Plants have evolved a multitude of mechanisms to actuate organ movement. The osmotic influx and efflux of water in living cells can cause a rapid movement of organs in a predetermined direction. Even dead tissue can be actuated by a swelling or drying of the plant cell walls. The deformation of the organ is controlled at different levels of tissue hierarchy by geometrical constraints at the micrometre level (e.g. cell shape and size) and cell wall polymer composition at the nanoscale (e.g. cellulose fibril orientation). This paper reviews different mechanisms of organ movement in plants and highlights recent research in the field. Particular attention is paid to systems that are activated without any metabolism. The design principles of such systems may be particularly useful for a biomimetic translation into active technical composites and moving devices.
Actuation and sensing systems in plants are increasingly recognized as models to derive biomimetic principles (Shahinpoor & Thompson 1995; Taya 2003; Skotheim & Mahadevan 2005; Matthews & Giurgiutiu 2006; Fratzl et al. 2008). Higher plants evolved a large variety of mechanisms to actuate the movement of their organs. The accompanying changes in geometry lead to several different functions, such as spatial reorientation, fixation, seed dispersal, organ locomotion, ingestion, etc. (Haupt 1977; Hart 1990). Some of the systems are associated with a rapid response to external stimuli, as well described for the leaf folding of Venus flytrap or mimosa (Weintraub 1952; Findlay 1984; Hodick & Sievers 1988; Braam 2005; Forterre et al. 2005; Volkov et al. 2008). Others perform very slow movements to adjust the spatial orientation of organs, as known for leaning stems and branches in trees (Wardrop 1965; Okuyama et al. 1994; Burgert et al. 2007; Coutand et al. 2007; Goswami et al. 2008) or to specifically deform organs, as analysed for pine cones and wheat awns, respectively (Dawson et al. 1997; Elbaum et al. 2007, 2008).
Plant biologists distinguish between nastic and tropic movements. The movement in nastic responses is independent of the spatial direction of a stimulus, whereas tropic movements are influenced by its direction (Braam 2005; Patil & Vaijapurkar 2007). Moreover, plant movements can be distinguished by means of active and passive systems (Hill & Findlay 1981). Active systems are triggered by action potentials, based on living cells that activate and control the response by moving ions and changing the permeability of membranes, etc. (Toriyama & Jaffe 1972; Simons 1981; Rea 1983; Hodick & Sievers 1988, 1989; Sibaoka 1991; Shimmen 2001; Fromm & Lautner 2007; Moran 2007; Uehlein & Kaldenhoff 2008). Passive systems are mainly based on already dead tissues that are designed to undergo predetermined changes upon changes in environmental conditions (Dawson et al. 1997; Elbaum et al. 2007). These systems do not require further control or energy supply by the organism once their growth is completed, which makes them quite interesting for biomimetic transfer.
In this paper, we concentrate on the structural and mechanical constraints enabling plants to actuate their organs, summarizing current knowledge and recent discoveries in how plants generate internal stresses and movement. Readers further interested in physiological aspects regarding stimulation and control are referred to some excellent reviews (Jaffe & Galston 1968; Roblin 1979; Satter & Galston 1981; Biddington 1986; Ueda & Yamamura 2000; Braam 2005; Telewski 2006). The first part of the paper provides an overview of the various mechanisms of movement at the macroscopic level by means of exemplary species that are well known for the actuation of their organs. The second part zooms in on micro- and nanostructural organizations, and deals with the question of how the actuation is controlled by cellulose fibril architectures in the cell wall. Here, the main focal point is on systems that actuate movements without any metabolism and generate stresses just by the influx and the efflux of water in cell walls. The final part deals with recent approaches to derive biomimetic principles from the analysed mechanisms of plant actuation.
2. Mechanisms of movement
Haupt (1977) categorized different mechanisms of movement, which, in terms of higher plants, ensure mainly organ bending and some of them in a rapid manner, ballistic locomotion. Upon all mechanisms of movement, unequal growth is probably the easiest way of actuating a bending movement. All other mechanisms that are based specifically on turgor pressure, cohesion forces or the swelling of cell walls require specialized cells or tissue types that have to be differentiated during the growth of the organ.
(a) Cell growth
Bending movements due to cell growth require unequal cell expansion on adverse sides of an organ. As per definition, growth corresponds to an irreversible expansion of cells, which starts with a selective loosening of load-bearing linkages between cellulose microfibrils (Fry et al. 1992; Cosgrove 2000, 2005; Schopfer 2001) and leads to stress relaxation in the cell wall. In consequence, the turgor pressure in the cell drops, because the internal stress in the cell wall and the turgor pressure are mechanically balanced. The reduction in turgor pressure results in a cellular water uptake, which finally causes the expansion of the cell. Lockhart (1965) was one of the first to describe this interrelation of wall expansion and water uptake in a biophysical model, which was the basis for further approaches to take cell wall properties and the influence of water flow into account (Sellen 1983; Cosgrove 1993; Veytsman & Cosgrove 1998; Proseus et al. 1999, 2000).
The basic concept of the osmoregulation of cell growth is that the extensibility of the cell wall is a measure of its (visco)plastic deformation behaviour. In this sense, cell growth can be clearly distinguished from other turgor-based actuation systems that are based on a pure elastic response of the cell wall. One of the most studied model systems of growth-induced bending are the coleoptiles of grasses (Hart 1990). A coleoptile is the first organ of a spear; its function is to protect the first leaf, the so-called cotyledon. The artificial stimulation of one side of a coleoptile by growth-promoting substances, such as phytohormones (auxin), results in a curved system because cell elongation on the stimulated side exceeds that on the untreated side.
(b) Turgor pressure
Variations in turgor pressure enable plants to perform rapid movements. In most cases, only a few cells, the so-called motor cells, create the driving force for the actuation of the whole organ (Toriyama & Jaffe 1972; Fleurat-Lessard et al. 1988; Morillon et al. 2001; Moran 2007). The cell walls of these cells are elastically deformed by the inner pressure exerted by the cells. The most abundant turgor-based actuations are stomatal movements, which enable the plant to control gas exchange with the environment via pores in the leaves (Hill & Findlay 1981; Simons 1981; Morillon et al. 2001; Roelfsema & Hedrich 2002).
In many cases, a whole organ is actuated by the changes in turgor pressure. Two prominent examples of rapid movements are the leaf folding of mimosa (Mimosa pudica) and of the Venus flytrap (Dionaea muscipula), which are discussed in more detail in the following.
Circadian and light-induced leaf movements of several plants are due to changes in turgor pressure in specialized motor cells that are part of flexible hinges (so-called pulvini) located at the base of the stalk of the leaf (Iino et al. 2001; Moran 2007). A similar mechanism, but with a rapid movement, is well known for mimosa, which folds its leaves upon external stimuli. Motor cells on one side of the pulvinus (so-called extensor cells) lose their turgor pressure upon stimulation (Weintraub 1952; Campell & Thomson 1977; Samejima & Sibaoka 1980). As a result, the motor cells on the opposite side (flexor cells) are stretched (Braam 2005), but without any further increase in turgor pressure, which reduces the cell stiffness dramatically. As a consequence, the entire hinge loses its bending stiffness, deflects and thereby triggers the folding movement (Haupt 1977).
The carnivorous Venus flytrap performs a movement of its leaves to trap and digest insects as a source of nitrogen (Findlay 1984; Fagerberg & Allain 1991). It is the fastest plant organ movement known, with a leaf closure time of approximately 100 ms in the rapid phase (Forterre et al. 2005). The movement is stimulated when the prey deflects sensory hairs on the inner surface of the leaf. This stimulation translates into a propagation of action potentials (Stuhlman & Darden 1950; Hodick & Sievers 1988, 1989), an instantaneous change of membrane permeability and, in consequence, in volume changes of leaf tissues that result in trap closure (Hart 1990; Sibaoka 1991). In contrast to the circadian and light-induced leaf movements and the rapid movement of mimosa, Venus flytrap leaves do not possess a pulvinus with a definite movement mechanism (Williams & Bennett 1982). The mechanism of Venus flytrap closure is still debated and different models have been proposed, e.g. a rapid loss of turgor pressure, acid growth of motor cells or a prestress by extensible mesophyll cells that are retained under compression in the open leaves (Brown 1916; Williams & Bennett 1982; Hodick & Sievers 1989; Fagerberg & Allain 1991; Fagerberg & Howe 1996). Although being contrary about the underlying mechanism, all models account for drastic volume changes of the leaf cells/tissue on the adaxial and abaxial side during trap closure. A conclusive biomechanical explanation of the enormous speed in leaf folding was recently presented by a detailed analysis of the mechanical constraints in the prestressed system (Forterre et al. 2005; figure 1).
In open condition, the leaves show a concave plane folding, maintained in a metastable state by means of highly turgorized cells. Upon contact to a certain number of sensor hairs by the prey, the metastable folding state becomes instable due to volume changes of the leaf tissues. It is then an elastic relaxation that closes the trap by converting the concave folding of the leaf to a convex folding.
Another well-known bending movement is due to the touch response of tendrils (Jaffe & Galston 1968; Engelberth 2003). It relies on a combined action of cell growth and of changes in turgor pressure (Haupt 1977). For instance, the bending movements of tendrils of peas are based on a shortening of the cells on the ventral side due to a loss in turgor pressure and on an elongation of cells on the dorsal side due to partial growth processes.
(c) Cohesion forces
A particularly impressive example for a movement caused by the cohesion forces of water is a catapult mechanism to distribute fern spores. Specialized cells, so-called annular cells, form a ring around a spore capsule. These cells have thicker cell walls on three sides, which give them a U-shaped appearance in cross section (figure 2).
Upon water removal from the cell lumina, the cohesion forces pull both the ends of the U-shaped cells together. This leads to a shortening of the outside perimeter of the annulus. The resulting transverse tensile stresses cause a rupture of the spore capsule at a predefined position, which releases some of the spores. Finally, the increasing stress in the annular cell walls and the ongoing stress generation due to the drying process eventually exceed the cohesion forces of the water. This leads to a spontaneous rebound of the system that catapults the fern spores out of the capsule (Haupt 1977).
(d) Cell wall swelling or shrinking
Plants can prepare tissues for a passive actuation of organs, even after cell senescence. This is possible by controlling the anisotropic deformation of cells upon swelling and shrinking by an elaborated adjustment of cell wall architecture. Plant cell walls are composites of stiff cellulose fibrils embedded in a pliant and highly swellable matrix consisting of hemicelluloses, pectin, structural proteins and/or lignin (Fengel & Wegener 1984; McNeil et al. 1984; Carpita & Gibeaut 1993; Fahlén & Salmén 2005). Hence, the orientation of the cellulose fibrils in the cell walls is a crucial adjustment mechanism, and different cellulose alignments in subsequent tissue layers can cause a substantial and directed bending of organs. It is just the influx and efflux of water in the cell walls that causes changes in cell and tissue geometry. The principle is based on the fact that, in a composite where the (partially crystalline) cellulose fibrils are not swelling, the extension upon water uptake becomes extremely anisotropic. Indeed, swelling will occur preferentially in the direction perpendicular to the fibrils.
This actuation principle is implemented by a wide variety of species in their seed dispersal units. A well-known example is the release of ripe seeds from conifer cones (Dawson et al. 1997). Upon drying in ambient air, pine cones open due to a bending movement of their scales (figure 3).
In these scales, differently structured tissue layers on the upper and lower sides can be distinguished. Fibres being part of the upper side of the scales consist of cell walls in which the cellulose microfibrils are oriented almost parallel to the cell axis, also corresponding to the long axis of the scale. Upon drying, these cell walls show an anisotropic shrinkage with minor deformation in the axial direction, but a pronounced shrinking perpendicular to the cell axis. In cells on the lower side of the scales, cellulose microfibrils are oriented almost perpendicular to the cell axis. Thus, the efflux of water forces the matrix polymers to shrink in the axial direction. The combined action of both strongly connected tissue layers leads to a bending of the scale, which opens the cone. In analogy, this system can be regarded as a kind of bimetal with two substantial differences: (i) the switch is hygrosensitive and not thermosensitive and (ii) while bimetals need different elements, the directed deformation upon moisture changes can be achieved with the same chemical composition.
It is simply the variation of cell wall architecture by means of cellulose fibril orientation that leads to the required anisotropic response. Further examples of moisture-dependent seed dispersal units are the spore capsules of some mosses (Ingold 1959; Haupt 1977), which show the same mechanism, but a hygrosensitive opening and closure indicating that these complex systems are even more moisture sensitive than the conifer cones.
3. Recent insight into the control of actuation by cellulose fibril architectures
A more detailed discussion of the influence of cell wall architecture on plant actuation is provided on the basis of recent studies on the locomotion of wheat dispersal units and the bending movements of trees to upright leaning stems and branches.
(a) Wheat awns
A swelling and drying mechanism provides motility to seed dispersal units of wild wheat. It has been shown, for wheat awns, that an optimized cell wall architecture of their cells allow them to perform swimming movements and to propel the seed along and into the ground (Elbaum et al. 2007). The awns are steering guides, supporting structure and engine at the same time. They are guiding a ripe grain to the Earth with the pointed end downwards. They also provide unidirectional motility for a movement along or into the ground. This action is powered solely by the daily change in humidity, which induces a curvature of the awns depending on moisture (figure 4).
Similar to the conifer cones, the awns possess tissue types with different cell wall architectures on the inner side, facing towards the other awn (called the cap) and the outer side, facing outwards (called the ridge). In the cap, the cell walls have cellulose fibrils that are oriented almost parallel to the cell axis (corresponding to the awn axis), which makes the axial direction less sensitive to moisture changes. However, in the lower section of the ridge, a random distribution of cellulose fibrils was measured. In fact, cell walls in this part of the awn correspond to a succession of thin layers with alternating cellulose orientation (Elbaum et al. 2008). Here, moisture changes result in an isotropic shrinkage in all directions. In consequence, in the dry daytime air, the awns bend outwards, while at night, dampened by the dew, they move towards each other. For a forward motion, the seed dispersal unit has to ensure that it is not moving backwards during the strokes. This is achieved by fine, barb-like silica hairs on the outside of the awn, which cause the device to act as a ratchet by preventing the awns from pushing themselves out of the soil when the awns move towards each other during the night. Hence, over several days, this unidirectional movement, similar to the swimming strokes of a frog, pushes the grain along and into the soil (Elbaum et al. 2007).
(b) Movements of stem and branches
Movements are also performed by trees when bending stems and branches upwards, but on much lager time scales. In a straight trunk, the peripheral part of the normal wood is prestressed in tension (Boyd 1950a; Kubler 1987). These longitudinal tensile stresses compensate for the comparatively low compressive strength of wood and strengthen the tree against high wind loads (Mattheck & Kubler 1996). To enable the tree to direct growing organs towards a predetermined position, additional stresses have to be generated by specific wood tissues, so-called reaction wood. Softwoods form compression wood on the underside of the organ that generates compressive stresses, whereas hardwoods form tension wood on the upper side, which exceeds the tensile stress generation in the normal wood (Wardrop 1965).
An interesting aspect of stress generation in trees is that the internal stresses have to be incorporated in the cell wall during a short differentiation process of the living cell (approx. 2 weeks). Accordingly, a retarded insertion of cellulose fibrils and matrix polymers during cell wall formation and, in particular, the temporally delayed incrustation of lignin has been proposed to play a crucial role in stress generation (Boyd 1950b; Bamber 1979; Yamamoto 1998). In recent studies, the underlying mechanisms of stress generation in compression wood and tension wood were further examined and quite different conclusions were reached.
Indeed, conifers are able to generate either tensile or compressive stresses in their tissues according to needs. The regularly formed cells (normal wood tracheids) are typically prestressed in tension along the longitudinal direction, but the compression wood cells generate longitudinal compressive stresses in their cell walls (figure 5).
It is very interesting to observe that the stress generation capabilities of the various tissue types are mainly controlled by their specific cell wall architecture. Normal wood tracheids possess a rectangular shape and the cellulose fibrils in their cell walls are oriented almost parallel to the cell axis (microfibril angle (MFA) approx. 5–20°). Compression wood tracheids are round in cross section and have a notably higher cellulose MFA (Côte & Day 1965).
Starting from their native state (wet), a further swelling of the different cell types in saturated sodium iodide solution results in a slight reduction in length of normal wood (approx. −0.3%) compared with a large increase in length of compression wood (approx. 2.5%). Hence, the same swelling treatment results in an opposite deformation pattern (i.e. either tensile or compressive stresses), depending only on the nano-architecture of the cell wall.
The reason for this humidity-driven actuation is that stiff cellulose fibrils act as rigid elements, while the matrix polymers are swollen by the sodium iodide treatment. It could be shown in a simple mathematical model that, if no torsion of the cell is allowed (as expected in a wood tissue comprising many parallel cells), the direction of the generated stress depends on the angle between the cellulose fibrils and the longitudinal cell axis. For an MFA below 45°, cells are shortening in the longitudinal direction during swelling, while they are elongating otherwise. However, if (even a slight) torsion of the cell is allowed, compressive stresses are always generated. Accordingly, the generation of either tensile or compressive stresses is mainly controlled by a clever arrangement of cellulose fibrils in the cell walls (Burgert et al. 2007).
Hardwoods have evolved an additional mechanism that allows them to generate very high tensile stresses on the upper side of a bending organ, such as to pull leaning stems and branches upwards. The tension wood fibres tend to contract longitudinally during differentiation, which generates longitudinal tensile stresses of up to approximately 70 MPa (Okuyama et al. 1994). Tension wood fibres of many species show an additional characteristic structural feature in comparison with regular fibres. A so-called G-layer appears in the normally empty lumen along the axis of the cylindrical cell. Apart from minor amounts of hemicelluloses and lignin (Joseleau et al. 2004; Lehringer et al. 2008), it consists of almost pure cellulose organized in aggregates of 30–40 nm diameter strictly parallel to the cell axis (MFA approx. 0°) and can fill the whole lumen (Daniel et al. 2006; Müller et al. 2006; Goswami et al. 2008). The G-layer is believed to be the operative part of the tension wood fibre (Côte & Day 1965; Clair et al. 2003, 2006) although the underlying mechanisms of stress generation are still discussed (Mellerowicz et al. 2008) and high tensile stresses can be generated also in its absence (Qiu et al. 2008).
The fundamental question is how the length of tension wood fibres can be reduced by a G-layer consisting of axially oriented almost non-contractile cellulose fibrils. A possible explanation has been proposed in a recent study whereby the contraction of the fibre is not caused by the G-layer directly, but by an interaction of the G-layer with the surrounding secondary cell wall (Goswami et al. 2008), a mechanism that had been first suggested by Münch (1938).
Indeed, measurements of dimensional changes of tension wood slices upon enzymatic G-layer removal revealed an elongation by 1.6 per cent in the longitudinal direction and shortening by approximately 1 per cent perpendicular to the fibre axis. The severe longitudinal elongation upon G-layer removal implies that the G-layer actively shortens the tension wood fibre as a result of the stresses generated during maturation. Moreover, cellulose fibrils in the cell wall of tension wood fibre surrounding the G-layer were found to have an MFA of approximately 36° with respect to the cell axis.
Tension wood tissue with the G-layers removed shows a mechanical behaviour typical for normal plant tissue, with a similarly high cellulose MFA in the main cell wall layer (Bodig & Jayne 1993; Köhler & Spatz 2002; Keckes et al. 2003; Burgert 2006). The presence of the G-layer increases the initial stiffness of the fibre by a factor of approximately 4, and further deformation is characterized by a jagged appearance of the stress–strain curve. This can be interpreted as a stick and slip movement created by a friction of the G-layer, which strongly swells laterally and exerts pressure onto the cell wall (figure 6).
Owing to the spiral arrangement of cellulose in the cell wall, the circumferential hoop stress is converted into a contraction of the cell along its length. It has been shown that the optimal spiral angle for the generation of longitudinal contractile stresses is close to the observed MFA of approximately 36° (Fratzl et al. 2008). Hence, the combination of an axially stiff and laterally swellable G-layer with a suitable cellulose MFA in the secondary cell wall is responsible for the generation of considerable tensile stresses pulling the branches of hardwood trees upwards towards the light.
4. Towards biomimetic actuating composites
As already discussed, there are many organ movements in plants due to moisture changes in cell walls that are controlled by the architecture of stiff cellulose fibrils embedded in a swellable matrix. This reveals a general principle by which active gels can be directed in their swelling behaviour by stiff fibres distributed in a suitable way. From a materials science point of view, it is quite interesting to note that these actuating movements function without the need of an active metabolism, very unlike molecular motors in the human muscle, for example. This makes the humidity-based actuation systems in plants particularly interesting for biomimetic materials research.
The main ingredients for a moving composite mimicking plant cells are a swelling matrix (e.g. a gel) with embedded stiff fibrils or needles. Assuming that the matrix is tightly bound to the fibrils (e.g. by covalent bonding), then isotropic swelling of the matrix is hindered by the presence of the essentially undeformable elements. The consequence is an anisotropic deformation and, in ideal cases, even a complex movement may result from the simple swelling of a gel. An example for a realization of such a device was given by Sidorenko et al. (2007), where a family of parallel silicon needles were embedded into a hydrogel and covalently bound to the molecules in the gel. Shrinking of the hydrogel is obviously not possible in the direction of the stiff needles, except if the needles are tilting. The contraction of the gel thus results in a collective tilting of the needles, leading to a lateral actuation of the needle–gel composite. This, and even more complex movements, could be obtained by controlling the degrees of freedom for the movement of the needles (Sidorenko et al. 2007).
More generally, one may think about the type of actuation to be generated. It is well known from lever arms that a given mechanical energy can be converted either into a large movement associated with a small force or into a large force combined with a small movement. The well-known devices for these conversions are the catapult and the pulley, respectively. It is interesting to note that similar devices can be generated at the microscopic scale just by designing appropriate fibre architectures into the swellable matrix. The example is given directly by the plant cells. The wheat awn discussed earlier is optimized for large movements, while the reaction wood cells in spruce are designed to generate large stresses but almost no movement.
To describe the underlying principle, we may imagine a hollow tube reinforced by a stiff fibre spiralling around the central channel. When the matrix around these fibres swells, axial stresses and strains are generated, with different magnitude and even sign, depending on the microfibril angle μ and on the stiffness ratio, f, between the fibres and the matrix (figure 7). At a stiffness ratio of 20 (typical for plant cells), for example, the strongest tensile stresses are generated for an angle of the order of 25°, while the largest deformation is obtained for an angle close to 35° (Fratzl et al. 2008). Above 45°, the tensile stress or deformation turns into compressive (figure 7). Such construction principles may eventually become useful for designing gel-based actuators for artificial muscles.
In summary, the observation of plant movement reveals a number of physical principles that may be interesting for the development of bioinspired actuators. In particular, the actuating by the swelling or shrinking of cell walls, such as in wood, pine cones or wheat awns, does not require an active metabolism and is entirely based on materials behaviour. These systems are particularly suitable as inspirations for technical actuating devices, based on fibre-reinforced swellable matrices, such as gels.
One contribution of 9 to a Theme Issue ‘Biomimetics I: functional biosurfaces’.
- © 2009 The Royal Society