bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 1 Cuticle architecture and mechanical properties: a functional 2 relationship delineated through correlated multimodal imaging 3 Nicolas Reynoud1, Nathalie Geneix1, Angelina D’Orlando1,2, Johann Petit3, Jeremie 4 Mathurin4, Ariane Deniset-Besseau4, Didier Marion1, Christophe Rothan3, Marc 5 Lahaye1, Bénédicte Bakan1* 6 1 7 Cedex3, France 8 2 INRAE PROBE research infrastructure, BIBS Facility, F- 44300, Nantes, France 9 3 INRAE, Univ. Bordeaux, UMR BFP, F-33140, Villenave d’Ornon, France. 10 4 Institut de Chimie Physique, UMR8000, Université Paris-Saclay, CNRS, 91405 11 Orsay, France 12 * Author for correspondence: benedicte.bakan@inrae.fr INRAE, Unité Biopolymères, Interactions, Assemblages, BP71627 44316, Nantes 13 14 Main text 5376 words 15 Abstract 179 words 16 Number of figures: 6 17 Abstract 18 • Cuticle are multifunctional hydrophobic biocomposites that protect aerial 19 organs of plants. Along plant development, plant cuticle must accommodate 20 different mechanical constraints combining extensibility and stiffness, the 21 corresponding structure-function relationships are unknown. Recent data 22 showed a fine architectural tuning of the cuticle architecture and the 23 corresponding chemical clusters along fruit development which raise the 24 question of their impact on the mechanical properties of the cuticle. 25 • We investigated the in-depth nanomechanical properties of tomato fruit cuticle 26 from early development to ripening, in relation to chemical and structural 27 heterogeneities by developing a correlative multimodal imaging approach. 28 • Unprecedented sharps heterogeneities were evidenced with the highlighting of 29 an in-depth mechanical gradient and a ‘soft’ central furrow that were 30 maintained throughout the plant development despite the overall increase in 31 elastic modulus. In addition, we demonstrated that these local mechanical 32 areas are correlated to chemical and structural gradients. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 33 • This study shed light on a fine tuning of mechanical properties of cuticle 34 through the modulation of their architecture, providing new insight for our 35 understanding of structure-function relationships of plant cuticle and for the 36 design of biosinpired material. 37 Keys words: AFM PF-QNM, correlated multimodal imaging, hyperspectral, 38 nanomechanical, plant cuticle, Raman, Solanum lycopersicum, 39 40 41 42 Introduction 43 Plant terrestrialization coincided with the development of the cuticle at the 44 surface of aerial organs, to cope with harsh desiccating and UV-rich conditions 45 (Niklas et al., 2017; Jiao et al., 2020) The cuticle fulfils multiple biological functions 46 including the regulation of water and gas exchanges and the protection against 47 environmental stresses (Martin & Rose, 2014; Fernández et al., 2021). Plant cuticle 48 is also critical during the plant development as it prevents organ fusion (Sieber et al., 49 2000; Ingram & Nawrath, 2017; Renault et al., 2017) and provides a biomechanical 50 support for the maintenance of the physical integrity of cuticle throughout the 51 development and expansion of plant organs (Bargel & Neinhuis, 2005; Knoche & 52 Lang, 2017). 53 54 Actually, the mechanical properties play a key role on the biological functions 55 of cuticles, especially by supporting plant growth and resistance to environmental 56 stress. Many studies have investigated these properties through tensile tests on 57 isolated skins or cuticles (Petracek & Bukovac, 1995; Wiedemann & Neinhuis, 1998; 58 Bargel & Neinhuis, 2004, 2005; Matas et al., 2004a,b; López-Casado et al., 2007; 59 Domínguez et al., 2009; Lopez-Casado et al., 2010; Takahashi et al., 2012; Tsubaki 60 et al., 2013; Khanal et al., 2013; Khanal & Knoche, 2014; Benítez et al., 2021). 61 Because of its thick, astomatous and easy-to-isolate cuticle, the tomato (Solanum 62 lycopersicum) fruit is a convenient model to study the mechanical properties of the 63 cuticles at different scales. In facts, most of the available knowledge has been 64 obtained on tomato fruit. The well-defined growth of tomato (Guillet et al., 2002; 65 Renaudin et al., 2017) is also advantageous for studying cuticle stiffness and bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 66 extensibility, the balance of which is necessary to avoid cuticle rupture in developing 67 organs. In particular, the cuticle must accommodate the massive increase in volume 68 during fruit expansion while ripening provokes the disassembly of cell walls, resulting 69 in higher mechanical stress in the cuticle (Domínguez et al., 2009; Knoche & Lang, 70 2017; Jiang et al., 2019). Taken together, it appears that the mechanical properties of 71 the cuticle depend primarily on the assembly and interactions of cuticular 72 components, which evolve during organ development. 73 74 Indeed, the plant cuticle is a biocomposite made of a complex supramolecular 75 assembly of monomeric and polymeric lipids, polysaccharides and phenolics 76 (Fernández et al., 2016; Philippe et al., 2020; Reynoud et al., 2022). Cutin, the 77 hydrophobic scaffold of cuticles is an insoluble polyester of oxygenated fatty acids 78 (Hunneman & Eglinton, 1972; Bhanot et al., 2021) whose polymerization index can 79 vary during the plant development (Philippe et al., 2016). The cutin matrix is further 80 filled and coated by waxes (Busta & Jetter, 2018; Lee & Suh, 2022). Besides, 81 polysaccharides are also entangled in the cuticle (CEP, cuticle-embedded 82 polysaccharides). The CEP comprise highly methyl- and acetyl-esterified pectins, 83 hemicelluloses and crystalline cellulose, which are structurally distinct from the non- 84 cutinized cell wall polysaccharides (NCP). (Philippe et al., 2020). Finally, phenolic 85 compounds including esterified hydroxycinnamic acids (Riley & Kolattukudy, 1975; 86 Graça & Lamosa, 2010) and free and bound flavonoids are also found in the cutin 87 polyester (Hunt & Baker, 1980; Reynoud et al., 2022). The composition and 88 proportion of plant cuticle components vary during fruit growth and ripening 89 (Domínguez et al., 2008; España et al., 2014b). Recently, fine chemical Raman 90 mapping of the tomato fruit cuticle revealed the in-depth spatial heterogeneity of its 91 components (lipids, polysaccharides, phenolics) (González Moreno et al., 2022) 92 including the cutin polymer matrix (Reynoud et al., 2022). The additional observation 93 of the fine architectural tuning of the corresponding chemical clusters along fruit 94 development raise the question of their impact on the mechanical properties of the 95 cuticle. To gain further insights into these relationships, detailed information, at the 96 nanoscale level, on the mechanical properties of the cuticle and their changes in the 97 developing fruit are required. 98 bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 99 To assess the in-depth mechanical properties of the Cutin Polymer Matrix 100 (CPM) over cherry tomato fruit development in relation to chemical and structural 101 heterogeneities, we designed a correlated multimodal imaging methodology. Atomic 102 Force Microscopy (AFM) provides the sensitivity to address the mechanical 103 properties of tissues at the nanometric scale (Dufrêne et al., 2017). This technique 104 has been successful in depicting the mechanical heterogeneity of a wide range of 105 biological tissues including cell wall and synthetic polymers (Xi et al., 2015; Melelli et 106 al., 2020; Stoica et al., 2021). AFM was also used to probe the surface of tomato 107 cuticle (Round et al., 2000; Isaacson et al., 2009). Focusing on the polymeric 108 structure of the plant cuticle, we combined Peak-Force Quantitative Nanomapping 109 (PF-QNM) mode of AFM with hyperspectral imaging techniques such as confocal 110 Raman microspectroscopy and Optical PhotoThermal InfraRed (OPTIR) (Zhang et 111 al., 2016). These approaches revealed that the CPM exhibits distinct in-depth 112 mechanical areas with specific dynamics during the phases of cell expansion and 113 fruit ripening that are correlated with chemical and structural gradients. 114 Material and methods 115 Sample preparation 116 Cherry tomato (S. lycopersicum var. cerasiforme WVa106) plants were grown in 117 controlled glasshouses as previously described (Alhagdow et al., 2007). Flowers were 118 tagged at anthesis and fruits at 10, 15, 20, 25, 30, 35 and 40 Days Post Anthesis (DPA) 119 were harvested. Exocarp was collected at the equatorial part of the fruits, chemically 120 fixed and impregnated in paraffin as previously described (Reynoud et al., 2022). 121 Paraffin blocks were cut with an ultramicrotome (Leica EM UC7, Leica Microsystems 122 SAS, Nanterre, France) equipped with diamond knifes (Histo and Ultra, Diatome, Nidau, 123 Switzerland) to obtain 1µm-thick cross-sections. Sections were placed on BaF2 windows 124 and paraffin was removed with successive baths of methylcyclohexane, ethanol and 125 water. In addition, to study the susceptibility to cutinase of the cutin polymer matrix 126 (CPM), 1µm-thick cross sections for 20 and 35 DPA stages were subjected to cutinase 127 (Humicola insulens NZ 51032, Chiral Vision) treatment for 4h at 37°C as previously 128 described (Philippe et al., 2020). 129 In parallel, tomato fruits were peeled off and skins prepared as isolated cuticle as 130 previously described (Philippe et al., 2016). Cutin samples were obtained after removing 131 waxes and non-bound phenolic with chloroforme:methanol (2/1, v/v). bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 132 133 Correlated Multimodal Imaging 134 Correlated multimodal imaging was performed on the same sample per 135 developmental stages and measurements were conducted in controlled environment 136 (ambient air pressure and temperature of 20-25°C). 137 138 Atomic Force Microscopy (AFM) 139 Measurements of the tomato CPM nanomechanical properties were performed 140 on a multimode 8 atomic force microscope (Bruker Nano Surface, Santa Barbara, 141 CA, USA) equipped with RTESPA-150 (Bruker AFM Probes, Camarillo, CA, USA) 142 and operated in Peak-Force Quantitative Nanoscale Mechanical mapping (PF-QNM). 143 A deflection sensitivity of 12 nm.v-1 was determined on a sapphire surface (12M PF- 144 QNM, Bruker AFM Probes, Camarillo, CA, USA) with an elastic modulus of 400GPa. 145 The spring constant was determined after the thermal tune procedure using the 146 Sader’s methods (https://sadermethod.org/; Sader et al., 1995). Before and after 147 acquisition of a set of samples, calibrations of tips radius were evaluated using PS- 148 LMDE-12M (Bruker AFM Probes, Camarillo, CA, USA), i.e., a blend of PolyStyrene 149 (PS) with an elastic modulus of 2 GPa and Low-Density Polyolefin Elastomer (LMDE) 150 with a module of elasticy of 0.16 GPa. The spring constant ranged between 10 and 151 13 N/m and the tip radius between 8 and 10 nm for the probes used in this study. 152 Nanomechanical imaging was performed with a Peak force setpoint set at 6nN, and 153 an oscillation of 160kHz and 40µm/s for scan rate. Hertz’s Model (Hertz, 1882) was 154 chosen over the Derjaguin-Muller-Toporov (DMT) (Derjaguin et al., 1975) to process 155 force-distance curve as the fit was better and adhesion forces were found to be quite 156 homogeneous in our samples. Moreover, the indentation applied was below 1% of 157 the sample height, meeting the rule for the Hertz model (Zdunek & Kurenda, 2013). 158 Several acquisitions per sample were done on 1µm-thick cross-sections and at the 159 surface of isolated cutin sample with at least two images per developmental stage 160 processed with Gwyddion software (http://gwyddion.net; Nečas and Klapetek, 2012). 161 Topographic images were levelled by least square method and mechanical 162 measurements such as profiles or sampling in CPM areas, were performed on 163 filtered elastic modulus maps to remove outliers, i.e., elastic modulus higher than the 164 upper limit of the AFM probes were discarded. After assessing normality and bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 165 homoscedasticity, mean comparison were achieved through two-way parametric 166 ANOVA and further evaluated through Tukey HSD using R software (www.r- 167 project.org) and “FactoMineR” package (Lê et al., 2008). 168 169 Confocal Raman microspectroscopy 170 Raman imaging was performed with an inViaTM Renishaw confocal Raman 171 microspectrophotometer and operated as previously described (Reynoud et al., 172 2022) with minor adjustments. Maps were recorded on the same area as acquired 173 with AFM, with spatial resolutions of 0.5 µm in both x- and y-directions. Cosmic rays 174 were removed from Raman spectra using the WiRE 4.23 software (Renishaw, UK). 175 To improve the signal to noise ratio and the specificity of Raman signals, an 176 Extended Multiplicative Scattered Correction (EMSC) method (Kerr & Hennelly, 177 2016) combined with Principal Components Analysis (PCA) noise filtering PCA (He 178 et al., 2020) was used. Preprocessing was executed in Quasar software (version 179 1.5.0, https://quasar.codes) and spectra further analysed in R software. 180 In parallel, molecular orientations within the CPM were investigated at 20 and 181 40 DPA stages as previously described (Reynoud et al., 2022) by acquiring the same 182 area with distinct laser polarization directions: one set parallel to the plane of 183 incidence (0°) and the other one set orthogonal to the plane of incidence (90°). 184 185 Optical PhotoThermal InfraRed (OPTIR) 186 OPTIR imaging was performed on a mIRageTM Infrared microscope 187 (Photothermal Spectroscopy Corp., Santa Barbara, CA, USA). Samples, i.e., 1µm- 188 thick cross-sections on BaF2 window (see ‘Sample preparation’ subsection), at 20, 30 189 and 40 DPA stage were placed on the motorized plate and the same area recorded 190 in Raman and AFM was mapped. The IR source was a pulsed, tunable four-stage 191 QCL device, scanning from 950 to 1950 cm-1, with the power set to 13% and duty 192 cycle of 1%. The probe laser was a visible laser at 532 nm with the power set at 193 5.7%. Before measurement, calibration was performed on carbon black reference to 194 optimize laser position for 1850, 1550, 1300 and 1050 cm-1, each wavenumber 195 corresponds to one chip of the QCL device. Mapping was performed with a 40× 196 objective (Schwarzschild, NA = 0.78) achieving a spatial resolution of around 0.5 µm bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 197 in both x and y directions with spectral data points spaced by 2 cm-1. Each data point 198 is an average of four spectra. 199 200 Correlative analysis 201 In order to associate mechanical properties with chemical information, Raman 202 or OPTIR maps were first subjected to pixel co-registration to spatially colocalize 203 chemical data with mechanical data from elastic modulus maps. The white light 204 images of the area recorded in Raman or OPTIR were superimposed onto AFM 205 topographic maps using at least 10 visual features as anchoring points. These linear 206 transformations were performed in Mountains Map 9 (Digital Surf, Besançon, France) 207 software and newly spatialized mechanical maps were extracted to be further 208 processed in R software. As AFM and Raman or OPTIR imaging have distinct spatial 209 resolution, degradation of AFM resolution by averaging was perform with “raster” 210 (https://cran.r-project.org/web/packages/raster/) 211 project.org/web/packages/terra/) packages to fit the 0.5 µm²-size of pixel in Raman 212 and OPTIR. Inter-modality Pearson correlations between hyperspectral data and 213 mechanical properties were conducted using R package “Hmisc” (https://cran.r- 214 project.org/web/packages/Hmisc/index.html). and “terra” (https://cran.r- 215 216 Results 217 The mechanical properties of the tomato cutin polymer matrix display 218 spatiotemporal heterogeneities 219 To assess the in-depth mechanical properties of the cutin polymer matrix of the 220 tomato fruit from early developmental stages (10 DPA) to ripening (40 DPA), cross- 221 sections of isolated exocarps were imaged by AFM with PF-QNM mapping mode. 222 First, to validate our experimental set-up, we compared at the same development 223 stages and with identical AFM conditions, the mechanical properties, i) at the surface 224 of isolated non-fixed tomato fruit cutin (Fig. 1a) and ii) on the outer edge of the fixed 225 cutin cross-sections (Fig. 1b). Young’s modulus of the surface was assessed on 226 cuticular ridges between two depressions (Fig. 1a, square), which represent 227 epidermal cell boundaries (Isaacson et al., 2009). From 15 DPA to 40 DPA a 228 significant increase of the Young’s modulus, from 145 ± 52 to 542 ± 302 MPa, was bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 229 observed. These results perfectly fit previous data obtained from isolated tomato 230 cuticles during the fruit growth and ripening (Andrews et al., 2002; Bargel & Neinhuis, 231 2005; Benítez et al., 2021). Likewise, from 15 to 40 DPA, Young’s modulus at the 232 edge of cuticle cross-sections significantly increased from 139 ± 115 to 697 ± 415 233 MPa (Fig. 1c). Despite these differences between absolute values, a similar increase 234 was observed with a positive correlation (0.45, p-value < 0.001). Therefore, we 235 further imaged the tomato fruit cross-sections from 10 to 40 DPA (Fig. 2). 236 Unprecedented sharp differences in the Young’s modulus within the CPM were 237 highlighted (Fig. 2c). For instance, a gradient in the elastic modulus was observed 238 from the external part of the cuticle to the cell wall. In addition, a clear central furrow 239 with a lower Young’s modulus was evidenced during fruit growth, from 15 DPA to 40 240 DPA (Fig. 2c). The absence of a central furrow at 10 DPA suggests that it is formed 241 later, concomitantly with the cuticular peg (Segado et al., 2016) and onset of rapid 242 deposition of cutin in the cuticle (Petit et al., 2014). At the 30-35 DPA stage, the 243 furrow did not span the entire in-depth of the CPM, but divided near the outermost 244 part of the cuticle facing the cuticular pegs (see arrows Fig 2c). Besides, during 245 tomato fruit development, Young’s modulus of the CPM was modified although not 246 homogeneously 247 spatiotemporal heterogeneities of elastic modulus in the cutin polymer matrix 248 focusing on these specific areas. (Fig. 2c). Accordingly, we further assessed the specific 249 250 251 252 Spatiotemporal dynamics of the mechanical properties of CPM The CPM mechanical properties were transversally and longitudinally measured during fruit development to assess their spatiotemporal dynamic (Fig. 3). 253 The in-depth mapping of the elastic modulus was done from the surface of the 254 epidermal cells to the surface of the cuticle, on the periclinal part of the CPM to avoid 255 any biases due to thickness heterogeneities of cell wall and cuticle (Fig. 3a). Linear 256 regressions were performed (Fig. 3b and S1) and a progressive increase of the CPM 257 elastic modulus was evidenced during development, but with distinct profiles (Fig. 258 3c). At 15 DPA, the mechanical gradient showed three phases, with two apparent 259 inflection points. From 15 to 25 DPA, and from 30 to 35 DPA, two phases were 260 observed with distinct slopes whereas, at 40 DPA, a single linear increase of the 261 elastic modulus was observed. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 262 Likewise, we looked through to Young’s modulus profiles of the furrow within 263 the CPM at the different developmental stages (Fig. 3d). Linear regressions were 264 performed on the maximum, the minimum and on both sides of the slope (Fig. 3e and 265 S2) to compare mechanical properties (Fig. 3f) and size of the furrow over fruit 266 development (Fig. 3g). The size of the furrow was evaluated according to Young’s 267 modulus value. Two different zones were highlighted within this furrow, i.e., the 268 central soft area (expressed as ‘Thalweg’) and the overall furrow size (hereafter 269 named ‘Valley’) (Fig. 3e). A significant increase of Young’s modulus was measured in 270 both the Thalweg (Fig. 3f, min) and the ‘hard’ sides (Fig. 3f, max) of the furrow. 271 Indeed, between 15 and 40 DPA, the elastic modulus increased from 122 ± 8 to 605 272 ± 72 MPa for the central furrow and from 276 ± 3 to 1842 ± 77 MPa for the sides of 273 the furrow. Interestingly, during fruit expansion, i.e., from 15 to 25 DPA, the difference 274 between the elastic modulus of the ‘Thalweg’ and the Sides’ of the furrow were 275 almost constant and started to increase between 25 and 30 DPA, i.e., at the end of 276 rapid fruit expansion and onset of ripening (Guillet et al., 2002). The size of this 277 central furrow was monitored during tomato fruit development. The overall size of the 278 furrow (‘Valley’) was quite homogeneous (around 1.5 µm) during development. 279 Conversely, the central part of the furrow (‘Thalweg’) was around 0.38 µm for all 280 developmental stages except for the 25 DPA which showed a wider Thalweg of 0.63 281 µm (Fig. 3g). At this stage, the difference between Valley and Thalweg sizes was 282 lower than that at other developmental stages indicating a more abrupt transition 283 from high-to-low elastic modulus in the furrow. 284 Taken together, these peculiar spatiotemporal dynamics of the nanomechanical 285 properties of the cutin matrix during fruit development led us to look for possible 286 relationships with the chemical composition of these areas. 287 288 Chemical gradients and in-depth mechanical heterogeneities are related. 289 To investigate these in-depth mechanical heterogeneities both Raman and 290 Optical PhotoThermal InfraRed (OPTIR) spectroscopies were used. Cluster analyses 291 from the OPTIR maps mainly revealed a gradient in the cutin/polysaccharides ratio 292 from the cuticle surface toward the surface of epidermal cells (Fig. 4a and S3) 293 (Philippe et al., 2020). Indeed, as already indicated from macroscopic analyses 294 (isolated peels or cuticle), the amount of embedded polysaccharides has a significant bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 295 impact on the elastic modulus (López-Casado et al., 2007). In this regard, Raman 296 mapping provides a finer chemical clustering of the CPM, i.e., lipids, polysaccharides 297 and phenolic compounds (Fig. 4a and S3) (Reynoud et al., 2022). From these 298 Raman maps, the radial profile of specific bands of CPM components were obtained 299 (Fig. 4b and S4). Then, Pearson correlations were calculated between elastic 300 modulus and relative Raman intensity of these specific bands (Fig. 4b, bold values in 301 panels). Statistically significant correlations were obtained at every developmental 302 stage, although the correlation values were lower at 35 and 40 DPA as a result of the 303 complexification of CPM with increased numbers of chemical clusters (Reynoud et 304 al., 2022). Positive correlations were observed between the elastic modulus and the 305 Raman intensity of crystalline cellulose (e.g., 0.76 at 30 DPA) and pectins (e.g., 0.75 306 at 30 DPA) (Fig. 4b). The same conclusion could be drawn using a generic band for 307 polysaccharides (1375 cm-1), due to the HCC, HCO and COH bending deformations 308 (Chylińska et al., 2014). At the opposite, negative correlations were evidenced with 309 the Raman intensity of bands assigned to lipids (e.g., -0.70 at 20 DPA) and p- 310 coumaric acid (pCA) (e.g., -0.69 at 20 DPA) (Fig. 4b). Upon ripening, CPM 311 accumulates phenolics as evidenced by a generic band for phenolics (1170 cm-1) that 312 is assigned to δ(CH) of aromatic rings (Ś wisłocka et al., 2012), related to phenolic 313 acids and flavonoids. 314 Flavonoids are specifically accumulated in the cuticle during ripening (Laguna 315 et al., 1999; España et al., 2014b) including a fraction associated to CPM (Fig. 4b) 316 (Hunt & Baker, 1980; Domínguez et al., 2009). Interestingly, negative correlations of 317 the intensity of both flavonoid specific bands with elastic modulus were highlighted. 318 Accordingly, at the submicron scale, our results indicate that the cutin-associated 319 flavonoids do not account for the modification of Young’s modulus related to 320 flavonoid accumulation of the isolated cuticles (España et al., 2014a). Together, 321 these results indicate that the nanomechanical heterogeneities revealed in the CPM 322 are correlated to chemical gradients. 323 324 The central furrow displays a fine structural arrangement and 325 progressive chemical inhomogeneity 326 To decipher the specific mechanical properties of the central furrow and its 327 relationships with the chemical composition and structure of the CPM, a multimodal bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 328 approach was conducted combining Raman imaging and Infrared mapping at three 329 developmental stages, i.e., 20, 30 and 40 DPA. Two areas of the CPM were 330 compared, the central furrow and the sides of the furrow (Fig. 3e and S5). At 20 and 331 30 DPA, principal component analysis (PCA) from Raman data showed no strong 332 specific spectral fingerprints of the central furrow until the 40 DPA stage whereas it 333 was more progressive for OPTIR data (Fig. 5a). PCA loadings of OPTIR data (Fig. 334 S6a) evidenced that the central furrow was rather composed of cutin whereas 335 polysaccharides were progressively excluded to the sides. Loading of Raman data 336 (Fig. S6b) indicated a change of the relative accumulation of phenolic compounds in 337 the furrow sides along with crystalline cellulose whereas pectins concentrated in the 338 central furrow. These multimodal spectroscopic approaches indicated that the central 339 furrow adopts a slight but distinct chemical composition than the furrow sides along 340 tomato fruit development. 341 Then, the specific molecular orientation of CPM components was further 342 assessed by Raman spectroscopy using a linear polarized laser. By following 343 changes in band intensity due to different settings of polarization, parallel (0°) and 344 orthogonal (90°) to the cuticle surface, specific molecular orientations could be 345 inferred (Fig. 5b). We focused on lipid bands, as the central furrow was mostly 346 composed of cutin. Two bands related to CH2 modes of vibrations: δ(CH2, CH3) and 347 τ(CH2), and one related to C-C stretching vibrations (Czamara et al., 2015) were 348 found to changes according to polarization (Fig. 5b). A higher intensity with the 0° 349 polarization direction suggested a preferential orientation of lipids parallel to the 350 cuticle surface for both 20 and 40 DPA stages (Fig. 5b and S7). Differences between 351 the central furrow and its sides were observed at both stages (Fig. 5b and S7) 352 illustrating a distinct macromolecular arrangement of lipids of these areas. 353 To further gain insight into the specific chemical composition and structure of 354 the central furrow, the susceptibility to cutinase was checked. AFM experiments were 355 conducted on this furrow area enabling AFM topography. Upon cutinase treatment, 356 the central furrow was sharply carved out compared to the furrow sides at 20 DPA 357 and 35 DPA as well (Fig. 6) This acute difference of cutinase susceptibility within the 358 central furrow could be related to higher accessibility of the enzyme, in agreement 359 with the difference in macromolecular arrangement highlighted by Raman mapping 360 (Fig. 5b). Moreover, cutinase is more active towards primary than secondary ester 361 bounds (Lin & Kolattukudy, 1978). Accordingly, the significant impact of cutinase on bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 362 the furrow topography could be due to a lower reticulation degree of the cutin 363 polyester (i.e., with less secondary ester bonds) in the furrow than on the sides of the 364 furrow. 365 Together our results showed variations in nanomechanical properties in the 366 central furrow that are associated not only with chemical modifications but also with 367 macromolecular arrangements within the CPM. 368 369 Discussion 370 Heterogeneity of local mechanical properties of the cutin polymer matrix: 371 a way to address challenges during fruit growth? 372 Elucidation of relationships between the architecture of the cutin- 373 polysaccharide assemblies and the functional properties of the cuticle is critical to 374 understand and modulate plant cuticle functionalities. All along fruit development, the 375 mechanical properties of the cuticle must adapt to the trade-off between fruit 376 expansion and maintenance of cuticle integrity (Knoche & Lang, 2017). 377 In the present study, in depth nanomechanical mapping of the cutin polymer 378 matrix evidenced unprecedented heterogeneities in the cuticle, including “soft” 379 (exhibiting elastic modulus around 300-500 MPa) and “hard” (modulus around 1.5-2.5 380 GPa) domains. Such spatial nanomechanical heterogeneity, maintained throughout 381 the development, was not observed at the surface of isolated cuticle (Fig. 1) (Round 382 et al., 2000; Isaacson et al., 2009). Most strikingly, a “central furrow” displaying a 383 lower elastic modulus was evidenced in the CPM between the anticlinal walls of 384 adjacent epidermal cells (Fig. 2). The biological function of this peculiar ‘soft’ central 385 furrow maintained during the fruit expansion phase has to be addressed. The lower 386 Young’s modulus suggests a higher mechanical compliance (Rusin & Kojs, 2011). A 387 similar situation is observed with some Algae where specialized soft and extensible 388 tissue (geniculum) are inserted between stiff and rigid calcified cell wall 389 (intergeniculum) (Denny and King 2016). In addition, this central region of plant 390 cuticle is currently described as a place for the deposition of cuticle material to form 391 the so-called ‘cuticular pegs’ (Deas & Holloway, 1977). From an anatomical point of 392 view, the ‘central furrow’ could be compared to the ‘middle lamella’ in-between two 393 primary cell walls. In accordance with this hypothesis, this area is enriched in pectins bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 394 and poor in cellulose (Fig. 5a and S6) as observed in the ‘middle lamella’ (Jarvis et 395 al., 2003). In this regard, middle lamella in plant cell wall acts not only in cellular 396 adhesion but also plays a role in the mechanical resistance under tension or 397 compression (Zamil & Geitmann, 2017). Previous studies suggest that the cuticle 398 above the anticlinal peg concentrate the mechanical stress (Knoche & Lang, 2017). 399 Accordingly, the presence of a ‘soft’ area might play a key role in the biomechanical 400 properties of the epidermis by “absorbing” mechanical stress, while allowing tissue to 401 extend. In line with this hypothesis, the size of the central furrow (‘Thalweg’) is 402 maximum at 25 DPA corresponding to the maxima of fruit expansion rate (Guillet et 403 al., 2002). Interestingly, at a macroscopic scale, tensile test of isolated cuticle 404 highlighted a higher extensibility at this development stage of the tomato fruit 405 (España et al., 2014b; Benítez et al., 2021). These mechanical features are probably 406 associated to biochemical modification within the CPM. Indeed, the 25 DPA stage 407 was also highlighted as a turning point in macromolecular rearrangement of 408 polysaccharides embedded in the cuticle (Reynoud et al., 2022). Furthermore, our 409 data indicate that the lipid polyester of this central ‘soft’ area has also peculiar 410 structural features including a specific lipid orientation revealed by Raman mapping 411 and higher sensitivity to cutinase hydrolyses (Fig. 5 and 6), suggesting a specific 412 macromolecular organization and a lower reticulation degree of the polyester. 413 Interestingly, in cutin-inspired polyesters, the lower polymerization index was also 414 associated with a lower elastic modulus and higher extensibility (Marc et al., 2021). 415 Altogether, our study highlighted a specific design of tomato cuticles architectures 416 leading to alternating soft and hard interfaces. Plant cuticle is therefore another 417 example of biological material featured by site specific mechanical properties to 418 address environmental or developmental constraints (Liu et al., 2017). 419 420 The cutin-polysaccharide continuum: toward the design of a structural 421 gradient during fruit development 422 The cutin-polysaccharide continuum plays a pivotal role in the mechanical 423 properties of the plant cuticle. Indeed, from tensile tests of either fruit skin or isolated 424 cuticles, it was suggested that the extent of cell-wall cutinization, is responsible for 425 the higher mechanical stiffness in a crack resistant tomato cultivar (Matas et al., 426 2004a). Likewise, the proportion of polysaccharides within the cuticle was positively bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 427 correlated to the elastic modulus (López-Casado et al., 2007; Takahashi et al., 2012), 428 although the contribution of each component to cuticle mechanical properties is 429 difficult to determine. 430 In the present study, correlated multimodal imaging enabled new insight in the 431 cutin-polysaccharide continuum and demonstrated a fine tuning of its 432 nanomechanical properties during fruit development. From fruit expansion to fruit 433 maturation phases, an overall increase of the elastic modulus (up to 6 fold) was 434 observed. Furthermore combining the hyperspectral and nanomechanical mapping, 435 the impact of the cutin-polysaccharide ratio on the mechanical properties of CPM 436 agrees with the strain-hardening behavior provided by tensile tests measurements of 437 isolated cuticles (Matas et al., 2004a; Bargel & Neinhuis, 2005; Benítez et al., 2021). 438 More precisely, our study highlighted three types of interface, i.e., transition zones, in 439 the cutin-polysaccharide continuum (Fig. 3c and S1) that coincides with sharply 440 contrasted phases of tomato fruit development, i.e., the fruit expansion (10-25 DPA) 441 and the ripening process (30-35 DPA) leading to the red ripe stage (40 DPA) (Guillet 442 et al., 2002; Renaudin et al., 2017). From 10 DPA to 30 DPA, in-depth elastic 443 modulus mapping highlighted a sharp interfacial zone between the cutin-rich and the 444 polysaccharides-rich areas. During maturation, a gradual broadening of this 445 interfacial zone turned into a continuous gradient transition (Fig. 3c). In composites, 446 structural transitional mechanical gradients are currently considered as a way to 447 accommodate mechanical properties mismatches (e.g. elastic modulus) by a smooth 448 transition and to provide stress relief at the interface between dissimilar materials 449 (i.e., lipid polyester and cell wall polysaccharides) (Liu et al., 2017). Such structural 450 gradients result in an increased toughness and have been observed in many 451 biological materials such as the dentin-enamel junction in mammalian teeth (Naleway 452 et al., 2015), the tendon-ligament (Lu & Thomopoulos, 2013), collagen/elastin of skin 453 (Labroo et al., 2021) and chitin/protein in squid beaks (Miserez et al., 2008). 454 Such mechanical gradients are mainly driven by changes in local chemical 455 composition or arrangements of the building blocks (Li et al., 2021). In this regard, 456 the stiffening of the CPM could be related to the higher cross-linking observed in 457 mature tomato cutin polyester (Philippe et al., 2016; Chatterjee et al., 2016) including 458 the contribution of pCA (Reynoud et al., 2021). In addition, pCA might participate in 459 the stiffening of cuticle through cross-linking by peroxidase action. Indeed, the 460 application of exogenous peroxidase on isolated cuticle results in mechanical bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 461 stiffening regardless of the developmental stages (Andrews et al., 2002). In cuticle, 462 phenolic acids are present from early developmental stages (González Moreno et al., 463 2022; Reynoud et al., 2022) and are known targets of peroxidase. The peroxidase 464 driven stiffening of the CPM is therefore possible, as peroxidase activity considerably 465 increases in the epidermis during fruit ripening along with the phenolic burst 466 (Thompson et al., 1998). From the isolated cuticle tensile tests, flavonoids 467 accumulation during the fruit maturation have been associated to the increase of 468 elastic modulus (Domínguez et al., 2009; Benítez et al., 2021). However, in-depth 469 imaging of the CPM during tomato fruit development showed that the spatial 470 accumulation of flavonoids at the maturation stage did not correlate with a high 471 elastic modulus. In cuticle, a fraction of the flavonoids is readily extracted with waxes 472 while another one is tightly embedded in the CPM (Hunt & Baker, 1980; Luque et al., 473 1995). Accordingly, our results suggest that stiffening impacts of flavonoids should 474 rather be attributed to the solvent soluble fraction than to the CPM-associated 475 fraction. 476 Besides, our recent data showed that polysaccharides embedded in the cutin 477 matrix are subjected to macromolecular modifications during tomato fruit 478 development (Reynoud et al., 2022). Indeed, the pectin to cellulose ratio within the 479 tomato cuticles varies spatially and temporally, while the cellulose crystallinity and 480 hemicellulose remodeling increase during fruit development. The contribution of the 481 embedded highly esterified pectins to the mechanical properties is still unclear 482 (Bidhendi & Geitmann, 2016) while the modification of the crystalline cellulose 483 distribution will likely affect the in-depth mechanical properties. Likewise, pectin 484 deposition appears essential for the assembly of cellulose microfibrils as abnormal 485 cellulose organization was observed for an Arabidopsis thaliana mutant impaired in 486 pectin biosynthesis (Du et al., 2020). Thus, the tight pectin-cellulose interactions 487 observed in the CPM (Reynoud et al., 2022) and in the cell wall (Wang et al., 2015) 488 should significantly impact the CPM mechanical properties. Accordingly, the higher 489 structural order of both pectin and cellulose during fruit maturation might contribute to 490 the progressive stiffening of the CPM. In the same extent, hemicelluloses were 491 largely remodeled in the CPM during development (Reynoud et al., 2022). This 492 should modify their binding to cellulose (Grantham et al., 2017; Jaafar et al., 2019) 493 which affects microfibrils organization (Cosgrove, 2022), hence supporting a 494 modification of the mechanical properties of CPM. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 495 Finally, specific molecular orientations were observed in the CPM area with 496 distinct mechanical properties (Fig. 5a). Such correlation could be inferred by analogy 497 to the contribution of microfibrils orientations of cellulose in the load-bearing capacity 498 of the cell wall (Cosgrove, 2022). It should be noted that in the CPM, specific 499 orientations of cellulose microfibrils were observed that were modified during tomato 500 fruit development (Reynoud et al., 2022). Due to periclinal expansion, progressive 501 microfibrils reorientation and straightening (Renaudin et al., 2017; Zhang et al., 2021) 502 would results in higher elastic modulus. Indeed, macromolecular straightening is 503 monitored through the persistence length, i.e., the distance over which molecular 504 chain is aligned to main tangential axis (Flory & Volkenstein, 1969). An increase in 505 the persistence length results in a more stiff rod (with a higher elastic modulus), than 506 a more coiled or bend structure (Usov et al., 2015; Zdunek et al., 2021). The same 507 conclusion could be drawn for pectins which harbored an ordered structures and 508 specific orientation in the CPM during fruit maturation (Reynoud et al., 2022). These 509 macromolecular modifications could account for the increase in elastic modulus over 510 fruit development. 511 Altogether, the locally heterogeneous mechanical properties of CPM are finely 512 tuned during the development of tomato fruit and are related to different local 513 variations of chemical compositions, macromolecular arrangements and distribution. 514 Such multiplicity of gradients will likely provide an architectural basis to fit the CPM 515 mechanical properties with the mechanical stress imposed during the different 516 phases of fruit expansion and ripening. This fine tuning of cuticle architecture and its 517 mechanical properties provide new insight for plant breeding as well as the design of 518 bioinspired functional material. 519 Acknowledgements 520 521 Raman and AFM were performed at the PROBE research infrastructure, Biopolymers 522 Interactions, Structural Biology (BIBS) facility, Nantes. OPTIR was performed at the 523 Institut de Chimie Physique. NR was supported by Ph.D. fellowship (SeaSCAPE) 524 granted by INRAE and the Region Pays de la Loire. This work was also supported by 525 the ANR (Agence National de la Recherche) grant COPLAnAR (ANR-21-CE11- 526 0035). 527 bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 528 Competing interest 529 Authors declare no conflict of interest. 530 531 Author contributions 532 CR, ML, DM and BB designed the research. NR, NG, JP, JM and ADO performed 533 experiments. NR, NG, ADO, JM, ADB, ML, DM, CR and BB analyzed the data. NR, 534 ML, DM, CR and BB wrote the paper. 535 536 ORCID 537 538 Nicolas Reynoud https://orcid.org/0000-0002-3447-9406 539 Nathalie Geneix https://orcid.org/0000-0003-0594-0088 540 Angélina D’Orlando https://orcid.org/0000-0002-8118-3819 541 Johann Petit https://orcid.org/0000-0002-6746-1755 542 Jérémie Mathurin https://orcid.org/0000-0002-6769-6394 543 Ariane Deniset-Besseau https://orcid.org/0000-0002-1923-4988 544 Didier Marion https://orcid.org/0000-0002-0672-6545 545 Christophe Rothan https://orcid.org/0000-0002-6831-2823 546 Marc Lahaye https://orcid.org/0000-0002-7752-4158 547 Bénédicte Bakan https://orcid.org/0000-0001-9088-1232 548 549 References 550 Alhagdow M, Mounet F, Gilbert L, Nunes-Nesi A, Garcia V, Just D, Petit J, 551 Beauvoit B, Fernie AR, Rothan C, et al. 2007. Silencing of the mitochondrial 552 ascorbate synthesizing enzyme L-galactono-1,4-lactone dehydrogenase affects plant 553 and fruit development in tomato. Plant Physiology 145: 1408–1422. 554 Andrews J, Adams SR, Burton KS, Edmondson RN. 2002. Partial purification of 555 tomato fruit peroxidase and its effect on the mechanical properties of tomato fruit 556 skin. Journal of Experimental Botany 53: 2393–2399. 557 Bargel H, Neinhuis C. 2004. Altered tomato (Lycopersicon esculentum Mill.) fruit 558 cuticle biomechanics of a pleiotropic non ripening mutant. Journal of Plant Growth bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 559 Regulation 23: 61–75. 560 Bargel H, Neinhuis C. 2005. Tomato (Lycopersicon esculentum Mill.) fruit growth 561 and ripening as related to the biomechanical properties of fruit skin and isolated 562 cuticle. Journal of Experimental Botany 56: 1049–1060. 563 Benítez JJ, Guzmán-Puyol S, Vilaplana F, Heredia-Guerrero JA, Domínguez E, 564 Heredia A. 2021. Mechanical Performances of Isolated Cuticles Along Tomato Fruit 565 Growth and Ripening. Frontiers in Plant Science 12: 1–15. 566 Bhanot V, Fadanavis SV, Panwar J. 2021. Revisiting the architecture, biosynthesis 567 and functional aspects of the plant cuticle: There is more scope. Environmental and 568 Experimental Botany 183: 104364. 569 Bidhendi AJ, Geitmann A. 2016. Relating the mechanics of the primary plant cell 570 wall to morphogenesis. Journal of Experimental Botany 67: 449–461. 571 Busta L, Jetter R. 2018. Moving beyond the ubiquitous: the diversity and 572 biosynthesis of specialty compounds in plant cuticular waxes. Phytochemistry 573 Reviews 17: 1275–1304. 574 Chatterjee S, Matas AJ, Isaacson T, Kehlet C, Rose JKC, Stark RE. 2016. Solid- 575 State 13 C NMR Delineates the Architectural Design of Biopolymers in Native and 576 Genetically Altered Tomato Fruit Cuticles. Biomacromolecules 17: 215–224. 577 Chylińska M, Szymańska-Chargot M, Zdunek A. 2014. Imaging of polysaccharides 578 in the tomato cell wall with Raman microspectroscopy. Plant Methods 10: 14. 579 Cosgrove DJ. 2022. Building an Extensible Cell Wall. Founders Review for Plant 580 Physiology: 1–32. 581 Czamara K, Majzner K, Pacia MZ, Kochan K, Kaczor A, Baranska M. 2015. 582 Raman spectroscopy of lipids: A review. Journal of Raman Spectroscopy 46: 4–20. 583 Deas AHB, Holloway PJ. 1977. The Intermolecular Structure of Some Plant Cutins. 584 Lipids and Lipid Polymers in Higher Plants: 293–299. 585 Derjaguin B., Muller V., Toporov Y. 1975. Effect of contact deformations on the 586 adhesion of particles. Journal of Colloid and Interface Science 53: 314–326. 587 Domínguez E, España L, López-Casado G, Cuartero J, Heredia A. 2009. 588 Biomechanics of isolated tomato (Solanum lycopersicum) fruit cuticles during 589 ripening: the role of flavonoids. Functional Plant Biology 36: 613. 590 Domínguez E, López-Casado G, Cuartero J, Heredia A. 2008. Development of 591 fruit cuticle in cherry tomato (Solanum lycopersicum). Functional Plant Biology 35: 592 403–411. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 593 Du J, Kirui A, Huang S, Wang L, Barnes WJ, Kiemle S, Zheng Y, Rui Y, Ruan M, 594 Qi S, et al. 2020. Mutations in the Pectin Methyltransferase QUASIMODO2 Influence 595 Cellulose Biosynthesis and Wall Integrity in Arabidopsis thaliana. The Plant Cell 32: 596 tpc.00252.2020. 597 Dufrêne YF, Ando T, Garcia R, Alsteens D, Martinez-Martin D, Engel A, Gerber 598 C, Müller DJ. 2017. Imaging modes of atomic force microscopy for application in 599 molecular and cell biology. Nature Nanotechnology 12: 295–307. 600 España L, Heredia-Guerrero JA, Reina-Pinto JJ, Fernández-Muñoz R, Heredia 601 A, Domínguez E. 2014a. Transient silencing of CHALCONE SYNTHASE during fruit 602 ripening modifies tomato epidermal cells and cuticle properties. Plant Physiology 166: 603 1371–1386. 604 España L, Heredia-Guerrero JA, Segado P, Benítez JJ, Heredia A, Domínguez 605 E. 2014b. Biomechanical properties of the tomato (Solanum lycopersicum) fruit 606 cuticle during development are modulated by changes in the relative amounts of its 607 components. New Phytologist 202: 790–802. 608 Fernández V, Gil-Pelegrín E, Eichert T. 2021. Foliar water and solute absorption: 609 an update. Plant Journal 105: 870–883. 610 Fernández V, Guzmán-Delgado P, Graça J, Santos S, Gil L. 2016. Cuticle 611 structure in relation to chemical composition: Re-assessing the prevailing model. 612 Frontiers in Plant Science 7: 1–14. 613 Flory PJ, Volkenstein M. 1969. Statistical mechanics of chain molecules. 614 Biopolymers 8: 699–700. 615 González Moreno A, Domínguez E, Mayer K, Xiao N, Bock P, Heredia A, 616 Gierlinger N. 2022. 3D (x-y-t) Raman imaging of tomato fruit cuticle: Microchemistry 617 during development. Plant Physiology. 618 Graça J, Lamosa P. 2010. Linear and branched poly(ω-hydroxyacid) esters in plant 619 cutins. Journal of Agricultural and Food Chemistry 58: 9666–9674. 620 Grantham NJ, Wurman-Rodrich J, Terrett OM, Lyczakowski JJ, Stott K, Iuga D, 621 Simmons TJ, Durand-Tardif M, Brown SP, Dupree R, et al. 2017. An even pattern 622 of xylan substitution is critical for interaction with cellulose in plant cell walls. Nature 623 Plants 3: 859–865. 624 Guillet C, Just D, Bénard N, Destrac-Irvine A, Baldet P, Hernould M, Causse M, 625 Raymond P, Rothan C. 2002. A fruit-specific phosphoenolpyruvate carboxylase is 626 related to rapid growth of tomato fruit. Planta 214: 717–726. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 627 He Q, Zabotina OA, Yu C. 2020. Principal component analysis facilitated fast and 628 noninvasive Raman spectroscopic imaging of plant cell wall pectin distribution and 629 interaction with enzymatic hydrolysis. Journal of Raman Spectroscopy: 1–10. 630 Hertz H. 1882. Ueber die Berührung fester elastischer Körper. crll 1882: 156–171. 631 Hunneman DH, Eglinton G. 1972. The constituent acids of gymnosperm cutins. 632 Phytochemistry 11: 1989–2001. 633 Hunt G, Baker D. 1980. Phenolic constituents of tomato fruit cuticles. 634 Phytochemistry 19: 1415–1419. 635 Ingram G, Nawrath C. 2017. The roles of the cuticle in plant development: organ 636 adhesions and beyond. Journal of experimental botany 68: 5307–5321. 637 Isaacson T, Kosma DK, Matas AJ, Buda GJ, He Y, Yu B, Pravitasari A, Batteas 638 JD, Stark RE, Jenks MA, et al. 2009. Cutin deficiency in the tomato fruit cuticle 639 consistently affects resistance to microbial infection and biomechanical properties, 640 but not transpirational water loss. Plant Journal 60: 363–377. 641 Jaafar Z, Mazeau K, Boissière A, Le Gall S, Villares A, Vigouroux J, Beury N, 642 Moreau C, Lahaye M, Cathala B. 2019. Meaning of xylan acetylation on xylan- 643 cellulose interactions: A quartz crystal microbalance with dissipation (QCM-D) and 644 molecular dynamic study. Carbohydrate Polymers 226: 115315. 645 Jarvis MC, Briggs SPH, Knox JP. 2003. Intercellular adhesion and cell separation 646 in plants. 44: 977–989. 647 Jiang F, Lopez A, Jeon S, de Freitas ST, Yu Q, Wu Z, Labavitch JM, Tian S, 648 Powell ALT, Mitcham E. 2019. Disassembly of the fruit cell wall by the ripening- 649 associated polygalacturonase and expansin influences tomato cracking. Horticulture 650 Research 6. 651 Jiao C, Sørensen I, Sun X, Sun H, Behar H, Alseekh S, Philippe G, Palacio 652 Lopez K, Sun L, Reed R, et al. 2020. The Penium margaritaceum Genome: 653 Hallmarks of the Origins of Land Plants. Cell 181: 1097-1111.e12. 654 Kerr LT, Hennelly BM. 2016. A multivariate statistical investigation of background 655 subtraction algorithms for Raman spectra of cytology samples recorded on glass 656 slides. Chemometrics and Intelligent Laboratory Systems 158: 61–68. 657 Khanal BP, Grimm E, Finger S, Blume A, Knoche M. 2013. Intracuticular wax fixes 658 and restricts strain in leaf and fruit cuticles. New Phytologist 200: 134–143. 659 Khanal BP, Knoche M. 2014. Mechanical properties of apple skin are determined by 660 epidermis and hypodermis. Journal of the American Society for Horticultural Science bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 661 139: 139–147. 662 Knoche M, Lang A. 2017. Ongoing Growth Challenges Fruit Skin Integrity. Critical 663 Reviews in Plant Sciences 36: 190–215. 664 Labroo P, Irvin J, Johnson J, Sieverts M, Miess J, Robinson I, Baetz N, Garrett 665 C, Sopko N. 2021. Physical characterization of swine and human skin: Correlations 666 between Raman spectroscopy, Tensile testing, Atomic force microscopy (AFM), 667 Scanning electron microscopy (SEM), and Multiphoton microscopy (MPM). Skin 668 Research and Technology 27: 501–510. 669 Laguna L, Casado CG, Heredia A. 1999. Flavonoid biosynthesis in tomato fruit 670 cuticles after in vivo incorporation of 3H-phenylalanine precursor. Physiologia 671 Plantarum 105: 491–498. 672 Lê S, Josse J, Rennes A, Husson F, Rennes A, Husson F. 2008. FactoMineR: An 673 R package for multivariate analysis. Journal of Statistical Software 25: 1–18. 674 Lee SB, Suh MC. 2022. Regulatory mechanisms underlying cuticular wax 675 biosynthesis. Journal of Experimental Botany 73: 2799–2816. 676 Li C, Ouyang L, Armstrong JPK, Stevens MM. 2021. Advances in the Fabrication 677 of Biomaterials for Gradient Tissue Engineering. Trends in Biotechnology 39: 150– 678 164. 679 Lin TS, Kolattukudy PE. 1978. Induction of a Biopolyester Hydrolase (Cutinase) by 680 Low Levels of Cutin Monomers in Fusarium solani f. sp. pisi†. Journal of Bacteriology 681 133: 942–951. 682 Liu Z, Meyers MA, Zhang Z, Ritchie RO. 2017. Functional gradients and 683 heterogeneities in biological materials: Design principles, functions, and bioinspired 684 applications. Progress in Materials Science 88: 467–498. 685 López-Casado G, Matas AJ, Domínguez E, Cuartero J, Heredia A. 2007. 686 Biomechanics of isolated tomato (Solanum lycopersicum L.) fruit cuticles: The role of 687 the cutin matrix and polysaccharides. Journal of Experimental Botany 58: 3875– 688 3883. 689 Lopez-Casado G, Salamanca A, Heredia A. 2010. Viscoelastic nature of isolated 690 tomato (Solanum lycopersicum) fruit cuticles: a mathematical model. Physiologia 691 Plantarum 140: 79–88. 692 Lu HH, Thomopoulos S. 2013. Functional attachment of soft tissues to bone: 693 Development, healing, and tissue engineering. Annual Review of Biomedical 694 Engineering 15: 201–226. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 695 Luque P, Bruque S, Heredia A. 1995. Water permeability of isolated cuticular 696 membranes: A structural analysis. Archives of Biochemistry and Biophysics 317: 697 417–422. 698 Marc M, Risani R, Desnoes E, Falourd X, Pontoire B, Rodrigues R, Escórcio R, 699 Batista AP, Valentin R, Gontard N, et al. 2021. Bioinspired co-polyesters of 700 hydroxy-fatty acids extracted from tomato peel agro-wastes and glycerol with tunable 701 mechanical, thermal and barrier properties. Industrial Crops and Products 170: 702 113718. 703 Martin LBB, Rose JKC. 2014. There’s more than one way to skin a fruit: Formation 704 and functions of fruit cuticles. Journal of Experimental Botany 65: 4639–4651. 705 Matas AJ, Cobb ED, Bartsch JA, Paolillo DJ, Niklas KJ. 2004a. Biomechanics and 706 anatomy of Lycopersicon esculentum fruit peels and enzyme-treated samples. 707 American Journal of Botany 91: 352–360. 708 Matas AJ, Cobb ED, Paolillo DJ, Niklas KJ. 2004b. Crack resistance in cherry 709 tomato fruit correlates with cuticular membrane thickness. HortScience 39: 1354– 710 1358. 711 Melelli A, Arnould O, Beaugrand J, Bourmaud A. 2020. The middle lamella of 712 plant fibers used as composite reinforcement: Investigation by atomic force 713 microscopy. Molecules 25. 714 Miserez A, Schneberk T, Sun C, Zok FW, Waite JH. 2008. The transition from stiff 715 to compliant materials in squid beaks. Science 319: 1816–1819. 716 Naleway SE, Porter MM, Mckittrick J, Meyers MA. 2015. Structural Design 717 Elements in Biological Materials: Application to Bioinspiration. : 5455–5476. 718 Nečas D, Klapetek P. 2012. Gwyddion: An open-source software for SPM data 719 analysis. Central European Journal of Physics 10: 181–188. 720 Niklas KJ, Cobb ED, Matas AJ. 2017. The evolution of hydrophobic cell wall 721 biopolymers: from algae to angiosperms. Journal of Experimental Botany 68: 5261– 722 5269. 723 Petit J, Bres C, Just D, Garcia V, Mauxion JP, Marion D, Bakan B, Joubès J, 724 Domergue F, Rothan C. 2014. Analyses of tomato fruit brightness mutants uncover 725 both cutin-deficient and cutin-abundant mutants and a new hypomorphic allele of 726 GDSL Lipase. Plant Physiology 164: 888–906. 727 Petracek PD, Bukovac MJ. 1995. Rheological properties of enzymatically isolated 728 tomato fruit cuticle. Plant Physiology 109: 675–679. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 729 Philippe G, Gaillard C, Petit J, Geneix N, Dalgalarrondo M, Bres C, Mauxion JP, 730 Franke R, Rothan C, Schreiber L, et al. 2016. Ester cross-link profiling of the cutin 731 polymer of wild-type and cutin synthase tomato mutants highlights different 732 mechanisms of polymerization. Plant Physiology 170: 807–820. 733 Philippe G, Geneix N, Petit J, Guillon F, Sandt C, Rothan C, Lahaye M, Marion 734 D, Bakan B. 2020. Assembly of tomato fruit cuticles: a cross-talk between the cutin 735 polyester and cell wall polysaccharides. New Phytologist 226: 809–822. 736 Renaudin JP, Deluche C, Cheniclet C, Chevalier C, Frangne N. 2017. Cell layer- 737 specific patterns of cell division and cell expansion during fruit set and fruit growth in 738 tomato pericarp. Journal of Experimental Botany 68: 1613–1623. 739 Renault H, Alber A, Horst NA, Basilio Lopes A, Fich EA, Kriegshauser L, 740 Wiedemann G, Ullmann P, Herrgott L, Erhardt M, et al. 2017. A phenol-enriched 741 cuticle is ancestral to lignin evolution in land plants. Nature Communications 8. 742 Reynoud N, Geneix N, Petit J, D’Orlando A, Fanuel M, Marion D, Rothan C, 743 Lahaye M, Bakan B. 2022. The cutin polymer matrix undergoes a fine architectural 744 tuning from early tomato fruit development to ripening. Plant Physiology: 1–20. 745 Reynoud N, Petit J, Bres C, Lahaye M, Rothan C, Marion D, Bakan B. 2021. The 746 Complex Architecture of Plant Cuticles and Its Relation to Multiple Biological 747 Functions. Frontiers in Plant Science 12: 1–7. 748 Riley RG, Kolattukudy PE. 1975. Evidence for Covalently Attached p -Coumaric 749 Acid and Ferulic Acid in Cutins and Suberins . Plant Physiology 56: 650–654. 750 Round AN, Van B, Dang S, Estephan R, Stark RE, Batteas JD. 2000. The 751 influence of water on the nanomechanical behavior of the plant biopolyester cutin as 752 studied by AFM and solid-state NMR. Biophysical Journal 79: 2761–2767. 753 Rusin T, Kojs P. 2011. Plant Biomechanics. In: Encyclopedia of Earth Sciences 754 Series. 602–604. 755 Sader JE, Larson I, Mulvaney P, White LR. 1995. Method for the calibration of 756 atomic force microscope cantilevers. Review of Scientific Instruments 66: 3789– 757 3798. 758 Segado P, Domínguez E, Heredia A. 2016. Ultrastructure of the epidermal cell wall 759 and cuticle of tomato fruit (Solanum lycopersicum L.) during development. Plant 760 Physiology 170: 935–946. 761 Sieber P, Schorderet M, Ryser U, Buchala A, Kolattukudy P, Métraux JP, 762 Nawrath C. 2000. Transgenic Arabidopsis plants expressing a fungal cutinase show bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 763 alterations in the structure and properties of the cuticle and postgenital organ fusions. 764 Plant Cell 12: 721–737. 765 Stoica I, Epure EL, Constantin CP, Damaceanu MD, Ursu EL, Mihaila I, Sava I. 766 2021. Evaluation of local mechanical and chemical properties via afm as a tool for 767 understanding the formation mechanism of pulsed uv laser‐nanoinduced patterns on 768 azo‐naphthalene‐based polyimide films. Nanomaterials 11: 1–24. 769 Świsłocka 770 Spectroscopic (FT-IR, FT-Raman, 1H and 13C NMR) and theoretical studies of p- 771 coumaric acid and alkali metal p-coumarates. Spectroscopy 27: 35–48. 772 Takahashi Y, Tsubaki S, Sakamoto M, Watanabe S, Azuma JI. 2012. Growth- 773 dependent chemical and mechanical properties of cuticular membranes from leaves 774 of Sonneratia alba. Plant, Cell and Environment 35: 1201–1210. 775 Thompson DS, Davies WJ, Ho LC. 1998. Regulation of tomato fruit growth by 776 epidermal cell wall enzymes. Plant, Cell and Environment 21: 589–599. 777 Tsubaki S, Sugimura K, Teramoto Y, Yonemori K, Azuma J ichi. 2013. Cuticular 778 Membrane of Fuyu Persimmon Fruit Is Strengthened by Triterpenoid Nano-Fillers. 779 PLoS ONE 8: 1–13. 780 Usov I, Nyström G, Adamcik J, Handschin S, Schütz C, Fall A, Bergström L, 781 Mezzenga R. 2015. Understanding nanocellulose chirality and structure-properties 782 relationship at the single fibril level. Nature Communications 6. 783 Wang T, Park YB, Cosgrove DJ, Hong M. 2015. Cellulose-pectin spatial contacts 784 are Inherent to never-dried Arabidopsis primary cell walls: Evidence from solid-state 785 nuclear magnetic resonance. Plant Physiology 168: 871–884. 786 Wiedemann P, Neinhuis C. 1998. Biomechanics of isolated plant cuticles. Botanica 787 Acta 111: 28–34. 788 Xi X, Kim SH, Tittmann B. 2015. Atomic force microscopy based nanoindentation 789 study of onion abaxial epidermis walls in aqueous environment. Journal of Applied 790 Physics 117: 1–10. 791 Zamil MS, Geitmann A. 2017. The middle lamella - More than a glue. Physical 792 Biology 14. 793 Zdunek A, Kurenda A. 2013. Determination of the elastic properties of tomato fruit 794 cells with an atomic force microscope. Sensors (Switzerland) 13: 12175–12191. 795 Zdunek A, Pieczywek PM, Cybulska J. 2021. The primary, secondary, and 796 structures of higher levels of pectin polysaccharides. Comprehensive Reviews in R, Kowczyk-Sadowy M, Kalinowska M, Lewandowski W. 2012. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 797 Food Science and Food Safety 20: 1101–1117. 798 Zhang D, Li C, Zhang C, Slipchenko MN, Eakins G, Cheng JX. 2016. Depth- 799 resolved mid-infrared photothermal imaging of living cells and organisms with 800 submicrometer spatial resolution. Science Advances 2: 1–8. 801 Zhang Y, Yu J, Wang X, Durachko DM, Zhang S, Cosgrove DJ. 2021. Molecular 802 insights into the complex mechanics of plant epidermal cell walls. Science 372: 607– 803 711. 804 805 Figure legends 806 807 Fig. 1 Mechanical properties of the surface of isolated cuticles and at the edge of 808 cuticle cross-sections along tomato fruit development. (a) Illustration of elastic 809 modulus map superimposed onto the volumetric topography of isolated and non-fixed 810 tomato cuticle at 20 Days Post Anthesis (DPA) stage. Red square represents a 811 window of 7x7 µm2. (b) Illustration of elastic modulus map of a fixed cutin cross- 812 section at 20 DPA. The red polygon delimits an area of 0.5 µm-width on the entire 813 length of the cross section. Bar, 5 µm. (c) Comparison of mechanical properties on 814 the surface and at the edge of the cutin polymer matrix at different developmental 815 stages of tomato fruit. The letters above error bars indicate a significant difference 816 between developmental stages and sampled regions (P < 0.05, ANOVA and Tukey 817 HSD). 818 819 Fig. 2 Peak-Force Quantitative Nanoscale Mechanical (PF-QNM) mapping of the 820 cutin polymer matrix during tomato fruit development. Two maps per developmental 821 stage were acquired, representative images are shown. (a) Light visible images of 822 cuticles with red rectangles highlighting areas that were investigated with PF-QNM. 823 (b) Topography. (c) Young's modulus. DPA, Days Post Anthesis. Bars represent 10 824 µm in (a) and 5 µm in (b) and (c). White arrows indicate the division of the central 825 furrow at 30 and 35 DPA. 826 827 Fig. 3 Spatiotemporal dynamics in mechanical properties of region of the cutin 828 polymer matrix (CPM). (a) Illustration of in-depth mechanical profile sampling (red 829 lines) of the CPM at 20 days post anthesis (DPA) from the surface of epidermal cell bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 830 (EC) to the surface of the cuticle (OE). Cut, Cuticle; CW, cell wall. Each profile 831 consisted of 160 to 235 equally interspersed points, according to the in-depth cuticle 832 thickness, with each point being a mean of 8 measurements. Two maps per 833 developmental stage were acquired with 7 sampling lines per stage. (b) Illustration of 834 linear regressions (blue lines) performed on the decreasing phase of the average 835 elastic modulus extracted from the in-depths profiles of a representative map at 20 836 DPA. Data are expressed as mean (solid lines) ± SD (shadowed area). To visualize 837 the complete profiles, readers are referred to Fig. S1. (c) Linear regression curves of 838 the in-depth elastic modulus of the CPM, according to the tomato developmental 839 stage. For proper comparison, the distance from the cell surface to the cuticle surface 840 was normalized and expressed as percentage. (d) Illustration of mechanical profile 841 sampling (red lines) of the central furrow of the CPM at 20 DPA. Ten profiles per 842 developmental stage were sampled with each profile consisting of 120 equally 843 interspersed points over 3µm; each point being a mean of 8 measurements. (e) 844 Illustration of linear regressions (blue lines) performed on the average elastic 845 modulus extracted from the transverse profile of a representative map at 20 DPA. 846 Data are expressed as mean (solid lines) ± SD (shadowed area). Regressions were 847 calculated on the minimum and maximum plateau and on both sides of the 848 depression. Size of the furrow was estimated on the upper part of the curve, between 849 1st and the 4th bending points of the modulus curve (‘Valley’) and on the lower part of 850 the curve, between the 2nd and the 3rd bending points (‘Thalweg’). To visualize the 851 complete profiles, readers are referred to Fig. S2. (f) Dynamics of elastic modulus of 852 the central furrow and besides, over the developmental stage of tomato fruit. Values 853 are mean of two maps acquired per developmental stages and calculated on the 854 ‘Thalweg’ (min) and on the ‘Sides’ (max). Letters above error bars indicate significant 855 difference (ANOVA, P <0.05 and Tukey HSD). (g) Dynamics of the central furrow 856 size (‘Thalweg’, ‘Valley’) of the CPM over tomato fruit development. Values are mean 857 ± SD (n=2). Bars, 5µm. 858 859 Fig. 4 Chemical profile of the cutin polymer matrix (CPM) over the in-depth thickness. 860 (a) Illustrations of sampling zone in cross-section of tomato cuticle at 25 DPA for 861 AFM maps (left), OPTIR (middle) and Raman (right). For both OPTIR and Raman 862 data, the sampled zone is represented by a k-means clustering map with three 863 clusters for OPTIR data and five clusters for Raman. To see the mean spectra bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 864 associated to the different clusters, readers are referred to Fig. S3. Bars, 5µm. (b) 865 Chemical composition assessed by Raman spectroscopy and elastic mechanical 866 properties determined by PF-QNM from the cell surface to the cuticle surface of the 867 CPM over tomato fruit development. Chemical profiles of CPM components were 868 obtained on area-mean normalized data to avoid biases due to difference in cross- 869 section thickness (see Fig. 2) and expressed as Raman intensity with arbitrary units 870 (a.u.). For each component a specific band was selected according to a homemade 871 reference database (Reynoud et al., 2022): Cutin (1440 cm-1), Crystalline cellulose 872 (380 cm-1), Pectin (852 cm-1), Polysaccharides (1375 cm-1), p-Coumaric acid (1605 873 cm-1), Flavonoid 1 (547 cm-1), Flavonoid 2 (1550 cm-1) and Phenolics (1170 cm-1). 874 Bold values in charts represent Pearson correlations (*P value < 0.05; ** P value < 875 0.01; *** P value < 0.001) between elastic modulus and Raman intensity. See Fig. S4 876 for full spectra. 877 878 Fig. 5 Chemical composition and macromolecular arrangement of the central furrow 879 over the development of the tomato CPM. (a) Principal component analysis (PCA) of 880 Raman (up) and OPTIR (down) datasets of the central furrow compared to furrow 881 sides at 20, 30 and 40 days post anthesis (DPA). Sampling zones and PCA loadings 882 are found in supplemental information (Fig. S5 and S6). (b) Macromolecular 883 orientation of lipids within the central furrow compared to the furrow sides at 20 DPA. 884 The laser polarization direction was modified from parallel (0°) to orthogonal (90°) to 885 the cuticle surface. For each polarization, two maps were acquired at 20 and 40 DPA 886 (Fig S7). The light image (upper left) represents the area of the cuticle that were 887 sampled to assess the molecular orientations within the central furrow (pink) and the 888 furrow sides (blue) with the corresponding elastic modulus (lower left). Bar, 5µm. 889 Data are expressed as mean (solid lines) ± SD (shadowed area). For both central 890 furrow and furrow sides, mean Raman spectra for each polarization direction were 891 calculated (right). The parallel (solid lines) to orthogonal (dashed lines) ratio of 892 intensity was illustrated by vertical bars for three bands related to lipids: two bands 893 attributed to CH2 deformations, i.e., τ(CH2) at 1305 cm-1 and δ(CH2,CH3) at 1440 cm-1 894 and one band assigned to ν(C-C) vibration at 1065 cm-1. Insets represents the lined 895 up different intensity ratio (bars) for proper comparison between cuticle areas. 896 bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 897 Fig. 6 Susceptibility of the cutin polymer matrix (CPM) to cutinase hydrolysis. 898 Topography of the CPM either subjected to cutinase treatment or not at 20 (a) and 35 899 (b) days post anthesis (DPA). The 3D topographies (left) illustrate the carving effect 900 of the cutinase in the central furrow with the blue line representing the sampled 901 transverse profiles within the CPM. The corresponding topographic profiles (right) 902 were calculated as a mean of 10 profiles over 3 µm, each profile being a mean of 8 903 measurements. Data are expressed as mean (solid lines) ± SD (shadowed area). 904 Two maps per developmental stages were acquired. 905 906 Supporting Information 907 908 Fig. S1 In-depth mechanical properties of the cutin polymer matrix over tomato fruit 909 development. 910 911 Fig. S2 Mechanical properties of the central furrow of the cutin polymer matrix over 912 tomato fruit development. 913 914 Fig. S3 Cluster analysis of hyperspectral maps of the tomato cutin polymer matrix at 915 25 DPA stage. 916 917 Fig. S4 In-depth Raman profile of the cutin polymer matrix over tomato fruit 918 development. 919 920 Fig. S5 Hyperspectral profiling of the central furrow and the furrow sides of the cutin 921 polymer matrix at 20, 30 and 40 days post anthesis (DPA). 922 923 Fig. S6 Loadings of principal component analysis of OPTIR and Raman datasets at 924 40 days post anthesis (DPA). 925 926 Fig. S7 Macromolecular orientation of lipids within the central furrow compared to the 927 furrow sides at 40 DPA. (a) (b) 3.0 GPa bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 2.5 2.0 4.0 µm 1.5 1.0 0.0 µm y: 50 0 x: 5 µm µm 0.5 0.0 1.50 Youngs modulus (GPa) (c) 1.25 Edge Surface f 1.00 g e 0.75 0.50 b b b 0.25 0.00 a a 15 a 20 d b c 25 30 Stage (DPA) 35 40 Fig. 1 Mechanical properties of the surface of isolated cuticles and at the edge of cuticle cross-sections along tomato fruit development. (a) Illustration of elastic modulus map superimposed onto the volumetric topography of isolated and non-fixed tomato cuticle at 20 Days Post Anthesis (DPA) stage. Red square represents a window of 7x7 µm2. (b) Illustration of elastic modulus map of a fixed cutin cross-section at 20 DPA. The red polygon delimits an area of 0.5 µm-width on the entire length of the cross-section. Bar, 5 µm. (c) Comparison of mechanical properties on the surface and at the edge of the cutin polymer matrix at different developmental stages of tomato fruit. The letters above error bars indicate a significant difference between developmental stages and sampled regions (P < 0.05, ANOVA and Tukey HSD). 10 DPA 15 DPA 20 DPA 25 DPA 30 DPA 35 DPA 40 DPA (a) 1.50 µm 1.25 1.00 (b) 0.75 0.50 0.25 0.00 3.0 GPa 2.5 2.0 (c) 1.5 1.0 0.5 0.0 Fig. 2 Peak-Force Quantitative Nanoscale Mechanical (PF-QNM) mapping of the cutin polymer matrix during tomato fruit development. Two maps per developmental stage were acquired, representative images are shown. (a) Light visible images of cuticles with red rectangles highlighting areas that were investigated with PF-QNM. (b) Topography. (c) Young's modulus. DPA, Days Post Anthesis. Bars represent 10 µm in (a) and 5 µm in (b) and (c). White arrows indicate the division of the central furrow at 30 and 35 DPA. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (a) was not certified by peer review) is the author/funder, (which (d) who has granted bioRxiv a license to display the preprint in perpetuity. It is made OE available under aCC-BY-ND 4.0 International license. Cut CW EC (e) Linear regression Linear regression 3.0 2.5 2.0 1.5 1.0 EC CW Cuticle OE 0.5 Young's modulus (GPa) 0.0 0.6 0.4 Valley 0.2 Min Thalweg 0.0 0 2 4 6 8 10 1.0 0.5 2.0 2.5 (f) Regression line 3.0 15 DPA 2.5 20 DPA 2.0 30 DPA 25 DPA 35 DPA 40 DPA 1.5 1.0 0.5 (g) 20 40 60 80 100 Relative distance from cell surface to cuticle surface (%) Thalweg Valley g 2.0 2.0 1.5 e d 1.0 1.5 1.0 f 0.5 0.0 0 min max 2.5 Young's modulus (GPa) Yougn's modulus(GPa) 1.5 Distance (µm) Distance from cell towards surface (µm) (c) Sides Max Width (µm) Young's modulus (GPa) (b) c b b a a a 15 20 25 bc 0.5 c 0.0 0.0 30 Stage (DPA) 35 40 15 20 25 30 35 Stage (DPA) 40 bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint Fig. was 3 Spatiotemporal dynamics in mechanical properties of region the cutin polymer (which not certified by peer review) is the author/funder, who has granted bioRxiv a license to display of the preprint in perpetuity. It is made available under a CC-BY-ND 4.0 International license . matrix (CPM). (a) Illustration of in-depth mechanical profile sampling (red lines) of the CPM at 20 days post anthesis (DPA) from the surface of epidermal cell (EC) to the surface of the cuticle (OE). Cut, Cuticle; CW, cell wall. Each profile consisted of 160 to 235 equally interspersed points, according to the in-depth cuticle thickness, with each point being a mean of 8 measurements. Two maps per developmental stage were acquired with 7 sampling lines per stage. (b) Illustration of linear regressions (blue lines) performed on the decreasing phase of the average elastic modulus extracted from the in-depths profiles of a representative map at 20 DPA. Data are expressed as mean (solid lines) ± SD (shadowed area). To visualize the complete profiles, readers are referred to Fig. S1. (c) Linear regression curves of the in-depth elastic modulus of the CPM, according to the tomato developmental stage. For proper comparison, the distance from the cell surface to the cuticle surface was normalized and expressed as percentage. (d) Illustration of mechanical profile sampling (red lines) of the central furrow of the CPM at 20 DPA. Ten profiles per developmental stage were sampled with each profile consisting of 120 equally interspersed points over 3 µm; each point being a mean of 8 measurements. (e) Illustration of linear regressions (blue lines) performed on the average elastic modulus extracted from the transverse profile of a representative map at 20 DPA. Data are expressed as mean (solid lines) ± SD (shadowed area). Regressions were calculated on the minimum and maximum plateau and on both sides of the depression. Size of the furrow was estimated on the upper part of the curve, between 1st and the 4th bending points of the modulus curve (‘Valley’) and on the lower part of the curve, between the 2nd and the 3rd bending points (‘Thalweg’). To visualize the complete profiles, readers are referred to Fig. S2. (f) Dynamics of elastic modulus of the central furrow and besides, over the developmental stage of tomato fruit. Values are mean of two maps acquired per developmental stages and calculated on the ‘Thalweg’ (min) and on the ‘Sides’ (max). Letters above error bars indicate significant difference (ANOVA, P <0.05 and Tukey HSD). (g) Dynamics of the central furrow size (‘Thalweg’, ‘Valley’) of the CPM over tomato fruit development. Values are mean ± SD (n=2). Bars, 5 µm. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. Raman OPTIR AFM (a) Cut CW (b) p-Coumaric acid Flavonoid 1 Flavonoid 2 Phenolics Raman intensity (a.u.) Raman intensity (a.u.) Raman intensity (a.u.) Raman intensity (a.u.) Polysaccharides Raman intensity (a.u.) Pectin Raman intensity (a.u.) Cristalline cellulose Raman intensity (a.u.) Cutin Raman intensity (a.u.) Mechanical properties Module(GPa) 15 DPA 20 DPA 25 DPA 3.0 2.5 2.0 1.5 1.0 0.5 0.0 3.0 2.5 2.0 1.5 1.0 0.5 0.0 3.0 2.5 2.0 1.5 1.0 0.5 0.0 5.5 6.0 5.0 30 DPA 35 DPA 3.0 2.5 2.0 1.5 1.0 0.5 0.0 40 DPA 3.0 2.5 2.0 1.5 1.0 0.5 0.0 3.0 2.5 2.0 1.5 1.0 0.5 0.0 1.6 5.0 4.0 3.5 2.4 1.2 5.0 4.5 cor = 0.56*** cor = -0.70*** 3.0 2.0 cor = 0.69*** cor = -0.42*** 3.0 3.0 1.5 0.8 cor = -0.42*** 2.75 cor = 0.52*** cor = -0.28*** cor = 0.76*** 2.5 cor = -0.34*** 2.0 cor = 0.31*** 2.0 2.0 2.25 cor = 0.32*** 2.00 1.5 2.25 2.0 1.75 1.5 2.00 1.0 1.0 1.8 3.0 2.5 2.50 1.50 1.75 1.25 1.0 1.6 cor = 0.43*** cor = 0.73*** 1.0 0.9 cor = 0.75*** 0.6 0.2 0.5 cor = 0.54*** 2.5 1.9 cor = 0.35*** cor = 0.34*** 3.5 2.0 2.5 1.5 1.3 2.8 2.4 2.2 2.0 0.4 0.8 0.2 cor = 0.48*** 12 cor = 0.40*** 0.6 3.0 2.5 2.4 14 1.2 1.0 0.8 2.0 1.3 cor = 0.26*** 1.0 cor = 0.62*** 16 1.2 3.5 0.7 0.4 cor = 0.52*** 1.4 4.0 0.8 0.6 1.0 cor = 0.54*** 1.0 0.9 0.8 1.1 1.5 3.5 3.5 4.0 4.0 cor = -0.47*** 2.2 1.2 4.0 1.6 1.0 2.4 2.0 1.2 2.2 1.7 1.6 1.2 cor = -0.49*** 1.5 1.0 2.6 2.0 1.1 cor = -0.48*** cor = -0.69*** 1.8 1.3 1.5 2.3 1.0 cor = -0.57*** 0.75 0.75 0.75 0.75 1.2 0.50 0.50 0.50 0.50 0.8 0.25 0.25 0.25 0.25 0.4 1.5 1.5 1.5 1.5 1.0 1.0 1.0 1.0 0.5 0.5 0.5 0.5 cor = -0.36*** cor = -0.34*** 2.0 1.5 1.0 cor = -0.27*** 0.5 cor = -0.24*** 1.8 1.8 1.6 1.6 1.4 cor = -0.25*** 1.2 2.25 cor = -0.11*** 0.8 0.5 1.5 1.75 1.4 1.2 0.4 4.0 0.9 3.5 0.6 0.6 1.0 1.25 0.3 0.75 cor = -0.53*** 2.0 4.0 6.0 Distance from CW to Cut (µm) 0.2 cor = -0.62*** 2.0 4.0 6.0 8.0 10.0 Distance from CW to Cut (µm) cor = -0.27*** 0.5 2.0 4.0 6.0 8.0 Distance from CW to Cut (µm) 0.4 cor = -0.50*** 2.0 4.0 6.0 8.0 Distance from CW to Cut (µm) cor = -0.29*** 2.5 2.0 4.0 6.0 8.0 Distance from CW to Cut (µm) 0.3 cor = -0.32*** 2.0 4.0 6.0 8.0 Distance from CW to Cut (µm) bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which not certified byprofile peer review) the author/funder, who has granted bioRxiv a license display the preprint in perpetuity. It(a) is made Fig. was 4 Chemical of isthe cutin polymer matrix (CPM) overtothe in-depth thickness. available under aCC-BY-ND 4.0 International license. Illustrations of sampling zone in cross-section of tomato cuticle at 25 DPA for AFM maps (left), OPTIR (middle) and Raman (right). For both OPTIR and Raman data, the sampled zone is represented by a K-means clustering map with three clusters for OPTIR data and five clusters for Raman. To see the mean spectra associated to the different clusters, readers are referred to Fig. S3. Charts represent means spectra per cluster for each dataset. Bars, 5 µm. (b) Chemical composition assessed by Raman spectroscopy and elastic mechanical properties determined by PF-QNM from the cell surface to the cuticle surface of the CPM over tomato fruit development. Chemical profiles of CPM components were obtained on area-mean normalized data to avoid biases due to difference in crosssection thickness (see Fig. 2) and expressed as Raman intensity with arbitrary units (a.u.). For each component a specific band was selected according to a homemade reference database (Reynoud et al., 2022): Cutin (1440 cm-1), Crystalline cellulose (380 cm-1), Pectin (852 cm-1), Polysaccharides (1375 cm-1), p-Coumaric acid (1605 cm-1), Flavonoid 1 (547 cm-1), Flavonoid 2 (1550 cm-1) and Phenolics (1170 cm-1). Bold values in charts represent Pearson correlations (*P value < 0.05; ** P value < 0.01; *** P value < 0.001) between elastic modulus and Raman intensity. See Fig. S4 for full spectra. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 20 DPA (a) 30 DPA furrow sides central furrow 4 4 2 2 40 DPA 0 Dim2 (30.5%) Dim2 (19.4%) Dim2 (10.2%) RAMAN 4 0 −2 0 −4 −2 −4 −4 0 4 8 Dim1 (73.4%) −6 −3 0 −2.5 3 Dim1 (46.5%) 0 2.5 5.0 Dim1 (32.7%) furrow sides central furrow 2 1 0 Dim2 (5.8%) 4 Dim2 (39.8%) OPTIR Dim2 (16.9%) 2 0 0 −1 −2 −4 −2 −3 −4 0 4 −3 8 Dim1 (65.5%) (b) Raman intensity (counts) 200 Young's modulus (GPa) 20 DPA 1.00 0.75 0.50 0 3 −5 6 Dim1 (48.8%) 0 5 Dim1 (89.9%) 1440 δ(CH2, CH3) Cutin 1065 1305 1440 v(C-C) τ(CH2) δ(CH2, CH3) Cutin Cutin Cutin 1305 τ(CH2) Cutin 150 1065 v(C-C) Cutin 100 50 0.25 0 0.00 −2.0 −1.0 0.0 1.0 Distance (µm) 2.0 400 600 800 1000 1200 −1 Wavenumber (cm ) 1400 1600 bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint Fig. was 5 Chemical macromolecular of thethecentral (which not certified bycomposition peer review) is the and author/funder, who has granted arrangement bioRxiv a license to display preprint infurrow perpetuity.over It is made available under aCC-BY-ND 4.0 International license. the development of the tomato CPM. (a) Principal component analysis (PCA) of Raman (up) and OPTIR (down) datasets of the central furrow compared to furrow sides at 20, 30 and 40 days post anthesis (DPA). Sampling zones and PCA loadings are found in supplemental information (Fig. S5 and S6). (b) Macromolecular orientation of lipids within the central furrow compared to the furrow sides at 20 DPA. The laser polarization direction was modified from parallel (0°) to orthogonal (90°) to the cuticle surface. For each polarization, two maps were acquired at 20 and 40 DPA (Fig S7). The light image (upper left) represents the area of the cuticle that were sampled to assess the molecular orientations within the central furrow (pink) and the furrow sides (blue) with the corresponding elastic modulus (lower left). Bar, 5 µm. Data are expressed as mean (solid lines) ± SD (shadowed area). For both central furrow and furrow sides, mean Raman spectra for each polarization direction were calculated (right). The parallel (solid lines) to orthogonal (dashed lines) ratio of intensity was illustrated by vertical bars for three bands related to lipids: two bands attributed to CH2 deformations, i.e., τ(CH2) at 1305 cm-1 and δ(CH2,CH3) at 1440 cm-1 and one band assigned to ν(C-C) vibration at 1065 cm-1. Inset represents the lined up different intensity ratio (bars) for proper comparison between cuticle areas. bioRxiv preprint doi: https://doi.org/10.1101/2022.12.19.521062; this version posted December 19, 2022. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. (a) 1.25 Control Cutinase Height (µm) 1.00 0.75 0.50 3 y: 0.25 m 0µ 1.5 µm 0.00 0.0 0.0 µm x: m 17 µ 1.0 2.0 3.0 4.0 5.0 Transverse profile (µm) (b) 1.25 1.3 µm y: 0.0 µm 30 µm x: 2 m 6µ Height (µm) 1.00 0.75 0.50 0.25 0.00 0.0 1.0 2.0 3.0 4.0 5.0 Transverse profile (µm) Fig. 6 Susceptibility of the cutin polymer matrix (CPM) to cutinase hydrolysis. Topography of the CPM either subjected to cutinase treatment or not at 20 (a) and 35 (b) days post anthesis (DPA). The 3D topographies (left) illustrate the carving effect of the cutinase in the central furrow with the blue line representing the sampled transverse profiles within the CPM. The corresponding topographic profiles (right) were calculated as a mean of 10 profiles over 3 µm, each profile being a mean of 8 measurements. Data are expressed as mean (solid lines) ± SD (shadowed area). Two maps per developmental stages were acquired.