Development of Sugar beet Guard Cell Enrichment Protocol
Since our experiment starts with salt stress treatment of the plant roots and then focuses on analyzing guard cells on the leaves, we need to use a fast and efficient guard cell enrichment method. The enrichment process of guard cells is divided into two steps: isolation of epidermal fragments and digestion of epidermal fragments. There are two commonly used methods for obtaining the epidermal fragments, the transparent tape-peel method  and the blender pulverization method [1, 25]. We first obtained the epidermal fragments of BvM14 line through these two methods (Figure 1A, 1B). The obtained epidermal fragments were then treated by 2.1% cellulase R-10, 0.075% macerozyme R-10 (Figure 1C, 1D). We observed that the guard cells obtained by the tape-peel method has a large amount of mesophyll cells contamination even after the enzyme treatment. In contrast, the guard cells obtained by the blender method are less contaminated, but this method requires enzymatic digestion for six hours to achieve high quality. This long processing time may cause artificial molecular changes, which can affect downstream proteomic analyses . Therefore, we chose to use the blender pulverization method to obtain epidermal peers and adjusted the type and concentration of enzymes to shorten the enzymatic hydrolysis time to 20 minutes. Finally, we found that the guard cells could be rapidly enriched by using 4.2% cellulase R-10, 0.075% macerozyme R-10 and 0.2% pectolyase Y-23 [1, 2, 27]. As shown in the Figure 1D, the pavement cells were digested away, and stomatal guard cells were of high quality. Purity was further confirmed by a purity assay of the guard cell samples using Real-Time PCR for a guard cell-specific transcript H+-ATPase (AHA1). As shown in Figure S1A, the transcript levels of sugar beet AHA1 (specifically expressed in guard cells) was 70 times higher in the enriched guard cells than in leaves, indicating little contamination of guard cell samples from mesophyll cells. In addition, high level distribution of AHA1 expression in different cells of leaves from ePlant (https://bar.utoronto.ca/ePlant/) has further proved that AHA1 is a guard cell-specific gene (Figure S1B). Our method greatly shortens the enrichment time of guard cells and reduces contamination from other types of cells, such as mesophyll cells and epidermal pavement cells. It is suitable for guard cell proteomics experiments.
Stomatal Movement and Activity of APX in Response to Salt Stress
Previous studies have shown that when plants are exposed to salt stress, stomata may close transiently, but over the long term adaptation, plants will not compromise diurnal stomatal movements in response to the day-night cycle [28, 29]. To investigate responses of stomatal movement to short-term salt stress, we measured how salt stress affected BvM14 line stomatal aperture at seven different time points (0, 10, 20, 30, 40, 50, 60 min), and selected the time points when stomatal aperture became the smallest for proteomic analysis. In order to determine the most suitable method for stomatal movement assays. We first referred to a method in a previous study for measuring the stomatal aperture of Triticum aestivum . We spread clear nail varnish on the abaxial side of the selected leaves. Then peeled the dried nail polish molds off with a tweezer. The slides were visualized under a laser scanning confocal microscope (Olympuj, Japan) (Figure S2A). This method clearly reflects the size of the stomata, but in subsequent experiments we found that this method has poor repeatability and a lot of damage to the leaves, thus is not suitable for our experiments. After consultation and literature searching, we tested a method that involves fixation of the leaves . This method turned out to be reproducible, but it took a long time. (Figure S2B). It still did not meet the need of our experiment. Finally, through our own optimization, we were able to capture the changes of stomatal apertures in a short period of time (Figure S2C). Briefly, after tearing off the abaxial epidermal peels from the sugarbeet M14 leaves, we fixed the leaves in a freshly prepared Carnoy fixative solution (3 parts 100% ethanol : 1 part acetic acid, v/v) for 45 seconds, and then dried and viewed under microscope. Using this method, we observed that the stomatal aperture reached the lowest level under 200 mM and 400 mM NaCl treatments at 20 and 30 minutes, respectively (Figure 2A). Since the closing of the stomata is caused by the perception of ABA and H2O2 signals triggered by salt stress, the changes in protein levels may precede the changes in the stomatal aperture. Thus, in order to accurately determine the optimal time points for proteomics of guard cell response to salt stress, we measured APX activities, which can reduce H2O2 contents in plant. As shown in Figure 2B, under 200 and 400 mM NaCl treatments, the APX activities reached the highest levels at 10 min and 20 min, respectively. The time for APX activities to reach its peak is 10 minutes earlier than the time for stomatal aperture to reach its smallest level, regardless of whether the plants were treated with 200 or 400 mM NaCl. Therefore, we used the APX assay results to determine the time points for the proteomics experiments.
iTRAQ Analysis of Differentially Abundant Proteins (DAPs)
To examine the proteomic changes in guard cells in response to the salt stress, three biological replicates were analyzed for each control and treatment. Under the 200 mM NaCl treatment, 12,672 peptide spectrum matches (PSMs) and 1796 proteins were identified against the Sugar beet database, of which 1069 of the proteins had at least two unique peptides. Under the 400 mM NaCl treatment, 13,170 PSMs and 1609 proteins were identified, where 989 of the proteins had at least two unique peptides. The peptide number distribution of proteins indicates that approximately 60% of identified proteins contained more than two unique peptides (Figure S3). The distribution of protein mass is mostly in the range of 10 to 60 kDa (Figure S4).
To identify proteins significantly affected by salt stress. Protein with at least two unique peptides were used to screen for DAPs with a p-value < 0.05 and a fold-change > 1.2. Under the 200 mM NaCl treatment, 80 DAPs were identified, of which 40 were increased and 40 were decreased. Under the 400 mM NaCl treatment, 72 DAPs were identified, including 17 increased and 55 decreased compared to control samples. The DAPs with fold-changes > 2 were marked in the volcano plot (Figure 3A and 3B). At the level of individual proteins, we found that a non-specific lipid-transfer protein (W6JNH5) was increased by 3- and 2-fold under 200 mM and 400 mM NaCl, respectively. Non-specific lipid-transfer protein (nsLTP) has been found to modulate plant tolerance to salt, drought, cold stresses, as well as defense against bacterial and fungal pathogens [32-34]. Our results indicate that nsLTP may play a vital role in the process of guard cells against salt stress. The overlap between the DAPs is shown in Figure 3C. In the two different salt concentration groups, five proteins were observed to be regulated in the same trend (both decreased or both increased in two groups). The other DAPs did not show this kind of concerted changes. The result indicates some shared components between the two salt treatment groups had conserved responses to the short-term salt stress treatments. A detailed description of DAPs is shown in Supplemental Tables S1 and S2.
The global expression of DAPs was further estimated using hierarchical clustering analysis. The results were displayed in the heat map (Figure 4A and 4B). The cluster analysis showed that the proteins of both groups were similar in distance and preferentially sorted together, indicating that the sample repeatability of the control groups and the treatment groups was good. At the same time, results also showed that the DAPs were well-distinguished and selected. Overall, the heat map provided a better visualization of the proteomic changes between control groups and treatment groups.
Bioinformatic Analysis of the DAPs
Bioinformatic analysis of the DAPs can provide an in-depth understanding of the proteomics result . To reveal potential molecular mechanisms underlying the guard cell response to the salt stress, gene ontology (GO) functional classification of the DAPs was conducted based on the molecular function, cellular components and biological processes. We compared the GO classification of DAPs under 200 mM and 400 mM NaCl treatments. As shown in Figure 5A, both salt stress conditions showed similar patterns in molecular function, cellular components, and biological process. In the molecular function analysis, proteins with binding and structural constituent of ribosome were differentially expressed under the 200 mM NaCl and 400 mM NaCl treatments. These proteins include photosynthesis-related proteins (oxygen-evolving enhancer protein 2 and oxygen-evolving enhancer protein 3), which may play an important role in energy balance in response to the salt stress. In the cellular component, the proteins were mainly located in the plastid. Compared with the 200 mM NaCl result, the 400 mM NaCl treatment had more proteins in the ribosome and cytosol. In the biological process, proteins were enriched in 36 processes, including response to stress, translation, catabolic process, carbohydrate metabolic process, etc., implying that severe salt stress induced DAPs in a larger number of biological processes. AgriGO functional enrichment results were consistent with those in Figure 5A. In the molecular function category, proteins with RNA binding and structural constituent of ribosome were enriched (Figure S5). In terms of cellular components, plastid, chloroplast part, cytoplasm and organelle part proteins were highly enriched (Figure S6). In biological processes, proteins involved in response to stress, response to abiotic stimulus and translation were significantly enriched (Figure S7).
To gain insights into the potential functions of the DAPs and their metabolic pathways, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was conducted (Figure 5B). Clearly, there were striking differences between the DAPs from the two different salt stress conditions. Under 200 mM NaCl treatment, most of DAPs involved in phenylpropanoid biosynthesis, glutathione metabolism, purine metabolism, and starch and sucrose metabolism pathways were differentially expressed. In contrast, under 400 mM NaCl treatment, most of DAPs involved in carbon fixation, glyoxylate and dicarboxylate metabolism, pyruvate metabolism, and methane metabolism pathways were differentially expressed. These results showed that different pathways were deployed under different salt stress conditions. Higher stress levels caused slowdown in biosynthesis and increase of cellular damage and catabolic activities. Under 200 mM NaCl treatment, five proteins (glutaredoxin-like isoform, ascorbate peroxidase, dehydroascorbate reductase, fasciclin-like arabinogalactan protein and glutathione s-transferase) associated with glutathione metabolism were identified. All the proteins except the glutaredoxin-like isoform, were increased, suggesting activation of glutathione metabolism under the 200mM NaCl. Among non-enzymatic antioxidants, glutathione is one of the most abundant soluble antioxidants in higher plants . It plays a vital role as electron donors and scavenge ROS directly through ASA-GSH cycle . However, under the 400 mM NaCl treatment, we didn’t identify any DAPs associated with glutathione metabolism. This result supports that different pathways were deployed, and may indicate disfunction of antioxidant systems under high salt stress .
Subcellular localization analysis revealed that most of the DAPs were targeted to the chloroplast, nucleus and cell wall under the 200 mM NaCl treatment (Figure 5C). Under the 400 mM NaCl treatment, most of the DAPs were targeted to the chloroplast, plasma membrane and cytoplasm (Figure 5D). The different subcellular locations of the guard cell DAPs from the 200 mM and 400 mM NaCl treatments highlight differential regulations and plasticity of the guard cell proteome in response to different stress conditions. To further comprehend potential interactions among the DAPs, the STRING protein interaction database was used to analyze protein-protein interaction (PPI). Among the DAPs, six pathways were significantly enriched including: cysteine and methionine metabolism, RNA transport, carbon fixation, glyoxylate and dicarboxylate metabolism, spliceosome and ribosome (Figure 6). The interactions between proteins associated with spliceosome and ribosome appear to be more complicated, and most of these proteins were decreased.
Transcriptional analysis of the Genes Encoding the DAPs in Guard Cells
The utility of using enriched stomatal guard cells is to correlate guard cell molecular changes to the salt stress-induced stomatal movement. Here we focused on 16 DAPs potentially involved in stress response and stomatal movement. Five proteins were identified in both the 200 mM NaCl and 400 mM NaCl samples (i.e., glycine-rich cell wall structural protein-like, non-specific lipid-transfer protein , translation initiation factor IF-3 , ATP-dependent Clp protease proteolytic subunit and glycine-rich RNA-binding protein RZ1A), 5 were identified only in the 200 mM NaCl samples (i.e., DUF642, protein aspartic protease in guard cell 1, temperature-induced lipocalin, salt tolerance protein 5 and UPF0603 protein), and 6 were identified only in the 400 mM NaCl samples (i.e., salinity-induced protein, jasmonate-induced protein homolog, salt tolerance protein 6, cation/calcium exchanger 5, fasciclin-like arabinogalactan protein 10 and V-type proton ATPase subunit G).
To compare and contrast the changes at the transcription level and protein level of the 16 DAPs, we analyzed their transcription level changes using real-time PCR (Figure 7). Among the 10 DAPs under 200 mM NaCl treatment, the transcriptional levels of 7 genes were consistent with the corresponding protein level trend. The aspartic protease in guard cell 1, salt tolerance protein 5 and UPF0603 protein displayed different transcriptional changes from the corresponding protein level changes. Under the 400 mM NaCl treatment, among the 11 DAPs, 6 genes showed transcriptional changes inconsistent with the corresponding protein level changes. The inconsistency between the changes at the transcriptional level and at the protein level may be attributed to differential stability of the molecules, as well as posttranscriptional and/or posttranslational modifications.