Changes in texture, cellular structure and cell wall composition in apple tissue as a result of freezing

Food Research International 42 (2009) 788–797

Contents lists available at ScienceDirect

Food Research International journal homepage: www.elsevier.com/locate/foodres

Changes in texture, cellular structure and cell wall composition in apple tissue as a result of freezing Sophie Chassagne-Berces a, Cécile Poirier b, Marie-Françoise Devaux b, Fernanda Fonseca a,*, Marc Lahaye b, Giuseppe Pigorini c, Christel Girault c, Michèle Marin a, Fabienne Guillon b a

UMR 782 Génie et Microbiologie des Procédés Alimentaires, INRA, AgroParisTech, F-78850 Thiverval-Grignon, France INRA-UR1268 Biopolymères Interactions Assemblages, F-44300 Nantes, Cedex 03, France c AIR LIQUIDE, Centre de Recherche Claude Delorme, 78354 Les Loges-en-Josas, France b

a r t i c l e

i n f o

Article history: Received 26 September 2008 Accepted 3 March 2009

Keywords: Fruit Freezing rates Physical parameters Macrovision CLSM Cryo-SEM Structure Cell wall Image texture analysis Granulometry

a b s t r a c t Apple texture is one of the critical quality features for the consumer. Texture depends on several factors that are difficult to control and which change with freezing. To better understand the mechanisms involved in the texture degradation of apple tissues following freezing/thawing, our approach was to combine mechanical properties, cellular structure and cell wall composition measurements on fresh and thawed apples (Granny Smith) after three different freezing protocols (at 20 °C, 80 °C and 196 °C). This work highlighted the interest of applying macrovision and image texture analysis to quantify the freezing effects on cellular structure and ice crystal size. Freezing at 20 °C and after immersion into liquid nitrogen were the protocols affecting the most fruit texture leading to cell membrane breakage resulting in cell wall collapse and tissue breakage, respectively, which accounted for the mechanical behaviour of the samples. All freezing protocols induced vacuole burst showing that the turgor pressure preservation remains critical during the freezing process. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction Freezing is used extensively to preserve food (fruit, vegetables, and meat). However, especially in the case of fruit, this preservation technique may result in textural changes leading to food softening (Brown, 1977). The quality of frozen/thawed fruit depends on a large number of factors including the type of fruit, the variety, raw quality and ripeness or time lapse between harvesting and processing (Marani, Agnelli, & Mascheroni, 2007; Phan & Mimault, 1980). Once the raw material and the harvesting conditions have been defined, the optimal freezing protocol for texture preservation has to be determined. In general, it is accepted that fast freezing better preserves local structure. It induces the production of a large number of small ice crystals that cause less migration of water and less breakage of cell walls, and consequently less texture deterioration (Brown, 1977; Delgado & Rubiolo, 2005; Marti & Aguilera, 1991). However, the kinetics of too fast a freezing can provoke breakage at the product level due to ice density differences with water which lead to texture modification. To improve stabilisation by freezing, a better understanding of the complex * Corresponding author. Tel.: +33 1 30 81 59 40; fax: +33 1 30 81 55 97. E-mail address: [email protected] (F. Fonseca). 0963-9969/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodres.2009.03.001

physical and chemical mechanisms taking place inside the fruit tissue during freezing/thawing is still needed. Texture of fruit is determined by different physical characteristics that arise from the structural organization at different levels: from molecular to tissue level. The cell is the elementary unit within the tissue and its integrity strongly impacts textural quality. Among the many factors involved in fruit texture, the structural integrity of the cell components (cell wall and middle lamella) and cell turgor pressure determined by water content in the vacuoles are the most important (Waldron, Smith, Parr, Ng, & Parker, 1997). Several instrumental techniques are required to investigate the changes of fruit textural properties after freezing/ thawing from different points of view: mechanical, microscopic and biochemical. Mechanical measurements, such as the Kramer–Shear test (Mastrocola, Pittia, & Lerici, 1996; Phan & Mimault, 1980), compression tests (Chiralt et al., 2001; Kim & Hung, 1994), puncture tests (Marani et al., 2007; Phan & Mimault, 1980; Zhang, Duan, Zhang, & Peng, 2004) or back extrusion (Robbers, Singh, & Cunha, 1997) are widely used to evaluate the firmness at the organ scale. Such testing makes it possible to measure the mechanical properties of fruit tissue before and after freezing.

789

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797

2. Materials and methods 2.1. Samples The apple variety Granny Smith was selected due to its availability and its good stability during storage conditions. The variety exhibits a lower texture degradation than other apple varieties (personal communication). Fruits were purchased from an agricultural cooperative (Dorléane, Saint Hilaire Saint Mesmin – Loiret, France), where they were stored at 1 °C under a modified atmosphere. Apples were studied at commercial maturity. In the lab, they were stored in a cold chamber at 4 °C, for a maximum duration of one month, until the moment of the experiment. 2.1.1. Sample preparation The apples were placed overnight at ambient temperature (21 ± 1 °C) before sampling. A 2 cm thick transverse cross section was cut at the equatorial level of each apple. Cylinders (1.2 cm in diameter and 2 cm in height) were taken equidistant from the surface and the seed sacs in the parenchyma region using a circular punch. 2.1.2. Freezing and thawing Three freezing protocols were applied: at 20 °C in a cold chamber, at 80 °C in gas nitrogen convection (Silversas, Air Liquide, Paris, France) and by immersion in liquid nitrogen (LN2, boiling point = 196 °C) until the core temperature reached the equilibrium value with the freezing temperature. The three protocols correspond to slow (0.9 °C/min), intermediate (8.1 °C/min) and very fast (310 °C/min) freezing rate, respectively. Once frozen, the samples were packed in polyethylene bags and thawed in a cold chamber at 4 °C overnight. They were finally placed at room temperature (21 ± 1 °C) until the sample cores reached room temperature. Fresh samples were used as reference samples for each experimental technique. The three freezing protocols and fresh samples are referred to as protocols in the following. 2.2. Mechanical properties The texture of apple cylinders was measured with a TA.XT2i texture analyser (Stable Micro Systems Ltd, Godalming, UK). All experiments were conducted at 21 °C. The fruit cylinders were kept at the same orientation because of fibrous non-isotropic properties of apple flesh (Khan & Vincent, 1993). For each protocol, two cylinders per fruit and ten fruit were analysed resulting in 20 measurements. Penetrometry tests were carried out as in Mehinagic et al., 2003 but with a 2 mm diameter cylindrical probe, penetrat-

zone where flesh rupture mean force is calculated

3.5 3

Force (N)

Microscopy is a useful tool to visualize food structure at the tissular and cellular scales and to study the influence of freezing. Structural modifications associated with freezing/thawing have been studied in several plant tissues through light microscopy (Buggenhout et al., 2006a; Khan & Vincent, 1996; Otero, Martino, Zaritzky, Solas, & Sanz, 2000; Sterling, 1968) and scanning electron microscopy (SEM) without (Delgado & Rubiolo, 2005; Sousa, Canet, Alvarez, & Tortosa, 2006) or with cryo-system (Cryo-SEM) (Alonso, Tortosa, Canet, & Rodriguez, 2005; Bomben, King, & Hayes, 1983; Martinez-Monzo, Martinez-Navarrete, Chiralt, & Fito, 1998; Tregunno & Goff, 1996). The impact of freezing on the microstructure of apples (Bomben et al., 1983), carrots (Buggenhout et al., 2006a), peaches (Otero et al., 2000), raspberries and blackberries (Sousa et al., 2006), blueberries and wild blackberries (Marti & Aguilera, 1991) and strawberries (Delgado & Rubiolo, 2005) have already been studied. The main drawback of microscopic techniques is sample preparation that is time consuming and the reduced field of view observed in a single image. Confocal laser-scanning microscopy (CLSM) enables non destructive optical sectioning of samples and make sample preparation easier and more rapid (Gray, Kolesik, Hoj, & Coombe, 1999; Kalab, Allan-Wojtas, & Miller, 1995). The observation of a representative number of cells requires the acquisition of a large number of images or several adjacent images and the reconstruction of the whole region as a mosaic image (Guillemin, Devaux, & Guillon, 2004). An alternative technique to microscopy is to use stereomicroscope or macrovision systems making it possible to observe a field of view of about 1 cm2. Plant tissue and cellular structures of tomato fruit can be characterised in this way (Devaux, Bouchet, Legland, Guillon, & Lahaye, 2008). The characterisation of samples by imaging techniques is completed by applying image analysis to quantify the structure observed. Cell size and shape can be measured from microscopic images after a segmentation of each individual cell. For macroscopic images, techniques based on image texture analysis can be envisioned to quantify information on object size. Image texture refers to local variation of grey levels and several methods have been proposed to quantify these variations. Grey level granulometry from mathematical morphology (Serra, 1982) has been successfully applied to extract quantitative information related to cell size in tomato tissues (Devaux et al., 2008). Biochemical changes of the cell wall (Waldron et al., 1997) are also related to texture changes. Fruit cell walls are composed of cellulose and hemicellulose embedded in a matrix of pectins (Kunzek, Kabbert, & Gloyna, 1999; Muhlethaler, 1967). During fruit ripening or storage, softening occurs as a result of enzymatic degradation of cell walls (Johnston, Hewett, & Hertog, 2002). Few authors (Alonso et al., 2005; Buggenhout et al., 2006b) have studied biochemical changes following fruit freezing. In this work, two hypotheses of fruit tissue degradation during freezing were kept. Freezing process influences ice crystal formation resulting in (1) vacuole rupture causing loss of turgor pressure and (2) structural damage of cells and cell walls, and, hence, the modification of tissue architecture. The aim of this work is to investigate the changes of textural properties during freezing/thawing at different levels of observation to better understand the mechanisms involved in the softening of apple tissue following freezing and thawing and to quantify the changes. Texture was measured by mechanical techniques at the organ scale. The cellular structure was investigated at the tissue scale using macrovision. The vacuolar integrity was observed at the cell scale using confocal microscopy. Ice crystals were analysed by cryo-scanning electron microscopy. The global composition of the cell wall was studied using biochemical techniques. Three freezing protocols were applied to study the different hypotheses of degradation. The apple was taken as a model of fruit due to its macroscopic flesh homogeneity.

(F)

2.5 2 1.5

(8)

1

(2)

0.5

(LN2)

0 0

2

4

6

8

10

Distance (mm) Fig. 1. Force–distance curves obtained during puncture test for fresh apples (F) and thawed apples after freezing: at 20 °C (2), at 80 °C (8) and immersion in liquid nitrogen (LN2).

790

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797

ing 10 mm into samples at 0.5 mm/s. Force vs. deformation curves were recorded and flesh rupture mean force between 7 and 9 mm penetration was calculated and corresponded to firmness expressed in Newton (Fig. 1.). Compression tests were performed with a 21 mm compression plate (P/21) at 0.1 mm/s until a 15% strain was reached. Stress vs. strain curves were analysed and Young’s modulus (E) was obtained from the slope of the loading curve at the point of its highest gradient. 2.3. Cell wall composition 2.3.1. Preparation of the alcohol insoluble material Three fruit were analysed per protocol. Six cylinders for each fruit were freeze-dried and ground using a Fast Prep-24 crusher (MP BIOmedicals; 10seconds at 6.5 m/s). Alcohol insoluble material was prepared under high pressure (100 bars) and high temperature (85 °C) with ASE 200 (Dionex) as follows. Ground materials were washed with 70% ethanol and dehydrated by solvent exchange (100% ethanol). The material was then dried with 100% ethanol. A Dubois test (Dubois, Gilles, Hamilton, Rebers, & Smith, 1956) was used to control the absence of free sugars in the last alcohol extract during the preparation of alcohol insoluble material (absence of color). 2.3.2. Sugar analysis Identification and quantification of neutral sugars were carried out by gas–liquid chromatography (GC) after sulphuric acid degradation (Hoebler, Barry, David, & Delort-Laval, 1989). Alcohol insoluble materials were dispersed in 13 M sulphuric acid for 30 min at 25 °C and then hydrolyzed in 1 M sulphuric acid (4 mol/L, 2 h, and 100 °C). Sugars were converted to alditol acetates (Blakeney, Harris, Henry, & Stone, 1983) and chromatographed on a BP 225 capillary column (SGE France SARL; temperature: 198 °C, carrier gas: H2). For calibration, a standard sugar solution (rhamnose, fucose, arabinose, xylose, mannose, galactose, and glucose) and inositol (the internal standard) were used. Uronic acids in acid hydrolyzates were quantified using the metahydroxydiphenyl colorimetric method (Thibault, 1979). Uronic acids were assumed to be composed solely of galacturonic acid. The degree of pectin-esterification by methanol and acetic acid in alcohol insoluble materials was measured by HPLC after alkaline hydrolysis according to the method used by Levigne et al. (Levigne, Thomas, Ralet, Quemener, & Thibault, 2002). The degree of methyland acetyl-esterification corresponds to the mol amount of methanol and acetic acid, respectively, per 100 mol of galacturonic acid measured by colorimetry (Thibault, 1979). 2.4. Physical structure The physical structure was studied at different scales. The cellular structure was investigated at the tissue scale using macrovision. The vacuolar integrity was observed at the cell scale using confocal microscopy. Ice crystals were analysed by cryo-scanning electron microscopy. Image analysis was applied to quantify the cellular structure and ice crystal size. 2.4.1. Imaging the cellular structure of fruit flesh Three apples and two cylinders per fruit were studied for each treatment. Three 150 lm thick transverse sections were obtained per cylinder using a vibrating blade microtome (HM 650 V, Microm International GmbH, Walldrof, Germany). Ultrasonic waves were applied on sections for 5 min to eliminate air bubbles. Images were acquired using a macrovision system composed of a CDD monochrome camera (Sony XC-8500CE) fitted with a 50 mm lens (AF Nikkor, 1:1.8) and a 36 mm extension tube. Sections were backlighted using a fiber optic ring-light (SCHOTT DCRILL, Polytec,

France) that sent light at oblique angles. The camera and lens were adjusted to observe a 5.50 mm  7.25 mm area with image resolution equal to 9.55 lm per pixel. Grey levels were coded between 0 (black) and 255 (white). Image acquisition was performed on a PC Station equipped with a Matrox METEOR II board and driven by Vision Stage software (Version 1.6fr Alliance Vision, France). 2.4.2. Imaging of the vacuole compartment Three apples and two cylinders per fruit were studied. To be able to observe the vacuole, cells must not be cut during section preparation. As apple cell diameter ranges between 200 and 300 lm (Khan & Vincent, 1990), the thickness of the transverse sections was fixed to 500 lm. Two sections per cylinder were cut using a vibrating blade microtome (HM 650 V, Microm International GmbH, Walldrof, Germany). Ultrasonic waves were applied on each section for 5 min. Sections were stained using neutral red (0.05% in 5 mM Hepes KOH buffer, pH 8–8.5) in order to make the vacuoles fluorescent. The neutral red was prepared in sucrose solution (20% w/w) corresponding to apple juice osmolarity (800 mOsm) to preserve the isotonic state of cells. The colored sections were twice washed in the iso-osmolarity sucrose solution for 5 min. Images were acquired using a Confocal laser-scanning microscope (Zeiss, LSM 410). Regions were observed for each section. The 10 lens was used to observe together small and large cells. The excitation wavelength was 543 nm and the light emitted over 570 nm was collected using a long pass filter. Eight scans were averaged for each image in order to reduce background noise. 3D images (about 18 images per x–y direction depending on the decrease in fluorescence brought about by the depth of the optical plane) were acquired for further 3D image processing. 2.4.3. Imaging ice crystals into cells Two apples and two cylinders per fruit were studied for each freezing protocol. Frozen samples at 20 °C and at 80 °C were immediately stored in liquid nitrogen in order to fix the ice structure formed during freezing. According to Bomben and King (1982) no changes in ice structure due to fixation were apparent. This step was not necessary for the last freezing protocol corresponding to immersion in liquid nitrogen. Frozen apple cylinders were transferred into the cold stage of a cryo-system (GATAN) for scanning electron microscopy (PHILIPS 525) and then cut with a knife (at 150 °C). After etching at 80 °C to reveal cell outlines, cylinders were cooled at 160 °C before being coated with gold, and the surface was examined at 160 °C under a low acceleration voltage (10 keV). 2.4.4. Image texture analysis using grey level morphological granulometries Cellular structure and ice crystal size were quantified using image texture analysis. Grey level granulometric methods from mathematical morphology were applied, as described by Devaux et al. (2008). Mathematical morphology consists of the successive application of two basic transformations called ‘‘opening” and ‘‘closing”. Opening and Closing are based on the comparison of images to ‘‘structuring elements” of given sizes and shapes. Only particles larger than the structuring element remain after an opening or a closing. Opening is applied to study the size of bright objects, whereas closing is applied to characterise dark objects. A size distribution can be obtained by applying a sequence of either openings or closings with structuring elements of increasing size. After each opening or closing step, the sum of grey level is computed. Variations between each step depend on the amount of bright or dark objects removed by opening or closing. They are drawn according to the size of the structuring element resulting in a granulometric curve for the image.

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797

The cellular structure was quantified from macrovision images by applying 35 successive opening and closing steps. Square structuring elements of size ranging from 28.65 lm to 678.05 lm by increments of 19.1 lm were used. Ice crystal size was quantified from Cryo-Scanning Electron Microscopy images by applying 100 successive opening and closing steps. Linear structuring elements were used instead of square to reveal the length of ice crystals

791

rather than their average diameter. As the orientation of crystals was random, four orientations were considered: 0, 45, 90 and 135°. The length of structuring elements ranged from 0.62 lm to 41.21 lm by increments of 0.41 lm. Image analysis procedures were developed using Matlab2007a software (The MATHWORKS, USA). 2.5. Statistical analysis

(a)

3.5 3.0

Firmness (N)

2.5 2.0 1.5 1.0 0.5 0.0

Fresh

Young's modulus (MPa)

(b)

LN2

(- 80 °C)

(- 20 °C)

2.5.1. Granulometric curve analysis Principal Component Analyses were applied to analyse the sets of granulometric curves that quantified the cellular structure and ice crystal size. Data tables were built from the curves by considering the images as individuals and the opening and closing steps as variables. Principal component analysis assesses synthetic uncorrelated principal components as linear combinations of all the variables, in the present case opening or closing steps (37). Similarity maps of images can be drawn using the component scores. Interpretation of components is obtained by looking at the linear combination coefficients, called loadings (Robert, Devaux, & Bertrand, 1996). Principal component analysis was applied within MATLAB environment (The MATHWORKS, USA).

2.5

2.5.2. Variance analysis Variance analyses were applied to the different sets of data: principal components extracted for cellular structure, principal components extracted for ice crystals, mechanical firmness values, sugar proportions in cell walls. Variance analyses were performed using statgraphics Plus 5.1 (Sigma-Plus, France).

2 1.5 0.1

3. Results

0.05

3.1. Mechanical properties

0

Fresh

LN2

(-80°C)

(-20°C)

Fig. 2. Effect of freezing protocols at 20 °C, at 80 °C and immersion in liquid nitrogen LN2 on (a) apple firmness and (b) Young’s modulus. Fresh apples were used as control. Means across repetitions are represented with a confidence interval.

Fig. 2 represents the mean values of two texture parameters (firmness and Young’s modulus) for the four protocols. For the two freezing parameters, the effect of freezing rates was significant (P < 0.01), as expected from results in the literature (Brown, 1977;

Fig. 3. Macroscopic images of apple parenchyma tissue sections before and after freezing. (a) fresh apple, (b) thawed apple after freezing at freezing at 80 °C (d) thawed apple after immersion in liquid nitrogen. A tissue crack is arrowed. The field of view was 5.5  7.25 mm.

20 °C (c) thawed apple after

792

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797

Delgado & Rubiolo, 2005). Values for frozen/thawed apples differed significantly from those of fresh apples. Nevertheless, freezing at 80 °C provoked less degradation (54%) in firmness than freezing at 20 °C (79%) or immersion in liquid nitrogen (91%). Freezing at 80 °C and immersion in liquid nitrogen exhibited less Young’s modulus decrease (97%) than freezing at 20 °C (99%). 3.2. Cellular structure Typical examples of images of fresh or frozen/thawed apple tissues acquired by Macrovision are presented in Fig. 3. Cells are clearly visible on the image of fresh apples. Some intercellular spaces could be seen. Large white spots correspond to vascular bundles. Freezing at 20 °C induced large changes of the cellular structure: cell walls seemed to be collapsed. More and larger intercellular spaces were observed. The cellular structure for the freezing at 80 °C did not differ much from that of the fresh samples. In the case of liquid nitrogen freezing a long and thin tissue crack differentiated apple tissue following this freezing condition from the fresh samples. Object size was quantified using greyscale mathematical morphology. Cells and intercellular spaces appeared as dark objects while cell walls appeared as bright objects. In order to characterise both bright and dark objects, opening and closing were applied. Opening also took into account vascular bundles. The relative amount of vascular bundles was assumed to be identical for each freezing protocol as samples were taken from the same apple regions. Average granulometric curves were calculated for each freezing protocol and are shown in Fig. 4. The left part of the curves represents the analysis of cells and intercellular spaces by closing. A maximum corresponding to the smallest dark objects, i.e. cells, was observed around 100 lm. Above 200 lm, the largest dark objects represented the intercellular spaces. Less cells and more intercellular spaces were found in apples frozen at 20 °C. The proportion of cells and intercellular spaces in thawed apples after freezing at 80 °C or after immersion in nitrogen liquid were similar to those in the fresh apples. The right part represents the analysis of cell walls, collapsed cell walls and vascular bundles by opening. The first peak (50 lm) corresponds to the removing of cell walls after the first opening. After freezing at 20 °C the thawed apples exhibited a higher amount of objects larger than 100 lm corresponding to collapsed cell walls. On the basis of image analysis by opening and closing, it was apparent that tissues are similar

Fig. 5. Cellular structure. (a) Principal component analysis of opening–closing curves: similarity map of components 1 and 2. Convex hulls were drawn for fresh apples (F) and thawed apple after freezing at 20 °C (2), at 80 °C (8) and immersion in liquid nitrogen (A). (b and c). Loadings corresponding to components 1 (b) and 2 (c).

with freezing at 80 °C, immersion in liquid nitrogen and fresh tissues. Principal component analysis was applied in order to compare the whole set of granulometric curves. The first two principal components accounted for 67.26% and 15.60% of the total variance respectively. The similarity map of the 72 granulometric curves is plotted on Fig. 5a. A freezing effect was found for components 1 and 2 by variance analysis. Freezing at 20 °C significantly differed from other protocols. On component 1, samples after freezing at 80 °C and by immersion in liquid nitrogen differed from fresh samples. Interpretation of the components was obtained by looking at the corresponding loadings (Fig. 5b,c). Loading of component 1 (Fig. 5b) mainly described size variations of bright objects. Positive values were observed for large opening sizes corresponding to collapsed cell walls and negative values for smaller opening sizes corresponding to cell walls. Images of thawed apples after freezing at 20 °C exhibited positive score values (Fig. 5a) associated with cell wall collapses. Fresh samples were found on the opposite side. Freezing at 80 °C and using liquid nitrogen were intermediate. Loading of component 2 (Fig. 5c) mainly described size variations of dark objects. Positive values were observed for small closing sizes corresponding to cells, and negative values for larger closing sizes corresponding to intercellular spaces. Again, freezing at 20 °C differed from other freezing protocols because of the larger amount of intercellular spaces and therefore a lower amount of intact cells. 3.3. Preservation of vacuoles

Fig. 4. Cellular structure. Image analysis. Average granulometric curves assessed for fresh apple tissue (F) and thawed apple tissue after freezing: at 20 °C (2), at 80 °C (8) and immersion in liquid nitrogen (A).

Confocal laser scanning microscopy was used to visualize cell vacuoles in fresh and frozen/thawed apple tissue. Fig. 6. shows examples of 3D images. For the fresh state, vacuoles were clearly visible and filled up the total cell volume (Fig. 6a). After the

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797

793

Fig. 6. 3D images of apple cells from fresh apples (a), thawed apples after freezing at objective: 40; image size: 1280  1280 lm).

20 °C (b), at

freezing/thawing protocol, no intact vacuole could be observed in cells whatever the freezing protocols (Fig. 6b–d). This result indicates that the tonoplast which surrounds the vacuole was never preserved after application of the freezing/thawing processes, in accordance with cell membrane damage following freezing/thawing in carrot tissue (Greve et al., 1994).

four directions in the image: 0, 45, 90 and 135°. The four resulting granulometric curves were summed to obtain a final curve that measure the length of crystal averaged for all directions. Closingopening curves were assessed to characterise both ice crystals and the freeze concentrated matrix. The three freezing protocols mainly differed through comparison of the ice crystals (dark objects), (Fig. 8). Thawed apples after freezing at 20 °C exhibited a very distinctive granulometric profile. Large ice crystals with a closing size mainly between 10 and 30 lm represented almost the total number of ice crystal. For the other two freezing protocols, small ice crystals were found with a closing size of less than 5 lm. After freezing at 80 °C, the crystals were slightly larger (about 3 lm) than those observed after immersion using liquid nitrogen (about 2 lm). With regards to bright objects, i.e. freeze concentrated matrix (between ice crystal area), after freezing at 20 °C some differences were observed. As a consequence of ice crystal size, the thickness of the freeze concentrated matrix was larger than after freezing at 80 °C or using liquid nitrogen. Principal component analysis was applied to the set of granulometric curves computed for all macroscopic images (Fig. 9). The first principal component described the ice crystal size with the larger size on the left and the smaller size on the right, differentiating freezing at 20 °C from the freezing at 80 °C followed by immersion in liquid nitrogen. Principal component 2 revealed the intermediate size of ice crystals after freezing at 80 °C.

3.4. Freezing effect on ice crystal size Cold-stage scanning electron microscopy was used to observe the ice crystal size. Fig. 7 shows examples of images obtained for frozen apple tissue with different freezing protocols. Thanks to the quality of the images especially at 100 lm, it was possible to carry out an interpretation based on descriptions reported by Bomben and King (Bomben & King, 1982). Bright regions (Fig. 7) correspond to the freeze concentrated matrix, the cytoplasmic membrane, and the cell walls. Darker regions corresponded to ghost of ice microcrystal that sublimed during etching. Differences in intensity were due to differences in height between ice crystal, freeze concentrated material and insoluble structures of the tissue. As expected, after freezing using liquid nitrogen, cells contained many small ice crystals which were evident as pits (Fig. 7c). After freezing at 80 °C, ice crystals in the cell appeared slightly larger (Fig. 7b). After freezing at 20 °C, ice crystals were much bigger (Fig. 7a). Furthermore the shape of cell walls seemed to be less regular for the freezing at 20 °C than for the other two freezing protocols. Grey level morphological granulometry was applied to quantify ice crystal size. Ice crystals exhibited elongated shapes and were orientated in different directions. Image texture analysis was therefore applied by using linear structuring elements oriented in

80 °C (c) or immersion in liquid nitrogen (d).(colouring: neutral red;

3.5. Chemical composition of cell wall Changes in average sugar composition, degrees of methyl- and acetyl- esterification (DM, DA, respectively) of cell wall polysaccharides in the alcohol insoluble material obtained from fresh

794

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797

Fig. 7. Cryo-Scanning Electron Microscopy images of apple tissue frozen in three different protocols. (a) Freezing at in liquid nitrogen. Cell membranes are arrowed. Left-scale bar = 500 lm and right-scale bar = 100 lm.

20 °C; (b) Freezing at

80 °C; (c) Freezing by immersion

Fig. 8. Ice crystal size. Image analysis. Average granulometric curves assessed for frozen apple tissue following freezing at 20 °C (2), at 80 °C (8) and immersion in liquid nitrogen (A).

Fig. 9. Ice crystal size. Principal component analysis of opening–closing curves: similarity map of components 1 and 2. Convex hulls were drawn for freezing at 20 °C (2), at 80 °C (8) and immersion in liquid nitrogen (A).

and frozen/thawed apples were studied (Table 1). Whatever the protocol, the total sugars accounted for 90% of alcohol insoluble material mass. The main neutral sugars of apple cell wall were glu-

cose (from 29% to 32%) and arabinose (from 9% to 11% dry weight). In addition to their cellulose and hemicellulose origins, some contaminating residual starch may contribute to the high glucose

795

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797 Table 1 Average sugar composition and degrees of methyl- and acetyl-esterification (DM and DA, respectively) of the alcohol insoluble material obtained from fresh apple tissue or frozen in three different protocols. Means with the same letter (a, b) indicate there is no significant difference (Student Newman Keul’s test: P < 0.01). Fresh

Thawed after freezing protocols At

Sugarsa Rhamnose Fucose Arabinose Xylose Mannose Galactose Glucose Neutral sugars Uronic acids Total sugars DMa DAa a

1.2 1.5 15.3 6.8 2.7 5.8 35.3 59.1 30.8 89.9 77.1 25.1

(a) (a) (b) (ab) (ab) (a) (a) (a)

20 °C

1.3 (a) 1.5 13.6 (b) 6.9 2.9 (a) 5.5 (b) 35.8 (a) 54.9 (b) 30.2 (ab) 85 (ab) 80.6 24

At

80 °C

Immersion in LN2

1.1 1.5 15.8 6.8 2.6 6.7 34.2 56.9 29.8 86.7 80.3 24.1

(b)

1.2 (a) 1.5 15 (ab) 6.8 2.8 (ab) 5.9 (ab) 35.9 (a) 54.2 (b) 27.4 (b) 81.6 (b) 80.5 25.6

(a) (b) (a) (b) (ab) (ab) (ab)

P

0.0001 0.8374 0.0001 0.6333 <0.0001 0.0134 0.0063 0.0034 0.0468 0.0055 0.2522 0.2461

% Dry weight and % molar.

content of the cell wall preparation. Uronic acid content accounted for 27 to 31% of dry weight. After freezing, the proportion of neutral sugars and uronic acids were lower in cell walls of frozen/ thawed apples than in cell walls of fresh apples. The alcohol insoluble materials from thawed apples after freezing at 20 °C or immersion in liquid nitrogen were significantly (p < 0.05) poorer in total sugars than those from fresh apples. These levels reflected a lower content in uronic acids (p < 0.05) and neutral sugars (p < 0.0034) in thawed apples after freezing. Among neutral sugars, the amount of arabinose was significantly (p < .0001) lower in thawed apples after freezing at 20 °C than in fresh apples, whereas the amount of mannose was significantly (p < 0.05) richer after freezing at 20 °C. Rhamnose was slightly lower (p < 0.0001) in apples frozen at 80 °C but the variation was extremely limited in all treatments. The other neutral sugars as well as the degrees of methyl- and acetyl- esterification did not exhibit significant variations between the fresh state and the frozen/thawed state. 4. Discussion 4.1. Imaging and grey level granulometry to quantify the cellular structure and ice crystal size The tissue of fresh ‘‘Granny Smith” apple is composed of cells regularly distributed and intercellular spaces. Macrovision makes it possible to observe the parenchymateous tissue from microtome sections in a single image. At this scale, resolution presents some limits: cells smaller than 50 lm are difficult to visualize. In addition, because of cell translucency, several cell layers could be visible in images contributing to grey levels in the images and making segmentation of individual cells irrelevant. Alternatively, grey level granulometry was applied to quantify the cellular structure. The

Table 2 Effect of freezing at

20 °C, at

Freezing protocols

At At At

20 °C (0.9 °C/min) 80 °C (8.1 °C/min) 196 °C (310 °C/min)

method allows the joint characterisation of cell and intercellular space sizes as well as cell wall collapse. The average grey level cell size in the apple tissue ranged from 80 to 130 lm and intercellular spaces were revealed for sizes over 200 lm. Results on cell size are in agreement with those obtained with other microscopy techniques and image analysis: between 75 and 150 lm by light microscopy (Lewicki & Porzecka-Pawlak, 2005) or between 50 and 300 lm by light or scanning electron microscopy (Khan and Vincent (1990). The largest intercellular spaces observed by Khan and Vincent (1990) were up to 3000 lm long and 100–200 lm large, which establishes that the segmentation of our images have been performed in the diameter direction. Grey level granulometry from mathematical morphology has also been applied to extract quantitative information on ice crystal size. Quantification from scanning electron microscopy images is a tricky task because grey level variations reveal both the wanted structures and artefacts caused by non-planar surfaces and orientations of the samples. This situation causes unwanted variability in the contrast between holes and edges. Despite these drawbacks, grey level granulometry was applied to attempt extraction of quantified information concerning the size of ice crystals. Whatever the freezing protocols, ice crystals inside cells were clearly visible; the size of ice crystals was between 10 and 30 lm after freezing at 20 °C and below 5 lm for faster freezing at 80 °C and by immersion in liquid nitrogen. Ice crystal sizes in apple tissue were quantified in few papers. Data were obtained on apple tissue images by manually counting intersections on circular test lines located at random positions by Bomben et al. (1983). For fast freezing rate (450 °C/min), they also observed small ice crystals between 3.6–2.7 lm. However for slow freezing rate (<1 °C/min), apple tissue was strongly deteriorated either by freezing protocol or by sample preparation and comparison with our results remained difficult. 4.2. Link between structure preservation (tissue, cell, cell wall) and mechanical texture modification of frozen/thawed tissue Whole texture of fruit is a complex characteristic resulting from several factors. Mechanical properties evaluated at the organ level depend on contributions from the different levels of structure and how they interact with one another. In the present work, the cellular structure, the integrity of vacuoles and the composition of cell walls were taken into account. Depending on the protocol, freezing caused different kinds of damage which we attempted to relate to differences in softening (Table 2). The highest deformation of texture at the organ level measured by mechanical methods was provoked after freezing at 20 °C. Breakage of vacuoles, a major modification of cell wall composition and a high number of collapsed cell walls, larger intercellular spaces and tearing of the tissue displayed on macrovision images were observed depending on the structural level studied. The water contained in the vacuoles determined the turgor pressure, which is recognized to have an effect on fruit texture properties (Alvarez, Saunders, & Vincent, 2000; Rojas, Gerschenson, &

80 °C and immersion in liquid nitrogen LN2 on different levels of structure. Structure levels Cell wall

Cell

Tissue

Organ

Composition modification

Vacuole integrity

Cellular structure

Softening

+++ + ++

destroyed destroyed destroyed

collapsed cell wall, tearing tissue preservation tissue preservation, breakage

Puncture

Compression

++ + +++

++ + +

796

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797

Marangoni, 2001). As a consequence of vacuole rupture, the interaction between cell walls and cellular contents (organic acids, phenols and hydrolytic enzymes) is facilitated and could allow cell wall enzymatic dissolution during thawing. The main sugars, the amounts of which were modified after freezing at 20 °C, concern arabinose representative of pectin and mannose, which is cellulosic–hemicellulosic sugar. The pectin substances are the main components of the middle lamella, a region considered important for maintaining cell to cell adhesion and cell packing in fruit tissue (Johnston et al., 2002; Kunzek et al., 1999). Moreover the lowest arabinose content could be explained by arabinan losses from the rhamnogalacturonan I domains of pectin which may participate in the cell wall mechanical characteristics by forming interaction with cellulose (Zykwinska, Ralet, Garnier, & Thibault, 2005). As a consequence, modification of pectins and hemicelluloses may contribute to the collapse of the cell walls, resulting in cell separation with the presence of larger intercellular spaces. The degradation of cell walls seems to be higher for slow freezing rates (at 20 °C) than for fast freezing rates. The combined effect of turgor pressure decrease and cell wall alteration may be responsible for tearing of tissue associated with the softening. According to the mechanical test applied at the organ scale, freezing by immersion in liquid nitrogen caused the highest (puncture) or the lowest (compression) deformation of texture compared to other freezing protocols. The presence of tissue breakage displayed on each macrovision image can explain the excessive texture degradation, resulting in softening observed by puncture. Puncture test evaluates local fracture behaviour (Bourne, 2002). This cracking related to a rapid formation of an external layer of ice has been described in the literature (Kim & Hung, 1994). It appears mainly at the level of vascular bundles, where the cell size is smaller. On the other hand, by compression texture analysis the lower deformations were observed following immersion in liquid nitrogen and freezing at 80 °C. This overall measure which evaluates the deformability of the tissue taken as a whole (Bourne, 2002) can certainly better reflect the lower cellular structure degradation obtained by macrovision analysis. Minimum texture degradation was observed after freezing at 80 °C, but the texture difference between the fresh tissue and the frozen/thawed tissue was still high. No intact vacuoles could be observed and a slight modification of cell wall composition was noticed. This work highlighted the importance of vacuole preservation and the turgor pressure involved in apple textural properties. The impact of the freezing rate on cell wall composition remains complex. The main changes observed after freezing at 20 °C compared to faster freezing (at 80 °C and liquid nitrogen) were probably caused by the osmotic imbalance due to ice crystal formation (Mazur, 1984). Freezing at 20 °C, corresponding to a slow cooling rate, was responsible for fewer ice crystals but crystals of larger size. It also resulted in high osmodehydration which damaged the vacuoles and the cell walls and degraded the cellular structure, resulting in overall structural degradation and thus a decrease in firmness. Thawing emphasizes the degradation which has already taken place in the frozen state. In contrast, freezing at 80 °C (and immersion in liquid nitrogen) which corresponds to faster freezing rates, produced a large number of small ice crystals which may make it possible to maintain cell compartments and the cellular structure in the frozen state. Thawing of rapidly cooled tissues may thus be the most important factor responsible for cell membrane disruption. The migratory recrystallisation of ice which takes place when molecular mobility increases, allows the crystallisation of residual water that is kinetically inhibited during cooling. Ice crystals which are formed can then grow and solute concentration increases thus leading to vacuole deterioration.

4.3. A link with process engineering At a macroscopic level, for the faster freezing kinetics (immersion in liquid nitrogen, about 300 °C/min), heat transfer kinetics could be well related to the tissue breakage observed at the external part of the apple cylinder. Based on the measurement of the temperature at the centre of the apple sample compared to a transient heat transfer simulation during cooling of a cylinder, an average value of the dimensionless Biot number, representing the ratio of the convection intensity outside the product over the conduction intensity inside the product, has been estimated, with: Biot = h .R/k, where h is the convection heat transfer coefficient (W/m2/K), k, the heat conductivity of apple tissue (W/m/K) and R the characteristic dimension of the cylinder (radius) (m). In freezing kinetics at 20 °C and at 80 °C, the Biot number remains below 0.1 (respectively from 0.03 to 0.1) but for immersion at 196 °C the estimated Biot number is higher than 0.1 and close to 1. According to the Biot number below 0.1, the temperature gradient inside the product is not significant and breakage is not induced. However, in spite of the small size (1.2 cm of diameter), when immersed in 196 °C, conduction intensity inside the product remains too low in connection with convection intensity and the resulting thermal gradient causes macroscopic structural damage. A limiting value of 0.1 for the Biot can be taken into account to extrapolate and optimize the cooling kinetics from lab scale to industrial scale to avoid breakage in the apple tissue studied. If for example the product size is higher than that of the sample studied, a lower convection intensity should be applied at the lab scale. Furthermore, the Biot estimation is not helpful for discriminating the effect of the freezing protocol between 20 °C and 80 °C at the cellular level. Further study of mass transfer (water mobility) is necessary to better quantify the evolution at the cellular level. 5. Conclusions Fs a conclusion, it is a real challenge to better identify when structure degradation takes place during processing in order to preserve the texture in complex food structures issuing from plants. The relationship between quantitative analysis of morphological changes in apple microstructure during freezing and final textural properties needs to be better understood to define new freezing kinetics and guide fruit quality engineering. Such better understanding and prediction tools may also be elaborated on the basis of a representation of the coupling of heat and mass transfer kinetics inside the tissues. Acknowledgment The authors wish to thank Air Liquide which financially supported this research and Dorleane SA who furnished apples and made possible their storage under modified atmosphere. We also thank to B. Bouchet, J. Vigouroux (BIA, INRA-Nantes) for their excellent technical assistance in microscopic measurements and biochemical analysis of cell walls, respectively. We would also like to thank B. Quemener (BIA, INRA-Nantes) for fruitful discussions and N. Wolff (PESSAC, INRA-Versailles) for her skilful technical support with Cryo-SEM. References Alonso, J., Tortosa, M. E., Canet, W., & Rodriguez, M. T. (2005). Ultrastructural and changes in pectin composition of sweet cherry from the application of prefreezing treatments. Journal of Food Science, 70(9), E526–E530. Alvarez, M. D., Saunders, D. E. J., & Vincent, J. F. V. (2000). Fracture properties of stored fresh and osmotically manipulated apple tissue. European Food Research and Technology, 211(4), 284–290.

S. Chassagne-Berces et al. / Food Research International 42 (2009) 788–797 Blakeney, A. B., Harris, P. J., Henry, R. J., & Stone, B. A. (1983). A simple and rapid preparation of alditol acetates for monosaccharide analysis. Carbohydrate Research, 113(2), 291–299. Bomben, J. L., & King, C. J. (1982). Heat mass transport freezing apple tissue. Journal of Food Technology, 17(5), 615–632. Bomben, J. L., King, C. J., & Hayes, T. L. (1983). Cold-stage scanning electron microscope measurement of ice morphology in apple tissue as a function of freezing rate. Cryobiology, 20(5), 574–586. Bourne, M. C. (2002). Food texture and viscosity: Concept and measurement (2nd ed.). London: Academic Press. pp. 427. Brown, M. S. (1977). Texture of frozen fruits and vegetables. Journal of Texture Studies, 7(4), 391–404. Buggenhout, S. V., Lille, M., Messagie, I., Loey, A., Autio, K., & Hendrickx, M. (2006a). Impact of pretreatment and freezing conditions on the microstructure of frozen carrots: Quantification and relation to texture loss. European Food Research and Technology, 222(5), 543–553. Buggenhout, S. V., Messagie, I., Maes, V., Duvetter, T., Van Loey, A., & Hendrickx, M. (2006b). Minimizing texture loss of frozen strawberries: Effect of infusion with pectinmethylesterase and calcium combined with different freezing conditions and effect of subsequent storage/thawing conditions. European Food Research and Technology, 223(3), 395–404. Chiralt, A., Martinez-Navarrete, N., Martinez-Monzo, J., Talens, P., Moraga, G., Ayala, A., & Fito, P. (2001). Changes in mechanical properties throughout osmotic processes: Cryoprotectant effect. Journal of Food Engineering, 49(2–3), 129–135. Delgado, A. E., & Rubiolo, A. C. (2005). Microstructural changes in strawberry after freezing and thawing processes. Lebensmittel-Wissenschaft und-Technologie, 38(2), 135–142. Devaux, M. F., Bouchet, B., Legland, D., Guillon, F., & Lahaye, M. (2008). Macro-vision and grey level granulometry for quantification of tomato pericarp structure. Postharvest Biology and Technology, 47(2), 199–209. Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., & Smith, F. (1956). Colorimetric method for determination of sugars and related substances. Analytical Chemistry(28), 350–356. Gray, J. D., Kolesik, P., Hoj, P. B., & Coombe, B. G. (1999). Confocal measurement of the three-dimensional size and shape of plant parenchyma cells in a developing fruit tissue. The Plant Journal, 19(2), 229–236. Greve, L. C., Shackel, K. A., Ahmadi, H., McArdle, R. N., Gohlke, J. R., & Labavitch, J. M. (1994). Impact of heating on carrot firmness – Contribution of cellular turgor. Journal of Agricultural and Food Chemistry, 42(12), 2896–2899. Guillemin, F., Devaux, M. F., & Guillon, F. (2004). Evaluation of plant histology by automatic clustering based on individual cell morphological features. Image Analytical Stereology, 23, 13–22. Hoebler, C., Barry, J. L., David, A., & Delort-Laval, J. (1989). Rapid acid hydrolysis of plant cell wall polysaccharides and simplified quantitative determination of their neutral monosaccharides by gas–liquid chromatography. Journal of Agricultural and Food Chemistry, 37(2), 360–367. Johnston, J. W., Hewett, E. W., & Hertog, M. L. A. T. M. (2002). Postharvest softening of apple (Malus domestica) fruit: A review. New Zealand Journal of Crop and Horticultural Science, 30(3), 145–160. Kalab, M., Allan-Wojtas, P., & Miller, S. S. (1995). Microscopy and other imaging techniques in food structure analysis. Trends in Food Science and Technology, 6(6), 177–186. Khan, A. A., & Vincent, J. F. V. (1990). Anisotropy of apple parenchyma. Journal of the Science of Food and Agriculture, 48, 455–466. Khan, A. A., & Vincent, J. F. V. (1993). Compressive stiffness and fracture properties of apple and potato parenchyma. Journal of Texture Studies, 24(4), 423–435. Khan, A. A., & Vincent, J. F. V. (1996). Mechanical damage induced by controlled freezing in apple and potato. Journal of Texture Studies, 27(2), 143–157. Kim, N. K., & Hung, Y. C. (1994). Freeze-cracking in foods as affected by physical properties. Journal of Food Science, 59(3), 669–674.

797

Kunzek, H., Kabbert, R., & Gloyna, D. (1999). Aspects of material science in food processing: Changes in plant cell walls of fruits and vegetables. Zeitschrift fur Lebensmittel-Untersuchung und -Forschung, 208(4), 233–250. Levigne, S., Thomas, M., Ralet, M. C., Quemener, B., & Thibault, J. F. (2002). Determination of the degrees of methylation and acetylation of pectins using a C18 column and internal standards. Food Hydrocolloids, 16(6), 547–550. Lewicki, P. P., & Porzecka-Pawlak, R. (2005). Effect of osmotic dewatering on apple tissue structure. Journal of Food Engineering, 66(1), 43–50. Marani, C. M., Agnelli, M. E., & Mascheroni, R. H. (2007). Osmo-frozen fruits: Mass transfer and quality evaluation. Journal of Food Engineering, 79(4), 1122–1130. Marti, J., & Aguilera, J. M. (1991). Effects of freezing rate on the mechanical characteristics and microstructure of blueberries and wild blackberries. Revista De Agroquimica Y Tecnologia De Alimentos, 31(4), 493–504. Martinez-Monzo, J., Martinez-Navarrete, N., Chiralt, A., & Fito, P. (1998). Mechanical and structural changes in apple (VarGranny Smith) due to vacuum impregnation with cryoprotectants. Journal of Food Science, 63(3), 499–503. Mastrocola, D., Pittia, P., & Lerici, C. R. (1996). Quality of apple slices processed by combined techniques. Journal of Food Quality, 19(2), 133–146. Mazur, P. (1984). Freezing of living cells: Mechanisms and implications. American Journal of Physiology–Cell Physiology, 247(3), C125–142. Mehinagic, E., Royer, G., Bertrand, D., Symoneaux, R., Laurens, F., & Jourjon, F. (2003). Relationship between sensory analysis, penetrometry and visible-NIR spectroscopy of apples belonging to different cultivars. Food Quality and Preference, 14(5–6), 473–484. Muhlethaler, K. (1967). Ultrastructure and formation of plant cell walls. Annual Review of Plant Physiology, 18(1), 1. Otero, L., Martino, M., Zaritzky, N., Solas, M., & Sanz, P. D. (2000). Preservation of microstructure in peach and mango during high-pressure-shift freezing. Journal of Food Science, 65(3), 466–470. Phan, P. A., & Mimault, J. (1980). Effects of freezing and thawing on fruit. Evaluation of some texture parameters and exudate. Relation to fruit quality.. International Journal of Refrigeration, 3(5), 255–260. Robbers, M., Singh, R. P., & Cunha, L. M. (1997). Osmotic–convective dehydrofreezing process for drying kiwifruit. Journal of Food Science, 62(5), 1039–1042. Robert, P., Devaux, M. F., & Bertrand, D. (1996). Beyond prediction: Extracting relevant information from near infrared spectra. Journal of Near Infrared Spectroscopy, 4, 75–84. Rojas, A. M., Gerschenson, L. N., & Marangoni, A. G. (2001). Contributions of cellular components to the rheological behaviour of kiwifruit. Food Research International, 34(2–3), 189–195. Serra, J. (1982). Image analysis and mathematical morphologies. London: Academic Press. Sousa, M., Canet, W., Alvarez, M., & Tortosa, M. (2006). Effect of processing on the texture and structure of raspberry (cv. Heritage) and blackberry (cv. Thornfree). European Food Research and Technology, 223(4), 517–532. Sterling, C. (1968). Effect of low temperature on structure and firmness of apple tissue. Journal of Food Science, 33(6), 577–580. Thibault, J. F. (1979). Automatisation du dosage des substances pectiques par la méthode au méta-hydroxydiphenyl. Lebensmittel-Wissenschaft und-Technologie, 12, 247–251. Tregunno, N. B., & Goff, H. D. (1996). Osmodehydrofreezing of apples: Structural and textural effects. Food Research International, 29(5–6), 471–479. Waldron, K. W., Smith, A. C., Parr, A. J., Ng, A., & Parker, M. L. (1997). New approaches to understanding and controlling cell separation in relation to fruit and vegetable texture. Trends in Food Science and Technology, 8(7), 213–221. Zhang, M., Duan, Z.-H., Zhang, J.-F., & Peng, J. (2004). Effects of freezing conditions on quality of areca fruits. Journal of Food Engineering, 61(3), 393–397. Zykwinska, A. W., Ralet, M. C. J., Garnier, C. D., & Thibault, J. F. J. (2005). Evidence for in vitro binding of pectin side chains to cellulose. Plant Physiology, 139(1), 397–407.