Two batches of seeds were used in the present study. Seeds collected in October 2013
were used for the determination of gas permeability and water permeability of seed
coat. The presence or absence of a gas barrier is identified by determining the difference
in respiration rate between the intact primary dormant seeds and the seeds with their
seed coats were cracked. By monitoring the changes in the masses of intact primary
dormant seeds under germination conditions, we determined whether the seed coat interferences
with water uptake. Seeds collected in 2014 were used for metabolomics analysis.
Fresh Korean pine cones were collected from 30 trees aged 50 years old in a Korean
pine plantation in October 2013 and October 2014, respectively, at the Qingyuan Forest CERN, Chinese Academy of Sciences (CAS), Northeast China (41º51.102’ N, 124º54.543’ E, 456-1116 m a.s.l.). The gymnosperm
cones were opened to release the seeds with the help of threshing machine. Then, these
fresh seeds are dried indoors until the moisture content of the seeds is around 10%.
Determination of water permeability and gas permeability of seed coat
Those seeds that collected in October 2013 were immediately used for the determination
of gas permeability and water permeability of seed coat. We determined the respiration
rates of intact primary dormant seeds and the seeds with their seed coats were cracked.
Every 25 seeds (intact primary dormant seeds or the seeds with their seed coats were
cracked) were placed in a 9-cm diameter Petri dish. There were eight layers of filter
paper in each Petri dish. These filter papers were previously moistened with 12 ml
deionized water. Each Petri dish was placed in a 1.3L air-tight plastic box. At the
same time, a beaker containing 10 ml NaOH (1 mol L-1) solution was also put in plastic box to trap CO2. The plastic box was then incubated in light/dark (14/10 h) at alternating temperature
regimes of 25/10°C in a controlled growth chamber (MGC-450HP-2, Bluepard, Shanghai,
China). This incubation temperature was selected according to the NFSSTC standards
for the Pinus. Photosynthetic photon flux density was about 120 μmol m-2 s-1 located at the top of Petri dishes. BaCl2 (1 mol L-1) was used to precipitate CO2 absorbed by NaOH in the beaker. Then the remaining NaOH in the beaker was titrated
with 1 mol L-1 HCL. The volume of HCL that consumed during the titration is used to calculate the
amount of CO2 produced by seed respiration. Deionized water was added at any time during incubation
to ensure suitable water conditions for seed respiration. The respiration rate was
measured every day and expressed as µmol CO2 g-1 minute-1. Three replicates were carried out separately for seed respiration.
Every 25 intact primary dormant dry seeds were placed in a 9-cm diameter Petri dish.
Eight layers of filter paper moistened with 12 ml deionized water were placed in each
Petri dish. Three Petri dishes were used as three replicates. The Petri dish was covered
with plastic wrap to reduce moisture loss and then incubated under germination conditions.
Every 24 h, seeds were removed from the filter paper, blotted dry, weighed to the
nearest 0.1mg and returned to the Petri dishes. The percentage of increase in seed
mass was calculated using the following formula. The percentage of increase in seed
mass = [(the masses of seeds incubated under germination conditions – the masses of
primary dormant dry seeds)/ the masses of primary dormant dry seeds] × 100.
Preparation of seeds that were used for metabolomics analysis
The seedlot collected in October 2014 was divided into two portions, one of which
was immediately stored at -20°C until the experiment examining the metabolism of primary dormancy. The other portion of the seedlot was moist chilled in early November. The specific
procedure of moist chilling was as follows: seeds were first given a 7-day running
water soak and then buried in soil at 50 cm depth for 6 months (November 2014 - April
2015) in the Korean pine plantation.
After moist chilling, the germination capacity of moist chilled seeds was assessed
to ascertain whether they had been released from primary dormancy. The primary dormancy
released seeds in the present study refer to those seeds after approximately 6 months
of moist chilling. In late April 2015, fresh dry seeds were removed from -20°C storage
conditions. The germination experiment was subsequently conducted with these seeds
under germination conditions to determine whether the seeds were primary dormant status.
The primary dormant seeds in the present study refer to those dry seeds after approximately
of 6 months of storage at -20°C. Primary dormancy released seeds and primary dormant
seeds were then used for metabolomics analysis.
Each germination test consisted of three replications of 20 seeds each. These 20 seeds were placed in a 9-cm diameter Petri dish with eight pieces of filter
paper (Xinxing, Hangzhou, China) moistened with deionized water. Deionized water was added to the Petri dish to ensure appropriate moisture required for seed
germination. All dishes were wrapped with plastic film (Miaojie, Shenyang, China)
to reduce water loss and then incubated in light/dark (14/10 h) at alternating temperature
regimes of 25/10°C in a controlled growth chamber (MGC-450HP-2, Bluepard, Shanghai,
China). Photosynthetic photon flux density was about 120 μmol m-2 s-1 located at the top of Petri dishes. Germination was considered completed and was
then recorded when radicle protrusion was greater than 2 mm and was assessed at two days intervals for 6 weeks. At the end of the germination
test, the seeds failing to complete germination were cut to test the seed embryo viability with a tetrazolium method.
Experimental design of metabolomics analysis
The results on moist chilled seeds germination revealed that the radicle protrusion
of the Korean pine seeds started at 11 days of imbibition. Primary dormancy released
seeds were incubated under a germination inductive temperature for 11 days. The following
seed samples were selected for metabolic analysis: primary dormancy released seeds,
5-day incubated primary dormancy released seeds corresponding to germination sensu stricto (i.e. none of the seeds showed visible germination at this stage), and 11-day incubated
primary dormancy released seeds corresponding to germination sensu stricto (conditions subsequently abbreviated PDRS, PDRS5, and PDRS11, respectively).
Primary dormant seeds were incubated under a germination inductive temperature for
11 days. The following seed samples were selected for metabolic analysis: primary
dormant seeds, 5-day incubated primary dormant seeds, and 11-day incubated primary
dormant seeds (conditions subsequently abbreviated PDS, PDS5, and PDS11, respectively).
After transfer of PDS to germination conditions for incubation, these seeds cannot
germinate. Therefore, we recognize that the primary dormancy of seed was maintained
during 11 days of incubation.
Specifically, every 20 PDRS were placed in a 9-cm diameter Petri dish. Every 20 PDS
were also put into a 9-cm diameter Petri dish. There were 12 Petri dishes for PDRS
and 12 Petri dishes for PDS, respectively. Those PDRS (PDS) in 12 Petri dishes were
then incubated under germination conditions for 11 days in a same manner that described
in germination test. Six Petri dishes containing PDRS and six Petri dishes containing
PDS were used as replicates both at 5 days after incubation and 11 days after incubation.
There were also six replications of 20 PDRS each and six replications of 20 PDS each
before incubation. Before incubation, after 5 days of incubation and 11 days of incubation,
those seeds in six Petri dishes were removed and washed three times with deionized
water and then dried with a filter. The embryos were dissected from the rest of the
seed structure, immediately frozen in liquid N2 and pulverized in liquid N2, lyophilized and stored at -20°C until metabolite analysis. Six replicates for each seed treatment (PDRS, PDS) and before sampling and at each
sampling time were analyzed separately for metabolite measurements.
The 100 mg embryo powder was put into a 2 mL Eppendorf tube and extracted with 1.5
mL of 80% methanol. Tridecanoic acid (0.4 μg mL-1) was applied as an internal standard. In order to adequately extract the metabolites,
the sample was first vortexed for 5 minutes and then centrifuged at 20598 x g for 10 minutes. After that, 800 μL of the supernatant was removed into another 1.5
mL Eppendorf tube and lyophilized for 10 h. Then, 100 μL of methoxyamine solution (20 mg mL-1) that was prepared with pyridine was added to the Eppendorf tube to dissolve the
dry residue. This solution sample was then incubated in a water bath for 90 minutes
at 37°C to perform the oximation reaction. Subsequently, 80 μL of MSTFA (N-Methy-N-(trimethyl-silyl)
trifluoroacetamide) was added to the sample and incubated in a water bath for 60 minutes
at 37°C to conduct the silylation reaction. Then, 200 μL of the supernatant was removed
into a sample vial the lid crimped in place and used for gas chromatography-mass spectrometry (GC-MS) analysis.
A QP 2010 GC-MS equipped with an AOC-20i automatic injector (Shimadzu, Japan) was
used in the present study. Chromatographic separation of metabolites was accomplished
on a 30 m×0.25 mm×0.25 μm DB-5 MS column (J&W Scientific, Folsom, CA, USA). The injection
temperature was 300°C. Helium (99.9995%, ShunTai, ShenYang, China) was used as a carrier
gas. The constant flow of the carrier gas was set as 2.4 mL minute-1. The injection volume was 1 μL. The split ratio was set as 10:1. In order to achieve
ionization of the metabolites, the electron impact model at 70 eV was used. The temperature
of the interface was 280°C. The temperature of the ion source was set to 230°C. The
mass spectra scan scope ranged from 33 to 500 m z-1. The scan speed was 5 scans second-1. The solvent cut time was 5.7 minutes. The column temperature was maintained at 70°C
for the first 3 minutes and then increased to 310°C at the rate of 5°C minute-1. The 310°C was maintained for 5 minutes.
The peak deconvolution and identification analysis were conducted with the AMDIS (automated
mass spectral deconvolution and identification system, National Institute of Standards
and Technology) software by mass spectra matching. Furthermore, the mass spectra,
retention time, and retention index of commercial standards were also used to identify
the structures of metabolites. As long as peak area is less than 1000 or signal-to-noise ratio is below 20, the peak will be excluded. In order to reduce the probability of false positives and eliminate artifact peaks,
methods described in several studies were also applied in the present study[48-50].
The germination percentage was defined as the number of seeds completing germination/total
number of viable seeds × 100%. Principal component analysis (PCA) was performed to
visualize the classification of 6 samples (PDS, PDS5, PDS11, PDRS, PDRS5 and PDRS11)
using the MetaboAnalyst (http://www.metaboanalyst.ca/). All metabolite data were normalized by using unit-variance (UV) scaling before
performing PCA. In UV-scaling, each variable was mean-centered and divided by the
standard deviation. In order to visualize the patterns of change in the differential
metabolites, hierarchical cluster analysis (HCA) was conducted with MetaboAnalyst
on the basis of Pearson correlation coefficients between different metabolites. The
HCA plot was then divided into seven groups to clearly show the relationships and
trends among the differential metabolites in the six samples. Partial least squares
discriminant analysis (PLS-DA) was then used to identify the differentially expressed
metabolites that contributed to the separation of each of the four pairs of samples.
Therefore, a total of four partial least squares discriminant analyses were conducted.
Specifically, those differentially expressed metabolites were identified by inspecting
loadings plots from PLS-DA. Only those metabolites with VIP (variable importance in
the projection) > 1 and significant changes (P < 0.05) in content were considered to have significant contributions to the classification
of each of the four pairs of samples[48, 51, 52]. These four pairs of samples are as follows: PDRS5 vs PDRS, PDRS11 vs PDRS5, PDS5
vs PDS and PDS11 vs PDS5. Fold changes of the metabolites with VIP > 1 between each
of the four pairs of samples were also calculated. Fold change was calculated as the
ratio between two group means and then log2 transformed. Specifically, they were calculated by the following formulas: log2(PDRS5/PDRS), log2(PDRS11/PDRS5), log2(PDS5/PDS) and log2(PDS11/PDS5). The calculation of fold changes was conducted with MetaboAnalyst. Those metabolites
with VIP > 1 were then subsequently subjected to metabolic pathway analysis to identify
and visualize the significantly altered metabolic pathways between PDRS5 vs PDRS,
PDRS11 vs PDRS5, PDS5 vs PDS and PDS11 vs PDS5. The pathway enrichment analysis and
pathway topological analysis in the MetaboAnalyst web tool were used together to perform
metabolic pathway analysis. The P value calculated with the‘Global Test’algorithm in pathway enrichment analysis was
used to indicate the significance of enriched metabolic pathways. The impact value calculated with the ‘Relative-Betweenness Centrality’ algorithm
in pathway topological analysis was used to estimate the relative importance of metabolic
pathways. The impact-value threshold was set to 0.1 to identify the most relevant metabolic
pathways[55, 56]. Finally, the result of the metabolic pathway analysis (metabolome view) was rendered
in a graph format. The P value of each metabolic pathway was log-transformed and then set as the Y-axis and
the pathway impact value was set as the X-axis. Each node in a graph represents a
metabolic pathway. The node color indicates the P value of each metabolic pathway, while the node radius indicates the impact value
of each metabolic pathway. Therefore, dark red, large circles located in the top right corner of the “metabolome
view” represent the main altered pathways compared to the yellow, small circles located
in the left of the graphs. The software VANTED was then used to visualize the pathway
map of the significantly altered metabolites.