III. Long-term effects of light on plant–herbivore or plant–pathogen interactions 679
IV. Mechanisms of responses to the light environment: the whole-plant perspective 684
V. Short-term responses to the light environment: induced defences 686
VI. Mechanisms for light-dependent induced defences 688
VII. Interpreting interactions between light and defence responses 690
Plants frequently suffer attack from herbivores and microbial pathogens, and have evolved a complex array of defence mechanisms to resist defoliation and disease. These include both preformed defences, ranging from structural features to stores of toxic secondary metabolites, and inducible defences, which are activated only after an attack is detected. It is well known that plant defences against pests and pathogens are commonly affected by environmental conditions, but the mechanisms by which responses to the biotic and abiotic environments interact are only poorly understood. In this review, we consider the impact of light on plant defence, in terms of both plant life histories and rapid scale molecular responses to biotic attack. We bring together evidence that illustrates that light not only modulates defence responses via its influence on biochemistry and plant development but, in some cases, is essential for the development of resistance. We suggest that the interaction between the light environment and plant defence is multifaceted, and extends across different temporal and biological scales.
Light is fundamental to the existence of plants. It affects all aspects of growth and development, as a primary requirement for plant fitness is optimization of light harvesting for photoautotrophic growth. Hence, photoreceptors such as phytochromes and cryptochromes sense quantitative and qualitative features of the light environment and, via associated signal transduction pathways, regulate plant physiology and development. Light, through photosynthesis, also controls much of the biochemical activity within plant tissues, something that is reflected by the fact that a wide array of genes are transcriptionally regulated by the circadian clock in Arabidopsis thaliana (Harmer et al., 2000). By sensing day length, plants also use light as a seasonal indicator which controls the transition to reproductive growth in many plant species. Although essential, light can also pose problems for plants. Increased doses of UV light can cause damage at the molecular level, and even simple changes in ambient sunlight can overload photosynthetic electron transport (PET), causing damaging reactive oxygen species (ROS) to accumulate. Plants have evolved many systems to minimize the impacts of such deleterious effects of light, including the production of photoprotective pigments, biochemical systems to rapidly modulate chloroplast electron transport, physiological responses such as the ability to re-orient chloroplasts, and photomorphogenic responses that optimize the interaction of the leaves with light over longer time scales.
As well as direct effects on plant metabolism, growth and development, light inevitably influences many other plant responses to the environment. These include defences against pests and pathogens. There is a wide range of information in the scientific literature on the effects of light on defence responses, ranging from ecological- to molecular-scale investigations of both short- and long-term responses. Our aim here is to draw some general conclusions about the impact of light on plant defence, and to attempt to suggest conceptual models that explain the observations in terms of both the molecular and the ecophysiological responses to light and biotic attack.
II. Light as an environmental variable
Light is an extremely dynamic component of the terrestrial environment. Changes are both quantitative (including variation in instantaneous irradiance, dose accumulated over time, and day length) and qualitative (in terms of light spectral balance). Plants and their associated herbivores and pathogens may respond to each of these different components of variation.
1. Variation in the quantity of light
The quantity of light falling on a surface at a given moment, usually referred to as ‘light intensity’, is formally defined in terms of either energy per unit area (= irradiance) or quanta per unit area (= photon flux; see Bjorn, 2002). Some elements of the variation in irradiance are predictable; for example, variation with time of day, season and latitude are all functions of the elevation of the sun in the sky (the higher the solar elevation, the higher the irradiance). As a result, irradiance reaches a maximum near the equator, at midday, and, at mid to high latitudes, in midsummer. Superimposed on these systematic geographical and seasonal variations in irradiance are variations resulting from factors such as cloud cover, aspect on a sloping site, and shade from nearby structures or plant canopies (Bjorn, 2002). Some of these factors affect all wavelengths of light more or less equally, whereas others are much more wavelength specific (section 11–2).
Many biological responses to light can be described as simple functions of irradiance. The rate of photosynthesis in plants is a typical example. Although photosynthesis is a function of irradiance, growth is determined by the sum of photosynthetic carbon fixation over time which is, in turn, a function of the amount of light received by the plant over that period. Thus, growth and yield, and many other long-term effects of light, are best described by the accumulated dose of photosynthetic radiation, for example by daily light integral (Kitaya et al., 1998; Korczynski et al., 2002; Dielen et al., 2004). Light damage is also often a function of accumulated dose, as with many whole-plant responses to UV radiation (Gonzalez et al., 1998; de la Rosa et al., 2001).
2. Variation in the quality of light
Light quality is the balance between different wavelengths. Different organisms perceive different wavelengths in different ways. The three primary colours of human vision define ‘visible’ light (approx. 400–700 nm), but other animals, including many invertebrate and some vertebrate herbivores, may perceive different wavebands, notably in the ultraviolet region (primarily UV-A: 315–400 nm). Thus, what is actually perceived as ‘visible light’ varies substantially amongst species. Photosynthetically active radiation (PAR) is usually defined as 400–700 nm, but plants also detect and utilize different wavelengths as environmental cues. Responses to red (R) and far-red (FR) (detected by phytochromes), blue and UV-A (detected by cryptochromes, phototropins and related photoreceptors) are well defined (Gyula et al., 2003; Spalding & Folta, 2005). The mechanistic basis for responses to UV-B (280–315 m) remains poorly defined: some responses may be a function of damage to DNA and other biological molecules, but there is also evidence for a specific UV-B photoreceptor (Jenkins et al., 2001).
The spectral balance of sunlight in the field is influenced by a range of factors. Temporal changes in the ratio of UV to longer wavelengths are largely driven by the increase in the ratio at high solar elevations. At temperate and high latitudes, sunlight is relatively enriched in UV, especially UV-B, in summer compared with winter. Similarly, the ratio of UV to PAR is highest near noon. There is some enrichment of FR relative to R at twilight (Salisbury, 1981). Spatially, UV:PAR ratios are typically higher at low latitudes. Cloud typically reduces all wavelengths of sunlight but shorter wavelength UV less than PAR, resulting in some increase in the UV:PAR ratio under cloud conditions (Calbo et al., 2005). Shade from plant canopies has major effects on spectral balance, notably in terms of R:FR (Ballaré, 1999; Gyula et al., 2003; Vandenbussche et al., 2005), but also in terms of the ratio of UV:PAR (Grant & Heisler, 2001; Heisler et al., 2003; Grant et al., 2005).
III. Long-term effects of light on plant–herbivore or plant–pathogen interactions
Studies of the effects of both shade and diurnal variation in light on plant interactions with their natural enemies deal mostly with herbivores; effects on disease remain relatively poorly understood. Studies of herbivory (Table 1) have mostly been in the context of variation in the light environment caused by plant canopies, such as the effect of position relative to neighbours, including gaps in woodland canopies, or woodland edge (individuals within gaps or at the forest edge receiving more sunlight than those within). Studies of woody plants have also considered the influence of vertical position in the canopy (foliage at or near the top of the canopy receiving greater insolation than that low down in the canopy), and the direction in which foliage faces (in the northern hemisphere, south-facing foliage receives higher irradiances than north-facing foliage). Experiments have either used these natural variations in light environment (for example taking foliage from, or placing plants in, different locations) or artificially manipulated light using shade cloth, etc. (Table 1). In some cases, the latter experiments have related to the use of shading as a tool in crop production. Of course, shading results in complex changes in the light environment, both quantitative and qualitative, which can differ depending on the source of shade. Thus, although some studies implicitly assume that shade influences plant–herbivore interactions through changes in photosynthesis driven simply by the reduction in PAR, there may be independent effects of altered spectral balance in shade (R:FR or UV:PAR; see previous section). Artificial shade treatments do not necessarily reproduce these spectral changes.
Table 1. Overview of field experiments on the effects of the light environment on (a) plant–herbivore interactions and (b) plant–pathogen interactions
These studies considered the effects of variation in total light, and in some cases responses have been attributed not just to photosynthetic radiation but to the longer wavelengths of sunlight, resulting in changes in the thermal environment. The potential role of UV wavelengths was not considered in these studies.
+ve, shade increases the leaf area eaten by a herbivore or infected by a pathogen, or has some beneficial effect on herbivore performance or behaviour (e.g. reduced mortality, increased growth rate or increased efficiency of food conversion); –ve, negative responses; 0, shade treatments had no significant effect; na, not assessed.
In the field, shade will also influence overall radiation balance with possible consequences for the abiotic environment of the host, the herbivore and potentially other organisms such as parasitoids or predators of the herbivore. Temperatures of the air and of organisms are typically lower in the shade, with direct effects on a wide range of processes, and indirect effects such as altered water balance, which may result in reduced plant water deficits compared with full sunlight. Equally, ‘shade’ in the field may alter the biotic environment through mechanisms unrelated to any effect on solar radiation. For example, there is a substantial literature on the role of tree canopies in determining the number and species richness of the community of insectivorous birds that can have a major influence on herbivory (Marquis & Whelan, 1994; Strong et al., 2000; Van Bael & Brawn, 2005). Certainly the effects of tree canopies on herbivory in crops such as coffee (Coffea sp.) can be interpreted in relation to greater predation by birds, not changes in the light environment (Perfecto et al., 2004). Such effects highlight the complexity of shade as an environmental variable. Clearly, many of the same arguments can be made in relation to comparisons between day and night, which differ in far more than simply the light environment. While these broader mechanisms are largely beyond the scope of this review, they provide an important context for any assessment of light-mediated changes.
1. Light and herbivory
Day/night Diurnal variation in herbivory has been viewed primarily as a function of the biology of the herbivore rather than the host. The general expectation is that most invertebrate herbivores are less active during the day than at night, at least partly because the risk of predation or parasitism is greater during the day (Hassell & Southwood, 1978). However, there are many exceptions to this pattern (Springate & Basset, 1996; VanLaerhoven et al., 2003). For example, Novotny et al., (1999) reported a three times greater risk of predation during the day compared with night, yet herbivores were more abundant during daylight. Some insect herbivores feed almost exclusively during the day (Kreuger & Potter, 2001), with the temperature dependence of behaviour perhaps being a key driver. One host characteristic that shows diurnal variation and which might influence both herbivores and higher trophic levels is the emission of volatiles. There are quantitative and qualitative differences in wound-induced volatiles between day and night (De Moraes et al., 2001; Gouinguené & Turlings, 2002; Martin et al., 2003; see also section V). Herbivores, and their parasites and predators, are able to detect and respond to such changes, and diurnal variation in the volatile signal may result in differential effects on different herbivores (De Moraes et al., 2001) as well as higher trophic levels (Maeda et al., 2000). However, some predators appear capable of isolating key information against this highly variable volatile signal (Meiners et al., 2003).
Shade The general hypothesis that herbivory would be suppressed in plants grown in full sun compared with those in shade has been shown to be correct in many systems, especially, but not exclusively, for leaf-chewing insects (Table 1a). This is true at least in the sense that, when consumed, leaf tissue from plants grown in shade is more favourable to herbivore growth and/or development. However, plants grown in full sunlight may suffer an increase in the leaf area consumed compared with shade-grown plants (Table 1a). This increased consumption may be a function of reduced food quality in full light, as insects often compensate for low food quality by increasing intake (Slansky & Wheeler, 1992). However, this mechanism may not fully explain increased consumption of high-light tissue, as preferences can persist even when extracts of sun- or shade-grown leaves are incorporated into artificial diets (Panzuto et al., 2001). These plant-mediated changes interact with herbivore responses. For example, adults of some insect herbivores may prefer high-light locations for certain activities, such as egg laying (Alonso, 1997). There are clear examples of such direct herbivore responses outweighing higher host quality of shade-grown plants (Sipura & Tahvanainen, 2000). Interestingly, an example where herbivore damage is more severe in plants grown at higher light is one of the few examples where light-dependent variation in herbivory has been proved to have significant effects on host population dynamics (Louda & Rodman, 1996). In that study, the native crucifer Cardamine cordifolia suffered significantly greater herbivory when natural shade was removed. Some components of host resistance were reduced in full sun, but many were increased, and some of these changes appeared to be related to the mild water deficits that occurred in plants growing in full sun. Insects were also more abundant in the sun. Overall, changes in herbivory were attributed to the combined effects of changes in host defence (with responses perhaps being partly to light and partly to water deficits) and herbivore abundance (Louda & Rodman, 1996).
Although canopy shade may have slightly different effects on PAR and UV wavelengths (Grant & Heisler, 2001; Heisler et al., 2003; Grant et al., 2005), in broad terms the two are highly correlated across natural gradients between sun and shade. Thus, the great majority of research into the effects of shade on herbivory will have manipulated both PAR and UV, although the possible role of the UV component of sunlight is generally ignored in interpreting results. The growing literature on the specific effects of UV wavelengths on plant–herbivore interactions demonstrates that variation in UV, or at least UV-B, can be significant in many systems (Table 2). Indeed, it has been suggested that cyclical variation in the population of both vertebrate and invertebrate herbivores may be driven by the effects of natural variation in solar UV-B on host defensive chemistry (Selas et al., 2004). The experimental manipulation of UV-B alone results in changes in plant–herbivore interactions that show many parallels with those seen with broad-spectrum shading. In most studies, foliage from reduced UV-B environments is generally found to be a higher quality resource for herbivores than foliage from unfiltered sunlight in terms of herbivore mortality, growth rates or the efficiency of food utilization (Table 2). In the field, defoliation caused by herbivory is often increased when ambient solar UV-B is reduced using wavelength-specific filters (Table 2). However, as with ‘total shade’ treatments, both laboratory and field studies show that these UV effects vary amongst host species, and perhaps genotypes, and also amongst herbivores (Table 2). The mechanisms by which exposure to UV could directly affect insect herbivores remain rather unclear, although the visual systems of many insects perceive longer wavelength UV. The consequent disruption of foraging and dispersal in UV-deficient conditions can be significant both in experimental studies (Mazza et al., 1999) and in the commercial application of UV-opaque plastics for the control of horticultural pests such as thrips and whiteflies (reviewed by Raviv & Antignus, 2004). In the field, UV might also influence herbivore populations through the suppression of entomopathogens, whether nematodes (Fujiie & Yokoyama, 1998), fungi (Braga et al., 2001, 2002), bacteria (Myasnik et al., 2001) or viruses (Shapiro & Domek, 2002).
Table 2. Overview of the effects of ultraviolet radiation on plant–herbivore interactions
Ipomoea batata/Bemisia tabaci, Frankliniella occidentalis or Aphis gossypii
Polythene tunnels with ambient or attenuated total solar UV
Substantial reductions in attack by all three insects
Zea mays/Ostrinia nubilalis
+ or – UV in the glasshouse
Larvae preferred leaves grown without UV-B
Oryza sativa/Helicoverpa armigera
Artificial UV-B irradiation
Extracts of irradiated leaves had anti-feedant, growth- inhibitory and antibiotic properties against larvae, and effects persisted in adults, which laid fewer, less viable eggs
Bemisia argentifolii/Frankliniella occidentalis
Polythene tunnels with ambient or attenuatedtotal solar UV
Insects dispersed preferentially into ambient UV environments, but UV had no effect on flight ability
Pisum sativum/Autographa gamma
CE room with a range of UV-B doses
Increased UV-B increased leaf nitrogen and when foliage was fed to larvae this was correlated with an increase in larval growth rate and a reduction in the amount of plant material consumed
Trifolium repens/Spodoptera litura or Graphania mutans
CE room with and without UV-B
36% reduction in weight of S. litura on foliage grown at high UV, but this depended on host genotype. G. mutans showed little response
Glycine max/Caliothrips phaseoli
Ambient or near-zero UV-B in the field
UV-B reduced thrip herbivory: insects preferred leaves from reduced UV-B and avoided solar UV
Ambient or near-zero UV-B in the field
Insects preferred low UV-B environment
Lolium perenne and Festuca spp./Schistocerca gregaria
Ambient and elevated UV-A or UV-B in the field
No herbivore responses to excised leaves from different UV-B treatments except in F. pratensis where responses varied with UV treatment and/or endophyte infection of the host
Plantago lanceolata/Precis coenia or Trichoplusia ni
CE room at high ambient UV-B or above
Growth of T. ni larvae was faster when they were fed excised leaves from plants treated with elevated UV-B. Direct exposure of larvae to the UV treatments increased mortality of T. ni. UV had no significant effects on P. coenia
Polythene tunnels with ambient or attenuated total solar UV
Attenuation of UV reduced whitefly dispersion, resulting in reduced populations in low-UV tunnels
Quercus robur/natural herbivore community
Ambient and elevated UV-A or UV-B in the field
Plants under elevated UV-B or UV-A suffered greater herbivory
Gunnera magellanica/natural herbivore community
Ambient or near-zero UV-B in the field
Leaf area damaged increased under reduced UV-B
Gunnera magellanica/natural herbivore community
Ambient or near-zero UV-B in the field
Leaf area consumed increased by 25–75% under attenuated UV-B
Nothofagus antarctica/natural herbivore community
Ambient or near-zero UV-B in the field, and sun-exposed and shaded branches
Solar UV-B reduced insect damage by at least 30%, and this occurred with foliage in both sunny and shaded positions
Calluna vulgaris/Strophingia ericae (Homoptera)
Ambient and elevated UV-B in the field
Increased UV-B reduced herbivore population density over two seasons
Salix myrsinifolia and S. phylicifolia/Phratora vitellinae or natural herbivore community
Ambient and elevated UV-B in the field
Herbivores were more abundant under elevated UV-B but host did not suffer greater herbivore damage. Excised leaves of S. phylicifolia from elevated UV-B treatment reduced growth of P. vitellinae larvae compared with control leaves, but there was no comparable effect with leaves of S. myrsinifolia
Populus trichocarpa/Chrysomela scripta
Zero, ambient and 2× ambient
Leaves from highest UV-B treatment significantly reduced larval consumption efficiency
Six plant species/Deroceras reticulatum (Mollusca)
Ambient or near-zero UV-B in the field
Significant effects in two of the six species. In Nothofagus antarctica, leaf area consumed reduced by two-thirds in foliage from under near-ambient UV-B. In Carex decidua, twice as much leaf area was consumed in reduced UV-B radiation
Glycine max/Anticarsia emmatalis or natural herbivore
Ambient or near-zero UV-B in the field
Leaves from reduced UV-B treatment were more attractive to larvae, and supported higher community growth rates and lower mortality. No direct effect of UV exposure on larval mortality. Attentuation of UV increased natural herbivore damage by 2-fold
Morus nigra/Bombyx mori
Artificial UV irradiation in CE rooms
UV treatments reduced consumption of foliage by larvae
The extent to which reductions in solar UV contribute to the overall effects of shade on plant–herbivore interactions remains unclear. So far as we are aware, the only study to explicitly consider the effects of both UV and shade is that of Rousseaux et al. (2004), who studied herbivory of Nothofagus antarctica. Both the number of sites attacked and the area of leaf removed by insect herbivores were reduced on the sun-exposed side of the canopy. This response occurred even when UV-opaque filters removed the UV-B component of sunlight. However, removing UV-B significantly reduced leaf area removed on both sun-exposed and shaded sites. These data suggest that the effects of UV-B and those of other components of natural shade can act independently, a contention that is supported by chemical changes induced (section IV–3).
2. Light and disease
Day/night Whilst defoliation by many herbivores is sufficiently rapid to allow differentiation of damage occurring during the day from that occurring at night, disease is a longer term process. Thus, it is not surprising that, so far as we are aware, investigations of diurnal changes in plant–pathogen interactions have dealt with specific aspects, such as sporulation, spore dispersal or infection. The concentration of airborne spores in and around plant canopies is far higher at night than during the day in a wide range of fungi (Schmale & Bergstrom, 2004; Gilbert, 2005; Zhang et al., 2005). However, in other fungal pathogens, spore concentrations peak during the day (Gadoury et al., 1998; Su et al., 2000) or show more complex diurnal patterns (Hock et al., 1995). These processes in plant–pathogen interactions may be influenced by the lower temperature, higher humidity or presence of leaf surface water from dew occurring at night and, as with herbivory, it is not always clear what role is played by the direct effects of light. However, there is clear evidence that spore release is initiated by light in some systems (Gadoury et al., 1998; Su et al., 2000). Light also directly inhibits spore germination and/or germ tube growth in many plant pathogenic fungi (Elison et al., 1992; Joseph & Hering, 1997; Tapsoba & Wilson, 1997; Mueller & Buck, 2003; Beyer et al., 2004), and this is certainly the case for UV (Paul, 2000). Overall, it is probably the case that plants are subject to greater challenge by many pathogens at night than during the day, but this is certainly not the case for all pathogens.
Shade The influence of shade on plant–pathogen interactions has been much less extensively studied than comparable effects on plant–herbivore interactions. However, a number of studies of noncrop systems have shown that shade increases infection by a range of pathogens (Table 1b). As with herbivory, there are exceptions to the usual expectation that disease is more severe in the shade, as seen with coffee rust (Hemileia vastatrix) (Soto-Pinto et al., 2002), anthracnose (Colletotrichum gloeosporioides) of Euonymus fortunei (Ningen et al., 2005) and powdery mildew (Microsphaera alphitoides) on oak (Quercus petraea) (Kelly, 2002).
For the most part, the mechanisms by which shade influences plant–pathogen interactions remain poorly understood, although plant pathologists have often attributed the effects of shade to factors such as humidity and leaf surface wetness, which are clearly central to the biology of many plant pathogens (Jarosz & Levy, 1988; Meijer & Leuchtmann, 2000; Koh et al., 2003). However, a number of studies have shown that infection by a range of pathogens can be affected by the light environment of the host before inoculation. While wheat (Triticum aestivum) seedlings exposed to low light intensity were more susceptible to subsequent inoculation by Puccinia striiformis than dark-grown seedlings (de Vallavieille-Pope et al., 2002), in other cases infection was found to be inversely proportional to preinoculation irradiances (Zhang et al., 1995; Shafia et al., 2001). This indicates direct effects of light on host resistance. Furthermore, Pennypacker (2000), showed that reduced light led to increased infection by Sclerotinia sclerotiorum in soybean (Glycine max), and Verticillium alboatrum but not Fusarium oxysporum in alfalfa (Medicago sativa). This was linked to host resistance mechanisms, as the effects of shade in both crops only occurred in resistant genotypes where resistance was quantitative (requiring a large investment of resources) rather than qualitative (based on the hypersensitive response, requiring a smaller investment of energy) (Pennypacker, 2000). These conclusions parallel much thinking concerning herbivore resistance (section IV).
Light quality as well as light quantity can affect disease. Red light suppressed powdery mildew of cucumber (Cucumis sativa), and the effect appeared to be reversed by far-red light (Schuerger & Brown, 1997). There are also indications that host resistance may be induced by preinoculation exposure to red light (Islam et al., 1998; Rahman et al., 2002; Khanam et al., 2005). Pathogenic fungi may respond directly to spectral balance, and this response is exploited in the use of plastic films which modify spectral balance as a component of disease control in horticulture. Films that transmit more blue light than longer or shorter wavelengths can be used to suppress sporulation in downy mildews and Botrytis cinerea (Reuveni & Raviv, 1992, 1997). Similarly, many plant pathogens use UV radiation as a cue to regulate sporulation, and films opaque to UV radiation can be used to reduce a wide range of crop diseases (reviewed by Raviv & Antignus, 2004). However, manipulating UV has complex effects on pathosystems. While UV-A may stimulate sporulation, exposed fungal tissues can be vulnerable to UV-B radiation, and solar UV-B is a major constraint on the spore survival of many pathogens (Paul, 2000). The effects of reduced UV-B may be sufficient to explain the overall increase in disease in shade (Gunasekera et al., 1997) or variation in cloud cover (Paul, 2000; Wu et al., 2000). Equally, prior exposure to UV can affect various components of host resistance. Exposure of the host before inoculation reduced subsequent infection in a range of pathosystems, but there are exceptions (reviewed by Gunasekera et al., 1997; Paul et al., 2000). Increases in infection with increased UV-B have sometimes been attributed to host injury providing sites for colonization by necrotrophic pathogens (Manning & von Tiedemann, 1995), but it is now recognised that this mechanism is probably confined to UV doses well above the ambient range (Paul, 2000). Contrasting responses amongst pathosystems are certainly not explained simply on the basis of biotrophic and necrotophic pathogens. Powdery mildews (Erysiphales) are biotrophic pathogens that grow on leaf surfaces exposed to incident radiation. There are several reports of UV-B exposure reducing powdery mildew infections, both in the laboratory (Willocquet et al., 1996; Paul, 1997) and in the field (Keller et al., 2003). However, exposure to increased UV-B led to increased powdery mildew (Microsphaera alphitoides) in oak (Newsham et al., 2000), which is consistent with the greater occurrence of this disease in open sites (Kelly, 2002). Overall, the contribution of UV to shade effects on plant–pathogen interactions is likely to be a function of interactions between the relative effects of UV-A and UV-B on direct damage and spore induction in the pathogen, and host resistance mechanisms.
IV. Mechanisms of responses to the light environment: the whole-plant perspective
As already discussed, the literature on the whole-plant biology or ecology of the influence of light on plant–herbivore or plant–pathogen interactions is diverse. Responses are attributed to a wide range of possible underlying mechanisms in the responses to light not only of the host plant, but also of the herbivore or pathogen, or higher trophic levels. Responses may also be associated with other environmental factors correlated with the light environment, rather than light per se. With this broad view of underlying mechanisms of response, light-mediated changes in the host plant are viewed as just one component of many. Furthermore, agronomists, and especially ecologists, consider a wide range of host characteristics as being significant in determining the overall effects of light on herbivory or disease. Chemical traits influencing herbivory include tissue nitrogen (N) chemistry [e.g. total N concentration, carbon:nitrogen (C:N) ratio, or protein or amino acid concentration], carbohydrate composition (total carbohydrates or components such as the soluble fraction), and water content. Aspects of morphology and physical properties such as leaf thickness, toughness and the possession of thorns or spines can also be significant for plant–herbivore interactions. In addition, the increase in specific leaf area with increasing shade that is commonly observed across a range of species (e.g. Crotser et al., 2003; Curt et al., 2005; Poorter et al., 2006) not only influences leaf physical properties but also may change how herbivores respond to chemical defence by changing the relationship between chemical contents and leaf area or biomass. Changes in host resistance, whether constitutive or induced by attack, certainly play an important role in coupling herbivory or disease to the light environment, but this is certainly not the only significant mechanism.
1. Host quality as a food resource for herbivores or pathogens
Many ecological studies of the mechanisms by which light influences herbivory (there is little comparable research on pathogens) have been conducted in the context of alternative theories of plant defence, such as the resource availability hypothesis (Coley et al., 1985), the growth differentiation balance (GDB) hypothesis (Herms & Mattson, 1992) and the carbon nutrient balance (CNB) hypothesis (Bryant et al., 1983). These hypotheses share in common the principle that plant allocation to defence is a function of competition between end-points (growth, storage and defence) for limited resources, such as photosynthate. A meta-analysis of almost 150 published experimental tests of CNB in woody species (Koricheva et al., 1998) revealed that the basic prediction of the hypothesis that shading would reduce concentrations of ‘carbon-based defensive chemicals’ (CBDCs) was broadly correct. Indeed, shading appeared to have a far stronger influence on such compounds than N supply, which the CNB hypothesis predicts will be inversely related to defence (Koricheva et al., 1998). When CBDCs were divided into three subgroups, phenylpropanoids, hydrolysable tannins and terpenoids, all three were reduced by shading, with phenylpropanoids showing the greatest response (Koricheva et al., 1998). More recent research confirms that shading reduces concentrations of CBDCs, in herbaceous as well as woody species (Jansen & Stamp, 1997; Crone & Jones, 1999; Hemming & Lindroth, 1999; Rowe & Potter, 2000; Tattini et al., 2000; Briskin & Gawienowski, 2001; Henriksson et al., 2003). In addition, it is now clear that shading may reduce concentrations of a wide range of secondary metabolites, not only of CBDCs, which have been the primary focus of studies associated with testing the CNB hypothesis. Shade reduced cyanogenic glycosides but not CBDCs in Eucalyptus cladocalyx (Burns et al., 2002), while in Prunus turneriana shade resulted in a change in the distribution of cyanogenic glycosides between older and younger leaves (Miller et al., 2004). However, shading did not affect the concentration of defensive amides in Piper cenocladum (Dyer et al., 2004). Exposure to UV-B increased cyanogenic alkaloids in some genotypes of Trifolium repens (Lindroth et al., 2000) and the effects of UV-B on plant phenolics are now very well established, and are not related to the ideas of resource limitation inherent in the CNB hypothesis. In general, increased exposure to UV-B results in increased concentrations of total phenolics (Bassman, 2004), although there are exceptions (Rousseaux et al., 1998; Salt et al., 1998; Levizou & Manetas, 2001). Specific phenolic compounds may show contrasting responses to UV-B, with flavonoids showing particularly consistent increases (Lavola et al., 1998; Tegelberg & Julkunen-Tiitto, 2001; Warren et al., 2002b; Lavola et al., 2003; Tegelberg et al., 2003; Warren et al., 2003; Rousseaux et al., 2004), with well-established dose–responses in some cases (de la Rosa et al., 2001).
Of course, it is certainly not the case that low light reduces the concentration of defensive chemicals in all plants (Burns et al., 2002), and a fundamental point is that not all compounds decline in concentration under low light. This specificity in the effect of shading, and its relationship to the responses of herbivores to putative defensive compounds, have been the subject of intense discussion in the context of alternative defence theories (Lerdau et al., 1994; Berenbaum, 1995; Hamilton et al., 2001; Close & McArthur, 2002; Koricheva, 2002; Nitao et al., 2002). Specificity is best characterized for phenolic compounds in woody species. For example, in Populus tremuloides, low light reduced proanthocyanidins (condensed tannins) but had less effect on phenolic glycosides, which were the main factor influencing herbivory (Hemming & Lindroth, 1999). In Betula pubescens, total phenolics and soluble proanthocyanidins were reduced by shade netting treatments, but gallotannins (hydrolysable tannins), cell-wall-bound proanthocyanidins and flavonoids (including kaempferols and quercetins) were not affected (Henriksson et al., 2003). The phenolic composition of another birch species (Betula pendula) is influenced by light spectral quality. Tegelberg et al. (2004) concluded that increasing R:FR shifted the balance of phenolics from chlorogenic acids to flavonoids, and that this effect was distinct from those of increasing UV-B, which increased concentrations of many flavonoids (kaempferols and quercetins) and chlorogenic acids. Spectral modification had no effect on proanthocyanidins in B. pendula (Tegelberg et al., 2004), unlike shading treatment in Betula pubescens (Henriksson et al., 2003). Weinig et al. (2004) found that increased R:FR increased total phenolics in seedlings of Impatiens capensis, although both these authors and Tegelberg et al. (2004) linked changes in phenolics with the reduced growth observed at higher R:FR. In Nothofagus antarctica, removal of solar UV-B radiation increased the concentration of hydrolysable tannins (gallic acid and its derivatives) but decreased the concentration of a flavonoid aglycone (Rousseaux et al., 2004). Flavonoid aglycone was also increased on the sun-exposed side of the canopy, as was quercetin-3-arabinopyranoside (Rousseaux et al., 2004).
The responses of herbivore to shade-induced change in host chemistry are less well explained by bulk chemistry (total phenolics for example) than concentrations of specific compounds (Crone & Jones, 1999; Ossipov et al., 2001; Henriksson et al., 2003; Lahtinen et al., 2004; Rousseaux et al., 2004). Overall, it is increasingly clear from the ecophysiological literature that the responses of defence-related chemicals to shade are far more subtle than can be explained by the bulk diversion of carbon into secondary metabolism that is predicted by the CNB hypothesis. The molecular and cellular literature is now beginning to shed light on some of the underlying mechanisms through which this fine-tuning of plant secondary metabolism is controlled (see section VII).
V. Short-term responses to the light environment: induced defences
In addition to the constitutive defences produced by plants that can be influenced by light, evidence is accumulating that induced defences may also be affected. Induced defences are those that involve rapid changes in biochemistry and gene expression in response to herbivore attack or pathogen infection. In the case of pathogen infection, such responses usually require molecular recognition events, such as classic gene-for-gene based resistance. Physical damage can also be sufficient to activate some responses, especially in the case of herbivore defence, although several elicitors of specific responses have been isolated from herbivore oral secretions. The term ‘induced resistance’ broadly refers to plant responses such as the hypersensitive response (HR), the biosynthesis of defensive secondary metabolites (e.g. phytoalexins), and the up-regulation of expression of defence genes [such as those encoding pathogenesis-related (PR) proteins and protease inhibitors].
There is anecdotal evidence that the development of plant resistance to microbial pathogens can often require illumination during the infection process. The scientific literature contains a number of reports supporting this idea. For example, light is necessary for development of resistance responses to Pseudomonas solanacearum in tobacco (Nicotiana tubacum) (Lozano & Sequeira, 1970), Xanthomonas oryzae in rice (Oryza sativa) (Guo et al., 1993), and Pseudomonas syringae and Peronospora parasitica in Arabidopsis (Mateo et al., 2004; Zeier et al., 2004). Furthermore, red light treatments were able to induce resistance to B. cinerea and Alternaria tenuissima in broad bean (Vicia faba) (Islam et al., 1998; Rahman et al., 2003). As well as these studies on interactions between plants and pathogens, there are also several examples of plant responses to isolated pathogenic elicitors that are also light-dependent. For example, leaf necrosis in tomato in response to an avirulence elicitor from Cladosporium fulvum is substantially reduced in the dark (Peever & Higgins, 1989), and cell death induced by the fungal toxins AAL from Alternaria alternata (Moussatos et al., 1993) and fumonisin B1 (Asai et al., 2000; Stone et al., 2000) requires light, as does the fumonisin B1-induced expression of the Systematic Acquired Resistance (SAR) marker gene Pathogenesis-Related 1 (PR1) (Asai et al., 2000). In addition, Tang et al. (1999) found that necrotic lesion formation activated by over-expression of the tomato Pto disease resistance gene also requires light, although the same authors found that HR mediated by the endogenous Pto gene in plants inoculated with an incompatible strain of P. syringae was light-independent (Tang et al., 1999). This contrasts with the light dependence of resistance to the same pathogen in Arabidopsis conferred through a different resistance–avirulence gene interaction (Zeier et al., 2004). Interestingly, programmed cell death caused by UV-C treatment also requires illumination with white light following a lethal UV-C dose in Arabidopsis (Danon et al., 2004). It is important to note, however, that, in addition to these examples, there are many inducible defence responses that are clearly not light-dependent. Indeed, responses to the same stimuli can involve light-dependent and independent elements. For example, whereas cell death in response to the C. fulvum elicitor in tomato was light-dependent, lipoxygenase enzyme activation was not (Peever & Higgins, 1989). Finally, it should be noted that these findings tend to be rather ad hoc and based on light/dark differences – very few studies have considered the qualitative or quantitative effects of light on resistance.
In green tissues, chloroplasts are an obvious target that can respond to changes in the light environment, although chloroplasts might not be considered an obvious part of a defence response. However, links between chloroplast function and disease resistance have been identified in several systems. For example, silencing of the 33K subunit of the oxygen-evolving complex of photosystem II (Abbink et al., 2002) or over-expression of the DS9 chloroplast metalloprotease (Seo et al., 2000) increases susceptibility of tobacco plants to tobacco mosaic virus infection. White leaves of the variegated albostrians barley (Hordeum vulgare) mutant support increased growth of the fungal pathogen Bipolaris sorokiniana (Schäfer et al., 2004) and fail to produce Salicylic acid (SA) in response to powdery mildew infection (Jain et al., 2004). In Arabidopsis, the presence of functional chloroplasts is also required for HR in leaves infected with an incompatible strain of P. syringae (Genoud et al., 2002). Thus, resistance in a number of different plant–pathogen interactions requires chloroplast function, although this does not necessarily mean that it requires light.
In contrast to pathogen defence, there are relatively few specific studies on the influence of light on induced resistance against herbivores or responses to wounding. One exception to this is the class of so-called indirect defences. These involve the generation of complex mixtures of volatile compounds that are used by predators and insect parasitoids, such as parasitic wasps, as cues to locate their prey and hosts, respectively (Pare & Tumlinson, 1999). As noted above, many investigations of herbivore-induced volatile production have shown that this response is largely light-dependent (e.g. Loughrin et al., 1994; Halitschke et al., 2000; Maeda et al., 2000; Gouinguené & Turlings, 2002). In general, volatile emission induced by herbivore feeding or by application of methyl jasmonate appears to follow a diurnal cycle, with emission being much stronger during the light period than the dark. However, other defence-related volatiles are also produced during the night (e.g. De Moraes et al., 2001).
The plant hormone jasmonic acid (JA) plays a central role in controlling responses to wounding and herbivore attack and to infection by some pathogens, especially necrotrophic fungi. The early steps of JA biosynthesis occur in the chloroplasts of wounded leaves (Turner et al., 2002), but JA synthesis is not necessarily light-dependent. Wound-induced JA biosynthesis was observed in soybean hypocotyls in the dark (Creelman et al., 1992) and also occurs in nonphotosynthetic tissues such as potato (Solanum tuberosum) tubers (Koda & Kikuta, 1994). Furthermore, Zeier et al. (2004) observed that pathogen-induced JA levels in Arabidopsis were higher in the dark than in the light. This suggests that induced responses to wounding might be largely light-independent, although it is important to note that, in the vast majority of studies, no direct comparison of JA biosynthesis or of responses to JA or wounding have been made between different light conditions. Where such comparisons have been made, there is evidence in some cases that wound- and JA-induced responses can in fact be light-dependent. Most notable amongst these are the indirect defences, but direct defence responses can also be light-dependent. For example, in a series of reports on the expression of stress-inducible genes from rice, several were identified that, in general, required light for their induction by wounding and by exogenous JA application (Agrawal et al., 2002a,b,c, 2003). In Arabidopsis, the ASCORBATE PEROXIDASE 2 (APX2) gene is also wound-induced, but by a JA-independent pathway. Instead, it appears to be regulated by changes in photosynthetic electron transport (PET) in wounded leaves, which results in increased levels of ROS (Chang et al., 2004). Interestingly, most of the light-dependent wound-induced genes from rice are also responsive to applied H2O2 and copper (a ROS generator), even in the dark (Agrawal et al., 2002b,c, 2003). These data suggest that light-driven generation of ROS in chloroplasts around sites of wounding might be responsible for the expression of a subset of wound-induced genes.
VI. Mechanisms for light-dependent induced defences
Whilst there is a large body of research defining the physiological basis for the light dependence of constitutive defences, the basis of the effect of light on induced resistance is less well understood. There are two general mechanisms by which light could regulate defence responses in plants. The first of these is based on the energetic status of light-driven chemical reactions (dependent on the ability of PET to generate ATP and reducing power), and the second is based on the direct perception of light and downstream light-responsive signalling pathways.
1. Photosynthesis and ROS
Photosynthesis uses light energy to drive electrons through complex electron transport chains in the thylakoid membranes, which harvest the energy from activated carriers to ultimately generate ATP and reducing power in the form of NADPH. These key metabolites are then used in carbon fixation in the Calvin cycle, as well as in various other metabolic reactions that take place in the chloroplasts, such as fatty acid biosynthesis and assimilation of N into amino acids. There are two ways in which these light-dependent processes in chloroplasts could impact on short-term, induced defence responses. First, major changes in gene expression, protein synthesis and defence metabolism could potentially be affected by the loss in the dark of substrates synthesized in chloroplasts. Interestingly, at least part of the biosynthetic pathways for three major defence-related hormones, JA, SA and abscisic acid (ABA), are also located in plastids. Secondly, as indicated in section V2, chloroplasts can be a significant source of ROS during stress conditions. Plant leaves acclimate to average ambient light intensities during their growth, such that the levels of light-harvesting complexes and Calvin cycle enzymes are optimized to make most efficient use of the available light. However, when light intensities transiently increase, or when carbon fixation is prevented, PET generates more electrons than can be accepted by the available electron acceptor NADP+. In these situations, free electrons from the electron transport chain can be transferred directly to oxygen to form ROS. Secondly, increased excitation energy can be dissipated via photorespiration, which ultimately results in the generation of H2O2 in the peroxisomes. Normally, a range of biochemical and physiological systems are activated to minimize over-reduction of the electron transport chain and to scavenge those ROS that are produced. However, under severe acute stress, ROS can accumulate to concentrations that exceed the array of antioxidant systems of the chloroplast (Apel & Hirt, 2004). Additionally, damage to the chloroplasts or disruption of chlorophyll biosynthesis can result in the accumulation of photosensitive pigments that can directly generate ROS in the light. As ROS are well known as important regulators of several defence responses (Apel & Hirt, 2004), significant perturbations in redox balance in the chloroplasts may contribute to ROS-regulated defence.
The implications of the requirement for light for chloroplast-derived ROS may extend beyond the direct signalling roles of ROS. For example, one consequence of ROS production under stress conditions is lipid peroxidation. Many of the products of lipid peroxidation reactions that occur following wounding or pathogen attack are also reactive electrophile species – molecules with reactive (electrophilic) carbonyl groups (Vollenweider et al., 2000). Many of these electrophiles are now known to act as important signalling molecules, eliciting a range of defence responses ranging from cell death to defence gene expression (Vollenweider et al., 2000; Alméras et al., 2003; Thoma et al., 2003; Cacas et al., 2005). Electrophiles produced as a consequence of stress may be derived either from direct attack of ROS on membrane lipids or from the activity of lipoxygenase enzymes. Light is therefore likely to directly influence the generation of ROS-derived electrophiles (and downstream responses), but not those generated by lipoxygenase activity. Interestingly, such effects have been noted in several interactions between plants and pathogens or their elicitors. For example, Montillet et al. (2005) found that, in response to the elicitor cryptogein, cell death was mediated by light-dependent ROS in the light, but in the dark cell death was independent of ROS and correlated with the activity of a specific lipoxygenase activity. Hence, different mechanisms for the production of bioactive electrophiles may be required to operate under different light environments.
2. Photosensitive pigments and ROS
During pathogen resistance responses, the primary source of ROS is not the chloroplast, but an enzyme found in the plasma membrane known as NADPH oxidase, or respiratory burst oxidase (Apel & Hirt, 2004). One might therefore assume that light-dependent, chloroplast-derived ROS are not likely to be important in pathogen defence. However, the situation is not necessarily clear-cut, as the importance of the NADPH oxidase does not preclude an additional role for chloroplast ROS. Many researchers have isolated mutants from various species, collectively termed lesion mimic mutants, that display spontaneous formation of necrotic lesions on their leaves (Lorrain et al., 2003). These lesions are similar to those formed during the HR (a key component of disease resistance responses) and are generally accompanied by the increased expression of PR genes and increased resistance to infection. Generally, lesion mimic mutants were isolated and characterized as part of an effort to understand the mechanisms of disease resistance signalling. However, it is likely that, in many cases, these mutants in fact highlight a more general link between chloroplast ROS and plant stress responses, including pathogen resistance. This idea is discussed in detail elsewhere by Mullineaux and colleagues (Karpinski et al., 2003; Bechtold et al., 2005), but is based on two findings. First is the observation that lesion formation in many of these mutants is light-dependent (e.g. Johal et al., 1995; Genoud et al., 1998; Brodersen et al., 2002). Secondly, cloning of several of the genes defined by these mutations has identified a number of genes involved in chlorophyll biosynthesis or degradation (e.g. Hu et al., 1998; Ishikawa et al., 2001; Mach et al., 2001; Pruinska et al., 2003). In addition, manipulation of the expression of several other genes involved in chlorophyll biosynthesis also results in light-dependent lesion mimic phenotypes and increased disease resistance (e.g. Kruse et al., 1995; Mock & Grimm, 1997; Mock et al., 1999; Molina et al., 1999). The most likely explanation for these observations is that ROS are produced by the action of light on chlorophyll intermediates that act as photosensitizers – that is, they absorb light energy which excites electrons which are subsequently transferred to molecular oxygen to form ROS. These ROS then act as signals to initiate plant defence responses, including pathogen resistance.
Clearly, then, the light-dependent generation of ROS from free photosensitive pigments or those present in the photosynthetic light-harvesting complexes can impact on defence in mutants and transgenic plants with altered chloroplast biology. The question, then, is whether they do so under normal circumstances. At present, it is not possible to answer this question, but it is likely that plants have evolved mechanisms to deal with the problems of light-dependent ROS generation in tissues under attack from pests and pathogens. For example, the Arabidopsis CHLOROPHYLLASE 1 (AtCHL1) gene is involved in chlorophyll degradation, and is required to remove photosensitive porphyrin ring intermediates. AtCHL1 is induced by wounding and infection with necrotrophic pathogens (Benedetti et al., 1998; Kariola et al., 2005), at which time it functions to prevent accumulation of ROS generated from breakdown products of chlorophyll released from damaged chloroplasts. Plants with reduced AtCHL1 gene expression show increased resistance to Erwinia carotovora, a necrotrphic bacterial pathogen, but increased susceptibility to Alternaria brassicicola, a fungal necrotroph (Kariola et al., 2005). Resistance to E. carotovora is conferred by an SA-dependent pathway, whilst resistance to A. brassicicola is normally regulated via JA-dependent signalling. As ROS can potentiate SA-dependent defences which in turn can antagonise JA-dependent resistance, it appears that AtCHL1 might modulate the balance between SA- and JA-dependent resistance pathways by controlling ROS generation from chlorophyll metabolites. Interestingly, over-expression of the ACD2 red chlorophyll catabolite reductase gene in Arabidopsis, which would be expected to reduce the accumulation of photosensitizers, generated increased tolerance to a virulent strain of P. syringae (Mach et al., 2001). In these plants, bacterial growth was not affected, but cell death symptoms were reduced.
Whilst beyond the scope of this review, it is also notable that many plant species synthesize photosensitizers that are thought to act as direct defences. In the presence of UV-B or white light, these so-called phototoxins generate ROS that function to directly inhibit herbivore or pathogen function (Downum, 1992). Conversely, several genera of fungal pathogens also produce photosensitive toxins, such as cercosporin, that result in plant cell necrosis (Daub & Ehrenshaft, 2000).
3. Light signalling
The second major mechanism suggested above (section VI) by which light may regulate defence is via direct light-responsive signalling pathways. Evidence for this type of regulation has been recently uncovered in Arabidopsis. Genoud et al. (1998) identified an Arabidopsis light signalling mutant, psi2 (phytochrome signalling), which, in addition to effects on light-dependent expression of photosynthetic genes, displayed light-dependent development of spontaneous necrotic lesions and increased PR1 gene expression. Further characterization of these phenotypes showed that light regulated the resistance responses at multiple levels. First, PSI2 is a regulator of phytochrome-mediated responses, and phytochrome A (PhyA) and phytochrome B (PhyB) are also required for light-dependent HR lesion formation and PR gene expression (Genoud et al., 2002). Consequently, resistance to P. syringae is reduced in phytochrome mutants and increased in the psi2 mutant. This is an example of light acting in a direct signalling role to modulate induced resistance. How and why phytochrome signalling might impact on disease resistance are unclear, although such an effect might represent a sensitive mechanism by which cytosolic and nuclear responses are matched with changes in chloroplast activity caused by variations in light intensity. Perhaps significantly, in these experiments, HR (although not PR gene expression) also required the presence of functional chloroplasts, as cell death was not observed in white sectors of variegated leaves. Hence, both metabolic and signalling roles for light may combine to co-ordinate a full resistance response.
In terms of induced defences we can therefore identify a range of different levels of interaction between light and responses to biotic attack. These include a range of effects on ROS generation, as well as direct signalling roles for light via phytochrome signalling, and are summarized in Fig. 1.
VII. Interpreting interactions between light and defence responses
In assessing the range of experimental systems we have discussed in this article, a general conclusion is that, where light has been found to modulate plant defence against herbivores or disease, its effect is usually to increase defence. A key question, therefore, is whether we can identify mechanistic explanations for this observation. As is often the case, ecologists and molecular biologists have taken very different approaches to the question of interactions between light and defence. Given that this is a complex interaction with different components, it is not surprising that such different approaches are possible. Clearly, the fundamental importance of light for plant growth and development means that there is no single explanation that can unite observations across widely different scales of organization. However, one way forward is to place the whole range of evidence, from molecular to ecophysiological, within the framework of optimal defence theory (Hamilton et al., 2001). Is a greater investment in defence in high light consistent with optimal defence theory, and, if so, do the molecular and cellular data provide insights into the mechanisms through which optimal defence is achieved? This relates to a second important point, which is the precise terminology used to describe defence. The semantics of defence in plant–pathogen or plant–herbivore interactions, which has been widely debated by ecologists and ecophysiologists (Clarke, 1986; Stowe et al., 2000), but less so by cell and molecular biologists, forms a pertinent background to these questions. Defence is defined as any mechanism that protects the plant from reductions in fitness in the presence of herbivores or pathogens and has two components. The first component is resistance, which reduces the severity of attack by inhibiting the activity or performance of the herbivore or pathogen. The second component is tolerance, which reduces the negative consequences of attack on host fitness. In our view, the clear differentiation between resistance and tolerance is essential to understanding mechanisms of interactions between light and defence.
The first requirement of optimal defence theory, that tissues that have the greatest value to the plant should be most defended, is clearly satisfied. Models of canopy photosynthesis are consistent in showing that leaves exposed to high light contribute most photosynthate (Leuning et al., 1995; dePury & Farquhar, 1997). Secondly, defence should be in proportion to the probability of attack. There are clearly many systems in which herbivores are more abundant and/or more active in high-light environments, for example as a result of higher temperatures (see section III). Arguably, the higher N concentration of high-light tissues may increase their potential palatability for herbivores, and so increase the risk of attack. There are certainly examples where exposed tissues suffer more herbivory although they are better defended (e.g. Louda & Rodman, 1996; Sipura & Tahvanainen, 2000). These arguments are harder to apply to pathogens and, if anything, it might be expected that the probability of infection might be lower under high-light conditions as a result partly of direct light effects (see section II) and partly of the correlated lower humidity and leaf surface water. The third requirement of optimal defence theory is that defence is a function of the balance between its benefits and its costs. The ‘broad-brush’ prediction of the CNB hypothesis, that defence is less costly under high-light conditions because substrates are more freely available, fails to explain the specificity in the responses of individual metabolites to the light environment. Nonetheless, there are a number of other mechanisms that could result in altered costs of defence under different light conditions.
One element of changed costs of defence may relate to the induction of shade-avoidance mechanisms under low-light conditions. The possible trade-offs between defence and shade avoidance responses at low light as they relate to competitive ability have recently been reviewed by Cipollini (2004), who argued that shade-avoidance responses could constrain defence via a number of mechanisms. First, the shift in allocation to extension growth under shade might directly compete with allocation to defence, although not necessarily by competition for resources. There may be direct interference between the signalling mechanisms controlling acclimation to the light environment and those regulating defence. Increased stem elongation in the shade response is under the control of auxins and gibberellins (Vandenbussche & Van Der Straeten, 2004). Auxin may interact with defence via cross-talk between indole acetic acid (IAA) and defence signalling, such that IAA reduces JA-induced production of defence compounds (Kernan & Thornburg, 1989; Baldwin et al., 1997). Conversely, the concentrations of active auxins and the expression of auxin response genes are reduced by wounding (Thornburg & Li, 1991; Cheong et al., 2002; Schmelz et al., 2003) and herbivory (Schmelz et al., 2003). Cipollini (2004) also suggested that cell wall stiffening might be a mechanism for antagonism between shade avoidance and defence, with the gibberellin-mediated cell wall loosening leading to increased cell expansion in the shade being incompatible with the cell wall stiffening that can be a significant component of defence.
Cipollini (2004) described the interference between the shade response and defence as an opportunity cost but, equally, there may be a range of ‘opportunity benefits’ that reduce the cost of defence in high light, because processes induced for photoprotection also confer protection against biotic attack. High-light stress, including UV-B irradiation, activates molecular responses that have much in common with pathogen and herbivore responses (Mackerness et al., 1999; Rossel et al., 2002; Izaguirre et al., 2003; Kimura et al., 2003; Stratmann, 2003). In fact, the increasing documentation of the kinds of responses induced by various biotic and abiotic stresses makes it clear that there are many overlaps in these responses. To try to understand the significance of these overlapping responses, it is useful to consider what the functions of induced responses to these different environmental factors might be. For example, many stress responses include increases in the accumulation of antioxidants and the expression of protective chaperone proteins (such as heat-shock proteins and osmoprotective proteins). Many forms of environmental insult will disrupt biochemistry leading to increased ROS generation, requiring increased antioxidant production to counteract their cytotoxic effects. While there may be many mechanisms for ‘opportunity benefits’, in our view, many may be based on the involvement of ROS in responses to light, herbivory and disease. Understanding these potential mechanisms requires careful differentiation between resistance and tolerance.
1. Light and the resistance components of defence against herbivore or pathogen attack
Light-driven generation of ROS in damaged plants may be central to interactions between light and the resistance components of defence against pathogens or herbivores. Photosensitive chlorophyll degradation intermediates formed as a result of cellular damage caused by herbivores and necrotrophic pathogens can contribute to ROS generation and defence signalling (Kariola et al., 2005), as can excess hydrogen peroxide derived from photorespiration (e.g. Chamnongpol et al., 1998; Mateo et al., 2004). Several studies described in section V also indicate a requirement for functional chloroplasts to activate the HR during pathogen resistance, which might also suggest a functional relationship between light-driven reactive oxygen chemistry and defence. NADPH oxidase is clearly an important source of ROS for defence signalling, but is metabolically costly (in terms of NADPH consumption). It is possible that, in some systems, ROS generation is supplemented by the action of light on photosensitive pigments such as chlorophyll. Potentially, ROS provides a basis for a ‘supply side’ hypothesis very different from the CNB hypothesis. Resistance is facilitated in (high) light tissue because ROS for signalling can be supplied at less cost via light-driven reactions than ROS generated in the dark. Interestingly, there is evidence that elevated UV-B can enhance wound-induced defensive chemicals (Levizou & Manetas, 2001).
There are also specific examples of proteins involved in both resistance and responses to light that may be directly involved in signalling cross-talk. The zinc finger transcriptional regulator LSD1 (Lesion Simulating Disease) is an Arabidopsis protein first identified through a genetic mutation which conferred a runaway cell death phenotype (Jabs et al., 1996). The LSD1 gene has been studied mainly with regard to its role as a negative regulator of pathogen-induced hypersensitive cell death. More recently, however, Mateo et al. (2004) showed that LSD1 is also involved in acclimation to high-light stress. Interestingly, the same authors showed that the effects of LSD1 on pathogen-induced cell death are mediated by ROS generated during light-dependent photorespiration. NPR1/NIM1 (NONEXPRESSER OF PR GENES1 NONINDUCIBLE IMMUNITY1) is another signalling protein identified as a key regulator of multiple pathogen resistance pathways. Over-expression of a rice NPR1 gene leads not only to elevated disease resistance, but also to hypersensitivity to light (Chern et al., 2005).
2. Light and the tolerance components of defence against herbivore or pathogen attack
As noted above, both biotic attack and light stress are sources of oxidative stress in plant tissues. Furthermore, light and biotic attack may also act synergistically to increase oxidative stress. Biotic stress can result in uncoupling of the light and dark reactions of photosynthesis, meaning that ‘normal’ ambient light intensities cause ROS generation from photosynthesis (Bechtold et al., 2005). One common feature of many stress responses is the down-regulation of genes encoding many components of the photosynthetic machinery (e.g. Izaguirre et al., 2003; Kimura et al., 2003). This may serve as a negative feedback loop to reduce ROS generation, but also to shift metabolism into areas that compete with photosynthesis, such as the oxidative pentose phosphate and shikimic acid pathways (Scharte et al., 2005). Plant mechanisms involved in protection against oxidative stress or repairing the damage it causes are known to be activated by both light and herbivore or pathogen attack (e.g. Rossel et al., 2002; Kimura et al., 2003; Apel & Hirt, 2004). A key point is that these are tolerance mechanisms, not resistance. Clear differentiation between such mechanisms and resistance (i.e. mechanisms that inhibit the herbivore or pathogen) is central to understanding interactions between light and defence, not least the widely discussed role of phenolic compounds in such interactions.
Plant phenolics are a highly diverse group of chemicals that fulfil a range of functions. Some phenolics have demonstrable roles in plant interactions with herbivores or pathogens, either as components of resistance (section IV–3) or as attractants for herbivores (e.g. Roininen et al., 1999; Ikonen et al., 2002). Other phenolics function as ‘sunscreens’ or antioxidants, and some authors have argued that photoprotection is the primary role of many plant phenolics (Close & McArthur, 2002). In considering interactions between light and defence, key points are (i) that plants in high-light conditions are potentially confronted with the risk of increased herbivory (section III–3) and the concurrent need for photoprotection, and (ii) that both light and attack can induce oxidative stress. Under such conditions, phenolic compounds might fulfil at least three functions: (a) as sun-screens reducing light penetration to vulnerable tissues (not selected for by herbivory or disease), (b) as antioxidants involved in reducing the damage caused by ROS (selected for by biotic attack as well as light), and (c) as resistance compounds inhibiting the activity of herbivore or pathogen (not selected for by light).
These multiple functions would be expected to result in the compound-specific changes in the concentration of phenolics evident in the recent ecophysiological literature (see section IV). They would also be expected to lead to different trade-offs in the production of phenolics. In terms of tolerance, the production of phenolic antioxidants in high-light tissue might be seen as an opportunity benefit for defence against biotic attack. Conversely, the synthesis of phenolics conferring resistance (sensuo stricto) against herbivory or disease may represent an opportunity cost on the production of phenolics acting as sun-screens, and vice versa.
These different trade-offs might be expected to be reflected in enzyme activity and gene expression. From this perspective, the three functions of phenolics noted above, while distinct, might all be expected to be associated with an elevated basal flux through the phenylpropanoid pathway. This may explain some of the parallels in terms of global gene expression between herbivory and light stress (e.g. Izaguirre et al., 2003; Gachon et al., 2005). Most commonly, it is the genes encoding the enzymes controlling entry of substrates into the phenylpropanoid pathway, such as phenylalanine ammonia lyase (PAL) and chalcone synthase (CHS), that are noted as responsive to multiple stresses. However, such induction of PAL or CHS is clearly only one element in the regulation of the phenylpropanoid pathways, and there are examples of competition between elements of phenylpropanoid metabolism delivering compounds with different functions. In Sorghum bicolor there is competition between the accumulation of anthocyanin in response to light and the synthesis of phytoalexins in response to challenge by the fungus Cochliobolus heterostrophus (Lo & Nicholson, 1998). This was attributed to the down-regulation of genes specific to anthocyanin biosynthesis and the corresponding up-regulation of genes encoding enzymes involved in phytoalexin synthesis (Lo & Nicholson, 1998). Similarly, in grapes, there appears to be competition between the production of anthocyanins of photoprotection and phytoalexins (resveratrol) for defence against pathogens (Jeandet et al., 1995).
These results show that the plant is able to ‘fine-tune’ phenolic metabolism as the balance of costs and benefits shifts in the face of competing end-points. Recent detailed analyses have begun to reveal the details of the regulation of the phenylpropanoid pathway. In the field, exposure of Vaccinium myrtillus to full sunlight up-regulates a whole series of phenylpropanoid pathway enzymes, but changes in PAL and CHS are much smaller than changes in ‘downstream’ enzymes involved in the synthesis of specific photoprotective compounds (Jaakola et al., 2004). It is clear that sets of several phenylpropanoid pathway genes, for example those involved in flavanol or monolignol biosynthesis, are coregulated during both development and stress responses (Gachon et al., 2005). In the case of light-responsive expression of flavanol biosynthesis, one mechanism for this coregulation was demonstrated to stem from the possession of common transcription factor binding sites in the promoters of coregulated genes (Hartmann et al., 2005). However, while there is clearly coregulation of major elements of the phenylpropanoid pathway, not all enzymes are represented in these gene expression clusters (Gachon et al., 2005). Furthermore, many key downstream enzymes exist in different isoforms with different substrates and products, fulfilling different functions (Kumar & Ellis, 2003). Thus, up-regulation of a single enzyme or even a cluster of coregulated elements of a pathway under high light or biotic attack may reveal little in the absence of understanding of the behaviour of those enzymes controlling pathway end-points.
In our view, there is no single answer to the question of how light alters the cost of defence against herbivory or pathogen attack. However, on the balance of the evidence, it seems likely that costs will often become lower with increasing light. This, taken with the greater value of high-light tissues and the greater risk of attack, at least by herbivores, suggests that increased defence is consistent with the predictions of optimal defence theory. The argument that plants have fine control of defence metabolism, which is a major contrast to ‘supply side’ theories such as the CNB hypothesis, is well established (e.g. Berenbaum, 1995), and molecular studies are increasingly revealing the nature of such fine control. Research at the scale of the transcriptome and metabolome has begun to provide information on the mechanisms by which optimum defence is achieved. However, it is clear that proper understanding of optimum defence cannot be gained through quantification of bulk changes at the whole-plant or whole-organ level, whether in global gene expression, or in bulk measures of defensive chemistry, such as total phenolics. What is required is more detailed temporal and spatial resolution of the responses of specific genes or compounds in the context of their function in the plant under biotic attack and different light conditions.
Whilst ecologists and molecular biologists have mostly taken different approaches to the question of interactions between light and defence, we feel that these approaches can provide an interface which can deliver benefits to both sets of disciplines. Work across these scales can be extremely effective in linking molecular responses with ‘real life’ ecological outcomes to stress (see, for example, work from the group of Ian Baldwin), and we strongly encourage efforts to integrate molecular and ecological studies in all areas of biology.
Work on this topic in the authors’ laboratories was supported by a Royal Society University Research Fellowship (MRR) and research grants from DEFRA (CSA6138) and the UK Horticultural Research Council (CP19) to NDP. We would also like to thank the anonymous referees for their encouraging and helpful comments.