Monitoring a metabolic profile of wheat by using FTIR spectroscopy and chemometric methods — concept studies

Changes in plants under the influence of a variety of chemical and physical factors are reflected in metabolomic changes. To date, there are very few methods that would allow studying metabolic changes occurring in single cells. Spectroscopic methods es - pecially combined with the chemometrics methods are a very good tool to investigate such changes in metabolomics. Tracking changes in plants is of particular importance in industry, especially when studying how the use of fertilizers affects plants. In this paper, we present preliminary research as concept of proof to examine whether the use of FTIR (Fourier Transform Infrared Spectroscopy) helps to monitor the changes in the metabolomic profile of the plants. For preliminary research, four species of ce - reals and cuckooflower were used. In this step, it was possible to verify the differences in metabolites that are produced by plants belonging to different families. Then one species of grain was selected and subjected to eleven different physical and chemical factors. Next, the research was expanded to determine the optimal concentration of hydrogen peroxide. FTIR spectra of leaves and extracts of the plants were obtained for all experimental groups and then analyzed with the use of chemometric methods: HCA (Hierarchical Component Analysis) and PCA (Principal Component Analysis). Those methods were used to help in the interpretation of metabolic changes resulting in the plant in response to external factors.


Introduction
Metabolomics is the study and analysis of chemicals (metabolites) present in cells and tissues, which are products or intermediates formed during metabolism [1].Thanks to these studies, it is possible to detect physiological changes in the body caused by external and toxic factors, it is also possible to determine metabolites that are characteristic of particular disease syndromes, e.g.tumors.Toxicity assessment of metabolites is used in pharmacology, the appropriate dose is assessed, which will be non-toxic and its effect on physiology is easily investigated.Substances produced in plants can be divided into two basic groups: primary and secondary metabolites [2].Metabolomics could also contribute to the study of defining the product of stress metabolism or molecules that can be assigned to the acclimation of plants [3,4].
The world of plants is very diverse.Currently, it has been described around 500,000 species [5], which differ in structure, origin, climate zone in which they grew, and many other factors.Due to their diversity in structure, a large variety of metabolites -chemical compounds are present in plants.They are mainly involved in the primary and secondary metabolism of the plant, and due to this feature, they are divided into two main groups: basic substances (primary metabolites) and specific substances (secondary metabolites) [6,7].Primary metabolites are formed by primary metabolism and are present in all plants, while secondary metabolites are produced by specific enzymes associated with primary metabolic pathways.Specific substances occur only in specific plant groups in response to environmental conditions for better adaptation [8].Their metabolic pathways include photosynthesis, hydrocarbon conversion, respiration, amino acid, protein, fat synthesis, and fat utilization [8].These substances are involved in growth and plant development, the process of respiration, the process of photosynthesis, and the synthesis of hormones and proteins.They can be divided into four basic groups: carbohydrates, protein, nucleic acid, and fatty acid [7].
Secondary metabolites are produced with the participation of specific enzymes in conjunction with the main metabolic pathways.Secondary metabolites only occur in specific groups of plants in response to environmental conditions to better adapt [8].Secondary metabolites are not involved in plant growth and development [7].They are responsible for the colour of plants and protect against herbivores, microorganisms, and UV radiation.They also attract insects to pollinate the plant [9].British Nutrition Foundation has adopted a division of these metabolites into four groups: phenolic compounds, terpenes, alkaloids, and sulfur compounds [7].Within those two basic groups, there is considerable variation.Several criteria divided them into subgroups, due to the unclear meaning of many of these compounds.A very large number of lesser-known so far, substances are used in medicine, although some are harmful to human health.
Metabolomics in plant research has many applications.It enables the improvement of genetically modified plants and the development of new safer pesticides.Plant metabolomics is of particular interest because it allows for determining the range and function of both first and secondary metabolites [3,10].Fourier Transformed Infrared Spectroscopy (FTIR) in combination with chemometric methods is ideal for the analysis of metabolites.FTIR spectroscopy is a non-invasive method that is relatively fast [11], has a low percentage of error [12], and the bands on the spectra can be assigned to specific groups of metabolites produced by the plant [13], hence its widespread use in both plant and human tissue metabolomic studies.The disadvantage of FTIR-ATR (Attenuated Total Reflectance) is high water absorption and the bands that come from it have high intensity [11].Spectrum, which is obtained by spectroscopic method contains a lot of data, to reduce multidimensionally, minimization of variations between the groups, and maximize differences apply chemometric methods [12].There are many examples in the literature of FTIR being used in this type of research.The analysis of fingerprint region allows to distinguish between plants belonging to different types [14] to detect mechanical damage [12] and specify differences in phenotypic and genotypic.It also allows to determine metabolic changes, identify changes in the major metabolic pathways [13], and monitor changes in metabolic or structural [14][15][16].Yeliana et al. [17] show that with the use of FTIR spectroscopy, it was possible to distinguish the total content of phenols and flavonoids in propolis from different geographical regions in Indonesia. Hussian et al. [18] have shown that with the use of FTIR spectroscopy combined with PCA (Principal Component Analysis), it is possible to determine the metabolic changes in fruit caused by climate changes.
Plants require the proper condition for the correct development i.e. temperature, irrigation, lighting, soil pH, fertilization, and the lack of interaction with the stress factors.At a time when at least one factor is in deficiency or excess, it causes changes in the metabolism of the plant.Plant stress can be caused by physical factors such as the lack of water, intense lighting, and low temperature (i.e.abiotic stress).A distinction is made between biotic stress, which is called biological agents such as the onslaught of pathogens or the presence of Science, Technology and Innovation, 2022, 17 (1-2), 41-75 Monitoring a metabolic profile of wheat by using FTIR spectroscopy and chemometric methods… other plants.Both types of stress are associated with the uncontrolled growth of active oxygen species (ROS) in a cell, to which the hydrogen peroxide belongs.The cumulative H₂O₂ can react with a variety of biomolecules, some of these reactions are irreversible and cause their inactivation (e.g.organelles dysfunction or necrosis).There are also reversible redox reactions, which protect the biomolecules to preserve their biological functions and modulate their activity.Hydrogen peroxide affects the expression of genes, the activity of phytohormones, sugar metabolism, and photoperiod (duration of light on plants daily) to obtain long-term resilience of the plants on a wide range of pathogens.It also contributes to the activation of PR proteins (called Pathogenesis Related), the antioxidant enzyme system, and the synthesis of secondary metabolites [19].
In this work, we present the preliminary research which shows the potential of the FTIR-ATR method combined with chemometric methods to distinguish the metabolic profile of plants belonging to different rows and to monitor the changes in the metabolic profile of plants induced by physical and chemical stress factors.The extract and leaves of plants were analyzed after 14 and 28 days after seeding and treated with selected factors and measured with the FTIR-ATR method.To analyze the metabolomic profile two chemometric methods: HCA (Hierarchical Cluster Analysis) and PCA (principal component analysis) were performed.

Materials and methods
To determine metabolic profiles in plants belonging to different rows; we planted 4 species of cereals: oats, wheat, barley, rye (Granum company), and cuckooflower (POLAN company) (Figure S1 in Supporting Information).Then the wheat was subjected to both chemical and physical stress.The reference sample was watered and kept at 20°C.One of them was grown at a lower temperature of about 5°C, another was kept without access to the light (in the following description this sample was called shadow), and at a lower temperature.Those plants were watered with normal water.The other plants were watered with additions of selected substances: coffee grounds (Tchibo®), extract of black tea (Remsey Earl Grey®), 3% sucrose solution (sugar -Slodka Lyzec-zka®), 1% H₂O₂ solution (pharmaceutical laboratory AV-ENA®), 2% salicylic spirit solution (pharmaceutical plant Amara®), 1% NaOH solution (POCh), 1% H₂SO₄ solution (POCh), fertilizer used for orchids (Planta company), solid fertilizer with eight component (POLIFOSKA 8, Grupa Azoty S.A.).The plants were watered with about 50 ml of the substance, every two days.
All plants were grown for 30 days and counted after germination.From each pot (experimental group) after 14 and 28 days the leaves were harvested.Then part of the collected leaves was crushed in a mortar and subjected to extraction in waterless ethanol, another part was retained unchanged.
In addition, the effect of H₂O₂ concentration on metabolic changes was determined.The wheat was watered and treated with hydrogen peroxide solutions at concentrations of 0.5%, 1.5%, 2.0%, 2.5%, and 3.0%.From these plots, the leaves and extract were obtained.

Infrared spectroscopy measurements
Extracts and leaves were measured with the use of FTIR-ATR spectroscopy (Thermo Scientific FTIR Nicolet iS5).All spectra were collected with 34 scans with 2 cm −1 spectral resolution.For each experimental group, one spectrum of leaves and extracts was collected after 14 and 28 days.Then the spectra were preprocessed which includes: trimming scale (4000-600 cm −1 ), automatic baseline correction, and ATR correction.To facilitate the analysis and readability of images the abbreviation was used (Figure S2 in SI).

Collection of plant material
Preprocessed spectra were subjected to the RStudio program and chemometric analyses were performed (PCA and HCA) using the ChemoSpec library [20].Due to a low number of data points for each experimental group the PCA shows only a graphical representation in scores plots.
To perform chemometric analysis, recorded spectra have been encoded according to the following algorithm: the first symbol denotes the plant type (p -wheat, ooats, j -barley, r -cuckooflower, and z -rye), the second symbol indicates whether the spectrum was recorded for extract or leaves and how many days after planting they were harvested (source: e -extract, l -leaves, time: 14 and 28 days), and the last position in the name refers to the conditions with which the plantings were treated at the time of farming (w -watered, t -kept at a lower temperature, ts -kept at a lower temperature and without exposure to light, f -watered with coffee grounds, h -watered with tea extract, c -watered with 3% sucrose solution, uX -watered with X% H₂O₂ solution, n -watered with 1% NaOH solution, k -watered with 1% H₂SO₄ solution, s -watered with orchid fertilizer, and a -watered with Polifoska 8 fertilizer) (Figure S2 in SI).

Organoleptic analysis
The changes caused by different stress factors were first measured based on visual inspection.The plants were photographed after 14 and 28 days after sowing (Figure 1).Significant differences in high, density, and leaf thickness of the plants were detected.
During the breeding of seedlings, the following observations were made.The first plants that sprouted were watering with 1% H₂O₂ solution and supplied with fertilizer for orchids, whereas, the last piercedout seeds watering with 1% H₂SO₄ solution.In pot watering by salicylic spirit solution, no seed ever germinated.The plant watered with H₂O₂ solution has the coarsest leaves and the most seeds had sprouted.During the watering the hydrogen peroxide has been decomposed which resulted in loosening the ground.After some time near the pot powered by sucrose, the fruit fly gathered and the plant was the lowest in comparison with the rest.The seedling supplied by fertilizers has got long but thin leaves which caused lodging, the plant watering by H₂SO₄ solution has got also thinner leaves and some of them wrap off the soil.The quickset kept without access to sunlight has got yellow and thin leaves, and only part of the seed has to be pierced out.

Identification of metabolomic markers
Analysis of the infrared spectrum can cause a lot of trouble, mainly due to the very large number of bands that occur.An additional difficulty is the presence of water in the analyzed samples.Therefore, spectral analyses were undertaken in terms of the presence or absence of selected bands on the spectrum, which may contribute to the differentiation of the experimental groups studied.The upsent of selected bands was marked as a minus sign (-) and presented in Table 1.
As can be seen in Table 1 more differences can be withdrawn based on the spectra of leaf extracts.Comparing the spectra of the extracts of leaves and some dependencies can be seen, some of the peaks are shown in only one of these groups, which is due to that some of the compounds are only visible after the extraction while others only analyze if pure leaf.Some peak appears after 4 weeks, which is probably because

Extracts
Leaves / 14 days after sowing Leaves / 28 days after sowing Jw -barley watering with water, Ow -oats watering with water, Zw -rye watering with water, Rw -cuckooflower watering with water, Pw -wheat watering with water, Pt -wheat kept in a lower temperature, Pts -wheat kept without access to light, Pf -wheat watering with coffee grounds, Ph -wheat watering with tea extract, Pc -wheat watering with 3% sucrose solution, Pu -wheat watering with 1% H₂O₂ solution, Pn -wheat watering with 1% NaOH solution, Pk -wheat watering with 1% H₂SO₄ solution, Pswheat watering with fertilizer for orchids, Pa -wheat supplied with fertilizer Polifoska, (+) -peak occurs, (-) -the peak is absent.
the metabolites which correspond to the wavelengths have been produced by the plant in response to environmental conditions.But others disappeared because they have been used for the production of other metabolites.There are also differences between the spectra and their first derivatives (data not shown) some peaks can be seen after examination of the first derivative.Further analyses were performed on both leaves and their extracts, but due to this fact, only extracts were present in this manuscript.The assignment of spectra presented in Table 1 was made based on the literature (Table T1 in SI.) data about primary and secondary metabolites groups to wavenumbers (Table 2, and Table T2 in SI).It is not possible to accurately determine which of the individual metabolites derived data bandwidth because within groups of metabolites is enormous structural diversity and limitations in the methods of detection.The number of bands after 28 days is less than after 14 days for both the extract and leaves.Analyzing the band for all the examined plants it is noted that there are small shifts of the bands compared to a reference sample.Monitoring a metabolic profile of wheat by using FTIR spectroscopy and chemometric methods…

Analysis of different plant species
Each species of plant has a preference condition for the best growth, including a temperature suitable for germination, or the temperature in which occurs the biggest weight gain of the plant.The spectra of different cereal species and cress (Figure 2) differ only in the intensity of the peaks, so it can be concluded that the metabolites produced by these plants do not differ significantly.After 14 days, cuckooflower showed the highest intensity for extracts ranging from 3700-3000 and 1750-1500 cm −1 and wheat extract showed the lowest intensity.Barley also showed high intensity, while oats and rye are more similar to wheat.In the ranges not mentioned above, the intensities are arranged oppositely.After 28 days, the order of intensity in the ranges 3700-3000 and 1750-1500 cm −1 is as follows: barley shows the highest intensity, followed by cress.Another extract is wheat and rye extract.The least intensity is characterized by oat extract.In the ranges not mentioned above, the situation is similar to that of extracts after 14 days -the intensities of extracts are reversed.Comparing the spectrum of leaves after 14 days also the greatest intensity in the ranges 3700-3000 and 1750-1500 cm −1 shows cuckooflower, followed by oats, barley, rye, and wheat.However, after 28 days the order changes.The greatest intensity is shown by cress.Next are the leaves of rye, oats, wheat, and barley.Both for leaves after 14 days and 28 in ranges except 3700-3000 and 1750-1500 cm −1 the order of the given leaves is reversed.Based on Table T1 in SI and Table 1, it can be noted that some metabolites are not present in all plants.In cress leaves after 14 days there are no peaks at wavelengths of about 880 and 720 cm −1 .The distinguishing feature of cress from the rest of cereals is the presence of a strand in extracts after 14 days at a wavelength of about 1634 cm −1 .Cuckooflower and oat extracts show peaks after 14 days at wavelengths of about 1156, 1127, and 1121 cm −1 , these peaks are not present in the rest of the studied species.Wheat and oats are distinguished by the absence of peaks for extracts after 28 days at wavelengths of about 705 and 730 cm −1 .Among cereals, barley is the most diverse species.
HCA analysis (Figure 3A) shows a division of spectra first according to the collection time of leaves (14 -marked in green and 28 -marked in purple) then they are divided into two groups: leaves (blue frames) and extracts (yellow frames).Among the lower levels, cress leaves after 28 days appear on a separate branch, which means that they differ most from the rest samples, the same situation is visible for wheat after 28 days.At a Euclidean distance of about 100 the branch was divided into pairs: barley leaf and wheat leaf and oat leaf and rye leaf.Among the extracts, oats occur on one branch as rye, and wheat on one branch with cress.Among the samples, 14 days after planting, the dendrogram also divides into two branches at a level of about 1300.On one of them, there are extracts of wheat, oats, and rye, the last two of which are similar to each other (as for extracts after 28 days).On the second branch at a height of about 250 can be extracted pairs: barley leaf and oat leaf and wheat leaf and rye leaf.The leaves of the cress are located on a separate branch, from which it follows that it is not homogeneous with the rest samples.
In a further step, the PCA analysis was applied (Figure 3B).It is visible that all samples tested have been grouped into 5 main clusters.The first in the lower right corner of the chart (purple circle) shows the leaves analyzed after 28 days: wheat, barley, oats, and cress.The second cluster is located in the upper right corner (orange circle) and contains data on extracts after 28 days of different cereal species and cress.The next cluster is about -1500 PC1 and there are extracts after 14 days of oats, wheat, and cuckooflower.A little higher and to the right are extracts after 14 days of barley and watercress.On the other hand, the last group contains leaves of all plant species after 14 days and a cress leaf after 28 days from planting.

Metabolomics changes in plants due to chemical and physical factors
To study metabolic changes under the influence of various chemical and physical factors, wheat that grew rapidly was selected for experiments.The physical methods resulted in a reduced temperature and lack of access to sunlight.The chemical agents were: coffee grounds, tea extract, sugar solution, solution of H₂O₂, NaOH, H₂SO₄, and two types of fertilizers liquid (fertilizer for orchids) and solid (Polifoska 8).From each plant, the leaves were collected and subjected to FTIR measurements, and also their extracts after 14 days (Figure 4A) and 28 (Figure 4B) days of spreading.Monitoring a metabolic profile of wheat by using FTIR spectroscopy and chemometric methods… The visual analysis of changes was very difficult, due to the richness of different metabolites inside plants.A summary of the range from which it was possible to infer plant changes is presented in Tables 1 and 2. The biggest differences occur between the extract after 14 days after the sowing of individual seedlings.Frequency response at wavelength 1646 and 1274 cm −1 occur only in pe14k, and the band corresponding wavelength 1630 cm −1 is absent only in pe14f.The peak at 1490 cm −1 is not present in spectra of pe14k, pe14f, and pe14h.The peak at a wavelength of 1472 cm −1 is absent in the spectra of the extracts mentioned above and also in pe14t, the peak at 1464 cm −1 is not visible in pe14k and pe14ts.There can be also distinguished peaks that only occur in certain extracts of seedlings after 14 days.A peak at wavelength 1206 cm −1 occurs in pe14w, pe14k, and 1164 cm −1 is present in spectra of pe14t, while a peak at 703 cm −1 is absent only in spectra of pe14ts.Out of all examined spectra, the extracts after 14 days are the most similar to the extracts of the control seedlings.The only difference is the absence of a peak at a wavelength of 1206 cm −1 .Between the extracts of 28 days after sowing there is also a difference in the occurrence of peaks.The band at a wavelength 1630 cm −1 is absent in pe28f, pe28w, and pe28a, the peak at 1616 cm −1 did not occur at pe28s.In the spectra of pe28u1_0 the peak at 1490 cm −1 is not present, while the band at 1472 cm −1 is not noticeable in the spectra of pe28w and pe28c.In the spectra of the control sample, seedlings are not apparent in the peaks at wavelength 707 cm −1 while in the rest spectra of extracts, this band is present.Some of the peaks occur in the specific spectra of the extract after 28 days.Peak at wavelength 1206 cm −1 is evident in: pe28k, pe28h, pe28w, pe28t and pe28a.However, the peaks at wavelengths 1180 and 1161 cm −1 are demonstrable in spectra of pe28k, pe28w, and pe28t.
On the spectrum of wheat extracts after 14 days (Figure 4A) under different stressful factors can be noticed that the physical factors such as reduction of temperature and lack of light have different spectra profiles from the profiles of chemical factors.The most visible difference is a disparity in the intensity of bands.Plants subjected to physical factors accumulated large amounts of water in their leaves.In the spectra of leaves after 14 days this fact is not well seen, in both types of analysis, the bands are close to each, have similar intensity, and the spectra almost overlap each other.The spectrum of the control plants watering with water and kept at a normal temperature differs from the other, it can be seen as a separate cluster on both HCA and PCA results.It can be easily seen from both types of analysis that both chemical and physical factors, can change the metabolomic profile of the plant.
Based on the FTIR spectra measured for extracts of all seedlings treated with different chemical and physical factors, the chemometric analysis was performed.HCA plot shows (Figure 5A) that the dispersion between samples is very large.One of the most striking results is the close grouping of plants watered with different concentrations of H₂O₂, which may indicate that the concentration has no major effect on the resulting changes.There is also no clear separation between 14 and 28 days of growing seedlings on the dendrogram.PCA on the other hand shows (Figure 5B) the grouping of spectra of plants watered with water.The number of spectra subjected to PCA in this study is too low to draw exact conclusions, and for this reason, no loadings were shown and conclusions were drawn only based on graphs (scores).

Dependence of hydrogen peroxide concentration on changes occurring in plants
In the last step, the wheat was watered with different concentrations of hydrogen peroxide (0-3.0%).The FTIR spectra of leaf extracts after 14 and 28 days are presented in Figure 6.
As can be seen, the spectrum profiles of leaf extracts after 14 and 28 days are similar, which can mean that the plants have similar metabolomic profiles, and in the spectra, there are no significant differences in the intensity of bands.The spectra of wheat watered with 1% of H₂O₂ seedlings in November are not compared to the others which were bred in February, due to their large diversity.Between these two months, there is a big difference in light intensity and probably it was the cause of so different spectra profiles of these plants.By comparing the spectra of plants watered with different concentrations of hydrogen peroxide (except 1%) there are small differences in the intensity of individual bands.There are no offsets of bands or they are so small that almost invisible.The chemometric analysis (HCA and PCA -data not shown) also confirms that the differences between the spectra are small, these plants always have been grouped as one class.Table T1.Ranges of wavenumbers corresponding to given groups of metabolites [1,[11][12][13][21][22][23][24][25][26]

Figure 1 .
Figure 1.The illustration of the changes in tested oats after 14 and 28 days as a result of different external stress factors

Figure 3 .Figure 4 .
Figure 3. Chemometric analysis of leaves and extracts across different species.HCA analysis of extracts and leaves from different plant species (A).PCA scores of extract and leaves after 14 and 28 days (B)

Figure 5 .Figure 6 . 1 Figure S1 .
Figure 5. HCA (A) and PCA (B) results of extracts of all seedlings treated with different chemical and physical factors

Table 1 .
Analysis of infrared spectra -a summary of the bands that distinguish individual seedlings

Wavenumber at which the peak attains maximum value [cm-1] Assignent to metabolic group Extract after 14 days Extract after 28 days Leaves after 14 days Leaves after 28 days
Monitoring a metabolic profile of wheat by using FTIR spectroscopy and chemometric methods… Science, Technologyand Innovation, 2022, 17 (1-2), 41-75