Introduction
Corn (Zea mays L.) is a crop extensively planted all over the
world, serving as raw materials of fodder, grains and energy (Byrt et al. 2011). Compared with
rice, wheat and other food crops, corn is better in drought, cold, barren
tolerance and excellent environmental adaptability. Corn is an excellent food
crop with a high nutritional value. It is an important source of feed in animal
husbandry, aquaculture, etc., and also one of the indispensable raw materials
for food, medical and health, light industry, chemical industry, etc. (Byrt et al. 2011; Ma 2019). It plays an important role in ensuring
national food security. 500 g of corn kernels contain 365 g of carbohydrates,
which is slightly lower than rice; 21.5 g fat, more than any other cereal
crops; 42.5 g protein, second to millet only; higher vitamin B than other crops
(Chen et al. 2004; Ou 2016). Corn
also contains more cellulose. As an important industrial raw material, more
than 500 kinds of industrial products have been made directly or indirectly
from corn kernels and by-products (Ge et
al. 2017; Miao 2018; Wang 2018). The main product of modern corn industry
is corn starch, which is widely used in food, medicine, textile and other
industrial fields (Liu 2014; Lu et al. 2018;
Tong 2019). Glucose, liquor, beer, acetone and so on can be made from corn
seeds (Wang 2016; Wang and Wang 2016). Corn stalks can be used to make
fiberboard, paper, rayon, electrical insulation and chemical rubber plate.
Besides, corn has a wide range of applications in medicine. For example, corn
starch is an important raw material to produce penicillin, streptomycin,
aureomycin and other antibiotics. Corn has high yield and wide use, and has
high economic value as food, feed, industrial and pharmaceutical raw materials.
In China, cultivated area of corns ranks only second to wheat and paddy rice.
Located in Southwest China, Yunnan is one of the provinces where corns are
introduced at the earliest (Huang 2012). According to statistics, the cultivated area and
yield of corn in Yunnan rank first and second, respectively. For instance, the
cultivated area of corn reached 1.7177 million hectares in 2016; and the yield
of the same year was up to 9.196 million tons (Li et al. 2018). As corns are extremely rich, cheap, easy to obtain, strongly
environment adaptable, and also have many biological activities, such as
anti-oxidation, anti-tumor, immunomodulation and bacteriostasis, etc., it has a broad development and
application prospect (Yang et al. 2019).
As a major medium in which crops are planted, soil contains not only
nutrients required by growth of crops, but also substances (e.g., plant hormones, amino acid and
antibiotics) secreted by soil microbials and secretions (e.g., polysaccharose and amino acids) (Kuang et al. 2003). All these substances have certain influences on metabolism of
soil microbials and plants. Soil resources are precious for human survival, and
soil fertility is its essential attribute (Wang and Fu 2007; Pang 2009; Huang et al. 2017). Soil fertility is an
indicator of the ability of the soil to provide various nutrients required for
crop growth. It is a comprehensive performance of various basic soil properties
and differentiates the soil from soil parent materials and other natural
bodies. The material basis of soil proves a natural resource and means of
agricultural production (Xiong 2001; Huang and Sun 2006; Marschner and Rengel 2007; Lazcano 2011). Soil fertility is the basic
property and essential characteristic indicating the ability of soil to supply
and coordinate nutrients, water, air and heat for plant growth, and the
comprehensive response of soil physical, chemical and biological properties
(Liu 2010; Buckland and Grime 2010; Ling-An et
al. 2011; Zhao and Bai 2013). The evaluation of soil fertility
contains qualitative and quantitative methods. Qualitative descriptions of soil
fertility quality are relatively simple, such as intuitive description of how
the soil looks, feels and smells. While the quantitative method refers to the
calculation of the "score" of soil quality according to the
quantitative soil properties, and the best soil usually gets the highest score
(Song 2011). Soil fertility index includes soil chemical, physical, biological
and environmental condition index, and all the factors are expressed in
numerical value. In this way, a series of scores will be involved in soil
fertility evaluation. It is difficult to find the internal relationship between
each index from these data, which is difficult to achieve by manual processing.
Therefore, the comprehensive evaluation of soil fertility must be carried out
from the perspective of multiple factors by means of mathematical analysis (Luo
et al. 2002; Yang et al. 2016). Different soil types
correspond to different vegetational forms and microbial compositions. High
throughput testing and data processing as its approaches, mass spectrometry
(MS) and nuclear magnetic resonance (NMR) as its analytical platforms, and
information modeling and system integration as its objectives. At present, metabolomics
has become one of the powerful research tools (Ram et al. 2005; Baker 2011; Geyer 2013; Suthar et al. 2013). Ultra-high
Performance Liquid Chromatography-Quadrupole-Time-of-Flight/Mass Spectrometry
(UPLC/Q-TOF-MS) is a chromatographic separation system that uses Ultra-high
Performance Liquid Chromatography (UPLC) as a chromatographic separation
system. TOF-MS is a mass spectrometry technology formed by series analyzers
(Zhang et al. 2017). UPLC/Q-TOF-MS is
one of the most effective methods for multi-component analysis and
identification of complex matrices in recent years (Lacina et al. 2010). Compared with traditional high-performance liquid
chromatography (High Performance Liquid Chromatography, HPLC), the speed,
sensitivity and resolution of UPLC are 9 times, 3 times and 1.7 times higher
respectively (Wren 2005; Lucie et al. 2006).
Q-TOF-MS tandem technology is an important breakthrough in mass spectrometry
technology. It has the advantages of high sensitivity, high selectivity,
multi-stage mass spectrometry, and high information acquisition speed.
Therefore, the UPLC/Q-TOF-MS combination technology can effectively solve the
problems of complex composition and quantitative difficulties in soil
composition analysis, and can analyze the metabolites in the soil quickly,
accurately, comprehensively, and reliably. For the past few years, metabolomics
techniques have been widely applied in the research of soil. It is also
increasingly combined with other system biology technology, such as proteomics,
transcriptomics and genomics (Chen et al.
2013; Mandal and Singh 2013). MS is deemed as an analytical method dependent on
mass-to-charge ratio determination for ions. Thanks to its sensitivity and specificity,
MS gradually surmounts NMR and becomes the most powerful tool for qualitative
and quantitative analysis on metabolites in metabolomics-related investigations
(Dunn and Ellis 2005; Griffiths and Wang 2009) Since red soils are extensively distributed in Yunnan Province (Zhou 1983) corns in
Yunnan are mostly planted in red soils. Moreover, red soils with high viscidity
contain metallic oxides such as iron and aluminum. Being highly acidic, red
soils also contain a few amounts of organic matters (Huang 2012; Zhao et al. 2019).
In this paper high performance liquid chromatography/ tandem
high-resolution mass spectrometry was performed to test metabolome of
rhizosphere soils of corn at milk stage in Dongchuan and Fumin regions. In
combination with bio-information analysis, MS data were interpreted for
presenting metabolites in such soils and providing references for improvement
of soils where corns are planted in both regions.
Materials and Methods
Sample collection
Dongchuang is located at the northeastern part of Yunnan, Fumin, as a
county, is situated in the northeast of central Yunnan. Being 200 km apart,
both regions have a subtropical monsoon plateau climate. Rhizosphere soils of
corns at a milking stage were sampled from red soils in Dongchuang District and
Fumin County of Yunnan Province. While those collected in Dongchuan were
labeled as DCMRS, those collected in Fumin were labeled as FMMRS. Corn plants
that grow well were collected, of which the roots about 5 cm below the surface
layer were cut off, with soils attached. After which, they were put in a
preservation box at 4°C and brought back to the laboratory. Totally, 6 soil
samples were gathered in Dongchuang and Fumin, respectively. On an ultra-clean
bench, excess soils were shaken off (Riley and Barber 1970; Hui et al. 2016) and the soils still
attached on the roots were collected. Subsequently, impurities (including plant
residues and fibrous root systems) in the soils were removed. Then, each soil
sample was further equally divided into two portions, viz. one was stored at 4°C
temporarily and used for general analysis on soil compositions, the other
portion was stored at -80°C and used for non-targeted metabolomic analysis.
Analysis soils physico-chemical properties
Soil pre-treatment was carried out according to requirements in determination
of total nitrogen, organic matters, ammonium nitrogen, pH, available phosphorus
and available potassium. To be specific, the method proposed by (Li et al. 2018) in Tropical
Forestry was adopted to implement soil determination. However, a semi-micro
Kjeldahl method was selected for total nitrogen determination. The organic
matter content in soils was measured by the potassium dichromate (external
heating) method. The contents of available phosphorus and potassium were
determined by the NaHCO3 (0.5 mol/L) leaching-molybdenum-antimony
colorimetric method and NH4OAc (1.0 mol/L) leaching-flame spectrometry.
By following Li (2011), the potentiometric method was utilized to measure the soil pH, in
which water–soil ratio was set to 2.5:1. At last, ammonium nitrogen in soils
was determined by the indophenol blue colorimetric method. It should be noted
that each test was repeated for three times.
Metabolite leaching
The collected samples were thawed on ice, and
metabolites were extracted with 50% methanol buffer. Briefly, 20 μL of sample was extracted with 120
μL of precooled 50% methanol,
vortexed for 1 min, and incubated at room temperature for 10 min. Then, the
extraction mixture was stored overnight at -20°C, followed by centrifugation at
4,000 g for 20 min. After which, the supernatants were transferred into new
96-well plates, stored at -80°C prior to the LC-MS analysis. In addition,
pooled QC samples were also prepared by mixing 10 μL of each extraction mixture.
LC/MS analysis
All samples were acquired by the LC-MS system
according to machine orders. Firstly, all chromatographic separations were
performed using an ultra-performance liquid chromatography (UPLC) system
(SCIEX, UK). An ACQUITY UPLC T3 column (100 mm*2.1 mm, 1.8 µm, Waters, U.K.) was used for the reversed phase separation. The
column oven was maintained at 35°C. The flow rate was 0.4 mL/min. The mobile
phase consisted of solvent A (water, 0.1% formic acid) and solvent B
(Acetonitrile, 0.1% formic acid). Gradient elution conditions were set as
follows: 0~0.5 min, 5% B; 0.5~7 min, 5% to 100% B; 7~8 min, 100% B; 8~8.1 min,
100 to 5% B; 8.1~10 min, 5% B. The injection volume for each sample was 4 µL.
A high-resolution tandem mass spectrometer
TripleTOF5600plus (SCIEX, UK) was used to detect metabolites eluted form the
column. The Q-TOF was operated in both positive and negative ion modes. The
curtain gas was set to 30 psi, Ion source gas1 was set to 60 PSI, Ion source
gas 2 was set to 60 PSI, and an interface heater temperature was to 650°C. For
positive ion mode, the Ionspray voltage floating was set to 5000 V. For
negative ion mode, the Ionspray voltage floating was set to -4500V. The mass
spectrometry data were acquired in IDA mode. The TOF mass range was from 60 to
1200 Da. The survey scans were acquired in 150 ms and a total of 12 product ion
scans were collected if exceeding a threshold of 100 counts per second
(counts/s) with a 1+ charge-state. Total cycle time was fixed to
0.56 s. Four time bins were summed for each scan at a pulser frequency value of
11 kHz through monitoring of the 40 GHz multichannel TDC detector with
four-anode/channel detection. Dynamic exclusion was operated for 4 s. During
the acquisition, the mass accuracy was calibrated once every 20 samples.
Furthermore, in order to evaluate the stability of the LC-MS during the whole
acquisition process, a quality control sample (Pool of all samples) was
acquired for every 10 samples.
Information analysis
The pretreatments of acquired MS data including peak picking, grouping,
retention time correction, second peak grouping, and annotation of isotopes and
adducts were performed using XCMS software (Want et al. 2006). LC−MS raw data files were converted into mzXML format and
then processed by the XCMS, CAMERA and metaX (Wen et al. 2017) toolbox of R software. Each ion was identified by
combining retention time (RT) and m/z data. Intensity of each peak was recorded
and a three-dimensional matrix consisting of arbitrarily assigned peak indices
(retention time-m/z pairs), sample names (observations) and ion intensity
information (variables) was generated.
The online KEGG database was used to annotate the metabolites by
matching the exact molecular mass data (m/z) of samples with those from
database. If a mass difference between observed value and the database value
was less than 10 ppm, the metabolite would be annotated and the molecular
formula of metabolites would further be identified and validated by the
isotopic distribution measurements. An in-house fragment spectrum library of
metabolites was also used to validate the metabolite identification.
The intensity of peak data was further preprocessed by metaX. Those
features that were detected in less than 50% of QC samples or 80% of biological
samples were removed; the remaining peaks with missing values were imputed with
the k-nearest neighbor algorithm to further improve the data quality. PCA was
performed for outlier detection and batch effects evaluation based on the
pre-processed dataset. Quality control-based robust LOESS signal correction was
fitted to the QC data according to the order of injection to minimize the drift
of signal intensity over time. In addition, the relative standard deviations of
the metabolic features among all QC samples were calculated, and those > 30%
were then removed. Before quantitative analysis on metabolites, quality control
over their intensity was carried out to obtain quantitative information of
high-quality ion peaks. Based on CV and hierarchical clustering of QC samples,
data quality control can be achieved.
Statistical analysis
SPSS 20.0 (SPSS, Inc., Chicago, IL, USA) was used
for data statistics and analysis, while GraphPad 8.0.2 was utilized to plot the
relevant histograms. SIMCA_P 13.0 was used to implement principal component
analysis (PCA) and orthogonal to partial least squares discriminant analysis
(OPLS-DA). The R software was adopted to plot heat maps of metabolites and
corresponding calculations were carried out by means of hierarchical clustering
(HCL). As for distance computing, the Euclidean distance was adopted. It is
worth noting that ward. D2 was used as the clustering method.
Student t-tests were conducted to detect difference in metabolite
concentration between 2 phenotypes. The P value was adjusted for
multiple tests using an FDR (Benjamini–Hochberg). Supervised PLS-DA was
conducted using metaX to discriminate the variables between groups. The VIP
value was calculated. A VIP cut-off value of 1.0 was used as a criterion to
select important features.
Results
Soil physico-chemical properties
Nitrogen, phosphorus and potassium are three major soil nutrients essential
to plant growth. These three nutrients are not only applied to the most during
agricultural production, but also consumed the most by plant growth. In this
study, the contents of total nitrogen, ammonium nitrogen, soil organic carbon,
available phosphorus and available potassium in DCMRS and FMMRS were measured
(Table 1; Fig. 2A–D).
Soil metabolic substances: The extracted
metabolic substances were detected in both positive and negative ion modes
using a high-resolution mass spectrometer. The total ion count and the number
of substances annotated by primary and secondary mass spectrometry data can be
found in positive and negative ion modes (Table 2).
Detection control of metabolic substances: The only mass-to-charge ratio and chromatographic retention time were
revealed in mass spectra of each metabolite. With the goal of controlling data
quality, quality control over substances obtained by XCMS was provided,
including total ion chromotogram, m/z-rt distribution of metabolites, m/z
difference ranges generated by peak alignment of each substance, and rt
difference ranges thus produced. As shown in the total ion chromotogram, the
detected metabolites can be well isolated in the chromatographic condition set
for this study. A great number of features were found when the retention time
reached 1–1.5 min, 2–3.5 min or 8 min approximately. Among them, the number of
features generated at a retention time of 2–3.5 min reached its peak. In terms
of m/z, the maximum number of features was produced when it ranged between 200
and 400. Triple TOF 5600, a high-resolution mass spectrometer (resolution: 30,
000 and above), was utilized to detect the metabolites. Subsequently, alignment
could be implemented with the help of XCMS. It turned out that the aligned m/z
had a range of -0.015–0.015 and the aligned retention time all varied between
-0.5 min and 0.5 min (Fig. 2A–D).
Identification and quantification of metabolic substances
Identification of metabolic substances: According to the detected substances, the primary m/z was matched in
the KEGG database to acquire primary identification results. During primary
identification, one m/z correspond to multiple metabolites. In order to produce
more accurate identification results, secondary mass spectrogram of in-house
metabolites was compared with their secondary mass spectrometry data (Table 2).
When several metabolites share identical chemical composition, a
substance may still have different isomerides due to different types of element
sorting. Considering that these isomerides have the same molecular weight, it is
less likely to distinguish them based on mass spectra. For this reason, one
substance may correspond to multiple possible metabolites during primary
identification of substances. Hence, one-to-many statistics were
made for metabolite identification, (Fig. 3A–B).
Table 1: Soil pH, and contents of total nitrogen, organic matters, ammonium
nitrogen, available phosphorus and available potassium in soils (Mean ± SD)
|
|
TN (Total nitrogen) (mg/kg) |
AN (Ammonium nitrogen) (mg/kg) |
SOC (Soil organic carbon) (mg/kg) |
AP (Available phosphorus) (mg/kg) |
AK (Available potassium) (mg/kg) |
pH |
|
DCMRS |
0.67 ± 0.14 |
39.22 ± 3.32 |
18.20 ± 0.89 |
51.31 ± 5.76 |
418.83 ± 17.92 |
5.79±0.43 |
|
FMMRS |
2.10 ± 0.42 |
19.493 ± 1.47 |
19.94 ± 2.27 |
23.93 ± 3.51 |
478.97 ± 12.93 |
6.77±0.76 |
Notes: DCYRS stands for rhizosphere soil for corns in a
seedling stage in Dongchuan District, DCMRS for rhizosphere soil for corns at a
milk stage in Dongchuan District, and FMMRS for rhizosphere soil for corns at a
milk stage in Fumin County
Table 2: The total ion count and identification statistics of metabolites
|
Mode |
All feature |
All annotated |
MS2 |
KEGG |
|
POS |
4189 |
2477 |
85 |
1899 |
|
NEG |
2739 |
1230 |
34 |
957 |
Notes: Mode: An ion mode in which the substance is
detected under a mass spectrometer; POS: Positive ion mode; NEG: Negative ion
mode; All feature: The number of substances extracted by the XCMS software; All
annotated: The number of substances obtained by primary and secondary mass
spectrometry data; and, MS2; The number of secondary ions identified, that is
the number of substances that can be matched with both primary m/s and
secondary fragment ions m/s of a substance in the database
%20555-566_files/image002.jpg)
Fig. 1: A: Total nitrogen; B:
Available phosphorus; C: Organic
matters; D: Ammonium nitrogen; E: Available potassium; and, F: Soil pH. DCMRS refers to rhizosphere
soil for corns at a milk stage in Dongchuan District, and FMMRS refers to
rhizosphere soil for corns at a milk stage in Fumin County
After metabolite matching in the KEGG database, the metabolites were
categorized (Fig. 3C–D). In line with global and overview maps of KEGG pathway
level 2, the number of metabolites associated with lipid metabolism,
xenobiotics biodegradation and metabolism, metabolism of terpenoids and
polyketides, and amino acid metabolism was above 160. More particularly, global
and overview maps corresponded to probably the most metabolites. Regarding
lipid metabolism, xenobiotics biodegradation and metabolism, metabolism of
terpenoids and polyketides, and amino acid metabolism, the number of potential
metabolites may range from 165 to 200. Among the first 20 KEGG pathways, there
were metabolic pathways, biosynthesis of secondary metabolites, microbial
metabolism in diverse environments, biosynthesis of antibiotics, degradation of
aromatic compounds, arachidonic acid metabolism, and sesquiterpenoid and triterpenoid
biosynthesis, etc.
According to peak areas of DCMRS and FMMRS, the top first 20
metabolites are respectively listed in Table 3 and 4, including their MZ, RT,
peak areas, names and formulas. It could be found that such metabolites had
similar rankings in both DCMRS and FMMRS. For example, 2-methyl-1-propylamine
had the largest peak area, followed by gamma-butyrolactone and
3-methyl-1-butylamine successively. Except proline and valine, peaks areas of
other metabolites in DCMRS were larger than those in FMMRS.
Quantitative analysis on metabolic substances: Quantitative information of metabolites was derived from primary
chromatographic peak areas of the substance. Moreover, no quantitative analysis
was made on differences of the ions with CV > 30% for a reason they may vary
substantially during the experiment. Quantitative statistical information of
metabolites is listed in Table 5. The number of features extracted by XCMS was
4,189 in a positive ion mode or 3,622 in a negative ion mode. After the number
of ions of a substance with a missing value < 80% in the sample or < 50%
in the QC sample was acquired, the number of high quality features can be
figured out, which was 3,622 in the positive ion mode or 2,389 in the negative
ion mode.
As shown in Fig. 4, quality control analysis was carried out prior to
quantitative analysis on metabolites. In this figure, A stands for CV
distribution of the sample. After normalization, such data were proven to be
consistent with post-processing requirements. In addition, C represents
distribution intensity evaluation of differential metabolites A Boxlot diagram
was plotted after log 2 normalization of relevant
intensity values, where log 2 (i.e.,
intensity) ranged from 10 to 12. B indicates that hierarchical clustering
analysis on metabolite intensity of FMMRS and DCMRS was implemented by means of
ward D2. At last, principal component analysis (PCA) was made on FMMRS, DCMRS
and QC samples. More particularly, FMMRS, DCMRS and QC samples can be
preferably differentiated in the PCA diagram. This indicates there was certain
difference in differential metabolites between FMMRS and DCMRS.
Table 3: First 20 substances with maximum peak areas in DCMRS
|
Number |
MZ |
RT |
Average peak area |
Name |
Formula |
|
1 |
228.1959023 |
3.04505 |
136183.8769 |
2-Methyl-1-propylamine |
C4H11N |
|
2 |
256.2631784 |
5.566516667 |
131652.8086 |
Gamma-Butyrolactone |
C4H6O2 |
|
3 |
284.2949678 |
6.489666667 |
125826.1855 |
3-Methyl-1-butylamine |
C5H13N |
|
4 |
132.1009235 |
1.766241667 |
112990.2737 |
2,5-Dihydro-2,4-dimethyloxazole |
C5H9NO |
|
5 |
229.1405197 |
2.7169 |
87674.70663 |
(2R,3S)-2-methyl-3-propyloxirane |
C6H12O |
|
6 |
135.0796726 |
5.277983333 |
76720.33604 |
Triethylamine |
C6H15N |
|
7 |
228.1956866 |
6.213366667 |
57819.67183 |
Triethylamine |
C6H15N |
|
8 |
279.1587938 |
4.5003 |
57116.42804 |
Choline |
C5H14NO |
|
9 |
172.1328279 |
2.528366667 |
50148.55729 |
1-Piperidinecarboxaldehyde |
C6H11NO |
|
10 |
87.04354563 |
3.111133333 |
46848.60368 |
Proline |
C5H9NO2 |
|
11 |
310.3105369 |
6.595866667 |
44256.53949 |
Valine |
C5H11NO2 |
|
12 |
250.1773982 |
6.213366667 |
43742.21084 |
Indane |
C9H10 |
|
13 |
136.0204317 |
6.441 |
36122.42364 |
Nicotinic acid |
C6H5NO2 |
|
14 |
118.0852842 |
0.889216667 |
32412.89428 |
Malonic acid |
C3H4O4 |
|
15 |
226.1799047 |
6.220416667 |
32183.90691 |
Malonic acid |
C3H4O4 |
|
16 |
413.266458 |
7.251366667 |
31642.44589 |
Hypotaurine |
C2H7NO2S |
|
17 |
353.2299268 |
2.996966667 |
28868.39866 |
Dihydrothymine |
C5H8N2O2 |
|
18 |
337.2349649 |
5.83 |
28080.3092 |
Pipecolate |
C6H11NO2 |
|
19 |
540.5357826 |
6.34005 |
27069.31943 |
Isoleucine |
C6H13NO2 |
|
20 |
149.0224496 |
4.50395 |
24963.68949 |
Leucine |
C6H13NO2 |
%20555-566_files/image004.jpg)
Fig. 2: Metabolite detection control. A: Total ion chromotogram (TIC) of metabolites, where the overall
mass spectrum signal strength of the sample is under control. In this map,
x-axis represents time and the y-axis stands for sums of all ion intensity at
each time point in the mass spectrogram; B:
A mz-rt distribution diagram, where x-axis refers to the substance retention
time, y-axis refers to m/z of the substance, each point in the diagram
represents a substance; C: The m/z
alignment ranges of metabolites; D:
The rt alignment ranges of metabolites
As shown in Fig. 5, mean intensity of metabolites
undergoing secondary mass spectrum identification
(A) is given. Here, bar length of the second ring represents diverse mean
intensity of the metabolites in the mass spectrum. Moreover, mean intensity of
which should be multiplied with a log10 function. A bar stands for a type of
metabolites color represents the type of metabolites. Fig. 5B shows the
correlation of metabolites subjected to secondary mass spectrum identification.
In this study, their correlation is embodied in different colors and various
shades. For example, the darker the red is, the stronger their correlation will
be; the darker the blue is, the weaker the correlation will be. In addition,
two correlated areas are marked by a green triangle in dashed lines. Fig. 5C
shows the intensity of metabolites for secondary mass spectrum identification,
where x-axis refers to DCMRS, FMMRS and QC samples, while y-axis represents the
type of metabolites. Each square in the figure represents a metabolite.
Likewise, darker red corresponds to a stronger correlation, but darker blue
represents a weaker correlation.
Differential metabolite screening & differential metabolic pathway
analysis
Table 4: Top 20 substances with maximum peak areas in FMMRS
|
Number |
MZ |
RT |
FMMRS Average peak area |
Metabolite |
Formula |
|
1 |
228.1959023 |
3.04505 |
134457.8221 |
2-Methyl-1-propylamine |
C4H11N |
|
2 |
256.2631784 |
5.566516667 |
127645.9067 |
Gamma-Butyrolactone |
C4H6O2 |
|
3 |
284.2949678 |
6.489666667 |
123988.4382 |
3-Methyl-1-butylamine |
C5H13N |
|
4 |
132.1009235 |
1.766241667 |
86018.61235 |
2,5-Dihydro-2,4-dimethyloxazole |
C5H9NO |
|
5 |
229.1405197 |
2.7169 |
55237.28711 |
(2R,3S)-2-methyl-3-propyloxirane |
C6H12O |
|
6 |
135.0796726 |
5.277983333 |
54484.58411 |
Triethylamine |
C6H15N |
|
7 |
228.1956866 |
6.213366667 |
53744.10721 |
Triethylamine |
C6H15N |
|
8 |
279.1587938 |
4.5003 |
50242.79421 |
Choline |
C5H14NO |
|
9 |
172.1328279 |
2.528366667 |
49425.78368 |
1-Piperidinecarboxaldehyde |
C6H11NO |
|
10 |
87.04354563 |
3.111133333 |
48750.38422 |
Proline |
C5H9NO2 |
|
11 |
310.3105369 |
6.595866667 |
44981.94301 |
Valine |
C5H11NO2 |
|
12 |
250.1773982 |
6.213366667 |
40121.6433 |
Indane |
C9H10 |
|
13 |
136.0204317 |
6.441 |
32067.15209 |
Nicotine acid |
C6H5NO2 |
|
14 |
118.0852842 |
0.889216667 |
30676.72433 |
Malonic acid |
C3H4O4 |
|
15 |
226.1799047 |
6.220416667 |
29604.93344 |
Malonic acid |
C3H4O4 |
|
16 |
413.266458 |
7.251366667 |
27758.16764 |
Hypotaurine |
C2H7NO2S |
|
17 |
353.2299268 |
2.996966667 |
26494.2727 |
Dihydrothymine |
C5H8N2O2 |
|
18 |
337.2349649 |
5.83 |
25551.76821 |
Pipecolate |
C6H11NO2 |
|
19 |
540.5357826 |
6.34005 |
24993.91453 |
Isoleucine |
C6H13NO2 |
|
20 |
149.0224496 |
4.50395 |
24497.43798 |
leucine |
C6H13NO2 |
%20555-566_files/image006.jpg)
Fig. 3: Primary m/z identification for metabolite. A: One-to-many statistical chart of
metabolite identification, where x-axis represents the number of identified
metabolites corresponding to one feature, and y-axis is the number of features;
B: Approximate categories of
metabolites identified, where x-axis stands for the metabolite category, y-axis
stands for a proportion taken by each category of metabolites in all
metabolites; C: Classification of
KEGG pathways, where x-axis refers to items of KEGG pathway level 2, y-axis
refers to the number of possible metabolites that involve such classification; D: First 20 pathways in which all
possible metabolites may participate
Differential metabolite screening: In this
paper, q-value was acquired after BH correction is conducted based on
differential folds of univariate analysis and t-test. Moreover, the value of Variable Important for the
Projection (VIP) was acquired through PLS-Discriminant Analysis (PLS-DA). The
number of differential metabolites selected is shown in Fig. 6A. It can be seen
that the number of up-regulated and down-regulated metabolites of FMMRS was 222
and 261, respectively (Fig. 6B).
Table 5: Quantitative statistics of metabolites
|
Mode |
All feature |
High quality feature |
|
POS |
4189 |
3622 |
|
NEG |
2739 |
2389 |
Notes: Mode: An ion mode in which the substance is
detected under a mass spectrometer; POS: Positive ion mode; NEG: Negative ion
mode; All features: The number of substances extracted by the XCMS software;
High quality feature: The number of ions in a substance with missing value <
80% in the sample or < 50% in the QC sample
%20555-566_files/image008.jpg)
Fig. 4: Quantitative quality control over metabolites. A: CV distribution diagram, where
CV=SD/Mean; and, the x-axis represents value of CV and the y-axis represents a
proportion taken by the number of ions. Generally, it is considered that if CV ≤
30%, the sample shows good repeatability; B:
Intensity of each metabolite in each sample is presented in the heat map,
where, each scale is obtained by log transformation of intensity values; the
Euclidean distance is introduced in calculation; and, ward.D2 serves as the
clustering method; C: Boxplot of
intensity distribution evaluation for metabolites in each sample, where, y-axis
represents log2 transformation of intensity in the mass spectrum; D: A chart of PCA scores, where, each
point represents a sample; if the points are rather close to each other, it
indicates that the samples are very similar; otherwise, it reveals that certain
differences can be found among these samples
Fig. 6C shows the volcano plot of univariate statistical test
evaluation. The green and red points stand for down-regulated and up-regulated
substances, respectively (Fig. 6D). As a supervised discriminant analytical
statistical method, PLS-DA has the potential to reveal differences among
different groups to the greatest extent. By using such method, a relation model
for metabolite expression quantity and sample category was constructed by means
of PLS regression (PLSR) so that modeling and prediction of sample categories
can be realized. Additionally, VIP was worked out for the purpose of evaluating
influential intensity and explanatory ability of various metabolite expression
patterns for sample classification and discrimination. This may contribute to
screening of metabolic markers (screening condition: VIP ≥ 1.0). R2
and Q2, as two parameters of the PLS-DA model, were calculated to be
0.99 and 0.94. The closer their values are to 1, the more reliable the PLS-DA
model will be. On this basis, it is deemed that the relation model constructed
is rather reliable.
Differential metabolism analysis: Differential
metabolic pathway analysis of differential substances and KEGG was carried out.
Through comparison between FMRMRS and DCMRS, 151 differential metabolic
pathways were obtained, including 76 up-regulated ones and 75 down-regulated
ones. It was found that some of these metabolic pathways were associated with
nitrogen cycling (Fig. 7), such as carbon metabolic pathway, amino
acid metabolic pathway, antibiotics anabolism, and biosynthesis of siderophore
group non-ribosomal peptides. Some differential metabolic pathways were in a
positive ion mode (Table 6).
Discussion
%20555-566_files/image010.jpg)
Fig. 5: Quantitative information of metabolites in DCMRS and
FMMRS. A: An annular chart of the
mean intensity of metabolites undergoing secondary mass spectrum
identification, where the color represents the type of metabolites; B: Correlation of metabolites subjected
to secondary mass spectrum identification, where correlation between two
metabolites is shown; the darker the red is, the stronger their correlation
will be; but if the blue is darker, the correlation will be weaker; C: Metabolite intensity heat map of
secondary mass spectrum identification, where color blocks above represent
different groups of DCMRS, FMMRS and QC samples, but color blocks on the left
refer to different types of metabolites; in the middle, each color block
represents a metabolite; similarly, darker red corresponds to stronger
correlation, but darker blue represents weaker correlation
Red soil containing metallic oxides such as iron and aluminum is widely
distributed in Yunnan Province, China (Huang and Fu 2002; Zhao et al. 2019). Corn is an important
fodder and food crop of Yunnan. In this study, soil conditions are
preliminarily evaluated by measuring soil pH and the contents of total
nitrogen, organic matters, ammonium nitrogen, available phosphorus and
potassium in DCMRS and FMMRS for corns at a milk stage in Dongchuan District of
Yunnan. Overall, FMMRS is more fertile than DCMRS, but less acidic (Table 1 and
Fig. 1). These components in the red soil may be produced by fertilizers
applied prior to the milk stage of corns. For example, chemical fertilizers can
increase the contents of available phosphorus and available potassium in the
soil. Moreover, once organic fertilizers (e.g.,
decaying manure from the livestock) are applied, the contents of total nitrogen
and organic matters in the soil may go up (Zhang 2011).
The source of metabolic substances in rhizosphere soil is in exudates of
rhizosphere microbes and plant root systems. Matters detected by soil
metabolomic analysis consist of these metabolic substances. Plant root exudates
%20555-566_files/image012.png)
have the capability
to enhance stress tolerance of plants and promote growth of soil microbes (Bashir et al. 2016). According to relevant findings, corn root exudates contain
asparaginic acid, tyrosine, starch and flavonoids (Kuang et al. 2003; Carvalhais et
al. 2011; Zhu et al. 2016; Luo et al. 2017). In this study, amino acid,
saccharides and flavonoids were detected in root soil DCMRS and FMMRS. The
influence of soil microbes on rhizosphere soil cannot be ignored either. The
reason is that microbes are able to produce secondary metabolites by using
plant root exudates, including antibiotics and acids. Through comparison
between FMMRS and DCMRS, multiple antibiotic synthesis associated metabolic
pathways have been annotated. In this course, the metabolic pathways associated
with iron transport were observed. In addition, another investigation on
microbiological compositions in corn rhizosphere soil was carried out as well.
Likewise, not only were bacteria associated with antibiotic synthesis detected,
but also functional genes and metabolic pathways associated with iron element
metabolism and transport in soil were also found. It is expected that some
significant information can be provided for improving the soil where corns are
planted. The pH of DCMRS was lower than of FMMRS. However, if the pH value of
rhizosphere soil were always low, it would affect the accumulation of crop
biomass. Before the mature stage of corn, alkaline chemical fertilizer can be
properly added to improve soil pH value, which is more conducive to the
accumulation of dry matter in corn. Rhizosphere soil metabolites affect many
aspects of crop growth. Proper application of organic fertilizer or microbial
soil inoculants can improve the content of soil metabolites by adjusting the
composition of soil microorganisms, thus improving the stress resistance and
yield of corn.
Conclusion
Concerning the non-targeted metabolome, metabolites in samples can be comprehensively
analyzed to confirm an analytical method for differential metabolites. Under
the circumstance that massive data are generated during
non-targeted metabolomic detection, the selected separation and detection
technique may affect the data. As an analytical approach, UPLC-Q/TOF-MS is
featured with high sensitivity and high through-put. Therefore, it is adopted
in this study to detect and analyze metabolites in DCMRS and FMMRS. As an
extraction solvent exerting a great influence on leaching of metabolites from
the samples, methanol-water was adopted in this study to extract most
water-soluble and fat-soluble substances from the samples. Although, these
substances can largely reveal components of metabolites in the soil, errors
caused by the detection technique or calculations may lead to data loss to a
certain extent. To solve such a defect, new detection techniques and
calculation methods are needed in the future research with the goal of reducing
errors.
Table 6: Top 20 metabolic pathways ranked according to difference significance
through comparison of FMRMRS and DCMRS
|
# |
Pathway |
Difft. metabolites (241) |
All metabolites (1176) |
P value |
Pathway ID |
|
1 |
ABC transporters |
23 |
34 |
1.95E-09 |
map02010 |
|
2 |
Microbial metabolism in diverse
environments |
79 |
238 |
1.45E-07 |
map01120 |
|
3 |
Starch and sucrose metabolism |
8 |
8 |
2.83E-06 |
map00500 |
|
4 |
Cyanoamino acid metabolism |
9 |
15 |
0.000871116 |
map00460 |
|
5 |
Aminoacyl-tRNA biosynthesis |
7 |
10 |
0.000954765 |
map00970 |
|
6 |
Ascorbate and aldarate metabolism |
7 |
10 |
0.000954765 |
map00053 |
|
7 |
Carbapenem biosynthesis |
4 |
4 |
0.001728987 |
map00332 |
|
8 |
Phenylalanine, tyrosine and
tryptophan biosynthesis |
5 |
6 |
0.001745478 |
map00400 |
|
9 |
Carbon metabolism |
7 |
13 |
0.00771802 |
map01200 |
|
10 |
Inositol phosphate metabolism |
3 |
3 |
0.008521438 |
map00562 |
|
11 |
Phenylalanine metabolism |
10 |
23 |
0.009946065 |
map00360 |
|
12 |
Caprolactam degradation |
6 |
12 |
0.02120928 |
map00930 |
|
13 |
Degradation of aromatic compounds |
22 |
73 |
0.02876643 |
map01220 |
|
14 |
Nitrogen metabolism |
3 |
4 |
0.02889879 |
map00910 |
|
15 |
Toluene degradation |
6 |
13 |
0.03277322 |
map00623 |
|
16 |
Tryptophan metabolism |
6 |
13 |
0.03277322 |
map00380 |
|
17 |
Caffeine metabolism |
4 |
7 |
0.0357948 |
map00232 |
|
18 |
Pentose and glucuronate
interconversions |
4 |
7 |
0.0357948 |
map00040 |
|
19 |
Bacterial chemotaxis |
2 |
2 |
0.04185845 |
map02030 |
|
20 |
Terpenoid backbone biosynthesis |
2 |
2 |
0.04185845 |
map00900 |
%20555-566_files/image014.jpg)
Fig. 7: Metabolic pathway Map 01120 annotated by the KEGG
pathway, a pathway of microorganism in different environments; those marked by
a red rectangle represent annotated nitrogen metabolic pathway
Author Contributions
Yunpeng
Luan proposed the experimental idea of the article. Yunpeng
Luan, Nadezda Vladimirovna Verkhovtseva and Vladimir Mikhailovich Goncharov
designed and supervised the experiment process. Yunqi Luan, Lifang He, Li
Zheng, Dechang Mao, Xiao-Guang Yue, Hongmin Wang and Peng Guo participated
Experiments and sorting out and analyzing the data. Yunpeng Luan mainly writes
the article. All authors participate in writing and revising the article.
References
Baker M (2011).
Metabolomics: From small molecules to big ideas. Nat Meth 8:117–121
Bashir O, K Khan, KR Hakeem, NA Amir, GH Rather, R Mohiuddin (2016). Soil microbe diversity and root exudates as important
aspects of rhizosphere ecosystem. In: Plant, Soil and Microbes, pp:337–357. Springer, Cham,
Switzerland
Buckland SM, JP Grime
(2010). The effects of trophic structure and soil fertility on the assembly of
plant communities: A microcosm experiment. Oikos 91:336–352
Byrt CS, CP Grof, RT
Furbank (2011). C4 Plants as biofuel feedstocks: Optimizing biomass production
and feedstock quality from a lignocellulosic perspective free access. J
Integr Plant Biol 53:120–135
Carvalhais LC, PG
Dennis, D Fedoseyenko (2011). Root exudation of sugars, amino acids, and
organic acids by corn as affected by nitrogen, phosphorus, potassium, and iron
deficiency. J Plant Nutr Soil Sci 174:3–11
Chen CX, J Wang, H Guo, WY Hou, N Yang, BR Biao, M Liu, HQ Dai, XT Liu, FH Song, LX Zhang (2013). Three
antimycobacterial metabolites identified from a marine-derived Streptomyces spp. MS100061. Appl Microbiol Biotechnol 97:3885–3892
Chen YX, DW Zhou, YF
Zhang (2004). Temporal and spatial variations of chemical constituents in
maize. J Appl Ecol 15:1589–1593
Dunn WB, DI Ellis
(2005). Metabolomics: Current analytical platforms and methodologies. TrAc Trend
Anal Chem 24:285–294
Ge WX, XW Niu, WJ Zhang (2017). The nutritional value, processing and
utilization of corn cob and its application in animal husbandry production.
Cer Feed Industr 11:51–54
Geyer T (2013). Modeling metabolic processes between
molecular and systems biology. Curr Opin Struct Biol 23:218–223
Griffiths WJ, Y Wang (2009). Mass spectrometry: From proteomics to
metabolomics and lipidomics. Chem Soc Rev 38:1882–1896
Huang J, XJ Jiang,
YH Zeng (2017). Research progress on evaluation methods and indexes of paddy
soil fertility. Soils Fert Sci Chin 6:1–8
Huang JM (2012).
Research on present situation and development strategies of corn industry in
Yunnan province. Hortic Seed 8:57–62
Huang Y, BC Fu
(2002). The change properties of the chemical compositions for Laterite. J
Kunming Univ Sci Technol 27:63–66
Huang Y, WJ Sun
(2006). Soil science-change trend of organic carbon content of farmland topsoil
in mainland China in the past 20 years. Abstr Chin Acad J 51:1
Hui J, XQ Wu, DX Cai
(2016). Soil nitrogen content and enzyme activities in rhizosphere and
non-rhizosphere of summer corn under different nitrogen application rates. J
Appl Ecol 27:1917–1924
Kuang YW, DZ Wen, CW Zhong, GY Zhou (2003). Root Exudates and Their Roles in Phytoremediation. Acta Phytoecol Sin 27:709–717
Lacina O, J Urbanova, J Poustka (2010).
Identification/quantification of multiple pesticide residues in food plants by
ultra-high-performance liquid chromatography-time-of-flight mass spectrometry. J
Chromatogr A 1217:648–659
Lazcano C (2011). The use of
vermicompost in sustainable agriculture: Impact on plant growth and soil
fertility. Soil Nutr 10:1–23
Li JF, HY Wan, XY Wang (2018). Analysis of soil chemical properties
of five different forest types in Wanning City. Trop For 46:22–25
Li P (2011).
Analyses on rice soil nutrients in Ningde Zhangwan town. J Xiamen Univ-Nat
Sci 50:49–52
Ling-An N, H Jin-Min, Z Bao-Zhong
(2011). Influences of long-term fertilizer and tillage management on soil
fertility of the North China plain. Pedosphere 21:813–820
Liu H (2014). Corn industry: A Multi-pronged Approach
to Get Out of Trouble. Agric Knowledge, China
Liu J (2010). Study on Soil Quality
Characteristics and Fertility Control Technology of Degraded Red Soil in the
Red Soil Region of Central and South Human. Hum Agric Univ, China
Lu QM, ZC Chen, AL He (2018). Application
and development of corn gluten meal, a by-product of corn starch processing. J Food Saf Qual 9:15–22
Lucie N, M Ludmila, P Solich (2006).
Advantages of application of UPLC in pharmaceutical analysis. Talanta
68:908–918
Luo DQ, J Bai, DT Xie (2002).
Discussion on soil fertility evaluation index and method. Ecol Environ Sci
11:202–205
Luo Q, SY Wang, LN Sun, H Wang (2017). Metabolic profiling of root exudates
from two ecotypes of Sedum alfredii treated with Pb based on GC-MS. Sci Rep 7; Article 39878
Ma
XH (2019). Research on quality
characteristics and comprehensive utilization of maize. Cer Oils 273:1–3
Mandal K, B Singh (2013). Persistence of fipronil and its
metabolites in sandy loam and clay loam soils under laboratory conditions. Chemosphere 91:1596–1603
Marschner
P, Z Rengel (2007). Contributions of the
rhizosphere interactions to soil biological fertility. In: Soil Biological Fertility, pp:81–98.
Springer, The Netherlands
Miao CR (2018).
Research on the Nutritional Value of Corn and Its Post-processed Products. J
Chin Food Saf 215:175
Ou JL (2016). An
overview of corn nutrition and key points of quality control. Guangdong Feed
25:38–40
Pang YM (2009).
Research Progress of Soil Fertility Evaluation. J Shanxi Agric Sci 2:87–89
Ram RJ, NC Verberkmoes, MP Thelen, GW Tyson, BJ Baker, RC Blake, M Shah, RH Hettich, JF Ban (2005). Community
proteomics of a natural microbial biofilm. Science 308:1915–1920
Riley D, SA Barber (1970). Salt accumulation at the soybean (Glycine max (L.) Merr)
root-soil interface. Soil Sci Soc Amer J 34:154–155
Song SS (2011). Research on Soil Fertility
Evaluation Method. Northwest A&F University, Xianyang, China
Suthar MS, MM Brassil, G Blahnik, M
Aimee, HJ Ramos, SC Proll, SE Belisle, MG Katze, MJ Gale (2013). A systems
biology approach reveals that tissue tropism to West Nile virus is regulated by
antiviral genes and innate immune cellular processes. PLoS Pathog 9; Article
e1003168
Tong Y (2019). The development course and prospect of
China's corn starch and starch sugar industry technology. Food Ferment Indust
45:294–298
Wang LM (2018). Thoughts on corn production and seed
industry development and breeding goals under the new situation. In: Proceedings
of the 2018 Shandong Crop Society Academic Annual Meeting, Shandong, China
Wang SD (2016). Using corn stover to produce
bio-butanol and its separation and purification. DaLian Univ Technol China
Wang Y, YJ Wang (2016). Application of Corn Pellets as Auxiliary
Material in Beer Brewing. Chinese Foreign Wine Industry and Beer
Technology, China
Wang ZL, Q Fu (2007). Research progress on comprehensive
evaluation of soil fertility. Soil Crop 23:15–18
Want EJ, G Maille, R Abagyan (2006). XCMS:
Processing mass spectrometry data for metabolite profiling using nonlinear peak
alignment, matching, and identification. Anal Chem 78:779–787
Wen B, Z Mei, C Zeng (2017). Meta X: A flexible and comprehensive
software for processing metabolomics data. BMC Bioinform 18; Article 183
Wren SAC (2005). Peak capacity in gradient
ultra-performance liquid chromatography (UPLC). J Pharmaceut Biomed 38:337–343
Xiong SG (2001). Basic Soil Science. Chin
Agric Univ Press, China
Yang XQ, H Zhi, H Zhang (2019). Research progress on chemical
constituents, pharmacological activity and utilization status of different
parts of corn. Jilin J Tradit Chin Med 6:837–840
Yang YQ, GF Chen, HL Guo (2016). Research on big data processing
technology in soil fertility evaluation. J Chin Agric Mech 4:233–236
Zhang Y, BM Feng, X Lu (2017).
Application progress of UPLC/Q-TOF-MS combined technology in drug analysis. Nat
Prod Res Dev 29:1992–1996
Zhang ZD (2011). Research on Soil
Fertilization and Changes of Components During Decomposition of Cow Manure
Mixed with Green Manures. Jilin Agric University, Jilin, Chin
Zhao H, HG Bai (2013). Erratum to: The
critical soil P levels for crop yield, soil fertility and environmental safety
in different soil types. Plant Soil 327:39–39
Zhao KL, BR Wang, MG Xu (2019). Changes in pH
with depths of soils derived from different parent materials and analysis of
acidification in Southern China. J Plant Nutr Fert 25:1308–1315
Zhou LF (1983). Characteristics and zonality of soil distribution in
Yunnan Province. J Mount Res 4:33–40
Zhu SS, JM Vivanco, DK Manter (2016). Nitrogen
fertilizer rate affects root exudation, the rhizosphere microbiome and
nitrogen-use-efficiency of corn. Appl Soil Ecol 107:324–333