Recombinant ArtinM activates mast cells

The objective of the present study was to characterize the effect of monomeric rArtinM on mast cells. Therefore, it was essential to confirm that rArtinM was indeed monomeric. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) was used to compare the homogeneity of native and recombinant ArtinM preparations. Under nondenaturing conditions or after thermal dissociation rArtinM exhibited a single protein band of approximately 13 kDa that corresponds to rArtinM monomers (Fig. 1a, lanes 1 and 2). jArtinM, the native tetrameric form, was used as a control. When undenatured jArtinM was loaded onto the gel, a protein band of approximately 60–80 kDa was observed. This band corresponds to jArtinM tetramers (Fig. 1a, lane 3). When jArtinM was submitted to thermal dissociation, a single protein band of approximately 13 kDa, corresponding to the dissociated tetramers (Fig. 1a, lane 4), was observed. These results indicate that E. coli expresses a monomeric form of ArtinM. It is also plausible that E. coli expresses oligomeric forms of ArtinM but these forms cannot be detected by electrophoresis, since their bonds could be dissociated by exposure to SDS.

jArtinM and rArtinM were also submitted to size exclusion chromatography on a Superdex 75 column, which was calibrated by using protein molecular weight standards. jArtinM presented two distinct peaks, the first with the apparent molecular mass of 42 kDa and the second peak with the apparent molecular mass of 22 kDa, together these two peaks had a molecular mass of 64 kDa (Table 1). This estimate is compatible with previous data from mass spectrometry analysis [28]. rArtinM had the lowest molecular mass, 13 kDa, thus reinforcing the hypothesis that rArtinM is expressed in a monomeric form (Table 1).

Because both of these analytical techniques could result in the dissociation of weakly associated monomers, an analytical ultracentrifugation assay (AUC) was performed to determine if rArtinM monomers are able to self-associate in solution. The AUC analysis, based on the sedimentation velocity and also on the sedimentation equilibrium, showed that in solution, rArtinM appeared to be trimeric, and exhibited a molecular weight of approximately 50 kDa (Fig. 1b). Minor peaks were absent from rArtinM. A previous AUC analysis of jArtinM demonstrated that it has an S value of 4.24 and is a 64 kDa tetramer. However, jArtinM is capable of forming higher order oligomeric species since there were minor peaks at S values of approximately 6, 7, 9, 10 [29]. This is in agreement with previous results using molecular modeling [30]. To understand these differences, crystal trials are being conducted on the two ArtinM forms using the crystallization conditions published for native artocarpin [30].

The molecular weights obtained from the sedimentation velocities were confirmed by sedimentation equilibrium, which makes no assumption regarding the hydrodynamic shape of the protein. This analysis shows that, in solution, rArtinM monomers are able to self-oligomerize, mostly into trimers.

It has been previously shown that rArtinM exhibits the same binding affinity for the trimannoside Manα1-3[Manα1-6] Man as does jArtinM. [25, 26] To assess the carbohydrate-binding properties of jArtinM and rArtinM at a higher resolution, glycan microarray analyses was carried out using a panel of 255 lipid-linked oligosaccharide probes representing diverse mammalian glycan sequences and their analogs. Fifty N-glycan related probes were included in the array. jArtinM (Fig. 2a) and rArtinM (Fig. 2b) bound exclusively to the N-glycan sequences, in agreement with published data on the specificity of the native form [31]. There was no significant difference in the specificities of the recombinant and native lectins, suggesting that the features that allow for the modified oligomerization of rArtinM do not affect the three-dimensional structure of its CRD, which is responsible for the glycan binding specificity.

The next step was to investigate if the recombinant form of ArtinM would be able to induce mast cell degranulation. A dose–response curve for the release of β-hexosaminidase and TNF-α was performed (Fig. 3). In the absence of IgE, only the higher concentrations of rArtinM were able to release β-hexosaminidase (20 and 40 μg/ml) and TNF-α (40 μg/ml) (Fig. 3a and b). On the other hand, in the presence of IgE, all concentrations of rArtinM induced β-hexosaminidase release (Fig. 3a), and for TNF-α there was a greater release at 10 and 40 μg/ml of rArtinM (Fig. 3b). The ability of rArtinM to induce the production of the newly formed lipid mediator, LTC4, was also examined. The production of LTC4 was observed only in sensitized cells and occurred in a dose-dependent manner (Fig. 3c). The synthesis of the newly-synthesized mediator, IL-4, was also investigated. Only the higher concentrations of rArtinM (20 and 40 μg/ml) induced the synthesis and release of IL-4 either in the presence or absence of IgE (Fig. 3d).

It was then of interest to determine if the degranulation and activation induced by rArtinM was dependent on its CRDs. Therefore, rArtinM (40 μg/ml) was pre-incubated with its specific sugar, D-mannose, and β-hexosaminidase release assay was performed. Pre-incubation of rArtinM with D-mannose abolished the release of β- hexosaminidase, in the absence or presence of IgE (Fig. 4), indicating that mast cell degranulation mediated by rArtinM is dependent on its CRDs. The effect of rArtinM on mast cell activation was also assessed by scanning electron microscopy (SEM). By SEM unstimulated RBL-2H3 cells are spindle shaped and their surface is covered with fine microvilli (Fig. 5). After stimulation via FcεRI the cells have deep ruffles on their surface and are spread over the substrate (Fig. 5). Incubation with rArtinM (40 μg/ml) produced similar changes, which are consistent with mast cell activation (Fig. 5). When rArtinM (40 μg/ml) was pre-incubated with D-mannose, the cells had the same morphological characteristics as the unstimulated cells (Fig. 5). Taken together, these results show that mast cell activation and degranulation induced by rArtinM is mediated by its CRDs.

Since the recombinant form of ArtinM is produced as monomers, it was of interest to investigate if rArtinM could act as a co-stimulatory molecule in the FcεRI signaling pathway. Co-stimulatory molecules act in parallel with a major signaling pathway, either by amplifying the intensity of the stimuli or by extending/prolonging its time-response [32]. For this, rArtinM was incubated for 30 min along with the IgE-TNP specific antigen, DNP₄₈-HSA, and β-hexosaminidase release assay was performed. Interestingly, when rArtinM was jointly incubated with DNP₄₈-HSA, the levels of β-hexosaminidase released were diminished by 37 % (Fig. 6) compared to cells stimulated with DNP₄₈-HSA alone, suggesting that rArtinM disrupts the IgE-DNP₄₈-HSA binding, probably blocking the interaction between DNP₄₈-HSA and IgE.

Since rArtinM was able to activate and degranulate mast cells, it was of interest to investigate if rArtinM is activating transcription factors such as NFkB and NFAT. For this, transfected RBL-2H3 cell lines with a GFP-gene reporter for NFkB and NFAT expression were used. rArtinM was able to activate NFkB in an IgE-dependent manner, but rArtinM did not activate NFAT (Fig. 7).

rArtinM was expressed in Escherichia coli BL21- CodonPlus(DE3)-RP and purified as previously reported [25]. rArtinM preparations containing less than 0.05 ng/ml of bacterial endotoxin, as determined by the Limulus amoebocyte lysate assay, were used in this study (Sigma-Aldrich., St. Louis, MO).

Native and recombinant forms of ArtinM were submitted to size exclusion chromatography for molecular weight determination, on a Superdex 75 column (Sigma Aldrich) coupled to an AKTA protein purification system (GE Healthcare, Uppsala, Sweden), which was calibrated by using protein molecular weight standards (Protein Mixture, GE Healthcare). The molecular weight of proteins was determined by partition coefficient (Kav) using this formula: Kav = Ve-Vo/Vt-Vo, where Ve is the elution volume of the samples, Vt is the total volume and Vo is the void volume of the gel bed. High molecular weight blue dextran was used to determine the void volume.

Sedimentation velocity measurements were performed using a Beckman XL-A analytical centrifuge equipped with both absorbance and interference optics. All data were acquired at a rotor-speed of 50,000 rpm at 20 °C using a Beckman An60Ti rotor. For each sample, 100 scans were acquired at 120 s intervals. Buffer density and viscosity as well as the partial specific volume of the protein were calculated using SEDNTERP (Alliance Protein Laboratories, Thousand Oaks, CA,).

The native and recombinant ArtinM forms were biotinylated as previously described [38] and quantified by determining their absorbance at 280 nm (OD280). Microarrays were composed of lipid-linked oligosaccharide probes robotically printed in duplicate on nitrocellulose-coated glass slides at 2 and 7 fmol per spot (in-house designation sets 18–21bis) using a non-contact instrument, as previously described [57]. The microarray binding assays of biotinylated ArtinM proteins were performed at 19 °C–20 °C, as previously described [58]. In brief, the slide arrays were blocked with 1 % w/v bovine serum albumin (Sigma Aldrich) in casein blocking solution (Pierce Chemical Co, Thermo Fisher, Waltham, MA) for 1 h. The biotinylated ArtinM (50 μg/mL) was overlaid, and binding was detected with streptavidin conjugated to Alexa Fluor 647 (Molecular Probes, Thermo Fisher, Waltham, MA) at 1 μg/mL in blocking solution. Glycoarray data analysis was performed with dedicated software [59]. The binding signals were probe-dose dependent.

RBL-2H3 cells, a rat mast cell line, were used in this study [60]. The stable transgenic RBL-2H3 derived cell lines, VB9 and NFkB 2, were also used. The VB9 cell line is a GFP- reporter cell line for NFAT activation [54]. The NFkB 2 cell line is a GFP-reporter cell line for NFkB activation [61], which presents a genome transduction with a reporter vector that possess 4 copies of the binding site for NFkB that regulate GFP expression. All cells were grown as monolayers in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Thermo Fisher, Waltham, MA) supplemented with 15 % fetal calf serum (Sigma-Aldrich) as previously described [60]. Transfected cells were selected with geneticin (0.4 mg/ml) (Sigma-Aldrich). All cell lines were generously provided by Dr. Reuben P. Siraganian, NIDCR, NIH, Bethesda, MD.

As a positive control for FcεRI stimulated cells, the cells were cultured in the presence of a 1:5000 dilution of mouse IgE anti-TNP ascites fluid (kindly provided by Dr. Reuben P. Siraganian) for 16 h, and then stimulated with the multivalent specific antigen DNP₄₈-HSA (Sigma-Aldrich) at 50 ng/ml. As positive control for FcεRI-independent stimulation, the cells were incubated with calcium ionophore-A23187 (Sigma-Aldrich) at 0.1 μM for degranulation assays and, at 0.7 μM for NFkB and NFAT activation assays. In experimental conditions, the cells were sensitized or not with IgE anti-TNP ascites fluid (1:5000) for 16 h and then stimulated with rArtinM. In some experiments, rArtinM was preincubated with D-mannose 100 mM (Sigma-Aldrich) for 1 h at 4 °C. To evaluate release of preformed mediators (β-hexosaminidase and TNF-α) and the lipid mediator LTC4, cells were stimulated for 45 min. For scanning electron microscopy, the cells were stimulated for 20 min. For newly-synthesized IL-4, the cells were stimulated for 12 h. For NFkB and NFAT activation, the cells were stimulated for 5 and 17 h respectively.

3.0 × 10⁴ cells/well were plated in 96 well tissue culture plates (Corning Life Sciences, Lowell, MA) in the absence or presence of IgE and cultured overnight. The cells were washed 2 times with Tyrode’s buffer (137 mM NaCl; 2.7 mM KCl; 12 mM NaHCO₃; 0.37 mM NaH₂PO₄; 0.1 mM MgCl₂; 1.3 mM CaCl₂; 10 mM Hepes, pH 7.3) supplemented with 0.1 % BSA (Sigma-Aldrich) and 0.01 % gelatin (Sigma-Aldrich) and then incubated with the stimulus diluted in Tyrode’s buffer for 45 min as described above in Cell sensitization and stimulation. After stimulation, the supernatants were transferred to clean wells and β-hexosaminidase activity measured as previously described [1]. All assays were run in triplicate.

IgE-sensitized or unsensitized cells were incubated with rArtinM for 45 min or 12 h. The concentrations of LTC4, TNF-α and IL-4 in the cell culture supernatants were measured by ELISA (Leukotriene C4 EIA kit, Cayman Chemical Company, MI, USA; OPTEIA™ Rat TNF ELISA kit II; OPTEIA™ Rat IL-4 ELISA kit II, BD Biosciences, San Jose, CA) according to the manufacturer’s instructions.

Cells (3x10⁴) were cultured on 13 mm diameter glass coverslips in 24 well plates (Corning Life Sciences). The cells were cultured in the presence or absence of IgE for 16 h. The cells were then stimulated as described in Cell sensitization and stimulation. The samples were prepared as previously described [62], and were examined with a JEOL JSM-6610 LV scanning electron microscope (JEOL, Ltd.; Tokyo, Japan).

NFkB2 and VB9 cells (1x10⁵) were sensitized or not with IgE, and then stimulated. Fluorescence levels were measured using a Guava Personal Cell Analysis-96 System and data were processed by Guava InCyte Software (Millipore Co., Billerica, MA).

Data was analyzed using GraphPad Prism (GraphPad Software, Inc., La Jolla, CA). Results were expressed as mean ± SEM. Differences between groups were assessed by one-way ANOVA followed by Tukey’s Multiple Comparison Test. *p < 0.05; **p < 0.01; ***p < 0.001.