THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.
VOl. 268, No. 3, Issue of January 25, pp. 2154-2159.1993
Printed in U.S.A.
Regulation of Human Tissue Factor Expressionby mRNA Turnover* (Received for publication, August 28, 1992)
Shawn M. Ahern$& ToshiyukiMiyatagll, and J. Evan SadlerSII** From the 11 Howard Hughes Medical Institute Laboratories and the$.Departments of Medicine and of Biochemistry and Molecular Biophysics, The JewishHospital of St. Louis, Washington University School of Medicine, St. Louis,Missouri 631 10
Tissue factor serves as the cellular receptor for cir- the hemostatic system to maintain a balance between the culating blood coagulation factor VI1 and is the prin- promotion and inhibition of blood clotting. Tissue factor is cipal physiological initiator of blood coagulation. Tis- expressed in many cell types but is not normally expressed in sue factor is not normally expressed in cells that con- cells that contact blood such as endothelial cells and monotact blood, such as endothelial cellsand monocytes, but cytes. Tissue factor canbe induced in these cells by a variety can be induced in these cells by tumor necrosis factor of agents including tumor necrosis factor, endotoxin, interleuo r tumor-promoting phorbol esters. Following induc- kin-1, or tumor-promoting phorbol esters such as phorbol 12tion, the human tissue factormRNA is degraded with myristate 13-acetate. In addition, tissue factor be can induced a half-life of -0.75-1.5 h. The cellular mechanisms in mouse and human fibroblastsby treatment with serum or responsible for this rapid mRNA turnover were invescertain growth factors (2-9). tigated with chimeric tissue factor*@-globin constructs Studies investigating the mechanisms responsible for the expressed in stably transfected mouse NIH/3T3 cells. These constructs were expressed with the transiently induction of tissue factor expression have shown that the inducible c-foe promoter which eliminated theneed to levels of tissue factor mRNA increaserapidly and transiently use transcriptional inhibitors to determine mRNA half- after addition of the inducing agent. The tissue factor tranlives. Sequences capableof conferring rapid turnover script is then rapidly degraded with a half-life of -48 min to to the normally stable @-globin transcript werelocal- 1.5 h. Simultaneoustreatment with theproteinsynthesis inhibitor cycloheximide “superinduces” the tissue factor gene ized to the last 600 nucleotides of the tissue factor mRNA. The 3‘ end of this fragment is similar to pre- and substantially increases thehalf-life of the transcript (9regulation, tissue viouslydescribedAU-rich mRNA destabilizing ele- 11). Thus, in addition to transcriptional expression is regulated post-transcriptionally by ments. Activity of the tissue factor element was de- factor pendent on its specific sequence and not simply a high mRNA degradation. AU nucleotide content. The degradation of unstable The significance of transcriptional regulation in cell growth, chimeric tissue factor-&globin mRNAs was prevented differentiation, andmetabolism is well documented. However, by inhibition of transcription with actinomycin D. Chi- there is increasing evidence thatmRNAturnoveris also meric tissuefactor*@-globin mRNAs weresuperinimportantinregulating geneexpression. Thesteady-state duced by the protein synthesis inhibitor cycloheximide, levels of many mRNAs do not reflect their transcriptional and this superinduction may be due in part to stabili- rates, suggesting that cellular metabolism is influenced by the zation of the mRNA. stability of individual mRNAs and by the ability of cells to regulate mRNA turnover (12). Selective mRNA stabilization or destabilization is important in regulating the expression of histones (13, 14), P-tubulin (15-18), transferrin receptor ( 5 , Blood coagulation is an important hostdefense mechanism 19-21), human growth hormone (22) vitellogenin (23,24), and that ensures the maintenance of vascular integrity. Tissue insulin (25). The cis-acting elements involved in regulating factor serves as the cellular receptor for blood coagulation theturnoverrates of thesemRNAs vary greatlyinboth factor VI1 and is the major initiator of blood clotting. Once sequence and location. bound to tissue factor, factor VI1 is rapidly activated and then The basal half-life of a transcript can also be an important catalyzes the activationof factor X to Xa.Activated factor X point of metabolic control. One example is a class of shortconverts prothrombin to thrombin, and thrombin converts lived mRNAs which share a common sequence motif with a fibrinogen to fibrin, leading to formation of the insoluble high content of A and U nucleotides in their 3”untranslated fibrin clot (1). regions (UTRs).’ This groupof mRNAs includes the lymphoProper regulation of tissue factor expression is critical for kines granulocyte-macrophage colony-stimulating factor (GM-CSF) and tumor necrosis factor and the proto-oncogenes * This work was supported in part by National Institutes of Health c-fos, c-my, and c-myb. Following a rapidandtransient Grant HLBI 14147 (Specialized Center of Research in Thrombosis). transcriptional induction, these transcripts are quicklydeThe costs of publication of this article were defrayed in part by the graded (26). When inserted into the 3’-UTR of a normally payment ofpage charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 stable mRNA,AU-rich regions from the 3’-UTR of the c-fos, interleukin-2, and GM-CSF mRNAs are able to destabilize solely to indicate this fact. f Supported by the Division of Biology and Biomedical Sciences, the chimeric transcript, suggesting that this motif functions Washington University, St. Louis, MO 63110. as a destabilizing element (21, 27, 28). The 3’-UTR of the ll Supported by an InternationalFellowship awarded by the Fogarty International Center, National Institutes of Health. Present address: Laboratory of Thrombosis Research, National Cardiovascular Center Research Institute, Fujishirodarai 5, Suita, Osaka 565, Japan. ** To whom correspondence should be addressed. Tel.: 314-3629067; Fax: 314-454-0175.
The abbreviations used are: UTR, untranslated region; GM, granulocyte-macrophage; CSF, colony-stimulating factor; PCR, polymerase chain reaction; kb, kilobase; IFN, inteferon; Pipes, 1,4-piperazinediethanesulfonic acid.
2154
Tissue Factor mRNA Turnover
2155
with XhoI plus3 pg of pSV2neo (35) linearized with EcoRI. Colonies were selected in Dulbecco's modified Eagle's medium,10% calf serum, and 0.5 mg/mlG418 sulfate(GIBCO)and pooled after14 days. Transfected cells were serum-starved for 36-48 h in Dulbecco's modified Eagle'smedium, 0.5% calf serumandthenstimulatedwith Dulbecco's modified Eagle's medium, 15% fetal calf serum.Stock solutions of cycloheximide and actinomycin D (Sigma) were added directly to the cultures to the desired final concentration. Nuclease SI Protection Analysis-Plasmid pRSVNEO was linearized by digestion with BglII. Plasmid pRSVMPG1.5, which contains the full length mouse P-globin gene cloned into the unique SalI site of the expression vector pUCSRSV, was linearized by digestion with BamHI. The linearized plasmids were treated with calf intestinal phosphatase (Boehringer Mannheim) and 5' end-labeled using [y32P]adenosine triphosphate (Amersham Corp.) and T4polynucleotide kinase (Pharmacia LKB Biotechnology Inc.). Total cellular RNA was isolated from NIH/3T3 cells using a guanidine isothiocyanate-Sarkosyl solution followed by phenol extraction and ethanol precipitation as describedpreviously (36). Approximately 200,000 cpm of each EXPERIMENTALPROCEDURES labeled probe was hybridized overnight with approximately 25 pg of Oligonucleotide Primers-The oligonucleotide primers used were: RNA a t 48 "C. Hybridizations were performed in 80% formamide, 1 2823, cgccgcggtcgtcgacatttgcttctgacatagttgtgttgact;2824, aatctagagat- mM EDTA, 40 mM Pipes-Na, pH 6.4, and 0.4 M NaCl. Nuclease S1 gatcagtttagtggtacttgtgagccaaggcagtg; 2825, catctagaccctttcctgctcttg- digestion was performed in 30 mM NaOAc, pH 4.4, 280 mM NaC1, cctgtgaacaatgg; 2826, cggtcgacacggtaccgaaaccatttgttgaatttttctctgagtg; and 4.5 mM Zn(OAc)2. Protected fragments of 330 nucleotides for 2829, gatctagataaagtatattaaataaattatctcaa; 8307, aatctagattggctgtc- neomycin mRNA and 200 nucleotides for chimeric transcripts were cgaggttt; 8210, attctagaccccatctctactaaaaa; 8211, attctagaggaagca- analyzed on 8% denaturing polyacrylamide gels as described previctgttggagct; 8212, taatctagatgaaacattcagtgggga; 8213, taatctagacaa- ously (37). Runoff Transcription Assay-For transcription studies, NIH/3T3 acctcggacagcca; 8307, aatctagattggctgtccgaggttt; 8308, aatctagacatggagacccct; 8309, aatctagaccacaccaagctaatt; 9095, aatctagaaatca- cells were harvested by scraping, collected by centrifugation at 800 X taaactctgt; 9096, aatctagaggcaaactttgtattaa; 9097, aatctagaaagctttt- g for 5 min a t 4 "C, resuspended in 5 ml of ice-cold 10 mM HEPES, gaggggct; 9098, aatctagagcccctcaaaagcttt; 9099, aatctagattaatacaaa- pH 8.0, 1.5 mMMgC12, and 10 mMKC1. After 15 min on ice, cells were disrupted using a 5-ml syringe with a 25-gauge needle. Nuclei gtttgcc; 9100, aatctagacacactttatgattt. Plasmid Constructions-Plasmid pBSMPG1.5 was constructed to were collected by centrifugation as above, resuspended in 0.25 ml of introduce unique cloning sites between the coding sequence and 3'- transcription buffer, and runoff assays were performed using [ C Y ~ ~ P ] UTR of the mouse P-globin gene. Using oligonucleotide primers 2825 G T P (3000 Ci/mmol, Amersham Corp.) as described previously (9). and 2826 which introduced a 5' XbaI site and a 3' SalI site, the 344 The labeled RNA was heated a t 80 "Cfor 10 min, chilled on ice, and base pair 3'-UTR of the full-lengthmouse P-globin gene was amplified hybridized to nitrocellulose filters that contained denatured target by polymerase chain reaction (PCR)from a murine P-globin genomic DNA. After restriction endonuclease digestion, target DNAs were DNA fragment in pBR322 provided by Dr. Timothy J. Ley (Wash- purifiedfromvectorsequences by agarose gel electrophoresis and ington University, St. Louis MO). This fragmentwas cloned into the subsequently denatured with0.3 M NaOH at 80 "Cfor 10 min. Filters creating plasmid vector pBluescript I1 KS+ (Stratagene) were prepared by slot blotting 2.5 pg of each target DNA followed by pBSMPG344. Using primers 2823 and 2824, exons 1 and 2 along with baking a t 80 "C for 20 min. DNA targets used were a 1.5-kb EcoRIthe coding portion of exon 3 from the mouse /3-globin gene were BamHI fragment from the rat c-fos cDNA (9), the 2.0-kb SalI fragamplified by PCR. These primers introduced 5' SacII and SalI sites ment from pTF4, a 1.9-kb BamHI fragment from the human @-actin along with3' BclII and XbaIsites. This fragment was cloned into the cDNA (38), anda 1.5-kb BamHI-Hind111 fragment from pRSVNEO SacIIand XbaI sites of plasmid pBSMPG344 tocreateplasmid containingthe neomycin resistance gene (39). PhiX174 DNA pBSMPG1.5. (GIBCO) was included as a negative control target. The expression vector pBSFOS was constructed by subcloning a Densitometry-Autoradiogramsfrom S1 nuclease protectionex719-base pair EcoRI-NarI fragment from the 5' end of plasmid pF711 periments were analyzed with an Ultrascan XL laser densitometer (31) into the vector pBluescript. Plasmid pF711 contains a complete (Pharmacia). To determine mRNA half-lives, an exponential decay copy of the human c-fos gene from 0.71 kb upstream of transcription function ( y = Ae-B") was fitted to the data using the RS/1 system initiation. The EcoRI-NarI fragment containsa functional c-fos pro- (BBN Software ProductsCorp., Cambridge, MA). moter including serum response elements. Plasmid pBSFOSMpGwas constructed as the parent construct for RESULTS all chimeric tissue fact0r.P-globin constructsby subcloning the1.63Tissue Factor 3'-UTR Contains a n mRNA Destabilizing kb SalI fragment from plasmid pBSMPG1.5 to the unique SalI site of vector pBSFOS, downstream of the c-/os promoter. Element-Previous studies in our laboratory demonstrated The tissue factor '0-globin chimeric genes were constructed as that tissue factorgene expression is regulated both transcripfollows. Various regions of the tissue factor cDNA were amplified tionally and post-transcriptionally. After treatment with tuXbaI sites. with the PCR using primers that introduced 5' and 3' mor necrosis factor or phorbol 12-myristate 13-acetate there T h e amplified fragments were subcloned into the XbaI site present is a dramatic and transient increase in tissue factor mRNA in the P-globin gene in plasmid pBSFOSMBG. The constructs and the PCR primersemployed to make them are: pTF1,8307 and 8308 levels. In untreated human umbilical vein endothelial cells, pTF2, 8212 and 8213; pTF3, 8210 and 8211; pTF4, 2829 and 8309; transcription of the gene is undetectable, but is rapidly and pD1, 2829 and 9097; pD2, 2829 and 9096; pD3, 2829 and 9100; pD4, transiently induced in response to phorbol 12-myristate 138309 and 9095; pD5,8309 and 9099; pD6, 8309 and 9098 pD7, 9097 acetate or tumor necrosis factor. Tissue factor mRNA levels and 9099; pD8,9097 and 9095; pD9,9096 and 9095. The DNA sequences of all segments derived from synthetic oligo- subsequently decrease with a half-life of -48 min suggesting that the tissue factor transcript is targeted for rapid degranucleotides or PCR were confirmed by dideoxy sequencing (32). dation (9). Cell Culture and Transfection-NIH/3T3cells (AmericanType Culture Collection, CRL 1658) were grown in Dulbecco's modified T o identify which regions of the tissue factor transcript are Eagle's medium (GIBCO) supplemented with 10% heat-inactivated responsible for rapid turnover, we screened the full-length bovine calf serum (GIBCO),2 mM glutamine, 100 units/ml penicillin, tissue factor mRNAfor sequences capableof destabilizing the 100 pg/ml streptomycin, and 250 ng/ml amphotericin B. Cells were normally long-lived mouse /?-globin transcript. To determine grown under 5%CO, and split using 0.05% trypsin, 0.02% EDTA (GIBCO). Approximately lo6 cells were seeded per 100-mm dish 24 h the stabilityof the chimeric transcripts, a transient induction prior to transfection by the calcium phosphate technique (33, 34). assay was utilized which took advantage of the serum inducTransfection mixtures contained 15pg of the test plasmid linearized ibility of the c-fos promoter in plasmid pBSFOSMPG. After
tissuefactortranscript also containsanAU-richdomain which is characteristicof this class of short-lived mRNAs(29, 30). We have investigated the molecular mechanisms responsible for the rapid turnover of the tissue factor transcript by stably expressing chimeric tissue factor. P-globin constructs in mouse NIH/3T3 cells. Our results indicate that sequences responsible for tissue factor mRNA turnover are located in the last 600 nucleotides of the tissue factor transcript. The most 3' 150 nucleotides, which bear some similarity to previously described AU-rich destabilizing elements, are necessary for destabilizing activity while at least 450 nucleotides are required for optimal activity. The mRNA destabilizing activity is sequence-specific and is sensitive to inhibitors of transcription and protein synthesis.
Tissue Factor mRNA Turnover
2156
a period of serum starvation, the c-fos promoter is transcriptionally inactive. Following serum addition, the promoter is rapidly and transiently induced, and transcriptional activity terminates after -30 min (40). Once mRNA levels are maximal, subsequent degradation of the mRNA can be determined by assaying mRNA concentrations over time. Use of the transiently induced c-fos promoter to express the chimeric transcripts avoids the need to use transcriptional inhibitors which are known to affect the half-lives of many mRNAs (28, 41). The full-length tissue factor cDNA was divided into four regions which were subclonedinto the 3'-UTR of the mouse @-globingenein plasmid pBSFOSM@G(Fig. 1). Chimeric constructs were transfected as XhoI-linearized fragments into NIH/3T3 cells and G418-resistant colonies were pooled to ensure a heterogeneity of integration sites. Pooled cells were serum-starved for 36-40 h and subsequently stimulated with 15% fetal calf serum for up to 6 h. Total RNA was isolated and chimeric mRNA levels were determined by S1 nuclease protection analysis with a probe directed against @-globin mRNA sequences. Levels of neomycin mRNA were assayed similarly as a control for RNA content and quality. Levels of transcripts under control of the c-fos promoter were consistently maximal -1.5 h after induction with serum. Neomycin mRNA was reproducibly stable for 2 6 h afterserum addition. Transcripts encoded by pTF1, pTF2, and pTF3were stable since mRNA levels didnot decline significantly for 2 6 h after induction. However, mRNA encoded by pTF4 was unstable, declining within 6 h to the level seen before serum addition (Fig. 2). For some samples in Fig. 2, such as the 4.5-h point with the pTF2 construct, the signal for the neomycin transcript is low. This probably reflects variation in the recovery of RNAand emphasizesthe importance of normalizing the 8globin signals to the control neomycin signals in such S1 nuclease protection assays. Other experiments do not show this variation and confirm the results of Fig. 2. A half-life of 2.4 f 0.5 h (S.E.) was calculated for TF4 mRNA (see Fig. 4). Thus, nucleotides 1520-2080 of the tissue factor mRNA contain one or more destabilizing elements. Deletion Analysisof the Tissue Factor mRNA Destabilizing Element-A series of pTF4 deletions were madeto investigate the sequence requirements of the mRNA destabilizing activity of TF4. 150-nucleotide sequential deletions were made from either the 5' or 3' end of the tissue factor sequence in pTF4. Three additional constructs were made in which sequences were deleted fromboth the 5' and 3' ends of the tissue factor
A
I
IF1
I
HUMAN FACTOR TISSUE IF2 I
IF3
B z:;"m---TFl
j
TF2
TF4
TF3
.&
Nee-
0-Globin
I
lF4
3' UTR
Coding Sequence
-
FIG. 2. Induction and decay of tissue factor.@-globinchimeric transcripts. A, Schematic representation of the regions of the tissue factor transcript screened destabilizing for activity;B, NIH/ 3T3 cells were stably transfected with the indicated constructs. At the indicated time intervals after serum stimulation, total cellular RNAwas isolated andanalyzedby S1 nuclease protection. Neo, fragment protected by neomycin transcripts; @-Globin,fragment protected by tissue factor-&globin chimeric transcripts.
01 02 03 05 04 w ~~~'Ol.534.5d'OU3456'0I53456' ~ ~ ~ ~ , , ~ 0 1 5 3 4 3 d S O l . 5 3 4 5 6 '
.. .
,
ins..", N.0-
T
--09
08
07
Hour,
.
0 15 3 4 5 6 01.5 3 4 . 5 6 0 1 5 3 4 5 6
..
.
P-Gbbin-
FIG. 3. Induction and decay of TF4 .&globin chimeric transcripts with deletions in TF4. NIH/3T3 cells were stably transfected with the indicated constructs. At the indicated time intervals after serumstimulation total cellular RNA was isolated and analyzed by S1 nuclease protection. Neo, fragmentprotectedbyneomycin transcripts; &Globin, fragmentprotected by chimerictranscripts. Schematic representation and analysis of TF4 deletions: A, Dl, D2, and D3; R, D4, D5, and D6; C, D7, D8, and D9.
sequence in pTF4. These nine fragments were amplified by the PCR and Nucleotides subcloned into the 3'-UTR of the mouse @F l globin gene in plasmid
[email protected] constructs were 70-490 lF2 400-mo stably transfected into NIH/3T3 cells and thestability of the TF3 960-1520 TF4 15202080 encoded transcripts was assayed. Transcripts encoded by pD1, pD2, and pD3 declined over the time course of serum stimulation indicating that each retained some mRNA destabilizing activity (Fig. 3A). HowX L1 ever, all three deleted mRNAs were more stable than pTF4 rw mRNA. Mean half-lives of the encoded mRNAs were calculated to be 4.5 k 1.2 h (S.E.) for pD1, 8.5 & 2.7 h (S.E.) for pD2, and 6.8 f 1.0 h (S.E.) for pD3 (Fig.4). Transcripts and pD9 wereall stable FIG. 1. Construction ofchimeric tissue factor .&globin plas- encoded by pD4, pD5, pD6, pD7, pD8, mids. The tissue factor cDNA and the regions insertedinto the XbaI and transcript levels remained elevated 26 h after serum site ofpBSFOSMpG to create pTF1, pTF2, pTF3, and pTF4 are addition (Fig. 3,B and C ) . This indicates that optimal destashown schematically. The human c-fos promoter is represented as a bilizing activity requires nucleotides 1520-2080 of the tissue dotted box with the transcription start site denoted by the horizontal factor mRNA. Deletions lacking nucleotides 1930-2080 enarrow. The polyadenylation site is indicated by An. j3-Globin exons are represented as empty boxes and introns as solid boxes. The coded transcripts that were stable, indicating that thisregion nucleotidesequence numbers refer to the tissue factor cDNA sequence is necessary for destabilizing activity. Interestingly, this region of the tissue factor mRNA has the highest content of A (30).
. . -Hnan Tlswe Factor cDNA
constructs
Tissue Turnover Factor mRNA
2157
- ACT D
+ACT D
0 1.5 3 4.5 6
0 1.5 3 4.5 6
"
Hours in Serum
Neo-
P-Globin-
FIG.6. Effect of actinomycin D on TF4 destabilizing activity. NIH/3T3 cells stably transfected with pTF4 were serum stimu-
2 3 HOURS
' O 0 1
4
5
FIG. 4. Decay of tissue factor .&globin chimeric transcripts. Autoradiographs from four separate experimentswere quantitated by laser densitometry. Chimeric transcript levels were normalized to neomycin mRNA and expressed as a percent of @-globinchimeric levels a t time 0 . 0 , TF4; 0, Dl; A, D2; A, D3; V, D4-D9.
lated. After 1.5 h, actinomycin D (10 pg/ml) was added and total cellular RNA was isolated at the indicated time intervalsand analyzed by SI nuclease protection. Neo, fragmentprotected by neomycin transcripts; @-Globin,fragment protected by chimeric transcripts.
-
Neo-
P-Globin-
~ ~0 1.5& 3 4.5 , 6 N
e
o
-
e
-
0 1.5 3 4.5 6
7
- CHX
*CHX
"
0 1.5 3 4.5 6
Hours in Serum 0 1.5 3 4.5 6
- -
"
-,-
r
-
FIG.7. Effect of cycloheximide on TF4 destabilizing activity. NIH/3T3 cells stably transfected with pTF4 were serum-stimulated. After 1.5 h, cycloheximide (10 pg/ml) was added. Total cellular RNA was isolated at theindicated time intervals and analyzed by S1 nuclease protection. Neo, fragmentprotected by neomycin transcripts; @-Globin,fragment protected by chimeric transcripts.
FIG. 5. Sequence-dependence of the TF4 destabilizing element. NIH/3T3 cells were transfected with pTF4R which is pBSFOSM@Gwith TF4 inserted in the reverse orientation. A t the indicated time intervals after serum stimulation, total cellular RNA was isolated and analyzed by SI nuclease protection. Neo, fragment protected by neomycin transcripts; @-Globin,fragment protected by chimeric transcripts. The data for pTF4 (left) from the experiment of Fig. 6 (left)are provided for comparison.
and U nucleotides and shows similarity to previously described AU-rich motifs that are responsible for rapid mRNA turnover (21,26-28,411. TF4 Is Dependent on Its Specific Sequencefor Destabilizing Function-To determine whether the destabilizing activity of TF4 was a function of its specific sequence or its high content of (A + U) nucleotides, pTF4R was constructed by subcloning TF4 in the reverse orientation into the 3'-UTR of the mouse P-globingene in pBSFOSMPG. This plasmid encodes an mRNA with the same content of A and U nucleotides but a different sequence. pTF4R was transfected into NIH/3T3 cells and the stability of the encoded transcript was determined (Fig. 5). pTF4R mRNA was stable and mRNA levels remained elevated for 2 6 h after addition of serum. This contrasts with the rapid turnover of transcripts encoded by pTF4 (Fig. 5) and suggests that the destabilizing activity of TF4 is dependent on specific mRNA sequences. Effect of Actinomycin D on the Tissue Factor mRNA Destabilizing Element-The transcriptional inhibitor actinomycin D stabilizes chimeric P-globin transcripts containing mRNA destabilizing elements from either the c-fos or theinterferon(3 transcripts (28, 41). To determine the effect of transcriptional inhibitors on the tissue factor mRNA destabilizing element, cells stably transfected with pTF4 were treated with 10 pg/ml actinomycin D 1.5 h after addition of serum. In contrast to untreated cells, TF4 transcript levels remained elevated for 24.5 h in actinomycin D (Fig. 6). Thus, thetissue
Hours inserum
0 .25 .5
c-Fos l3-Globin
PhiX Actin
-
Neo
.
+ 3 5 9 1.5
-- -mm
I
-,-
0 1 ) -
FIG. 8. Effect of serum and cycloheximide on runoff transcription. Cells were treated with serum and cycloheximide ( C H X ) for the indicated times as described under"Experimental Procedures." For each hybridization, a sampleof 4 X 10' cpm of radiolabeled RNA transcripts was used and theautoradiogram was exposed for 15 h.
factor destabilizing element requires ongoing transcription for activity. Effect of Cycloheximide on the Tissue Factor mRNA Destabilizing Element-The endogenous tissue factor transcript in endothelial cells and monocytes is stabilized in response to the protein synthesis inhibitor cycloheximide (9, 10, 42). To determine whetherthe sequences present in TF4 are sufficient to recapitulate this response, cells stably transfectedwith pT4 were treated with 10pg/mlcycloheximide1.5 h after the addition of serum. In contrast to untreated cells, TF4 transcript levels remained elevated 24.5h in cycloheximide treated cells (Fig. 7). In Fig. 7 (right) thecontrol experiment performed without cycloheximide shows a loss of the signal for neomycin transcripts that may be due to decreased RNA recovery. This experiment is similar to those of Fig. 5 (left) and Fig. 2 (pTF4) which do not show a loss of neomycin transcripts; when normalized to the neomycin signals all of these experiments yield similar values for the half-life of pTF4 transcripts. Previous work demonstrates that thec-fos promoter can be
2158
Tissue Factor mRNA Turnover 1490
1500
1510
1520
1530
1!%0
1550
zation by TF4 was sequence-specific and not due to disruption of the @-globinmRNA. Deletion analysis was performed to determine the specific sequence requirements of this destabilizing activity. The last 150 nucleotides of TF4 corresponding to nucleotides 19301640 1650 1660 1670 1680 1690 1700 ~ ~ T C C A T G T ~ A A A A " " 2080 of the tissue factor cDNA (30) were necessary for destat bilizing activity. Any chimeric mRNA lacking this region was 1715 1725 1735 1745 1755 1765 1775 ~ ~ T A A T M T A T ~ ~ ~ T A C A ~ ~region ~ T TF4 contains 79% (A U) nucleotides stable. This of t as well as four copies of the pentamerAUUUA (Fig. 9). These 1790 1800 1810 1820 1830 1840 1850 ~ ~ A l T A A ~ ~ C A ~ A l T ~ A A T A ~ sequences A C A are A T characteristicof a previously described AU-rich motif which is important in regulating the turnover of a class 1865 1875 1885 1895 1905 1915 1925 interleukin-2, ~ ~ ~ ~ l T ~ ~ ~ ~ ~ A T A of short-lived A T mRNAs ~ ~ thatAincludes T -GM-CSF, ~ IFN-0,and c-fos (26). Deletions from the5'end of TF4 1940 1950 1960 1970 1980 1990 2000 reduced itsA mRNA destabilizing activity (Fig. 4). ~ A ~ A ~ ~ ~ ~ A ~ ~ ~ T A T A T M ~ ~ t The tissue factor destabilizing element is unique in that it 2035 2025 2045 2055 2065 2075 appears to require -450 nucleotides of sequence in addition to the most AU-richregion for optimal destabilizing activity. FIG. 9. Nucleotide sequence of TF4. Nucleotide numbers cor- This is in contrast previously to described AU-rich destabilizrespond to the sequence of Scarpati et al. (30). Ouerlined sequence ing elements isolatedfrom the c-fos, GM-CSF, and IFN-P refers to the 3' end of the inverted Alu repeat present in the tissue transcripts which are all 20 h to -2.4 h. This value is comparable to responsetoactinomycin Dbetween the isolatedAU-rich the previously reported half-life of 48 min to 1.5 h for the elements and the full-length IFN-P and tissue factor tranfull-length endogenous tissue factor transcript in endothelial scripts. Differentcell types were used to study thefull-length cells and peripheralblood monocytes (9, 11).Other regions of tissue factor transcript and its isolated mRNA destabilizing the tissue factor transcript of similar size did not affect the elements. Thus, thedifference could be due to differences in D. Alternatively, the differstability of the P-globin transcript indicating that destabili- cellular responses to actinomycin
+
+
Tissue Factor mLRNATurnover
2159
ences may be due to actinomycin D-independent destabilizing and for providing the mouse @-globin gene, and Lisa Westfield for elementspresent inother regions of the mRNAs. These preparing syntheticoligonucleotide primers. sequences may be required in additionto thealready described REFERENCES destabilizing elements for properly regulated turnover of the 1. Furie, B., and Furie, B. C. (1988) Cell 53,505-518 2. Bevilacqua, M. P., Pober, J. S., Majeau, G. R., Cotran, R. S., and Gimhrone, full-length mRNAs. M. A. (1984) J. Exp. Med. 160,618-623 The full-length endogenous tissue factor mRNA in human 3. Bevilacqua, M. P., Pober, J. S., Majeau, G. R., Fiers, W., Cotran, R. S., and Gimhrone, M. A. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,4533-4537 endothelial cells (9, 10) and monocytes (42) and mouse fibro4. Hartzell, S., Ryder, K., Lanahan, A,, Lau, L. F., and Nathans, D. (1989) blasts (45) apparently is stabilized in response to the protein Mol. Cell. Biol. 9 , 2567-2573 5. Moore, K. L., Andreoli, S. P., Esmon, N. L., Esmon, C. T., and Bang, N. synthesis inhibitor cycloheximide. Our results demonstrate U. (1987) J. Clin. Invest. 7 9 , 124-130 that mRNA encoded by pTF4 is superinduced in response to 6. Naworth, P. P., and Stern, D. M. (1986) J . Exp. Med. 163,740-745 7. Naworth, P. P., Handley, D. A,, Esmon, C. T., and Stern, D. M. (1986) cycloheximide (Fig. 7), in partdue to stimulation of the human Proc. Natl. Acad. Sci. U. S. A. 83,3460-3464 c-fos promoter. Previous studies have also reported that the 8. Ranganathan, G., Blatti, S. P., Suhramaniam, M., Fass, D. N., Maihle, N. J., and Getz,M. J. (1991) J. Biol. Chern. 266,496-501 c-fos promoter is stimulated by cycloheximide in mouse fibro9. Scarpati, E. M., and Sadler, J. E. (1989) J. Biol. Chem. 264,20705-20713 blasts and human endothelial cells (9, 40, 43) (Fig. 8). Al- 10. Crossman, D.C., Carr, D. P., Tuddenhum, G. D., Pearson, J. D., and McVey, J. H. (1990) J . Biol. Chem. 265,9782-9787 though the tissue factor promoter is also induced in response 11. Gregory, S. A., Morrissey, J. H., and Edgington,T. S. (1989) Mol. Cell. Biol. t o cycloheximide, Scarpati andSadler (9) were able to circum9 , 2752-2755 vent thisinduction by simultaneous treatment with the tran- 12. Carniero, M., and Schihler,U.(1984) J . Mol. Eiol. 1 7 8 , 869-880 13. Graves, R. A., Pandey, N. B., Chodchoy, N., and Marzluff, W. F. (1987) scriptional inhibitor actinomycin D. In the presentsystem, it Cell 48,615-626 was not feasible to inhibit transcription from the c-fos pro- 14. Heintz, N., Sive, H. L., and Roeder, R. G. (1983) Mol. Cell. Biol. 3 , 539550 moter after induction with cycloheximide due to thesensitiv- 15. Cleveland, D. W. (1988) Trends Biochem. Sci. 13,339-343 Gay, D. A,, Yen, T. J., Lau, J. T. Y., and Cleveland, D. W. (1987) Cell 60, 16. ity of TF4 mRNA stability to actinomycin D. Thus, it is 671-679 . - ~ . . difficult to ascertain the relative contributions of c-fos tran- 17. Gay, D. A., Sisodia, S. S., and Cleveland, D. W. (1989) Proc. Natl. Acad. Sei. U. S. A. 86,5763-5767 scriptional induction and TF4 message stabilization to the 18. Pachter, J. S., Yen, T. J., and Cleveland, D.W. (1987) Cell 51,283-292 superinduction of TF4 tocycloheximide. 19. Casey, J. L., Hentze, M. W., Koeller, D. M., Caughman, S. W., Rouault,T. A., Klausner, R. D., and Harford, J. B. (1988) Science 240,924-928 Several possible mechanisms could mediate cycloheximide20. E. W., and Kiihn, L. C. (1988) Cell 53,815-825 dependentstabilization of TF4. The inhibition of protein 21. Mullner, Reeves, R., Elton, T.S., Nissen, M. S., Lehn, D., and Johnson, K.R. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,6531-6535 synthesis may decrease the concentration of a labile protein Paek, I., and Axel, R. (1987) Mol. Cell. B i d . 7 , 1496-1507 which normally mediates turnover of the tissue factor tran- 22. 23. Brock, M. L., and Shapiro, D.J. (1983) J. Biol. Chem. 258,5449-5455 script; this would be consistent with the observed ability of 24. Brock, M. L., and Shapiro, D. J. (1983) Cell 34,207-214 25. Welsh, M., Nielson, D. A., MacKrell, A. J., and Steiner, D. F. (1985) J . actinomycin D to stabilized TF4 mRNA (Fig. 6). AlternaBiol. Chem. 260. 13590-13594 Caput, D., Beutleg B., Hartog, K., Thayer,R., Brown-Shimer, S., and 26. tively, degradation of the tissue factor mRNA may require Cerami, A. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 , 1670-1674 ongoing translation. Additional experiments are necessary to 27. Shaw, G., and Kamen, R. (1986) Cell 46,659-667 28. Shyu, A., Greenherg, M. E., and Belasco, J. G. (1989) Genes & Deu. 3 , 60distinguish between these possibilities. 72 In summary, we have identified a -600-nucleotide desta- 29. Mackman, N., Morrissey, J. H., Fowler, B., and Edgington, T. S. (1989) Biochemistry 2 8 , 1755-1762 bilizing element in the 3’-UTR of the tissue factor mRNA. A 30. Scarpati, E. M. Wen D., Broze G. J. Jr., Miletich J. P., Flandermeyer, portion of this sequence is similar to previously described R. R. Siege1,’N. R.,’and Sadle; J. E.’(1987) B i o c h e h t r y 26,5234-5238 mRNA destabilizing elements isolated from many rapidly 31. Treismk, R. (1985) Cell 42,8891902 32. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. degraded lymphokine and proto-oncogene transcripts. HowU. S. A. 74,5463-5467 ever, the tissue factor destabilizing element appears to be 33. Graham, F. L., and Van der Eb, A. J. (1973) Virology 52,456-467 34. Parker, B. A., and Stark, G. R. (1979) J. Virol. 31,360-369 significantly larger than these previously described elements. 35. Southern, P. J., and Berg, P. (1982) J . Mol. Appl. Genet. 1,327-341 The tissue factor destabilizing element is sequence-specific 36. Chomczynski, P., and Sacchi, N. (1987) Anal. Biochem. 162,156-159 37. Dittman, W. A,, Kumada, T., Sadler, J. E., and Majerus, P. W. (1988) J. and sensitive to inhibitors of bothproteinsynthesis and Biol. Chem. 263,15815-15822 38. Ley, T. J., Connolly,-N. L., Katamine S., Cheah, M. S. C., Senior, R. M., transcription. A better understanding ofhow cells regulate and Rohhins, K. C. (1989) Mol. Cell.’ Biol. 9,92-99 tissue factor expression by mRNA stability will increase our 39. Moon, A. M., and Ley, T. J. (1991) Blood 77,2272-2284 Greenherg, and Ziff, E. B. (1984) Nature 311,433-438 40. understanding of how the hemostatic system is able to main- 41. Whittemore,M.L.,E.,and Maniatis, T. (1990) Mol. Cell. Biol. 6 4 , 1329-1337 tain the appropriate balance between promotion and inhibi- 42. Brand, K., Fowler, B. J., Edgington, T. S., and Mackman, N. (1991) Mol. Cell. Biol. 11, 4732-4738 tion of blood coagulation. 43. Wilson, T., and Treisman, R. (1988) Nature 3 3 6 , 396-399 Acknowledgments-We thank Dr. Richard Treisman for providing plasmid pF711, Dr. Timothy J. Ley for many valuable discussions
44. Lindsten, T., June, C. H., Ledbetter, J. A., Stella, G., and Thompson, C. B. (1989) Science 2 4 4 , 339-342 45. Lau, L. F., and Nathans, D. (1987) Proc. Natl. Acad. Sci. U. S . A. 84,11821186
