Plant Reddening in Silene Germana Is Due to Anthocyanin Accumulation in Response to Visible Light E

Plant Biology ISSN 1435-8603

RESEARCH PAPER Whole-plant reddening in Silene germana is due to anthocyanin accumulation in response to visible light E. Narbona1, J. Jaca2, J. C. del Valle1, F. Valladares3 & M. L. Buide1 1 Department of Molecular Biology and Biochemical Engineering, Pablo de Olavide University, Sevilla, Spain 2 Mediterranean Institute for Advanced Studies (CSIC-UIB) Mallorca, Balearic Islands, Spain 3 Centro de Ciencias Medioambientales, Instituto de Recursos Naturales, CSIC, Madrid, Spain

Keywords ABSTRACT Anthocyanins; flavonoids; florivory; flower abortion; intra-individual variation; secondary • The phenology of anthocyanin accumulation in leaves has been widely studied in metabolites. perennial plants; several hypotheses have been proposed to explain their adaptive significance. Here, we explored the photoprotection hypothesis in Silene germana,a Correspondence Mediterranean annual plant with late-spring/summer flowering. E. Narbona, Department of Molecular Biology • We analysed the temporal patterns of anthocyanin accumulation in photosynthetic and Biochemical Engineering, Pablo de calyces, leaves and stems and throughout the reproductive season, and their relation- Olavide University, Carretera de Utrera km 1, ship with flower abortion, florivory and plant mortality due to drought. In addition, 41013 Sevilla, Spain. the flavonoid production and the photoinhibitory response were measured in a shad- E-mail: [email protected] ing experiment. • The whole plant becomes red at the end of the flowering and remains red until fruiting Editor and senescence. Calyces were redder on the side with more sun exposition. Aborted J. Arroyo flowers showed redder calyces than those of fruiting flowers. No effect of plant redness on florivory or plant mortality was found. The shading experiment showed a positive Received: 15 May 2018; Accepted: 11 July relationship between anthocyanin accumulation and intensity of solar radiation, but 2018 plants growing in absence of UV showed similar redness than full sunlight plants. Plants growing in natural shade lack anthocyanins but produced the same amount of doi:10.1111/plb.12875 non-anthocyanin flavonoids. Anthocyanic and non-anthocyanic plants showed similar photochemical efficiency (Fv/Fm) after sun exposition, but in early morning, the former showed lower Fv/Fm values. Plants growing in full sunlight produced more fruits than those of natural shade plants. • Whole-plant reddening during fruiting and senescence appears to be a property of S. germana. Our results suggest that anthocyanin accumulation depends on sunlight intensity, but non-anthocyanin flavonoids are produced constitutively.

excess light, cold temperature, drought, salinity, pathogens or INTRODUCTION herbivores (Lee & Gould 2002). However, their adaptive signif- Anthocyanins are the largest group of flavonoids, and the most icance remains unclear (Archetti et al. 2009; Landi et al. 2015). diverse colour-producing pigments in plants (Andersen & Traditionally, five chief hypotheses have been postulated to Jordheim 2006). In angiosperms, anthocyanins may be accu- ascribe a function for the accumulation of anthocyanins in mulated in all organs, from roots to flowers, and are usually leaves and stems: (i) relieving photoinhibition by acting as sun- sequestered in vacuoles of the epidermis or mesophyll (Whel- screen to shield plant tissues against excess light (photoinhibi- dale 1916; Hatier & Gould 2009). In reproductive organs, tion hypothesis); (ii) absorbing UV-B radiation and reducing anthocyanins may produce orange, blue, pink, purple and its injurious effects such as photoinhibition and DNA damage black colorations, whereas in vegetative organs mostly red or (UV-B protection hypothesis); (iii) mitigating oxidative dam- purple are observed (Andersen & Jordheim 2006; Manetas age in leaves subjected to different abiotic stressors, by scaveng- 2006). Anthocyanin production should represent some adap- ing reactive oxygen species (ROS) (antioxidant protection tive benefits, or functions, for the plants that compensate their hypothesis); (iv) conferring resistance to water stress, i.e., metabolic and photosynthetic cost (Chalker-Scott 2002; Arche- drought, salinity, flooding and freezing via osmotic adjustment tti et al. 2009). of the vacuolar sap (osmoregulatory hypothesis); and (v) direct The functions usually attributed to anthocyanins depend on or indirect defence from biotic stressors such as herbivores or whether they are accumulated in vegetative or reproductive pathogens, including functions such as warning signals, apose- organs. In the latter, it is broadly accepted that the main func- matics or camouflage, among others (biotic defence hypothe- tion of anthocyanins is to attract pollinators or seed dispersers sis) (Chalker-Scott 2002; Manetas 2006; Hatier & Gould 2009; to flowers or fruits respectively (Schaefer & Ruxton 2011). In Lev-Yadun & Gould 2009; Llorens et al. 2015). The three-first vegetative organs, plants often produce anthocyanins in hypotheses can be grouped under a ‘photoprotection hypothe- response to environmental stressors such as UV-B radiation, sis’, since anthocyanins can directly or indirectly protect

968 Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Narbona, Jaca, del Valle, Valladares & Buide Anthocyanin accumulation in Silene germana against excessive irradiance, including UV light (Landi et al. flavonoids, (ii) the photoinhibitory response and (iii) the 2015). Concerning photoprotection and osmoregulatory female reproductive performance. In view of the photoprotec- hypothesis, some authors advocated that other non-anthocya- tion hypothesis for the accumulation of anthocyanins in vege- nin flavonoids such as flavonols, flavones and flavanones are tative and reproductive organs (Steyn et al. 2002; Landi et al. better molecules than anthocyanins to carry out these functions 2015), we postulate that plants of S. germana should increase (Manetas 2006). These non-anthocyanin flavonoids are colour- anthocyanin accumulation with time of exposure and/or with less and their accumulation is usually correlated with those of intensity of sunlight, and anthocyanins should decrease pho- anthocyanins (Falcone Ferreyra et al. 2012; Del Valle et al. toinhibition. In addition, the spatiotemporal patterns of antho- 2015; Berardi et al. 2016). cyanin accumulation in relation with florivory and flower The regulation of anthocyanin biosynthesis is usually cell- or abortion may provide clues to other possible function of antho- tissue-specific, and this enables the plants to control their pro- cyanins (Steyn et al. 2002; Archetti 2009). duction in space (i.e., different parts of the plants) and time (i.e., specific period of plant life) (Albert et al. 2014). Vegetative anthocyanins may be accumulated in the whole plant or in MATERIALS AND METHODS specific parts, such as apical or basal sites on the leaves or stems Study system (Wheldale 1916; Neumann & Schwemmle 1994; Gould et al. 2010). In addition, some species remain red throughout their Silene germana is an annual winter species occurring in dolo- lives (Gould et al. 2000; Manetas 2006), but most plants turn mitic outcrops on mountains from 700 to 1800 m (Talavera red only at certain period of their lives (e.g., Del Valle et al. 1990). It is endemic to the south and southeast of Spain. This 2018a; Perea et al. 2018). Three main patterns have been small plant (<20 cm) produces one to six flowers acropetally detected regarding the appearance of anthocyanins in leaves (usually 1–2), arranged in a dichasial inflorescence. Flowers are during the plant’s life (i.e., the phenology of anthocyanin accu- characterized by small white petals and a calyx that swells dur- mulation): new-formed leaves produced in spring or summer ing flower anthesis, totally covering the ovary, and is persistent (juvenile or spring reddening), leaves that turn red only in win- in fructification (Fig. 1). Flowers are protandrous, but anthers ter (winter reddening) and those that become red just before and stigmas are closed and its maturation overlaps by some being shed (autumn reddening) (Chalker-Scott 2002; Garcıa- days, causing spontaneous self-pollination (E. Narbona, Plazaola et al. 2003; Hughes 2011; but see Fernandez-Marın unpublished data), as has occurred in other Silene species et al. 2015). These time and space patterns of anthocyanin pro- (Buide et al. 2015). duction in vegetative organs have most often been investigated The study was carried out in Sierra de Grazalema Natural in shrubs and trees. Thus, functions of anthocyanins are mostly Park, located in the province of Cadiz, southern Spain. We tested in perennial species (e.g., Archetti et al. 2009; Hughes studied the ‘Horno de la Miera’ population (36°470 N, 2011). Annual species may also accumulate anthocyanins, per- 5°240 W, 800 a.m.s.l.) situated on Jurassic dolomites. In this manently or transiently, in parts or in the whole plant (Whel- population, the number of individuals of S. germana varies dale 1916; Figure S1). Nevertheless, information regarding the from ca. 80 to 200 (dry and wet years respectively). The site has adaptive significance of anthocyanins is scarce and almost a typical dry summer Mediterranean climate, with a mean exclusively based on crop and ornamental species (Boldt et al. annual temperature of 15.3 °C and a mean annual rainfall of 2014; Tattini et al. 2014). Thus, an interesting open question is 966 mm. Vegetation consists of sparse scrub dominated by whether the functional roles proposed for anthocyanin accu- Ceratonia siliqua L., Juniperus oxycedrus L., J. phoenicea L., mulation in vegetative organs in perennial species are also Lavandula stoechas L. and Ulex baeticus Boiss. Plants were com- applicable to annual species. In this regard, studies of wild pop- pletely exposed to sun from approximately 09:00 to 17:00 h ulations in which plants are under the influence of natural because of the eastern exposure of the slope and the scarcity of selective agents and interconnected ecological factors are desir- tree canopy. able (e.g., Narbona et al. 2010; Menzies et al. 2015). In this paper, the anthocyanin accumulation in different Monitoring temporal redness variation, flowering, herbivory plant organs is examined in a wild population of Silene ger- and plant mortality mana J. Gay (Caryophyllaceae), a tiny annual plant occurring in mountains of the southern Iberian Peninsula. Flowering and In May 2014, we randomly selected and tagged 53 plants of dispersal take place from May to July; thus, S. germana plants S. germana. Population censuses were carried out once a week are exposed to the high temperatures, high solar radiation and throughout the reproductive season; from before flowering (6 drought of the Mediterranean summer, which commonly cause May 2014), until all fruits matured (10 June 2014) and the oxidative stress (Werner et al. 2002; Llorens et al. 2003). This plants died 1 week later. Daily climatic conditions during the species may accumulate anthocyanins in leaves and stems, but period of study were assessed using HelioClim-3 database (pro- also in the swollen photosynthetic calyx (Fig. 1). Caterpillars of vided by SoDa service) and are shown in Figure S2. In the third Geometridae sp. and Hadena sp. have been observed predating week, two plants died as a result of trampling and were not the ovary and calyces of the flowers (i.e., florivory) (E. Nar- considered in some analyses. Censuses were carried out at the bona, personal observation; Fig. 1C). To study the possible role same time (between 10:00 and 13:30 h) on sunny days. In each of anthocyanins in S. germana, we analysed the phenology of census, a photograph of each plant was taken to analyse the anthocyanin production and their relationship with florivory, plant’s redness. All photos were taken from the same angle flower abortion and plant mortality. In addition, a shading (90°) and distance from the plants with graph paper as a back- experiment was performed in the field to determine: (i) the ground. As the population is on an eastern slope, plants accumulation of anthocyanins and non-anthocyanin received more light hours on the east surface than that on the

Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 969 Anthocyanin accumulation in Silene germana Narbona, Jaca, del Valle, Valladares & Buide

Fig. 1. Details of Silene germana plants showing the different anthocyanin accumulation. (A, B) An individual sampled during weeks 2 and 5 respectively. Within-plant position (bottom, middle and upper) and inflorescence levels (inflorescence level 1 and 2) are shown; note that the plant is mostly green on week 2 (except for bottom stem and leaves), whereas on week 5 it is totally red. (C) Hadena sp. (yellow arrow) predating flower at bud stage of inflorescence level 2. (D, E) A inflorescence sampled during weeks 3 and 6 respectively. In D, red calyx of aborted flower at inflorescence level 2 is showed. In E, calyces, leaves and stems are still red in the last stage of senescence. (F) microscopic observation of transverse section of the abaxial surface of a leave; note that anthocyanins are accumulated in epidermal cells. west. We took pictures from the east side of the plants because accurately estimates the anthocyanin concentration in photo- we did not observe variation of redness on east and west sides synthetic tissues (Del Valle et al. 2018b). We analysed the entire of the plant. The exception to this occurs in the calyces from area of the leaves and calyces and 1 cm of the stems. We differ- the fourth week; thus, east and west sides were photographed entiated three parts of the plant’s architecture in leaves and in these censuses. stems (bottom, intermediate and upper parts), and two levels In each census, we also noted the number of flowers and of flowers in the inflorescence (Fig. 1A and B). fruits and their developmental stages (bud, flower anthesis, Repeatability of measurement (precision) for red/green aborted flower, immature and mature fruit; Fig. 1D). Florivory index of a specific plant area was assessed by means of an and plant mortality by drought were also surveyed; in the lat- ANOVA-based method according to Krebs (1999). We randomly ter, the plants appeared totally dry (yellow-brownish). Micro- selected six S. germana plants (six photographs), and each pho- scopic analysis of plants indicated that anthocyanins are tograph was measured six times. Repeatability was extremely accumulated in epidermal cells of leaves (adaxial and abaxial high at all plant organs and within-plant positions (Table S1). surfaces), calyces and stems (Fig. 1F). In addition, to ensure the accuracy of our anthocyanin estimates via redness of digital images, we carried out linear regressions between redness and anthocyanin concentration Redness quantification to estimate anthocyanin concentration measured spectrophotometrically (see below) in each plant All photographs were taken with a Canon G10 digital camera, organ. In leaf and stem samples, the relationship was highly with an effective pixel count of 14.7 megapixels. This is a mid- significant with moderate predictive power (R2 = 0.69, 2 priced equipment with high quality optics and full regulation F1,13 = 29.30, P < 0.001 and R = 0.63, F1,17 = 29.18, P < 0.0001 of exposure and metering, which is recommended for unbiased respectively). data acquisition (Stevens et al. 2007). We follow the methodology of image acquisition and processing detailed in Del Valle Shading experiment et al. (2018b), with the exception that calibration was assessed with the Chart White Balance macro (Haeghen et al. 2000) In May 2016, an experiment with four light environments was designed for ImageJ software (version 1.48; Schneider et al. performed in the same population: (i) ‘natural shading’ (91% 2012). Once images were calibrated, average red and green transmitted sunlight reduction, 45% UVA/B reduction), in channel values were measured in the leaves, stems and calyces which we sampled plants that were situated under dense shade using the same software. Afterwards, the red/green index (re- of an Ceratonia siliqua shrub; (ii) ‘mild shading’ (65% trans- ferred to below as redness) was calculated because this index mitted sunlight reduction, 92% UVA/B reduction), in which a

970 Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Narbona, Jaca, del Valle, Valladares & Buide Anthocyanin accumulation in Silene germana

60 9 60 cm high density polyethylene mesh nailed to the each plant, we measured the stem and the calyx of a flower; in ground at 50 cm height was used to cover the plants; (iii) ‘UV most cases, leaves were too small to put the leaf clip and pro- exclusion’ (9% transmitted sunlight reduction, 100% UVA/B duce accurate measurements. As we mentioned above, natural reduction), the same system than the later but using a polycar- shaded plants began to be exposed to solar radiation at noon, bonate sheet (4 mm thickness; SEDPA ind., Perenchies, whereas plants from the full sunlight treatment were exposed France); and (iv) full sunlight exposure plants (control). In for the whole day. In order to differentiate between the effects each treatment, total solar radiance and UVA/B were measured of anthocyanins and shade on photoinhibition, we performed by means of Megger PVM210 irradiance meter (Megger Co., another experiment in a group of plants growing in partial Dallas, USA) and PCE-UV34 UV light meter (PCE Inst., Dur- shading conditions. Four plants mostly red and other four ham, UK) respectively. Measures were taken at 10:00 and mostly green were covered from 07:00 to 12:00 h with a black 13:00 h of a sunny day (13th June), using the mean to calculate nylon mesh at 80 cm height (75% transmitted sunlight reduc- the radiation reduction of each treatment with respect to the tion, 98% UVA/B reduction) and photochemical efficiency was control plants (maximum solar radiance was 1153 and measured as described above. 1336 Wm 2, and UVA/B was 32 and 47 Wm 2 at 10:00 and 13:00 h respectively). In the natural shading experiment, plants Female reproductive success were exposed to full sunlight from approx. 12:00 to 15:30 h due to sun’s movement. Structures of mild shading and UV As an estimate of female reproductive success, we counted the exclusion treatments were placed in April, when plants are in total number of flowers, aborted flowers and mature fruits in seedling stage (1–2 cm height). Seven to 19 plants per treat- 34 and 78 plants from natural shading and full sunlight treatment were marked and the redness of calyces, leaves and stems ments respectively. In addition, the mature seeds and undevel- were measured at the flowering peak stage using the digital oped ovules in one fruit of six plants of each treatment were image method. On the other hand, the presence and location analysed. We then calculated the fruit and seed sets. of anthocyanins were examined through microscopic observations of hand sections. Statistical analysis In each plant organ, differences in redness between date and Spectrophotometric analysis of flavonoids within-plant position were analysed using linear mixed models The content of anthocyanins and non-anthocyanin flavonoids with repeated measurements. We considered ‘date’ as the fixed of S. germana plants from full sunlight and natural shading repeated variable, ‘within-plant position’ as a fixed variable treatments was determined by spectrophotometry using a within subjects, and ‘individual’ was a random factor. Because Multiskan GO microplate spectrophotometer (Thermo Fisher data were unbalanced, restricted methods of maximum likeli- Scientific Inc., MA, USA). One to three leaves, calyces and hood estimations were used in the model (Quinn & Keough samples of the stem (one centimetre long) were taken from 2002). Mean values for each factor level were compared using eight plants per treatment. Each plant material was placed the Bonferroni post-hoc test (Quinn & Keough 2002). In the into a 1.5 ml of MeOH:HCl (99:1% v:v), and subsequent con- analysis of calyces, aborted flowers were not considered (see servation and homogenisation were performed as described in below). To meet the requirements of normality and Del Valle et al. (2015). Weights of each sample were measured homoscedasticity, variables were transformed using boxcox using a precision scale. Three replicas of 200 ll per sample function in R (MASS library; Venables & Ripley 2002). were measured at A350 for combined non-anthocyanin flavo- To ascertain if calyx redness affects the probability of flower noids (flavones and flavonols) and at A530 nm for antho- predation or abortion, we used logistic regressions with a logit cyanins (Del Valle et al. 2015). We previously measured the link function and binomial error distribution. Flower abortion absorbance spectra of the samples from 200 to 800 nm to ver- mainly occurred at inflorescence level 2 (see Results); thus, ify the peaks of maximum absorbance. Anthocyanin content calyx redness of aborted flowers was compared with fruiting was adjusted using the formula A5300.24 9 A653, and total flowers at this inflorescence level. The relationship between anthocyanin and non-anthocyanin flavonoid concentrations redness and plant mortality by drought was also compared were quantified using five-point calibration curves of cyani- with logistic regressions. Because dead plants lose pigments din-3-glucoside and isoorientin standards respectively (see Del upon dying, we used the redness value of the census before the Valle et al. 2018a). event for the analyses, i.e., for dead plants on a given census day, the redness of these plants on the previous census were compared with those of the rest of plants on the previous cen- Chlorophyll fluorescence measurements sus. A similar procedure was used for predated flowers. Gener- To estimate the effect of anthocyanins in photoinhibition, we alized linear models (GLMs) with Gaussian link functions were analysed the photochemical efficiency of photosystem II (Fv/ used to test the effect of sun exposure (sunny and shade sides) Fm) in plants from full sunlight and natural shade treatments on the redness of the calyces. The same analyses were per- (N = 9 and 11 respectively) using a pulse-modulated fluorome- formed to test the differences of shading treatments in redness ter (FMS2, Hansatech Instruments, Norfolk, UK). Measure- and anthocyanin and non-anthocyanin flavonoids concentra- ments were carried out on two sunny days (6th and 13th June tions, considering ‘treatment’ and ‘plant part’ as fixed factors 2016) in early morning, before noon and afternoon (08:30, (Gaussian link functions). Variables were log transformed to 11:30 and 13:30 h GMT + 2 respectively). Before each measure- fulfil normality. In each plant part, the relationships between ment, samples were dark-acclimated for 30 min using Handy- redness and total solar and UVA/B were performed by mean of PEA leaf clips with a mobile shutter (Aragon et al. 2008). In linear regressions. Differences in Fv/Fm between plants from full

Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 971 Anthocyanin accumulation in Silene germana Narbona, Jaca, del Valle, Valladares & Buide sunlight and natural shade treatments were performed by (1.23 0.027 for sunny sides and 1.15 0.026 for shade sides t-tests in each time of the day; measurements of stems and (t1,13 = 2.22; P < 0.05). calyces were merged to increase statistical power. Finally, GLMs with Poisson and logit link functions were used to test the Relationship among redness and flower abortion, herbivory effect of these two treatments in female reproductive compo- and plant mortality nents. In logistic regressions and GLMs, model selection was car- Most of the plants produced one or two flowers (58.8% and ried out using the Akaike’s information criterion (Crawley 31.4% respectively) and only 7.8% and 2.0% produced three 2007). We used quasibinomial error structure to correct for and six flowers. Six plants (11.8%) did not reproduce because data overdispersion when necessary (Crawley 2007). Linear the only flower produced was predated by caterpillars. A total regressions and GLMs were performed in R v3.1.1 (R Core of 15 flowers on 15 different plants (29.4% of plants studied) Team 2016) using R-studio v 0.99.486, whereas mixed-model showed predation during the study. Florivory occurred on ANOVAs were carried out using SPSS version 20.0 (SPSS, Inc., weeks 2 (13.3%), 3 (66.7%), 4, 5 and 6 (6.7% each day). Using 2011, Chicago, IL, USA). Multiple comparisons between the data set for week 3, logistic regression yielded no effect on morphs were carried out using the ‘multcomp’ R-package with the redness of the calyx upon occurrence of florivory Bonferroni adjustment (Hothorn et al. 2008). (0.96 0.007 for predated flowers and 0.97 0.015 for unpre- dated flowers; Z1,52 = 0.421, P = 0.67). The percentage of flowers setting fruits was 100% in plants RESULTS with only one flower, but flower abortion represented 62.5%, 16.7% and 33.3% of the flowers of plants with two, three and Spatial and temporal redness variation six flowers respectively. Flower abortion occurred on weeks 3 Redness of leaves and stems significantly varied with both plant (71.4% of total aborted flowers), 4 and 6 (14.3% each week). positions and date (Table 1). In general, leaves and stems Results of logistic regression for the data set of week 3, showed showed a decrease of redness from the bottom up, and an that aborted flowers had statistically higher calyx redness than increase of redness during plant life (Fig. 2A and B). Interac- fruiting flowers (1.23 0.024 and 1.07 0.013 respectively; tion between position and date was also significant, i.e., the t1,19 = 3.08; P < 0.01; Fig. 1E). increase of redness over time and at each within-plant position Plants started to die due to drought on weeks 3 and 4, but was not homogeneous (Table 1). Thus, at the lowest part of the the main events of mortality occurred on the last weeks (Fig- plant, redness of leaves and stems significantly increased from ure S3). Plant mortality was not related to the redness of stems week 1 to 3, and then remained constant (Fig. 2A and B). How- nor leaves; in fact, dead and live plants showed similar redness ever, in the middle and upper level leaves, redness significantly of stems and leaves (Table 2). Plant size was not related to increased from week 1 to 5 or 6. In general, leaves and stems mortality, except for week 6, on which live plants were taller from bottom levels showed higher redness than middle and than dead plants (149.6 8.6 cm and 112.5 6.3 cm respec- upper levels from weeks 1 to 4 (Table S2); on week 5, the mid- tively). dle and upper levels reached the same redness as the bottom level leaves. Shading experiment Calyx redness of inflorescence levels 1 and 2 increased from the bud and flower anthesis stage (weeks 1 and 2) to the fruit Anthocyanin accumulation was significantly different across dispersal stage (weeks 5 and 6) (Fig. 2C). Analysis showed a the shading environments (t2,179 = 2.23; P < 0.01). In calyces, significant effect on date, but not on inflorescence levels leaves and stems, full sunlight plants showed the higher redness (Table 1); at both first and second inflorescence levels, redness whereas natural shaded plants displayed the lowest redness statistically increased from week 1 to 5. In general, the whole (Fig. 3; Table S3 and Figure S4). Redness showed a positive plant becomes red at the end of flowering and remains red relationship with solar radiation intensity in each plant part until senescence (Fig. 1E). (r2 = 0.93, t = 5.30, P < 0.05 for calyces; r2 = 0.99 t = 34.79, On week 4, the sides of the calyces of the inflorescence level P < 0.0001 for leaves; and r2 = 0.93, t = 5.16, P < 0.05 for 1 with greater exposure to sunlight were significantly redder stems). Conversely, redness was statistically unrelated to UV than those less exposed (mean redness SE = 1.17 0.019 radiation intensity (r2 = 0.51, t = 1.44, P = 0.29 for calyces; 2 2 and 1.12 0.015 respectively; t1,45 = 2.23; P < 0.05). Similar r = 0.25, t = 0.82, P = 0.50 for leaves; and r = 0.32, t = 0.96, results were found for the calyces of the inflorescence level 2 P = 0.43 for stems). In fact, plants in the UV exclusion

Table 1. Results of the linear mixed models testing for the effects of within-plant position and phenology on the redness of leaves, stems and calyces.

leaves stems calyces variables df FP-value df FP-value df FP-value

Position 2/388 115.2 <0.0001 2/618 91.4 <0.0001 1/193 3.4 0.068 Date 5/133 270.9 <0.0001 5/202 321.6 <0.0001 5/58 282.1 <0.0001 Position 9 date 9/160 7.0 <0.0001 10/202 14.9 <0.0001 5/57 0.6 0.693 df, degrees of freedom of the numerator by the denominator.

972 Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Narbona, Jaca, del Valle, Valladares & Buide Anthocyanin accumulation in Silene germana

Table 2. Results of logistic regression analyses on the inﬂuence of redness of stems and leaves and plant size on plant mortality.

stems leaves

variables df tP-value tP-value

Week 4 Redness 50 1.417 0.163 0.393 0.693 Height 50 0.221 0.826 0.007 0.995 Week 5 Redness 43 0.858 0.396 0.571 0.571 Height 43 1.736 0.090 1.352 0.182 Week 6 Redness 30 0.344 0.733 0.589 0.560 Height 30 2.148 0.041 2.131 0.042

For calculation of stems and leaves redness, the value of bottom, middle and upper positions were averaged. Quasibinomial distribution in logistic regression model was used to correct for data overdispersion.

a a ab a a

ab ab ab

ab b b b Redness (red/green index) Redness (red/green

Fig. 2. Spatial and temporal variation of redness in leaves (A), stems (B) and calyces (C) during the study. Within-plant position levels (bottom, middle and upper) are shown for leaves and stems, and within-inflorescence levels (first and second) are shown for calyces. Values are means standard Fig. 3. Redness of leaves, stems and calyces of S. germana plants growing errors. Redness was measured on calibrated photographs using the red/ under four light environments. For calculation of stems and leaves redness, green index (see Methods). Horizontal greyscale boxes represent different the value of bottom, middle and upper positions were averaged. Different developmental stages of flowering: bud and flower anthesis (black), fruit letters indicate significant differences among treatments within each plant (grey) and fruit dispersal (white). Different letters indicate significant differ- part using GLMs and Bonferroni post-hoc test (a < 0.05). ences among dates within each position, using the Bonferroni post-hoc test (a < 0.05). Note that there are no values for bottom leaves on week 6 exposed to sunlight, and they showed statistically similar Fv/Fm because all these leaves became senescent. levels than full sunlight plants (Fig. 5A). In artificially shaded plants, red plants showed an early morning significant decrease in Fv/Fm compared to green plants; before noon, this trend is treatment showed statistically similar redness levels than full maintained but differences were not statistically significant, sunlight plants (Fig. 3). and in afternoon anthocyanic and non-anthocyanic plants Spectrophotometric analyses showed that plants from full showed statistically similar low Fv/Fm values (Fig. 5B). sunlight treatment had significantly higher content of antho- Full sunlight plants produced a significantly higher number cyanins than plants from natural shade treatment (t1,29 = of flowers than plants of the natural shade treatment (2.45 ver- 4.57, P < 0.0001; Fig. 4A); in fact, the latter produced negligi- sus 1.43 flowers per plant respectively; Table 3). The propor- ble amounts of anthocyanins in leaves, stems and calyces. tion of flowers setting fruits was statistically similar between Regarding non-anthocyanin flavonoids, plants from both treatments; as a result, the total number of fruits per plant was treatments produced statistically similar amounts of these higher in full sunlight than in natural shade (2.14 versus 1.29 compounds (t1,29 = 0.084, P = 0.934), and non-significant fruits per plant respectively). With respect to number of ovules, differences were observed in each plant part (Fig. 4B). seeds and seed set, plants from both treatments exhibited statis- In early morning, photochemical efficiency was 14% lower tically similar values (Table 3). in full sunlight anthocyanic plants than natural shaded non- anthocyanic plants (Fig. 5A). Before noon, plants from full DISCUSSION sunlight were exposed to sunlight and showed a high decrease of Fv/Fm with respect to the natural shaded plants. In the after- Vegetative organs of Silene germana gradually become red from noon, plants from the natural shade treatment were fully the beginning to the end of flowering. Calyces also become red

Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 973 Anthocyanin accumulation in Silene germana Narbona, Jaca, del Valle, Valladares & Buide

A B

FW) *** -1 *** ns **

ns Flavonoids content (mg g

Fig. 4. Concentration of anthocyanins (A) and non-anthocyanin flavonoid (B) in leaves, stems and calyces of S. germana plants growing under full sunlight (dark bars) and natural shade (white bars) treatments. Asterisks indicate significant differences between treatments within each plant part using GLMs. ns, not significant; **P < 0.01; ***P < 0.001.

during development, being fully red in early fruit and dispersal ** A stages. At the fruiting period, the whole plant becomes red, and *** maintains this colour during senescence. Transient reddening of leaves, or the whole plant, in annual species is relative common (Wheldale 1916, pp 22). However, in some annual species like S. germana, the reddening may be permanent, occurring from the reproductive period to the last stages of senescence. ns For instance, southern populations of the annual S. littorea showed dark red leaves, stems and calyces during the flowering period (Del Valle et al. 2015), and a similar reddening pattern is found in S. vulgaris and S. latifolia when plants grow in stressful soils (Kamsteeg et al. 1979; Ernst et al. 2000). In addition, other Mediterranean annual species also exhibit this red- Fv/Fm B dening pattern during reproduction and senescence, such as * ns Rumex bucephalophorus and Crassula tillaea (Figure S1). Thus, whole-plant reddening seems common among annual or short perennial, herbaceous species, at least in some groups of angiosperms. We have found that in both leaves and stems, anthocyanins ns are gradually accumulated from the bottom up. Higher anthocyanin content in bottom parts has been reported in branches of Quintinia serrata, in which the most apical, youngest leaves bear significantly less anthocyanins than older leaves (Gould et al. 2000). The higher redness in both bottom leaves and stems compared with the younger apical ones, could simply be Time of the day (GMT +2) linked to the longer light exposure of older tissues, given the acropetal growth of S. germana. Heterogeneity in cell response to stimuli is possible, given that anthocyanin production is cell- Fig. 5. Variation of photochemical efficiency (Fv/Fm) from early morning to and tissue-specifically regulated (Albert et al. 2014), which may afternoon in (A) plants from full sunlight (filled circles) and natural shade lead to a gradual increase in redness at certain parts of the plant (empty circles) treatments, and (B) anthocyanic plants (filled circles) and that correspond to increasing light intensity or sun exposure non-anthocyanic plants (empty circles) from artificially shaded plants. In A, (Del Valle et al. 2015). This may also explain why the sun- plants from the full sunlight treatment were exposed to solar radiation for the whole day, whereas natural shaded plants began to be exposed at noon. exposed sides of the calyces were redder than the shaded sides, In B, both anthocyanic and non-anthocyanic plants were covered with a as is found in stems of Cornus stolonifera and in peduncles and black nylon mesh from 07:00 to 12:00 h. Data of stems and calyces were rays of inflorescences of Sambucus nigra (Gould et al. 2010; merged to increase statistical power. Asterisks indicate significant differ- Cooney et al. 2015). On the other hand, our shading experi- ences between treatments within each time of day using t-tests. ns, not ment showed a clear positive relationship between redness and significant; *P < 0.05; **P < 0.01; ***P < 0.001. intensity of solar radiation, in which plants growing in natural shade lack anthocyanins. It is common that production of anthocyanins is activated by sun exposition; in fact, the regula- the wavelength of the light spectrum, visible or UV, that is tory genes of the anthocyanin biosynthetic pathway are acti- involved in the activation the anthocyanin production seems vated by light (Albert et al. 2014; Zoratti et al. 2014). However, variable among species (Zoratti et al. 2014; Llorens et al. 2015).

974 Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Narbona, Jaca, del Valle, Valladares & Buide Anthocyanin accumulation in Silene germana

Table 3. The effect of light environment on female reproductive success of synthesis in senescent leaves is hypothesized to extend their Silene germana. lifespan (Schaberg et al. 2008) or increase nutrient resorption (Hoch et al. 2003). Because calyces of aborted flowers of S. ger- variables mean SE df tP-value mana turn red before becoming senescent, anthocyanins may Number of flowers Full sunlight 2.45 0.19 112 2.95 <0.01 play a role in increasing nutrient resorption, which could be Natural shade 1.43 0.19 relocated from the calyx to the ovary (Lim et al. 2007). In this Number of fruits Full sunlight 2.14 0.18 112 3.04 <0.01 vein, we have found that the whole plant becomes red during Natural shade 1.29 0.21 reproduction period and remains red during senescence. Thus, Fruit set (%) Full sunlight 90.5 2.79 112 0.82 0.413 anthocyanins may play a role in increasing nutrient recovery at Natural shade 85.7 6.09 the whole-plant level during senescence (Peng et al. 2008; Number of ovules Full sunlight 41.7 3.52 11 0.13 0.897 Landi et al. 2015), which coincides with the important stage of Natural shade 42.2 1.33 fruit maturation. Recently, it has also been suggested a similar Number of seeds Full sunlight 30.0 5.47 11 0.06 0.954 role of anthocyanins in senescing peduncles of Sambucus nigra Natural shade 29.7 1.65 (Cooney et al. 2015, 2018). Seed set (%) Full sunlight 70.7 8.02 11 0.05 0.956 The calyces of herbivorized and non-herbivorized flowers Natural shade 70.3 2.88 showed a similar redness. This suggests that caterpillars do not Differences between treatments were performed by GLMs with quasibino- differentially select calyces with different redness. The direct mial distribution to correct for data overdispersion. effect of anthocyanins as deterrents of florivorous insects is controversial (Lev-Yadun & Gould 2009). In some study sys- tems, insects avoid the plant parts containing anthocyanins Our results suggest that visible light, but not UVA/B, is neces- (Cooney et al. 2012; Menzies et al. 2015), whereas in others this sary for anthocyanin production in leaves, stems and calyces of does not occur (Archetti et al. 2009). Although our result does S. germana. not support a direct effect of anthocyanins against herbivores All the above lines of evidence suggest that the accumulation in S. germana, other anti-herbivory functions of anthocyanins, of anthocyanins in S. germana may be related to a photopro- such as aposematism or camouflage against more generalist tective function. Within the photoprotection hypothesis, the herbivores, cannot be discarded (Lev-Yadun & Gould 2009). UV protection role is barely probable given that anthocyanins In conclusion, our results showed a spatiotemporal pattern are produced in the absence of UV radiation. On the other of anthocyanin accumulation in S. germana. In general, the hand, the photoinhibitory role is usually tested by exposing whole plant becomes red at the end of flowering and remains vegetative tissues to high light fluxes (Hatier & Gould 2009); red until fruiting and senescence. Our shading experiment plant parts with anthocyanins commonly show higher Fv/Fm showed that the accumulation of anthocyanins, but not non- values than green parts, and then better high-stress recovery anthocyanin flavonoids, is related to visible radiation intensity. (e.g., Gould et al. 2010; Cooney et al. 2015; but see Fernandez- Conversely, we have found no effect of redness on flower pre- Marın et al. 2015). In contrast, we found that anthocyanic and dation. Taken together, these findings indicate that the light non-anthocyanic plants showed similar levels of Fv/Fm after intensity may play a decisive role in the accumulation of antho- two hours of sun exposition, but in early morning conditions, cyanins in S. germana, but their specific photoprotective role the former showed lower Fv/Fm values. The fact that antho- needs further research. It is particularly interesting that, despite cyanic plants of artificial shading experiment also showed lower the extreme solar radiation of the Mediterranean summer Fv/Fm values than non-anthocyanic plants suggests that this (Martınez-Ferri et al. 2000), S. germana plants growing at full reduction in photochemical efficiency is chronic. Similar sunlight conditions showed higher female reproductive success reduction in Fv/Fm in early morning has been documented in than plants under natural shade. The fact that anthocyanin red versus green leaves of other species (reviewed in Gould et al. accumulation was produced in different plant organs, with a 2018) and may be explained because green leaves obtain photo- temporal within-plant position variation, would allow multiple protection through other mechanisms (Williams et al. 2003), roles for anthocyanins (Kovinich et al. 2014). which does not imply absence of photoprotective role of anthocyanins in red leaves (Logan et al. 2015). Like other ACKNOWLEDGEMENTS perennial species of the Mediterranean (Martınez-Ferri et al. 2000; Aragon et al. 2008), plants of S. germana are simultane- We thank Ana Prado for her assistance both in field and lab, ously exposed to excess radiation, drought and high tempera- Luis Gimenez-Benavides for assistance in identification of ture (Valladares & Pearcy 2002), and even though they showed caterpillars, Anna Grandell for English edition, Claire Thomas a higher female reproductive success in full sunlight than in from SoDa Service for providing climatic data of the study site, natural shade. Anthocyanins and constitutive non-anthocyanin and the Consejerıa de Medio Ambiente from Cadiz for allow- flavonoids, may help this species deal with the oxidative stress ing the research at the Sierra de Grazalema Natural Park. This of the excessive summer sunlight (Agati et al. 2009; Landi et al. work was supported by FEDER funds and grants from the 2015), but further studies are needed to confirm this role. Spanish Ministry of Economy and Competitiveness through Calyces of aborted flowers were much redder than those of the research projects CGL2012-37646 and CGL2015-63827-P. fruiting flowers. Because flowers show spontaneous self-pollination (E. Narbona, unpublished data), flower abortion is SUPPORTING INFORMATION probably caused by lack of resources or architectural effects (Stephenson 1981), as occurs in other Silene species with cym- Additional supporting information may be found online in the ose inflorescences (Buide 2008; Buide et al. 2015). Anthocyanin Supporting Information section at the end of the article.

Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 975 Anthocyanin accumulation in Silene germana Narbona, Jaca, del Valle, Valladares & Buide

Figure S1. Examples of annual species of the Mediterranean Figure S4. Surface and microscopic observations of leaves pasture showing reddening in leaves, stems, calyces or fruits (Las and stems of Silene germana from the shading experiment. A: Pajanosas population, Seville, Spain; °3401500 N, 6°0601600 W). Stems of plants from full sunlight (left) and natural shade treat- Photos were taken during winter-spring 2015. (A) Erodium prim- ments (right). B: Transverse section of a stem from a full sun- ulaceum Welw. ex Lange, Geraniaceae; (B) Tripodion tetraphyllum light plant. C: Leaves of plants from full sunlight (left) and (L.) Fourr. (Fabaceae); (C) Stachys arvensis L. (Lamiaceae); (D) natural shade treatments (right). D: Microscopic observations Borago ofﬁcinalis L. (Boraginaceae); (E) Crassula tillaea Lester- of transverse section of the abaxial surface of a leave from full Garland (Crassulaceae); (F) Bromus driandrus Roth (Gramineae); sunlight treatment; note that anthocyanins are accumulated in (G) Sherardia arvensis L. (Rubiaceae); (H) Rumex bucephalopho- vacuoles of epidermal cells. rus L. (Polygonaceae). Table S1. Repeatability of measurement analyses of the red/ Figure S2. Climatic conditions during the phenology of green ratio for leaves, stems and calyces in each within-plant anthocyanin accumulation study (2014) at La Camilla popula- position. tion (P. N. Grazalema, Spain). Squares, daily maximum tempera- Table S2. Estimated means, standard errors and conﬁdence ture; triangles, daily minimum temperature; circles, UV radiation intervals (95%) for each plant position and census day (Jcm 2); bars, global horizontal irradiation (Whm 2). Census obtained for the linear mixed models analyses performed for days are highlighted in gray bars. Precipitation occurs on 20th leaves, stems and calyces. May (6.8 mm), 31 May (1.8 mm) and 1 June (10.4 mm). Data Table S3. Results of Bonferroni multiple comparisons tests from HelioClim-3 database. comparing redness of leaves, stems and calyces among the Figure S3. Plant survival during the study period of Silene shading treatment. germana at the La Camilla population.

REFERENCES Chalker-Scott L. (2002) Do anthocyanins function as biotechnological applications. Frontiers in Plant osmoregulators in leaf tissues? Advances in Botanical Science, 3, 222. Agati G., Stefano G., Biricolti S., Tattini M. (2009) Research, 37, 103–127. Fernandez-Marın B., Esteban R., Mıguez F., Artetxe U., Mesophyll distribution of ‘antioxidant’ flavonoid Cooney L.J., van Klink J.W., Hughes N.M., Perry N.B., Castaneda~ V., Pinto-Marijuan M., Becerril J.M., glycosides in Ligustrum vulgare leaves under con- Schaefer H.M., Menzies I.J., Gould K.S. (2012) Red Garcıa-Plazaola J.I. (2015) Ecophysiological roles of trasting sunlight irradiance. Annals of Botany, 104, leaf margins indicate increased polygodial content abaxial anthocyanins in a perennial understorey herb 853–861. and function as visual signals to reduce herbivory in from temperate deciduous forests. AoB Plants, 7, Albert N.W., Davies K.M., Lewis D.H., Zhang H., Pseudowintera colorata. New Phytologist, 194, 488– plv042. Montefiori M., Brendolise C., Boase M.R., Ngo H., 497. Garcıa-Plazaola J.I., Hernandez A., Becerril J.M. (2003) Jameson P.E., Schwinn K.E. (2014) A conserved net- Cooney L.J., Schaefer H.M., Logan B.A., Cox B., Gould Antioxidant and pigment composition during work of transcriptional activators and repressors reg- K.S. (2015) Functional significance of anthocyanins autumnal leaf senescence in woody deciduous spe- ulates anthocyanin pigmentation in Eudicots. The in peduncles of Sambucus nigra. Environmental and cies differing in their ecological traits. Plant Biology, Plant Cell, 26, 962–980. Experimental Botany, 119,18–26. 5, 557–566. Andersen Ø.M., Jordheim M. (2006) The antho- Cooney L.J., Logan B.A., Walsh M.J., Nnatubeugo Gould K.S., Markham K.R., Smith R.H., Goris J.J. cyanins. In: Andersen Ø. M., Markham K. R. (Eds), N.B., Reblin J.S., Gould K.S. (2018) Photoprotection (2000) Functional role of anthocyanins in the leaves Flavonoids: chemistry, biochemistry and applica- from anthocyanins and thermal energy dissipation of Quintinia serrata A. Cunn. Journal of Experimental tions2nd edn. CRC Press, Boca Raton, FL, pp 452– in senescing red and green Sambucus canadensis Botany, 51,1107–1115. 471. peduncles. Environmental and Experimental Botany, Gould K.S., Dudle D.A., Neufeld H.S. (2010) Why Aragon C.F., Escudero A., Valladares F. (2008) Stress- 148,27–34. some stems are red: cauline anthocyanins shield induced dynamic adjustments of reproduction dif- Crawley M.J. (2007) The R book. John Wiley and Sons, photosystem II against high light stress. Journal of ferentially affect fitness components of a semi-arid Chichester, UK. Experimental Botany, 61, 2707–2717. plant. Journal of Ecology, 96, 222–229. Del Valle J.C., Buide M.L., Casimiro-Soriguer I., Whit- Gould K.S., Jay-Allemand C., Logan B.A., Baissac Y., Archetti M. (2009) Classification of hypotheses on the tall J.B., Narbona E. (2015) On flavonoid accumula- Bidel L.P. (2018) When are foliar anthocyanins evolution of autumn colours. Oikos, 118, 328–333. tion in different plant parts: variation patterns useful to plants? Re-evaluation of the photo- Archetti M., Dorin€ T.F., Hagen S.B., Hughes N.M., among individuals and populations in the shore protection hypothesis using Arabidopsis thaliana Leather S.R., Lee D.W., Lev-Yadun S., Manetas Y., campion (Silene littorea). Frontiers in Plant Science, mutants that differ in anthocyanin accumula- Ougham H.J., Schaberg P.G., Thomas H. (2009) 6, 939. tion. Environmental and Experimental Botany. Unravelling the evolution of autumn colours: an Del Valle J.C., Buide M.L., Whittall J.B., Narbona E. https://doi.org/10.1016/j.envexpbot.2018.02.006 interdisciplinary approach. Trends in Ecology & Evo- (2018a) Phenotypic plasticity in light-induced flavo- Haeghen Y.V., Naeyaert J.M.A.D., Lemahieu I., Philips lution, 24, 166–173. noids varies among tissues in Silene littorea W. (2000) An imaging system with calibrated color Berardi A.E., Hildreth S.B., Helm R.F., Winkel B.S., (Caryophyllaceae). Environmental and Experimental image acquisition for use in dermatology. IEEE Smith S.D. (2016) Evolutionary correlations in fla- Botany, 153, 100–107. Transactions on Medical Imaging, 19, 722–730. vonoid production across flowers and leaves in the Del Valle J.C., Gallardo-Lopez A., Buide M.L., Whittall Hatier J.H.B., Gould K.S. (2009) Anthocyanin function Iochrominae (Solanaceae). Phytochemistry, 130, 119– J.B., Narbona E. (2018b) Digital photography as a in vegetative organs. In: Gould K. S., Davies K. M., 127. reliable, non-destructive method to estimate antho- Winefield C. (Eds), Anthocyanins. Springer, New Boldt J.K., Meyer M.H., Erwin J.E. (2014) Foliar anthocyanin pigment content in vegetative and reproduc- York, NY, USA, pp 1–19. cyanins: a horticultural review. Horticultural tive plant organs. Ecology and Evolution, 8, 3064– Hoch W.A., Singsaas E.L., McCown B.H. (2003) Reviews, 42, 209–252. 3076. Resorption protection. Anthocyanins facilitate nutri- Buide M.L. (2008) Disentangling the causes of intrain- Ernst W.H., Nelissen H.J., Ten Bookum W.M. ent recovery in autumn by shielding leaves from florescence variation in floral traits and fecundity in (2000) Combination toxicology of metal-enriched potentially damaging light levels. Plant Physiology, the hermaphrodite Silene acutifolia. American Jour- soils: physiological responses of a Zn-and Cd- 133, 1296–1305. nal of Botany, 95, 490–497. resistant ecotype of Silene vulgaris on polymetallic Hothorn T., Bretz F., Westfall P. (2008) Simultaneous Buide M.L., del Valle J.C., Pissatto M., Narbona E. soils. Environmental and Experimental Botany, 43, inference in general parametric models. Biometrical (2015) Night life on the beach: selfing to avoid polli- 55–71. Journal, 50,346–363. nator competition between two sympatric Silene spe- Falcone Ferreyra M.L., Rius S., Casati P. (2012) Flavo- Hughes N.M. (2011) Winter leaf reddening in ‘ever- cies. Annals of Botany, 116, 201–211. noids: biosynthesis, biological functions, and green’ species. New Phytologist, 190, 573–581.

976 Plant Biology 20 (2018) 968–977 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Narbona, Jaca, del Valle, Valladares & Buide Anthocyanin accumulation in Silene germana

Kamsteeg J., Brederode J.V., Nigtevecht G.V. (1979) at midday during the Mediterranean summer. Tree Review of Ecology, Evolution and Systematics, 12, Genetics of anthocyanin formation in petals of the red Physiology, 20, 131–138. 253–279. campion (Silene dioica (L) Clairv.). Genetica, 51,5–13. Menzies I.J., Youard L.W., Lord J.M., Carpenter K.L., Stevens M., Parraga C.A., Cuthill I.C., Partridge J.C., Kovinich N., Kayanja G., Chanoca A., Riedl K., Otegui van Klink J.W., Perry N.B., Schaefer H.M., Gould Troscianko T.S. (2007) Using digital photography to M.S., Grotewold E. (2014) Not all anthocyanins are K.S. (2015) Leaf colour polymorphisms: a balance study animal coloration. Biological Journal of Lin- born equal: distinct patterns induced by stress in between plant defence and photosynthesis. Journal of nean Society, 90, 211–237. Arabidopsis. Planta, 240, 931–940. Ecology, 104, 104–113. Steyn W.J., Wand S.J.E., Holcroft D.M., Jacobs G. Krebs C.J. (1999) Ecological methodology. Benjamin/ Narbona E., Guzman B., Arroyo J., Vargas P. (2010) (2002) Anthocyanins in vegetative tissues: a pro- Cummings, Menlo Park, CA, USA. Why are fruit traits of Cistus ladanifer (Cistaceae) so posed unified function in photoprotection. New Landi M., Tattini M., Gould K.S. (2015) Multiple func- variable: a multi-level study across the western Phytologist, 155, 349–361. tional roles of anthocyanins in plant-environment Mediterranean region. Perspectives in Plant Ecology, Talavera S. (1990) Silene L. In: Castroviejo S., Laınz interactions. Environmental and Experimental Bot- Evolution and Systematics, 12, 305–315. M., Lopez Gonzalez G., Monserrat P., Munoz~ Gar- any, 119,4–17. Neumann G., Schwemmle B. (1994) Organ-and tissue- mendia F., Paiva J., Villar L. (Eds), Flora iberica: Lee D.W., Gould K.S. (2002) Anthocyanins in leaves specific biosynthesis of flavonoids in seedlings of plantas vasculares de la penınsula iberica e islas and other vegetative organs: an introduction. Oenothera odorata (Onagraceae). Plant Biology, 107, baleares. Real Jardın Botanico-CSIC, Madrid, pp Advances in Botanical Research, 37,1–16. 95–102. 313–406. Lev-Yadun S., Gould K.S. (2009) Role of anthocyanins Peng M., Hudson D., Schofield A., Tsao R., Yang R., Tattini M., Landi M., Brunetti C., Giordano C., in plant defence. In: Gould K. S., Davies K. M., Gu H., Rothstein S.J. (2008) Adaptation of Ara- Remorini D., Gould K.S., Guidi L. (2014) Epidermal Winefield C. (Eds), Anthocyanins. Springer, New bidopsis to nitrogen limitation involves induction of coumaroyl anthocyanins protect sweet basil against York, NY, USA, pp 22–28. anthocyanin synthesis which is controlled by the excess light stress: multiple consequences of light Lim P.O., Kim H.J., Nam H.G. (2007) Leaf senescence. NLA gene. Journal of Experimental Botany, 59, 2933– attenuation. Physiologia Plantarum, 152,585–598. Annual Reviews in Plant Biology, 58, 115–136. 2944. Valladares F., Pearcy R.W. (2002) Drought can be Llorens L., Penuelas~ J., Filella I. (2003) Diurnal and Perea R., Fernandes G.W., Dirzo R. (2018) Embryo size more critical in the shade than in the sun: a field seasonal variations in photosynthetic performance as a tolerance trait against seed predation: contribu- study of carbon gain and photo-inhibition in a Cali- and water relations of two co-occurring Mediter- tion of embryo-damaged seeds to plant regeneration. fornian shrub during El Nino~ year. Plant, Cell & ranean shrubs, Erica multiflora and Globularia alpi- Perspectives in Plant Ecology, Evolution and Systemat- Environment, 25, 749–759. num. Physiologia Plantarum, 118,84–95. ics, 31,7–16. Venables W.N., Ripley B.D. (2002) Modern applied Llorens L., Badenes-Perez F.R., Julkunen-Tiittoc R., Quinn G.P., Keough M.J. (2002) Experimental design statistics with S. Springer, New York, NY, USA. Zidornd C., Fereres A., Jansen M.A.K. (2015) The and data analysis for biologists. Cambridge University Werner C., Correia O., Beyschlag W. (2002) Character- role of UV-B radiation in plant sexual reproduction. Press, Cambridge, UK. istic patterns of chronic and dynamic photoinhibi- Perspectives in Plant Ecology, Evolution and Systemat- R Core Team (2016) R: a language and environment tion of different functional groups in a ics, 17, 243–254. for statistical computing. Available from http:// Mediterranean ecosystem. Functional Plant Biology, Logan B.A., Stafstrom W.C., Walsh M.J., Reblin J.S., www.R-project.org (accessed 12 March 2016). 29, 999–1011. Gould K.S. (2015) Examining the photoprotection Schaberg P.G., Murakami P.F., Turner M.R., Heitz Wheldale M. (1916) The anthocyanin pigments of hypothesis for adaxial foliar anthocyanin accumula- H.K., Hawley G.J. (2008) Association of red col- plants. University Press Cambridge, UK, Cambridge. tion by revisiting comparisons of green-and red- oration with senescence of sugar maple leaves in Williams E.L., Hovenden M.J., Close D.C. (2003) leafed varieties of coleus (Solenostemon scutellari- autumn. Trees, 22, 573–578. Strategies of light energy utilisation, dissipation and oides). Photosynthesis Research, 124,267–274. Schaefer H.M., Ruxton G.D. (2011) Plant-animal com- attenuation in six co-occurring alpine heath species Manetas Y. (2006) Why some leaves are anthocyanic munication. Oxford University Press, Oxford, UK. in Tasmania. Functional Plant Biology, 30, 1205– and why most anthocyanic leaves are red? Flora, 201, Schneider C., Rasband W., Eliceri K. (2012) NIH 1218. 163–177. Image to ImageJ: 25 years of image analysis. Nature Zoratti L., Karppinen K., Luengo Escobar A., H€aggman Martınez-Ferri E., Balaguer L., Valladares F., Chico Methods, 9, 671–675. H., Jaakola L. (2014) Light-controlled flavonoid J.M., Manrique E. (2000) Energy dissipation in Stephenson A.G. (1981) Flower and fruit abortion: biosynthesis in fruits. Frontiers in Plant Science, 5, drought-avoiding and drought-tolerant tree species proximate causes and ultimate functions. Annual 534.