Randy S . Haun-f, Joel Moss, and Martha Vaughan. From the Laboratory of Cellular Metabolism, National Heart, Lung, and Blood Institute, National Institutes of ...
THEJOURNAL OF
Vol. 268, No. 12, Issue of April 25, pp.Printed 8793-8800,1993 in U.S.A.
BIOLOCtCAL CHEMISTRY
Characterization of the Human ADP-ribosylation Factor3 Promoter TRANSCRIPTIONAL REGULATION OF A TATA-LESSPROMOTER* (Received for publication, November 2, 1992, and in revised form, January 26, 1993)
Randy S . Haun-f,Joel Moss, and Martha Vaughan From the Laboratory of Cellular Metabolism, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892
The 5“flanking region of the human ADP-ribosyla- the similarities of their structures, individual ARFs appear to tion factor 3 gene contains the features of a housebehave independently and selectively in their GTP-dependent keepinggene. It lacks a TATAorCAAT box, has association with Golgi in vitro (Tsai et al., 1992). several GC boxes within a highly GC-rich region, and ARFs are highly conserved proteins that have been detected utilizes multiple transcriptioninitiation sites. The cis- in all eukaryotic cells from Giardia to mammals (Price et al., acting elements involved in regulating expression of 1988;Sewell and Kahn, 1988; Tsuchiya et al., 1989,1991; the gene were identified by transient transfections of Murtagh et al., 1992). At present, based on molecular cloning, IMR-32 neuroblastomacells. Reporter plasmids were at least six mammalian ARFs are known (Price et al., 1988; modified tofacilitate construction of defined promoter deletions linked to chloramphenicol acetyltransferase Sewell and Kahn, 1988; Bobak et al., 1989; Monaco et al., 1990; Tsuchiya et al., 1991).These fall into threeclasses based or luciferase using ligation-independent cloning. Transfection analyses indicated that sequences within on deduced amino acid sequence, size, phylogenetic analysis, 68 base pairs of the transcription initiation site were and gene structure (Tsuchlya et al., 1991; Tsai et al., 1991b; necessary forfull expression, in particular a sequenceLee et al., 1992; Serventi et al., 1993). Class I is composed of ARF1, ARF2, and ARFB, with 181 amino acids, which differ containing the 10-base palindrome pair TCTCGCGAGA. Electrophoretic mobility shift assays chiefly near the amino and carboxyl termini. Class I1 consists performed with IMR-32nuclear extracts demon- of ARF4 and ARFB, each with 180 amino acids that differ stratedthataDNA-bindingprotein,termedTLTF, from class I ARFs near the amino terminusand in the boundtoanoligonucleotidecontaining this palin- carboxyl half of the protein. ARF6, the sole member of class drome.Competitionexperimentsshowedthatmuta111, contains 175 amino acids and differs from the other ARFs tions within the core of the palindrome abolished in throughout the coding region with extensive differences in to a 5’- the carboxyl half of the protein. vitro binding and that the same protein bound proximal sequence. Expression of the promoter conThe ARFs from the three classes differ in tissue-specific taining a mutated palindrome was reduced dramati- expression and developmental regulation (Tsuchiya et al., cally, consistent with the conclusion that this region 1989, 1991; Tsai et al., 1991a). In particular, on immunoblots functions in vivo to control expression of the ARF3 of proteins from bovine, rat, frog, and chicken tissues, levels gene. of ARF were higher in brain than in non-neural tissues (Tsai et al., 1991a). In rat brain, the amounts of sARF I and sARF 11, the products of the ARFl and ARF3 genes, respectively (Tsai et al., 1992), were similar on the second postnatal day, ADP-ribosylation factors (ARFs)’ are a family of -20-kDa day and thereafter, sARF I1 predomguanine nucleotide-binding proteins that were originally iden- but by the 10th postnatal tified by their ability to stimulate i n vitro cholera toxin- inated (Tsai et al., 1991a). Similarly, on Northern blots hycatalyzed ADP-ribosylation of GSa,the subunit of the stim- bridized with ARF-specific oligonucleotides, ARF3 mRNA ulatory heterotrimeric GTP-binding protein of the adenylyl- increased from the 2nd to the 27th postnatal day, whereas cyclase system (Kahn andGilman, 1984,1986; Serventi et al., mRNAs for other ARFs either declined or remained constant 1992). More recently, ARFs have been implicated in intracel- (Tsai et al., 1991a). It seems likely, therefore, that functional lular proteintrafficking and are thought to be involved in the specificity is achieved, in part, by regulated expression of assembly of non-clathrin-coated vesicles (Stearns et al., 1990; these genes in different cells or at different times in cellular Donaldson et al., 1991; Serafini et al., 1991). In contrast to development. This study was undertaken to define elements responsible for regulating expression of the ARF3 gene dueto * The costs of publication of this article were defrayed in part by its predominance in neuronal tissues and increasing expresthe payment of page charges. This article must therefore be hereby sion during brain development. (Y
marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(‘) reported in thispaper has been submitted to the GenBankm/EMBL Data Bank withaccessionnumber(s) M74492. $ T o whom correspondence should be addressed Bldg.10, Rm. 5N307, NIH, Bethesda, MD 20892. Tel.: 301-496-5193;Fax: 301-4021610. The abbreviations used are: ARFs, ADP-ribosylation factors; CAT, chloramphenicol acetyltransferase; GRE, glucocorticoid-responsive element; LIC, ligation-independent cloning; PCR, polymerase chain reaction; TLTF, TATA-less transcription factor; bp, base pair(s); RSV-LTR, Rous sarcoma virus long terminal repeat.
MATERIALS ANDMETHODS
Oligonucleotide Synthesis-Oligonucleotides were synthesized on an Applied Biosystems (Foster City, CA)380BDNA synthesizer. After ammonium hydroxide deprotection, oligonucleotideswere evaporated to dryness in a Savant Speed-Vac (Farmingdale, NJ), suspended in water, and desalted over gel filtration columns (NAP-5, Pharmacia LKB Biotechnology Inc.). DNA Sequence Analysis-DNA sequences of alkaline-denatured double-stranded plasmid DNA (Chen and Seeburg, 1985) were determined by the chain termination method (Sanger et al., 1977) using Sequenase kits (U.S. Biochemical Corp.) and sequence-specific oli-
8793
8794
ADP-ribosylation Factor 3 Promoter
gonucleotide primers. Primary structure analyses were carried out using PC/Gene sequence analysis software (IntelliGenetics, Mountain View, CA). Data base searches using BLAST (Altschul et al., 1990) were performed at the NationalCenter for Biotechnology Information using the BLAST network service. Searches for transcription factor consensus sites were performed using MacVector (IBI, New Haven, CT) and the data base supplied with version 3.5. Polymerase Chain Reaction Amplifications-Polymerase chain reactions (PCR) were carried out in atotal volume of 100pl containing 100 ng of each primer, 25 ng of pG51 DNA (Tsai et al., 1991b), 200 p M of each dNTP, Taq DNA polymerase buffer (Promega, Madison, WI), and 10% glycerol. Samples were heated at 95 “C for 5 min to denature the template DNA, and Taq DNA polymerase (2.5 units, Promega) was added; samples were then overlaid with 75 plof mineral oil (U. S. Biochemical Corp.) and subjected to 30 cycles of 94 “C for 1 min, 52 “C for 1 min, and 72 “C for 2 min using a DNA thermal cycler (Perkin-Elmer Cetus).After the final cycle, incubation a t 72 “C was continued for 7 min to extend incomplete products. PCR products were precipitated by the addition of 0.1 volume of 3 M sodium acetate, pH 5.5, and 2 volumes of ethanol. The PCR products were purified after electrophoresis through a 1%agarose gel containing ethidium bromide (0.2 pg/ml) using QIAEX (Qiagen, Studio City, CA) as described by the manufacturer. Pkwmid Construction-To simplify cloning of the 5”flanking region of the ARF3 gene into a chloramphenicol acetyltransferase (CAT) reporter plasmid, the multiple cloning region of the plasmid pCAT-Basic (Promega) was replaced with synthetic oligonucleotides containing additional restriction endonuclease sites. These complementary oligonucleotideswere synthesized to produce single-stranded ends compatible with HindIII- and XbaI-digested vector. The pCATBasic plasmid was digested with HindIII and XbaI, and ligated with the annealed oligonucleotides resulting in plasmid pCAT (H3IXba). The humanARF3 genomic clone pG51(Tsai et al., 1991b)was digested with BamHI and NotI; the ARF3 gene fragment containing part of exon 1and the5’-flanking region was isolated using QIAEX following electrophoresis in a 1%agarose gel and ligated into pCAT (H3/Xba) digested with BglII and NotI. The resulting plasmid, pARF3Not/ CAT, contained 191 bp of exon 1 and -2.5 kilobases of 5”flanking sequences cloned 5’ to theCAT gene.Initial deletions of the upstream sequences were produced by digestion of the plasmid pARF3Not/ CAT with either SphI or PstI, removal of the upstream sequences by gel electrophoresis, and then religation of the plasmid. The products of this procedure contained 330 and 72 bp of 5”flanking sequences, respectively.The fragment containing 716 bp of 5’-flanking sequences and exon 1, obtained by XbaI digestion of pARFBNot/CAT, was ligated to XbaI-digested pCAT/Basic yielding the promoter construct in the forward and reverse orientations. To facilitate the construction of further deletion mutations, the pCAT/Basic and pGL/Basic (for luciferase expression) vectors (Promega) were modified for ligation-independent cloning (LIC). Oligonucleotides containing LIC sequences were synthesized to produce HindIII and XbaI (CAT vector)-compatible or KpnI and HindIII (luciferase vector)-compatible ends when annealed. The doublestranded oligonucleotideswere ligated into thepCAT-Basic and pGLBasic plasmids that had been digested with the appropriate restriction enzymes generating pCAT/LIC and pLUC/LIC, respectively. Defined ARF3 promoter deletions were then prepared by ligation-independent cloning using the primers indicated in Fig. 2. The forward primers contained 1 2 nucleotides at their 5’ ends (5‘-CCTGCTCGTCTG-3’) that were complementary to the single-stranded tails 5’ to Apaldigested LIC reporter plasmids treated with T4 DNA polymerase in the presence of dATP. Similarly, the reverse primer contained 12 nucleotides at its 5’ end (5’-GGTGGTGCTCTG-3’) that were complementary to the single-stranded tails 3’ to the ApaI site. Defined deletions were generated by synthesizing forward primers targeted to specific regions of the promoter and amplifying the region from a genomic subclone containing exon 1 and the5’-flanking region. Mutated ARF3/CAT constructs were constructed by amplifying the ARF3 promoter with mutated forward primers containing additional LIC sequences described above at their 5’ ends and the same reverse primer used above. The mutated promoters were cloned into the pCAT/LIC vector as described. A positive control vector expressing the luciferase gene under the control of the Rous sarcoma virus long terminal repeat (RSV-LTR) was constructed similarly by amplifying the RSV-LTR from the plasmid DRSV/L (de Wet et al.. 1987) using the primers 5’CCTGCTCGTCTGAAGCTGCTCCCTGCTT-3” and 5”GGTGGTGCTCTGAACCCAGGTGCACAC-3’. (Underlined nucleotides
correspond to RSV-LTR sequences; LIC sequences are not underlined.) The plasmid pRSV-LUC/LIC was then prepared by cloning the amplified viral promoter into thepLUC/LIC vector as described. Ligation-independent Cloning Reactions-PCR-amplified DNA was cloned into the pCAT/LIC or pLUC/LIC vectors essentially as described (Haun et al., 1992). Briefly, 100-200 ng of amplified DNA or ApaI-linearized plasmid vector were incubated at 37 “C for 20 min in a total volume of 20 p1 containing 33 mM Tris acetate, pH 8.0, 66 mM potassium acetate, 10 mM magnesium acetate, 0.5 mM dithiothreitol, 10 mMMgClZ, T4 DNA polymerase (3 units, New England Biolabs or 1 unit, Boehringer Mannheim), and 2.5 mM dATP (linearized vector) or 2.5 mM TTP (amplified DNA). The reactions were terminated by heating a t 75 “C for 10 min. The T4DNA polymerasetreated vector and PCR fragment were mixed and precipitated with 0.1 volume of 3 M sodium acetate (pH 5.5) and 2 volumes of ethanol. The precipitated DNA was suspended in 15 pl of 10 mM Tris-HC1, pH 8.0,l mM EDTA, 100 mM NaCl preheated to 65 “C and annealed at room temperature for 1-2h. A sample (5 pl) was then used to transform competent DH5a bacteria. Plasmid DNAwas prepared from randomly selected ampicillin-resistant colonies, and the construct was verified by double-stranded DNA sequencing. Cell Transfections and Enzyme Assays-IMR-32 human neuroblastoma cells were obtained from the American Type Culture Collection (Rockville, MD). Cells were maintained in Eagle’s minimum essential medium (Biofluids, Rockville, MD) supplemented with 10% fetal bovine serum, 2 mM glutamine, and non-essential amino acids in a 5% COz/air environment. For DNA transfections, cells were seeded at 7.5 X 10’ cells/lO-cm plate. After 2 days, 15 pg of cesium chloride-purified ARF3/CAT and 5 pg of pRSV-LUC/LIC plasmid DNA were co-precipitated with calcium phosphate by overnight incubation in 3% COz (Chen and Okayama, 1987). The following day the plates were rinsed with Eagle’s minimum essential medium, growth medium containing penicillin and streptomycin (10 units and 10 pg, respectively, per ml) was added, and the plates were returned to 5% COz. After 2 days, cells were harvested by repeated pipetting using cold Dulbecco’s phosphate-buffered saline (BioWhittaker, Walkersville, MD), washed, and suspended in 100 mM potassium phosphate, pH 7.8, 1 mM dithiothreitol. Lysates were prepared by disrupting the cells with three cycles of freeze-thawing ( d r y ice, 5 min/37 “C, 3 min) and thenremoving cell debris by centrifugation in a microcentrifuge (16,000 X g, 5 min, 4 “C). Luciferase activity was measured using a Lumat LB 9501 luminometer (EG&G Berthold, Nashua, NH) and D-lUCiferin as a substrate (potassium salt, Analytical Luminescence Laboratory, San Diego, CA) as described (de Wet et al., 1987). CAT activity was determined by the phase extraction procedure of Seed and Sheen (1988) using [3H]chloramphenicol (Du Pont-New England Nuclear) after endogenous deacetylating activity was destroyed bv heating the lysates for 10 min at 65 “C (Crabb and Dixon, 1987). Nuclear Extracts and Mobilitv Shift DNA Bindiw Assay-Nuclear extracts were prepared from IMR-32 cells accordingto the procedure of Dignam et al. (1983). Protein concentration was determined using a Bio-Rad protein assay kit with bovine serum albumin as a standard. Complementary oligonucleotides were5’ end-labeled with [y-”P] ATP using T4 polynucleotide kinase, annealed, purified by polyacrylamide gel electrophoresis, and isolated by electroelution onto NA45 membrane (Schleicher & Schuell). DNA binding assays were performed as described by Singh et al. (1986) using 30,000 cpm (-0.25 pmol) of radiolabeled oligonucleotide and 8-12 pg of IMR-32 nuclear extract. After a 30-min incubation at room temperature, the resulting complexes were subjected to electrophoresis in 4% polyacrylamide gels using a low ionic strength Tris acetate buffer (Strauss and Varshavsky, 1984). Competition experiments were performed using 1 pmol of the indicated unlabeled ARF3 complementary oligonucleotides or 1.75 pmol of double-stranded transcription factor consensus oligonucleotides (AP1, AP2, TFIID, Spl,Octl,CTF/NFl, GRE, CREB, and NF-KB (Promega)). Extractsof bacterially expressed AP2 were obtained from Promega and used according to the supplier’s instructions. I
~~~
RESULTS
Construction of ARF3ICAT Plasmids-To identify the sequencesresponsiblefor controllingthe expression of the human ARF3 gene, a BamHIINotI fragment of the genomic subclone pG51 (Tsai et al., 1991b), which contains 191 bp of exon 1and -2.5 kilobases of 5”flanking sequence, was ligated
ADP-ribosylation Factor 3 Promoter
+
into aBglII NotI-digested modified expression vector, pCAT (H3/Xba), encoding the bacterial CAT gene (Fig. 1A). Initial deletions were made from the resulting plasmid, pARF3Not/ CAT, by using restriction sites XbaI, SphI, and PstI located within the 5"flanking region. A search of the promoter region for consensus sequences of binding sites for known transcription factors identified severalpotentialsites within the ARFB sequence (datanot shown). Deletion mutations were targeted systematically to eliminate potential binding sites to determine whether the sites were utilized in regulating ARF3 gene expression, using a LIC expression vector (pCAT/LIC, Fig. 1B). SequenceAnalysis of 5"Flanking Region-The sequence within approximately 300 bp of the major transcription start site is GC-rich (66.9%) with a high concentration (79.6%) between -151 and -59 (Fig. 2). This region also contains three consensus sequences for binding the transcription factor Spl. In addition, the proximal 5"flanking sequence lacks consensus TATA or CAAT sequences. A search of the GenBank data base identified two partial Alu repeats. One, located between -945 and -791, is homologous (87% identity) to an Alu repeat within the first intron of the human hypoxanthine phosphoribosyltransferase gene (Edwards et aE. (1990); GenBank accession no. M26434) and corresponds to part of a second monomer unit of a consensus Alu repeat (Kariya et al., 1987). The second repeat (-1298 to -1136) is homologous tothe repeat within intron L of the human prothrombin gene (83% identity) reported byDegen and
A
8795
Davie (1987, GenBank accession no. M17262). This repeat is complementary to partof the first monomer unit of a consensus Alu repeatandcontainsacentral A-rich region (ArnTACA,, where m and n are usually between 4 and 8) located between -1258 and -1244 (Kariya et al., 1987). Identification of cis-Acting Elements-To determine which sequences of the ARF3 5'-flanking region could direct expression of the reporter gene, plasmids were transfected into the human neuroblastoma cell line IMR-32. Initial transfections using the plasmids constructed using unique restriction sites (ABam, AXba(+), ASph, and APst) indicated that each was capable of directing high levels of expression of the CAT gene (Fig. 3, lanes I , 3,8, and 1 4 ) , whereas a construct containing the 5"flanking region in the opposite orientation, AXba(-), only elicited CAT activity at the level of transfecting vector alone (Fig. 3, lanes 20-22). Potential cis-acting elements identified by sequence analysis were characterized using a series of targeted deletions (Figs. 2 and 3). These constructs indicated that deletion of sequences to within 100 bp of the transcription start site had little effect on promoter activity (Fig. 3, lunes 1-13). Three potential S p l binding sites were identified in the search of the promoter(underlinedin Fig. 2 ) . To examine whether these sites affect the transcription of the ARF3 promoter in the transient assays, LIC primers E, D, and N were used to construct deletion mutants starting at position -168 (containing all three S p l sites), -119 (containing only the proximal Spl site), and -96 (all Spl sites removed), respectively
-
Hmdlll
Sac1
Soil
Sph I
PsrI
BglII
Nor1
Kpn I
" " " "
XhaI
-x -
_" $$% 23: -"
a-=
Eco RI
1 -
AGCTTGAGCTCGTCGACGCATGCTGCAGAGATCTGCGGCCGCGGTACCT ACTCGAGCAGCTGCGTACGACGTCTCTAGACGCCGGCGCCATGGAGATC
-=
gxzr: 4*:2s
I I \ IY
1
pCAT (H3iXba)
BgI I1 +Nor I digested
Barn HI +Nor I digested
i
S O 0 bp
pARF3Not/CAT
pCAT/LIC
-
PLUCLIC
Apu I digested T4 DNA polymerase + dATP
-AGCTTAA "ATTGGACGAGCAGAC
-
mlCAA
xho I
Hot dl11
M
G
Z
CAT
"ATGGTTGGACGAGCAGAC
I
Apa I d~gesled T4 DNA polymerase + dATP
MTTa H m dl11
CAGAGCACCACCTTA
~uclferase
FIG. 1. Construction of CAT and luciferase reporter plasmids. A , sequence of modified multiple cloning region of CAT reporter plasmid, pCAT (H3/Xba), is in upper left. Schematic diagram of human ARF3 genomic subclone pG51 and partial restriction enzyme map is in upper right. Construction of ARFB reporter plasmid pARWNot/CAT is shown for the ligation of the BurnHI + Not1 fragment of ARF3 gene into BgZII + NotI-digested modified CAT plasmid. Solid lines represent 5"flanking region and intron 1 of the ARF3 gene, solid box
corresponds to exon 1, and thin lines indicate vector sequences. B, top, sequences of ligation-independent cloning region of modified CAT (pCAT/LIC) and luciferase (pLUC/LIC) vectors; Bottom, modified reporter plasmids digested with ApuI and treated with T4 DNA polymerase in the presence of dATP to generate noncomplementary single-stranded tails.
8796 -1318
ADP-ribosylation Factor 3 Promoter AAGCATCTCTGGCCTCAGTTCTTATAGGTA~CXCATCCCCG
between -96 and -72 (Fig. 3, lanes 13 and 14) and returned to the basal level after the sequence between -72 and -58 -1190 ~ ~ ~ C ~ ~ A ~ C A G G C C T G G C C C C A was deleted (Fig. 3, lanes 14 and 15). Althougha slight -1120 G C T T A G T T C T T C T T G G A A G C . U G G ~ C A A A C ~ C ~ T C A A C C A A G G C A A T ~ T T C A T . U G A G C A A elevation in CAT activity was observed for the -72 mutant - 1 0 5 0 CAATTTGGGCTCAAGCCTTAGGGAAGAGCCATGTGTTCGTrCCCTTArACTGCTGT~TGAGGAGAGAG (APst), a deletion of this region constructed by the ligation-980 G G C A C T T A G A G A A C C T T C T A T T G G C C G G G C G C G G T C C T C / independentcloning procedure, APst/LIC, produced CAT activity similar to thatobserved with the -96 and -58 dele-910 T T T C G C C i G A G G l X T C l X V n C W K S G . L APst/LIC, 0.93; -58, - 0 4 0 ~ 1 W W L C ~ G G . U C C T T C T G T T C T T A G tions (relative CAT activity: -96, 0.96; xhu I -170 CATTACTAGTGCCCTTC~RTC~CAGC~A~~TTTGTTTCAAGGCAGTTCTGCATGT~ACCAGGTGCAA1.07). A significant drop in activity, however, was observed K with deletion of the region between -58 and -17 (Fig. 3, lanes -700 AACGGAGTGTGAIGGGAAGTAGGAG~GGGTAGG~GCAG~~ATGGGTGC~GTCCTTCACCTTTCCT 15 and 17). Inspection of this sequencerevealeda 10-bp -630 TGGGATACTAAAATCCCTACTCATC~AGGCACACTTCCTC~CCA~ACTTCTG~TAT~AGG~TTCCAC sequence with dyad symmetry (-47, 5’-TCTCGCGAGA-3’, A - 5 6 0 TCTCCGTCTTTCAAGTTCCACTT~CTGGCTCAGAG~C~AGCTGT~~TCCT~TAGCCTCTTAAGTTCAT -38). As some previously characterized transcription factors -490 C ~ C T C C T T C C T C ~ T C C T ~ A ~ G G G A A A G G A ~ C ~ ~ ~ T G A ~ ~ C A G ~ C C C C A C A G ~ ~ ~ ~ T A G ~ C ~ ~ A G G ~ bind tosequences with dyadsymmetry (e.g. CAMP-responsive . H - 4 2 0 TAGTGAGCAACAACAATGGTGCGG~GAGGGC~TTCAGGGGATGGCAGCTAGCCCAGA~TAAG~T~G element, TGACGTCA (Montminy et al., 1986)), itwas temptp hI -350 A T G A A G T C T G G C r G G G T A A G G ~ C C A C A G C A A C G A C T G T T C A C C C A ~ A T C C C ~ C T T T G C C C ing to speculate that this palindrome was involved in binding . F -280 IIAGGAGCCCCTGGC.CCCCTCAGCCCAGCCCAGCCCG~CTCCTACGCAGCCAACTTGCCGCCTCAGTCCTT~TT a trans-acting factor regulatingARF3 expression. The role of E -210 CCAGGCACAACCCTGGAGCT~GGGA~TCTCACCTCAG~CCTCACTG~C~C~~CCCCATTC~CGGCT~ this sequence, therefore, was examined by constructing a ‘ N -140 C C T C C G C C G s C C A C R G C C C C ~ C T C ~ A G C T C C A ~ C G C C C G G C T G C C G G T G ~ fusion gene with the 5’ half of the palindromedeleted (Fig. 2, PS, I P). Consistent with this hypothesis, CAT activity of -10 ~ G C C G C T G C ~ A T G G T G ~ T G G G T C T ~ ~ G A G A A ~ T G ~ C G C T A G C T A C ~ G C G ~ A G ~ ~ T C G C G C Gprimer ATCGG smu I wasdecreased to the 1 ~ A G G T G G G A A G T A T C ~ G C T G G G G G ? G ~ G G G * A ~ G C ~ G C A ~ ~ G C ~ ~ A G r ~ C C G ~ G C G G C Tcells T C C G transfected with this construct . A extent as observed for the construct with the entire 71 G T G C T A G G T G G A G G G A A G G A A G G A G G ~ A G C C G G G G A G ~ G G G C ~ A G G G C A ~ ~ A ~ C ~ G G G ~ ~ G ~ G ~ Csame GGC~ Not 1 sequence deleted (Fig. 3, lanes 16 and 17). A further decline 141 TGGGGRGCCAGAGGGAGCGG~GCGGAACCTGCGGGGCAGAGGCGGCGGCCGCAG~GGCGCAGCT R . in CAT activity, to the level of transfecting vector alone, was FIG. 2. Nucleotide sequence of proximal 5‘-flanking region elicited with deletion constructs extending beyond the cap and exon 1 of ARF3 gene. Sequence(GenBankaccessionno. site (Fig. 3, lanes 18 and 19). M74492) is numbered with the major transcription initiation site at Detection of DNAIProtein Interactions Using Electropho+1 (indicated by bent arrow). Primers used togeneratedeletion retic Mobility Shift Assay-Deletion analysis of the 5“flankconstructs are indicatedby arrows above and below the sequence; tails to 5’ends of ing region identified regions within the sequence from -96 to represent ligation-independent cloning sequences added primers as described under “Materials and Methods.”Partial repeti- -17 that could serve as potential sitesfor regulatory regions tive Alu sequences are in boldface italicized typeface. Unique restric- involved in DNA/protein interactions.To examine the ability tion endonuclease sitesare indicated above the sequence. of these sequences to interact with trans-acting factors,complementary oligonucleotides that encompass thisregion were (Fig. 2). CAT activity determined after transfection of IMR- synthesized and used in mobility shift assays with nuclear extracts prepared from IMR-32 cells (Fig. 4A).After electro32 cells indicatedthat removal of theSplsitesdidnot significantlyaffect promoter activity (Fig. 3, lanes 12-13). phoresis of the products of binding reactions in low ionic Transcription was enhanced by the removal of the sequence strength polyacrylamide gels, two complexes were separated - 1 260 O C m m T T G T A T T A T l G S 2 - W C X C T i Z X
.
J
.
-
FIG. 3. Transient expression of ARF31CAT deletion constructs in IMR-32 neuroblastoma cells. Left, diagram of 5’-flankingsequences and exon 1 of ARFS/CAT deletion mutants. Open boxes represent ARF3 5“flanking sequences, hatched box represents repetitive Alu sequences, solid boxes represent 5’ portion of exon 1, and thick horizontal lines represent CAT vector sequences. Locations of consensussequences for binding sites forSpl transcription factor are indicated by vertical lines. Names of
the constructs usedin text and corresponding primers shown in Fig. 2 are listed to the left, deletion end points relative to the major transcription initiation site (+I) are to the right, and scale (in base pairs) is depicted below the diagram. Right, 15 ~g of the indicated ARFS/CAT construct were co-transfected with 5 pg of pRSV-LUC/LIC plasmid into IMR-32 cells. CAT and luciferase enzyme activities weremeasured in duplicate after transient expression. AverageCAT enzyme activity was divided by the average luciferase activity of the same cell extracts and is presented relative to that of the ABam construct (openboxes).Thin lines represent the S.E.; n = 4-7.
Relative CAT activity
I
2500
I
w
-7x6
2
~716
3
b
“X
4 5
4x9
H
-49x
6
i”
-412
H
~295 -235 -16K
I-
-119
-96
-72
-17
+5 I +I14
71hR
-
9
t”
IO II I?
13 14
IS
-5X -4 I
7
x
~330
E Pb
16 17 In IY
20 21
22
ADP-ribosylation Factor 3 Promoter
A
8797
-
-29 -98 I I CCAGCTCCAGCGCCCGGCTGCCGGTGCTGCAGCCGCTGCCATGGTGATGGGTCTCGCGAG~CTGCCGCT 2 1 3 3mut I AA 1 mut 1 CT 3mut2 AG
B FIG. 4. Electrophoretic mobility shift assays. A , DNA sequence of ARF3 promoter region between -98 and -29. Arrows indicate sequence with dyad symmetry; lines represent oligonucleotides used for mobility shift assays; letters indicate mutations introduced into oligonucleotides used as competitors. B, mobility shift assay performed with labeled complementary oligonucleotides 1-3 (-0.25 pmol) using varying amounts of IMR-32 nuclear extracts. Arrow indicates position of ARF3-protein complex. C, competitor oligonucleotides (2 pmol) were incubated in DNA binding assay prior to addition of labeled oligonucleotide. Competitor DNAs are indicated above appropriate lanes; hyphens indicate control lanes without addition of competitor. Arrows indicate position of ARF3-protein complex using labeled oligo 1 (upper panel) or oligo 3 (lower panel).
C
competitor
using either labeled oligo 1 (Fig. 4B,lanes 2 and 3) or oligo 3 (Fig. 4B,lanes 8 and 9),whereas only a single strong band was detected using oligo 2 (Fig. 4B,lanes 5 and 6 ) . Competition experiments indicatedthat thelower band, detected with all three labeled oligonucleotides, was due to nonspecific binding. The upper band, however, migrated with the same mobility regardless of whether oligo 1 or 3 served as the labeled oligonucleotide (Fig. 4B, arrow). Furthermore, formation of the specific complex with labeled oligo 1 was efficiently inhibited by excess unlabeled oligo 1 or oligo 3, but not oligo 2 (Fig. 4C,upper panel, lanes 2-4). Similarly, oligo 1and oligo 3, but not oligo 2, could compete for the complex formed with labeled oligo 3 (Fig. 4C,lower panel, lanes 2 4 ) . As oligo 3, which contains the 10-bp palindrome, could form a specific complex, complementary oligonucleotides were synthesized containing point mutations to identify residues involved in the DNA/protein interaction. When these oligonucleotides were used in competition experiments, only oligo-
nucleotide 3mut3, with themutationatthe core of the palindrome (GC + AA), failed to compete, indicating that these residues were critical to the formation of this complex (Fig. 4C, lower panel, lanes 5-7). Again, the mutated oligos produced the same pattern of competition with labeled oligo 1 and oligo 3 (compare Fig. 4C,upper and lower panels, lanes 5-7). These results are consistent with the notion that the same factorthat binds to thepalindrome within oligo 3 binds to a region within oligo 1. To define residues responsible for binding to oligo 1, a mutation similar to that in the 3mut3 oligo wasintroduced into thissequence, which was designated oligo lmutl. It failed to compete for binding to labeled oligo 1,indicating that these nucleotides are necessary for the factor to bind this region (Fig. 4C,upper panel,lane 9). In a search of consensus sequences for binding sites of known transcription factors, no sequences in oligo 1 or oligo 3 were found to contain sites for previously characterized factors. The factor binding to this region, therefore, appears
8798
ADP-ribosylation Factor 3 Promoter
by these criteria to represent a novel DNA-binding protein. sequences was able to compete. The ability of labeled oligo 1 To test the specificity of these sequencesto bind the factor inor oligo 3 to be competed in a similar manner, however, is IMR-32 nuclear extracts,oligonucleotides containing the spe- consistent with the results obtainedin the previous competicific binding sites for a number of known factors were used tion experiments (Fig. 5A, compare upper and lower panels). Although oligo 1 and oligo 3 do not containobvious consenin competition experiments (Fig. 5A and Table I). Unexpectexperedly,oligonucleotides containing an AP2-binding site or a sus sequences for either AP2 or GRE, the competition IMR-32 extracts that glucocorticoid-responsive element (GRE) were able to com- iments suggested that the factor in Pete for binding (Fig. 5A, lanes 3 and 8); none of the other binds to these ARF3 sequences might indeed have been described previously. Using an extract of bacterially expressed AP2, a specific complex could only be formed with the labeled AP2 oligo (Fig. 5B, lanes 3-5, open triangle). Furthermore, when the DNA-proteincomplexes were resolved on the same gel, theAP2-protein complexformed using a 26-bp oligo migrated witha lower mobility than thecomplex formed using the IMR-32 extractandthe30-bpARF oligos (Fig. 5B; compare position of ARF3-protein complex, arrow in lanes 1 and 2, to AP2-protein complex, open triangle in lane 5). Transfection of Mutated ARF3ICATConstructs-Although the mobility shift assays indicated the ability of the ARF3 promoter sequences between -98 and -69 (oligo 1) and -58 and -29 (oligo 3) to form a specific complex with a factor in the IMR-32 nuclear extracts, these in vitro assays did not address the functional consequence of these DNA/protein interactions. Two ARF3/CAT expression constructs, therefore, were prepared that contained the mutations in oligos 1 2 3 4 5 6 7 x 9 1 0 lmutl and 3mut3 that failed to compete in thevitro in binding assays. These constructs, NMut and "ut, respectively, were transfected into IMR-32cells in parallelwith the neighboring unmutated ARF3/CAT constructs (Fig. 6). The unmutated constructs produced the same pattern of transcriptional activity as observed in previous experiments (compare Fig. 6 to Fig. 3). The NMut construct, containing the mutation corresponding tooligo 1, had onlymarginally less activity than the unmutated construct (Fig. 6, compare N and NMut). The mutationin oligo 3, however, caused a dramaticdropin These resultssuggest activity (Fig. 6, compareM and "ut). that the DNA-protein complex betweenthe factor in the IMR32 extracts and the sequence containing the 10-bp palindrome within oligo 3 is critical for the expression of the ARF3gene.
B
DISCUSSION
FIG.5. Electrophoretic mobility shift assays with oligonucleotides containing binding sites for known transcription The ARF3gene 5"flanking region lacks a TATA orCAAT factors. A , competitor oligonucleotides (1.75 pmol) were incubated consensus sequence immediately 5' to the initiation site, the with IMR-32 nuclear extracts prior to addition of labeled ARF3 oligonucleotide (-0.25 pmol). Hyphens indicate control lanes without addition of competitor; arrows indicate position of ARF3-protein complex, as in Fig. 4, using labeled oligo 1 (upper panel) or oligo 3 (lowerpanel). B, labeled ARF3 oligonucleotides were tested for their ability to bind bacterially expressed AP2. Arrow indicates position of ARF3-protein complex identified incontrol reactionsperformed using IMR-32nuclear extracts (lanes 1 and 2 ) . A specific ARF3-AP2 complex with similar mobility was not observed using labeled ARF3 oligonucleotides andanAP2extract (lanes 3 and 4 ) , whereasa complex with a different mobility was formed (open triangle) with a labeled AP2 oligonucleotide in the presence of an AP2 extract (lane 5).
TABLE I OliPonucleotides usedin mobilitv shift comDetition studies Factor
AP1 AP2 TFIID SP 1 Oct 1 CTF/NFl GRE CREB NF-KB
Sequence (5' + 3')
CGCTTGATGAGTCAGCCGGAA
GATCGAACTGACCGCCCGCGGCCCGT GCAGAGCATATAAGGTGAGGTAGGA ATTCGATCGGGGCGGGGCGAGC TGTCGAATGCAAATCACTAGAA CCTTTGGCATGCTGCCAATATG TCGACTGTACAGGATGTTCTAGCTACT AGAGATTGCCTGACGTCAGAGAGCTAG AGTTGAGGGGACTTTCCCAGGC
I
I1
\
\\I,,,
\I,*,
\I
\l\l,,,
I'
1
H
ARF K A T conslruch
FIG. 6. Transient expression of ARF3 promoter constructs containing mutations incapable of competing for in vitro binding. Mutations identified by mobility shift assays that were not able to compete for in vitro binding were used to generate mutated ARF3 promoter/CAT constructs. Mutation NMut represents oligonucleotide lmutl, and MMutcorresponds to oligonucleotide 3mut3 used in the competition experiments. CAT and luciferase enzyme activities were determined in duplicate after transient expression in IMR-32 cells. Average CAT activities were divided by average luciferase activities, and values are represented relative to the APst construct (solid boxes). Thin lines represent S.E.; n = 2-4.
ADP-ribosylation Factor 3 Promoter proximal sequence is GC-rich and contains several GC boxes, and ARF3 mRNA initiates at multiple sites(Tsai et al., 1991b). Thesecharacteristicsare common among “housekeeping” genes, whichare expressed in all cells and notsubject to environmentalcontrol(Dynan, 1986). Thesestructural features of the ARF3 promoter are consistent with recent evidence that ARFs are involved in intracellular protein trafficking (Serafini et al., 1991) and that the ARF3 gene is expressed in all tissues (Tsuchiya et al., 1991). There are, however, differences in expression during development (Tsai et al., 1991a), in the levels of expression among the ARFs, and among different tissues for any given ARF (Tsuchiya et al., 1989, 1991; Tsai et al., 1991a), with predominant expression of ARF3 in neural tissues (Tsaiet al., 1991a). To identify &-acting elements responsible for the expression of the human ARF3 gene, fusion genes were constructed with variable amounts of the 5”flanking region of the ARF3 gene linked to thebacterial CAT reporter gene. The chimeric genes were transfected into human IMR-32 neuroblastoma cells, which express the endogenous ARF3 gene (Tsai et al., 1991b). In generating deletion mutants, unique internal restriction sites were utilized to produce gross deletions within the 5’-flanking sequences. To examine the promoter in greater detail, the common procedure of producing random deletions over a large segment of DNA by relatively uncontrolled methods (e.g. unidirectional exonuclease I11 digestion (Henikoff, 1984)) is clearly not effective in producing a series of defined mutants that differ by small increments. As it was desirable to target deletions to specific sites andexpression vectors had been modified successfully for a similar purpose (Haun and Moss, 1992),two reporter plasmids (CAT and luciferase) were altered to allow ligation-independent cloning of targeted DNA fragments into theirpromoter regions (Fig. 1B). In this study, the CAT vector was used for the deletion constructs and an RSV-driven luciferase vector, generated with the modified pLUC/LIC plasmid, as the control. Sincy the same cloning strategy is used with both vectors, the same DNA fragment can be cloned into either of the plasmids, thereby allowing the luciferase vector to be used for analysis of a weak promoter. Transfection of the ARF3 deletion mutants revealed that with removal of the Alu sequences in thedistal portion of the 5’-flanking region there was only a modest increase in transcriptional activity suggesting that these repetitive elements do not adversely affect ARF3 transcription. In addition, removal of consensus sequences for several potential transcription factors by targeted deletions dispersed along the promoter region indicated that these sequences were not required for expression of the gene in transient assays. Among elements removed in the deletion analysis was a cluster of potentialSpl-bindingsites within the GC-rich region of the 5’-flanking sequence. In some genes with TATAless promoters containing multiple GC boxes, binding of S p l has been shown to be critical for transcription initiation (Pugh and Tjian, 1990,1991). Although these promotersdo not bind directly theTFIID complex, atranscriptional component containing the TATA box-binding factor thought tobe a key link between promoter-specific activators and the RNA polymerase I1 initiation machinery, it is proposed that promoterbound Spl acts through protein-protein interactions via a protein tether to anchor TATA box-binding factor to the TATA-less template(Pugh and Tjian, 1990, 1991). Since removal of the GC boxes did not affect the transcriptional activity of the fusion gene and binding of Spl to thisregion was not detected in DNase I footprinting experiments (data not shown), it could be reasoned that these elements are not
8799
necessary due to the utilization of a cryptic TATA box. Inspection of the sequences surrounding the initiation site, however, does not support this explanation suggesting that an Spl-independent mechanism activates ARF3 expression in the absence of a TATA box. Recent data have also identified a factor(s) that binds at, or near, the initiation element of some promoters and may interact with the TFIIDcomplex in a mannersimilar to the Spl tether toactivate transcription (Nakatani et al., 1990; Set0 et al., 1991; Roy et al., 1991). Comparison of the degenerate consensus sequence, YAYTCYYY, where Y represents a pyrimidine nucleotide proposed for binding the initiator factor (Roy et al., 1991), with the sequence identified in this study revealed no significant homology. Although the 5‘ half of the palindrome is pyrimidine-rich, it seems unlikely that the same factor is involved in activating transcription due to the poor overall homology between these elements. A region between -58 and -17 was identified that was critical for full expression of the promoter. This region contained a 10-bppalindromic sequence, TCTCGCGAGA, which was a potential candidatefor binding a trans-acting factor. A deletion construct, P, that contained only the 3’ portion of the palindrome exhibited the same activity as the construct with the entireelement removed, consistent with the proposal that this sequence motif is critical for expression of the gene. To determine whether proteins in nuclear extracts prepared from IMR-32 cells could specifically interact with the critical sequences identified by the transfection analyses, electrophoretic mobility shift assays were conducted using complementary oligonucleotides targeted to the-98 to -29 region. These experiments revealed that a nuclear factor, termed TLTF (TATA-less transcription factor), could specifically bind to an oligonucleotide containing the palindromic sequence, that this DNA/protein interaction could be interrupted by competition with an oligonucleotide containing a mutation at the core of the palindrome, and that AP2- and GRE-containing oligonucleotides could compete for binding. Furthermore, an upstream element, represented by oligo 1, was able to bind a factor in the nuclear extracts. Competition experiments are consistent with the notion that thesame factor binds to both promoter elements, but the possibility of having two factors with similar binding sites cannot be ruled out. As both DNA regions do not contain the palindromic sequence, the entire element may not be necessary for binding the factor. Alignment of the two elements, however, 5’-CCAGCTCCAsGCCCGGCTGCCGGTGCTGC- 3’
I II
Ill
IIIIII I
5”ATGGTGATGGGTCTCgGAGAA-CTGCCGCT-3’
(oligo 1) (oligo 3)
does reveal identities between the potential binding sites, including the GC nucleotides (underlined) that when mutated no longer compete for complex formation. Consistent with the expression of ARF3 mRNA in all tissues, a similar ARF3protein complex was observed using HeLa cell nuclear extracts (data notshown). Whether there is a direct correlation between ARF3 expression in different tissues and the levels of TLTF awaits further characterizationof this factor. Mobility shift assays performed with bacterially expressed AP2 indicated that theARF3 gene sequences do not interact with AP2 in uitro. As ARFs have been detected in all cell types, whereas induction by steroid hormones is restricted to cells possessing the appropriate hormone receptor, it seems unlikely that the ARF3 gene, encoding a protein involved in intracellular protein trafficking, would be controlled by steroid hormones, but this possibility has not been excluded. Thefactthat AP2- and GRE-containing oligonucleotides
8800
ADP-ribosylation Factor 3 Promoter
could compete for binding, however, is a further indication that the factor may have a broader binding spectrum than represented by the ARF3 promoter. A dramatic decrease in transcription was observed after introducing the oligo 3 mutation into a CAT reporter construct, consistent with the view that failure of TLTF, identified by in vitro mobility shift assays, to bind in vivo results in the loss of transcriptional activity. Mutation of the upstream binding site, however, did not produce a significant change in transcription. This observation is reminiscent of the deletion analysis of the closely spaced S p l binding sites in the SV40 promoter, which concluded that the binding site proximal to the start site of transcription has the greatest influence in activating RNA synthesis (Gidoni et al., 1985).The upstream binding site may not directly affect transcription; rather it may serve to stabilize binding of TLTF to the adjacent downstream site via protein/protein interactions. Since the ARF3 gene appears to possess an Spl-independentpromoter, TLTF may be a novel activator for TATA-less transcription and may represent an additional mechanism for tethering the TFIID complex. Further purification and characterization of the factor will be necessary to understand its mode of action and to determine whether it plays a more general role in transcriptional regulation of TATA-less genes. Acknowledgments-We thank Dr. Inez Serventi for helpful discussions, Dr. Carolyn Minth for critically reading the manuscript, and Dr. Su-Chen Tsai for the human ARF3 genomic clone pG5,. REFERENCES Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) J.Mol. Biol. 2 1 5 , 403-410 Bobak, D. A., Nightingale, M. S., Murtagh, J. J., Price, S. R., Moss, J., and Vaughan, M. (1989) Proc. Natl. A c d . Sci. U. S. A. 86,6101-6105 Chen, C., and Okayama, H. (1987) Mol. CelL Bwl. 7,2745-2752 Chen, E. Y., and Seeburg, P. H. (1985) DNA fNyI 4 , 165-170 Crabb, D. W., and Dixon, J. E. (1987) Anal. Biochem. 163,88-92 Degen, S. J. F., and Davie, E. W. (1987) Biochemistry 26,6165-6177 de Wet. J. R.. Wood. K. V.. DeLuca., M.., Helinski. D. R.. and Subramani. S. (1987) Mol.’ceti. B ~ 7,725-737 L Dimam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983) Nucleic Acids Res. ~~~
Pugh, B. F., and Tjian, R. (1990) Cell 6 1 , 1187-1197 Pugh, B. F.,and Tjian, R. (1991) Gems & Deu. 5,1935-1945 Roy, A. L., Meisterernst, M., Pognonec, P., and Roeder, R. G. (1991) Nature RR4. - - -,-24.5-248 -- - --
Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74. - -,.54li.?-5Ali7 - --- - -- .
Seed, B., and Sheen, J.-Y. (1988) Gem (Amst.) 67,271-277 Serafini, T., Orci, L., Amherdt, M., Brunner, M., Kahn, R. A., and Rothman, J. E. (1991) Cell 67,239-253 Serventi, I. M., Moss, J., and Vaughan, M. (1992) Curr. Top. Microbiol. Immurwl. 175,43-67 Serventi, I. M., Cavanaugh, E., Moss, J., and Vaughan, M. (1993) J. Biol. Chem. 268,4863-4872
Set0 E., Shi, Y., and Shenk, T. (1991) Nature 354,241-245 Sew& J. L., and Kahn, R. A. (1988) Proc. Natl. Acad. Sei. U. S. A. 85,4620AWA
Si&h,*H., Sen, R., Baltimore, D., and Sharp, P. A. (1986) Nature 3 1 9 , 1541.552
St&&, T., Willin ham, M.C., Botstein, D., and Kahn, R. A. (1990) Proc. Natl. Acad. S a . S. A. 8 7 , 1238-1242 Strauss, F.,and Varshavsky, A. (1984) Cell 37,889-901 Tsai, S.-C., Adamik, R., Tsuchi a, M., Chang, P. P., Moss, J., and Vaughan, M. (1991a) J. Biol. Chem. 26$: 8213-8219 Tsai, S.-C., Haun, R. S., Tsuchiya, M., Moss, J., and Vaughan, M. (1991b) J. Biol. Chem. 266,23053-23059 Tsai, S.-C., Adamik, R., Haun, R. S., Moss, J., and Vaughan, M. (1992) Proc. Natl. A e d . SCL.U. S. A. 89,9272-9276 Tsuchiya, M., Price, S. R., Nightingale, M. S., Moss, J., and Vaughan, M. (1989) Blochemistry 2 8 , 9668-9673 Tsuchiya, M., Price, S. R., Tsai, S.-C., Moss, J., and Vaughan, M. (1991) J. Biol. Chem. 266,2772-2777
8
