Differential Effect of B Lymphocyte–induced Maturation Protein (Blimp-1) Expression on Cell Fate during B Cell Development

BCL1 3B3 (ATCC CRL 1669) cells were purchased from the American Type Culture Collection (Rockville, MD) and grown as previously described (21). BCL1 5B1b cells were provided by Dr. Samuel Strober (Stanford University) and were grown as previously described (22), except for lowering the concentration of FCS to 10%. Where indicated, IL-2 and IL-5 (10% supernatants from respective IL-2– and IL-5–producing cell lines; reference 23) were added. 18-81 cells were obtained from Dr. Matthias Wabl (UCSF, San Francisco, CA; reference 24), and WEHI 231 1/8 cells from Dr. Noel Warner (Becton Dickinson, San Jose, CA; reference 25). L10A (26) and BAL-17.7.1 (27) cells were a gift of Dr. Richard Asofsky (NIAID, Bethesda, MD). The COS-7 (ATCC CRL 1651) cells were obtained from the American Type Culture Collection (28). Human embryo epithelial cells 293T were provided by Dr. Gary Nolan (Stanford University) and were grown as previously described (29).

The Blimp-1 open reading frame (ORF) was subcloned into the EcoRI site of the pyDF30 vector (provided by Stephen Ho and Jerry Crabtree, Stanford University), a derivative of the pcD-SRα expression vector (30) that contains a sequence encoding for a flu antigen (FLAG) octapeptide (31) located 3′ from the multiple cloning site. The last 112 aa of Blimp-1 were removed by using a unique Bsp120I site.

The region encoding for aa 381–450 of Blimp-1 was removed by using the unique AflII and SplI restriction sites in the pyDF30/Blimp-1 vector. The ends of the restriction were blunted and religated, resulting with the generation of an in-frame truncation of 69 aa in Blimp-1 proline-rich domain.

The XbaI–MspI fragment of the pyDF30/ Blimp1 that contains the entire Blimp-1 ORF and flanking sequences was subcloned into the same sites of the pEJM1 vector. This vector was derived from the pSH160s/NFAT vector (provided by Stephen Ho and Jerry Crabtree), and contains a sequence encoding for a FLAG octapeptide followed by three repeats of the SV40 large T antigen nuclear localization signals (NLS) (32). The first 179 aa of Blimp-1's ORF were removed by digesting the pyDF30/Blimp-1 vector with XbaI and XcmI. The resulting construct encoding for an N′ truncated version of Blimp-1 was fused to three repeats of the SV40 NLS and a FLAG tag. This manipulation prevented the deletion of any potential NLS sequences from Blimp-1 NH₂ terminus required for its translocation.

Keller and Maniatis (33) found that the 65 aa that located NH₂ terminal to the zinc fingers were important for protein stability. Therefore, we included the equivalent region of Blimp-1 together with its zinc fingers (hereafter named the NZ domain) in the “minimal Blimp-1 protein”. Nucleotides 1615– 2289 of the Blimp-1 ORF (13) were previously cloned into a pET24d(+) expression vector (Novagene, Madison, WI). In brief, the sequence was amplified for 15 cycles with a high fidelity PCR enzyme (the ULTIMA™DNA polymerase; Roche Molecular Systems, Inc., Branchburg, NJ) and cloned between the BamHI and HinDIII sites of the vector pET24d(+). The NZ fragment was generated from HindIII digestion of the Blimp-1 containing pET24d(+) vector followed by blunting and was then cloned into the EcoRI site of pEJM1 vector.

The green fluorescent protein (GFP) expression vector was based on the pG310 expression vector provided by Dr. Edward Mocarski (Stanford University). In brief, the promoter and enhancer of ie1, exon 1, and intron 1 of CMV (−1022 to +947) were inserted between the XbaI and HinDIII sites of the polylinker of pGEM-2 (Promega, Madison, WI). A polyadenylation signal from the ie1 gene was inserted between the BamHI and ClaI sites (+2756 to +2916). Finally, the NheI–XbaI portion of the pG310 was removed. A vector containing a mutated GFP ORF (provided by Dr. Stanley Falkow, Stanford University), which produces an enhanced fluorescent protein (34), was blunt-ended and ligated into the EcoRV site of LITMUS™ 28 (New England Biolabs, Beverly, MA). An EcoRI–BamHI fragment that contained the GFP insert was cloned into the same sites of pG310. The resulting GFP expression vector was named pTYG-1.

All cells (except 293T) were transfected by electroporation according to Chu et al. (35). BCL1 and COS-7 cells were suspended at 2.5 × 10⁷/ml in DMEM plus 10% FCS. Ten million cells (400 μl) were placed in a 0.4-cm electrode gap gene pulser cuvette (Bio-Rad, Hercules, CA). After the addition of 10 μg of the vector(s), the samples were gently shaken and subjected to electroporation in a GENEPULSER™ apparatus (Bio-Rad) set to 960 μF and 230 mV for COS-7 cells or 250 mV for BCL1 5B1b and 3B3 cells. The cells were then incubated for 10 min at room temperature and resuspended in 10 ml of complete media. Approximately 10% of the cells were plated on a coverslip placed in a 6-well tissue culture dish and used for immunostaining 24 h after transfection. The rest of the cells were plated in a 10-cm dish and incubated for 48 h before lysis and Western blot analysis. 18-81 cells were suspended at 1.7 × 10⁷/ml in RPMI plus 10% FCS. Five million cells (300 μl) were subjected to electroporation set to 960 μF and 240 mV as described above. The transfected cells were cultured in 10 ml RPMI media for 18 h. WEHI 231 1/8 and L10A cells were suspended at 2 × 10⁷/ml DMEM plus 20% FCS. Five million cells (250 μl) were incubated for 10 min on ice with the vector(s) before being electroporated (960 μF, 240 mV). After electroporation, the cells were kept on ice for 5 min, and then 1.75 ml DMEM plus 20% FCS were added and the suspensions were kept at room temperature for 10 min. The transfected cells were transferred to a 6-well petri dish and cultured for 18 h. 293T cells were transfected by the calcium phosphate method according to Pear et al. (36). In brief, 7.5 × 10⁶ cells were transfected with 5 μg of truncated Blimp-1 expression vectors. 48 h after transfection the cells were harvested and nuclear extracts were prepared.

The TUNEL (TdT-mediated dUTP nick end labeling) assay was performed with an in situ cell death detection kit following the manufacturer's instruction (Boehringer Mannheim GmbH, Mannheim, Germany). In brief, DP1/GFP double transfected and GFP transfected control 3B3 and 5B1b cells were sorted by FACS^® (Becton Dickinson) and incubated for an additional 48 h. The cells were then centrifuged onto glass slides using a Shandon cytospin (Pittsburgh, PA), air dried, and fixed for 30 min in 4% paraformaldehyde diluted in PBS (pH 7.4). After permeabilization with 0.1% Triton X-100 diluted in 0.1% sodium citrate, the cells were labeled with 50 μl of the TUNEL reaction mixture for 60 min at 37°C, washed, and incubated for 30 min with 50 μl of converter-alkaline phosphatase. The slides were rinsed and incubated for 10 min at room temperature with 100 μl of Fast Red TR/Naphthol AS-MX solution (Sigma Chemical Co., St. Louis, MO).

pEJM1/NZ transfected cells were harvested 24 h after plating, fixed with 4% paraformaldehyde, and permeabilized with 0.1% Triton X-100 diluted in PBS. The cells were then stained for 1 h with a mixture of 1% BSA, 2% normal rat serum, and a 1:1,000 dilution of anti-FLAG^® M2 antibody (Eastman Kodak Co., Rochester, NY). Cells were then washed with PBS and stained with a 1:1,000 dilution of a biotinylated goat anti–mouse antibody (PharMingen, San Diego, CA) diluted in 1% BSA solution in PBS. The slides were washed again with PBS and stained with 1% BSA and a 1:1,000 dilution of FITC-conjugated avidin diluted in 1% BSA solution in PBS. The cover slips were mounted on slides with 10 μl of VECTASHIELD^® (Vector Labs., Inc., Burlingame, CA), fixed with nail polish, and examined under a fluorescence microscope.

Cells were stained with mAbs for 30 min on ice in 50 μl of PBS supplemented with 5% FCS (FACS buffer). The stained cells were washed twice in FACS buffer, then resuspended in 200 μl of buffer and analyzed on a FACScan^® with the FACSDesk software (Beckman Center Shared FACS^® Facility, Stanford University). The antibodies used for the staining were the monoclonal 6A5.1 anti-BCL1 idiotype (gift of Ellen Vitetta, Southwestern Medical Center, Dallas, TX), FITC-conjugated anti-IgD, PE-conjugated anti–B7-2, FITC-conjugated anti-CD40, and biotinylated anti–Syndecan-1 (PharMingen), and biotinylated anti-IgM (Sigma Chemical Co.). Biotinylated antibodies were used as a second step with either PE- or FITC-conjugated avidin (PharMingen). Dead cells were excluded in each analysis by propidium iodide (PI) staining (37). GFP⁺/PI⁻ transfected 5B1b cells were sorted 15–24 h after transfection using a preset compensation of the GFP for the PI channel. Cells were sorted into RPMI media, washed once with media, and cultured with or without stimuli for an additional 48 h.

293T cells transfected with truncated Blimp-1 mutants were used to prepare nuclear extracts (38). In brief, cells were lysed in 0.5% NP-40, 30% sucrose, 25 mM Tris-HCl, pH 7.5, 25 mM KCl, and 7.5 mM MgCl₂ supplemented with proteinase inhibitors (leupeptin, pepstatin, PMSF, and aprotinin). The nuclei were spun and rinsed with RSB (10 mM NaCl, 10 mM Tris-HCl, pH 7.5, 10 mM MgCl₂), and nuclear proteins were extracted. Nuclei were resuspended first in 200 mM NaCl, 50 mM Tris-HCl, 0.2 mM EDTA, 0.2 mM EGTA, 2 mM MgCl₂, 0.5 mM dithiothreitol (DTT), 5% glycerol, and protease inhibitors. Subsequently, an equal volume of buffer composed of 600 mM NaCl, 50 mM Tris-HCl, 0.2 mM EDTA, 0.2 mM EGTA, 2 mM MgCl₂, 0.5 mM DTT, 5% glycerol, and protease inhibitors was added. The nuclear protein was extracted with rotation at 4°C for 30 min and spun for 60 min at 100,000 g. The protein concentration was determined using a Bradford assay. 5 μg of each nuclear extract was used for EMSA (electromobility shift assay) (38). In brief, the extracts were mixed with 10⁵ cpm of ³²P–end-labeled PRFI (CGCGTACAGAAAGGAAAGGACTAG) (17) or PRDI (GTGAAAGGGAGAAGTGAAAGTGGGAAATTCC) (14) probe, 3 μg poly (dI-dC) in binding buffer (10 mM Tris-HCl, pH 7.5, 1 mM DTT, 1 mM EDTA, 5% glycerol, and 10 μM ZnSO₄). After 30 min incubation at room temperature, the reactions were subjected to electrophoresis on a 6% polyacrylamide gel at 180V for 1.5–2 h. Competitors (100×) were incubated for 30 min before the addition of the labeled probe. After separation, the gel was fixed, dried, and exposed for 2 h to a BIOMAX MS film at −80°C (Eastman Kodak Co.).

The amount of IgM in the supernatants was quantitated using a sandwich ELISA (39). In brief, 50 μl of a 10 μg/ml solution of goat anti–mouse IgM (Sigma Chemical Co.) was plated on an IMMULON^® 4 plate and the plate was incubated for 2 h at 37°C. Excessive antibody solution was removed and 200 μl of 5% BSA in PBS was added to each well and incubated overnight at 4°C. The plate was washed three times with 200 μl of 1% BSA/ PBS, and incubated for 2 h at 30°C with 200 μl of decreasing dilutions of culture supernatants that had been centrifuged at 1,000 g to remove remaining cell debris. A standard curve was determined with different concentrations of mouse anti-MOPC104e IgM (Sigma Chemical Co.). The plates were washed three times with 1% BSA/PBS and incubated for 2 h with 50 μl of 1:14,000 dilution of alkaline phosphatase–conjugated goat–anti mouse IgM (Sigma Chemical Co.). After three washes the plates were incubated with 50 μl of SIGMA 104^® phosphatase substrate (Sigma Chemical Co.) diluted in zinc-buffer (100 mM Glycine NaOH, pH 10.0, 1 mM ZnSO₄, 1 mM MgCl₂) until a visible color change was seen. The plates were analyzed using a 96-well plate reader and VMAX software according to the manufacturer's recommendation (Molecular Devices, Menlo Park, CA).

Cell cycle analysis was performed according to a modified protocol for PI staining of cell nuclei (40). In brief, cell pellets were resuspended in 100 μl 85.5 mg/ml sucrose, 11.7 mg/ml sodium citrate, and 5% DMSO, pH 7.6. The cells were trypsinized and permeabilized by adding 900 μl of 0.003% trypsin diluted in SS buffer (0.1% sodium citrate, 0.1% NP-40, 1.5 mM spermine-4HCl in 50 mM Tris base [pH 7.6]) for 20 min. The trypsin and cell RNA were inhibited and degraded by adding 0.05% trypsin inhibitor and 0.01% RNase diluted in SS buffer for 20 min at room temperature. Finally, the cells were stained with 0.01% PI and 3.3 mM spermine-4HCl in SS buffer for 20 min. The cells were then analyzed on a FACScan^® or an EPICS^® ELITE fluorescense-activated cell sorter (Coulter, Miami, FL) by adjusting the voltage of the instrument to distinguish between the different phases of the cell cycle and data were collected using the FACSDesk software. The collected data were further analyzed by ModFit^® software (Verity Software House, Inc., Topsham, ME) and the distribution of the cells between the different phases of the cell cycle and apoptosis percentages was adjusted to fit the software developer's suggestions.

The mature leukemic 3B3 B cell subline was derived from the leukemia-lymphoma BCL1 5B1b (41) and displays a lower spontaneous secretion level of IgM as compared with the parental line (42). Blackman et al. (21) showed that this low level of secretion is due to deficient J chain expression and suggested that these cells are partially activated due to the lack of a second signaling event for complete maturation. Although the different BCL1 sublines were cloned over a decade ago, the maturation differences between them have not been characterized extensively. When the 5B1b and 3B3 sublines were stained with an antiidiotypic mAb, 6A5.1. (43), both lines were strongly positive (data not shown), confirming that they were derivatives of the same clone. However, they differed in a number of morphological, biochemical, and developmental parameters. The 3B3 line exhibited a more regular and rounded morphology (Fig. 1,D) with a slow growth rate. Cell cycle distribution of the line was characteristic of relatively differentiated cells, with >50% cells in the G0/G1-phase (Fig. 1,b). The 5B1b cells, on the other hand, were larger and more dendritic (Fig. 1,A), and had a majority (54.4%) of cells in the S-phase of the cell cycle (Fig. 1 a).

In addition to these differences in morphology and cell cycle distribution, the expression of several cell surface markers also differed. 40–50% of the 5B1b line exhibited a relatively nonactivated phenotype (IgM^BrightIgD^BrightCD86^Dull) (Fig. 1, B and C), whereas the 3B3 line appeared to be fully activated (IgM^BrightIgD^DullCD86^Bright) (Fig. 1, E and F). This indicates that although both sublines originated from the same mature B cell clone, they differed from each other in their activation states.

Differentiation of the 3B3 cells into IgM secreting cells was characterized by their ability to secrete IgM and upregulate J chain expression (44). We followed the changes in the size and expression of several surface expression markers for 72 h after stimulation with IL-2/IL-5. Before cytokine treatment, the partially activated 5B1b cells were IA^{d Bright}CD19^BrightCD44^Bright Syndecan^Neg (references 10, 45; Fig. 2, B and C), with a cell size peaking at 240 afU (arbitrary fluorescence units) and cell granularity at 160 afU (Fig. 2,A). After 72 h of IL-2/IL-5 stimulation, the cell size increased slightly and peaked at 280 afU, and granularity increased to 185 afU (Fig. 2,D). This change was accompanied by a 100-fold downregulation of IA^d expression on the cell surface of ∼30% of the cells (Fig. 2,E). There were no noticeable changes in the cell surface expression of CD19, CD44, and Syndecan-1 (Fig. 2 F).

In contrast to these relatively minor changes in the 5B1b cells, the 3B3 line responded much more dramatically to IL-2/IL-5 stimulation. In particular, a small percentage of the resting 3B3 cells were IA^{d Dull}CD44^DullSyndecan1^Bright and expressed ∼50% less CD19 (Fig. 2, H and I) than 5B1b. After IL-2/IL-5 stimulation, this minor population gradually became the dominant phenotype (Fig. 2, K and L) concomitant with a substantial decrease in cell size from a peak at 220 afU to a peak at 100 afU (Fig. 2, G and J). These changes suggest that the 3B3 cells differentiated from an activated mature phase to a centrocyte-like cell after IL-2/IL-5 stimulation.

Lin et al. (17) demonstrated that Blimp-1 induces apoptosis in the 18-81 pre–B cell line. This effect was shown to correlate with the downregulation of c-myc expression by Blimp-1. Suppression of c-myc was also reported to induce apoptosis in the immature B cell line WEHI-231 (46). These data suggest that Blimp-1 may induce an apoptotic response at several stages of the B cell development. We therefore compared the effect of the transfection of Blimp-1 into the pre–B 18-81, immature WEHI-231, mature L10A, and the BAL17 cell lines. The results showed that by as early as 18 h after transfection, 10–80 μg of Blimp-1 induced 11–38% apoptosis in WEHI-231 in a dose-dependent manner (Fig. 3, a–d). A similar effect was also found in the L10A and the 18-81 cell lines (Fig. 3,e and reference 17). In BAL17, Blimp-1 transfection induced apoptosis but the effect was not dose dependent (Fig. 3,e). We also compared the long-term effect of the transfection of a low amount of Blimp-1 into 5B1b and 3B3 cells. Both sublines were transfected with 10 μg of Blimp-1 and 10 μg of GFP constructs, and selected for GFP expression by FACS^® after a 15 h incubation. Approximately 5–20% of the cells were positive for GFP, and these were cultured for an additional 72 h and then centrifuged onto slides for in situ TUNEL assay. Apoptosis as determined by the intense nuclear TUNEL staining and condensed nuclear morphology were clearly evident in Blimp-1–transfected 5B1b (Fig. 4,B) but not 3B3 (Fig. 4 A).

To identify the domain(s) in Blimp-1 that was involved in the induction of apoptosis, we examined the proapoptotic ability of four truncation mutants in WEHI-231 cells. In the ΔN′ truncation mutant, the N′ acidic domain (aa 71–87) and the first 49 aa of the PRD1-BF1-RIZ homology (PR) domain (aa 131–179) were removed. To compensate for the possible depletion of a necessary NLS, three SV40 NLS were introduced 5′ to the truncated ORF. The results show that this truncation had minimal effect on the ability of Blimp-1 to induce apoptosis in WEHI-231 cells (Fig. 4). On the other hand, the depletion of the C′ acidic domain (aa 752–810) had a modest effect on the ability of Blimp-1 to induce apoptosis in WEHI-231 cells (Fig. 5,B). The transfection of this ΔC′ truncation mutant reduced the apoptotic ability of Blimp-1 by ∼20%. The most significant effect on the ability of Blimp-1 to kill WEHI-231 cells was noticed when a portion of the proline-rich domain was removed (aa 381–450). The truncation of this domain reduced the apoptotic ability of Blimp-1 on WEHI-231 by 82%. A comparable loss of the proapoptotic activity was observed when most of the putative effector domains of Blimp-1 were removed, as in the NZ construct. The loss of the proapoptotic activity of NZ construct was not due to its inability to translocate into the nuclei because the fusion of the SV40 NLS to the NZ fragment enables proper nuclear translocation of the protein. In situ staining of transfectants with the M2 anti-FLAG antibodies showed that the protein clearly localized to the nucleus (Fig. 5,D). Furthermore, the deletion of a major portion of the proline-rich domain in the Δproline construct did not alter the DNA binding ability of the recombinant protein to the β-interferon (PRDI) or the c-myc (PRFI) elements (Fig. 5 C). A similar pattern was observed when most Blimp-1 effector domains were removed in the NZ truncation mutant (results not shown). These results indicate that the loss of the proapoptotic function in the Δproline and the NZ recombinant proteins was probably due to the inability of these truncated proteins to interact with other nuclear factors that are required for the induction of apoptosis.

To examine whether or not the NZ recombinant protein was able to block the induction of apoptosis in B cells and the development of Blimp-1–induced changes, such as IgM secretion and blasting, we cotransfected 5B1b cells with different combinations of Blimp-1 and/or the NZ construct and with the GFP expression vector as a selection marker. 24 h after transfection, the viable GFP⁺ 5B1b cells were sorted and incubated for an additional 48 h with or without IL-2/IL-5 stimulation. At the end of the cell culture, the level of IgM, percentage of blasting (cells gated on forward scatter between 160 and 250), and percentage of dying cells (determined by GFP⁺/PI⁺) were examined. The results showed that the transfection of NZ suppressed the low spontaneous cell death of 5B1b. The ectopic overexpression of Blimp-1 induced a robust apoptotic response. This response could be suppressed in cells cotransfected with the NZ and Blimp-1 expression vectors (Fig. 6,B). Although insufficient for driving the 5B1b cells to a fully differentiated phenotype (Fig. 2), IL-2/IL-5 treatment induced an increase in IgM secretion and in the percentage of blasts in these cells (Fig. 6,A). Ectopic expression of the NZ protein alone resulted in an approximately twofold reduction in both IgM secretion and the percentage of blasting cells in response to IL stimulation (Fig. 6,A). However, IL-2/IL-5 treatment of NZ transfected cells produced an apoptotic response that was almost sevenfold higher than control 5B1b cells. On the other hand, Blimp-1 transfection-induced apoptosis of these cells was slightly reduced by IL-2/IL-5 treatment (Fig. 6 B). These results suggest that the blasting and IgM secretion effects triggered by IL-2/IL-5 in 5B1b might be dependent on an adequate expression level of Blimp-1. Nevertheless, these effects did not seem to be connected with the apoptotic ability of Blimp-1. Therefore, it seems that in partially activated B cells, Blimp-1 may have a balanced effect between the induction of apoptosis and the induction of maturation. One possible explanation for this effect is that the contribution of Blimp-1 to either pathway requires the coexpression of stage-specific factors that would shift Blimp-1 function from induction of apoptosis to promotion of differentiation.

Blimp-1 was originally shown to regulate the expression of structural genes (J chain and Syndecan-1) during the differentiation of mature B lymphocytes into plasma cells (13). Recently, it has been found that the expression of Blimp-1 in pre–B cells triggers an apoptotic response that appears to be mediated by the downregulation of c-myc expression (17). Here we report that Blimp-1 can induce apoptosis across a broad range of immature and mature B cell lymphoma lines and can only promote differentiation in fully activated B cells. We also have identified the functional domain required for the induction of apoptosis and anticipate that Blimp-1 functions may depend upon the stage-specific expression of a cofactor.

The developmental stages represented by the 5B1b and 3B3 sublines appear to define a temporal window of B cell differentiation within which the function of Blimp-1 shifts. Partially activated 5B1b cells are impaired in their responses to IL-2/IL-5 stimulation and arrested at the centroblast stage. On the other hand, fully activated 3B3 cells respond to IL-2/IL-5 stimulation and differentiate into centroblast and, later, centrocyte-like cells (Fig. 2). The transfection of 5B1b cells with the NZ dominant negative suppresses blasting and IgM secretion induced by IL-2/IL-5 treatment (Fig. 6), thus indicating that Blimp-1 is an essential component in IL-2/IL-5 signaling during B cell maturation. In contrast to the antidifferentiation effect of NZ, overexpression of Blimp-1 in 5B1b but not in 3B3 cells results in marked apoptotic death that can be suppressed by the NZ dominant negative. This, in turn, indicates that the Blimp-1–mediated apoptotic machinery is still functional in the 5B1b cells. Similarly, the transfection of Blimp-1 in earlier stages of the B cell development, such as the pre–B (18-81), immature B (WEHI-231) and mature B (L10A) cell stages induces an apoptotic response. This effect is mainly dependent on a 69-aa region within the proline-rich domain (Fig. 5). It is conceivable that the binding of an additional factor to this domain is required to elicit a response. For example, Blimp-1 may interact in vitro with YY1, a positive regulator of c-myc (17), whose expression plays an important role in controlling the cell fate. Since the downregulation of c-myc was shown to be involved in the induction of apoptosis in both 18-81 (17) and WEHI-231 (46), its suppression by Blimp-1 is anticipated to play a role in the elimination of partially activated B cells. However, it appears that other genes that are involved in the regulation of the cell cycle may be targeted by Blimp-1.

The PR domain (aa 131-231) (16) was suggested to modulate the repressor ability of RIZ (47). Our results show that the deletion of 50% of this domain in the ΔN′ truncated protein had no significant effect on the ability of this protein to induce apoptosis in WEHI-231 cells. Therefore, the integrity of the PR domain may be unnecessary for the induction of apoptosis. In fact, the truncation of either the N′ or C′ acidic domains had only a minor effect on the apoptotic ability of Blimp-1 (Fig. 5 B).

Given that Blimp-1 is able to exert such different effects in closely related cells, the question arises as to how this is accomplished. One possible explanation is that different combinations of the zinc finger motifs may recognize distinct gene targets. A mechanism of this type was recently shown in the case of NeP1/CTCF, an 11–zinc finger protein that uses different combination of its zinc fingers to bind to either the c-myc gene or the lysozyme silencer, F1 (48). The partial use of Blimp-1 zinc fingers for the recognition of potential targets was demonstrated in the work of Keller and Maniatis (32). In that report the authors show that two out of the five PRDI-BF1 zinc fingers (and a stretch of 65 aa located N′ to these fingers) are sufficient for its binding to the PRDI element. Recently, we have found that the binding of NZ to the PRF1 element, but not to the PRD1 element, is sensitive to posttranslational modifications or conformational changes (Chi, J.-T., and E.J. Messika, unpublished results). Thus, Blimp-1 may be modulated to switch specificity to different targets due to interactions with stage-specific proteins, using different combinations of its zinc fingers, as the cell matures.

In summary, we propose that Blimp-1 may be part of a “gate keeper” mechanism in the differentiation process, as its upregulation causes apoptosis of immature or partially stimulated B cells. This may constitute an important selection mechanism to eliminate self-reactive B cells that have escaped earlier negative selection mechanisms or have been inappropriately (and partially) activated.

We thank Dr. Samuel Strober for the BCL1 5B1b line; Drs. Chan Beals, Stephen Ho, and Jerry Crabtree for the pyDF30 vector and SV40 NLS; Dr. Edward Mocarski for the pG310 vector; Dr. Stanley Falkow for the GFP construct; Dr. Ellen Vitetta for the anti-BCL1 idiotypic antibodies; and Dr. Ziv Reich for critical reading of the manuscript.

This work was supported by a grant from the National Institutes of Health (AI-19512) and by the Howard Hughes Medical Institute.

aa

amino acid(s)

afU

arbitrary fluorescence units

Blimp

B lymphocyte–induced maturation protein

FLAG

flu antigen

GFP

green fluorescent protein

NLS

nuclear localization signal(s)

ORF

open reading frame

PI

propidium iodide

PR

PRD1-BF1-RIZ homology

TUNEL

TdT-mediated dUTP nick end labeling