In vivo analysis of local wall stiffness at the shoot apical meristem in Arabidopsis using atomic force microscopy
Summary
Whereas the morphogenesis of developing organisms is relatively well understood at the molecular level, the contribution of the mechanical properties of the cells to shape changes remains largely unknown, mainly because of the lack of quantified biophysical parameters at cellular or subcellular resolution. Here we designed an atomic force microscopy approach to investigate the elastic modulus of the outer cell wall in living shoot apical meristems (SAMs). SAMs are highly organized structures that contain the plant stem cells, and generate all of the aerial organs of the plant. Building on modeling and experimental data, we designed a protocol that is able to measure very local properties, i.e. within 40–100 nm deep into the wall of living meristematic cells. We identified three levels of complexity at the meristem surface, with significant heterogeneity in stiffness at regional, cellular and even subcellular levels. Strikingly, we found that the outer cell wall was much stiffer at the tip of the meristem (5 ± 2 MPa on average), covering the stem cell pool, than on the flanks of the meristem (1.5 ± 0.7 MPa on average). Altogether, these results demonstrate the existence of a multiscale spatialization of the mechanical properties of the meristem surface, in addition to the previously established molecular and cytological zonation of the SAM, correlating with regional growth rate distribution.
Introduction
The growth and shape of the plant is entirely defined by two parameters: namely growth rate and growth anisotropy (Erickson, 1976; Coen et al., 2004). At the cellular level, growth is driven by internal turgor pressure, and is restricted by the ability of the cell wall to extend under this pressure. The cell wall is a complex composite material consisting of stiff cellulose microfibrils, which are tethered by xyloglucans and embedded in a matrix of hydrosoluble pectins and structural proteins (Cosgrove, 2005; Burton et al., 2010). The presence of cellulose microfibrils in highly oriented arrays creates anisotropy in the wall, which is crucial for the growth direction of cells. Although it has been proposed that weakening of the cell wall plays a major role in the control of cell expansion rates (McQueen-Mason and Cosgrove, 1994; Pien et al., 2001; Cosgrove, 2005; Fleming, 2006; Peaucelle et al., 2008), very little is known on the dynamics and mechanics of wall properties at (sub)cellular resolution and in living cells.
Shoot apical meristems (SAMs) are populations of dividing, undifferentiated cells that generate organs at the tips of stems and branches throughout the life of a plant (Aida and Tasaka, 2006; Sablowski, 2007; Traas and Hamant, 2009; Barton, 2010). The SAM is also an intriguing morphogenetic model. Although meristematic cells undergo mitosis, expansion and identity switches, the global shape of the SAM remains relatively stable. Nevertheless, the physics behind this apparent stability remains poorly understood. Note that this is not specific to the SAM, as the root apex also exhibits a stable dome shape for instance.
Based on various microsurgical experiments, the predominant model proposes that the epidermis is under tension, and is limiting for growth in the aerial parts of the plants (Kutschera and Niklas, 2007). In this scenario, the mechanical properties of the outer wall of the epidermis would be the main point of control for shape changes in plants. This hypothesis is also supported by the fact that a dwarf mutant affected in brassinosteroid synthesis or signaling can restore its normal growth if the corresponding wild-type gene is expressed in the epidermis only (Savaldi-Goldstein and Chory, 2008). Because no tool has been able to quantify the mechanical properties of the outer wall with sufficient resolution in living cells, this is still a matter of debate. More generally, although computer simulations can provide plausible mechanistic hypotheses behind plant morphogenesis (e.g. Cui et al., 2010; Green et al., 2010), testing those hypotheses requires a quantitative assessment of the mechanical properties of the tissue.
To address these points, a range of micromechanical approaches have been developed or revived in recent years. Extensometers have been used on hypocotyls and roots to measure creep rates and extract wall extensibility and yield threshold (Cosgrove, 2011). For instance, acid-induced wall extension in maize roots was monitored and correlated with expansin activity using extensometers (Wu et al., 1996). Microindentation techniques have been particularly successful in addressing the mechanical properties of single cell systems, like pollen tubes (Geitmann, 2006a,b). These methods notably showed that wall stiffness decreases right before a phase of pollen tube growth, thus correlating local mechanical properties of the wall with growth (Zerzour et al., 2009). Results from indentations in tissues are more complex to interpret, as the measured output depends on the distribution of wall stiffnesses and turgor pressures, as well as tissue geometry (Geitmann and Ortega, 2009). To understand the mechanics of a growing tissue, it is therefore crucial to obtain quantifications that are as local as possible to uncouple the contribution of the local mechanical properties of the cell wall from that of the tissue.
The spatial resolution of the measure of wall mechanics depends on the size of the indentor. In this respect, atomic force microscopy (AFM), which usually involves the use of a nanometric cantilever, is attracting growing interest in the field. AFM is notably able to provide direct measurement of the mechanical properties of a sample surface, and at any scale. For instance, AFM studies on bacteria, yeast and animal cells in culture have uncovered differences in elasticity between these organisms, and have also shown the impact of chemical or enzymatic treatments on the mechanical properties of the cell (Alonso and Goldmann, 2003; Kumar and Weaver, 2009; Jacot et al., 2010; Scheuring and Dufrene, 2010). Local subcellular heterogeneities have also been revealed thanks to the resolution of the method. For example, it has been shown in budding yeast that the bud scar is 10 times stiffer than in the rest of the wall (Touhami et al., 2003). In plants, AFM has been mainly used to image wall structure at high resolution (Kirby, 2011). For instance, using this technique, parallel cellulose microfibrils were observed in native pollen tube walls after application of the plant hormone auxin (Wu et al., 2008). Mechanical properties of certain plant polymers have been measured (e.g. the adhesive properties of secreted glycoproteins from the green alga Enteromorpha; Callow et al., 2000). AFM has also been used to measure the elastic modulus of extracted walls from Gossypium spp. (cotton), Glycine max (soybean), Oryza sativa (rice) and Triticum spp. (wheat) (Wu et al., 2010). A limitation of these pioneering studies in the context of plant development is that they were not conducted on living tissues.
Here we aim at putting the AFM technique on a firm basis to monitor the mechanical properties of walls in living tissue. More specifically, we developed an AFM protocol to measure the mechanical properties of the SAM surface. The quantifications were local enough to reveal the presence of subcellular heterogeneities in elasticity. The AFM also revealed that the outer wall stiffness is relatively well defined at the cell level. At a regional scale, coherent mechanical properties at the SAM surface were also observed. In particular the presence of very stiff outer walls at the tip of the SAM suggest that, at least in this domain, local growth rates could depend on the mechanical properties of the outer wall.
Results
AFM imaging of the meristem surface at high resolution
To access the surface of the SAM more easily, we used the pin1-6 mutant line, which exhibits the typical dome shape of the SAM (Figure 1a–d). In this mutant, the central zone (CZ) is not severely affected, whereas the peripheral zone (PZ) exhibits a hybrid identity between organ and boundary, based on the expression of marker genes (Vernoux et al., 2000; Kwiatkowska, 2004). The meristems were cut from the plant, glued on a coverslip with adhesive silicon and placed in water at room temperature (20°C) (see Experimental Procedures). Using the p35S::GFP-MBD lines (a microtubule marker line), we were able to check that the meristematic cells were growing in these conditions (Figure 1a–e). Notably, cell divisions were still visible more than 2 h after cutting the meristem from the stem (Figure S1). The topographical features were consistent with SEM images of the SAM surface (see e.g. Laufs et al., 1998): we observed small cells (usually 5–10 μm wide) with a smooth and homogeneous surface, separated by creases between cells (Figure 1f,g). In contrast to SEM qualitative data, the AFM allowed us to quantify this topography: typically, right after cell division, we found a height difference between the top of a cell and the crease of 50–100 nm, and this difference reached 300 nm in older cells (Figure 1h).

Atomic force microscopy (AFM) imaging of the surface of the shoot apical meristem (SAM) at high resolution.(a) FM4-64-stained pin1-6 meristem, (b) GFP-MBD pin1-6 meristem, (c, d) close-ups from (a) and (b), respectively.(e) Diagram of the set-up used to image the SAM and to obtain force curves with AFM.(f) Topography of a small area of the SAM using the tapping mode. Note that the height scale has been magnified by a factor of seven to show the high resolution that can be achieved using AFM to measure tissue topography.(g) Friction image of the surface of a few cells at the tip of the SAM.(h) The topography of two cells at the tip of the SAM is plotted. The resolution of the height axis allows for a precise quantification of the shape of the cell surface. Note that the height scale has been magnified by a factor of 25 in this panel.
As a standard for the rest of the paper, we always made sure that the group of analyzed cells formed a rather flat surface, thus allowing the cantilever to be normal to the surface, and preventing artefacts from a sloping surface.
Interpreting the force curve on the meristem
As we aimed to obtain high-resolution and very local quantifications of the mechanical properties of the outer wall, we extracted the raw force curves that the AFM produced, and developed a protocol to analyze them for a maximum indentation of 100 nm (Figure 2a–c).

Interpreting the force curve on the shoot apical meristem (SAM).(a) A typical deflection curve obtained from the SAM. The abscissae axis (z on the figure) describes the upward motion of the sample towards the cantilever until it reaches it (red dot). The 0 position on this axis is determined empirically by the user, and corresponds to the bending of the cantilever for a deflection maximum imposed by the user. As the surface resists penetration, the cantilever bends and a deflection is recorded. In blue, a similar curve is obtained when approaching the cantilever on mica, a flat and stiff material that serves as a maximal stiffness reference.(b) The standard deviation of the deflection is plotted in order to detect the point of contact (when the standard deviation of the deflection exceeds a noise maximum).(c) Various force–indentation curves can be drawn depending on the exact position of the point of contact.(d) In a logarithmic representation, the green line has a slope of 2, consistent with a pyramidal contact: F = kI2 (Equation 1). The curves obtained in (c) are compared with the theoretical prediction, ln(F) = ln(k) + 2ln(I), to select the model corresponding to a true pyramidal contact (red curve), and from which we estimate the intercept ln(k), and in turn the elastic modulus E = 0.75(1 – n2)k/tan(θ) (Equation 2; see text).(e) Close-up from (d) near the best fit.
(1)
(2)Several curves were excluded based on the following criteria: in a few cases, we could not obtain an acceptable quadratic fit (Equation 1), or the window for the fit with the experimental curve was narrower than 50 nm. In order to compare wall properties at approximately the same position between samples, we also excluded the curves where a quadratic fit could be obtained, but with a range of I values outside the average 40–100-nm window. Altogether, only 16% of the curves were excluded for our analysis.
The measure of the stiffness of the surface of living meristematic cells is local
To measure the local stiffness of the outer wall, we used a cantilever with sufficient stiffness to penetrate into the wall before saturation (approximately 100 nm), but with a low enough stiffness to avoid too dramatic a deformation of the outer wall, which would inevitably be dependent on the internal turgor pressure and/or the presence of an anticlinal wall. Our initial tests led us to use a cantilever with a stiffness of 0.16 N m−1.
To investigate how local our measure was, we first used the finite-element method to model the indentation of an elastic material (details in Appendix S1). We considered the tip to be very stiff, and took into account its exact geometry. We explored a full range of elastic properties for the cell wall, assuming a linear orthotropic material with different elastic moduli Ex, Ey, and Ez in the x, y and z directions, respectively, where z is the direction normal to the wall. In the simulation presented in Figure 3a, the cell wall was 250 nm thick and, according to the orientation of cellulose microfibrils, we considered the wall to be softer in the normal direction (Ex = Ey = 4Ez). We found that for an indentation of about 100 nm, the deformation impacts a volume of radius of the order of 150 nm. Thus the simulations predict a measure of the outer wall properties averaged over a small volume.

Local measurement of the stiffness of the surface of living shoot apical meristem (SAM) cells.(a) A finite-element model assuming that the wall is four times softer in the normal direction, and taking into account the parameters of the atomic force miscroscopy (AFM), predicts a very local deformation of the wall, spanning a volume of radius of approximately 150 nm.(b) Apparent elastic modulus for a turgid cell, measured as ln(k) (see Figure 2).(c) Apparent elastic modulus for a turgid cell at t = 0 and undergoing plasmolysis afterwards, measured as ln(k). The apparent drop in ln(k) at t = 2 min corresponds to the collapse of the tissue following plasmolysis. Because the surface of the meristem moved as the cantilever approached, values between t = 2 min and t = 8 min are meaningless (but they illustrate what the user would observe in a real experiment). When the meristem was stable again (after t = 8 min), values were comparable with those from before plasmolysis.(d) Heterogeneity in elastic moduli and the standard deviation (in MPa) within a cell from the tip of the SAM. At least five force curves were recorded per position on each cell.(e) Topographical image of cell from the stem showing patches of cuticle.(f) Shape of the raw deflection curve during the approach of the cantilever in the meristem (purple) and on the stem (blue).(g) Shape of the raw deflection curve during the retraction of the cantilever in the meristem (purple) and on the stem (blue).
To test this prediction experimentally and in order to exclude the role of turgor in our measurement, we investigated the impact of plasmolysis in meristematic cells from the tip of the SAM. First, we obtained force curves on meristems immersed in water, from which we extracted k and deduced the elastic modulus E (Equations 1 and 2; Figure 3b). Five minutes after the first series of measurements, we replaced the water solution with a hypertonic solution, by injecting a half volume of 0.5 m NaCl solution. The injection of the NaCl solution led to the collapse of the meristem as a result of the flow of water that such treatment induces. Nevertheless, when the cantilever reached the surface again, the elastic modulus was very similar to that obtained before the osmotic stress (Figure 3c). This result is thus consistent with a measure of the intrinsic and local properties of the wall. Incidentally, this shows that we are unable to detect changes in modulus caused by structural modifications following a decrease in wall tension.
To further address this point, we next measured the elastic modulus of the outer wall at different locations of the cell surface. We reasoned that if the deformation encompassed the entire outer wall of a cell, the center of the cell would bend more than the sides of the cells, and thus the apparent elastic modulus should be systematically lower in the center of the cells. As shown in Figure 3d, this was not the case: we obtained a variable elastic modulus within a relatively narrow window in each cell, but we could not correlate the values of the elastic modulus with a position in the cell. In several cases, we actually measured the maximum elastic modulus near the center of the cell, in contradiction with the measurement of the entire outer wall (Figure 3d, and see Figure 4b).

The outer wall of the shoot apical meristem (SAM) exhibits local and regional differences in elastic moduli.(a) Distribution of the elastic moduli of the outer wall at the tip (orange) and flanks (green) of the SAM.(b) Heterogeneity in elastic moduli and their standard deviation (in MPa) within a stiff cell from the tip of the SAM. At least five force curves were recorded per position on each cell.(c) Heterogeneity in elastic moduli and the standard deviation (in MPa) within a cell from the flank of the SAM. At least five force curves were recorded per position on each cell.(d) Distribution of the elastic moduli of two representative cells from the tip of the SAM.(e) Distribution of the elastic moduli of two representative cells from the flanks of the SAM.
We also checked whether our measurement was not too local. Typically, if the measurements were too local, we would only reveal differences in the cuticle thickness or composition. That scenario would be consistent with a relatively constant elasticity in cells, with or without plasmolysis. Given that the thickness of the cuticle reaches 30–50 nm in the epidermis of fully differentiated leaves of Arabidopsis, this seems very unlikely (Kurdyukov et al., 2006; Kosma et al., 2009). To further address the putative impact of the cuticle on the AFM output, we obtained several force curves on the stem, approximately 1 cm below the SAM, where patches of cuticle could be observed (Figure 3e). Interestingly, the shape of the force curve was dramatically affected for those cells, most notably with a strong attraction during the approach, and with an even stronger adhesion during retraction (Figure 3f,g). In fact, the presence of cuticle with these settings, prevented us from obtaining an interpretable force curve. Although we cannot exclude that some of our measurement on the meristem reflects the presence of cuticle, we believe that this can be neglected. To conclude, we think that we are measuring the local mechanical properties of the wall, mostly in the normal direction (see Appendix S1).
The outer wall of the SAM exhibits local and regional differences in elastic moduli
We next investigated whether differences in stiffness could be measured between cells in the meristem, depending on their localization. We found that the outer walls were usually very stiff at the tip (5 ± 2 MPa; n = 487 curves), whereas on the flanks the top walls were much softer on average (1.5 ± 0.7 MPa; n = 1380 curves) (Figure 4a). Data obtained on a wild-type dissected meristem are consistent with these measurements in the pin1-6 background (Figure S2A,B).
In addition to the clear opposition between both domains, the distribution of elastic moduli appeared rather large at the tip (Figure 4a). More specifically, a significant number of elastic moduli were very high (11 ± 2.3 Mpa; Figure 4a). A careful examination of the data showed that these were not randomly distributed, but corresponded to a subset of cells at the tip of the meristem (Figure 4b). A comparison between two representative cells of these populations is represented in Figure 4d. This strongly suggests that the stiffness of the outer wall is regulated at the cellular level, and that individual cells at the tip can maintain specific stiffness values, within a certain window, relatively independently of their neighbours.
The outer walls on the flanks of the meristem seemed rather soft in comparison, and the range of elastic moduli was narrower in this domain (Figure 4c,e). Interestingly, it is now well established that the tip of the meristem grows more slowly than the flanks (Kwiatkowska, 2004). Our quantifications thus suggest a clear correlation between the stiffness of the outer wall and the spatialization of growth rates in the SAM.
Discussion
We describe here an AFM-based method that allows the quantification of the local stiffness of the cell wall in living plant tissues, in a non-invasive manner. Note that the technique was also successfully applied to hypocotyls from 7-day-old in vitro-grown Arabidopsis seedlings (Figure S2C,D). Using this protocol, we reveal a complex spatialization of the mechanical properties of the outer wall in the SAM, at subcellular, cellular and supracellular scales. The SAM is classically pictured as a binary structure, with a tip where slow-dividing stem cells are maintained, and with flanks where cell division and elongation is stimulated, resulting in organogenesis and differentiation. In addition to these histological features, cells at the tip of the SAM exhibit some unique features, including the expression of specific genetic markers (Sablowski, 2007; Barton, 2010), highly dynamic microtubule orientations (Hamant et al., 2008), and a predicted insensitivity to the plant hormone auxin (de Reuille et al., 2006). The high stiffness found at the tip of the SAM further highlights this zonation.
A careful analysis of growth rates in the epidermis has been performed in the pin1 mutant (Kwiatkowska, 2004). Strikingly, it demonstrates the presence of a ring of fast-growing cells surrounding slow-growing cells at the tip. This suggests that the growth rate in the epidermis of the SAM can be correlated with the mechanical properties of the outer wall. An interesting prospect for the AFM method will be to develop correlative microscopy approaches, to precisely assess the position of wall-stiffness measurements with fluorescent landmarks, and investigate whether the boundary between stiff and soft domains forms a gradient or a sharp separation in the meristem. Such an opto-mechanical coupling would also allow us to assess more directly whether local growth rates can be associated with mechanical properties of the outer cell wall, consistent with the hypothesis that the epidermis acts as the main limiting factor for growth in the meristem. In this respect, it has recently been proposed that the alignment of cortical microtubules next to the outer wall is sensitive to the elongation rate (Chan et al., 2011). It would be interesting to check whether wall stiffness could act as an instructional signal in this process.
More generally, at this stage, we cannot rule out the possibility that mechanical properties of the inner walls and internal tissues have a major role in growth patterns. It is therefore crucial to develop other, more invasive, indentation techniques, to compare their outputs with those given with AFM, thus deciphering the contribution of the internal tissues from the outer wall in controlling local growth rates and morphogenesis.
Within the tip, the variability we observed between individual cells could be explained by differences in the position of the cells in the cell cycle. Softer tip cells could also represent a subpopulation of cells switching identity from the central zone to the peripheral zone, and thus becoming competent for organ initiation. In the flanks of the meristem, the outer walls were significantly softer, and their elastic modulus was somewhat comparable with what we observed in 7-day-old hypocotyls (Figure S2C,D). As we now have a way to quantify the mechanical properties at a subcellular resolution in living cells, this opens the path for further studies on mutants with patterning defects in order to dissect the link between the genetic control of cell identity and the mechanical control of growth.
Beyond the implication for SAM biology, this study demonstrates how the quantification of tissue mechanics is essential to understand morphogenesis in growing organisms. AFM certainly has the potential to address other important questions in other tissues, like cell elongation in hypocotyls, patterns of vascularization or the mechanics of guard cells in the leaf epidermis.
Experimental procedures
Plant material and growth conditions
The pin1-6 and p35::GFP-MBD lines are in the WS-4 ecotype, and have been described previously (Vernoux et al., 2000; Hamant et al., 2008).
Confocal microscopy
The SAMs were examined with an inverted or an upright LSM-510 Laser-Scanning Confocal Microscope (Zeiss, http://www.zeiss.com), as described previously (Grandjean et al., 2004). Projections of the signal in the L1 layer were obtained using merryproj (de Reuille et al., 2005).
Atomic force microscopy
The AFM measurements were performed by means of a Nanoscope IIIa (Bruker, http://www.bruker.com) using standard oxide-sharpened silicon nitride probes (NP-S; Bruker). The deflection sensitivity of the cantilever was calibrated on fused silica. Calibration curves of resonant frequency versus spring constant were used to derive a value of 0.16 N m−1 (thermal tuning in PeakForce QNM mode; Bruker). Given the limitations of the scanner in the z direction (7 μm), we only imaged and measured forces on the flat area of the sample. Shoot apices (approximately 1 mm high) were cut from the pin1.6 mutant and stuck on a slide using MED-1356 adhesive silicon (NuSil Silicone Technology, http://www.nusil.com). To access the flanks of the SAM, we either scanned a flat area at the periphery of a meristem or tilted the meristems before fixing them on the slide. Measurements were performed in water at room temperature (20°C). Topographical images were recorded, and subsequently force measurements were performed. Analysis was performed on data from 13 pin1-6 meristems, representing about 3250 force curves.
Acknowledgements
We thank UMS 3444 Platim and Benoit Landrein for help with imaging. This study was supported by the Laboratoire Joliot Curie (ENS/CNRS) and by a grant from Agence Nationale de la Recherche ANR-10-BLAN-1516 ‘Mechastem’.




