Mechanosensing of stem bending and its interspecific variability in five ...

Annals of Botany 105: 341 –347, 2010 doi:10.1093/aob/mcp286, available online at www.aob.oxfordjournals.org

SHORT COMMUNICATION

Mechanosensing of stem bending and its interspecific variability in five neotropical rainforest species Catherine Coutand1,2,*, Malia Chevolot3, Andre´ Lacointe1,2, Nick Rowe4,5 and Ivan Scotti3 1

INRA, UMR 547 PIAF, F-63000 Clermont-Ferrand, France, 2Universite´ Blaise Pascal, UMR 547 PIAF, INRA-, F-63000 Clermont-Ferrand, France, 3UMR ECoFoG, BP 709, F-97387 Kourou Cedex, France, 4Universite´ de Montpellier 2, UMR AMAP, F-34000 Montpellier, France and 5CNRS, UMR AMAP, F-34000 Montpellier, France * For correspondence. E-mail [email protected] Received: 17 July 2009 Returned for revision: 14 September 2009 Accepted: 28 October 2009 Published electronically: 8 December 2009

† Background and Aims In rain forests, sapling survival is highly dependent on the regulation of trunk slenderness (height/diameter ratio): shade-intolerant species have to grow in height as fast as possible to reach the canopy but also have to withstand mechanical loadings (wind and their own weight) to avoid buckling. Recent studies suggest that mechanosensing is essential to control tree dimensions and stability-related morphogenesis. Differences in species slenderness have been observed among rainforest trees; the present study thus investigates whether species with different slenderness and growth habits exhibit differences in mechanosensitivity. † Methods Recent studies have led to a model of mechanosensing (sum-of-strains model) that predicts a quantitative relationship between the applied sum of longitudinal strains and the plant’s responses in the case of a single bending. Saplings of five different neotropical species (Eperua falcata, E. grandiflora, Tachigali melinonii, Symphonia globulifera and Bauhinia guianensis) were subjected to a regimen of controlled mechanical loading phases (bending) alternating with still phases over a period of 2 months. Mechanical loading was controlled in terms of strains and the five species were subjected to the same range of sum of strains. The application of the sum-of-strain model led to a dose–response curve for each species. Dose–response curves were then compared between tested species. † Key Results The model of mechanosensing (sum-of-strain model) applied in the case of multiple bending as long as the bending frequency was low. A comparison of dose –response curves for each species demonstrated differences in the stimulus threshold, suggesting two groups of responses among the species. Interestingly, the liana species B. guianensis exhibited a higher threshold than other Leguminosae species tested. † Conclusions This study provides a conceptual framework to study variability in plant mechanosensing and demonstrated interspecific variability in mechanosensing. Key words: Mechanosensing, interspecific variability, trees, lianas, rain forest, neotropical species, bending, biomechanics, Bauhinia, Eperua, Symphonia, Tachigali.

IN T RO DU C T IO N Interest in mechanical signals is increasing because of their implication in the control of plant morphogenesis (Moulia et al., 2006; Hamant et al., 2008); their effects, which have been described as thigmomorphogenesis since the 1970s (Boyer, 1967; Jaffe, 1973), are also the focus of increasing interest. External mechanical signals generally induce a decrease of elongation and a stimulation of diameter growth. Thigmomorphogenesis has been demonstrated for plants including both herbaceous (for a review see Biddington, 1985) and woody species (Jacobs, 1954; Larson, 1965; Telewski and Pruyn, 1998; Meng et al., 2006; Coutand et al., 2008; reviewed by Telewski, 1995). Although thigmomorphogenetic responses have been attributed to a range of external mechanical stimuli, internal mechanical signals produced by deformations induced by gravity and self-weight lead to responses similar to those of thigmomorphogenesis but have been referred to as gravity resistance (Soga et al., 2006). Telewski (2006) proposed a unified hypothesis of plant mechanosensing including both gravimorphism and thigmomorphogenesis.

From an ecological point of view, growth in height is an important functional trait for sapling survival that is linked to light requirements and behaviour. In contrast to shadetolerant species, shade-intolerant species have to reach the canopy as fast as possible to survive, which means that they have to grow in height with a minimum investment of material in diameter growth. However, their survival also depends on their capacity to withstand mechanical loadings due to wind or to their own weight and avoid buckling (Mc Mahon, 1973; Fournier et al., 2006). The control of trunk slenderness (height/diameter ratio) is thus an important, if not essential, factor in the survival of saplings. Biomechanical studies suggest that it is the perception of mechanical signals that is a necessary prerequisite for plants to control their dimensions and their stability-related morphogenesis (Fournier et al., 2006). It is therefore of interest to know how sensitivity of mechanosensing may vary and if it can account for differences in slenderness of different species observed in natural conditions (Jaouen, 2008). Jaffe (1973) found different responses to internode rubbing in a variety of herbaceous species: species such as Hordeum vulgare, Bryonia dioica, Cucumis sativus,

# The Author 2009. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. For Permissions, please email: [email protected]

342

Coutand et al. — Interspecific variability of mechanosensing

Phaesolus vulgaris, Mimosa pudica and Ricinus communis exhibited a reduced elongation whereas plants of Cucurbita pepo, Pisum sativum and Triticum aestivum did not exhibit such a response, suggesting that interspecific variability exists in mechanosensing. However, interspecific variability of mechanosensing is difficult to assess from these experiments because of differences in the duration of mechanical stress and plant size. Therefore, the strains resulting from stem rubbing might have been different among species with the result that the mechanical stimulus was not properly quantified. Finally, to our knowledge, interspecific variability of mechanoperception has not yet been investigated among sympatric tropical forest species. The results presented in this study are part of a wider project (2006 – 2009) known as ‘Woodiversity’ investigating the diversity of biomechanical strategies of tree growth in tropical forest plants in French Guiana. When variation in thigmomorphogenetic responses is observed between species, at least two sources of variability can be inferred: one due to the variation in mechanical perturbation (Coutand and Moulia, 2000; Coutand et al., 2009), the other due to intrinsic variation of sensitivity to mechanical loading between species. Recent quantitative studies of thigmomorphogenesis have shown that these two sources of variability can be disentangled: a quantitative study of thigmomorphogenesis found that the stimulus that is perceived by plants during mechanical stimulation are the strains experienced by the plant and not forces or stresses (Coutand and Moulia, 2000). Second, a model of mechanoperception has been developed that assumes that (1) each volume of tissue that undergoes deformation generates a signal that is proportional to the strain, and (2) the sum of these local signals contributes to the observed thigmomorphogenetic response. The model predicts a relationship between the observed plant response and the applied sum of longitudinal strains. In tomato (Solanum lycopersicum) it was shown that a single temporary bending led to a transitory modification of elongation rate. Within minutes of undergoing bending, tomato stems show a cessation of growth followed by a progressive recovery of the pre-bending elongation rate (Coutand et al., 2000). Based on these experimental results, the model shows a loglinear relationship between the duration of the growth response and the sum of imposed longitudinal strains (Coutand and Moulia, 2000): D ¼ alnðSumstrains=Sumstrains0 Þ

ð1Þ

where D is the plant response duration, a the species sensitivity, Sumstrains is the sum of longitudinal applied strains and Sumstrains0 is the stimulus threshold beyond which the level of the sum of strain induces a thigmomorphogenetic response. More recently, the underlying assumptions of the model have been validated by measuring two localized responses of plant stem after a single bending treatment at the loaded zone: the growth in diameter and the level of expression of a primary mechanosensitive gene, PtaZFP2 (Coutand et al., 2009). As it has been validated for different responses and at different scales, the model can thus be used to control the source of variability due to intensity of the mechanical stimulus.

The objective of this model is to obtain two parameters, which are independent of stem geometry and mechanical properties and which provide an indication of a species’ intrinsic mechanosensitivity. If a quantitative relationship between the level of applied strains and the plant response could be established for different species and different growth forms, comparisons of dose –response curves could then indicate whether interspecific variability of mechanosensing exists. In addition to quantitative studies based on single bending treatments, other analyses have suggested an acclimation with mechanical loading at the entire plant level (Coutand et al., 2008) as well as responses at the organ level (ethylene production) (Telewski, 1990) or at the cellular level (Arteca and Arteca, 1999); characterization of these acclimation processes has just begun. A recent study on poplar revealed that a single bending treatment induces a modification of diameter growth rate for about 4 – 5 d (Coutand et al., 2009). Results from Martin (2009) on young poplars have revealed that the application of one bending per day for 3 d increased diameter growth rate. However, continued bending treatment for more than 3 d had no additional effect: diameter growth returned progressively to levels observed prior to bending stimulation. The results also demonstrated that increased diameter growth activity stimulated by bending can be recovered if a sufficient ‘rest’ period is observed between bending treatments (Martin, 2009). These recent findings suggest that the mechanosensing model should apply in the case of successive loadings if the frequency of bending is sufficiently low to enable the plant to recover its mechanosensitivity fully between two loadings. The results also suggest that the maximal thigmomorphogenetic response is not obtained by stimulating the plant continuously but by alternating periods of bending with still periods. This investigation had two aims: first, to verify the mechanosensing model when successive bending treatments are applied and, second, to determine whether interspecific variability of mechanosensing exists. The experimental set-up incorporated mechanical loading by successive controlled bending treatments. Groups of saplings from five tropical species were submitted to a regimen of alternating mechanical loading phases with still phases for 2 months. Bending was controlled via an experimental device, which allowed control over the level of strains applied to the stem (Coutand et al., 2009). Bending treatments were used to investigate (1) the responses of stems of each species to a range of applied strains and (2) the responses of stems of different species within the same range of applied strains. The first objective was to assess whether the mechanosensitivity model led to a sum-of-strains – growth response curve for each species. A second objective was to investigate whether characteristics of the strain – response curves varied between species and thus indicated interspecific variability of the mechanosensitivity. M AT E R I A L S A N D M E T H O D S Saplings

Five species were investigated: four Caesalpiniaceae [Eperua grandiflora (Aubl.) Benth., Eperua falcata Aubl., Bauhinia guianensis Aubl., Tachigali melinonii (Harms) Zarucchi &

Coutand et al. — Interspecific variability of mechanosensing Herand] and one Clusiaceae (Symphonia globulifera L. f.). All saplings were approximately 1 year old. Eperua plants were issued from seeds. T. melinonii and S. globulifera were harvested from Paracou experimental research station (58180 N, 528230 W, French Guyana) while other species were collected from similar rainforest habitats in this surrounding area of north-east French Guiana. The species were selected on the basis of several factors, including (1) representative and common sympatric species of this lowland rain forest, (2) a range of different growth forms including sapling representatives of trees and one liana, (3) likelihood of successful cultivation in controlled conditions and (4) confidence of correct identification of young saplings. All species were transplanted and grown in individual 20-litre pots filled with a mixture of 30 % sand and 70 % of A-horizon natural soil from the forest during January – March 2007. Soil water content (approx. 25 – 30 m3 m23) was maintained at field capacity by watering the plants three times a week. Plants were cultivated in a screened tunnel under ambient conditions for 1 year; then 1.5 months before the beginning of the bending experiments, they were moved to a protected experimental screened tunnel, also under ambient conditions, where the experiment took place. Saplings were watered regularly by individual irrigation devices throughout the experiment. Total number of individuals per species and initial dimensions of plants are summarized in Table 1. Plants were grouped in pairs with one individual of each pair taken as a control and the other individual, of approximately equal size, subjected to bending treatment.

343

day and a second phase where stems were left without being stimulated. Results on poplars have shown that the average duration of the diameter growth response induced by a single bending was about 4 – 5 d (Coutand et al., 2009). Plants were thus bent once a day for 3 d (maximization of the effect of bending) and then left without being bent for 4 d, i.e. the assumed minimum average time for the plant to recover its sensitivity to mechanical loading. Plants were then bent again for 3 d, . . . , etc. This regimen of alternating bending treatments and ‘rest’ periods was applied for 2 months. The sum of strains due to each bending was computed as detailed in Coutand et al. (2009). The sum of strains perceived during the whole treatment was computed as the sum of spatial sums of longitudinal strains induced by each bending. Growth measurements

Stem diameter was measured weekly and in the plane of bending by using a Vernier caliper (0.01-mm precision) made of plastic so as to avoid wounding the stem. Measurements were made at precisely marked positions on the stems at approx. 150 mm from the stem base in the middle of the bent part of the stem. Stem length was measured weekly using a measuring tape ( precision 0.5 cm). Quantification of growth responses

The growth responses due to the mechanical treatment were first quantified by the difference in length and diameter between the beginning and end of the experiment. Statistical

Bending experiments

In order to control the level of applied strain, each stem was bent around a rigid plastic tube of known diameter (Fig. 1). To create the range of mechanical loadings, stem diameters were taken into account and the appropriate tube diameter selected so that all the species underwent the same range of sum of strains [details of computations of strains imposed to stems when bent around a plastic tube can be found in Coutand et al. (2009)]. The tubes used for creating a range of imposed strains were 40, 100, 200 and 300 mm in diameter. Following the results obtained using poplar, an appropriate regimen of mechanical loading was chosen, i.e. that maximized the effect of bending on plant growth. This allowed plant responses to be measured while avoiding too many interactions between successive loadings; each cycle of the regimen comprised two phases: a first phase where stems were bent once a

rstem

Stem

rtube

Bent part of the stem

Plastic tube Central line F I G . 1 Scheme of the stem in the bent state (from Coutand et al., 2009). Locally, the maximal longitudinal strain is given by the product of the central line curvature by the stem radius (rstem). The curvature of the central line is given by the inverse of the sum of the radius of the plastic tube (rtube) and the stem radius.

TA B L E 1. Average dimensions of plant stems before treatment

Species Bauhinia guianensis Eperua falcata Eperua grandiflora Symphonia globulifera Tachigali melinonii

Total number of individuals 12 20 16 20 12

Height (cm)

Stem diameter, perpendicular to the plane of bending that was about to be applied (mm)

85.1 (+51.0) 84.0 (+16.5) 41.8 (+25.8) 38.6 (+7.9) 71.7 (+26.6)

5.6 (+1.4) 6.6 (+1.2) 4.7 (+2.5) 4.7 (+0.6) 6.5 (+1)

Stem diameter, parallel to the plane of bending that was to be applied (mm) 5.5 6.7 4.6 4.7 6.5

(+1.6) (+1.2) (+2.4) (+0.6) (+0.8)

344


analyses revealed significant differences in diameter increment between bent and control plants but no difference in terms of elongation. There was a marked interspecific heterogeneity in the diameter growth of control plants; in order to quantify the growth response in diameter better, the increment of diameter of each bent plant was subsequently normalized by dividing it by the average diameter increment of control plants for each species. Average diameter increment of control plants was used instead of the increment of diameter of the paired control because, except for E. grandiflora, no relationship was found between the increment of diameter and the initial stem diameter in control plants.

significant difference in height growth between control and bent plants. Diameter growth of control plants was relatively homogeneous among tree saplings (between 0.5 and 0.7 mm and over twice as much for the liana, Table 2). Significant increases in diameter in the plane of bending were observed between bent and control stems, indicating that cambial growth was stimulated by the bending treatment. For each species, the average increase of diameter growth in comparison with the controls varied greatly with species: B. guianensis exhibited the least increase in diameter growth (þ25 %), followed by E. falcata (þ51 %), E. grandiflora (þ86 %), S. globulifera (þ120 %) and T. melinonii (þ123 %). Relationship between stem diameter growth responses and applied strains

Differences of growth in height and diameter between bent and unbent plants were analysed for each species by comparison of means using a classical Student’s t-test at the level of 5 % (Dagne´lie, 1975), regardless of the level of strain applied. Diameter growth response data were then analysed by covariance analysis (ANCOVA) with mechanical stimulus as the covariable, using type III sums of squares as computed by the general linear model (GLM) procedure (Statistica software v.5.5, StatSoft, Tulsa, OK, USA). In a first step, homogeneity of slopes among species was assessed with the full-factorial GLM. If slopes were found to be homogeneous, a second step was performed by removing the species slope interaction from the model to test for differences in y-intercepts among species. Where relevant, homogeneous groups of y-intercepts were constructed using the Newman – Keuls test at a significance threshold of P ¼ 0.05. For E. falcata two data points were considered outliers and removed from the data set for statistical analyses. RES ULT S Effects of bending on stem elongation and diameter growth

Height growth of control plants varied between approx. 3 and 9 cm for the tree saplings and about three times as much (over 30 cm) for the liana B. guianensis (Table 2). There was no

The relationships between the applied level of strain and diameter growth of the stem for each species are presented in Fig. 2. The parameters of the linear fits are given in Table 3. The relationship was loglinear. This could be written as: IncDiam ¼ alnðSumstrains=Sumstrains0 Þ

IncDiam ¼ alnðSumstrainsÞ alnðSumstrains0 Þ:

Treatment

Growth in height (cm)

Bauhinia guianensis

Control Bent Control Bent Control Bent Control Bent Control Bent

31.2 (+36.5) 18.2 (+15.7)ns 9.1 (+9.9) 10.3 (+10.0)ns 3.9 (+4.1) 6.1 (+3.6)ns 2.95 (+14.6) 5.5 (+2.2)ns 9.1 (+2.2) 8.2 (+5.2)ns

Eperua falcata Eperua grandiflora Symphonia globulifera Tachigali melinonii

Growth in diameter (mm) 1.7 (+0.6) 2.2 (+0.6)* 0.6 (+0.2) 0.9 (+0.5)ns 0.5 (+0.3) 0.9 (+0.3)* 0.7 (+0.2) 1.6 (+0.6)* 0.7 (+0.6) 1.6 (+0.3)*

Statistical results of Student’s t-test at the level of 5 %, comparing growth (growth in height and growth in diameter) between control and bent plants for each species. ns, not significant; * significant, P , 0.05.

ð3Þ

After checking for slope homogeneity (data not shown), the ANCOVA (Table 4) shows that the common slope and the average value of the intercept both significantly differed 3·5 Bauhinia guianensis Eperua falcata

3·0

Eperua grandiflora Symphonia globulifera

2·5

Tachigali melinonii

2·0 1·5 1·0 0·5 0

TA B L E 2. Average growth observed after 2 months of treatment Species

ð2Þ

and can also be written as:

Diameter increment (mm)

Statistical analyses

0

0·5

1·0 Ln(Sum of strains)

1·5

2·0

F I G . 2 Diameter increment according to logarithm of the applied sum of longitudinal strains per species and the linear fitting per species. The values of parameters of the linear fit are given in Table 4.

TA B L E 3. Parameters of the linear fit of the increment of diameter according to the logarithm of the applied sum of longitudinal strains Species

Slope

Intercept

R2

Bauhinia guianensis Eperua falcata Eperua grandiflora Symphonia globulifera Tachigali melinonii

1.46 0.59 0.73 1.25 0.93

0.84 0.15 0.33 0.44 0.66

0.57 0.38 0.78 0.52 0.49

Coutand et al. — Interspecific variability of mechanosensing TA B L E 4. ANCOVA of diameter growth response

Average intercept Species LogSumstrains Error

d.f.

F

P

1 4 1 31

5.50 10.2 37.3

0.025 0.000023 0.000001

The model was computed as additive after checking for slope homogeneity.

345

value of 1 gives a cumulative sum of strains of approximately 1.22 mm3 for E. grandiflora, S. globulifera, T. melinonii and of 1.91 mm3 for E. falcata and B. guianensis. These values corresponded to the cumulative sum of strains experienced by the stem during the 9 weeks of treatment, i.e. nine bouts of three bending so a total of 27 bending. The average stimulus threshold for one bending is thus approx. 0.045 mm3 for E. grandiflora, S. Globulifera and T. melinonii and approx. 0.071 mm3 for E. falcata and B. guianensis. DISCUSSION

TA B L E 5. Results of the post-hoc test (Newman –Keuls at the confidence level of 5 %) Species Eperua falcata Bauhinia guianensis Eperua grandiflora Symphonia globulifera Tachigali melinonii

Mean intercept

Group 1

1.106 1.257 1.864 2.198 2.238

**** ****

Group 2

**** **** ****

Standardized diameter increment

4·5 Bauhinia + Eperua falcata Eperua grandi. + Symphonia + Tachigali

4·0 3·5 3·0

y = 1·6091x + 0·6947 R 2 = 0·6242

2·5 2·0

y = 0·9085x + 0·3562 R 2 = 0·4314

1·5 1·0 0·5 0

0

0·5

1·0 Ln(Sum of strains)

1·5

2·0

F I G . 3 Standardized diameter increment according to the logarithm of the applied sum of longitudinal strains.

from 0, with significant differences in the intercept value between species. Two groups of species emerged from the post-hoc analysis of the intercepts (Table 5): one group composed of T. melinonii, E. grandiflora and S. globulifera and the other of E. falcata and B. guianensis. The second term in eqn (3) corresponds to the constant of the loglinear relationship. As the slope (a) was common to all the species, variability in the constant can be ascribed to variability in mechanical stimulus threshold (Sumstrains0). A normalized increment value of 1 would correspond to an increment of diameter observed after bending that is equal to that observed for controls, i.e. bending has no effect on diameter growth. Comparisons of normalized increment values therefore enabled us to determine threshold values of strain above which diameter growth was stimulated. Threshold values can be determined from Fig. 3: a normalized response

Effect of bending on diameter growth and elongation

Following the definition of thigmomorphogenesis given by Jaffe (1973), mechanical loadings lead to a reduction of elongation and a stimulation of diameter growth. In the present experiments, the bending of the basal part of the stem induced an increase of growth in diameter but did not lead to a statistically significant reduction of elongation. Three hypotheses can be evoked to explain this. (1) High variability of elongation growth even in control plants masked the statistical difference between bent and control plants. (2) Bending applied to the basal part of the stem is not sufficient to induce a thigmomorphogenetic signal that is able to affect the elongation zone located at the top of plants. This second hypothesis contrasts with results found on tomato stems where a light bending of the basal part of the stem induced a significant reduction of elongation of the apical part (Coutand et al., 2000). (3) Contrasting with herbaceous plants, in woody plants a mechanical signal applied to a limited region of the stem might not induce a systemic transfer of information. Conversely to what was found for elongation, and as expected from earlier studies of thigmomorphogenesis, the bending treatment (regardless of the level of strain applied) induced an increase in stem diameter growth for the studied species. The interest here was to quantify the strain imposed on the stems, thus enabling the relationships between the applied sum of strains and the plant growth response to be examined. As mechanical treatment led to variation (between control and bent plants) of diameter growth only, the study of interspecific variability was processed on diameter growth data. Quantitative data: dose – response curves for diameter growth responses

The results revealed correlations between normalized diameter increment and the sum of applied strains for all tested species. This demonstrates that the model of mechanoperception established for a single bending treatment can also be applied to cases of multiple bending, as long as the bending frequency is sufficiently low. The dose – response curves, as well strain threshold values, provide the basic information necessary to begin to model plant growth under mechanical loading. Within the range of experimental data, some of the points of the relationship between standardized diameter increment and sum of applied strains were below 1, indicating that the

346


bending regimen has led to a reduction of diameter growth. In general, the thigmomorphogenetic response is known to stimulate diameter growth rather than inhibit it, but there are some exceptions; a study of the effect of brushing on cauliflower growth revealed a reduction of diameter growth by mechanical loading (Biddington and Dearman, 1985). Similar results have also been reported in woody species (Telewski and Jaffe, 1980; Telewski, 1995; Cordero, 1999). It should be noted that in the study of Cordero (1999) comparing the growth of saplings exposed to wind or protected from wind, the average wind speed of the wind-exposed site was very low (2.5 – 6.8 km h21), which may have led to light deformations of the tree saplings even if not quantified. It is thus possible that mechanical loading could lead to a reduction of diameter growth in the case of a very small sum of strains. The present results also highlight the need for a more precise definition of what a ‘normal’ or ‘average’ mechanical environment is and for a better definition of a ‘control’ plant in such experiments. As plants are always submitted to a background of more-or-less intense mechanical stimuli, the growth observed is the result of possibly numerous sources of stimulation at variable times and for variable periods. The intensity of average wind will depend on the location. Saplings of species studied here grow habitually in the rainforest understorey where they are rarely exposed to strong and continuous wind but will receive mechanical stimulation from rainfall, debris fall and sporadic wind, particularly in more open situations or tree-fall gaps. A possible ecological role of interspecific variability of mechanosensing

The lianoid species B. guianensis and the tree saplings of Eperua exhibit the highest slenderness ratios prior to tests, with E. falcata having the highest threshold of strain of the tested species. Interestingly, the relatively high threshold to strain observed in Bauhinia occurred in a species that showed significantly higher total growth in length and diameter, both in controls and after bending treatments, than the four other species. The result appears to be coherent with the growth strategy of the liana: a higher threshold may mean that the species will not respond to low mechanical signals and so maintain relatively high growth in length rather than diameter. High slenderness ratios combined with relatively high stiffness (Speck and Rowe, 1999) in B. guianansis most probably allow the plant to explore wider distances to detect structures to which to anchor (Isnard and Silk, 2009). It is nevertheless intriguing to notice that among the tested species, stems of Bauhinia exhibited the highest threshold, yet tendrils of B. guianensis appear to be highly mechanosensitive. During the present experiments, these structures twisted rapidly when they encountered a support. It would be thus interesting to compare the mechanosensing parameters of the tendrils with those of stems. The method developed in this paper to study interspecific variability could be equally applied to explore diversity of mechanosensing between organs. Following the idea that mechanosensing may be a regulator of stem slenderness, the result found in tree saplings of E. falcata is surprising in comparison with the result of tree

saplings of E. grandiflora: under natural conditions, E. falcata exhibits a lower slenderness than E. grandiflora (Jaouen, 2008) so that a lower threshold might have been expected. In the present experiments, Tachigali saplings were selfsupporting, but Jaouen et al. (2007) have reported that Tachigali can develop a non-self-supporting habit by leaning against neighbouring trees. This has been interpreted as a biomechanical strategy of growth enabling maximal growth in length over growth in diameter, allowing the saplings to reach the canopy as soon as possible. It would be interesting to study the effect of bending on self-supporting and non-self-supporting individuals in order to determine how the model parameters change with changes in habit. If our interpretation of the Bauhinia liana threshold is valid, one might expect an increase in stimulus threshold among nonself-supporting Tachigali. What are the structural bases of mechanosensing interspecific variability?

The mechanisms underlying the observed interspecific differences in terms of strain threshold remain to be discovered. Variability of responses to shaking was observed in different strains of Pinus taeda (Telewski, 1990) as well as differences in ethylene production (even in control plants), suggesting that the level of ethylene production might affect the thigmomorphogenetic responses. Based on results from recent studies, mechanosensitive channels are thought to be involved in the perception –transduction phase of mechanosensing in plants and some of these mechanosensitive channels have been recently cloned in Arabidopsis roots (Haswell et al., 2008). Channel conductance is different when channels are in homomeric complexes than when they are in heteromeric complexes (Haswell et al., 2008). It is thus possible that differences in strain threshold might rely on different complexes of channels but the link between mechanosensitive channel activities and plant mechanosensing parameters remains to be established. Conclusions

This present study has provided quantitative data about the effect of successive mechanical loading on tree and liana saplings via mechanosensing. The results demonstrated that the ‘sum-of-strains’ model, quantitatively linking the plant growth response and the sum of applied strains and originally developed for cases of single bending treatments, can also be applied for successive loading as long as the loading frequency is sufficiently low. A comparison of parameter values among four tree species and one liana species gave evidence of interspecific variability in mechanosensing. It appears that the species studied exhibited the same sensitivity but differed in terms of the threshold to the mechanosensing stimulus; this might at least partly be explained by differences in growth habit. The results also revealed that construction of models aimed at investigating plants under successive loading will have to take into account a species parameter. Finally, the study provides a conceptual tool for investigating the

Coutand et al. — Interspecific variability of mechanosensing interspecific variability of mechanosensing, which has, to our knowledge, not previously been provided. ACK N OW L E DG E M E N T S We thank Mr J. Weigel (Kourou) for transplanting and taking such great care of the plants used in this study. We are grateful to two anonymous reviewers for their suggestions. This study was funded by the ANR, Woodiversity project ANR-05-BDIV-012-03. L I T E R AT U R E C I T E D Arteca JM, Arteca RN. 1999. A multi-responsive gene encoding 1-aminocyclopropane-1-carboxylate synthase (ACS6) in mature Arabidopsis leaves. Plant Molecular Biology 39: 209 –219. Biddington NL. 1985. A review of mechanically induced stress in plants. Science Horticulturae 36: 12–20. Biddington NL, Dearman AS. 1985. The effect of mechanically induced stress on the growth of cauliflower, lettuce and celery seedlings. Annals of Botany 55: 109 –119. Boyer N. 1967. Modification de la croissance de la tige de bryone (Bryonia dioica) a` la suite d’irritations tactiles. Compte-Rendu de l’Acade´mie des Sciences de Paris 267: 2114–2117. Cordero RA. 1999. Ecophysiology of Cecropia schreberiana saplings in two wind regimes in an elfin cloud forest: growth, gas exchange, architecture and stem biomechanics. Tree Physiology 19: 153– 163. Coutand C, Moulia B. 2000. Biomechanical study of the effect of a controlled bending on tomato stem elongation: local strain sensing and spatial integration of the signal. Journal of Experimental Botany 51: 1825– 1842. Coutand C, Julien JL, Moulia B, Mauget JC, Guitard D. 2000. Biomechanical study of the effect of a controlled bending on tomato stem elongation: global mechanical analysis. Journal of Experimental Botany 51: 1813–1824. Coutand C, Dupraz C, Jaouen G, Ploquin S, Adam B. 2008. Mechanical stimuli regulate the allocation of biomass in trees: demonstration with young Prunus avium trees. Annals of Botany 101: 1421–1432. Coutand C, Martin L, Leblanc-Fournier N, Decourteix M, Moulia B, Julien J-L. 2009. Strain mechanosensing quantitatively controls diameter growth and PtaZFP2 gene expression in poplar. Plant Physiology 151: 223-232. Dagne´lie P. 1975. Theórie et methodes statistiques II. Applications agronomiques, 2nd edn. Gembloux: Les presses agronomiques de Gembloux. Fournier M, Stokes A, Coutand C, Fourcaud T, Moulia B. 2006. Tree biomechanics and growth strategies in the context of forest functional ecology. In: Herrel A, Speck T, Rowe NP. eds. Ecology and biomechanics: a mechanical approach to the ecology of animals and plants. London: Taylor & Francis, 1–33. Hamant O, Heisler MG, Jonsson H, et al. 2008. Developmental patterning by mechanical signals in Arabidopsis. Science 322: 1650– 1655.

347

Haswell ES, Peyronnet R, Barbier-Brygoo H, Meyerowitz EM, Frachisse J-M. 2008. Two MscS homologs provide mechanosensitive channel activities in the Arabidopsis root. Current Biology 18: 730– 734. Isnard S, Silk WK. 2009. Moving with climbing plants from Charles Darwin’s time into the 21st century. American Journal of Botany 96: 1205– 1221. Jaffe MJ. 1973. Thigmomorphogensis: the response of plant growth and development to mechanical stimulation. Planta 114: 143–157. Jacobs MR. 1954. The effect of wind sway on the form and development of Pinus radiata D.Don. Australian Journal of Botany 2: 35–51. Jaouen G, Almeras T, Coutand C, Fournier M. 2007. How to determine sapling buckling risk with only a few measurements. American Journal of Botany 94: 1583–1593. Jaouen J. 2008. Etude des strate´gies biomećaniques de croissance des jeunes arbres en peuplement he´te´roge`ne tropical humide. PhD thesis, Universite´ de Nancy I, France. Larson PR. 1965. Stem forms of young Larix as influenced by wind and pruning. Forest Science 11: 412–424. Martin L. 2009. Etudes e´tapes prećoces de la mećano-perception chez le peuplier : approches biomećanique et molećulaire. PhD thesis, Universite´ Blaise Pascal, Clermont-ferrand, France. Mc Mahon TA. 1973. Size and shape in biology. Elastic criteria impose limits on biological proportions, and consequently on metabolic rates. Science 179: 1201–1204. Meng SX, Lieffers VJ, Reid DEB, Rudnicki M, Silins U, Jin M. 2006. Reducing stem bending increases the height growth of tall pines. Journal of Experimental Botany 57: 3175– 3182. Moulia B, Coutand C, Lenne C. 2006. Posture control and skeletal mechanical acclimation in terrestrial plants: implications for mechanical modeling of plant architecture. American Journal of Botany 93: 1477–1489. Soga K, Wakabayashi K, Kamisaka S, Hoson T. 2006. Hypergravity induces reorientation of cortical microtubules and modifies growth anisotropy in azuki bean epicotyls. Planta 224: 1485–1494. Speck T, Rowe NP. 1999. A quantitative approach for analytically defining, growth form and habit in living and fossil plants. In: Kurmann MH, Hemsley AR. eds. The evolution of plant architecture. Kew: Royal Botanic Gardens, 447–479. Telewski FW. 1990. Growth, wood density, and ethylene production in response to mechanical perturbation in Pinus taeda. Canadian Journal of Forest Research 20: 1277– 1282. Telewski FW. 1995. Wind-induced physiological and developmental responses in trees. In: Coutts MP, Grace J. eds. Wind and trees. Cambridge, Cambridge University Press, 237–263. Telewski FW. 2006. A unified hypothesis of mechanoperception in plants. American Journal of Botany 93: 1466–1476. Telewski FW, Jaffe MJ. 1981. Thigmomorphogenesis: changes in the morphology and chemical composition induced by mechanical perturbation in 6-month-old Pinus taeda seedlings. Canadian Journal of Forest Research 11: 380–387. Telewski FW, Pruyn ML. 1998. Thigmomorphogenesis: a dose response to flexing in Ulmus americana seedlings. Tree Physiology 18: 65–68.