INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Eukaryotic gene expression is determined not only by the rate of gene transcription, but also by the rate at which the transcript is degraded. The steady state concentration of a particular mRNA, as well as the rate at which a new transcriptionally induced steady state is attained, is directly related to the message half-life (1, 2). Numerous recent studies have been aimed at understanding the role of specific sequences within the mRNA in determining basal transcript stability and have led to the identification of consensus sequences that confer stability or instability to a transcript (2, 3). Although it is known that many stimuli alter mRNA stability, little is known about either the cis-acting sequences in the RNA or the cellular factors involved in regulation of message decay.

Plasminogen activators (PAs)¹ are serine proteases that are critical for thrombolysis and also play a key role in other physiological functions involving tissue remodeling. Plasminogen activator-inhibitors (PAIs) are specific inhibitors of PAs and are members of the serine protease inhibitor (serpin) family of proteins. Type-1 PAI (PAI-1) is a 50 kDa glycoprotein found in plasma, platelets, and a variety of cell types; its expression is regulated by growth factors, cytokines, and hormones, including agents that alter cellular cAMP levels (4).

HTC rat hepatoma cells synthesize and secrete tissue-type plasminogen activator (tPA) and PAI-1 (5). We have previously reported that incubation with the synthetic glucocorticoid, dexamethasone, decreases HTC cell tPA activity by increasing PAI-1 mRNA and protein 5-10-fold, an effect that is entirely explained by an increase in PAI-1 gene transcription (5-7). In contrast, these cells respond to cyclic nucleotides with a dramatic (50-fold) increase in tPA activity that is primarily the result of a decrease in PAI-1 mRNA and protein (8). This effect is time- and concentration-dependent; a 90% decrease in PAI-1 mRNA occurs after 12 h of incubation with 1 mM 8-bromo-cAMP plus 1 mM isobutyl-1-methylxanthine (cA) (8, 9). Nuclear run-on experiments have shown that cA causes about a 60% decrease in PAI-1 transcription, not sufficient to account for the decrease in message accumulation. We have determined the half-life of HTC cell PAI-1 mRNA by following the decrease in message level either after inhibition of transcription with actinomycin D or after washing out the transcriptional inducer, dexamethasone. By either method, PAI-1 mRNA displays a half-life of 4.0-4.5 h. Incubation of cells with cA accelerates PAI-1 mRNA decay, decreasing the half-life to 1.5 h. Thus, cyclic nucleotides regulate PAI-1 gene expression at both a transcriptional and posttranscriptional level. Interestingly, although actinomycin D does not alter the basal rate of PAI-1 mRNA degradation, it prevents the cA-induced acceleration of decay; incubation with cyclic nucleotides in the presence of actinomycin D results in a decay rate identical to that seen in the presence of actinomycin D alone (7).

The aim of this study is to understand the mechanism by which cyclic nucleotides regulate PAI-1 mRNA decay by determining the sequences within the PAI-1 transcript that are required for 8-bromo-cAMP regulation. We report here that the 3'-UTR of PAI-1 mRNA can confer cyclic nucleotide regulation on the otherwise stable -globin mRNA. Deletion analysis of the PAI-1 3'-UTR reveals the presence of at least two cA-responsive regions; one of these is in the 3'-most 134 nt of PAI-1 mRNA. Furthermore, the 3'-most 134-nt sequence by itself is able to confer cA responsiveness on the heterologous -globin mRNA, suggesting that this sequence plays a major role in the hormonal regulation of PAI-1 mRNA stability.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- Eagle's minimal essential medium, glutamine, fetal calf serum, calf serum, G418, restriction endonucleases (REs), Taq DNA polymerase, T4 ligase, and TRIzol^® were purchased from Life Technologies, Inc. Bovine serum albumin was purchased from Intergen Co. (Purchase, NY). Cycloheximide, 8-bromo-cAMP, and isobutyl-1-methylxanthine were purchased from Sigma. T3 and T7 RNA polymerases, RNase inhibitor, RNase A, and RNase T1 were purchased from Boehringer Mannheim. -³²P-labeled UTP (specific activity 800 Ci/mmol) was obtained from Amersham Pharmacia Biotech. The vector pSVL was purchased from Amersham Pharmacia Biotech, and pBluescriptKS⁽⁾ and Escherichia coli AG-1 competent bacteria were purchased from Stratagene (La Jolla, CA). Rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA (10) was a kind gift from Dr. Rodrigo Bravo (Bristol-Myers Squibb). Oligonucleotide primers used in PCR were synthesized by the University of Michigan Biomedical Research Core Facilities.

Constructs for Transfection Analysis-- The vector pSVL, which has an SV40 late promoter upstream and an SV40 polyadenylation signal downstream from the multicloning site (MCS), was used to drive expression of the DNA constructs used in these studies. Because the SV40 late promoter does not appear to be regulated by cyclic nucleotides, this vector allowed us to examine the effects of cA on message decay independent of transcriptional effects. All DNA vectors described below were amplified in E. coli AG-1 competent bacteria by established procedures (11).

()

()

()

Constructs for Riboprobes-- A PCR-amplified fragment of m-globin from bp 1256 to bp 1436 (which includes coding and intronic sequence) was subcloned into the PstI and BamHI sites of the pBluescriptKS⁽⁾ MCS. When linearized by digestion with HindIII (site in the pBluescriptKS⁽⁾) and transcribed using T7 RNA polymerase, this plasmid generates an antisense riboprobe complementary to 119 nt of the m-globin coding region mRNA (Fig. 1).

()

Preparation of Radiolabeled RNA-- Riboprobes were prepared by published methods (11). Plasmid DNA (approximately 500 ng) linearized with the appropriate RE was incubated with T7 or T3 RNA polymerase in transcription buffer (Boehringer Mannheim) containing RNase inhibitor, ATP, CTP, GTP, and [³²P]UTP at 37 °C for 60 min. RNase-free DNase I was added, and, after 15 min at 37 °C, the reaction was extracted with phenol/chloroform (1:1) and the RNA was precipitated in the presence of ammonium acetate and ethanol. The labeled RNA was purified by electrophoresis through a 6% polyacrylamide, 8 M urea gel, eluted from the gel, and ethanol-precipitated.

Cell Culture-- HTC cells were maintained in monolayer cultures as described previously (5). In preparation for experiments, cells were plated in 60-mm tissue culture dishes and grown to confluence. Cultures were washed in sterile phosphate-buffered saline (PBS) and incubated in serum-free medium containing 0.1% bovine serum albumin with or without 0.1 mM cycloheximide. After 10-16 h, the medium was removed, the monolayers washed twice with PBS to remove cycloheximide, and fresh medium with or without cA was added. Cells were harvested at the times indicated for each experiment. For experiments in which the RNA from the transfected genes was to be analyzed, incubation with cycloheximide provides two major advantages. First, cycloheximide significantly enhances expression from the SV40 late promoter (5-50-fold), resulting in easier detection of transcript from the transfected gene (13). Second, cycloheximide can be readily removed and its effects completely reversed by washing the monolayers with PBS and adding fresh medium (15, 16), resulting in a dramatic decrease in SV40 promoter activity. This protocol allowed us to measure mRNA degradation without using actinomycin D, which not only inhibits transcription but also blocks the cA effect on PAI-1 mRNA (7). Cycloheximide is known to alter the decay rates of some transcripts (2); however, the degradation rate of endogenous PAI-1 mRNA (determined using actinomycin D to block transcription) in cells from which cycloheximide has been removed after a 12-h incubation with the inhibitor, is nearly identical to that in cells not treated with cycloheximide. Furthermore, in cells transiently treated with cycloheximide, as in untreated cells, endogenous PAI-1 mRNA decay is accelerated after incubation of the cells with cA.

Transfections-- HTC cells were transfected using the Ca₂PO₄ DNA precipitation technique, and stably transfected cells were selected in media containing 1 mg/ml G418 as described previously (17). Colonies of stably transfected cells were pooled to minimize the possible effects of different integration sites.

Ribonuclease Protection Analysis-- Total cellular RNA was isolated either by the method of Chomczynski and Sacchi (7, 18) or by the TRIzol^® reagent method as described by Life Technologies, Inc. Ribonuclease protection analysis was carried out essentially as described (11). Total cellular RNA (20-40 µg/sample) was precipitated and resuspended in 30 µl of hybridization buffer (40 mM Pipes, pH 6.4, 400 mM NaCl, 1 mM EDTA, 80% formamide) containing 1-5 × 10⁵ cpm of ³²P-labeled riboprobe. After 5 min at 85 °C, the samples were placed at 50 °C for 16-20 h. RNase digestion buffer (350 µl of 10 mM Tris-Cl, pH 7.4, 300 mM NaCl, 5 mM EDTA, 20 µg/ml RNase A, and 100 units/ml RNase T1) was added to each reaction. After 30 min at 30 °C, SDS (0.5% final concentration) and proteinase K (125 µg/ml) were added and incubation continued at 37 °C for 30 min. The samples were then extracted with phenol/chloroform and ethanol-precipitated. The RNA pellets were washed with 80% ethanol, allowed to air-dry, and resuspended in an 80% formamide gel-loading buffer. Samples were subjected to electrophoresis through 6% polyacrylamide, 8 M urea gels and the gels were dried under vacuum. The amount of radioactivity in each protected fragment was determined by PhosphorImager analysis (Molecular Dynamics, Sunnyvale, CA). For each sample the signal in the specific band of interest was normalized to the signal in the 80-nt protected fragment of GAPDH.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

The 3'-UTR of PAI-1 mRNA Confers Cyclic Nucleotide Responsiveness onto the -Globin Transcript-- To determine the sequences in the PAI-1 mRNA involved in the cA regulation of message stability, we stably transfected HTC cells with DNA constructs carrying portions of the PAI-1 sequence and examined the degradation rates of the transcribed mRNAs. The pSVL vector was chosen because the SV40-late promoter, unlike some other viral promoters, does not appear to be regulated by 8-bromo-cAMP.² To confirm that the cA regulation can be faithfully reproduced in a transfection system, we first prepared a PAI-1 mini-gene using the SstI recognition site to delete 530 bp of the coding region, and subcloned it into the vector pSVLneo. The mRNA transcribed from the PAI mini-gene can be distinguished from the endogenous PAI-1 mRNA in ribonuclease protection analysis using a PAI-1 riboprobe that overlaps the deleted sequence. The half-life of the transcript from the PAI-1 mini-gene is 4.0 h and is decreased to 2.0 h in cells treated with cA (data not shown). These results are nearly identical to those observed for endogenous PAI-1 mRNA (7) and confirm that the cyclic nucleotide regulation of the transfected mini-gene faithfully reflects the regulation of the endogenous gene. These results also confirm that under these conditions cA regulation of message stability can be reproduced in cells that have been transiently treated with cycloheximide.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Schematic representation of DNA constructs. Mouse -globin genomic DNA from bp 233 to bp 1669 was amplified by PCR, and the 1430-bp product was subcloned into the pSVL MCS to give pSVL G. The m-globin coding sequence (1200 bp including intronic sequences, solid bar) and the PAI-1 3'-UTR (1730 bp, open bar) were amplified by PCR to include the restriction endonuclease sites shown. The PCR products were ligated into the pSVL plasmid as indicated to give pSVL G/P. A template for a m-globin riboprobe that will recognize 119 bp of the coding region (5'G riboprobe) was prepared by subcloning a PCR-amplified fragment from bp 1256 to bp 1436 into the MCS of pBluescriptKS(). , poly(A) signal.

View larger version (49K): [in this window] [in a new window]   Fig. 2.   Effect of cA on decay rates of transcripts from chimeric constructs. Following a 10-h incubation in serum-free medium containing 0.1 mM cycloheximide, duplicate monolayer cultures of HTC cells stably transfected with pSVL G (A) or pSVL G/P (B) were washed in PBS and incubated in serum-free medium with or without 1 mM 8-bromo-cAMP and 1 mM isobutyl-1-methylxanthine. At the times indicated, the cells were harvested, total RNA isolated, and ribonuclease protection analysis carried out on 20 µg (pSVL G) or 40 µg (pSVL G/P) RNA per sample using 32P-labeled riboprobes for 5' globin and for GAPDH (GAP). The sizes of the protected fragments are: globin, 119 nt; GAPDH, 80 nt. Size markers in lane M are DNA from HpaII-digested and 32P-labeled pBR322. Duplicate lanes represent duplicate cultures. C and D, the experiments shown in panels A and B were analyzed using the PhosphorImager. For each sample the signal in the globin protected fragment was normalized to that in the GAPDH fragment, and the data are presented as a percentage of the amount at time zero. , control; , cA.

Deletion Analysis of the PAI-1 3'-UTR-- In an effort to locate the region within the PAI-1 3'-UTR responsible for cyclic nucleotide responsiveness, we used convenient restriction endonuclease sites to delete selected portions of the PAI-1 cDNA in the context of the G/P construct. The 3'-UTR of PAI-1 mRNA has several sequences of potential interest (14) illustrated in the top portion of Fig. 3. AU-rich elements (AREs) that include AUUUA motifs flanked with U-rich stretches have been implicated in the instability of mRNAs of cytokines and oncogenes (3). The PAI-1 3'-UTR has three AUUUA motifs (ATTTA in the cDNA), two in close proximity at nt 2053 and 2117, and a third farther downstream at position 2728, close to the two 70-nt U-rich stretches (T-rich in the cDNA) (nt 2821-3023). PAI-1 has a highly conserved region in its 3'-UTR; the 127-nt sequence in rat PAI-1 from nt 2506 to nt 2632 is 80% identical in mink, murine, bovine, porcine, and human PAI-1 and shares 85-97% identity when compared pairwise with each of these species (14, 22-26). Finally, PAI-1 has an unusual stretch of 39 GpA dinucleotides between nt 2363 and 2442. The location of these regions within the 3'-UTR, the RE sites that flank them and the designations given them are shown in the top of Fig. 3. Regions "A" and "C" are so named for lack of putative regulatory sequences.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Schematic representation of deletion constructs and effect of 8-bromo-cAMP on mRNA half-lives. The chimeric gene pSVL G/P described in Fig. 1 was the parent plasmid used to prepare constructs with deletions in the PAI-1 3'-UTR. The positions of several unique restriction sites and of certain potentially interesting sequences in the 3'-UTR of PAI-1 are shown at the top of the figure. The closed triangles () indicate the location of ATTTA pentamers, and the closed bar () represents the 127-bp region that is 80% identical in murine, rat, bovine, mink, porcine, and human PAI-1. GpA and TTTs show the locations of the 39 GpA-repeat and the T-rich regions, respectively. The G/P deletion constructs shown on the lower left side of the figure were prepared using the RE sites as described under "Experimental Procedures" and are designated by the PAI-1 sequence deleted as indicated below the schematic in the top portion of the figure. HTC cells stably transfected with each of the deletion constructs were examined for the effect of cA on the degradation rates of the transcripts as described in the legend to Fig. 2. The mRNA half-lives calculated from several experiments are summarized. The means ± S.E. (for the Globin and G/P constructs, n = 8) or ± S.D. (for the deletion constructs, n = 2 or 3) are shown. The ratio of the message half-life in untreated and cA-treated cells was calculated separately for each experiment and the mean of these ratios is reported in column Control/cA. , poly(A) signal.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Effect of 8-bromo-cAMP on decay rates of transcripts from deletion constructs. HTC cells transfected with pSVL G/P U II (nt 2923-3060, panel A), pSVL G/P 1600 (nt 1331-2926, panel B), or pSVL G/P U I (nt 2718-2926, panel C) were analyzed for cyclic nucleotide regulation of mRNA degradation as described in the legend to Fig. 2. Ribonuclease protection analysis was carried out on 40 µg of RNA per sample. In each graph duplicate cultures from a single experiment are shown. , control; , cA.

Insertion Analysis: The 3'-most 134 nt of PAI-1 Is Sufficient to Confer Cyclic Nucleotide Regulation-- To more clearly define the cis-acting regions involved in the cA regulation, we prepared constructs in which selected portions of the PAI-1 3'-UTR sequence were inserted into the 3'-UTR of the -globin gene (Fig. 5). These constructs were then stably transfected into HTC cells and experiments performed as described above for the deletion constructs. Results from experiments with four of these constructs are shown in Fig. 6. The mRNA transcribed from pSVL G/G, like the -globin transcript, is stable and not regulated by cyclic nucleotides (Fig. 6A). However, when the 3'-most 130 nt of PAI-1 sequence (lacking the final 6 A nucleotides; see Fig. 8) was inserted into the globin 3'-UTR (pSVL G/G + U II), the resultant construct was regulated by cA. Fig. 6B shows a composite of three such experiments; the mean half-life of 9 h was decreased to 4 h upon incubation with cA, a 2.2-fold change. The construct pSVL G/G + 1400 nt has an insertion of 1378 nt of the 5' end of the PAI-1 3'-UTR. The transcript from this construct displayed a half-life of 4.7 h, which was decreased to 2.4 h upon incubation with cA (Fig. 6D). In contrast, the message transcribed from pSVL G/G+U I, which has the upstream U-rich region, is not cA-regulated (Fig. 6C). These results confirm that the 3'-most 134-nt region of PAI-1 3'-UTR is sufficient to confer cyclic nucleotide regulation and that there is at least one other regulatory element in the upstream region of the 3'-UTR. The PAI-1 sequence between bp 2770 and bp 2923, containing the upstream U-rich stretch (U I), does not appear to confer cyclic nucleotide regulation.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Schematic representation of PAI-1 insertion constructs. The m-globin sequence from bp 233 to bp 1470 was prepared by PCR to include a 5' XhoI recognition site and a BglII site 3' to bp 1470 (solid bar). The m-globin 3'-UTR from bp 1471 to bp 1669 that includes the polyadenylation signal () and a 5' BglII site and a 3' XbaI site was amplified by PCR (hatched bar). The products were cloned into the XhoI and XbaI sites of pSVLneo to give the construct pSVL G/G that has a BglII recognition site 25 bp downstream from the translation stop codon. PAI-1 fragments of interest were prepared by PCR to include BglII recognition sites on both the 5' and 3' ends and were ligated into the BglII site of pSVL G/G. The resultant constructs are designated by the region of the PAI-1 sequence inserted.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effect of 8-bromo-cAMP on decay rates of transcripts from insertion constructs. Experiments were carried out as described in the legend to Fig. 2 using HTC cells transfected with pSVL G/G (panel A), pSVL G/G + UII (bp 2925-3054, panel B), pSVL G/G + U I (bp 2714-2926, panel C) or pSVL G/G + 1400 nt (1331-2709, panel D). Ribonuclease protection analysis was carried out on 40 µg of RNA per sample. In panel A the average ± S.E. of five experiments (10 cultures) and in panel B the average ± S.E. of three experiments (6 cultures) are shown. In panels C and D, duplicate cultures from a single experiment are shown. , control;  or , cA.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Effect of 8-bromo-cAMP on decay rates of transcripts from insertion constructs. PAI-1 insertion constructs were prepared as described in Fig. 5. The left side of the figure shows the PAI-1 sequence numbers and regions of potential interest (as defined in the legend to Fig. 3) included in each insertion. The ability of the insert to confer cA-induced destabilization of the transcript is shown on the right, where a "+" indicates a control/cA ratio of mRNA half-lives greater than 2.2 and a "" indicates a ratio of less than 1.4.

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Sequence of the 3'-most 134 nt of the rat PAI-1 3'-UTR. The sequence to PAI-1 mRNA from nt 2927 to nt 3060 as reported by Zeheb et al. (14) is shown. The U-rich region is underlined, and the A-rich region is designated by a bold underline.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

In this study, we have examined the mechanism by which cyclic nucleotides accelerate the rate of rat HTC cell PAI-1 mRNA degradation. We have shown that sequences in the 3'-UTR of PAI-1 can confer both instability and cA regulation of degradation on the otherwise stable and non-regulated mouse -globin mRNA. The PAI-1 3'-UTR has at least two cA-responsive elements; one of these elements is in the 3'-most 134 nt, which by itself is able to confer cA regulation on the globin message.

Studies from several laboratories have defined specific elements in the 3'-UTR of transcripts that are responsible for their relative stability. With a few exceptions (2, 27), these have all been instability determinants. The most extensively studied instability determinants are the AU-rich elements, or AREs, found in the 3'-UTR of many highly labile transcripts including those of proto-oncogenes (28, 29), growth factors (30), and cytokines (31, 32). AREs may have either multiple AUUUA motifs resulting in the nonomer UUAUUUA(U/A)(U/A) (28), AUUUA motifs scattered in a generally U-rich sequence, or an AU-rich region without an AUUUA motif (3). In addition, two U-rich domains in the yeast MFA2 3'-UTR appear to be involved in regulation of poly(A) nuclease activity (33) and a 29-nt U-rich sequence in human amyloid precursor protein mRNA has been implicated in regulation of its degradation (34).

Rat PAI-1 has two AUUUAs in close proximity at nt positions 2053 and 2117, but these are not in an otherwise AU-rich region. A third AUUUA sequence is near the two 70-nt U-rich stretches (nt 2821-2892 is 57% U and nt 2952-3023 is 51% U, separated by a 60-nt sequence that is only 20% U), making this region a candidate as an instability determinant. Interestingly, one of these U-rich regions is included in the 3'-most 134-nt fragment that is able to confer both an increase in basal degradation rate and cA regulation of decay on the -globin transcript. Although the U-rich region may be involved in the cyclic nucleotide responsiveness, it is not by itself sufficient as seen by the results shown in Fig. 7. Decay rates of transcripts from G/G constructs carrying either one or both of the U-rich sequences, but lacking the last 30 nt, are not regulated by cA. In addition, the upstream functional element(s), is not simply the upstream U-rich sequence, as evidenced by the fact that constructs lacking both U I and U II (G/P U I + U II (Fig. 3) and G/G + 1400 nt (Figs. 6 and 7)) display cA regulation of mRNA decay.

The growing number of reports of hormone, growth factor, and cytokine regulation of mRNA stability reflects the growing realization of the importance of this process in overall regulation of gene expression. Rapid changes in the abundance of a short-lived transcript can be effected by a small change in the half-life (2). There are now several studies that have identified cis-acting sequences and/or trans-acting factors potentially involved in hormonal regulation of degradation. The dramatic stabilization of the frog liver vitellogenin mRNA is mediated by binding of an estrogen-inducible protein, recently identified as the KH domain protein, vigilin, to a 27-nt sequence within the 3'-UTR (35, 36). Tumor necrosis factor- induces a 4-fold stabilization of GLUT1 mRNA in 3T3-L1 pre-adipocytes, which appears to be mediated by a 105-nt GC-rich portion of the 3'-UTR and is correlated with increased protein binding to the 3'-UTR (37). Stabilization of amyloid precursor protein mRNA in activated peripheral blood mononuclear cells is accompanied by an increase in binding of a protein to a 29-nt U-rich element in the 3'-UTR (38). Finally, in rat pituitary cultures, thyroid hormone accelerates degradation of thyrotropin  message and causes an increased binding of a cellular factor to a 41-nt portion of the 3'-UTR of the transcript (39). This 41 nt includes a 12-nt consensus sequence found in several unstable mRNAs, including the 3'-most 134 nt of PAI-1 mRNA (11 of the 12-nt consensus at nt 3010-3020).

Agents that elevate cellular cAMP levels also have been shown to regulate mRNA degradation, in some cases stabilizing message (40-42), but more often causing down-regulation of mRNAs. Treatment of rat Sertoli cells with follicle-stimulating hormone causes destabilization of androgen receptor, follicle-stimulating hormone receptor, and G-protein -subunit mRNAs (43-45). In both rat ovary and human endometrial stromal cells, luteinizing hormone/human chorionic gonadotropin receptor mRNA is down-regulated by human chorionic gonadotropin, apparently through accelerated degradation (46, 47). Cyclic nucleotide analogues also destabilize asialoglycoprotein receptor mRNA in HepG2 cells (48), tyrosine aminotransferase in H4 rat hepatoma cells (49), angiotensin type I and type II receptor mRNA (50, 51) and the kidney-specific transcription factor, LFB3, mRNA in porcine kidney cells in culture (52). In two cases, cA regulation of message stability has been associated with binding of proteins to the 3'-UTR of the transcript. Chlorophenylthio-cAMP both stabilizes rat hepatoma cell phosphoenolpyruvate carboxykinase mRNA 10-fold and decreases the binding of a 100-kDa protein to the 3'-UTR of this transcript (53). In hamster smooth muscle cells, isoproterenol destabilizes ₂-adrenergic receptor mRNA and increases the binding of both an AUF1 related protein and a more specific ARB protein to the 3'-UTR of ₂-adrenergic receptor mRNA (54, 55). Sequences in human and hamster ₂-adrenergic receptor mRNAs that are A+U-rich have been implicated in both protein binding and regulation of message stability (56, 57). The 134-nt region that we find to confer cA regulation of PAI-1 mRNA stability has both a U-rich and an A-rich region (Fig. 8), similar but not identical to those in ₂-adrenergic receptor mRNA.

The cyclic nucleotide regulation of PAI-1 mRNA accumulation, first observed in HTC cells (8), has also been demonstrated in rat testicular peritubular cells (58), astrocytes (59) and osteoblasts (60), mink lung epithelial cells (61), human fibrosarcoma cells (62), umbilical vein endothelial cells (63), and synovial cells (64). Our previous experiments have shown that in rat HTC cells, whereas transcriptional regulation plays a role in the cA-induced decrease in PAI-1, the major effect is on message stability. The studies reported here demonstrate that at least two elements in the 3'-UTR of PAI-1 mRNA are involved in regulation of PAI-1 message decay. One of these regulatory elements is the 3'-most 134 nt. We have found that a number of cellular proteins interact with the same PAI-1 sequence that can confer cA regulation,³ suggesting that one or more of these factors may play a role in cyclic nucleotide regulation. Our current studies are aimed at delineating the exact sequences responsible for the cyclic nucleotide regulation of mRNA degradation and defining the binding proteins involved in this regulation, as well as identifying the other cis-acting element(s) in the PAI-1 3'-UTR.