From ancient genes to modern communities: the cellular stress response and the evolution of plant strategies


  • S. PIERCE,

    Corresponding author
    1. Dipartimento di Biologia Strutturale e Funzionale, Università degli Studi dell’Insubria, Via J.H. Dunant 3, 21100 Varese, Italy
    Search for more papers by this author

    1. Dipartimento di Biologia Strutturale e Funzionale, Università degli Studi dell’Insubria, Via J.H. Dunant 3, 21100 Varese, Italy
    Search for more papers by this author

    1. Dipartimento di Biologia Strutturale e Funzionale, Università degli Studi dell’Insubria, Via J.H. Dunant 3, 21100 Varese, Italy
    Search for more papers by this author

†Author to whom correspondence should be addressed. E-mail:


  • 1Two major plant strategy theories attempt to explain how phenotype determines community structure. Crucially, CSR plant strategy theory suggests that stress and sporadic resource availability favour conservative phenotypes, whereas the resource-ratio hypothesis views the spatial heterogeneity of resources as selecting for optimal foraging in chronically unproductive habitats. Which view is most realistic?
  • 2The ecophysiology literature demonstrates that stress is comprised of two processes: (1) limitation of resource supply to metabolism; and (2) damage to biomembranes, proteins and genetic material (chronic stress). Thus stress is defined mechanistically as the suboptimal performance of metabolism.
  • 3Adaptations to limitation buffer metabolism against variability in external resource supply; internal storage pools are more consistent. Chronic stress elicits the same ancient cellular stress response in all cellular life: investment in stress metabolites that preserve the integrity and compartmentalization of metabolic components in concert with molecular damage-repair mechanisms.
  • 4The cellular stress response was augmented by morphological innovations during the Silurian–Devonian terrestrial radiation, during which nutrient limitation appears to have been a principal selection pressure (sensu CSR theory).
  • 5The modern stress–tolerator syndrome is conservative and supports metabolism in limiting or fluctuating environmental conditions: standing resource pools with high investment/maintenance costs impose high internal diffusion resistances and limit inherent growth rate (sensu CSR theory).
  • 6The resource-ratio hypothesis cannot account for the cellular stress response or the crucial role of ombrotrophy in primary succession. CSR theory agrees with current understanding of the cellular stress response, terrestrial radiation and modern adaptations recorded in chronically unproductive habitats, and is applicable as CSR classification.


CSR plant strategy theory (Grime 1974, 1977, 1979, 2001) and the resource-ratio hypothesis (Tilman 1985, 1988) both aim to provide a functional explanation of plant phenotypes and their role in community processes. Although the two theories advocate profoundly opposing views, particularly concerning the nature of stress and plant responses, indecisive analysis of ‘the debate between Grime and Tilman’ (e.g. Grace 1991) has prolonged the impasse that hinders community ecology research. Specifically, Grime invokes stress as one of the principal forces influencing plant communities, with plant responses being conservative in nature. Tilman does not believe stress is a concept applicable to vegetation (Grace 1991), and advocates foraging and optimized resource acquisition as a survival strategy in chronically unproductive habitats. Both theories assume that community composition is a product of the strategies of component species. Thus the aim of this article is to review experimental evidence of the mechanisms underpinning survival in chronically unproductive habitats, to consider which theory is most realistic. We start by summarizing the two theories before discussing the concept of stress and reviewing plant adaptations in chronically unproductive habitats and the evolution of phenotype throughout the terrestrial radiation.

the debate between grime and tilman

Grime (1979, 2001) suggests that while plants in productive habitats acquire resources via foraging, chronically unproductive environments select for stricter control of growth and development, including the resistance of suboptimal periods for resource acquisition and growth. Nutrient limitation, characterized by long periods of stress punctuated by resource pulses, is proposed as central to the evolution of economic resource use and persistent, unpalatable organs that can rapidly resume assimilation when resources become available. Insufficient stress tolerance, and destructive events such as herbivory, fire, soil erosion or the most extreme climatic conditions, result in tissue death. This is a state from which tissues cannot recover, representing the threshold between stress and disturbance. Regular disturbance events elect for rapid completion of the life cycle within temporal windows of opportunity and survival as seeds (ruderalism). These concepts are distilled into CSR theory, which considers the interplay between competition (autogenic inhibition of biomass production); stress (allogenic inhibition); and disturbance (tissue destruction) in shaping phenotype. Characteristic developmental traits are inherent to competitive (C), stress-tolerant (S) and ruderal (R) strategists, with intermediate strategies apparent.

Tilman (1985, 1988) suggests that plants mediate limitations to growth by optimal foraging – allocating photosynthate between root and shoot meristems in a highly plastic manner, thereby regulating the biomass of assimilatory organs to optimize resource acquisition. Two environmental factors, nitrogen availability and light intensity, are considered in this resource-ratio hypothesis – allocation to roots is favoured when N is the most limiting resource; when light is the most limiting resource shoot allocation is favoured. The resource-ratio hypothesis is primarily concerned with plant succession and thus encompasses both unproductive and productive habitats. It assumes that resource availability gradients and competition for resources are the prime determinants of phenotype and thus vegetation change through succession.

A fundamental difference between CSR theory and the resource-ratio hypothesis is that the former can account for the influence of intermittent resource acquisition on the evolution of phenotype (phenotypes that can ‘stand by’ ready to intercept resource pulses), whereas the latter does not consider short-term temporal fluctuation in resource availability as a selection pressure influencing phenotype. Surprisingly, this has not previously entered into the debate between Grime and Tilman, or into evaluations of CSR theory (Tilman 1988; Grace 1991; Oksanen & Ranta 1992; Wilson & Lee 2000).

the nature of stress

Environmental conditions change daily and seasonally due to the movement of the planet about its axis and the precession of the axis throughout the year. Thus habitats are characterized by combinations of environmental variables that may be interdependent, or that change in intensity relative to one another throughout the day and as the year progresses. For example, ‘drought’ is not a lack of water, but a period where high evaporation rates outweigh rainfall, dependent on temperature and ultimately light intensity, all of which interact to affect plant growth (Lüttge 1997). In terms of plant growth, it is perhaps more valuable to consider ‘arid’ habitats as characterized by brief periods of optimal water availability, rather than by drought. Similarly, forest understorey plants experience extensive low light limitation briefly interrupted by sunflecks. Even then, high temperatures associated with direct sunlight limit potential carbon gain (Leakey et al. 2003). Indeed, plants rarely experience optimum conditions – even in productive habitats, light intensities may vary between zero (night) and excessively high over diel periods, with only a brief window when optimal light intensity, temperature and resource availability allow entirely unfettered photosynthesis. Thus stress is often the rule, and optimal conditions the ephemeral exception. However, environmental stress factors appear to take radically different forms. What exactly is stress, and is a single underlying adaptive strategy really appropriate?

Grime (1979, 2001) defines stress as ‘the external constraints which limit the rate of dry matter production of all or part of the vegetation.’Larcher (2003) views stress as ‘a significant deviation from the optimal condition of life.’ Crucially, these definitions are not restricted to the resources needed for growth, but also encompass stress factors (stressors) such as temperature or salinity that impose metabolic injury (that is, that damage metabolic machinery, mainly enzymes, biomembranes and genetic material). In the light of the evidence presented in Table 1, stress can be subdivided into two processes: (1) low internal resource supply rates to metabolism; and (2) metabolic injury. These can be interpreted, respectively, as ‘limitation’ and ‘chronic stress’, although ‘stress’ is a useful umbrella term covering both. Thus stress is experienced within cells, and can be defined mechanistically as the suboptimal performance of metabolism. Larcher's (2003) clear distinction between stress (an internal state) and stressor (the stress stimulus) is essential.

Table 1.  Examples of inherent plant adaptations to abiotic stress and underlying mechanisms
Stressor/stress-tolerance or avoidance mechanismsHow does the adaptation work?Example taxaCitation
  • *

    , Non-angiosperm taxa;

  • †, taxa that may have undergone artificial selection.

High temperature
Plastic expression of heat-shock proteinsStabilization of membranes, proteins and DNA (via chromatin stabilization)Retama raetam (Fabaceae)Merquiol et al. (2002)
Evaporative coolingMaintenance of a temperature suitable for metabolismPitcairnia integrifolia (Bromeliaceae), and probably all plantsLüttge (1997); Beerling et al. (2001)
Expression of cold-shock proteins (CSP)RNA stabilization during translation, membrane stabilizationPinus taeda (Pinaceae), Lotus japonica (Fabaceae), Mesembryanthemumcrystalinum (Aizoaceae), Ceratopterisrichardii* (Parkariaceae), Chlamydomonasreinhardtii* (Chlamydomonadaceae)Graumann & Marahiel (1998);Karlson & Imai (2003);Gimalov et al. (2004)
Production of phospholipids with a greater degree of unsaturation in their fatty acid tailsMaintenance of membrane fluidity and thus the compartmentalization and function of metabolismArabidopsis thaliana (Brassicaceae), Oryza sativa (Poaceae)Sung et al. (2003) and references therein
Meristems surrounded by protective tissuesA stress-avoidance mechanism that produces a favourable microclimate, buffers meristematic tissue against low temperatures (that would otherwise decrease membrane fluidity)Espeletia schultzii (Asteraceae)Smith (1974)
Expression of antifreeze proteins (AFP)Inhibition of the growth and recrystallization of ice crystals, thereby protecting membranes, proteins and DNAAmmopiptanthus mongolicus (Fabaceae), Lolium perenne (Poaceae)Wang et al. (2003); Griffith & Yaish (2004) and references therein
Expression of cryoprotectinBinds to thylakoid membranes, providing stability and integrityBrassica oleracea (Brassicaceae)Sror et al. (2003)
Supercooling – tough bud scales and pectin decrease apoplast permeability, preventing ice crystal penetration into meristematic tissueProtection of metabolic components against intracellular ice crystal formationVitis riparia (Vitaceae)Jones et al. (2000)
Reactive oxygen species
Antioxidant enzyme production (e.g. superoxide dismutase (SOD), catalase, peroxidase, ascorbate peroxidase) and increased SOD activityControls free radicals, protecting nucleic acids, polysaccharides, proteins and membrane lipidsPoa sphyondylodes, Bromus inermis, Elymus nutans (Poaceae), Xerophyta viscosa (Velloziaceae), Mesembryanthemum crystalinum (Aizoaceae)Hurst et al. (2004); Vicréet al. (2004) and references therein; Zhou & Zhao (2004)
Antioxidant production (carotenoids and anthocyanin)Blocks light, inhibits free radicals and protects components of metabolismCraterostigma spp. (Scrophulariaceae)Vicréet al. (2004) and references therein
A variable proportion of root aerenchyma tissue (lacunar gas transport system)Redistribution of O2 from photosynthesis to roots, maintaining internal supplySenecio aquaticus (Asteraceae), 10 Poaceae and two Cyperaceae from AustraliaCrawford (1989); McDonald et al. (2002)
Pneumatophores with lenticelsStress avoidance – enhancement of external resource supply (O2)Avicennia spp. (Verbenaceae),Crawford (1989)
Sonneratia spp. (Sonneratiaceae), Xylocarpus spp. (Meliaceae) 
Slow glycolytic metabolism. Expression of anaerobic proteins (including alcohol dehydrogenase)Delayed exposure to hypoxia. Maintenance of energy supply via fermentation, breakdown of harmful waste products (ethanol)Phleum pratense (Poaceae), Isoetes alpinus* (Isoetaceae)Bertrand et al. (2001); Sorrell (2004)
Low CO2 supply (submergence or water on leaves)
Change of photosynthetic pathway (C3 to C4)CO2-concentration mechanism provides a consistent supply of substrate for metabolismEleocharis baldwinii (Cyperaceae)Ueno (2004)
Recycling of CO2 from respiration via CAM, plastic CAM expressionMaintains internal CO2 supply during and after stomatal blockage (when external supply is unavailable)Aechmea dactylina, Tillandsia usneoides (Bromeliaceae) Werauhia capitata (Bromeliaceae)Haslam et al. (2002); Pierce et al. (2002) Pierce et al. (2001, 2002)
Water-repellent epicuticular wax powderWater repellency keeps stomata free of water and maintains CO2 supply  
High light intensity (photosynthetically active radiation)
Energy dissipation via thermal relaxation of excited electrons in chlorophyll (a.k.a. radiationless transition – measured using associated chlorophyll fluorescence)Photoprotection (diversion of energy away from photosystem proteins and pigments)Geum montanum (Rosaceae), Laminaria digitata, L. saccharina* (Laminariaceae), Delesseria sanguinea* (Delesseriaceae), Phyllophora pseudoceranoides* (Phyllophoraceae), Plocamium cartilagineum* (Plocamiaceae)Manuel et al. (1999);
Dring et al. (2001)
Variable expression of xanthophyll cycle pigments, thermal relaxationVariable photoprotection in response to a fluctuating stressorYucca brevifolia (Agavaceae),Ensminger et al. (2001); Barker et al. (2002);Gevaert et al. (2002); Lovelock & Robinson (2002)
Laminaria saccarina* (Laminariaceae),
Cladophora glomerata* (Cladophoraceae),
Bryum pseudotriquetrum* (Bryaceae),
Ceratodon purpureus* (Ditrichaceae), Grimmia antarctici* (Grimmiaceae) 
Light screening via transient anthocyanin pigment productionShields photosystems and provides photoprotection (particularly for young leaves lacking a well developed xanthophyll cycle)Bauhinia variegata, Ricinus communis (Fabaceae)Smillie & Hetherington (1999); Manetas et al. (2002); reviewed by Steyn et al. (2002)
Light screening via anthraquinone pigment (parietin) synthesisStress-avoidance involving photoprotectionXanthoria parietina* (Teloschistaceae)Solhaug & Gauslaa (2004)
Light reflectance by trichomes or epicuticular wax, thermal relaxationCombined stress-tolerance (energy dissipation) and avoidance (light reflectance) ultimately providing photoprotectionEncelia farinosa (Asteraceae), Cotyledon orbiculata, Cotyledonpaniculata (Crassulaceae), Tillandsia spp. (Bromeliaceae)Ehleringer (1976); Robinson et al. (1993); Martin et al. (1999);Pierce (2005)
Low/variable light intensity
Seasonal internal reorganization of leaf anatomy and cell structure (palisade extent, chloroplast number, chloroplast size, thylakoid size and structure)Maintains appropriate light interceptionGuzmania monostachia (Bromeliaceae)Maxwell et al. (1999)
Investment in high chlorophyll and Rubisco contentsMaintenance of photosynthetic capacity – photosynthesis ‘stands by’ for rapid utilization of high light intensities during sunflecksShorea leprosula (Dipterocarpaceae)Leakey et al. (2003)
CAM provides a high photosynthetic capacityPhotoprotection and efficient resource use – photosynthesis ‘stands by’ for rapid utilization of high light energy during sunflecksAechmea magdalenae (Bromeliaceae)Skillman et al. (1999)
Leaf ‘pattern-gene’ colour variegationLeaves have areas of differential light absorbance and light responseTillandsia butzii (Bromeliaceae)Pierce (2003)
Ultra-violet radiation
Expression of betacyanin and acylated flavonol glycosidesPigments shield underlying metabolismMesembryanthemum crystalinum (Aizoaceae)Ibdah et al. (2002)
PhotoreactivationDNA repairSanionia uncinata* (Amblystegiaceae),Batschauer (1993);
  Sinapis alba (Brassicaceae)Lud et al. (2002)
Nucleotide excision repairDNA repairLycopersicon esculentum,Xu et al. (1998)
  Nicotiana tabacum (Solanaceae), 
  Oryza sativa, Zea mays (Poaceae) 
Regulation of stomatal and cuticular transpirationMaintenance of cell water potential and thus functioning of metabolic componentsLarrea tridentata (Zygophyllaceae)Ogle & Reynolds (2002)
Change of photosynthetic pathway (C3 to CAM)Increased water-use efficiency maintains cell water potential and thus functioning of metabolic components, maintenance of resource supply (CO2)Umbilicus rupestris (Crassulaceae), Clusia minor (Clusiaceae), Sedum integrifolium (Crassulaceae), Peperomia obtusifolia (Piperaceae)Cushman & Borland (2002) and references therein
Plasticity in CAM expressionMaintenance of resource supply (CO2) following periods of stomatal closure. Increased water-use efficiency maintains cell water potential and thus functioning of metabolic componentsKalanchoë diagremontiana (Crassulaceae), Polypodium crassifolium, Platycerium veitchii, Microsorium punctatum (all Polypodiaceae)*Holtum & Winter (1999); Griffiths et al. (2002)
Expression of stress proteins (dehydrin, osmatin) and sugars (raffinose, trehalose)Stabilization of membranes and proteinsAnastatica hierochuntica (Brassicaceae)Hoekstra et al. (2001)
Expression of late embryogenesis-abundantStabilization of membranes and proteins.Craterostigma plantagineumVicréet al. (2004)and references therein
(LEA) proteins, ABA-regulated expression of LEA-like CdeT27-45. Regulation of transpiration by leaf foldingMaintenance of cell water potential and thus metabolic context(Scrophulariaceae)
Expression of compatible solutes (proline, glycine betaine, polyols)Stabilization of membranes and enzymesAmmophila arenaria (Poaceae), Puccinellia martima (Poaceae)Smirnoff & Stewart (1985); Crawford (1989) and references therein
Aquaporin-mediated regulation of root hydraulic conductanceMaintenance of cell water potential and thus metabolic contextAgave deserti (Agavaceae)North et al. (2004)
Differences in Rubisco large subunit amino acid sequencesAltered Rubisco specificity factor results in forms that have a greater carboxylase activity (cf. oxygenase activity), increasing carboxylation efficiency and thus water-use efficiency24 species of diverse taxonomy and ecology native to the Balearic IslandsGalmés et al. (2005)
Low nutrient availability
Biotypes with slower respiratory rate and growth have longer-lived leaves, retaining and conserving more nutrientsMaintenance of consistent internal resource supply (N)Lolium perenne (Poaceae)Pilbeam (1992)
Dauciform root productionMaintenance of consistent internal resource supply (P)Schoenus unispiculatus (Cyperaceae)Shane et al. (2005)
Carnivorous syndromeStress avoidance by enhancement of external resource supply (N) – carnivory supplies ≈50% of the plant's NDrosera rotundifolia (Droseraceae)Millett et al. (2003)
Na+ exclusion by a suberized root endodermis, ion channels with greater affinity for K+ than for Na+Maintenance of cytoplasm ionic balance and thereby enzyme functionPuccinellia tenuiflora (Poaceae)Peng et al. (2004)
Selective Na+ and K+ uptake by root plasma membrane channelsMaintenance of cytoplasm ionic balanceThellungiella halophila (Brassicaceae)Volkov et al. (2003)
Change of photosynthetic pathway (C3 to CAM). Salt dilution in succulent tissues and foliar bladder cellsMaintenance of cytoplasm ionic balance and thereby enzyme function, increased water-use efficiency (maintenance of cell water potential and thus metabolic context)Mesembryanthemum crystalinum (Aizoaceae)Winter & Gademann (1991);Adams et al. (1998)
Excretory salt glandsMaintenance of cytoplasm ionic balanceAtriplex spp. (Chenopodiaceae)Crawford (1989) and references therein
Salt removal – NaCl accumulation in older leaves followed by leaf dehiscenceMaintenance of enzyme function in younger tissuesJuncus maritimus (Juncaceae)Crawford (1989)
Metal toxicity
Removal of ions (Zn) from cytosol to the vacuole using elevated expression of high-affinity transportersAvoids damage to plasma membrane, avoids enzyme inhibition (e.g. ATPase, nitrate reductase)Thlaspi caerulescens (Brassicaceae)Assuncao et al. (2001)
High cationic exchange capacity of cell walls and high constitutive expression of peroxidases and chitinases limit Al diffusion through cell walls. Root mucilage binds AlAvoids inhibition of cell division, regulates cell wall assembly, ion fluxes, plasma membrane stability, lipid peroxidationPicea abies (Pinaceae)Nagy et al. (2004)

Clearly, stress is a state of cells that affects individuals, but can this extend to communities? If we accept that the behaviour of individuals (e.g. growth rate, mortality) can affect the composition of the community (implicit in both theories), then the stress responses of individuals will inevitably result in community-scale effects. Indeed, variation in the prevalence of different stressors favours different species, resulting in a characteristic community structure (Körner 2003). As the community is composed of species suited to the particular regime of stressors, it should not be regarded as ‘stressed’ under a typical environmental regime (Körner 2003, 2004). However, should the environmental regime change too rapidly for community structure to follow suit, then the existing community will include stressed individuals that exhibit lower productivity (conforming to Grime's definition). Chronically stressed individuals may not die immediately, introducing a lag between the imposition of an unusual environmental regime and species replacement: for woody vegetation a lag of up to ‘several hundred years’ could be expected, during which the vegetation is exposed to a suboptimal environmental regime that favours the formation of a different community (e.g. tree-line–climate relationships; Körner 1999). Thus, although stable climax communities should not be described as ‘stressed’, stress ultimately influences the character of communities and is a driver of temporal community-change processes such as desertification or succession. Stress is therefore a valid, indeed vital, concept for community ecology. Furthermore, Grime's community-scale definition is entirely compatible with a molecular view of stress: ‘the external constraints which limit [metabolism and thus] the rate of dry matter production of all or part of the vegetation’.

adaptation to stress

Different stressors act in essentially the same way: by reducing metabolic performance. Do apparently dissimilar stress adaptations also share a common principle? As demonstrated by the examples in Table 1, adaptations to stress involve investment of resources into three main areas: (1) maintenance of internal resource supply to primary metabolism (limitation avoidance); (2) protection of primary metabolism against structural and functional disruption; and (3) the repair or removal of damaged metabolic components. These are discussed in turn in the following paragraphs, although they are not mutually exclusive and typically occur together.

limitation avoidance

Adaptations that maintain resource supply do not aid resource acquisition per se, but rather guarantee consistency of supply to metabolism. For example, CO2 supply to leaves is not constant, as rainwater on leaf surfaces can obstruct stomata. In humid habitats such as tropical montane cloud forests, plants may use acuminate ‘drip-tips’, hypostomy and water-repellent indumenta to avoid stomatal blockage (Pierce et al. 2001), but some rely on physiological plasticity in C uptake and storage. Crassulacean acid metabolism (CAM) provides a C-storage pool and flexible CO2 uptake in response to leaf surface wetness, and thus a more consistent internal supply of CO2 to carboxylation (Pierce et al. 2002). Additionally, the main enzyme used in CAM CO2 uptake, phosphoenolpyruvate carboxylase, has a higher affinity for its substrate than Rubisco. This is advantageous for aquatic taxa, such as Eleocharis spp. and Littorella uniflora, that use CAM or C4 photosynthesis to maintain a consistently high CO2 concentration in chloroplasts (Robe & Griffiths 1990; Ueno 2004). Lüttge (2004) points out that CAM is primarily ‘a strategy for variable, flexible and plastic niche occupation rather than lush productivity.’ Lack of lush productivity for CAM plants can be ascribed to denser, succulent leaves (low specific leaf area, SLA) with higher internal diffusion resistances that lead to slower growth rates (Maxwell et al. 1997).

C3 species have a similar approach to maintaining a consistent internal C supply: the grasses Festuca ovina and Nardus stricta maintain a continuous flux of C through storage carbohydrate pools, rather than simply depositing carbohydrates for use in future growth (Atkinson & Farrar 1983). Continuous access to stored carbohydrates buffers growth against variation in instantaneous photosynthesis, but requires leaves to be substantial enough to contain standing storage pools. This leads to high maintenance costs, denser tissues (low SLA) with high internal resistance to CO2 diffusion, and thus low growth rates (Atkinson & Farrar 1983). Low growth rates also result from investment in structural tissues (41–51% of C assimilated each day was used for structural growth; Atkinson & Farrar 1983). Thus the internal supply of resources to metabolism may periodically be divorced from external resource supply, and the ability to grow rapidly is sacrificed for the ability to grow consistently in an inconsistent environment.

protection of primary metabolism

The function of biomembranes, enzymes and genetic material is intrinsically linked to their structure. Stressors induce stress by distorting these structures, thereby changing the properties of biomolecules and inhibiting metabolic processes. For example, chilling acts mainly on cell membranes, decreasing phospholipid movement and membrane fluidity, ultimately decreasing transport across the membrane (Lyons et al. 1964). Chloroplasts are the most sensitive site of chilling, characterized by distorted thylakoid membranes (Kratsch & Wise 2000). Acclimation to chilling involves the production of phospholipids with a greater degree of unsaturation in their fatty acid tails (phospholipids are more widely spaced and the membrane is more fluid at low temperature; Sung et al. 2003); elaboration of the chloroplast envelope to enhance transport (Kratsch & Wise 2000); and cold-shock proteins that act as chaperones – physically holding and stabilizing biomembranes (Gimalov et al. 2004).

This is essentially the same situation as heat stress, also characterized by membrane and protein disruption and a similar plant response; heat-shock proteins bond weakly with specific portions of proteins and DNA to form a supporting ‘exoskeleton’ (visualized by Littlefield & Nelson 1999). Expression of heat-shock proteins responds to temperature over hourly time-scales (Merquiol et al. 2002), but these chaperones are also expressed in response to chilling, reactive oxygen species and salinity (Sabehat et al. 1996; Wang et al. 2004).

Similarly, cryoprotectin protects against freezing by binding and stabilizing thylakoid membranes (Sror et al. 2003). However, antifreeze proteins, recorded in organisms as diverse as bacteria, fish, ferns, gymnosperms and angiosperms, work in a different way: not by binding metabolic components, but by binding ice crystals and preventing crystal growth (Griffith & Yaish 2004). Nonetheless, the role of antifreeze proteins is, in effect, the same as cold-shock and heat-shock proteins: the protection of metabolic components. Metabolic protection is also a key characteristic of other freezing-tolerance mechanisms such as supercooling (Table 1).

Stress responses traditionally involve two processes: stress tolerance (the ability of the protoplasm to resist stress); and stress avoidance (the isolation of metabolism from stressors) (Larcher 2003). Both protect metabolism. For example, direct sunlight is extremely energetic and exposed habitats are a hostile environment for photosystem proteins and pigments. All photosynthetic organisms rely on energy dissipation systems, such as thermal relaxation and the xanthophyll cycle, to provide photoprotection (Table 1), and these may be particularly well developed in exposed situations (Barker & Adams 1997). Highly specialized foliar trichomes or epicuticular waxes may also reflect sufficient light to photoprotect leaves (Robinson et al. 1993; Pierce 2005). Stress avoidance is also common in habitats experiencing temperature extremes: Raunkiaer (1910) recognized that plant architecture and the position of buds is important, determining the extent to which extremities are exposed during adverse seasons – buds are defensive structures that shield delicate, dormant meristematic tissues from stressors. These stress-avoidance mechanisms typically form part of a syndrome including stress-tolerance traits, all of which protects metabolism.

repair of metabolism

No protective mechanism can be 100% effective. Thus plants use a variety of mechanisms to prevent the accumulation of defective macromolecules, repairing DNA (Xu et al. 1998) and also repairing or removing damaged proteins (Klütz 2005). Nucleotide excision repair is one of the most flexible DNA-repair mechanisms, operating by recognizing, excising and replacing damaged portions of DNA (Hoeijmakers 1994), strong evidence for which has been demonstrated in a range of higher plant families (Xu et al. 1998; Table 1). However, the most common repair mechanism is photoreactivation, during which photolyase, energized by blue light or UV-A, repairs the pyrimidine dimers of DNA (Sancar 1990). This has been demonstrated for both lower and higher plants (Batschauer 1993; Lud et al. 2002). Additionally, enzymes such as FtsH may be expressed that aid in the repair or removal of partially unfolded proteins (Herman et al. 2003).

the principle of stress tolerance

In the light of the above evidence, the common principle shared by apparently disparate stress-tolerance mechanisms is the safeguarding of metabolic function. This includes tolerance to chronic stress, which requires a more detailed definition: the preservation of the integrity and arrangement of molecules that comprise metabolism. Clearly, cells become stressed by more than simply a lack of resources, and plants in chronically unproductive habitats cannot rely solely on instantaneous resource acquisition for survival. Although the resource-ratio hypothesis and CSR theory both essentially have resource-based views of stress, the importance of cellular acclimation mechanisms is recognized by Grime (1979, 2001) and is consistent with CSR theory's conservative stress tolerance. CSR theory is thus compatible with existing life forms in real situations (Table 1), whereas the resource-ratio hypothesis is not. Stresses that occur over time-scales as short as minutes or hours, such as those encountered during sunflecks or cold shock, cannot be countered by growing new leaves or roots. When plants experience pervasive ‘non-resource’ stressors such as chilling, growing out of trouble is clearly not an option – the only viable strategy is resilience and readiness to act when conditions allow. Indeed, ecophysiologists have expressed essentially the same view of stress and a general, conservative stress response (Larcher 2003) – although lacking the wider context of community ecology. What events selected for this common stress response?

the cellular stress response and ancient life

Metabolism in all cellular life is based on the same underlying biophysical principles, depending ultimately on the bonds between atoms, molecules and parts of molecules. The process of energy capture during photosynthesis illustrates that life relies on interactions between energy and matter, and the consequent state of matter. Energy can be either used or dissipated, or will alter matter in a way that could damage metabolism. Conversely, a lack of energy (or suitable matter) stymies biochemical reactions and affects structure. Exposure of metabolism to energy sources is thus a necessary risk that requires regulation – molecular protective mechanisms are an inevitable characteristic of metabolism.

Contemporary molecular mechanisms of stress tolerance have been inherited from the earliest forms of life. Chronic stress was a hallmark of life in ocean surface layers and shallow-water benthic habitats during the Archean era (3·8–2·5 Ga b.p.). As the Earth lacked substantial atmospheric oxygen and had no ozone layer, higher UV-B intensities and UV-C exposure probably caused DNA damage rates three orders of magnitude greater than today (Cockell 2000), and were probably a major influence on the evolution of early life (Sagan 1973; Pierson et al. 1993). Modern analogues of Archean micro-organisms produce a range of UV-screening and antioxidant pigments (Garcia-Pichel & Castenholz 1991; Wynn-Williams et al. 2002), detoxifying enzymes (Miyake et al. 1991; Lancaster et al. 2004), and suites of enzymes facilitating DNA excision repair or photoreactivation (Karentz et al. 1991; Jeffrey et al. 1996) – all mechanisms demonstrated by modern higher plants that fall within our definition of stress tolerance. Chloroflexus aurantiacus (a contemporary analogue of Archean shallow-water stromatolites) is extremely resistant to UV-C exposure, clearly a relict trait as the ozone layer now shields the biosphere from UV-C (Pierson et al. 1993). Stromatolites experience periodic UV stress, depending on the inclination of the sun throughout the day and variation in sea level, determined by tidal forces. Thus the motion of the solar system imposes inescapable periodic stresses that undoubtedly influenced the evolution of ancient life.

Ancient origins for stress tolerance are confirmed by the fact that all cellular life shares ≈300 stress proteins: the ‘minimal stress proteome’ conserved as part of a ‘cellular stress response’ (Klütz 2005). The cellular stress response also includes numerous DNA repair mechanisms, the genes and proteins of which are similar for a range of plant and animal families (Ahmed et al. 1997; Xu et al. 1998). Indeed, the minimal stress proteome includes five chaperone proteins that assist in the recognition, removal or refolding of damaged proteins (cylophilin, DnaJ/Hsp40, DnaK/Hsp70, GrpE & Hsp60; Klütz 2005). Plant cold-shock proteins are ‘nearly identical to prokaryotic CSPs in size and sequence’, supporting an early origin for these chaperones (Karlson & Imai 2003). The ubiquity and operation of these proteins illustrates that the cellular stress response is a general response expressed regardless of the nature of the stress (Klütz 2005). Indeed, ‘overlaps among stress responses extend from the transcriptional level to the intracellular signalling pathways that regulate gene expression … a primary stress exposure might lead to increased resistance to a subsequent stress, a phenomenon known as cross-tolerance’ (Stratmann 2003). Archean life was probably subject to a combination of chronic stresses necessitating cross-tolerance, including oxidative stress, high ion densities, high and variable temperatures and extreme pH gradients (Garcia-Pichel 1998; Klütz 2005). Thus the cellular stress response is an ancient, general response that forms the basis of stress tolerance for all cellular life.

evolution of the modern stress-tolerator syndrome

Multicellular plants differ from their single-celled progenitors in that cellular stress tolerance is only a part of a greater stress syndrome involving specialized tissues, organs and whole-plant architectures, characterized by ‘inherently slow rates of growth, the evergreen habit, long-lived organs, sequestration, and slow turnover of C, mineral nutrients, water, infrequent flowering and the presence of mechanisms which allow the intake of resources during temporarily favourable conditions’ (Grime 2001). The cellular stress response alone does not appear to be sufficient to account for this higher level of organization. For instance, Grime (2001) notes that fast-growing species may also be found in unproductive habitats if local soil conditions provide sufficient nutrients, even in habitats characterized by stressors such as chilling. Furthermore, mineralization occurs in pulses over time-scales as short as hours, and necessitates maintenance of long-lived roots that can stand by, ready to assimilate resources (Campbell & Grime 1989). Mineral nutrient limitation thus appears to be a more important constraint to community productivity than factors such as temperature, and one that Grime (2001) argues is the principal pressure selecting for long tissue life span. Did nutrient limitation have a central role in the evolution of modern phenotypes?

Basal land plants were related to modern liverworts (Bryophyta, Hepaticae), evolving ≈490 Ma b.p. (Ordovician period; Kenrick & Crane 1997 and references therein; Sanderson et al. 2004). Extant liverworts are intimately associated with N-fixing cyanobacterial symbionts (Wong & Meeks 2002) and mycorrhizae (Russell & Bulman 2005), suggesting that mineral nutrition could have been a major selection pressure acting on early land plants. Nitrogen is a critical resource that plants never evolved the ability to assimilate directly from plentiful atmospheric N2, with local microbial mineralization being a major bottleneck to productivity. However, basal land plants were not ‘stress-tolerators’ as they relied on turgor for support, lacking the specialized vascular and supporting tissues that also confer constitutive defence (toughness) against herbivory. They also lacked stomata and were restricted to humid, probably riparian, habitats.

The Silurian–Devonian terrestrial radiation (439–363 Ma b.p.) marks the transition from a reliance on the cellular stress response (e.g. the minimum stress proteome, energy-dissipation systems) to a combination of cellular and morphological stress-tolerance mechanisms. Mosses (Bryopsida) appeared in the early/mid-Silurian (Kenrick & Crane 1997) and were probably capable of suspending metabolism during drought. Proctor (2000) warns that this is often misinterpreted as the ‘ultimate expression of vascular-plant drought tolerance’ and that mosses are actually mesomorphic, ‘ectohydric’ plants that must maintain contact with external water for growth. Thus the cellular stress response presides during drought, but novel morphological adaptations govern growth, specifically the ‘ombrotrophic’ ability of the shoot to hold rainwater and absorb dissolved inline image. Moss ombrotrophy may dominate N cycling in primary succession, as N is assimilated and concentrated in shoot tissues to be released following senescence decades later, and areas devoid of moss-derived organic matter experience severe leaching (Bowden 1991; Jónsdóttir et al. 1995). (Note that optimal foraging is highly improbable for these plants as the shoot assimilates all resources, enlargement of which simultaneously increases both light and nutrient interception.) Ombrotrophic ecosystems are more dependent on climatic inputs than on substrate nutrient/water status, allowing colonization of exposed rock. Bowden (1991) measured inputs of 6·1 kg N ha−1 year−1 from precipitation and 0·6 kg N ha−1 a−1 from associated cyanobacterial N2 fixation – small amounts, but highly substantial if accumulated for millions of years. Thus the Siluro–Devonian radiation could be viewed as a similar primary succession, with advancing moss communities producing a suitable edaphic environment for later life forms. By the Mid Silurian, Cooksonia spp. had stomata and a simple vascular system (all advantageous for controlled nutrient uptake and transport), and probably lived in nutrient-limited ecosystems (Edwards 1996 and references therein).

Until the Devonian (409–363 Ma b.p.), ‘competition among plants was restricted by their tenuous hold on the abiotic landscape’ (Bateman et al. 1998). However, by this time soils were extensive enough to support more substantial life forms such as Equisetaceae. Extant Equisetum species in unproductive Arctic habitats have deep roots that convey nutrients from soil horizons beyond the reach of sympatric species, concentrating nutrients in tissues and forming nutrient-rich litter (Marsh et al. 2000). Equisetum also has mycorrhizal partners and rhizomes (Dhillion 1993; Andersson & Lundegardh 1999) – mycorrhizae were undoubtedly present by the Early Devonian, and probably much earlier (Remy et al. 1994). Although this genus has a later Cenozoic origin (Des Marais et al. 2003), and several species are successful in productive habitats, Equisetum demonstrates that a Devonian body plan had the potential for effective nutrient management and could strongly influence nutrient cycling. Equisetum species also exhibit enhanced constitutive defence in the form of a tough stele and silica deposits in epidermal cell walls (Holzhüter et al. 2003), augmenting the chemical defences also found in liverworts and algae (Perry et al. 2003). Devonian ecosystems may have included arthropod herbivores, although detritivores and their predators were probably the main animal component (Edwards 1996; Kenrick & Crane 1997).

Thus vascular plant strategies probably represent the culmination of a protracted liberation from severe nutrient limitation – this can explain why the terrestrial radiation took so long to get fully under way. Soil development was undoubtedly a prerequisite for the Devonian explosion of life forms, during which burgeoning global vegetation cover depleted atmospheric CO2 and increased O2 concentrations (Beerling & Woodward 1997). Limiting CO2 concentrations prompted higher stomatal densities, also allowing greater evaporative cooling and facilitating the evolution of laminate leaves that did not overheat despite greater interception of sunlight (Beerling et al. 2001). By the end of the Middle Devonian, competition for light probably favoured the evolution of lignin, lateral meristems (cambium), and thus tree life forms (Kenrick & Crane 1997). Subsequent to these physiologically driven radiations, regenerative innovations such as seeds (Late Devonian) and flowers (a Cretaceous response to earlier winged insects) facilitated incursions of novel tree life forms into an established canopy.

Nutrient limitation has also been a principal selection pressure during Cenozoic radiations (65 Ma b.p. to present). For the African grass genus Ehrharta (crown age 11·3–4·5 Ma b.p.), molecular phylogenetics demonstrates that all ancestral forms are slow-growing stress-tolerators (sensu Grime) native to oligotrophic (nutrient-deficient) habitats, from which faster-growing ruderal species later radiated, partly in response to more fertile substrates (Verboom et al. 2004). Similarly, primitive Bromeliaceae (≈65 Ma b.p.) are tough-leaved C3 herbs of humid, oligotrophic habitats, whereas all xerophytic CAM epiphytes are extant (Benzing 2000 and references therein). In this case, stress-tolerators from a chronically unproductive habitat laid the foundation for successful adaptive radiation into a range of even more unproductive niches that required further specialization. Indeed, nutrient-absorbing foliar trichomes (a defining feature of Bromeliaceae) maintain an absorptive function despite modification to mediate high light stress for extant taxa (Pierce 2005). Bromeliads could not have evolved the epiphytic/lithophytic habit were it not for this ability to absorb mineral nutrients via foliar trichomes and to use these meagre resources efficiently (Benzing & Renfrow 1980). These are also ombrotrophic plants for which optimal foraging cannot operate. Indeed, Benzing (2000) concludes that nutrient limitation encourages investment in persistent leaves at the expense of growth rate and reproduction, with this robust habit preadapting many species to further stressors such as drought and salinity. Furthermore, slow tissue replacement following disturbance places bromeliads at particular disadvantage, necessitating investment in defensive chemistry and ‘strengthening polymers’ (sensu CSR theory). This constitutive defence may be exaggerated in extant life forms such as the spiny subfamily Bromelioideae (reviewed by Benzing 2000).

Thus the cellular stress response and adaptations to nutrient limitation are both conservative responses, but nutrient limitation appears to have played a further role in the evolution of the stress-tolerant syndrome for modern terrestrial plants. This is an appealing hypothesis, because it can explain radiation from basal taxa relying on cellular stress-tolerance mechanisms and chemical defence, through true stress-tolerator phenotypes with additional constitutive defence, to highly specialized stress-tolerators with flexible photosynthetic modes and more effective physical defence (e.g. Cactaceae).

resolving the functional basis of real community processes

CSR theory is well supported by the above functional evidence. However, are CSR theory and the resource-ratio hypothesis useful? It could be argued that the resource-ratio hypothesis has a focused goal – aiming to explain how resource availability governs species composition along environmental gradients – and appears to provide a coherent solution. However, the environment includes short-term fluctuations that can be significant survival risks, and includes stressors other than resource limitation. The failure of the resource-ratio hypothesis to consider these aspects of the environment as selection pressures can explain why it conflicts with real phenotypes in chronically unproductive habitats. Additionally, the model of root competition inherent to the resource-ratio hypothesis is too simple to approximate real systems (Craine et al. 2005). More importantly, it lacks unambiguous experimental support (Gleeson & Tilman 1994), with optimality models contradicting real plant responses to nutrient addition in chronically unproductive habitats (partitioning to roots is favoured in response to higher nutrient availability, enhancing luxury consumption in support of future growth; Van Wijk et al. 2003). The only conceivable situation in which the resource-ratio hypothesis could be useful is on planets that do not experience environmental fluctuations imposed by rotation and precession.

CSR theory appears to provide a realistic theoretical framework for the functional classification of terrestrial plants: species endemic to unproductive habitats rely on augmenting metabolism (S) or, in extremis, avoiding disturbance (R), and those of productive habitats rely on foraging (C). However, in nature all plants possess a minimal stress response. Conversely, stress-tolerant species are capable of some morphological plasticity, as illustrated by root proliferation of the late-successional alpine Carex curvula in response to nutrient patches (Körner 1999). Thus absolutely pure primary strategies, such as the extreme C, S and R corners of Grime's triangle, are unlikely to exist in nature – actual adaptive strategies represent variations within this theoretical framework. Can real plants, growing in situ, be classified according to this system? CSR classification (Hodgson et al. 1999) achieves just this, and has been used successfully to investigate the mechanisms underpinning a primary succession on a glacier foreland (Caccianiga et al. 2005). This developing methodology undoubtedly provides the most promising opportunity for future research into vegetation processes.


The resource-ratio hypothesis disagrees with a vast body of literature representing a contemporary, multidisciplinary view of stress, stress responses and terrestrial adaptive radiation. In contrast, CSR theory includes a realistic view of adaptive radiation in response to stress and is useful for understanding community processes.


We thank Roberta M. Ceriani for discussion during the conception of this article, and Alessandra Luzzaro and Rossella De Andreis for pointing out several references. S.P. was supported by the Native Flora Centre (Centro Flora Autoctona), via the University of Insubria.