Amaranthus leucocarpus lectin recognizes a moesin-like O-glycoprotein and costimulates murine CD3-activated CD4+ T cells

The Galβ1,3GalNAcα1,O-Ser/Thr specific lectin from Amaranthus leucocarpus (ALL) binds a ∼70 kDa glycoprotein on murine T cell surface. We show that in the absence of antigen presenting cells, murine CD4+ T cells activated by an anti-CD3 antibody plus ALL enhanced cell proliferation similar to those cells activated via CD3/CD28 at 48 h of culture. Moreover, ALL induced the production of IL-4, IL-10, TNF-alpha, and TGF-beta in CD3-activated cells. Proteomic assay using two-dimensional electrophoresis and far-Western blotting, ALL recognized two prominent proteins associated to the lipid raft microdomains in CD3/CD28-activated CD4+ T cells. By mass spectrometry, the peptide fragments from ALL-recognized proteins showed sequences with 33% homology to matricin (gi|347839 NCBInr) and 41% identity to an unnamed protein related to moesin (gi|74186081 NCBInr). Confocal microscopy analysis of CD3/CD28-activated CD4+ T cells confirmed that staining by ALL colocalized with anti-moesin FERM domain antibody along the plasma membrane and in the intercellular contact sites. Our findings suggest that a moesin-like O-glycoprotein is the ALL-recognized molecule in lipid rats, which induces costimulatory signals on CD4+ T cells.


Introduction
Activation of CD4 þ T cells is a crucial event in the adaptive immune response, since these cells control humoral and cellular immunity through the recognition of specific antigens [1]. The activation process initially involves interaction between T cell receptor (TCR) and MHCpeptide complexes on antigen-presenting cells (APCs) resulting in the formation of an immunological synapse [1]. Receptor-ligand complexes implicated in the immunological synapse are glycoproteins, which help both cell junction and lipid raft mobility on the membrane surface [2]. Lipid rafts are membrane microdomains that concentrate signalling molecules required for cellular costimulation and activation [3] in which the band 4.1, ezrin, radixin, and moesin (FERM) family proteins have an essential role in organizing the interactions among transmembrane proteins and the underlying cytoskeleton [4].
T cell activation requires stimulation via TCR and costimulation signals; a powerful costimulatory pathway is provided through the CD28 molecule; however, other lipid raft-proteins have been shown to generate enough costimulation to activate T cells [5]. Some studies have shown that the saccharide portion of glycosylated proteins, such as CD2 and CD8, have an important role in the co-receptor interactions as reviewed [6]. For this reason, it is important to know how the saccharide structure influences many of these processes.
Cell surface oligosaccharides have been shown to play an important role in recognition events [6]. Oligosaccharides

Mice
Six-to eight-week-old male Balb/c mice were killed by cervical disruption in accordance to the guidelines accredited by the Institutional Research and Ethics Committee. Cells from the axial lymph nodes were obtained, washed, and suspended in RPMI-1640 medium culture. Cell viability was assessed by the trypan blue dye exclusion test.
Purified CD4 þ T cell culture and proliferation assays CD4 þ T cells were isolated from lymph nodes using a MACS negative selection kit according to the manufacturer's instructions (Miltenyi Biotec). Purity of CD4 þ T cells was assessed by flow cytometry using PE-labeled anti-CD4 mAb and CyChr-streptavidin after biotin-labeled anti-CD3 mAb.
The purity and viability percentages of cell suspensions were 95-98%.
Purified CD4 þ T cells at a density 1 Â 10 7 /mL of RPMI-1640 medium were stained with CFSE according to described method [17]. In brief, 1 mL of suspended cells in RPMI medium, were incubated with 15 mL of 0.5 mM CFSE (prepared from a 5 mM stock solution dissolved in dimethyl sulfoxide) for 10 min at 378C in darkness. After incubation, cells were washed twice in 8 mL of RPMI medium supplemented with 10% heat-inactivated fetal calf serum, 1% L-glutamine, 100 mM sodium pyruvate, 100 IU penicillin, 50 mg/mL gentamicin, 100 mg/mL streptomycin, and 50 mM b-mercaptoethanol (supplemented-RPMI medium), and cell viability was evaluated by trypan blue dye exclusion test. CFSE-treated cells were cultured in a 96-well plate (Nunc, Thermo Fisher Scientific, Inc. Waltham, MA) with a supplemented-RPMI medium and stimulated with either 0.1, 1, or 5 mg/mL of plate-immobilized anti-CD3 mAb alone or in the presence of 1 mg/mL soluble anti-CD28 mAb or ALL at different concentrations (5, 10, 15, or 20 mg/mL) during 24, 48, 72, and 96 h at 378C in a 5% CO 2 atmosphere. Cells cultured with anti-CD28 mAb plus ALL or with these reagents separately were used as controls. Both CFSE-untreated and non-stimulated cells were also used as controls. At the end of the culture periods, the cells were acquired on a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) and analyzed by the FlowJo software (Tree Star, Inc. Ashland, OR). To evaluate the cell divisions that have occurred under stimulation, the FlowJo proliferation platform was used. A histogram based on the fluorescence intensity of unstimulated CFSE-stained cells, which lay within a scatter gate, was performed to differentiate the divided cells of lower intensity. The subsequent gates enclosed cell populations with progressive twofold decreases in fluorescent intensity. Cell proliferation was assessed by percentage of divided cells, cell proliferation index, and division index. Additionally, images from activated cells were acquired on a Zeiss Axivert 25 inverted microscope (Carl Zeiss, G€ ottingen, Germany).

Membrane cholesterol depletion by methyl-bcyclodextrin
Purified CD4 þ T cells were stimulated by a 1 mg/mL immobilized anti-CD3 mAb alone or in the presence of 1 mg/ mL soluble anti-CD28 mAb or 5 mg/mL ALL, for 48 h of culture. Once cultured, cells (5 Â 10 5 ) were washed in phosphate-buffered saline (PBS) and incubated with 10 mM MbCD for 30 min at 378C [18,19]. Then, cells were washed in PBS containing 0.1% BSA and 0.1% sodium azide (PBS-BSA buffer), incubated with biotin-ALL (15 mg/mL) for 30 min at 48C, followed by a second incubation with CyChr-streptavidin at dilution 1:400 and analyzed by flow cytometry. Non-activated cells incubated with CyChrstreptavidin after biotin-ALL, and activated-cells incubated only with CyChr-streptavidin were used as controls.

Flow cytometry analysis
The percentage of cell surface molecules on purified CD4 þ T cells was analyzed by direct or indirect immunofluorescence. In brief, cells were washed in a PBS-BSA buffer and stained with fluorescent reagents during 15 min at 48C for incubation. Afterwards, cells were washed in PBS-BSA buffer, suspended in FACS-Flow buffer and acquired in the FACSCalibur flow cytometer (BD Biosciences). In each case, 25 Â 10 3 cells were counted in linear mode for side and forward scatter and by log amplification for fluorescent cells. Fluorochrome-labeled isotype-matched control mAbs or fluorescent-second reagents were used to evaluate background staining. Data from cytometer were analyzed by the FlowJo software (Tree Star, Inc.).

Cytokine mRNA production
Purified CD4 þ T cells (5 Â 10 6 ) were simultaneously activated by 1 mg/mL immobilized anti-CD3 mAb plus 1 mg/mL anti-CD28 mAb or 1 mg/mL immobilized anti-CD3 mAb plus 5 mg/mL ALL (optimal concentrations) for 48 h of culture. Non-stimulated cells were used as control. To evaluate cytokine transcripts, cells were washed in sterile PBS, and total RNA was isolated by using the mini RNeasy kit according to manufacturer's instructions (Qiagen) and quantified by a spectrophotometer at 285 nm. One microgram of an RNA sample was reverse-transcribed through the Omniscript RT kit and 1 mg from the resulting cDNA was used to amplify the IL-2, IL-4, IL-10, TNF-a, IFN-g, and TGF-b genes by polymerase chain reaction (PCR). Each cDNA sample was amplified in duplicate as described [20]; the sets of primers and the cDNA concentration were calibrated for a number of cycles to obtain amplicons in the linear phase of amplification. The following gene specific primer sequences were used: ( PCR reactions were performed in a total volume of 30 mL buffer containing 50 mM KCl, 10 mM Tris-HCl, pH 8.3; 0.1 mg/mL gelatin, 2 mM MgCl 2 , 100 nM of each primer, 200 mM dNTPs, and 0.5 U of ampli-Taq polymerase/75 ng of cDNA. After 29-36 cycles, the PCR products were analyzed in 10% acrylamide gels and stained with ethidium bromide. The amount of the corresponding cytokine mRNA per microgram of total RNA, normalized to the amount of b-actin mRNA, was determined by scanning densitometry and expressed as arbitrary units.
For intracellular cytokine detection, cells were activated for 48 h as described and 4 h before brefeldin-A was added (1 mg/mL). At the end of the incubation period, cells were harvested, washed in PBS-BSA buffer, and fixed with 4% p-formaldehyde in PBS for 10 min at 48C. Then, cells were washed twice in PBS and permeabilized with saponine buffer (0.1% saponine, 0.01% pig IgG, 10 mM HEPES, 10% BSA in PBS), shaking gently for 10 min at 48C. Subsequently, cells were incubated with fluorochrome-labeled Abs against IL-2, IL-4, IL-10, TNF, IFN-gamma, and TGF-beta for 30 min at 48C. Cells were washed in 0.1% saponine, 100 mM HEPES, 10% BSA in PBS, then washed in PBS-BSA buffer, and analyzed by flow cytometry.

Isolation of lipid rafts from CD4 þ T cell activated via CD3/CD28
Purified CD4 þ T cells were stimulated by optimal concentrations of anti-CD3 mAb plus anti-CD28 mAb for 48 h of culture. Non-stimulated cells were used as control. Once cultured, cells were treated using a slightly modified method [21]. In brief, cells (1 Â 10 8 ) were incubated in 200 mL lysis buffer (1% Triton X-100 in base buffer TEN containing 10 mM Tris-HCl, pH 7.5, 5 mM EDTA, and 150 mM NaCl supplemented with a protease inhibitor mixture from Roche) and shaken for 30 min at 48C. The cell lysate was sonicated four times Â 15 s at 14 Hz frequency with intervals of 30 s on ice and clarified (1500 g for 5 min at 48C). The resulting supernatant was mixed in OptiPrep in base buffer TEN to obtain a solution of 40% OptiPrep; 0.75 mL of the solution were transferred to a 2.5 mL ultracentrifuge tube (Beckman, Palo Alto, CA). Next, 1.2 mL of 30%, 0.25 mL of 5%-OptiPrep, and 0.25-mL base buffer TEN were sequentially overlaid on top of the clarified cell lysate. The sample was then centrifuged at 60,000 g in a TL555 rotor (Beckman) movable angle for 2 h at 48C. Seven 0.34 mL fractions were collected gradually from the top of the gradient and stored at À708C. The protein concentration in each OptiPrep gradient fraction was determined by a DC protein assay kit according to manufacturer's instructions (Bio-Rad). BSA was used as the standard.

Nano LC-ESI-MS/MS
Protein spots of interest in 2-DE were identified by comparing proteins stained in Coomassie blue and those recognized through ALL in far-Western blotting. Gel spots were excised from the 2-DE with a sterile scalpel. Gel pieces were washed in 50% (v/v) acetonitrile and 25 mM sodium bicarbonate (pH 8.5) for 15 min twice to remove Coomassie dye. Next, samples were dried in a vacuum after dehydration with 100% acetonitrile for 10 min at room temperature, followed by rehydration with sequencing-grade modified trypsin (Promega, Madison, WI) in 25 mM ammonium bicarbonate (pH 8.5) at 378C overnight. In-gel tryptic digested samples were trapped and desalted isocratically on an LC-Packing PepMap C 18 m-pre-column cartridge (Dionex, Sunnyvale, CA) and loaded into an integrated nano-LC-ESI-MS/MS system by an analytical C 18 capillary column connected online to a quadrupole acceleration time-of-flight, Ultima API, and mass spectrometer (Micromass, Manchester, UK). Instrumental operation, data acquisition, and analysis were carried out under the full control of Mass-Lynx 4.0 (Micromass). The 1-s survey scans were run over in the m/z mass range of 400-2000. A maximum of three concurrent MS/MS acquisitions were triggered for 2þ, 3þ, and 4þ charged precursor detection at an intensity above the predefined threshold. Product ions were analyzed by the Mascot software (www. matrixscience.com) using both NCBInr and EST databases. Parameters for the Mascot search were peptide mass tolerance of 1 Da, MS/MS ion mass tolerance of 1 Da, maximally one missed cleavage and tryptic digestion. Variable modifications included cysteine carbamidomethylation and methionine oxidation [25]. Only proteins with ion scores >30 were reported.

Confocal microscopy
Purified CD4 þ T cells (2 Â 10 6 ) were cultured in chamber slides (BD Falcon, San Diego, CA) and activated by anti-CD3 plus anti-CD28 mAbs for 48 h. After removing the supernatant, cells with or without treatment in 1% p-formaldehyde (for 5 min), were washed twice in PBS. Cells were incubated with 1% BSA in PBS for 20 min, washed in PBS, and incubated again either with biotin-ALL (15 mg/mL) for 2 h, followed by a second incubation with FITC-streptavidin at dilution 1:200, or with rabbit antimoesin FERM domain mAb, followed by an Alexa Fluor 546-labeled donkey anti-rabbit IgG secondary antibody for 30 min at 48C in humidity chamber. Cells were then washed in PBS and slides were mounted in Vectashield (containing DAPI dye) at dilution 1:3 in PBS. Cells incubated with FITCstreptavidin or Alexa Fluor 546-secondary antibody were used as controls. For colocalization assays, cells were fixed in 1% p-formaldehyde and visualized by microscopy equipped with the ApoTome.2 confocal system (Carl Zeiss, Oberkochen, Germany) using a 63x/1.4 objective lens. The digital images were processed by AxioVision based on Release 4.8.2 SP1 software (Carl Zeiss, Jena, Germany). The images were taken under the same exposure, magnification, and intensification; the processing was the same for all the images shown.

Statistical analysis
Data were analyzed by Graphpad Prism 5 software (La Jolla, CA) and Origin 8.5.1 sr2 software (Northampton, UK), using a Shapiro-Wilk test to reveal population distributions. Student's t-test was performed for comparison of variables that were symmetrically distributed, and values are shown as mean AE standard deviation (SD). To compare groups, an analysis of variance test followed by Bonferroni's multiple comparison test was carried out. Values were considered statistically significant at P < 0.05.

ALL enhanced proliferation of CD4 þ T cells activated by anti-CD3 antibody
To confirm whether ALL can act as a costimulatory molecule, purified CD4 þ T cells were simultaneously stimulated via immobilized anti-CD3 mAb plus soluble ALL. The results showed 39.5 AE 0.6% CD4 þ T cell proliferation at 1 mg/mL anti-CD3 mAb, which increased 1.7-fold (69.1 AE 0.9%) in the presence of 5 mg/mL ALL (P ¼ 0.03) only at 48 h of cell culture. However, under these conditions, at 10, 15, or 20 mg/mL ALL did not show the highest increase in cell proliferation. Likewise, ALL alone was unable to induce CD4 þ T cell proliferation, showing (even at high doses) behavior similar to that of non-stimulated cells (data not shown). To compare the effect of ALL on CD3-substimulated CD4 þ T cell proliferation versus that mediated by the CD28 Figure 1. The CD4 þ T cell proliferation percentage induced by CD3/A. leucocarpus lectin is comparable to that induced by CD3/CD28. Murine lymphoid node CD4 þ T cells were purified, cultured, and analyzed by flow cytometry and microscopy. (a) Histograms representing cells stained by carboxyfluorescein succimidyl ester (CFSE) and cultured at indicated times. Cell proliferation (sequential halving of fluorescence intensity) was induced under indicated conditions using 1 mg/mL anti-CD3 mAb and either 1 mg/mL anti-CD28 mAb, or 5 mg/mL ALL. molecule, cells were treated as above, at different times of culture. In Figure 1a, the analysis by histograms showed that CD4 þ T cell proliferation percentage induced via CD3 plus ALL (54.8 AE 2.0%) was comparable to that induced by CD3 plus CD28 at 48 h of culture (63.6 AE 8.1%). The effect of ALL on CD3-substimulated CD4 þ T cells was maintained until 72 h, whereas costimulation by CD28 persisted until 96 h of culture (Fig. 1b). On the contrary, no difference was detected at 24 h of cell culture under any conditions tested (Fig. 1a). The quantitative analysis showed that proliferation index was 1.1-fold higher in cells activated by anti-CD3/CD28 than those cells activated by anti-CD3/ALL at 48 h of culture (Table 1). In both cell stimulations, similar results were observed through the division index at 48 h; and the values were the same for cell proliferation index and division index at 72 h of cell activation (Table 1). Analysis by microscopy showed typical cell clusters formed through sub-stimulation with immobilized anti-CD3 mAb (1 mg/mL) alone on purified CD4 þ T cells at 48 h of culture (Fig. 1c). The massive cell clusters increased in size and number in culture of CD4 þ T cells stimulated via CD3/CD28, which were similar shape in those cells stimulated via CD3/ALL (Fig. 1c, upper panels). The quantitative measurement on area occupied by cell clusters showed a same average value between cells stimulated by CD3/CD28 and CD3/ALL (Fig. 1d). In contrast, cells incubated with ALL or anti-CD28 mAb or ALL plus anti-CD28 mAb exhibited a similar shape than non-stimulated cells (Fig. 1c, lower panels).

CD4 þ T cells stimulated by anti-CD3 mAb plus ALL produce cytokines
To evaluate cell function, CD4 þ T cells were simultaneously stimulated by anti-CD3 mAb plus ALL. The results showed similar levels of transcripts for IL-2, TNF-a, TGF-b, and IFN-g compared to those shown by CD3/CD28-stimulated CD4 þ T cells (Fig. 2a). However, cells activated via CD3/ALL showed a lower production of mRNA for IL-4 and IL-10 than those cells activated by CD3/CD28. The difference was statistically significant for both IL-4 (0.06 AE 0.05 vs. 0.64 AE 0.19; P ¼ 0.005, respectively) and IL-10 (0.19 AE 0.08 vs. 0.39 AE 0.07; P ¼ 0.03, respectively) (Fig. 2a). With regard to the levels of soluble cytokines measured in cell culture supernatant, the IL-2 was 67-fold higher in cells activated by CD3/CD28 than in those cells activated by CD3/ALL (P < 0.005). On the order hand, IL-10 was 1.5-fold higher in the culture supernatant of cells activated by CD3/ALL than in cells activated by CD3/CD28 (P < 0.05) as shown in Figure 2b. Other cytokines did not show significant differences in their amounts, the levels of both IFN-g and IL-4 were almost undetectable (Fig. 2b). Concerning to the measurement of intracellular cytokines, there were a higher percentage of cells positive to IL-2 and TGF-b in the culture of cells activated by CD3/ALL than in cells activated by CD3/ CD28 (Fig. 2c).

A. leucocarpus lectin recognizes a receptor associated to lipid rafts
Because freshly obtained mouse CD4 þ T cells incubated with FITC-labeled ALL showed formation of fluorescent patching and capping on cell surface (data not shown), we examine whether cholesterol depletion in lipid rafts affects cell membrane recognition by ALL. For this, CD4 þ T cells were treated with MbCD after activation via CD3/CD28 or CD3/ ALL for 48 h. Results showed that cells sub-stimulated by immobilized anti-CD3 antibody alone at a dose of 1 mg/mL, a low cell proliferation was observed (Fig. 3, lower left panel). Under this condition, ALL recognized 27% of the cells, which was lost after treatment with MbCD (Fig. 3, upper left  panel). In contrast, cells stimulated by anti-CD3 plus anti-CD28 mAbs, the percentage of ALL-positive cells increased 2.8-fold (Fig. 3, upper middle panel). Similarly, cells stimulated by anti-CD3 mAb plus ALL, the percentage of cells positive to ALL increased 2.3-fold (Fig. 3, upper right  panel). Furthermore, the recognition of proliferating cells by 188 ALL was returned to a background level after treatment with MbCD (Fig. 3, upper middle and right panels).

A. leucocarpus lectin recognizes two prominent proteins from lipid rafts
The far-Western blotting analysis of the resulting lipid raft fractions (isolated from CD3/CD28-activated CD4 þ T cells during 48 h) showed a marked reactivity of ALL on two bands from fraction 7 (Fig. 4a, middle panel), which were also recognized by the cholera toxin B subunit (Fig. 4a, lower  panel). Additionally, ALL recognized bands in fractions 5 and 8 but at a lesser extent (Fig. 4a, middle panel). Here, the cholera toxin B subunit was used as a reagent to confirm the isolation of lipid rafts. Subsequently, to identify the cholesterol-rich membrane proteins recognized by ALL, a far-Western blotting of fractions 5 and 7 was again performed after 2-DE. In fraction 7, around 40 protein spots were observed in Coomassie staining (Fig. 4b), where two prominent protein spots with mobility relatively close to 70 kDa were selected (amplified region in Fig. 4c). The selected protein spots in turn were recognized by ALL (Fig. 4d). Same results were obtained from fraction 5, but at a lower intensity than from fraction 7 (data not shown). The ALL-recognized protein spots resulting from fraction 7 were subjected to mass spectrometry and compared with the protein database. The analyzed protein spots are shown in Table 2. The number one exhibited a 33% homology to matricin (gi|347839 NCBInr) with a theoretical mass of 61 kDa, whereas the number two showed 41% identity to an unnamed protein closely related to moesin, with a theoretical mass of 67.8 kDa (gi|74186081 NCBInr) and 6.3 pI. This last protein sequence was also analyzed by both the protein structure prediction system, Phyre 2 [26] and NetOGlyc v 4.0 software [27]. The Phyre 2 system showed 90% identity to the FERM domain of moesin, whereas the second showed nine potential O-glycosylation sites; which the positions 165, 205, and 206 exhibited 0.68, 0.78, and 0.86 scores, respectively.
Staining by ALL colocalizes with anti-moesin FERM domain antibody in activated CD4 þ T cell surface To test whether ALL recognizes a FERM family-bearing protein, CD4 þ T cells were activated via CD3/CD28 for 48 h, after, stained with both ALL and anti-moesin FERM domain mAb. Staining was performed on undamaged cells, since they had not been treated with permeabilizing reagents. The fluorescent staining by ALL was observed along of cell surface only in ALL-positive cells, which seem be concentrated at intercellular contact sites (Fig. 5A, D). Interestingly, the vast majority of cells were recognized by anti-moesin mAb, and in some cells, the fluorescent staining was also intense at cellular interaction sites (Fig. 5B, E). Cells incubated with secondary reagents did not showed fluorescent staining (Fig. 5G-I). The colocalization events were observed at intercellular contact sites in the agglutinated ALL-positive cells (Fig. 5C, F).

CD4 þ T cells activated by anti-CD3/anti-moesin mAbs secrete cytokines
To examine whether the anti-moesin FERM domain mAb costimulates CD3-substimulated CD4 þ T cells, purified CD4 þ T cells were simultaneously activated by anti-CD3 mAb plus anti-moesin mAb for 48 h. Results showed activated cell clusters similar to those from cells stimulated  Only proteins that showed significant ion scores (>33) are indicated. *ALL-positive spots identified in Figure 3e. **Represents the identity or extensive homology (P < 0.05) with the indicated protein.
by CD3/CD28 or CD3/ALL. The measurement in pixels 2 was 312.9 AE 435.8 to non-stimulated cells; 2492 AE 1275 to cells upon stimulation by CD3/CD28; 2629 AE 1150 to CD3/ALL; and 2393 AE 1246 to CD3/moesin. Likewise, level of cytokines in culture supernatant from cells activated by CD3/moesin was very similar to those cells activated via anti-CD3 mAb plus ALL (Table 3). Nevertheless, the concentration of IL-17A in CD3/ALL-activated cells was higher than CD3/ moesin-activated cells (Table 3).

Discussion
The role of glycans to induce T cell activation can be shown by the use of lectins, which are proteins or glycoproteins with  reversible binding to their saccharide ligands [28]. The lectins have at least one non-catalytic domain that recognizes oligosaccharides in a specific spatial conformation [28]. One of these lectins, ALL with specificity for GalNAc in either the Thomsen-Friedenreich antigen (Galb1,3GalNAc a1,O-Ser/ Thr) or Tn antigen (GalNAc a1,O-Ser/Thr) [8] was used in this work. The role of ALL in T cell costimulation has been reported [15]; however, we showed that in the absence of APCs, ALL enhanced CD4 þ T cell proliferation when given a CD3-based stimulation, which was comparable to stimulation via CD3/CD28. The CD28 molecule is known as the major T cell costimulatory receptor for the production of IL-2 [29]. Thus, the production of cytokines was determined in ALL/CD3-stimulated cells; in this sense, non-CD28 T cell costimulatory molecules have been found to induce cytokines production. Among these molecules, the stromal cell-derived factor-1 a costimulates cell proliferation and production of IL-2, IL-4, IL-10, and IFN-g on activated CD4 þ T cells [30]. Monoclonal antibodies to CD5 or CD9 induce also potent [ 3 H]TdR uptake and IL-2 production on activated T cells [31]. Likewise, concurrent CD2 and CD48 costimulation enhances CD3-driven T cell activation concerning both proliferation and synthesis of transcript for IL-2 and IFN-g [32]. Nevertheless, T cell costimulation by CD5, CD9, or CD2-CD48 interaction is less efficient than CD28 [31][32]. In our case, the production of transcripts for IL-2, TNF-a, TGF-b, and IFN-g in CD4 þ T cells activated via CD3/ALL did not show statistical differences comparing with those activated via CD3/CD28. Concerning to secreted cytokines, we observed that IL-2 is more abundant in the culture of cells activated through CD3/CD28 in comparison to the other conditions of culture, possibly by positive feedback during the activation [1]. Although the costimulation via ALL appears to be insufficient for production of IL-2, the number of cells positive for IL-2 and TGF-b were higher than the costimulation by CD3/CD28. Altogether our results suggest that the major ALL-recognized glycoprotein is able to costimulate the cytokine production, reflecting the T cell functional activity.
On the basis of that costimulatory molecules are found in lipid raft during T cell activation and because T cells incubated with ALL display formation of patching and capping on cell surface, we used MbCD to deplete cholesterol from lipid raft after activation of CD4 þ T cells. Previous studies of our laboratory have been shown that ALL binds 22% non-stimulated CD4 þ T cells, and the number of ALL-positive cells increases three-fold after 24-48 h activation [13]. The disruption of lipid rafts by treatment with MbCD caused loss in ALL recognition in cells activated by 48 h, suggesting that ALL binds to glycoproteins recruited or located in lipid rafts from activated CD4 þ T cells. The membrane raft microdomains are known as the signaling complexes that process intracellular reaction cascades, where diverse molecules participate in signal transduction [1]. We selected two prominent protein spots with molecular mass close to 70 kDa from lipid raft membranes resolved by twodimensional electrophoresis. The two protein spots were recognized by ALL, one of them showed the highest score of identity (41%) with an unnamed protein (gi|74186081) closely related to moesin. This protein sequence was also analyzed by NetOGlyc 4.0 software [27] showing nine potential sites for O-glycosylation, of which four of them exhibited a very high probability of being recognized by ALL. Previous studies have shown that the binding site of ALL interacts with inner core of the Thomsen-Friedenreich antigen, and that ALL binds GalNAc residues, which are separated by four amino acids, whereas GalNAc residue clusters prevent recognition by ALL [33]. Another analysis on protein sequence was performed using Phyre 2 [26] that showed 90% identity with the FERM domain of moesin. The FERM domain defines members of the band 4.1 family, which includes cytoskeletal proteins, such as the ezrinradixin-moesin (ERM) protein family, as well as several tyrosine kinases and phosphatases [34]. The ERM proteins play an important role in the interface between plasma membrane proteins and major cytoskeletal structures that regulate membrane-associated signalling pathways, as reviewed [4]. Data from confocal immunofluorescence staining showed that ALL and anti-moesin FERM domain mAb colocalized in cellular contact sites, indicating the presence the O-glycans recognized by ALL in these zones. Changes in glycosylation of cell surface O-glycoproteins during T cell activation have been reviewed [2,35]. The decrease of acid sialic on activated T cells [36], allows exposition of GalNAc or Galb1,3GalNAc moieties on cell surface glycoproteins, which could be recognized by ALL. However, as observed in immunofluorescence image, some agglutinated cells were strongly stained by ALL and weakly bound with anti-moesin mAb and viceversa. Moreover, other agglutinated cells were not recognized either by ALL or anti-moesin mAb, suggesting that not all moesin-like proteins display the same glycosylation profile and that different O-glycoproteins could be recognized by ALL. Diverse studies have evidenced the importance of O-glycans on cell surface because they seem be involved in activation process of the T cell. Comelli E.M. et al. reported that the activation of murine T cells lead to a marked decrease of N-glycans carrying NeuAca2,6Gal and the increase of O-glycans carrying the disaccharide Galb1,3GalNAc on the cell surface [37]. Similarly, van Vliet et al. showed that the activation of human T cells via CD3/CD28 was accompanied by an increase in the expression of Galb1,3-GalNAc on the cell surface [38]. We reported in this study direct visualization of a moesin-like protein on the activated CD4 þ T cell surface. Although moesin is localized in the submembranous cytoskeleton, some authors have reported moesin-like proteins on the leukocyte surface [39][40]. Ariel A. et al. demonstrated that a moesin-like protein is expressed on a membrane of human resting T cells, which significantly increased on cell surface after 48 h activation [41]. Interestingly, the anti-moesin FERM domain antibody was able to enhance CD3-based stimulation in CD4 þ T cells after 48 h of culture, being the first time that an anti-moesin FERM domain antibody displays a costimulatory function.

Conclusions
Our findings pointed out that the major glycoprotein recognized by ALL was able to costimulate the production of cytokines; A. leucocarpus lectin binds two glycoproteins situated in lipid raft microdomains from activated CD4 þ T cells; one of which showed identity to a moesin-like protein that was visualised on activated-CD4 þ T cell surface; and that anti-moesin antibody was able to costimulate CD4 þ T cells activated by CD3.