Non-targeted Metabolomic Profiling of Maize Landraces ( Zea mays L . ) Combined with Chemometric Tools

Grain samples of maize landraces were collected and subjected to Fourier Transform Infrared spectroscopy (FTIR) analysis combined with chemometric tools in other to discriminate them regarding the chemical composition. Principal component analysis (PCA) and hierarchical clustering (HCA) were applied on selected peaks of the spectral data. The most important chemical groups found in all maize landrace samples were monoterpenes, sesquiterpenes, tetraterpenes, aminoacids, polysaccharides, lipids and proteins. Trace signals of secondary metabolites were also found in cultivars according to year of harvest. Original Research Article Uarrota et al.; IJBCRR, 20(1): 1-9, 2017; Article no.IJBCRR.35832 2


INTRODUCTION
Commonly referred to as maize or corn, Zea mays ssp.mays is one of the world's most important crop plants, achieving a multibillion dollar annual revenue.In addition to its agronomic importance, maize has been a keystone model organism for basic research for nearly a century [1][2].According to Dwivedi [1], the current industrial agriculture system may be the single most important threat to maize biodiversity.A serious consequence of biodiversity loss is the displacement of locally adapted maize landraces with adaptation traits to future climates by mono-cropping with genetically uniform hybrids and improved cultivars.
Maize landraces represent heterogeneous, local adaptations of domesticated species, and thereby provide genetic resources that meet current and new challenges for farming in stressful environments.These local ecotypes can show variable phenology and low-to-moderate edible yield, but are often highly nutritious.The main contributions of maize landraces to plant breeding have been traits for more efficient nutrient uptake and utilization, as well as useful genes for adaptation to stressful environments such as water stress, salinity, and high temperatures [1,[3][4][5].
Systematic maize landraces evaluation may define patterns of diversity, which will facilitate identifying alleles for enhancing yield and abiotic stress adaptation, thus raising the productivity and stability of staple crops in vulnerable environments [1,[3][4][5].In other hand, regarding the evaluation tools, metabolomics has recently been claimed as a promising concept and valuable tool in biotechnology, given its extensive range of applications in functional genomics and, more globally, in the characterization of biological systems [6].Indeed, the possible tasks include studying metabolic systems, measuring biochemical phenotypes, understanding and reconstructing networks and discriminating between samples [6].In the targeted metabolomics approach, specific metabolites of known identity are profiled; good quantitative precision is typically obtained.On the other hand, untargeted metabolomics aims to simultaneously measure as many metabolites as possible in a biological specimen [7].
Taking into account that maize landraces are currently being replaced by new improved cultivars with a narrow genetic basis in Brazilian agriculture and with a purpose to add value of those important genetic materials, the present study was designed with the main goals of investigating the chemical composition of maize landraces using non-targeted metabolomic approach (Fourier Transformed Infrared Spectroscopy-FTIR) combined with chemometric tools (PCA and HCA)as a rapid diagnostic tool to ascertain changes in chemical composition.

Maize Landraces
Maize landraces (Zea mays L.) were kindly provided by small farmers of the Anchieta County (the western part of Santa Catarina state, southern Brazil, 26°31′11″ S, 53°20′26″ W) and produced under agro-ecological management practices during three different harvests (2007,2008,2009) Table 1 presents the details of the cultivars used each year).Maize grains were collected, dried and crushed to obtain fine powder for later analysis by Fourier transform infrared spectroscopy (FT-IR).

Flour Sample Preparation
Maize grains (250g -dry weight) were selected from each maize genotype and ground to pass a 0.5mm sieve using a laboratory cyclone mill (MB Braeski C.Q).

Fourier Transform Infrared Spectroscopy
FTIR spectra of maize flours were recorded in a Bruker IFS-55 (Model Opus v. 5.0, Bruker Biospin, Germany) spectrometer with a DTGS detector equipped with a golden gate single reflection diamond attenuated total reflectance (ATR) accessory (45° incidence-angle).A background spectrum of the clean crystal was acquired and samples (100 mg) were spread and measured directly after pressing them on the crystal.The spectra were recorded at the absorbance mode from 4000 to 500 cm .Five replicate spectra were collected for each sample.For pre-processing, the spectra were normalized, baseline-corrected in the region of interest by drawing a straight line before resolution enhancement (k factor of 1.7).The assumed line shape was Lorentzian with a half width of 19 cm −1 [8][9][10][11].

Chemometric Analysis
FT-IR spectra were acquired and subjected to chemometric analysis using the R software [12].Non-targeted metabolomics was used to find variations in the chemical composition of different cultivars, cultivated under different harvests and environmental conditions.FT-IR spectra were acquired in transmittance mode and then transformed to absorbance mode, as spectra objects in the ChemoSpec package [13] using the equation 1.
(Eq. 1) where A is the absorbance and T the transmittance The spectra object was then converted to a hyperSpec [14] object using the bridging package hyperChemoBridge [15].Before applying chemometric tools, the spectra were baseline corrected, intensity vector normalized and smoothed for subsequent analysis of PCA and HCA [6] for sample classification according to their biochemical status.

Peak Selection and Multivariate Analysis
Results of the assignments for the most characteristic FTIR bands found in maize landraces and related chemical group are summarized in Table 2.The most important chemical groups found in all maize landrace samples were monoterpenes, sesquiterpenes, tetraterpenes, aminoacids, polysaccharides, lipids and proteins (Table 2).
The intensity of peaks is also shown in the Fig. 1  (A-C) for maize samples from 2007, 2008 and 2009 respectively.The selected peaks found in Fig. 1 were subjected for further multivariate analysis aiming to find similarities or differences between maize samples.showed to be similar duo to their lower levels of proteins and lipids.
When FTIR data of 2007, 2008 and 2009 were subjected to hierarchical clustering (Fig. 3A-C) aiming to find similarities or differences between samples and look for important wave bands related to sample clustering, an interesting result can be observed and the main bands responsible for sample differences were found.Differences in the 2007 dataset were mainly due to the region of carbohydrates (998, 1012 and 3000).Cultivar 15 grouped alone, and the most similar in chemical composition were cultivars 7 and 17, 11 and 12 respectively (Fig. 3A).
For samples of 2008 harvest, the region of proteins, amylose and amylopectin was most important in sample dissimilarity.PR0 and MPA10 were the most dissimilar cultivars (Fig. 3B).Two visible groups were found; being the first composed by LP0 and RXE0 and second group by RX0, MG0, RJO and R8C0.For maize landraces of2009 (Fig. 3C), proteins derived the sample clustering.MPA11 and LP1 were the most dissimilar cultivars confirming that observed previously in PCA analysis.