Project Topics
Get Complete Project Material File(s) Now! »
Sampling and acquisition
Two days after MRI measurements, samples were taken from each fruit for macrovision imaging. The sampling protocol consisted of cutting a 1 cm slice, corresponding to twice the MRI virtual slice, at the equator of the fruit. Two 1 cm wide rectangular samples were then taken from the cortex (see Figure 2-2B, gray line) and kept for two months at 4°C in a solution composed of 85:10:5% (v/v) ethanol 96%, formaldehyde 37% solution and acetic acid. Before imaging, tissue was rehydrated by placing the sample in successive ethanol baths (70%, 50% and 30%). Finally, 200 μm thick sections were cut from the middle of each sample using a vibrating blade microtome (MICROM, HM 650V, Microm International GmbH, Walldrof, Germany). They were degassed for 30 s under mild vacuum to remove remaining air bubbles. Sections were then imaged under water using a macro-vision system (Devaux et al., 2008) comprising a CCD camera (Sony XC 8500 CE, Alliance Vision, Montélimar, France) fitted with a 50 mm lens (f 1:1.8 Nikon) and a 20 mm extension tube. Samples were back-lit using a fiber-optic ring-light supplied by Polytec (Pantin, France). The camera and lens were adjusted to observe a 10.7 mm x 14.4 mm area and images were digitized in 1620 x1220 pixels (pixels of 3.6×3.6 μm²).
Image processing
Two to four images per fruit were selected for image analysis. Images were considered for visual texture analysis according to both cell morphology and arrangement as they did not allow segmentation of cells. Image pixels were coded from 0 (black) to 255 (white). Gray level granulometric methods were applied using MatLab software to extract overall information concerning distribution of cell dimensions (Devaux et al., 2008) on ROIs of the inner and outer regions of the cortex (see Figure 2-2B). ROIs were selected (MatLab software) to correspond to MRI ROIs. Outer ROIs began at 1.5 mm from the cuticle and ended at 40% of the image length. Inner ROIs began at 60% of the image length from the cuticle and were 80% of the outer ROI length wide. By applying successive morphological closings on ROIs, dark objects on the image smaller than the structuring element were filled by the mean pixel value under the mask. A curve Vi was constructed by measuring the sum of gray levels after each closing step versus the size of the structuring element. The raw curve was normalized according to the sum of the initial (Vinitial) and final (Vfinal) gray levels and written as: gi = (Vi−Vi+1)/(Vinitial−Vfinal) where g(i) is the percentage of variation in gray level for the ist step. The maximum size of the structuring element was set at 200 pixels, corresponding to 726 μm.
Apparent microporosity and its impact on T2
The heterogeneity of the apparent microporosity in apple tissues of Fuji and Ariane cultivars is depicted in Figure 2-4. Vascular bundles were the least porous parts of the fruit and were situated in a low microporosity region, highlighting the border between the core and the cortex tissues. Microporosity increased from this border to the outer cortex region and then decreased near the cuticle. For example, apparent microporosity values were about 15% of gas in the tissue for the core, from about 30 to 40 % for the cortex and about 10% in the vascular bundles of Ariane fruit. The same pattern was observed for the Jonagored cultivar (Supplementary Figure 2-8). The apparent microporosity measurements on ROIs from the outer and inner cortex are presented in Table 2-1 and confirm results from apparent microporosity maps. Indeed, apparent microporosity in the outer cortex region was higher than in the inner cortex region. The result of a MANOVA-test (Table 2-2) showed this variation to be significant.
Tissue histology and MRI measurements
Figure 2-5 shows the macrovision images of cortex of Fuji and Ariane cultivars with zooms on the regions corresponding to the ROIs used for MRI analyses in the outer and inner cortex tissue. The cortex tissue was heterogeneous in terms both of cell size and shape. Under the cuticle, an approximately 2 mm wide region was composed of small cells with sizes increasing with distance from the cuticle. This region matched the first peripheral pixels of T2 and microporosity maps (Figure 2-3 and Figure 2-4) characterized by lower apparent microporosity and T2 values than the rest of the cortex tissue. As the cells were small, there was more light diffusion on the image and the region appears whiter compared to other regions. The outer cortex tissue was characterized by round cells of roughly 180 μm. For most fruit studied, the cells tended to elongate in a direction perpendicular to the cuticle when the distance from the cuticle increased. Some inner cortex tissues included non-elongated cells (Figure 2-5D). The elongated cells observed close to the vascular bundles were often oriented (Figure 2-5A, B) in the direction of the closest vascular bundle. The border between the cortex and core tissues (Figure 2-5A, B and C, arrows) was a compact tissue with smaller cells, appearing slightly brighter as in mature fruit (Esau, 1977; Fisk, 1962). These findings agreed with previous studies on apple parenchyma histology, showing the tissue cell morphology dependency on location (Bain and Robertson, 1951; Khan and Vincent, 1990; Schotsmans et al., 2004). They highlighted heterogeneity of the higher tissue structure of the inner parenchyma compared to the outer parenchyma tissue for fruit of the same size and cultivar. When comparing fruit of different sizes, large fruit appeared visually to be richer in intercellular spaces in the outer part of the pericarp (Figure 2-5). The cell shape of small fruit tended to be more heterogeneous in the inner parenchyma tissue.
The cell morphology of the Jonagored cultivar was similar and varied in the same way between the outer and inner regions (Supplementary Figure 2-9). The image texture algorithms used to analyze macrovision images yielded size distributions of dark objects, which were used to estimate mean cell sizes, assuming that dark objects were all cells. Cell volume was estimated by computing the estimations of horizontal and vertical cell sizes (Equation 2-3). A MANOVA revealed a significant variation in cell volume between the inner and the outer cortex tissues (F = 18.40, p = 0.000), the outer cortex cells being larger than the inner cortex cells, and between the outer cortex of fruit of different sizes (F = 10.17, p = 0.000). The tissue heterogeneity within apple cultivars reduced the differences between the different apple varieties. The heterogeneity was greater in the inner cortex tissue due to cell elongation and the presence of vascular bundles. Further analysis was therefore focused on the outer cortex. As the thickness of the section was 200 μm (although cell size varied), fruit with small cells had more cell layers, which resulted in underestimation of cell size for such fruit (Figure 2-5C). However, underestimation was limited as the granulometric analysis was based on gray levels.
Figure 2-6 depicts cell volume of the outer cortex region of Fuji cultivar as a function of fruit weight. As the relationship between weight (w) and volume (V) was linear ( , ), fruit weight was used instead of estimation of fruit volume, the measurement of which was less precise.
Acquisition
MRI measurements were carried out with a 1.5 T MRI scanner (Magnetom, Avanto, Siemens, Erlangen, Germany). The fruit were maintained at a controlled temperature (20 ± 0.5 °C) during acquisitions; they were placed in the Faraday cage of the MRI scanner ten hours before measurement to stabilize the temperature. The median plane A 2 cm B of each fruit (transverse section at middle height of fruit) was imaged with a pixel size of 1.19 mm x 1.19 mm, slice thickness of 5 mm and repetition time of 10 s. The choice of spatial resolution was governed by the trade-off between signal to noise ratio (SNR) of the images and spatial heterogeneity of the apple fruit. Two MRI sequences were performed:
(i) a multi spin echo (MSE) sequence, with the parameters chosen according to (Adriaensen et al., 2013): non-selective refocusing pulse, inter-echo spacing (ΔTE) of 7.1 ms, bandwidth 260 Hz/pixel, 1 scan and 512 echoes per echo train for multi exponential T2 estimation.
(ii) a multi gradient echo (MGE) sequence, with first echo time (TE1) of 2.8 ms and ΔTE of 1.6 ms, bandwidth 1096 Hz/pixel, 12 echoes and 2 scans for T2* estimation. The parameters were determined in a preliminary study as the trade-off between the shortest echo times (TE1 and ΔTE) and the SNR ratio.
The SNR of the images obtained by the MSE and MGE sequences at the shortest echo time were approximately 120 and 80, respectively.
Table of contents :
1. Etude bibliographique et objectifs
I. La pomme
1) Anatomie
2) La texture et son évolution, vues par les mesures instrumentales
II. RMN bas champ, IRM
1) Théorie
2) Méthodologie
3) Mécanismes de relaxation
4) Relaxation dans les tissus végétaux
5) Applications en biologie végétale
III. Contexte et objectifs de l’étude
2. MRI investigation of subcellular water compartmentalization and gas distribution in apples
I. Introduction
II. Materials and Methods
1) Fruit and sampling
2) MRI
3) Macro-vision imaging
4) Statistics
III. Results and discussion
1) Spatial distribution of T2
2) Apparent microporosity and its impact on T2
3) Tissue histology and MRI measurements
IV. Conclusion
V. Supplementary data
3. Analysis of the dynamic mechanical properties of apple tissue and relationships with the intracellular water status, gas distribution as measured by MRI, histological properties and chemical composition
I. Introduction
II. Materials and Methods
1) Fruit
2) MRI
3) Mechanical measurements
4) Macrovision Imaging
5) Chemical analyses
6) Statistics
III. Results and discussion
1) Macro-vision imaging
2) Mechanical measurements
3) MRI measurements
4) Chemical analyses
5) Principal component analysis and overall correlations
6) Cell wall polysaccharide chemistry
IV. Conclusion
V. Supplementary data
4. Effect of storage on the NMR signal of the different tissues of apple fruit
I. Material and Methods
1) Fruit
2) MRI
II. Results
1) Spatial distribution of monoexponential T2, proton density and apparent microporosity
2) Multi-exponential relaxation and apparent microporosity in specific regions
3) Evolution of the relaxation signal during storage
III. Discussion
IV. Conclusion
5. General discussion and perspectives
I. MRI measurements and data analysis
II. Relaxation mechanism in apple
1) Cell size
2) Chemical composition
3) Temperature
III. Apple mechanical properties and physiology
References .
