Dec 17, 1992 - Christopher Hovens and Pamela Mitchell for critical reading of the manuscript and ... Church,G.M. and Gilbert,W. (1984) Proc. Natl. Acad.
The EMBO Journal vol.12 no.4 pp.1355- 1362, 1993
Cloned transcription factor MTF-1 activates the mouse metallothionein I promoter
Freddy Radtke, Rainer Heuchel, Oleg Georgiev, Martin Hergersberg, Marisa Gariglio1, Zlatko Dembic1 and Walter Schaffner Universitat Zurich, Institut fuir Molekularbiologie II, Winterthurerstrasse 190, CH-8057 Zurich and IF.Hoffmnann-La Roche AG, PRTB, Grenzacherstrasse, CH-4002 Basel, Switzerland Communicated by W.Schaffner
Metaliothioneins (MTs) are small cysteine-rich proteins whose structure is conserved from fungi to man. MTs strongly bind heavy metals, notably zinc, copper and cadmium. Upon exposure of cells to heavy metal and other adverse treatments, MT gene transcription is strongly enhanced. Metal induction is mediated by several copies of a 15 bp consensus sequence (metal-responsive element, MRE) present in the promoter region of MT genes. We and others have demonstrated the presence of an MRE-binding factor in HeLa cell nuclear extracts. We found that this factor, termed MTF-1 (MRE-binding transcription factor) is inactivated/reactivated in vitro by zinc withdrawal/addition. Here we report that the amounts of MTF-1 - DNA complexes are elevated severalfold in zinc-treated cells, as measured by bandshift assay. We have also cloned the cDNA of mouse MTF-1, a 72.5 kDa protein. MTF-1 contains six zinc fingers and separate transcriptional activation domains with high contents of acidic and proline residues. Ectopic expression of MTF-1 in primate or rodent cells strongly enhances transcription of a reporter gene that is driven by four consensus MREd sites, or by the complete mouse MT-I promoter, even at normal zinc levels. Key words: metallothionein-I gene/metal responsive element (MRE)/transcriptional regulation/transcription factor/zinc finger protein
et al., 1984; Stuart et al., 1984; Andersen et al., 1986; Karin et al., 1987; Otto, 1987; Harlow et al., 1989; Zafarullah, 1988). The mouse MT-I promoter contains six MREs (MRE,a-f) within the first 200 bp 5' of the transcriptional start site. MREaa-d confer metal-responsive transcription when tested independently in front of a reporter gene. The ability to mediate metal-activated transcription varies between the different MREs; MREd is the strongest MRE of the mMTI promoter (Stuart et al., 1985) and responds to the same spectrum of heavy metals as does the complete promoter (Cizewski Culotta and Hamer, 1989). Several groups have characterized MRE-binding proteins, by bandshift assays (Westin and Schaffner, 1988; Searle, 1990, Koizumi et al., 1992), in vitro and in vivo footprints (Seguin, 1991; Andersen et al., 1986; Miller et al., 1988), UV cross-linking experiments (Andersen et al., 1990) and Southwestern analysis (Seguin and Prevost, 1988, Czupryn et al., 1992). These factors specifically bind in a metaldependent manner to the MREs of the mMT-I promoter and therefore might be responsible for metal-induced transcription. Previously, we showed in bandshift assays the binding of a zinc-dependent factor, designated MTF-1, to the mouse MREd sequence using HeLa nuclear extracts (Westin and Schaffner, 1988). Here we report the isolation and characterization of a mouse cDNA encoding the MREd-binding protein MTF- 1. MTF-1 is a zinc finger protein of the TFIIIA type (C2H2) (Brown et al., 1985; Miller et al., 1985). Zinc treatment of mouse cells increases the level of MTF-1-DNA complexes several-fold, probably due to higher affinity and/or concentration of the factor. Expression of MTF- 1 in HeLa cells activates transcription of the wild-type mouse MT-I promoter that contains all MREs, as well as a synthetic promoter consisting of four MREd elements fused to the ( globin TATA box.
Introduction Metallothionein (MT) genes encode low molecular weight, cysteine-rich proteins, involved in multiple cellular processes such as metal homeostasis, adaptation to stress and heavy metal detoxification (Kagi and Schaffer, 1988; for review see Hunziker and Kagi, 1985; Karin, 1985; Hamer, 1986). The synthesis of MTs is controlled at the level of transcription and can be induced by a wide range of different stimuli, including exposure to heavy metals such as Zn2+ Cd2+ and Cu2+ or other stimuli such as treatment with steroid hormones, interleukins, cAMP, diacylglycerol, phorbol ester and interferon (Hunziker and Kagi, 1985; Kagi, 1991). Transcriptional activation by heavy metals is mediated by several copies of a 15 bp consensus sequence (metalresponsive element, MRE) present in the promoter region of all MT genes analyzed to date (Carter et al., 1984; Karin Oxford University Press
Results Cloning of MTF-1 The mouse metallothionein I (mMT-I) promoter contains several MRE sequences (Carter et al., 1984; Karin et al., 1984; Stuart et al., 1984). Of these, MREd is by itself most effective at mediating metal-induced transcription (Stuart et al., 1984). This motif also appears to be the strongest binding site for MTF-l factor (Westin and Schaffner, 1988), therefore MREd was a candidate binding site for screening a cDNA expression library for MTF-1. However, MREd also binds the ubiquitous transcription factor Spl. Thus, using a compilation of known MRE sequences (Stuart et al., 1985) we designed an MRE oligonucleotide (MRE-s), which has the same high binding affinity for MTF-l as the strong MREd-binding site, but lacks any Spl-binding activity 1355
F.Radtke et al.
copies of MRE-s. One clone was found in 2.0 x 106 plaques and characterized by restriction digestion and sequencing. This clone encoded a protein of 69 kDa containing six zinc fingers of the TFIIIA type (C2H2) (Brown et al., 1985; Miller et al., 1985), followed by a putative activation domain
(Figure 3). MRE-s was found to confer metal-induced transcription on a reporter gene as does the MREd-binding site (data not shown). A Xgtl 1 cDNA expression library from the murine lymphocytic leukemia cell line, L1012, was screened with an oligonucleotide probe containing multiple
A 1 1
103 3 205 37 307 71 409 105 511 139 613 173 715 207 817 241 919 275 1021 309 1123 343 1225 377 1327 411 1429 445 1531 479 1633 513 1735 547 1837 581 1939 614 2041
GGAGGGGAGAAGCTGCTGCTGCCGCCGTTGCCGGGAGCCGCGGAGAAAGGCCGTTACCTTTCCATTTCTCACAATTGGGCTGAGCACACGTGAGCCATGGGG 102 M
*
2
G
GAACACAGTCCAGACGACAATATCATCTTCTTTAAGGGAGAGGAGGATGACCTAACCCCGCATGACAAGATGCTCAGGTTTGTGGATGACAACGGACTGGTG E
H
S
P
D
D
N
I
I
F
F
K
G
E
E
D
D
L
T
P
H
D
K
M
L
F
R
V
D
D
N
G
L
204 36
V
CCTTCCTCATCTGGAACTGTTTATGATAGGACCACTGTTCTCATAGAACAGGACCCTGGCACTTTGGAGGATGATGAAGACGATGGACAGTGTGGAGAACCC 306 P
S
S
S
G
T
V
Y
D
R
T
T
V
L
I
E
0
D
P
G
T
L
E
D
D
D E
D
G
0
C
G
E
70
P
TTGCCTTTTCTGGTGGAAGGTGAAGAGGGCTTCTTAATAGATCAGGAAGCAATGTCCCAGGGTTACGTACAACACATTATCTCACCAGATCAGATTCATTTG L
P
F
L
V
E
G
E
E
G
F
L
I
D
Q
E
A
M
S
0
G
Y
V
0
H
I
I
S
P
D
Q
I
H
408 104
L
ACTATAAACCCTGGTTCCACACCCATGCCAAGAAACATTGAAGGTGCAACTCTTACTCTGCAGTCGGAATGTCCAGAAACGAAACGGAAAGAAGTA AAGCGG 510 T
I
N
P
G
S
T
P
M
P
R
N
I
E
G
A
T
L
T
L
0
S
E
C
P
T
E
K
R
K
E
V
K
138
R
TACCAGTGCACCTTTGAGGGATGCCCTCGCACCTACAGCACAGCAGGCAACCTGCGCACCCACCAGAAGACTCACCGAGGAGAGTACACCTTCGTCTGTAAT
612 172
CAGGAGGGCTGTGGCAAGGCCTTCCTCACCTCCTACAGCCTCAGGATCCATGTGCGAGTGCACACAAAGGAGAAGCCATTTGAGTGTGACGTGCAGGGCTGT
714 206
Y
0
0
E
C
G
T C
F
G
E
K
G
A
C
F
P
L
R
T
Y
T
S
Y
S
S
T L
A
R
G I
N
H
L
V
R R
T V
H
H
0
T
K
K
T E
H
K
G
R
P
F
E
E
Y
C
T D
F
V
V
Q
C
G
N
C
GAGAAGGCGTTCAACACACTGTACAGGTTGAAAGCACATCAGAGGCTTCACACAGGGAAAACGTTTAACTGTGAATCTCAAGGCTGCAGCAAATACTTCACC 816 E
K
A
F
N
T
L
Y
R
L
K
A
H
0
R
L
H
T
G
K
T
F
N
C
E
0
S
G
C
S
K
Y
F
240
T
ACACTCAGTGACCTGCGGAAGCACATTCGGACCCATACAGGAGAAAGCCATTTCGATGTGATCACGACGGCTGTGGGAAAGCGTTTGCAGCGAGCCACCAC
T
L
S
D
L
R
K
H
I
R
T
H
T
G
E
K
P
F
R
C
D
H
D
G
C
G
K
A
F
A
A
S
H
918 274
H
CTTAAAACTCACGTCCGGACGCACACTGGTGAAAGACCCTTCTTCTGCCCCAGCAACGGCTGTGAGAAGACGTTTAGCACTCAGTACAGTCTCAAAAGTCAC 1020
308 L K T H V R T H T G E R P F F C P S N G C E -K - T - F- -S- -T -Q- -Y- - S- L - K S H -7 ----.r- - . . . ~ ~~~~~~~~~~~~~ . . ATGAAAGGTCATGATAACAA GGCACTGCATACAGTGCACTTCCACAGCACAATGGATCAGAGGATACAAATCACTCACTTTATCTGAGTGAGCTGGGCCTTi 1122 342 IE D T N H S L Y L S E L G L T A Y S A L P 0 H --N G- -S '... H D N__-K G__M K G _____------_-- _-__- __-- __ -----1CTGTCCACAGATTCTGAACTGCAAGAAAATTCCAGTTCGACACAGGACCAGGACCTCAGCACAATTTCACCAGCAATCATCTTTGAATCAATGTTCCAGAAT'1224 L S T D S E L 0 E N S S S T 0 D 0 D L S T I S P A I I F E S M F 0 N ' 376 -- -
--
1TCCGA3GACCCTGGG3TTCAGGACG2CCCTCTGCA6ACAGCTGCCTTGATTGACAGTTTTAATGGTGATGCAGAGTCCGTCATTGATIGTCCCACCGCCTGCAl326 S
D
D
P
G
I
0
D
D
P
0
L
T
A
A
L
I
D
S
F
N
G
D
A
E
S
V
I
D
'V
P
P
P
'410
A
IGGAAATTCGGCGTCTTTATCTCTTCCACTCGTACTGCAGTCTGGCATCTCCGAGCCACCCCAGCCTCTGCTACCAGCAACAGCTCCGTCCGCTCCTCCACCT'1428
'G N S A S L S L P L V L 0 S G I S E P P Q P L L P A T A P S A P P P '444
1G5TCCCTCCCTAGGACCTGGTTCCCAGCCAGCTGCATTTGGCAGCCCCCCTGCTCTCTTACAACCTCCAGAAGTGCCTGTTCCCCACAGCACACAGTTTGCTl1530 0 478 A
P
S
L
G
P
G
S
Q
P
A
A
F
G
S
P
P
A
L
L
Q
P
P
E
V
P
V
P
H
S
F
T
A
sGCTAATCATCAAGAGTTTCTTCCGCACCCCCAGGCGCCACCCCAGACCATCGTACCAGGACTTTCTGTTGTTGCTGGGGCTCCTIGCATCAGCAGCAACAGTG 1632
iA
649
N
H O
E
F
L
P
H
P 0
A PP
0
T
I
V
P
G
L
S
V
V
A
G
A
P
'A
S
A
A
T
512
V
GCGTCAGCCGTGGCAGCACCAGCCCCACCACAAAGTACTACTGAGCCCCTGCCTGCTATGGTCCAGACTCTGCCCCTGGGTGCCAACTCTGTCCTAACTAAT
A
S
A
V
A
A
P
A
P
P
0
S
T
T
E
P
L
P
A
M
V
0
T
L
P
G
L
A
N
S
V
L
T
N
1734 546
AATCCCACCATAACCATCACCCCAACTCCTAACACGGCAATCCTGCAGTCCAGCCTAGTCATGGGAGAACAGAACTTACAGTGGATATTGAACGGTGCCACC 1836 N
P
T
I
T
I
T
P
T
P
N
T
A
I
L
0
S
S
L
V
M
G
E
0
N
0
L
W
I
L
N
G
A
580
T
AGT;CACCACAAAACCAAGAACAAATTCAGCAAGCATCGAAAGTGGAGCAGGTGTACTTCGCCACCACTGTACCAGTGGCCAGTGGCACAGGGAGCTCTGTT 1938 S
S
P
0
N
0
E
0
I
0
0
A
S
K
V
E
0
V
Y
F
A
T
T
V
P
V
A
S
G
T
G
S
S
V
614
CAGCAGATTGGCCTCAGTGAACCTGTGATCATCATCAAACAAGAGGAGGCCTGTCATGTCAGTCGCGTGCCGGGACTCTGCGAAGGAGCGGGCGGCCGGCCG 2040 0
0
I
G
L
S
E
P
V
I
I
I
K
0
E
E
A
C
H
V
S
R
V
P
G
L
C
E
G
A
G
G
R
648
P
GAGAAAGGGATGCCCCACCCCCTCCAGAGCCCGAACCCCCAGCCTCCTGATGGGCCCAGCCTGCAGCTGCCACCCTAGACTTCTCCCTCATCCCTGCCCCCC 2142 E
K
G
M
P
H
P
L
0
S
P
N
P
0
P
P
D
G
P
S
L
0
L
P
P
673
*
2245
CCCCCTCCGCTGGACCCTCTTCCTGTGAGCACAGCAGACGGCTGAGGCTCCTCAGACCCTCCGAAACGTTAAGTGCCATGGATGTGTCAGAGTTCCTGTCCC 2244 CTCCAGAGCCTGGACACCCCGTCCAGTCCGGTTCACAGTGAAGCACTACTGCAGGGGTAGAGGAAAGACCTGGCCGGCAGCTTCTCCAGGGGAAGGGCCTGT 2346
2347 2449
GCGCACACCCCTCCAGGAACAGGGTGAGCAGGAAGTACCGGCTGTGATGCTGCCGTCATGGGGTCAGAAATTGGAAGGATGAAGAAACCTGCCATCTGAAAG 2448 CTCACCTTTCAGACGTATTTTCTTTACTCGTATCCCAGGAACATCCATTTTAAGGAACTGATCTTTGGAGGAAAAAACAAAAAACAAAAACAAAAAAAAAAA 2550
2143
B 1
315 329
137
zinc finger domain
405/406
506
673
acidic proline-rich region region
Fig. 1. DNA and amino acid sequence of the mouse MTF-1 cDNA. (A) The nucleotide sequence of the mMTF-1 cDNA is shown along with the deduced amino acid sequence, using the single letter code. Position 1 represents the first nucleotide of the cloned cDNA insert. For amino acids, the first methionine was assigned number 1. Stars designate in-frame stop codons. Some features of the sequence are highlighted as follows: the putative methionine initiation codon (underlined), the zinc finger region (boxed) containing six zinc fingers of the TFEIIIA type (C2H2) and two putative activation domains, namely an acidic (aa 329-405) and a proline-rich domain (aa 406-506; dashes). The cDNA terminates in an A-rich region which, however, does not appear to be the poly(A) tract, since we did not find a poly(A) signal. A putative nuclear localization signal is present 5' of the zinc finger region. (B) Schematic diagram showing zinc finger and putative transcription activation domains of mMTF-1.
1356
Metallothionein promoter is activated by cloned MTF-1
with a high representation of acidic amino acids, followed by proline-rich sequences (Figure 1). This cDNA, termed mMTF-la, was used for most binding and activation studies. Subsequently we isolated another clone from 1.2 x 106 plaques of a murine lung cDNA library. This cDNA, which is 1.2 kb in length, overlaps the MTF-la clone and extends further than it towards the 5' end. This clone contains another, better AUG start site 25 amino acids (aa) upstream of the previously utilized start site of MTF- la. Within 6 bp upstream of the more N-terminal AUG there is an in-frame stop codon. A composite cDNA clone including this more N-terminal region was then constructed. Utilization of the more N-terminal AUG start site, present in the longer composite clone, yielded a protein with a predicted mass of 72.5 kDa. Expression of this cDNA clone in HeLa and COS cells revealed the same binding and activation properties as those of the slightly shorter MTF-la. The protein encoded by the clone was termed mMTF-1, since it was indistinguishable from the factor previously identified in nuclear extracts (Westin and Schaffner, 1988; see below).
Sp1HSV oligonucleotide or MRE-s oligonucleotide, respectively (Figure 2A, lanes 6 and 7). Bandshift analysis with nuclear extracts from transfected COS cells and MREd oligonucleotide resulted in a strong mMTF-la complex (Figure 2A, lane 1) migrating at the same position as the endogenous mMTF-1 complex. The competition behavior of the mMTF-la complex was identical to that of the endogenous mMTF-I complex; it was not competed by an Spl HSV oligonucleotide (Figure 2A, lane 2) but was competed by an MRE-s oligonucleotide (Figure 2A, lane 3). Not unexpectedly, the metal chelator EDTA inactivated the binding of Spl, MTF-1 and cloned mMTF-la (Figure 2A, lanes 4 and 8). This inhibition was also observed with o-phenanthroline (Figure 2A, lanes 11 and 12). The specificities with which mMTF-la and endogenous mMTF-1 bind DNA were also compared using mutant MRE oligonucleotides in bandshift analysis (Figure 3). Besides MRE-s, five single point mutations were tested, designated MRE Mutl -Mut5. As shown in Figure 3, both endogenous and the cloned recombinant MTF-la factor had the ability to bind all mutants, though to a variable extent. Most importantly, the relative affinities for the different mutant DNA sequences are identical for endogenous mouse MTF-1 (Figure 3A) and cloned MTF-la (Figure 3B).
Endogenous MTF- 1 and recombinant MTF- la exhibit identical binding properties The cDNA of mMTF-la was recloned into two different expression vectors to allow translation from the heterologous AUG site of the thymidine kinase gene, or under the control of the first AUG in the cDNA itself. When driven by the strong enhancer/promoter of cytomegalovirus in transfected monkey COS cells, large quantities of factor were produced as assessed by bandshift analysis using MRE oligonucleotides [Figure 2A, compare lane 1 (COS/recomb.MTF-1a) with lane 9 (COS cells)]. Bandshift analysis using nuclear extracts from mouse 3T6 cells and MREd oligonucleotides identified an MTF-1 complex and a more slowly migrating Spl complex (Figure 2A, lane 5). These two complexes could be competed specifically using a 400-fold excess of unlabeled
A
Endogenous MTF- 1 and recombinant MTF- 1 have indistinguishable digestion patterns in the proteolytic clipping bandshift assay Binding reactions were carried out as described in Materials and methods using nuclear extracts as indicated in Figure 2B and C. The nuclear extracts were incubated with the endoproteinase Glu C and subsequently analyzed in a bandshift experiment using end-labeled MREd oligonucleotides. The patterns obtained for the proteinDNA complexes of the recombinant MTF-la and the endogenous mouse MTF- 1 were similar but not identical.
C
B -I.
I.A + "., . 2,;-C
+
-
r r @+
+
Fig. 2. (A) Bandshift analysis using nuclear extracts from COS cells alone (lane 9), from COS cells transfected with cloned mouse MTF-la (lanes 1-4, 11), or from mouse 3T6 cells (lanes 5-8, 12). A 39 bp end-labeled MREd oligonucleotide was used as probe. A 400-fold molar excess of SplHSV oligonucleotide (lanes 2 and 6) or MRE-s oligonucleotide (lanes 3 and 7) was added prior to addition of extracts. Nuclear extracts were treated with 2 mM EDTA (lanes 4 and 8) or with 2 mM o-phenanthroline (lanes 11 and 12) before adding to zinc-free binding buffer. The positions of the protein-DNA complexes are indicated as Spl, pMTF (primate MTF) and mMTF (mouse MTF). (B and C) Proteolytic clipping bandshift analysis of MREd DNA-protein complexes using nuclear extracts from mouse 3T6 cells (lanes 13 and 19), from mouse L1210 cells (lane 17), from COS cells containing recombinant mMTF-la (lanes 15 and 21), from COS cells containing recombinant mMTF-l (lane 18), from HeLa cells containing recombinant mMTF-1 (lane 20), from COS cells (lanes 16 and 22), or a mixture of nuclear extracts from 3T6 cells and COS cells containing recombinant mouse MTF-la (lane 14). After an initial binding reaction, endoproteinase Glu C (V8-protease) was added, resulting in specific MTF DNA proteolytic cleavage products as indicated by the arrows.
1357
F.Radtke et al.
']i'.. 2
::~W
M
W
*
'3
B
I\
AN\
MRE-s
Mt.t' Mu'
2
M,i3
MREa
m
R: 1j
M RFc
MARL .,-
. i.
..
-,
I"
M.1, q :.'-; H)
Fig. 4. South-western blot analysis using nuclear extracts from COS cells transfected with an expression vector of the cloned mMTF-l cDNA (lane 1), 3T6 cells (lane 2) and COS cells (lane 3). End-labeled MREd or MRE-s oligonucleotides were used as probes. Competitions were done as indicated below using a 100-fold molar excess of unlabeled Sp1HSV or MRE-s oligonucleotides.
One MTF -DNA complex of the recombinant protein migrated more rapidly than the corresponding complex from the endogenous factor (Figure 2B, lanes 15 and 13). No differences between recombinant MTF-la and endogenous mouse MTF- 1 were seen when proteolytic patterns generated by Arg C endoproteinase were compared (data not shown). The second cDNA clone with the N-terminal extension, by contrast, was indistinguishable in the proteolytic clipping experiments from endogenous mouse MTF-l (Figure 2C, lanes 18 + 20 and 17 + 19) indicating that these proteins are identical. In another binding study, South-western blot analysis was used to determine the nature of the protein oligonucleotide interaction. As shown in Figure 4, filters probed with MRE oligonucleotides revealed specific binding of a single protein in nuclear extracts from mouse 3T6 cells (lane 2) and a major band in COS cell nuclear extracts (lane 3). There is a species difference, in that the mouse MTF-l (lanes 1 and 2) migrates more quickly than the corresponding primate MTF-l (lanes -
1358
,O
V0
