Feb 5, 2016 - From the Molecular Biology Research Program, Henry Ford Hospital, Detroit, Michigan 48202 and the Department of Biological. Chemistry ...
Vol . 268, No. 4:, Issue of February 5, pp. 2493-2499,1993 Printed in U. S.A.
THEJOURNALOF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.
Contribution ofSpecific cis-Acting Elements toActivity of the Mouse Pro-a2(1) Collagen Enhancer* (Received for publication, August 7, 1992)
Robert J. Pogulis and Svend 0.FreytagS From the Molecular Biology Research Program, Henry Ford Hospital, Detroit, Michigan 48202 and the Departmentof Biological Chemistry, Universityof Michigan, Ann Arbor, Michigan 48109
standing of the mechanisms governing collagen gene transcription shouldlead to a better understanding of the myriad of physiological and pathologicalprocesses that involve changes in collagen synthesis. To this end, numerous studies have been aimed at identifying cis-acting regulatory elements in mammalian collagen genes, and the trans-acting factors which bind to them (3, 12). A complex picture is emerging, one which involves regulatory elements residing on both sides of the transcription initiation site. A common feature of several collagen genes is the presenceof a transcriptional regulatory region in the first intron. The al(I), a2(I), al(II), and al(1V) pro-collagen genes all containintronic regions that modulate transcriptional activity (13-18). To date,only two specific cis-acting intronic elements and the trans-acting factors that bind to them have been characterized. An AP-1 site is required for transcriptional stimulation by an intronic region of the human prod ( 1 ) collagen gene (19). In addition, the binding of a novel nuclear factor is necessary for optimal activity of the proal(1V) enhancer (20). Similarly, intronic regions are required for maximal expression of the mouse pro-a2(1) collagen gene, although nospecific cis-acting elements havebeen defined. A 1.1-kb’ intronic segment fulfills the criteriaof a classicalenhancer in transient transfection experiments (13). This segment is also required formaximal transcription from the collagen promoter in transgenic mice. While a transgene containing 2 kb of 5 ’ Type I collagen is a major component of the extracellular flanking sequence from the mouse pro-a2(1) collagen gene is matrix and one of the most abundantly synthesized proteins expressedin the appropriate tissues,expression is signifiin mammals. Collagen levels are modulated during the course cantly lower than that of the endogenous gene (21). Inclusion of numerous biological processes, both normal and patholog- of the 1.1-kb intronic segment results in much higher transical, suchas cell adhesion, cellular differentiation, wound gene expression (22). In the present study,we have identified healing, fibrosis, inflammation, andoncogenic transformation two cis-actingelementswithinthepro-a2(1) collagen en(1-6). A variety of cytokines, growth factors, oncogene prodhancer that arerequired for transcriptional stimulationof the ucts,andother effector molecules can influence collagen collagen promoter in NIH/3T3fibroblasts. One element binds synthesis (2-8). A number of these effects involve changes in a nuclear factor that we have designated “collagen intron specific gene transcription (2, 4, 6-8), although the specific binding factor (CIBF-I) I” and contains motif a cis-acting elements and trans-acting factors responsible have (TGTTTTAA) found in the c-mos and humanpapillomavirus not beenidentifiedin many cases. It is likely that under enhancers. The second site resembles a “GT box” and binds normal circumstances transcription of the al(1) and aZ(1) affinity-purified Spl. pro-collagengenes istightly regulated, sincesteady-state mRNA levels precisely reflect the 2:1 stoichiometry of the MATERIALSANDMETHODS corresponding chains in the type I collagen triple helix (9). Plasmid Construction-Plasmids pR36 (13)and pAZ1003 (21) were Indeed, 2-3-fold changes in type I collagen levels are associated with several disease states ( 5 , 10, 11).Increased under- a gift of B. de Cromhrugghe.pR36 containsan 1106 hp XmnI
Intronic transcriptional regulatory regions are found in several collagen genes. We demonstrate here the importance of two distinct cis-acting elements for the activity of a transcriptional enhancer located in the first intron of the mouse pro-a2(1) collagen gene. Enhancer subfragments were tested for their ability to stimulate a linked promoter following transient transfection into NIH/3T3 fibroblasts. A 92-base pair subfragment retained significant enhancer activity. DNase I footprinting identified a binding site for a nuclear factor (designated CIBF-I)within thissubfragment. Electrophoretic mobility-shift experiments demonstrated that an oligonucleotide containing 18 base pairs from the protected region was sufficient for specific binding of CIBF-I andsuggested that CIBF-I may be related to a protein(s) that binds to the rat c-mos enhancer. A second cis-acting element, identified by electrophoretic mobility-shift and methylation interference experiments, bound affinity-purified Spl and an Spl-likeprotein present in crude nuclear extracts. Small deletion mutations in either the CIBF-I or Spl site abolished factor binding in vitro and reduced enhancer activity in vivo. Our results represent the first identification of specific cis-acting elements required for full activity of the pro-aZ(1) collagen enhancer.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact. $ To whom correspondence should he addressed Molecular Biology Research Program, Education & Research Bldg., Rm. 7041, Henry Ford Hospital, 2799 W. Grand Blvd., Detroit, MI 48202. Tel.: 313876-1949,876-1973; Fax: 313-876-2380.
fragment from the first intron of the mouse pro-a2(I) collagen gene inserted in the BglII site of pAlOCAT2. pAZ1003 contains the proa2(I)collagen promoter (-2000 to +54) fused to the bacterial chloramphenicol acetyltransferase(CAT) gene. PlasmidspRP36and pRP40 were generated by amplifying the X m n I fragment from a pR36 The abbreviations used are: kh, kilohase pairs, hp, base pair(s); CAT, chloramphenicol acetyltransferase; SV40, simian virus 40; HPV, human papillomavirus.
2493
2494
cis-Acting Elements
in the (u2(I)Collagen Enhancer
template using the polymerase chainreaction(23)withprimers containing BglII linkers and cloning thepurified fragments into the RglII site of pAlOCAT2 and the BamHI site of pAZ1003, respectively. pRP40 contains this insert in its natural orientation withrespect to the direction of transcription, whereas in pRP36 the orientation is reversed. Preparations of pRP36 isolated from two independent bacterial clones displayed activity equal to that of pR36 in transient transfection assays. Plasmids pG3ES and pG3SE were constructed by isolating an EcoRI-SphI subfragment of the intronic XmnI fragment, blunting a t both ends with T4 polymerase, and cloning the blunt-ended fragment into the HincII site of plasmid pGEM3 (Promega).pG3ESandpG3SEcontaintheEcoRI-SphIfragmentin opposite orientations. pRP92 contains the HindIII-SphI fragment (excised as a HindIII fragment from pG3SE), which was blunted and inserted into the BglII site of pAlOCAT2. Plasmids pRP292 and pRP146 contain the fragments indicatedFig. in 1,blunted andcloned in the BglII site of pAlOCAT2. All constructions were checked by restriction analysis. Plasmids used in transient transfections were purified by cesium chloride gradientcentrifugation(24),andthe preparations were verified to contain greater than 90% supercoiled DNA by agarose gel electrophoresis. Generation of Deletion Mutations-Site-specific deletion mutations were introduced into the enhancer region using a polymerase chain reaction technique described by Ho et al. (25). Two sets of primers wereused two amplifyoverlapping fragmentsandintroducethe mutation.Thesefragments were purified and fused in a second amplification, to generate a mutant XmnI fragment. Mutant fragments were cloned intotheparental vectors togeneratemutant plasmids. These mutant constructs were checked by restriction analysis and Maxam-Gilbert (26) sequencing. Large scale preparations of eachmutant plasmid were made from two independentbacterial clones,andinall cases, theactivities were similarintransient transfection assays. Cell Culture, Transient Transfections, and CAT Assays”NIH/3T3 cells (ATCC CRL 1658) and HeLa cells (ATCC CCL 2) were grown in Dulbecco’s modified Eagle’s medium supplemented with 10 mM HEPES, pH 7.4, 2 mM glutamine, 100 pg of streptomycin, and 100 units of penicillin/ml, and either 10% calf (NIH-3T3) or fetal (HeLa) bovine serum (growth medium). Cells were grown in ahumidified incubator in anatmosphere of90% airand10% CO,. Transient transfections performed by the CaPO4-DNA coprecipitation method (27). Cells were plated approximately 16 h prior to transfection a t a density of 3 X IO5cells/90-mm (diameter) dish. On the following day, the growth medium was replaced, and a CaP04-DNA coprecipitate containing 10 pg of the test plasmid and 5 pg of pRSVPGal (28), to control for transfection efficiency, was applied to thecells. The cells were shocked with 15% glycerol 5 h later, washed, and the growth medium replaced. Cells were harvested 40 h after transfection. CAT assays and p-galactosidase assays were performed as previously described (24). CAT activity was quantified with an Ambis Radioanalytic Imaging System (SanDiego, CA) and normalized to p-galactosidase activity. DNase I Footprinting-Plasmid pG3ES was digested with HindIII and 5’end-labeledaccording tostandard procedures (24). The HindIII-SphIenhancerfragment wasreleasedbydigestion with BarnHI andgel-purified. NIH/3T3 nuclear extractswere prepared by the method of Dignam et al. (29). DNaseI footprinting was performed as previously described (30, 31). Binding reactions were carried out in a final volume of 50 pl. 10,000 cpm (-2.5 fmol) [32P]DNAwas incubated with 20 pg of nuclear ext,ract and5 pg of poly(dI.dC) in a mixture containing 12.5 mM HEPES, pH 7.5, 50 mM KC], 10% (w/ v) glycerol, 0.005% Nonidet P-40, 1.0 pM ZnSO,, and 0.5 mM dithlothreitol. The binding reaction was allowed to proceed on ice for 20 min, combined with 50 pl of 10 mM MgC12-5 mM CaCI,, and incubated at room temperature for 1 min. Samples were thentreatedwith 1 minandthe freshly diluted DNase I (CooperBiomedical)for reaction stopped by the addition of 90 p1 of 20 mM EDTA, 1%SDS, 0.2 M NaCI, 250 pg/ml yeast tRNA. Extraction with phenol/chloroform (1:l) and ethanol precipitation followed. The precipitate was washed with 70% ethanol, dried under vacuum, and resuspended in 5 pl of a loading buffer containing 80% formamide. Samples were then heat-denatured and electrophoresed through an 8% denaturing polyacrylamide gel. Electrophoretic Mobility-shift Assay-End-labeled restriction fragments or double-stranded synthetic oligonucleotides were incubated with 1-5pgof nuclear extract and 1 fig of poly(dI.dC) in a final volume of 20 pI containing 10 mM HEPES, pH 7.9, 10% glycerol, 0.005% Nonidet P-40, 50 mM KCI, 0.25 mM dithiothreitol, and 0.25
mM phenylmethylsulfonyl fluoride. Binding was allowed to proceed a t room temperature for 15 min, and samples were loaded directly onto a 4.5%, low ionic strength, non-denaturing polyacrylamide gel. Assays containing affinity purified Spl (a gift ofA. Courey and R. Tjian)) were performed similarly, except no poly(dI.dC)was added. Methylation InterferenceAssay-The intronic PvuII-Hind111 fragment was labeled at the HindIII site and partially methylated at the guanine residues (26). This labeled fragment was used to perform methylation interference analysis aspreviously described (32). Binding reactions were set up as above, but scaled up IO-fold. The free and bound [32P]DNA were resolved on a non-denaturing gel and recovered by electroelution. The DNA was then ethanol precipitated and subjected to piperidine cleavage (26). Piperidine-cleaved DNA corresponding to the shiftedcomplex, free fragment, and inputDNA were subsequently electrophoresed through an 8%denaturing polyacrylamide gel. RESULTS
Transcriptional Stimulationby Subfragments from the First Intron of the Pro-aS(I) Collagen Gene-An intronic fragment spanning nucleotides +418 to +1524 (XmnI-XmnI fragment, Fig. 1)was previously demonstrated to enhance transcription when linked to either the pro-a2(1) collagen or SV40 early promoter (13). Two findingssuggested that sequences between +748 (EcoRI) and +lo40 (SphI)may be important for enhancer activity. First, no significantdecrease in stimulatory activity was observedupondeletion of sequencesbetween t1040 and +1524, and second, noenhanceractivity was detected between +134 and +748 (13). Therefore, we focused on the region between nucleotides +748 and +lo40 in our attempts to identifyspecific cis-acting elements that contribute to enhanceractivity. Thefull-length 1106-bp XmnIfragmentand various subfragments were tested for their ability to stimulate the SV40 and pro-aB(1) collagen promoters following transient transfection into NIH/3T3fibroblasts. The full-length XmnI fragment enhanced transcriptionfrom the SV40 promoter 9fold. Wheninserted 1.6 kb downstream of thepro-a2(I) collagen promoter, this fragment enhanced transcription 3.5fold (Fig. 1, pRP36 and pRP40). These results are in qualitative agreement with the results of Rossi and de Crombrugghe (13). Two nested subfragments retained significant stimulatory activity. Plasmids pRP292 and pRP92, containing 292 bp (EcoRI-SphI) and 92 bp(HindIII-SphI)fragments,respectively, displayedtranscriptional activities3-4-fold greater than that of pAlOCAT2 (Fig. 1). Sequencesfrom +802 to +948 (pRP146)had no stimulatoryeffect. These datasuggest that a positive cis-acting element(s) is located between nucleotides +948 and +lo40 (HindIII-SphI). The Intronic HindIII-SphI Fragment Contains a Binding Site for a Nuclear Factor-Based on the results described above, we examined the HindIII-SphI fragment for binding by specific nuclear factors. DNase I footprinting experiments identified a 25-bp protected region extending from +983 to +lo07 (Fig. 2, bracketed region I ) , indicating the bindingof a nuclear factor(s) to sequences within this region. Although the 3’ boundary of the footprint was not well demarcated, subsequent experiments(below) demonstrated thatsequences between+983 and +lo07 are necessary and sufficient for specific binding of a nuclear factor. The nucleotide sequence spanning footprintI is shown in thelower panel of Fig. 2. The absence of cleavage near the top of theDNase I ladders (bracket marked ns) was evident across a range of DNase I concentrations and spanned a region that included only 5 bp of intronic sequence before extending into thecloning vector. Numerous other experiments failed to demonstrate any specific interaction with this region (data not shown). Consequently, we more thoroughly examined the DNA-protein interaction involving sequences within footprint I.
cis-Acting Elements
in the a2(I) Collagen Enhancer
2495
BoI I1
pAl OCAT
SV40 e a r l y
CAT gene
promoter
Barn HI
pAZl003 CAT gene
mouse alpha 20) collagen promoter
FIG. 1. Transcriptional stimulation by various intronic subfragments of the pro-a2(1) collagen gene. Upper panel,structure of the vectors used to construct various test plasmids. Vertical arrows, sites in which intronic fragments were inserted. Middle panel, representative CAT assays. Lower panel, schematic representation of the intronic fragments contained in the various test plasmids and tabulation of relative transcriptional activity. The CAT activity (+ S.E.) of each test plasmid is expressed relative to thebasal activity of the parental vector (either pAZ1003 or pAlOCAT2), which was set equal to 1.0. The data shown are averages of four or more independent transfection assays. Numbers represent nucleotide positions with respect to the transcription initiation site. X , XmnI, E, EcoRI, P,PuuII, H,HindIII, S, SphI.
Ttt?T* Test Plasmid Vector
Activity
Fragment
Relative CAT
lntronic
"
8 pRP40
pAZ1003
pRP36
pAlOCAT
pRP292
pA1OCAT
pRP146
pAlOCAT
pRP92
PA1 OCAT
To determine whether sequences within footprint I were required for binding, the effect of a 20-bp deletion (AZ,Fig. 2, lower panel) on factor binding was assessed by an electrophoretic mobility-shift assay. Incubation of the wild-type HindIII-SphI fragment with NIH/3T3 nuclear extracts generated two major complexes (Fig. 3, lanes 2 and 6). The AI mutation abolished formation of one of these complexes (Fig. 3, complex Z, compare lanes 2 and 4 ) . Competition experiments demonstrated that formation of complex I was specific (Fig. 3, lanes 5-10),while formation of the other major complex (ns)was due to nonspecific interactions (data not shown). Thus, a site within the HindIII-SphI region is necessary for binding of a specific nuclear factor which we designate here as collagen intron-binding factor I (CIBF-I). Furthermore, an 18-bp oligonucleotide containing nucleotides +987 to +lo05 from this site is sufficient for specific binding of CIBF-I (see below). Inspection of the CIBF-I site revealed that it contains an 8-bp motif (TGTTTTAA) that is also found in footprinted regions of the rat c-mos (33) and HPV (34) enhancers. To determine whether CIBF-I was similar to the protein(s) that interacts with this motif in the c-mos enhancer, we performed electrophoretic mobility-shift experiments with synthetic oligonucleotides. The sequences of the oligonucleotidesused are shown in Table I. Incubation of the CIBF-I oligonucleotide
\@$
*@8
f
II 1 I
I I
I I
X
E P
H
s
I
$I
I H H H
3.5
f
0.4
9.3
f
1.0
2.9
2
0.7
0.9
f
0.1
3.8
+ 0.4
X
with NIH/3T3 nuclear extracts generated a complex that was specifically competed away bythe CIBF-I and DS3 (from the c-mos enhancer, see Table I) oligonucleotides, but not by oligonucleotidescorresponding to several other factor-binding sites (Fig. 4A). Moreover, nucleoprotein complexes formed with the CIBF-I and DS3 oligonucleotides had similar mobilities in a non-denaturing gel (Fig. 4B). Together these data suggest that the DS3 site is capable of binding CIBF-I and raise the possibility that CIBF-I is closely related to a protein previously shown to bind the c-mos and HPV enhancers. The second complex (complex ZZ) in Fig. 4B was detected only at higher protein concentrations and has notbeen characterized. The observations that the CIBF-I oligonucleotide generated more of the CIBF-I complex (Fig. 4B)and competed at least 4-fold more effectively than DS3 for CIBF-I binding (Fig. 4A) suggest that additional sequences outside the TGTTTTAA motif influence binding affinity. We also examined nuclear extracts from several different cell lines for the presence of CIBF-I and found it to be present in Swiss 3T3 and 3T3-Ll fibroblasts, HeLa cells, RPMI 2650 cells, CV-1 cells,and EL4 murine leukemic cells, indicating that CIBF-I is ubiquitously distributed (data not shown). An Spl-like ProteinBinds to a GT Box in the PvuZZ-HindZZZ Fragment-Although the PvuII-Hind111 fragment alone was not sufficient to confer transcriptional enhancement (Fig. 1,
cis-Acting Elements the in
2496
A A
DKaw I: Lane:
1
2
3
4
5
N E : - - -+ +
6 P + &
(u2(I)Collagen Enhancer Robe:
wt
amp:
- -
~
AI ”
NE: - +- +
wt ”
s s n n
- +
+ + + +
5 6
7 8 9 1 0
ns
+I003
N,
I
”
M
+983
F, 1 2 1
.9
I
11.1:
1
I
.
T;ACI~IFI~C’JL:;I:I’I~A/~A~A~%AI~,~CAGI A ~ : , ,: A, b. . A , . ~ . . . > ~ A : ~ l V l 4 l : : .lA ’ ; A l ; l t A ~
I
FIG.2. DNaseI
Ai
I
footprinting of the92-bpHindIII-SphI
fragment. Upper panel, The HindIII-SphI fragment was labeled a t the HindIII 5’ end, incubated without (-, lanes 1-3), or with (+, lanes 4-6), nuclear extract prepared from NIH-3T3 cells. Samples were treated with increasing amounts of DNase I as indicated. The area of specific protection ( I ) ismarked by a bracket. Numbers represent nucleotide position with respect to the transcription initiation site. The bracketlabeled n9. marks aregion of nonspecific interaction. CIA, a Maxam-Gilbert sequencing ladder. Lower panel, nucleotidesequencefrom +978 to +lo12 of the mouse pr0-02(1) collagen gene. The region the region protected from DNaseI digestion ( I ) is marked by a bracket above the sequence. A bracket below the sequence identifies the bases deleted to generate the site-specific deletion mutation, AI.
3 4
FIG.3. Sequences within footprint I are required for specific bindingof a nuclear factor to the HindIII-SphI fragment. A 5’ end-labeled wild-type HindIII-SphI fragment ( u J ~lanes , I-2,5IO) and one with 20 bases deleted from footprint I ( A I , lanes 3-4), were incubated without (-1 and with (+) nuclear extract ( N E ) from NIH/3T3 fibroblasts. Binding reactions were subsequently fractionated on a 4.5% nondenaturing polyacrylamide gel. Competition experiments were performed with 10 and 30 ng of a specific (s, lanes 7 and 8, respectively) or nonspecific ( n s , lanes 9 and 1 0 ) unlabeled competitor fragment.The specific competitor (s)was a 292-bp EcoRISphI fragment from the first intron of the mouse pro-02(1) collagen gene. The nonspecific competitor (ns) was a 248-bp NsiI-PstI fragment frommouse aP2 promoter (42). The specific complex ( I ) , a nonspecific complex (ns),and the free fragment ( F ) are indicated.
factor to thePuuII-Hind111 fragment. Binding of CIBF-I and Spl Is Required for Full Enhancer Activity-To determine whether CIBF-I and/or S p l binding was requiredforenhanceractivity,the full-length XmnI fragment (1106 bp) containing either the AI or AS mutation pRPl46), two different intronic fragments (+802 to +1800 was inserted into pAZ1003 and pAlOCAT2 and tested for the and +54 to +958) which overlap only in the PuuII-Hind111 ability to enhance transcription in NIH/3T3fibroblasts. The region were previously shown topossess significant enhancer relative transcriptional activities of these mutant constructs activity (13).Inspection of this region revealed that it contains are summarized in Table 11. Binding of both proteins was a “GT box” (GGGGTGG), a novel Spl-binding motif origi- required for maximal transcriptional enhancement when the (35). nally identified in the P4 and P5 BPV promoters XmnI fragment was positioned downstream of the pro-a2(1) Electrophoreticmobility-shiftexperimentsdemonstrated collagen promoter, a context resembling that of the endogethat the GT motif was required for binding of the PuuIInous gene. The AI mutation (CIBF-I site) virtuallyabolished HindIII fragment by affinity-purified S p l as well as an Spl- enhancer activity, and theAS mutation effected a significant like factor presentin nuclear extractsof NIH/3T3 and HeLa reduction. The AS mutation also reduced stimulation of the cells (Fig. 5 A ) . Affinity-purified S p l formed a total of three heterologous SV40 early promoterby the XmnI fragment, but complexes with this fragment; the migration pattern varied the AI mutationhadno effect inthiscontext(Table 11, dependingonthepreparation of S p l used. Three of the pRP36AS and pRP36A1, respectively).Becausedeletion of complexes formed with crude nuclear extracts from NIH/3T3the CIBF-I site did not affect stimulation of the SV40 proo r HeLa cells comigrated with the complexes formed with moter by the 1.1-kb XmnI fragment,we considered itpossible affinity-purified S p l (Fig. 5A, complexes I, 11, and 111, lanes thatthisfragmentcontainedotherelement(s)capable of 2-7). The other complexesobserved (ns)are the result of compensating for theAI mutation. Therefore, we deleted the nonspecific interactions (data not shown). Methylation inter-CIBF-Isite from the92-bpHindIII-SphIfragmentinan ference analysis showed that the protein responsible for the attempt to isolate the effect of CIBF-I. In contrast to the formation of complex I1 makes contact withspecific guanine results obtained with the 1.1-kb XmnI fragment, the stimunucleotideswithinthe G T motif (Fig. 5B). Furthermore, latory effect of the smaller HindIII-SphI fragment on the oligonucleotides corresponding to knownS p l sites effectively SV40 promoter was effectively eliminated by the AI mutation competed for the formation of all three specific complexes (Table 11, pRP92AI). Thus, binding of both CIBF-I and Spl (data not shown). Based on these results we conclude that the contribute to stimulationof the native pro-a2(1)collagen and protein detected in NIH/3T3 and HeLa nuclear extracts is heterologous SV40promoters by fragmentsfromthefirst likely to be Spl. Deletion of a 10 bpregion encompassing the intron of the pro-a2(1) collagen gene. G T box (AS, Fig. 5B) greatlyreduced or abolished bindingof DISCUSSION both purified Spl and the factor(s) detected in nuclear extracts (Fig. 5A, lanes9-14), demonstrating that sequences The activityof the mouse pro-a2(1) collagen enhancer, and from +871 to +880 are required for binding of an Spl-like of other collagen enhancers as well, is probably mediated by
cis-Acting Elements in the 012(I) Collagen Enhancer
2497
TABLE I Oligonucleotides usedin electrophoretic mobility-shift experiments The abbreviation used is: MSV LTR, murine sarcoma virus long terminal repeat. Oligo Sequence" Origin of sequence 5' - TCCCCTGTTTTAAATAGG - 3' aS(1)collagen enhancer, CIBF-I 3 ' AGGGGACAAAATTTATCC 5' mouse
5' - CACAGTGTTTTAACTGTA-3'
DS3
rat c-mos enhancer
3' -GTGTCACAAAATTGACAT-5'
C/EBP
5' - GAACTAAGCAATCAGTTCGCT- 3' 3'-CTTGATTCGTTAGTCAAGCGA-5'
MSV LTR
SPl
5'-GCTCCAGGCGGGGGCGGGGCCCGGGTTCGG-3'
Human AS promoter
3'-CGAGGACCGCCCCCGCCCCGGGCCCAAGCC-5'
AP-1 a
Human collagenase promoter
5'-TAAAGCATGAGTCAGACACACCT-3' 3'-ATTTCGTACTCAGTCTGTGTGGA-5'
Boldface type identifies positions a t which the sequences of CIBF-I and DS3 are identical.
m
+871
I
+881
II
os
a
'"
.
w.
1 2 3 os AS
I
2 3
4
5 6 7
8 9
lOll12
1 2 3 4
FIG. 4. Binding of a similar nuclear factor(s) to the CIBF-I and DS3 oligonucleotides.A , competitive electrophoretic mobilityshift assay. The CIBF-I oligonucleotide was incubated without (-, lune 1) and with (+, lunes 2-12) nuclear extract from NIH-3T3 cells. Lane 2, no competitor; lunes 3-22, reactions contained a 10 (IO) or 40-fold (40) molar excess of various unlabeled oligonucleotides, as indicated. The sequences of the oligonucleotides used are shown in Table I. B, comparison of the complexes formed with the DS3 and CIBF-I oligonucleotides. Lanes 1 and 2, DS3 oligonucleotide incubated without (-) and with (+) of NIH/3T3 nuclear extract; lunes 3 and 4, CIBF-I oligonucleotide incubated without and with nuclear extract. The specific CIBF-I complex (CIBF-I),a second, uncharacterized complex ( I I ) , and the free fragment (F)are indicated.
t h e binding of multiple trans-acting factors. The minimum amount of DNA sequence required for full activity of several collagen enhancers (a2(1), al(II), and al(1V)) is substantial (13, 20, 36), and in some cases, two or more large fragments with onlya modestamount of commonsequencedisplay significant stimulatory activity (13,20). Thus, conceivable it is that these intronic regulatory regions contain multiple cisacting elements, some of which may be functionally redundant. To date, onlytwo specificcis-acting elements havebeen identified in collagen gene introns. In the experiments described here, we have demonstrated the importance of two distinct cis-acting sequence elements for the activity of the intronic mousepro-aB(1)collagen enhancer. These results significantly expand the repetoire of factors that have been demonstratedtomodulate collagengene transcription by binding to specific intronic elements. Binding of CIBF-I was required for enhancer activity when the XmnI fragment was inserted 1.6 kb downstream of the pro-a2(1) collagen promoter, a spatial arrangement resem-
FIG. 5. Binding of an Spl-like protein to the 146-bp PuuIIHindIII fragment. A, comparison of the DNA binding activity of NIH-3T3 and HeLa cell nuclear extracts to that of affinity-purified Spl. Thewild-type PuuII-Hind111fragment (lunes 1-7) and one with a 10 base deletion encompassing nucleotides +871 to +880 (AS, lunes 8-14) were incubated with nuclear extract (NE) or varying amounts of two different preparations of affinity-purified S p l (Spl),as indicated. Lunes 1 and 8, no protein (-). The specific S p l complexes (II I I ) , two nonspecific complexes (ns), andthe free fragment are indicated. B, upper panel, methylation-interference analysis of complex 11. A partially methylated PuuII-Hind111 fragment was 5' endlabeled at the HindIIIsite and incubated with NIH-3T3 nuclear extract. F, free, B, bound, I, input piperidine-cleaved t3*P]DNA.A region in which multiple G residues are undermethylatedin the bound fraction is indicated by a bracket. Numbers represent nucleotide position with respect to thetranscription initiation site. Lower panel, nucleotide sequence of the region containing the GT box. Contacts identified by methylation interference are marked by solid circles(0) below the sequence. The position of a site-specific deletion ( A S ) introduced into thisregion is marked by a bracket.
bling that of the endogenous gene. This leads us to believe that CIBF-I is necessary for maximal expression or the proa2(I) collagen gene in vivo. Although stimulation of the heterologous SV40 promoter by the 292-bp HindIII-SphI fragment was entirely dependent on the CIBF-I site, thiswas site not required for stimulation of the SV40 promoter by the 1.1kb XmnI fragment. These results suggest the presenceof one or more additional cis-acting elements within the XmnI fragment that can compensate for absence of the CIBF-I site under certain conditions. The more stringent requirement for this site when the XmnI fragmentis positioned downstream of its native promoter may be due to thespecific set of trunsacting factors that bind to this particular promoter, and/or
2498
cis-Acting Elements
in the a2(1) Collagen E n h a n c e r
TABLEI1 Effect of mutations on enhancer uctivity Transient transfection of NIH/3T3 fibroblasts was performed as describedunder"MaterialsandMethods." Mutant plasmidsare derived from the constructsdepictedinFig. 1 and contain the deletions shown in Fig. 2 (AI) and Fig. 5 (AS).
boxes in two different bovine papillomavirus promoters mediatestransactivation by the E2 enhancerproteininthe absence of a functional E2-binding site, suggesting that Spl can productively interact with other regulatory proteins (35). A transcriptional regulatory region in the first intron of the human pro-al(1) gene also contains an Spl motif, which is Relative Site protected in DNase I footprintingexperiments(16).The Enhancer Plasmid CAT deleted activityb functional significance of this element has not been estabactivitp lished. % The mouse pro-a2(I) collagen enhancer was previously None 3.5 f 0.4 100 pRP40 demonstrated to be active in NIH/3T3 cells, but inactive in CIBF-I 1.2 f 0.1 8 pRP40AI the lymphoid cell line, S194 (13). Based on these results, it 1.6 f 0.5 24 pRP40AS SPl was concluded that the pro-a2(I) collagen enhancer is cell None 9.3 f 1.0 100 pRP36 type-specific. We have found that the mouse pro-a2(I) colla7.8 f 1.0 82 CIBF-I pRP36AI gen enhancer is active in several cell types including mouse 2.4 f 0.1 17 pRP36AS SPl fibroblasts(NIH/3T3and3T3-L1),human cervicalcarcinoma cells (HeLa), and African Green Monkey kidney cells None 3.8 f 0.4 100 pRP92 (CV-I), all of which express CIBF-I and Spl.' The pro-a2(1) CIBF-I 1.1 f 0.1 4 pRP92AI ' CAT activity (k S.E.) is expressed relative to the basal activity collagen enhancer is also active in mouse kidney proximal of the corresponding parental plasmid (either pAZ1003 or pA10CAT2) tubule epithelial cells grown in high glucose medium (39). possessing a promoter,but no intronic enhancer sequences. The Together, these results are consistent with the fact that a relative activity of each parent plasmid was assigned a value of 1.0. number of non-fibroblastic cell types can synthesize type I Data represent averages of three or more independent transfection collagen,includingkidney epithelial cells, smooth muscle assays. Results are expressed as a percentage of the stimulation (above cells, and macrophages (39-41). Thus,theactivity of the basal promoter activity) obtained from the corresponding wild-type mouse pro-a2(I) collagen enhancer is not limited to fibroblasts, although it may function only in cell types that are enhancer fragment. capable of synthesizing typeI collagen. In conclusion, we have identified two specific factors, CIBFspatialrelationships between between promoter-andenhancer-bound factors. Consistent with this view, an intronic I and Spl, which contribute to maximal activity of the proregulatory region in the human pro-al(1) collagen gene dis- d ( 1 ) collagen enhancer. Further characterization of these and other factors that bind to intonic regulatory regions in plays promoter-, position-, and orientation-dependent activity the collagen genes may provide insight into the biochemical (14, 16, 37, 38). The CIBF-I site containsa motif (TGTTTTAA) found in mechanisms governing collagen expression under normal and footprinted regions of the rat c-mos (33) and human papillo- pathological conditions. mavirus enhancers (34). We have provided the first mutaAcknowledgments-We are grateful to B. de Crombrugghe for tional evidence that this motif can contribute to enhancer providing plasmids pR36 andpAZ1003 and toA. Courey and R.Tjian function. Competition experiments with synthetic oligonucle- for providing affinity-purified Spl. We thank M. LaPointe and T. otides suggested that the CIBF-I site and homologous the site Geddes for comments on the manuscript. fromthec-nosenhancerinteract with thesame or very REFERENCES similar factor(s). Resultsfrom previous studies raise thepos1. Dhawan, J., and Farmer, S. R. (1990) J . Biol. Chem. 266,9015-9021 sibility that binding to the TGTTTTAA motif may involve 2. Weiner. F. R.. Shah. A,. Smith. P. J.. Ruhin., C. S... and Zern. M. A. (1989) . . Bioc&m&tb 28,4094-4099' several proteins. Three or more heparin-Sepharose fractions 3. de Crombrugghe, B., Vuorio, T., Karsenty, G., Maity, S., Rutheshouser, E. from mouse L cell nuclear extracts are necessary for binding C., and Goldberg, H. (1991) Ann. Rheum. Dis. 50,872-876 4. Goldstein, R. H. (1991) Am. J. Physlol. 2 6 1 , L29-L40 to this motif in the c-mos enhancer (33). In addition, UV 5. Phan, S. H. (1989) in Lung Cell Biology (Massaro, D., ed) pp. 907-980, cross-linking studies indicate that upfour to polypeptides (80, Marcell Decker, New York 6. Schmidt, A., Setoyama, C., and de Crombrugghe, B. (1985) Nature 3 1 4 , 59,54, and50 kDa) bindto the HPV TGTTTTAA motif (34). 286-289 We have also demonstrated that binding of Spl to a G T 7. Rossi, P., Karsenty, G., Roberts, A. B., Roche, N. S., Sporn, M. B., and de Crombrugghe, B. (1988) Cell 52,405-414 box in the PuuII-Hind111 fragment is required for maximal 8. Yang, B.-S., Geddes, T. J., Pogulis, R. J., de Crornbrugghe,B., and Freytag, activity of the pro-a2(1)collagen enhancer. In contrast to the S. 0. (1991) Mol. Cell. Biol. 11, 2291-2295 9. Vuust, J., Sobel, M. E., and Martin, G. R. (1985) Eur. J. Biochem. 1 5 1 , CIBF-I site, thisG T box is necessary for stimulation of both 449-453 the natural pro-a2(I)collagen promoter and theheterologous 10. Bonadio, J., Saunders, T.L., Tsai, E., Goldstein, S. A,, Morris-Wiman, J., Brinkley, L., Dolan, D. F., Altschuler, R. A,, Hawkins, Jr., J. E., Bateman, SV40earlypromoter.Theinability of the146-bp PuuIIJ. F., Mascara, T.. and Jaenisch. R. (1990)Proe. NatL Acad. Sci. U. S. A. Hind111 fragment to stimulatetranscription suggests that 8 7 , 7145-7149 11. Herrmann, K., Heckmann, M., Kulozik, M., Haustein, U. F.,and Krieg, T. binding of Spl to the GT box in thisregion is productiveonly (1991) J.Inuest. Dermatol. 97,219-222 when other factors are boundelsewhere within the XmnI 12. Vuorio, E., and de Crombrugghe, B. (1990) Annu. Reu. Biochem. 59,837872 fragment. To the best of our knowledge, this GT box is the 13. Rossl, P., and deCrombrugghe, B. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,5590-5594 first functional Spl-binding site identified within a distally 14. Boast, S., Su, M-W., Rarnirez, F., Sanchez, M., and Avvedimento, E. V. located, naturallyoccurring transcriptional regulatory region. (1990) J. Biol. Chem. 2 6 5 , 13351-13356 15. Bornstein, P., and McKay, J. (1988) J. Biol. Chem. 2 6 3 , 1603-1606 Corey et al. (30) previously demonstrated that distally and 16. Bornstein. S. (1987) ~,P.., McKav. J.. Morishima. J. K.. and Devaravalu. ~. . proximally bound S p l can synergistically activate transcrip- .~ Proc. Natl. Acad. Si; d 5'. A. 84,669-8873 tion over a 1.7-kb distance. Although S p l itself has not been 17. Horton, W., Miyashita, T., Kohno, K., Hassell, J.R., andYamada,Y. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,8864-8868 demonstratedtobindtothepro-a2(I) collagen promoter, 18. Killen, P. D., Burhelo, P. D., Martin, G. R., and Yamada, Y. (1988) J. Biol. Chem. 263,12310-12314 binding of Spl to the thecollagen enhancer may contribute 19. Liska, D. J., Slack, J. L., and Bornstein, P. (1990) Cell Reg. 1 , 487-498 t o transcriptional stimulation by interacting with other pro- 20. Burhelo, P. D., Bruggernan,L. A., Gahrlel, G. C., Klotman, P. E., and moter-bound factors or attracting additional stimulatoryfac* R. J. Pogulis and S. 0. Freytag, unpublished data. tors via protein-protein interactions. Binding of Spl to GT '
"" ~
""
~
~
~~~~
~~~
~~~
cis-Acting Elementsin the a2(I) Collagen Enhancer Yamada, Y. (1991) J. Biol. Chem. 266,22297-22302 21. Khillan, J. S., Schmidt, A,, Overbeek, P. A,, and deCrombrugghe, B. (1986) Proc. Natl. Acad. Sei. U. S. A. 83, 725-729 22. de Crombruggbe, B., Karsenty, G., Maity, S., Vuorio, T., Rossi, P., Ruteshouser, E. C., McKinney, S. H., and Lozano, G. (1990) Ann. N. Y. Acad. Sci. 580,88-96 23. Saiki, R. K., Gelfand,D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B., and Erlich,H. A. (1988) Science 239, 487-491 24. Sambrook, J., Fritsch, E. F., and Maniatis,T. (1989) Molecular Cloning:a Laborotory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 25. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K., and Pease, L. R. (1989) Gene (Amst.) 77,51-59 26. Maxam, A. M., and Gilbert, W. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 560-564 27. Parker, B., and Stark, G. (1979) J. Viral. 3 1 , 360-369 28. Karsenty, G., and de Crombrugghe, B. (1990) J . Biol. Chem. 265, 99349942 29. Dignam, J. D., Lehowitz, R. M., and Roeder, R. G. (1983) Nucleic Acids Res. 11,1475-1489
2499
30. Courey, A. J., Holtzman, D. A., Jackson, S. P., and Tjian, R. (1989) Cell 59,827-836 31. Anderson. G. M.. and Frevtae. S. 0. (1991) Mol. Cell. Biol. 11. 1935-1943 32. Sen, R., and Baltimore, D: ( 3 8 6 ) Cell 46,'705-716 33. van der Hoorn, F. A. (1987) J. Mol. Biol. 193, 255-266 34. Nakshatri, H., Pater, M. M., and Pater, A. (1990) Virolo y 178 92 103 35. Li, R., Knight. J. D., Jackson,. S. P.,.~Tiian, R.. and Botctan, M: R.11991) Cell 65,293-505 36. Yamada Y. Miyashita T. Sava ner P. Horton W., Brown K. S. Abramczuk J., Hou-Xiank, X., EohAo, K., Bolanher, M., and Brugge! man, L. (19'90) Ann. N. Y. Acad. Sci. 580,81-87 37. Bornstem, P., McKay, J., Llska, D. J., Apone, S., and Devarayalu,S. (1988)
_"_
Mnl Cell Rinl -, A. A R 5 1 - 4"l. R57
38. Sherwood, A. L., and Bornstein P. (1990) Biochern. J. 265,895-897 Alvarez, R. J., Goldfarb, S., 39. Ziyadeh, F. N., Sni es, E. R., %'atanabe,.M., and Haverty, T. (1990) Am. J.Phystol. 2 5 9 , F704-F714 40. Kindy, M. S., Chang, C.-J., and Sonenshein, G. E. (1988) J. Biol. Chem.
8.
"_, __."
2G:R 1 1 A%-1 1 AWI
_I_"
41. Vaage J. and Lindblad W. J. (1990) J. teuk. Biol. 48 274-280 42. Chr& R . J. YanV 'W Ntambi J. M. Geiman D: E. Landsbulz W. H., griedmk, A.%., Nacabeppu, $., Keliy, T. J., i n d Lahe, M.D. (lh89) Genes & Deu 3,1323-1335
