A METHOD FOR TESTING CELLULAR-LEVEL WATER CONTENT AND DISTRIBUTION IN FRUIT AND VEGETABLE TISSUES BASED ON RAMAN SPECTROSCOPY

Abstract

A method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy comprises preprocessing of samples, acquisition and preprocessing of imaging spectra, Gaussian peak-separation fitting of imaging spectra, pseudocolor imaging according to the fitting results, and visualization of distribution of water content and water bonding state at the cell level. The distribution of water content and water binding state is visualized at the cellular level in the fruit and vegetable tissues for the first time, and relatively reliable quantitative analysis results of the content of water with different bonding states according to the visualization imaging results is obtained. The new method for testing cellular level water content in fruit and vegetable tissues solves the current problem of not being able to detect cellular level water changes in fruit and vegetable processing, and has a good prospect for the research on fruit and vegetables processing.

Claims

1. A method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy, the method comprising: (1) cutting fruits and vegetables for testing into a sample; (2) placing the sample in a stage of a laser confocal microscope for imaging spectrum acquisition; specifically, selecting a cell region from the sample by an objective lens of the laser confocal microscope, then meshing the selected cell region to obtain uniformly distributed intersection points, then marking the corresponding coordinate information of each intersection point in the selected cell region, and then scanning each intersection point to obtain the corresponding Raman spectrum of water at each intersection point in the cell region; wherein the step size of the grid is 3-5 μm; (3) processing the Raman spectra obtained in step (2) for smoothing noise reduction and removing fluorescence background, and then performing the Gaussian peak fitting; wherein five sub-peaks at 3000-3800 cm.sup.−1 and two or three sub-peaks at 2700-3000 cm.sup.−1 are obtained for each Raman spectrum; (4) summing up the areas of the five sub-peaks at 3000-3800 cm.sup.−1 after the peak fitting of each Raman spectrum to obtain the corresponding water content A at the intersection point in the cell region, and then determining the bonding state of the corresponding water molecules at the intersection point according to the ratio R of the peak area of the sub-peak centered at 3410-3440 cm.sup.−1 to the peak area of the sub-peak centered at 3200-3220 cm.sup.−1; and (5) by combining with the coordinate information of each intersection point, using the corresponding water content A and ratio R at all the intersection points as pixels for pseudocolor imaging to obtain the distribution of water content and water bonding state at the cell level in the fruit and vegetable tissues.

2. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 1, wherein the laser used for the imaging spectrum acquisition in step (2) is a 532 nm laser, wherein the size of the grid is 1200 gr/mm, the Hole is 500, and the scanning range is 2700-3800 cm.sup.−1; the acquisition conditions are as follows: the acquisition time is 3-5 s, the accumulation times are 2-3 times, and the laser energy attenuation is 25% to 50%.

3. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 1, wherein the imaging spectrum acquisition in step (2) is performed at a depth of 50-100 μm.

4. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 1, wherein the “Gaussian peak fitting” in step (3) is carried out by using the Matlab software with the Peakfit function.

5. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 4, wherein in step (3), the “smoothing noise reduction” adopts the Savitzky-golay convolution smoothing algorithm in the Matlab software; and the “removing fluorescence background” adopts an adaptive iteratively reweighted penalized least squares background subtraction algorithm.

6. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 4, wherein the “Gaussian peak fitting” in step (3) is a fixed-peak-position Gaussian curve fitting, that is, the Gaussian iterative curve fitting algorithm is used to perform the Gaussian curve fitting on the spectrum in the case of fixed peak position; the method for determining the peak position is as follows: randomly selecting fifty Raman spectra, and using the Peakfit software to perform iterative peak fitting to decompose each Raman spectrum into seven or eight sub-peaks; then averaging the peak position information of the sub-peaks obtained from the fifty Raman spectra to get the average peak position information of the seven or eight sub-peaks, which can be used as the peak-position of the fixed-peak-position Gaussian peak separation.

7. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 4, wherein the size of the cell region in step (2) is determined by the size of the cells, and the average diameter of the conventional fruit and vegetable cells is 100-300 μm; the magnification of the objective lens is 10×.

8. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 4, wherein the fruits and vegetables in step (1) are one of apples, potatoes, grapes, pears and cabbage stems.

9. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 8, wherein the sample in step (1) is peeled fruits and vegetables; the shape of the sample is of a disc, having a size of diameter×thickness=12 mm×2 mm; before and during the test, the sample is stored in a quartz chamber that, sealed with a quartz cover glass of 0.3 mm in thickness, has a temperature of 2° C. to 10° C. and a humidity greater than 80%.

10. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 4, wherein the “pseudocolor imaging” in step (5) is carried out by using the Matlab software with the Pcolor and Colormap functions, wherein Shading interp is used for shading.

11. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 2, wherein the “Gaussian peak fitting” in step (3) is carried out by using the Matlab software with the Peakfit function.

12. The method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy according to claim 3, wherein the “Gaussian peak fitting” in step (3) is carried out by using the Matlab software with the Peakfit function.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0035] FIG. 1 shows the hydrogen bonding modes of water molecules.

[0036] FIG. 2 shows a Raman spectrum and the peak fitting result of water molecules with different hydrogen bonding modes at 293 K and 0.1 MPa.

[0037] FIG. 3 is an optical microscopy image of apple tissue at a depth of 50 μm in Example 1 of the present invention under a 10× objective lens.

[0038] FIG. 4 is an optical microscope image of a single cell in a test region at a depth of 50 μm of the apple tissues tested in Example 1 of the present invention (an enlarged view of the central test region in FIG. 3).

[0039] FIG. 5 shows all the raw Raman spectra of each intersection point in the cell region of the apple tested in Example 1 of the present invention after being scanned.

[0040] FIG. 6 is the raw, smoothed and baseline-corrected Raman spectrum of a certain original spectrum obtained in Example 1 of the present invention (apple tissue).

[0041] FIG. 7 shows the sub-peaks and total peak of a certain preprocessed spectrum obtained by the FPGICF algorithm in Example 1 of the present invention (apple tissue).

[0042] FIG. 8 shows the distribution of the determination coefficients (number of spectra) of the preprocessed spectrum for the Gaussian peak fitting in Example 1 of the present invention (apple tissue).

[0043] FIG. 9 is a histogram of the distribution of the determination coefficient of the preprocessed spectrum for the Gaussian peak fitting in Example 1 of the present invention (apple tissue).

[0044] FIG. 10 shows the cellular level water content distribution of the apple tissue in Example 1 of the present invention.

[0045] FIG. 11 shows the cellular level location distribution of water with different bonding state of the apple tissue in Example 1 of the present invention.

[0046] FIG. 12 shows the inversion results of the apple tissue obtained by the NMR method.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The present invention will be further described in detail below with reference to examples and drawings, but the embodiments of the present invention are not limited thereto.

Example 1

[0048] A method for testing the distribution of content and state of water in plant tissues by Raman spectroscopy is provided, comprising the following steps:

[0049] (1) Using a sampler to take a 12 mm×15 mm (diameter×height) sample column from an apple in the radial direction, then using a self-made slicing device to cut the sample column into slices of 2 mm in thickness, then placing the slices immediately in a quartz chamber with constant temperature and humidity, and then sealing the chamber with a quartz cover glass of 0.3 mm in thickness; setting the humidity of the chamber to 80% and the temperature to 4° C., and then placing the chamber on the stage of a laser confocal Raman microscope for testing.

[0050] (2) Aligning the 10× objective lens of the laser confocal Raman microscope to observe the sample, and adjusting the focal length to obtain a clear microscopic image of the tissue structure; selecting cells with a complete structure as the measurement regions (200 μm×200 μm in size, as shown in FIG. 4), and then dividing the selected regions into a grid with a step size of 5 μm to obtain a total of 1681 intersection points. The light source of the laser confocal Raman microscope, as a point light source, scanned from top to bottom and from left to right along the intersection points of the grid to acquire the Raman spectrum generated at each intersection point, so each spectrum had the corresponding coordinate information in the measurement region; for the spectrum acquisition, the spectral range was set to 2700-3800 cm.sup.−1, a 532 nm laser was used, the size of the grid was 600 gr/mm, the Hole was 500, the laser energy attenuation was 25%, the acquisition time was 3 s, the accumulation times were three times, and the acquisition depth was 50 μm; the time required to complete all the scanning was about 3 h. All the original Raman spectra acquired were shown in FIG. 5.

[0051] (3) All the acquired original Raman spectra were smoothed by the Savitzky-golay smoothing method and removed the baseline by the airPLS method; the Savitzky-golay Smooth function in the Matlab software was used for smoothing, wherein the window size of the Savitzky-golay algorithm used to smooth the spectrum was set to 15, and the degree of polynomial in the Savitzky-golay algorithm was 1; airPLS (adaptive iteratively reweighted penalized least squares), used to remove the baseline of the spectrum, which could well remove the baseline of the spectrum and retain the effective information of the spectrum to the maximum extent; in the airPLS algorithm, the parameter) was set as 10e.sup.12, the order of the differential penalized term was 3, the weight anomaly ratio at the start and end of the spectrum was set to 0.08, the asymmetry parameter at the start and end was 50, and the maximum number of iteration times was set to 10. FIG. 6 shows the spectrum of a certain original Raman spectrum after being preprocessed by smoothing and removing the baseline.

[0052] (4) Performing the Gaussian peak fitting on the preprocessed Raman spectra. Because the components of the sample was complex, it would have a certain impact on the spectrum, which made the peak positions of the spectrums different after the Gaussian peak fitting, not conducive to the subsequent data processing; therefore, the method of fixed-peak-position peak-separation fitting was used to fit all the spectra. The method to determine the peak positions was as follows: Randomly selecting fifty preprocessed Raman spectra, then using the software Peakfit v4.12 (Seasolve software Inc.) to decompose the spectrum into several sub-peaks, and obtaining 7 sub-peaks in the spectral range of 2700-3800 cm.sup.−1 according to the peak fitting determination coefficient and the number of iterations, wherein five sub-peaks in the spectral range of 3000-3800 cm.sup.−1 were the OH stretching vibration peaks; averaging the peak positions of the 7 sub-peaks of the fifty preselected spectra to obtain the peak position information of the fixed-peak-position peak fitting; performing peak fitting on all the preprocessed spectra through the Matlab software with the Peakfit function according to the obtained peak position information, so as to obtain the peak position, peak height, peak width and peak area of the sub-peaks of all the spectra after the peak separation, as well as the determination coefficient of the fixed-peak-position peak fitting of each spectrum, the relative error between the fitted spectrum and the original spectrum, the fitting residual error, and other information. The peak fitting results of a certain Raman spectrum are shown in FIG. 7 and Table 1.

TABLE-US-00001 TABLE 1 The peak position, peak height, peak width and peak area of the sub-peaks of a certain Raman spectrum after peak separation Peak Peak Peak Peak Sub-peak No. position height width area 1 2895 148.9 73.2 11636.29 2 2950 249.8 83.7 22228.41 3 3053 30.61 69.17 2178.08 4 3219 978 218.8 227970.64 5 3428 1159 223 270103.66 6 3583 134.4 121.4 17502.96 7 3633 81.28 76.72 6648.71

[0053] (5) Summing up the peak areas of the five sub-peaks at 3000-3800 cm.sup.−1 after the peak fitting to obtain the sum A of the peak areas of each Raman spectrum at 3000-3800 cm.sup.−1, then the A was the content of water molecules. In the five sub-peaks at 3000-3800 cm.sup.−1, the ratio R of the peak area of the third sub-peak (centered at 3428 cm.sup.−1) to the peak area of the second sub-peak (centered at 3219 cm.sup.−1) could be used to determine the bonding state of water molecules; the number of the peak-fitting spectra of all the spectra was 1681. The fitting determination coefficient obtained was shown in FIG. 8, and the histogram of the distribution of the determination coefficient was shown in FIG. 9.

[0054] (6) combining the content of water molecules A and the corresponding coordinate information to perform pseudocolor imaging by using the Matlab software as shown in FIG. 10, to obtain the water distribution map at the cellular level in the selected range. The larger the value of A was, the higher the water content at this point was; while the smaller the value of A was, the lower the water content at this point was. Besides, the ratio R could be used as the pixel to judge the hydrogen bonding state of water at each point, as shown in FIG. 11. The larger the value of R was, the lower the bonding degree was; while the smaller the value of R was, the higher the binding degree was. Thus, the visualization of the distribution of water content and water binding state at the cellular level in fruit and vegetable tissues could be realized. The Matlab pseudocolor imaging used its own Pcolor and Colormap functions; and Shading interp was used for shading.

[0055] According to the value of R of water in FIG. 11, water with R value smaller than 1.2 was defined as bound water, water with R value greater than 1.4 was defined as free water, and water with R value between 1.2 and 1.4 was defined as immobilized water. The content of water with different states was calculated based on the corresponding coordinate information according to the water content in FIG. 10. Taking three cylinders in each apple and then three slices in each cylinder, then testing all the slices by the method of this application to obtain 9 groups of data, and then analyzing and calculating these data; averaging the values of free water, bound water and immobilized water, respectively, and comparing these average values with the results measured by the NMR method (measuring three identical cylinders on the same apple and averaging the values, i.e., the test sample was the same as the sample used in the method of this application) and the Marlin Chick experiment (two groups of the same apple were tested, and three identical cylinders were obtained from each group, i.e., the test samples were the same as the samples used in the method of this application, one group used for the sucrose dipping test, the other used for the experiment of water measurement by drying, and then the obtained results were averaged), respectively. The results were shown in Table 1.

[0056] The NMR method was as follows:

[0057] {circle around (1)} Taking three sample columns with a size of 12 mm×15 mm (diameter×height) from different parts of the same apple along the radial direction, then placing the sample columns in an NMR test tube with a diameter of 20 mm, and then sealing the tube with a parafilm and placing in a refrigerator to stay at a constant temperature of 4° C. for 2 h.

[0058] {circle around (2)} A low-field NMR instrument was used in this experiment; before the test, a standard sample was used to calibrate the instrument, and an FID sequence was used to find the center frequency, then a CPMG test sequence was selected, and the testing parameters are as follows: SW=200 kHz, RFD=0.02 μs, RG1=5, DRG1=3, DR=1, PRG=0, NS=3, TW=8000 ms, TE=0.5 ms, NECH=2500.

[0059] {circle around (3)} Putting the sample into a magnet box, and selecting the cumulative sampling to collect the T.sub.2 relaxation signal of the sample; after the sampling, the data would be automatically saved in the database; selecting the measured data to perform data inversion to obtain the final results. In the inversion results, the peak of the T.sub.2 relaxation peak in the range of 100-1000 ms was considered as the peak of free water in the sample, and the corresponding percentage of the peak area was the percentage of free water; the peak of the T.sub.2 relaxation peak in the range of 10-100 ms was considered as the peak of immobilized water, and the corresponding percentage of the peak area was the percentage of immobilized water; the peak of the T.sub.2 relaxation peak in the range of 0-10 ms was considered as the peak of bound water, and the corresponding percentage of the peak area was the percentage of bound water. The results were shown in FIG. 12.

[0060] The Marlin Chick experiment was performed as follows:

[0061] {circle around (1)} Taking six sample columns with a size of 12 mm×15 mm (diameter×height) from different parts of the same apple along the radial direction; cutting three of the sample columns into sample discs with a thickness of 2 mm, then respectively placing the sample discs in three weighing dishes with a known mass of m.sub.0, then respectively weighing the total mass m.sub.1 of the three weighing dishes, then placing the three weighing dishes in an oven to dry at 105° C. for 10 h to a constant weight, and then weighing the total mass m.sub.2; calculating the water content of the tissue according to the following formula:

[00001]$\mathrm{water}\mathrm{content}\mathrm{of}\mathrm{tissue}\left(\%\right)=\frac{\left(m_1-m_2\right)}{\left(m_1-m_0\right)}\times 100.$

[0062] {circle around (2)} Also cutting the other three sample columns into sample discs with a thickness of 2 mm, and respectively placing the sample discs in the other three weighing dishes with the known mass of M.sub.0 to obtain the total mass M.sub.1; using a pipette to add a sucrose solution with a mass percent concentration of 60% respectively into the three weighing dishes, then gently shaking the weighing dish to make the solution and sample mixed uniformly, and then weighing its mass M.sub.2.

[0063] {circle around (3)} Placing the weighing dish on a rotary oscillator to oscillate for 6 h, and setting the rotation rate of the oscillator so that the solution in the weighing dish could be gently shaken in one direction without spilling out of the weighing dish.

[0064] {circle around (4)} After the oscillation, fully shaking the solution, then using a pipette to drop 200 μL of the sample on the ground glass surface of the Abbe refractometer, and then screwing the prism tightly; measuring the sugar concentration D.sub.2 of the solution at 20° C., then measuring the original sugar concentration D.sub.1, and then calculating the free water content (%) in the tissue according to the following formula:

[00002]$\mathrm{free}\mathrm{water}\mathrm{content}\mathrm{of}\mathrm{tissue}\left(\%\right)=\frac{\left(M_2-M_1\right)\times \left(D_1-D_2\right)}{\left(M_1-M_0\right)\times D_2}\times 100.$

[0065] Thus, the percentage (%) of the bound water content of the total water content=(water content of tissue−free water content of tissue)×100/water content of tissue, this percentage is the bound water content; and the proportion of free water in total water content=1−bound water content of the total water.

[0066] (7) Table 2 shows the comparison between the contents of bound water and free water measured by the method of this application and the contents of bound water and free water measured by the Marlin Chick experiment and NMR.

TABLE-US-00002 TABLE 2 The contents of bound water and free water in an apple measured by different test methods Water state Free Immobilized Bound Test method water water water The method of 8.8 ± 2.1% 83.1 ± 4.7% 8.1 ± 1.8% this application NMR method 8.5 ± 1.7% 85.6 ± 6.2% 5.9 ± 1.3% Marlin Chick method 89.9 ± 7.2% 10.1 ± 5.1% 

[0067] The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited thereto, and any other alterations, modifications, replacements, combinations and simplifications should be equivalent substitutions and included in the scope of protection of the present invention.