Proteome Analysis of the Wild and YX-1 Male Sterile Mutant Anthers of Wolfberry (Lycium barbarum L.)

Pollen development is disturbed in the early tetrad stage of the YX-1 male sterile mutant of wolfberry (Lycium barbarum L.). The present study aimed to identify differentially expressed anther proteins and to reveal their possible roles in pollen development and male sterility. To address this question, the proteomes of the wild-type (WT) and YX-1 mutant were compared. Approximately 1760 protein spots on two-dimensional differential gel electrophoresis (2D-DIGE) gels were detected. A number of proteins whose accumulation levels were altered in YX-1 compared with WT were identified by mass spectrometry and the NCBInr and Viridiplantae EST databases. Proteins down-regulated in YX-1 anthers include ascorbate peroxidase (APX), putative glutamine synthetase (GS), ATP synthase subunits, chalcone synthase (CHS), CHS-like, putative callose synthase catalytic subunit, cysteine protease, 5B protein, enoyl-ACP reductase, 14-3-3 protein and basic transcription factor 3 (BTF3). Meanwhile, activities of APX and GS, RNA expression levels of apx and atp synthase beta subunit were low in YX-1 anthers which correlated with the expression of male sterility. In addition, several carbohydrate metabolism-related and photosynthesis-related enzymes were also present at lower levels in the mutant anthers. In contrast, 26S proteasome regulatory subunits, cysteine protease inhibitor, putative S-phase Kinase association Protein 1(SKP1), and aspartic protease, were expressed at higher levels in YX-1 anthers relative to WT anthers. Regulation of wolfberry pollen development involves a complex network of differentially expressed genes. The present study lays the foundation for future investigations of gene function linked with wolfberry pollen development and male sterility.


Introduction
In flowering plants, male reproductive processes take place in the stamen, a part of the anther, which contains diploid sporogenous cells that experience meiosis to form haploid microspores and finally develop into pollen grains or the male gametophyte [1][2][3][4][5]. Detailed analysis of anther development has shown that cell differentiation occurs in a precise chronological order, with distinct stages, which can be related to bud size [6]. Once this process becomes disordered, pollen production might be aborted, resulting in male sterility. Many genes controlling stamen and pollen development have been identified, and their specific roles characterized. These functions appear to be conserved in higher plants [7][8][9][10][11].
In recent years there has been an increasing application of proteomic approaches to study anther development and pollen reproduction, such as in Arabidopsis [12][13][14][15], rice [16][17][18][19], tomato [2,20], and Brassica napus [21][22][23]. By applying proteomic analyses, many proteins specifically expressed in anthers with roles in pollen development [24], tapetum degradation [25], programmed cell death (PCD) [26], and callose hydrolyzation [27] were identified. Proteins involved in metabolic process, stress resistance [16,28], and several transcription and translational regulating factors [29][30][31] were also characterized. It is difficult to find male sterile related genes affecting pollen development directly; therefore, methods for globally detecting different expression patterns between the male sterile mutant and the WT must be applied. The exploitation of plant male sterility is enhanced by combining transcriptome with proteome analysis of developing pollen [32]. In tomato, proteomic analysis of malesterile 7B-1mutant anthers at the stage when tetrads formed revealed that proteasome and 5B protein, with potential roles in tapetum degeneration, were down-regulated. Cystatin, regulator of endogenous proteolytic activities during seed maturation and germination and in PCD, were up-regulated and correlated with the male sterility [2]. In addition, proteins associated with carbohydrate and energy metabolism, photosynthesis and flavonoid synthesis were also down-regulated in CMS anthers of Brassica napus, all of which might have roles in pollen development [23]. Moreover, several proteins correlated with male sterility were identified in rice, with roles in protein synthesis, signal transduction, cell death and carbohydrate metabolism. The occurrence of male sterile mutants is usually explained by the lack of certain proteins/enzymes involved in pollen development. In the gametophytic male-sterile mutant gaMS-2 of maize, reduced Zea m1 level is associated with the sterility [33]. GS was inactivated in tobacco anthers and microspores by a dominant-negative mutant approach, resulting in male sterility [34]. The conditional male sterile mutant of Arabidopsis was associated with the FLP1 protein, likely playing a role in the synthesis of the components of tryphine, sporopollenin of exine and the wax of stems and siliques. [35]. In addition, several housekeeping proteins with potential roles in microspore development also showed altered abundance. All of these identified proteins have important functions in pollen development in higher plants.
Lycium barbarum L, a woody bush, is a famous traditional Chinese herbal medicine that nourishes the kidneys and liver, brightens eyes, reduces blood glucose and serum lipids, and has anti-aging, immunomodulating, anticancer, anti-fatigue, and male fertility-facilitating properties [36,37]. It has been widely used as a health food for 2300 years, with its fruit being used to produce various types of health products and foods, such as medicinal beverages and dietary soups [38,39]. However, it is difficult to meet the need for substantial improvement of existing varieties of wolfberry by natural selection.
Plant male sterility is important for developmental and molecular studies and in hybrid seed programs [40]; therefore, many male sterile mutants in higher plants have been characterized. To the best of our knowledge, few studies on male sterility of wolfberry have been reported. The spontaneous male sterile mutant YX-1 produces flowers with shrunken stamens, in which pollen production is aborted in early tetrad stage. YX-1 mutants also show reduced stamen filament length relative to the WT, and appear beneath the receptive stigma at flower opening [41]. As the male sterile varieties are valuable resources that greatly facilitate the production of hybrids via cross-pollination, in-depth study of anther and pollen development, YX-1 male sterile mutant has important application value in hybrid breeding programs and important theoretical significance.
In this present study, we report the proteomes of the anthers of the WT and YX-1 wolfberry and identify differentially expressed proteins (in terms of protein spot volume). We discuss their possible biological roles and their potential effects on anther development and pollen fertility, with the aim of understanding the molecular mechanism of the biological process at the proteomic level. The protein profiles of anthers at early tetrad stage, when tetrad and tapetum show normal development in WT anthers, but tapetum and tetrads degeneration are observed in YX-1, were compared by 2D-DIGE, and interesting proteins were identified by mass spectrometry. More than 1760 spots were observed on DIGE gels. Compared with the WT, in YX-1 mutant anthers, proteins related to energy, carbohydrate and amino acid metabolism, pollen development, stress response, and signaling were down-regulated. In contrast, proteins involved in proteolysis were up-regulated. The possible functions of these proteins in pollen development and in male sterility are discussed. The results will provide important information for further studies on the molecular mechanism of male sterility in wolfberry.

Plant Growth and Anther Collection
The YX-1 male sterile mutant and the WT were grown in the wolfberry orchard of Yinchuan Yuxin Wolfberry seed industry Co., Ltd., located in Yinchuan, Ningxia Municipality, China. Plants were watered regularly and fertilized weekly with a commercial fertilizer (40% Compound Fertilizer, Nantong Hailing Fertilizer Co., Ltd., Nantong, China).
Flower buds of about 4.8-5.0 mm(WT)and 5.0-5.8 mm(YX-1)length (Fig. 1A, corresponding the early stage of tetrad) were picked using a combination of the bud size and anther cytological investigation, as previously reported [42], dissected under a microscope. Stamens from buds at the same stage,when tetrads and tapetum cell were degrading in YX-1 anthers but showed normal development in WT anthers(confirmed by anther squashes, Fig. 1D)were isolated for 2D-DIGE analysis. Stamen samples were either used immediately or frozen in liquid nitrogen and stored at 280uC until further analysis.

Protein Sample Preparation
Anther protein extractions were performed using a trichloroacetic acid (TCA)-acetone protocol [23] with some modifications. The samples of fresh or frozen wolfberry anthers were finely powdered in liquid nitrogen and homogenized with chilled acetone/10% TCA for 30 min, and then precipitated overnight at 220uC. Precipitated proteins were centrifuged at 120006g for 45 min at 4uC. After three washes with acetone, the pellets were vacuum dried and proteins extracted with buffer containing 7 mol?L -1 urea, 2 mol?L -1 thiourea, 150 mmol?L -1 Tris-HCl (pH8.5), 4% CHAPS, and 1 mmol?L -1 PMSF, by vortexing for 1 h at room temperature, and centrifugation at 12 000 g for 45 min at 4uC. The supernatants were collected and the protein content was determined according to the Bradford method, using the Bio-Rad protein assay reagent (Bio-Rad, Hercules, CA, USA), and samples were stored at 280uC until 2-DE.

Fluorescence Labeling Protein with CyDyes
Protein labeling with fluorescent cyanine dyes was performed according to the manufacturer's instructions (GE Amersham, Fairfield, CT, USA). Individual samples from three groups (pooled internal standard, YX-1, and the WT) were labeled with Cy2, Cy3, and Cy5, respectively. The three dyes were designed to ensure that proteins common to each sample have the same relative mobility regardless of the dye used to tag them. CyDyes were reconstituted in anhydrous DMF and combined with samples at a ratio of 400 pmol of CyDye to 50 mg of protein. Labeling was performed on ice and in the dark for 30 min. The reaction was then quenched by incubating with 1.5 mL of 10 mM lysine on ice in the dark for 10 min.

DIGE Analysis
Relative protein quantification across male sterile mutant YX-1 and WT samples was performed using DeCyder 2-D Differential Analysis Software (v 6.05.11, GE Amersham). The Cy2-labeled pooled internal standard on every gel allowed accurate relative quantitation of protein spot features across different gels. The spots that were present on at least two gels based on the image analysis were considered as expressed protein spots. Student's t-test (p,0.05) and one-way ANOVA were used to calculate significant differences in relative abundances of protein spot-features in the male sterile anther compared with the WT anthers. Spots with reproducible and significant variations, at least 1.5-fold up-or down-regulated, were considered differentially expressed proteins.

In-gel Digestion
Protein spots were cut from gels, destained for 20 min in 30 mM potassium ferricyanide/100 mM sodium thiosulfate (1:1 v/v), washed in Milli-Q water until the gels were colorless, and then lyophilized. Each spot was digested in 5 mL 10 ng/mL trypsin (sequencing-grade reagent, Promega, Fitchburg, WI, USA) at 37uC overnight. The peptides were extracted three times with 60% ACN/0.1% TFA. The extracts were dried completely by centrifugal lyophilization. The resulting tryptic digests were concentrated and desalted using C18 ZipTips (Millipore Corporation, Bedford, MA, USA) according to the manufacturer's protocol.

Mass Spectrometry and Data Analysis
Samples were mixed (1:1 v/v) with 5 mg/ml HCCA matrix and analyzed by a 4800 Plus MALDI-TOF/TOF TM Analyzer (Applied Biosystems, Carlsbad, CA, USA). Calibration for MS/ MS mode was carried out using fragment ion masses from Glufibrinopeptide. Initial analysis of protein digests on the 4800 instrument was carried out by MS in positive ion reflectron mode, using trypsin autolysis products (m/z 842.510, 1045.564 and/or 2211.105) for internal mass calibration where possible, or else the default calibration. Parent mass peaks with a mass range of 800-4000 Da and minimum S/N 50 were picked out for tandem TOF/TOF-MS/MS analysis. The 10 most abundant precursor ions in each spectrum (excluding trypsin autolysis fragments) were subsequently selected by Protein Pilot TM v4.0 software for MS/ MS analysis, with collision induced dissociation (CID) closed, increased laser fluence, and 2500 laser shots collected per sample. The UV laser was operated at a 25 Hz repetition rate with a wavelength of 355 nm and the accelerated voltage was 2 kV.
Protein identifications were conducted by combining search (MS plus MS/MS) to the entries of a non-redundant protein and/ or EST-viridiplantae database downloaded from the National Center for Biotechnology Information using MASCOT open source (http://www.matrixscience.com) in the NCBI non-redundant database. Search parameters were the enzyme trypsin, taxonomy restrictions to Viridiplantae, 650 ppm peptide mass tolerance in MS, 60.2 Da for MS/MS data, one missed cleavage, carbamidomethyl (C) as a fixed modification and methionine oxidation as a variable modification. The confidence in the peptide mass fingerprinting matches (p,0.05) was based on the MOWSE score and confirmed by the accurate overlapping of the matched peptides with the major peaks of the mass spectrum. Only significant hits, as defined by the MASCOT probability analysis (p,0.05), were accepted. The functional classification of the identified ESTs was performed following a BLAST search, and the protein sequences of matching ESTs were then searched against the NCBInr protein database. Only BLAST matches with E values #10 230 were selected [19].

Quantitative Real-time RT-PCR Analysis
Quantitative real-time RT-PCR was used to assay gene expression levels for apx and atp synthase beta subunit. Primers for quantitative real-time RT-PCR analysis are shown in Table 1. Wolfberry actin (Accession number HQ415754.1) was used for RNA normalization. All wolfberry RNA samples were diluted to 200 ng ml -1 . The SuperScript TM III platinumH two-step qRT-PCR kit with SYBRH Green (Invitrogen, Carlsbad, CA, USA) was used for detecting the expression levels of the genes. Quantitative real-time RT-PCR was carried out in a final volume of 20 ml containing 10 ml PlatinumH SYBRH Green qPCR SuperMix-UDG, 10 mM forward and reverse primers in a 7500 Real time PCR System (Applied Biosystems). The relative expression levels of all the samples were analyzed according to recommendations in the User Bulletin for the 7500 Real time PCR System. All reactions were performed in three biological replicates. The threshold cycles (Ct value) of the target genes and actin in different samples were obtained by quantitative real-time RT-PCR.

Enzyme Activity Assay
To monitor the activity of APX and GS, protein was extracted with 100 mmol L 21 sodium phosphate buffer (pH 7.0) containing 5 mmol L 21 ascorbate and 1 mmol L 21 EDTA, and 10 mmol L 21 Tris-HCl buffer (pH 7.6) containing 1 mmolL 21 MgCl 2 , 1 mmol L 21 EDTA and 1 mmol L 21 b-mercaptoethanol ), respectively from WT and mutant anthers. APX activity was determined by the decrease of absorbance at 290 nm (extinction coefficient 2.8 mM cm 21 ) as described by Chen and Asada [43].The reaction mixture was composed of 50 mM potassium phosphate buffer (pH 7.0), 0.5 mM ascorbate, 0.2 mM H 2 O 2 and the appropriate volume of protein extract.
Determination of GS activity was performed by the method of Oaks et al [44]. The reaction mixture contained 80 mmol L 21 glutamate, 1 mmol L 21 hydroxylamine, 8 mmol L 21 ATP, 0.2 mol L 21 N-tris-(hydroxymethyl)methyl glycine (Tricine) (pH 7.8), 4 mmol L 21 MgSO 4 and 0.2 mmol L 21 EDTA and the appropriate volume of protein extract, and the absorbance of the hydroxamate derivative of glutamic acid measured at 540 nm. One unit of GS activity was defined as 1 mmol L-glutamate cmonohydroxamate formed per min. Protein in enzyme extract was estimated using Bradford method [45]. Samples were performed in three biological replicates.

Anther Development
Under the growth conditions as mentioned in ''Plant growth and anther collection'', buds at the early tetrad stage from YX-1 mutant were longer (approximate 5.0-5.8 mm length) than that from the WT (approximate 4.8-5.0 mm length, Fig. 1A); and the stamens of the YX-1 appeared less slender compared with those of the WT (Fig. 1B ). Also, YX-1 mutants showed reduced stamen filament length relative to the WT, and appear beneath the receptive stigma at flower opening (Fig. 1C). Corresponding to the stage and bud lengths outlined in Fig. 1A the tapetal cells were intact, with a dense cytoplasm containing a few small vacuoles, both tetrads and tapetum showed normal development in WT anthers (Fig. 1D). In contrast, in YX-1 anthers (Fig. 1E), tetrads and tapetal cells had started to degenerate, and numerous small vacuoles were present in the tapetum,also small vacuoles appeared in tetraspores. Anther squashes at an earlier stage (buds of 3.0-5.0 length mm) in both WT and YX-1 showed tetrads enveloped within a thick callose wall, a rich cytoplasm in the tetraspores and tapetal cells radially enlarged with a dense cytoplasm. At later stages (buds of 6.0-7.0 mm length), microspores were observed in WT anthers, but in YX-1 anthers, tetrads and the tapetal cells had completely disintegrated and disappeared, resulting in an empty anther chamber (not shown). This confirmed our earlier observation that in the YX-1 mutant, pollen development breaks down at the early tetrad stage [42].

2D-DIGE Analysis of Anther Proteomes
Soluble proteins of wolfberry anthers from YX-1 (approximate 5.0-5.8 mm length) and the WT buds (approximate 4.8-5.0 mm length) were extracted. After preparative experiments, we selected pH 3-11 strips for proteome analysis. Due to the limitations of conventional 2-DE for reproducibility and sensitivity, we used the DIGE technology to study differentially expressed proteins in WT and YX-1 anthers. Equal amounts (50 mg) of protein samples of WT and YX-1 anthers were labeled with Cy2 (internal standard), Cy3, or Cy5 dyes. An overlay of the Cy3 and Cy5 images from the 2D-DIGE gels is shown in Fig. 2. The protein expression patterns of YX-1 mutant were generally similar to those of WT anthers, and more than 1760 spots were observed by DIGE methodology.
From the result of 2-DE image analysis, we found a number of spots with lower or higher protein abundances (measured as the relative spot volume) in YX-1 anthers compared with those in the WT. A threshold limit of 1.5-fold was set in this study as previously reported [2] and three replicates were performed to reduce the number of potential false positives because of the sensitivity and reproducibility of DIGE technology. Fig. 2 shows a representative DIGE image of WT and YX-1 anther protein extracts labeled with Cy3 and Cy5 and separated with IPG 3-11 strips and the numbered spots used for mass spectrometry analysis. Some of these differential spots are shown in the enlarged portion of gels in Fig. 3. Fifty-two spots showed at least a 1.5-fold change in protein abundance (p,0.05), among which, 13 showed an increase in YX-1 mutant anthers and 39 showed decreased abundance compared with their levels in WT anthers.
As can be seen from Table 2, between the WT and YX-1 anthers, three differently expressed proteins were identified as mitochondrial ATP synthase related subunits (spots No. 10, 11 and 12), which are involved in energy metabolism. Three proteins correspond to APX (spots No. 26, 27, and 28), which is involved in stress response processes. GS (spot No. 15, Table 2) was one of the proteins that showed a major expression difference ( Fig. 2 and 3) between the WT and YX-1 anthers in DIGE gel images. Therefore, we selected these three enzymes to further confirm changes using enzyme activities or mRNA expression.

Expression of apx and atp Synthase Beta Subunit mRNA
To further understand the differential expressions of APX and component of the mitochondrial ATP synthase complex at the level of gene expression between the WT and YX-1 anthers, quantitative real-time RT-PCR was used to analyze their expression levels at the stage of the early tetrad development (the same stage as the proteome analysis) between the male sterile anthers and the WT. It should be noted that these two genes had been isolated by 59-and 39-race methods in our previous experiments [46]. Fig. 5 shows the apx and atp synthase beta subunit mRNA expression levels in the wolfberry anthers. At the early tetrad stage of anther development, the expression level of apx and atp synthase beta subunit mRNA in YX-1 anthers was only 11% and 31% to that in WT anthers, respectively. The results clearly indicated that expression of apx and atp synthase beta subunit mRNA is significantly reduced during the process of anthers abortion.
Enzyme Activity GS (spot No. 14) showed maximum difference in expression between the WT and the mutant in DIGE gel images (Fig. 2); in addition, three spots (spot No. 26, 27 and 28) were identified to be APX, which displayed low abundance in mutant anthers. Therefore, we analyzed the activity of these two enzymes in WT and YX-1 anthers. The activity of GS was nearly 5 times, and that of APX approximately 3 times, higher in WT compared with that in YX-1 anthers (Fig. 6A and 6B), which correlated with the 2-D DIGE results.

Discussion
There has been little detailed proteomic characterization of the male sterile wolfberry. Therefore, we conducted a comprehensive proteomics analysis between mutant YX-1 and WT anthers to gain an understanding of the mechanisms of wolfberry male sterility. The potential roles of some differentially expressed proteins in anther development and pollen fertility are discussed below.
The tapetum is universally present in higher plant anthers. In addition to its main function of supplying nutrition for meiocytes/ spores, the tapetum has other roles, such as the production of the locular fluid, callase, pollenkitt/tryphine, sporophytic proteins and enzymes, and the formation of exine precursors, all of which are required for the normal development of microspores to pollen  grains [4]. The tapetum undergoes cellular degradation during the late stage of anther development, which is considered a programmed cell death (PCD) event. Furthermore, tapetal cell disintegration accords well with the post-meiotic anther development processes. A premature or delayed degradation of the tapetum can result in male sterility [47]. In normal anthers of wolfberry, both tetrads and tapetum show normal development, with a dense cytoplasm and only a few small vacuoles present at the early tetrad stage (the stage used in the present study, Fig. 1D). In contrast, in the YX-1 mutant anthers at the same stage, the tapetum begins to degenerate, coupled to numerous small vacuoles which also appear in tetraspore (Fig. 1E), which suggests that Mascot protein score for ions complemented by the percentage of the confidence index (C.I.). f Sequence Coverage. g Spot volume ratios are the average for each spot from three replicate gels. h Protein spots with a significant change in abundance (1.5-fold or above) between the WT and YX-1, and a P-value of#0.05 were considered statistically significant. *Spots from EST database. Unidentified differently spots not listed. doi:10.1371/journal.pone.0041861.t002 premature tapetum degeneration in YX-1 could contribute to male sterility.
A number of enzymes, belonging to the carbohydrate and energy metabolism group, are reduced in YX-1 anthers, relative to the WT. These down-regulated enzymes include mitochondrial ATP synthase subunits, fructokinase-like protein, MDH, aldolase, and GAPDH ( Fig. 2 and Table 2). In CMS lines of rice and B. napus, some of the enzymes involved in energy and carbohydrate metabolism are also down-regulated [16,23]. In particular, three protein spots were identified to be mitochondrial ATP synthase D chain (spot No.10, Fig. 2 and Table 2), putative ATP synthase subunit E (spot No.11, Fig. 2, 3 and Table 2), and mitochondrial ATP synthase beta subunit (spot No.12, Fig. 2 and Table 2). Cytoplasmic male sterility in plants is generally the consequence of dysfunction of mitochondria in the pollen, and several mitochondrion DNA regions encoding F0F1-ATPase (i.e. ATP synthase) subunits have been identified associated with CMS [48]. Plant mitochondrial genomes contain approximately 60 open reading frames (ORFs), among which, orf25 is implicated playing a role in the CMS of T cytoplasm maize [49], and orfB is involved in CMS in several plant species [50][51][52]. An accepted hypothesis on the mechanism of CMS is that the increased demand for respiratory function and cellular energy in the form of ATP during anther development may be compromised by expression of the aberrant mitochondria genes. ATP synthase b-subunit was observed in pollen mitochondria and was found to be generally important for male gametophytic development [53]. If the b-subunit is defective, it will cause the dysfunction of F0F1-ATPase, which may impact the energy output of mitochondria, resulting in abnormal anther development with non-functional pollens [48,54,55]. In the present study, protein spot No.12 was identified as beta subunit of ATP synthase F1 sector and it is inferred to be a defective protein, which may lead to the dysfunction of F0F1-ATPase by incorporation into the ATP synthase complex. Quantitative realtime RT-PCR showed that the expression of atp synthase beta subunit RNA was reduced by approximately 70% in YX-1 anthers, relative to the WT (Fig. 5), which is consistent with the difference in protein abundance. In fact, it has been suggested that during microspore development, demand for energy is especially high. The reduced level of ATP synthase beta subunit suggests that the male sterile mutant plants are in an energy starved state. This is in accordance with down-regulation of genes controlling enzymes associated with energy in a CMS line of B. napus [23,56].
Four down-regulated proteins are involved in carbohydrate metabolism: fructokinase-like protein, MDH, aldolase, and GAPDH. Starch is synthesized in anthers before meiosis and subsequently hydrolyzed to provide energy for lipid synthesis in both tapetum and microspores [57]. Decreased abundance of these enzymes in YX-1 anthers could alter levels of sugar and starch, two molecules key to biosynthesis and energy balance.
Several other proteins/enzymes, including GS, APX, putative callose synthase catalytic subunit, CHS, CHS-like and enoyl-ACP reductase, were down-regulated in YX-1 anthers, and these  proteins/enzymes may have a role in tapetum and pollen development. One of the most notable difference in WT and YX-1 anther gels was the location of the protein putative GS (spot No.14, Fig. 2, 3, and Table 2), and may be involved in pollen development. GS is found as two isoforms, cytoplasmic GS1 and chloroplastic GS2, catalyzing the ATP-dependent conversion of glutamine to glutamate. Isolated and in vitro-cultured microspores were unable to develop into functional pollen grains in a medium lacking glutamine [58], which indicated that glutamine plays a key role in plant amino acid metabolism and pollen development. The importance of GS1 in pollen reproduction has been shown in rice [59] and maize [60]. In tobacco, GS1 was inhibited by introducing mutated tobacco GS genes fused to the tapetum-specific TA29 and microspore-specific NTM19 promoters, and pollen aborted close to the first pollen mitosis in the transgenic plants, resulting in male sterility [61]. YX-1 anthers also showed lower GS activity relative to WT (Fig. 6A). The decreased GS activity in YX-1 anthers could cause a reduction in glutamine, which is required for pollen development thereby resulting male sterility.
It is notable that three spots (spot No.26, 27 and 28; Fig. 2 and Table 2) were identified as APXs and all of them showed lower amounts in YX-1 relative to WT anthers. Just as in cotton and rice [62,63], lower activity/amounts of oxidative stress enzymes in cytoplasmic male-sterile anthers was detected compared with fertile anthers. Moreover, the expression of apx RNA and the APX activity were lower in YX-1 anthers, respectively, relative to the WT ( Fig. 5 and Fig. 6B), during the process of anther abortion, when a great deal of ROS might be generated in the anther cell.
Callose synthase is responsible for the synthesis of callose deposited at the primary cell wall of meiocytes, tetrads and microspores in Arabidopsis, and T-DNA insertion mutations of the CalS5 gene resulted in degeneration of microspores, thereby, male sterility [64]. Several studies have also described mutations in callose wall formation and dissolution in petunia [65] and tobacco [66] that disrupt fertility. Putative callose synthase catalytic subunit (spot No.35; Fig. 2 and Table 2) showed low spot volume in YX-1 anthers. Collectively, the evidence indicates that the timing of callose formation and dissolution are critical for normal fertility.
Most plant phenolics, including flavonoids, are products of phenylpropanoid metabolism. CHS is one of the main enzymes in the flavonoid biosynthesis pathway, and an alteration in CHS abundance would be expected to affect the accumulation of all classes of phenolic compounds. Generally, tapetal cells produce proteins and lipids, as well as flavonoids, which are secreted into the pollen sac and form part of the exine [4]. Several enzymes involved in secondary metabolism, including CHS, are specifically or predominantly expressed in the tapetum [67]. It is reported that CHS is essential for pollen development and fertility in several plant species, and disruptions to CHS activity in the anthers resulted in the production of sterile pollen [68][69][70][71][72]. In addition, in recent research in Arabidopsis anthers, LAP5 and LAP6, encoding anther-specific proteins with similarity to CHS, were suggested to play a role in pollen development and exine formation [73]. All of the above results suggest that flavonoids play an important role in the development of functional pollen. This study showed that two proteins, CHS (spot No.37, Fig. 2, 3, and Table 2) and CHS-like protein (spot 36, Fig. 2 and Table 2) are down-regulated in the mutant, indicating the premature degradation of the tapetum in YX-1 mutant is concomitant with the reduction of anther specific CHS abundance. Thus, the level of flavonoids might decrease to below the level required to generate the pollen exine, leading to male sterility.
As a catalytic component of the fatty acid synthetase system in plants, enoyl-ACP reductase is prominently expressed in the tapetum, developing pollen grains, and vascular tissue of anthers. In the Arabidopsis mod1 mutant, reduced activity of enoyl-ACP reductase led to abnormal development of various organs and reduced fertility [74]. It is also reported that the DPW gene, encoding a fatty-ACP reductase, is expressed in both tapetal cells and microspores during anther development in rice, and in a dpw mutant, defective anther development and degenerated pollen grains with an irregular exine appeared [75]. In YX-1, a protein identified as enoyl-ACP reductase (spot No. 38, Fig. 2, and Table 2) showed reduced abundance, which might affect fatty acid synthesis and anther development.
Besides metabolic pathways, proper anther development requires diverse regulatory processes. 14-3-3 proteins, being conserved phosphopeptide binding proteins in eukaryotic organisms [76,77], regulate diverse biological processes in plants, such as metabolism, transcription, organellar protein trafficking, and stress responses [78,79]. There have been some reports that 14-3-3 proteins are associated with ATP synthases in a phosphorylationdependent style, playing a regulatory role in starch accumulation [80], regulation of PCD as a MAPKKKa-interacting protein in pollen development [81], In maize, reduced abundance of the14-3-3 protein led to temporal gene expression changes and, ultimately, pollen sterility [31]. In YX-1 anthers, a 14-3-3 protein (spot No.34, Fig. 2, 3, and Table 2) abundance was downregulated compared with the WT. The aberrant abundances of such 14-3-3 factors could contribute directly to YX-1 defects. Another protein, BTF3 (spot No.30, Fig. 2 and Table 2) was also detected as having a reduced abundance level in YX-1 anthers relative to the WT. BTF3 is the b-subunit of the nascentpolypeptide-associated complex, with a conserved role in regulating protein localization during translation in plants [82]. In a photoperiod-sensitive male-sterile mutant of rice, defects in pollen development were related to abnormal protein localization in anther tissue layers, including the tapetum [83]. The reduced abundance of BTF3 in male sterile anthers was similar to results obtained in tomato [2], which was considered to affect protein localization in the anther and hence affect pollen development.
In YX-1 anthers at the early tetrad stage, some of proteolytic enzymes, including aspartic protease, 26S proteasome regulatory subunit and SKP1,as well as cysteine protease inhibitor, were upregulated and these proteins may have a role in tapetum degeneration. Aspartic protease acts as an anti-cell-death factor participating in PCD, and overexpression of the gene encoding aspartic protease resulted in male sterility in Arabidopsis [84]. In common with the observation in the 7B-1 male sterile mutant of tomato [2], spot No.22 (Fig. 2 and Table 2) was identified as aspartic protease with increased amounts in mutant anther. The higher abundance of aspartic protease in the YX-1 anther could disturb PCD of the tapetum and pollen development, causing male sterility.
Another protein with higher spot volume in YX-1, relative to WT anthers, was 26S proteasome regulatory subunit (spot No.20, Fig. 2, 3 and Table 2). Proteasomes are regulators of many processes such as the cell cycle, embryogenesis, metabolism, gametophyte survival, hormone signaling, senescence and defense [85,86], and have been identified in plant reproductive organs, such as anthers [2,86]. During PCD, proteasomes are released into the extracellular space and have the potential to damage nearby cells. The higher level of this protein in YX-1 compared with WT anthers might disturb the degradation of regulatory proteins in anther tissues, hence leading to premature degradation of the tapetum and male sterility. Selective proteolysis of proteins mediated by the ubiquitin pathway is an important pathway for controlling many biological events. The SCF class of E3 ubiquitin ligases controls the ubiquitination of a wide variety of substrates, thereby mediating their degradation by the 26S proteasome. In Arabidopsis, it was reported that the Skp1 homologue ASK1 involved in the regulation of pollen development, and the ask1-1 mutant produces polyads containing microspores of variable number and size, leading to non-viable pollen grains and male sterility [87]. In this study, the reduced level of protein spot putative SKP1 (spot No.19; Fig. 2 and Table 2) in YX-1 anthers may affect pollen development.
In plants, cysteine protease inhibitor act as regulators of endogenous proteolytic activities. In rice, the TDR gene controls tapetum degeneration by targeting anther specific cysteine protease and protease inhibitor genes [88]. The male sterile mutant showed a reduced activity of cysteine protease (spot No.17; Fig. 2 and Table 2) but higher activity of cysteine inhibitor (spot No.23; Fig. 2 and Table 2) and this would inhibit cysteine protease activity and disturb tapetum development thereby effecting male sterility.
Other proteins were also reduced in YX-1 anthers, such as 5B protein, considered to be related to tapetum degradation by inhibiting proteasome activity, are cysteine-rich and are specifically expressed in the tapetum and stamen in plants [25,89]. In tomato, the 5B protein showed lower abundance at the tetrad stage in male sterile mutant 7B-1 [2]. In our study, the abundance of the 5B protein (spot No.18, Fig. 2 and Table 2) was reduced in the YX-1 anther relative to the WT, which indicated that tapetum development was disturbed by abnormal abundance of the 5B protein, leading to pollen abortion. In addition, a number of other proteins showed altered abundance, including calmodulin-like protein 1 (spot No.32; Fig. 2 and Table 2) and putative calciumbinding protein(spot No.33; Fig. 2 and Table 2), which translates a signal of cytosolic Ca 2+ elevation to downstream protein targets in numerous signal transduction cascades [90]. The altered abundance level of these enzymes/proteins in YX-1 anthers might affect the abundance of regulatory proteins in anther tissues and, ultimately, pollen development.

Conclusions
This study applied a proteomic approach to identify regulating proteins in the anthers of a male sterile mutant of wolfberry. We conclude that the breakdown of pollen development at the early tetrad stage of YX-1 mutant anthers is associated with the differential expression of several proteins, including energy conversion related (e. g., ATP synthase subunits), amino acid metabolism related (e.g., GS), stress response related (e.g., APX), proteins with roles in signaling (e.g., 14-3-3 protein), anther development (e.g., putative callose synthase catalytic subunit), as well as proteases and protease inhibitor (e.g., 5B protein, 26S proteasome regulatory subunits, aspartic protease, cysteine protease, cysteine protease inhibitor and putative SKP1). Significantly, the abnormal protein complex of ATP synthase may cause the dysfunction of mitochondrion, and alterations of some protein abundances (such as APX, GS) could be the consequences of mitochondrial dysfunction in the pollen as indicated by changes in abundance of subunits of ATP synthase. These data indicate regulating patterns of wolfberry pollen development is a complex network. The significant impact on pollen fertility in YX-1 appears to be the result of regulation of multiple metabolic pathways with differentially expressed genes potentially involved, which is supported by the results in male-sterile mutants of Arabidopsis, rice, tomato and maize [2,32,91,92].