Proteomics of Salt Gland–Secreted Sap Indicates a Pivotal Role for Vesicle Transport and Energy Metabolism in Plant Salt Secretion

Soil salinization is one of the major factors restricting crop growth and agricultural production worldwide. Recretohalophytes have developed unique epidermal structures in their aboveground tissues, such as salt glands or salt bladders, to secrete excess salt out of the plant body as a protective mechanism from ion damage. Three hypotheses were proposed to explain how salt glands secrete salts: the osmotic hypothesis, a hypothesis similar to animal fluid transport, and vesicle-mediated exocytosis. However, there is no direct evidence to show whether the salt gland–secreted liquid contains landmark proteins or peptides which would elucidate the salt secretion mechanism. In this study, we collected the secreted liquid of salt glands from Limonium bicolor, followed by extraction and identification of its constituent proteins and peptides by SDS-PAGE and mass spectrometry. We detected 214 proteins and 440 polypeptides in the salt gland-secreted droplets of plants grown under control conditions. Unexpectedly, the proportion of energy metabolism-related proteins increased significantly though only 16 proteins and 35 polypeptides in the droplets of salt-treated plants were detected. In addition, vesicle transport proteins such as the Golgi marker enzyme glycosyltransferase were present in the secreted sap of salt glands from both control and salt-treated plants. These results suggest that trans-Golgi network-mediated vesicular transport and energy production contributes to salt secretion in salt glands.


Introduction
Soil salinization severely restricts global agricultural production and plant growth. The rapid expansion of saline-alkali land has become one of the most urgent problems in the world. In addition to naturally occurring salinized land, secondary soil salinization resulting from unreasonable irrigation, climate change, and seawater intrusion is also becoming more widespread, leading to a marked decrease in farmland and total grain production [1]. The growth of most land plants (named glycophytes or nonhalophytes) is significantly inhibited under salt stress, such as reduced plant biomass, delayed flowering, and altered root system architecture. A concentration of 50 mM NaCl will significantly inhibit the growth and development of glycophytes such as maize (Zea mays), soybean (Glycine max), and rice (Oryza sativa) [2,3]. More importantly, some plants (named halophytes) can grow and complete their entire life cycle on saline soils containing at least 200 mM NaCl. Only about 1% of all land plants are halophytes [4,5], and recretohalophytes, such as L. bicolor [5], and salt-dilution halophytes, such as Suaeda salsa [6,7] belong to the typical euhalophytes. Only recretohalophytes have developed unique structures that distinguish them from glycophytes and other halophytes: salt glands and salt bladders [5,8,9]. explored their possible functions using bioinformatics analysis to decipher the possible mechanism behind salt secretion.

Characteristics of L. bicolor Salt Glands
Salt glands of L. bicolor possess spontaneous blue fluorescence from ferulic acid, a constituent of cutin [30]. The density of salt glands was stable, at about 17 glands in each 100× field of view of the microscope [31]. Most salt glands were located near the veins and were scattered over the leaf surface rather than in clusters. Salt glands had four fluorescent foci (Figure 1a,b) and four secretory pores (Figure 1c). Salt crystals were observed after the excretory pores of salt glands excreted liquid treated by salt treatment (Figure 1d).

SDS-PAGE Analysis of Proteins in the Secretion Liquids
The leaf disc method was used to analyze the proteins in the secretion liquids. In order to verify the effect of leaf disc injury on the protein secretion of the salt gland, we performed Evans blue staining on leaf discs treated with different treatments and found that isolated leaf discs only showed marginal cell damage, while different treatments did not affect the relative activity of salt gland cells ( Figure S2). When the leaf discs were treated with 200 mM NaCl, the concentration of Na + , K + , and Cl − in secretion sap secreted from salt glands were 4.08 μM and 4.64 μM, but the concentration of Na + and Cl − were 0.54 μM and 0.71 μM under the control, which were significantly lower than NaCl condition. While the concentration of K + was 0.12 μM significantly decreased under NaCl

SDS-PAGE Analysis of Proteins in the Secretion Liquids
The leaf disc method was used to analyze the proteins in the secretion liquids. In order to verify the effect of leaf disc injury on the protein secretion of the salt gland, we performed Evans blue staining on leaf discs treated with different treatments and found that isolated leaf discs only showed marginal cell damage, while different treatments did not affect the relative activity of salt gland cells ( Figure S2). When the leaf discs were treated with 200 mM NaCl, the concentration of Na + , K + , and Cl − in secretion sap secreted from salt glands were 4.08 µM and 4.64 µM, but the concentration of Na + and Cl − were 0.54 µM and 0.71 µM under the control, which were significantly lower than NaCl condition. While the concentration of K + was 0.12 µM significantly decreased under NaCl condition compared with 0.55 µM under control condition ( Figure S3). To determine whether proteins were secreted by salt glands, we separated the sap collected from salt glands using SDS-PAGE, followed by silver staining to detect proteins. Indeed, we detected many proteins in the secretion liquids from both control and NaCl-treated salt glands, with most proteins showing molecular weights ranging from 15 to 65 kDa ( Figure S4). Unexpectedly, NaCl treatment led to a marked drop in the protein numbers and amounts.

Protein Identification by Mass Spectrometry
We determined the protein identity and their individual mass in the secreted liquids from salt glands using mass spectrometry. We observed that the protein mass in the secretion sap from control and NaCl-treated leaf discs was similar ( Figure 2), with most proteins having a molecular weight of 30~50 kDa. However, we detected far more proteins in the secretion sap of control leaf discs than from NaCl-treated leaf discs, which was consistent with the SDS-PAGE results.
condition compared with 0.55 μM under control condition ( Figure S3). To determine whether proteins were secreted by salt glands, we separated the sap collected from salt glands using SDS-PAGE, followed by silver staining to detect proteins. Indeed, we detected many proteins in the secretion liquids from both control and NaCl-treated salt glands, with most proteins showing molecular weights ranging from 15 to 65 kDa ( Figure  S4). Unexpectedly, NaCl treatment led to a marked drop in the protein numbers and amounts.

Protein Identification by Mass Spectrometry
We determined the protein identity and their individual mass in the secreted liquids from salt glands using mass spectrometry. We observed that the protein mass in the secretion sap from control and NaCl-treated leaf discs was similar ( Figure 2), with most proteins having a molecular weight of 30~50 kDa. However, we detected far more proteins in the secretion sap of control leaf discs than from NaCl-treated leaf discs, which was consistent with the SDS-PAGE results.

Annotation of Proteins and Peptides
To explore the possible functions of the proteins and peptides detected in the secreted liquid of salt glands, we subjected them to Gene Ontology annotation. We thus determined that the proteins in the secreted sap of control leaf discs have different subcellular localizations, such as chloroplasts, mitochondria, peroxisomes, the cytosol, and the Golgi apparatus. About 17% of all control proteins localized to chloroplasts, 26% in cytoplasm, 7% in mitochondria, and 5% in peroxisomes ( Figure 3). In addition, 38% of the proteins were uncharacterized or unknown. We noticed several organelle marker enzymes such as glycosyltransferases (Golgi marker), catalase (peroxisome marker), and cytochrome oxidase (mitochondrion marker) among salt gland-secreted proteins.

Annotation of Proteins and Peptides
To explore the possible functions of the proteins and peptides detected in the secreted liquid of salt glands, we subjected them to Gene Ontology annotation. We thus determined that the proteins in the secreted sap of control leaf discs have different subcellular localizations, such as chloroplasts, mitochondria, peroxisomes, the cytosol, and the Golgi apparatus. About 17% of all control proteins localized to chloroplasts, 26% in cytoplasm, 7% in mitochondria, and 5% in peroxisomes ( Figure 3). In addition, 38% of the proteins were uncharacterized or unknown. We noticed several organelle marker enzymes such as glycosyltransferases (Golgi marker), catalase (peroxisome marker), and cytochrome oxidase (mitochondrion marker) among salt gland-secreted proteins.
To deeply explore the potential functions and differences of proteins and peptides secreted by salt glands under salt stress and control conditions, we took a closer look at the possible functions of proteins in secreted sap. For control leaf discs, 40 of the identified proteins were involved in photosynthesis. Of these, 36 proteins were not present in the sap of NaCl-treated salt glands, with three Rubisco large chains and one Rubisco small chain subunit remaining (Table S1). In addition, we detected 26 proteins involved in respiration in the sap of control leaf discs, of which only four (two ATP synthase alpha subunits, one ATP synthase beta subunit, and one xanthine dehydrogenase) remained in the sap of NaCl-treated salt glands (Table S2). To deeply explore the potential functions and differences of proteins and peptides secreted by salt glands under salt stress and control conditions, we took a closer look at the possible functions of proteins in secreted sap. For control leaf discs, 40 of the identified proteins were involved in photosynthesis. Of these, 36 proteins were not present in the sap of NaCl-treated salt glands, with three Rubisco large chains and one Rubisco small chain subunit remaining (Table S1). In addition, we detected 26 proteins involved in respiration in the sap of control leaf discs, of which only four (two ATP synthase alpha subunits, one ATP synthase beta subunit, and one xanthine dehydrogenase) remained in the sap of NaCl-treated salt glands (Table S2).
After bioinformatics analysis, proteins related to vesicle transport were mainly involved in cytoskeleton-associated proteins, protein kinases, and proteins with functional domains (Table 1). Proteins related to photosynthesis and energy metabolism accounted for 30.8% of the number of total proteins secreted in the control group, while vesicle transport-related proteins accounted for 3.7% of the total number of proteins secreted in the same samples. Proteins related to photosynthesis and energy metabolism accounted for 50% of the total number of proteins secreted by leaf discs exposed to salt stress. Notably, enzymes related to reactive oxygen species (ROS) scavenging, such as peroxidase, superoxide dismutase, and catalase, were no longer detected in the sap from salt-treated salt glands.  After bioinformatics analysis, proteins related to vesicle transport were mainly involved in cytoskeleton-associated proteins, protein kinases, and proteins with functional domains (Table 1). Proteins related to photosynthesis and energy metabolism accounted for 30.8% of the number of total proteins secreted in the control group, while vesicle transportrelated proteins accounted for 3.7% of the total number of proteins secreted in the same samples. Proteins related to photosynthesis and energy metabolism accounted for 50% of the total number of proteins secreted by leaf discs exposed to salt stress. Notably, enzymes related to reactive oxygen species (ROS) scavenging, such as peroxidase, superoxide dismutase, and catalase, were no longer detected in the sap from salt-treated salt glands. We also identified various peptides in the secreted sap of salt glands, which are both degradation products and peptides with biological functions. Thirty-eight peptides with potential functions were randomly chosen for annotation (Table 2). These peptides were mainly derived from photosynthesis-related enzyme chains, elongation factor fragments, ATP synthase subunits, and special domain fragments, which play an important role in plant growth and development, apoptosis, signal transduction, metabolism, and response to stress (Table 2). Elongation factor 1-alpha, phosphoglycerate kinase (fragment), ATP synthase (alpha and beta subunits), and phosphoglycerate kinase were related to the salt stress response. Similar to the protein results, the number and type of peptides decreased significantly under 200 mM NaCl treatment. However, the Ribulose bisphosphate carboxylase large chain, as well as the alpha and beta subunits of ATP synthase, were still present in sap after salt treatment.

Discussion
L. bicolor has special salt-secreting structures on its leaf surface called salt glands. In this study, we showed that the liquid secreted from these salt glands contained a variety of proteins and polypeptides, whose number and variety sharply declined under salt stress compared to control conditions. We believe that the presence of peptides of secretion sap under control conditions may be due to a sudden enhancement of vesicular transport carrying a large number of immature protein fragments. However, under 2 weeks of salt condition, the L. bicolor could adapt the activity of salt glands to secrete Na + and Cl − . We conclude that the decrease in the number of proteins and peptides is related to the adaptation of the salt gland secretion salt of the halophytes L. bicolor under salt treatment, and the remaining proteins and peptides after 2 weeks of salt treatment may be involved in the normal salt secretion under salt stress. According to our result, these proteins and polypep-tides participate in vesicular transport, energy production, and material metabolism. In order to better understand the role of proteins and peptides in salt-gland secretion, we tried to discuss the functions of vesicular transport-related proteins, energy metabolizing proteins, cargo proteins, stress-related proteins, and their roles in salt-gland secretion. In particular, we identified Golgi marker enzymes, such as glycosyltransferases and vesicle transport cargo proteins, among the secreted proteins. These results indicate that vesicle transport plays an important role in salt gland secretion in the recretohalophyte L. bicolor and that many salt secretion-unrelated metabolic processes, such as ATP production, are put on hold or reduced under high-salinity conditions.
As a conventional membrane transport pathway, the steps of vesicular transport (budding, transport, tethering, docking, and fusion) engage several organelles and require the participation of multiple protein complexes [49,50]. The TGN coordinates cell organs by receiving, secreting, and directing vesicles to their intended destinations [51,52]. Salt secretion from salt glands was recently linked to vesicle transport. For instance, Smaoui et al.
(2011) undertook a lanthanum tracing study in the salt glands of the recretohalophyte saltbush (Atriplex), which revealed that lanthanum is deposited in endocytic vesicles at the nuclear membrane, in the rough endoplasmic reticulum, and the Golgi lumen [53]. After treatment of L. bicolor leaf discs with the vesicle transport inhibitor BFA, the Golgi apparatus was reported to be deformed, the cis-Golgi apparatus disappeared, so-called BFA compartments formed in the cytoplasm, and the secretion capacity of leaf discs dropped sharply. A threshold concentration of BFA resulted in almost no fluid secretion by leaf discs [23]. BFA is a fungal-derived macrolide that specifically inhibits Golgi-mediated secretion and is a powerful tool for dissecting the plant secretion and endophagocytic pathway [54][55][56]. These studies indicate that the Golgi apparatus and TGN play a key role in salt gland secretion. Hamaji et al. (2009) found that vesicles were more active in Arabidopsis (Arabidopsis thaliana) under salt treatment, with salt ions not only accumulating in vacuoles but also in these active vesicles [57]. This report also provided evidence for the participation of vesicle transport in salt ion movements [57]. Our experimental results showed for the first time that the secretions of salt glands contain proteins and peptides, especially the Golgi marker protein glycosyltransferase, as well as enzymes associated with energy metabolism (Table 1 and Table S2).
Glycosyltransferases are usually located at the Golgi membrane or lumen [58]. The presence of glycosyltransferases and other proteins in the secreted liquid may reflect a lack of selectivity in vesicle transport. The selectivity of ion channels and transporters depends on the structural properties of ion channel and transporter proteins; however, the vesicle transport depends on the local solute concentration near the site of vesicle formation and the charge of the vesicle membrane and is not selective [22,59]. In addition, we also detected many cargo proteins in the secreted liquid transported by vesicles, such as chitinases and xyloglucan endotransglucosylases/hydrolases (XTHs). This observation also suggests that salt glands may be involved in a rich vesicle transport.
We identified many energy-related proteins in the secretions, which supports the notion that vesicle transport activities of salt secretion need a lot of energy. Vesicle transport mainly assists in the transport and exchange of substances (such as proteins and lipids) between the organellar membranes [60]. Here, we found that the liquid from salt gland secretions also contained proteins related to photosynthesis and energy metabolism (30.8% of all proteins), ROS metabolism, protein synthesis, and some uncharacterized proteins (21.5%). Most of these proteins were detected in the form of polypeptides and originated from various organelles. The secreted proteins included Ribulose bisphosphate carboxylase and ATPase (Tables S1 and S2), indicating that large amounts of energy and an H + gradient are required to sustain salt secretion from salt glands, with abundant photosynthesis producing multiple metabolic substrates. Many reports have also shown that the many mitochondria and Golgi in the salt glands provide the necessary energy for salt secretion [1,61].
Our results showed that the control leaf discs of L. bicolor plants secreted many proteins after floating on 200 mM NaCl solution for 24 h, which was profoundly different from those of the plants watered with 200 mM NaCl solution for 2 weeks. After 2 weeks of treatment with 200 mM NaCl, the local ion concentration is known to increase near the vesicles, as does the speed of vesicle movement [1,57]. After 2 weeks of salt treatment, plants adapted to salt stress; the vesicles perform their normal transport function within the cell while also secreting ions ( Figure S2). Compared to salt secretion by salt glands under normal conditions, the secreted liquid after salt treatment contained 16 proteins, which were divided into three main categories: proteins related to photosynthesis and respiration (50%), protein synthesis, and uncharacterized proteins (18.8%). Moreover, the proportion of photosynthesis-and energy metabolism-related proteins increased significantly in saltstressed samples compared to control samples, indicating that the increased salt secretion exhibited by salt-stressed leaf discs may require more energy expenditure than control conditions. This higher energy quota may be the result of normal vesicle transport after adaptation to salt stress.
In addition, we identified several proteins typically associated with stress responses in the control samples, such as ROS-scavenging enzymes, chitinases, and XTHs. Notably, we did not detect these proteins upon salt treatment. We believe that ROS-scavenging enzymes may be related to plant response to salt stress and resistance to oxidative stress in the leaf cytoplasm, so it is kept inside the plant cells instead of being secreted [62]. XTHs also respond to biotic/abiotic stresses. Forty-two SeXTH genes were upregulated in the succulent halophyte glasswort (Salicornia europaea L.) under salt and drought stress [63,64]. We also identified elongation factor 1-alpha (EF1A), an indicator of cell growth and development, which regulates protein synthesis by altering the expression of eEF1A to rapidly respond to external environmental stress and cellular signals [65]. We detected several uncharacterized proteins in samples from both control and salt-treated leaf discs; more attention should therefore be paid to the functional dissection of these uncharacterized proteins secreted by plant salt glands, as they may contribute to the regulation of salt secretion or salt gland formation. Many known polypeptides contribute to plant responses to salt stress. The Arabidopsis homolog to tomato CAP-derived peptide 1 (CAPE1) and Cysteine-rich transmembrane module3 (CYSTM3) raises salt sensitivity and negatively regulates salt tolerance by inhibiting Na + efflux in Arabidopsis roots [66,67].
In addition, we detected fragments from the large subunit of Ribulose bisphosphate carboxylase, as well as the alpha and beta subunits of ATP synthase, under both control and salt treatment, hinting at their role in maintaining active salt secretion and providing energy under prolonged salt stress. These results also suggest that plants respond quickly to salt stress through functional peptides when suddenly exposed to salt stress. Since vesicle transport is determined by the local ion concentration near the vesicle and the charge of the vesicle membrane, the complement of secreted proteins appeared to be poorly reproducible ( Figure 2). We detected the Golgi marker enzyme glycosyltransferase in the secretion sap, which confirmed that salt ions were excreted in vivo through the classical secretion pathway, with the involvement of TGN-mediated vesicle transport in salt glands. Lu et al. (2020) showed that the SYNTAXIN OF PLANTS61 (SYP61) protein located at the TGN plays an important role in salt secretions from salt glands via virus-induced gene silencing of its encoding gene [23]. Our results provide support for the endocytosis hypothesis in the secretion mechanism by salt glands [5,23,28]. However, which proteins participate in the secretion of salt ions, as well as their mode of action, needs to be studied in more detail. On the other hand, under salt treatment after two weeks, the Na + and Cl − concentration in secretion sap significantly increased, compared with the amount of protein and peptides (Figures 2 and S3). We speculated that the main function of the salt gland is secreting ions, and this kind of ion and protein secretion balance is how to control under salt conditions. Whether there are proteins that regulate salt gland secretion, we do not yet know.

Plant Materials and Culture Conditions
L. bicolor seeds were collected from the Yellow River Delta, Dongying, China (37 • 20 N; 118 • 36 E) in 2020 and stored in a refrigerator at 4 • C for 6 months before use. Healthy, uniform seeds were selected and planted in a Petri plate with 30 mL of Murashige and Skoog (MS) medium. After germinating for 7 days (22 ± 3 • C), seedlings were transplanted into well-mixed soil (soil: vermiculite: perlite, 3:1:1, w/w/w) and grown in a greenhouse (28 ± 3 • C/23 ± 3 • C, day/night) at a light intensity of 600 mol/m 2 /s (with a 16-h-light/8-hdark photoperiod) and 70% relative humidity. When the sixth true leaf of the seedling grew, uniform seedlings were used for further experiments. The control plants were irrigated with Hoagland's nutrient solution containing 5 mM KNO 3

Salt Gland Observations
Salt glands were observed using the method described by Li et al. (2020) [68]. Leaves were observed under a scanning electron microscope according to the method of Shah et al. (2018) [69]. The true leaves of L. bicolor seedlings were placed in enzymolysis solution and shaken at 240 g for 2 h in the dark at 27 • C. Then, 1% (w/v) diselase, 0.1% (w/v) pectolyase Y-23, and 1% (w/v) cellulase Y-C were dissolved in PH 5.7 buffer (20 mM MES, 20 mM KCl, 0.4 M mannitol) to make enzymolysis solution. The leaves were transferred to a homogenizer containing phosphate-buffered saline of pH 7.4 for grinding. The homogenate was filtered through a 70-µm cell screen and stored on ice. The salt glands were observed under UV excitation (330-380 nm) using differential interference contrast microscopy (DIC). In addition, we calculated the ions concentration of secretion sap under 200 mM NaCl treated and control conditions using the method of Lu et al. (2020) [23]. Moreover, the relative activity of cells was measured using the method of Gaff et al. (1971) [70].

Extraction of Proteins from the Secretion Liquid of Salt Glands
Uniform L. bicolor leaves (usually the sixth true leaf) were used to collect the secretion liquid, as described by Lu et al. (2020) [23] ( Figure S1). Leaf discs without a main vasculature were collected with a hole puncher (diameter 1 cm), and each leaf disc was placed in 200 mM NaCl solution for 24 h in a growth chamber, with the abaxial surface facing up. Mineral oil was layered onto the leaf discs immediately after floating on the NaCl solution. After 24 h, many droplets formed on the leaf disc surface and were collected with a micropipette. The collected secretion liquid was centrifuged at 4000× g for 10 min at room temperature. The lower aqueous phase was used to extract the proteins by trichloroacetic acid (TCA)-acetone precipitation after removing the upper mineral oil phase. The aqueous phase was divided into 10-mL centrifuge tubes, and three volumes of pre-cooled TCAacetone solution (20%, w/v) were added. After fully mixing, the samples were incubated at −20 • C for 2 h to precipitate proteins, which were collected using centrifugation at 10,000× g at 4 • C for 10 min. The pellets were then rinsed twice with pre-cooled acetone and air-dried. The pellet was then resuspended in 2× loading buffer before SDS-PAGE. Protein separation was carried out according to Laemmli. (1970) [71]. A PAGE gel quick preparation kit (10%) was used to prepare the gels (Yeasen, Shanghai, China). The separation gel consisted of 4 mL of 10% separation gel buffer, 4 mL of 10% separation gel solution, and 70 µL of 10% modified ammonium persulfate solution. The stacking gel comprised 1 mL of 10% color-concentrated gel buffer, 1 mL of 10% concentrated gel solution, and 18 µL of 10% modified ammonium persulfate solution. A protein silver stain plus kit was used to stain proteins (SK6020, Coolaber, Beijing, China).

Protein Identification, Annotation, and Characterization
The protein analysis was performed with liquid chromatography-mass spectrometry (Q Exactive HF-X, Thermo Fisher, Waltham, MA, USA). The samples were separated using a nano liquid chromatographic system with a flow rate of 5-300 nL/min (UltiMate 3000 RSL, Thermo Fisher, USA). The protein samples were dissolved in loading buffer, injected with an automatic sampler, bound to a C18 capture column (3 µm, 120 A, 100 µm × 20 mm), and then eluted into an analysis column (2 µm, 120 A, 750 µm × 250 mm) for separation. An analytical gradient was established using two mobile phases ( ACN]). The flow rate of the liquid phase was set to 300 nL/min. The MS analysis was performed in DDA mode, with each scan cycle consisting of a full MS scan (R = 60 K, AGC = 3 × 10 6 , Max IT = 20 ms, scan range = 350-1800 m/z), followed by 20 MS/MS scans (R = 15 K, AGC = 2 × 10 5 , max IT = 100 ms).

Data Analysis
The mass spectrometry data were retrieved by MaxQuant (V1.6.2.10) using the MaxLFQ retrieval algorithm [72]. The retrieval database was the Proteome Reference database of Plumbaginaceae in UniProt. Protein fragments or subunits with molecular weight less than 10,000 Da were selected for collation and functional annotation.