INTRODUCTION

Granule-associated serine proteinases are major constituents of polymorphonuclear leukocytes (PMNL),¹ cytotoxic lymphocytes, and mast cells, accounting for up to 30% of the cellular protein in these cells. They are involved in many physiologic and pathologic processes. Unlike the more extensively characterized serine proteinases such as trypsin, chymotrypsin, or pancreatic elastase that are stored as inactive zymogens in the secretory vesicles of cells and activated only after secretion into the intestinal lumen, granule-associated serine proteinases of leukocytes and mast cells are stored as fully active enzymes. Nevertheless, based on their known cDNA sequences, these immune/inflammatory cell proteinases are initially translated as zymogens and then processed in several steps including the cleavage of signal peptides and the subsequent removal of short propeptides that typically consist of two amino acid residues (1, 2). The cleavage of the propeptides is unusual in that it occurs at an acidic residue in contrast to most proteinase zymogens that are processed at a basic or, rarely, an aromatic residue. Thus, a major mechanism of control of leukocyte or mast cell granule-associated serine proteinases occurs at the level of dipeptidase activation.

Dipeptidyl-peptidase I (DPP-I, EC 3.4.14.1), a cysteine proteinase, was recently demonstrated to play a requisite role in removing the activation dipeptide from many of the leukocyte and mast cell granule-associated proteinases including human cathepsin G, leukocyte elastase, mast cell chymase and tryptase, and lymphocyte granzymes B and H (3-5). DPP-I, originally called cathepsin C, was discovered when extracts of kidney were found to catalyze the hydrolysis of Gly-Phe--naphthylamide (6). It is a lysosomal enzyme widely expressed in many tissues that is felt to be important in intracellular degradation of proteins. The enzyme was purified from human spleen and characterized as a glycoprotein with a pI of 5.4, a molecular mass of 200 kDa as determined by gel filtration under non-denaturing conditions, and a subunit size of 24 kDa (7).

The strong circumstantial evidence that DPP-I is the central coordinator for activation of many serine proteinases contained in immune/inflammatory cells and that it is differentially expressed in human tissues (8) emphasizes the need for in-depth studies to define factors regulating its expression. A human and a rat DPP-I cDNA have recently been cloned (9, 10), but the reported sequences contained only a portion of the 5-untranslated regions (UTRs). Moreover, information on gene expression and regulation has not been reported. In the present investigation we describe the structure, localization, and expression of the gene for DPP-I. We also demonstrate its regulation in cytokine-stimulated lymphocytes.

EXPERIMENTAL PROCEDURES

Maximum strength Nytran membranes were from Schleicher & Schuell, Inc. Multiple tissue Northern blots, ExpressHyb hybridization solution, human spleen total RNA, and Marathon cDNA amplification kits were from CLONTECH. MicroFastTrack kits, cDNA cycle kits, and TA Cloning kits were from Invitrogen. LA PCR kits were from Panvera. TRI REAGENT was from Molecular Research Center Inc. [-³²P]ATP (3000 Ci/mmol) and [-³²P]dCTP (3000 Ci/mmol) were from Amersham Life Science, Inc. Sequenase DNA sequencing kits were from U. S. Biochemical Corp. RPMI 1640, minimum Eagle's medium, nonessential amino acids, and Cot1 DNA were from Life Technologies, Inc. Defined fetal bovine serum was from Hyclone Laboratories (Logan, UT). Hybrisol VI was from Oncor Inc. Actinomycin D, cycloheximide, and other chemicals not specifically mentioned were high quality grade from Sigma.

Poly(A)⁺ RNA was isolated from human spleen using a MicroFastTrack kit. The first strand cDNA was synthesized with a combination of oligo(dT) and random primers. A portion of the cDNA was amplified using the PCR primers that represent the 5 (743-760 nt) and 3 (complementary to 1429-1446 nt) termini for the rat mature protein coding region (10). The PCR product represented 756-1455 nt of human DPP-I cDNA and was used as a probe to screen a human genomic PAC (1 bacteriophage-derived rtificial hromosome) library (GenomeSystems, Inc.). Phage DNA from the PAC clones was purified by alkali lysis, and the insert was released with NotI followed by digestion with EcoRI. The digested DNA was analyzed by Southern blot hybridization with the human DPP-I cDNA probe used in the library screening, and the relevant DNA fragments were purified from the gel using a Prep-A-Gene kit. To determine the exon and intron organization, fragments of genomic DNA were amplified by PCR and sequenced. To locate intron sequences, the following oligonucleotide primer pairs were selected to amplify overlapping regions spanning the entire length of the human DPP-I cDNA: 1) 13-37 nt and complementary to 855-878 nt, 2) 855-872 nt and complementary to 1438-1455 nt, and 3) 1385-1408 nt and complementary to 1661-1684 nt. The primer pairs used to obtain the 5-flanking sequence and identify the polyadenylation site were 1) T7 sequencing primer and complementary to 139-162 nt and 2) 1621-1641 nt and SP6 sequencing primer.

To obtain data on the intron size and splice junction site, long and accurate PCR was performed to amplify the fragments of genomic DNA using a GeneAmp^TMPCR system 9600. The PCR amplification reaction consisted of an initial denaturation at 94 °C for 1 min, followed by 30 cycles of denaturation at 94 °C for 15 s, annealing at 62 °C for 15 s, and extension at 72 °C for 2 min. Each PCR product was analyzed on an agarose gel, directly subcloned into the pCR^TMII vector using the TA cloning kit, and sequenced.

The genomic plasmid clone was used as a probe for chromosomal localization of DPP-I by fluorescence in situ hybridization. The genomic clone was nick translation-labeled with biotin, hybridized to metaphase chromosomes, and detected with Cy3-conjugated streptavidin. Human metaphase chromosome spreads were prepared by standard procedures and G-banded after trypsin treatment and Wright's staining. Hybridization and detection conditions on metaphase chromosomes were performed as described previously (11). Briefly, the G-banded preparations were destained with a fixative containing methanol and glacial acetic acid (3:1), dehydrated by successive washings in 70, 90, and 100% ice-cold ethanol, and dried at 37 °C. Probe signals were detected with Cy3 conjugate viewed through a triple pass filter using an epifluorescence microscope. The fluorescence in situ hybridization image was overlaid on the G-banded metaphase image to localize the gene.

To determine the transcription start site, anchored PCR was performed. Briefly, the first strand cDNA was synthesized at 50 °C using spleen total RNA (1 µg) and a gene-specific primer that was complementary to 855-878 nt. The 5-end of purified cDNA was (dA)-tailed with terminal deoxynucleotidyltransferase and anchored. The second strand cDNA was synthesized using oligo(dT) containing a 3 rapid amplification of cDNA ends adapter primer. After purification, the double-stranded cDNA was amplified by PCR using a primer pair of an abridged universal amplification primer (Life Technologies, Inc.) and a gene-specific primer complementary to 828-851 nt. The PCR product was re-amplified with an abridged universal amplification primer and a nested gene-specific primer complementary to 751-774 nt. The PCR product was analyzed on an agarose gel, subcloned into PCR^TMII vector, and sequenced.

Primer extension was performed to confirm the transcription start site. Briefly, the 18-nt primer complementary to 17-34 nt was end-labeled with [-³²P]ATP using T4 polynucleotide kinase. An annealing reaction was carried out with 100 fmol of labeled primer and 25 µg of human spleen total RNA at 58 °C for 20 min. To this reaction, avian myeloblastosis virus reverse transcriptase (1 unit) was added to a 20-µl final volume containing 1 mM dNTPs, 50 mM Tris, pH 8.3, 50 mM KCl, 10 mM MgCl₂, 10 mM dithiothreitol, and 0.5 mM spermidine. The extension reaction was carried out at 42 °C for 30 min and terminated by adding 20 µl of loading dye. The primer extension samples were boiled for 10 min and loaded onto a 6% sequencing gel. After electrophoresis, the gel was dried and developed.

Multitissue blots were used to determine expression of the DPP-I gene within human tissues. For studies of gene expression in immune/inflammatory cells, U937, PLB 985, or HL-60 cells were grown as described previously (12). Studies of gene regulation were conducted in IL-2-stimulated lymphocytes. Experiments were performed on cells having viabilities of >95% as judged by trypan blue exclusion. Cells were changed to fresh medium before the exposure to possible agonists/modulators at concentrations specified for the indicated periods of time. The cells were harvested at the end of each time period and kept frozen at 80 °C until further analysis. Actinomycin D stock (5 mg/ml) was prepared in 95% (v/v) ethanol. Cycloheximide stock (10 mg/ml) was prepared in phosphate-buffered saline.

Total RNA was prepared by the acid guanidinium thiocyanate phenol/chloroform method (13) and quantified by measuring absorbance at 260 nm. RNA (10 µg/lane) was size-fractionated on 1% agarose, 0.4 M formaldehyde gels containing formamide and transferred to Nytran membranes by capillary action. The RNA was cross-linked to the membrane by exposure to UV light, prehybridized at 68 °C for 30 min, and hybridized with an [-³²P]dCTP-labeled probe at 68 °C for 1 h. After several low stringency washes, the blot was washed twice at high stringency and developed.

RESULTS

Overall, the human DPP-I gene spans 3.5 kb and contains two exons separated by 1645 nt of intronic DNA (Fig. 1). The first exon comprises the 5-UTR followed by 889 nt that encode the signal peptide, propeptide, and partial mature protein region. The second exon contains the remainder of the mature protein-coding sequence of 501 nt, the stop codon, and the 3-UTR including a polyadenylation signal. The location of the intron was confirmed by sequence analysis. As shown in Fig. 1, the exon-intron boundaries conform to classical splice donor and acceptor consensus sequences (14). The single intron splice site occurred between nucleotides 952 and 953, the first and second nucleotides of the codon for Gly²⁹⁷ in the cDNA sequence, which was indicative of a phase 1 intron. The exon sequence agrees with that determined for the cDNA, indicating that the obtained cDNA using reverse transcription PCR is free of PCR artifacts.

The cDNA is similar in size and composition to that recently reported for human ileum DPP-I (9) but contains an additional 30 nt of the 5-UTR, including the transcription initiation site (see below). The DPP-I cDNA sequence spans 1888 nucleotides and includes a 1392-nt open reading frame that encodes for 463 amino acids. This includes a 24-amino acid residue signal peptide, a 206-amino acid residue propeptide region, and a 233-amino acid residue mature enzyme. The coding region of human DPP-I shows 78% identity to the rat at nucleotide and amino acid levels. The mature protein shows 88% identity at nucleotide and amino acid levels with the respective rat sequences. The cDNA sequence differs from that reported for the ileum at nucleotides 276 (C  G, Leu⁷³ unchanged) and 1440 (G  A, Pro⁴⁵⁹ unchanged) in the protein coding region and at nucleotides 1461 (C  G), 1503 (A  G), and 1861 (G  A) in the 3-UTR. In addition, there is a five-nucleotide (ACTGC) deletion immediately 5 to the poly(A)⁺ tail in the spleen cDNA when compared with that of the ileum. These five nucleotides (ACTGC) preceding the poly(A)⁺ tail reported for human ileum DPP-I by Paris et al. (9) are identical to those we observed in the genomic sequence. The basis for this difference is not presently known but may result from the existence of limited genetic variability.

Of the 20 metaphase cells that were located, all showed Cy3 signals on the long arms of chromosome 11. Fourteen of 20 showed four hybridization signals (one per chromatid, two on each chromosome 11) whereas 6 showed only one signal on one chromosome 11 and two signals on the other chromosome 11. No other chromosomes showed signals with the genomic probe, suggesting a single genomic sequence with high homology to the DPP-I gene locus. Imaging techniques further localized DPP-I to 11q14.1-q14.3 (Fig. 2) with a 92.5% efficiency of hybridization.

Sequence analysis of independent subclones generated by anchored PCR indicated that transcription initiates at the A nucleotide located 63 nt upstream from the ATG that represents the translation initiation codon. Primer extension analysis with the antisense primer positioned 30 nt 5 to the ATG yielded a 34-nt-long product (Fig. 3) calculated to end at the A nucleotide located 63 nt 5 from the ATG and is, therefore, consistent with the results obtained by anchored PCR. The sequence of the region encompassing the transcription initiation site revealed that the A nucleotide is preceded by an invariant C nucleotide and matches the consensus cap signal that has been found in the majority of eukaryotic promoters (15).

Identification of Putative Regulatory Elements in the 5-Flanking Sequence of Human DPP-I Gene

The 5-regulatory region was determined by sequencing the ~3-kb PCR fragment that contained a portion of the first exon and the adjoining upstream region (Fig. 4). Computer analysis of this region revealed several features that are characteristic of promoters. The first 200 nt of the immediate 5-region relative to the transcription initiation site is GC-rich (65%) as compared with the GC composition (50%) of the whole 5-region. Further analysis of this region revealed neither a classical TATA box nor a CCAAT box. However, a potential cis-acting DNA element, GC-box/imian virus 40 rotein (Sp1) binding site (position 55 in reverse orientation) (16) was identified. The 5-region contains recognition sequences for several other transcription factors including three sites for the yclic AMP esponse lement inding rotein (CRE-BP) (17) at positions 519, 950 (reverse orientation), and 953; five sites for CAAT/nhancer inding rotein (C/EBP) (18) at positions 480, 531 (reverse orientation), 660, 732 (reverse orientation), and 665; two sites for NFB/c-Rel (19) at positions 637 (reverse orientation) and 797; and two sites for Oct-1 (20) at positions 897 and 1115 (reverse orientation). Other sites of interest in the 5-region include binding sites for several cell-specific transcription factors involved in the proliferation and differentiation of hematopoietic cells. These include five yeloid inc inger (MZF1) sites (21) at positions 73 (reverse orientation), 116 (reverse orientation), 349 (reverse orientation), 435, and 1070; nine GATA (22) family binding sites including four GATA-1 sites at positions 76 (reverse orientation), 731 (reverse orientation), 939, and 1055 (reverse orientation), four GATA-2 sites at positions 76 (reverse orientation), 550, 731 (reverse orientation), and 939, and one site for GATA-3 at position 77 (reverse orientation); two sites for aros (IK-2) (23) at positions 691 and 1074 (reverse orientation); one site for the mphoid transcription actor (Lyf-1) (24) at position 1075 (reverse orientation); three sites for v-Myb (25, 26) at positions 563, 654, and 969; one site for the uclear espiratory actor (NRF-2) (27) at position 1054 (reverse orientation); one site for Pbx-1 (28) at position 1111; one site for E1A (arly region of adenovirus)-associated -kDa rotein (p300) (29) at position 357 (reverse orientation); and several sites for CdxA (30). Potential recognition sequences for the Ets (26 ransformation pecific) family of transcription factors c-Ets-1 (31) and ts-ie (Elk-1) (32) are also present in reverse orientation at positions +29, 42, 1052, and 1054 and at position 39, respectively.

Nucleotide sequence of the 5-flanking region of human DPP-I gene.

The expression pattern of DPP-I in adult human tissues was determined by Northern blot analysis. The human DPP-I cDNA probe hybridized to a transcript of ~2 kb in all the tissues but with varying intensities. The strongest signal for human DPP-I mRNA was detected in the lung, kidney, and placenta. A signal of moderate intensity was detected in the small intestine, colon, spleen, and pancreas. A low intensity signal was observed in the heart, reproductive organs (testis and ovary), and peripheral blood leukocytes. A weak signal was present in the thymus, prostate, liver, and skeletal muscle. Transcripts were barely detectable in the brain.

Because of our interest in DPP-I as a central coordinator in activating granule-associated serine proteinases, we determined the expression of human DPP-I mRNA in immune/inflammatory cells and their precursors. A representative Northern blot is shown in Fig. 5. Among the fully differentiated cells, the strongest hybridization signal was observed in PMNL and alveolar macrophages. A weak signal was detected in unstimulated lymphocytes and monocytes. Among precursor cells, strong signals were observed in PLB 985, a myelomonoblastic cell line, U937, a myelomonocytic cell line, and HL-60, a promyelocytic cell line.

Because granzymes present in lymphokine-activated lymphocytes are presumably activated by DPP-I that is present in only low levels in unstimulated lymphocytes, we studied the regulation of the human DPP-I gene in lymphocytes stimulated by IL-2. As shown in Fig. 6, low levels of human DPP-I mRNA were present in lymphocytes not exposed to IL-2. When lymphocytes were exposed to IL-2, induction of human DPP-I mRNA occurred as early as 12 h, peaked at 48 h, and then declined by 72 h. Actinomycin D prevented the induction of human DPP-I mRNA in lymphocytes stimulated by IL-2. Treatment of lymphocytes with cycloheximide also prevented the induction of human DPP-I mRNA by IL-2 (data not shown). These results indicate that the induction of human DPP-I mRNA observed in lymphocytes treated with IL-2 most likely occurred at the level of gene transcription and was dependent on protein synthesis.

DISCUSSION

DPP-I is a lysosomal cysteine proteinase differentially expressed in a variety of tissues and thought to play an important role in intracellular protein degradation. To date, it is the only cysteine proteinase that has been demonstrated in PMNL, and recent studies have focused on the importance of this enzyme as a central coordinator for activation of granule-associated serine proteinases contained in PMNL and mast cells. In this investigation, to better understand the molecular basis for the regulation and physiologic effects of DPP-I and to gain insight into its tissue-specific and cytokine-induced expression, we have determined the organization, chromosomal location, and expression of the human DPP-I gene.

DPP-I has been classified as a member of the lysosomal papain-type cysteine proteinase family that also includes cathepsins B, H, L, O, and S. This classification is based on its localization within the cell, acidic pH optimum for enzyme activity, and conserved amino acid sequence with respect to the NH₂-terminal and COOH-terminal regions that form the substrate-binding pocket of the enzyme. However, unlike cathepsin B, H, L, O, and S, which are monomeric proteins (molecular mass 20-30 kDa) with endopeptidase activity, DPP-I is an oligomeric protein (200 kDa) with exopeptidase activity. In addition, the overall amino acid sequence homology of DPP-I shows relatively little identity with other members of this group of proteinases.

We now report that the organization of the human DPP-I gene contrasts strikingly with that of the other enzymes contained within the papain group. The human DPP-I gene is of limited size and complexity, existing as a single copy that spans approximately 3.5 kb, contains two exons divided by a single intron, and is expressed as a single transcript. Reports of genes previously described for cathepsin B (33), cathepsin H (34), cathepsin L (35), and cathepsin S (36) emphasize complex structures consisting of multiple exons and introns, some undergoing alternative splicing that gives rise to multiple transcripts that are differentially expressed. Recently, an alignment/phylogeny of the papain superfamily of cysteine proteases was created (37) in which cathepsin B and DPP-I were placed in the same class, appearing to have diverged from the other papain group of sequences before the origin of kinetoplastids. However, the grouping of cathepsin B and DPP-I was not well supported statistically. Based on the results of the current investigation demonstrating the strikingly different genomic organization, we question the grouping of cathepsin B and DPP-I and speculate that rather than having a common ancestral origin with the other mammalian cysteine proteinases of the papain superfamily, DPP-I may have evolved into the class through convergence by selective evolutionary pressure. Also of note, human DPP-I is neither located on the chromosomes of other cysteine proteinase groups (36, 38, 39) in which it is classified nor on chromosomes of granule-associated serine proteinases (40) to which it functions as a processing enzyme.

To begin to address the regulation of human DPP-I gene expression, we analyzed the 5-flanking sequence for potential upstream regulatory elements. The transcription initiation site is located 63 nt from the translation initiation site and is surrounded by a canonical cap signal sequence. Consensus transcription sequences such as TATA and CCAAT are notably absent in the GC-rich upstream region of the cap site. Eukaryotic promoters lacking a TATA or CCAAT sequence frequently encode proteins or enzymes with housekeeping functions. Most of these constitutively expressed genes exhibit multiple transcription initiation sites distributed over a limited region (41) and multiple Sp1 binding sequences. In contrast, the human DPP-I promoter has a single transcription initiation site and a single Sp1 site. This is similar to the cathepsin S (36), the thrombin receptor (42), and the nerve growth factor receptor (43) genes that are also subject to regulated expression.

The 5-flanking region of the human DPP-I gene contains putative regulatory elements. Many of these elements (e.g. MZF1, v-Myb, GATA) have been shown to be important in proliferation and differentiation of hematopoietic cells. The promoter region contains T cell-specific transcription factor binding sites (e.g. IK-2/Lyf-1, p300) as well as NF-B recognition sites reported to be involved in the cytokine-stimulated gene expression (44). Further functional analysis of the promoter region will be necessary to determine which of the factors are involved in the regulation of the human DPP-I gene.

It has been reported previously that DPP-I mRNA is widely expressed in all rat tissues (10) and that the relative level of message in different tissues mirrored the protein content (45) as well as the enzyme activity (8). This also appears true for the human, where the level of expression of DPP-I transcripts observed in the current investigation is in close agreement with reports of the distribution of DPP-I enzyme activity in tissues (8) and hematopoietic cells (3, 7). However, there are notable differences in the expression of DPP-I in the human and the rat. In the human, for example, DPP-I is expressed at low or barely detectable levels in the liver and brain, whereas in the rat DPP-I is highly expressed in the liver and moderately expressed in the brain. These results suggest species-specific expression of the enzyme.

Importantly, the pattern of expression of DPP-I transcripts in immune/inflammatory cells is distinct from that observed for granule-associated serine proteinases. Results from the current investigation suggest that DPP-I is expressed at all stages of myeloid cell development. In contrast, mRNA expression of granule-associated serine proteinases is restricted to specific stages of myeloid cell development (46-51). This suggests a role for DPP-I in immune/inflammatory cells that extends beyond that of a processing enzyme for the granule-associated enzymes.

In summary, we isolated and characterized the human DPP-I gene including a 1.2-kb promoter region. The gene contains a single intron and maps to chromosome 11q14.1-q14.3. The putative promoter region has neither consensus TATA nor CCAAT sequences, a characteristic of housekeeping genes, but it appears to be regulated, at least in certain settings. Further studies are needed to determine the basis for this regulated expression.