Cleavage of transcription factor SP1 by caspases during anti-IgM-induced B-cell apoptosis

Authors


K. Bommert, Max Delbrück Center for Molecular Medicine, Robert Rössle Str. 10, D-13092 Berlin-Buch, Germany. Fax: +49 30 9406 3124, Tel.: +49 30 9406 3817, E-mail: bommert@mdc-berlin.de

Abstract

Apoptosis is instrumental in the processes generating the diversity of the B-cell repertoire. Autoreactive B-cells are eliminated by anti-IgM crosslinking after encountering self-antigens, but precise mechanisms leading to B-cell apoptosis are still not well understood. We report here the cleavage of the transcription factor SP1 in the human Burkitt lymphoma cell line BL60 during anti-IgM-induced apoptosis. Western blot analysis revealed two cleavage products of approximately 68 kDa and 45 kDa after induction of apoptosis. Cleavage could be completely inhibited by zDEVD-fmk, an inhibitor specific for caspase 3-like proteases. In-vitro cleavage of recombinant SP1 by recombinant caspase 3 (CPP32) or caspase 7 (Mch 3) results in similar cleavage products as those observed in vivo. Recombinant caspase 6 (Mch 2) primarily generates a 68-kDa cleavage product, as observed after calcium ionophore (CaI) induced B-cell apoptosis. In contrast, caspase 1 (ICE) did not cleave SP1 in vitro. The time course of SP1 cleavage during anti-IgM-induced apoptosis is paralleled by an increase of caspase activity measured by DEVD-p-nitroanilide (DEVD-pNA) cleavage. DNA band-shift assays revealed a decrease in the intensity of the full length SP1/DNA complex and an increase in the intensity of a smaller complex due to the binding of one SP1 cleavage product. By Edman sequencing we could identify a caspase 3 cleavage site after Asp584 (D584AQPQAGR), generating a 22-kDa C-terminal SP1 protein fragment which still contains the DNA binding site. Our results show the cleavage of the human transcription factor SP1 in vivo and in vitro, underlining the central role of caspase 3-like proteases during the process of anti-IgM-induced apoptosis.

Abbreviations
CaI

calcium ionophore

DEVD-pNA

DEVD-P-nitroanilide

CAD

caspase activated DNase

ICAD

inhibitor of caspase activated DNase

FITC

fluoresceine isothiocyanate

DNA-PK

DNA-dependent protein kinase

Apoptosis is a fundamental process in lymphocyte development [1,2]. Immune tolerance can be achieved by apoptosis or receptor editing. Autoreactive B-cells are eliminated by clonal deletion via surface IgM signaling after encountering self-antigen [3–9]. It is still not completely understood how the death signal is transmitted from the B-cell receptor into the cell, finally leading to cell death. It was recently shown that Igα and Igβ, part of the B-cell receptor complex, are both necessary for B-cell apoptosis [10].

Some proteins of the bcl-2 family are shown to be involved in anti-IgM-induced B-cell apoptosis. We have already shown that bax-α expression is upregulated in BL41, a Burkitt lymphoma B-cell line, after anti-IgM-induced apoptosis and overexpression of bax-α leads to increased sensitivity towards anti-IgM [11,12]. Overexpression of bcl-xL rescues WEHI-231 cells, an immature mouse B-cell line, from anti-IgM-induced cell death [13].

Recent publications have reviewed the discovery of a growing family of proteases involved in the process of apoptosis. Identification of these cysteine-dependent aspartate-specific proteases (caspases) is a major concern in apoptosis research at the present time, not least because mechanisms regulating apoptosis might provide new targets in cancer therapies [14–20]. It is hypothesized that caspases are activated in a cell-specific manner depending on the stimulus, and are thought to act via a cascade effect. This has been demonstrated in vitro by certain caspases activating others. Thus, pathways exist to transmit signals via sequential caspase activation [21]. Caspase-mediated proteolysis results in a small number of cuts. The identification and evaluation of caspase substrates and the extent of their action in vivo is of central importance in understanding the apoptotic process and its regulation [22,23]. It is still unclear whether the caspases themselves are the real ‘executioners’ during the cell death program, or rather the target proteins which are irreversibly cleaved by these caspases. The cleavage can result in activation or inactivation of the protein, but never in degradation, reflecting the importance of their action in the well organized process of apoptosis. A number of caspase substrates have been identified, including the caspases themselves, protein kinase Cδ (PKCδ), p21-activated kinase 2 (PAK2), PITSLRE kinase or the recently identified ICAD, an inhibitor of the caspase activated DNase (CAD) [24–28]. Cleavage of ICAD leads to the release of CAD, responsible for DNA fragmentation, which is characteristic of most cells undergoing apoptosis. Other substrates, e.g. the RB tumor suppressor gene product, become inactivated as a result of caspase cleavage [29]. Structural proteins, like lamin-A and -B or Gas 2, alter their assembly/disassembly properties after cleavage, suggesting a potential role in the morphological changes during the process of apoptosis [30,31].

It has already been demonstrated that caspase 3 plays an important role in B-cell apoptosis. D4-GDI, an abundant hematopoietic cell GDP dissociation inhibitor for the Ras-related Rho family GTPases, is cleaved by caspase 3 during anti-IgM-induced B-cell apoptosis. Cleavage as well as apoptosis can be blocked by zDEVD-fmk, a specific inhibitor for caspase 3-like proteases [32].

We report here the cleavage of the transcription factor SP1 in the human Burkitt lymphoma cell line BL60-2 after induction of apoptosis by anti-IgM crosslinking, UV-irradiation and action of calcium ionophore (CaI). Cleavage occurs by caspase 3, 6 or 7, demonstrated in vitro and by inhibitor studies. We identified a new caspase 3 cleavage site, generating a C-terminal 22-kDa cleavage product. DNA band-shift assays show the decreased binding activity of the full-length protein and increased binding of at least one cleavage product. These results further emphasize the central role for caspase 3-like proteases in the process of anti-IgM-induced apoptosis.

Materials and methods

Cell culture

The human Burkitt lymphoma cell line BL60-2 was cultured in RPMI-1640, 10% heat-inactivated fetal calf serum, penicillin-streptomycin, 1 mm sodium pyruvate, 2 mm glutamine, 10 mm Hepes pH 7.4, 20 nm bathocuproine disulfonic acid, 50 µmα-thioglycerol at 37 °C and 5% CO2 as described previously [32]. Apoptosis was induced by adding 1.3 µg·mL−1 anti-IgM F(ab)2 (Dianova, Hamburg, Germany) or 1 µg·mL−1 CaI (Sigma) to the cell culture medium. UV-irradiation was performed in a 10-cm culture dish containing 25 mL RPMI medium (240 mJ·cm−2, Stratalinker 2400). The rate of apoptosis was measured by acridine orange staining or AnnexinV-FITC (fluoresceine isothiocyanate) labeling as described previously [32]. Inhibition of apoptosis was performed by incubating the cells with 0–200 µm zDEVD-fmk 30 min prior to addition of anti-IgM F(ab)2.

Immunoblotting

BL60-2 cells from different time points after incubation with anti-IgM were centrifuged for 10 min at 1000 g and washed in 1 × NaCl/Pi. Cell pellets were lysed (20 mm Tris/acetate pH 7.0, 10 mm sodium glycerophosphate, 50 mm sodium fluoride, 5 mm sodium pyrophosphate, 1% Triton-X 100, 0.1 mm EDTA, 1 mm EGTA, 0.2 mm phenylmethanesulfonyl fluoride, 0.27 mm sucrose, 2 µg·mL−1 leupeptin) for 40 min at 4 °C. Proteins (30 µg·lane−1) were electrophoretically separated using 10% SDS/PAGE. After transfer to nitrocellulose and blocking, filters were incubated with 1 : 1000 diluted anti-SP1 (Santa Cruz) or anti-caspase 3 (Transduction Laboratories). Thereafter, filters were incubated with HRP-conjugated goat anti-rabbit antiserum (1 : 15 000, Santa Cruz). Membranes were developed using the ECL system (Amersham).

Measurement of caspase activity

1 × 106 BL60-2 cells from various time points after stimulation were harvested and lysed in lysis buffer according to the Apo Alert protocol (Clontech). Cell lysates were incubated with the substrate DEVD-p-nitroanilide (DEVD-pNA) for 50 min at 37 °C. Cleavage of the chromophore pNA was measured at OD400 using a spectrophotometer.

In vitro cleavage-assay

Cleavage reactions were performed in a reaction volume of 20 µL, containing 0.1% Chaps, 50 mm Hepes pH 7.5, 1 mm phenylmethanesulfonyl fluoride, 50 mm Leupeptin, 20 µg Aprotinin, 5 mm dithiothrietol, 60 ng recombinant caspase 1, 3, 6 or 7, 1 µL recombinant SP1 (1 footprint unit, Promega), with or without 200 µm zDEVD-fmk (Calbiochem). Reactions were incubated at 37 °C for 2 h and subsequent immunoblotting was performed as described above. Bacterial expression and purification of caspases was performed as described [24]. Caspase 6 was obtained from Pharmingen. Recombinant SP1 (450 µL) was concentrated on spin columns (Eppendorf) and incubated with 300 ng recombinant caspase 3 (Pharmingen) for 1.5 h at 37 °C in reaction buffer as described above. Reaction mixture was separated on 10% SDS/PAGE, blotted on a poly(vinylidene difluoride) (PVDF) membrane, bands were visualized by Coomassie staining. Edman Sequencing was performed as described previously [33].

DNA binding assay

Nuclear extracts were prepared at different timepoints after stimulation with anti-IgM F(ab)2 according to the protocol of Schreiber et al.[34]. Oligonucleotides resembling the SP1 recognition sequence and a mutant SP1 recognition sequence (Santa Cruz), or Oct-2 recognition sequence were end-labeled with [α-32P]dATP using T4 polynucleotide kinase. Five micrograms of nuclear extract were incubated in 20 mm Hepes pH 7.9, 60 mm KCl, 4% Ficoll, 1 µg poly(dIdC), 2 mm dithiothreitol, 2 µg BSA in a 20-µL reaction volume. Radioactive labeled oligonucleotide was added (50 000 c.p.m. per reaction) and samples were incubated for 30 min at 30 °C. In the supershift assay, 0.5 µg of the antibody (Santa Cruz) was added to the reaction mixture 15 min after incubation. Peptide competition was performed using 0.1 µg PEP2 (Santa Cruz). The peptide was added to the reaction buffer containing the antibody 30 min prior to addition of the nuclear extract. DNA/protein complexes were resolved by electrophoresis on 4% polyacrylamide gels, containing 50 mm Tris, 50 mm boric acid and 0.1 mm EDTA.

Results and discussion

We report here the cleavage of the transcription factor SP1 during anti-IgM-induced apoptosis in the human Burkitt lymphoma cell line BL60. Cells from a highly sensitive subclone of BL60 were incubated with anti-IgM F(ab)2 fragments for the indicated time intervals. SP1 expression and cleavage was examined by Western blotting using an SP1-specific polyclonal antiserum recognizing an epitope corresponding to amino acids 436–454 (Fig. 1). Two cleavage products could be detected. The first cleavage product of 68 kDa appears after 8 h and a second cleavage product of 45 kDa appears after 16 h anti-IgM treatment. The full-length protein, of 95/106 kDa, is almost entirely cleaved 24 h after induction of apoptosis. Both cleavage products were detected after UV-induced apoptosis, whereas CaI induced apoptosis generated mainly the 68-kDa cleavage product (Fig. 1).

Figure 1.

Figure 1.

Timecourse of SP1 cleavage in BL60-2 cells. BL60-2 cells were treated with anti-IgM F(ab)2, Calcium ionophore or UV irradiation for various time intervals. Cells were harvested and lysates from total cells were analyzed by Western blotting using a SP1-specific polyclonal antiserum (Santa Cruz). Arrows on the right designate the 95/106 kDa SP1 full length protein and the 68- and 45-kDa cleavage products appearing after induction of apoptosis.

Cleavage of SP1 could be blocked by zDEVD-fmk in a dose-dependent manner (Fig. 2). Cells were incubated with different concentrations of the irreversible caspase 3-like inhibitor zDEVD-fmk prior to stimulation with anti-IgM. After 24 h, cells were lysed and protein extracts were separated by SDS/PAGE. Western blotting revealed that SP1 cleavage at both sites is almost completely inhibited at a concentration of 200 µm zDEVD-fmk, as shown in Fig. 2. In contrast acYVAD-cmk, a specific irreversible inhibitor for caspase 1-like proteases, could not block the cleavage even at a concentration of 200 µm (data not shown). These findings strongly suggest that caspase 3, or another zDEVD-fmk sensitive caspase, may be responsible for SP1 cleavage during anti-IgM-induced apoptosis.

Figure 2.

Figure 2.

Dose-dependent inhibition of SP1 cleavage by zDEVD-fmk. BL602 cells were incubated with 0–200 µm zDEVD-fmk an inhibitor for caspase 3-like proteases prior to incubation with anti-IgM F(ab)2. SP1 cleavage was determined by Western blotting using an anti-SP1-specific antiserum. Arrowheads indicate the 95/106 kDa full length SP1 protein and the 68- and 45-kDa cleavage products.

The growing family of cysteine proteases comprises 11 different members in the human system [14–18]. Caspases are synthesized as inactive precursor molecules, and are converted by proteolytic cleavage in cells undergoing apoptosis to active heterodimers, composed of two chains of ≈20 kDa and 10 kDa. The active caspase contains two heterodimers and two catalytic sites [35]. Caspase 1-like proteases seem to play a major role in inflammatory processes, whereas caspase 3-like proteases seem to be important for the process of apoptosis [36]. Caspase 3 and caspase 6 are considered the major active caspases in apoptosis. Caspase 3 may be activated by upstream caspases, demonstrated by in-vitro assays and the discovery of proteins of the complex death transduction pathway in cells induced to die via the FAS and TNF receptors [37].

The signaling after B-cell receptor crosslinking and the mechanisms finally leading to cell death are currently not well understood. To date, caspase 3 is the only caspase shown to be involved in anti-IgM-induced apoptosis, and nothing is known about the upstream activators involved. Caspase 6 and caspase 7 are described as caspases which act downstream of caspase 3, and both can be cleaved and activated by caspase 3 in vitro[21]. This suggests that either caspase 3 or a downstream caspase, such as caspase 6 or 7, may be responsible for the cleavage of the transcription factor SP1 during anti-IgM-induced B-cell apoptosis. We could not identify a specific caspase 3 consensus cleavage site (DXXD) in the SP1 amino acid sequence To analyze which caspase is able to cleave SP1 we performed an in-vitro assay. Recombinant SP1 (Promega) was incubated with recombinant caspase 1, 3, 6 or 7, with or without prior incubation with zDEVD-fmk. Western blotting revealed the cleavage of SP1 by recombinant caspase 3 and 7, resulting in two distinct fragments of approximately 68 kDa and 45 kDa. In vitro cleavage by recombinant caspase 6 resulted in mainly one 68 kDa fragment. In contrast, caspase 1 did not cleave SP1 in vitro(Fig. 3). In-vitro cleavage by caspase 3, 6 and 7 could be inhibited by zDEVD-fmk. Our results show that cleavage of the transcription factor SP1 by caspase 3 or 7 results in the same cleavage products as observed in vivo after anti-IgM-induced apoptosis, but with different efficiency. Caspase 3 cleaves SP1 more efficiently than caspase 7. Cleavage by caspase 6 leads to just one 68-kDa fragment, correlating with the SP1 cleavage observed after UV-irradiation. These findings lead us to conclude that the transcription factor SP1 is cleaved by caspase 3 or 7 in Burkitt lymphoma cells after anti-IgM-induced apoptosis.

Figure 3.

Figure 3.

In-vitrocleavage of recombinant SP1. Recombinant SP1 protein (2 footprint units, Promega) was incubated with purified caspase 1, 3, 6 or 7 (60 ng) for 2 h with or without caspase 3 inhibitor zDEVD-fmk. Samples were separated on 10% SDS/PAGE and Western blotting was performed as described in Fig. 1. Arrowheads indicate the 95/106 kDa full length SP1 protein and the 68- and 45-kDa cleavage products.

Caspase activity was measured in cell extracts by cleavage of the substrate DEVD-pNA leading to the release of the chromophore pNA, which is detectable at OD400. This assay revealed an increase of caspase activity during anti-IgM stimulation with a maximum activity after 16 h. Increased caspase 3 activity is due to proteolytic cleavage of the inactive caspase 3 precursor (32 kDa) into the active enzyme consisting of two subunits of 17 kDa and 12 kDa. This process is a sequential two step process, with a 24-kDa intermediate. The 24-kDa and 17-kDa fragments were recognized by the antibody in a Western blot analysis (Fig. 4b). The timecourse of caspase activity parallels that of SP1 cleavage into a 68-kDa fragment after 8 h and a 45-kDa fragment after 16 h of anti-IgM induction (Fig. 1 and Fig. 4).

Figure 4.

Figure 4.

Caspase activity is increased after anti-IgM-induced apoptosis. (A) Caspase activity was measured by cleavage of the substrate DEVD-pNA. Cells from different time points of incubation with anti-IgM F(ab)2 were lysed and incubated with DEVD-pNA. The released chromophore pNA was measured at OD400. Cell lysates incubated with zDEVD-fmk prior to incubation with DEVD-pNA show strong reduction of caspase 3 activity as marked by an arrow. (B) Western blot analysis indicating the cleavage of caspase 3 during anti-IgM induced apoptosis. BL60-2 cells were treated with anti-IgM F(ab)2 for various time intervals, cells were harvested and lysates from total cells were analyzed with a polyclonal antibody. Cleavage of caspase 3 generates the p24 intermediate autocatalytically cleaved to p17.

It was recently demonstrated by Piedrafita et al. [38] that SP1 is cleaved into a 68-kDa fragment after retinoic acid-induced apoptosis in human T cells. Cleavage could be blocked by zVADfmk, a broad-spectrum inhibitor for caspases, as well as by zDEVD-fmk. Our in-vivo and in-vitro results show that SP1 is cleaved into a 68-kDa and a 45-kDa fragment by caspase 3 and caspase 7 or mainly the 68-kDa fragment by caspase 6. Comparison of these results and our own show that, depending on the cell type and stimulus, cleavage of SP1 results in either one 68-kDa fragment via caspase 6 or two fragments of 68 and 45 kDa via caspase 3 or 7.

In order to investigate if cleavage of the transcription factor SP1 has functional implications during the process of anti-IgM-induced apoptosis, we performed DNA band-shift assays using a SP1-specific binding site oligonucleotide. Three SP1-specific complexes were observed, reflecting the binding of the 95/106 kDa-full-length protein to the SP1-specific binding site. Signal intensity of these SP1/DNA complexes decreases with time after induction of apoptosis (Fig. 5, lanes 1–3). This decrease is confirmed in the supershift assay using a SP1-specific antibody (Fig. 5, lanes 4–6). Specificity was demonstrated by competition of the antibody binding using the corresponding peptide or a mutant SP1 oligonucleotide (Fig. 5, lanes 7–12). A smaller complex appears in the band-shift assay after 10 h of anti-IgM treatment. This complex increases in intensity after 24 h. We suggest that this smaller DNA/protein complex resembles one of the cleavage products binding to the SP1-specific oligonucleotide. DNA band-shift assays, using the specific binding sites for the ubiquitous transcription factor Oct-2, showed equal levels of the DNA/protein complex after induction of apoptosis (Fig. 5, lanes 13–15).

Figure 5.

Figure 5.

SP1 binding activity is decreased after anti-IgM-induced apoptosis. DNA band-shift assays were performed to analyze nuclear extracts prepared from BL60–2 cells after 0, 10 and 24 h after incubation with anti-IgM F(ab)2. Five micrograms of the nuclear extracts were incubated with a 32P-labeled SP1 DNA binding site oligonucleotide (lane 1–9). Three specific SP1 protein/DNA complexes were detected (lefthand bracket) with decreasing intensity 10 and 24 h after anti-IgM-induction. A smaller complex is formed which corresponds to one of the cleavage products binding to the SP1 binding site after 10 h of stimulation increasing after 24 h (lane 1–3). The decrease of the SP1/DNA complex was also shown by supershift analysis (SP1* in lane 4–6). The antibody/protein-DNA complex can be competed by a specific peptide (lane 7–9). Specificity of the SP1/DNA complexes was demonstrated using mutant oligonucleotide (lane 10–12). Incubation with an Oct-2-specific oligonucleotide shows two complexes (righthand bracket) with equal amounts in each lane after different time points anti-IgM stimulation (lane 13–15).

SP1 is an abundant nuclear protein in most cells, but the level of expression changes during development and varies in different cell types [39]. SP1 is shown to bind to many promoters of tissue-specific and ubiquitous genes, containing the methyl-CpG-binding sequence and binds to DNA by interaction of contiguous Zn(II)finger motifs [40]. N-terminal deletion assays have shown that the last 168 amino acids of the C-terminus, containing the Zn finger domains are sufficient for sequence-specific DNA binding [40]. The last 327 amino acid residues from the C-terminus appear to be important for transcriptional activity in vitro, whereas the last 168 amino acid residues are still able to activate transcription but less efficient [41]. Recent results showing the SP1 activity is in addition regulated by phosphorylation at serine residues. Double-stranded DNA-dependent protein kinase (DNA-PK) has been identified as an SP1 kinase. Interestingly DNA-PK itself is a substrate for caspase 3 [42]. Transactivating activity of SP1 is increased after phoshorylation, therefore inactivation of DNA-PK might lead to reduced SP1 activity [43,44].

To analyse the caspase 3-cleavage sites, recombinant SP1 protein was cleaved in vitro by caspase 3. Protein fragments were separated on 10% SDS/PAGE, and then blotted onto a PVDF membrane. After subsequent staining with Coomassie blue we detected a band with the apparent relative molecular mass of 22 kDa which was not detectable in the Western blot analysis. Edman sequencing of the fragment revealed a new caspase 3 cleavage site after Asp583, with the sequence ENSPD(583)AQPQAGR. This fragment contains only the DNA binding domain of SP1 and lacks the epitope of the SP1 antibody.

These results lead us to the hypothesis that SP1 cleavage during the process of anti-IgM-induced B-cell apoptosis might lead to a nonfunctional transcription factor fragment which is still able to bind to DNA but is no longer able to activate transcription. It has already been shown in deletion analysis of SP1 that a similar fragment still contains the DNA binding activity but has strongly reduced transactivation activity [41]. Therefore, we hypothezise that SP1 cleavage by recombinant caspase 3 might generate a dominant negative SP1 transcription factor fragment.

Our findings, along with recently published results, demonstrate that caspase 3 is one of the main ‘executioners’ in anti-IgM-induced apoptosis resulting in the cleavage of several cellular proteins with different functions, such as D4-GDI and the transcription factor SP1 [32]. Cleavage of the transcription factor SP1 represents a new substrate for caspase 3-family proteases. Investigation of the functional relevance of the 22-kDa cleavage product in SP1-dependent gene transcription and apoptosis has yet to be performed, as has exploration of whether overexpression of the 22-kDa cleavage product leads to apoptosis by acting in a dominant-negative manner, or if uncleavable SP1 mutant protein affects the process of programmed cell death.

Conclusions

We have shown by Western blot analysis that the human transcription factor SP1 is cleaved into a 68-kDa and a 45-kDa fragment during anti-IgM-induced B-cell apoptosis. Cleavage does not occur if apoptosis is inhibited by zDEVD-fmk, an inhibitor specific for caspase 3-like proteases, but still occurs in the presence of ac-YVAD-cmk, an inhibitor which is specific for caspase 1-like proteases. The timecourse of SP1 cleavage during anti-IgM-induced apoptosis is paralleled with the increase in caspase activity as measured by DEVD-pNA cleavage and Western blot analysis. In vitro cleavage of recombinant SP1 with recombinant caspase 3 and caspase 7 leads to the same cleavage products as those observed in vivo, whereas caspase 7 cleavage results in just one 68-kDa cleavage product as observed after CaI induced apoptosis leading to the conclusion, that different apoptotic stimuli lead to different cleavage of the same protein. DNA band-shift assays revealed a reduced formation of the full-length SP1 protein/DNA complex, and an increased formation of a smaller complex, which is due to the DNA binding of a SP1 cleavage product. We could also identify a new caspase 3 cleavage-site motif at position D538, generating a 22-kDa C-terminal SP1 fragment still containing the DNA binding site but with reduced transactivation activity.

Acknowledgements

We like to thank Mrs. Ute Nitschke for excellent technical assistance. This work was supported by the Deutsche Forschungsgemeinschaft (DFG-Ma 1664/1-2) as well as by an EC Biomed Program Grant (BMH4-CT96-0300). R.B. and P.V. are postdoctoral researchers with the FWO-Vlaanderen.

Ancillary