Induced mutations in tomato SlExp1 alter cell wall metabolism and delay fruit softening

Accepted Manuscript Title: Induced mutations in tomato SlExp1 alter cell wall metabolism and delay fruit softening Author: Silvia Minoia Adnane Boualem Fabien Marcel Christelle Troadec Bernard Quemener Francesco Cellini Angelo Petrozza Jacqueline Vigouroux Marc Lahaye Filomena Carriero Abdelhafid Bendahmane PII: DOI: Reference:

S0168-9452(15)30008-X http://dx.doi.org/doi:10.1016/j.plantsci.2015.07.001 PSL 9225

To appear in:

Plant Science

Received date: Revised date: Accepted date:

30-4-2015 29-6-2015 2-7-2015

Please cite this article as: Silvia Minoia, Adnane Boualem, Fabien Marcel, Christelle Troadec, Bernard Quemener, Francesco Cellini, Angelo Petrozza, Jacqueline Vigouroux, Marc Lahaye, Filomena Carriero, Abdelhafid Bendahmane, Induced mutations in tomato SlExp1 alter cell wall metabolism and delay fruit softening, Plant Science http://dx.doi.org/10.1016/j.plantsci.2015.07.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Induced mutations in tomato SlExp1 alter cell wall metabolism and delay fruit softening.

Silvia Minoiaa,c, Adnane Boualema, Fabien Marcela, Christelle Troadeca, Bernard Quemenerb, Francesco Cellinic, Angelo Petrozzac, Jacqueline Vigourouxb, Marc Lahayeb, Filomena Carrieroc, and Abdelhafid Bendahmanea§

a

INRA, UMR1403, IPS2, CNRS-UMR 9213, Université Paris-Sud, Université d'Evry,

Université Paris-Diderot, Sorbonne Paris-Cité, 2 rue Gaston Crémieux, 91057 Evry, France. b

INRA, UR1268 Biopolymers, Interactions and Assemblies, rue de la Géraudière, F-44316

Nantes, France. c

ALSIA, Centro Ricerche Metapontum Agrobios, SS Jonica 106 Km 448.2, 75012 Metaponto

(MT), Italy.

Email addresses: Silvia Minoia: [email protected] Adnane Boualem: [email protected] Fabien Marcel: [email protected] Christelle Troadec: [email protected] Bernard Quemener: [email protected] Francesco Cellini: [email protected] Angelo Petrozza: [email protected] Jacqueline Vigouroux: [email protected] Marc Lahaye: [email protected] Filomena Carriero: [email protected] Abdelhafid Bendahmane: [email protected]

§

to whom correspondence should be addressed : [email protected]

Highlights

• TILLING mutants were isolated in SlExp1 gene • Slexp1-6_W211S and Slexp1-7_Q213Stop mutants affect the polysaccharide-binding site • Mutations in SlExp1 gene increase fruit firmness and enhance fruit shelf-life. • Loss of SlExp1 activity modifies cell wall polysaccharide chemical composition. Abstract Fruit ripening and softening are key traits for many fleshy fruit. Since cell walls play a key role in the softening process, expansins have been investigated to control fruit over ripening and deterioration. In tomato, expression of Expansin 1 gene, SlExp1, during fruit ripening was associated with fruit softening. To engineer tomato plants with long shelf life, we screened for mutant plants impaired in SlExp1 function. Characterization of two induced mutations, Slexp1-6_W211S, and Slexp1-7_Q213Stop, showed that SlExp1 loss of function leads to enhanced fruit firmness and delayed fruit ripening. Analysis of cell wall polysaccharide composition of Slexp1-7_Q213Stop mutant pointed out significant differences for uronic acid, neutral sugar and total sugar contents. Hemicelluloses chemistry analysis by endo-β-1,4-Dglucanase hydrolysis and MALDI-TOF spectrometry revealed that xyloglucan structures were affected in the fruit pericarp of Slexp1-7_Q213Stop mutant. Altogether these results demonstrated that SlExp1 loss of function mutants yield firmer and late ripening fruits through modification of hemicellulose structure. These SlExp1 mutants represent good tools for breeding long shelf life tomato lines with contrasted fruit texture as well as for the understanding of the cell wall polysaccharide assembly dynamics in fleshy fruits. Keywords

Tomato, expansin gene, fruit ripening, fruit firmness, long shelf life, TILLING, cell wall structure 1. Introduction

Agriculture in the 21st century faces multiple challenges among which improving productivity and reducing food loss and food waste, to feed over 9.1 billion people by 2050 [1]. It is estimated that about 32 percent of all food produced in the world is lost or wasted [2]. Tomato (Solanum lycopersicum L.) is the second most consumed vegetable crop in the world, after potato, and a model species to investigate fruit maturation and shelf life [3-5]. Tomato fruits go through a development stage known as ripening. This occurs when the fruit has finished its growth. Ripeness is followed by fruit softening that leads to deterioration of the fruit. Both fruit ripening and softening are major attributes that contribute to fruit perishability. Excessive fruit softening also leads to changes in fruit quality and causes significant losses to farmers, retailers and consumers. Even though ripening can be delayed through several external procedures, once initiated, the physiological changes associated with ripening are irreversible and lead to alterations that are detrimental to long storage [6,7].

One of the tomato breeding challenges is to deliver to consumers, tomato fruits that are desirable in terms of taste, texture, and color, with delayed post-harvest ripening-related softening. Different strategies have been employed to extend the shelf life of tomato fruits. One major approach has been the manipulation of ethylene biosynthesis which is the key signal of the onset of ripening in tomato [4]. Several mutants affected in ethylene production or perception were identified and used in tomato breeding to delay softening [5]. A number of studies have also focused on the manipulation of the polyamine or anthocyanin pathways because the accumulation of antioxidants has been proved to extended tomato shelf life [8,9].

An alternative pathway to control fruit softening is to control cell wall integrity. During ripening, the cell wall architecture that consists of rigid, inextensible cellulose microfibrils held together by networks of matrix glycans, pectins and structural glycoproteins, is progressively modified [10]. However, for many of these investigations, in spite of their contribution in delaying fruit softening, they often had deleterious consequences on fruit quality traits such as flavor, texture and aroma [8]. Quantitative quality trait loci, such as firmness, elasticity, color, fruit weight, diameter, meltiness, and aroma volatiles were also identified and introgressed in elite lines [11,12].

Expansins are plant cell-wall loosening proteins involved in cell enlargement and in a variety of other developmental processes in which cell-wall modification occurs [13,14]. Expansins are thought to loosen cell wall hemicellulose-cellulose interactions, with as consequence an improvement of cellulose microfibril slippage, as cell expand during organ development. They also favor cell wall polysaccharide degradation by enzymes, likely by a pore-opening mechanism resulting from the modifications of the hemicellulose-cellulose assembly.

During fruit ripening in tomato, a highly expressed expansin gene, SlExp1, was shown to contribute to fruit softening [15]. Suppression of SlExp1 expression reduced fruit softening in the early ripening stages. SlExp1-silenced fruits also displayed limited cell wall pectin degradation without any effect on other cell wall matrix polysaccharides [16]. In contrast, over-expression of SlExp1 resulted in an increase of fruit softening and in the degradation of matrix glycans including xyloglucans in green fruits [16]. Suppression of both SlExp1 and the endo-polygalacturonase

gene,

SlPG,

coding

the

major

enzyme

degrading

the

homogalacturonan domains of pectin during tomato softening, further improved fruit firmness during ripening [17].

The role of expansins on cell wall elasticity, through the modulation of hemicellulosecellulose interactions, is supported by studies on synthetic cell walls [18] and isolated tomato cells [19]. At the fruit scale, softening results from a complex combination of several factors such as modifications of cell wall polysaccharide assemblies [15] and cell-cell debonding [20], with possible tissue reorganization under stress and reduction of turgor pressure [21,22].

Improving fruit shelf life is still a major target for tomato breeding. However, breeders have only access to few natural alleles in genes implicated in fruit ripening and softening. Engineering new alleles in different components of the fruit softening pathways will likely extend breeder’s tool box to improve tomato shelf life. Modulation of SlExp1 activity in tomato cultivar could be a convenient mean to control post-harvest fruit softening and to improve organoleptic fruit quality. As an alternative to genetically modified tomato [10,16], we screened for induced mutations in SlExp1 gene, in Ethyl methanesulfonate (EMS)mutagenised tomato collections of 9980 M2 plants, using the TILLING concept [23-25]. The phenotypic analysis of the induced SlExp1 alleles led to the identification of two Exp1 variants that enhance fruit shelf-life and delay fruit softening. These novel Exp1 alleles could be useful in breeding programs in combination with other alleles in the antioxidant or ethylene pathway aiming at developing new tomato commercial lines improved in fruit firmness as well as to investigate cell wall polysaccharides metabolism of fleshy fruits.

2. Materials and methods

2.1. TILLING screenings

EMS mutant populations generated in the genetic background of two processing tomato varieties M82 [25] and Red Setter [24] were used for TILLING screens. PCR amplification was based on nested-PCR and was carried out using two couples of SlExp1 target-specific primers. 4 ng of eight-fold pooled genomic DNA were used for the first PCR with the primer For1 (5’AGG AAG AAT CCC TGG TGT TTA CTC T3’) and Rev1 (5’TAT CTG TAT TTT CAG TGA GGA CTC G3’). Internal primers For2 (5’TTA TAC AGC CAA GGA TAC GGA GTT A3’) and Rev2 (5’CCT AAG GTG AAC AAC ACT CTG AAA T3’), 5’-end labelled with IRDye 700 and IRDye 800 dye (LI-COR®, Lincoln, NE, USA) respectively, were employed in the second PCR in the conditions described in [26]. Mutation detection was performed with the mismatch specific endonuclease ENDO I [27] and the LI-COR 4300 DNA analyzer (LI-COR®, Lincoln, NE, USA). Gel images were analyzed using Adobe Photoshop software (Adobe Systems Inc., San José, CA, USA). Finally, identified mutations were validated by Sanger sequencing.

2.2. Selection of mutant lines

Two mutant lines, Slexp1-6 and Slexp1-7, were selected to perform detailed firmness analyses. For each mutant, M3 seeds were planted and grown in greenhouse conditions under standard cultivation conditions and genotyped to assess the mutation allelic condition. Plants homozygous for the mutated allele were isolated. Plants homozygous for the “wild type” allele were used as strict control of homozygous mutant allele since it’s assumed that, within the same M3 family, they share most of the background mutations but lack the mutation of interest. Parental lines of the EMS populations, M82 and Red Setter lines, were also used as control.

2.3. Measure of fruit firmness

For each selected genotype, roughly 10 M3 plants were grown in greenhouse conditions. At the breaker stage (B0), tomato fruit were collected, stored in continuous light at 19°C and analyzed in post-harvest condition at different ripening stages: breaker (B0), breaker + 5 days (B+5), breaker + 10 days (B+10), breaker + 15 days (B+15) and breaker + 25 days (B+25). Firmness analysis was performed on whole tomato fruit without peel by using a pressure tester (Forge Gauge, Lutron GF-5000-A) fitted with a cylindrical plunger of 6 mm in diameter. For each ripening stage, seven to nine fruits were analysed and three measures were taken on the equatorial zone for each tomato fruit. Data were statistically treated by t-test. We considered significant only differences with p < 0.05.

2.4. Analysis of histological and biochemical properties of pericarp tissue

The histological and biochemical analyses were only performed on fruits collected from M4 plants at mature green stage and analyzed at mature green (G) stage, breaker (B), breaker + 5 days (pink, P), breaker + 10 days (red, R) and breaker + 25 days (over-ripe, S).

2.4.1. Pericarp tissue histology

For each of five tomato fruits at the mature green stage, two pieces of pericarp tissue of about 1cm3 were sampled at the equatorial region opposite to each other. Each sample was divided into two pieces from the external to the internal epidermis. One 200 µm thick transverse section was obtained per piece using a vibrating blade microtome (HM 650 V, Microm

International GmbH, Walldrof, Germany). Images (total of four images per fruit) were acquired using a tri CCD Colour video camera (JVC KY F55B) equipped with a 50 mm Nikon (AF Nikkor, 1:1.8) lens and 12mm extension tube. Cell size distribution in the pericarp sections of the different genotypes was established by closing image analysis using a square structuring element according to [28].

2.4.2. Cell wall polysaccharides chemical composition

Sampling : Five to eight pericarp tissue cylinders were sampled with a cork borer of 0.96 cm diameter from an about 1 cm thick equatorial slice of five fruits per genotype and ripening stage. Samples of pericarp tissue were frozen in liquid nitrogen and then stored at -20 °C prior to biochemical analyses.

Cell wall preparation : Frozen samples were freeze-dried, ground to a powder and extracted by 70% ethanol at 100 °C and 100 bars (ASE 200, Dionex, Synnuvale, CA, USA) to inactivate endogenous enzymes and remove free sugars until no detectable color was observed in the extract by the phenolsulfuric acid assay [29]. The resulting cell wall preparation is referred to as alcohol insoluble material (AIM). Cell wall polysaccharides composition : All measurements were performed on dry samples. Identification and quantification of neutral sugars by gas–liquid chromatography (GC) and quantification of uronic acid were performed after sulfuric acid degradation as described [30]. AIM was dispersed in 13 M sulfuric acid for 30 min at 25 °C and then hydrolyzed in 1 M sulfuric acid (2 h, 100 ° C). Sugars were

converted to alditol acetates and chromatographed on a DB 225 capillary column (J&W Scientific, Folsorn, CA, USA; temperature 205 ° C, carrier gas H2). For calibration, standard sugars solution and inositol as internal standard were used. Uronic acids in acid hydrolyzates were quantified using the metahydroxydiphenyl colorimetric method [31]. Data were corrected for residual starch determined after amylolysis [32] and quantification of the released glucose by HPAEC (PA1 analytical column, 250 x 4 mm, Dionex). The degree of pectin-esterification by methanol (DM) and weight % of acetic acid ester (AA) was measured by HPLC after alkaline hydrolysis [33].

Fine structure of hemicelluloses : Five mg of cell wall preparations were hydrolyzed by endo-β-D-1,4-glucanase (Trichoderma longibrachiatum, Megazyme, Bray, Ireland) as described [30]. Oligosaccharides were identified by MALDI-TOF MS on M@ldi LR (Waters) spectrometer using DHB/6ATT matrix. The matrix used was a mixture (V/V) of 2-5 dihydroxybenzoic acid (DHB) at 21mg /ml and of 6-Aza-2-thiothymine (6ATT) at 12 mg/ml in 80% of CH3CN containing 2µl of TFA 10% in water [34]. A mixture (1 µL, V/V) of hydrolysate and of matrix was deposited in duplicates on the MS probe. The analysis was performed in the positive ion mode on a Micro mass M@LDI-TOF LR mass spectrometer (Waters, USA) equipped with a 337 nm nitrogen laser and delayed extraction. Analyses were carried out in the reflector mode using a laser frequency of 5Hz, an accelerating voltage of 15kV and a delay time of 250 ns. A total of 10 spectra were recorded in the mass range m/z 500-2000 per genotype at a given ripening stage. The instrument was externally calibrated using the mono-isotopic masses of typical xyloglucan oligosaccharides from the laboratory collection (XXG of m/z 791.243; XXXG of m/z 1085.338; mono-acetylated XXFG of m/z 1435.459; mono-acetylated XLFG of m/z 1597.512).

Oligosaccharide nomenclature followed that in the literature for xyloglucans [Supplementary Table 1; 35]. In brief, it uses uppercase letters representing an individual 1−>4 linked β-Dglucose residue and its pendant side chains. Accordingly, bare glucose residue is designated by the letter G while when branched by a α-D-xylosyl residue on O-6, it is refers to X. With further extension of the branch by one β-D-galactosyl, by one α-L-arabinosyl residue or by the disaccharide α-L-arabinosyl (1->3) α-L-arabinoside linked on xylose O-2, the structure is then referred to by the letters L, S and T, respectively. When acetyl-esterification occurs, the coding is extended by the small letter a and by a number corresponding to the total of group present. Other oligosaccharides based on hexoses or pentoses are referred to as by the code Hex or Pen, respectively, followed by the number of specific sugar and ended, when required, by the letter code a and the number of acetyl groups. Substitutions by mono-methylated- and non-methylated uronic acid are referred to as m and u followed by the respective number of these groups. Data treatment and statistic (ANOVA, pairwise t-test) were performed using R software version 2.7.2 software (The R Foundation for Statistical Computing; http://www.Rproject.org) or MATLAB (histology). When not mentioned in the text significant differences were set at p < 0.05.

3. Results

3.1. Identification of SlExp1 alleles in the polysaccharide-binding site Expansins comprise a large plant multigene families [14,36-38], mainly involved in cell-wall loosening, leading to cell enlargement [13,16,39,40]. In tomato, Expansin1 gene (SlExp1, Solyc06g051800) is the main expansin family member highly expressed during fruit ripening [16,41]. To get new insights into the role of SlExp1 in the softening processes during fruit ripening, we screened for induced mutations in SlExp1 in two tomato EMS mutant

collections. In total, we screened 9980 M3 families, among which 4759 and 5221 M3 families were in M82 and Red Setter background, respectively [24,25] . As Expansins belong to a large multigene family, we designed gene-specific primers that are able to discriminate SlExp1 from other members of the gene family. The selected primers amplified a DNA fragment of 1025 bp consisting of 342 bp of intronic region and 683 bp of coding sequence (Fig. 1A). In this TILLING screen, we identified 28 independent point mutations among which 19 were intronic and 9 were exonic. Out of the 9 exonic mutations, 3 were silent, 5 were missense and one induced a premature stop codon at the 213th amino acid position leading to a truncated SlExp1 protein (Table 1, Fig. 1A).

Among the missense mutations, the L145F (Slexp1-1), P154L (Slexp1-4), V197L (Slexp1-5) and S238F (Slexp1-9) changes occurred in non-conserved amino acid positions and X-ray crystallography studies did not specify any role for those residues (Fig. 1B) [42]. Therefore L145F, P154L, V197L and S238F mutations are predicted to not affect the function of the protein. In contrast, the W211S (Slexp1-6) mutation affects a highly conserved residue and in the three-dimensional cristal structure, W211 residue is located in the D2 functional domain and is one of the residues that form the polysaccharide-binding site (Fig. 1B) [42]. Therefore, W211S mutation is predicted to impair the protein function. 3.2. Mutations in SlExp1 gene increase fruit firmness

To test whether induced mutations in SlExp1 could affect fruit softening, the Slexp1-6 (W211S) and Slexp1-7 (Q213Stop) mutations, predicted to impair the protein function, were analyzed for firmness. As Slexp1-6 and Slexp1-7 derive from Red Setter and M82 mutant collections, sister plants segregating for the induced mutations (homozygous WT) and Red

Setter and M82 parental lines were used as control (Fig. 2). As expected, for all the studied genotypes, fruit firmness decreases progressively during fruit ripening. At breaker stage (B0), the highest firmness values were measured for the Slexp1-6 and Slexp1-7 mutant lines with, respectively, 41% and 46% fruit firmness enhancement as compared to the control plants (Fig. 2). The improved fruit firmness observed at B0 in Slexp1-6 and Slexp1-7 was maintained at all ripening stages. At the overripe stage, twenty five days after breaker (B+25), Slexp1-6 and Slexp1-7 mutant lines still have a significantly higher fruit firmness compared to the relative WT and parental control lines (Fig. 2). From this we concluded that Slexp1-6 and Slexp1-7 have enhanced fruit firmness during ripening. 3.3. Loss of SlExp1 activity modifies the cell wall polysaccharide chemical composition

To determine the cell wall modifications that may explain changes observed for fruit firmness, first histological analyses were performed in the pericarp tissue of the Slexp1-7 and Slexp1-6 mutants at different ripening stages. Pericarp thickness and cell size distribution were measured on mature green and breaker fruits by image analysis [28]. In this analysis no significant difference between the Slexp1-7, Slexp1-6 mutants and the WT control was detected (data not shown).

Cell wall polysaccharide composition and molar proportion of individual sugars were characterized and are within values reported in the literature (Table 2 and Supplementary Table 2) [43]. Typical losses of galactose and arabinose proportions were observed during fruit ripening. Comparisons of Slexp1-7 and WT fruits pointed out significant differences for uronic acids, neutral sugars and total sugar content. At the green stage, Slexp1-7 mutant fruits have lower uronic acids content whereas at the breaker stage, neutral sugars and total sugar contents were higher than the wild-type. Moreover, in the Slexp1-7 mutant fruits, a significant

reduction of the methyl esterification of pectin was observed at the overripe stage (Table 2). However, these differences were not observed on the basis of individual sugar molar proportions (Supplementary Table 2), specifying that they arose from an overall difference in polysaccharide content in the cell wall preparation (ie less proteins) rather than from specific modifications of the polysaccharide sugar composition.

Variations in hemicellulose chemistry and cell wall interactions were assessed by endo-β-1,4-D-glucanase hydrolysis. Degradation products were taken as representative of the effect of the fruit ripening process and the Slexp1-7 mutation on the fine structure and accessibility of hemicelluloses to the endo-β-1,4-D-glucanase. MALDI-TOF spectra of the glucanase hydrolyzate were as described in Supplementary Fig. 1 [43] and showed the presence of 4 series of oligosaccharides identified on the basis of their mass as sodium adducts (M + Na)+ and literature data [44-48]. A major series corresponded to xyloglucans with major ions attributed to XSGG, XSGGa1, XXGGa1 and moderate or low contribution of SGG, SGGa1, XXGG, XLGGa1 and LSGGa1 structures. Two ions could not be distinguished between XTGG and Hex7a1 and XTGGa1 and Hex7a2. The second major series corresponded to acetylated hexoses structures (Hexa) attributed to glucomannans oligomers released by the glucanase. A third series of ions corresponded to structures based on non-acetylated hexoses and was attributed to pectic galactan side chains. Finally the ion of mass attributed to Pen5u2 resulted from the glucuronoxylan. The impact of fruit ripening process and Slexp1-7 mutation on the intensities of the most common MALDI-TOF MS peaks is reported on Fig. 3. For the WT fruits, the intensity of oligosaccharides attributed to pectic galactan oligosaccharides (Hex) markedly decreased during the ripening process (i.e. Hex5 as a representative oligomer on Fig. 3). Mono or diacetylated Hex structures attributed to glucomannan oligomers increased up to the

breaker/pink stage and decreased thereafter. Among xyloglucan structures SGGa1, XXGGa1 and to a lesser extent, XXGG, LSGGa1 were also affected by ripening (Fig. 3).

In the Slexp1-7 mutant fruits, pectic galactan oligosaccharides (Hex) content is similar to the oligosaccharides content in the WT fruits and decrease during fruit ripening. In contrast, xyloglucan structures were affected in fruits of the Slexp1-7 loss of function mutant. Particularly SGGa1 ion intensity was higher in the Slexp1-7 mutant starting at the breaker stage and significantly different from WT at the pink and red stages (Fig. 3). Other significant differences but less marked were also noted at the mature green stage for SGG and XLGGa1, and at the overripe stage with XXGGa1. LSGGa1 tended to be also affected at the pink and red stages, though not statistically significant (Fig. 3).

Discussion

In this study, new allelic variants of SlExp1gene were isolated trough TILLING by screening two tomato EMS mutant collections generated in the M82 and Red Setter genetic background. Nine induced point mutations were identified in the SlExp1 coding region. Based on the predicted impact on SlExp1 enzyme activity two missense mutations, Slexp1-6_W211S and Slexp1-7_Q213Stop, were selected for detailed phenotypic analysis of the fruit firmness and the cell wall polysaccharide composition, during ripening. Fruits of the Slexp1-6 and Slexp1-7 mutants were firmer than the controls and firmness was increased at all ripening stages including the overripe stage (B+25). The enhanced firmness delays fruit softening and extends fruit shelf-life. The observed fruit firmness enhancement in Slexp1-6 and Slexp1-7 mutants confirmed previous results reported in transgenic suppression of Exp1 [16]. However, in [16] the relative fruit firmness enhancement was in a range of 13% to 23%, likely due to

incomplete co-suppression of the SlExp1 gene. In the SlExp1 TILLING mutants, we observed 41% to 46% fruit firmness increase. To bring new insight on how SlExp1 loss-of-function mutant improves fruit firmness, we analysed the cell wall of the mutant and the wild type plants during ripening. Although fleshy fruit mechanical characteristics can be affected by histological parameters [43,49-51], pericarp tissue thickness and cell size distributions did not differ between Slexp1-6 and Slexp1-7 mutants and the controls at both mature green and breaker stages. In the light of the synergistic effect of expansin and polysaccharide hydrolases on the fruit softening related cell wall deconstruction, the origin of the mechanical alteration was searched in the kinetics of cell wall polysaccharide chemical modifications during fruit ripening. Although suppression of tomato expansin was shown to particularly limit degradation of chelator soluble pectin [16], in the present work the polysaccharide composition was not affected in the Slexp1-6 and Slexp1-7 mutants and the pectic galactan and arabinan side chain metabolism related to ripening occurred similarly both in Slexp1-6 and Slexp1-7 mutants and wild type genotypes. However, reduction in methyl esterification of pectins in ripe or overripe mutant fruits reveals an unclear mechanism of methyl-esterification/de-esterification that needs to be further substantiated. As expansin is thought to be primarily targeted on xyloglucan-cellulose interactions and particularly those in the “biomechanical hotspots” [52], a particular focus was made on hemicellulose chemistry. Ripening is known to partly depolymerize xyloglucan [16,53,54] and affect the fine structure or accessibility of tomato fruit xyloglucan, glucomannan and xylan to endo-β-1,4-D-glucanase [43]. The profile of the oligosaccharides structure released in the hydrolyzate reflects the complex metabolic processes involved in tomato fruit ripening. The latter include both biosynthesis of xyloglucan [55] and catabolism of cell wall polysaccharides including xyloglucan reshuffling by xyloglucan transglucosylases/hydrolases

(XTH) [56,57]. In this work, ripening and genetic related hemicelluloses structural modifications were also observed. The SlExp1 loss-of-function in the Slexp1-7 mutant was hypothesized to induce modifications in the interaction of xyloglucans to cellulose and thus, in their ability to be hydrolyzed by exogenous endo-β-1,4-D-glucanase [58]. The glucanase hydrolyzates were assessed by MALDI-TOF MS analysis, a technique which was successful in the identification of plant variants for cell wall xyloglucan structures [59-62]. Although great care must be taken when interpreting MS ions intensity with regard to the impact of sample preparation on ionization [63], variations in the normalized intensities relative to XSGG (100%) were taken as representative of differences in oligosaccharides concentration in the hydrolyzate. Considering a lower efficiency in the xyloglucan-cellulose interaction breaking capacity in the Slexp1-7 mutant, differences in MALDI-TOF MS profiles with WT were expected to reveal structures closely involved in cellulose interactions. Only xyloglucan structures in the glucanase hydrolyzate were affected in the impaired Slexp1-7 mutant. In the Slexp1-7 mutant line, the mutation had a remarkable impact on the proportion of the SGGa1 structure starting at the breaker stage and persisting until the overripe stage. The modification of xyloglucan structure occurring at the known maximum gene expression affecting firmness suggests a link between mutation, texture and hemicellulose fine structure. The overall limited modification of xyloglucan structure is in keeping with the observation that down regulated expansin in tomato had limited impact on its depolymerization [16]. Further work should evaluate if the SGGa1 xyloglucan structure in tomato fruit cell wall glucanase hydrolyzate represents markers of expansin activity and should elucidate their function in the xyloglucan-cellulose interactions and/or in the regulation of cell wall enzymes activities, such as XTH. Nevertheless, the Slexp1-7 mutant line shows ripening related alteration in the softening process and represent good tool for breeding program aiming at developing texture contrasted

lines of tomato as well as for basic studies on the mechanisms of cell wall polysaccharides assembly dynamics and its consequence on the mechanical properties of fleshy fruits. Acknowledgements We thank C. Clepet and V. Gomez for discussions and comments on the manuscript. We thank people of the BIBS platform at INRA Nantes for the help on mass spectrometry experiments. This work was supported by the BAP (Biology and Plant Breeding) department of INRA, the Italian Ministry of University and Research (MIUR, ITALYCO Project, DD n° 603/RIC), the European project EU-SOL (Contract number: FOOD-CT-2006-016214), the SPS Labex (Saclay Plants Sciences), the IDEX Paris-Saclay (project 3P) and by a Marie Curie fellowship. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. References [1] J. Parfitt, M. Barthel, S. Macnaughton, Food waste within food supply chains: quantification and potential for change to 2050, Philos Trans R Soc Lond B Biol Sci, 365 (2010) 3065-3081. [2] FAO World Bank, Reducing post-harvest losses in grain supply chains in Africa., Report of FAO–World Bank workshop (2010) 120p. [3] R. Alba, P. Payton, Z. Fei, R. McQuinn, P. Debbie, G.B. Martin, S.D. Tanksley, J.J. Giovannoni, Transcriptome and selected metabolite analyses reveal multiple points of ethylene control during tomato fruit development, Plant Cell, 17 (2005) 2954-2965. [4] J. Giovannoni, Molecular Biology of Fruit Maturation and Ripening, Annual review of plant physiology and plant molecular biology, 52 (2001) 725-749. [5] J.J. Giovannoni, Fruit ripening mutants yield insights into ripening control, Curr Opin Plant Biol, 10 (2007) 283-289.

[6] V. Prasanna, T.N. Prabha, R.N. Tharanathan, Fruit ripening phenomena--an overview, Crit Rev Food Sci Nutr, 47 (2007) 1-19. [7] D. Martinez-Romero, G. Bailen, M. Serrano, F. Guillen, J.M. Valverde, P. Zapata, S. Castillo, D. Valero, Tools to maintain postharvest fruit and vegetable quality through the inhibition of ethylene action: a review, Crit Rev Food Sci Nutr, 47 (2007) 543-560. [8] L. Bassolino, Y. Zhang, H.J. Schoonbeek, C. Kiferle, P. Perata, C. Martin, Accumulation of anthocyanins in tomato skin extends shelf life, The New phytologist, 200 (2013) 650-655. [9] Y. Zhang, E. Butelli, R. De Stefano, H.J. Schoonbeek, A. Magusin, C. Pagliarani, N. Wellner, L. Hill, D. Orzaez, A. Granell, J.D. Jones, C. Martin, Anthocyanins double the shelf life of tomatoes by delaying overripening and reducing susceptibility to gray mold, Current biology : CB, 23 (2013) 1094-1100. [10] D.A. Brummell, M.H. Harpster, Cell wall metabolism in fruit softening and quality and its manipulation in transgenic plants, Plant molecular biology, 47 (2001) 311-340. [11] M. Causse, V. Saliba-Colombani, L. Lecomte, P. Duffe, P. Rousselle, M. Buret, QTL analysis of fruit quality in fresh market tomato: a few chromosome regions control the variation of sensory and instrumental traits, J Exp Bot, 53 (2002) 2089-2098. [12] J.J. Giovannoni, Breeding new life into plant metabolism, Nature biotechnology, 24 (2006) 418-419. [13] D.J. Cosgrove, Loosening of plant cell walls by expansins, Nature, 407 (2000) 321-326. [14] J. Sampedro, D.J. Cosgrove, The expansin superfamily, Genome Biol, 6 (2005) 242. [15] D.A. Brummell, Cell wall disassembly in ripening fruit., Functional Plant Biology, 33 (2006) 103-119. [16] D.A. Brummell, M.H. Harpster, P.M. Civello, J.M. Palys, A.B. Bennett, P. Dunsmuir, Modification of expansin protein abundance in tomato fruit alters softening and cell wall polymer metabolism during ripening, Plant Cell, 11 (1999) 2203-2216.

[17] A.L. Powell, M.S. Kalamaki, P.A. Kurien, S. Gurrieri, A.B. Bennett, Simultaneous transgenic suppression of LePG and LeExp1 influences fruit texture and juice viscosity in a fresh market tomato variety, J Agric Food Chem, 51 (2003) 7450-7455. [18] S.E. Whitney, M.J. Gidley, S.J. McQueen-Mason, Probing expansin action using cellulose/hemicellulose composites, Plant J, 22 (2000) 327-334. [19] C.X. Wang, L. Wang, S.J. McQueen-Mason, J. Pritchard, C.R. Thomas, pH and expansin action on single suspension-cultured tomato (Lycopersicon esculentum) cells, J Plant Res, 121 (2008) 527-534. [20] M.C. Jarvis, S.P.H. Briggs, J.P. Knox, Intercellular adhesion and cell separation in plants., Plant, Cell and Environment, 26 (2003) 977-989. [21] K.A. Shackel, C. Greve, J.M. Labavitch, H. Ahmadi, Cell turgor changes associated with ripening in tomato pericarp tissue, Plant Physiol, 97 (1991) 814-816. [22] W. Wei, C. Yang, J. Luo, C. Lu, Y. Wu, S. Yuan, Synergism between cucumber alphaexpansin, fungal endoglucanase and pectin lyase, J Plant Physiol, 167 (2010) 1204-1210. [23] N. Menda, Y. Semel, D. Peled, Y. Eshed, D. Zamir, In silico screening of a saturated mutation library of tomato, Plant J, 38 (2004) 861-872. [24] S. Minoia, A. Petrozza, O. D'Onofrio, F. Piron, G. Mosca, G. Sozio, F. Cellini, A. Bendahmane, F. Carriero, A new mutant genetic resource for tomato crop improvement by TILLING technology, BMC Res Notes, 3 (2010) 69. [25] F. Piron, M. Nicolai, S. Minoia, E. Piednoir, A. Moretti, A. Salgues, D. Zamir, C. Caranta, A. Bendahmane, An induced mutation in tomato eIF4E leads to immunity to two potyviruses, PLoS One, 5 (2010) e11313. [26] M. Dalmais, J. Schmidt, C. Le Signor, F. Moussy, J. Burstin, V. Savois, G. Aubert, V. Brunaud, Y. de Oliveira, C. Guichard, R. Thompson, A. Bendahmane, UTILLdb, a Pisum sativum in silico forward and reverse genetics tool, Genome Biol, 9 (2008) R43.

[27] K. Triques, B. Sturbois, S. Gallais, M. Dalmais, S. Chauvin, C. Clepet, S. Aubourg, C. Rameau, M. Caboche, A. Bendahmane, Characterization of Arabidopsis thaliana mismatch specific endonucleases: application to mutation discovery by TILLING in pea, Plant J, 51 (2007) 1116-1125. [28] M.F. Devaux, B. Bouchet, D. Legland, F. Guillon, M. Lahaye, Macro-vision and grey level granulometry for quantification of tomato pericarp structure., Postharvest Biology and Technology, 47 (2008) 199-209. [29] M. Dubois, K. Gilles, J.K. Hamilton, P.A. Rebers, F. Smith, A colorimetric method for the determination of sugars, Nature, 168 (1951) 167. [30] B. Quemener, D. Bertrand, I. Marty, M. Causse, M. Lahaye, Fast data preprocessing for chromatographic fingerprints of tomato cell wall polysaccharides using chemometric methods, J Chromatogr A, 1141 (2007) 41-49. [31] N. Blumenkrantz, G. Asboe-Hansen, New method for quantitative determination of uronic acids, Anal Biochem, 54 (1973) 484-489. [32] B.V. McCleary, T.S. Gibson, D.C. Mugford, Measurement of total starch incereal products by amyloglucosidase-alpha-amylase method: collaborative study., Journal of AOAC International, 80 (1997) 571-579. [33] S. Levigne, M. Thomas, M.C. Ralet, B. Quemener, J.F. Thibault, Determination of the degrees of methylation and acetylation of pectins using a C18 column and internal standards. , Food Hydrocolloids, 16 (2002) 547-550. [34] U. Lewandrowski, A. Resemann, A. Sickmann, Laser-induced dissociation/high-energy collision-induced dissociation fragmentation using MALDI-TOF/TOF-MS instrumentation for the analysis of neutral and acidic oligosaccharides, Anal Chem, 77 (2005) 3274-3283.

[35] S.C. Fry, W.S. York, P. Albersheim, A. Darvill, T. Hayashi, J.P. Joseleau, Y. Kato, E. Perez-Lorences, G.A. MacLachlan, M. McNeil, e. al., An unambiguous nomenclature for xyloglucan-derived oligosaccharides., Physiologia Plantarum, 89 (1993) 1-3. [36] T.Y. Shcherban, J. Shi, D.M. Durachko, M.J. Guiltinan, S.J. McQueen-Mason, M. Shieh, D.J. Cosgrove, Molecular cloning and sequence analysis of expansins--a highly conserved, multigene family of proteins that mediate cell wall extension in plants, Proc Natl Acad Sci U S A, 92 (1995) 9245-9249. [37] H.T. Cho, H. Kende, Expression of expansin genes is correlated with growth in deepwater rice, Plant Cell, 9 (1997) 1661-1671. [38] D.J. Cosgrove, Creeping walls, softening fruit, and penetrating pollen tubes: the growing roles of expansins, Proc Natl Acad Sci U S A, 94 (1997) 5504-5505. [39] S. McQueen-Mason, D.M. Durachko, D.J. Cosgrove, Two endogenous proteins that induce cell wall extension in plants, Plant Cell, 4 (1992) 1425-1433. [40] E. Keller, D.J. Cosgrove, Expansins in growing tomato leaves, Plant J, 8 (1995) 795-802. [41] J.K. Rose, H.H. Lee, A.B. Bennett, Expression of a divergent expansin gene is fruitspecific and ripening-regulated, Proc Natl Acad Sci U S A, 94 (1997) 5955-5960. [42] N.H. Yennawar, L.C. Li, D.M. Dudzinski, A. Tabuchi, D.J. Cosgrove, Crystal structure and activities of EXPB1 (Zea m 1), a beta-expansin and group-1 pollen allergen from maize, Proc Natl Acad Sci U S A, 103 (2006) 14664-14671. [43] M. Lahaye, B. Quemener, M. Causse, G.B. Seymour, Hemicellulose fine structure is affected differently during ripening of tomato lines with contrasted texture, Int J Biol Macromol, 51 (2012) 462-470. [44] W.S. York, V.S. Kumar Kolli, R. Orlando, P. Albersheim, A.G. Darvill, The structures of arabinoxyloglucans produced by solanaceous plants, Carbohydr Res, 285 (1996) 99-128.

[45] Z. Jia, Q. Qin, A.G. Darvill, W.S. York, Structure of the xyloglucan produced by suspension-cultured tomato cells, Carbohydr Res, 338 (2003) 1197-1208. [46] Z. Jia, M. Cash, A.G. Darvill, W.S. York, NMR characterization of endogenously Oacetylated oligosaccharides isolated from tomato (Lycopersicon esculentum) xyloglucan, Carbohydr Res, 340 (2005) 1818-1825. [47] J.P. Vincken, W.S. York, G. Beldman, A.G. Voragen, Two general branching patterns of xyloglucan, XXXG and XXGG, Plant Physiol, 114 (1997) 9-13. [48] C. Assor, B. Quemener, J. Vigouroux, M. Lahaye, Fractionation and structural characterization of LiCl-DMSO soluble hemicelluloses from tomato, Carbohydr Polym, 94 (2013) 46-55. [49] A.A. Khan, J.F.V. Vincent, Anisotropy in the fracture properties of apple flesh as investigated by crack-opening tests., Journal of Materials Science, 28 (1993) 45-51. [50] J.A. Abbott, R. Lu, Anisotropic mechanical properties of apples. , Trans ASAE 39 (1996) 1451-1459. [51] D. Galvez-Lopez, F. Laurens, M.F. Devaux, M. Lahaye, Deciphering apple texture through instrumental, sensory and histological phenotyping in an apple progeny during storage., Euphytica, 185 (2012) 171-183. [52] Y.B. Park, D.J. Cosgrove, Xyloglucan and its interactions with other components of the growing cell wall, Plant Cell Physiol, 56 (2015) 180-194. [53] N. Sakurai, D.J. Nevins, Changes in physical-properties and cell-wall polysaccharides of tomato (Lycopersicon esculentum) pericarp tissues. , Physiologia Plantarum, 89 (1993) 681686. [54] G. Maclachlan, C. Brady, Endo-1,4-[beta]-Glucanase, Xyloglucanase, and Xyloglucan Endo-Transglycosylase Activities Versus Potential Substrates in Ripening Tomatoes, Plant Physiol, 105 (1994) 965-974.

[55] M. Huysamer, L.C. Greve, J.M. Labavitch, Cell Wall Metabolism in Ripening Fruit (IX. Synthesis of Pectic and Hemicellulosic Cell Wall Polymers in the Outer Pericarp of Mature Green Tomatoes (cv XMT-22), Plant Physiol, 114 (1997) 1523-1531. [56] J.M. Eklof, H. Brumer, The XTH gene family: an update on enzyme structure, function, and phylogeny in xyloglucan remodeling, Plant Physiol, 153 (2010) 456-466. [57] E. Miedes, K. Herbers, U. Sonnewald, E.P. Lorences, Overexpression of a Cell Wall Enzyme Reduces Xyloglucan Depolymerization and Softening of Transgenic Tomato Fruits. , Journal of Agricultural and Food Chemistry, 58 (2010) 5708-5713. [58] M. Pauly, P. Albersheim, A. Darvill, W.S. York, Molecular domains of the cellulose/xyloglucan network in the cell walls of higher plants, Plant J, 20 (1999) 629-639. [59] O. Lerouxel, T.S. Choo, M. Seveno, B. Usadel, L. Faye, P. Lerouge, M. Pauly, Rapid structural phenotyping of plant cell wall mutants by enzymatic oligosaccharide fingerprinting, Plant Physiol, 130 (2002) 1754-1763. [60] G. Mouille, H. Witucka-Wall, M.P. Bruyant, O. Loudet, S. Pelletier, C. Rihouey, O. Lerouxel, P. Lerouge, H. Hofte, M. Pauly, Quantitative trait loci analysis of primary cell wall composition in Arabidopsis, Plant Physiol, 141 (2006) 1035-1044. [61] N. Obel, V. Erben, T. Schwarz, S. Kuhnel, A. Fodor, M. Pauly, Microanalysis of plant cell wall polysaccharides, Mol Plant, 2 (2009) 922-932. [62] Y. Westphal, H.A. Schols, A.G. Voragen, H. Gruppen, MALDI-TOF MS and CE-LIF Fingerprinting of plant cell wall polysaccharide digests as a screening tool for arabidopsis cell wall mutants, J Agric Food Chem, 58 (2010) 4644-4652. [63] M.J. Kang, A. Tholey, E. Heinzle, Quantitation of low molecular mass substrates and products of enzyme catalyzed reactions using matrix-assisted laser desorption/ionization timeof-flight mass spectrometry, Rapid Commun Mass Spectrom, 14 (2000) 1972-1978.

Figure legends

Fig. 1. Gene structure and 3D structural analysis of SlExp1. (A) Schematic diagram of SlExp1 gene structure. The numbers indicate the size of the 2 exons (filled boxes) and the intron (black lines) in bps. Intronic, silent and missense mutations are represented by black, grey and orange triangles, respectively.

The green dashed line

represents the screened amplicon (B-C) Superposition of the maize EXPB1 structure determined by X-ray crystallography [42], indicated in beige and the 3D model of SlExp1, indicated in light blue. The tomato SlExp1 model was determined using the Geno3D server (http://geno3d-pbil.ibcp.fr). Residues involved in the long potential polysaccharide-binding site, W26, Y27, G40, G44, Y160 and W194 are represented in green, pink, blue, purple, yellow and red sticks, respectively.

Fig. 2. Firmness analysis of SlEXP1 TILLING mutants. Firmness measurement for Slexp1-6 (A) and Slexp1-7 (B) mutants. Tomato samples were collected at the breaker stage and analyzed in post-harvest conditions at fixed ripening stages: breaker stage (B0), five (B+5), ten (B+10), fifteen (B+15) and twenty five (B+25) days after breaker stage. Bars represent standard errors and asterisks indicate significative difference at P < 0.05.

Fig. 3. Boxplot of glucanase hydrolyzate. Boxplots represent MALDI-TOF MS ion intensities of hemicellulose oligosaccharides from the glucanase hydrolyzate of the tomato WT and Slexp1-7 mutant. Ion nomenclature is based on [35].

Tables

Mutant allele

Nucleotide substitution

Amino acid substitution

Slexp1-1

C433T

L145F

Slexp1-2 Slexp1-3 Slexp1-4 Slexp1-5 Slexp1-6 Slexp1-7(1) Slexp1-8 Slexp1-9(1)

C447T T450A C461T G931A G974C C979T T1053A C1055T

R149R A150A P154L V197I W211S Q213Stop S237S S238F

Table 1. Induced point mutations identified in SlExp1 coding region.

List of tomato mutant lines along with the identity and position of the mutated nucleotides and induced amino acids changes. (1) Mutant lines identified in the M82 genetic background.

Genotype Ripening stage

WT Slexp1-7 WT Slexp1-7 WT Slexp1-7 WT Slexp1-7 WT Slexp1-7

G B P R S

UA

21.4 ±1.7 18.1 ±2.9 22.3 ±2.1 23.0 ±1.8 23.3 ±0.9 22.7 ±1.4 24.5 ±0.9 24.2 ±1.4 24.6 ±1.0 24.9 ±1.1

NS

Total

% dry weight 55.8 ±8.9 77.2 ±8.6 56.5 ±8.2 74.6 ±6.2 48.7 ±3.3 71.0 ±4.9 57.4 ±5.6 80.4 ±5.4 45.3 ±6.1 68.6 ±6.8 43.4 ±3.3 66.0 ±3.6 48.8 ±10.2 73.3 ±11.1 50.8 ±3.5 75.0 ±3.1 43.3 ±3.9 67.8 ±4.7 45.9 ±3.3 70.8 ±3.3

AA

1.8 1.6 1.8 1.7 1.7 1.7 1.6 1.6 1.5 1.5

±0.1 ±0.1 ±0.1 ±0.2 ±0.1 ±0.1 ±0.2 ±0.1 ±0.1 ±0.1

DM Mol / 100 mol UA 44.8 ±5.9 41.8 ±10.7 47.3 ±8.1 50.7 ±9.6 48.0 ±5.3 50.6 ±4.4 46.9 ±5.1 49.7 ±8.6 46.4 ±2.0 40.0 ±2.4

N

6 5 5 5 5 5 5 5 5 5

Table 2: Cell wall polysaccharide composition and molar proportion of individual sugars. Uronic acids (UA), total neutral sugars (NS) and total sugars (Total), acetic acid ester content (AA) (± standard deviation) on the weight basis of dry alcohol insoluble material and degree

of methyl esterification of pectin (DM) in cell walls recovered from pericarp tissue of wild type (WT) and Slexp1-7 mutant at the green (G), breaker (B), pink (P), red (R) and overripe (S) stages. Significant difference between WT and Slexp1-7 mutant are in bold (pairwise ttest; p < 0.05); N= number of fruits. Supporting Information

Supplementary Figure 1: MALDI-TOF mass spectra of xyloglucan oligosaccharides The pics represent glucanase hydrolyzate of cell wall preparations of green M82 wild type pericarp tissue. See text and Supplementary Table 1 for nomenclature ; number above code refers to m/z value

Supplementary Table 1: Nomenclature used to refer to hemicelluloses structures 1

Elementary building block of xyloglucan structures: XLGG is made of the linkage of X, L

and 2 G elements;

2

Number after code refers to the number of residues in the

oligosaccharide: Pen5u2 is made of 5 pentoses (xylosyl residues) and 2 glucuronosyl residues. The number following the code a refers to the number of acetyl groups.

Genotype Ripening

Sugar (mol%)

N

stage Rha

Ara

Xyl

Man

Gal

Glc

UA

WT

G

Slexp1-7 WT

0.7 ±0.4 5.3 ±0.6 5.5 ±0.2 5.9 ±0.6 13.3 ±0.9 41.7 ±3.4 27.6 ±3.4 5 B

Slexp1-7 WT

P

R

Slexp1-7 WT Slexp1-7

1.0 ±0.3 4.9 ±0.7 5.8 ±0.4 5.8 ±0.5 12.4 ±1.3 40.6 ±0.6 29.6 ±1.0 5 1.1 ±0.1 5.0 ±0.3 5.8 ±0.2 5.8 ±0.2 11.8 ±1.2 43.7 ±2.8 26.7 ±2.7 5

Slexp1-7 WT

1.0 ±0.0 4.9 ±0.4 5.4 ±0.2 5.6 ±0.2 13.1 ±0.9 41.9 ±3.2 28.1 ±3.9 6

S

0.9 ±0.3 3.9 ±0.2 6.8 ±0.2 6.0 ±0.2

6.3 ±0.8 44.5 ±2.6 31.7 ±2.3 5

0.9 ±0.5 3.9 ±0.2 6.7 ±0.2 6.2 ±0.1

6.8 ±0.7 43.5 ±2.0 32.0 ±2.0 5

1.2 ±0.1 3.1 ±0.5 6.4 ±0.6 5.7 ±0.3

5.0 ±0.5 47.1 ±3.0 31.6 ±4.1 5

1.2 ±0.1 3.0 ±0.3 6.6 ±0.4 5.8 ±0.3

4.5 ±0.5 48.7 ±1.4 30.2 ±2.3 5

1.1 ±0.0 2.6 ±0.2 6.5 ±0.3 5.8 ±0.2

3.3 ±0.4 46.8 ±0.9 33.9 ±1.5 5

1.0 ±0.1 3.2 ±1.0 6.7 ±0.3 5.6 ±0.3

4.6 ±1.4 46.0 ±1.8 32.9 ±1.9 5

Supplementary Table 2: Molar proportions (± standard deviation) of rhamnose (Rha), arabinose (Ara), xylose (Xyl), mannose (Man), galactose (Gal), glucose (Glc) and uronic acids (UA) in the cell wall polysaccharides from pericarp tissue of wild type (WT) and Slexp1-7 mutant at the green (G), breaker (B), pink (P), red (R) and overripe (S) stages. N: number of fruits.