Introduction

Understanding phenotypic plasticity of perennial plants like trees in response to wind is challenging in the context of climatic change (Wang et al. 2022). Trees respond to transient mechanical stresses such as wind by adjusting their growth and wood formation through a process called thigmomorphogenesis (Jaffe 1973; Telewski and Jaffe 1981). Locally induced responses of trees to stem bending are well characterized. Transient stem flexures induce significant increase in secondary growth accompanied by the formation of a wood showing different mechanical and anatomical characteristics (Kern et al. 2005; Roignant et al. 2018). All these adjustments provide a mechanical benefit for the behavior of the stem (Niez et al. 2020). These modifications are controlled by rapid and high remodeling of transcriptional responses in the stem bending area (Martin et al. 2010; Pomiès et al. 2017). At a distance from the stimulated area, several experiments conducted under natural conditions showed that longitudinal growth is reduced under wind exposure. This impact on longitudinal growth has been enlightened for Larix laricina (Larson 1965), Prunus avium (Coutand et al. 2008), and the herb Medicago sativa (Moulia 2004). Investigations of the effects of artificial mechanical stimuli, including bending, also revealed the decrease of longitudinal growth in Lycopersicum esculentum (Depege et al. 1997; Coutand et al. 2000), Juglans regia (Leblanc-Fournier et al. 2008), and Populus tremula × alba (Niez et al. 2019). This longitudinal growth change involves a long-distance transfer of information from the mechanically stimulated zone to the plant apex. In plants, long-range signaling in response to several external factors involves diverse actors such as transport of chemical molecules (hormones, proteins, peptides, RNA, microRNAs, reactive oxygen species, etc.) through the plant's vascular system and propagation of electrical signals (Choi et al. 2016). The nature of the information vector that may propagate within a plant after the bending of an organ is still under investigations. In various tree species, after bending of branches or stems, a hydraulic signal, characterized as a hydraulic pressure wave, has been observed (Lopez et al. 2014; Louf et al. 2017). Recently, in poplar, we described the propagation of an electrical response exhibiting distinctive characteristics to other electrical responses characterized to date (Tinturier et al. 2021). The shape of this new type of electrical signal induced by stem bending, a rapid depolarization followed by a repolarization at the initial potential, showed similarities with AP as described by Pickard (1973). These patterns may suggest that this electrical response could be an action potential. However, several experimental points contradicted this analysis: first, the amplitude of the electrical signal peaks at approximately 35 mV near the deformation area of the stem and decreases almost linearly over a 20 cm distance. Second, its propagation velocity is high, at around 15 cm s−1, and decreases with distance. These characteristics were not consistent with the hypothesis of an action potential, for which the amplitude and propagation velocity are both assumed to be constant (Król et al. 2010; Sukhov et al. 2011). Because of these different characteristics, this bending-induced electrical response has been named the “Gradual Potential” (Tinturier et al. 2021).

Two hypotheses have been formulated to explain the generation and propagation of the Gradual Potential: The GP may result from either cell deformation during bending, followed by electrotonic propagation through local current loops via excitable cells, or the passage of the hydraulic pressure wave induced by bending. (Lopez et al. 2014; Louf et al. 2017). The hydraulic pressure wave would likely deform the walls of parenchyma cells in contact with the xylem vessels and/or stretch the plasma membrane thus, initiating the GP by activating mechanosensors like mechanosensitive ion channels (Hamilton et al. 2015; Frachisse et al. 2020). The generation of a pressure wave in plants requires specific parameters. This involves the presence of a hydraulic network for conducting fluids within a rigid structure (Louf et al. 2017) and the mechanical properties of the hydraulic network driving the amplitude of the signal. This behavior was confirmed at the inter-specific level for natural branches of Populus, Pinus, and Quercus and validated with a biomimetic approach on artificial branches (Louf et al. 2017). However, the possible attenuation of the pressure wave was not investigated. Nevertheless, it was proposed that the pressure wave could propagate initially in a ballistic regime and then could shift to a diffusive regime when viscous effects become dominant over inertia, accompanied by damping (Louf et al. 2017). In the context of a hydro-electric coupling (second hypothesis), the rapid attenuation of the GP following stem bending in Populus as it exits the deformed zone would suggest an immediate transition of the pressure wave into a diffusive regime.

In this context, the present study aimed to perform an in-depth analysis of propagation properties of GP to confirm the high velocity described by Tinturier et al. (2021). Moreover, we tested the genericity of this typical bending-induced electrical response by comparing the effect of stem bending between poplar and a gymnosperm species, Pseudotsuga menziesii, (Douglas-fir), characterized by the presence of a different conducting hydraulic network.

Materials and methods

Plant material and culture conditions

Young poplar (Populus tremula × alba, clone INRA 717-1B4) were obtained by in vitro micropropagation (Leple et al. 1992). Once they reached a height of about 4 cm, the acclimation process from in vitro to hydroponic solution began through decreasing relative humidity (Martin et al. 2009). Trees were then placed in a growth chamber (16 h/8 h light/dark cycle at 40 μmol m−2 s−1 and 22 °C/18 °C with air relative humidity of 60%). Four months after micropropagation, 40 poplar were used for the experiments. At this stage, stems were about 77.8 cm (± 1.5 SE) tall with an average diameter of 5.8 mm (± 0.3 SE).

Douglas-fir trees (Pseudotsuga menziesii, variety PME-VG-04-France 1 VG) were grown in pots by the company Dubost Foret, France. For the stem bending experiments, 9 plants were used. The trees were 4 years old. At this stage, stems were about 82 cm (± 5.3 SE).

Prior to electrophysiology experiments, each plant was moved into a Faraday cage in ambient laboratory conditions (16 h/8 h light/dark cycle at 20 μmol m−2 s−1 and 22 °C/20 °C). Plants were fixed in two points of the stem with clamping rings (Fig. 1). Foam was rolled around each part of the stem before tightening the clamping rings to avoid stem wounds and to allow possible stem diameter variations. In the case of poplars, the root system was immersed in a 20 L tank filled with hydroponic solution.

Fig. 1
figure 1

Experimental setup for electrophysiological measurements after stem bending of A Populus tremula × alba and B Douglas-fir. The plants are placed in Faraday cages and fixed at two points on the stem using clamping rings (tree support). Two centimeters above, a circular template (constant radius) is positioned against the stem. The stem is bent by a motorized arm (motor + translation stage). The electrodes are placed in the immobile portion of the stem. Electrode e1 is inserted 0.2 cm below the stem–template contact. Subsequently, the distances are as follows: e1–e2 = 2 cm, e2–e3 = 3 cm, e3–e4 = 5 cm, e4–e5 = 5 cm, and e5–e6 = 5 cm. The potential difference is monitored between each measurement electrode and a reference electrode (ref) placed in the bath for poplar and in the soil for Douglas-fir

Full size image

Stem bending

The mechanical treatment consisted of a transient curvature of a 12 cm-long stem segment placed 35 cm below the apex (Fig. 1). The leaves on this bent segment were previously removed with a razor blade to avoid uncontrolled mechanical stimuli and leaf wounding. The speed and the magnitude of the bending stimulation were controlled by a motorized arm (hybrid stepping motors 17PM-H311-P1, Minebea.Co) that pushed the stem against a plastic template, which had a constant radius of curvature (Tinturier et al. 2021). The speed of the motor was fixed to apply a bending time of 2 s (go and return). The strain magnitude on the stem periphery was controlled by the radius of curvature of the plastic template. The setup was adjusted to apply a maximal peripheral longitudinal strain εmax of around 2% according to the following equation (Coutand et al. 2009; Moulia et al. 2015):

$${\varepsilon_{\max }}\;( \%) = 100\; \times {\frac{D} {D + 2 \times \rho }},$$

where ρ is the radius of curvature of the plastic template and D is the mean diameter of the stem in the direction of the bending. This value is high enough to generate significant thigmomorphogenetic responses like secondary growth increase (Niez et al. 2019).

Monitoring of extracellular electrical signals

The measurement setup followed the general method described in Tinturier et al. (2021). Shortly, for each experiment, only one single tree was placed in a Faraday cage to minimize interferences. All electronic materials were located outside the cage. A needle with an approximate diameter of 0.5 mm was manually used to create a pilot hole of approximately 2.5–3 mm in depth prior to the insertion of the measurement electrode. Six measuring electrodes [tinned copper wire (359–835), 0.25 mm diameter, RS Components] were inserted into the stem at different heights, passing through all the tissue to the pith. They measured electrical potential simultaneously near the bending area and up to 20 cm distant. After the installation was completed, the plant was left undisturbed for stabilization for 24–48 h.

In the case of experiments on poplars, the reference electrode (RC3 model, World Precision Instruments) was made of an Ag/AgCl wire immersed in the nutrient solution. In the case of the Douglas-fir experiment, the reference electrode was placed directly in the ground of the pot. The electrical potential is the difference between a measuring electrode in the stem and the reference electrode. In contrast to intracellular measurements, extracellular measurements show potential changes downward for depolarization and upward for hyperpolarization. The first electrode (e1) was inserted 0.2 cm below the stem-to-template contact (Fig. 1). The other respective distances of each electrode from e1 were 2 cm (e2), 5 cm (e3), 10 cm (e4), 15 cm (e5), and 20 cm (e6). The propagation velocity of GPs was determined by carefully measuring the fall time of each measurement electrode and precisely accounting for the inter-electrode distance (Fig. 2A). Data analysis included fitting exponential decay curves to quantify the velocity changes of GPs as a function of distance from the bending zone. All the electrodes, including the reference electrode, were connected to a cDAQ-9171 (National Instrument) electronic card. The electronic card was used as an impedance amplifier (10 GΩ) and A/D converter. DAQ Express 1.0.1 software (National Instrument) recorded the potential difference with a sampling rate of 200 Hz. The graphs (Figs. 2A, 3A) were built using MATLAB® software.

Fig. 2
figure 2

Gradual potential generated by stem bending in Populus tremula × alba. A Representative example of electrophysiological recording using 6 measurement electrodes showing a gradual potential induced by a transient stem bending applied on the stem (black arrow). The potential difference of each measurement electrode decreases sequentially according to the distance from the bending site. In the zoomed part of the figure, colored arrows indicate the depolarization time used to calculate the average speed of the gradual potential. B Evolution of the average speed of the gradual potential as a function of the distance from the bending area. C Evolution of the amplitude of the gradual potential as a function of the distance from the bending site. D Evolution of the half-life of the gradual potential as a function of the distance from the bending area. The half-life time of the signal was defined as the duration at half amplitude. Data from 40 biological replicates were analyzed using the non-parametric Kruskal–Wallis test followed by a post hoc Conover–Iman test (P < 0.05). Letters above boxplots indicate significant differences between groups. For each box, the black line represents the median and the red dot represents the mean

Full size image
Fig. 3
figure 3

Gradual potential generated by stem bending in Douglas-fir. A Representative example of electrophysiological recording using 6 measurement electrodes showing a gradual potential induced by bending in Douglas-fir. In the zoomed part of the figure, colored arrows indicate the depolarization time used to calculate the average speed of the gradual potential. B The depolarization time is used to estimate the average speed of the gradual potential between each electrode. C Evolution of the amplitude of the gradual potential as a function of the distance from the bending area. D Evolution of the half-life of the gradual potential as a function of the distance from the bending area. Data from ten biological replicates were analyzed using the non-parametric Kruskal–Wallis test followed by a post hoc Conover–Iman test (P < 0.05). Letters above boxplots indicate significant differences between groups. For each box, the black line represents the median and the red dot represents the mean

Full size image

Statistical analysis

All data were statistically analyzed using R software. Kruskal and Wallis’s tests and post hoc Conover–Iman tests were performed to compare results in terms of amplitude and half-life. The significance of differences was evaluated according to the P value (P < 0.05).

Results

GP generation in response to poplar stem bending

The transient flexure of poplar stem immediately triggered the propagation of a GP over a distance of 20 cm beyond the deformed area (Fig. 2A). In the proximity of the bent zone, GP exhibited its highest propagation velocity, averaging approximately 15 cm.s−1 (Fig. 2B). This velocity decreased exponentially (R2 = 0.979) to approximately 3 cm.s−1 at 17.5 cm from the bent zone. The amplitude of the GP was also at its maximum near the deformation area, averaging 36 mV (± 1 SE) (Fig. 2C). Subsequently, it declined almost linearly (R2 = 0.764) to an average of 0.9 mV (± 0.4 SE) at 20 cm from the bending site. The GP was further characterized by its half-life, corresponding to the duration at half amplitude. On average, it was approximately 49.7 s (± 3.5 SE) at its maximum near the bent zone and gradually decreased to an average of 4.2 s (± 1.5 SE) at 20 cm (Fig. 2D). This in-depth analysis of velocity and amplitude of the bending-induced electrical signal highlights the marked differences from an action potential. These results confirmed our previous results and the need to classify this electrical signal as a Gradual Potential (Tinturier et al. 2021).

Comparison of electrophysiological responses after stem bending: Douglas-fir versus poplar

A total of ten Douglas-fir trees were successively placed in a Faraday cage and experienced a bending stimulation (Fig. 1B). Every bending event of Douglas-fir stems induced the propagation of a depolarization wave that exhibited the typical GP characteristics (Fig. 3A). Similar to the GP in poplar, the propagation velocity of this wave reached its maximum in the proximity of the deformation (Fig. 3B), at approximately 10 cm s−1, and then decreased exponentially (R2 = 1). However, the amplitude and half-life of the GP measured in Douglas-fir were lower than in poplar. As shown in Fig. 3C, the amplitude of the response was at its maximum near the deformation area, with an average of 3.9 mV (± 0.8 SE), decreased to 1.4 mV (± 0.2 SE) at 2 cm from the stressed zone and further to 0.5 mV (± 0.2 SE) at 5 cm. The propagation distance was consequently reduced: only one GP signal reached 10 cm, while the nine others dissipated in a range between 5 and 10 cm. The reduced amplitude and shorter propagation distance of the GP signal in Douglas-fir resulted in a maximum propagation distance two-to-four times shorter than that observed in poplar. The half-life of the electrical response induced by bending averaged 9 s near the deformation and gradually damped with distance (Fig. 3D). The average propagation velocity was also lower in Douglas-fir than in poplar, with a maximum of 10 cm s−1 in Douglas-fir compared to approximately 15 cm s−1 in poplar.

Discussion

The comparison of electrical signal propagation after stem bending in two tree species confirmed that bending generates a signal for which velocity propagation still exceeds the electrical responses observed in non-motor plants by a factor of 10–100 (Fromm and Lautner 2007; Sukhov et al. 2011; Huber and Bauerle 2016). Moreover, in both species, the average velocity decreased exponentially with the distance from the bent area (Figs. 2B, 3B). However, exploring the difference of amplitude and propagation of the gradual potentials observed in the two species after an identical mechanical stimulation is of great interest.

One first possible hypothesis could be technical and linked to our experimental setup that slightly differed between Douglas-fir and poplar. In the case of Douglas-fir, trees were potted, and the reference electrode was inserted into the moist soil near the stem. In the case of poplars, the trees had their roots immersed in a nutrient solution, where the reference electrode was placed. The path between the reference electrode and the measurement electrodes may potentially generate less resistance in the solution than in the soil, leading to a possible attenuated GP in Douglas-fir. The hydroponic growing conditions for poplar were chosen previously because of the rapid growth of poplars in these conditions and the future possibility of easily adding chemical substances (ROS inhibitors, calcium channels, etc.) to understand the molecular players behind this signal. However, Douglas-fir did not seem to tolerate these growing conditions in our context. To check this hypothesis, a poplar plant was grown in a pot and the experiment was repeated with the reference electrode inserted in the pot (supplemental Fig. 1). It turned out that the GP was totally similar in amplitude to the case of trees grown hydroponically with roots immersed in the solution. This result suggests that we can rule out this hypothesis based on set-up differences.

A second hypothesis is that the ability to generate and propagate electrical signals could follow different processes in gymnosperms and angiosperms. The electrical responses generated and transmitted by gymnosperms have been scarcely studied. Only one publication reports electrophysiological measurements in gymnosperms (Asher 1968). Asher measured the electrical response to the bending of pine needles using extracellular electrophysiological measurement methods. Using two measurement electrodes placed on either side of the stimulation, the author recorded a potential variation with a shape similar to the GP. The average amplitude was approximately 0.76 mV (400 technical repetitions on 8 Pinus seedlings), a value close to that of the GP we measured in the Douglas-fir stem at 2 and 5 cm from the bent area. Unlike angiosperms, for which electrophysiological studies in response to a wide range of stimuli have been conducted (Sukhov et al. 2019), no typical plant signals, such as action potentials, slow waves, and system potentials, have been measured in gymnosperms and could be compared to the described GP.

The differences between poplar and Douglas-fir GPs could be explained using the hydraulic–electric coupling hypothesis. In 2021, we proposed that the GP generated by stem bending resulted from the hydraulic pressure wave induced by this mechanical stimulation (Lopez et al. 2014; Louf et al. 2017). The hydraulic pressure wave (PW) would trigger depolarization by activating mechanoreceptors in the neighboring cells of the xylem hydraulic network through which it propagates. The amplitude of depolarization would be proportional to that of the PW. The amplitude of PW initially depends on the longitudinal deformations applied during bending (Louf et al. 2017) and then propagates following a diffusive regime that is dependent on the viscosity of the xylem and the geometric structure of its hydraulic conduits. The geometry of anatomical structures of wood differs between gymnosperms and angiosperms and can affect the attenuation of the PW. For Douglas-fir, the xylem is composed of small tracheids that are typically 1–2 mm in length and have an average diameter of 10 µm (Dunham et al. 2007; Peterson et al. 2007). These tracheids overlap at their ends and are connected by pit pores, allowing for slow sap conduction. In Populus tremula × alba, xylem includes large vessels about 150 mm long and 40 µm in diameter (Lemaire et al. 2021). Pit pores connect the vessels together at their ends. Other pits, on the radial walls, link the vessels to adjacent parenchyma cells. In both systems, the PW needs to pass through the small pits that would generate an energy loss leading to the attenuation of the signal. It can be assumed that, for a given distance, the PW needs to pass through a greater number of tracheids and pits in Douglas-fir compared to the hydraulic network of poplar with long vessels. Therefore, if as hypothesized, the GP is generated through PW signaling, these anatomical differences between Douglas-fir and poplar could explain the higher attenuation of the GP propagation we observed in Douglas-fir.

Conclusion and prospects

This study confirms the spreading of a Gradual Potential in both poplar and Douglas-fir stems after a bending stimulation. Differences in amplitude and propagation distance between GPs in Douglas-fir and poplar suggest the importance of stem anatomical structures in GP generation and propagation.

These findings provide a basis for further investigation. Future research could expand this inquiry to a wider range of gymnosperms species to determine the universality or specificity of these observed trends. Additionally, a comparative study with diverse anatomical features even in angiosperms or even using a biomimetic approach linked with physical modeling of fluid propagation could provide a more comprehensive understanding of the interplay between tree structure and GP dynamics. Furthermore, to enhance our comprehension of GP mechanisms, it is essential to conduct a detailed investigation of different flexion parameters. This includes deformation intensity, bent stem length, and flexion speed rate. Such an analysis would be instrumental in improving our models. This task aims to explore the potential function of the electrical signal as a means of long-distance communication. It raises intriguing questions about the inherent limitations or specificities in the propagation of the GP signal over extended distances. Understanding these nuances could shed light on the broader implications of such signaling mechanisms in tree physiology, potentially revealing adaptive strategies or constraints related to growth regulation across different environments and species.