Dipeptidyl peptidase I and granzyme A are coordinately expressed during CD8+ T cell development and differentiation.

Dipeptidyl peptidase I (DPPI) is a granule protease that plays a requisite role in processing the proenzyme form of the CTL granule serine proteases (granzymes). This study assesses DPPI mRNA and enzyme expression during T lymphocyte ontogeny and CTL differentiation. The most immature CD3- CD4- CD8- thymocytes were found to express >40-fold higher levels of DPPI mRNA, although levels of DPPI enzymatic activity in CD3- CD4- CD8- thymocytes were only modestly higher than those seen for CD4+ CD8+ or CD4+ CD8- thymocytes. More mature CD8+ CD4- thymocytes and CD8+ splenocytes expressed significantly higher levels of DPPI mRNA and enzymatic activity than CD4+ CD8+ or CD4+ CD8- thymocytes. Granzyme A mRNA expression was observed in DPPI expressing CD3- CD4- CD8- and CD8+ CD4- thymocytes and was also observed in CD8+ CD4- splenocytes; however, expression was not observed in CD4+ CD8+ or CD4+ CD8- thymocytes. Both DPPI mRNA and granzyme A mRNA expression in CD8+ T cells decreased to very low or undetectable levels during the first 48 h after allostimulation in MLCs. However, peak levels of both DPPI and granzyme A expression were observed later in the course of CD8+ T cell responses to alloantigen, with DPPI mRNA expression peaking on either day 3 or day 4 and granzyme A expression peaking at the end of a 5-day MLR. These data indicate that DPPI is expressed at all stages of T cell ontogeny and differentiation in which granzyme A mRNA is detected; consequently, DPPI appears to be available for the processing and activation of granzyme A during both CD8+ T cell development and differentiation.

A selective inhibitor of dipeptidyl peptidase I impairs generation of CD8+ T cell cytotoxic effector function.

CTL express high levels of dipeptidyl peptidase I (DPPI), a granule thiol protease able to convert the zymogen precursors of granzymes A and B into active proteases. In the present studies, the effects of specific inhibition of DPPI on generation of CTL effector functions were examined. When T cell DPPI activity was inhibited by >95% throughout 5-day MLC, a significant reduction in the generation of CD8+ T cell BLT esterase activity (<30% of control) and cytolytic activity (<10% of control) was observed. DPPI inhibition during the second to fourth days of 5-day MLC also was associated with reduced proliferation of CD8+ T cells, but had no effect on CD4+ T cell proliferation or IL-2 production by either population. CTL generated in the continuous presence of DPPI inhibition also exhibited impaired lysis of anuclear erythrocyte targets and diminished killing of nucleated targets by perforin-independent pathways. In contrast, inhibition of DPPI during only the last 24 h of 5-day MLC was associated only with reduced generation of BLT esterase activity and reduced lysis of nucleated targets by perforin-dependent pathways. Repeated or delayed inhibition of DPPI in MLC containing granzyme B-deficient responder cells also impaired generation of cytotoxic activity. These results indicate that DPPI or other DPPI-like protease activities not only are required for the activation of granzymes, but also play a role in the expansion and differentiation of full CD8+ T cell cytolytic activity.

Cloning and characterization of the cDNA encoding mouse dipeptidyl peptidase I (cathepsin C).

The mouse dipeptidyl peptidase I cDNA has been cloned and sequenced and patterns of DPPI mRNA expression in various tissues characterized. Sequences encoding DPPI are highly conserved among mouse, rat and human and include an unusually long propeptide and a distinctive tyrosine adjacent to the active site cysteine.

IL-1 beta convertase (ICE) does not play a requisite role in apoptosis induced in T lymphoblasts by Fas-dependent or Fas-independent CTL effector mechanisms.

Both IL-1beta convertase (ICE) and other members of the ICE-like family of proteases have been reported to play a role in Fas-mediated apoptosis. Con A-stimulated T lymphoblasts generated from splenocytes isolated from ICE-deficient H-2b mice were found to be more susceptible than wild-type lymphoblasts to DNA fragmentation induced by H-2b-specific CTL derived from normal or Fas ligand-deficient gld/gld mice. Trinitrophenyl (TNP)-modified, H-2b target cell-specific CTL were generated from perforin-deficient mice and were found to induce DNA fragmentation only in target cells expressing functional Fas receptors. Similar rates of DNA fragmentation were induced in TNP-modified ICE -/- and ICE +/+ T lymphoblast targets by perforin -/- TNP-modified, H-2b target cell-specific CTL. In addition, anti-Fas Abs induced apoptosis in thymocytes, Con A-stimulated spleen T cells, LPS-stimulated spleen B cells, and thymocytes from ICE -/- mice. However, DNA fragmentation induced by either allospecific FasL-defective CTL, or by perforin-deficient, TNP-modified, H-2b target cell-specific CTL was prevented in ICE -/- target cells loaded by electroporation with Ac-DEVD-CHO, an inhibitor of CPP32 and related ICE family proteases. These findings indicate that ICE does not play a requisite role in Fas-dependent or Fas-independent mechanisms of apoptosis induced in peripheral T lymphoblasts by CTL. However, both major pathways of CTL-induced apoptosis appear to be dependent on the enzymatic activity of other ICE family proteases.

Functional roles for granzymes in murine epidermal gamma(delta) T-cell-mediated killing of tumor targets.

Granzymes, a family of serine proteases contained in cytoplasmic granules of cytotoxic T lymphocytes and natural killer cells, play a critical role in killing tumor targets by triggering rapid breakdown of DNA and subsequent apoptosis. We have reported previously that dendritic epidermal T cells, which are skin-specific members of the tissue-type gamma(delta) T-cell family in mice, are capable of killing selected tumor cell lines. Here we report that short-term cultured dendritic epidermal T-cell lines contain significant N-alpha-benzyloxycarbonyl-L-Lys-thiobenzyl esterase activity, produce granzyme A protein, and express constitutively mRNA for granzymes A and B. Messenger RNA expression for granzyme B was also confirmed in freshly procured Thy-1+ epidermal cells (i.e., dendritic epidermal T cells). Finally, preincubation of dendritic epidermal T cell lines with a granzyme inhibitor, dichloroisocoumarin, but not with a cysteine protease inhibitor, E-64, abrogated completely their capacity to trigger DNA breakdown in YAC-1 target cells. These results reinforce the concept that dendritic epidermal T cells represent skin-resident killer cells that share several functional properties with conventional killer leukocytes, thereby playing a local immunosurveillance role against tumor development.

Interleukin-15 expression by human endothelial cells: up-regulation by ultraviolet B and psoralen plus ultraviolet A treatment.

The purposes of the present study were to determine whether endothelial cells express IL-15 and to evaluate effects of ultraviolet B (UVB) and 8-methoxypsoralens plus UVA (PUVA) on such expression. Cultured human endothelial cells derived from dermis or umbilical veins were subjected to reverse transcriptase-polymerase chain reaction (RT-PCR) and immunoblot analyses for the detection of IL-15 mRNA and protein, respectively. Both dermal and umbilical vein endothelial cells were shown to express IL-15 mRNA and protein, and these markers were upregulated following UVB or PUVA treatment (but not by UVA or 8-methoxypsoralens alone). Also using RT-PCR, dermal and umbilical vein endothelial cells were shown to express IL-2R gamma c mRNA. These results expand the sources of IL-15 in skin to include keratinocytes, dermal fibroblasts, and now endothelial cells. That IL-15 from all three skin cells can be upregulated by UV treatment suggests a role for this cytokine in photosensitive disorders. Finally, the possibility of an autocrine effect of IL-15 on endothelial cells is raised by the expression of IL-2R gamma c in these cells.

Dipeptidyl peptidase I is enriched in granules of in vitro- and in vivo-activated cytotoxic T lymphocytes.

Recent studies have suggested that dipeptidyl peptidase I (DPPI) is the major post-translational processing enzyme responsible for generating activated myeloid and lymphoid granule serine proteases. The current studies assessed the relative levels of DPPI and granzyme A (BLT esterase) in B6 anti-H-2d-specific CTL generated in mixed lymphocyte cultures (in vitro-activated CTL), by infusion of B6 spleen cells into irradiated H-2d mice (graft-vs-host, GVH CTL) or by 1 degree and 2 degrees peritoneal immunization of B6 mice with P815 (H-2d) cells (PE CTL). In contrast to low levels of DPPI activity in unstimulated CD4+ spleen T cells, both unstimulated CD8+ spleen T cells and in vitro-activated CTL populations were several-fold enriched in DPPI activity, while PE CTL and GVH CTL expressed even higher levels of DPPI. Depletion of DPPI-enriched cells by treatment with Leu-Leu-OMe resulted in loss of cytolytic effector function from each CTL population. However, PE CTL and GVH CTL were more sensitive to the toxicity of Leu-Leu-OMe than were in vitro-activated CTL. While standard BLT esterase assays detected much higher levels of this serine protease activity in GVH CTL or in vitro-activated CTL than in PE CTL, levels of BLT esterase activity significantly above the basal levels present in unstimulated CD8+ or CD4+ T lymphocytes were found in association with immunoreactive granzyme A in lysates of PE CTL. In both PE CTL and in vitro-activated CTL, DPPI and BLT esterase activity co-localized in the granule fraction of cell lysates, and similar percentages of total cellular BLT esterase and DPPI were exocytosed upon cross-linking of surface CD3. Thus, both in vivo- and in vitro-activated CTL were found to possess functional granules containing readily detectable albeit somewhat different levels of DPPI and granzyme A activity.

Generation of active myeloid and lymphoid granule serine proteases requires processing by the granule thiol protease dipeptidyl peptidase I.

The proenzyme activation peptides predicted from cDNAs encoding each of the granule serine proteases synthesized by cytotoxic lymphocytes and myeloid and mast cells are composed of 2 residues. The mechanism by which these amino-terminal dipeptides are cleaved to generate the active enzymes has not been elucidated. The comparable distribution of dipeptidyl peptidase I (DPPI) and serine proteases and the ability of DPPI to hydrolyze relevant dipeptide sequences suggested a role for DPPI in the processing and activation of granule serine proteases. This study demonstrates that inhibition of DPPI activity is associated with impairment of the generation of granule serine protease activity in CD8(+) T cells, lymphokine-activated killer cells, P815 mastocytoma cells, and U-937 myeloid cells. Inhibition of DPPI resulted in impairment of the generation of cathepsin G enzymatic activity without reduction in the amount of immunoreactive cathepsin G produced. In U-937 cells pulsed with [3H]isoleucine, inhibition of DPPI activity was associated with the accumulation of the inactive proenzyme form of cathepsin G bearing an amino-terminal dipeptide extension to the isoleucine residue that normally occupies the amino terminus of the enzymatically active protein. These results indicate that DPPI plays a requisite role in the post-translational processing and activation of members of the family of granule serine proteases expressed in bone marrow-derived effector cells.

ATP-stimulated degradation of endogenous proteins in cell-free extracts of BHK 21/C13 fibroblasts. A key role for the proteinase, macropain, in the ubiquitin-dependent degradation of short-lived proteins.

Baby hamster kidney (BHK) 21/C13 cell proteins, labeled with [35S]methionine, [14C]leucine or [3H]leucine in intact cells, were degraded in soluble, cell-free extracts by an ATP-stimulated process. The stimulatory effect of ATP appeared to require ATP hydrolysis and was mediated to a large extent by ubiquitin. Although the cell extracts contained endogenous ubiquitin, supplementation with exogenous ubiquitin increased ATP-dependent proteolysis by up to 2-fold. Furthermore, antibodies against the E1 ubiquitin conjugating enzyme specifically inhibited both conjugation of [125I]ubiquitin to endogenous proteins and ATP/ubiquitin-dependent proteolysis. Addition of purified E1 to antibody-treated extracts restored conjugation and proteolysis. Proteins containing the amino acid analogues canavanine and azatryptophan were also degraded in vitro by an ATP/ubiquitin-dependent process but at a rate up to 2-fold faster than normal proteins. These results indicate that soluble, cell-free extracts of BHK cells can selectively degrade proteins whose rates of degradation are increased in intact cells. Treatment of cell-free extracts with antibodies against the high molecular weight proteinase, macropain, also greatly inhibited the ATP/ubiquitin-dependent degradation of endogenous proteins. Proteolysis was specifically restored when purified macropain L was added to the antibody-treated extracts. Treatment of cell extracts with both anti-macropain and anti-E1 antibodies reduced ATP/ubiquitin-dependent proteolysis to the same extent as treatment with either antibody alone. Furthermore, proteolysis could be restored to the double antibody treated extracts only after addition of both purified E1 and macropain. These results provide strong evidence for an important role for macropain in the ATP/ubiquitin-dependent degradation of endogenous proteins in BHK cell extracts.

Relationships among the subunits of the high molecular weight proteinase, macropain (proteasome).

An analysis of the subunits of the high molecular weight proteinase, macropain (multicatalytic proteinase or proteasome) from human erythrocytes has been conducted using N-terminal amino acid sequencing, gel electrophoresis and reverse-phase peptide mapping. This analysis provided evidence for the existence of 13 subunits of different primary structure. Five subunits were susceptible to the Edman degradation and yielded unique N-terminal sequences. Similarities among these sequences, however, indicated that the subunits are homologous. Two-dimensional gel electrophoresis discriminated 10 major components, which included two of the subunits for which N-terminal sequences had been determined and eight N-terminally blocked subunits. Tryptic peptide mapping indicated that all 10 of these components have a different amino acid sequence. Tryptic peptides from some of the subunits were subjected to amino acid sequence analysis, and the data indicated that all the subunits tested in this way are related by common ancestry. The data suggest that at least nine of the total of 13 subunits are encoded by members of the same gene family; the remaining four subunits have not yet been investigated in sufficient detail to establish their relationships. No evidence for a close relationship with any previously investigated proteinase family has been found. Finally, through a comparison of the 'latent' and 'active' forms of macropain, the study established a close similarity in the subunit composition of these catalytically very different species, although proteolytic degradation of selected subunits appears in the active form of the enzyme.

The latent form of macropain (high molecular weight multicatalytic protease) restores ATP-dependent proteolysis to soluble extracts of BHK fibroblasts pretreated with anti-macropain antibodies.

Specific immunoadsorption of the high molecular weight multicatalytic protease, macropain, from postmicrosomal extracts of BHK fibroblasts inhibited ATP-dependent proteolysis of exogenous protein substrates. The immunoprecipitated macropain represented the latent (L) form of the protease because it had low protease activity but was activated by methods that activate purified macropain L. Reconstitution of the antibody-treated extracts with purified macropain L, but not macropain A, from bovine heart or human erythrocytes, completely restored ATP-dependent proteolysis, even though ATP did not directly activate either purified macropain L or the immunoprecipitated protease. Reconstituted ATP-dependent proteolysis was saturable with respect to added macropain and never exceeded the level of proteolysis present in the original extract. These results indicate that macropain L plays a key role in ATP-dependent proteolysis but suggest that the protease may require interaction with or modification by another cellular component to demonstrate this effect.

The high molecular weight multicatalytic proteinase, macropain, exists in a latent form in human erythrocytes.

The high molecular weight multicatalytic proteinase, macropain, has been purified from human erythrocytes in two forms that differ in caseinolytic activity up to 100-fold. Each form has a native molecular weight of 600,000 and is composed of a number of subunits ranging in molecular weights from 35,000 to 21,000. Although the two proteinase forms share a number of electrophoretically indistinguishable subunits, there are also subunits unique to the respective forms. The less active proteinase represents a latent enzyme because it was fully activated by two procedures including dialysis against water and pretreatment with low concentrations of sodium dodecyl sulfate. These procedures caused differential changes in the caseinolytic and two peptidase activities of the proteinase. An Mr 35,000 subunit, characteristic of latent macropain, is immunologically related to at least one of the other components of active macropain and disappeared after proteinase activation by dialysis. Nevertheless, loss of this subunit was not the cause of the increased activity. These results suggest that the proteolytic activity of cells may be regulated by the activation of the latent form of macropain.

ATP-dependent mechanisms for protein degradation in mammalian cells.

The biochemical basis for ATP-dependent protein degradation observed in intact cells has been studied in cell-free extracts of baby hamster kidney fibroblasts. ATP plays at least two distinct roles in proteolysis. First, ATP is required for the covalent conjugation of ubiquitin to protein substrates. This modification markedly enhances the rates of degradation of some, but not all, proteins. Second, ATP appears to stimulate the activity of a protease capable of degrading both ubiquitinated and non-ubiquitinated proteins. This protease has several biochemical and catalytic features that resemble those of the previously described high molecular weight protease, macropain. Furthermore, antibodies against highly purified human erythrocyte macropain inhibit both ubiquitin-dependent and ubiquitin-independent pathways of ATP-dependent proteolysis. Such results provide evidence for an important role for macropain in ATP-dependent proteolysis in mammalian cells.

An enzyme related to the high molecular weight multicatalytic proteinase, macropain, participates in a ubiquitin-mediated, ATP-stimulated proteolytic pathway in soluble extracts of BHK 21/C13 fibroblasts.

Soluble, cell-free extracts of BHK 21/C13 fibroblasts degraded a variety of exogenous proteins to acid-soluble peptides at pH 8.0. ATP stimulated the rates of proteolysis. Both the absolute rate of proteolysis and the magnitude of the ATP effect depended on the specific substrate. For example, casein was degraded approximately 10-fold faster than lysozyme, but lysozyme degradation was more highly stimulated by ATP than was casein degradation. Ubiquitin enhanced the ATP-stimulated proteolysis of each substrate in both postmicrosomal extracts and DEAE-cellulose fractionated extracts. In each extract, ubiquitin enhanced the ATP-stimulated degradation of lysozyme to a greater degree than that of casein. These results suggested that lysozyme was degraded by a pathway that was more dependent upon ubiquitin than was casein. Further evidence for this conclusion was obtained in studies using substrates whose amino groups were blocked by extensive methylation or carbamoylation. The high molecular weight proteinase, macropain, appears to be involved in the ATP-stimulated degradation of both substrates. Specific immunoprecipitation of macropain with polyclonal antibodies resulted in the inhibition of ATP-stimulated proteinase activity both in the absence and presence of ubiquitin. These results indicate that macropain plays a role in both ubiquitin-mediated and ubiquitin-independent ATP-stimulated proteolysis in BHK cell extracts.

ATP-stimulated proteolysis in soluble extracts of BHK 21/C13 cells. Evidence for multiple pathways and a role for an enzyme related to the high-molecular-weight protease, macropain.

Soluble extracts of cultured cells (BHK 21/C13) degraded a variety of exogenous proteins to acid-soluble peptides at pH 8.0. ATP stimulated this proteolytic activity up to 10-fold. The ATP effect was dependent on Mg2+ and was not elicited by nonhydrolyzable analogs of ATP. After the extract was fractionated on DEAE-cellulose, ATP-stimulated protease activity was in the fraction that bound to the resin and eluted in buffer containing 0.4 M NaCl. This activity had characteristics that were indistinguishable from those of the unfractionated extract but the degree of ATP stimulation was two- to three-fold lower. Although no protease activity was detected in the unbound fraction, reconstitution of this material with the bound fraction enhanced the ATP stimulation up to twofold. The component responsible for the enhancement of the ATP stimulation had properties similar to ubiquitin and purified ubiquitin enhanced the ATP-stimulated protease activity in the fractionated extract. Substrates whose amino groups were almost completely blocked by various chemical modifications were still degraded in an ATP-stimulated fashion, but the degradation of these substrates was not affected by ubiquitin. The protease activity isolated by ion-exchange chromatography was fractionated further by gel filtration chromatography on Sephacryl S-300. ATP-stimulated protease activity eluted with an apparent molecular weight of 750,000. Protease activity was enhanced up to eightfold by Mg2+-ATP but was not increased further by ubiquitin. An activity that hydrolyzed the synthetic peptide Z-Val-Leu-Arg-MNA coeluted with ATP-stimulated protease activity, but peptide hydrolysis was not affected by ATP. These and other catalytic and biochemical characteristics suggested that the protease might be related to the high-molecular-weight protease, macropain, recently purified by us from human erythrocytes (M. J. McGuire and G. N. DeMartino Biochim. Biophys. Acta (1986) 873, 279-289). Antibodies raised against macropain specifically reacted with proteins characteristic of macropain in the column fractions containing ATP-stimulated protease activity. These antibodies also specifically immunoprecipitated 70-100% of the ATP-stimulated protease activity as well as Z-Val-Leu-Arg-MNA hydrolyzing activity. Thus BHK cell extracts appear to contain both ubiquitin-mediated and ubiquitin-independent pathways for the ATP-stimulated degradation of proteins. Furthermore, at least one of these pathways appears to involve a high-molecular-weight, ATP-stimulated protease related to macropain.

Purification and characterization of a high molecular weight proteinase (macropain) from human erythrocytes.

An alkaline proteinase, previously identified in rat liver and heart, has been purified from the soluble fraction of human erythrocytes. The proteinase has an apparent molecular weight of 600 000 and is composed of eight subunits with molecular weights ranging from 32 000 to 21 000. The proteinase degrades both protein and synthetic peptide substrates with a broad pH optimum of 7.5-11.0. Among the synthetic peptides tested, tripeptides with arginine at the P1 position (e.g. Z-Val-Leu-Arg-4-methoxy-2-napthylamine and Boc-Leu-Gly-Arg-4-methylcoumarin-7-amide) are particularly good substrates. The proteinase appears to be sulfhydryl-dependent and is inhibited completely by mersalyl acid and by hemin; inhibitors of serine and metallo-type proteinases have no effect on proteinase activity. Interestingly, a variety of other proteinase inhibitors such as leupeptin, chymostatin and N-ethylmaleimide failed to completely inhibit protein-hydrolyzing activities of the enzyme. These results indicate that these activities may be accounted for by at least two different catalytic sites. Proteinase activity is stable in the presence of 1 M urea, 0.5% Triton X-100 or 0.03% SDS and is not affected by ATP. Based on the high molecular weight and sulfhydryl-dependence, we have named this proteinase macropain.