Identification of Apoptosis-associated Proteins in a Human Burkitt Lymphoma Cell Line

Apoptosis or programmed cell death is essential in the process of controlling lymphocyte growth and selection. We identified proteins that are involved in anti-IgM antibody-mediated apoptosis using a subclone of the human Burkitt lymphoma cell line BL60. Apoptosis-associated proteins were detected by high resolution two-dimensional gel electrophoresis on a micropreparative scale. Comparison of the high resolution two-dimensional gel electrophoresis protein patterns from apoptotic and non-apoptotic cells showed differences in ∼80 spots including protein modifications. Analysis of the predominantly altered proteins was performed by internal Edman microsequencing and/or by peptide mass fingerprinting using matrix-assisted laser desorption/ionization mass spectrometry. Analysis was significantly improved by using new micropreparative high resolution two-dimensional gels employing high protein concentrations. The following 12 apoptosis-associated proteins were identified: heterogeneous nuclear ribonucleoprotein (hnRNP) A1, hnRNP C1/C2, FUSE-binding protein, dUTPase, lymphocyte-specific protein LSP1, UV excision repair protein RAD23 homologue B (HHR23B), 60 S acidic ribosomal protein P0 (L10E), heterochromatin protein 1 homologue α (HP1α), nucleolin, lamin, neutral calponin, and actin. Fragmentation of actin, hnRNP A1, hnRNP C1/C2, 60 S acidic ribosomal protein P0, lamin, and nucleolin could be inhibited by benzyloxycarbonyl-Asp(OMe)-Glu(OMe)-Val-Asp(OMe)-fluoromethyl ketone, a selective irreversible inhibitor of CPP32 (caspase 3).

volved in apoptosis by conventional biochemical and molecular approaches at the mRNA level. However, the process of apoptosis can be initiated by a variety of stimuli and results in defined morphological and biochemical changes (3) that may be easier studied at the protein level. Apoptosis is characterized by cellular and nuclear shrinkage, cytoplasmic blebbing, condensation of nuclear chromatin, and fragmentation of nuclear DNA (3,4). Analysis of the apoptotic pathway in lymphocytes revealed several molecules as key controllers in the execution of apoptosis (5). Anti-IgM antibody-mediated apoptosis is thought to be controlled on at least two levels by the members of the bcl-2 gene family and the interleukin-1␤-converting enzyme/Ced3-like cysteine proteases now called caspases. A large number of caspases have been described (6 -8), but the individual roles of most intracellular proteases and their substrates remain to be elucidated. A number of proteins have been shown to be involved and specifically cleaved during apoptosis (16). Death substrates that are cleaved by caspases include poly(ADP-ribose) polymerase (9), DNA-dependent protein kinase (10), protein kinase C␦ (11), hnRNP 1 C1/C2 (12), lamin (13), nucleolin (14), and actin (15). We have already shown that caspase 3 plays a crucial role in anti-IgM antibody-induced B cell apoptosis. Cleavage of the Rho GDP dissociation inhibitor D4-GDI as well as apoptosis can be completely inhibited by Z-DEVD-FMK, a specific inhibitor of caspase 3-like proteases (16).
In this study, we provide the first evidence that the protein hnRNP A1 is specifically cleaved into three distinct fragments of the apparent sizes 32, 29, and 16 kDa, respectively, during anti-IgM antibody-induced apoptosis visualized by Western blot analysis. Cleavage can be completely inhibited by Z-DEVD-FMK, leading to the conclusion that hnRNP A1 is specifically cleaved by caspase 3-like proteases. Different proteins like HP1␣ (heterochromatin protein 1 homologue ␣), FBP, dUTPase, HHR23B (UV excision repair protein RAD23 homologue B), LSP1 (lymphocyte-specific protein 1), ribosomal protein P0, and neutral calponin seem also to be involved in anti-IgM antibody-induced apoptosis since alterations of these proteins could also be inhibited by Z-DEVD-FMK. This was shown by high resolution two-dimensional gel electrophoresis analysis (2-DE) of protein extracts from cells, which prior to anti-IgM antibody stimulation were treated with Z-DEVD-FMK. A new micropreparative approach for the two-dimensional gel separation of the complex protein mixture is described that facilitated the analysis of these proteins.

EXPERIMENTAL PROCEDURES
Cell Culture-The human Burkitt lymphoma cell line BL60-2 was cultured in RPMI 1640 medium, 10% heat-inactivated fetal calf serum, penicillin/streptomycin, 1 mM sodium pyruvate, 2 mM glutamine, 10 mM HEPES (pH 7.4), 20 nM bathocuproinedisulfonic acid, and 50 M ␣-thioglycerol at 37°C and 5% CO 2 as described previously (16). Apoptosis was induced by adding 1.3 g/ml anti-IgM F(ab) 2 (Dianova, Hamburg, Germany) to the cell culture medium. The rate of apoptosis was measured by acridine orange staining or FITC-conjugated annexin V (Bender Medical System, Vienna, Austria) labeling as described previously. After 24 h of anti-IgM antibody treatment, the separation of apoptotic cells from non-apoptotic cells by FITC-conjugated annexin labeling and subsequent magnetic separation with anti-FITC magnetic microbeads (Miltenyi Biotec, Gladbach, Germany) were performed. Inhibition of apoptosis was performed by incubating the cells with 200 M Z-DEVD-FMK (Calbiochem-Novabiochem, Bad Soden/Taunus, Germany) 30 min prior to addition of anti-IgM F(ab) 2 essentially as described previously (16).
Preparation of Protein Samples-The cell pellets were thawed and, during this procedure, rapidly mixed with protease inhibitor solutions (17). The final concentrations of protease inhibitors in the protein sample were 1.4 M pepstatin A, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 2.1 M leupeptin, 1 mM EDTA, 1 mM KCl, and 20 mM Tris-HCl (pH 7.1) according to Klose and Kobalz (18). Additionally, the pellet was rapidly mixed with urea (9 M final concentration), then with dithiothreitol (70 mM final concentration), and finally with 2% carrier ampholytes for analytical gels or 6% carrier ampholytes for micropreparative gels (Servalyt pH 2-4, Serva, Heidelberg, Germany). After 30 min of gentle stirring at room temperature, the samples were centrifuged in an ultracentrifuge at 100,000 ϫ g for 25 min. The clear supernatant was frozen at Ϫ70°C. The resulting protein concentration was determined by amino acid analysis (19,20) and was 30 -35 mg/ml. 1-2 l of these samples were loaded at the anodic side on an analytical gel, and 100 -120 l were applied on a micropreparative gel.
High Resolution Two-dimensional Gel Electrophoresis-The proteins were separated by the large gel 2-DE technique (18). The complete 2-DE equipment and ready-to-use gel solutions were purchased from Wittmann Institute of Technology and Analysis of Biomolecules GmbH (Teltow, Germany). The 20-cm isoelectric focusing rod gels (inner diameter of analytical and micropreparative gels: 0.09 and 0.25 cm, respectively) contained 3.5% acrylamide, 0.3% piperazine diacrylamide (Bio-Rad, Munich, Germany), and a total of 4% (w/v) carrier ampholytes (pH 2-11) (WITA GmbH). The electrophoretic conditions of the analytical gels during the isoelectric focusing were according to Klose and Kobalz (18). The electrophoretic conditions of the micropreparative gels during isoelectric focusing were as follows: 1 h at 100 V, 1 h at 250 V, 17.5 h at 570 V, 1 h at 1000 V, 30 min at 1500 V, and 10 min at 2000 V. After focusing, the rod gels were equilibrated for 10 min in a buffer containing 125 mM Tris phosphate (pH 6.9), 40% glycerol, 70 mM dithiothreitol, and 3% SDS. The equilibrated rod gels were frozen at Ϫ70°C. After thawing, the isoelectric focusing rod gels were immediately applied to an SDS-polyacrylamide gel that contained 15% (w/v) acrylamide and 0.2% N,NЈ-methylenebisacrylamide. The SDS-polyacrylamide gel electrophoresis system of Laemmli (21) was used, but the stacking gel was replaced by an equilibrated isoelectric focusing gel. Electrophoresis of the analytical SDS-polyacrylamide gel (30 ϫ 23 ϫ 0.075 cm) was performed using a two-step increase of current, starting with 15 min at 65 mA, followed by a run of 7-8 h at 85 mA, until the dye reached the end of the gel. Electrophoresis of the micropreparative SDS-polyacrylamide gel (30 ϫ 23 ϫ 0.15 cm) was also performed using a two-step increase of current, starting with 15 min at 130 mA, followed by a run of 7-8 h at 170 mA, until the dye reached the end of the gel. For analytical gels, the proteins were silver-stained based on the procedure of Klose and Kobalz (18). For micropreparative gels, the proteins were stained by Coomassie Blue as described by Eckerskorn et al. (22).
Enzymatic In-gel Digestion, Extraction of the Peptides from the Gel, and Concentration by Reversed-phase Material-For micropreparative investigations, up to 10 Coomassie Blue-stained protein spots of the same protein were combined from the micropreparative gels. Enzymatic in-gel digestion with trypsin (Promega, Madison, WI), extraction of the peptides from the gel, desalting, and concentration by reversed-phase material (Serva) were performed according to Otto et al. (23).
Reversed-phase HPLC and Edman Microsequencing-The separation of the peptides was performed using a microbore HPLC system and a column of 2.1-mm inner diameter and 10-cm length (RPC C2/C18 SC 2.1/10, Smart System, Pharmacia Biotech, Freiburg, Germany). A gradient with an increasing concentration of acetonitrile in 0.1% (v/v) trifluoroacetic acid and a flow rate of 100 l/min at room temperature was used. The peptide fractions were concentrated in a SpeedVac concentrator (Savant, Hicksville, NY) and loaded onto a Biobrene-coated glass fiber filter of a Procise or 477A pulsed liquid-phase sequencer with an on-line 120A phenylthiohydantoin-derivative analyzer (Perkin-Elmer/ Applied Biosystems, Foster City, CA). Sequencing reagents were purchased from Perkin-Elmer/Applied Biosystems (Weiterstadt, Germany).
MALDI-MS-MALDI-MS was performed with a VG Tof Spec (Fisons, Manchester, United Kingdom) equipped with a nitrogen laser (337 nm; pulse duration of 4 ns) and a VAX 400 VLC station using OPUS software (Version 3.1). Spectra were obtained in the linear or reflectron mode by summing 20 -50 laser shots. A saturated solution of ␣-cyano-4-hydroxycinnamic acid in aqueous 50% acetonitrile and 0.1% trifluoroacetic acid was used as matrix (24). The matrix (1.2 l) and sample (0.8 l) were mixed on the target and air-dried. For mass determination, the dried target was inserted into the mass spectrometer. Measurement took place under vacuum at 22 or 24 kV.
Protein Identification by Computer Analysis-The peptide mass maps produced by MALDI-MS were searched using the program FRAG-MOD (E.-C. Mü ller, Max-Delbrü ck-Centrum fü r Molekulare Medizin, Berlin-Buch, Germany), a modified version of the FRAGFIT program (25). Amino acid sequences were compared with the SWISS-PROT Data Bank (release of June 1997) and with the deduced amino acid sequences of the GenBank TM /EMBL Data Bank (release of June 97).

Induction of Apoptosis by Anti-IgM Antibody and Separation of Apoptotic Cells by Magnetic Cell
Sorting-To determine proteins involved in the regulation of anti-IgM antibody-induced apoptosis, we used a subclone of a Burkitt lymphoma cell line (BL60-2). After 24 h of treatment with anti-IgM antibody, 40 -50% of the BL60-2 cells bound FITC-conjugated annexin V as measured by fluorescence-activated cell sorting analysis, and 50 -60% showed fragmented nuclei after acridine orange staining (16). Both measurements are sensitive and reliable indicators for programmed cell death. 24 h after stimulation with anti-IgM antibody, cells were incubated with FITC-conjugated annexin V and subsequently with anti-FITC microbeads. Separation of apoptotic and non-apoptotic cells was achieved using magnetic separation columns. By this method, we could increase the sensitivity and facilitate the detection of proteins that were cleaved or down-regulated.
2-DE Technique-The investigations of apoptosis-associated proteins were aimed at elucidating protein changes between an apoptotic state and a non-apoptotic state by means of subtractive protein analysis (17,26). 28) with a resolution of up to 10,000 proteins (18) can be used to investigate complex protein mixtures. Protein spot positions and intensities were compared between patterns of apoptotic and non-apoptotic cells on analytical gels. Protein spots that were significantly different in intensity were analyzed by tryptic in-gel digestion, subsequent mass analysis of the eluted peptide mixture, and/or Edman microsequencing of the microbore HPLC-separated peptides (23). However, a prerequisite for successful protein analysis using 2-DE is the ability to apply high protein concen-trations to the gel without losing the high resolution power of analytical gels. We increased the protein concentration up to four times on micropreparative gels compared with earlier investigations (23). We used Coomassie Blue-stained micropreparative gels as described under "Experimental Procedures" with comparable resolution and reproducibility to silverstained analytical gels for microanalysis. In this way, we could avoid the protein in-gel concentration procedure (29).
Identification of Apoptosis-associated Protein Spots-The 2-DE protein map of non-apoptotic cells containing ϳ3250 protein spots is shown in Fig. 1. Spots X (actin, 42 kDa, pI 5.3), Y (fatty acid-binding protein, 15 kDa, pI 6.6), and Z (glyceraldehyde-3-phosphate dehydrogenase, 36 kDa, pI 8.6) were identified by mass fingerprinting and were used as markers for molecular mass and pI on the gel. The visual comparison between 10 2-DE patterns of non-apoptotic versus apoptotic Bcells revealed ϳ80 protein spots that were significantly altered in intensity. After IgM-mediated apoptosis, spots of increasing intensity should be due to up-regulation of the expression of a protein. In addition, protein fragments occurred newly, e.g. when cleaved by a protease or due to specific cleavage by a caspase activity. Spots decreasing in intensity could be due to down-regulation of a protein or to degradation or fragmentation of a protein. The framed areas 1-9 in the 2-DE pattern of non-apoptotic cells in Fig. 1 were then compared with the corresponding areas in the 2-DE pattern of apoptotic cells. Differences in the protein spot pattern are marked by letters (Figs. 2 and 3). To identify an altered protein spot unequivocally, we employed two different approaches, namely internal Edman microsequencing and peptide mass fingerprinting. An example of the identification of the protein dUTPase by mass fingerprinting is shown in Fig. 4. The identification of the protein neutral calponin by internal Edman microsequencing and an HPLC-separated peptide pattern are demonstrated in Fig. 5. These methods resulted in the identification of 12 proteins, as demonstrated in Table I, that are associated with anti-IgM antibody-mediated apoptosis. We found that actin, lamin B1, nucleolin, and hnRNP C1/C2 are fragmented. Neutral calponin, LSP1, HHR23B, ribosomal protein P0, FBP, dUTPase, hnRNP A1, and HP1␣ are either up-or down-regulated during the apoptotic process of anti-IgM antibody-in- duced apoptosis. These proteins and their functional role in the apoptotic process will be discussed below.
Cleavage of hnRNP A1-In the protein pattern of apoptotic cells (Fig. 3, panel 5*), a new spot (KЈ) could be detected. This spot was not present in the protein pattern of non-apoptotic cells (Fig. 3, panel 5).
To investigate the onset of cleavage, we performed Western blot analysis utilizing the monoclonal antibody 4B10. Fig. 6 shows that hnRNP A1 was cleaved into at least three distinct fragments. Fragments A and B appeared after 8 -12 h of incubation with anti-IgM antibodies. The smallest fragment (C) occurred after 24 h and corresponded in size to spot KЈ, which was identified on the 2-DE gel. Recently, we showed that caspase 3 is a crucial protease in anti-IgM antibody-induced apoptosis (16). In the global 2-DE approach used here, we found that by incubation with Z-DEVD-FMK, a cell-permeable irreversible inhibitor specific for caspase 3-like proteases, prior to stimulation with anti-IgM antibody, apoptosis could be completely inhibited. We therefore analyzed if hnRNP A1 cleavage in the cells could be blocked by Z-DEVD-FMK as investigated by Western blotting with antibodies raised against this protein. Fig. 7A shows that after addition of 200 M Z-DEVD-FMK prior to anti-IgM antibody incubation, the generation of all three cleavage products was inhibited. (Fragment C could only be detected after overexposure of the Western blot (Fig. 7B).) We could not identify a potential caspase 3 cleavage site (DXXD) in the hnRNP A1 amino acid sequence, indicating that caspase 3 probably has still other recognition sites that are currently unknown; another possibility would be that hnRNP A1 is cleaved by a caspase located downstream of caspase 3.
The hnRNP A1 protein consists of 371 amino acids and has a molecular mass of 38,715 Da and a theoretical pI of 9.26. The hnRNP A1 spot KЈ has a mass of ϳ16,000 Da and a pI of ϳ7.1. We analyzed the sequence of this fragment and found that it corresponds to an N-terminal fragment of the entire protein ( Table I). The hnRNP A1 protein is one of the major pre-mRNA/mRNA-binding proteins in eukaryotic cells and an abundant protein in the nucleus. hnRNP A1 is likely to be involved in mRNA transport from the nucleus to the cytoplasm (30).
hnRNP C1/C2-In our 2-DE pattern of non-apoptotic cells (Fig. 2, panel 1), the hnRNP C1/C2 proteins were resolved in six protein spots, D1-6 (Table I). These six spots (D1-6) were reduced to four spots (DЈ1-4) in the protein pattern of apoptotic cells (Fig. 2, panel 1*). The peptide mass fingerprints of these four DЈ spots identified them as members of the hnRNP C1/C2 protein family as well (Table I). However, the shift of the hnRNP C1/C2 protein spot group in the pattern of the apoptotic cells (Fig. 2, panel 1*) toward a neutral pH and lower mass range is evidence for the fragmentation of these proteins in the apoptotic process. The hnRNP C1/C2 proteins (33,299 Da, pI 5.1) are abundant nuclear proteins and thought to be involved in RNA splicing. Waterhouse et al. (12) have given a detailed report on hnRNP C1/C2 proteins that are cleaved by interleukin-1␤-converting enzyme-like proteases during apoptosis in a Burkitt lymphoma cell line induced by ionizing radiation, etoposide, and ceramide.
Actin-Two new spots (AЈ1 and AЈ2) appeared in the 2-DE pattern of apoptotic cells (Fig. 3, panels 2* and 3*) with molecular masses of ϳ31 and 15 kDa, respectively, according to their migration on the SDS gel. They were identified as fragments of actin (Table I) (15), and we could find fragments of these masses. Therefore, it is likely that actin was cleaved by a caspase activity.
Lamin B1-The new spot BЈ (Fig. 3, panel 4*) in the 2-DE pattern of apoptotic cells was identified as an N-terminal fragment of the lamin B1 protein (Table I). Lamin B1 (66,277 Da, pI 5.1) is a component of the nuclear lamina, and lamins are substrates for caspases (13). Rao et al. (13) suggested that the lamin cleavage by caspases facilitates the fragmentation of the nucleus.
Nucleolin-The new spot CЈ (Fig. 3, panel 5*) in the 2-DE pattern of apoptotic cells was identified as a C-terminal fragment of the protein nucleolin (Table I). Nucleolin (76,212 Da, pI 4.6) is a major RNA-binding protein of the nucleolus, and it is found to be associated mainly with the pre-ribosomal particles (31). Nucleolin has also been implicated in nuclear transport and in a variety of transcriptional processes (32). The interaction of granzyme A and nucleolin in vitro (14) may be important in the process of apoptosis, which accompanies cytotoxic T lymphocyte-mediated lysis of target cells. Pasternack et al. (14) demonstrated that the serine protease granzyme A binds to and cleaves nucleolin in vitro.
Neutral Calponin-The high intensity protein spot E (Fig. 3,  panel 6) in the 2-DE pattern of non-apoptotic cells was determined as neutral calponin (Table I). The intensity decreased significantly in the 2-DE pattern of apoptotic cells (Fig. 3, panel 6*). According to Masuda et al. (33), human neutral calponin (34,220 Da, pI 7.1) is a nonmuscle isoform of calponin and may play a physiological role in cytoskeletal organization. The fragmentation of human neutral calponin in the apoptotic process could be essential for cytoskeletal reorganization.
Lymphocyte-Specific Protein LSP1-Protein spot F (Fig. 2,  panel 1) belongs to a group of five protein spots in the 2-DE pattern of non-apoptotic cells and was identified as LSP1 (Table I). In the 2-DE pattern of apoptotic cells (Fig. 2, panel 1*), the five protein spots including protein spot F were significantly decreased in intensity. LSP1 (37,192, pI 4.7) binds to F-actin and to the cytoskeleton through its carboxyl-terminal basic domain (34). Jongstra-Bilen et al. (34) suggested that LSP1 interacts with the cytoskeleton by directly binding to F-actin. The degradation of LSP1 seen here would be compatible with the cytoskeletal reorganization observed in the apoptotic process.
HHR23B-Protein spot G (Fig. 2, panel 1) was identified in the 2-DE pattern of non-apoptotic cells as the HHR23B protein (Table I). In the 2-DE pattern of apoptotic cells (Fig. 2, panel  1*), a group of six spots including protein spot G were significantly decreased in intensity. The HHR23B protein (43,171 Da, pI 4.9) stimulates and is therefore involved in DNA excision repair in vitro (35). HHR23B is part of a complex consisting of the xeroderma pigmentosum complementation group C protein and one of the two human homologues of the Saccharomyces cerevisiae repair gene product Rad23 (36). The nucleotide excision repair consists of the removal of the damaged nucleotide(s) from DNA by dual incision of the damaged strand on both sides of the lesion, followed by filling of the resulting gap and ligation. DNA repair plays a crucial role in the prevention of mutagenesis.
60 S Acidic Ribosomal Protein P0 (L10E)-In the protein pattern of non-apoptotic cells (Fig. 3, panel 7), we characterized two close H spots as 60 S acidic ribosomal protein P0 (Table I).
In the pattern of the apoptotic cells (Fig. 3, panel 7*), two close HЈ spots were identified as 60 S acidic ribosomal protein P0 as well (Table I). Nevertheless, in apoptotic cells, the spots were remarkably shifted toward neutral pH, and we therefore conclude that 60 S acidic ribosomal protein P0 (34,274 Da, pI 5.7) is modified after anti-IgM antibody-induced apoptosis. Further investigations are needed to analyze the modification and/or fragmentation of 60 S acidic ribosomal protein P0. Grabowski et al. (37) suggested an extraribosomal function of human 60 S acidic ribosomal protein P0 as a DNA repair protein. Preliminary studies in our laboratory have shown that the ribosomal protein P0 can be detected in the nucleus as well.
FBP-Protein spot I belongs to a group of three protein spots in the 2-DE pattern of non-apoptotic cells (Fig. 3, panel 8) and was identified as FUSE-binding protein (Table I). In the 2-DE pattern of apoptotic cells (Fig. 3, panel 8*), the three protein spots were significantly decreased in intensity. FBP (67,535 Da, pI 7.4) plays a role as a growth-dependent regulator of c-myc expression (38) correlating with the c-myc expression that is rapidly down-regulated in BL60-2 cells after anti-IgM antibody-induced apoptosis (data not shown). dUTPase-The high intensity protein spot J (Fig. 3, panel 9) in the 2-DE pattern of non-apoptotic cells was characterized as dUTPase (Table I). Recently, this protein has been described as DUT-N in human cells (40). The intensity decreased significantly in the 2-DE pattern of apoptotic cells (Fig. 3, panel 9*). The enzyme dUTPase (15,367 Da, pI 6.1) catalyzes the hydrolysis of dUTP to dUMP and pyrophosphate, thereby preventing a deleterious incorporation of uracil into DNA (40,41). According to Gadsden et al. (40), analysis with isogenic uracil-DNA glycosylase (UNG1)-deficient or -proficient strains of S. cerevisiae indicated that the absence of dUTPase results in cell death. In addition, studies in Drosophila indicate that the regulated inhibition of dUTPase, combined with differential expression of uracil glycosylase, may lead to programmed cell death during larval development (42).
HP1␣-The expression of protein spot L was increased in apoptotic cells, as seen when comparing the 2-DE pattern of apoptotic cells (Fig. 3, panel 9*) and the 2-DE pattern of nonapoptotic cells (panel 9). This protein spot L was characterized as HP1␣ (Table I). This was confirmed by the molecular mass and pI of the wild-type HP1␣ protein (22,225 Da, pI 5.7).  Sugimoto et al. (43) showed that human HP1 is a DNA-binding protein and suggested that HP1 may play a role in gene silencing during apoptosis by modifying the active euchromatin into inactive heterochromatin. Recently, it has been shown that HP1 can be associated with the transcriptional coactivators TIF1␣ and TIF1␤ as well as with the inner nuclear membrane protein LBR (44). Inhibition of Cleavage and Modification by Z-DEVD-FMK-To verify the protein fragments due to cleavage of caspase 3-like proteases, specific enzymes involved in the apoptotic process, the cells were treated with Z-DEVD-FMK, a selective irreversible inhibitor of caspase 3, before induction of apoptosis using anti-IgM antibody. By comparison of the 2-DE pattern of cells 22 h after apoptosis induction with that of cells pretreated with Z-DEVD-FMK, fragmentation of actin, hnRNP C1/C2, hnRNP A1, lamin B1, and nucleolin could be completely inhibited, and no alterations were observed with proteins like HP1␣, 60 S acidic ribosomal protein P0, FBP, dUTPase, HHR23B, LSP1, and neutral calponin. In Fig. 8, the effect of Z-DEVD-FMK on fragmentation of actin is shown as an example. The formation of a 15-kDa fragment (marked with an arrow) was absent in Z-DEVD-FMK-treated cells. This result reveals that actin is a target protein of caspase 3, in accordance with Mashima et al. (15), and it is clearly involved in the process of anti-IgM antibody-mediated apoptosis.
Conclusion-A prerequisite for understanding programmed cell death is the knowledge of the molecular events controlling this process. We investigated the total protein fraction of human Burkitt lymphoma BL60 cells and found considerable changes in the protein patterns after two-dimensional high resolution gel electrophoresis when comparing apoptotic and non-apoptotic cells. Among the 80 detectable protein spot al-terations, we characterized 12 proteins and their variants by chemical and mass spectrometric approaches. Most proteins in the patterns seem to remain the same even when cells are undergoing apoptosis. However, a small number of critical proteins seemed to be involved. To identify these proteins, we have successfully applied 2-DE in a new micropreparative manner. To prove the direct involvement of the identified proteins in the apoptosis process, we treated the cells with a specific caspase inhibitor prior to apoptosis induction. Cleavage of the identified fragmented proteins and apoptosis could be inhibited for hnRNP A1, hnRNP C1/C2, actin, 60 S acidic ribosomal protein P0, lamin, and nucleolin by Z-DEVD-FMK, a selective inhibitor of caspase 3-like proteases. Our analysis resulted in the identification of several new proteins altered after anti-IgM antibody-induced apoptosis, namely neutral calponin, LSP1, HHR23B, ribosomal protein P0, FBP, dUTPase, hnRNP A1, and HP1␣. These proteins are potential candidates for controlling apoptosis in B lymphocytes. Using the same approach, we already identified the protein D4-GDI as a caspase substrate during anti-IgM antibody-induced apoptosis (16). The results and the strategy of global protein analysis presented here demonstrate that this approach is a very useful tool to clarify the molecular events in apoptosis.