show that a domain of PU.1 that activates transcription from multimerized PU.1 binding sites is not required to activate the 1 enhancer together with Ets-1. In.
The EMBO Journal vol.15 no.17 pp.4665-4675, 1996
Context dependent transactivation domains activate the immunoglobulin g heavy chain gene enhancer
Batu Erman and Ranjan Sen1 Rosenstiel Basic Medical Sciences Research Center and Department of Biology, Brandeis University, Waltham, MA 02254, USA
'Corresponding author
Enhancers and promoters nucleate the assembly of multiprotein complexes that are required for the transcriptional activation of eukaryotic genes. Although multimerized binding sites of individual transcription factors sometimes mimic the properties of an enhancer, the combinatorial use of factors is considered to be crucial for achieving biological specificity. The minimal B cell specific immunoglobulin j heavy chain gene enhancer is activated by a combination of tissuerestricted ETS proteins and ubiquitously expressed basic helix-loop-helix transcription factors. Here we show that a domain of PU.1 that activates transcription from multimerized PU.1 binding sites is not required to activate the 1 enhancer together with Ets-1. In contrast, a transactivation domain in Ets-1 is necessary to activate this enhancer synergistically with PU.1. Furthermore, the Ets-1 activation domain functions only when tethered to the pA site of the enhancer. These observations illuminate two forms of context dependence: first, all possible transcription activation domains may not be required to achieve combinatorial specificity; second, functional transcription activation domains may require appropriate positioning on DNA. Keywords: context dependent transactivation/enhancer/ ETS domain/immunoglobulin heavy chain
Introduction The immunoglobulin ,u heavy chain (IgH) gene intron enhancer is a tissue-specific regulatory element that is necessary for expression (Libermann and Baltimore, 1990; Nelsen and Sen, 1992) and rearrangement (Yancopoulos and Alt, 1985; Alt et al., 1986; Oltz et al., 1993) of the IgH gene. The correlation between the transcription and the recombination of this gene has been verified in mice containing disrupted p enhancer alleles (Chen et al., 1993; Serwe and Sablitzky, 1993). Moreover, experiments in transgenic mice indicate that this enhancer is sufficient to direct expression of heterologous genes in B and T lymphoid cells (Adams et al., 1985; Langdon et al., 1986; Reik et al., 1987; Schmidt et al., 1988). Thus, the enhancer is activated at very early stages of B cell differentiation, prior to the onset of DH to JH rearrangements, and serves as a useful probe for regulatory events that direct early B lymphoid cell differentiation. The Ig j enhancer (p70) contains binding sites for multiple trans-acting nuclear factors. Proteins that bind to ( Oxford University Press
the gEl-gE5 sites of the enhancer belong to the basic helix-loop-helix (bHLH) family of factors and are expressed ubiquitously. Proteins with more restricted tissue distribution bind to the ,A, ,B and octamer motifs of the enhancer. A noteworthy feature of the enhancer is that mutation in single protein binding elements typically does not significantly reduce enhancer activity. This observation has been interpreted to mean that the various motifs of the enhancer are functionally redundant. To assist in the mechanistic analysis of enhancer function, we have previously defined a minimal domain of this enhancer (g70) in which all the elements present are simultaneously required for enhancer activity. This domain contains three elements: the ,uB site which binds the ETS domain protein, PU. 1; the ,A site which binds several ETS proteins such as Ets-1, Erg-3, Fli-l and ERP (Nelsen et al., 1993; Rivera et al., 1993; Lopez et al., 1994); and the gE3 site which binds bHLH-leucine zipper proteins (bHLH-LZ) such as TFE3, USF and the Myc/ Max heterodimer (Beckmann et al., 1990; Gregor et al., 1990; Roman et al., 1992; Blackwell et al., 1993). It is noteworthy that the minimal p enhancer domain we have identified contains binding sites for both ubiquitous and tissue-restricted enhancer binding proteins, analogous to the components of the full enhancer. Our working hypothesis, therefore, is that understanding the mechanism of activation of the minimal enhancer represents a valid starting point for the analysis of the full enhancer. In transient transfection assays in S 194 B cells, the ,70 activates transcription as a monomer or a dimer; however we usually assay the dimeric form because of its higher activity. B cell-specific activation of ,70 dimers closely parallels the larger monomeric enhancer, insofar as mutation of either the ,uA or the jB elements essentially abolishes the activity of both enhancers. However, mutation of the ,uE3 element significantly affects the activity of the g70 enhancer, but not the larger enhancer. This is most likely because the loss of the ,uE3 element in the context of the larger enhancer is compensated for by other bHLH binding motifs such as uE 1, E2 and E5. In the minimal domain iE3 is the only such element present and therefore is critical for enhancer function. Furthermore, it is interesting to note that multimerized copies of neither ,uB nor ,uE3 activate a basal promoter in B cells or non-lymphoid cells (Beckmann et al., 1990; Nelsen et al., 1990), whereas multimerized gA elements activate transcription in pre-B cells, but not in mature B cells (Libermann and Baltimore, 1993). Thus, the activity of the minimal enhancer in B cells most likely represents a combinatorial specificity achieved by the juxtaposition of binding sites for the three factors. Consistent with this hypothesis, the (g70)2 enhancer can be transactivated in non-lymphoid cells by co-expressing the ETS domain genes PU. I and Ets-1, but not by either gene alone (Nelsen 4665
B.Erman and R.Sen
et al., 1993). Furthermore, in non-lymphoid cells as well, all three sites are simultaneously required for optimum activity, thus closely mimicking the characteristics of the enhancer in B cells. Based on these results, we have previously proposed that expression of PU.] and Ets-] in non-lymphoid cells complements the lack of tissue restricted genes required for g enhancer activation and provides a system for dissecting the mechanism of activation of this enhancer by ETS domain genes (Nelsen et al., 1993). In this study we sought to identify the domains of PU. 1 and Ets- 1 proteins that are required to activate the p70 enhancer. We found that a transactivation domain in PU. 1 that activates transcription from multimerized PU. 1 binding sites is not required to activate the ,u enhancer together with Ets-1. In contrast, a transactivation domain in Ets- 1 is necessary to activate this enhancer synergistically with PU. 1. Furthermore, the Ets- 1 activation domain functions only when tethered to the gA site of the enhancer, and not to the gB site. The PU. 1 activation domain can substitute for this activity when tethered to the ,uA site with a DNA binding domain that recognizes the gA site. These results demonstrate that functional transcription activation domains may require appropriate positioning on DNA to achieve combinatorial specificity.
Results Analysis of PU. 1 deletion mutants Several functional domains have been previously characterized in the PU. 1 protein (Figure lA). These are an N-terminal transcription activation domain necessary to activate multimerized PU. 1 binding sites (Hagemeier et al., 1993; Shin and Koshland, 1993; Kominato et al., 1995), a PEST domain containing two casein kinase II phosphorylation sites necessary for interaction with NF-EM5 (Pongubala et al., 1992, 1993), and the DNA binding ETS domain (Klemsz et al., 1990). We have previously shown that in non-lymphoid cells, a dimer of the minimal B cellspecific Ig g enhancer [(470)2] is weakly transactivated by PU. 1 alone, but strongly activated when PU. 1 and Ets-1 are co-expressed (Nelsen et al., 1993). To determine which portion of PU. 1 was required to activate the minimal g enhancer in non-lymphoid cells, deletion mutants of PU. 1 were co-expressed with full-length Ets- 1 protein in COS cells and assayed for their ability to transactivate a reporter plasmid containing a ,70 dimer. Sequential deletion of the previously identified transactivation domain (PU.1A97) and the PEST domain (PU.1A162) did not affect Ets-1 dependent synergistic transactivation of the minimal t enhancer (Figure 1B). The largest deletion in PU.1A200 removed a part of the PU.1 ETS domain and consequently encoded a protein that did not bind DNA (data not shown) or activate transcription (Figure 1 B). Expression levels of the six functional PU. 1 mutants were shown to be comparable by electrophoretic mobility shift assays (EMSA) using a PU. 1 binding site derived from the SV40 enhancer, and whole cell extracts prepared from transfected COS cells (Figure 1C). Full-length PU. 1 generated the most-slowly-migrating complex, specific to transfected cells, indicated by the topmost arrow (Figure 1C, lane 1) and several faster-migrating complexes that we attribute to partially proteolyzed PU. 1 protein. A similar pattern was seen with the PU.1 deletion mutants
4666
(Figure 1 C, lanes 2-6), showing that expression of the total DNA binding PU.1 derivatives was comparable in these experiments. We conclude that the previously identified transactivation domain of PU. 1 is not necessary to synergistically transactivate the minimal g enhancer with Ets- 1. Conclusions of all the transfection data in this paper are presented schematically in Figure 7. Line A shows that enhancer activity is obtained by co-transfecting full-length PU. 1 and Ets- 1, which are shown bound to the gB and gA sites respectively. Note that for simplicity, only one of the ,70 fragments is shown in the figure. Furthermore, although the gE3 site is shown in the figure without any protein bound, it is likely that this site is occupied by an endogenous protein since it is required for enhancer activity in these assays. The PU.1 deletion study is summarized in line B.
Analysis of Ets-1 deletion mutants Like PU.1, several functional domains have been previously described in the chicken c-Ets-1 protein. A transactivation domain between the residues 124-243 in the c-Ets-1 protein was identified by making fusion proteins with heterologous DNA binding domains (Gegonne et al., 1992; Schneikert et al., 1992; Wasylyk et al., 1993), a region of the murine protein between residues 254 and 293 was shown to inhibit DNA binding by the Ets- 1 ETS domain [ETS(Ets-l)] (Lim et al., 1992; Nye et al., 1992; Wasylyk et al., 1992), the C-terminal ETS domain itself (Ho et al., 1990; Nye et al., 1992; Wasylyk et al., 1992) and a C-terminal tail which is also inhibits DNA binding (Hagman and Grosschedl, 1992; Lim et al., 1992). We constructed expression vectors containing N-terminal truncations of the gene encoding murine Ets-1 (Figure 2A) to identify which regions of Ets-1 were required to synergistically transactivate the ,u enhancer together with PU. 1. The first three mutants lack 167, 231 and 286 amino acids from the N-terminus of Ets- 1 respectively. The last mutation shown [ETS(Ets-1)] contains only the 100 amino acid ETS domain of Ets- 1 (residues 325-426) and lacks all other sequences, including 13 amino acids from the carboxy terminus. In COS cell co-transfection experiments, only the full-length Ets- 1 and the first N-terminal deletion (Ets-IA 167) synergized with PU. 1 to activate the g enhancer (Figure 2B). In the presence of Ets-1A231, activity of the reporter plasmid was .... r!'.
...
H
fi
ff
Hw
4
:- -
;) F > r
95% identical to the human and murine Ets-I proteins (Watson et af., 1988; Gunther et al., 1990). Cross hatched regions extend from residues 254 to 293, indicating the position of the domain shown to inhibit sequence specific DNA binding (Lim et al., 1992; Nye et al., 1992; Wasylyk et al., 1992) and from residues 415 to 440, indicating the C-terminal tail which is also inhibitory to DNA binding (Hagman and Grosschedl, 1992; Lim et al., 1992). The shaded carboxyterminal domain (residue 336 to 415) is the DNA binding ETS domain (Ho et al., 1990; Nye et al., 1992; Wasylyk et al., 1992). Within this domain the dark hatched regions mark the highly conserved regions I and 2 (Wasylyk et al., 1993). (B) Transient transfection analysis of Ets-1 deletion mutants in COS cells. (p70) dimer reporter plasmids were co-transfected with full-length PU. 1 and deletion mutants of Ets- 1 as indicated in the figure, and CAT assays performed as described in the legend to Figure 1. CAT enzyme activity is shown on the X axis as the fold activation normalized to the activity of the reporter plasmid in the absence of co-transfected transactivators (last bar). All results shown represent the average of at least four transfections carried out in duplicate. (C) EMSA analysis of Ets-1 deletion mutant expression in transfected COS cells. A consensus Ets-l binding site (SC-l) (Nye et al.. 1992) was used in EMSA with 20 gg of whole cell extract prepared from the cells transfected with expression vectors as indicated. Nucleoprotein complexes formed by full-length Ets-1, or deletion mutants, are indicated by arrows on the left; asterisks indicate complexes attributed to proteolyzed proteins. Topmost band is present in extracts from un-transfected cells. (D) Western blot analysis of Ets-I deletion mutant expression in transfected cells. Arrows indicate the migration of Ets-1, or deletion mutants, recognized by a C-terminal antibody.
strengthen the results obtained above, we tested the ability of key PU. I and Ets- 1 deletion mutants to activate a monomeric minimal Rt enhancer in the same reporter plasmid. Co-expression of full-length PU. 1 and Ets- 1 transactivated the monomer enhancer -5-fold (Figure 3) compared with the 15- to 20-fold increase in transcription observed with the t70 dimer containing reporter plasmid (see for example, Figures lB and 2B). The shortest version of PU. 1 (PU.1A162), that retained the DNA binding ETS domain and the ability to activate g70 dimers synergistically with Ets- 1 (Figure IB), efficiently activated the monomeric enhancer together with Ets- 1 (Figure 3). In contrast, the smallest Ets- 1 deletion mutant (Ets- 1A286,
4668
see Figure 2A) did not synergize with full-length PU. 1 to transactivate the monomeric enhancer (Figure 3). Note that for these transfections, ELISA was used to quantitate the levels of CAT enzyme expressed and the data are normalized to the levels expressed in the presence of fulllength PU.1 and Ets-1, which is assigned a value of 100. We conclude that the domains of the ETS proteins required to activate the monomeric enhancer are the same as those that activate p70 dimers.
Analysis of PU. 1-Ets- 1 interactions The lack of requirement of the previously identified PU. 1 transactivation domain suggests that the PU. 1 part of the
Context dependent transactivation domains 200
-
acn -I0 U,
0D 0
1 00
cn LU
70'tn C)
a:
C/)
O/)
00O
CL U)
0
-
_112 84
C-)
His-Ets-1
1
*
-53 34
-29 PU.1
+
Ets-1 PUlA1 62 Ets-1 A286
+
+
Fig. 3. Activation of a ilmonomiler-ic miniimal enhancer by PU.Ia1nd Ets- deletion1 mItutants. (p55) motinomerci reportel plaisimid (2 pL) was co-transfecteCd With fLIl 1-lelnltl PU.1 (2' [t) anid Ets-1 (2 p ) 01 deletioni mutanits ther-eof aLs ilidicated in the fi-Lire. aind CAT assavs per-for-miled by ELISA as described in Materials a;nd methods. CAT enl/Vime activitv is shown oni thc axis as the percentage ot the activitv of the reporter plasmild in the presenice of co-trailsfected PU.I anid EtsIlist b;ar). ResuIlts shown ripresens t the avxra2e of tx-o tr-anisfections cair-ied out in duplicaLte.
multi-proteiin complex on the enhancer mav not interact w,ith RNA polymerase 11 and the basal trainscription machinery,. Several models foI the possible functioin of the ETS domain of PU. 1 on the p ei-haincer are considered in the discussion section. We tested whether PU. 1 mav contribute to enlhaincer activity by interacting directly 'with Ets- 1. Equivalent amouints of glutathione S-transferase (GST) fusioni proteinis conta.ining full-lenath PU. 1. the ETS domain of PU. 1. a fraament containingc, the SH2/SH3 domains of Crk II. or the GST protein alone were bound to alutathione-agarose beads. followed by incubation with full-length Ets- I protein. Matrix-bound proteills 'Were separated by SDS-PAGE and associated Ets-1 was detected by immunoblottingL with an anti-Ets-1 antibody. Ets- 1 protein bound to GST-PU. 1 or the GST-ETS(PU. 1) but not efficiently to either the beads alone oI beads containine GST or GST-Crk II (Figure 4). We confirmed that the beads contained comiipar-able amounts of GST protein derivatives by probing the same Western blot with anti-GST antiserum (data niot shown). Our results showv that Ets-1 canl interact directly with PU.1 i)l vitro. and that the ETS domain of PU.1 is sufficient for this interaction. These observations suggest that the ETS domain of PU.1 may activ!ate the pt enhanicer by interacting with DNA bound Ets- 1.
Analysis of PU. 1-Ets- 1 chimeric proteins Our wvorkina model is that the functional minimal enhancer binds PU. I at the pB site. Ets- 1 at the pA site and a basic helix-loop-helix protein at the interveninlg pE3 sequence. as diaCrammed in Ficure 7A. The studies described above sucyest that the PU. I activation domain at the pB site is not required for enhancer ftunction, but it is necessary to hav-e the Ets- I activation domaini at the pA site. Two interpi-etations are possible foI- these observations: either that the Ets- I transactivation domain has special properties
Fig. 4. PU. 1 ETS domaini inter-acts with Ets- i71 ivitro. GST fusion containtine full-lenLth PU.1 (GST-PL.1). the ETS domain of PU.I [GST-ETS(PU. I). CrII (GST-Crkill). or GST protein alone (GST)] were bound to Llutathione-agarose beads. followed bh inicubation with histiditne tagged Ets- pr-otein (His-Ets-I ). The mratrixbound proteinis were fractionated by SDS-PAGE and adsor-bed Ets-I was detected by imm1untioblottin- wvith ain anti-Ets-I antibodv. proteins
that ai-e required to activate this enhancer, or alternatively. that the location of the tranisactiv-ation domain on the enhancer DNA is important for the function. To distinauish between these possibilities, we constructed two chimeric proteins. In the fusion protein PU.l(Ets). we replaced the ETS domain of PU. I with the ETS domaini of Ets- 1. and in the fusioIn protein Ets-l(PU), we replaced the ETS domain of Ets- I with the ETS domain of PU.1 (Figure 5A). Because the DNA binding specificity of ETS domain proteins is believed to be determined by the ETS domain itself. we reasoned that PU.l(Ets) would intei-act primarily with the p.A site and therefore brincg with it to that site the PU.1 transactivation domain. Conversely, Ets-l(PU) would interact primarily with the pB site and therefore brincg with it to the ftB site the rest of Ets- 1. We first determined that the chimeric proteins had the expected DNA binding specificities. The PU.l(Ets) chimera was expressed as a GST fusion protein anid anialyzed by EMSA. GST-PU.I(Ets) bound to a wild-type enhancer probe that contains the tA, tE3 and ptB motifs (Figure 5B lane 2). and to a probe in which the ptB site had been mutated (Figure 5B. lane 4). but not to a probe containing a mutant pA site (Fiaure 5B. lane 3). The binding pattern showed that this chimeric protein recognized primarily the p.A site. Additional evidence for altered specificity was obtained by using the PU. binding site from the SV40 enhancer and a consensus Ets-l binding site (SC-1) in EMSA. The SC-I probe. but not the SV4O probe. bound GST-PU.l(Ets) (Figure 5B. lanes S and 6). The Ets-I(PU) chimera was expressed as a hexa-histidine tagged protein [His-Ets- I (PU)] in bacteria. I, vitro bindina assays with p enhancer-derived probes showed that this protein bound the wild-type and p.A mutated enhancer DNA. but not to the probe mutated at the pB site (Figure 5C. lanes 2-4). showing that it recognized the p.B site more efficiently than the pA site. These results indicated that the PU. I (Ets) chimera recognized the p enhancer via the p.A site and the Etsl(PU) chimera via the p.B site. We then tested both chimeras for their ability to activate the minimal p. enhancer in COS cells. Like wild-type PUL. PU.I(Ets) 4669
B.Erman and R.Sen
Fig.
5. Altered DNA
(residues indicated above)
ETS domain (hatched was
PU.1-Ets-I
region)
was
Ets-I
to 267 of
PUl1)
is shaded with dots, and the
replaced by
representation of PU.1-EtsI1chimeras. In the PU.1(Ets) chimera, replaced by the ETS domain of Ets-I (102 amino acids, residues PU.1 portion is hatched. In the Ets-1(PU) chimera, the ETS domain of (residues indicated above). This resulted in a protein containing the PU.I1
chimeras. (A) Schematic
(107 amino acids, residues 160
325-427 of Ets-1). The ETS domain of
Ets-I1
of
binding specificity
the ETS domain of PU.1
the ETS domain of PU. 1
was
Ets-I1 (dotted region). The five amino acid difference in size between Ets-1I to maximize homology with PU.lI (Wasylyk et al., 1993). (B) DNA
within the framework of
due to gaps introduced in the ETS domain of
the two ETS domains
binding analysis
of the
expressed as a GST fusion protein [GST-PU.1(Ets)] and purified from bacterial extracts. EMSA was enhancer containing the g.A, l.B and l.t3 motifs carried out with this protein and the following probes: lane 2, Pstl-BamHl fragment of the (Nelsen et al., 1990); lane 3, same fragment with a mutated ItA element; lane 4, same fragment with a mutated .tBl element; lane 5, PU.1 binding site from the SV40 enhancer; lane 6, optimal Ets-I1 binding site, SC-i. Specific nucleoprotein complexes are indicated by arrows. Lane contains the wild-type probe with no added protein. (C) DNA binding analysis of the Ets- l(PU) chimera. The chimeric protein was expressed as a hexahistidine tagged protein [His-Ets- l(PU)] and purified from bacterial extracts by nickel affinity chromatography. EMSA used g. enhancer Pstl-Ba,nHI PU.1(Ets) chimera. The chimeric protein
fragment probes wild-type probe
wild-type protein.
obtained from the with
no
added
was
enhancer
(lane 2),
l.A mutant enhancer
alone activated the g7O dimer weakly (Figure 6A). Unlike PU. 1, which synergizes with Ets- 1 to activate the i enhancer fragment, PU. 1 (Ets) transactivated the enhancer with PU. 1 (Figure 6A, and summarized schematically in Figure 7D), but not with Ets- 1 (data not shown). To obtain further evidence that transactivation seen with the chimera was mechanistically similar to the synergy between PU. 1 and Ets- 1, we tested the domain requirements of PU. 1 in this context. We found that the ETS domain of PU. 1 was sufficient to activate the ,u enhancer fragment when coexpressed with PU. 1 (Ets) (Figure 6A, and summarized in Figure 7E). We conclude that PU. 1 residues present in this chimera can functionally substitute for the Ets- 1 transactivation domain when tethered to the ,uA site via the ETS domain of Ets-1. These experiments suggest that the Ets- 1 activation domain is not unique in being able to activate the g70 enhancer. The observations described above show that domains of PU. 1, other than the DNA binding ETS domain, can contribute to transcriptional activity of the g70 dimer, only when bound to the j.A site, but not to the gB site (for schematic summary compare Figure 7C and E). However, based on these experiments, we could not tell whether the Ets-1 activation domain was also subject to strict position effects. We therefore examined whether the location of the Ets-1 activation domain at the ,uA site was
4670
(lane 3) and .tB
mutant
enhancer
(lane 4). Lane
contains the
important for function. To tether the Ets- 1 activation domain to the gB site we used the Ets- 1 (PU) chimera that binds primarily to the ,uB site in vitro. Expression of Ets1 (PU) alone, or together with the ETS domain of Ets- 1 (Ets-1A286) did not activate the enhancer in COS cells (Figure 6B). These results suggest that an enhancer with Ets-1(PU) bound at jiB and Ets-1A286 bound at gA is not active (Figure 7F). To confirm that the Ets-1(PU) chimera was expressed, we assayed whole cell extracts from transfected cells by EMSA (Figure 6C). The SV40 PU.1 probe generated a unique nucleoprotein complex in cells transfected with the Ets-1(PU) expression construct (Figure 6C, lanes 1 and 2, indicated by upper arrow). PU. 1 expression in these cells generated a faster mobility complex indicated by the lower arrow (Figure 6C, compare lanes 2 and 3). Thus the chimeric protein was expressed at levels comparable with PU. 1. To demonstrate that the Ets- 1 (PU) chimera was a transcriptional activator, we assayed its function on a synthetic enhancer in which the jA sequence was replaced by a second ,uB element. A dimer of this enhancer (containing four [tB elements) was strongly transactivated by Ets- I (PU) alone in COS cells (Figure 6D), thereby confirming that this protein could activate transcription. We conclude that the previously identified activation domain of Ets- 1, when tethered to the ,uB site, cannot activate transcription as it does from
Context dependent transactivation domains
B A
2cO ::S -
.5
z
-
:
ul
-
c ;)
u
0
2
PU.i
PU.1
EtsPU.! Et.:; P U.11 e 2
Ets-.
_-U
-_-_
Ets-1. PU; Ets-1 286h
C probe: SV40
Pu.1 Ets-1(PU)
D
+ -
+ -
+
-
30
-
--
20 A
4
I--w 1.
C)
am
C)
10-
-
0-
Pu.1 ETSI PU.1, 12
34
Ets-< PU
Fig. 6. Functional analysis of PU.l-Ets-l chimeric proteins. (A) Analysis of the PU.l(Ets) chimera. p70 dimer reporter plasmid was co-transfected into COS cells with expression vectors for various proteins as indicated below the graph; CAT enzyme activity was assayed. and is normalized to the activity of 170 dimer reporter in the absence of co-transfected transactivator (last bar). Results shown represent the average of at least four transfections carried out in duplicate. (B) Analysis of the Ets-l(PU) chimera. COS cell transfections were carried out as described above. Results shown are normalized to the activity of .00 dimer reporter in the absence of co-transfected transactivator (last bar), and are the average of at least four transfections carried out in duplicate. (C) Expression of Ets-l(PU) protein in transfected COS cells. The SV40 PU.1 binding site was used in EMSA with extracts obtained from COS cells transfected with expression vectors encoding the proteins as indicated in the figure. Upper arrow indicates the position of Ets-l(PU)-DNA complex and the lower arrow indicates the position of PU.1-DNA complex. Bands labeled with asterisks are due to partially proteolyzed proteins. These complexes were not observed in non-transfected cells (lane 4). (D) Activity of Ets-l(PU) chimera assayed with a minimal enhancer in which the ,uA site was replaced with a second 1B element. resulting in the configuration pBPE3.tB instead of the normal ,uA,uE3tB. tBtE3tB dimer synthetic reporter plasmid was co-transfected into COS cells with expression vectors for PU. 1. the ETS domain of PU.1 and the Ets-1(PU) chimera; CAT enzyme activity was assayed, as described in the legend to Figure 1 and is normalized to the activity of the gBgE3gB dimer reporter in the absence of co-transfected transactivator (last bar). Results shown represent the average of three transfections carried out in duplicate.
the iiA site. Therefore, positioning of the transactivation domain on DNA is critical for generating enhancer activity.
Discussion In this study we have identified the domains of factors required to activate the minimal Ig ,t heavy chain gene enhancer. This enhancer contains three sequence elements:
,uA, gB and ,uE3, that bind Ets- 1, PU. 1 and several ubiquitously expressed bHLH-ZIP proteins respectively (Kadesch, 1992; Nelsen et al., 1993). Here we show that the ETS domain of PU.1 is sufficient to cooperate with full-length Ets- 1 to transactivate the minimal enhancer in COS cells, suggesting that the previously identified activation domain of PU.1 is not required for . enhancer activity. In contrast, a transactivation domain in Ets-1 is necessary for this protein to cooperate with full-length
4671
B.Erman and R.Sen
Fig. 7. Schematic summary of transfection analyses. The ,uA, ,uB and ,uE3 elements of the tripartite t70 enhancer are represented as the light and dark single headed arrows, and the double headed arrow respectively. This is to signify the directionality of the ,uA and ,tB sites and the partial palindromic character of the tE3 element. For simplicity, protein-bound enhancers are not shown as dimers. PU.1 protein is shown as the hatched object in A, with the circular portion (touching the tB arrow) representing the DNA binding ETS domain. The hatched ellipse represents the non-ETS domain portions of PU. 1. Ets-1 protein is shown as the shaded object in A, with the circular portion (touching the ,uA arrow) representing the DNA binding ETS domain. The other domains of Ets-1 are depicted as the shaded ellipse. Note that although the ,uE3 site is shown in the figure without any bound protein, in these assays, an intact ,uE3 site is required for enhancer activation reflecting the use of an endogenous COS cell protein (Nelsen et al., 1993). Each line summarizes the results of transfection analyses in COS cells using a ,t70 dimer reporter plasmid described in the text, with expression vectors encoding wild type or mutant proteins, as indicated. (A) PU.1 and Ets-1. (B) PU.1A162 and Ets-1. (C) PU.1 and ETS(Ets-1). (D) PU.1 and the PU.l(Ets) chimera. (E) PU.1A162 and PU.l(Ets). (F) The Ets-l(PU) chimera and ETS(Ets-l). + and - in the column marked 'Activity' summarizes the results of co-transfection experiments with the appropriate combination of ETS proteins.
PU. 1 to activate the enhancer. These observations directly demonstrate that transcription activation domains identified by assaying the activity of a protein on multimerized binding sites need not correspond to domains of the proteins required to activate a complex, cell-specific regulatory sequence such as the one being assayed here. Furthermore, we show that the Ets-1 activation domain is not unique, and can be functionally substituted by residues from PU.1 that lie outside of the DNA binding domain. Lastly, our studies suggest that a critical requirement for activity is the placement of the activation domain at the gA, but not the ,uB, site of the enhancer. These results illustrate two forms of context dependence. First, all possible transcription activation domains that are bound to a complex regulatory element may not be utilized to achieve combinatorial specificity. Secondly, functional transcription activation domains may require appropriate positioning on the DNA.
4672
Some of our conclusions are based in part on the analysis of chimeric proteins. Because it is possible that creation of these novel molecules alters the structural features that are important for protein function, our interpretations have to be taken with this note of caution. However, we consider this to be an unlikely possibility, because in generating the chimeras reported here, we have swapped ETS domains that are likely to be structurally related. In doing this, we have therefore not needed to take into account variables that are often inherent in generating fusion proteins, such as the composition and length of the spacer that links two functional domains. Furthermore, there is increasing evidence for the structural similarity between different ETS domains. For example, recent studies of the ETS domains of Fli-1 and Ets-1 revealed similar secondary structural features (Donaldson et al., 1994; Liang et al., 1994) and our own analysis of PU. 1-DNA interactions show that the nucleotide and backbone contacts made by PU. 1 are very similar to those defined for Ets- 1 (Y.Rao and R.Sen, unpublished data), despite these ETS domains being the most divergent in the family (Wasylyk et al., 1993). We have previously shown that all three sites in the minimal enhancer are necessary for B cell-specific enhancer activity, suggesting that proteins binding to each of these sites are functionally important. What is the role of PU. 1 in this assembly that does not require a transcriptional activation domain? First, there is some precedence for the DNA binding domain of transcription factors to have weak transactivation potential (Hollenberg and Evans, 1988; Davis and Weintraub, 1992; Miner and Yamamoto, 1992; Pomerantz et al., 1992; Hagman et al., 1995) and perhaps the ETS domain of PU. 1 falls into this category. Our use of a regulatory element in which three different proteins come together to activate transcription may have accentuated the otherwise weak transactivation potential of this domain. Alternatively, the ETS domain of PU. 1 may interact with Ets- 1 bound at the ,uA site. This interaction can potentially alter the structure of either Ets- 1 or PU. 1 (or both) and thereby affect transactivation potential. Thirdly, PU.1 may function by altering DNA conformation that is necessary for appropriate juxtaposition of the three factors in a functional complex. Although the experiments described in this paper do not distinguish between these possibilities, we favor a combination of the latter two models based on several considerations. In support of the second model, ETS family proteins have been previously shown to direct protein-protein interactions that are templated by DNA (Hipskind et al., 1991; Thompson et al., 1991; Dalton and Treisman, 1992; Pongubala et al., 1992; Wotton et al., 1994; Giese et al., 1995). We have not observed significant cooperative DNA binding by Ets- 1 and PU. 1 to the enhancer, usually the method of choice for assaying such interactions. However, we have shown that PU. 1 and Ets- 1 can associate in vitro in a GST pull-down assay (Figure 4), as well as in a yeast two-hybrid assay (W.Dang and R.Sen, unpublished data). These observations suggest that the two proteins may interact after binding to DNA independently. The third model is supported by our recent experiments that show that PU.1 binding bends DNA and the PU.1 ETS domain is sufficient to induce this conformational change (Nikolajczyk et al., 1996). If the
Context dependent transactivation domains
only role of PU. 1 in the context of the .t enhancer is to bend DNA, this may explain the lack of requirement of its own transactivation domain. The distinct transactivation domains of PU.1 that are required to activate different regulatory sequences is reminiscent of the properties of the transcription factor LEF-I (Carlsson et al., 1993; Giese and Grosschedl, 1993). Unlike PU.l, LEF-1 does not activate transcription from a regulatory sequence composed of multimerized LEF-i binding sites. However, in the context of other proximal protein binding sites, LEF-i does transactivate the T cell receptor oc chain gene enhancer, and a transactivation domain different from its DNA binding domain is required for activity. The results with PU. 1 presented here offer a complementary view of context-dependent transcription activation that differs from that noted previously with LEF- 1. Our observations thus accentuate the diverse mechanisms that may be used to achieve combinatorial specificity. Another interesting aspect of our studies is the implication that simply bringing a transcriptional activation domain to the DNA is not sufficient for function; rather, the domain may require appropriate positioning on the DNA. This conclusion is based on two complementary observations. First, that the PU.1 activation domain can work when brought to the ptA site [as in the PU. I (Ets) chimera], but not at the RtB site (as in the full-length PU. l plus a truncated Ets- 1). Secondly, that the Ets- 1 activation domain works when brought to the iA site (as in fulllength Ets- 1 plus a truncated PU. 1), but not when brought to the pftB site [as in the Ets-l(PU) chimera plus a truncated Ets-l]. Both observations suggest that the transactivation domain has to be present on the tA site and cannot function from the pftB site. Why is the location of the transactivation domain important? We hypothesize that the minimal pt enhancer presents a combined transcription activation domain to the basal transcriptional machinery, composed of the activation domains of Ets-1 and a protein bound at the adjacent iE3 site. This is consistent with the ,uE3 site being necessary for activity of this enhancer fragment in both B and non-lymphoid cell lines, and the known strong transcription activation domain of TFE3, one of the ,uE3 binding proteins (Beckmann et al., 1990; Artandi et al., 1994). In this scenario, location of the transcription activation domain at the gB site may place it either too far away, or inappropriately positioned, to associate with the domain of the .tE3 binding protein to activate transcription. Combining effector domains of more than one DNA bound factor reveals another level of regulatory complexity.
Materials and methods Construction of mammalian and bacterial expression plasmids The PU. 1 expression vector (pEVRF-PU. 1) has been previously described (Nelsen et al.. 1993) and PU. I deletion mutants were constructed as follows: PU.1A33. NsiI (filled in with Klenow) to XbaI fragment from pBS-PU. I which contains the PU. I cDNA in the EcoRI site, was ligated into pEVRF2 (Matthias et al., 1989) cut with StinaI and XbclI: PU.1A69, Xnnl1-XbaI fragment from pBS-PU. I was ligated into pEVRFO cut with Snial and Xbal: PU. 1A97 TaiqI (filled in with Klenow) to Xbail fragment from pBS-PU. 1 uas ligated into pEVRF2 cut with SinaI and XbhIl PU.lA133. SaicI (treated with T4 polymerase) to Xbal fragment from pBS-PU.l was ligated into pEVRFO cut with SinatI and XbaI: PU.1A162.
the ETS domain of PU. I was amplified by PCR using, the oligonucleotides 5'-GGG GGA TCC CAC GGG GAG ACA-3' and 5'-GGG GAA TTC CTC GCC GCT GAA-3' (where the underlined codons correspond to amino acids 162 and 256). dioested with BainiHI and EcoRI (filled in with Klenow) and ligated into pEVRFO. cut with BamtiHI and Soial. The C-terminal segment of PU. I and the 3' untranslated region was added to this molecule by replacing a KpnI-Xbal fragment in this construct with a KpnI-Xba fragment from the parental pBS-PU.I plasmid; PU.1A200. Asp718 (filled in with Klenow) to XbaI fragment from pBS-PU.I was ligated into pEVRFO. cut with SinaI and XbaI. The GST-tagged fulllength PU.I(GST-PU.1) and ETS domain of PU.I [GST-ETS(PU.l)] were constructed as follows: The PU.1 cDNA was amplified by PCR using the oligonucleotides 5'-G GGT AGG CCT ATG TTA CAG GCG TGC AAA ATG-3' (where the underlined codon corresponds to the start codon of PU. 1) and 5'-CTG CAC GCT CTG CAG CTC TGT G-3'. The product of this reaction was digested with PstI. and ligated in a three-way ligation reaction into pGEX2T cut with Soiil and EcoRI, along with a 963 bp PstI-EcoRI fragment digested and purified from pBS-PU. 1. The 5' primer introduces three additional amino acids between the coding regions of GST and PU.1. The ETS domain of PU.l was amplified by PCR using the same oligonucleotides used to generate PU.1A162, digested with BamiiHI and EcoRI and ligated into pGEX2T cut with Ba,lnHI and EcoRI. The Ets- I expression vector contained the full-length gene cloned into the BamiiHI site of pEVRFO (pEVR-Ets-1) (Nelsen et al.. 1993). Expression vectors for Ets- 1 deletion mutants were constructed as follows: Ets1A167, SphlI (T4 polymerase) to XbaI fragment from pEVR-Ets-l was ligated into pEVRFO cut with Sinial and Xbal Ets-1A231. HincII-Xbal fragment from pEVR-Ets- I was ligated into pEVRF2 cut with SinaI and XbaI: Ets-1A286, TcoI (filled in with Klenow) to XbaI fragment from pEVR-Ets-1 was ligated into pEVRF2 cut with SinaI and XbaI: ETS(Ets-1). the ETS domain of Ets-l was amplified by PCR with the oligonucleotides 5'-CCT GGG CCC GGG GCC CTG GCT GGC TAC ACA-3' and 5'-CGG GAG TCG ACG CTC AGG GGT GTA TCC CAG CAG-3' (where the underlined codons correspond to amino acids 325 and 427) (Gunther et al.. 1990), digested with SinaI and SalI (filled in with Klenow) and ligated into pEVRF3S, cut with SinalI. The A nomenclature indicates that residues 1 to the number indicated were missing in a particular deletion mutant. The full-length Ets-l cDNA was inserted into the pET14 vector (pET14-Ets-l) by digesting the pEVREts-l plasmid with BamiiHI (filled in with Klenow) and ligated into the pET14 vector digested with Ba,nHI (filled in with Klenow). The general strategy for generating the chimeric proteins was to introduce unique Smial and Sail sites on either sides of the ETS domains of both the PU. I and Ets- I genes. while deleting the sequences encoding the ETS domains. These mutant genes will be referred to as PU.IAETS or Ets-IAPU respectively. DNA fragments with Soial-SalI ends and encoding the ETS domains of either proteins were obtained by PCR and cloned into the appropriate ETS domain deleted gene to generate the required fusion protein. PU. IAETS was generated using pEVRF-PU. I as the template, PU. I primers. PUl.l1 and PU1.2 and primers that anneal to pEVRF vector sequences outside the PU.1 coding region, EVRI and EVR2. The 5' and 3' ends of the PU. I gene were obtained by PCR amplification using PUL.1 and EVR1, and PU1.2 and EVR2, digested with BaimHI and EcoRI respectively, and cloned into pGEX2T (Pharmacia Biotech, Inc.) cut with BainiHI and EcoRI to generate pGEX 2T PU. 1AETS. The ETS domain of Ets-l. generated by using pEVRF-Ets-l as the template and primers ETS 1.1 and ETS 1.2, was digested with SmnaI and SaII, and cloned into pGEX 2T PU. 1AETS cut with the same enzymes. to produce pGEX PU.l(Ets). The mammalian expression vector pEVR PU.l(Ets) was generated by ligating a Bal,nHI-DraI fragment from pGEX PU. I (Ets) into pEVRFl ditested with BainHI and SmalI. Ets-IAPU was generated using pEVRF-Ets-I as the template, Ets-l primers. ETS1.3. and ETS1.4 and primers that anneal to pEVRF vector sequences outside the Ets-l coding region, EVR1 and EVR2. The 5' and 3' ends of the Ets-l gene were obtained by PCR amplification using ETS1.3 and EVRI. and ETSI.4 and EVR2. digested with Xbal and Ba,niHI respectively, and cloned into pGEM7Zf cut with Xbal and BamiHI to generate pGEM7Zf Ets-1AETS. The ETS domain of PU.1, generated by using pEVRF-PU.1 as the template and primers 5END and SALI. was digested with SmnaI and Sall, and cloned into pGEM7Zf EtsIAETS cut with the same enzymes, to produce pGEM7Zf Ets-1(PU). The mammalian expression vect)r pEVR Ets-1(PU) was generated by ligating a 380 bp B-o1,HI-XbAI fragment from the Ets-l1 cDNA and a 1 kb XbaI-BamnfHI fragment from pGEM7Zf Ets-l(PU) into pEVRFO cut with Bat?zHI. in a three-way ligation reaction. The bacterial expression
4673
B.Erman and R.Sen plasmid pET28 Ets- I (PU) was generated by ligating the BaniHI fragment from pEVR Ets-l(PU) into pET28a (Novagen, Inc.) cut with BamnHI. All expression plasmids were sequenced to ensure that the appropriate reading frame was maintained. The sequences of primers used for mutagenesis and generation of the ETS domains flanked by restriction sites were as follows: (EVRl) 5'-GGG GGA TCT TGG TGG CGT G-3'; (EVR2) 5'-CCC TGA AAA CTT TGC CCC CTC C-3'; (PUl.l) 5'-CCC GTG CAG AAG CCC GGG CCC AGG C-3'; (PUI.2) 5'-GGC CTG GCC GAG CGT CGA CTC CCG CC-3'; (ETS 1.3) 5'-GGC AGC CCC GGG AAT GAC AGG CTT GTC CTT G-3'; (ETS1.4) 5'-CTG AAC GTC GAC ACG CCA TGC TGG ATG-3'; (5END) 5'-GGG CCC GGG CTT CTG CAC GGG GAG-3'; (SALl ) 5'-CGG GAG TCG ACG CTC GGC CAG GCC CCC ACG GCC CAG CAC CTC GCC GCT GAA CTG-3'; (ETSl.l) 5'-CCT GGG CCC GGG GCC CTG GCT GGC TAC AC-3': (ETS 1.2) 5'-CGG GAG TCG ACG CTC AGG GGT GTA TCC CAG CAG-3'.
Construction of reporter plasmids The p70 dimer reporter was described previously (Nelsen et al., 1993). The p55 monomer reporter was constructed by ligating a 55 bp PstIBamHI fragment of the murine intronic p enhancer into the pSP72 plasmid digested with PstI and BamiiHI. A fragment containing the enhancer was digested from this plasmid with HindIll and Asp718 (filled in with Klenow), and ligated into the A56CAT enhancerless reporter plasmid digested with Sall (filled in with Klenow). The ltBgpE3B dimer reporter was constructed by ligation of two tandem repeats of the annealed oligonucleotides 5'-TCG ACA TTT GGG GAA GGA GGT CAT GTG GCA AGG CTA TTT GGG GGA GGG AAC-3' and 5'TCG AGT TCC CTT CCC CAA ATA GCC TTG CCA CAT GAC CTC CTT CCC CAA ATG-3' into the A56CAT enhancerless reporter plasmid digested with Sall (the l.B sites are underlined).
extracts), 70 mm NaCl, and 10% glycerol for 10 min at room temperature, and reactions were resolved on a 4% polyacrylamide gel. Wild-type, pA- or ,uB- probes [described in Nelsen et al. (1993)] were of equal specific activity. The PU.1 binding site probe from the SV40 enhancer was generated by annealing the synthetic oligonucleotides: 5'-GAT CCC TCT GAA AGA GGA ACT TGG-3' and 5'-GAT CCC AAG TTC CTC TTT CAG AGG-3'. The consensus Ets- I binding site probe was obtained by annealing the synthetic oligonucleotides 5'-TCG ACA AGG CAC TCA CTT CCG GCT TGG CCG-3' and 3'-CCG TGA GTG AAG GCC GAA CCG GCA GCT-5' (Nye et al., 1992). For Western blot analysis of Ets-I deletion mutants, 20 pg of whole cell extracts from transfected cells were separated by SDS-PAGE (12%), transferred to Hybond paper by electro-blotting and probed with antiEts-I antiserum (SC 350, 1:500) (Santa Cruz Biotechnology, Inc.). Proteins were visualized by chemiluminescent detection according to the manufacturer's specifications (Amersham Life Science, Inc.) For GST pull-down experiments, -8 pg of GST fusion proteins or GST protein were incubated with 20 ,l of glutathione-agarose (Molecular Probes, Inc.) beads (pre-equilibrated with a TTBS solution containing 0.2% BSA) in 600 p1 of this TTBS solution for 2 h at 4°C. The beads were centrifuged at 500 g for 2 min, washed once with 1 ml of TTBSBSA, and 2 pg of His-Ets-1 protein was added and incubated in 600 p1 of TTBS-BSA for 2 h at 4°C. The beads were washed three times with 1 ml of a TTBS solution that did not contain BSA, resuspended in 2X loading buffer (125 mM Tris-HCI pH 6.8, 20% glycerol, 2% SDS, 2% 2-mercaptoethanol, 10 pg/ml bromophenol blue) boiled for 3 min at 100°C, and separated by SDS-PAGE (12%). A Western analysis was performed as previously described using anti-Ets-1 serum. The presence of equivalent amounts of GST fusion proteins on the glutathione-agarose beads was determined by probing the same immunoblot with an antiGST antibody (SC-459 1:500) (Santa Cruz Biotechnology, Inc.).
Cell culture, transfections, and CAT assays COS cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% newborn serum, and 50 tg/ml each of penicillin and streptomycin. p70 dimer or p55 monomer enhancer containing CAT reporter plasmids (2 .tg) were co-transfected with PU. 1 and Ets-1 expression vectors (2 pg each) (or expression vectors containing deletion mutants of either gene, as shown) into COS cells using the calcium phosphate method as previously described (Nelsen et al., 1993). Forty to forty-eight hours after transfection, whole cell extracts were prepared by three rounds of freeze-thaw, and CAT enzyme activity was assayed as previously described (Nelsen et al., 1993) or the level of CAT protein in the extracts were determined by a CAT ELISA Kit (Boehringer Mannheim, Corp.) according to manufacturer's instructions. Briefly, 10 or 100 pg of extracts from p70 reporter transfected or p55 reporter transfected cells respectively were incubated in the wells of a 96-well plate coated with anti-CAT antibody at 37°C for h. The wells were washed five times with the supplied wash buffer, incubated with an antiCAT-DIG antibody at 37°C for 1 h, washed again, and incubated with an anti-DIG-POD antibody at 37°C for 1 h, and assayed with the supplied substrate solution at room temperature from 10 min to I h. In transfections that did not contain one or both transactivators, the total DNA was held constant at 6 .tg by including the empty expression vector, pEVRF2.
In vitro protein expression, mobility shift, Western and GST pull-down assays The PU.1(Ets) chimeric protein was expressed as a GST fusion protein [GST-PU. I(Ets)] from the pGEX2T PU.1(Ets) plasmid and purified from
bacterial extracts as described by the manufacturer (Pharmacia Biotech, Inc.). Similarly, the full-length PU. 1 protein (GST-PU. 1), the ETS domain of PU.1 [GST-ETS(PU.1)] and the GST-CrkII protein were expressed as GST fusion proteins from the pGEXPU. 1, pGEXETS(PU. I) and pGEXCrkII plasmids respectively, and purified in a similar fashion. The GST protein contained the 219 amino acids encoded in the pGEX2T plasmid and was expressed and purified similarly. The Ets- 1 (PU) chimeric protein was expressed as a hexa-histidine tagged protein [His-Ets-l(PU)] from the pET28 Ets- 1 (PU) plasmid and purified from bacterial extracts by nickel affinity chromatography, as described by the manufacturer (Novagen, Inc.). The full-length Ets-1 protein was expressed as a hexahistidine tagged protein (His-Ets-1) from the pET14-Ets-l plasmid and purified in a similar fashion. For mobility shift experiments, either bacterially expressed and purified chimeric proteins, or 20 pg of transfected whole cell extracts were incubated with 32P-labeled oligonucleotide DNA probes (10-20 000 c.p.m.) in the presence of 25 ng poly(dl-dC).poly(dI-dC) (I p1g for
4674
Acknowledgements We thank M.Cortes, B.Nikolajczyk, L.Zeng, H.Ishii, E.Rao, W.Dang and G.Tian for discussions, assistance with plasmid constructions and gifts of purified proteins and Drs M.Rosbash, S.Smale, N.Speck, M.Baron and B.Nikolajczyk for critical comments on the manuscript. The pGEXCrkIl plasmid was a gift of Dr Ruibao Ren. Supported by grants from the NIH(GM38925) to R.S.
References Adams,J.M., Harris,A.W., Pinkert,C.A., Corcoran,L.M., Alexander,W.S., Cory,S., Palmiter,R.D. and Brinster,R.L. (1985) The c-myc oncogene driven by immunoglobulin enhancers induces lymphoid malignancy in transgenic mice. Nature, 318, 533-538. Alt,F.W., Blackwell,T.K., DePinho,R.A., Reith,M.G. and Yancopoulos, G.D. (1986) Regulation of genome rearrangement events during lymphocyte differentiation. Immunol. Rev., 89, 5-30. Artandi,S.E., Cooper,C., Shrivastava,A. and Calame,K. (1994) The basic helix-loop-helix-zipper domain of TFE3 mediates enhancer-promoter interaction. Mol. Cell. Biol., 14, 7704-7716. Beckmann,H., SuL. and Kadesch,T. (1990) TFE3: a helix-loop-helix protein that activates transcription through the immunoglobulin enhancer pE3 motif. Genes Des:, 4, 167-179. Blackwell,T.K., Huang,J., Ma,A., Kretzner,L., Alt,F.W., Eisenman,R.N. and Weintraub,H. (1993) Binding of myc proteins to canonical and noncanonical DNA sequences. Mol. Cell. Biol., 13, 5216-5224. Carlsson,P., Waterman,M.L. and Jones,K.A. (1993) The hLEF/TCF-la HMG protein contains a context-dependent transcriptional activation domain that induces the TCRoc enhancer in T cells. Genes Dev., 7, 2418-2430. Chen,J., Young,F., Bottaro,A., Stewart,V., Smith,R.K. and Alt,F. (1993) Mutations of the intronic IgH enhancer and its flanking sequences differentially affect accessibility of the JH locus. EMBO J., 12, 4635-4645. Dalton,S. and Treisman,R. (1992) Characterization of SAP-1, a protein recruited by serum response factor to the c-fos serum response element. Cell, 68, 597-612. Davis,R.L. and Weintraub,H. (1992) Acquisition of myogenic specificity by replacement of three amino acid residues from MyoD into E12. Science, 256, 1027-1030.
Context dependent transactivation domains
Donaldson.L.W., Petersen.J.M.. Gra-ves,B.J. and Mclntosh,L.P. (1994) Secondary structure of the ETS domain places murine Ets- I in the superfamily of winged helix-turn-helix DNA-binding proteins. Biochemistry. 33. 13509-13516.
Gegonne.A.. Punyammalee.B.. Rabault.B., Bosselut.R.. Seneca,S.. Crabeel,M. and Ghysdael,J. (1992) Analysis of the DNA binding and transcriptional activation properties of the Ets- I oncoprotein. Nel Biologist, 4. 512-519. Giese,K. and Grosschedl,R. (1993) LEF-I contains an activation domain that stimulates transcription only in a specific context of factor-binding sites.
EMBO J., 12. 4667-4676.
Giese.K.. Kingsley.C.. Kirshner.J.R. and Grosschedl.R. (1995) Assembly and function of a TCRa enhancer complex is dependent on LEF-l induced DNA bending and multiple protein-protein interactions.
Genies Del:. 9. 995-1008. Gregor,P.D., Sawadogo.M. and Roeder.R.G. (1990).
The adenovirus major late transcription factor USF is a member of the helix-loophelix group of regulatory proteins and binds to DNA as a dimer. Genies Deic.. 4. 1730-1740. Gunther.C.V., Nye.J.A.. Bryner.R.S. and Graves,B. (1990) Sequencespecific DNA binding of the proto-oncoprotein ets- 1 defines a transactivational activator sequence within the long terminal repeat of the moloney murine sarcoma virus. Genies De:. 4, 667-679. Hagemeier,C., Bannister.A.J., Cook,A. and Kouzarides,T. (1993) The activation domain of transcription factor PU. I binds the retinablastoma (RB) protein and the transcription factor TFIID in vitro: RB shows sequence similarity to TFIID and TFIIB. Proc. Natl Acadl. Sci. USA. 90. 1580-1584. Hagman.J. and Grosschedl,R. (1992) An inhibitory carobxyl-terminal domain in Ets- I and Ets-2 mediates differential binding of ETS family factors to promoter sequences of the imib- 1 gene. Proc. Ncatl Acad. Sci. USA. 89. 8889-8893. Hagman,J.. Gutch.M.J.. Lin,H. and Grosschedl.R. (1995) EBF contains a novel zinc coordination motif and multiple dimerization and transcriptional activation domains. EMBO J.. 14. 2907-2916. Hipskind,R.A., Rao.V.N.. Mueller,C.G.F., Reddy.E.S.P. and Nordheim.A. (1991) Ets-related protein Elk-I is homologous to the c-fos regulatory factor p62TCF. Nature. 354, 531-534.
Ho,l.. Bhat.N.K.. Gottschalk,L.R., Lindsten.T.. Thompson,C.B., Papas,T.S. and Leiden,J.M. (1990) Sequence specific binding of human Ets-l to the T cell receptor ax gene enhancer. Science. 250. 814-818. Hollenberg.S.M. and Evans.R.M. (1988) Multiple and cooperative tramisactivation domains of the human glucocorticoid receptor. Cell. 55. 899-906.
Kadesch,T. (1992) Helix-loop-helix proteins in the regulation of immunoglobulin gene transcription. Inoninal. Toclday, 13, 31-36. Klemsz.M.J., McKercher,S.R.. Celada.A., Be\veren,C.V. and Maki.R.A. (1990) The macrophage and B cell-specific transcription factor PU.I is related to the ets oncogene. Cell. 61. 113-124.
Kominato,Y., Galson.D.L., Waterman.W.R., Webb,A.C. and Auron.P.E. (1995) Monocyte expression of the human prointerleukin IB gene (ILIB) is dependent on promoter sequences which bind the hematopoietic transcription factor Spi-1/PU.1. Mol. Cell. Biol.. 15. 58-68.
Langdon.W.Y., Harris.A.W., Cory,S. and Adams.J.M. (1986) The c-omyc oncogene perturbs B lymphocyte development in Ep-mvc transeenic mice. Cell. 47. 11-18.
Liang.H., Olejniczak.E.T., Mao.X., Nettesheim.D.G..
Yu.L.. Thompson, C.B. and Fesik,S.W. (1994) The secondary structure of the ets domain of human Fli- i resembles that of the helix-turn-helix DNA-binding motif of the Eschzerichiai caoli catabolite gene activator protein. Pr-oc. Naitl Acacl. Sci. USA. 91. 11655-11659. Libermann.T.A. and Baltimore,D. (1990) Transcriptional regulation of immunoglobulin gene expression. Mol. A.spects Cell. Regul.. 6. 399421. Libermann,T.A. and Baltimore.D. (1993) iT. a pre-B-cell-specific enhancer element in the imimlunoglobulin heavy-chain enhancer. Mol. Cell. Biol.. 13. 5957-5969. Lim,F., Kraut,N., Frampton,J. and GrafT. ( 1992) DNA binding by c-EtsI but not v-Ets is repressed by an intramolecular mechanism. EMBO J.. 11, 643-652. Lopez.M.. Oettgen.P.. Akbarali.Y., Dendorfer.U. and Libermann.T.A. (1994) ERP. a new member of the ets transcription factor/oncoprotein family: cloning, characterization, and differential expression during B-lymphocyte development. Mol. Cell. Biol.. 14. 3292-3309. Miner.J.N. and Yamamoto.K.R. (1992) The basic region of AP- I specifies
glucocorticoid receptor activity at a composite response element. Genies Dec. 6, 2491-2501. Nelsen.B. and Sen.R. (1992) Reaulation of immunoglobulin gene transcription. Itit. Rem: Cy tol.. 133. 121-149. Nelsen.B.. Kadesch.T. and Sen.R. (1990) Complex regulation of the immunoglobulin p heavy-chain gene enhancer: pB. a new determinant of enhancer function. Mol. Cell. Biol.. 10, 3145-3154. Nelsen,B., Tian,G.. Erman.B.. Gregoire.J.. Maki.R.. Graves.B. and Sen,R. (1993) Regulation of lymphoid-specific immunoglobulin , heavy chain gene enhancer by ETS-domain proteins. Science. 261, 82-86. Nikolajczyk.B.S.. Nelsen.B. and Sen.R. (1996) Precise alignment of sites required for p enhancer activation in B cells. Mol. Cell. Biol.. 16, in press. Nye.J.A.. TerseniJ.. Gunther.C.V.. Jonsen.M.D. and Graves.B.J. (1992) Interaction of murine Ets-1 with GGA-binding sites establishes the ETS domain as a new DNA-binding motif. Genies Dev:. 6. 975-990. Oltz.E.M., AltF.W., Lin.W., Chen.J., Taccioli,G.. Desiderio,S. and Rathbun,G. (1993) A V(D)J recombinase-inducible B-cell line: role of transcriptional enhancer elements in directing V(D)J recombination. Mol. Cell. Biol.. 13. 6223-6230. Pomerantz.J.L.. Kristie,T.M. and Sharp.P.A. (1992) Recognition of the surface of a homeo domain protein. Genzes Dev:. 6. 2047-2057. Pongubala.J.M.R.. Nagulapalli.S.. Klemsz,M.J.. McKercher.S.R., Maki. R.A. and Atchison,M.L. (1992) PU.1 recruits a second nuclear factor to a site important for immunoglobulin K 3' enhancer activity. Mol. Cell. Biol.. 12. 368-378. Pongubala.J.M.R.. VanBeveren.C.. Nagulapalli.S.. Klemsz.M.J.. McKercher,S.R., Maki.R. and Atchison.M.L. (1993) Effect of PU.1 phosphorylation on interaction with NF-EM5 and transcription activation. Scienzce. 259, 1622-1625. Reik.W.. Williams,G., Barton.S.. Norris,M., Neuberger,M. and Surani,M.A. (1987) Provision of the immunoclobulin heavy chain enhancer downstream of a test gene is sufficient to confer lymphoidspecific expression in transgenic mice. Elit: J. Ilii1/7i00ol., 17. 465-469. Rivera.R.R.. Stuiver.M.H.. Steenbergen.R. and Murre.C. (1993) Ets proteins: New factors that regulate immunoglobulin heavy chain gene expression. Mol. Cell. Biol.. 13. 7163-7169. Roman.C., Matera.A.G.. Cooper.C.. Artandi,S., Blain,S.. Ward.D.C., anld Calame.K. (1992). mTFE3. an X-linked transcriptional activator containing basic helix-loop-helix and zipper domains, utilizes the zipper to stabilize both DNA binding and multimerization. Mol. Cell. Biol.. 12. 817-827. Schmidt.E.V., Pattengale.P.K.. Weir.L. and Leder.P. (1988) Transgenic mice bearing the human c-nmvc gene activated by an immunoglobulin enhancer: A pre-B-cell lymphoma model. Proc. Natil Acad. Sci. USA. 85, 6047-6051. Schneikert.J., Lutz.Y. and Wasylyk.B. (1992) Two independent activation domains in c-Ets-1 and c-Ets-2 located in non-conserved sequences of the ets aene family. Oncogene. 7. 249-256. Serwe.M. and Sablitzky.F. (1993) V(D)J recombination in B cells is impaired but not blocked by tarreted deletion of the immunoglobulin heavy chain intron enhancer. EMBO J.. 12. 2321-2327. Shin.M.K. and Koshland,M.E. (1993) Ets-related protein PU.1 regulates expression of the immunoglobulin J-chain gene through a novel Etsbinding element. Genies Devc, 7, 2006-2015. Thompson.C.C., Brown.T.A. and McKnight,S.L. (1991) Convergence of Ets- and notch-related structural motifs in a heteromeric DNA binding complex. Scienice. 253. 762-768. Wasylyk.B.. Hahn.S.L. and Giovane.A. (1993) The Ets family of transcription factors. Em,,: J. Biolein.. 211. 7-18. Wasylyk.C., Kerckaert.J. and Wasylyk,B. (1992) A novel modulator domain of Ets transcription factors. Genies Dec:, 6. 965-974. Watson.D.K.. McWilliams.M.J.. Lapis.P., Lautenberger.J.A.. Sc\veinfest,C.W. and Papas.T. (1988) Mammalian ets-I and ets-2 genes encode highly conserved proteins. Proc. Natl Acad. Sci. USA. 85. 7862-7866. Wotton,D.. Ghysdael,J.. Wang.S.. Speck.N. and Owen.M. (1994) Cooperative binding of Ets-1 and core binding factor to DNA. Mol. Cell. Biol.. 14, 840-850. Yancopoulos,G.D. and AltF.W. (1985) Developmentally controlled and tissue-specific expression of unrearranged VH gene segments. Cell. 40, 271-281.
Receimedl on Deceinber 21. 1995: remisedl
oii
41Ail 22. 1996
4675
