I. Isolation of Cell Walls

II. Fractionation of Cell Wall Polysaccharides

III. Total Sugar and Uronic Acids

IV. Methyl Esterification of Uronic Acids

V. Monosaccharide Composites

VI. Linkage (methylation) Analysis

VII. Fourier Transform Infrared (FTIR)

VIII. Protocol for the screening of the UniformMu maize population with near infrared reflectance spectroscopy


VII. Fourier Transform Infrared (FTIR)

Plant growth

Arabidopsis seeds were surface-sterilized with 70% (v/v) ethanol in water, then 10% Chlorox (6.25% NaClO3) containing 0.05% Tween 20, and washed with sterile water. The seeds were incubated at 4°C for three days, then plated axenically onto 0.8% Phytoagar (Sigma) containing one-half strength Murashige and Skoog salts (Sigma), pH 5.7, supplemented with 1% sucrose, and transferred to 23°C. Plantlets were grown for two weeks at 23°C under 16 h of light each day of 240 µEm–2s–1 generated by GE F21 “pink” fluorescent lamps. The plantlets were incubated for the last two days in darkness at 23°C to lower starch content, a contaminating feature in cell wall isolations that can give rise to high absorbances in the infrared spectra.

 

Cell wall isolation

The entire shoots of two-week-old Arabidopsis plantlets were harvested into liquid nitrogen, crushed into a powder in 2-ml Eppendorf centrifuge tubes, and then suspended in 50 mM Tris[HCl], pH 7.0, containing 1% SDS, and heated to 70°C to extract protein and other non-wall components, then collected on a nylon mesh filter (47-µm square, Nitex, Briarcliff Manor, NY), and washed with water. Three 5-mm stainless steel balls were added to the tubes and the samples were homogenized at 1200 cycles min–1 in a reciprocating shaker (Geno-Grinder, SPEX Certi-Prep) for 10 min. The cell walls from the homogenate were collected on nylon mesh filters, and washed with 50% hot ethanol and then water at 70°C. The cell wall material in water was returned to the 2-ml Eppendorf tube, with the three 5-mm stainless steel balls, and homogenized a second time at 1200 cycles min–1 in the reciprocating shaker for 10 min. The cell walls were then collected on the nylon mesh and washed sequentially with water (70°C), 50% (v/v) ethanol (70°C), and water at ambient temperature. The walls were suspended in deionized water and allowed to settle. Aliquots of these wall preparations were then plated on gold-plated IR reflective slides (EZ-Spot, Spectra-Tech) and air-dried.

 

Infrared microspectroscopy

Mid-infrared spectroscopy detects, within underivatized cell walls, the vibrations of all molecular bonds for which the component atoms differ in electronegativity, including asymmetric bonds such as C-H and O-H and particular functional groups, such as esters, amides, and carboxylates. The frequencies of molecular vibrations are modified by the local environment of molecular bonds, as influenced by hydration state, the structure and conformation of the molecule, and interactions with other molecules. Therefore, the IR spectrum is characteristic of the architecture as well as the composition of the wall.

Cell walls were spotted on to 48-well gold-plated microscope slides, and IR spectra were acquired in “transflectance” mode. At least 35 spectra, from populations of 40 plantlets, were obtained for each genotype. The slides with cell wall preparations were supported on the stage of a Nicolet Continuum series microscope accessory to a 670 IR spectrophotometer with a liquid nitrogen-cooled mercury-cadium telluride detector (Thermo-Electron, Madison, WI). An area of wall (up to 125 x 125 µm), excluding vascular cells, was selected for spectral collection. In transflectance, the beam is transmitted through the wall sample, reflected off the gold-plated slide, and then is transmitted through the sample a second time. One hundred and twenty-eight interferograms were collected with 8 cm–1 resolution and co-added to improve the signal-to-noise ratio for each sample. Spectra were collected from multiple areas of each sample and then area-averaged and baseline-corrected. Baseline-corrected and area-normalized data sets of spectra are then used in the multivariate data analysis.

The data sets of mutant and wild type spectra can be downloaded as Excel files (file extension .xls) compatible with Excel95 and above. For each gene of interest in the tables listed under the gene families, there are two Excel files in the column titled “Infrared spectra”, one containing mutant spectra, and the other containing the appropriate wild type control. Both files have up to 40 columns of data. The first column represents the infrared range 800 to 1800 cm–1. Each of the other columns represents the absorbance values over that range for each infrared spectrum. To plot the spectra, use the first column as the X axis and plot the other column values on the Y axis. It is then possible to generate average spectra, digital subtraction spectra, or to carry out your own multivariate analyses (see Figure 1 for examples). We have already carried out an exploratory principal components analysis and partial least squares analysis as described below.

(Figure 1)

Multivariate data analysis

Spectra are constituted by a discrete series of measurements at defined intervals; by nature, they are multivariate. Statistical algorithms for treating multivariate data such as principal components analysis (PCA) or partial least squares (PLS) (Kemsley, 1998) can be applied to reveal molecular features that are the basis for discriminating among populations (Chen et al., 1998). PCA reduces the high dimensionality of spectral data to a smaller set of computer-derived variables, termed principal components, that together account for all of the variance in a set of (sometimes apparently similar) spectra (Kemsley, 1998). Each spectrum has an associated value for each new variable, the PC score, representing its relative distance from the mean of the population (see Figure 2).


(Figure 2)

Partial Least Squares is another spectral decomposition technique for spectroscopic data and closely related to PCA. However, in PLS, the decomposition is performed in a slightly different fashion. PLS factors are the latent variables extracted as linear combinations of the manifest independent variables (the variates of the spectrum). Ordinarily the first 3-7 will account for 99% of the variance in the populations of spectra.

We have used up to 5 of these new, computer-derived, variables, to classify spectra as mutant or wild type in pairwise comparisons of the populations of 40 spectra of each genotypic class. We have used equal numbers of spectra, minimum 35, for each mutant versus wild type comparison, and we have calculated the distance to the centre of each class (mutant or wild type) using a distance metric called Mahalanobis distance. The Mahalanobis distance is calculated using, 1, 2, 3, 4 or 5 of the PCs or PLS factors and the % of correctly classified spectra are noted. To be classified as wild type, for example, the distance of the individual spectrum to the mean of the wild type population should be less than the distance to the mean of the mutant population. The results are presented in the final two columns of the table entitled “PCs, % classified” and “PLS, % classified”. The first number in each column is the number of new variables, either principal components or PLS factors, that were used to obtain the second number, the % of correctly classified spectra. The PCs and PLS factors are ranked in order of the proportion of variance in the spectra that they account for, therefore, the fewer variables used to obtain a high % correct classification, the stronger is the phenotype. In some cases, only 2 or 3 of these variables were required to discriminate the populations of mutant spectra from wild type spectra. The percentage of correct classification reflects the probability of correctly identifying an unknown as mutant or wild type solely from an infrared spectrum. We recommend, purely on an empirical basis, that this score should be above 80% in both columns to consider that the mutant has a spectroscopic phenotype. We used WIN-DAS software (Kemsley, 1998) for PCA and Matlab 6.5.1 (The MathWorks, Inc., Natick MA, USA) for PLS.

 

Chen, L-M., Carpita, N.C., Reiter, W-D., Wilson, R.W., Jeffries, C., and McCann, M.C. (1998). A rapid method to screen for cell wall mutants using discriminant analysis of Fourier transform infrared spectra. Plant J. 8, 375-382. [link to pdf of article]

Kemsley, E.K. (1998). Discriminant Analysis of Spectroscopic Data. ( Chichester, UK: John Wiley and Sons).



Back to top