SeaSonal variation of leaf ecophySiological traitS of IrIs varIegata obServed in two conSecutive yearS in natural habitatS with contraSting light conditionS

The amount and pattern of individual phenotypic responses to seasonal changes in environmental conditions were determined in clones of Iris variegata growing in differing light habitats. For the purpose of the study, 97 clonal plants of the rhizomatous herb I. variegata that experienced different light conditions in their two native habitats were selected: one along the top and slope of sand dunes and one in woodland understories. Two fully expanded leaves that had developed during spring, summer and fall in two consecutive years were sampled from each of these clones. Six leaf traits affecting the photosynthetic rate of a plant − morphological (specific leaf area), anatomical (stomatal density) and physiological (total chlorophyll concentration, chlorophyll a/chlorophyll b ratio, carotenoid concentration, chlorophyll a/carotenoid ratio) exhibited significant plastic responses in the two different light habitats. To test whether these traits differ between exposed and shaded habitats as well as during different vegetation periods, we used the repeated model analysis of variance (ANOVA). Results of the repeated ANOVA revealed statistically significant effects of year, habitat and period of vegetation season. Patterns of changes during growing seasons were year-specific for almost all analyzed traits.

The evolution of adaptation depends not only on the genetic variation that leads to between-population differentiation for the traits in question, but also on their phenotypic plasticity that can lead to emergence of the same phenomenon (Schlichting and Pigliucci, 1998;Pigliucci, 2001;de Jong, 2005).Since morphological plasticity in leaf size may reflect underlying anatomical and physiological changes and because the magnitude and the direction of plasticity to the environment can differ among congeneric species (Sultan, 2003), or conspecific populations (Pemac and Tucić, 1998) with different environmental distributions, it seems important to examine which component of leaf phenotype is the most significant for plastic adjustment to ambient conditions that significantly change through vegetative season (Valladares, 2012;Ferreira, 2013;Ballare, 2014).
For plants, light is one of crucial environmental factors because it provides energy for photosynthesis and controls growth and development.Under suboptimal irradiance conditions, plastic changes occur at the level of individual leaves.They include different modifications in leaf structure and physiology, which lead to a specific leaf display commonly named "shade leaves" (Lichtenthaler, 1996;Ruberti, 2012;Ciolfi, 2013).Shade leaves have a larger leaf area, a reduced stomatal density, are thinner and possess a smaller sclerenchyma size and vascular bundle number, and a lower light saturation point of the overall photosynthesis compared to leaves exposed to full sunlight (Bjorkman, 1981;Lichtenthaler, 1996;Cao and Booth, 2001;Avramov et al., 2007).Since photosynthesis, and therefore plant growth, directly depends on the level of available solar energy, a shade-elicited increase in the size of light-capturing leaf area effectively compensates for the inevitable reduction in photosynthetic rate in light deprived environments (Wild and Wolf, 1980;Sultan and Bazzaz, 1993;Nicotra et al., 1997;Tucić et al., 1998;Oguchi et al., 2003).
The objective of this study was to quantify the relationship between the phenotypic characteristics of genotypes/clones of Iris variegata and their growth conditions through comparison of the amount and pattern of phenotypic responses to seasonal changes in environmental conditions.Three characters that directly affect the photosynthetic rate of an individual were analyzed: specific leaf area (SLA), which influences light energy interception, stomatal density, which controls gaseous exchange, and chlorophyll and carotenoids concentrations, which act as indicators of plant stress.In this paper, we analyzed plastic response to seasonal changes of environmental factors in two successive years.The focus was also on the determination of the suitability of Iris variegata for more detailed research in the course of the aforementioned forms of trait variability in response to environmental heterogeneity.
In order to fulfill those objectives we addressed the following: (i) how did the monitored ecophysiological traits chang during vegetational seasons; (ii) whether genotypes originating from different light habitats differ in patterns of changes during seasons, and (iii) whether the mode of change is the same or different during two successive years?

MaterialS and MethodS
Iris variegata L. (Hungarian, variegated or multicolored iris) inhabits areas of central and southeastern Europe.It can be found in Germany, southern Czech Republic, Austria, southern Slovakia, Hungary, western Ukraine, Bulgaria, southern Romania and northern Serbia.I. variegata inhabits grassy and open forest habitats.Since its natural habitats differ in a number of environmental parameters, most prominent is the divergence in quality and intensity of light, and therefore habitats can generally be divided into open habitats where plants are exposed to full sunlight, and shaded habitats, such as the ground floor of different forest stands with a reduced intensity and changed quality of the light.
Two groups of genotypes of Iris variegata inhabiting contrasting light conditions in the dune system at Deliblato Sands (44°48' N, 38°58' E, Serbia) were selected for this study.The first group of genotypes ("Open") occupied exposed areas on the dunes covered with annual and perennial herbs and low shrubs, while the second group of genotypes ("Shaded") was situated in the understories of different forest stands.These habitats are very different in terms of light intensity, soil characteristics and vegetation cover; nonetheless, natural growing clones were present in both.The estimated mean values for daily photosynthetic active radiation (PAR) and red:far-red (R:FR) light ratio in these habitats, under clear sky conditions, were 1173.1 μmol m -2 s -1 ; R:FR = 1.025 at the exposed dune site and 128.36 μmol m -2 s -1 ; R:FR = 0.78 under the canopy, respectively.Forty-seven clonal plants from the open and 50 from the shaded habitat were randomly selected, marked and leaves were taken for analysis in May, July and September of 2012 and in the same periods of 2013.
In order to assess the plant response to different light conditions, the following quantitative traits were measured: (1) stomatal density (SD; number of stomata/μm 2 ); (2) specific leaf area (SLA; leaf area/leaf mass, cm 2 /g); (3) total chlorophyll concentration (ChlT; mg/g); (4) the ratio of chlorophyll a to chlorophyll b (Chla/b); (5) carotenoid concentration (Car; mg/g) and ( 6) the ratio of chlorophyll a to carotenoids (Chla/Car).Specific leaf area was determined as the ratio of leaf surface area and dry leaf biomass (48 h at 60ºC) using the last fully developed leaf.Stomatal density was estimated by applying a microrelief method (Pazourek, 1970).For slide preparation, each leaf was painted across the middle of its surface with a 0.5cmwide band of clear nail polish over the entire width, in a direction perpendicular to the longer axis of its blade.After allowing it to dry, the dry polish copy of the leaf surface was peeled off with a piece of transparent adhesive tape and mounted on a microscope slide.Stomatal counts were made in 10 randomly chosen microscopic fields (X40 magnification) on each microscope slide per leaf.
Total chlorophyll concentration was estimated according to Hiscox and Israelstam (1979).Chlorophyll extracts were prepared using the upper half of the second fully developed leaf.An average of 10-15 mg of leaf tissue were placed in a vial containing 1 mL of DMSO (dimethyl sulfoxide) and incubated for 6 h at 65ºC.After incubation, chlorophyll and carotenoid absorptions were measured at wavelengths of 665, 649 nm and 480 nm, using a Multiscan Spectrum v1.2.
To estimate phenotypic variability of several leaf traits in Iris variegata in the environmental conditions of its natural habitats during spring, summer and fall, analysis of variance was applied.Since the experimental measurements of the same traits were repeated on each clonal individual over time (from the beginning until the end of the vegetation period), obtained data were analyzed using a repeated-measures analysis of variance (REPEATED option in SAS GLM procedure; SAS Institute, 2011).In our experimental model, year and habitat were defined as classification variables.In this approach, each repeated measure of a trait on the same individual is treated as a dependent variable.According to measures terminology, the measured individuals are referred to as subjects, the repeated observations on each individual as the within-subject or repeated factor, and the habitats from which the subjects originated and the experimental treatments to which they were exposed as the between-subject factor.In this particular experiment, the Year and the Habitat would be considered as the between-subject factors, while the Season, the Season x Year and the Season x Habitat interaction would be considered as within-subject factors.
The pattern of response of the within-subject factor (Season) was analyzed using the profile analysis (REPEATED/PROFILE option in the SAS GLM procedure; SAS Institute, 2011), which transforms the within-subject repeated observations to a set of differences (contrast) and then makes the univariate analyses on the contrast.The results of a profile analysis of the within-subject data show whether there are significant changes in trait values in the following time intervals: between spring and summer, summer and fall, and spring and fall.

reSultS
In the sampled clones growing in different habitats, all analyzed traits were affected by prevailing light conditions.Seasonal changes in the analyzed traits are displayed in the results of repeated model analysis (ANOVA) and PROFILE analysis (Tables 1 and  2).While genotypes from light habitats did not show different patterns of change across seasons, the mode of change during two successive years was clearly different (Figs. 1 and 2).
There was a general tendency for stomatal density to decrease from spring to summer and to increase from summer to fall in genotypes from both open and shaded habitats in the first observed year.A somewhat different pattern was detected in the following year, showing a continuous decrease in the same trait values through three seasons (Fig. 1).
By observing all seasons, the mean stomatal densities differed between two years (significant yeareffect in ANOVA, P=0.0003, Table 1).Similar results were registered when stomatal density were compared between clones from different light habitats (habitat effect, P<0.0001, Table 1).Genotypes from the open habitat had significantly higher values of stomatal densities compared to the shaded habitat (Fig. 1).
The results of ANOVA on the differences of consecutive seasons revealed a statistically significant season effect (P<0.0001) indicating that the mean stomatal density significantly changed over seasons.Since the season x year interaction was also significant (P<0.0001), it suggests that the ways in which stomatal density changed between the seasons were not the same in successive years (Table 1).
Significant season x habitat interaction indicates that the stomatal densities of genotypes originating from different habitats changed in different ways during the three seasons within the year (P<0.0001)(Table 1).In order to identify a particular time-interval in which the within-subject effects were different, profile analyses were computed for each of three contrasts: spring-summer; summer-fall and spring-fall.In all three time-intervals there was significant mean effect (P<0.0001), which clearly shows that there were significant differences in the mean stomatal density between all three seasons (Table 1).The mean SLAs differ between the two years (P<0.0001).Similar results were obtained when SLAs were compared between genotypes occurring in differently lit habitats (significant habitat effect, P<0.0001) (Table 1).Genotypes from the open habitat had significantly lower values of SLA in all three seasons and in both years (Fig. 1).The repeated ANOVA of the specific leaf area between seasons revealed a statistically significant season effect (P<0.0001), indicating significant change in the mean values of SLA through different parts of the vegetation periods.The mean response curves of the levels of the between-subject factors (year and habitat) revealed statistically significant season x habitat interaction (P=0.0030) as well as season x year interaction (P<0.0001).The first-mentioned interaction testifies that SLA values change in rather different ways in genotypes from different habitats through all seasons.The time x year interaction table 1. Repeated model analysis of variance (ANOVA) and PROFILE analysis for three Iris variegata leaf traits (stomatal density; specific leaf area; total chlorophyll concentration) in genotypes from contrasting light habitats observed across seasons (spring, summer, fall) in two consecutive years.shows significant differences in the pattern of SLA change through seasons between the two years (Table 1).Individual ANOVAs implemented on each of the three contrast variables (spring-summer; summer-fall and spring-fall) found significant mean effect for all three time-periods, which indicates that there were significant differences between all three parts of the vegetative seasons (Table 1, Fig. 1).

Source of variation
The total chlorophyll concentrations were higher in the shaded than in the open habitats across all seasons in both monitored years (significant habitat effect, P<0.0001) (Fig. 1, Table 1).Unlike the previously analyzed morphological and anatomical traits (SLA and SD), for total chlorophyll concentration there was not found a significant year-effect, which led to the conclusion that ChlT did not differ between adjacent years (Table 1).Univariate tests for within-subject effects revealed two significance sources of differences: season (P<0.0001) and season x habitat (P<0.0001)(Table 1).The season effect showed significant change in mean values of total chlorophyll concentrations throughout the seasons.Also, significant season x habitat interaction indicates the different nature of changes in mean values of ChlT for genotypes originating from different habitats through all seasons.Season x year interaction was not statistically significant (Table 1).
Individual ANOVAs computed on each of the three contrast variables (spring-summer; summerfall and spring-fall) showed a significant mean effect (P<0.0001), which indicates that the change rates of total chlorophyll concentration were significantly different across seasons (Table 1).
One of the important features of plant stress physiology beside total chlorophyll concentration is chlorophyll a/b ratio.When comparing overall mean values of chlorophyll a/b ratio, it was clearly shown that the mentioned concentrations were significantly higher in the shaded than in the open habitats (habitat effect, P<0.0001, Table 2) (Fig. 2).By observing all seasons, the mean chlorophyll a/b ratio differed between the two years (significant year-effect, P<0.0001) (Table 2).
Results of repeated ANOVA on the Chla/b differences of consecutive seasons revealed a statistically significant season-effect (P<0.0001),indicating a different pattern of change of mean values for the observed trait between seasons.Remaining sources of differences in within-subject data were also statistically significant.Significant season x year interaction (P<0.0001)(Table 2) testified that mean values of chlorophyll a/b ratio changed in different ways during the two years.Significant season x habitat interaction (P=0.0038)revealed significant differences in seasonal changes of the mean values of Chla/b in genotypes from different habitats (Table 2).The results of a profile analysis of the within-subject data showed that differences in the chlorophyll a/b ratios were not the same at all times during the year.There was a general tendency for the average chlorophyll a/b ratio to decrease from spring to summer and to maintain the reached level from the summer to fall timeperiod (Fig. 1).Individual ANOVAs implemented on each of the three contrast variables (spring-summer; summer-fall and spring-fall) found significant mean effect for two time-periods: the spring-summer and spring-fall, which explains the significant differences between spring and summer as well as between spring and fall, but not between summer and fall (Table 2).
The mean reaction norms for carotenoid concentration showed that mean values were higher in genotypes from the shaded in comparison with genotypes from the open habitat through all three seasons (Table 2).An identical pattern was registered in both analyzed years (Fig. 2).
The obtained data revealed a statistically significant habitat-effect (P=0.0015)indicating that the mean carotenoid concentrations significantly differed between clones occurring in different habitats while the mean response curve for the second betweensubject factor (year) showed no statistical significant outcome (Table 2).
The repeated ANOVA analysis on Car uncovered that all sources of differences in within-subject data were statistically significant.Season-effect (P<0.0001)revealed a significant change in the mean values of carotenoid concentrations through different parts of the vegetation periods (Table 2).Season x habitat interaction (P=0.0003) as well as season x year interaction (P=0.0446) were statistically significant.The first interaction indicates that mean values of the  observed trait change in different ways in genotypes from different habitats across all seasons.The second interaction uncovers a significantly different pattern of change in carotenoid concentration in the different years (Table 2).Individual ANOVAs implemented on each of the three contrast variables (spring-summer; summer-fall and spring-fall) found a significant mean effect for all three time-periods, which explains why there were significant differences in the carotenoid concentrations in Iris variegata plants between spring and summer, between spring and fall, and between summer and fall (Table 2).
The last analyzed data determining plant physiological response to alternative light regimes were the chlorophyll a/carotenoids ratios.Data showed that clones from shaded habitat had higher mean values of this trait in spring and summer in both years.In the last vegetation period, fall clones from the open habitat had higher mean values of chlorophyll a/carotenoids ratio compared to clones from the shaded habitat in both years.
Results from repeated ANOVA suggested significant year-effect (P<0.0003), which means that the observed trait mean values differed between years (Table 2).Habitat-effect did not demonstrate a significant impact on the observed subjects (Table 2).Results of repeated ANOVA of the differences of consecutive seasons revealed a statistically significant season-effect (P<0.0001),indicating different mean values for the observed trait during the three seasons.Other sources of differences in the within-subject data were also statistically significant (Table 2).Season x year interaction (P<0.0001)testified that mean values of the chlorophyll a/carotenoids ratio changed in different ways in the two years during the vegetative seasons.Season x habitat interaction (P=0.0083)revealed a significant change in the mean values in genotypes from different habitats (Table 2).The results of a profile analysis of the within-subject data have shown that differences in Chla/Car were statistically significant at all times during the year, implying significant differences between all three parts of the vegetative seasons.

diScuSSion
Mean daily radiation prevailing in Deliblato Sands inevitably decreases during the course of the year due to an increase in vegetation cover -both under the forested understory (shaded habitats) and on the open dune sites (open habitats).It would be reasonable to expect that the stomatal density of successive leaves during the vegetative season alters in parallel with changes in the amount of photosynthetically active radiation experienced during development (Tucić et al., 1999).In this study, it is quite clear that stomatal density in I. variegata changed in an expected manner in the second analyzed year.The increase in the number of stomata per leaf area that was found in the fall season of the first observed year despite a continuous decrease in light intensity could be ascribed to the impacts of other attributes in their environment.The humidity of the external and internal environments of a leaf can be important in influencing its stomatal density and size.The more arid the conditions, the greater number of stomata per leaf area and the steeper the insertion gradient in stomatal density between successive leaves (Fanourakis, 2013).Since annual changes in air temperature and dryness of the soil in the Deliblato Sands can significantly differ over the years, it is possible that the effects of these changes can override the effects of reduced light, resulting in a higher number of stomata per leaf area than expected.
As well as stomatal density, specific leaf area in I. variegata changed in accordance to the ecophysiological predictions for adaptive phenotypic responses to light changes (Vandenbussche, 2005;Franklin, 2008).Previously conducted research on the related species Iris pumila L. inhabiting the same territory but to some degree different habitats than Iris variegata, showed the advantages of a particular plant model system for various genetic, ecological and evolutionary studies (Tucić et al., 1998;Avramov et al., 2007;Tarasjev, 2012).It has been previously assessed that genotypes of the congeneric Iris pumila display a remarkable variability in response to variation in the light conditions they happen to encounter (Tucić et al., 1998;Avramov et al., 2007).Unlike this congeneric species that occurs mostly in sun-exposed sites and can be classified as a shade-avoiding species, I. variegata occurs equally in sun-exposed and understory sites.These previous data were used to compare the phenotypic responses of the two congeneric Iris species.Prevailing radiation levels imply that shadeinduced modifications in photosynthetically active surface area relative to plant biomass are indeed beneficial for light inception in Iris pumila (Avramov, 2007).In an earlier study, it was reported that under reduced light intensity Iris pumila plants exhibited increased leaf expansions, as well as extended leaf longevity relative to plants grown in exposed habitats (Tucić et al., 1998).The mean values of SLA of Iris variegata were higher in the shaded than in the open habitat, across all seasons.The decrease in SLA that was found in the summer period of the first observed year, in spite of a continuous decrease in light intensity, may be explained by the impacts of other environmental factors (Schmitt, 2003).
During their growth, plants experience different light conditions that may have an important effect on stomatal characteristics (Al Afas et al., 2007;Poorter et al., 2009;Loranger and Shipley, 2010).A positive relationship between irradiance level and stomatal density is also well documented (Cai, 2004).Results from this study have shown that the magnitude of the mean phenotypic response changed during the time course.Primarily, the mean values of stomatal densities were higher in the open compared to the shaded habitats in both analyzed years, which corresponds to already mentioned positive connection between radiance level and stomatal density.
Previous reports showed a quantitative loss in photosynthetic pigments, differential changes in chlorophyll a/b ratio, remarkable changes in carotenoid composition in general, and alterations in the composition of xanthophyll cycle pigments in particular, in response to stress factors (Vandenbussche, 2005;Biswal, 2011).
Under stressful conditions, concentrations of chlorophyll pigments decline in relation to carotenoid concentration.The concentration of chlorophyll b also declines in comparison to chlorophyll a con-centration, which influences their ratio substantially (Ricks and Williams, 1975;Franklin, 2005).Phenotypic variation in the physiological traits (Chla/b; ChlT) of Iris variegata in response to different light conditions showed in many ways a similar pattern of change to those of the mentioned reports on the differential change of photosynthetic pigments under the same stress.
Carotenoid concentrations were higher in genotypes originating from the shaded compared to those from the open habitat.The mode of change is not the same in the two years, but the pattern of change in carotenoid concentrations between genotypes from different habitats was the same during seasons in a single year (Nikolić, 2005;Kostić, 2012).
Data from previous studies clearly show that in a light-stress environment the chlorophyll a/carotenoids ratio will be changed in favor of carotenoids, which serve as photoprotective pigments (Sanchez-Gomez, 2006).This has to be considered with caution because other biochemical and physiological mechanisms of photoprotection are commonly found in plants in environments with high irradiance (Goncalves et al., 2001, Franco et al., 2007, Lichtenthaler et al., 2007).
In this study, results deviating from expectations occurred in the fall of both years, with particular differences noticeable in the second year.The opposite trend can be explained by the fact that plants are already adapted to the existing light conditions.(Capuzzo, 2012).However, change in photosynthetic pigment content is not exclusively an indication of a stressful light environment.For example, contents of chlorophyll a, chlorophyll b, total chlorophyll and carotenoid pigment could also be changed due to the decrease of soil moisture content (Yan, 2011).
Since the results obtained from this experiment largely confirmed theoretical expectations concerning adaptation to light changes (with the exception of the results for SD obtained in the fall of the first analyzed year; results for SLA that significantly differed between alternative habitats in the same seasons in the two different years, and results for Chla/Car for the fall of the second observed year), we found this species interesting for further study of the plasticity and variability in plasticity of Iris variegata genotypes in response to defined and controlled experimental light conditions, and to gain insight into potential differences in the selection regimes between natural habitats.conflict of interest disclosure: None of the authors of this manuscript has personal or financial relationship with people or organizations that could influence or bias the work and results presented in this paper.

acknowledgments:
This work was funded by the Ministry of Education, Science and Technological Development of Serbia (project OI 173025 "Evolution in heterogeneous environments: mechanisms of adaptation, biomonitoring and conservation of biodiversity").authors' contributions: Study conception and design were done by Stevan Avramov, Aleksej Tarasjev and Uroš Živković.Experimental work was performed by Uroš Živković.Statistical analysis and interpretation of the result as well as drafting of manuscript were done by all authors.

table 2 .
Repeated model analysis of variance (ANOVA) and PROFILE analysis for three Iris variegata leaf traits (chlorophyll a/b ratio; carotenoid concentration; Chla/Carotenoids ratio) in genotypes from contrasting light habitats observed across seasons (spring, summer, fall) in two consecutive years.