Reproductive output, seed anatomy and germination under water stress in the seeder Cistus ladanifer subjected to experimental drought

Environmental and Experimental Botany 123 (2016) 59–67

Contents lists available at ScienceDirect

Environmental and Experimental Botany journal homepage: www.elsevier.com/locate/envexpbot

Reproductive output, seed anatomy and germination under water stress in the seeder Cistus ladanifer subjected to experimental drought D. Chamorro, A. Parra, J.M. Moreno* Departamento de Ciencias Ambientales, Universidad de Castilla-La Mancha, Avenida Carlos III s/n, 45071 Toledo, Spain

A R T I C L E I N F O

A B S T R A C T

Article history: Received 15 September 2015 Received in revised form 2 November 2015 Accepted 4 November 2015 Available online 23 November 2015

Reproductive output and seed traits can be altered by water availability during seed formation and maturation, which could affect the population recovery after fire of seeders (i.e., species regenerating from seeds). This is important for species in fire-prone, dry areas that are projected to encounter reduced total precipitation and longer annual drought with climate change. Here we determine the sensitivity of several reproductive processes to drought in Cistus ladanifer, a plant widely distributed in the shrublands of the western part of the Mediterranean Basin. Three levels of annual drought were simulated in a shrubland by means of a rainout shelter and an irrigation system in 6
6 m plots. Fruits and seeds from drought-exposed mother plants were collected, and reproductive output, seed size and anatomy studied. Seeds non-exposed/exposed to fire cues (heat plus smoke) were germinated at five levels of water stress (Cs = 0.0 to 0.50 MPa). Hydrotime modeling was applied to germination under water stress. Plant growth was sensitive to drought, but reproductive output, seed size, dormancy and viability were not. Drought significantly affected seed anatomy, increasing the palisade layer at the micropyle. Drought in the maternal plants, in interaction with seed exposure to fire cues, significantly reduced final germination. Water stress during germination decreased final germination, independent of maternal plant drought, and interacted with fire cues to decrease germination when exposed. Hydrotime modeling confirmed that fire cues made seeds highly sensitive to water stress (Cb (50) = 0.25 MPa). Postgermination viability was reduced in seeds from drought-treated maternal plants that were exposed to fire cues and germinated under water stress. Reproductive output showed low plasticity in response to drought. However, the effects of drought in the mother plant affected seed anatomy and germination in interaction with fire cues. The conclusion is that exposing C. ladanifer maternal plants to drought arguably increases seed sensitivity to water limitations during germination after fire. ã 2015 Elsevier B.V. All rights reserved.

Keywords: Maternal environment effects Seed coat Micropyle Physical dormancy Fire cues Hydrotime model Climate change

1. Introduction Seed production and seed characteristics, including size, dormancy and germination responses to various environmental cues, are key plant reproduction that have evolved driven by habitat characteristics, and make a major contribution to plant fitness (Venable and Brown, 1988; Leishman et al., 2000; Donohue et al., 2005). Water availability in the mother plant at the time of seed formation (i.e., maternal effects) (Roach and Wulff, 1987) can affect both these components (Fenner, 1991, 1992; Gutterman, 2000; Donohue, 2009). It is well known that, in general, water stress negatively affects plant growth, ovule formation and maturation, and thus fruit and seed production effort. Moreover,

* Corresponding author. E-mail addresses: [email protected] (D. Chamorro), [email protected] (A. Parra), [email protected] (J.M. Moreno). http://dx.doi.org/10.1016/j.envexpbot.2015.11.002 0098-8472/ ã 2015 Elsevier B.V. All rights reserved.

not all seeds produced under water stress are viable, so the overall reduction in regenerative potential from seed can be large (Albert et al., 2001). Water stress in maternal plants can also affect important seed characteristics including provisioning and size, usually negatively (Fenner, 1992). Alteration of seed formation and traits by drought could affect germination rates and ultimately have consequences on plant regeneration, especially under future climate conditions (Walck et al., 2011). Therefore, understanding the effects of drought on plant reproduction is of utmost importance in water-limited areas around the world that are projected to encounter further limitations due to climate change. Not all plants are equally sensitive to water stress. Some species are able to partially modify allocation patterns between growth and reproduction, relatively increasing the latter to favor greater reproductive outputs in spite of water limitations (Sultan, 2003). In addition, some species can reduce the number of fruits produced and/or the number of seeds per fruit (and therefore total seeds), but maintain seed size or even increase it under water stress; these

60

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67

include both annual (Sultan, 1996; Gorecki et al., 2012) and perennial species (Stromberg and Patten, 1990). This plasticity in seed size is important: seed size has often been related to germinability and seedling survival once germination has occurred (Sultan, 1996; Moles and Westoby, 2004). A reduction in the number of seeds but an increase in their size could still ensure a relatively large number of established seedlings. Water availability of the maternal plant can also affect seed dormancy, depending on its type. While seeds with physiological dormancy (i.e., a physiological inhibiting mechanism in the embryo prevents germination) often exhibit a negative relationship with plant water availability (i.e., fewer seeds are dormant as water stress increases) (Meyer and Allen, 1999; Tielbörger and Valleriani, 2005), seeds with physical dormancy (i.e., a waterimpermeable coat prevents imbibition and thus germination) exhibit the opposite trend (i.e., more seeds are dormant at higher water stress) (Clua et al., 2006; D’hondt et al., 2010). Physical seed dormancy can also be related to seed size, whereby larger seeds have lower dormancy than smaller seeds (Delgado et al., 2001). In fire-prone environments, where a long-lived seed bank is advantageous to produce a large flush of germination after fire (Moreno et al., 2011), reduced dormancy would imply prompt germination once the seed is moistened, which would be detrimental to this strategy. Missing the opportunity to regenerate after fire to outcompete other plants would threaten the dominance and/or persistence of plants in such environments. Determining the impact of water stress on dormancy is therefore of utmost importance in plants that respond to fire by means of a long-lasting soil seed bank. The testa of the seed originates from the mother plant, and starts differentiating from the ovule teguments right after pollination (Beeckman et al., 2000). Arguably, variations in the environment experienced by the mother plant until seed maturation could affect seed anatomy. There is evidence that changes in maternal temperature can affect total seed mass and testa mass, but not endosperm or embryo mass (Lacey et al., 1997). Moreover, reduced seed germination can be related to structural changes in the testa, such as greater thickness (Wada et al., 2011). Release from physical dormancy is dependent on the breaking of the testa to allow seed imbibition (Gama-Arachchige et al., 2013). In fireprone environments, exposure of seeds to the heat of fire cracks the testa in various places (Aronne and Mazzoleni, 1989), making imbibition possible. However, little is known about the possible effects of variations in the anatomical characteristics and dormancy break factors in relation to the conditions during seed formation. The germination process is also highly sensitive to water stress (Baskin and Baskin, 2014), but sensitivity varies within species and between species (Fady, 1992; Thomas et al., 2010). In studies on gradients, variations within species have been related to the environment of the mother plant (Fady, 1992; Tilki and Dirik, 2007); however, the cause of such relationships is not known. They could be due to trait selection driven by the local environment along the gradient (i.e., be genetically fixed), and/or variations in the maternal environment during seed development and maturation along the gradient (i.e., phenotypic plasticity) (Roach and Wulff, 1987). In seasonal habitats with large interannual variability in precipitation, a plastic germination response to water stress due to genetic or phenotypic effects could be advantageous to cope with variations in water availability from seed formation through seed germination and initial establishment. However, maternal effects have been studied mainly in annual plants. These being short-lived, the sensitivity of the mother plant to the environment could have a large adaptive value (Herman and Sultan, 2011). Little is known, however, about maternal effects on the sensitivity of the

seeds of long-lived perennial species, including shrubs, to water stress in germination. Mediterranean environments are fire-prone and have limited water, and precipitation variability is common, particularly where precipitation is low (Lionello et al., 2006). With climate change, reduced precipitation is projected for this region in spring, at the time of flowering, and fall, lengthening the summer drought (Christensen et al., 2013). Changes in precipitation patterns consistent with these projections have already been observed (De Luis et al., 2010). Understanding the factors that control plant regeneration under limited water availability is very important for species that regenerate from seeds (i.e., seeders) after fire, since this is the only mechanism they have for recovering their population after the blaze. The objective of this work was to determine the response in terms of reproductive output (fruit and seed biomass), seed number, size, anatomy, dormancy, viability, and germination to water stress in the germination environment in seeds of Cistus ladanifer non-exposed/exposed to fire cues (heat and smoke) at a Mediterranean-climate continental site in Central Spain. To assess the sensitivity of the various reproductive processes in mother plants to different precipitation regimes, a mature shrubland was subjected to a fully replicated rainfall manipulation experiment during the growing season, exposing the vegetation, including mature Cistus ladanifer plants, to various levels of drought. 2. Material and methods 2.1. Study area and rainfall manipulation experiment The rainfall manipulation experiment was carried out in a mature shrubland dominated by several woody perennial shrubs (mainly Cistus ladanifer and Erica species), located at the Quintos de Mora range station (39 250 N, 4 040 W). The climate is typically Mediterranean, with a mean annual temperature of 14.9  C and a mean annual precipitation of 622 mm (Los Cortijos meteorological station, 39190 N, 4 040 W) (AEMET, Spain). The experiment was implemented using a set of automatic rainout shelters and an irrigation system that allowed the desired level of rainfall in each treatment to be controlled at two-week intervals. Four rainfall treatments were implemented: (1) EC: environmental control (i.e., natural rainfall); (2) HC: historical control (simulation of long-term rainfall patterns, including two months of drought [July and August]); (3) MD: moderate drought (total rainfall reduction of 25% from HC [i.e., to percentile 8 of the historical record], with five months of drought from May to September); and (4) SD: severe drought (reduction of 45% from HC [to percentile 2 of the longterm historical record], with seven months of drought from April to October). Treatments were implemented as follows: at the beginning of each two-week period shelters were open to natural rainfall until the established levels for the period were reached. After these were reached, the shelters were deployed at every rain event to preclude rainfall. If rainfall was insufficient to match the set target at the end of the period, irrigation was applied up to that level. These treatments were applied on a total of 16 plots (6
6 m), four plots per treatment, following a randomized complete block design with four blocks. Rainfall manipulations started in late March, 2009, and continued until late September of that year, when the shelters and tubing were removed for a few days to allow experimental burning of the plots. The system was reinstalled shortly after burning and the treatments continued for several years. Here, we report data from the first growing season after treatment implementation on the pre-fire mature shrubland. C. ladanifer flowers in May, and seed dispersal starts in mid-July. The precipitation since the beginning of the hydrological year (October 1, 2008) until the start of rainfall manipulations (March

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67

61

2.3. Testa thickness The testa is composed of an outer waxy layer and an inner lignified palisade layer (Corral et al., 1989). Thickness of the palisade layer was measured in both the chalaza and micropyle regions from micrographs obtained by scanning electron microscopy (SEM). Ten seeds per rainfall treatment were longitudinally cut and mounted on SEM specimen stubs using double-sided carbon tape. Seeds were scanned with an FEI QUANTA 200 Scanning Electron Microscope at an acceleration voltage of 20.0 kV and low vacuum mode, and micrographs were taken, from which the various dimensions were measured. 2.4. Germination experiment

Fig. 1. Accumulated rainfall (mm) from the beginning of the hydrological year until fruit harvesting in plants of Cistus ladanifer under various rainfall treatments: environmental control (EC, natural rainfall), historical control (HC, natural rainfall and/or irrigation to mimic long-term average, drought during 2 months starting in July), moderate drought (MD, drought during 5 months, starting in May) and severe drought (SD, drought during 7 months, starting in April). Light dash-line indicates the long-term average rainfall recorded in the study area.

15, 2009) was 313 mm, which was lower than the long-term average for that time period (396 mm) (Fig. 1). Rainfall recorded until fruit-harvesting (July 24, 2009) was as follows: (i) historical control (HC)—544 mm; (ii) moderate drought (MD)—442 mm; (iii) severe drought (SD)—383 mm; and (iv) environmental control (EC)—367 mm (Fig. 1). Note that rainfall at the EC treatment plot was much lower than the long-term historical average (HC), due to the spring of 2009 being very dry; EC yielded even lower values that the most severe treatment (SD) (Fig. 1). For further details and a full description of the experimental set-up see Parra et al. (2012). 2.2. Plant growth, fruit collection and fructification features Three individuals of C. ladanifer were selected and tagged in the central area of each plot, a total of 48 individuals (12 individuals per treatment). Each individual was measured for its maximum length and maximum and minimum crown diameter; based on these results we estimated plant biovolume by approximating plant shape to an ellipsoid (V = 3/4p r1 r2 r3). Growth was measured in three apical shoots, which were tagged on April 1 and collected on July 15. Then, the shoots were dried (70  C for 48 h) and weighed, and the relative growth rate (RGR; mg g1 d1) was calculated following Harper (1977): RGR ¼ ðlnB1  lnB0 Þðt1  t0 Þ1 where B1 is the dry biomass of the labeled apical shoot at the end of the experiment (t1; July 15), and B0 is the dry biomass of the labeled apical shoot at the beginning of the experiment (t0; April 1). The number of fruits was counted for each individual, from each of which six fruits were randomly selected and harvested in late July, just before the beginning of seed dispersal. In the laboratory, fruit and seed mass were determined, and the number of seeds per fruit counted. Fruits are globular, woody capsules, 10–15 mm in size, with 8–10 valves, containing 329–1390 seeds. Fruit size was estimated by weighing the six fruits per plant. Fruit mass per plant was calculated from fruit number per plant and mean fruit size per plant. Number of seeds per plant was estimated from fruit number per plant and mean number of seeds per plant. Mean single seed size was estimated by weighing two groups of 25 seeds from each fruit (totaling 12 measures for each individual), dividing each by the number of seeds counted. Fruit tissue mass and seed mass per fruit were estimated from seed counts and mean seed size.

Seed samples were mixed to obtain one sample per treatment. Prior to germination, half of the chosen seeds were subjected to a heat shock of 100  C during 10 min in a forced-draft oven, immediately followed by exposure to smoke to simulate the effects of fire. Smoke was generated by burning a mixture of the dominant species at the site, and was conducted into a closed chamber where the seeds were maintained for 20 min at ambient temperature. After this, seeds were incubated in six solutions of different levels of water potential: (Cs): 0 MPa, 0.1 MPa, 0.2 MPa, 0.3 MPa, 0.4 MPa and0.5 MPa. Deionized water was used as control, and different solutions with the desired water potential were made using polyethylene glycol in deionized water (PEG 6000). Concentrations of PEG were determined using a standard formula (Michel and Kaufmann, 1973). Seeds were set to germinate in Petri dishes (5.5 cm in diameter) over two sheets of filter paper (Whatman no. 1). We used six petri dishes, with 25 seeds each, for each rainfall, fire-cues and water potential treatment. All Petri dishes were sealed with parafilm to prevent them from desiccating. The incubation lasted 60 days, with temperature set at 20  C and a photoperiod of 12 h. Dishes were placed at random in a temperature-controlled chamber (Model G-21, Ibercex), and the position of each dish was changed every three days to avoid position effects. During the first 30 days, germination was recorded daily. From day 30 onward, germination was counted every three days. Radicle emergence from the seed was used as the criterion for germination. Germinated seeds were removed. At the end of the experiment, the viability of the non-germinated seeds was checked by means of the tetrazolium chloride test (Baskin and Baskin, 2014). Testing was done after the seeds were cut into two halves and incubated in a 1% solution of 2,3,5-triphenyl tetrazolium chloride for 48 h in dark conditions. Final germination was corrected by viability. 2.5. Data analysis The effect of rainfall treatments (EC, HC, MD and SD) on plant growth (RGR, mg g1 day 1), number of fruits per plant, fruit tissue mass per plant (mg), number of seeds per plant, seed mass per plant (mg), and mean seed mass (mg) was analyzed by ANOVA, with rainfall treatments as a fixed factor. We first tested block effect as a random factor and plant biovolume as covariate, but the results were not significant and were therefore discarded from the analysis. The mean value of each variable per plot was used, the plot being the sampling unit (n = 4). The effect of rainfall treatments on the thickness of the palisade layer in the chalazal and micropyle regions was tested using a oneway ANOVA, with rainfall treatments as the factor. In this case, n = 10, corresponding to 10 seeds analyzed per treatment. Germinability was characterized by the final germination at the end of the experiment (FG) and by germination speed, using the number of days until the first germination occurred (T0) and the

62

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67

number of days required to reach 50% of the total germination (T50). Viability was calculated as the difference between germinated seeds plus viable seeds after the tetrazolium test and the initial number. Two viability measures were used: (1) after germination under control conditions (i.e., deionized water) ( V0); (2) after germination under the full range of water potential used (Vc). Differences in V0 between seeds exposed and nonexposed to fire cues would account for mortality of seeds caused by these manipulations. Differences between V0 and Vc would account for the effects of water stress during germination. Seed dormancy (i.e., the proportion of viable seeds that did not imbibe without further treatment) was calculated by subtracting the seeds that germinated without fire-cues treatment from all viable seeds. Generalized Linear Models with maximum likelihood were used for testing differences in seed dormancy, V0 non-exposed, V0 exposed, FG, T0, T50, and Vc among rainfall, fire-cues and water potential treatments. Seed dormancy, V0 non-exposed and V0 exposed were analyzed for differences between rainfall treatments. The fire-cues effect was tested for V0. Rainfall, fire cues and water potential were defined as fixed factors for other variables. Based on error structure we assumed a binomial error distribution for seed dormancy, V0 non-exposed, V0 exposed, FG and Vc, and a Poisson error distribution for T0 and T50. Based on Akaike’s Information Criteria, different link functions were fitted, whereby logit link functions were the best fit for seed dormancy, V0 nonexposed, V0 exposed, FG and Vc, identity functions were the best function for T0 and T50. Post-hoc pair-wise comparisons among treatments were performed using Bonferroni corrections. 2.6. Hydrotime modeling We modeled the germination response at the various water potentials using a hydrotime model (Bradford and Still, 2004), following the equation:

uH ¼

ðC  Cb ðgÞÞ tg

where C (MPa) is the ambient water potential, Cb(g) (MPa) is the base water potential for a germination fraction g, tg (h) is the time required for the fraction g to germinate, and uH (MPa h) is the hydrotime required for the proportion g to germinate. For a given population, uH is a constant while Cb(g) follows a normal distribution that is characterized by s cb(g) (Bradford and Still,

2004). The application of the model requires high germination values for the control conditions (i.e., 0 MPa); modeling was therefore only possible for fire-cues-treated seeds. To calculate u H, Cb(50) and s cb(50), the data on the time-course of germination for each of the treatments was modeled using the function Cb  (uH/tg), based on the following equation (Bradford and Still, 2004): Probit ¼

ðCb  ðu H =tg Þ  Cb ð50ÞÞ : s Cb ð50Þ

The uH values obtained were those that best fitted each rainfall treatment following an iterative procedure. Once such a value was obtained, we proceeded to calculate the remaining parameters (Cb(50) and s cb(50)). 3. Results The plants selected for this study were not significantly different in biovolume across treatments (Table 1). Plant relative growth rate (RGR) was significantly affected by rainfall, with the highest RGR being obtained in the moister treatment (HC) and the lowest in the severe drought treatment (SD). Note the severity of the natural drought during the year of experimentation (Fig. 1), with EC having low RGR, similarly to the MD treatment (Table 1). Rainfall treatments did not significantly affect reproductive output (fruit mass or seed mass per plant). Fruit and seed size was equally unaffected by rainfall treatments; seed mass per fruit accounted for 34–37% of the total fruit mass (Table 1). Similarly, rainfall treatments did not affect the number of fruits or seeds per plant (Table 1). On average, 91% of the viable seeds were dormant – that is, 9% were impermeable (i.e., non-dormant) – and this proportion was not significantly affected by rainfall treatments (Table 1). Seed viability (V0) was 94% and 92% in seeds non-exposed or exposed to fire cues, respectively, and was not significantly affected by rainfall treatments (Table 1). Exposure to fire cues did not affect the proportion of viable seeds (x2 = 0.636; P = 0.425). Palisade thickness in the chalazal region did not significantly differ among seeds from different rainfall treatments, but did vary in the micropyle region (Table 1). Palisade thickness was greater in the treatments that had less rainfall (EC, SD) in comparison with higher rainfall MD or HC treatments (Table 1; Fig. 2).

Table 1 Mean  SE plant biovolume (n = 4), relative growth rate (RGR) (n = 4), reproductive output (n = 4), thickness of the palisade layer of the testa (n = 10), seed dormancy (n = 6) (i.e., non-permeable fraction), seed viability (n = 6) based on germination under deionized water for non-exposed and exposed seeds to fire cues (FC) for each rainfall treatment (HC: historical control, MD: moderate drought, SD: severe drought, EC: environmental control). P values of statistical tests (see Section 2) are shown in bold when statistically significant; in this case, different letters show significant differences among rainfall treatments based on post-hoc Tukey test (P < 0.05).

3

Plant biovolume (m ) RGR (mg g1 day1) Fruit mass per plant Fruit size (mg) Tissue mass per fruit (mg) Seed mass per fruit (mg) Fruits per plant (No.) Seeds per plant (No.) Seeds per fruit (No.) Seed size (mg) Seed dormancy (%) Seed viability (%) Non-exposed to FC Exposed to FC Palisade layer thickness Chalaza region (mm) Micropyle region (mm)

HC

MD

SD

EC

P

0.4  0.1 30  1a 14.4  3.8 485  23 319  15 167  9 31  9 24533  7228 815  49 0.200  0.004 90  3

0.6  0.2 29  1ab 26.5  6.1 519  33 337  23 183  12 52  12 46333  12112 889  38 0.208  0.006 95  2

0.4  0.1 25  1b 28.2  11.7 540  18 338  13 202  8 51  17 46045  16032 941  59 0.217  0.008 90  3

0.7  0.3 27  2ab 34.9  9.6 518  22 323  19 194  8 66  17 65064  18297 976  86 0.201  0.004 89  3

0.436 0.035 0.153 0.563 0.839 0.096 0.493 0.108 0.520 0.202 0.227

97  1 97  3

91  2 93  2

94  2 91  4

94  2 88  4

0.410 0.199

19.9  0.5 27.6  2.1a

20.1  0.5 33.3  1.2ab

19.7  0.7 34.1  1.8b

18.3  0.8 37.6  1.9b

0.201 0.001

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67

63

Fig. 2. Scanning electron micrographs of the longitudinal section of the mycropilar region of C. ladanifer seeds from different rainfall treatments. (A) Environmental control, (B) Severe drought, (C) Moderate drought and (D) Historical control. Abbreviations: wl, waxy layer; pl, palisade layer; en, endosperm.

Rainfall treatments significantly affected neither final germination (FG) nor time to initiate germination (T0), nor time to reach 50% of the final germination (T50) (Table 2). However, viability after incubation under the various water potentials (Vc) was sensitive to changes in rainfall (Table 2). Seeds from the drier EC treatment had the lowest mean viability (61%), and showed significant difference from the other three treatments (SD, MD and HC), which had similar Vc (68%, 69% and 72%, respectively) (Fig. 3). On the other hand, rainfall treatments interacted with exposure to fire cues to

affect FG (Table 2). FG of seeds subjected to fire cues showed significant differences among rainfall treatments; by contrast, seeds not exposed to fire cues did not show differences between rainfall treatments. Fire-cues-exposed seeds from HC and MD showed the highest mean FG (52 and 53%), while seeds from SD and EC treatments showed the lowest (43 and 42%, respectively). No other significant interaction was detected between rainfall treatments and any of the other factors (Table 2).

Table 2 Results of GLM for rainfall (R), fire cues (FC) and water potential (C) effects on final germination (FG), time to germination (T0), time to 50% germination (T50) and postgermination viability (Vc) (n = 6). Significant P values are marked in bold (P < 0.05).

Rainfall Water potential Fire cues R
C R
Fc C
Fc C
R
Fc

FG

T0

T50

x

2

2.742 208.513 74.225 11.050 9.431 35.749 11.143

df

P

x

3 5 1 15 3 5 15

0.433 <0.001 <0.001 0.749 0.024 <0.001 0.742

4.789 103.813 23.295 7.044 6.824 2.603 14.356

2

VC

df

P

x

df

P

x2

df

P

3 3 1 9 3 3 9

0.188 <0.001 <0.001 0.633 0.078 0.457 0.110

1.542 112.708 6.109 11.203 2.184 1.968 11.933

3 3 1 9 3 3 9

0.673 <0.001 0.013 0.262 0.535 0.579 0.217

14.445 255.742 25.600 19.075 1.028 9.534 20.225

3 5 1 15 3 5 15

0.002 <0.001 <0.001 0.210 0.795 0.090 0.163

2

64

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67

Fig. 3. Final germination (FG), time to germination (T0), time to 50% germination (T50) and post germination viability (Vc) (n = 6) for seeds non-exposed to fire cues (heat + smoke) (empty bars and circles, left panels) or exposed to them (solid bars and circles, right panels) and germinated with water at different water potentials (0 to 0.5 MPa), from plants that had been grown in plots subjected to different rainfall manipulations: HC (historical control), MD (moderate drought), SD (severe drought), EC (environmental control). Note that EC rainfall was below SD (Fig. 1). Letters show significant differences among rainfall manipulations within fire cues non-exposed or exposed seeds based on pairwise comparisons with Bonferroni correction (P < 0.05) after GLM analysis (see Table 2).

The water potential (Cs) of the germinating solution significantly affected FG, T0, T50 and Vc (Table 2; Fig. 3). In the case of FG, there was also a significant interaction between Cs and exposure to fire cues. The net result was that FG decreased as Cs decreased, but this effect was mainly in the seeds that were exposed to fire cues, and was less prevalent in those not treated. Germination speed (both T0 and T50) responded only to the main effects of the water potential of the germination solution, with values increasing

(delayed initiation of germination, longer time to germinate 50% of the seeds) at the lower Cs values. Hence, T0 increased from five days under control conditions to 14 days in 0.2 MPa, and 20 days in 0.3 MPa. T50 increased from 8.5 days to 11.5 in 0.1 MPa. T0 and T50 were not calculated at the lowest treatments (0.4 and 0.5 MPa), owing to many dishes not having any germination. Vc significantly decreased with Cs, without further interactions with seed exposure to fire cues.

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67

Exposing the seeds to fire cues significantly increased FG and decreased T0, T50 and Vc (Table 2; Fig. 3). Non-exposed seeds germinated 5% on average, while exposed seeds germinated 47% on average. T0 and T50 decreased, on average, by four days with exposure to fire cues. Vc of seeds subjected to fire cues was significantly lower (60%) than that of non-exposed seeds (74%). Exposing the seeds to fire cues had some significant interactions, as commented upon earlier (Table 2). No significant triple interactions among variables were found (Table 2). The hydrotime model fitted well with the germination patterns of fire-cues-exposed seeds, with r2 between 0.83 and 0.92. u H ranged between 32 and 50 MPa (Table 3) for the various rainfall treatments. Similarly, Cb(50) values ranged between 0.21 and 0.30 MPa, but the values of s cb(50) were large, indicating overlapping among treatments. The mean Cb(50) of fire-cuesexposed seeds was 0.25 MPa. 4. Discussion In our study, reproductive output, either in terms of fruit and seed mass or seed size, as well as seed output per plant, was not affected by the extreme drought imposed by our experimentation, including the most extreme natural drought experienced during the year of the experiment. This finding contrasts with the fact that vegetative growth was significantly affected by drought (under both natural and experimental conditions). Both fruit and seed output in Mediterranean woody plants are commonly responsive to rainfall variability from year to year (Keeley, 1977; Martin and Puech, 2001). Manipulative experiments like the one implemented here often find that number and quality of fruits and seeds are greater under less stressful conditions (Breen and Richards, 2008), with few species being capable of maintaining fruit or seed quantity and quality under water stress (Sultan, 1996; Gorecki et al., 2012). Our results, however, coincide with those reported by Del Cacho et al. (2013), who found a lack of significant effects of experimental drought in another Cistaceae, albeit a chamaephyte, in seeds per fruit and seed size. This indicates that, at least during the drought occurring during one season, C. ladanifer has a lowplasticity reproductive output. This species adjusts itself to variations in water availability by reducing growth when water is limited (Parra et al., 2012; Ramírez et al., 2012). Notwithstanding, Ramírez et al. (2012) also found that C assimilation was not affected by drought, which could explain the maintained reproductive output found here. Cistus seeds are well known for their physical dormancy (Baskin and Baskin, 2014), but, as in other families with this kind of dormancy, species in this genus, including C. ladanifer, have a fraction of seeds that are innately permeable. Being dormant is important to ensure a large soil seed bank that, after fire, will allow the species to recolonize the burned area (Céspedes et al., 2012). The fraction of dormant seeds (or its non-dormant and permeable counterparts) in this species varies quite significantly (23–96%) between and within populations (Pérez-García, 1997; Chamorro et al., 2013). Variations in dormancy between populations are independent of the prevailing climate conditions of the seed’s origin (Chamorro et al., 2013), but differences in habitat conditions Table 3 Hydrotime (uH), base water potential for 50% fraction (Cb(50)), standard deviation (s cb (50)) and r2 values for estimates of hydrotime parameters in different rainfall treatments.

uH (MPa h) Cb (50) (MPa) s cb (50) (MPa) r2

HC

MD

SD

EC

36 0.27 0.13 0.91

50 0.30 0.14 0.92

32 0.21 0.13 0.92

42 0.23 0.17 0.83

65

among populations during the year of seed harvesting have not been considered (i.e., maternal effects were not tested). In our study, dormancy was not related to the maternal environment. The severity of the treatments and of the growing season experienced during the experimentation allows us to argue that changes towards drier climate conditions are unlikely to affect dormancy in this species. Since our treatments covered at best the meanaverage rainfall, it cannot be said whether variations in seed dormancy as a result of years with above-average rainfall could affect this seed trait. In seeds with physical dormancy, the width of the chalazal region has been related to variations in the innately permeable fraction, with permeable seeds having a larger pore (Meisert et al., 1999). However, pore width and changes in the anatomy of the surrounding tissue of the testa have not been related to one another. Here, the thickness of the palisade layer at the margins of the pore did not differ as a result of the drought treatments imposed on the mother plants. This is consistent with the finding that the proportion of permeable seeds is not affected by treatments. Pore width, however, was not measured due to the destructive procedure of SEM of the seed. Notwithstanding, maternal effects had an impact on the anatomy of the seed in the micropyle region. Although the micropyle is involved in the water gap of a number of species with physical dormancy (GamaArachchige et al., 2013), such a role has not been found for the Cistaceae, where imbibition in permeable intact seeds occurs only through the chalazal region. The micropyle is the location through which the radicle emerges. It is not known whether variations in the anatomy of the seed in this region, such as the thickening found here, play a role in germination in this species. We found that germination in fire-cues-treated seeds interacted with the maternal growing conditions and the water stress of the incubation environment, so that seeds germinating at lower water potential in the more severe drought plots (i.e., thicker micropyle region) had reduced germination. Since exposure to heat in these seeds ruptures the seed coat at various locations (Aronne and Mazzoleni, 1989), prompting imbibition at these locations, lack of imbibition cannot be the cause of such reduced germination. This allows us to argue that other factors must be involved in causing the reduced germination, but whether the thickening of the micropyle or some other characteristic is instrumental in this is something that needs to be researched. In the study, the effects of the drought treatments imposed on the mother plants on final germination resulted from an interaction with fire cues. While permeable seeds had similar final germination independent of the level of drought endured by the mother plant, seeds exposed to fire cues had reduced final germination for the treatments with the highest levels of drought. Viability post germination under water stress was very sensitive to rainfall treatments, fire cues and water stress of the germinating solution. This means that many seeds died during the germination experiment, and that more died when the seeds were exposed to fire cues and originated from mother plants that had been droughttreated. Given the variability of conditions that seeds could encounter in an area with highly variable precipitation, this indicates that the potential for the seeds of a drought-exposed mother plant to germinate diminishes after fire. As water stress in the germinating environment increased, final germination was much reduced. This effect manifested itself in the seeds exposed to fire cues. Post-fire environments are subject to greater fluctuations in temperature and moisture at the soil surface (Auld and Bradstock, 1996) due to lack of vegetative cover and the black color of the upper soil surface for many months after fire. All of this means that the potential for protracted maternal effects on successful germination and establishment after fire could be significant, but this needs further quantification.

66

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67

This negative maternal effect in seeds exposed to fire cues contrasts with that observed in seeds that germinate readily (i.e., were non-dormant). Reduced sensitivity to decreasing water potential in germination of untreated seeds of C. ladanifer has been documented previously (Pérez-Fernández et al., 2006). In our case, the effect of water stress in the germinating environment was not manifest, so that germination continued much without change, irrespective of the water potential of the solution used. This was therefore independent of the drought endured by the mother plant. This means that germination of seeds that are permeable proceeds with great resistance to the germinating conditions and independently of the conditions during seed formation. This is to the opposite of what one would expect for a species that is ‘adapted’ to fire (Keeley et al., 2011) in a highly variable hydric environment. Our findings allow us to argue that the non-dormant fraction will be favored to some extent in comparison with the dormant one (i.e., exposed to fire cues) in conditions of drought experienced by the mother plant. This type of response would facilitate this species’ colonization of new areas without fires. Indeed, C. ladanifer is well known for its ability to colonize abandoned fields, and to develop dense shrublands in the absence of fire (Torres et al., 2012). Our results document that the persistence of this species in more drought-prone conditions and in the absence of fire could be facilitated via the non-dormant fraction. Although the hydrotime model could not be calculated for the non-dormant fraction due to low germination, calculation of Cb(50) for the fire-cues-exposed seeds yielded rather high values, much higher than have been described for other species in fireprone environments (Thomas et al., 2010). This reinforces the idea that C. ladanifer is highly sensitive to water stress during germination once the physical barrier of the seed has been broken by heat, and smoke has had its effect. The above effects are likely to be reinforced by the fact that germination proceeded more rapidly in fire-cues-exposed seeds (less time to start germination, less time to reach 50% of total germination) than in non-exposed (i.e., non-dormant) seeds. For a plant to persist in a fire-driven ecosystem, quick initiation of the germination in response to a first rain could be detrimental, as the process could suffer if the next rainfall does not occur within a short time. This is further reinforced due to the particular stressful environment of burned areas. The process of imbibition has three phases, and is irreversible beyond a certain point (Finch-Savage and Leubner-Metzger, 2006). It is not known when this point is reached, but it is probable that this happens more rapidly the quicker the initiation. A faster process would mean the seeds could be more exposed to subsequent drought hazard than non-dormant seeds. Again, this strategy is not fully consistent with adaptation to a fire and drought-prone environment, since the greatest benefits are reserved for the non-dormant seeds whose initiation of germination is delayed. A caveat in this discussion is that our experiment did not test whether seeds subjected to desiccation after the germination process was initiated would die, since conditions during incubation were stable. Moreover, the germination temperature was fixed, and even though C. ladanifer has a broad germination temperature niche (Chamorro et al., 2013), it is not known whether variations in temperature could affect the patterns found here. In any case, the germination biology of this species and its adaptation to fire is something that merits further investigation. In summary, in our work the reproductive output of C. ladanifer showed low plasticity in response to the drought experienced by the mother plants. However, effects of drought in the mother plant affected seed anatomy and germination when seeds were exposed to fire cues (heat and smoke). Severe drought in the mother plants significantly increased the palisade layer in the micropyle, and

negatively affected final germination of seeds exposed to fire cues. Additionally, viability post germination under water stress was very sensitive to the drought treatments in the mother plants, fire cues and water stress of the germinating solution. These results indicate that drought in the mother plant could increase the sensitivity of C. ladanifer to further drought hazard during germination after fire. Acknowledgments This work was funded by Spanish Ministry of Research (CGL2006-06974, project SECCIA) and the EC 7th FP (project FUME, GA 243888). We thank the Quintos de Mora personnel for their continuous support; V. Marugal, N. López and A. Velasco for their technical assistance; and three anonymous reviewers for useful comments and suggestions. We are grateful to K. J. Bradford for generously providing the software to calculate the parameters of the hydrotime model. References Albert, M.J., Escudero, A., Iriondo, J.M., 2001. Female reproductive success of narrow endemic Erodium paularense in contrasting microhabitats. Ecology 82, 1734– 1747. Aronne, G., Mazzoleni, S., 1989. The effects of heat exposure on seeds of Cistus incanus L. and Cistus monspeliensis L. G. Botanico Ital. 123, 283–289. Auld, T.D., Bradstock, R.A., 1996. Soil temperatures after the passage of a fire: do they influence the germination of buried seeds? Austral Ecol. 21, 106–109. Baskin, C.C., Baskin, J.M., 2014. Seeds Ecology, Biogeography, and, Evolution of Dormancy and Germination. Academic Press, San Diego. Beeckman, T., De Rycke, R., Viane, R., Inzé, D., 2000. Histological study of seed coat development in Arabidopsis thaliana. J. Plant Res. 113, 139–148. Bradford, K.J., Still, D.W., 2004. Applications of hydrotime analysis in seed testing. Seed Technol. 26, 75–85. Breen, A.N., Richards, J.H., 2008. Irrigation and fertilization effects on seed number, size, germination and seedling growth: implications for desert shrub establishment. Oecologia 157, 13–19. Del Cacho, M., Peñuelas, J., Lloret, F., 2013. Reproductive output in Mediterranean shrubs under climate change experimentally induced by drought and warming. Perspect. Plant Ecol. Evol. Syst. 15, 319–327. Céspedes, B., Torres, I., Luna, B., Pérez, B., Moreno, J.M., 2012. Soil seed bank, fire season, and temporal patterns of germination in a seeder-dominated Mediterranean shrubland. Plant Ecol. 213, 383–393. Chamorro, D., Luna, B., Moreno, J.M., 2013. Germination response to various temperature regimes of four Mediterranean seeder shrubs across a range of altitudes. Plant Ecol. 214, 1431–1441. Christensen, J.H., Krishna Kumar, K., Aldrian, E., et al., 2013. Climate phenomena and their relevance for future regional climate change. In: Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M. (Eds.), The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Clua, A., Fernández, G., Ferro, L., Dietrich, M., 2006. Drought stress conditions during seed development of narrowleaf birdsfoot trefoil (Lotus glaber) influences seed production and subsequent dormancy and germination. Lotus 36, 58–63. Corral, R., Pérez-García, F., Pita, J.M., 1989. Seed morphology and histology in four species of Cistus L. (Cistaceae). Phytomorphology 39, 75–80. D’hondt, B., Brys, R., Hoffmann, M., 2010. The incidence, field performance and heritability of non-dormant seeds in white clover (Trifolium repens L.). Seed Sci. Res. 20, 169–177. Delgado, J.A., Serrano, J.M., López, F., Acosta, F.J., 2001. Heat shock, mass-dependent germination, and seed yield as related components of fitness in Cistus ladanifer. Environ. Exp. Bot. 46, 11–20. Donohue, K., 2009. Completing the cycle: maternal effects as the missing link in plant life histories. Philos. Trans. R. Soc. Lond. 364, 1059–1074. Donohue, K., Dorn, L., Griffith, C., Kim, E., Aguilera, A., Polisetty, C.R., Schmitt, J., 2005. The evolutionary ecology of seed germination of Arabidopsis thaliana: variable natural selection on germination timing. Evolution 59, 758–770. Fady, B., 1992. Effect of osmotic stress on germination and radicle growth in five provenances of Abies cephalonica Loud. Acta Oecol. 13, 67–79. Fenner, M., 1991. The effects of the parent environment on seed germinability. Seed Sci. Res. 1, 75–84. Fenner, M., 1992. Environmental influences on seed size and composition. Hortic. Rev. 13, 183–213. Finch-Savage, W.E., Leubner-Metzger, G., 2006. Seed dormancy and the control of germination. New Phytol. 171, 501–523. Gama-Arachchige, N.S., Baskin, J.M., Geneve, R.L., Baskin, C.C., 2013. Identification and characterization of ten new water gaps in seeds and fruits with physical dormancy and classification of water-gap complexes. Ann. Bot. 112, 69–84.

D. Chamorro et al. / Environmental and Experimental Botany 123 (2016) 59–67 Gorecki, M.J., Long, R.L., Flematti, G.R., Stevens, J.C., 2012. Parental environment changes the dormancy state and karrikinolide response of Brassica tournefortii seeds. Ann. Bot. 109, 1369–1378. Gutterman, Y., 2000. Maternal effects on seeds during development. In: Fenner, T. (Ed.), Seeds: The Ecology of Regeneration in Plant Communities. Cabi Publishing, Oxford, pp. 59–84. Herman, J.J., Sultan, S.E., 2011. Adaptive transgenerational plasticity in plants: case studies, mechanisms, and implications for natural populations. Front. Plant Sci. 2 (102) . Harper, J.L., 1977. Population Biology of Plants. Academic Press Inc, San Diego. Keeley, J.E., 1977. Seed production, seed populations in soil, and seedling production after fire for two congeneric pairs of sprouting and nonsprouting chaparal shrubs. Ecology 58, 820–829. Keeley, J.E., Pausas, J.G., Rundel, P.W., Bond, W.J., Bradstock, R.A., 2011. Fire as an evolutionary pressure shaping plant traits. Trends Plant Sci. 16, 406–411. Lacey, E., Smith, S., Case, A., 1997. Parental effects on seed mass: seed coat but not embryo/endorsperm effects. Am. J. Bot. 84, 1617. Leishman, M.R., Wright, I.J., Moles, A.T., Westoby, M., 2000. The evolutionary ecology of seed size. In: Fenner, F. (Ed.), Seeds: The Ecology of Regeneration in Plant Communities. Cabi Pusblishing, Oxford, pp. 31–58. Lionello, P., Boscoso, R., Malonotte-Rizzoli, P., 2006. Mediterranean Climate Variability. Elsevier, Amsterdan. De Luis, M., Brunetti, M., González-Hidalgo, J.C., Longares, L.A., Martín-Vide, J., 2010. Changes in seasonal precipitation in the Iberian Peninsula during 1946–2005. Global Planet. Change 74, 27–33. Martin, A., Puech, S., 2001. Interannual and interpopulation variation in Helichrysum stoechas (Asteraceae), a species of disturbed habitats in the Mediterranean region. Plant Species Biol. 16, 29–37. Meisert, A., Schulz, D., Lehmann, H., 1999. Structural features underlying hardseededness in Geraniaceae. Plant Biol. 1, 311–314. Meyer, S.E., Allen, P.S., 1999. Ecological genetics of seed germination regulation in Bromus tectorum L. II. Reaction norms in response to a water stress gradient imposed during seed maturation. Oecologia 120, 35–43. Michel, B.E., Kaufmann, M.R., 1973. The osmotic potential of polyethylene glycol 6000. Plant Physiol. 51, 914–916. Moles, A.T., Westoby, M., 2004. Seedling survival and seed size: a synthesis of the literature. J. Ecol. 92, 372–383. Moreno, J.M., Zuazua, E., Pérez, B., Luna, B., Velasco, A., Resco, V., 2011. Rainfall patterns after fire differentially affect the recruitment of three Mediterranean shrubs. Biogeoscience 8, 3721–3732.

67

Parra, A., Ramírez, D.A., Resco, V., Velasco, A., Moreno, J.M., 2012. Modifying rainfall patterns in a Mediterranean shrubland: system design, plant responses, and experimental burning. Int. J. Biometeorol. 56, 1033–1043. Pérez-Fernández, M., Calvo-Magro, E., Ferrer-Castán, D., 2006. Simulation of germination of pioneer species along an experimental drought gradient. J. Environ. Biol. 27, 669–695. Pérez-García, F., 1997. Germination of Cistus ladanifer seeds in relation to parent material. Plant Ecol. 133, 57–62. Ramírez, D.A., Parra, A., Resco, V., Moreno, J.M., 2012. Differences in morphophysiological leaf traits reflect the response of growth to drought in a seeder but not in a resprouter Mediterranean species. Funct. Plant Biol. 39, 332–341. Roach, D.A., Wulff, R., 1987. Maternal effects in plants. Annu. Rev. Ecol. Syst. 18, 209–235. Stromberg, J., Patten, D., 1990. Variation in seed size of a Southwestern riparian tree, Arizona Walnut (Juglans major). Am. Midl. Nat. 124, 269–277. Sultan, S.E., 1996. Phenotypic plasticity for offspring traits in Polygonum persicaria. Ecology 77, 1791–1807. Sultan, S.E., 2003. Phenotypic plasticity in plants: a case study in ecological development. Evol. Dev. 5, 25–33. Thomas, P.B., Morris, E.C., Auld, T.D., Haigh, A.M., 2010. The interaction of temperature, water availability and fire cues regulates seed germination in a fire-prone landscape. Oecologia 162, 293–302. Tielbörger, K., Valleriani, A., 2005. Can seeds predict their future? Germination strategies of density-regulated desert annuals. Oikos 111, 235–244. Tilki, F., Dirik, H., 2007. Seed germination of three provenances of Pinus brutia (Ten.) as influenced by stratification, temperature and water stress. J. Environ. Biol. 28, 133–136. Torres, I., Céspedes, B., Pérez, B., Moreno, J.M., 2012. Spatial relationships between the standing vegetation and the soil seed bank in a fire-prone encroached dehesa in Central Spain. Plant Ecol. 214, 195–206. Venable, L., Brown, J.S., 1988. The selective interactions of dispersal dormancy and seed size. Am. Nat. 131, 360–384. Wada, S., Kennedy, J.A., Reed, B.M., 2011. Seed-coat anatomy and proanthocyanidins contribute to the dormancy of Rubus seed. Sci. Hortic. 130, 762–768. Walck, J.L., Hidayati, S., Dixon, K.W., Thompson, K., Poschlod, P., 2011. Climate and plant regeneration from seed. Glob. Change Biol. 17, 2145–2161.