Mitosis-specific phosphorylation of vimentin by protein kinase C coupled with reorganization of intracellular membranes.

Using two types of anti-phosphopeptide antibodies which specifically recognize vimentin phosphorylated by protein kinase C (PKC) at two distinct PKC sites, we found that PKC acted as a mitotic vimentin kinase. Temporal change of vimentin phosphorylation by PKC differed form changes by cdc2 kinase. The mitosis-specific vimentin phosphorylation by PKC was dramatically enhanced by treatment with a PKC activator, 12-O-tetradecanoylphorbol-13-acetate (TPA), while no phosphorylation of vimentin by PKC was observed in interphase cells treated with TPA. By contrast, the disruption of subcellular compartmentalization of interphase cells led to vimentin phosphorylation by PKC. Cytoplasmic and nuclear membranes are fragmented and dispersed in the cytoplasm and some bind to vimentin during mitosis. Thus, targeting of activated PKC, coupled with the reorganization of intracellular membranes which contain phospholipids essential for activation, leads to the mitosis-specific phosphorylation of vimentin. We propose that during mitosis, PKC may phosphorylate an additional subset of proteins not phosphorylated in interphase.

Differential targeting of protein kinase C and CaM kinase II signalings to vimentin.

Hydrolysis of inositol phospholipids by receptor stimulation activates two separate signaling pathways, one leading to the activation of protein kinase C (C kinase) via formation of diacylglycerol. The other is the inositol trisphosphate (IP3)/Ca2+ pathway and a major downstream kinase which is activated is Ca2+/calmodulin-dependent protein kinase II (CaM kinase II). To examine signaling pathways of C kinase and CaM kinase II to the cytoskeletal protein vimentin, we prepared monoclonal antibodies YT33 and MO82 which recognize the phosphorylation state of vimentin by C kinase and by CaM kinase II, respectively. Ectopic expression of constitutively active C kinase or CaM kinase II in primary cultured astrocytes by microinjection of the corresponding expression vectors induced phosphorylation of vimentin at each specific phosphorylation site, followed by reorganization of vimentin filament networks. In contrast, simultaneous activation of C kinase and CaM kinase II by inositol phospholipid hydrolysis with receptor stimulation led to an exclusive phosphorylation of vimentin at the CaM kinase II site, not at the site of C kinase. These results indicate that the intracellular targeting of C kinase and CaM kinase II signalings to vimentin is regulated separately, under physiological conditions.

Antibodies specific for proteolyzed forms of protein kinase C alpha.

The activation of protein kinase C (PKC) is irreversibly regulated by limited proteolysis catalyzed by a calcium-activated neutral cysteine protease, calpain. Calpain cleaves PKC alpha at specific sites in the hinge region between the catalytic and the regulatory domains of this kinase. Here we show a novel method for production of antibodies that bind specifically to the catalytic fragment of PKC alpha but not to the unproteolyzed protein. To detect proteolyzed PKC alpha 'cleavage site-directed antibodies,' which recognize amino-terminal regions in the nascent catalytic fragments and do not cross-react with the unproteolyzed enzymes, were raised using synthetic peptides corresponding to the amino-terminal sequences. The synthetic peptides used in this study were the sequences of human PKC alpha at the cleavage sites by m- and mu-types of calpains (LGPAGNKV and VISPSEDRKQPSNNLDRVKLT, respectively) and they are designated as CF alpha 2, CF alpha 4, in this order. Each synthetic peptide was injected into rabbit after conjugation with a carrier protein. The antibodies thus obtained (anti-CF alpha 2 or -CF alpha 4) specifically reacted with either the 46- or 45-kDa catalytic fragment of PKC alpha, respectively, whereas they did not cross-react with other fragments. Furthermore, the antibodies did not bind to the unproteolyzed enzyme nor fragments of PKC alpha obtained by treatment with other proteinases unless the fragment carried the same amino-terminal sequence. When human platelets were treated with calcium ionophore, the catalytic fragments of PKC alpha (45- and 46-kDa) were detected in the cytosol by immunoblotting with the antibodies. However, these antibodies did not bind unproteolyzed 80-kDa PKC alpha, although this form was dominant in the cytosol of the calcium ionophore-treated human platelets. In addition, the 45-kDa catalytic fragment of PKC alpha was detected in an apoptotic human fibroblast TIG-3 cells cultured in serum-free medium. Our method is applicable for analysis of proteolysis in various cellular states.

Novel antibodies specific for proteolyzed forms of protein kinase C: production of anti-peptide antibodies available for in situ analysis of intracellular limited proteolysis.

We show here a novel method for the in situ analysis of proteolyzed proteins in a cell. As a model, we focused on protein kinase C (PKC) beta, which is cleaved at a specific site between the catalytic and regulatory domains by calpain, the intracellular calcium-activated neutral proteinase. To detect proteolyzed PKC beta 'cleavage-site-directed antibodies', which specifically recognize the amino-terminal region of the catalytic fragment but do not cross-react with the unproteolyzed enzymes, were raised using synthetic peptide. The synthetic peptide used in this study was QGTKVPEEKTT, corresponding to the amino-terminal region of the catalytic fragment from human PKC beta generated by calpain. Rabbits were immunized with the synthetic peptide after conjugation with a carrier protein. Antibodies obtained reacted with the 46-kDa catalytic fragment of PKC beta, whereas they did not cross-react with unproteolyzed enzyme nor other fragments with different amino-termini. Thus, our antibody is specific to the amino-terminal sequence QGTKVPEEKTT, but does not recognize the same sequence located internally in native PKC beta. When human monoblast U937 cells were treated with calcium ionophore, the catalytic fragment of PKC beta was detected in the cytosol by immunoblotting with the antibody. However, this antibody did not bind unproteolyzed 80-kDa PKC beta, although this form was dominant in the cytosol of the calcium ionophore-treated cells. We could also detect comparable amounts of catalytic fragment in the calcium ionophore-treated cells by immunocytochemical staining with the same antibody. Our method was applied to examine the proteolysis of PKC beta in neutrophils stimulated with various reagents.

A variety of calpain/calpastatin systems in mammalian erythrocytes.

Calpain and its endogenous inhibitor, calpastatin, were isolated from erythrocytes of various mammals and their properties were compared. It has been widely believed that mammalian erythrocytes contain only mu-calpain. However, rat and human erythrocytes were found to contain two species of calpain, identified as mu-calpain and m-calpain from their elution positions on DEAE-cellulose column chromatography and their Ca(2+)-requirements. Thus, it is apparent that rat and human erythrocytes contain not only mu-calpain, but m-calpain as well. On the other hand, rabbit erythrocytes contain only mu-calpain. Western blot analysis showed that human and rabbit erythrocytes contain predominantly 70-kDa calpastatin (erythrocyte-type), but unnegligible amounts of 110-kDa calpastatin (tissue-type) are also present. Rat erythrocytes were shown to contain a calpastatin with a molecular mass of approx. 100 kDa almost exclusively; this molecular mass was in perfect coincidence with the mass of the calpastatin in rat lung. These results strongly suggest that rat erythrocytes contain a tissue-type calpastatin. No essential change in the calpain/calpastatin system during maturation of rabbit reticulocytes into mature erythrocytes was observed.

Purification and characterization of recombinant hamster tissue complement C1s.

Hamster complement C1s cDNA was inserted into expression plasmid BCMGSNeo, and transfected to SEA7 cells, A31 mouse fibroblasts transformed by polyoma virus. The transfectant secreted a large amount of recombinant C1s that was activated in the serum free culture medium and hydrolyzed acetyl-Gly-L-Lys-naphthyl ester (AGLNE). C1s was purified to a homogeneity from the culture medium of the transfectant by DEAE-Sephadex, Dymatrex orange A and size-exclusion HPLC. Purified hamster C1s consumed human complement in hemolytic assay and hydrolyzed gelatin in enzymography. To investigate the enzymic action of C1s at molecular levels, several antibodies were prepared against hamster C1s. One peptide (amino-acid residues 379-391) and two peptides (amino-acid residues 478-496 and 560-583) corresponding to the heavy and the light chain, respectively, were synthesized. The amino-acid sequences of these regions is not conserved between hamster and human C1s. Antibodies against these peptides were raised in rabbits. The anti-peptide antibodies bound specifically to hamster serum and recombinant C1s but not to human C1s. They inhibited the esterase activity of recombinant C1s to varying degrees depending on each antibody's binding site.

Activation and possible involvement of calpain, a calcium-activated cysteine protease, in down-regulation of apoptosis of human monoblast U937 cells.

An active form of calpain mu, a low-Ca(2+)-requiring intracellular cysteine protease, was detected using a cleavage site-directed antibody in apoptotic human monoblast U937 cells treated with tumor necrosis factor-alpha and interferon-gamma. Membrane-permeable calpain inhibitors accelerated apoptosis of U937 cells thus induced and suppressed the activation of procalpain. These findings suggest that calpain down-regulates apoptosis by shutting off the intracellular signals for cell death.

Monocytic differentiation modulates apoptotic response to cytotoxic anti-Fas antibody and tumor necrosis factor alpha in human monoblast U937 cells.

Interferon-gamma (IFN-gamma), vitamin D3 (VD), and retinoic acid (RA) induce differentiation of human monoblastic leukemia U937 cells to macrophage-like cells with potential superoxide anion-generating activity upon further stimulation. Here we report that U937 cells thus differentiated show various responses to apoptotic induction with a cytotoxic anti-Fas antibody and tumor necrosis factor (TNF). VD-or RA-treated U937 cells acquired resistance against Fas- or TNF receptor (TNFR)-mediated apoptosis, whereas apoptotic cell death was accelerated in IFN-gamma-treated cells. By flow cytometric analyses, no decrease in expression of surface Fas antigen or p55 TNFR was observed in differentiated U937 cells. Cell surface expression of CD11b was seen only when differentiation was induced with VD or RA but not with IFN-gamma. The growth of VD- or RA-treated cells was retarded but IFN-gamma-treated cells were prolific. These findings suggest that the differentiation state differs with the inducer and that the cellular response to apoptotic induction is closely related to the state including the cell cycle.

Absence of down-regulation and translocation of the eta isoform of protein kinase C in normal human keratinocytes.

Among 11 isoforms of protein kinase C (PKC), we previously reported that the eta isoform of PKC plays a crucial role in mediating differentiation of keratinocytes. Activation of PKC is associated with its intracellular translocation from the cytoplasm to the plasma membrane, followed by down-regulation through proteolytic cleavage of the PKC molecules. In the present study, we demonstrated that the eta isoform of PKC is unique in that it is not translocated nor down-regulated upon stimulation. The level of the eta isoform, assayed by immunoblotting, remained unchanged during the first 12 h and then increased slightly up to 24 h when treated with tumor promoters or activators of PKC in constitutively expressing normal human keratinocytes. The activity of the eta isoform also remained unchanged after the 12-O-tetradecanoyl-phorbol-13-acetate treatment, as judged by binding ATP analog, autophosphorylation, and phosphorylation of an exogenous substrate. The alpha isoform of PKC, however, was rapidly down-regulated and was undetectable by 6 h after the treatment. These observations were further confirmed by immunohistochemical staining of normal human keratinocytes and transiently expressing COS1 cells. In addition, although the alpha isoform rapidly translocated to the plasma membrane, the eta isoform remained in the cytoplasm.

Activation of protein kinases in canine basilar artery in vasospasm.

Subarachnoid hemorrhage (SAH) often leads to a long-term narrowing of cerebra! artery called vasospasm. To understand the molecular mechanisms in vasospasm, signal transduction of tyrosine kinase pathway and phosphorylation of myosin light chain (MLC) and calponin (CaP) in the basilar artery were studied. Vasospasm was produced in the canine basilar artery by a two-hemorrhage method, and vasocontraction was induced by a local application of KCI or serotonin to the basilar artery after a transclival exposure. Intracellular substrates of tyrosine kinase pathway, including Shc, Rafl, and extracellular-regulated kinases in the basilar artery, were activated after SAH, and the activation of Shc suggests stimulation of signal transductions from tyrosine kinase receptors, G-coupled receptors, or both. The activation of tyrosine kinase pathway in vasospasm also was supported by dose-dependent dilation of the spastic basilar artery on days 0 and 7 by topical application of genistein, a tyrosine kinase inhibitor, and associated marked inhibition of tyrosine phosphorylation of intracellular substrates, including Shc. In addition, the generation of protein kinase M, catalytic fragment of protein kinase C(alpha) (PKC alpha), in vasospasm on days 0 and 7 was inhibited in response to genistein, indicating an inactivation of mu-calpain. It is suggested, therefore, that the reversal of vasospasm by genistein is closely associated with the restoration of intracellular Ca2+ levels. However, the increased activities of Raf1 and extracellular-regulated kinases in vasospasm were declined on day 7 compared with those on day 0 or 2, suggesting that the activation of tyrosine kinase pathway is more closely associated with the early stage of vasospasm than with the late stage of vasospasm. The analysis by pyrophosphate polyacrylamide gel electrophoresis (PPi-PAGE) demonstrated three MLC bands in vasospasm on days 2 and 7, as well as in KCI- and serotonin-induced vasocontraction. Since PPi-PAGE resolves smooth muscle MLC into three bands in the MLC kinase (MLCK)-mediated phosphorylation and into a single band in the PKC-mediated phosphorylation based on the phosphorylation state, the current results suggest that MLC in vasospasm is phosphorylated by MLCK but not by PKC. In basilar artery, CaP was significantly down-regulated, and in addition, significantly phosphorylated on serine and threonine residues only in vasospasm on days 2 and 7. Although the significance of CaP phosphorylations in vivo still is controversial, CaP down-regulation and phosphorylation may attenuate the inhibition of Mg(2+)-ATPase activity by CaP and induce a potential enhancement of smooth muscle contractility in delayed vasospasm. Since CaP is phosphorylated in vivo by PKC, activated PKC in vasospasm may phosphorylate CaP. Thus, SAH stimulates tyrosine kinase pathway to increase intracellular Ca2+ and activate PKC, and the former activates MLCK to phosphorylate MLC, whereas the latter phosphorylates CaP but not MLC.

Complete primary structure of calcium-dependent serine proteinase capable of degrading extracellular matrix proteins.

A novel calcium-dependent serine proteinase (CASP) secreted from malignant hamster embryo fibroblast Ni 12C2 degrades extracellular matrix proteins. A complementary DNA encoding CASP has been isolated with the use of oligonucleotide probes synthesized based on partial amino acid sequences of CASP. The complete amino acid sequences of CASP revealed that it has a active site at the C-terminal side. Glu rich and proEGF homologous sites are found at the N-terminal site suggesting that it is structurally similar to blood coagulation factors such as IX, X and an anticoagulation factor, protein C.

The inhibitory effect of prostaglandin E1 on oxidative stress-induced hepatocyte injury evaluated by calpain-mu activation.

BACKGROUND: Prostaglandin E1 (PGE1) is known to inhibit ischemia-reperfusion injury of the liver. The calcium-dependent neutral proteinase, calpain-mu, is involved in oxidative stress-induced hepatocyte injury. We investigated the mechanisms of cytoprotection by PGE1, focusing on the elevation of intracellular calcium ([Ca2+]i), activation of calpain-mu, and calpain-mu-mediated activation of protein kinase C-alpha (PKC-alpha). METHODS: Cultured hepatocytes were treated with various amounts of PGE1 (0, 0.1, 1.0, 10, and 100 ng/ml) for 30 min and subsequently with 0.5 mM tert-butyl hydroperoxide (TBHP). Cell injury was evaluated by the release of lactate dehydrogenase. Plasma membrane bleb formation was examined by phase contrast microscopy. Activation of calpain-mu and limited degradation of PKC-alpha was evaluated by Western blotting using antibodies that specifically recognize the amino-terminal regions of calpain-mu and PKC-alpha. [Ca2+]i was measured by confocal microscopy using Fluo-3AM. RESULTS: LDH release from cells treated with 10 ng/ml PGE1 was significantly lower than from untreated cells (135 +/- 12 vs. 258 +/- 18 IU/L, respectively; P < 0.05). Morphologically, many blebs were observed in untreated cells, but very few were seen in those treated with 10 ng/ml PGE1. Western blotting revealed that the amount of activated calpain-mu and [Ca2+]i increased up to 1,300 nM at 35 min after the addition of TBHP (0.5 mmol/L) in control experiments (without PGE1). PGE1 (10 ng/ml) delayed the rise in [Ca2+]i for about 30 min, but did not suppress it completely. PKC-alpha decreased in experiments using PGE1 (10 ng/ml). CONCLUSION: PGE1 exerts its cytoprotective effect in TBHP-induced hepatocyte injury partly by inhibiting Ca2+-calpain-mu-mediated mechanisms.

A novel anti-apoptosis serum factor that down-regulates fas-mediated apoptosis.

The Fas antigen (CD95) is a receptor for apoptotic cell death signals. Human T Jurkat cells that express the Fas molecule on the cell surface are sensitive to cytotoxic anti-Fas monoclonal antibodies (mAbs). We report here that fetal calf serum (FCS) contains a novel factor (SAF: Serum Anti-Apoptosis Factor) that suppresses Fas-mediated apoptosis of Jurkat cells. The apoptotic cell death of Jurkat cells induced by an anti-Fas mAb was accelerated when FCS was removed from the medium. Depletion of serum itself did not induce apoptosis without the mAb. The apoptosis-suppressing activity was not seen in neonatal calf or adult bovine serum, suggesting that the SAF activity was developmentally regulated. The anti-apoptosis activity was not detected in any fraction of FCS obtained on gel filtration chromatography. However, the activity was recovered in a certain fraction on the addition of calf serum, suggesting that the anti-apoptotic activity of SAF requires other serum factor(s). We purified SAF from FCS using an assay system reconstituted with calf serum, and raised antibodies (Abs) to it in rabbits. The anti-SAF Ab inhibited the apoptosis-suppressing activity, indicating that SAF is a key factor for the anti-apoptotic activity of FCS.

Caspases cleave the amino-terminal calpain inhibitory unit of calpastatin during apoptosis in human Jurkat T cells.

We have previously reported the activation of procalpain mu (precursor for low-calcium-requiring calpain) in apoptotic cells using a cleavage-site-directed antibody specific to active calpain [Kikuchi, H. and Imajoh-Ohmi, S. (1995) Cell Death Differ. 2, 195-199]. In this study, calpastatin, the endogenous inhibitor protein for calpain, was cleaved to a 90-kDa polypeptide during apoptosis in human Jurkat T cells. The limited proteolysis of calpastatin preceded the autolytic activation of procalpain. Inhibitors for caspases rescued the cells from apoptosis and simultaneously inhibited the cleavage of calpastatin. The full-length recombinant calpastatin was also cleaved by caspase-3 or caspase-7 at Asp-233 into the same size fragment. Cys-241 was also targeted by these caspases in vitro but not in apoptotic cells. Caspase-digested calpastatin lost its amino-terminal inhibitory unit, and inhibited three moles of calpain per mole. Our findings suggest that caspases trigger the decontrol of calpain activity suppression by degrading calpastatin.

Erythropoietin 5'-flanking sequence-binding protein induced during hypoxia and cobalt exposure.

The 5'-flanking region of the human erythropoietin (Epo) gene contains a 0.14-kb sequence that is conserved in the Epo gene from mouse and located within a promoter that is activated under hypoxic conditions such as anemia. Using a fragment containing this sequence in DNA mobility shift assays, we found that specific DNA-binding proteins were induced in mouse kidney nuclei under anemic hypoxia. Using synthetic double-stranded oligonucleotides that contain this sequence, the essential binding site was defined to be the -40 to -20 region upstream of the transcription initiation site in the human Epo gene. By DNA affinity chromatography using a column with the immobilized 5'-flanking sequence, two inducible binding proteins with apparent molecular masses of 55 and 45 kDa were identified in the nuclei of mouse kidney and liver under anemic hypoxia. These binding proteins were also induced during cobalt exposure.

Suppression of superoxide-generating ability during differentiation of monocytes to dendritic cells.

Human peripheral monocytes cultured with GM-CSF and IL-4 differentiated to dendritic cells (DCs) and with GM-CSF alone to macrophages. Superoxide-generating ability in such DCs was found to be suppressed whereas that in macrophages remained constant. To examine the reason for the suppression in DCs, we evaluated by immunoblotting the levels of essential components of the superoxide generating system in the cells during the differentiation. In contrast to the levels of cytosolic 47- and 65-kDa components and Rac-p21, which remained constant throughout cultivation, those of the large and the small subunits of cytochrome b558 were found to decrease quickly by day 2 during cultivation of monocytes with GM-CSF and IL-4. DCs obtained after 7 days of cultivation had lost the large subunit almost completely and most of the small subunit. A cell surface epitope of the cytochrome detected by a monoclonal antibody also decreased during the differentiation. On the other hand, these components, including both subunits of cytochrome b558, were maintained in the cells during differentiation of monocytes to macrophages. These results indicate that the decreased levels of cytochrome b558, especially that of the large subunit, is responsible for the low level of superoxide-generating ability of DCs and that the suppression is caused by IL-4.

Carboxyl-terminal truncation and site-directed mutagenesis of the EF hand structure-domain of the small subunit of rabbit calcium-dependent protease.

A mutant of the small subunit of rabbit calcium-dependent protease lacking the amino-terminal one-fourth produced in Escherichia coli could associate with the native large subunit to exert protease activity. Deletion of a few carboxyl-terminal residues of this variant small subunit caused a significant decrease in the protease activity after reconstitution with the native large subunit. Loss of the fourth EF hand loop region by further truncation of the variant small subunit made interaction with the large subunit impossible. The calcium binding assay revealed that the fourth EF hand structure of the rabbit small subunit, which has been previously demonstrated to possess two calcium-binding sites, can bind calcium ions. Furthermore it was established by site-directed mutagenesis that the first EF hand structure, in addition to the fourth one, is capable of binding calcium ions. Replacement of amino acids in the EF hand structure affected interaction with the native large subunit or the calcium sensitivity of the reconstituted product. These findings indicate that the EF hand structure-domain of the small subunit is essential for the full protease activity.

Identification and characterization of inhibitory sequences in four repeating domains of the endogenous inhibitor for calcium-dependent protease.

We reported previously the cDNA cloning of the endogenous inhibitor for calcium-dependent protease (CANP inhibitor, calpastatin) and the expression of its fragments in Escherichia coli. The CANP inhibitor has four internal repeating domains each spanning about 140 amino acid residues. The inhibitory activity arises from these domains which have a well-conserved sequence, TIPPXYR, in their central positions. The inhibitory activities of various fragments expressed in E. coli suggest the involvement of the regions around the well-conserved sequences. In this report, we describe further detailed investigation on the interaction site of the CANP inhibitor with CANP by truncating inhibitor fragments and by using chemically synthesized peptides. The results clearly indicate that the sequence around the well-conserved sequence, TIPPXYR, is an interaction site. A peptide as short as 23 amino acid residues retained inhibitory activity, but a 9-residue peptide corresponding to the conserved sequence, VTIPPKYRE had none. The inhibitory sequence is suggested as LGXKDREXTIPPXYRXLL. The analysis of the competition between an inhibitor peptide and an irreversible inhibitor, E-64 for the reaction with the active site suggests no involvement of the active site cysteine residue of CANP in the inhibitory interaction between CANP and the CANP inhibitor. The high specificity of the CANP inhibitor to CANP arises from its interaction with residues other than the active site cysteine residue, possibly the subsite for substrate-binding of CANP.

Induction of essential components of the superoxide generating system in human monoblastic leukemia U937 cells.

Cytochrome b558 composed of large and small subunits, and cytosolic 47- and 65-kDa proteins are important constituents of the superoxide (O2-) generating system in phagocytes and B lymphocytes. In this paper, we describe changes in O2(-)-generating activity and expression of O2(-)-generating components during differentiation of human monoblastic leukemia U937 cells to macrophage-like cells. Undifferentiated U937 cells generated no O2- in response to a stimulation, although they expressed the three components other than the cytochrome b558 large subunit. When U937 cells were cultured with agents that induced the cell differentiation, such as vitamin D3, retinoic acid, interferon-gamma, and tumor necrosis factor, O2(-)-generating activity increased 5- to 200-fold depending on the agent used. Immunoblotting analysis revealed that the amounts of the four protein components essential for O2- generation increased, although their induction levels were significantly different between inducers. Among the four protein components, the cytochrome subunits were induced in low levels by all agents tested, which may explain why the O2(-)-generating activity of differentiated U937 cells was much lower than that of neutrophils from peripheral blood.

Characterization of the superoxide-generating system in human peripheral lymphocytes and lymphoid cell lines.

In B lymphocytes, but not T lymphocytes, isolated from human peripheral blood, we detected the four protein components essential for "the respiratory burst" by immunoblot analyses using peptide-directed antibodies. These are two membrane proteins, namely, 91- and 22-kDa subunits of cytochrome b558, and two cytosolic proteins with molecular masses of 47 and 65 kDa. Like in neutrophils, cytochrome b558 was expressed on the cell surface of peripheral B lymphocytes. Mean amounts (n = 8) of the 91-, 22-, 47-, and 65-kDa proteins, respectively, in peripheral B lymphocytes calculated from intensity of the blots were 0.011 +/- 0.003, 0.026 +/- 0.006, 0.179 +/- 0.022, and 0.039 +/- 0.013 relative to those in neutrophils on the basis of cell number. Epstein-Barr virus (EBV)-transformed cell lines derived from normal B lymphocytes and some B cell lines also possessed cytochrome b558 and two cytosolic proteins. Isolated human peripheral B lymphocytes generated the superoxide anion upon cross-linking of surface antigens such as IgM, IgD, IgG, HLA-DR, and CD19. EBV-transformants derived from normal peripheral B lymphocytes and B lymphoid cell lines also generated the superoxide anion when stimulated with various antibodies against surface antigens. These results indicate that peripheral B lymphocytes have substantial amounts of a superoxide-generating system identical to that in phagocytes and that the system is stimulated to generate the superoxide anion by the cross-linking of clonally expressed surface immunoglobulins or of certain surface antigens.