Comparative proteomic analysis between mature and germinating seeds in Paris polyphylla var. yunnanensis

Plant materials

The seedlings of Chonglou were planted in Ludian, Yunnan Province, southwest China. The seeds were collected after the fruits matured. The collected mature seeds were primed with 200 ppm GA₃ and were kept at 20 °C to germinate, according to our previously described method (Zhang & Ling, 2019). The priming resulted in the morphological post-maturation of the globular seed embryo on the 20th day of seed incubation (Zhang & Ling, 2019). The mature seeds were collected in the field and the seeds that were germinated until the 20th day of incubation were used as experimental materials. All materials were immediately snap-frozen in liquid nitrogen and stored at −80 °C until used. Three biological replicates of these experimental materials were used to detect the proteomics, the transcriptome and the levels of GAs and ABA, respectively.

The measurement of plant hormones

In this study, two major plant hormones, ABA and GAs, were measured in mature and germinating seeds, respectively. For ABA, the extraction and quantification were carried out using an HPLC-MS/MS (LCMS-8040 System; Shimadzu, Somerset, NJ, USA) (Thameur, Ferchichi & López-Carbonell, 2011). The determination of GAs were performed on an HPLC-MS/MS (LCMS-8040 system; Shimadzu, Somerset, NJ, USA) (Kojima & Sakakibara, 2012).

Protein extraction, digestion, and iTRAQ labeling

Protein samples were prepared as previously described (Yang et al., 2019) and the detailed protocol of protein extraction was shown in Fig. S1. Each sample was ground in liquid nitrogen with the pallet and then mixed with 1.5 ml lysis buffer containing 7 M urea, 2 M thiourea, 0.1% CHAPS and a protease inhibitor. The supernatant was collected after centrifugation at 25,000g for 20 min at 4 °C and then precipitated by adding five times the sample volume of 10% (w/v) trichloroacetic acid/acetone (−20 °C) to the tube overnight. The samples were centrifugated at 15,000g for 20 min at 4 °C and the pellet was rinsed three times with pre-cooled acetone. The precipitate was air-dried and dissolved in lysis buffer, which was used to measure the protein concentration with the Bradford assay.

The 100 μg protein aliquots were digested with 2.5 μg of trypsin at 37 °C overnight. The digested peptides were desalted with Strata X column and then vacuum-dried. Approximately 100 μg of desalted peptides were labelled with isobaric tags from the iTRAQ Reagent-8plex Multiple Kit according to the manufacturer’s instructions. The labelled peptides were pooled in equal amounts and dried by vacuum centrifugation.

LC-MS/MS analysis

The labelled peptide mixtures were redissolved in solution A (5% ACN and 95% ddH₂O, pH 9.8) and then fractionated with a Gemini C18 column in an LC-20AB HPLC system (Shimadzu, Kyoto, Japan). Next, retained peptides were eluted with solution B (95% ACN and 5% ddH₂O, pH 9.8). After separation, the collected fractions were concentrated by vacuum centrifugation and resuspended in solution C (2% ACN, 0.1% formic acid). The fractions were analyzed with an LC-20AD HPLC system (Shimadzu, Kyoto, Japan) coupled to a Q-Exactive mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA). The Q-Exactive mass spectrometer was operated in a data-dependent acquisition mode to switch automatically between MS and MS/MS acquisitions. The survey of the full-scan MS spectra (m/z 350-1,500) was set for 30 s dynamic exclusion. The mass spectrometry proteomics data have deposited to the Zenodo data set (DOI 10.5281/zenodo.6370960) and available.

The bioinformatics analyses of protein data

The raw mass data were used to search against the ‘non-redundant protein sequences (nr)’ database by the blast program. Protein identification was performed using the Mascot sever (Version 2.3.02) against the transcriptome data published in our previous study (Ling et al., 2017; Zhang & Ling, 2019). The confident protein identification involved at least one matched unique peptide with a confidence interval ≥95% according to the mascot probability scores. Meanwhile, iQuant software (v2.2.1, BGI-Shenzhen, Guangdong, China) (Wen et al., 2014) was used to quantify proteins in this study. The false discovery rate (FDR) analyses were conducted using the pocked protein FDR strategy (Savitski et al., 2015). To ensure the high confidence, the peptides with a global FDR_?% were removed in the protein analysis. The significant differentially expressed proteins (DEPs) were selected by fold-change ratio (>1.5 or <0.67, p < 0.05). Functional category analyses were performed with Blast2GO and the cluster of orthologous groups of proteins (COG). The metabolic pathway of DEPs was predicted using the Kyoto Encyclopedia of Genes and Genomes (KEGG) (http://www.genome.jp/kegg/) database. A p-value ≤ 0.05 was used as the threshold of significant enrichment of GO and KEGG pathways.

Transcriptome analysis

Transcriptome data of mature and germinating seeds were assembled and annotated in our previous study (Ling et al., 2017; Zhang & Ling, 2019). Each sample had three biological replicates. The transcriptome sequencing data for mature and germinating seeds used in this study are available in the NCBI SRA database under accession numbers SRX2335704 and SRP126881, respectively. The mRNA abundance was estimated using the number of uniquely mapped fragments per kilobase of transcript per million mapped reads (FPKM) method.

Differentially expressed genes (DEGs) were detected using NOIseq with a threshold of 0.8 (Tarazona et al., 2011). Genes were regarded as significantly differentially expressed when a p-value < 0.001, a FDR < 0.01 and an absolute log2 ratio > 1 in sequence counts across sample transcriptomes were achieved. The FDR was calculated to correct the p-value; the smaller the FDR, the smaller the error in judgment of the p-value. The p-value was calculated as follows:

$p=1-\sum _{i=0}^{m-1}{\displaystyle \frac{\left(\begin{array}{c}M\\ i\end{array}\right)\left(\begin{array}{c}N-M\\ n-i\end{array}\right)}{\left(\begin{array}{c}N\\ n\end{array}\right)}}$where m indicates the DEG number in each GO/KO term; M indicates all gene number in each GO/KO term; N represents the number of all transcripts with GO/KO annotation. The corrected p-value ≤ 0.01 as the threshold determined the significant enrichment of the gene sets.

iTRAQ analyses of proteins in mature and germinating seeds

In total, 66,482 spectra were generated from mature and germinating seeds of the 20th day based on the iTRAQ-LC-MS/MS proteomic analyses (Table 1). After filtering out low-scoring spectra, 43,762 unique spectra were matched to 4,488 proteins at the given thresholds (FDR < 1%, 95% confidence, at least one unique peptide in each positive protein identification) (Table 1). The length of these identified proteins ranged from 34 to 5,078 AA, of which proteins of less than 1,000 AA were the most abundant (94.81%) (Table S1). The average length of the proteins was 441 AA and a few of proteins (0.47%) were greater than 2,000 AA in length (Table S1). Among the identified proteins in mature and germinating seeds, the proteins with a relative abundance of >1.5-fold or <0.67-fold were considered to be DEPs. In this study, a total of 1,305 (642 down-regulated and 663 up-regulated) DEPs were identified in germinating seeds, when compared with mature seeds (Table 1, Table S2 and Fig. S2).

Functional annotation of the DEPs

To understand the functions of the identified DEPs, COG, GO and KEGG enrichment analysis were performed. COG analysis showed that these DEPs were grouped into 24 functionally categories. Most of the DEPs were clustered into the “Carbohydrate transport and metabolism” category, followed by “Posttranslational modification, protein turnover, chaperones” and “General functional prediction only” (Fig. 1). In addition, “Energy production and conversion”, “Amino acid transport and metabolism” and “Lipid transport and metabolism” clusters were found to be the other abundant categories (Fig. 1). The GO enrichment analysis revealed that the DEPs were associated with the different functional processes. At the molecular level, the majority of DEPs were involved in “catalytic activity” and “binding activity” (Fig. 2). Meanwhile, the DEPs were linked to “cell”, “cell part”, “organelles” and “membrane” at the cellular component level (Fig. 2). Most of the DEPs were involved in “cellular and metabolic processes” in biological process. These results indicated that the DEPs were involved in the different metabolic and cellular processes and may localize in the different cell parts and organelles.

Pathway-based analysis is helpful to further understand the biological functions and gene interactions. KEGG pathway analysis revealed that the DEPs were enriched in 132 pathways at the third level. Of these, “metabolic pathways” (359), “biosynthesis of secondary metabolites” (237), “carbon metabolism” (56), “starch and sucrose metabolism” (50), “protein processing in endoplasmic reticulum” (47), “biosynthesis of amino acids” (40), and “phenylpropanoid biosynthesis” (37) were found to be the top seven annotated KEGG pathways for the DEPs (Fig. 3A). At the second level, “metabolic pathways”, “biosynthesis of secondary metabolites” and “carbon metabolism” were classified into “Global and overview maps”, which accounted for the largest proportion (Fig. 3B). However, the DEPs in “metabolic pathways” and “biosynthesis of secondary metabolites” had a small enrichment factor (<0.4) although they had a large enrichment number (Figs. 3A and 4). This data meant that no less than 40% proteins in these above-mentioned pathways showed the different expression patterns between two samples. By contrast, “anthocyanin biosynthesis” exhibited a largest value of rich factor (= 0.8), followed by “photosynthesis”, “folate biosynthesis” and “flavonoid biosynthesis” (Fig. 4).

The comparison of ABA and GAs levels between mature and germinating seeds in Chonglou

Previous studies have revealed that ABA and GA play important roles in seed germination in Chonglou (Chen et al., 2011; Meng et al., 2006; Zhao, Luo & Li, 2014). Figure 5A shows that the average ABA content of mature seeds was 55 ng/g FW, whereas that of germinating seeds of the 20th day was significantly decreased to 9.43 ng/g FW (p < 0.05). In this study, nine endogenous GAs (GA₁, GA₃, GA₄, GA₇, GA₉, GA₁₂, GA₁₉, GA₂₄ and GA₄₄) were unequivocally quantified in mature and germinating seeds of the 20th day. Our results indicated high amounts of GA₁ and GA₃ in the mature and germinating seeds. The content of GA₃ was 179.62 ng/g FW in germinating seeds, and significantly higher than that of mature seeds (p < 0.05) (Table 2). Although the content of GA₁ showed the similar change pattern as that of GA₃ and no significant increase occurred in GA₁ (Table 2). The levels of other GAs (GA₄, GA₇, GA₉, GA₁₂, GA₁₉, GA₂₄ and GA₄₄) showed low amounts in mature and germinating seeds. For example, GA₄ and GA₁₂ were not detected in mature seeds, whereas GA₇, GA₉ and GA₂₄ were not detected in germinating seeds of the 20th day (Table 2). We found that GA₇ showed the significant reduction during seed germination. Altogether, the concentration of the GA₃ and GA₇ showed the significant and opposite changes during seed germination. Meanwhile, we separately calculated the ratio of GA₃/ABA and GA₇/ABA and found they were similar to the respective GAs levels (Fig. 5B).

Characterization and expressional analyses of carotenoid catabolic pathway genes

In plants, the carotenoids biosynthetic pathway provides the precursor for ABA synthesis (Finkelstein, 2013). To understand the molecular basis of ABA level changes during seed germination, we identified the candidate genes involving in carotenoid biosynthetic/catabolic pathway. In this study, a total of 29 candidate genes encoding 26 enzymes were found and their detailed information were summarized in Table 3 and Fig. 6. Of these candidate genes, only six enzymes, including carlactone synthase/all-trans-10′-apo-beta-carotenal 13,14-cleaving dioxygenase (CCD8), fabG, lycopene epsilon-cyclase (LCYE), peroxidase, VPS54 and UDP-glucose:glucosyl (waaD) had only one copy, whereas the rest of carotenoid genes had more than two copies (Table 3 and Table S3).

When germinating seeds were compared with mature seeds, 14 enzymes exhibited the different expressional patterns at the transcriptional level (Table 3 and Table S3). The mRNA levels of abscisic-aldehyde oxidase (AAO3), xanthoxin dehydrogenase (ABA2), beta-carotene isomerase (DWARF27), FMN2, alpha-L-fucosidase (FUCA), CYP97A3, PAF1, 15-cis-phytoene desaturase (PDS), tropinone reductase I (TR1) and VPS54 in germinating seeds were significantly higher than those in the mature seeds. In contrast, 9-cis-epoxycarotenoid dioxygenase (NCED) which is a rate limited enzyme responsible for ABA biosynthesis showed a significant decrease in the transcriptional level in germinating seeds. In addition, 11 transcripts encoding three enzymes, zeta-carotene desaturase (ZDS), zeaxanthin epoxidas (ZEP) and CYP707A were observed as up-regulated or down-regulated significantly (Table S3), which indicated that there are multiple ZDSs as well as ZEPs and CYP707As involved in ABA biosynthesis.

The pattern changes of the carotenoid candidate genes were analyzed at the translational level. Our results indicated that a total of eight candidate genes, including AAO3, beta-carotene 3-hydroxylase (BHY), 15-cis-phytoene synthase (crtB), carotenoid isomerase (CRTISO), violaxanthin de-epoxidase (VDE), ZDS, PAF1 and ZEP (Table 3, Table S3 and Fig. 6) were significantly changed. Of them, only the latter two genes were significantly up-regulated at the translational level during seed germination (Table 3, Table S3 and Fig. 6). In principle, the more mRNA molecules that are present in the cell, the more proteins can be synthesized. Therefore, the correlation of between mRNA and protein expression profiles of the carotenoid candidate genes was analyzed. However, we found that almost none of the unigenes were simultaneously and significantly changed at the translational and transcriptional levels (Table S3). There was one exception: one unigene encoded ZDS was significantly down-regulated at the mRNA and protein levels (Table S3 and Fig. 6). These results indicated that the mRNA and protein changes of the carotenoid candidate genes were not correlated.

Expressional analyses of GA biosynthesis genes at the different levels

We examined the expression of the key genes involving in GA biosynthesis in mature and germinating seeds at the different levels. The proteomic data analysis indicated that the enzymes of gibberellin 20-oxidase (GA20ox) and gibberellin 2-oxidase (GA2ox) exhibited a significant difference (Table S4). Of the three unigenes of GA20ox, two demonstrated slight decreases whereas one increased 3.6 times increase in germinating seeds compared to mature seeds (Table S4). By contrast, one GA2ox unigene exhibited a significant decrease in the germinating seeds (Table S4). Transcriptomic data analysis showed that only GA20ox and GA2ox had significantly changed patterns. Two copies of GA20ox had contrary expressional patterns and three copies of GA2ox were down-regulated during seed germination (Table S4). However, the GA20ox or GA2ox copies did not simultaneously show the significant changes at transcriptional and translational levels and there was still no correlation between the transcriptional and translational levels in GA20ox or GA2ox, two key enzymes involving in GA metabolic pathway (Table S4).