JFC1, a novel tandem C2 domain-containing protein associated with the leukocyte NADPH oxidase.

We have employed a yeast two-hybrid system to screen a B lymphoblast-derived cDNA library, searching for regulatory components of the NADPH oxidase. Using as bait the C-terminal half of p67(phox), which contains both Src homology 3 domains, we have cloned JFC1, a novel human 62-kDa protein. JFC1 possesses two C2 domains in tandem. The C2A domain shows homology with the C2B domain of synaptotagmins. JFC1 mRNA was abundantly expressed in bone marrow and leukocytes. The expression of JFC1 in neutrophils was restricted to the plasma membrane/secretory vesicle fraction. We confirmed JFC1-p67(phox) association by affinity chromatography. JFC1-containing beads pulled down both p67(phox) and p47(phox) subunits from neutrophil cytosol, but when the recombinant proteins were used, only p67(phox) bound to JFC1, indicating that JFC1 binds to the cytosolic complex via p67(phox) without affecting the interaction between p67(phox) and p47(phox). In contrast to synaptotagmins, JFC1 was unable to bind to inositol 1,3,4,5-tetrakisphosphate but did bind to phosphatidylinositol 3,4,5-trisphosphate and to a lesser extent to phosphatidylinositol 3,4-diphosphate. From the data presented here, it is proposed that JFC1 is acting as an adaptor protein between phosphatidylinositol 3-kinase products and the oxidase cytosolic complex.

Purification and characterization of a C5a-inactivating enzyme from human peritoneal fluid.

Earlier work has suggested that familial Mediterranean fever, an inherited disorder characterized by sporadic episodes of inflammation involving the pleural and peritoneal cavities and the joints, is caused by the lack of a C5a inactivator normally found in serosal fluid. We have purified this inactivator from ascites fluid and obtained a protein of molecular weight 53 to 56 kD with a specific activity 10,000-fold greater than the crude material. On Western blot, an inhibitory antibody recognized a single antigenic species at the same molecular weight. The enzyme had no activity against denatured bovine serum albumin. With recombinant C5a as substrate, the Km and Vm were 3.4 mumol/L and 52 nmol C5a/min/mg protein, respectively.

Cell-free activation of the respiratory burst oxidase by protein kinase C.

In intact neutrophils, phorbol ester treatment activates the respiratory burst oxidase, the enzyme responsible for O2-production by phagocytes. This effect is thought to be dependent on protein kinase C and on the phosphorylation of p47phox. In this paper, we report that protein kinase C activates the respiratory burst oxidase in a cell-free system consisting of isolated neutrophil cytosol and membrane. Oxidase activation required a highly active protein kinase C, recombinant p47phow and ATP, and was inhibited by the protein kinase C inhibitors H-7 and GF-109203X. PERIl depletion of cytosolic ATP by dialysis reduced oxidase activation by over 50% In contrast, neither protein kinase C inhibitors nor ATP depletion affected oxidase activation by SDS. These findings strongly suggest that in the cell-free system, the oxidase can be activated by the phosphorylation of p47phox.

Cytosolic guanine nucleotide-binding protein Rac2 operates in vivo as a component of the neutrophil respiratory burst oxidase. Transfer of Rac2 and the cytosolic oxidase components p47phox and p67phox to the submembranous actin cytoskeleton during oxidase activation.

The respiratory burst oxidase is responsible for O2- production in stimulated neutrophils and B lymphocytes. Components of this oxidase include cytochrome b558, a membrane-bound flavohemoprotein; the cytosolic polypeptides p47phox and p67phox; and one or more small G proteins including Rac1, Rac2, and/or Rap1A. We found that when normal neutrophils were activated, small percentages of each of the cytosolic proteins p47phox, p67phox, and Rac2 were transferred to the membrane cytoskeleton. However, Rac2 was not transferred to the membrane during activation of p47phox-deficient neutrophils. In normal cells, some p47phox also became associated with the non-cytoskeletal portion of the plasma membrane, but p67phox, Rac2, and O(2-)-forming activity were restricted to the cytoskeleton. Neutrophil activation also causes the phosphorylation of multiple serines in p47phox. The most heavily phosphorylated forms of p47phox were found solely in the membrane cytoskeleton. These results suggest that 1) the membrane cytoskeleton participates in respiratory burst oxidase activation, 2) the fully phosphorylated p47phox is located in the active oxidase, which resides in the membrane cytoskeleton, and 3) Rac2 acts like a dedicated component of the respiratory burst oxidase.

The reduction of cytochrome b558 and the activity of the respiratory burst oxidase from human neutrophils.

It is known that in respiratory burst oxidase preparations engaged in O2- production, cytochrome b558, a characteristic oxidase component, is partly reduced. This result has been interpreted in terms of a mechanism in which cytochrome b558 functions as an electron-carrying component of the respiratory burst oxidase, its level of reduction reflecting a steady-state partitioning of the cytochrome between reduced and oxidized forms as it ferries electrons from NADPH to oxygen. Kinetic arguments based on this interpretation have supported the proposal that the cytochrome is reduced at a rate sufficient to account for the rate of O2- production by activated neutrophils. We have confirmed the partial reduction of cytochrome b558 in neutrophil cytoplasts and in oxidase preparations exposed to NADPH, but have found that the reduction of the cytochrome bears no apparent relation to the activity of the oxidase, and can occur when NADPH is added to neutrophil membrane preparations that are unable to manufacture O2-. We therefore conclude that the NADPH-dependent reduction of cytochrome b558 seen in these preparations is unlikely to be a reflection of a catalysis-related steady state and that inferences drawn from such observations regarding the kinetic competence of the cytochrome may need to be reconsidered.

O2- production by B lymphocytes lacking the respiratory burst oxidase subunit p47phox after transfection with an expression vector containing a p47phox cDNA.

The respiratory burst oxidase of phagocytes and B lymphocytes is a complicated enzyme that catalyzes the one-electron reduction of oxygen by NADPH. It is responsible for the O2- production that occurs when these cells are exposed to phorbol 12-myristate 13-acetate or other appropriate stimuli. The activity of this enzyme is greatly decreased or absent in patients with chronic granulomatous disease, an inherited disorder characterized by a severe defect in host defense against bacteria and fungi. In every chronic granulomatous disease patient studied to date, an abnormality has been found in a gene encoding one of four components of the respiratory burst oxidase: the membrane protein p22phox or gp91phox, or the cytosolic protein p47phox or p67phox. We report here that O2- production was partly restored to phorbol 12-myristate 13-acetate-stimulated Epstein-Barr virus-transformed B lymphocytes from a patient with p47phox-deficient chronic granulomatous disease by transfection with an expression plasmid containing a p47phox cDNA inserted in the sense direction. No detectable O2- was produced by untransfected p47phox-deficient lymphocytes or by p47phox-deficient lymphocytes transfected with an antisense plasmid. The finding that O2- can be produced by p47phox-deficient B lymphocytes after the transfer of a p47phox cDNA into the deficient cells suggests that this system could be useful for studying the function of mutant p47phox proteins in whole cells.

The cytosolic components of the respiratory burst oxidase exist as a M(r) approximately 240,000 complex that acquires a membrane-binding site during activation of the oxidase in a cell-free system.

Sodium dodecyl sulfate (SDS) treatment of a mixture of cytosol and plasma membranes from resting neutrophils resulted in the activation of the respiratory burst oxidase, a complicated enzyme that catalyzes the production of O2- from NADPH and oxygen. Activation was accompanied by translocation to the plasma membranes of the oxidase components p47phox and p67phox, which in resting cytosol were found in a M(r) approximately 240,000 complex. This translocation, which appeared to take place without a major change in the size of the cytosolic complex, did not occur if the membranes lacked cytochrome b558, and was inhibited by the peptide PRGV-HFIFNK, a sequence found near the carboxyl terminus of cytochrome b558 that was known from earlier work to inhibit O2- production by the cell-free system (Rotrosen, D., Kleinberg, M. E., Nunoi, H., Leto T., Gallin, J. I., and Malech H. L. (1990) J. Biol. Chem. 265, 8745-8750). Cytosols pretreated with the cross-linking agents 3,3'-dithiobis(sulfosuccinimidyl) propionate (DTSSP) (cleavable by 2-mercaptoethanol) and bis-(sulfosuccinimidyl) suberate (not cleavable by 2-mercaptoethanol) lost most of their ability to support O2- production in the cell-free system, and oxidase components from DTSSP-treated cytosol failed to translocate to the plasma membrane. When DTSSP-treated cytosols were incubated with 2-mercaptoethanol, however, both O2- production and translocation were partly restored, indicating that the functional impairment in DTSSP-treated cytosols was probably due at least in part to a restriction in the conformational mobility of the cross-linked peptide chains in the approximately 240,000 complex. These findings provide further support for the idea that the cytosolic components of the respiratory burst oxidase exist in the form of a approximately 240,000 complex, and suggest that the exposure of this complex to SDS induces a structural change that may or may not be associated with the loss of an inhibitory subunit too small to cause a detectable change in the size of the complex. This SDS-induced change allows translocation to take place by creating a membrane-binding site on the surface of the complex.

Respiratory burst oxidase and three of four oxidase-related polypeptides are associated with the cytoskeleton of human neutrophils.

Resting and phorbol-activated human neutrophils were separated by treatment with Triton X-100 into detergent-extractable and cytoskeleton fractions. Respiratory burst oxidase activity was restricted entirely to the cytoskeleton. The cytoskeleton also contained approximately 15% of the neutrophil cytochrome b558, an oxidase-associated heme protein, as well as most of the oxidase-related cytosolic polypeptide p67phox. In contrast, the components of the oxidase-associated phosphoprotein family p47phox were found almost exclusively in the detergent extract, suggesting that p47phox is needed for oxidase activation but not for O2- production by the activated oxidase. Activation of the oxidase had no apparent effect on the distribution of any of these species between the cytoskeleton and the detergent extract. Our results support earlier studies implying that the cytoskeleton participates in an important way in regulating the activity of the O2(-)-forming respiratory burst oxidase of neutrophils.

Characterization of soluble vs membrane-bound human placental 5'-nucleotidase.

Three forms of 5'-nucleotidase purified from human placenta (two membrane-bound forms, one sensitive and one resistant to cleavage by phosphatidylinositol-specific phospholipase C, as well as a soluble form) had the same molecular weight before (73,000 Da) and after (56,000 Da) digestion with N-glycosidase F and showed similar amino acid compositions, N-terminal amino acid sequences, and KMs for IMP (9.6 to 11.9 microM). Thus, these three forms of 5'-nucleotidase appear to have very similar structures. The form sensitive to phosphatidylinositol-specific phospholipase C contained nearly 1 mol myo-inositol/mol of protein as determined by mass spectrometry, indicating a glycosyl phosphatidylinositol membrane anchor. Soluble 5'-nucleotidase contained a similar quantity of myo-inositol, suggesting that it was previously membrane-anchored via glycosyl phosphatidylinositol. The form resistant to phosphatidylinositol-specific phospholipase C contained less myo-inositol, leaving open the possibility of a third form of 5'-nucleotidase with a conventional transmembrane anchor.

Human T cell activation. Synergy between CD73 (ecto-5'-nucleotidase) and signals delivered through CD3 and CD2 molecules.

Interaction of the glycosyl phosphatidylinositol-linked differentiation Ag CD73 (ecto-5'-nucleotidase) with the CD73-specific mAb 1E9 generates agonistic signals that strongly synergize with T cell activation induced by CD3 and CD2 mAb. This synergy is observed only when 1E9 is immobilized on plastic and occurs in the absence of accessory cells or exogenous lymphokines. 1E9 induces a rapid (though transient) increase in [Ca2+]i in a minor proportion (20 to 30%) of unfractionated T lymphocytes (presumably CD73+ cells). However, this [Ca2+]i mobilization is not sufficient to fully activate CD73+ T cells, as shown by the requirement of additional signals such as CD3 or CD2 stimulation to initiate T cell proliferation. These signals cannot be substituted by the exogenous lymphokines, rIL-1, rIL-2, or rIL-4, or PMA (when T cells are rigorously depleted of monocytes). These data indicate that CD73 may behave as an accessory molecule regulating interactions between T cells and antigens or APC. A comparison was carried out with mAb 9.3 to the differentiation Ag CD28, another agonistic molecule with activating properties similar to CD73. Despite their lower percentage, the ability of CD73+ T cells to amplify the proliferation induced by CD3 or CD2 mAb was equivalent or even greater than that of CD28+ T cells. Once activated, CD73+ cells may recruit the remaining (CD73-) cells primed by CD3 or CD2 stimulation. Based on these data, we suggest that CD73+ T lymphocytes may be a specialized subset to amplify immune responses originated by the CD3 and CD2 activation pathways. Finally, the functional association between CD73 and integral membrane molecules like CD3 and CD2 suggests that GPI-anchored molecules may play a role in transmembrane signaling mediated by conventional second messenger systems.

Production and characterization of monoclonal antibodies to the glycosyl phosphatidylinositol-anchored lymphocyte differentiation antigen ecto-5'-nucleotidase (CD73).

A panel of monoclonal antibodies to the 69 kDa glycosyl phosphatidylinositol anchored lymphocyte differentiation antigen ecto-5'-nucleotidase (ecto-5'-NT, CD73) was produced using highly purified human placental 5'-NT as immunogen. Antibodies 1E9.28.1 and 7G2.2.11 inhibit soluble placental 5'-NT activity and recognize lymphocyte CD73 in indirect immunofluorescence and immunoprecipitation assays. In addition, 1E9.28.1 induces vigorous T cell proliferation in the presence of submitogenic doses of phorbol myristate and F(ab')2 goat anti-mouse Ig. Both antibodies can be used to purify the three major forms of placental 5'-NT by affinity chromatography. By two-color immunofluorescence, CD73 was found to be expressed on 19 +/- 5% of CD3+, 11 +/- 4% of CD4+, 51 +/- 14% of CD8+, 25 +/- 8% of CD28+, 15 +/- 5% of CD29+, 27 +/- 7% of CD45RA+, and 70 +/- 6% of CD19+ lymphocytes. Within T cells, CD73 expression is restricted to the CD28+ subset. Thus, CD73 is found on subsets of both T and B lymphocytes, with the highest expression on B cells and CD8+ T cells. In sections of hyperplastic tonsil, CD73 expression is restricted to the small lymphocytes of the follicular mantle zone, a small subset of extrafollicular lymphocytes situated within the epithelium of the tonsillar crypt, and to follicular dendritic cells within the lower part of the "light-zone." CD73 is also detected on subsets of endothelial cells of capillaries and venules and the basal layer of non-keratinizing squamous epithelium and transitional cell type mucosa of many tissues. Given the tissue distribution of CD73, along with its glycosyl phosphatidylinositol membrane anchoring and the observation that some CD73 antibodies are mitogenic, we propose that this interesting antigen may play a role in cell activation, lymphocyte homing, and/or cell adhesion.

Antibodies to 5'-nucleotidase (CD73), a glycosyl-phosphatidylinositol-anchored protein, cause human peripheral blood T cells to proliferate.

Human peripheral blood T cells were stimulated to proliferate when cultured with submitogenic doses of PMA and goat antibodies to 5'-nucleotidase (5'-NT). The degree of proliferation, as measured by [3H]TdR incorporation on day 3, was similar to that achieved by stimulation with PHA. Anti-5'-NT antibodies had no effect on PHA-induced proliferation. Maximal stimulation was achieved with 0.6 to 1.0 ng/ml of PMA and 125 micrograms/ml of IgG isolated from a goat anti-5'-NT antiserum. Both intact IgG and F(ab')2 fragments were stimulatory. IL-2R expression and IL-2 secretion were also induced by anti-5'-NT antibodies and PMA. Anti-5'-NT-induced proliferation was inhibited greater than 95% by a murine anti-IL-2 receptor mAb and required less than 0.3% monocytes. Similar results have been obtained with a murine mAb specific for 5'-NT. As expected, anti-5'-NT antibodies and PMA did not induce the proliferation of ecto-5'-NT-T cells isolated by cell sorting. Pretreatment of total T cells with phosphatidylinositol-specific phospholipase C removed an average of 89% of the 5'-NT activity from the cell surface and also inhibited by 83% the ability of the cells to proliferate in response to anti-5'-NT antibodies and PMA. Thus, the activation signal provided by anti-5'-NT antibodies is apparently transduced, in large part, by a form of the enzyme that is attached to the membrane via glycosyl-phosphatidylinositol linkage. These data suggest that 5'-NT may play a role in lymphocyte activation as has been proposed for other glycosyl-phosphatidylinositol-anchored lymphocyte surface proteins.

Functional characterization of ecto-5'-nucleotidase-positive and -negative human T lymphocytes.

Functional studies were performed on human peripheral blood T lymphocytes stained with goat anti-5'-nucleotidase antibodies and separated into ecto-5'-nucleotidase (ecto-5'-NT)-positive and -negative populations using the FACSTAR fluorescence-activated cell sorter. On the average, ecto-5'-NT+ T cells contained 34 +/- 13% CD4+ and 55 +/- 15% CD8+ cells, whereas ecto-5'-NT-T cells contained 65 +/- 12% CD4+ and 23 +/- 8% CD8+ cells. Staining with anti-5'-NT antibodies did not significantly alter the ability of unseparated T cells to proliferate in response to PHA or PMA, or in a MLR. However, prior incubation with anti-5'-NT antibodies did inhibit the ability of irradiated T cells to provide help for PWM-stimulated Ig synthesis by as much as 55%. In five separate experiments, ecto-5'-NT-T cells demonstrated an equal or better ability to incorporate [3H]TdR after PHA stimulation or in a MLR, as compared with ecto-5'-NT+ T cells. Similarly, ecto-5'-NT- T cells were not diminished in their ability to provide help for autologous B cells in a PWM-driven system. Clearly, the inability of ecto-5'-NT- T cells from patients with a variety of immunodeficiency diseases to function in these assays cannot be explained solely by their lack of ecto-5'-NT activity. In contrast, ecto-5'-NT-positive and -negative T cells showed markedly different dose-response curves for proliferation in response to PMA. Ecto-5'-NT+ T cells responded to lower doses of PMA (1.0 ng/ml) than did ecto-5'-NT- T cells and showed a two- to eight-fold greater rate of [3H]TdR incorporation at 3 to 10 ng of PMA per ml. Ecto-5'-NT+ T cells may have a protein kinase C that is more accessible or more easily activated or may utilize an alternate pathway of activation when stimulated with low concentrations of PMA.

Synthesis of immunoglobulin G by pokeweed mitogen- or Epstein-Barr virus-stimulated human B cells in vitro is restricted to the ecto-5'-nucleotidase positive subset.

Ecto-5'-nucleotidase (ecto-5'-NT) is believed to be a maturation marker for human B lymphocytes because its expression increases during normal development and is reduced in many patients with B cell immunodeficiencies. To determine whether this enzyme defines functional subsets of B lymphocytes, human peripheral blood B cells, separated into ecto-5'-NT positive and negative populations by using goat anti-5'-NT antibodies and the fluorescence-activated cell sorter, were compared for their ability to secrete polyclonal immunoglobulin. Both populations synthesized equivalent quantities of IgM in response to a T cell-dependent (PWM) or T cell-independent (EBV) stimulator of polyclonal immunoglobulin biosynthesis. However, ecto-5'-NT+ B lymphocytes synthesized 8- to 26-fold more IgG per cell than ecto-5'-NT- B cells. These data provide the first direct evidence that ecto-5'-NT is a marker for the functional maturation of human B cells and support the hypothesis that ecto-5'-NT deficiency in patients with hypogammaglobulinemia results from a block in B lymphocyte maturation.

Distribution of ecto-5'-nucleotidase on subsets of human T and B lymphocytes as detected by indirect immunofluorescence using goat antibodies.

Human T and B lymphocyte subsets were characterized for ecto-5'-nucleotidase (ecto-5'-NT) expression by two-color immunofluorescence by using polyclonal goat antibodies to 5'-NT and murine monoclonal antibodies to T and B cell subsets. Anti-5'-NT antibodies were prepared by immunizing a goat with purified human placental 5'-NT. Lymphocyte surface 5'-NT was detected with F(ab')2 fragments of immune goat IgG followed by biotinylated F(ab')2 rabbit anti-goat IgG and fluorescein isothiocyanate-avidin. Lymphocyte cell surface antigens were detected with phycoerythrin (PE)-conjugated anti-CD3, anti-CD4, anti-CD8, anti-CD16, and anti-CD19. HB-4, an antigen present on a major subset of human peripheral blood B cells, was detected with murine monoclonal anti-HB-4 and PE-anti-mouse-kappa. Analysis showed that ecto-5'-NT was expressed on 32 +/- 7% of CD3+, 19 +/- 6% of CD4+, and 50 +/- 21% of CD8+ T cells, but not on CD16+ lymphocytes. Ecto-5'-NT was also expressed on 81 +/- 8% of adult peripheral blood B cells as defined by PE-anti-CD19; HB-4 was expressed on 84 +/- 7% of CD19+ cells. The two populations of B cells were not identical, however, because HB-4 was co-expressed on only 79 +/- 18% of ecto-5'-NT+ B cells. Two-color immunofluorescent staining of T cells from a patient with congenital agammaglobulinemia and low T cell ecto-5'-NT activity revealed reduced percentages of ecto-5'-NT+ cells in his CD3+, CD4+, and CD8+ populations. Thus, reduced ecto-5'-NT activity by enzyme assay was paralleled by reduced numbers of 5'-NT molecules on the cell surface. Two-color immunofluorescent staining of B cells from a patient with hypogammaglobulinemia and low B cell ecto-5'-NT activity also revealed markedly reduced expression of 5'-NT. HB-4 expression was normal, however, suggesting that the patient's B cells were blocked in maturation subsequent to the acquisition of HB-4 but prior to that of ecto-5'-NT. These results demonstrate that anti-5'-NT antibodies will be valuable tools for analyzing ecto-5'-NT expression and lymphocyte maturation in patients with immuno-deficiency diseases.

Purification of 5'-nucleotidase from human placenta after release from plasma membranes by phosphatidylinositol-specific phospholipase C.

5'-Nucleotidase was purified greater than 1000-fold from human placenta by treatment of plasma membranes with S. aureus phosphatidylinositol-specific phospholipase C and affinity chromatography on Con A Sepharose and AMP-Sepharose. The resulting enzyme had a specific activity of greater than 5000 mumol/hr/mg protein and a subunit molecular weight of 73,000. Goat antibodies against 5'-nucleotidase inhibited enzyme activity and detected 5'-nucleotidase after Western blotting. These antibodies also recognized a soluble form of 5'-nucleotidase and residual membrane-bound 5'-nucleotidase which could not be released by phosphatidylinositol-specific phospholipase C treatment, suggesting that the three forms of the enzyme are structurally related. The soluble 5'-nucleotidase may be derived from the membrane-bound form by the action of an endogenous phospholipase C. The structural basis for the inability of some of the membrane-bound 5'-nucleotidase to be released by phosphatidylinositol-specific phospholipase C is unknown.

Ecto-5'-nucleotidase expression during human B cell development. An explanation for the heterogeneity in B lymphocyte ecto-5'-nucleotidase activity in patients with hypogammaglobulinemia.

Ecto-5'-nucleotidase (ecto-5'-NT) activity was measured in human B cells at different stages of development. Ecto-5'-NT activity of B cell preparations from fetal spleen and cord blood was 5.08 and 5.59 +/- 2.8 nmol/hr/10(6) cells, respectively; that of B cell preparations from adult peripheral blood, spleen, or lymph node was fivefold to sixfold higher (27.9 +/- 12, 29.2 and 33.8 nmol/hr/10(6) cells, respectively). The increased enzyme activity in B cell preparations from adult peripheral blood as compared with cord blood paralleled increased percentages of 5'-NT+ cells (69 +/- 12% vs 32 +/- 17%) and an average of twice as much enzyme activity per positive cell. Small, resting B cells that cannot synthesize Ig in vitro in response to pokeweed mitogen (PWM) were isolated from adult peripheral blood by mouse erythrocyte rosetting. Total ecto-5'-NT activity and the percentage of 5'-NT+ cells were equivalent in total B cells and the mouse erythrocyte rosette-positive subpopulation. Thus, ecto-5'-NT activity is acquired before B cells gain the ability to differentiate into Ig-secreting plasma cells in response to PWM. Ecto-5'-NT activity was also measured in B cell preparations from eight patients with common variable immunodeficiency. Six had reduced ecto-5'-NT activity (2.83 to 15.4 nmol/hr/10(6) cells), and two had normal activity (34.7 and 58.2 nmol/hr/10(6) cells). B cells from all six patients with low ecto-5'-NT activity failed to synthesize Ig when cultured with PWM and normal irradiated T cells. Of the two patients with normal B cell ecto-5'-NT activity, one also had B cells unresponsive to PWM, but B cells from the other patient appeared to more normal, in that they synthesized IgM and IgG when cultured with PWM plus irradiated allogeneic T cells. Thus, measurement of B cell ecto-5'-NT activity allows the subclassification of patients who have a common inability to synthesize immunoglobulin in vitro response to PWM. B cells with low ecto-5'-NT activity are presumably blocked at an earlier stage in development than B cells with normal ecto-5'-NT activity. Evaluation of ecto-5'-NT activity along with the expression of other B cell surface antigens should aid in the definition of discrete stages of B cell development.