Asymmetric responses to simulated global warming by populations of Colobanthus quitensis along a latitudinal gradient

Target species and study sites

Commonly known as the Antarctic pearlwort, Colobanthus quitensis (Kunth) Bartl. (Caryophyllaceae), is a small-sized cushion-like perennial herb with an extremely wide distribution range spanning from Mexico (17°N) to the Antarctic continent (68°S); sporadic populations can be found in different islands of Maritime Antarctica, as well as along the coast of the Antarctic Peninsula (Smith, 2003). C. quitensis individuals in Antarctic ecosystems are mostly distributed near seashores which are frequently associated with D. antarctica and mosses. Clonal reproduction is the more common means of propagation in the Antarctic populations of C. quitensis, but it is also capable of self-compatible sexual reproduction (Convey, 1996).

Individuals of C. quitensis were collected from three locations along a simple latitudinal transect: South America (SA), close to the city of Punta Arenas (53.1°S–70.9°W); South Shetland Islands (SI), close to the Polish Antarctic Station in Admiralty Bay (62.1°S–58.3°W), and the Antarctic Peninsula (AP), in Lagotellerie Island (67.5°S–67.2°W) (Fig. 1). Sixty plants were dug out from each site during the 2014/2015 growing season with enough soil around the roots (ca. 250 g) and kept well-watered in a plastic box under natural conditions of light and temperature until their transportation to the growth chambers at the Universidad de Talca, Chile (35.2°S). All plants were collected under permission of the Chilean Antarctic Institute (INACH; authorization number: 1060/2014).

Latitudinal trait variation in pigments and microstructures

Variations in anatomical (foliar microstructure) and physiological (photoprotective pigment) traits were assessed in individuals from all locations along the latitudinal gradient. Specifically, in 25 individuals from each location (one leaf per plant) we measured and compared five microstructure foliar traits: cuticle, mesophyll, palisade parenchyma and spongy mesophyll width (mm), and leaf transversal area (mm²). In the laboratory, leaves were sectioned in 2–3 mm in length and dehydrated in a graded ethanol series. Thin fragments of 75–100 nm sections were prepared on an ultra-microtome. Sections were stained with toluidine blue and viewed with a light microscope to analyse morphological attributes of the leaf. Thin sections were cut with a diamond knife and were stained on grids with uranyl acetate and lead citrate, and then examined in transmission electron microscope.

In addition, in a subset of 15 individuals from each location (one leaf per plant), we estimated the xanthophyll-cycle pigment content (violaxanthin, antheraxanthin and zeaxanthin) and the de-epoxidation state of the xanthophyll cycle pigments (DEPS). Leaves were repeatedly extracted (three times) with ice-cold 85% (v/v) acetone and 100% acetone using sonication for 45 min at 4 °C. Pigments were separated on a Dupont non-end capped Zorbax ODS-5 µmol column at 30 °C at a flow rate of 1 ml min-1. The solvents consisted of (A) acetonitrile/methanol (85:15, v/v) and (B) methanol/ethyl acetate (68:32, v/v). The HPLC gradient used was: 0–14 min 100% A, 14–16 min decreasing to 0% A, 16–28 min 0% A, 28–30 min increasing to 100% A, and 30–38 min 100% A. Detection was carried out at 445 nm. The de-epoxidation state (DEPS) of the pigments involved in the xanthophyll cycle was quantified as Z + (0.5 × A)∕(V + A + Z) (for more details see Molina-Montenegro et al., 2012b).

Latitudinal responses to simulated global warming experiment

To evaluate whether different latitudinal origins of C. quitensis differ in the magnitude of their responses to global warming, we measured the ecophysiological performance of plants from each origin (South America, Shetland Islands and Antarctic Peninsula) under current and future temperature conditions (warming) predicted by climate change models (IPCC, 2014). Based on the available predictions, a 4 °C increase in temperature was applied as warming treatment for all populations. Thus, “current” temperatures were 10 °C, 5 °C and 3 °C for SA, SI and AP, and “warming” temperatures were 14 °C, 9 °C and 7 °C for SA, SI and AP, respectively.

We established six experimental abiotic conditions using six automatic air-cooling growth chambers (model: LTJ300LY; Tianyi Cool, China) to simulate current and future environmental conditions during summer months of each latitudinal origin. The current conditions for each location were: South America (mean temperature: 10 °C, photoperiod: 18/6 h light/dark), Shetland Islands (mean temperature: 5 °C, photoperiod: 19/5 h light/dark) and Antarctic Peninsula (mean temperature: 3 °C, photoperiod: 21/3 h light/dark). In addition, future conditions for each location were: South America (mean temperature: 14 °C, photoperiod: 18/6 h light/dark), Shetland Islands (mean temperature: 9 °C, photoperiod: 19/5 h light/dark) and Antarctic Peninsula (mean temperature: 7 °C, photoperiod: 21/3 h light/dark). During all experimental time every chamber was maintained in the temperature and photoperiod indicated above, with a constant intensity of the photosynthetic active radiation (PAR) of 275 µmol m⁻²s⁻¹. But PAR was lowered to 20 µmol m⁻²s⁻¹ during seven months in order to mimic the natural variation in the solar radiation that affect at high latitude populations. A total of 30 plants per each latitudinal origin (SA, SI and AP) were randomly assigned to the current or warming conditions. Hence, in each growth chamber a total of 15 plants were maintained during three years, and once per year in January (middle of the growing season), photosynthetic rate was recorded on every individual in both current and warming conditions. The photosynthetic rate was recorded using an infra-red gas analyzer (IRGA, model Ciras-2; PP-System, Amesbury, MA, USA) under the same temperature of each growth chamber. In addition, at the end of the third year two fitness-related traits were measured as response variables, total biomass accumulation and reproductive effort. Total plant growth was measured as the average individual total biomass increase (final biomass—initial biomass, g) during the three years of the experiment. At the beginning of the experiments, we selected similar size plants (from each origin). These plants were randomly assigned to current and warming treatments. There were no differences in the initial weight of plants between current and warming treatments for any latitude (one-way ANOVA, F_2,87 = 29, 4; p < 0.001). At the end of the experiments, whole plants were harvested (root plus shoot), and individually weighed using a digital balance. Reproductive effort was measured by counting the number of flowers produced by each plant during the last year of the experiment. Since there were no differences in initial plant biomass (fresh weight) between current and future thermal regimes, all individuals were weighed before the start of the experiment. The magnitude of the responses of each latitudinal origin of C. quitensis to experimental warming was estimated as the average difference between future and current thermal conditions (i.e., delta = future—current values).

Statistical analysis

We used one-way ANOVAs to assess the significance of the observed differences between origins for all microstructural traits, pigment content and the DEPS state of foliar samples of C. quitensis. The magnitude of the responses of C. quitensis to experimental warming on the delta values (future—current conditions) for each trait (photosynthetic rate, total biomass accumulation, and reproductive effort) in different populations along of the latitudinal gradient was tested using one-way ANOVAs. In all analyses, significant differences between treatments were estimated using Tukey tests (HSD) as post-hoc comparisons. Due to the number of multiple comparisons the sequential Bonferroni correction was applied to all a posteriori contrasts. All statistical tests were made using R v.3.1.3 (R Core Team, 2015).

Latitudinal trait variation in pigments and microstructures

As predicted, pigment contents (xanthophyll-cycle pigments), de-epoxidation state of pigments (DEPS) and micro-structural foliar traits showed a clinal pattern of variation (Table 1 and Table S1). With the exception of antheraxanthin that did not have a clear relationship with latitude, concentration of the other pigments and the DEPS significantly increased with latitude (Table 1). Leaves became significantly thicker and more cylindrical towards higher latitudes (Table 1 and Table S1). This pattern of clinal variation is consistent with the increase in the photo-inhibitory conditions (lower temperatures and moisture and longer daylight exposition) at higher latitudes.

Latitudinal responses to simulated global warming experiment

The photosynthetic rates of C. quitensis exposed to experimental warming were significantly different among latitudinal origins (Figs. 2A–2C; Tables S2, S3). The greatest response to experimental warming was found in individuals from the southernmost population (AP), while individuals from South America showed the lowest response to experimental warming (Fig. 2; Tables S2, S3).

At the end of third year, experimental warming also had a greater positive and significant effect on total biomass accumulation and reproductive effort of Antarctic Peninsula C. quitensis individuals than those from South America or even from the Shetland Islands (Fig. 3; Tables S2, S3), suggesting that the observed differences in the photosynthetic response are also effectively translated into important fitness-related traits.

Latitudinal trait variation in pigments and microstructures

The prediction of a progressive latitudinal change in both physiological and morphological traits was supported by seven of the eight traits measured in C. quitensis. The direction of the trait variation was consistent with the functional physiological responses expected to cope with increasingly stressful abiotic conditions such as cold temperatures, low water availability and longer daylight periods found in the Antarctic continent (Alberdi et al., 2002; Convey, 2006). Despite the clear clinal variation in functional traits shown by C. quitensis that cope with photoinhibitory conditions, the biological performance (biomass and flower production) was lower in the southern margin. This result seems to agree with the core of the “spatial ecology theory” (Gaston, 2009; Sexton, Hangartner & Hoffman, 2014), which states that peripheral populations reduce their biological performance as a consequence of the reductions in habitat suitability and increased isolation (see, Sagarin & Gaines, 2002; Vaupel & Matthies, 2012).

Several studies have documented that high levels of the xanthophyll cycle pool could be considered as a pivotal mechanisms to avoid photoinhibition under high radiation and/or low temperature (Demmig-Adams & Adams, 1996; García-Plazaola, Matsubara & Osmond, 2007; Molina-Montenegro et al., 2012b). In addition, high levels of the xanthophyll cycle pool have been correlated with thermal dissipation of excess energy, being a dynamic and reversible mechanism in plants that inhabit cold environments (Bravo et al., 2007; Bascuñan-Godoy et al., 2010; Molina-Montenegro et al., 2012b; Míguez et al., 2015). In our study, those individuals of C. quitensis living in the south edge of distribution showed a significantly higher xanthophyll cycle pool, suggesting that in this population the capacity to avoid photoinhibition and to maintain its growth rate could be explained -at least in part- by the presence of this mechanism. The de-epoxidation state was also higher in the southern populations, where environmental stress is greater. The high de-epoxidation state found in all C. quitensis populations may be related to the capacity to dissipate excess energy, avoiding photoinhibition; this was more evident in the southern population where a combination of environmental stress conditions (colder, dryer, and photo-inhibitory) is found.

Morphological adaptations have been documented as one of the main mechanisms in plants to cope with environments (Mooney et al., 1991; and references therein). Foliar modifications in C. quitensis exposed to different abiotic conditions have been previously reported (Bascuñan-Godoy et al., 2010; Cavieres et al., 2016). We found that toward higher latitudes, C. quitensis individuals showed lower leaf transversal area and greater mesophyll thickness. It has been reported that increases in leaf thickness under cold acclimation may be beneficial for leaf survival under frost-induced cell dehydration, reducing the mechanical stress experienced during thawing (Stefanowska et al., 1999). The thicker palisade parenchyma shown by the southern population is likely to be associated with the higher photosynthesis values, since this foliar adaptation can influence CO₂ diffusion and assimilation (Vieira & Mantovani, 1995). These variations in foliar micro-structures seem to be a constitutive adaptation, since trait differences between populations were partially maintained in a common garden under controlled laboratory condition, suggesting that these functional adaptations could have a genetic basis.

Latitudinal responses to global warming

Altitudinal and latitudinal shifts in distribution are among the earlier impacts of global warming observed in plants (e.g., Menzel et al., 2006; Kelly & Goulden, 2008; Kopp & Cleland, 2014). Nonetheless, there are still few studies focusing on the potential adaptive responses of plant species to climate change (but see Nicotra et al., 2010; Merilä & Hendry, 2014). If increased temperature associated with climate change exceeds the thermal tolerance of a plant species, the species may either: (1) adjust to the new environmental conditions through acclimation (through phenotypic plasticity) or evolve (ecotype differentiation due to natural selection) (Bellard et al., 2012); or (2) track suitable environmental conditions in space (changing their geographical distribution) and/or time (adjusting their phenology and physiology). In contrast, if increased temperatures do not exceed the thermal tolerance, plastic responses rather than ecotype differentiation should be expected. Overall, our results indicate that C. quitensis will mainly respond to future warming through plastic adjustment and probably through range shifts after local scale population expansions in response to global change. However, biotic interactions (mostly competitive) will be also relevant to predict how this species will deal with future climate change. In a recent study, Torres-Díaz et al. (2016) showed that responses of C. quitensis to future climate change will be not only modulated by abiotic factors such as temperature and soil moisture, but also by biotic interactions with fungal endophytes. Moreover, plant–plant interactions between C. quitensis and the native and invasive plants could be even more critical to forecast the impacts of climate change on Antarctic vegetation. As shown by Molina-Montenegro et al. (2016) the alien P. annua shows high adaptive plasticity and greater competitive ability than the native D. antarctica. Therefore future studies should address how climate change will affect the interactions between the native species and also between the native C. quitensis and the alien P. annua.

Although environmental stress releasing due to future warming may promote anatomical changes that would increase traits such as light absorption, photosynthesis, growth and reproductive effort, warming may also compromise plant freezing tolerance and photoprotection. Warming has been reported to reduce the ability of perennial plants to resist freezing events through cold de-acclimation (reviewed in Pagter & Arora, 2013). Similar effects have been found in Andean plants exposed to experimental warming (e.g., Sierra-Almeida & Cavieres, 2010). Thus, future studies should address these potentially negative effects of stress releasing on Antarctic vascular plants.

Antarctic vascular plants are distributed along one of the most stressful gradients of the world, being affected by extreme environmental pressures that often limit their individual performance and population dynamics (Peck, Convey & Barnes, 2006). Nevertheless, recent warming during the last decades has already increased their frequency and distribution (Cannone et al., 2016). In a recent review, Valladares et al. (2014) showed how plastic responses are highly variable between populations of the same species, and how the spatial distribution of these responses are key to cope with rapid anthropogenic climate change. Understanding the adaptive response landscape of both plastic and selective processes would greatly improve the forecasting of the local and regional effects of global warming on plant species, particularly those with widespread ranges of distribution such Antarctic vascular plants.

It is important to mention that future climate change will not only raise temperature but will also increase liquid precipitation. For instance, Turner et al. (2009) documented increases in precipitation in the Antarctic Peninsula. It is important to acknowledge that in our experimental design plants we only evaluated the response of plants to warming, maintaining irrigation similar for both treatments. As shown in Torres-Díaz et al. (2016) warming tend have lower effects than watering or combined warming plus watering on C. quitensis fitness. Therefore, our experimental design might have underestimated the effects of future climate change on C. quitensis net photosynthesis, growth and reproductive effort.

Final remarks

It is important to note that our study is not free of limitations. Natural environmental conditions are almost impossible to reproduce in growth chambers, thus our estimations of plant performance could differ from those we would find in the field. Natural conditions were not fully mimicked in our experimental setup and future experimental research under field conditions could provide more accurate predictions of the effects of warming on C. quitensis performance. Field experiments using open top chambers (OTCs) would have been the more realistic way to study the future responses of C. quitensis to environmental warming. Nonetheless, as shown by Torres-Díaz et al. (2016) the responses of C. quitensis to future climate change can be complex, modulated by abiotic and biotic interactions, and additive and even antagonistic effects may affect the results of laboratory experiments.