Post-transcriptional regulation of chymase expression in mast cells. A cytokine-dependent mechanism for controlling the expression of granule neutral proteases of hematopoietic cells.

Although all mouse mast cells are derived from a common progenitor, these effector cells exhibit tissue-specific differences in their expression of the chymase family of serine proteases whose genes reside on chromosome 14. Immature bone marrow-derived mast cells (mBMMC), developed in vitro with interleukin (IL) 3-enriched medium, were cultured in the presence or absence of IL-10 to determine at the molecular level how the expression of the individual chymases is differentially regulated. As assessed by RNA blot analysis, mBMMC contain high steady-state levels of the transcript that encodes mouse mast cell protease (mMCP) 5, but not the homologous chymase transcripts that encode mMCP-1, mMCP-2, or mMCP-4. Nevertheless, nuclear run-on analysis revealed that these cells transcribe all four mast cell chymase genes. IL-10 elicited high steady-state levels of the mMCP-2 transcript, and pulse-chase experiments revealed that the half-life of the mMCP-2 transcript in mBMMC maintained in the presence of IL-10 is 4-fold longer than that in replicate cells subsequently cultured in medium without IL-10. Reverse transcription-polymerase chain reaction/nucleotide sequence analysis demonstrated that mBMMC cultured in the absence or presence of IL-10 correctly process mMCP-2 pre-mRNA. Experiments with cycloheximide and actinomycin D indicated that IL-10 induces expression of a trans-acting factor(s) that stabilizes the mMCP-2 transcript or facilitates its processing. The discovery that the expression of certain chymases in mBMMC is regulated primarily at the post-transcriptional level provides a basis for understanding the mechanism by which specific cytokines dictate expression of the chromosome 14 family of serine proteases in cells that participate in inflammatory processes.

Although all mouse mast cells are derived from a common progenitor, these effector cells exhibit tissue-specific differences in their expression of the chymase family of serine proteases whose genes reside on chromosome 14. Immature bone marrow-derived mast cells (mBMMC), developed in vitro with interleukin (IL) 3-enriched medium, were cultured in the presence or absence of IL-10 to determine at the molecular level how the expression of the individual chymases is differentially regulated. As assessed by RNA blot analysis, mBMMC contain high steady-state levels of the transcript that encodes mouse mast cell protease (mMCP) 5, but not the homologous chymase transcripts that encode mMCP-1, mMCP-2, or mMCP-4. Nevertheless, nuclear run-on analysis revealed that these cells transcribe all four mast cell chymase genes. IL-10 elicited high steady-state levels of the mMCP-2 transcript, and pulse-chase experiments revealed that the half-life of the mMCP-2 transcript in mBMMC maintained in the presence of IL-10 is ϳ4-fold longer than that in replicate cells subsequently cultured in medium without IL-10. Reverse transcription-polymerase chain reaction/ nucleotide sequence analysis demonstrated that mBMMC cultured in the absence or presence of IL-10 correctly process mMCP-2 pre-mRNA. Experiments with cycloheximide and actinomycin D indicated that IL-10 induces expression of a trans-acting factor(s) that stabilizes the mMCP-2 transcript or facilitates its processing. The discovery that the expression of certain chymases in mBMMC is regulated primarily at the post-transcriptional level provides a basis for understanding the mechanism by which specific cytokines dictate expression of the chromosome 14 family of serine proteases in cells that participate in inflammatory processes.
Mast cells are derived from multipotential hematopoietic stem cells, circulate in blood as immature cells, and undergo the final stages of their differentiation and maturation after their agranular progenitors become lodged in tissues (1)(2)(3)(4)(5). As assessed by their expression of three different classes of granule proteases, at least four phenotypically different types of mast cells are present in the tissues of the BALB/c mouse. Mast cells express at least four chymotrypsin-like neutral proteases (designated chymases) whose genes reside at a complex on chromosome 14 (6), along with the genes that encode cathepsin G and granzymes B, C, E, and F (7)(8)(9). Mast cells in the intestinal mucosa preferentially express mouse mast cell protease (mMCP) 1 1 and mMCP-2 (5, 10 -14), the mast cells in the peritoneal cavity and skin preferentially express mMCP-4 and mMCP-5 (15)(16)(17), and those in the spleen express every known granule chymase (5).
Based on in vitro studies, it appears that much of the granule protease pleiotropism of the mast cell is cytokine regulated. The immature mast cells (mBMMC) generated by culturing BALB/c mouse bone marrow cells in interleukin (IL) 3-enriched medium (18 -20) contain high steady-state levels of the transcripts that encode mMCP-5 (17) but not mMCP-1 (13), mMCP-2 (12), or mMCP-4 (16). However, these mBMMC will also express high steady-state levels of the mMCP-1 and mMCP-2 transcripts when cultured in the presence of IL-9 (21) or IL-10 (14,22), or high steady-state levels of the mMCP-4 transcript when cultured in the presence of c-kit ligand (23). Furthermore, the steady-state levels of the mMCP-1 and mMCP-2 transcripts in BALB/c mBMMC can be reversibly altered by adding or withdrawing IL-10 from the culture medium (14,22). Half-maximal levels of both chymase transcripts are achieved after the cells are cultured with IL-10 for only ϳ24 h, and plateau levels of the two chymase transcripts are reduced to half-maximum within 24 h of IL-10 withdrawal. The rate by which the steady-state levels of the two protease transcripts in these nontransformed mast cells can be increased or decreased raised the possibility that post-transcriptional control mechanisms might contribute to mast cell protease heterogeneity.
We now demonstrate that the IL-10-regulated expression of the chymase mMCP-2 in BALB/c mBMMC occurs primarily by a post-transcriptional mechanism. This discovery has broad implications for understanding how the expression of the chromosome 14 superfamily of serine proteases is regulated in mast cells and other hematopoietic cells during inflammation.

EXPERIMENTAL PROCEDURES
Cell Culture-mBMMC were obtained by culturing bone marrow cells from the femurs and tibias of 6 -10-week-old female BALB/c mice (Jackson Laboratory, Bar Harbor, ME) in IL-3-enriched 50% WEHI-3 * This work was supported by Grants AI-23483, AI-22531, AI-31599, AI-07306, AR-07530, AR-36308, and HL-36110 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  cell (line TIB-68; American Type Culture Collection, Rockville, MD) conditioned medium and 50% enriched medium (RPMI 1640 containing 100 units/ml penicillin, 100 g/ml streptomycin, 100 g/ml gentamicin, 2 mM L-glutamine, 0.1 mM nonessential amino acids, 50 M ␤-mercaptoethanol, and 10% fetal calf serum) for 2-6 weeks at 37°C in a humidified atmosphere of 5% CO 2 (20). Every 7 days, the nonadherent cells in each culture were transferred to a new flask containing fresh medium. In all experiments in which BALB/c mBMMC were treated with IL-10, cells were cultured in WEHI-3 cell-conditioned medium supplemented with 100 units of COS cell-derived IL-10 (22,24). The MC-9 cell line was used to standardize the level of biologically active IL-10 in the COS cell supernatant; 1 unit of IL-10 elicits a half-maximal incorporation of [ 3 H]thymidine into DNA. Cell viability was assessed by a dye exclusion assay. Routinely, cultured mBMMC (10 5 -10 6 cells/ml) were mixed with an equal volume of trypan blue (Life Technologies, Inc.), and the numbers of cells that excluded the dye were quantitated with a microscope equipped with a ϫ 40 objective.
RNA Blot and Reverse Transcription-Polymerase Chain Reaction (RT-PCR) Analyses-Total RNA was extracted (25) from BALB/c mBMMC cultured in the presence or absence of IL-10 with Tri Reagent™ (Molecular Research Center, Cincinnati, OH) and was quantitated by measuring the optical density at 260 nm. For blot analysis, RNA samples were denatured in formaldehyde/formamide and electrophoresed in 1.3% formaldehyde-agarose gels. The separated RNAs were transferred to Magna-Graph membranes (Micron Separations, Westover, MA) (26), and the resulting blots were individually probed with [ 32 P]dCTP (ϳ3000 Ci/mmol; DuPont NEN) random prime-labeled gene-specific probes for mMCP-1 2 (13,27), mMCP-2 (12), mMCP-4 (13, 16), mMCP-5 (17), or mouse ␤-actin (28). Hybridization reactions were carried out at 68°C in QuikHyb™ solution (Stratagene, La Jolla, CA) containing 200 g/ml of salmon sperm DNA and one of the radiolabeled probes (2 ϫ 10 6 cpm/ml) for 2 h. The membranes were washed twice (20 min each) at 65°C with 2 ϫ SSC (1 ϫ SSC ϭ 150 mM NaCl and 15 mM sodium citrate, pH 7.0) and 0.1% SDS, and then twice at 65°C (20 min each) with 0.1 ϫ SSC and 0.1% SDS. Kodak XAR-5 film and intensifying screens were used for autoradiography. Film was exposed at Ϫ80°C for 15 h to 3 days. The relative amount of each mMCP transcript was estimated by quantitating the associated radioactivity with a Betascope 603 Blot Analyzer (Betagen, Waltham, MA). To ensure that comparable amounts of total RNA were loaded in each lane, the gel was stained with ethidium bromide. In some instances, the blot was reprobed with the radiolabeled 18 S-specific oligonucleotide, 5-ACGGTATCTGATCGTCTTCGAACC-3Ј.
RT-PCRs were performed with total or cytoplasmic RNA isolated from 3-6-week-old BALB/c mBMMC. Samples were incubated with RNase-free DNase I for 10 min at room temperature to remove residual genomic DNA. Using the cDNA cycle kit from Invitrogen (San Diego, CA), ϳ2 g of RNA and 0.2 g of oligo(dT) primer (18-mer) were placed in a mirocentrifuge tube and incubated at 65°C for 10 min to disrupt the secondary structure of the transcript. Each RNA sample was reverse-transcribed at 42°C for 1 h in 20 l of a solution containing 1 ϫ reverse transcription buffer, 10 units of RNase inhibitor, 20 mM deoxynucleotide triphosphates, 20 mM sodium pyrophosphate, and 5 units of avian myeloblastosis virus reverse transcriptase. The reaction mixture was extracted with 20 l of phenol-chloroform, precipitated with 2 volumes of ethanol, and resuspended in 20 l of distilled water. A 2-l sample of the resulting cDNA preparation was mixed with 48 l of 1 ϫ PCR buffer containing 2 mM MgCl 2 , 2.5 units of AmpliTaq polymerase (Perkin-Elmer), 0.2 g of a primer corresponding to a 5Ј region in exon 1 (5Ј-ACTGGCAAAATGCAGGCC-3Ј) of the mMCP-2 transcript, and 0.2 g of a primer corresponding to a 3Ј region in exon 5 (5Ј-CATCAT-CACAGACATGTG-3Ј) of the mMCP-2 transcript. Thirty to 35 cycles of PCR were performed. Each cycle consisted of a 1-min denaturation step at 94°C, a 2-min annealing step at 55°C, and a 3-min extension step at 72°C. To verify the authenticity of their nucleotide sequences, the RT-PCR products were purified by electrophoresis on a 1% agarose gel. The appropriate bands were excised, purified with Geneclean™ (BIO 101, Vista, CA), and defined by Taq cycle-sequencing technology at the Biocore Facility of the Dana Farber Cancer Institute.
Nuclear Run-on Assays-Nuclear run-on assays were performed essentially as described by Greenberg and Bender (29). Briefly, pBluescript II KS (pBS) plasmid DNA samples lacking or containing a gene-specific probe for mMCP-1, mMCP-2, mMCP-4, mMCP-5, mouse mast cell carboxypeptidase A (mMC-CPA) (30), or ␤-actin were incubated at 65°C for 30 min in 0.2 N NaOH. The denatured DNA samples were neutralized with 2 M ammonium sulfate and then were slot-blotted (31) at ϳ5 g/lane onto nylon membranes. The resulting DNA blots were baked and UV-cross-linked. For an additional negative control, a 700base pair (bp) fragment of a type IX collagen cDNA in pBR322 DNA (32) was included in the slot blots. mBMMC (1-5 ϫ 10 7 ) maintained in medium containing or lacking IL-10 were washed with cold phosphate-buffered saline and disrupted by two treatments with 5 ml of ice-cold lysis buffer (10 mM NaCl, 3 mM MgCl 2 , 0.5% Nonidet P-40, and 10 mM Tris-HCl, pH 7.4). After centrifugation at 800 ϫ g for 5 min at 4°C, the nuclei-enriched pellets were resuspended in 200 l of storage buffer (40% glycerol, 5 mM MgCl 2 , 0.1 mM EDTA, and 50 mM Tris-HCl, pH 7.5). Each nuclear run-on reaction was initiated in a 400-l reaction buffer containing 0.5 mM ATP, 0.5 mM CTP, 0.5 mM GTP, 1 mM dithiothreitol, 10 units of RNase inhibitor, 100 Ci of [ 32 P]UTP, 20% glycerol, 5 mM MgCl 2 , 150 mM KCl, and 30 mM Tris-HCl, pH 8.0. The reaction mixture was incubated at 30°C for 30 min and then for ϳ5 min at room temperature with 25 g/ml DNase I (RNase-free). Tri Reagent (750 l) and CHCl 3 (200 l) were added, and the radiolabeled nascent nuclear RNAs in the aqueous phase were precipitated with isopropanol, resuspended in 100 l of buffer (10 mM Tes, 10 mM EDTA, and 0.2% SDS), purified with a gel filtration spin column (Clontech, Palo Alto, CA) and incubated with one of the DNA slot blots. Each hybridization reaction was carried out at 65°C for 36 -48 h in a small vial containing 1-2 ml of hybridization buffer (10 mM Tes, 10 mM EDTA, 0.2% SDS, 0.6 NaCl, 1 ϫ Denhardt's solution, and 300 g/ml salmon sperm DNA) and 1 to 5 ϫ 10 6 cpm of radiolabeled nuclear RNA. After hybridization, the blot was treated with 10 g/ml RNase A for 20 min at 37°C, washed three times (20 min each) with 2 ϫ SSC at 65°C, and exposed to x-ray film for 1-3 days.
Plasmids containing either a full-length ␤-actin, mMCP-2, or mMCP-4 cDNA were also digested with a restriction enzyme, and the digests (10 g of DNA/lane) were electrophoresed in a 1.2% agarose gel. The separated DNA was transferred to a nylon membrane, and the resulting blot was evaluated for hybridization to the appropriately sized insert with a replicate sample of radiolabeled nuclear RNA.
Pulse-Chase Experiments-Harrold et al. (33) found that a pulsechase assay is the most accurate method for evaluating the rate of turnover of a transcript that has a relatively long half-life. For the pulse-chase assay, 3-6-wk-old BALB/c mBMMC (10 8 ) were pretreated with IL-10 overnight in 50 ml of 50% WEHI-3 cell-conditioned medium to induce high steady-state levels of the mMCP-2 transcript. The resulting cells were then radiolabeled with 100 Ci of [5,6-3 H]uridine (40 Ci/mmol, DuPont NEN)/ml of medium for 20 h in the presence IL-10. After this pulse, the radiolabeled cells were washed twice with nonradioactive enriched medium. One-half of the cells were resuspended in 50 ml of 50% WEHI-3 cell-conditioned medium supplemented with IL-10, 100 g/ml unlabeled uridine, and 50 g/ml unlabeled cytidine. The other half of the cells were resuspended in the chase medium lacking IL-10. At various time points in the chase, samples of cells (ϳ10 7 ) were removed from the culture, total RNA was isolated, and the steady-state levels of the different transcripts were quantitated by blot analysis. In this analytical procedure, plasmid DNA (5 g lacking or containing a mMCP-2, mMCP-5, or ␤-actin gene-specific probe) was immobilized onto a separate lane of a slot blot, as described above. The filters were incubated 1-2 h at 65°C in 1 ml of hybridization buffer. Samples (0.5 or 1 ϫ 10 7 cpm) of radiolabeled RNA were denatured at 85°C for 5 min, cooled on ice, and then placed in the hybridization solution. The blots were incubated with the radiolabeled RNA samples for 2 days at 65°C. After hybridization, each blot was washed with two changes of 2 ϫ SSC and 0.1% SDS at 65°C for 10 min each and then with 2 changes of 0.2 ϫ SSC and 0.1% SDS at 65°C. The amount of blot-bound radioactivity was quantitated by liquid scintillation counting. pBS DNA was used to quantitate nonspecific binding of 3 H-labeled RNA. The amount of radioactivity bound to pBS DNA was subtracted from each set of readings to determine the amount of radiolabeled RNA that specifically bound to the mMCP-2, mMCP-5, and ␤-actin cDNA probes.
Effects of Actinomycin D and Cycloheximide on the Steady-state Levels of the mMCP-2 Transcript-To examine the effects of actinomycin D, BALB/c mBMMC were cultured in the presence of IL-10 for 48 h, washed twice, and resuspended in 50% WEHI-3 cell-conditioned medium with or without IL-10. Actinomycin D (Sigma) was then added to both cultures of cells at a final concentration of 1 g/ml. Samples of cells (ϳ5 ϫ 10 6 ) were removed immediately, or 5, 10, 15, or 24 h after the addition of actinomycin D. Total cellular RNA was isolated from each 2 The mMCP-1, mMCP-2, mMCP-4, and mMCP-5 gene-specific probes correspond to the 131-, 149-, 121-, and 200-bp 3Ј-UTRs of their respective cDNAs. The mMC-CPA gene-specific probe is ϳ1.3 kb in size and corresponds to the full-length cDNA; the ␤-actin gene-specific probe is 750 bp in size.
sample, and equal amounts of total RNA were loaded in the individual lanes of the gel. After electrophoresis, the steady-state level of the mMCP-2 transcript in the samples was determined by blot analysis.
To determine if an inhibitor of protein synthesis can block the IL-10induced expression of the mMCP-2 transcript, mBMMC (10 7 ) were washed and resuspended in 50% WEHI-3 cell-conditioned medium containing IL-10 and varying doses (0, 1, 5, and 20 g/ml) of cycloheximide (Sigma). After ϳ24 h of culture, total RNA was isolated from the four populations of mBMMC, and the amount of mMCP-2 mRNA in each was quantitated by blot analysis as described above. To determine if cycloheximide can influence the steady-state level of the established mMCP-2 transcript, 10 8 mBMMC were pretreated with IL-10, washed in enriched medium, and resuspended in 50 ml of 50% WEHI-3 cellconditioned medium containing 1 g/ml cycloheximide with or without IL-10. At various time points, total RNA was isolated and the steadystate levels of the mMCP-2 transcript in the resulting 10 samples were quantitated by blot analysis.

RESULTS
Transcription of Chymase Genes-As assessed by RNA blot analysis, BALB/c mBMMC developed with IL-3-enriched conditioned medium possess high steady-state levels of the mMCP-5 transcript, but not the mMCP-1, mMCP-2, and mMCP-4 transcripts (Fig. 1A). Nonetheless, nuclear run-on assays revealed that the mMCP-1, mMCP-2, and mMCP-4 genes are transcribed efficiently in BALB/c mBMMC whether these cells are cultured in the absence or presence of IL-10 (Fig.  1B). To confirm the slot blot data obtained with gene-specific probes of limited length, full-length mMCP-2 (n ϭ 1), mMCP-4 (n ϭ 2), and ␤-actin (n ϭ 2) cDNAs were resolved from plasmid DNA by gel electrophoresis and transferred to nylon membranes. Blot analysis demonstrated hybridization of radiolabeled transcripts in the nuclear RNA preparation to all three cDNAs but not to plasmid vector DNA (data not shown).
RT-PCR and Nucleotide Sequence Analysis of the mMCP-2 Transcript-The detection of mMCP-2 pre-mRNA in a nuclear run-on analysis of mBMMC before these mast cells were exposed to IL-10 prompted an RT-PCR investigation to determine whether or not the cells contain small amounts of a properly processed mMCP-2 transcript. With mMCP-2-specific primers, a 910-bp RT-PCR product was obtained from mBMMC cultured in the absence or presence of IL-10 (Fig. 2). As assessed by ethidium bromide staining, the 910-bp product generated from the RNA sample obtained from untreated mBMMC (lane 2) was substantially less than that obtained from mBMMC treated with IL-10 (lane 1) even though 30 cycles were used in the depicted RT-PCR analysis. Nevertheless, sequence analysis confirmed that both products corresponded to a properly processed mMCP-2 transcript (data not shown). Thus, mBMMC process mMCP-2 pre-mRNA identically whether or not they are exposed to IL-10.
Determination of the Half-life of the mMCP-2 Transcript by Pulse-Chase Analysis-With the use of a pulse-chase approach to monitor the half-lives of the mMCP-2, mMCP-5, and ␤-actin transcripts, mast cells were sequentially cultured in medium containing IL-10 and [ 3 H]uridine, washed, and chased in medium containing or lacking IL-10. The levels of the radiolabeled transcripts were then quantitated by slot blot analysis. Whereas the ϳ34-h half-life of the mMCP-5 transcript and the ϳ10-h half-life of the ␤-actin transcript were unaffected by treatment of the cells with IL-10, the half-life of the mMCP-2 transcript increased ϳ4-fold from ϳ8 h to ϳ34 h in cells treated with IL-10 (Fig. 3). In a second experiment carried out with a different lot of mBMMC, the half-life of the mMCP-2 transcript increased 3.4-fold, from ϳ8 h to ϳ27 h, in cells treated with IL-10.
Effect of Actinomycin D and Cycloheximide on the Steadystate Levels of the mMCP-2 Transcript-The capacity of IL-10 to slow down the rate of degradation of the mMCP-2 transcript was examined in mBMMC subsequently exposed to the inhibitor of gene transcription, actinomycin D. Preliminary doseresponse experiments revealed that actinomycin D at a final concentration of 1 g/ml effectively blocked new RNA synthesis in mBMMC that had been cultured with or without IL-10 (data not shown). In the presence of actinomycin D, the half-life of the mMCP-2 transcript in IL-10-treated mBMMC with or without continued IL-10 stimulation was ϳ13 h (Fig. 4).
To determine whether or not the IL-10-regulated expression of the mMCP-2 transcript in BALB/c mBMMC requires de novo protein synthesis, cells were treated with IL-10 for 24 h in the absence or presence of 1 to 20 g/ml cycloheximide. The IL-10- induced expression of the mMCP-2 transcript was effectively blocked by concomitant exposure of the cells to 1 g/ml cycloheximide (Fig. 5A). In a control experiment, when mBMMC were cultured in WEHI-3 cell-conditioned medium containing IL-10 and 1 g/ml cycloheximide for 24 h, washed, and then cultured in the presence of IL-10, but in the absence of cycloheximide for 48 h, the cells regained their ability to express high steady-state levels of the mMCP-2 transcript (data not shown). Thus, the cycloheximide effect on the steady-state level of the mMCP-2 transcript in mBMMC is reversible.
To determine if the half-life of the mMCP-2 transcript in  4, 6, 8, and 10) or just cycloheximide (lanes 3, 5, 7, and 9). RNA was isolated from the various populations of mBMMC immediately (lanes 3 and 4), or at 8 h ( lanes 5 and 6), 16 h (lanes 7 and 8), and 32 h (lanes 9 and 10) after the initiation of the cycloheximide treatment. For controls, replicate mBMMC were cultured in medium containing IL-10 but not cycloheximide (lane 1) or in medium containing neither IL-10 nor cycloheximide (lane 2). A blot was prepared and analyzed for the presence of the mMCP-2 transcript (top panel) and 18 S ribosomal RNA (middle panel). Results are expressed in the bottom panel as a percent of the starting level of the transcript. Similar findings were obtained in another experiment. IL-3-developed mBMMC depends on the synthesis of protein, mBMMC were exposed to IL-10 for 48 h to induce high steadystate levels of the mMCP-2 transcript. The cells were washed and then cultured for up to 32 h in WEHI-3 cell-conditioned medium containing 1 g/ml cycloheximide with or without IL-10. The steady-state levels of the mMCP-2 transcript decreased in the two populations of mBMMC at comparable rates (Fig. 5B).

DISCUSSION
Mast cell heterogeneity is now well established in the mouse, rat, and human, and the chymase superfamily of granule proteases has served as the principal marker for identifying tissue-specific mast cell phenotypes in these three species. Because the individual members of the chymase superfamily are differentially regulated in cultured mouse mast cells by cytokines, we examined how the expression of the chymase mMCP-2 is controlled in BALB/c mBMMC. We now report that the expression of mMCP-2 in these mast cells is regulated primarily by a post-transcriptional mechanism.
Based on the amino acids in their putative substrate binding pockets (17,34), it is likely that each mast cell chymase whose gene resides at the chromosome 14 complex (6) has a preferred substrate in vivo. In terms of their deduced amino acid sequences and the nucleotide sequences of their genes and transcripts, mMCP-1, mMCP-2, and mMCP-4 are more similar to one another than to mMCP-5. BALB/c mBMMC developed with recombinant IL-3-or IL-3-enriched conditioned medium contain high steady-state levels of the mMCP-5 transcript, but the levels of the transcripts that encode the other members of the chymase superfamily are all below detection by RNA blot analysis ( Fig. 1A and Refs. 12, 13, 16, and 17). Nonetheless, we now show by nuclear run-on analysis with gene-specific probes that all chymase genes are transcribed in these BALB/c mBMMC (Fig. 1B). Although it is possible that differential rates of transcription contribute somewhat to the overall steady-state levels of cytokine-inducible chymase transcripts, the nuclear run-on data indicate that the levels of these transcripts are controlled primarily by a post-transcriptional mechanism. When mBMMC cultured in the absence of IL-10 were compared to mBMMC cultured in the presence of IL-10, there was not much difference in the rates of transcription of the mMCP-1 and mMCP-2 genes despite the fact that there was a vast difference in the steady-state levels of these two chymase transcripts. Furthermore, even though the mMCP-4 gene is transcribed at a high rate with or without IL-10-stimulation, no mMCP-4 transcript could be detected in these cells by RNA blot analysis. RT-PCR analysis (Fig. 2) confirmed that the transcript that hybridizes to the mMCP-2 cDNA in the nuclear run-on analysis is mMCP-2. Moreover, nucleotide sequence analysis of the RT-PCR product revealed that the mMCP-2 transcript is properly processed in mBMMC whether these cells are cultured in the absence or presence of IL-10, a cytokine that induces high steady-state levels of mMCP-2 mRNA in these mast cells.
Because the levels of the mMCP-1, mMCP-2, or mMCP-4 transcripts in IL-3-developed mBMMC are below detection by blot analysis (Fig. 1A), the post-transcriptional control of any one of these mMCP transcripts can be studied only after its expression has been induced by cytokine treatment of the cells. We previously reported (14) that if IL-10-stimulated mBMMC are subsequently cultured in the absence of IL-10, there is a slow, but steady, decrease in the steady-state level of the mMCP-2 transcript. Pulse-chase experiments revealed that the mMCP-2 transcript has an ϳ4-fold longer half-life in IL-10treated mBMMC that continue to be cultured in the IL-10enriched medium than in replicate cells that have been transferred into IL-10-free medium (Fig. 3). IL-10 did not influence the half-life of the mMCP-5 or ␤-actin transcripts in these cells. However, the acquisition of high steady-state levels of the mMCP-2 transcript in IL-10-treated cells was fully blocked by cycloheximide (Fig. 5A), indicating that the inductive phase of the regulation depends on the synthesis of new protein. Inasmuch as the mMCP-2 gene is transcribed in mBMMC that have been cultured in the absence of IL-10 (Figs. 1B and 2), IL-10 must induce synthesis of a trans-acting protein that directly or indirectly stabilizes the mature mMCP-2 transcript or the processing of its precursor. If the mMCP-2 transcript is already expressed in abundance in mBMMC, the half-life of the transcript decreases from a rate of ϳ34 h in nondrug-treated cells (Fig. 3) to 13-15 h in cells treated with actinomycin D (Fig. 4) or cycloheximide (Fig. 5B). These data support the conclusion that IL-10 induces expression of a trans-acting factor that stabilizes the mMCP-2 transcript.
The fact that the half-life of the mMCP-2 transcript in cells exposed to IL-10 and actinomycin D (Fig. 4) or cycloheximide (Fig. 5B) does not decrease to 8 h could be a consequence of a secondary drug effect on the catabolism machinery itself. Alternatively, both drugs could effect the movement and processing of newly synthesized mMCP-2 pre-mRNA, as has been found with the IL-2 precursor transcript in cycloheximidetreated tonsil cells (35). If IL-10 induces the expression of a protein that stabilizes the mMCP-2 transcript, then the rate of turnover of the mMCP-2 transcript in the pulse-chase assay (Fig. 3) depends, in part, on the rate of turnover of the stabilizing protein. Because time would be required to deplete the intracellular level of the stabilizing protein following the removal of IL-10 from the culture medium, the ϳ4-fold difference in the turnover of the mMCP-2 transcript in the two populations of mBMMC is a minimum estimate. It, therefore, is likely that the rate of decay of the mMCP-2 transcript is extremely rapid in IL-3-developed mBMMC that have never been stimulated with IL-10.
The half-life of a transcript, which can differ widely in cells, depends on the presence or absence of certain structural features in the transcript and trans-acting factors in the cell (36). Some differentiating cells respond to developmental signals by changing the stability of their transcripts. For example, during erythrocyte development, the accumulation of globin is attributed mainly to the preferential stabilization of globin mRNA (37). When progenitor cells differentiate into mature B lymphocytes, high levels of immunoglobulin are produced due to increased stabilization of the immunoglobulin transcript (38). mBMMC are nontransformed mast cell-committed progenitors. The post-transcriptional regulatory pathway for protease expression in these cells may reflect their granule immaturity.
It is known that cis-acting elements in the 3Ј-UTRs of numerous transcripts regulate their half-life in cells. The most thoroughly studied RNA motif that controls the steady-state levels of mRNA in cells is the adenylate-and uridylate-rich cis-acting element (termed AU motif) (39) present in multiple copies in the 3Ј-UTRs of the transcripts that encode most cytokines and many early-response proteins. Because the mMCP transcripts do not possess an AU motif in their 3Ј-UTRs (Fig.  6), a different mechanism must control the steady-state levels of the varied chymase transcripts. The presence of the repetitive non-AU nucleotide sequence, ACCUACCUACCCAACUA, in 3Ј-UTRs of some Xenopus and Drosophila transcripts has been reported to influence their stability (40). The mMCP-1, mMCP-2, and mMCP-4 transcripts all have multiple "UGXC-CCC" sequences in their 3Ј-UTRs that are not present in the mMCP-5 transcript (Fig. 6). These non-AU repetitive sequences in the 3Ј-UTRs may be the cis-acting elements that regulate the steady-state levels of the mMCP-1, mMCP-2, and mMCP-4 transcripts.
It has been established with use of recombinant inbred mouse strains and interspecific backcrosses that the genes that encode mMCP-1, mMCP-2, mMCP-4, mMCP-5, neutrophil cathepsin G, and four cytotoxic lymphocyte granzymes all reside at a complex on chromosome 14 (6 -9). Inasmuch as no crossover has been detected by restriction-enzyme fragment length polymorphism analysis of chromosomal DNA, either these ϳ3-kb serine protease genes reside close to one another or some physical restraint in the particular region of chromosome 14 where they are located hinders recombination. Pulsed field gel electrophoresis of genomic DNA revealed that the mMCP-1, mMCP-2, mMCP-4, and mMCP-5 genes reside within 850 kb of one another (6). Nucleotide sequencing studies revealed that another chymase gene, encoding mMCP-8, is located only 5-7 kb away from the mMCP-1 gene (41). Recently, it was reported that a mouse T cell line (derived with IL-3 and IL-9 but maintained with IL-9 alone) expresses high steady-state levels of the transcripts that encode granzyme B, mMCP-1, mMCP-2, and mMCP-5 (42). Taken together, these findings raise the possibility that a common cis-acting enhancer element induces the transcription of many of the serine protease genes located at the chromosome 14 complex and that the steady-state level of each protease transcript is regulated by tissue-specific factors that instruct the cell which protease transcripts should be processed and/or protected from rapid degradation. Pertinent to this hypothesis is the fact that many genes that reside on chromosome 11 in the mouse (43,44) encode cytokines whose expression is also post-transcriptionally regulated during inflammation.
Because the mast cell chymases are enzymatically active at neutral pH, it would be advantageous to limit their expression to a single cell type. Any cell could transcribe the mMCP-2 gene if it expresses the appropriate DNA binding protein that induces the transcription of the gene in mast cells or if a mutation was to occur in the promoter of the gene so that a new cis-acting motif is created that is recognized by a different transcription factor in that cell. Many hematopoietic effector cells (e.g. neu-trophils, eosinophils, macrophages, natural killer cells, and cytotoxic lymphocytes) possess the chymase-zymogen processing enzyme dipeptidyl peptidase I (45) and therefore have the ability to convert any translated pro-mMCP-2 into mature, enzymatically active neutral protease. The post-transcriptional regulatory mechanism observed in the mast cell not only would accommodate cytokine regulation of protease expression during an inflammatory response but would also prevent the expression of the chymase family of mast cell proteases in other cell types if their genes were aberrantly transcribed.