B cell activation leads to shedding of complement receptor type II (CR2/CD21)

2.1 Mass-spectrometric protein identification and structure characterization

Purified sCD21 from human plasma was visualized with colloidal Coomassie stain after SDS-PAGE separation. After in gel tryptic digestion, mass-spectrometric analyses were performed. The high-quality mass spectra were subjected to a database search against the Swiss-Prot sequence database and this identified the protein as CD21 (Swiss-Prot entry: P20023) with sufficient confidence (probability-based score: 137) and a sequence coverage of 33%. Close inspection of the mass spectra showed that the most N-terminal peptide was detected with an ion signal at m/z 1367.71 (Table 1). This result shows that the first 20 amino-acid residues of the sequence deposited in the database form the signal peptide that is removed upon membrane-passage of the translation product.

Inspection of the CD21 sequence covered by mass-spectroscopy showed that only tryptic peptides in which no glycosylation sites are located were detected. Hence, one is tempted to speculate that most if not all glycosylation sites are in fact glycosylated in sCD21. Glycosylated peptides in many cases form ion signals with low abundance and thus may not be detectable in the mass fingerprints.

Furthermore, the recorded mass spectra show ion signals at m/z 2584.18 (A646–R666) and at m/z 2872.42 (C644–R666) that are attributed to the presence of the "short isoform" of CD21 lacking exon 11 (Fig. 1). From these results it seems likely that the majority of the present protein belongs to the "short form" of CD21. However, the presence of the "long isoform" of CD21 cannot be ruled out completely. The most C-terminal peptide that was detected included amino acids G768–R781 with m/z 1525.70 (Table 1). This result indicates that the soluble form of CD21 consists of the extracellular portion of the protein.

It should be noted that the peptide ion signal at m/z 1639.89 can be assigned to two peptides, one covering amino acids 34–48, and one consisting of amino acids 1001–1013 (Table 1). However, as no other peptide ion signals covering peptides in the vicinity of the partial sequence 1001–1013 are detected, the assignment of the mass value of m/z 1639.89 to this peptide seems less likely. By contrast, peptide ion signals to both flanking partial sequences of the peptide with amino acids 34–48 were detected in the mass spectra.

2.2 Peripheral blood B lymphocytes contribute to the plasma sCD21 pool

Soluble CD21 occurs at about 300 ng/ml in the blood of healthy persons and levels change in pathological conditions. However, it is not clear which class of immune cells shed CD21. To investigate the source of sCD21 in plasma, peripheral blood lymphocytes were isolated from two healthy individuals; B and T lymphocytes were sorted and taken into culture. Cell supernatants were collected after 36 h of incubation and sCD21-concentrations estimated by ELISA (Fig. 2). Only B cells were found to shed CD21. T cells shed very little or no detectable amounts of CD21 under these culture conditions. Thus although T cells outnumber B cells in the peripheral blood and express CD21 they do not appear to contribute to the sCD21 serum pool.

2.3 PMA activation of B lymphocytes leads to CD21 shedding

To dissect the mechanism of CD21 shedding we were interested to see if CD21 shedding is associated with lymphocyte activation. Peripheral blood lymphocytes from three healthy individuals were activated with the phorbol ester PMA and CaI for 5 h. Peripheral blood lymphocytes from all the three donors that were tested shed higher amounts of sCD21 when activated with PMA. Therefore, it is tempting to speculate that PMA+CaI activation directly acts on the activity of the protease.

Raji B cells were found to shed at least 10–50-times more CD21 than peripheral blood B cells. Raji B cells also shed more CD21 when activated with PMA+CaI for 5 h (Fig. 3A). The shedding was more pronounced during activation in the presence of PMA+CaI than with PMA or ionophore alone, indicating a synergistic effect of the ionophore on PMA action. A parallel decrease of membrane-associated CD21 was found by cytometric analyses (Fig. 3C).

As PMA-stimulation is pan-specific, we were interested in identifying specific signal transduction mechanism that leads to regulated shedding of CD21. Since CD21 may participate in particular in T cell dependent responses, as suggested by the analysis of CD21-deficient mice 26, 27, we investigated shedding under such conditions. Cross-linking Raji B cells with anti-IgM+anti-CD40 resulted in a similar phenomenon (Fig. 3B and D), suggesting a role of anti-IgM+anti-CD40 stimulation and signal transduction in CD21 shedding.

4.1 Cells, antibodies and reagents

The mature human B cell line, Raji (ATCC, Manassas, USA) was maintained in Iscove's DMEM (Invitrogen, Karlsruhe, Germany) supplemented with 10% FCS (Linaris, Bettingen, Germany), 1000 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen) at 37°C in 7.5% CO₂. The cells were transferred to serum-free medium (Invitrogen) for 24 h prior to shedding-experiments. Monoclonal anti-CD21 antibody clones BU32 (IgG1) 36 and THB5 (ATCC) were grown in serum-free hybridoma medium (Invitrogen) at 37°C in 7.5% CO₂. Peripheral blood B and T cells wereisolated from fresh blood collected from healthy volunteers and cultured in serum-free medium.

4.2 Lymphocyte isolation, sorting and activation

Lymphocytes were purified by ficoll density-gradient centrifugation of fresh human blood. B and T cells were enriched by magnetic sorting as previously described 22. The relative purity of the cells was determined by cytometric analysis using anti-CD4–PE/anti-CD8–FITC as markers for T cells, and anti-CD19–PE/anti-CD21–FITC as markers for B cells. Cells were stained in microtiter plates and loaded into a FACScan (Becton Dickinson, Heidelberg, Germany) using a self-constructed loader as previously described 37. B and T cells (1×10⁷ per ml) were cultured in Iscove's DMEM in tissue culture wells. For Raji B lymphocyte activation, cells were either incubated with PMA+CaI or activated by cross-linking IgM and CD40 in the presence of IL-2 (20 U/ml).

4.3 Purification of sCD21

Soluble CD21 from pooled human plasma was purified to homogeneity by affinity chromatography and density-gradient centrifugation as described before 22 and the purity was checked by 4–12% gradient SDS-PAGE (Invitrogen) and subsequent Western blotting.

4.4 Mass-spectrometric protein characterization

Gel pieces from SDS-PAGE containing sCD21 were excised and subjected to in gel digestion using an Investigator ProGest system (Genomic Solutions Inc., Ann Arbor, MI, USA). Proteins were reduced, alkylated, and digested with trypsin as previously described 28, except that 15 μl of 25 mM NH₄HCO₃ containing 72 ng of modified sequencing gradetrypsin (Promega, Madison, WI, USA) was used for digestion. After extraction of peptides from the gel, the samples were dried in a vacuum centrifuge. The peptides were resuspended in 5 μl of 60% acetonitrile / 0.1% TFA, and 0.5 μl of this peptide mixture was spotted simultaneously with 0.5 λ matrix (a saturated solution of α-cyano-4-hydroxy cinnamic acid in 35% acetonitrile / 0.1% TFA) on a MALDI target plate. Peptides were analyzed by MALDI-TOF MS using a Reflex III mass spectrometer (Bruker Daltonik, Bremen, Germany), equipped with the SCOUT source and delayed extraction, and operated in positive ion reflector mode. The spectra were first calibrated externally 28 and then internally recalibrated using three peptides arising from trypsin autoproteolysis([M+H]⁺ 842.50, [M+H]⁺ 2211.10, and [M+H]⁺ 3323.77). Tryptic, monoisotopic peptide masses were searched against the MSDB database using the Mascot software 1.8 (Matrix Science, Oxford, GB) setting a mass tolerance of 50 ppm. Spectra were analyzed using the BioTools software, version 2.0 (Bruker Daltonik, Bremen, Germany).

4.5 Quantification of sCD21 by ELISA

A sandwich ELISA was developed to quantitate sCD21 in tissue culture supernatants. The monoclonal antibodies THB5 and biotinylated BU32 were used as capture and revealing antibody, respectively. Titration was performed with purified sCD21 22 and a standard curve was plotted. Briefly, THB5 was coated onto an ELISA plate (TPP, Trassadingen, Switzerland) at a concentration of 5 μg/ml in coating buffer (0.1 M Na₂HPO₄ / NaH₂PO₄, pH 9.0) for 12–15 h at 4°C. After two washes with PBS containing 0.1% Tween-20 (PBST), the plate was blocked with 1% milk powder in PBS for 2 h at room temperature. The samples at appropriate dilutions in triplicates were added to the plates along with the standard. The plates were incubated at 4°Cfor 12–15 h, washed twice with PBST and then incubated for 2 h with BU32–biotin at room temperature. After two washes, an appropriate dilution of streptavidin coupled to horseradish peroxidase was added, the plates incubated for 1 h at room temperature, washed twice and H₂O₂/o-phenylenediamine was added as substrate/coloring agent. Enzymatic reactions were quantified measuring optical density at 492 nm in an ELISA reader (Anthos Microsystems, Krefeld, Germany) and sCD21 concentrations were calculated by extrapolating from the standard graph.