Functional dynamics of plant growth and photosynthesis – from steady-state to dynamics – from homogeneity to heterogeneity

Authors


U. Schurr. Fax: +49 2461612492; e-mail: u.schurr@fz-juelich.de

ABSTRACT

Plants are much more dynamic than we usually expect them to be. This dynamic behaviour is of paramount importance for their performance under natural conditions, when resources are distributed heterogeneously in space and time. However, plants are not only the cue ball of their physical and chemical environment. Endogenous rhythms and networks controlling photosynthesis and growth buffer plant processes from external fluctuations. This review highlights recent evidence of the importance of dynamic temporal and spatial organization of photosynthesis and of growth in leaves and roots. These central processes for plant performance differ strongly in their dependence on environmental impact and endogenous properties, respectively. Growth involves a wealth of processes ranging from the supply of resources from external and internal sources to the growth processes themselves. In contrast, photosynthesis can only take place when light and CO2 are present and thus clearly requires ‘input from the environment’. Nevertheless, growth and photosynthesis are connected to each other via mechanisms that are still not fully understood. Recent advances in imaging technology have provided new insights into the dynamics of plant–environment interactions. Such processes do not only play a crucial role in understanding stress response of plants under extreme environmental conditions. Dynamics of plants under modest growth conditions rise from endogenous mechanisms as well as exogenous impact too. It is thus an important task for future research to identify how dynamic external conditions interact with plant-internal signalling networks to optimize plant behaviour in real time and to understand how plants have adapted to characteristic spatial and temporal properties of the resources from their environment, on which they depend on.

INTRODUCTION

Plants are sessile organisms that lack homeostatic mechanisms and thus cannot usually avoid exposure to their spatially and temporally fluctuating environment. For extreme environmental conditions, plants form durable structures such as seeds or position-resilient organs at more sheltered places. This strategy has led to the typology of life forms (Raunkiaer 1934) and is the basis of adaptation to specific habitats in many plant species. Besides these adaptations of their habitus, plants depend on special anatomical, morphological or physiological features such as drought-adapted xeromorphic leaves or Crassulacean acid metabolism (CAM) photosynthesis to withstand stressful environmental conditions.

However, even in conditions that do not exert immediate stress, plants perform in a temporally and spatially heterogeneous environment. Selection of more efficient structures or functions to exploit heterogeneously distributed resources in the evolution is evident. Complex interactions of the fluctuating environment with the dynamics of plant processes are crucial for the performance of plants in natural biogeosystems as well as for crops. Despite the enormously important role of the dynamics of plant behaviour in natural and managed biogeosystems, the understanding of plant functional dynamics is far from being systematic, quantitative or sufficient to predict plant performance in future climatic scenarios (Osmond et al. 2004) or to be exploited to reach higher efficiency of resource use in crops.

This review highlights recent analyses of the dynamics of growth and photosynthesis, which are two plant processes that are of major importance for adaptation of plants to their environment. Both processes are highly dynamic in their nature, depend strongly on environmental conditions and are regulated by complex feedback processes linking environmental conditions with endogenous status. Some subprocesses of growth and photosynthesis are controlled by endogenous rhythms, and recent results support the hypothesis that correct tuning of plant internal processes to the external day–night rhythm is beneficial for plant performance and even causes competitive advantages relative to less-tuned specimens (Dodd et al. 2005).

However, both processes do also show distinct differences in their dynamic response to fluctuating environments. Growth is much more buffered from direct environmental impact than resource acquisition processes such as photosynthesis. While adaptation to environmental conditions and growth dynamics of above-ground parts seem to be dominated by endogenous rhythms in most situations (Walter & Schurr 2005), photosynthesis is dependent on the immediate photon flux. Photosynthetic efficiency dynamically acclimates to fluctuations in the environment, and non-photochemical dissipation mechanisms protect the photosynthetic apparatus from excessive irradiance. Growth, on the other hand, is responsible for the geometry of plants and of canopies to varying resource availability and thus plays a major role for the efficiency of resource acquisition. Differences in the spatial and temporal fluctuations of below-ground and above-ground resources can thus explain the contrasting sensitivity of growth of leaves and roots to environmental cues.

Analysing and understanding dynamic processes in plants requires methods that quantify dynamics within the functional and intact system. Major steps forward in recent years in understanding the dynamics of growth and photosynthesis were linked to the development and application of new methods to quantify their spatial and temporal heterogeneity. These methods have allowed plant research to take further steps to understand the dynamic nature of plants and the intimacy of the interaction with their fluctuating environment. It will be of crucial importance for basic research and – increasingly – for applied plant sciences to use this knowledge of how plants perceive dynamic cues from the environment and how they use their internal signalling networks to produce integrated and coordinated short-term and long-term responses. We predict that strategies to cope better with spatially and temporarily heterogeneous environments are the basis for higher efficiency that have been selected during evolution and which will improve the performance of crop varieties in the future.

GROWTH

Acclimation and phenotypic plasticity

Plants have an enormous plasticity to respond to less predictable and short-term fluctuations of environmental factors by altering growth of individual organs or by altering the distribution of growth among their modules. They can acclimate towards nutrient deficiency by altering the shoot–root ratio (Scheible et al. 1997) or by increasing shoot elongation in response to shade (Smith 1995). Their ability to reorient growth by differential growth of roots, leaves or internodes in response to gradients of nutrients, light, chemicals or even gravity has fascinated scientists for a very long time (Darwin 1880; Perrin et al. 2005). Such responses are important factors in the ‘Darwinian war’ for optimized fitness, finally leading to the selection of the most efficient and plastic plant species over generations. However, it has to be emphasized that the efficiency of a plant can only be defined properly within a certain spatial and temporal environmental setting.

Growth dynamics

The dynamic nature of the developing plant in response to the temporally changing and spatially heterogeneous environment and its interaction with the permanently developing internal status of the plant has to be quantified as a first step towards identification of the rules of phenotypic plasticity. With the development in recent years of new methods for quantifying growth processes in spatial and temporal scales relevant for molecular responses of plants, the task is to combine modern genomic, proteomic and metabolomic techniques with quantitative physiology to build a truly systemic understanding of plant acclimatization and adaptation to the dynamically changing environment.

Describing ‘growth’ and ‘expansion’ of plant tissues or plant organs can use a wide range of different properties: most studies investigating plant growth via destructive measurements (i.e. harvesting plants or plant organs) relate growth to an increase in dry matter of the plant. This approach cannot be applied in non-destructive measurements that quantify short-term changes. At high resolution irreversible increase of tissue, organ or plant biomass is commonly measured in relation to properties such as organ length (root), area (leaf) or volume. In this paper, we chose the relative elemental growth rate (increase in area per unit area per time; Silk 1984) as the adequate measure for plant growth activity.

Temporal dynamics of leaf growth

The timing of leaf growth during the day differs considerably between plant species and presumably reflects the integration of the particular photosynthetic resource capture and redistribution (source–sink) properties of the species. While a number of plants such as Ricinus communis (Walter, Feil & Schurr 2002) or Nicotiana tabacum (Walter & Schurr 2005) show maximal leaf growth at dawn, Populus deltoides grows strongest at dusk (Walter et al. 2005; Matsubara et al. 2005a) and Glycine max shows maximal growth rates in the middle of the night (Ainsworth, Walter & Schurr 2005). On the basis of transcriptomic data, it has recently been hypothesized that differences in the phasing of those rhythms might be due to differences in the timing of cytoplasmic versus vacuolar growth activity (Matsubara et al. 2005a). Leaves of monocot species generally show highest growth rates in the middle of the day (Watts 1974; Seneweera et al. 1995); the only study dealing with diel growth phenomena in CAM plants so far showed maximal growth rates around noon (Gouws et al. 2005). However, the maximum of growth rate in these plants might shift with water availability. In the latter case, it is likely that there is an explicit relationship between the distinctive photosynthetic carbon acquisition patterns of CAM and other requirements for growth, such as turgor and cytoplasmic pH.

Environmental conditions can alter the diel pattern of leaf expansion. Drought stress (Schurr et al. 2000), nutrient deficiency and various temperature regimes (Ainsworth et al. 2005) affect only the amplitude but neither the phasing nor the shape of the diel course of leaf expansion. In contrast, plants growing in elevated atmospheric CO2 concentrations showed a transient growth reduction in the afternoon in growing cottonwood leaves, which was not present when plants were grown in ambient CO2 concentrations (Walter et al. 2005). This response is not a direct physical response to higher CO2 concentrations, but most likely a case in which altered photosynthetic influx and source–sink carbon relations affect growth patterns of plants.

Short-term drastic alterations of growing conditions cause rapid but transient changes in leaf expansion. Such direct impact of, for example, external light variations, on growth processes have been observed in many species (e.g. Hsiao, Acevedo & Henderson 1970; Christ 1978): When lights are switched off, expansion strongly increases within minutes (relative expansion rates of up to 10%/h (Walter & Schurr 2005). This disturbance effect is followed by an exponential decrease to the original growth rate with a decay time of ≈ 15 min. When lights are switched on, the opposite behaviour is observed. In contrast to the diel growth activity, these transient growth spikes are brought about by a uniform variation throughout the entire leaf (Walter & Schurr 2005). Such spatial and temporal characteristics of the response are indicative of direct interactions with growth processes. For example, light intensity changes cause rapid alterations in cell wall pH (Mühling et al. 1995) and stomatal conductance (Mott & Buckley 2000) and thus may affect expansion directly via cell wall extensibility and leaf water relations, respectively. Transient responses of growth with similar kinetic features have even been observed, when turgor in the growing cells was artificially increased by means of the root pressure chamber. Despite the fact that turgor was doubled, leaf growth was only transiently increased and reached previous growth rates ≈ 20 min after increasing turgor (Schmundt et al. 1998).

However, ecological consequences of such short-term responses of leaf growth to short-term environmental fluctuations can be significant. Two congener Chamaecyparis species that are native to habitats of different light fluctuation regimes in the cloud forest of Taiwan have been shown to differ in their response to short-term light variations (Lai et al. 2005). The species adapted to grow in the understorey of the cloud forest was able to respond very rapidly to changes in intercepted light intensity, but leaves shrank significantly throughout morning hours when this species was exposed to constantly high light intensities throughout the day. The other species, adapted to grow in large forest gaps, could make better use of constant high light conditions, but grew worse in the low light conditions of the understorey. Such growth response patterns need to be evaluated further in terms of the photosynthetic responses of these ecotypes.

Spatial heterogeneity of leaf growth

Historically, time-consuming measurements were necessary to quantify base–tip gradients of leaf growth over timescales of days, and they were thus only established for a relatively small number of plant species (Smirnov & Zhelochovtsev 1932; Avery 1933). These classical observations influenced generations of plant scientists and led to the paradigm of post-emergent leaf development progressing as a basipetal wave in leaves of all dicot plants (Van Volkenburgh 1987). The base–tip gradient develops in parallel to the maturation of cells that occurs first in the leaf tip and then proceeds to the leaf base (Avery 1933; Foster 1936). Additional measurements of the development of gradients of cell size along the lamina (e.g. Walter, Roggatz & Schurr 2003), of cell division rate or meristematic activity by kinematic growth analyses (Silk 1984) and even on cellular differentiation, such as endoreduplication (Beemster et al. 2005), have supported tip-to-base gradients as a characteristic feature of dicot leaf growth.

Significant technological progress has eased the analysis of spatial patterns of growth in recent years. Traditional studies of growth dynamics were based on length measurement devices that registered the linear increase of leaf length precisely, but could not deliver quantitative data on areal growth and could not resolve spatial variations in expansion activity within the leaf lamina. Modern digital image sequence processing methods made it feasible to study growth of leaves (Schmundt et al. 1998) and roots (Walter et al. 2002; van der Weele et al. 2003) with high spatial and temporal resolution at the same time. For example, a base–tip gradient is the most prominent spatial pattern to be observed throughout the entire diel course of tobacco leaf growth (Walter & Schurr 2005).

On the basis of these new techniques, recent work has shown that a base–tip gradient is not a general characteristic of dicot leaf growth. In several species, including important crops such as soybean and poplar, no tip-to-base gradients of expansion growth have been observed (Walter, Rascher & Osmond 2004; Ainsworth et al. 2005; Matsubara et al. 2005a). The paradigm of a basipetal wave of dicot leaf growth as based on investigations in a small number of species will need to be revised on the basis of the evidence gained with the new techniques.

Small-scale patchiness and temporal inhomogeneities of expansion have been resolved by digital image sequence analysis with temporal resolutions of minutes, but smooth base–tip gradients can be observed when mean growth rates over several hours are calculated (Walter et al. 2002; Matsubara et al. 2005a; Walter & Schurr 2005). It is clear that small-scale expansion processes give rise to strong mechanical strains inside the growing tissues. Such small-scale changes in local expansion are highly relevant in the context of growth control mechanisms. Any irregularity of expansion in a leaf growing in a plane will result in strains that will eventually cause buckling of the leaf surface, if they are not compensated by appropriate spatial and temporal expansion in neighbouring parts of the leaf. Leaf veins most likely play a crucial role in the biomechanics of leaf growth, as they are the major biomechanical feature in the tissue. Biomechanics might thus provide an interesting alternative or contributor in controlling growth and shape in planar plant organs such as leaves in a similar way as previously proposed for meristems (Green 1992). External straining forces have already been shown to affect those small-scale growth variations (Walter et al. 2002). In addition, the heterogeneity of growth rate revealed by these methods shows spatial and temporal differences in source–sink relationships and local photosynthetic activities. This might indicate connections with stomatal patchiness, internal CO2 diffusion in homobaric and heterobaric leaves, and vascular delivery of sugars to sink leaves from source leaves.

Growth of the entire shoot leaves and acclimation of the three-dimenstional (3D) structure of the leaf canopy

The individual growth of leaves is embedded into the sequence of growth events occurring along the entire stem of a plant (Walter & Schurr 1999). At present, the link between growth dynamics of individual leaves and the entire leaves of a shoot is technically difficult to address. However, there is a clear need for future measurements to quantitatively analyse the behaviour of the entire leaf system of a plant or even the entire canopy at the stand level. Such quantitative analysis of the dynamics of the 3D geometry of plants will not only contribute to understanding 3D growth processes, but will also be most important to better understanding how a growing canopy acclimates to optimal light harvesting (Küppers 2003). An important task will be to identify how growth dynamics along the 3D geometry of a plant shoot addresses the requirement to harvest light, how this is adapted to variable light conditions at the stand level and how canopies balance their resources between structural stability and reaching high-light regions. It has been shown recently that parameters related to growth such as specific leaf area respond in a very flexible manner to the within-canopy variation of the light climate in different species (Niinemets, Kull & Tenhunen 2004; Niinemets 2004; Niinemets, Tenhunen & Beyschlag 2004).

Techniques for measuring total leaf area quantitatively without indirect approaches of models have been developed. For rosette plants such as Arabidopsis thaliana, it is feasible to deduce approximations of leaf area from single images of the leaf rosette. Such approaches can also be used to set up high- to mid-throughput screening systems to determine daily growth rates (Leister et al. 1999). Quantification of the true leaf area of rosette plants, however, clearly needs to address the problems of leaf overlap and non-horizontal leaf angles. Both problems are of paramount importance for the youngest, most actively growing leaves. They are positioned above older leaves and they are oriented more vertically than older leaves. The only practical approach to address those problems is to analyse plant surfaces in three dimensions by stereoscopic or tomographic methods.

The 3D structure of a plant can be analysed using a device based on three perpendicular magnetic fields (Sinoquet & Rivet 1997). Although the resolution of this method for measuring growth rates is too low to quantify short-term acclimations, it is a valuable tool to determine foliage distribution and light interception in 3D digitized trees (Sinoquet et al. 2005). Higher resolutions can be obtained via stereoscopic approaches that calculate the morphology of surface structures from matching object patterns on images that are taken from two or more cameras or camera positions at slightly different angles (Scharr & Küsters 2002). Another method for quantifying 3D data of plant surfaces uses laser range finders that scan the distance from the device to each spot of a plant surface. Such systems are used at different scales to analyse forest canopies from aircraft (Lefsky et al. 1999) or to analyse Arabidopsis surfaces (Kaminuma et al. 2004). Tomographic techniques like X-ray computer tomography or nuclear magnetic resonance imaging revolutionised medical diagnostics. However their potential to quantify changes of plant biomass and the growth of above- and below ground organs in plants has yet to be fully established (Heeraman, Hopmans & Clausnitzer 1997; Peuke et al. 2001).

Root growth dynamics and heterogeneities of the soil environment

Roots are exposed to an environment that is very different from the atmospheric environment of the shoot with respect to the spatial and temporal fragmentation of their habitat. Root growth is highly sensitive to environmental fluctuations and is organized in a different way than leaf growth: endogenous patterns of root growth are not present in growing root tips of all species that have hitherto been investigated (cherry, rice, sorghum, maize and tobacco as observed by Head 1965; Iijima et al. 1998; Walter et al. 2002; Walter & Schurr 2005). We can thus anticipate that relationships between root growth and the photosynthetic resource utilization may often be complex and that these will vary with the development of photosynthetic capacities and storage systems. In contrast to growing leaves, root growth is strongly and continuously affected by changing environmental factors such as temperature (Pahlavanian & Silk 1988; Walter et al. 2002), water deficit (Fan & Neumann 2004) or nutrients (Walter, Feil & Schurr 2003).

The spatial distribution of relative elemental growth rate within the root growth zone is more complex, but also more conserved between species than in the dicot leaf. Within the root meristem, the expansion growth rate is relatively low as repetitive cell divisions take place at rates between one division per day and one per week (Silk 1984). Cell division and regrowth of daughter cells to the initial mother cell size result in a biomass increase of a factor of two. This is the reason for expansion growth rates of less than 5%/h in the meristematic region. Expansion growth rate increases with the onset of gross cell elongation, soon reaching a maximum of ≈ 30%/h in the middle of the growth zone for most species (Brumfield 1942; Goodwin & Stepka 1945; Pahlavanian & Silk 1988; Walter et al. 2002). Basal to the zone of maximal expansion growth activity, the relative elemental growth rates decline steadily as the cells develop to maturity. The basic shape of this growth rate distribution can be modelled a priori using an approach in which a cell file is exposed to two counteracting virtual hormones entering the root growth zone from the apical and the basal side (Chavarria-Krauser & Schurr 2004; Chavarria-Krauser, Jäger & Schurr 2005).

Environmental impact on growth rates and distribution of relative elemental growth rate within the root growth zone have been identified in many cases (Pritchard 1994). In maize roots, often two transient maxima can be distinguished that alter their relative intensity with altering external nutrient availability (Walter et al. 2003). While in nutrient-free solution the apical peak is higher, full-strength nutrient solution leads to a more pronounced basal peak of the growth rate distribution.

Growth and development of the entire root system

The role of the growth dynamics of individual root tips in the performance of the entire root system is of enormous importance for the overall efficiency of the root system. Root systems have high developmental plasticity controlled by intrinsic and response pathways (Malamy 2005), and the degree to which they are branching can be altered easily by, for example, local nutrient availability (Zhang & Forde 2000). Root growth intensity is mostly dependent on hexose contents within the root (Freixes et al. 2002); root system architecture determines the ability of the plant to capture and transport resources from the soil (Thaler & Pages 1998). Thus, the 3D arrangement of root structures and associated functions such as nutrient and water uptake are of general importance for the efficiency of plant root systems to utilize spatially and temporally fluctuating resources.

Control mechanisms of growth dynamics in a fluctuating environment

Growth processes are regulated in the plant on a plethora of system levels, ranging from biomechanical restraints governing leaf form (Niklas 1999) to the endogenous genetic control at the transcriptome level in roots (Birnbaum et al. 2003; Bassani, Neumann & Gepstein 2004) and leaves (Trainotti, Pavanello & Casadoro 2004; Matsubara et al. 2005a), and finally on a whole system level by long-range signals (Heckenberger, Roggatz & Schurr 1998). It will be necessary to identify how environmental cues affect the network of endogenous growth-controlling processes to understand plant performance in fluctuating environments.

Control processes at different levels of the system

Molecular control processes of cell division (e.g. Beemster, Fiorani & Inzé 2003) and cell expansion (e.g. Vissenberg et al. 2000; Cosgrove et al. 2002) have been studied intensively in recent years. However, the interaction of these internal processes with the dynamics of external conditions is still scarcely known. Recent results propose that oscillations of cytoplasmic free calcium play an important role at second messenger level, linking external parameters such as day length with endogenous signalling pathways (Love, Dodd & Webb 2004), which are important in synchronizing growth processes with the circadian clock (Dodd et al. 2005).

The biophysical control of leaf growth responses to light involves increasing wall extensibility in Phaseolus vulgaris rather than changes in osmotic potential in the cells (Van Volkenburgh & Cleland 1980, 1981). Aquaporins might also play an important role in the biophysical control of leaf development, by regulating hydraulic conductance and turgor of the symplast (Siefritz et al. 2004).

Insights gained at different levels of investigation have to be collected and condensed in coming years to establish a conclusive view of the regulation of growth dynamics. Plant architectural models can provide insights into mechanisms of plant development and are beginning to address the relationship between patterns of gene expression and the resulting plant form (Coen et al. 2004; Prusinkiewicz 2004).

While at each scale of the control, a large number of results have been obtained, most studies refer to steady-state conditions and do not take into account the spatial heterogeneity of growth processes described above. Therefore, it will be of crucial importance to address questions such as:

  • • Why do diel cycles of leaf expansion differ between species?
  • • Why are spatial distributions of growth rates in root growth zones similar despite different sizes of organs – and hence different numbers of cells forming them – in different plant species?
  • • What is the role of growth heterogeneities and oscillating movements in roots and leaves, and how does the signalling network eventually guarantee, for example, the symmetrical growth of organs?
  • • How does a leaf manage to expand with balanced intensity in the vein and interveinal area although the tissues are at different developmental stages?
  • • What is the link between short-time growth fluctuations in response to strong changes of environmental conditions and the cellular and tissue mechanisms of growth?

Steps towards an integrated understanding of the regulation of growth dynamics and carbon relations

It would by far exceed the scope of this review to attempt a conclusive view of the regulatory network that controls growth. Yet, in the context of connecting growth with photosynthesis, as a starting point it makes sense to address linkages between the metabolism of carbohydrates – the transportable energy currency produced in photosynthesis – and growth dynamics of leaves and other tissues.

In extreme cases, it has been shown that alterations in the diurnal course of carbohydrate metabolism correlate with changes in the diurnal growth patterns in transgenic potato plants (Kehr et al. 1998). Walter et al. (2005) showed that elevated CO2 supply increased photosynthesis in growing leaves, temporarily depleted carbohydrate pools and changed the diurnal pattern of leaf growth. Combining high-resolution growth analyses with the wealth of information on diurnal carbon metabolism in growing tissues and the source tissues supplying them (e.g. Kemp & Blacklow 1980; Matt et al. 1998; Geiger, Servaites & Fuchs 2000; Chia et al. 2004; Walter & Schurr 2005) will be challenging. A possible connection between the carbohydrate pool of a growing leaf and the extensibility of the cell wall was indicated recently by experiments showing that pronounced internal strain forces and energy supplied from starch breakdown may be a requirement for regular growth patterns (Walter et al. 2002). Moreover, the observation that CAM and C4 plants seem to grow strongest at daytime, while growth in C3 plants seems to take place preferentially at dusk and dawn, indicates that the diel variation of photosynthesis, CO2 assimilation and carbohydrate production is systematically interacting with diel variations of growth activity (Gouws et al. 2005; Walter & Schurr 2005). The coincidence of the constant supply of carbohydrates to the root (Walter et al. 2003a) with the absence of diurnal growth patterns is another aspect for future investigations on the link between growth and carbon relations.

Especially small-scale growth fluctuations might be governed by immediate carbohydrate availability produced via photosynthesis. When Populus deltoides leaf growth was analysed either at a full sun or under a shaded branch, the overall diel growth patterns were similar in both conditions. However, heterogeneities of growth rates were higher in the shade, indicating that the local production of carbohydrates in the growing leaf might be beneficial for stabilizing leaf growth (Walter et al. 2005). However, these examples make clear that no simple direct link between carbohydrate metabolism and growth dynamics can be expected.

Photosynthesis

Photosynthetic light conversion – regulation at various levels

Photosynthesizing leaves are often exposed to a quite variable stream of photons. The efficiency with which absorbed photons are finally used for photosynthetic electron transport and carbon fixation is highly regulated. It is necessary to define and clarify terminology and processes before analysing the dynamic processes involved in this central process for plant performance.

Efficiency of light reaction is commonly measured using chlorophyll fluorescence techniques or oxygen evolution and denoted photosynthetic ‘quantum efficiency’ (also ΔF/Fm′ using fluorescence terminology), which quantifies the conversion of absorbed photons to transported electrons. Quantum efficiency of photosynthesis varies between 0.83 and close to zero at leaves of dark-adapted higher plants and depends primarily on light intensity (Björkman & Demmig 1987; Rascher, Liebig & Lüttge 2000). Environmental constraints additionally may directly affect quantum efficiency of light reaction as recently shown in a study analysing the effects of drought in the tropical rainforest mesocosm of Biosphere 2 (Rascher et al. 2004).

Efficiency of carbon fixation is commonly termed ‘photosynthetic yield’ or ‘photosynthetic efficiency’ and describes the moles of photons needed to fix one mole of CO2. Photosynthetic efficiency varies greatly between the theoretical minimum of 4 and values more than 50. In addition to being light driven, stomatal effects, leaf-internal CO2 concentration, and the amount and activity of ribulose 1·5-bisphosphate carboxylase/oxygenase (Rubisco) become important. In the past decades of photosynthetic research, a large number of studies investigating photosynthetic efficiency of chloroplast, cells and leaves in relation to physiological traits and environmental constraints have become available, and thus this topic will not be further reviewed here (see various review articles and books, e.g. Schulze & Caldwell 1995).

At the canopy level, efficiency of carbon fixation is termed ‘light use efficiency’ (LUE). LUE characterises whole canopy and refers to the projected ground surface rather than leaf area and describes the net canopy CO2 fixation per incoming photosynthetic radiation above the canopy. The spatial variation of LUE results in enormous variations of net photosynthetic productivity, which ranges from 30 to 1000 g C m−2 in different ecosystems (Scurlock et al. 1999).

All three processes (electron transport rate, leaf-level carbon uptake rate and net ecosystem photosynthesis) are ultimately light driven and provide carbohydrates, but their efficiency and regulatory properties are not linearly connected. Despite some results, obtained under controlled conditions, showing that electron transport and carbon uptake rate can be tightly coupled (Weis & Berry 1987), no general transfer functions between light absorption and carbon gain can be developed.

Photosynthesis in fluctuating light

Most of the mechanistic studies on photosynthesis were performed under laboratory conditions and at constant light conditions. Natural light regimes, however, are far from being constant. Light regimes at leaves can change within milliseconds if, for example, plants are exposed to light flecks in an understorey – not only in forests, but also in crop stands. Plants dynamically acclimate their photosynthetic system to these fluctuations in light and are well adapted to such natural conditions. On the one hand, they optimize their energy gain for carbon fixation; on the other, the sensitivity of the photosystem to over-energetization requires protection of the light harvesting systems from photo-damage at high irradiances. Thus, plants have evolved a variety of non-photochemical quenching mechanisms that are either constitutively active or activated on demand (Demmig-Adams & Adams 1992; Niyogi 2000).

The balance between photochemical charge separation and non-photochemical protection is also an ideal example how physiological mechanisms can interact and respond to fluctuations on different timescales. Very short changes in light intensity, for example, light flecks in understoreys, can be buffered directly by the primary processes in the electron transport chain: energy is stored in the proton gradient across the thylacoid membrane (Pearcy 1990) and can even be used for post-illumination carbon fixation (Kirschbaum et al. 1998). Light intensity fluctuations on the order of minutes may cause over-energetization and damage of the photosynthetic apparatus and lead to the activation of a biochemical pigment protection system. The xantophyll cycle dissipates excessive electrons during periods of high light, but can also be inactivated within minutes to allow maximum photosynthetic electron transport at low light conditions (Demmig et al. 1987; Demmig-Adams & Adams 1992, 1996; Horton, Ruben & Walters 1996). In shade leaves, a second pigment cycle was discovered recently and it involved the conversion from lutein to lutein-epoxide. The conversion and energy quenching operates in analogy with the xanthophyll cycle, but this so-called Lx-cycle is fully activated after hours of high light exposure and was described to lock-in this high light state (Matsubara et al. 2005b). The Lx-cycle is thus discussed as an adaptive mechanism to protect shade leaves from longer-lasting high light conditions, which may occur, for example, during gap formation. Flexible reaction to variable environmental constraints is also discussed as the main adaptive feature of CAM photosynthesis, which ‘is noted to be a strategy for variable, flexible and plastic niche occupation rather than lush productivity’ (Lüttge 2004).

At the canopy level, additional processes superimpose the physiological regulation of leaf photosynthesis. New regulatory properties emerge because of the 3D structure of natural canopies. Generally, physical processes such as gradients and spatio-temporal fluctuations in vapour pressure difference (VPD), temperature, turbulent air movement and light intensities become important. Leaf structure and morphology adapts to the different environmental conditions in a canopy by, for example, developing sun and shade leaves as morphological adaptations and acclimations on various timescales (Terashima & Hikosaka 1995). Typical sun or shade characteristics generally cannot be altered after leaf development is completed and thus biophysical and biochemical mechanisms such as the xanthophyll cycle have been evolved to dynamically activate photoprotective means to rapidly changing environments. Linked to the different demands during development, these protection mechanisms are organized differently in mature and growing tissues (Greer & Halligan 2001).

Spatial heterogeneity of photosynthetic efficiency on single leaves

The photosynthesis of higher plants was long regarded as being homogeneously distributed over single leaves. This dogma was destroyed during recent years with the development and application of chlorophyll fluorescence techniques, which allowed non-destructive quantification of quantum efficiency of photosystem II and of non-photochemical energy dissipation in attached leaves (Genty, Briantais & Baker 1989; Schreiber & Bilger 1993; Schreiber, Bilger & Neubauer 1995; Maxwell & Johnson 2000). Chlorophyll fluorescence imaging techniques were first developed in the late 1980s (Daley et al. 1989). Factors causing heterogeneity of photosynthesis in leaves are environmental factors with local impact, for example, plant pathogens (Balachandran, Osmond & Daley 1994; Balachandran et al. 1997; Lichtenthaler et al. 1997) and severe water stress (Osmond, Kramer & Lüttge 1999). Endogenous structures and processes also contribute to heterogeneity of photosynthesis in leaves and range from variegation of leaves (Osmond et al. 1998), heterogeneity of phloem loading and unloading (Siebke & Weis 1995b) to patchy stomatal opening (Cardon, Mott & Berry 1994; Siebke & Weis 1995a). An extraordinary example of dynamic, non-anatomy related heterogeneity was shown during the endogenous circadian rhythm of the CAM plant Kalanchoë daigremontiana. Chlorophyll fluorescence imaging techniques showed that the metabolic CAM cycle is expressed in dynamic patterns of independently initiated variations in photosynthetic efficiency across single leaves. Randomly initiated patches of varying photosynthetic efficiency, which were propagated over the leaf in wave fronts, dynamically expanding and contracting clusters, and clearly dephased regions were observed (Rascher et al. 2001; Rascher & Lüttge 2002; Rascher 2003).

Spatial variations of photosynthesis also occur in growing leaves either in the form of a distinct base–tip gradient (Meng et al. 2001) or as development-related differences between veinal and interveinal tissues (Walter et al. 2004).

Commercial fluorescence imaging systems are currently available, and chlorophyll fluorescence imaging has proven to be an accepted method for monitoring the spatial variations of the physiological status of light reaction and electron transport rates under a wide range of conditions (Nedbal et al. 2000; Chaerle & van der Straeten 2001; Leipner, Oxborough & Baker 2001; Osmond & Park 2001; Rolfe & Scholes 2002). However, the links to spatial variations in carbon assimilation and leaf growth remain sketchy at the best. Stomatal patchiness (Mott & Buckley 2000; Genty & Meyer 1995; West et al. 2005) and the emerging awareness of lateral CO2 diffusion inside leaves (Pieruschka, Schurr & Jahnke 2005) makes it difficult to transfer data on the localisation of electron transport processes to spatio-temporal variations in carbon uptake.

Heterogeneity of canopy photosynthesis

Molecular methods and spatially explicit physiological measurements are currently revolutionizing our understanding from the nanosecond processes of biochemical and biophysical regulation on the scale of molecules to the spatio-temporal processes on living leaves. In contrast, our knowledge of variations of photosynthesis on the canopy or ecosystem level is still fragmentary. Heterogeneity arises from diverse contributions of plant-mediated carbon and water vapour fluxes to core ecosystem functions such as energy transfer, responses to changing environment, stress and nutrients (Osmond et al. 2004; Rascher et al. 2004).

Limited canopy access and the limited scale of observations possible with portable instrumentation make it difficult to scale the inherent heterogeneity of leaf photosynthesis to canopy or ecosystem units. Eddy covariance measurements have contributed greatly to our understanding of plant-mediated exchange processes and photosynthetic carbon fixation of different ecosystems in the course of several years. However, global maps of the spatial and temporal distribution of photosynthetic carbon fixation (and evapo-transpiration) are needed to account for the observed variability of photosynthetic carbon fixation and biomass production of terrestrial plant ecosystems. Such maps of LUE cannot be extrapolated from spot measurements at single eddy sites (high temporal resolution, but no spatial resolution), but may be derived from remote sensing data. These methods have the potential to quantify the LUE of plant ecosystems directly and provide information about their spatial dynamics and temporal heterogeneity (Field, Gamon & Peñuelas 1995; Asner 1998).

Hyperspectral reflectance measurements show increasing potential for remote sensing of plant ecosystems. The photochemical reflectance index (PRI) was developed to serve as an estimate of photosynthetic LUE (Gamon, Filella & Peñuelas 1993; Peñuelas, Baret & Filella 1995a; Gamon, Serrano & Surfus 1997). This normalized difference reflectance index uses two wavebands: 531 nm, which is affected by the de-epoxidation of violaxanthin to zeaxanthin, two pigments of the xantophyll cycle (Demmig et al. 1987; Demmig-Adams & Adams 1992, 1996; Horton et al. 1996), and 570 nm, which remains unaffected by the de-epoxidation reaction (Gamon, Peñuelas & Field 1992). The PRI has been successfully used to detect changes in photosynthetic efficiency from the leaf to the ecosystem level (for an overview, see literature cited in Nichol et al. 2006; Rascher et al. in press). Nichol et al. (2000, 2002) showed that the PRI of a homogeneous forest correlates with ecosystem CO2 uptake. However, substantial effort is still needed to scale these results to heterogeneous canopies, as absolute PRI values may vary greatly between species (Guo & Trotter 2004) and are affected by leaf angle distribution and seasonal changes in canopy structure (Barton & North 2000; Filella et al. 2004).

Direct measurement of chlorophyll fluorescence at a distance has been an objective of many remote sensing programs (Field et al. 1995) and both passive (Carter, Theisen & Mitchell 1990; Carter et al. 1996), and laser-based active methods (Cerovic et al. 1996; Rosema et al. 1998; Corp et al. 2003) have been developed. Early progress, especially with the ratio of fluorescence F690/F730 (Günther, Dahn & Lüdeker 1994), and evaluation of excitation energy requirements (Rosema & Zahn 1997) promoted further developments based on steady-state levels of fluorescence (Flexas et al. 2000). Another active technology, laser-induced fluorescence transients (LIFT), uses a fast repetition rate technique (Kolber & Falkowski 1993; Kolber, Prasil & Falkowski 1998) to actively excite the photosynthetic apparatus and to realize the full potential of chlorophyll fluorescence analysis. Recently, the first prototype of a LIFT apparatus was developed with the aim of remotely quantifying the spatial and temporal distribution of photosynthetic LUE, and non-photochemical energy dissipation of terrestrial plants (Kolber et al. 2005) was successfully tested to quantify photosynthetic properties in the outer canopy of trees, which is not easily accessible (Osmond et al. 2004; Ananyev et al. 2005).

All these methods predominantly analyse the spatial and temporal heterogeneity of the outer surface of the stand. The outer canopy environment differs markedly from that of the protected understorey, in which leaves are much better buffered against water loss. However, recent evidence from an experiment at the Biosphere 2 shows that whole canopy photosynthesis is readily reflected in the leaves of the outer canopy (Rascher et al. 2004). Thus, we think that direct mapping of the LUE of the outer canopy can provide indispensable information on the spatial and temporal heterogeneity of canopy carbon gain. It will help to account for the inherent fluctuations in environmental conditions and plant metabolism to improve models and our understanding on growth, biomass allocation and carbon budgets.

System-level integration of growth, photosynthesis and carbon relations across scales

System-level integration of these insights into the dynamics of leaf and root growth and photosynthesis can have profound consequences for understanding plant production, soil biology and ecosystem functions. The consequences of plant growth in elevated atmospheric CO2, for example, demonstrate many dimensions to this problem, as articulated in recent overviews (e.g. Norby 1994; Poorter & Navas 2003; Long et al. 2004; Ainsworth & Long 2005).

An interesting test case for the impact of growth dynamics on stand-level carbon dynamics is the long-term, controlled environment, forest-stand scale experiment with coppiced cottonwood plantations in the Biosphere 2 laboratory. The system showed little evidence for ‘acclimation’ of above- or below-ground growth during four successive years of exposure to elevated CO2 and that system respiration, especially soil respiration, was dramatically stimulated by this treatment (Barron-Gafford et al. 2005). The biomass response above ground, and interaction with water deficit was dominated by enhanced leaf area development in elevated CO2 (Murthy et al. 2005). Working back to the leaf level, we demonstrated that higher leaf area in elevated CO2 was attributable to a larger population of faster growing leaves compared with controls (Walter et al. 2005). Diel growth rate analyses revealed a late-season afternoon minimum in elevated CO2 that corresponded with an afternoon glucose deficit in expanding leaves, presumably associated with other sink demands in the plants. These almost certainly included the large stimulation of fine-root production and soil respiration in elevated CO2 observed each year (Trueman & Gonzalez-Meler 2005). In addition, stable isotope analysis of soil-respired CO2 clearly indicated a stimulation of metabolism of the recalcitrant C-reserves in the soil prior to the experiment, presumably because of priming effects of enhanced rhizosphere deposition of carbon under elevated CO2.

CONCLUSIONS

Plants are highly dynamic systems and are exposed to spatially and temporally varying environmental conditions. Plants have developed protection mechanisms against unfavourable fluctuations and efficient mechanisms to use the heterogeneous conditions of their environment, from which they obtain their resources. This review highlighted some of the recent evidence in growth and photosynthesis research showing that the dynamics of central processes in plants are tightly linked to the spatial and temporal heterogeneity of the environment, in which plants have evolved. On the basis of these results, we propose that endogenous rhythms protect plants from their noisy environment. Plants actively maintain leaf growth rate within the norm of the diurnal variation when strong external conditions and even driving factors of growth change. Dynamic mechanisms in photosynthesis try to acquire as much light as possible, but protect the photosynthetic apparatus from being overexcited by fluctuating light conditions above certain thresholds.

New methods developed in recent years have opened new avenues for the understanding of spatial and temporal processes and have already shown that previous dogmas in plant research such as tip-to-base gradients of growth or homogeneity of photosynthesis in leaves have to be revised. These technologies also provide opportunities to place knowledge on molecular control in plants into the arena of natural fluctuating conditions, in which they have evolved and in which they act in intact plants. Addressing spatial heterogeneities and temporal dynamics at various scales will be of utmost importance to understanding plant behaviour, providing more insight into plant evolution in the past, breeding adequate crops for fluctuating environments at present and understanding how plants might adapt to global changes in the future.

ACKNOWLEDGEMENTS

We acknowledge the financial support provided by the Forschungszentrum Jülich in the Helmholtz Society, Deutsche Forschungsgemeinschaft, Deutscher Akademischer Austauschdienst, Alexander von Humboldt Foundation and National Science Foundation. A large part of our group's work related to the topic of this review was made possible by this support. We are grateful for the input of many group members and colleagues, especially Barry Osmond.

Ancillary