Molecular cloning, chromosomal localization, and expression of murine dipeptidyl peptidase I.

Dipeptidyl peptidase I (DPPI) is a lysosomal cysteine protease that catalyzes the sequential removal of dipeptides from the amino termini of various protein substrates. We have isolated a cDNA coding for murine DPPI from mouse thymus and spleen cDNA libraries. The deduced amino acid sequence codes for a protein of 462 amino acid residues; comparison of this deduced sequence with that of rat and human DPPI revealed 90.1% and 77.8% identity, respectively. Using DPPI cDNA, we obtained two BAC (Bacterial Artificial Chromosome) clones that contained the murine DPPI locus. The DPPI gene consists of seven exons and 6 introns, and spans approximately 20 kilobases. Using fluorescence in situ chromosome hybridization, we localized murine DPPI to chromosome 7D3-E1.1. We determined that DPPI protein is widely distributed in mouse tissues, although its relative abundance varies from tissue to tissue. In contrast to previous reports, we show here that DPPI mRNA and protein levels and enzymatic activity are unchanged during in vitro T cell activation, implying that this enzyme is not rate-limiting for granzyme processing.

Dipeptidyl peptidase I (DPPI, 1 or cathepsin C) is a lysosomal cysteine protease that belongs to the papain superfamily of proteases (1,2). DPPI is capable of sequentially removing dipeptides from the amino termini of various peptides and protein substrates (3)(4)(5)(6)(7) and, along with other cysteine proteases, is thought to play an important role in intracellular protein degradation and turnover (4,8,9). This enzyme has other postulated functions, including involvement in cell growth (10), neuraminidase activation (11), and platelet factor XIII activation (12). Several reports have also suggested that DPPI activity, although present in a wide variety of tissues, is significantly higher in cytotoxic lymphocytes and myeloid cells (13)(14)(15); in these cells, DPPI localizes to the secretory granule compartment and may be the major enzyme responsible for processing serine proteases from the proenzyme form to the catalytically active form (14,15). Finally, two different reports have shown that DPPI can directly cleave the activation prodipeptides of granzymes A and B (16,17), serine proteases that are found in the secretory granules of activated cytotoxic lymphocytes (CTL) and natural killer cells (18).
DPPI has been purified from several species (human, rat, bovine, porcine, and goat) (4, 19 -23), and the rat and human cDNA sequences have been described (1)(2)24). The amino acid sequences deduced from these cDNAs encode a protein of 462 amino acids (rat) or 463 amino acids (human) with three functional domains: a signal peptide (28 versus 24 residues in rat and human, respectively), an unusually long propeptide (201 versus 206 residues), and the mature region (233 residues in both species). The mature proteins share 87.5% identity with each other, and 28.5-39.5% identity with other members of the papain family (cathepsins H, L, B, O, and K, and papain).
DPPI is initially synthesized as a 55-kDa proenzyme, and is rapidly converted to the two-chain form of the mature enzyme (the enzymes responsible for the proteolytic processing of DPPI are currently unknown). The two chains have molecular masses of 25 kDa (heavy chain) and Ͻ10 kDa (light chain) (25). Unlike most other cysteine proteases, which are small monomeric proteins with molecular masses of 20 -30 kDa, DPPI exists as an oligomeric enzyme of about 200 kDa, as determined by gel filtration (4,6,26,27). There are differing reports concerning the subunit composition of the enzyme, ranging from dimers to hexamers. Most studies suggest that the monomeric subunit is composed of a heavy chain and light chain (both originating from the mature region of the protein); however, a recent report presented evidence that active human DPPI consists of four identical subunits, each composed of three different polypeptide chains, two of them disulfide-linked (28). Two of the chains showed pronounced similarity to the heavy and light chains of the other papain-like cysteine proteases, whereas the third chain corresponds to the proregion of the enzyme. Thus, it appears that a substantial portion of the proregion remains bound to the mature complex. The regulation of DPPI processing and assembly are unknown.
In this paper, we report the sequence of murine DPPI cDNA, and the genomic organization and chromosomal localization of the gene. We determined that murine DPPI protein is widely distributed, with the highest levels found in the spleen, lung, liver, and small and large intestines. Additionally, our studies revealed that DPPI expression and activity are unchanged following in vitro T cell activation.

EXPERIMENTAL PROCEDURES
Molecular Cloning and Sequencing of Murine cDNA-Total RNAs from mouse thymus and spleen were isolated by the guanidinium thio-cynate method as described previously (29). The first strand of cDNA was obtained with reverse transcriptase (Promega) using random hexamer primers. Double-stranded cDNAs were amplified by PCR using Taq polymerase (Perkin-Elmer) using two sets of primers: 1) 5Ј-AACT-GCACTTACCCTGACCTG-3Ј and 5Ј-GGCTTCATTGCAGCCACCATA-G-3Ј (to amplify the 5Ј portion of the cDNA); 2) 5Ј-TGCCCAAGGTTG- The putative initiation methionine is assigned to the first in-frame ATG codon at position 25. The oligonucleotide primer sets 1 and 2 were used to initially screen thymus and spleen cDNAs. Oligonucleotide primer 3 was used to obtain the 5Јuntranslated region. This sequence has been submitted to GenBank™ and has been assigned accession number U89269. B, comparison of the deduced amino acid sequence of murine DPPI with rat and human sequences. Gaps, indicated by dashes, are introduced to optimize the alignment. Asterisks indicate the conserved cysteine, histidine, and asparagine residues in the active site. Underlined sequences indicate the NH 2 -terminal sequences of the pro, mature, and light chains obtained by partial protein sequencing of DPPI (isolated from human spleen) (28). Arrowheads indicate the possible amino termini of the proregion, heavy, and light chains. Bold residues indicate possible glycosylation sites.
TGATGGTGGA-3Ј and 5Ј-CTACAATTTAGGAATCGGTATGGC-3Ј (to amplify the 3Ј portion of the cDNA). These primers were derived from the rat cDNA sequence. The amplified products were subcloned into the pCR2.0 vector (Invitrogen) and transformed into JM109. Plasmids with appropriate insert lengths were sequenced on an Applied Biosystems model 373 automated sequencer, using vector-based external primers and cDNA-specific primers constructed on the basis of previously obtained sequences.
To obtain the 5Ј-untranslated region and leader sequence, a mouse thymus cDNA library in ZAP was screened using PCR, with primers based on ZAP vector sequences and the reverse primer 5Ј-CGGAG-GCAATTCTCCCTTGGT-3Ј, based on the murine cDNA sequence. The amplified products were subcloned into pCR2.0 (Invitrogen) and sequenced.
Molecular Cloning of Murine Genomic DPPI-To obtain the full length genomic DPPI clone, the first 833 bp of the mouse cDNA sequence was used to screen a murine BAC library (Genome Systems, St. Louis, MO). Two clones (BAC 7930 and BAC 7931) hybridized strongly with the cDNA probe. DNA obtained from these clones were digested with EcoRI, and four fragments were subcloned into pUC19 based vectors. These plasmids were sequenced using a series of oligonucleotide primers defined initially by the cDNA sequences and subsequently by derived sequences. To establish the distance between exons, primers flanking the intron/exon borders were designed and PCR amplifications were carried out on the BAC clones using KlenTaqLA (Wayne Barnes, Washington University, St. Louis, MO) and prolonged (5-10 min) extension cycles. Single specific bands derived from each set of primers were directly subcloned into pCR2.0 (Invitrogen) and sequenced to confirm the intron/exon borders.
Southern Analysis-DNA obtained from 129/SvJ ES cells, mouse tails, or from BAC clones was digested with EcoRI and BamHI, electrophoresed on 1% agarose gels, transferred onto Hybond (Amersham) membranes, and probed with either a 833-bp 5Ј cDNA or a 645-bp 3Ј cDNA probe, as described previously (30).
Chromosomes were counterstained with prodidium iodide or 4,6diamino-2-phenylindole (DAPI) and examined with a Zeiss Axiophot epifluorescent microscope with a 100-watt mercury lamp. Digital images of selected metaphase spreads were recorded using cooled chargecouple device camera CH250 (Photometrics, Tuscon, AZ) and a filter system consisting of a triple-band-pass beam splitter and emission filters. Excitation of each of fluochromes used was accomplished by single-band-pass excitation filters in a computer-controlled motorized filter wheel. This made it possible to acquire sequential, registration shift-free gray-scale images of two or three fluorochromes (DAPI, fluorescein isothiocyanate, and/or rhodamine). Images were processed and analyzed on an Apple Power Macintosh 8100/100 computer using Oncor's recording and the analytic program Image, as well as NIH's Image and Yale University's Gene Join. To identify individual chromosomes and to assign the location of signal to specific chromosome regions, a method for direct visualization of fluorescent spots on look-up-table (LUT)-inverted digital images of DAPI banded chromosomes was used (33). To confirm the identity of chromosomes, preparations were hybridized (34) with a mouse chromosome 7 painting probe (Cambio, Cambridge, United Kingdom) and previously observed labeled metaphases were rerecorded.
RNA Preparations and S1 Nuclease Protection Assays-RNA was purified from adult 129/SvJ mouse organs or from activated splenocytes using a guanidinium thiocyanate miniprep protocol, as described (29). Since the first exon of DPPI did not contain any unique restriction sites that would permit the construction of a specific S1 probe, we designed a PCR probe spanning the 5Ј-flanking sequence and the putative first exon, using the primer set 5Ј-CCCAAAGGCCACTTTCCTTC-3Ј (forward) and 5Ј-GAACTTCTAGGGCCCACCTGG-3Ј (reverse). The 478-bp product was gel-purified and end-labeled using [␥-32 P]ATP and polynucleotide kinase as described previously (35). 20 g of total RNA was hybridized with 100,000 cpm of this labeled probe, and an S1 nuclease protection assay was performed as described previously (35). The specific probes for murine granzyme B and murine ␤ 2 microglobulin (m␤ 2 M) have been described previously (36).
Primer Extension Analysis-The reverse primer 5Ј-CAGGAGTGTC-GGAGCGCACGGTGCAGACTCCCAAAAGCAC-3Ј (directed against sequences in exon 1) was end-labeled with [␥-32 P]ATP and polynucleotide kinase. 50 -100 g of total cellular RNA was lyophilized with 100,000 cpm of the labeled oligonucleotide, and primer extension was carried out as described previously (37).

FIG. 1-continued
Production of Activated Lymphocytes-Splenocytes were either activated with 5 g/ml ConA (Sigma) supplemented with 50 units/ml recombinant human IL-2 or in a one-way mixed lymphocyte cultures as described previously (38). For the generation of Ad-Lak cells, splenocytes were cultured in the presence of high dose IL-2 (1000 units/ml) for 10 days as described previously (38).
Western Analysis-Organs from 129/SvJ mice and splenocytes harvested from IL-2/ConA activation were suspended in phosphate-buffered saline, and then an equal volume of 2 ϫ extraction buffer (2 M NaCl, 50 mM Tris, pH 8.0, 0.2% Triton X-100) was added. The extract was sonicated, and crude cellular debris was removed by centrifugation at 15,000 rpm at 4°C ϫ 15 min. Thirty l of this extract was used in each lane of an SDS-PAGE analysis using reducing conditions. Immunoblotting was carried out using a rabbit anti-murine DPPI. Murine anti-DPPI was generated by immunizing rabbits (EL Labs, Soquel, CA) with recombinant DPPI protein (amino acids 230 -462) made in Escherichia coli. The IgG fraction of the antiserum was purified and was shown to react specifically with recombinant DPPI protein (data not shown) and tissue extracts. The antibody recognizes the 25-kDa mature form of the DPPI protein, in agreement with a previous report (2). Murine granzyme B antiserum (38) and murine ␤-actin antiserum (Sigma) were used in control blots. Antibody-protein complexes were detected with a peroxidase-labeled anti-rabbit (DPPI and granzyme B) or anti-mouse (␤-actin) antibody, followed by enhanced chemiluminescence (ECL, Amersham).
DPPI Activity Assay-Bovine DPPI, hydroxylamine hydrochloride, Gly-Phe amide, 2-mercaptoethylamine hydrochloride, trichloroacetic acid, and ferric chloride solutions (all from Sigma) were prepared according to manufacturer's directions. Protein concentrations of cell lysates were assayed using the Bio-Rad method (39) and bovine serum albumin as a standard. DPPI activity was assayed in a 0.50-ml reaction mix (pH 6.8) containing 400 mM hydroxylamine, 50 mM Gly-Phe amide, 25 mM 2-mercaptoethylamine, and 100 l of cell lysate. The reaction was incubated at 37°C overnight and stopped by the addition of 0.50 ml of 20% trichloroacetic acid and 0.50 ml of 5% ferric chloride solution.
The absorbance values at 510 nm for control and test lysates were recorded, and units of DPPI activity/mg of protein were calculated using purified bovine DPPI (Sigma) (assayed simultaneously by the same method over a concentration range of 0.0001-0.1 units) as a standard.

RESULTS
Isolation and Analysis of Murine DPPI cDNA-We used a series of primers based on the rat cDNA sequence and RT-PCR to isolate the cDNA for murine DPPI. We were unable to obtain a full-length cDNA using primers that encompassed the entire mature rat coding region. However, using additional primer sets based on the rat sequence, we were able to amplify several 5Ј and 3Ј fragments from mouse and thymus cDNA, which together encompass most of the mature coding region. PCR fragments from at least three separate reactions for each primer set were subcloned and sequenced to confirm their identity. To obtain the 5Ј-untranslated region and leader se- quence, we used primers based on ZAP sequences and primers derived from the mouse sequence to screen a mouse thymus cDNA library. Fig. 1A shows the mouse cDNA and deduced amino acid sequences. The cDNA contains an open reading frame of 1386 nucleotides. The cDNA sequence isolated did not contain a poly(A) tail; however, a putative polyadenylation signal (AATAAA) is found 385 bp from the TAG stop codon in the genomic sequence ( Fig. 2A). The mouse sequence in the coding region shares 92.8% and 82.1% identity with the rat and human cDNAs, respectively. The percentage of identity for all three species is highest in the region coding for the mature protein (Table I). The initiation codon was assigned to the first in-frame ATG at position 25; the nucleotide sequence around it (AGCATGG) is nearly a perfect match with the Kozak consensus sequence. The deduced amino acid sequence of the open reading frame encodes 462 amino acids, yielding a calculated molecular mass of 52.41 kDa. Alignment of the deduced amino acid sequence of the mouse DPPI with rat and human DPPI revealed 90.1% and 77.8% sequence identity, respectively. The three amino acid residues that form part of the active site of all papain family members (Cys 25 , His 159 , and Asn 197 , according to the papain numbering system) are conserved in the mouse sequence and across species (Fig. 1B). The deduced protein sequences 24 DTPANCTYPD 33 , 230 LPESWDWRN 238 , and 394 DPFNPFELTN 403 (which in human and rat correspond to NH 2 -terminal sequences of the proregion, mature, and light chains; Ref. 28) are also well conserved in the mouse cDNA clone and align exactly with the human and rat sequences. A search for protein modification sites in the deduced amino acid sequence revealed three potential N-glycosylation sites (see Fig. 1B): two in the proregion (at residues 29 and 53) and one in the mature peptide (at residue 275). The positions and sequences of these sites are also conserved in all three species, suggesting that these sites may be glycosylated, potentially serving as recognition signals for targeting the enzyme to lysosomes.
Genomic Organization of Murine DPPI-We used the first 833 bp of the coding sequence to screen a mouse BAC library derived from 129/SvJ mice and obtained two clones (BAC 7930 and 7931) that hybridized strongly with the cDNA probe. An EcoRI and BamHI digest and Southern analysis of these two clones indicated that these BAC clones contain the entire coding sequence (data not shown). EcoRI-cleaved genomic DNA and BAC DNA probed with the 5Ј cDNA fragment revealed three hybridizing fragments (5, 3.8, and 2.3 kb). These fragments were subcloned into pUC19 vectors and subjected to DNA sequencing using a series of oligonucleotides based on the cDNA sequence. The 5-kb EcoRI fragment contains the 5Јflanking sequence, the 5Ј-untranslated sequence, and the ATG start codon with exons 1 and 2 (see Fig. 2). The 2.3-kb EcoRI fragment contains exon 3. The 3.8-kb EcoRI fragment contains exons 4 and 5. A 3Ј cDNA probe identified a 5-kb EcoRI fragment. This fragment contains the 3Ј portion of exons 6 and 7, and the 3Ј-untranslated region. Exon 7 contains a single inframe TAG stop codon, and an AATAAA polyadenylation motif. To confirm the 5Ј portion of exon 6, we used forward primer derived from exon 5 and a reverse primer starting from the internal EcoRI site in exon 6, and obtained a single PCR fragment of 5 kb. This fragment contains intron 5 and the remainder of exon 6 coding region. Once the intron/exon borders were established, we designed primers that flanked the rest of the exon borders and subjected the BAC clones to PCR to establish the intron sizes. The results are shown in Fig. 2; all the intron donor and acceptor sites conform to the consensus rules (40). The entire gene therefore consists of 7 exons and 6 introns, and spans over 20 kb (Fig. 2B).
Murine DPPI Is Located on Chromosome 7-The DPPI gene was localized by FISH using chromosomes prepared from mouse spleen cultures. Symmetrical fluorescent signals on sister chromatids of a medium-size chromosome was observed in 180 out of 200 metaphases hybridized with DPPI probe (Fig. 3,  A and C). The probe had a high specificity for this site, since symmetrical signals were not observed on other chromosomes. In 35 metaphases with satisfactory DAPI G-like banding, the FISH signals were localized relative to the chromosome bands, and the DPPI gene locus was assigned to chromosome 7D3-E1.1 (Fig. 3B). Since the resolution of DAPI banded mouse chromosomes is not always satisfactory, recorded metaphases were also rehybridized with mouse chromosome 7 painting probe. All the chromosomes with the specific hybridization signal were also positive for painting (Fig. 3D). This region is homologous with human chromosomes 15q, 6p, and 11q (41-43).
Defining the 5Ј End of Murine DPPI mRNA-To define the transcription initiation site of DPPI gene, primer extension analysis and S1 nuclease protection assays were performed (Fig. 4). For primer extension analysis, a 40-mer complemen-  tary to mRNA (located in the middle of exon 1) was end-labeled and annealed with 50 -100 g of total cellular RNA derived from mouse spleen or lung tissues. The primer extension product revealed a major band of 150 nucleotides (nt). The strongest band corresponds to the G residue 68 nt upstream from the ATG site, and was designated residue ϩ1 (Fig. 4A). To confirm the location of the 5Ј end of mRNA, an S1 nuclease protection assay was performed. For a DPPI-specific S1 probe, we generated a double-stranded DNA fragment by PCR using a reverse primer derived from exon 1 and a forward primer derived from the 5Ј-flanking region. This probe was end-labeled and annealed to 20 g of total cellular RNA derived from mouse spleen. A probe fragment of 212 nt was protected from S1 digestion by DPPI mRNA. A DNA sequencing ladder, generated using the same exon 1 reverse primer, confirmed that the initiation site is located at the same G residue identified by primer extension (Fig. 4B). These results imply that there are no additional exons and no alternatively spliced 5Ј-untranslated exons. DPPI Is Constitutively Expressed in Many Tissues, and Is Not Up-regulated following Splenocyte Activation-We next examined DPPI protein levels in adult mouse organs using Western blot analysis (Fig. 5) with a polyclonal rabbit anti-DPPI antibody generated against recombinant murine DPPI (see "Experimental Procedures"). The Western blots showed that the 25-kDa mature form of DPPI is detected in almost all tissues, but at varying levels. DPPI is most abundant in lung, liver, spleen, and small and large intestines; the abundance is intermediate in bone marrow, thymus, and stomach, but low in kidney. Heart and brain have very little detectable DPPI protein. S1 nuclease protection assays revealed a similar wide tissue distribution of DPPI mRNA (data not shown).
To determine whether the expression of DPPI is regulated following in vitro lymphocyte activation, we examined DPPI mRNA levels in resting, day 10 Ad-Lak cells (obtained from splenocyte cultures in the presence of high dose Il-2), day 1-5 MLR (Fig. 6), as well as day 2 and day 4 CTL derived from IL-2/ConA-activated splenocytes (Fig. 7A). These RNAs were analyzed using S1 nuclease protection assays with the exon 1 DPPI probe described above, a granzyme B probe (36) (as a marker of lymphocyte activation), and a m␤ 2 M probe (as a control for quality and content of mRNA). The results indicate that granzyme B mRNA is up-regulated following T cell activation, as expected; however, the level of DPPI mRNA is either unchanged (compared with resting levels) or actually decreased following activation.
To extend these results, the same day 0, 2, and 4 IL-2/ConA cells were used to make whole cell lysates that were subjected The probe was hybridized with 50 -100 g of RNA from spleen, lung, or a control sample without RNA. Primer extension was performed as described under "Experimental Procedures" and electrophoresed on an 8% acrylamide gel next to a sequencing ladder generated from the same exon 1 primer. RNA from both spleen and lung gave rise to a major cDNA product of 150 nt. In panel B, an S1 nuclease protection assay was performed. A PCR-generated fragment derived from exon 1 was used in the probe, as depicted at the bottom of the figure. A probe fragment of 212 nt is present in the spleen mRNA lane. This band corresponds to the one identified by primer extension, indicating that the 5Ј end of DPPI mRNA is located at this site.
FIG. 5. Tissue distribution of DPPI. Tissue lysates were obtained as described under "Experimental Procedures," subjected to SDS gel analysis using reducing conditions, and immunoblotted using a polyclonal rabbit anti-mDPPI antibody. Hybridization with a ␤-actin antibody (Sigma) was used to control for protein content; this antibody does not detect cardiac ␤-actin. The origin of additional lower molecular weight bands found in the lane containing stomach lysate is not yet understood.

FIG. 6. DPPI expression in Ad-Lak cells and MLR cultures.
RNAs derived from resting spleen, day 10 Lak cells, and day 1-5 MLR cultures were used in an S1 nuclease protection assay. Total RNA (20 g) was analyzed simultaneously with specific probes for DPPI, granzyme B (as an indicator for lymphocyte activation), and mouse ␤ 2 microglobulin (␤ 2 M) (as a control for RNA content and quality).
to Western blot analysis. Fig. 7B shows these blots sequentially probed for DPPI, granzyme B, and actin (as a control for protein loading). DPPI protein is present in resting splenocytes, but the level did not change substantially with activation. Granzyme B, on the other hand, was not detected in resting splenocytes, but was markedly up-regulated following T cell activation, as expected.
Finally, DPPI activity in these lysates was determined by examining levels of hydrolysis using the specific substrate Gly-Phe amide. IL-2/ConA-activated splenocyte lysates were assayed for DPPI activity and compared with standards prepared from commercially available purified bovine DPPI. As shown in Fig. 7C, approximately equal amounts of DPPI activity were present in all three lysates. DISCUSSION In this report, we characterized the murine DPPI gene and determined that this gene is highly conserved among species.
The murine DPPI gene is localized on chromosome 7 and is a single copy gene. DPPI protein is widely distributed in mouse tissues; its expression is not up-regulated by mitogens or allogeneic stimuli known to activate T lymphocytes.
DPPI exhibits a number of similarities with other lysosomal papain-like cysteine proteases. First, in the mature region of the enzyme, the amino acid sequence of DPPI shares significant homology with the other members of the papain family of cysteine proteases (up to 39% in the case of cathepsin H). It also retains the three amino acid residues that form the active sites of all papain-like proteases. However, unlike other cysteine proteases, DPPI contains a very long proregion, the function of which is still unknown. Although most other cysteine proteases are small monomeric proteins with molecular masses of 20 -30 kDa, DPPI exists as an oligomeric enzyme of about 200 kDa. DPPI therefore represents a distinct class of papain-like proteases. In panel A, RNAs derived from day 0, 2, and 4 IL-2/ConAactivated splenocytes were used in an S1 nuclease protection assay. Total RNA (20 g) was analyzed simultaneously with specific probes for DPPI, granzyme B, and ␤ 2 M. The same day 0, 2, and 4 cells were also used in a Western analysis (panel B). The same blot was probed sequentially with polyclonal antisera specific for DPPI, granzyme B, and actin (as a control for protein content). In panel C, these lysates were analyzed for DPPI activity, using purified bovine DPPI as a standard. The day 0 analysis was repeated six times, and the day 2 and 4 analyses were performed using two separate sets of activated lymphocytes. Enzyme activities were normalized for total protein content.
In this report, we have defined the genomic organization of the murine DPPI gene. Southern blot analysis and chromosomal analysis by FISH revealed that DPPI represents a single locus in the mouse genome. The gene was localized by FISH to chromosome 7 at region D3-E1.1. This region has limited homology to human chromosome 15q 11-ter (41-43), but synteny with this region has not yet been established. Several genes have been localized within the same region of chromosome 7. The oncogene product of feline sarcoma virus and the gene for mitochondrial isocitrate dehydrogenase have been localized to the 7 B3-E1 region; their human homologs were mapped to the distal segment of chromosome 15. The genes for malic enzyme and tyrosinase were localized at 7 E1-E2; their human counterparts are located on chromosome 6p25-24 and 11q21, respectively (41-43). Finally, the mouse hepatic fusion mesoderm development and taupe genes (41-43) are localized within the same region as DPPI.
The murine DPPI gene has seven exons and spans more than 20 kb. Cathepsins B and H both also span more than 20 kb, and comprise 10 and more than 12 exons, respectively (44,45). In contrast, cathepsin L is a relatively compact gene of 8.5 kb, comprising 8 exons (46). The intron-exon organization of cysteine proteases is only partially conserved (Fig. 8). DPPI, and cathepsins H and L share the common intron insertion position that isolates the exon containing the active site cysteine residue (47). This highly conserved domain is split by an intron in cathepsin B (44). DPPI also shares one common intron insertion site with cathepsin B (intron 6-exon 7 for DPPI and intron 5-exon 6 for cathepsin B). In all four genes, the active site histidine and asparagine residues are found within one exon, although the exon on which they reside, and the positions of the intron-exon insertion sites, differ from gene to gene (44,47). These similarities suggest that these cysteine proteases arose from a common ancestral gene, but that their structures have evolved to include intron losses and insertions.
Several studies have suggested that DPPI activity is enriched in in vitro and in vivo activated cytotoxic lymphocytes (14). In addition, DPPI activity was found to colocalize with serine protease activity (such as granzyme A tryptase activity) in the specialized granule fraction of CTL (14). Several studies have also demonstrated the ability of DPPI to directly cleave the prodipeptide of various serine proteases, thereby activating them (16,17). To examine whether the expression of DPPI is regulated during in vitro lymphocyte activation, we examined DPPI mRNA and protein levels, and DPPI activity (using a defined synthetic substrate) in resting splenocytes, or splenocytes activated with various protocols (Figs. 6 and 7). To our surprise, we found that DPPI mRNA were not up-regulated in Ad-Lak cells, MLR-derived cells, or Il-2/ConA-activated cells. Moreover, the protein levels and enzymatic activities seem to correlate with mRNA expression.
The discrepancy between these findings and previously published data (13-15) (measuring DPPI enzymatic activity only) may be due to several factors. In our study, we used unfractionated splenocytes; activated CD8ϩ T cells could potentially contain higher levels of DPPI activity than resting splenocytes or activated CD4ϩ T cells and B lymphocytes. However, we were also unable to find a significant increase in DPPI levels in either MLR or Ad-Lak cells, both of which contain relatively pure populations of activated cells. Also, in one study, the elevated levels of DPPI activity were not detected in all activated populations of CD8ϩ T cells (14). In fact, the level of DPPI activity in in vitro activated CD8ϩ T cells was unchanged compared with resting CD8ϩ T cells, which agrees with our data. Furthermore, despite the wide variations in levels of DPPI activity in various populations of in vitro and in vivo activated CTLs, these T cells all exhibited similar levels of cytolytic activity (14). These results suggest that DPPI is not up-regulated during the activation of T cells, but that its expression is constitutive.
These data do not alter the hypothesis that DPPI plays an important role in the posttranslational processing of serine proteases. Lytic granules (and the serine proteases contained within them) are not found in resting T cells, but are synthesized after T cell activation. The serine proteases undergo proteolytic processing during their intracellular transport and are packaged into secretory granules (with the prodipeptides presumably already cleaved). DPPI therefore may interact with the serine proteases shortly after they leave the Golgi apparatus, although the exact subcellular compartment where this processing takes place is not yet established.
The availability of DPPI probes and antisera will enable us to further study the functions of DPPI in normal and pathological conditions. Expression of purified rDPPI protein will also help to establish the role of this protease in granzyme processing. Finally, the creation of a loss of function mutation in mice, using homologous recombination techniques, should clarify the importance of this enzyme in vivo.