Biomechanical properties of the tomato (Solanum lycopersicum) fruit cuticle during development are modulated by changes in the relative amounts of its components

Authors


Summary

  • In this study, growth-dependent changes in the mechanical properties of the tomato (Solanum lycopersicum) cuticle during fruit development were investigated in two cultivars with different patterns of cuticle growth and accumulation.
  • The mechanical properties were determined in uniaxial tensile tests using strips of isolated cuticles. Changes in the functional groups of the cuticle chemical components were analysed by attenuated total reflectance–Fourier transform infrared (ATR-FTIR).
  • The early stages of fruit growth are characterized by an elastic cuticle, and viscoelastic behaviour only appeared at the beginning of cell enlargement. Changes in the cutin:polysaccharide ratio during development affected the strength required to achieve viscoelastic deformation. The increase in stiffness and decrease in extensibility during ripening, related to flavonoid accumulation, were accompanied by an increase in cutin depolymerization as a result of a reduction in the overall number of ester bonds.
  • Quantitative changes in cuticle components influence the elastic/viscoelastic behaviour of the cuticle. The cutin:polysaccharide ratio modulates the stress required to permanently deform the cuticle and allow cell enlargement. Flavonoids stiffen the elastic phase and reduce permanent viscoelastic deformation. Ripening is accompanied by a chemical cleavage of cutin ester bonds. An infrared (IR) band related to phenolic accumulation can be used to monitor changes in the cutin esterification index.

Introduction

The shape of an organism relies on a complex network of genetic regulation together with the contribution of mechanical forces (Mirabet et al., 2011). Thus, understanding shape changes requires the assessment of both molecular control and mechanical status. In plants, modifications in the mechanical properties of the cell wall are the main determinant of shape changes and growth (Cosgrove, 2005; Mirabet et al., 2011). The epidermal tissue, as the interface with the environment, has been shown to act as a barrier that limits and controls organ growth (Savaldi-Goldstein et al., 2007). The epidermis is in tension in planta as excised epidermal peels become shorter (Skene, 1980; Kutschera et al., 1987; Grimm et al., 2012) and cuts in the epidermis tend to remain open (Dumais & Steele, 2000). Epidermal cells form a continuous and uniform layer of strongly attached cells with thicker cell walls which are expected to be less extensible (Javelle et al., 2011; Mirabet et al., 2011). Hence, for an organ to grow it must overcome the restriction imposed by the epidermis. In this sense, the cuticle deposited on the outer surface of organs plays an important structural role as it stiffens the epidermal cell walls while they, in turn, give strength to the cuticle (Niklas, 1992; Matas et al., 2004; Bargel & Neinhuis, 2005).

Epidermal cells of the aerial parts of higher plants are covered by a cuticle, a continuous extracellular lipid membrane located at the boundary with the external environment. Cuticles are chemically heterogeneous in nature, basically consisting of an insoluble cutin matrix deposited on and intertwined with the outer epidermal cell wall. Hence, the cuticle can be regarded as a special modification of the cell wall, a cutinized cell wall (Riederer, 2006). Cutin is a high-molecular-weight polyester mainly composed of hydroxylated and epoxy-hydroxylated C16 and C18 esterified fatty acids (Heredia, 2003; Domínguez et al., 2011a). In addition, waxes and phenolics are also present. Waxes are mixtures of homologous series of long-chain aliphatics, such as alkanes, alcohols, aldehydes, fatty acids and esters, together with variable amounts of cyclic compounds such as triterpenoids (Bianchi, 1995). These waxes can be deposited on the surface (epicuticular) or embedded in the cutin matrix (intracuticular). The main constituents of the phenolic fraction are hydroxycinnamic acid derivatives and flavonoids (Hunt & Baker, 1980; Pfündel et al., 2006), although in some gymnosperms a lignin-like fraction has also been detected (Reina et al., 2001). In the tomato (Solanum lycopersicum) fruit cuticle the main phenolic compounds are the flavonoids naringenin and chalconaringenin, which accumulate during ripening (Schönherr & Bukovac, 1973; Hunt & Baker, 1980; Baker et al., 1982; Luque et al., 1995).

Tomato is considered a model of fleshy climacteric fruit for both basic and applied research (Foolad, 2007). Studies carried out in this species include studies investigating genes involved in fruit growth and ripening (Seymour et al., 2013), fruit anatomy development (Gillaspy et al., 1993), changes in cell wall chemistry (Lunn et al., 2013) and physiological disorders of economic importance such as cracking (Dorais et al., 2004). The tomato fruit has also become a model for cuticle biophysical studies, including biomechanics (for a review, see Domínguez et al., 2011a,b).

There is little scientific information on the development of the mechanical properties of the cuticle during plant growth. Knoche et al. (2004) studied changes in the mechanical properties of sweet cherry (Prunus avium) and found that the cuticle increased its elastic deformation during growth. More recently, Tsubaki et al. (2012) reported an increase in the stiffness and a decrease in the deformability of the Sonneratia alba leaf cuticle during development. In tomato, the rheological properties of the cuticle have been studied at only three stages of fruit development (Bargel & Neinhuis, 2005; López-Casado et al., 2007; Domínguez et al., 2009). Whereas no biomechanical differences were observed between immature and mature green tomato cuticles, a significant increase in stiffness and a decrease in strain were observed during ripening (Bargel & Neinhuis, 2005; López-Casado et al., 2007; Domínguez et al., 2009). However, a detailed biomechanical study of the tomato fruit cuticle during development is still needed.

The present work attempts to fill this gap by investigating the biomechanical behaviour of the tomato fruit cuticle together with the relationship among cuticle components, the macromolecular and structural characteristics of the cuticle and its mechanical properties. ‘Cascada’ and ‘Moneymaker’, two genotypes that represent the different patterns of cuticle accumulation described in this species, were employed.

Materials and Methods

Plant material

Two Solanum lycopersicum L. genotypes, ‘Cascada’ and ‘Moneymaker’, were used in this study. ‘Cascada’ is a cherry genotype whereas ‘Moneymaker’ is a medium-sized tomato. Plants were grown in a polyethylene glasshouse at the Estación Experimental La Mayora, Consejo Superior de Investigaciones Científica (CSIC), in the south-east of Spain. Tomato seedlings were grown in an insect-proof glasshouse, and transplanted into soil in a plastic-house at the four true-leaf growth stage. The within-row and between-row spacing was 0.5 and 1.5 m, respectively. Plants were watered when necessary, fertilized with the nutrient solution recommended by Cánovas (1995), supported by strings and pruned to a single stem. The harvesting period lasted from early May until mid-July.

Fruits were harvested at different stages of development – ‘Cascada’ every 5 d, starting at 10 d after anthesis (daa), until red ripe (Domínguez et al., 2008) and ‘Moneymaker’ every 10 d until red ripe, starting at 15 daa. Fruits with sizes far from the average were discarded. A minimum of 30 fruits per genotype and developmental stage were collected except for early stages of development when 100 fruits were collected. For both genotypes, 35 daa corresponded to mature green fruits, 45 daa to breaker and 55 daa to red ripe.

Tissue sectioning and staining

Small pericarp pieces from three fruits per genotype and developmental stage were fixed in a formaldehyde, acetic acid and ethanol solution (1 : 1 : 18); and later dehydrated in an ethanol dilution series (70–95%) and embedded in a commercial resin (Historesin Embedding Kit; Leica, Heidelberg, Germany). Samples were cross-sectioned into slices 4 μm thick using a Leica microtome (RM2125; Leica) and stained with Sudan IV (Jensen, 1962) to visualize the cuticle. Cuticle thickness was estimated from a minimum of 30–50 measurements on the cross-sectioned samples using an image capture analysis program (Visilog-Noesis 6.3; Noesis SA, Orsay Cedex, France). Ten sections per pericarp sample were analysed and one to two measurements taken in each section. The cuticle surface area was calculated using the same program and a minimum of 20 measurements per genotype and stage. Samples were then weighed in order to calculate the amount of cuticle per surface area. Cuticle density was calculated from the area of cross-sectioned samples using an image capture analysis program (Visilog-Noesis 6.3) and referred to the weight of the cuticular surface area already estimated.

Cuticle isolation

Cuticles were enzymatically isolated from tomato fruits following the protocol of Petracek & Bukovac (1995) using an aqueous solution of a mixture of fungal cellulase (0.2% w/v; Sigma, St Louis, MO, USA) and pectinase (2.0% w/v; Sigma), and 1 mM NaN3 to prevent microbial growth, in sodium citrate buffer (50 mM; pH 3.7). Vacuum was used to facilitate enzyme penetration, and fruit samples were incubated with continuous agitation at 35°C for at least 14 d. The cuticle was then separated from the epidermis, rinsed in distilled water and stored under dry conditions.

Cuticle components

Cuticular waxes were removed by heating (50°C) 0.1 g of isolated cuticles in 100 ml of chloroform : methanol (2 : 1, v/v) for 2 h. Cutin isolates were obtained after refluxing the dewaxed cuticles in a 6 M HCl aqueous solution for 12 h. Phenolic components were estimated after cutin depolymerization in a UV-VIS spectrophotometer (Pharmacia Biotech, Piscataway, NJ, USA) following the protocol described by Domínguez et al. (2009).

Cuticular components were estimated from 10 samples per genotype with each sample corresponding to cuticle pieces from a minimum of 30 fruits.

Ultra-high-performance liquid chromatography–mass spectrometry (UPLC-MS)

Cutin samples were depolymerized according to the protocol described in Domínguez et al. (2009) and neutralized with trifluoracetic acid to pH 3–5. Samples were then centrifuged at 13 000 g for 10 min and the supernatant was filtered with a PVDF membrane of 0.2 μm pore diameter. An aliquot (4 μl) of each sample was injected into the UPLC-MS apparatus. Separation by UPLC was conducted using an Acquity UPLC-PDA coupled to a Q-TOF mass microspectrometer (Waters Corp., Milford, MA, USA) with an Acquity BEH C18 column (50 mm × 2.1 mm). The column temperature was 40°C and the sample was injected at 20°C. Phenolic compounds were separated using gradient elution with mobile phases of 0.1% (v/v) formic acid in purified water (phase A) and 0.1% (v/v) formic acid in acetonitrile (phase B). A constant flow rate of 0.4 ml min−1 was used throughout. Mass spectra were determined in the positive-ion-detecting mode at the electrospray ionization intersurface. Two replicates per developmental stage were analysed.

Attenuated total reflectance–Fourier transform infrared spectroscopy (ATR-FTIR)

Infrared spectra of samples were obtained with an ATR accessory (MIRacle ATR; PIKE Technologies, Madison, WI, USA) coupled to an FTIR spectrometer (FT/IR-4100; JASCO, Tokyo, Japan). All spectra were recorded in the range from 3800 to 600 cm−1 with 4 cm−1 resolution and accumulating 50 scans. Samples were gently collocated on the spot of the ATR accessory and slowly pressed. Three replicates per genotype and stage were analysed on both the external and internal sides of isolated cuticle and cutin.

The penetration depth, dp, of ATR-FTIR depends on several factors such as the wave number (λ), the angle of incidence (θ), and the refractive index of the ATR crystal (n1) and the sample (n2), as indicated in the equation:

display math(Eqn 1)

Considering 1.5 the refractive index of the cuticle (Holloway, 1982), and θ = 45° and n1 = 2.4 for the equipment used, the penetration depth of ATR into the samples varied between 0.53 μm at 3800 cm−1 and 2.86 μm at 700 cm−1.

The degree of cutin esterification was estimated from the ratio between absorbance at 1730 cm−1 and absorbance at 2900 cm−1, A1730 : A2900 cm−1. Deconvolution was applied to the C=O stretching region (1770–1650 cm−1) using PeakFit v4.11 software (Scientific Software Solutions Int., Pella, Greece).

Mechanical tests

The mechanical properties of the cuticle were measured with an extensometer customized to work with cuticular membrane samples (resolution of ± 1 μm) following the method described by López-Casado et al. (2007). The mechanical tests were performed as transient creep tests maintaining samples in uniaxial tension, under a constant load for 1200 s, with the longitudinal extension being recorded every 3 s. Each sample was tested using an ascending sequence of sustained tensile forces (from 0.098 N to breaking-point by 0.098-N load increments) without recovery time (Matas et al., 2005). A minimum of five to seven samples per genotype and stage were analysed at 25°C and 40% relative humidity (RH). To obtain the corresponding stress–strain curves and Young's modulus (E), the applied stress was plotted against the total change in length (%) for every tensile force. The Young's modulus, an indicator of the sample stiffness, was calculated from the slope of the linear elastic phase. The breaking stress (stress needed to break the sample) and maximum strain at the breaking stress were also determined for each sample.

Statistics

Data are expressed as mean ± 1 SE. One-way ANOVA analyses were used to compare means between stages of development for each genotype, and Pearson's coefficient was used to assess the correlations between the variables studied in each genotype (SPSS, 2008). Letters in the figures and tables indicate significant differences (< 0.05).

Results

Biomechanical characterization

Stress–strain diagrams of the ‘Cascada’ cuticle throughout development showed a different mechanical behaviour at early stages of growth compared to later stages (Supporting Information Fig. S1). At 10 daa the cuticle was monophasic, with only elastic deformation. Viscoelastic deformation appeared at 15 daa, with several curves already showing a biphasic behaviour, although some still had only elastic deformation. From 20 daa until red ripe, the typical presence of two phases, an elastic behaviour at slow stresses and a viscoelastic behaviour at higher stresses, could be observed (Fig. 1a). In all the cases, the slope of the elastic phase was higher than that of the viscoelastic phase.

Figure 1.

Elastic (closed bars) and viscoelastic (open bars) strain (a, c) and stress (b, d) of the tomato (Solanum lycopersicum) fruit cuticle during development in the genotypes ‘Cascada’ (a, b) and ‘Moneymaker’ (c, d). Here, 35 d after anthesis (daa) corresponds to the mature green stage, 45 daa to breaker and 55 daa to red ripe. Error bars represent ± 1 SE.

Table 1 shows the biomechanical parameters calculated from these stress–strain curves, together with the mean thickness of the ‘Cascada’ fruit cuticle during growth and ripening. Young's modulus did not significantly change throughout the growth period (10–35 daa) except for a significant but transient increase at 15–20 daa. At the onset of ripening (40 daa), the modulus started to significantly increase until it reached its maximum at red ripe. During this short period, the Young's modulus exhibited a threefold increase. Breaking stress showed a similar behaviour to the Young's modulus, being low during fruit growth with a temporary increase at 15–20 daa. Later, breaking stress started to increase during ripening until red ripe. Maximum strain showed the opposite behaviour: an increase in the growth period until a maximum at mature green (35 daa) followed by a decrease during ripening. Most of the deformation suffered by the cuticle was viscoelastic, except at 10 daa where only elastic deformation was observed (Fig. 1a). Hence, the decrease in maximum strain observed during ripening was attributable only to a decrease in viscoelastic deformation, with no change in elastic deformation (Fig. 1a). In contrast, the increase in breaking stress during ripening was mostly attributable to an increase in the maximum stress withstood in the elastic phase (Fig. 1b).

Table 1. Biomechanical parameters of the tomato (Solanum lycopersicum) ‘Cascada’ fruit cuticle throughout development
Stage (daa)Young's modulus (MPa)Breaking stress (MPa)Maximum strain (%)Thickness (μm)
  1. Young's modulus is an indicator of sample stiffness, and breaking stress is the stress needed to break the sample. Maximum strain (maximum elongation) is expressed in %. Data are presented as mean ± SE. Means with different letters are significantly different (< 0.05). Thickness data are taken from Domínguez et al. (2008). daa, days after anthesis.

10195.1 ± 29.1 d20.9 ± 2.3 e10.6 ± 0.6 c1.9 ± 0.2 e
15384.7 ± 54.8 bc39.7 ± 5.3 bcd16.4 ± 3.2 bc4.4 ± 0.1 d
20398.1 ± 31.5 bc44.6 ± 2.7 abc18.1 ± 1.7 bc5.7 ± 0.1 c
25259.7 ± 18.0 cd25.8 ± 5.4 de19.9 ± 5.3 bc6.7 ± 0.1 b
30211.5 ± 31.5 d30.7 ± 1.8 cd23.0 ± 2.7 abc7.4 ± 0.1 a
35265.5 ± 31.4 cd33.4 ± 1.8 bcd31.5 ± 1.5 a7.2 ± 0.1 a
40415.1 ± 33.9 bc40.2 ± 2.6 bcd26.2 ± 1.3 ab7.3 ± 0.1 a
45553.0 ± 11.5 b49.6 ± 1.0 ab17.2 ± 1.4 bc6.5 ± 0.1 b
50546.1 ± 31.7 b44.1 ± 2.6 abc17.2 ± 3.2 bc6.1 ± 0.2 bc
55884.1 ± 70.1 a58.6 ± 6.7 a14.4 ± 2.8 bc6.2 ± 0.1 b

A similar analysis was carried out in ‘Moneymaker’, a medium-sized tomato genotype. Similarly, at the earliest stage of development, 15 daa, the cuticle exhibited only elastic behaviour, whereas in the next stages the typical biphasic behaviour was observed (Figs 1c, S2). Table 2 shows the biomechanical parameters together with the mean thickness of the ‘Moneymaker’ fruit cuticle during growth and ripening. Young's modulus significantly increased between 15 and 25 daa, did not change until the rest of fruit growth and then significantly increased (almost threefold) between breaker and red ripe. Breaking stress significantly increased until breaker and then decreased at red ripe to values similar to those of mature green. Maximum strain increased during growth and significantly dropped at red ripe. Similarly to ‘Cascada’, most of the cuticle strain was viscoelastic with only a little elastic deformation except at 15 daa, where only elastic deformation was observed. The decrease observed during ripening in maximum strain was attributable to a decrease in both elastic and viscoelastic deformation (Fig. 1c). Changes in the breaking stress were attributable to modifications of the viscoelastic phase, but did not affect the elastic phase (Fig. 1d).

Table 2. Biomechanical parameters and thickness of the tomato (Solanum lycopersicum) ‘Moneymaker’ fruit cuticle throughout development
Stage (daa)Young's modulus (MPa)Breaking stress (MPa)Maximum strain (%)Thickness (μm)
  1. Young's modulus is an indicator of sample stiffness, and breaking stress is the stress needed to break the sample. Maximum strain (maximum elongation) is expressed in %. Thickness data are taken from Domínguez et al. (2008). Data are presented as mean ± SE. Means with different letters are significantly different (< 0.05). daa, days after anthesis.

1562.1 ± 4.4 c9.7 ± 1.4 c15.1 ± 2.0 bc2.4 ± 0.1 e
25201.4 ± 12.6 b40.7 ± 1.8 b27.7 ± 3.5 a4.6 ± 0.1 d
35246.7 ± 15.3 b34.3 ± 4.3 b21.2 ± 2.1 ab5.1 ± 0.1 c
45216.6 ± 19.5 b55.3 ± 1.8 a30.1 ± 1.9 a6.1 ± 0.1 b
55617.8 ± 20.2 a32.4 ± 2.7 b6.4 ± 1.4 c7.1 ± 0.1 a

Changes in cuticle thickness were observed during growth and development in both genotypes. In ‘Moneymaker’, cuticle thickness steadily increased during growth and ripening until it reached its maximum at red ripe (Table 2). However, in ‘Cascada’, cuticle thickness increased during the immature green to mature green period (until 30–40 daa) and then significantly decreased during ripening (Table 1; Domínguez et al., 2008). Cuticle density throughout development is shown in Fig. 2. For both genotypes, a significant increase was observed at early stages of development (15 and 25 daa for ‘Cascada’ and ‘Moneymaker’, respectively) which corresponded to a significant increase in the accumulation of cuticle material. In the next stages, density decreased and remained constant until the onset of ripening. During ripening, ‘Cascada’ density increased while in ‘Moneymaker’ there was a marked decrease.

Figure 2.

Changes in cuticle density throughout fruit development in tomato (Solanum lycopersicum) genotypes ‘Cascada’ (closed circles) and ‘Moneymaker’ (open circles). Here, 35 d after anthesis (daa) corresponds to the mature green stage, 45 daa to breaker and 55 daa to red ripe. Error bars represent ± 1 SE.

Fig. 3 shows the changes in the amount of phenolic compounds present in ‘Cascada’ and ‘Moneymaker’ fruit cuticles throughout growth and ripening. In both genotypes, the amount of phenolics was very low during growth and significantly increased during ripening. In ‘Cascada’ (Fig. 3a), the increase was gradual and started at 35 daa until red ripe, with an almost twofold increase between the light orange (50 daa) and red ripe stages. However, in ‘Moneymaker’ the amount of phenolics remained low until breaker and showed a 10-fold increase at red ripe (Fig. 3b). A significant correlation could be observed in ‘Cascada’ between the amount of phenolics and the stiffness (= 0.927; < 0.01) and breaking stress (= 0.926; < 0.01). Similarly, in ‘Moneymaker’, the amount of phenolics correlated with the Young's modulus (= 0.930; < 0.05) but not with the breaking stress. These correlations were positive, suggesting a contribution of phenolics to the increase in stiffness and stress needed to break the cuticle.

Figure 3.

Changes in cuticle phenolics throughout fruit development in tomato (Solanum lycopersicum) genotypes ‘Cascada’ (a) and ‘Moneymaker’ (b). Error bars represent ± 1 SE. Letters indicate significant differences (< 0.05). Relative changes in p-coumaric acid (c) and p-hydroxy-benzoic acid (d) in the ‘Cascada’ cuticle during fruit development were identified by ultra-performance liquid chromatography–mass spectrometry (UPLC-MS). Data are presented as fold changes between the earliest stage studied (15 daa) and the other stages. Here, 35 d after anthesis (daa) corresponds to the mature green stage, 45 daa to breaker and 55 daa to red ripe.

The previous methodology applied to estimate phenolics did not allow discrimination between hydroxycinnamic acid derivatives and the flavonoid naringenin present in the tomato fruit cuticle. Hence, a UPLC-MS analysis of the phenolic compounds present in the isolated cuticles was performed. However, the extreme conditions of low pH and high temperature needed to depolymerize the cuticle prevented the analysis of naringenin and chalconaringenin, as they are chemically labile and suffer irreversible modifications such as oxidization and partial dimerization under these conditions. These conditions, nevertheless, did not seem to affect the hydroxycinnamic acids present in the cutin matrix. Thus, two phenolic acids, p-coumaric acid and p-hydroxybenzoic acid, could be detected in the cuticle during growth and ripening. The amount of p-coumaric acid did not change during growth and only increased slightly during ripening (Fig. 3c). The amount of p-hydroxybenzoic acid also remained constant during growth but showed an increase during ripening with a significant 40-fold maximum increase at breaker (Fig. 3d).

Infrared spectroscopy analysis

The development of plant cuticles and isolated cutins was also characterized by ATR-FTIR. Table S1 summarizes the main chemical functional groups associated with the different vibrations present in the outer and inner surfaces of ‘Cascada’ and ‘Moneymaker’ cuticle and cutin at 15 and 55 daa. The corresponding spectra are shown in Fig. S3. The 15-daa stage was chosen as an example of cuticles from growing fruits, as no significant band shifts were observed in either genotype during this period of development. More information about the assignments described in Table S1 can be found in the literature (Ramírez et al., 1992; Luque et al., 1995; Villena et al., 2000; Mazurek et al., 2013). Chemical differences observed in the cuticle spectra between the two surfaces are a consequence of the spatial and asymmetrical distribution of cuticle components. The outer surface of the cuticle is rich in waxes, aliphatic compounds, while in the inner surface polysaccharides from the epidermal cell wall (cellulose, pectin and hemicelluloses) are abundant.

Cutin from growing fruits (10–35 daa) showed the typical features of an aliphatic polyester with the characteristics bands for the hydroxyl (3403 cm−1), methylene (mainly νa(CH2) 2926 cm−1, νs(CH2) 2854 cm−1, δ(CH2) scissoring 1463 cm−1 and δ(CH2) rocking 724 cm−1) and ester functional groups (ν(C=O) 1729 cm−1) together with two well-defined shoulders ascribed to carboxylic acids with H bonds (ν(C=O···H) weak 1707 cm−1 and strong 1686 cm−1). These features are representative of the cutin chemical composition: polyester formed from polyhydroxy fatty acids. A similar spectroscopic pattern was observed for the cuticle outer surface but with subtle differences. The band associated with hydroxyl stretching was placed at 3390 cm−1 and the absorptions of the methylene stretching vibrations (νa and νs(CH2) 2919 and 2850 cm−1, respectively) were stronger, better defined and located at the typical values of alkyl chains in an ordered all-trans conformation (Douliez et al., 2005). Also, an improvement of the pyranose stretching (1054 cm−1) spectral region was observed. However, the cuticle inner surface showed a high increase in the bands ascribed to the polysaccharide functional groups (hydroxyl and glycosidic groups, 3340 and 1053 cm−1, respectively) in comparison to the outer surface.

Infrared spectra remained mostly unchanged during ripening, although three significant differences were observed compared with the growing period. These differences were visible in the cutin and cuticle outer and inner surfaces and corresponded to an increase in the intensity of the O-H stretching vibration and the relative intensity of the 1705 cm−1 shoulder of the C=O stretching band. Also, the presence of absorptions in the 1650–1550 cm−1 region and c. 834 cm−1 (associated with double bonds and aromatic molecules, Table S1) were not present in earlier stages of development. Such vibrations have been assigned to the phenolic compounds present in the cuticle of ripe tomato fruits. The above differences in the 1760–1485 cm−1 spectral region between cuticles of unripe and ripe fruits are shown in Fig. 4.

Figure 4.

Attenuated total reflectance–Fourier transform infrared (ATR-FTIR) spectra of the C=O and C=C stretching vibration region (1780–1490 cm−1) of tomato (Solanum lycopersicum) ‘Cascada’ (a) cutin, (b) cuticle outer surface and (c) cuticle inner surface. Red line: 55 d after anthesis (daa); red ripe. Black line: 15 daa; immature green. au, arbitrary units.

The presence of IR bands associated with phenolic compounds could be used to follow their incorporation in the cuticle during fruit ripening. For this purpose, changes in the area of the γ(C-H) aromatic band (c. 834 cm−1) were monitored during fruit development. This band was chosen because it is located in a spectral region without interference from other vibrations. Fig. 5 shows these changes in ‘Cascada’, although a similar behaviour was observed in ‘Moneymaker’. A significant increase in area was observed during ripening. This started at the mature green stage and reached a maximum at red ripe. Again, for ‘Moneymaker’ this change was less gradual and started at breaker. This change was more marked in the cutin and cuticle outer surface, while in the cuticle inner surface was barely detected.

Figure 5.

Area of the γ(C-H) aromatic band associated with the presence of phenolic compounds in the tomato (Solanum lycopersicum) ‘Cascada’ fruit cuticle during development. Closed circles: cuticle outer surface; open circles: cuticle inner surface; closed squares: cutin. Here, 35 d after anthesis (daa) corresponds to the mature green stage, 45 daa to breaker, and 55 daa to red ripe. Error bars represent ± 1 SE. au, arbitrary units.

Fig. 6 shows changes in the cutin esterification index during ‘Cascada’ and ‘Moneymaker’ fruit development. This parameter indicates the relationship between the relative intensities of the methylene groups, the most repeated structural unit in the cutin, and the ester functional group, the bond that links cutin monomers (Benítez et al., 2004; Girard et al., 2012). In our case, the esterification index was calculated as the ratio between the relative intensities of the C=O stretching vibration of the ester (c. 1730 cm−1) and the asymmetric vibration of the methylene (c. 2925 cm−1) functional groups. Results indicated a marked decrease of the esterification index during fruit ripening. In ‘Cascada’, esterification values were high at the immature green stages and gradually decreased from the mature green stage until red ripe, where the lowest value was observed (Fig. 6a). In ‘Moneymaker’, the esterification index remained high and did not change between the immature green stage and early breaker but suddenly decreased at red ripe (Fig. 6b). A linear relationship between the γ(C-H) aromatic band and cutin esterification index could be observed (Fig. 6c,d). Both parameters showed a significant and negative correlation: an increase in the area of the γ band accompanied by a decrease in the esterification index. This decrease in the esterification index was also negatively correlated with the amount of phenolic compounds, as shown in Fig. 3(a,b) (‘Moneymaker’, = −0.926; < 0.05; ‘Cascada’, = −0.939; < 0.05), and cuticle stiffness (‘Moneymaker’, = −0.844; < 0.01; ‘Cascada’, = −0.855; < 0.01).

Figure 6.

Cutin esterification index during fruit development in tomato (Solanum lycopersicum) genotypes ‘Cascada’ (a) and ‘Moneymaker’ (b). The esterification index was calculated as the ratio between the intensities of the C=O stretching vibration of the ester (1730 cm−1) and the asymmetric vibration of the methylene (2925 cm−1) functional groups. There was a linear relationship between the esterification index and γ band area in ‘Cascada’ (c) and ‘Moneymaker’ (d) cutin during fruit development. Here, 35 d after anthesis (daa) corresponds to the mature green stage, 45 daa to breaker, and 55 daa to red ripe. Error bars represent ± 1 SE. au, arbitrary units.

Spectral modifications in the C=O stretching region were monitored through the deconvolution of this band in three different components (ester groups at c. 1735 cm−1, carboxylic acid groups with weak H bonds at c. 1705 cm−1 and carboxylic acid groups with strong H bonds at c. 1685 cm−1) during fruit development (Fig. 7). Fig. S4 shows an example of this type of deconvolution. The highest percentage of area corresponded to C=O absorption of ester groups, which showed high values at the immature and mature green stages and decreased from the mature green stage until a minimum at red ripe. The area ascribed to C=O absorption of carboxylic acids with weak H bonds showed the opposite behaviour. This behaviour could be indicative of a chemical transformation of the ester groups into carboxylic acids with weak H bonds. Finally, C=O bands of carboxylic acids with strong H bonds showed the lowest percentage and, similarly to ester groups, remained constant during fruit growth and decreased during ripening.

Figure 7.

Deconvoluted area of the C=O stretching vibration of ester groups (closed squares), carboxylic acid groups with weak H bonds (open circles), and carboxylic acid groups with strong H bonds (closed circles) in tomato (Solanum lycopersicum) ‘Cascada’ cutin during fruit development. Here, 35 d after anthesis (daa) corresponds to the mature stage. Error bars represent ± 1 SE.

Discussion

Cuticle biomechanics during fruit growth: the importance of the cutin:polysaccharide ratio

Cuticle deposition during growth has been shown to follow different patterns in different species. For example, cuticle deposition in cherry (Prunus avium) occurred at early stages of development and the amount of cuticle decreased during fruit expansion (Peschel et al., 2007). In tomato, two different patterns have been identified, one for cherry tomatoes and another for medium-sized tomatoes. In cherry tomatoes, such as ‘Cascada’, cuticle is accumulated in the early stages of development, and a constant amount of cuticle is then maintained throughout development (Domínguez et al., 2008). In contrast, in medium tomatoes such as ‘Moneymaker’ the amount of cuticle increases during the whole growing and ripening period (Baker et al., 1982; Luque & Heredia, 1994). Table S2 shows the amount of ‘Moneymaker’ cuticle and its components during fruit growth and ripening. In accordance with the findings of previous studies, the amount (in micrograms) of cuticle, cutin, polysaccharides and waxes increased during the whole period until they reached their maxima at the red ripe stage. The cutin:polysaccharide ratio increased during development, indicating a preferential accumulation of lipid material over polysaccharides. At 15 daa, an unusual 1 : 1 ratio between cutin and polysaccharides was observed, but it was modified in the following stages mainly as a result of the accumulation of cutin and waxes. Although there is scarce literature on this topic, our results agree with those of Baker et al. (1982) and Luque & Heredia (1994), where the ratio between cutin and polysaccharides was close to 1: 1 at early stages of tomato fruit development but dramatically changed in later stages. In cherry tomato, as expected from the different pattern of cuticle accumulation, no significant changes in the cutin:polysaccharide ratio were observed at the early stages of development (Domínguez et al., 2008).

These differences in cuticle deposition during development between medium-sized and cherry tomatoes can lead to differences in their biomechanical behaviour. Several parameters are known to influence extensibility, such as density, thickness, the chemical nature and fraction of each component, and the bonds linking the different molecules (Geitmann & Ortega, 2009). Organ growth is the result of two processes: cell division that, in the case of tomato fruit, is mostly confined to the earliest stages of development, and cell expansion (Ho & Hewitt, 1986; Bertin et al., 2007). The latter is believed to be responsible for most of the organ growth, and is driven by the yielding of the cell wall that generates an irreversible increase in cell volume (Schopfer, 2006). The cuticle exhibited only elastic, reversible deformation at the earliest stage of development studied in both genotypes. This is in agreement with a period of cell division with no cell enlargement (Bertin et al., 2007; Domínguez et al., 2008). In this period, the cuticle was very thin with low amounts of material accumulated (Domínguez et al., 2008; Table S2). A viscoelastic, irreversible behaviour appeared at the next stage studied (15 daa in ‘Cascada’, and 25 daa in ‘Moneymaker’) and corresponded with the beginning of cell enlargement. Although Bertin et al. (2007) reported that cell division ceased in medium-sized tomatoes c. 30 daa, in ‘Moneymaker’, epidermal cells had already significantly decreased their number per unit surface area and increased in size at 25 daa (data not shown). In contrast, Knoche et al. (2004) found only plastic deformation in the cuticle at early stages of cherry growth while elastic deformation appeared later in development. The mechanism by which the cell wall switches between elastic and viscoelastic deformation and its genetic control are still unclear (Burgert & Keplinger, 2013). The presence of viscoelastic deformation was related to a significant accumulation of cuticle material, with twofold increases being found in both ‘Cascada’ and ‘Moneymaker’. This accumulation was accompanied by an increase in cuticle thickness, density, stiffness and deformability, suggesting a more compacted and stiff yet deformable material. This increase in cuticle density was transient, as the cuticle density decreased later, and seemed to mark a period of massive accumulation of material that was rearranged in the following stages.

Once the stage of biphasic behaviour (elastic at low stresses and viscoelastic at higher stresses) had been reached, the cuticle exhibited little biomechanical change during the remainder of the fruit expansion period: its stiffness, breaking stress and maximum strain remained constant. These results are in agreement with those of Bargel & Neinhuis (2005), who found no biomechanical differences between the cuticles of immature and mature green tomato fruits. Takahashi et al. (2012), in contrast, observed a continued increase in cuticle stiffness and breaking stress together with a decrease in maximum strain during S. alba leaf expansion. However, some differences in the yield stress and strain were observed during fruit expansion (see Fig. 1). It is known that the yield threshold is not a static mechanical property but is spatially and temporally modulated by the cell (Dumais et al., 2006). These changes corresponded to a decrease in the elastic phase at 25 and 35 daa in ‘Cascada’ and 35 daa in ‘Moneymaker’ which did not affect the final breaking stress or maximum strain; hence the duration of the viscoelastic phase was increased. It should be pointed out that Young's modulus was not modified, as both the stress and the strain at the yield point were decreased. Interestingly, at these stages the cutin : polysaccharide ratio was temporarily increased and the cuticle had more cutin than polysaccharides (see data in Domínguez et al. (2008), and Table S2). Cutin has been shown to be mostly viscoelastic, whereas polysaccharides confer stiffness to the cuticle (López-Casado et al., 2007). Thus, our data indicate that the elastic phase is susceptible not only to the amount of polysaccharides (López-Casado et al., 2007; Tsubaki et al., 2012) but also to subtle changes in the ratio between the main cuticle components. This implies a role of the cutin:polysaccharide ratio in the modulation of cuticle viscoelastic deformation, thus contributing to cell enlargement. This is an important topic with clear biological implications that deserves further research.

As was mentioned in the Discussion, density is important in determining polymer extensibility. Nevertheless, there is very little literature on this topic. Schreiber & Schonherr (1990) used a pycnometer to calculate the density of the tomato fruit cuticle. This method has the inconvenience of the substantial amount of cuticle needed per replicate. Takahashi et al. (2012) estimated cuticle density in S. alba leaves using the average cuticle thickness. Yet, in tomato fruit, cuticle pegs and cell invaginations are of significant size and vary between stages and genotypes; hence the use of thickness would lead to an overestimation of cuticle density. The methodology employed in this study takes both invaginations and pegs into account and needs only small amounts of cuticle, allowing a large number of replicates. Our results showed that tomato fruit cuticle density varied between 0.7 and 1.3 g cm−3 depending on the genotype and stage of development. These results are in accordance with those obtained by Schreiber & Schonherr (1990), who calculated a density of 1.0–1.3 g cm−3 for tomato fruit.

The ‘Cascada’ cuticle showed a significant but transient increase in cuticle stiffness and breaking stress at the first stages of cell expansion (15 and 20 dda) that was not observed in ‘Moneymaker’. This increase was related to a higher density at 15 and 20 daa compared with the following stages, which suggests that a more compacted material could be responsible for this change in stiffness and breaking stress. The gradual pattern of cuticle accumulation present in ‘Moneymaker’ probably prevented a transient stage of high density and stiffness, despite the small increase in density observed in ‘Moneymaker’ at 25 daa. Waxes, mostly intracuticular ones, have been shown to confer stiffness to plant cuticles (Petracek & Bukovac, 1995; Takahashi et al., 2012; Khanal et al., 2013) and could also participate in this temporal increase. Wax accumulation in ‘Cascada’ reached its maximum at 20 daa (Domínguez et al., 2008), but removal of waxes reduced Young's modulus similarly at 20 and 25 daa (c. 30%; data not shown), thus maintaining the differences in stiffness observed with waxes.

Cuticle biomechanics during fruit ripening: flavonoid accumulation and cutin depolymerization

Changes in tomato cuticle biomechanics during ripening have already been reported, such as an increase in stiffness and a decrease in maximum strain (Bargel & Neinhuis, 2004; López-Casado et al., 2007), and were attributed to the incorporation of flavonoids in the tomato fruit cuticle (Domínguez et al., 2009), supporting the hypothesis suggested by Bargel et al. (2006) that phenolic compounds could be related to a rigid cutin matrix at full maturity. Our results support this increase in strength and decrease in deformability. The correlation in both genotypes of cuticle stiffness with the amount of phenolic compounds present in the cuticle clearly shows their involvement in providing rigidity to the cuticle once cell walls start their degradation during ripening. Moreover, the higher stiffness during ripening can be explained as an increment in the stress needed to achieve the same elastic deformation. This increase in the length of the elastic phase is responsible for the higher breaking stress, as the length of the viscoelastic phase did not change during ripening. The decrease in the maximum strain is then attributable to a reduced viscoelastic deformation. Thus, flavonoids play two roles: one, providing strength during the elastic phase and two, reducing irreversible viscoelastic deformation. In other words, flavonoids allow the cuticle to withstand transient changes in fruit volume with little expansion. In fact, Domínguez et al. (2012) observed a significant decrease in tomato fruit growth rate during ripening.

Some differences in the mechanical responses of ‘Moneymaker’ and ‘Cascada’ cuticles were observed in this period. Flavonoid accumulation in ‘Moneymaker’ was not associated with an increase in breaking stress. In this case, there was an increase in the yield stress but the viscoelastic phase was clearly shorter. Similarly, the decrease in the viscoelastic deformation was more pronounced than in ‘Cascada’. These differences could be related to differences in the changes in cuticle density observed in the two genotypes. Whereas in ‘Cascada’ density increased during ripening and could contribute to the strengthening of the cuticle, in ‘Moneymaker’ the significant decrease in density would promote the opposite effect: lowering the breaking stress despite flavonoid accumulation.

The flavonoids naringenin and chalconaringenin are the main phenolic constituents present in the tomato fruit cuticle at red ripe, albeit some phenolic acids such as p-coumaric and ferulic acids have also been reported in low quantities (Hunt & Baker, 1980; Laguna et al., 1999). In the ‘Cascada’ cuticle, p-coumaric and p-hydroxybenzoic acids were present throughout fruit development and flavonoids only started their accumulation at 45 daa. p-hydroxybenzoic acid has been reported to be a minor component of both angiosperm and gymnosperm primary cell walls (Hartley & Harris, 1981; Carnachan & Harris, 2000; El Modafar & El Boustani, 2001). Hence, the gradual increase in stiffness between 35 and 45 daa could be attributable to the subsequent increase in p-hydroxybenzoic acid that occurred during this period in the cutin matrix. According to the literature, phenolic acids esterified to cell walls contribute to their mechanical properties, conferring stiffness and resistance (Peyron et al., 2001; Antoine et al., 2003). However, the maximum increase in cuticle stiffness and breaking stress occurred between breaker and red ripe, where p-hydroxybenzoic acid decreased and flavonoids accumulated. This suggests that the flavonoids naringenin flavanone and chalcone are the major contributors to the almost 2.5-fold increase in phenolics observed during this period.

Significant changes in the cutin matrix were observed during ripening. They corresponded to a decrease in the overall number of ester bonds and an increase in hydroxyl and carboxylic acid functional groups, which implies a rearrangement of the polymer chains into a different conformation with fewer ester or strong H bonds. These results indicate a depolymerization of the cutin matrix that did not reduce the mechanical strength of the cuticle. On the contrary, these changes were associated with an increase in cuticle stiffness, probably as a result of the role of phenolics as fillers restraining the mobility of cutin chains. A similar role as fillers has recently been reported for triterpenoids in the persimmon fruit (Diospyros kaki) cuticle (Tsubaki et al., 2013). Moreover, the ATR-FTIR spectroscopic profile showed a heterogeneous distribution of phenolic compounds within the cuticle. Thus, the highest amount of phenolics was present in the cutin, followed by the cuticle outer surface, and the lowest amount was detected in the cuticle inner surface. This heterogeneous allocation indicates that phenolics mainly interact and modify the cutin matrix and very little the polysaccharides derived from the cell wall. The significant correlation observed between the γ(C-H) aromatic band, the amount of phenolics (mostly flavonoids), and the degree of cutin esterification suggests that this band can be used to monitor these changes in the cutin matrix.

Ripening-related changes in the cuticle showed a co-regulation of flavonoid incorporation with depolymerization of the cutin matrix and modification of cuticle density. It would be of great interest to ascertain if these changes are regulated by ripening genes or are a direct/indirect consequence of flavonoid incorporation. The study of transgenic plants that do not accumulate flavonoids would be a first step towards identifying the regulatory pathway and eventually the different genes involved and their contribution to cuticle mechanics. In conclusion, cuticles from fruits at the initial stages of development are mechanically weak and only exhibit elastic deformation. Once a significant amount of cutin has been accumulated, viscoelastic behaviour appears and the mechanical behaviour of the cuticle remains mostly unaltered during the fruit expansion period. Ripening is accompanied by several changes in the cuticle, such as chemical cleavage of cutin ester bonds and incorporation of phenolic compounds. These changes modify both the elastic and viscoelastic behaviour of the cuticle which becomes much stiffer and less deformable.

Acknowledgements

This work was supported in part by grants AGL2009-12134 and AGL2012-32613 from the Plan Nacional de I+D, Ministerio de Educación y Ciencia, Spain.

Ancillary