Portal de Periódicos da CAPES

Coevolution and photoprotection as complementary hypotheses for autumn leaf reddening: a nutrient‐centered perspective

2021; Wiley; Volume: 233; Issue: 1 Linguagem: Inglês

10.1111/nph.17735

1469-8137

Nicole M. Hughes, Christian O. George, Corinne B. Gumpman, Howard S. Neufeld,

Light effects on plants

There has been recent discussion in the Forum of New Phytologist on the topic of adaptive functions of red, anthocyanin pigments in senescing leaves of temperate deciduous trees species in autumn. In this Letter, we (1) respond to some points raised previously in the Forum, (2) present the argument that soil nutrient characteristics should also be considered among the suite of environmental factors potentially driving evolution of autumn leaf color, and (3) provide evidence that nutrient deficiencies in soils of eastern North America (ENA) could also potentially account for the greater abundance of red-leafed species compared with Europe, especially when considered in combination with temperature and irradiance trends described by others (Zohner & Renner, 2017; Zohner et al., 2017; Renner & Zohner, 2019). This nutrient-centered perspective dovetails with both photoprotection and signaling hypotheses, as nutrient deficiencies both increase the need for photoprotection as a result of reduced photosynthetic capacity (Evans & Seemann, 1989; Martin et al., 2002; Hikosaka, 2004) and render the plant a lower-quality food source for herbivores (Ball et al., 2000; reviewed in Awmack & Leather, 2002). Cited as evidence against the photoprotection hypothesis, Pena-Novas & Archetti (2020a) reiterated Manetas's (2006) question of why plants would evolve anthocyanins as light-attenuating pigments if they 'absorb mainly green light, which is relatively harmless'. Agati et al. (2021) addressed the blue and red light-absorbing properties of anthocyanins in their Forum letter (also see Landi et al., 2021); here we defend the importance of green light to photosynthesis, and the value of a green light-attenuating pigment for photoprotection. First, while it is true that chlorophyll in solution absorbs green light less efficiently than red or blue wavelengths, whole-leaf absorptance of green wavelengths is not zero but, rather, closer to 70–75% (e.g. Hughes & Smith, 2007; Liu & van Iersel, 2021). This is because plants possess accessory pigments including carotenoids, which expand the photosynthetic action spectrum into blue-green wavelengths; also, absorption of blue and red wavelengths by the uppermost mesophyll cells renders green light the primary driver of photosynthesis in the lower mesophyll (Sun et al., 1998; Liu & van Iersel, 2021). A similar process to the latter occurs at larger structural scales in the forest, as the overstory preferentially absorbs red and blue light, leaving the understory enriched in green light (Smith et al., 2017). There is also evidence that plants can grow exclusively on green light (Johkan et al., 2012), that green light is used more efficiently than red and blue light at high photosynthetic photon flux densities (PPFDs) (Landi et al., 2020; Liu & van Iersel, 2021), that photosynthesis increases with increasing intensity of green light (Johkan et al., 2012), and that green light can induce photoinhibition of photosynthesis (Hughes et al., 2005; Zhang et al., 2010; Cooney et al., 2015). More pertinent to this discussion, however, is the question of whether attenuation of green photons by anthocyanins translates into photoprotection. This issue has been directly addressed by studies comparing changes in maximum quantum yield of photosystem II (or Fv/Fm), a popular metric for quantifying photo-oxidative stress, in tissues with or without anthocyanin, under green vs red light (which anthocyanins do vs do not absorb efficiently, respectively). Consistent with a photoprotective function of anthocyanins, these studies show more dramatic declines in Fv/Fm and/or total chlorophyll fluorescence in green than red tissues under green light, but similar declines under red light (Hughes et al., 2005, 2008; Zhang et al., 2010). However, as Pena-Novas & Archetti (2020a) emphasized, although studies such as these have demonstrated a photoprotective effect of anthocyanins, others have not. Accordingly, in the past decade, researchers have begun to systematically re-examine these seemingly contradictory cases, and have generally ruled in favor of the photoprotection hypothesis. For example, Logan et al. (2015) repeated Burger and Edwards' (1996) experiment using red- and green-leafed cultivars of Coleus. While Burger and Edwards reported no difference in quantum yield of O2 evolution following exposure to high light stress, a more detailed fluorescence and biochemical analysis by Logan et al. (2015) revealed that green-leafed varieties sustained similar photosynthetic efficiency by upregulating nonphotochemical quenching (NPQ, e.g. the xanthophyll cycle), an alternative form of photoprotection; under red light (which anthocyanins do not absorb efficiently), both varieties utilized NPQ to the same degree. This finding is consistent with those of other studies which concluded that anthocyanins serve as an alternative photoprotective mechanism to increased NPQ (Hughes et al., 2012; Moy et al., 2015; Ramírez-Valiente et al., 2015). Importantly, Logan et al. (2015) also demonstrated that this difference in NPQ does not necessarily translate into reductions in dark-adapted Fv/Fm or ΦPSII (quantum yield efficiency of photosystem II (PSII) in the light-adapted state). These were the primary metrics used by several studies cited by Pena-Novas & Archetti (2020a) as evidence against the photoprotection hypothesis, including Lee et al. (2003), Manetas et al. (2003) and Karageorgou & Manetas (2006). We note, however, that the findings of these latter two studies actually supported the photoprotection hypothesis, as both showed significantly lower Fv/Fm in green than in red leaves under white light (although this result was de-emphasized by Karageorgou & Manetas, 2006). Gould et al. (2018) also directly addressed conflicting reports of photoprotection by anthocyanins by teasing apart the effects of leaf age, temperature, duration and intensity of light source on Fv/Fm and xanthophyll cycle pigments in wild-type (WT), anthocyanin-deficient ttg1-1 and anthocyanin-rich pap1-D mutants of Arabidopsis. Their results showed that photoprotective effects of anthocyanins were observed only when PPFDs were both high-intensity (above saturating values) and sustained (> 2 h); similar results were reported for anthocyanic vs acyanic tissues of Colocasia esculenta, which exhibited no difference in Fv′/Fm′ or NPQ until after 2 h of sunlight exposure (Hughes et al., 2014). These results help to explain those of Kyparissis et al. (2007), who in fact noted that, 'Our [red and green] test plants displayed progressive reductions in PSII effective yield and gs, and increases in NPQ; however, their photoprotective capacity was not exceeded, as indicated by the absence of both a plateau in NPQ development and photoinhibitory damage, probably because the imposed stress was insufficient to drive the photoprotective potential of our test plants to their limits.' Similarly, in Esteban et al. (2008), leaves were exposed to relatively low-intensity PPFDs (300 μmol m−2 s−1) and for a short duration (30 min). The remaining study cited by Pena-Novas & Archetti (2020a) as contradicting the photoprotection hypothesis for leaf reddening (Hormaetxe et al., 2005) examined photoprotection by red carotenoids sequestered in chromoplasts of winter-red evergreens, and is therefore not relevant to the discussion of autumn leaf reddening (which involves vacuolar anthocyanins and chloroplasts/gerontoplasts). Another point made by Karageorgou & Manetas (2006) and evoked by Pena-Novas & Archetti (2020a) regarded the seemingly ill-positioned location of anthocyanins in the vacuole, relative to the source of photo-oxidative reactive oxygen species (ROS), presumably the chloroplast. It is our view that the position of anthocyanins in the vacuole is optimal for its function as a sunscreen, as the vacuole is the largest organelle in the plant cell, comprising up to 90% cytoplasmic volume (Tan et al., 2019). One large umbrella is more effective than many tiny umbrellas (i.e. chloroplasts), especially given that chloroplasts often migrate to vertical cell walls under high light stress (Davis & Hangarter, 2012), leaving underlying cells vulnerable when photoprotection is needed most. This photoprotective strategy is analogous to structural shading reported in the desert plant Retama raetam, which uses outer-canopy stems to shield photosynthetically active lower canopy stems during the dry season (Mittler et al., 2001). There is also ample evidence that anthocyanins can neutralize ROS from the vacuole. While some ROS (e.g. superoxide) cannot cross the vacuolar tonoplast in their original form (Takahashi & Asada, 1983), these molecules are short-lived, and are either rapidly protonated into OH˙ or converted by superoxide dismutase to H2O2, which can freely penetrate the tonoplast and enter the vacuole (Takahashi & Asada, 1983; Yamasaki et al., 1997). Accordingly, the vacuole also contains high concentrations of peroxidases (Zipor & Oren-Shamir, 2013; Zipor et al., 2015). Anthocyaninless tissues and mutants lacking anthocyanins have also been shown to accumulate more H2O2 and O2˙ compared with red tissues or WT plants (Gould et al., 2002; Kytridis & Manetas, 2006; Zhang et al., 2012), indicating that anthocyanins reduce ROS accumulation (Zhang et al., 2012). Manetas himself reported (in another 2006 paper) that 'vacuolar anthocyanins may be an effective in vivo target for oxy-radicals, provided that the oxy-radical source and the anthocyanic detoxifying sink are in close vicinity' (i.e. in mesophyll rather than epidermal cells; Kytridis & Manetas, 2006). Renner & Zohner (2020) criticized Pena-Novas & Archetti (2020a) for including 'green' leafed species in their analysis, claiming that 'any test of whether species from North America are more likely to turn red than non-American species during leaf senescence, must include only species that exhibit autumn leaf senescence, i.e. deciduous species'. Species defined as 'green' by Archetti (2009) included both evergreen species and deciduous species, which the authors observed to senesce green. Our suggestion is to continue to include evergreens, but to score autumn color in deciduous species based on their color during autumn in the tree's native range. Archetti (2009) scored autumn color based on personal observation in one geographic region, Gloucestershire, in the UK. Although the authors indicated that field guides were consulted in cases of conflicting observations, many deciduous species were nevertheless scored as 'green', despite exhibiting colorful autumn hues in their native locations (e.g. Betula nigra, Cornus canadensis and Quercus falcata, among many others). This may be a result of the relatively mild climate of the UK, or from abundant fertilization, which is known to delay the onset of leaf senescence (Fu et al., 2019). Fertilizing late in the season has even been reported to result in retention of leaves until killed by frost (Sakai & Larcher, 1987). Regardless, we feel strongly that the deciduous species marked 'green' by Archetti (2009) should be carefully re-evaluated to ensure the color scoring reflects that of their native locations, especially before they are applied to any additional analyses. We do, however, agree with Pena-Novas & Archetti (2020a) that evergreen species should be included in the overall study of leaf color change. However, we suggest that, in addition to scoring these species as 'evergreen', they could also be scored based on leaf color during senescence, whenever it occurs. Evergreen leaves often senesce with vibrant yellow and red hues (Fig. 1), although this tends to occur in spring and summer rather than the autumn (e.g. Poudyal et al., 2012), and receives far less study. There is also evidence that the color of young leaves of deciduous species in spring mirrors that of sensing leaves in autumn (Lev-Yadun et al., 2012), suggesting that selective pressures driving red leaf color are present during other seasons as well. A second reason we argue that evergreens should be included in the discussion of autumn leaf color is the possibility that plants with evergreen, red-deciduous and yellow-deciduous leaves could represent strategies along a continuous resource spectrum (Wright et al., 2004; Onoda et al., 2017). We discuss this possibility further in the next section. Lastly, while on the topic of scoring leaf color, Renner & Zohner (2019) stated: 'Following Archetti (2009), orange was grouped with yellow because both colours are caused by xanthophylls.' However, our own (unpublished) observations of autumn-orange leaf cross sections (as well as those of D. W. Lee, pers. comm.) are consistent with findings of Junker & Ensminger (2016), and suggest that orange color is imparted by low levels of anthocyanins, rather than increased carotenoids. Although not clearly communicated in their methods, Archetti (2009) also scored orange leaves as red (M. Archetti, pers. comm.). For these reasons, we suggest that future studies group orange leaves with red. The only abiotic environmental variables tested so far as potential selective pressures driving evolution of autumn leaf reddening have been related to climate – namely, high irradiance during autumn, cold temperatures and precipitation (Zohner & Renner, 2017; Zohner et al., 2017; Renner & Zohner, 2019; Pena-Novas & Archetti, 2020b). Yet, soil variables are strong, and in some cases stronger, predictors of leaf traits than climatic variables (Maire et al., 2015). The relative availability of plant soil nutrients varies by orders of magnitude over biogeographical gradients (Vitousek, 2004; Huston, 2012) and can be affected by several factors, including climate, topography, organic matter content, pH, soil particle size and parent material. Low soil fertility has long been known to favor evolution of evergreen over deciduous leaves (Monk, 1966; Chapin, 1980; Aerts, 1995). This trend has been supported in a broad range of biomes, including the temperate deciduous forest (citations in Givnish, 2002). We believe that an important clue to understanding why some deciduous species evolved to synthesize anthocyanins in autumn leaves lies in the fact that anthocyanin production in leaves is known to correlate with many of the same variables associated with evergreen leaves, for example, low leaf nitrogen (N), low maximum photosynthesis, and high leaf mass per unit area (LMA) (Bongue-Bartelsman & Phillips, 1995; Hodges & Nozzolillo, 1996; Kumar & Shanna, 1999; Kytridis et al., 2008; Peng et al., 2008; Nikiforou et al., 2011; Larbat et al., 2012; Carpenter et al., 2014; Meng et al., 2020), including during autumn (Lee et al., 2003; Schaberg et al., 2003; Anderson & Ryser, 2015). Studies in Arabidopsis have also demonstrated that anthocyanin production is part of the N limitation response, as is increased lignin production (which would contribute to increased LMA) (Peng et al., 2008). Increased LMA under N deficiency corresponds with greater leaf thickness and toughness, which are achieved via reallocation of leaf N to structural (rather than photosynthetic) proteins, cell wall thickening, denser cell packing, increased mesophyll layers, greater sclerophylly and increased allocation to major veins (Takashima et al., 2004; Wright et al., 2004; John et al., 2017). These characteristics maximize leaf longevity by making leaves more capable of withstanding environmental stress (Chabot & Hicks, 1982) and tougher for herbivores to consume and digest (Pérez-Harguindeguy et al., 2003; Poorter et al., 2009). Low-N leaves also often upregulate carbon-based chemical defenses (Stout et al., 1998; Karageorgou et al., 2008) and are fundamentally lower-quality food sources for herbivores (Ball et al., 2000; Awmack & Leather, 2002). From a physiological perspective, low leaf N also corresponds to reduced photosynthetic capacity per unit leaf mass (Amass) owing to reduction in photosynthetic enzymes (Evans & Seemann, 1989; Martin et al., 2002; Hikosaka, 2004) and reallocation of N to structural proteins (Takashima et al., 2004). Because low leaf N is associated with reduced nutrient quality, mechanically tougher leaves and reduced photosynthetic efficiency, anthocyanins could be simultaneously functioning in photoprotection and signaling low leaf quality. Perhaps autumn-leaf reddening reflects an evolutionary adaptation to environments that have moderately poor soils (i.e. poor, but not poor enough to warrant an evergreen life history)? This explanation is consistent with observations that senescing leaves of nitrogen-fixing species tend to be yellow, and are sometimes even retained green until killed by frost (Koike, 1990; Archetti, 2009). Leaf N, LMA, photosynthesis and leaf longevity represent the major axes of the leaf economics spectrum, which account for an estimated 75% of interspecific variation in key traits related to carbon fixation and N use (Wright et al., 2004; Onoda et al., 2017). We therefore feel that these trends in leaf characteristics could represent important clues and warrant further study. The greater number and proportion of red-deciduous species in eastern North America than in Europe (reported by Hoch et al., 2001; Renner & Zohner, 2019, 2020) have thus far been attributed to two primary factors: glacial extinction patterns (Lev-Yadun & Holopainen, 2009) and climate. Regarding the latter, forests of ENA feature more dramatic temperature fluctuations in autumn compared with Europe and East Asia (Zohner et al., 2017), higher September irradiance (Renner & Zohner, 2019) and shorter growing seasons (Zohner & Renner, 2017), which these authors argue supports a photoprotective hypothesis for autumn leaf reddening. In addition, Pena-Novas & Archetti (2020b) showed that the native ranges of red-deciduous species in North America are characterized by greater precipitation on average than green-leafed species, and that yellow-deciduous species tend to occur in colder areas on average than green-leafed species as well; however, their analysis could not resolve any factors that could account for divergence of red- vs yellow-deciduous species. We suggest that considering belowground factors (e.g. soil nutrient content, pH or even root structural limitations) could help to further resolve observed patterns. As a rough test of the hypothesis that low soil fertility might drive evolution of red autumn colors, we compared the following parameters for temperate deciduous forests of ENA, Europe and East Asia: pH, organic carbon, effective cation exchange capacity, total N and phosphorus. Consistent with our hypothesis, soils of deciduous forests in ENA were significantly less fertile in all metrics examined compared with Europe, with East Asia exhibiting a mixed range of values (Table 1). Average total soil N in Europe, for example, was nearly double the averages measured for ENA and East Asia (Fig. 2). Although the substantial land-use change that has occurred over previous centuries has no doubt altered these values from what they would be under natural conditions, they are nevertheless interesting to consider (assuming one agrees with the original premise that there are more red-leafed species in ENA than in Europe). These data can also be useful for answering unresolved questions. For example, Renner & Zohner (2020) questioned how the coevolution hypothesis could possibly explain the greater occurrence of autumn leaf reddening in ENA than in Europe; our data suggest that reduced nutrient availability should render ENA trees lower-quality hosts to parasites, resulting in more intense signaling. Similarly, the yellow-deciduous trees of northern Europe and Scandinavia noted by Holopainen & Peltonen (2002) may not warrant anthocyanins in autumn as a result of high soil N (Fig. 2), despite photoinhibitory conditions during autumn. Finally, it is possible that shorter growing seasons in ENA than in Europe reported by Zohner & Renner (2017) could be related to soil fertility, because, as previously mentioned, nutrient stress advances leaf senescence dates (Anderson & Ryser, 2015; Fu et al., 2019). If photoprotection by anthocyanins functions to protect nutrient translocation during autumn, should red-deciduous species necessarily exhibit enhanced nutrient resorption efficiency relative to yellow? It is our opinion that, if anthocyanins function as an alternative photoprotective strategy to increased NPQ (as suggested in Hughes et al., 2012; Moy et al., 2015; Ramírez-Valiente et al., 2015), then similar (rather than superior) nutrient translocation in red- vs yellow-deciduous species should serve as adequate support for a photoprotective hypothesis, especially if nutrient translocation diminishes in anthocyanin-deficient mutants (as shown by Hoch et al., 2003). A separate question is whether nutrient-deficient individuals (or species adapted to chronic nutrient deficiency) should exhibit greater resorption efficiency than those with more optimal nutrient status. While some authors have failed to demonstrate a relationship between leaf nutrient status and resorption efficiency (e.g. Aerts, 1996), Vergutz et al. (2012) showed a strong global, inverse relationship between nutrient status and resorption efficiency for C, N, P, K, and Mg, but only after correcting for leaf mass loss during senescence. Hence, if autumn leaf reddening has indeed evolved in response to low leaf nutrients, then red-deciduous species might exhibit more efficient nutrient resorption than yellow species, but not directly because of anthocyanins per se. Unfortunately, none of the studies comparing nutrient resorption efficiencies of red- and yellow-deciduous species properly controlled for leaf mass loss during senescence (Hoch et al., 2003; Duan et al., 2014; Pena-Novas & Archetti, 2021a). We therefore recommend that future studies include leaf mass correction factors in their resorption efficiency calculations (see Wang et al., 2020 for suggested approaches). Furthermore, these studies also tend to occur in common gardens, where soil nutrients are presumably equal for both red- and yellow-deciduous species; this too could potentially mask differences. As recently articulated by Pena-Novas & Archetti (2021b), 'The question is not just why autumn colours evolved, but why they evolved only in some species.' Proponents of the signaling hypothesis argue that species with red autumn leaves evolved in regions where selection pressure by insect parasites seeking out autumn host trees is high, while physiologists argue that red-leafed species are selected for in temperate regions with especially cold and bright autumns, and, we suggest here, lower-quality soils. But if other photoprotective strategies exist that plants may employ to mitigate photo-oxidative stress besides anthocyanins (e.g. NPQ), what selective pressures might drive a species to use anthocyanins over these other mechanisms? Unfortunately, as repeatedly emphasized by Pena-Novas and Archetti (2020a, 2021b), few physiological studies directly address this question. Nevertheless, we offer a tentative answer here, based on the limited studies available. First, Hoch et al. (2003) showed that acyanic mutants of three red-deciduous species resorbed significantly less N than wild-type (WT) conspecifics, suggesting that anthocyanins are part of an integrated, photoprotective strategy which NPQ cannot compensate for once anthocyanins have been removed. More recently, Moy et al. (2015) explored differences in photoprotection and photosynthetic protein catabolism during leaf senescence in a red-deciduous maple (Acer saccharum) and yellow-deciduous oak (Quercus bicolor). Their results showed that the red-deciduous maple began mobilizing chlorophyll and proteins associated with PSI and light-harvesting complexes (e.g. LHCa) earlier in the season (by nearly a month) compared with the yellow-deciduous oak, and also exhibited smaller pools of photoprotective xanthophyll cycle pigments and PsbS proteins in the month before senescence, consistent with a greater reliance on anthocyanins for photoprotection than NPQ. It is possible that the early dismantling of photosynthetic structures within the chloroplast (i.e. the site of NPQ) render this mechanism of energy dissipation less effective than vacuolar anthocyanins. Consistent with this explanation, Schaberg et al. (2003) showed that sugar maples (Acer saccharum) with low leaf N in June senesced earlier and produced redder autumn color, compared with individuals with elevated nutrient status. Perhaps evolving in low-nutrient soils favors a senescence strategy characterized by early remobilization of leaf nutrients, including those used in photosynthesis, rendering vacuolar anthocyanins more appropriate for photoprotection than NPQ. A larger study comparing the timing of chlorophyll breakdown and concerted photoprotective strategies for red- vs yellow-deciduous species is needed to test this hypothesis further. One strength of the nutrient-limitation hypothesis is that it neatly accounts for both proximate and ultimate causation of autumn leaf reddening. A comprehensive study of factors influencing timing and intensity of red coloration in sugar maple by Schaberg et al. (2003) showed that the strongest predictors of earlier and more intense autumn leaf reddening were low leaf N and high starch during the growing season (and high soluble sugars in autumn, accordingly). It is known that N deficiency promotes starch accumulation in leaves during the growing season (Marschner, 1995) as a result of limitations in metabolism associated with reduced growth (Paul & Foyer, 2001), and that starch is catabolized into soluble sugars in the autumn (Schaberg et al., 2003). The synergistic effects of high soluble sugars (especially sucrose) and low N on anthocyanin production in the presence of light have been demonstrated in numerous plant systems (e.g. Lea et al., 2007; Boussadia et al., 2010; Su et al., 2016; Zhou et al., 2020), and appear to result from upregulation of multiple independent transcription factors related to anthocyanin synthesis (e.g. PAP1 and PAP2), which together result in greater anthocyanin synthesis than individual stimuli alone (Lea et al., 2007). Hence, we propose that low leaf N and high soluble sugars may serve as the proximate causes for autumn leaf reddening, and that the ultimate function is both signaling and photoprotection. We would like to thank Dr Marco Landi and reviewers for their helpful comments on this manuscript. NMH, CBG and COG planned and designed the research; COG performed GIS measurements; NMH analyzed the data; and NMH, CBG and HSN wrote the manuscript.