Microspore collection and culture
We attempted to collect only late uninucleate microspores as the initial materials for in vitro heat treatment and culture by bud selection. A highly embryogenic cultivar Zhonggan 628 and a low yield cultivar in microspore-derived embryos 87-534 were chosen for the analysis. Treatments of 32 °C and 25 °C on the isolated microspores for 24 h were adopted in our experiments. The results showed that after two to three weeks, embryoids were produced from the isolated microspore, and the difference in the embryo rates between the 32 °C and 25 °C treatments in Zhonggan 628 are significant. In Zhonggan 628, the average number of embryos per bud is as high as 19.7 after 32 °C treatment, while the average number of embryos per bud is 0 without 32 ℃ treatment, in 87-534, there was no embryo at both 32 °C and 25 °C. (Fig. 1). We found that in vitro 32 °C heat treatment can significantly improve the embryo rate.
Identification of differentially expressed proteins in embryogenesis caused by high temperature
To decipher the proteome variations after treatments at 25 °C and 32 °C for 24 h in cultured microspores, a highly embryogenic cultivar Zhonggan 628 (A) and 87-534 (low yield in microspore-derived embryos, B) after treatments at 25 °C and 32 °C for 24 h were chosen as treatment and control group for the proteome comparisons. Using the label-free analysis, a total of 33434 peptides and 5961 protein species were identified (FDR < 0.01) in the 12 samples (Table S1). Principal component analysis showed clear separation between A and B (Fig. 2). Proteins with more than 1.5-fold changes in abundance (p ≤ 0.05) between the HS treatment (32 °C) and CK (25 °C) were selected as differentially expressed. A total of 363 showed expression differentials in mass spectrometry were identified from Zhonggan 628 (A), 25 °C vs 32 °C, among the 363 proteins, 115 were upregulated and 248 were downregulated (Table S1). In 87-534, 25 °C vs 32 °C, a total of 282 showed expression differentials in mass spectrometry were identified, among the 282 proteins, 162 were upregulated and 120 were downregulated (Table S1). We found that Y1 (A 25 °C vs 32 °C), Y2 (B 25 °C vs 32 °C), and Y3 (A 32 °C vs B 32 °C) have 37 proteins in common, there were 210, 166 and 357 proteins that could only be detected in Y1, Y2, and Y3, respectively, in contrast, there were 97 proteins in only Y1 and Y3, 60 proteins in only Y2 and Y3, and 19 proteins in only Y1 and Y2 (Fig. 3). We can speculate that those 97 proteins specially identified only in Zhonggan 628 after 32 °C heat-shock treatment were the proteins that will require more attention.
GO function classification of differentially expressed proteins
To further understand the function of the differentially expressed proteins, 97 key DEPs were searched as described in the materials and methods to obtain their function information. Using GO functional annotation, differentially expressed proteins could be divided into three categories: cellular component, biological process and molecular function (Fig. 4, Table S2). Molecular function and biological process, with 5 and 14 GO terms, respectively; 11 were assigned to cellular component (Fig. 4). Proteins with roles as the cellular component focused on cell and cell part, proteins related to molecular function were primarily the binding protein and catalytic protease, and the proteins related to biological process were primarily involved in cellular process and metabolic process.
KEGG pathway analysis of differentially expressed proteins
According to KEGG analysis with p ≤ 0.05, 24 DEPs were enriched in carbon metabolism, glyoxylate and dicarboxylate metabolism, starch and sucrose metabolism, protein processing in endoplasmic reticulum, glycolysis / gluconeogenesis, carbon fixation in photosynthetic organisms, plant-pathogen interaction and cutin, suberine and wax biosynthesis. The results in the KEGG pathway analysis of the differentially expressed proteins after HS responsive embryogenesis (Fig. 5, Table S2) showed that carbon metabolism included nine DEPs, protein processing in the endoplasmic reticulum included eight DEPs, glyoxylate and dicarboxylate metabolism included six DEPs.
The pathway of the different proteins was primarily involved in protein process and carbon metabolism, while some proteins will also be enriched to a different pathway. There were a number of proteins involved in many pathways, such as ribulose bisphosphate carboxylase small chain that were involved in three pathways (carbon metabolism, glyoxylate and dicarboxylate metabolism and carbon fixation in photosynthetic organisms), while participating in more than two pathways of protein as well as sugar isomerase (SIS) family protein and glucose-6-phosphate isomerase.
Interaction of differentially expressed proteins in embryogenesis caused by high temperature.
Plant proteins play interrelated roles together in the context of networks. Detailed analyses of the functional and physical protein interactions of 97 differentially expressed proteins (DEPs, 64 downregulated proteins and 33 upregulated proteins) were validated by constructing an Arabidopsis association model, which was performed to investigate the effects of heat shock treatment on microspore embryogenesis in vitro culture. Although these predicted interaction networks need to be verified, they have provided a narrow pool of protein–protein interactions in embryogenesis caused by high temperature in vitro in cabbage for further investigations.
We applied the STRING database to construct the interaction networks of those 97 DEPs in Arabidopsis. The results showed that 97 DEPs (64 downregulated and 33 upregulated proteins) interacted with 32 and 14 target proteins, respectively (Table S3 and Fig. 6). As expected, the network showed general and complex interactions between the 64 downregulated proteins and 33 upregulated proteins, for example, among those 64 downregulated proteins: hsp21 description (HSP21, corresponding to Bo1g050780.1), putative ankyrin repeat protein RF (TPR10, corresponding to Bo1g151900.1), heat stress transcription factor B-4b-like (HSFA7A, corresponding to Bo2g165560.1), sgt1 homolog B (SGT1, corresponding to Bo3g045210.1), chaperone protein ClpB1 (HSP101, corresponding to Bo6g118620.1) and cell division control protein 48 homolog A (CDC48, corresponding to Bo5g138000.1) closely interact with heat shock 70 kDa protein 5 (HSP70, corresponding to Bo5g021500.1, Bo8g066630.1). In addition, 20 kDa chaperonin (CPN20, corresponding to Bo2g023100.1), malate dehydrogenase (mMDH, corresponding to Bo6g031300.1), protein SCO1 homolog 1 (HCC, corresponding to Bo5g139630.1) showed very positive interactions with protein disulfide-isomerase (PDIL1-1, corresponding to Bo8g071060.1). For those 33 upregulated proteins: cytochrome P450, family 86, subfamily C (CYP86C2, corresponding to Bo9g076040.1), non-specific lipid-transfer protein (LTP12, corresponding to Bo1g059070.1), polygalacturonase QRT3 (QRT3, corresponding to Bo1g022160.1) and AT1G06260 (corresponding to Bo1g055000.1) closely interacted with AT1G20120 (corresponding to Bo5g029180.1, Bo7g061770.1, and Bo7g061790.1).
Niehl et al. found that CDC48 function as a cellular factor to regulate the movement of protein accumulation patterns in plant cells, suggesting a general role of CDC48 in endoplasmic reticulum membrane maintenance upon ER stress (ER) [34]. The most commonly described role of mitochondrial malate dehydrogenase is its catalysis of the interconversion of malate and oxaloacetate in the tricarboxylic acid cycle [35]. The roles of mMDH in Arabidopsis seed development and germination were investigated in mMDH1 and mMDH2 double knockout plants; a significant proportion of mMDH1mMDH2 seeds were nonviable and developed only to torpedo-shaped embryos, indicative of arrested seed embryo growth during embryogenesis [36]. Arabidopsis contains two SGT1 isoforms, SGT1a and SGT1b, and a double mutant would lead to lethal embryos in Arabidopsis, suggesting that the SGT1 proteins are essential for plant development; the SGT1 genes are also involved in enhancing the plant responses to various biotic and abiotic stresses [37,38]. The developmental induction of sHSP has been related to the potential for stress-induced microspore embryogenesis, and the ability of late bicellular pollen to respond to embryogenic induction treatment was accompanied by rearrangements of the microtubular cytoskeleton and the nuclear localization of the 70 kDa heat shock proteins [39,40]. These major interacting proteins may play an important role in cabbage embryogenesis.
Correlation of protein fold changes with transcripts.
The protein data and mRNA expression were correlated to further investigate protein relative abundance profiles of embryogenesis caused by high temperature treatment. The mRNA expression levels were obtained using the qRT-PCR analysis of the corresponding nine genes (Those are more nodes in the interaction network). Of the selected proteins, four of the genes showed similar change trends as the result of being label-free. As shown in Fig.7, we chose heat shock 70 kDa protein 5 (Bo5g021500, Bo8g066630), one of the most abundant proteins isolated from the microspore at 32 ºC, and, as expected, a significant increase was observed in the expression of HSP70 gene in response to high temperature, similar to HSP70, two genes encoding protein SCO1 homolog 1 (HCC) and heat stress transcription factor B-4b-like (HSF) showed a relative increased in abundance in Zhonggan 628 under high temperature (Fig. 7). Differently, the gene encoding protein HSP21, CPN20, mMDH, sSGT1and CDC48 showed different changes as a result of being label-free. Those genes showed a low correlation between the mRNA and protein abundance profiles, which displayed a discrepancy between label-free proteomics and qRT-PCR that could be due to sensitivity between the two analytical methods. The other possible reason for this discrepancy maybe introduced by protein posttranscriptional, translational, and posttranslational mechanisms or feedback loops between the processes of mRNA translation and protein degradation.