;De Micco and Aronne, 2009;Gratani et al., 2018;Parra and Moreno, 2018;Gori et al., 2019). Furthermore, morpho-physiological and metabolic responses of this species to extreme drought remain poorly investigated despite the increasing frequencies of heat waves hitting the Mediterranean region as a consequence of climate change.
In this context, the contrasting climatic conditions of the summers 2014 and 2015 provided an optimum opportunity to study the modifications in this maquis shrub behavior in response to ongoing climate change in its natural environment. In fact, the summer 2014 showed irregular rainfalls and moderate drought, while the summer 2015 was one of the driest and hottest in Europe in the last 70 years as a result of an extreme heatwave that lasted about 30 days until the end of July (Russo et al., 2015).
We took advantage of contrasting climatic conditions of summers 2014 and 2015 to test if changes in morpho-anatomical traits allow to maintain physiological performances and enhance the biosynthesis of secondary metabolites in C. incanus. We hypothesized that semi-deciduous Mediterranean shrubs species adapted to harsh environment, such as C. incanus, are able to face drought and heat stress thank to a coordination of morphoanatomical, physiological and biochemical traits. We expected that higher temperature in 2015 may have promoted a major investment in leaf construction and boosted the production of secondary metabolites, thus reducing water loss and damages to photosynthetic apparatus caused by oxidative stress. Overall, this study provides the substantial contribution of secondary metabolites, both condensed tannins and monoterpenes, to the suite of traits conferring high plasticity to C. incanus and, consequently to its high resilience against extreme environmental pressures associated to climate change.

Plant Material, Study Area, Sampling Design and Meteorological Data
In May 2014, eight healthy individuals of C. incanus were selected in the population previously described in Gori et al. (2019) and screened for size and uniformity. C. incanus plants were 0.6 ± 0.07 m height with a canopy area of around 0.6 ± 0.08 m 2 . The study was performed over 2 consecutive years on cloudless days in summer 2014 and 2015 (8-9 July 2014 and 14-15 July 2015) on the coastal sand dunes located in Castiglione della Pescaia (GR, Italy, 42 • 46 N, 10 • 53 E).

Water Relations
Two leaves per plant were collected in the field and transported to the laboratory in sealed zip-lock, tared plastic bags stored in a fridge bag. The leaves were weighed to determine fresh mass (FW) and then allowed to hydrate until saturation (constant weight) for 48 h in darkness to determine turgid mass (TW). Then, leaves were dried in an oven at 80 • C for 48 h to determine dry mass (DW). Finally, Relative Water Content (RWC) was calculated using the following equation: Leaf water potentials (Ψ w ) were directly measured in field, after the cut of two twigs per plant, using a Scholandertype pressure chamber (PMS Instruments, Corvallis, OR). The measurements of RWC and Ψ w obtained from the same plant were combined to make an individual replicate.

Morphological and Anatomical Measurements
From the upper part of the canopy 10 adult leaves per plant were sampled at midday to measure LMA by the calculation of the dry mass (drying the leaves in oven at 80 • C for 48 h) to leaf area ratio (mg cm −2 ). One twig per plant was collected and transported to the laboratory for anatomical measurements. Light microscopy segments of leaf tissue were taken from the basal-mid-lamina regions of eight adult leaves and fixed for 2 h at room temperature in 0.2 M phosphate buffer (pH 7.2) with 2.5% glutaraldehyde (Sigma Aldrich). Then, they were washed twice in the same buffer and post-fixed in 2% osmium tetroxide (Sigma Aldrich) for additional 2 h. Following dehydration in a graded ethanol series (30, 40, 50, 60, 70, 80, 95, and 100%), the samples were gradually embedded in Spurr resin (Sigma Aldrich) and polymerized at 70 • C for 24 h (Spurr, 1969). Semi-thin sections, 1-2 µm, were affixed to glass slides and observations were carried out in a Leica DM LB2 Light Microscope (Leica Microsystem, Germany).

Gas Exchanges and Chlorophyll Fluorescence
Measurements of gas exchanges and photosystem II maximum efficiency ( PSII ) were performed on two fully expanded and healthy leaves per plant, selected in the most illuminated part of the canopy, utilizing a LI-6400 portable photosynthesis system (Li-Cor, Lincoln, NE, United States) equipped with a cuvette of 2 cm 2 . Photosynthesis (P n ), stomatal conductance (g s ), 1 http://lamma.rete.toscana.it intercellular CO 2 concentration (C i ) and PSII were measured at (CO 2 ) of 400 µmol mol −1 and at the photosynthetic photon flux density recorded in the environment. The integral of daily photosynthesis (P n ), calculated using Sigma Plot 11.0 (Systat Software, Inc., San Jose, CA), has been employed to quantify the amount of carbon fixed during the day. Then, mesophyll conductance to CO 2 (g m ) was calculated using the variable J method (Harley et al., 1992) as follows: where , representing the CO 2 compensation point to photorespiration, was obtained from data reported in Galle et al. (2011), while light respiration (R d ) was calculated using the Kok method at PPFD steps of 150, 100, 80, 60, and 30 µmol m −2 s −1 (Kok, 1948).
Electron transport rate (J F ) was calculated from chlorophyll fluorescence: where PSII is the actual photochemical efficiency of Photosystem II (Genty et al., 1989), the partitioning factor (β) between photosystems I and II was considered to be 0.5 and leaf absorbance (α) 0.84. Moreover, we use the ratio J F /P n as an indicator of the excess of electron transport which can be diverted to other alternative electron sinks rather than photosynthesis (Arena and Vitale, 2018;Brunetti et al., 2018).
After gas exchanges, F v /F m was measured on the same leaves using a portable chlorophyll fluorescence system (Handy Pea, Hansatech, Norflolk, United Kingdom). Before measurements, leaves were adapted to the dark for 30 min. Gas exchange and chlorophyll fluorescence measurements obtained from the same plant were combined to make an individual replicate.

Non-destructive Determination of Chlorophyll and Epidermal Flavonols
At each daily sampling hour, four leaves per plant were measured with a Dualex R device (Dualex R Research, FORCE-A R , Orsay, France) on adaxial and abaxial side. Chlorophyll index (Chl i ) and epidermal Flavonol index (Flav i ) were obtained as described in Agati et al. (2016). In particular, the Chl i was calculated as the average of the adaxial and abaxial measurements, while the Flav i was calculated as the sum of both measurements. All measurements obtained from the same plant were combined to make an individual replicate.

HPLC Quantification of Polyphenols
Two leaves per plant were collected, immediately frozen in liquid nitrogen, stored at −80 • C and then lyophilized. In detail, lyophilized samples (150 mg of DW) were extracted twice with 5 mL of ethanol/water (75/25) adjusted at pH2.5 with formic acid and the supernatant partitioned with 3 × 5 mL of n-hexane. The ethanol fraction was reduced to dryness, and the residue was rinsed with 1 mL of methanol/water (90/10). Aliquots of 10µL were injected into the Flexar liquid chromatography equipped with a quaternary 200Q/410 pump and an LC 200 diode array detector (DAD) (all from Perkin Elmer R , Bradford R , CT, United States). Identification and quantification of these metabolites was carried out using retention times and UV spectral characteristics of authentic standards, as well as based on literature data (Gori et al., 2016). The total polyphenols content (Pol Tot ) consists in the sum of individual polyphenols detected, namely tannins (Tan) and flavonols (Flav).

GC-MS Quantification of Biogenic Volatile Organic Compounds (BVOCs)
The emission of BVOCs was measured on the same leaves used for gas exchange analyses by partially diverting the outlet of the cuvette of the Li-Cor system described above into a silico-steel cartridge packed with 200 mg of Tenax (Agilent, Cernusco sul Naviglio, Italy). A volume of 4 dm 3 of air was pumped through the trap at a rate of 200 cm 3 min −1 utilizing an environmental air sampling equipment (A.P. BUCK, Inc., Orlando, FL). Prior to the first enclosure of the leaves, blank samples from the empty cuvettes were taken every sampling day. The cartridge was analyzed using a Perkin Elmer Clarus 580 gas chromatograph coupled with a Clarus 560 Mass-Selective-Detector and a thermal desorber TurboMatrix (Perkin Elmer Inc., Waltham, MA, United States) operating for 25 min at 250 • C. The desorbed compounds were separated in a 30-m Elite-5-MS capillary column. The temperature of the column was first maintained at 40 • C (5 min), then increased with a 5 • C min −1 175 ramp to 250 • C, and maintained at 250 • C for 2 min. BVOCs were identified using the NIST library provided with the GC/MS Turbomass software. The authentic standards of α-pinene, β-pinene and myrcene (Sigma Aldrich, Milan, Italy) were used to prepare calibration curves as well as to compare the peak retention time and the peak fragmentation of monoterpenes found in the samples. Authentic standards of other compounds collected in the traps (tujene, α-phellandrene, and β-phellandrene) were not available, therefore the amount of these compounds was calculated using the calibration curve of α-pinene. Total BVOC emission (BVOCs Tot ) consists in the sum of individual monoterpenes detected in the samples.

Statistical Analysis
Statistical analysis was performed using the SPSS software program (SPSS, Inc., Chicago, IL, United States). To test the daily trend of parameters, we performed a two way ANOVA followed by Tukey's post-hoc test. When a mean value was reported for each summer, Student's t-tests were performed to evaluate differences between the two years. Linear regression analyses were performed to determine the effect of air temperature on gas exchanges and chlorophyll fluorescence parameters (P n , g s , F v /F m and PSII ) and water relations (Ψ w and RWC) considering the data collected in four different sampling hours (8:00, 12:00, 15:00, and 18:00) and on total polyphenols content (Pol Tot ) and BVOCs emission considering the data collected at 12:00. The linear regression analysis was also applied to determine the differential contribution of the morpho-anatomical traits (upper and lower epidermis, palisade and spongy parenchyma) to the total leaf thickness (LT) and to the gas exchange and chlorophyll fluorescence parameters (P n , g s , F v /F m , and PSII ). With the same set of data, a principal component analysis (PCA) was applied to gas exchange and chlorophyll fluorescence parameters (P n , g s , F v /F m , and PSII ), water relations (RWC and Ψ w ) and morpho-anatomical (leaf thickness) or total polyphenols content (Pol Tot ), biogenic organic compounds (BVOCs) emission and morpho-anatomical (leaf thickness) in order to distinguish the trait syndromes between the year of sampling. The PCA was calculated based on a correlation matrix using OriginPro 2020 software.

Meteorological Data
In July 2014, the average maximum temperature (T max ) was 29 • C, almost 5 • C lower than that recorded in July 2015. Average minimum temperatures (T min ) were 17.7 and 21.1 • C in July 2014 and 2015, respectively (Table 1). Hence, summer 2015 was unusually dry and hot as a consequence of a severe heat wave that hit the Mediterranean regions for about 30 days between the end of June and July (Russo et al., 2015). During the 2 months prior measurements cumulative rainfall was 84 mm in 2014 and 29 mm in 2015 ( Table 1). During the days of measurements, temperatures ( Figure 1A) and global irradiance ( Figure 1B) had a typically daily trend, showing higher values during the central hours. The daily irradiance did not change between the two summers ( Figure 1B). Whereas, in 2014 the daily time course of temperature was consistently lower than in 2015 ( Figure 1A), with minimum and maximum temperatures of 18 • C and 30 • C in 2014, and 22 • C, and 33 • C in 2015, respectively ( Figure 1A).

Morpho-Anatomical Leaf Traits
The microscopy analysis revealed significant variations among the samples collected in summer 2014 and 2015 (Figure 2A and Table 2). In both years, leaves showed a typical structure of a heliophylous plant: a dorso-ventral orientation, a high mesophyll density, with a well-developed palisade consisting of two/three layers of elongate-prismatic cells and a spongy parenchyma with a few airspaces (Figure 2A). In the leaves collected in 2015, the vacuolar accumulation of tannins, as evidenced through osmium staining, as spherical deposits or adjacent to the internal surface of the vacuolar membrane; condensed tannins fill almost completely the vacuoles of adaxial epidermis and palisade  Average maximum and minimum temperatures ( • C) and cumulative rainfall (mm) referred to May, June, and July prior the days of measurements. Data of the weather station "Grosseto" (Italy) from the regional archive of the LaMMA Consortium (http://lamma.rete.toscana.it).
Frontiers in Ecology and Evolution | www.frontiersin.org parenchyma cells, whereas occur in fewer cells or in smaller quantities in spongy and abaxial epidermis cells (Figure 2A). In addition, leaves collected in 2015 showed a significant increase in the thickness of almost all leaf tissues compared to those collected in 2014, i.e., increases of 44.2% (t-test; p ≤ 0.001) in the palisade and of 28.2% (t-test; p = 0.024) in the lower epidermis ( Table 2). A slight, but not significant increment, was observed also in the spongy parenchyma. The palisade parenchyma was the leaf tissue that mostly contributed to increase the leaf thickness in 2015 presenting a higher coefficient of determination (R 2 = 0.66) when compared to the other tissues (spongy mesophyll R 2 = 0.35, upper epidermis R 2 = 0.44 and lower epidermis = regression not significant) ( Figure 2B and Supplementary Table S1B). A strong, significant increase in LMA (+34.5%) (t-test; p = 0.002) was observed in 2015, as LMA ranged from 11.2 mg cm −2 in July 2014 to 17.1 mg cm −2 in July 2015 ( Table 2).
way ANOVA; p < 0.001) and at 18:00 (two way ANOVA; p < 0.001), respectively. The PCA identified groups of plants with similar water relations, gas exchange and morpho-anatomical features, with the first two components of the PCA explaining 41.58 and 27.74% of the variances. In the biplot (Supplementary Figure S1A) samples collected in different years were segregated in distinct groups.
The samples collected in 2015 were mainly grouped at the negative score of PC1 presenting lower values of w and RWC than samples collected in 2014 and grouped at the positive score of PC2 for g s and leaf thickness (LT). The P n vector contributed less for distinguishing both years.

Chlorophyll Fluorescence and Chlorophyll Index
Photosystem II maximum efficiency (F v /F m ) showed a typical midday depression at 12:00 in both years ( Figure 5A) (two way ANOVA; p < 0.001) with a negative correlation with air temperature along the day (Supplementary Table S1A). However, while in summer 2014 F v /F m recovered optimal values at 15:00, in 2015, this parameter fully recovered only at 18:00. In addition, F v /F m was significantly higher in 2014 than in 2015 at 8:00 (two way ANOVA; p = 0.024) and at 15:00 (two way ANOVA; p < 0.001) (Figure 5A). Photosystem II actual efficiency ( PSII ) differed between the 2 years, with significant lower values in 2014 than in 2015 (two way ANOVA; p < 0.001).
The minimum values of PSII were observed from 12:00 to 15:00 in both summers ( Figure 5B). The chlorophyll index (Chl i ) was significantly higher in 2014 than in 2015 at all sampling hours (two way ANOVA; p < 0.001) ( Figure 5C). However, the daily trends of this parameter were similar in both years, with a reduction during the central hours of the day followed by a recovery at 18:00 ( Figure 5C). The PCA identified groups of plant species with similar chlorophyll fluorescence and chlorophyll index as previously described for gas exchanges and water relations, with samples collected in different years segregated in distinct groups (Supplementary Figure S1A). According to the PCA analysis the samples collected in 2015 were mainly grouped at the negative score of PC1 presenting the lower values of F v /F m however, PSII vector contributed less for distinguishing both years.

Flavonol Index and Polyphenols
The Flavonol index (Flav i ) exhibited very similar values and daily trends in the two summers, increasing from 8:00 to 12:00 and then declining from 12:00 to 15:00 until 18:00 ( Figure 6A). Moreover, at 12:00, the value of Flav i was significantly higher in summer 2014 than in 2015 (two way ANOVA; p = 0.026) ( Figure 6A). Total polyphenols (Pol Tot ) were higher in 2015 than in 2014 (t-test; p = 0.022) ( Figure 6B) with a positive relation with air temperature (Supplementary Table S1), because of the significant difference in the total content of tannins, passing from ∼ 51 µmol g −1 dw −1 in 2014 to ∼70 µmol g −1 dw −1 in 2015 (t-test; p = 0.002) (Figure 6C). In contrast, the total content of flavonols did not change from 2014 to 2015 (t-test; p = 0.183) (Figure 6C).

Biogenic Volatile Organic Compounds
The total emission of BVOCs increased about fivefold in 2015 compared to 2014 (t-test; p < 0.001) ( Figure 7A) presenting a positive relation with air temperature (Supplementary Table S1). All BVOCs detected in the study were monoterpenes. The most important monoterpenes emitted by C. incanus leaves were αand β-pinene, their sum accounted for about 69% of the total emitted compounds in 2014 and for 60% in 2015 ( Figure 7B). Other monoterpenes detected were: α-, βphellandrene, thujene, and myrcene. In summer 2015, the emission of α-pinene, β-pinene and myrcene were about five times higher compared to summer 2014 (t-test; p < 0.001; p = 0.013; p = 0.011). The emissions of β-phellandrene and thujene increased three and nine times, respectively (t-test; p = 0.005; and p < 0.001), while α-phellandrene emission was about 15 times higher in July 2015 than July 2014 2014 (t-test; p < 0.001) (Figure 7B). The PCA identified groups of plant species with similar BVOCs (and also for Pol Tot ), with samples collected in different years segregated in distinct groups (Supplementary Figure S1B). The samples collected in 2015 were grouped at the positive score of PC1 (84.16% of the total variation), presenting the higher values for Pol Tot and BVOCs as well as for the leaf thickness (LT).

DISCUSSION
Leaves of summer 2014 maintained a better water status compared to the drier and hotter summer 2015 (Table 3). In particular, the higher leaf dehydration observed during the heat wave in July 2015, as estimated by the lower RWC, led to a consequent drop in Ψ w ( Table 3). Leaf morphoanatomical traits may have played an important role to avoid leaf shrinkage associated to dehydration during drought. Indeed, when water is a limiting factor, plants undergo anatomical alterations, particularly in their leaves, in which the transpiration flow is regulated by stomatal closure (Bosabalidis and Kofidis, 2002;Scoffoni et al., 2014). It has already been reported that the dimorphic C. incanus develops xerophytic leaves during summer Puglielli, 2019). In our study, total lamina thickness was increased by 28% in 2015, as compared to 2014, mainly due to a higher palisade parenchyma (+44%) and to a greater epidermal thickness (+28%) ( Table 2 and Supplementary Table S1B). The smaller leaf size (Supplementary Figure S2) and the greater contribution of upper epidermis and of mesophyll tissues, in particular palisade parenchyma, to the whole leaf structure (Figure 2B and Supplementary Table S1B), could have been important to increase resistance against dehydration and cell collapsing induced by the drier conditions of summer 2015 (Bosabalidis and Kofidis, 2002;De Micco and Aronne, 2012;Mansoor et al., 2019). We hypothesize that accumulation of leaf condensed tannins, both in upper epidermis and mesophyll cells (Figure 2A), may have contributed to strengthen the leaf structure of C. incanus in 2015 as previously reported for other species grown in xeric environments Gravano et al., 2000;Bussotti, 2008). We suggest that the increment in leaf condensed tannins and the higher LT observed in the drier FIGURE 5 | Diurnal trends of maximal photochemical efficiency of PSII (F v /F m ) (A), actual efficiency of PSII ( PSII ) (B), and Chlorophyll index (Chl i ) (C) of C. incanus leaves in July 2014 and 2015. Two way ANOVA followed by Tukey's pairwise comparison was performed to compare data (means ± SD; n = 8). Letters indicate significant differences (p ≤ 0.05) among sampling hour for each year, whereas asterisks represent significant differences among years for each sampling hour, *(p ≤ 0.05), **(p ≤ 0.01), ***(p ≤ 0.001).
FIGURE 6 | Epidermal Flavonol index (Flav i ) (A), total polyphenols content (Pol Tot ) (B), condensed tannins (Tannins) and flavonols leaves content (Flavonols) (C) of C. incanus leaves in July 2014 and 2015. Two way ANOVA followed by Tukey's pairwise comparison was performed to compare data (means ± SD; n = 8). Letters indicate significant differences (p ≤ 0.05) among sampling hour for each year, whereas asterisks represent significant differences among years for each sampling hour, *(p ≤ 0.05), **(p ≤ 0.01), ***(p ≤ 0.001). and hotter summer 2015 resulted in a significant increase in LMA (+36%). Indeed, previous studies have clearly shown a positive relationship between LMA and LT (Puglielli et al., 2017c(Puglielli et al., , 2019bGratani et al., 2018) and a possible contribution of soluble phenolics to LMA (Poorter et al., 2009). The higher LMA may have provided both a lower evaporative surface area and a lower light harvesting capability per unit of leaf mass ( Table 2; Boughalleb and Hajlaoui, 2011;De Micco et al., 2011;Perez-Martin et al., 2014;De la Riva et al., 2016), enabling plants to tolerate lower Ψ w without significant losses in leaf turgor (Munné-Bosch et al., 2003). LMA may have also an influence on leaf physiology by affecting the diffusion of CO 2 from substomatal cavities through intercellular air space to sites of carboxylation in the mesophyll tissues (Loreto and Centritto, 2008;Flexas et al., 2014;Peguero-Pina et al., 2017). In particular, several studies on non-stressed tree species have shown that a higher LMA constrains leaf mesophyll conductance (g m ), thus limiting photosynthetic assimilation rates (Syvertsen et al., 1995;Hassiotou et al., 2009). Nevertheless, recent evidences have shown that, when a higher LMA is accompanied to an increment in thickness of palisade parenchyma, this anatomical adaptation may be accompanied by a higher diffusion and availability of CO 2 across the leaf, facilitating the assimilation process (Evans, 1999;Milla-Moreno et al., 2016;Peguero-Pina et al., 2017). Similarly, in our study, the increment in LT and LMA in July 2015 may have led to a higher g m and, in turn, resulted in a higher CO 2 uptake at midday compared to July 2014 ( Figure 3A and Table 3). This result is in accordance with previous research on droughtacclimated plants, in which a positive relationship between g m and mesophyll thickness was found, supporting the hypothesis that, under water stress, a thicker leaf increases the surface area of chloroplasts exposed to intercellular airspace per unit leaf area (Hanba et al., 1999;Terashima et al., 2001;De Micco et al., 2011;Galmés et al., 2013;Flexas et al., 2014). The higher LMA may have also contributed to lower leaf temperature in summer 2015 (Gratani and Varone, 2004), as similar daily leaf temperatures were observed between the 2 years (Supplementary Figure S3), despite the higher air temperatures caused by the heatwave in July 2015 ( Figure 1A; Russo et al., 2015). In addition, the peculiar location of stomata inside epidermal crypts may have allowed the higher g s values in 2015, which contributed to the regulation of leaf temperature and the maintenance of similar P n between the 2 years ( Figure 3B and Supplementary Figure S1A; Aronne and De Micco, 2001). We cannot exclude that other leaf morpho-anatomical adjustments (e.g., leaf rolling, variations in leaf angles and density in leaf pubescence) may have contributed to protect the photosynthetic apparatus under the drier and hotter conditions of July 2015 (Figures 3A, 4A and Supplementary Figure S2; De Micco and Aronne, 2012;Arena et al., 2019). Particularly, paraheliotropism is an adaptative strategy of C. incanus to counteract the deleterious impact of high irradiance to protect leaves under stress condition, especially in summer, thus avoiding photoinhibition (Puglielli et al., 2017b;Pérez-Llorca et al., 2019). However, the P n time course in 2015 ( Figure 3A) showed the typical trend of a severe water-stressed plants, with a higher value in the morning followed by significant drop as the day progressed (Marino et al., 2014). In contrast, in 2014 at 15:00, the peak in P n coupled with the drop of C i (Figures 3A, 4B) indicated the lack of significant biochemical limitations to photosynthesis (Lawlor and Cornic, 2002). Conversely, in summer 2015, C i did not change significantly throughout the day with respect to the morning value, despite the significant drop in P n . This result might be associated to leaf rolling and suggests that, despite C. incanus leaf photochemistry is tolerant to extreme drought as well as different environmental conditions Gori et al., 2019;Puglielli et al., 2019a), RUBP (ribulose 1, 5bisphosphate) regeneration was partially impaired in July 2015, particularly during the hottest hours of the day, thus determining a P n decline in 2015 when temperature was higher (Gratani et al., 2018). It is noteworthy that many studies have shown that enzymes involved in RUBP regeneration are impaired by water stress at g s lower than 0.1 mol H 2 O m −2 s −1 (Sharkey and Seemann, 1989;Sánchez-Rodríguez et al., 1997;Grassi and Magnani, 2005). In our case, even if g s values were below this threshold in both years, we can hypothesize a significant metabolic impairment of P n only in July 2015, especially during the hottest hours of the day (Figure 4). However, in C. incanus leaves, RUBP regeneration is likely not limited by a decrease in J f and NADPH (nicotinamide adenine dinucleotide phosphate) synthesis (Flexas et al., 2004). In fact, the observed increase of J f /P n at 15:00 in July 2015 ( Figure 4C) indicates an excess of reducing power along the electron transport chain and a lower ability to use electrons for carbon assimilation (Sebastiani et al., 2019). In 2015, the excess of reducing power may have resulted in suboptimal F v /F m values (Figure 5A; Demmig and Björkman, 1987), which together with higher C i ( Figure 4B) and PSII (Figure 5B), may suggest higher utilization of alternative electron sinks (Grant et al., 2015;Puglielli et al., 2017b;Brunetti et al., 2018). At the same time, the daily down-regulation of PSII and the reduction of leaf chlorophyll content ( Figure 5C) observed in both years may have worked as a photoprotection mechanism to dissipate the high amount of light intercepted during the central hours of the day (Figure 1B; Baquedano and Castillo, 2007;Gori et al., 2019). When the capacity of plants to use the radiant energy is severely constrained, the excess of reducing power may be utilized for the biosynthesis of secondary metabolites, such as isoprenoids and polyphenols (Niinemets et al., 2002a,b;Harrison et al., 2013;Tattini et al., 2015). Indeed, secondary metabolites have been suggested to act as alternative electron sinks by consuming trioses phosphate, ATP (adenosine triphosphate) and NADPH, thus playing a role as an energy escape valve (Hernández and Van Breusegem, 2010;Morfopoulos et al., 2013). Among secondary metabolites, polyphenols perform different functions depending on their location in the leaf, acting as both radiation screeners, quenchers of ROS and key molecules involved in the acclimation processes to diurnal changes in UV-B radiation (Agati and Tattini, 2010;Agati et al., 2012;Jansen et al., 2012;Barnes et al., 2016;Tohge and Fernie, 2017;Neugart and Schreiner, 2018). Indeed, the comparable trends of flavonol index (Flav i ) observed in both years ( Figure 6A) could be related to similar diurnal changes in UV radiations (Stephanou and Manetas, 1997;Brossa et al., 2009). Furthermore, the reduction of chlorophyll content (Figure 5C), concomitantly with the increase in flavonols in the epidermis (Figure 6A) observed at 12:00 in both summers, may indicate that reversible changes in leaf biochemistry have helped C. incanus overcome stressful conditions on a daily basis (Mittler, 2006). Polyphenols of C. incanus leaves mainly belong to the classes of flavonols (quercetin and myricetin derivatives) and condensed tannins (Gori et al., 2016(Gori et al., , 2020. Both classes of molecules possess antioxidant function and may have played a role in the inhibition of radical formation under severe stress conditions ( Figure 6C; Hernández et al., 2004;Barbehenn et al., 2006;Sebastiani et al., 2019). In addition, flavonols may have contributed to both UV-B shielding and improve photo-protection at 12:00 (Burchard et al., 2000). As above mentioned, tannins may have helped increase leaf structure under the drier and hotter conditions of summer 2015 (Figure 2 and Supplementary Figure S1B; Ishida et al., 2008;Tharayil et al., 2011;Sumbele et al., 2012;Top et al., 2017). In support of this hypothesis, we found a statistically significant correlation between total polyphenols, mainly constituted by tannins, and air temperature ( Table 2 and Supplementary  Table S1A). The increase in tannin production observed in July 2015 could be also interpreted as a defense trait of C. incanus to protect the resources acquired during the growing season, thus contributing to retard leaf senescence and abscission and optimizing resource acquisition in nutrient-poor soils conditions (Mediavilla et al., 2001;Wright and Westoby, 2002;Poorter et al., 2009;Massad et al., 2014;Gratani et al., 2016;Puglielli et al., 2017a). Finally, the increment of the fiber-bound proportion of tannins in the leaf cell walls during leaf maturation may have contributed to the lignification process, thus protecting against pathogens and herbivores attack (Zucker, 1983;Bussotti et al., 1998). Another mechanism for counteracting leaf photooxidative damage under drought is the biosynthesis of BVOCs (Loreto and Schnitzler, 2010). The emission of monoterpenes is largely widespread among Mediterranean shrub vegetation exposed to concomitant stresses. In fact, these BOVCs have multiple protective functions (Rivoal et al., 2010;Fares et al., 2013;Fineschi et al., 2013;Llusià et al., 2016), ranging from antioxidant Llusià et al., 2006), protection against high temperatures (Owen et al., 2002;Haberstroh et al., 2018), defense against pathogens and herbivores to wound sealing after damage (Pichersky and Gershenzon, 2002). We observed a strong increment in monoterpene emissions in July 2015 compared to July 2014 (Figure 7), likely as a consequence of the higher air temperatures due to the 2015 heat wave ( Figure 1A and Supplementary Figure S1B). This is because, being highly volatile, the higher air temperature in 2015 increased the monoterpene partial pressure gradient between leaf and the atmosphere (Niinemets et al., 2010), resulting in their higher emissions than in 2014 (Supplementary Table S1A). However, no qualitative differences in the blend of monoterpenes emerged between the two years. Our results are consistent with those of Owen et al. (2001Owen et al. ( , 2002 who found α-pinene as the major monoterpene produced by this species. In addition, α-pinene and β-pinene, β-myrcene and α-phellandrene emissions were already detected in others Cistus spp. (Ormeño et al., 2007;Rivoal et al., 2010;Haberstroh et al., 2018). A high production of these compounds has been shown to help the leaves to withstand drought, which supports the assumption that they improve membrane stability, thus conferring thermal protection (Loreto et al., 1998;Copolovici et al., 2005). Hence, the diverse blend of terpenes detected in the emissions of this species may likely contributes to the drought tolerance of C. incanus under field conditions.

CONCLUSION
In conclusion, the results of our study indicate that C. incanus is a climate resilient species able to tolerate different environmental conditions, including threats caused by extreme weather events, as revealed from the new evidence on its capacity to adjust morpho-physiological and biochemical traits. In particular, the major investment in leaf construction observed in the drier and hotter conditions caused by the summer 2015 heat wave helped avoid leaf shrinkage associated to dehydration. Interestingly, the stimulation of the biosynthesis of secondary metabolites, in particular of monoterpenes and tannins, may have increased leaf thermo-tolerance and may have counteracted photooxidative damage, thus protecting C. incanus photosynthetic apparatus and allowing the maintenance of carbon assimilation despite the severe water and temperature stresses. Therefore, our results suggest that C. incanus might have greater potential to acclimate to climate change in its natural environment.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.