Cell, Vol. 60, 461-472,

February 9, 1990, Copyright

0 1990 by Cell Press

A Novel Octamer Binding Transcription Factor Is Differentially Expressed in Mouse Embryonic Cells Koji Okamoto, Hitoshi Okazawa, Akihiko Masaharu Sakai,’ Masami Muramatsu, and Hiroshi Hamada Department of Biochemistry Faculty of Medicine University of Tokyo 7-3-l Hongo, Bunkyo-ku Tokyo 113 Japan

Okuda,

Summary We have identified a novel octamer binding factor (Ott-3) in P19 embryonal carcinoma cells. Ott-3, which recognizes the typical octamer motif (ATTTGCAT) as well as the AT-rich sequence TTAAAATTCA, is present in P19 stem cells but disappears when the cells are induced to differentiate by retinoic acid (RA). Cloned cDNA corresponding to Ott-3 encodes a protein of 377 amino acids. Ott-3 has a conserved POU domain, but the remaining part is distinct from other POU domaincontaining proteins such as Ott-1 and Ott-2. mFlNA of 1.5 kb coding for Ott-3 is abundant in P19 stem cells but is dramatically repressed during RMnduced differentiation. Repression of the 1.5 kb mRNA is rapid and specific to RA. In mouse, act-3 mRNA is undetectable in all the adult organs examined. The N-terminal proline-rich region of Ott-3, when fused to the DNA binding domain of c&n, functions as a transcriptional activating domain. We suggest that Ott-3 is a novel octamer binding transcription factor that is developmentally regulated during mouse embryogenesis. Introduction Embryogenesis and morphogenesis of multicellular organisms are probably controlled by a group of regulatory genes which form a cascade with a determined hierarchy. Although such regulatory genes have been identified in Drosophila, their identities are not established in mammalian systems. In Drosophila many of the genes that are involved in embryogenesis appear to code for DNA binding, transcriptional factors (reviewed in Levine and Hoey, 1988). Thus, the way gene X controls gene Y is often such that gene X encodes a sequence-specific DNA binding factor that activates or represses transcription of gene Y, although other types of control mechanisms, such as one at the RNA splicing level (reviewed in Bandziulis et al., 1989), are known to operate. It is therefore quite reasonable to assume that the genes that control mammalian embryogenesis are those that encode transcriptional factors, which appear in a stage-specific and/or cell lineage-specific fashion. ’ Present Address: Department of Biochemistry, Hokkaido University School of Medicine, Kita-ku Sapporo 060, Japan.

With these viewpoints, we began to search for transacting factors that appear specifically at the embryonic stage. Instead of using mouse embryos, we employed embryonal carcinoma (EC) cell lines, since EC cells resemble embryonic stem cells (inner cell mass cells, in particular) in many ways and have been used as a model system of embryonic stem cell differentiation. As the first step we have previously (Bhat et al., 1988) isolated mouse genomic fragments containing enhancers specific to embryonic stem cells by using the enhancer-trap tagging method established earlier (Hamada, 1986a, 1986b). We next analyzed the stem cell-specific enhancers in more detail (Okamoto et al., submitted). The enhancers were found to be derived from endogenous retrotransposon elements, which have been shown to be expressed at an early embryonic stage (Brulet et al., 1983). The enhancers were active in EC stem cells but inactive when the cells were induced to differentiate. Nucleotide sequences critical for the enhancing activity were determined for two enhancer elements, which revealed that both enhancers contain two common sequences: one is the typical octamer motif (ATTTGCAT or ATTAGCAT), the other is an ATrich sequence (TTAAAATTCA) that resembles the recognition sequence of Drosophila En protein. In this report, we searched for DNA binding factors interacting with these enhancer sequences and found novel octamer binding factors in EC cells. cDNA corresponding to one of those factors was cloned. Its structure, binding specificity, activity as a transposon factor, and expression pattern were studied. Results Identification of Novel Octamer Binding Factors By using the enhancer-trap tagging method (Hamada, 1986a, 1986b), we previously selected mouse DNA fragments from the mouse genome that contained embryonic stem cell-specific enhancers (Bhat et al., 1988). Further analysis of these enhancers confirmed their specificities; they were active in all the EC stem cells tested but inactive when the cells were induced to differentiate. For two of these enhancers (El and E2 in Figure l), nucleotide sequences critical for the enhancing activity were determined by deletion analysis (Okamoto et al., submitted). As summarized in Figure 1, two sequences contain the enhancer activity; these are the only sequences common to El and E2. One of them is the octamer motif sequence, ATTTGCAT in E2 and ATTAGCAT in El (designated OCTA). The other sequence is an AT-rich sequence, TTAAAATTCA, which is 60% identical to the Drosophila En recognition sequence (designated En-like sequence). We first searched for DNA binding proteins that recognize nucleotide sequences within the El enhancer domain, by DNAase I footprint assay. Nuclear extract was prepared from undifferentiated (D-) and differentiated (D+) P19 cells and used for the assay. When the 300 bp region encompassing the enhancer domain was sur-

Cell 462

A El

E2

B : OCTA El

C:En-like El

E2 en

TTCTTTGI\GC piisiiq GCTGAGAOCA . .... . ATTTAATTGA

Figure 1. Structure of Two Embryonic El and E2

Stem Cell-Specific

Enhancers,

El and E2 are derived from mouse genomic DNA fragments (clones 015 and 052) we previously isolated by the enhancer-trap tagging method (Bhat et al., 1966). (A) DNA regions containing the enhancer activity were determined by deletion analyses (Okamoto et al., submitted) and are shown here. Two sequences common to El and E2 are indicated by an open box (OCTA) and a solid box (En-like). (B) Nucleotide sequences of OCTA in El and E2. together with OCTA found in the immunoglobulin heavy chain enhancer and OCTA in the histone 2b gene promoter, are presented. The octamer motif in each sequence is boxed. DNA regions of El and E2 that were protected in DNAase I footprint assays (Figure 2) are indicated by the lines. (C) Nucleotide sequences of the En-like motif found in El and E2 are boxed. DNA regions protected in DNAase I footprint assays are indicated by the lines. The consensus sequence for the Drosophila En protein binding sites is shown at the bottom. The En-like motif in El and E2 is identical to the consensus sequence at 6 of the 10 bases.

veyed, protection was observed in the two conserved regions described above; i.e., the 18 bp region encompassing OCTA (Figure 2) and the 23 bp region containing the En-like sequence. Similar, if not identical, protection patterns were observed with D- and D+ extracts. This protection seems to reflect specific DNA-protein interaction, since a homologous competitor abolished the protection (data not shown), and both DNA strands of the same region were protected (Figure 2). Similar protection was also observed in the OCTA and En-like regions of E2 (data not shown). At that time, two kinds of octamer binding factors were known: Ott-1 and Ott-2. While Ott-1 is ubiquitous, Ott-2 is specific to 6 cell (Fletcher et al., 1987; Scheidereit et al., 1987). To learn which type or what kind of octamer binding factor(s) was present in the EC cell extracts, the same set of extracts were next subjected to gel-shift assays. Two oli-

78

910

Figure 2. DNAase I Footprint Assay The 300 bp restriction fragment containing El (shown in Figure 1A) was subjected to DNAase I footprint assay. Two protected regions are indicated. TOP and BOTTOM are the strands shown at the top and the bottom, respectively, in Figure 1B. The 3’ends of both strands were labeled by the end-filling reaction. Lanes 1, 6, 7, and IO: no proteins added. Lanes 2, 3, and 6: nuclear extract from undifferentiated (D-) P19 cells. Lanes 4, 5, and 9: nuclear extract from differentiated (D+) P19 cells. Lanes 2 and 4: 5 pg of the indicated extract. Lanes 3, 5, 8, and 9: 10 pg of the indicated extract.

gonucleotides (OCTA26 and EN26, corresponding to the OCTA and En-like regions of El, respectively) were used as probes. To our surprise, the OCTA26 probe detected a novel octamer binding factor (Ott-3) as well as Ott-1 in D- P19 cells (Figure 3, right panel). The mobility of Ott-3 in the gel-shift assay was higher than that of Ott-1 or Oct2. Ott-3 was also detected by the EN26 probe (Figure 3, left panel), which was verified by the following competition experiments (Figure 4A). Binding of Ott-3 to OCTA26 was competed by EN26 as well as by OCTA26 itself. Furthermore, binding of Ott-3 to EN26 was competed by OCTA26 and by EN26. It should be noted that Ott-3 (and also Ott-1) shows a higher affinity to OCTA26 than to EN26, since OCTA26 competed the binding more efficiently than EN26 (Figure 4A). Therefore, Ott-3 should recognize not only OCTA26 but also EN26. When the nuclear extracts from D+ P19 cells were examined, Ott-3 was not detected; instead another binding factor (Ott-4) was observed (see Figure 3). Competition experiments (Figure 48) indicated that Ott-4 recognizes both OCTA26 and EN26. Compared with Ott-1, Ott-4 appears to have a higher affinity to EN26. The results described so far indicate two interesting points. First, novel octamer binding factors are present in

$%3Embryonic

Octamer-Binding

Factor

D- D+

-0CTl

0CTl-b

OCTl--)

+OCT4

OCT4-b .-OCT3-.

OCTf)

I

PROBE : PROBE : Figure 3. Detection Assays

12

EN2 6 of Novel Octamer

-

3

OCTAP 6

EN26

J

4

OCTAZ 6 Binding

Factors by Gel-Shift

Two double-stranded oligonucleotides, OCTA26 and EN26 were end labeled and used as probes. Binding factors detected are indicated by the arrows. D- and D+: see text.

0CTl-b OCT4--)

P19 EC cells. Second, these new factors are differentially expressed during stem cell differentiation. I

PROBE :

Structure of Ott-3, a New Member of the POU Family cDNAs for Ott-I and Ott-2 have been recently cloned (Sturm et al., 1988; Scheidereit et al., 1988; Muller et al., 1988; Ko et al., 1988). The structure of Ott-1 and Ott-2 indicated the presence of the POU domain (Herr et al., 1988) an ~150 amino acid sequence conserved among Pit-l (Ingraham et al., 1988; Bodner et al., 1988) uric-86 (Finney et al., 1988) Ott-I, and Ott-2. The POU domain consists of three subdomains: POU-specific A and B subdomains and the homeo-like subdomain. Since the POU domain appears to determine the binding specificity (Sturm et al., 1988), we presumed that Ott-3 and Ott-4 would also contain the POU domain. Therefore, we chemically synthesized two oligonucleotide probes: the 69 mer and 63 mer, corresponding to the POU-specific B subdomain and the homeo-like subdomain, respectively. When the cDNA libraries prepared from D- P19 cells and D+ P19 cells were screened by hybridization with those probes, positive clones were obtained only from the DP19 library. More than 20 positive clones were purified and characterized. Restriction mapping and partial sequencing indicated that all (except for one clone coding for Ott-1) were derived from the same mRNA; therefore, the one with the largest insert (Xl) was further characterized in detail. We first examined binding activity of the protein encoded in Xl. The nucleotide sequence of the Xl insert indicated that it contained the entire coding sequence including the initiation codon (Figure 58). Therefore, sense RNA was transcribed in vitro from the hC1 insert, which was subsequently translated in a rabbit reticulocyte translation system. The in vitro translated products were analyzed by a gel-shift assay with OCTA26 as probe (Figure 6A). The in vitro translation products reproducibly gave

I

OCTA26

Figure 4. Ott-3, Ott-4, and O&l like Sequences

EN26

Can Recognize

Both OCKA and En-

Binding of O&l, Ott-3, and Oct4 to OCTA26 or EN26 was competed by various unlabeled competitors. Amounts of competitors relative to that of the indicated probe are shown in fold (x). (A) PI9 D- nuclear extracts were used. (B) PI9 D+ nuclear extracts were used.

rise to a rather broad band that nearly, but not exactly, comigrated with the in vivo Ott-3. The discrepancy in the mobility between the in vitro products and the in vivo Ott-3 may be due to the lack of or incomplete protein modification in the reticulocyte lysates. Similar discrepancy has been reported for other DNA binding proteins such as Oct1 (Sturm et al., 1989). The in vitro products showed binding specificity similar to that of in vivo Ott-3; their binding to OClA26 was competed by unlabeled OcTA26 and EN26, and the former competitor was more efficient than the latter competitor (Figure 6A). To confirm that the cloned cDNA encodes Ott-3, we established several PlS-derived cell lines transformed with a vector that can express antisense RNA of hC1 (detailed properties of those cell lines will be described elsewhere). In parental P19 cells, binding activities of Ott-1 and Ott-3 are almost equal. On the other hand, the cell lines transformed with the antisense vector showed a greatly reduced level of Ott-3 relative to that of Ott-I. From these results, it is clear that the cloned cDNA encodes Ott-3. The nucleotide sequence of the EcoRl insert cloned in Xl revealed an open reading frame plus 5’ and 3’ untranslated regions. A putative initiation codon is found at the +l site (see Figure 58) which would make Ott-3 a protein of 377 amino acid residues. As expected, the conserved POU domain is found in the region from amino

Cell 464

A

Figure 5. Nucleotide Sequence of act-3 cDNA

EcoRI

Pst

B GTGAGCCGTCTTTCCACCAGGCCCCCGGCTCGGGGTGCCCACCTTCCCC 0045

001

ATGGCTGGACACCTGGCTTCAGACTTCGCCTCCTCACCCCCACCAGGTGGGGGTGATGGG 0109 MA G H LA S D F A S S LEE G G G Y G

TCAGCAGGGCTGGAGCCGGGCTGGGTGGATTCTCGAACCTGGCCT WVDSRTWLSFQG& 021 SAGLEEG

(A) Structure of 01%3 cDNA is shown. The entire coding sequence is indicated by the box. Three POU subdomains (A, B, and f-f) are also indicated. (B) Nucleotide sequence of the entrre coding region plus the 3’ untranslated region and a part of the 5’ untranslated region are shown. The nucleotide sequence is numbered on the right. 221 contains nucleotide sequences from 1 to 1315. The remaining part of the sequences (i.e., from nucleotide 1316 to the poly(A) tail) is derived from other overlapping cDNA clones, The predicted amino acid sequence is presented under the nucleotide sequence. The amino acid residues are numbered on the left. The three POU subdomains are underlined. Proline residues are underlined with bold. See text for details.

3169

CCAGGTGGGCCTGGAATCGGACCAGGCTCAGAGGTATTGGGGATCTCCCCATGTCCGCCC 0223 041 ~GG&GIGEGSEVLGISEC~& GCATACGAGTTCTGCGGAGGGATGGCATACTGTGGACCTCAGGTTGGACTGGGCCTAGTC 0289 V G L G L V 061 A Y E F C G G M A Y C G EQ CCCCAAGTTGGCGTGGAGACTTTGCAGCCTGAGGGCCAGGCAGGAGCACGAGTGGAAAGC 0349 081 EQVGVETLQEEGQAGARVES AACTCAGAGGGAACCTCCTCTGAGCCCTGTGCCGACCGCCCCAATGCCGTGAAGTTGGAG 0405 101 NSEGTSSEeCADR&NAVKlE AAGGTGGRACCAACTCCCGAGGAGTCCCAGGACATGAAAGCCCTGCAGAAGGAGCTAGAA 0465 SQDMKALOKELE 121 KVELTPEE CAGTTTGCCAAGCTGCTGAAGCAGAAGAGGATCACCTTGGGGTACACCCAGGCCGACGTG 0529 141 DFAKLLKQKBITLGYTOADV GGGCTCACCCTGGGCGTTCTCTTTGGAAAGGTGTTCAGCCAGACCACCATCTGTCGCTTC 0589 161 GLTLGVLFGKVFSOTTICKL GAGGCCTTGCAGCTCAGCCTTAAGAACATGTGTAAGCTGCGGCCCCTGCTGGAGAAGTGG 0645 181 E A L 0 I. S I. B NMC K L R P L L E K W GTGGAGGAAGCCGACRACAATGAGAACCTTCAGGAGATATGCAAATCGGAGACCCTGGTG 0709 201 "E E A QN N E N L Q E I C K S E T 5 V CAGGCCCGGAAGAGAAAGCGAArTAGCATTGAGCATTGAG~CCGTGTGAGGTGGAGTCT(,GA~;ACC 0763 221QAR&RKRTSIENR"RWSLET ATGTTTCTGAAGTGCCCGAAGCCCTCCCTACAGCAGATCACTCACATCGCCAATCAGCTT 0829 241MFLKCPK~SLOOITHIANOL GGGCTAGAGAAGGATGTGGTTCGAGTATGGTTCTGTAACCGGCGCCAGAAGGGCAAAAGA 0889 261GLEKDVVRVWFCNRROKGKR TCAAGTATTGAGTATTCCCACGAGRAGRGTATGAGTATGAGGCTACAGGACACCTTTCCCAGGGG 0949 281~IEYYQQEEYEATGH L S Q G GGGCTGTATCCTTTCCTCTGCCCCCAGGTCCCCACTTTGGCACCCCAGGCTATGGAAGCC 1009 AM E A 301GLYeFLCgQVETLAeQ CCCACTTCACCACACTCTACTCAGTCCCTTTTCCTGAGGGCGAGGCCTTTCCCTC~GTTC 1069 321~TS~HSTQSLFLRAR~FgLF 341g

CCGTCACTGCTCTGGGCTCTCCCATGCATTCAllACTGAGGCACCAGCCCTCCCTGGGGAT 1129 S L L W A L & C IQTEAgALgGD

GCTGTGAGCCAAGGCAAGGGAGGTAGACAAGAGAACC'rGGAGCTTTGGGGTTAAATTCTT ;189 I. w G 361AVSQGKGGRQENLE TTACTGAGGAGGGATTAAAAGCACAACAGGGGTGGGGGGTGGGATGGGGAAAGAAGCTCA 1249 GTGATGCTGTTGATCAGGAGCCTGGCCTGTCTGTCACTCATCATTTTG'~TCTTAAATAAA 1309 1338 GACTGGACACACAGTAAAAAAAAAAAAAA

acid residue 132 to 279 (see Figures 5B and 7). Like other POU domain-containing proteins such as Ott-1 and Ott-2, the POU domain of Ott-3 consists of three subdomains: a POU-specific A subdomain of 33 amino acids, a POUspecific subdomain 6 of 33 amino acids, and a homeo-like

subdomain of 60 amino acids. Subdomains A and B show 60% and 67% homology, respectively, to the corresponding regions of Ott-1 and Ott-2 (Figure 7). The homeobox subdomain of Ott-3 shares identity with the homeobox subdomains of Ott-1, Ott-2, Pit-l, and uric-86 at 35 of 60,

An Embryonic 465

Octamer-Binding

Factor

Eve, Antp, and Dfd at 20 of 60, 18 of 60, and 18 of 59 residues, respectively. Therefore, Ott-3 is more closely related to the POU family than to other homeobox-containing proteins. The remaining part of Ott-3, on the other hand, shows no significant sequence homology to any of the homeoboxcontaining genes including o&l, act-2, pit-l, and uric-86. Unlike Ott-1 and Ott-2, Ott-3 does not possess a glutamine-rich or a serine- and threonine-rich region. However, it should be noted that the N-terminal region of Ott-3 is very rich in proline; of residues 13 to 60, 23% are proline. The C-terminal region of Ott-3 (residues 304 to 358) is also rich in proline (22%). A proline-rich region is known to be one of the structural features that characterizes transcriptional activation domains (Mitchell and Tjian, 1989; Mermod et al., 1989). The central portion of Ott-3 is very rich in both basic and acidic amino acids. From residue 91 to 146,230~ (13 of 56) of the residues are either arginine or lysine, and 20% (11 of 56) of the residues are acidic residues. The region between residues 178 and 223 is also rich in charged amino acids (17 of 46). Other notable features such as the leucine zipper (Landschulz et al., 1988a) and the zinc-finger domain are not found in Ott-3.

A

OCT3--.

OCT3--L

1234567

8

9

10 II

12

OCTA2 6

probe:OCTA26 Figure 6. The Cloned cDNA Encodes

Oct.3

(A) The EcoRl insert derived from AC1 was subcloned in the Bluescript SK vector. The resulting plasmid was linearized and was used for the production of sense RNA. The sense RNA was translated into proteins as described in Experimental Procedures. The in vitro synthesized proteins were analyzed by gel-shift assay (lanes 2-5). Lane 2: no competitor added. The indicated competitor was added in lanes 3, 4. and 5. Lanes 1 and 7: P19 D- nuclear extract containing authentic Ott-3. Lane 6: reticulocyte lysate without added RNA. (B) Nuclear extracts were prepared from P19 cells and four PlS-derived cell lines (R2, R6, R9, and R14) that were stably transformed with an antisense expression vector (p6Aoct3:AS shown in Figure 10). Each extract was analyzed by the gel shift assay.

Expression Pattern of the act3 Gene We next examined the expression pattern of the act-3 gene. First, RNA was prepared from D- and D+ P19 cells and was analyzed by Northern blots. A single band of 1.5 kb mRNA was detected in D- P19 cells (Figure 8A). Therefore, it must be the 1.5 kb mRNA that encodes Ott-3. On the other hand, the 1.5 kb mRNA disappears dramatically when the cells are induced to differentiate by RA. This is consistent with the observation that Ott-3 disappears during cell differentiation (see Figures 3 and 4). Various organs of adult mouse and mouse embryos at several different stages (days 8, 10, 12, 14, and 16 postcoiturn) were also examined for expression of the o&3 gene. However, the act-3 mRNA was not detectable in any of the samples tested (Figure 88; RNA from day 8 embryos is not included in this Figure). This was predictable in a way because Pi9 cells, in which a high level of o&3 mRNA

35 of 60,29 of 60, and 32 of 60 positions, respectively (Figure 7). A rather low homology in the POU domain (about 60%) between Ott-3 and other POU-containing proteins indicate that Ott-3 is a distant member of the POU family. However, comparison of the homeobox subdomain of Oct3 with those of other homeobox-containing genes shows a considerably lower level (about 30%) of homology; this region of Ott-3 is identical to the homeobox domains of

A domain

IgECT1;

I-

~1

OCT-2 Pit-l ?iNC-86

ICONSENSVS

LE

FA

K

RI

LG TO

VG

G

lp CC’?- i CCT-l CC?-i p:' -1 UN:‘-86

B domain -4

QDMKALQK LEEPSDLE PEEPSDLE DMDSPEIR DMDT-DPR SQ TI

RFE

L LS

Homeodomain

N

L

I.

w

A

~1

NLQEICKSETLVQA----~-----RKR SDSSLSSPSALNSP--GIEGL-SKRRK VDSSIPSPNQLSSPSLGFDGLPGRRRK ALYNEK-~----~-----VGANERKRK A~KQKDT:GDIN----GIL?-NTDKKR

CIINSE’~~lS

Figure 7. Amino Acid Sequence

RT

Homology

I

LEF

?

I

AL

between Ott-3 and Other POU-Containing

K i' 9"XFCN

XC

'(,

Proteins

Regions of homology between Ott-3 and other members of the POU family are shown. Ott-1 (Sturm et al., 1988) Ott-2 (Muller et al., 1988; Scheidereit et al., 1988) Pit-l (Ingraham et al., 1988; Bodner et al., 1988) uric-86 (Finney et al., 1988). Amino acids are boxed if they are present at a given position in all five proteins.

Cell 466

A

A

DMSO + Aqgraqation c-------15 4 3

2

1 05

Aggrextion 0 05

1

2

3

4 S(days)

L

*

‘_

tl.5kb

1.5kb-W

B

aggregation 4 3

2

fRA 105-o-051 II

l.Skb+

Figure 9. The act-3 Gene Is Regulated

12 Figure 8. Expression

34

5

6 7 8

910

Pattern of the ocf-3 Gene

Poly(A)+ RNA (10 rg) prepared from the indicated samples was run on a formamide-agarose gel and was analyzed by Northern blot hybridization. (A) For P19 D+ cells, the cells were induced to differentiate by RA plus cell aggregation, and RNA was extracted on day 8 after the induction. The probe used in this experiment was the 220 bp fragment corresponding to the A and B subdomains of Ott-3 (this fragment was derived from a partial cDNA clone, 185). Three sets of filter were prepared, and they were washed at 50% under three different salt conditins, as indicated. (B) The probe used was the Avalll-EcoRI 220 bp fragment, which includes the C-terminal coding sequence and the 3’ untranslated region of the act-3 cDNA (from nucleotide residue 1095 to 1315). The filter was washed at 50% in 0.1 x SSC.

was detected, are probably equivalent to inner cell mass cells found in blastocysts. Therefore, one would expect to see expression of the act-3 gene at the blastocyst stage (day 3.5 postcoitum). We wished to examine embryos of earlier stages, but day 8 embryos were the earliest ones that could be used for Northern blot analysis. Expression of the act-3 Gene Is Regulated by RA The previous Northern blots (Figure 8A) indicated that Ott-3 was abundant in P19 stem cells but absent in RAinduced D+ cells. In that experiment, RNA was extracted from D- P19 cells or from D+ cells 8 days after the initiation of RA-induced differentiation. We next examined the time course of the act-3 mRNA, to learn when the act-3 mRNA disappeared during the cell differentiation. We also asked whether the expression of the act-3 gene would change during dimethyl sulfoxide (DMSO)-induced cell differentiation. It has been known that when P19 stem cells are treated with DMSO, they differentiate into distinct cell types such as cardiac muscles and skeletal muscles (Edwards et al., 1983; Rudnicki and McBurney, 1987).

2 3 (days)

i

by RA

Cytoplasmic RNA (10 Kg) prepared from the indicated cells was analyzed by Northern blot hybridization. The probe was the same AvalllEcoRl 220 bp fragment used in Figure 86. In (A) PI9 stem cells were induced to differentiate either by DMSO plus cell aggregation or by RA plus cell aggregation. 0: uninduced cells. The cells were harvested at various days after each induction. act-3 mRNA (1.5 kb) is indicated by the arrow. In (B), P19 stem cells were either plated on bacterial dishes without RA (Aggregation) or plated on tissue culture-grade dishes in the presence of 500 nM RA (RA). The cells were harvested for RNA extraction at various days after each treatment.

P19 stem cells were induced to differentiate by cell aggregation in the presence of RA or DMSO, and RNA was extracted from the induced cells at various days after the initial treatment. The level of act-3 mRNA was examined by Northern blots (Figure 9A). When the cells were induced in the presence of RA, the level of act-3 mRNA was already reduced to about 10% 12 hr after the induction and was undetectable at 1 day. Since this assay measures a steady state level of mRNA, not a synthesis rate, it is conceivable that de novo synthesis of ocr-3 mRNA is repressed much earlier than 12 hr after induction. In contrast to RAinduced differentiation, the level of act-3 mRNA did not change during DMSO-induced differentiation, indicating that the repression of the act-3 gene is specific for RAinduced differentiation. RA-induced differentiation of P19 cells into neurons and glial cells requires both cell aggregation and RA (JonesVilleneuve et al., 1983; Rudnicki and McBurney, 1987). In the RA-induced differentiation experiments described above, P19 stem cells were first plated in a bacterial-grade dish, allowing the cells to make aggregates, in the presence of RA. Perhaps two signals are necessary for the cell differentiation: one mediated by cell-cell interaction, which could be achieved by cell aggregation, and the other mediated by RA and the RA receptor. We next examined which signal (or if both) was required for the repression of act-3 mRNA. P19 stem cells were plated on bacterial-grade dishes in the absence of RA or plated on tissue

An Embryonic 467

&tamer-Binding

Factor

REPORTFRS pSScat pOCTAcat EFFECTORS pBAoct3 pBAAN:S pl3AoctJ:AS pt3A~N:As pBAoct1

:S

p6Aoctl

:AS

Figure il. Effect of Ott-3 SV40 Enhancer

Level on the GCTA26 Enhancer

and the

Relative CAT activity induced by each effector is summarized. CAT activity induced by pUC was set to 1.0. The right part, where the reporter is pOCTAcat, was taken from two experiments such as that shown in Figure 10. In the left part, the reporter is pSV2cat. The same set of the effecters were similarly examined. This is a summary of two transfection experiments.

UL

REPORTER:

pBScat

Figure 10. Stimulation

pOCTAcat

of the OCTA26 Enhancer

by Ott-3

The indicated reporter plasmid (IO ug) and 15 ug of each effector plasmid were cotransfected to D- P19 cells. Equivalent amounts of cell extracts were used for the CAT assay. Structures of reporters and effectors are summarized at the top. The open circle in the reporter plasmid is the enhancerless SV40 early promoter (including a TATA box and GC boxes). Each solid circle indicates one copy of the OCTA26 oligonucleotide.

culture-grade dishes at a low density (thus not allowing the ceils to form aggregates) in the presence of RA. The cells were harvested at various days after each treatment, and the 01%3 mRNA level was determined. As shown in Figure 96, it is now clear that RA alone is enough to repress the expression of ocf-3 mRNA, while cell aggregation alone does not have any effect on the level of 01%3 mFlNA. Ott-3 Is a Transcription Factor We wished to determine whether Ott-3 is a transcription factor as well as a DNA binding factor. For this purpose, we constructed an act-3 expression vector (pfMoct3, shown in Figure 10) in which the act-3 cDNA is placed under the control of the human 6-actin promoter. A reporter plasmid (pOCTAcat, shown in Figure 10) was also constructed, in which six tandem repeats of OCTA26 were placed upstream of an enhancerless SV40 early promoter linked to a cat gene. The six tandem repeats of OCTA26 have activity as a stem cell-specific enhancer; expression of pOCTAcat, relative to that of pBScat (Figure lo), is stimulated in P19 cells but not in D+ cell types such as L cells (Okamoto et al., submitted) (Figure 10). Initially, the o&3 expression vector and the reporter gene were cotransfected

into various D+ cells, and it was determined whether the reporter gene could be stimulated. However, no significant induction was observed in any of the D+ cell types tested (including L cells, HeLa cells, and several hybrid cell lines between P19 cells and L cells). Possible reasons for this observation are described later. We then took another approach. The level of Ott-3 in P19 cells was altered (increased or reduced) by introducing the ocf-3 expression vector or ocf-3 antisense vectors (p6AocD:AS and pPAAN:AS), and we examined how the activity of the stem cell-specific enhancer would respond to the alteration. The ocr-3 sense vector (pPAAN:S), act-7 sense vector (ppAoctl:S), and o&l antisense vector (p6Aoctl:AS) served as controls (only p6Aoct3 would produce a functional protein). When pOCTAcat was used as a reporter, the act-3antisense expression vectors reduced the activity of the OCTA26 enhancer while the act-3 expression vector increased the enhancer activity (Figures 10 and 11). p6Aoct3, when compared with pf3Aoct3:AS, induced CAT activity la-fold (Figure 11). Therefore, the increase in Ott-3 level in P19 cells appeared to stimulate the enhancer activity. (It is difficult to examine the alteration of the Ott-3 level, since the transfected cells comprised only 10%-E% of the the total cells when they were harvested for the CAT assay; data not shown.) The control plasmids including the ocf-7 antisense vector did not show such effect, suggesting that the observed effect was due to changes in the Ott-3 level in the transfected cells. When pSV2cat was used as a reporter, on the other hand, we observed an opposite effect: pf3Aoct3 reduced the activity of the SV40 enhancer, while two act-3 antisense expression vectors increased the activity (Figure 11). It seemed that the increase in Ott-3 level resulted in reduction of the SV40 enhancer activity. This effect is probably mediated through an octamer motif that is known

Cell 466

c-Jun system would be ideal for this purpose, since EC cells are known to possess a very low level of endogenous c-Jun (Chiu et al., 1988). The results (Figure 12) indicated that the act-3-c-dun hybrid protein behaved much like the wild-type c-Jun; both proteins could stimulate transcription of a target gene in a TRE-dependent manner. Unexpectedly, plFNcat was also activated considerably by the hybrid protein and by c-Jun and Ac-Jun to a lesser degree. It is possible that the specificity of the c-Jun binding domain was slightly loosened owing to the fusion. It is not certain why plFNcat was activated by c-Jun and Ac-Jun, but this may be ascribed to TRE-like sequences present in pBR322. Nonetheless, the results described here demonstrate that Ott-3 has a transcriptional activation domain.

pIFNcat pIFNcat-1TRE pIFNcat-3TRE

c-

jun

Ac-

jun

oct3-jun

Discussion

% ACETYLATION: ~0. 1 1 1 2 REPORTER:

1 3

pIFNcat

Figure 12. Trans-Activation

5 so.1 0.4 I , ,5 6

3 7

pIFNcat-1TRE by an Ott-3-c-Jun

IO 2 '0.1 70 60 8, ,9 10 11 12, pIFNcat-3TRE Hybrid Protein

Structures of the reporters and effecters used in this experiment are shown on the top. The open circle in the reporters indicates a TATA box derived from a human b-interferon gene. Each solid circle indicates one copy of a 16 bp oligonucleotide containing TRE. Five micrograms of the indicated reporter and IO ug of the indicated effector were mixed and transfected to P19 cells. Percent of acetylated chloramphenicol is shown for each lane, under the radioautogram.

to exist in the SV40 enhancer. These results suggest that, at least in P19 cells, Ott-3 activates the OCTA26 enhancer, while it rather represses the SV40 enhancer. Finally, we wished to demonstrate that Ott-3 has a transcriptional activation domain as well as a DNA binding domain. The deduced amino acid sequence of Ott-3 indicated the presence of a remarkably proline-rich region in its N-terminal portion (see Figure 5). Since several other transcription factors such as CTWNF-1 and ClEBP also contain a proline-rich activation domain (Landschulz et al., 1988b; Mitchell and Tjian, 1989; Mermod et al., 1989) we examined whether the proline-rich region of Ott-3 can function as a transcription activation domain. We constructed a vector that can express an Ott-3-c-Jun hybrid protein (Figure 12) in which the N-terminal half of Ott-3 (which includes the proline-rich region but excludes the POU domain) is fused to a DNA binding domain of c-Jun. The minimal DNA binding domain of c-Jun defined by others (Nakabeppu et al., 1988) was used. This vector, along with various reporter plasmids, was transfected to P19 cells, and we asked whether the ocr-3-c-Jun hybrid protein could rrans-activate a reporter gene in a TRE (TPAresponsive element)-dependent fashion. Wild-type c-Jun and Ac-Jun (lacking the N-terminal half of c-Jun) served as positive and negative controls, respectively. This TRE/

Novel Octamer Binding Factors In this paper, we have reported a novel octamer binding factor, Ott-3, which appears to be specifically expressed in embryonic stem cells. Several lines of evidence indicate that the cDNA described in this study indeed encodes Ott-3. First, only two kinds of octamer binding factors (Oct1 and Ott-3) were detected in D- P19 cells by the gel-shift assays (Figure 3). When a cDNA library constructed from D- P19 cells was screened, all the positive clones (except for one clone coding for Ott-1) were derived from the same mRNA. Second, the probes derived from the cloned cDNA detected a single 1.5 kb mRNA in D- P19 cells. This mRNA was absent in D+ P19 cells (Figure 8A), which was in good agreement with the gel-shift data (Figure 3). Finally, the Ott-3 level was greatly reduced or abolished in the PlS-derived cell lines that were stably transformed with the o&3 antisense expression vector (Figure 6B). In addition to Ott-3, we have also detected another octamer binding factor (tentatively designated Ott-4), which was absent in D- P19 cells but was induced in D+ P19 cells. Since the probe derived from the Ott-3 POU region failed to detect any related mRNA in D+ P19 cells, even under the least stringent condition (Figure 8A), Ott-4 is probably encoded by a POU family member distant from the act-3 gene. Very recently, we have employed the polymerase chain reaction technique described by He et al. (1989) and have cloned two POU genes from D+ P19 mRNA (Okamoto et al., unpublished data); one cDNA clone encodes Ott-2, the other codes for a protein that closely resembles Brn-1 and Brn-2 (He et al., 1989). Therefore, Ott-4 described in this report may be encoded by one of these genes. Lenardo et al. (1989) have recently detected, by gel-shift assay, a unique octamer binding factor called NF-A3 in F9 cells. NF-A3 formed a gel-shift complex similar to that of Ott-3; therefore it is possible that Ott-3 is identical to NFA3. Furthermore, after our manuscript was submitted, Scholer et al. (1989) have reported the presence of several new species of octamer binding factors (designated Ott-3 to Ott-10 by them) in mouse embryos and EC cells. Identities of those factors are not established at this moment since corresponding cDNAs have not been cloned. However, our Ott-3 seems to resemble their Ott-4 in the gel-

An Embryonic 469

Octamer-Binding

Factor

shift mobility and expression pattern. In any case, it appears that several different species of octamer binding factors exist, and they are expressed differentially during mammalian development. Ott-3 as a Transcription Factor Two lines of evidence have indicated that ocf-3 cDNA encodes a transcription factor. First, when the level of Ott-3 in Pi9 cells was altered by the act-3 expression vector or the act-3 antisense expression vectors, the increase in the Ott-3 level resulted in stimulation of the OCTA26 enhancer activity (Figures 10 and 11). Second, the N-terminal half of Ott-3 functioned as a transcriptional activating domain when fused to a DNA binding domain of c-Jun (Figure 12). In light of the recent observations by others (Mermod et al., 1989; Mitchell and Tjian, 1989) that other transcription factors such as CTWNF-I contain a proline-rich region in their activation domains, it was probably the proline-rich region included in the N-terminal half of Ott3 that functioned as a transcriptional activation domain. We noticed that a C-terminal region of Ott-3 (residues 304 to 358) is also rich in proline (12 of 55 residues; 22%). Whether this region can function as an activation domain as well remains to be determined. In fact, transcription factors often have more than one activation domain; Spl, for example (Courey and Tjian, 1988). The results shown in Figures 10 and 11 suggest that Ott-3 is involved in activation of the OCTA26 enhancer, while it rather represses the SV40 enhancer. The observed effect of Ott-3 on the SV40 enhancer is probably mediated through its binding to an octamer motif present in the SV40 enhancer. Perhaps Ott-3 represses the SV40 enhancer activity by competitive binding to the octamer motif. It has been pointed out that some EC cells possess repressor-like activity that inhibits the SV40 enhancer (Gorman et al., 1985). Ott-3 may account for such repressor activity present in EC cells. Initially, we examined whether Ott-3 can trans-activate the OCTA26 enhancer in various D+ cells. The D+ Cells tested included L cells, HeLa cells, and several hybrid cell lines between P19 and L cells. A number of different reporters were also tested, in which the OCTA enhancer or the original El enhancer was linked to the enhancerless SV40 promoter or other promoters. However, the act-3 expression vector failed to trans-activate any of the reporters in any of the cell types tested. Several explanations are possible for this. First and most importantly, when the Ott-3 protein produced by the act-3 expression vector in D+ cells was examined by the gel-shift assay, it showed a slightly faster mobility than Ott-3 prepared from D-- P19 cells (data not shown; it should be reminded that the in vitro synthesized protein also showed a slightly faster mobility). The difference in the mobility is certainly not due to the cDNA that was used to construct the act-3 expression vector. We have recently cloned and characterized a chromosomal act-3 gene (Okazawa et al., unpublished data). The nucleotide sequence of the genomic act3 gene perfectly matched the sequences of the act-3 cDNA clones including Xl. Furthermore, we have determined the transcription initiation sites for the act-3 gene.

The initiation takes place at multiple sites, all of which are clustered in a 50 bp region (the hC1 clone lacks only lo-60 bp of the 5’ untranslated sequence), indicating that there are no additional protein coding sequences other than those cloned in hC1. Therefore, the observed difference in the mobility must be due to some type of protein modification. It is not certain at this time what type of modification was responsible. However, some transcription factors are known to be regulated by phosphorylation; phosphorylation affects DNA binding or tfans-activation (for example, heat shock transcription factor; Sorger and Pelham, 1988). Whatever type of modification was responsible for the difference between the Ott-3 molecule in DP19 cells and the Ott-3 molecule produced in D+ cells, it is conceivable that the modification affected the function of Ott-3 (not DNA binding but transcriptional activating function). Alternatively, it is possible that the level of Ott-3 produced in the transfected cells was not enough. In fact, the level of Ott-3 produced was somehow low; only l/l0 of the endogeneous Ott-1 (data not shown). It is also possible that Ott-3 may have been able to frans-activate the enhancer if it has been linked to a different type of promoter. Repression of the ocf-3 Gene by RA RA has been implicated in cell differentiation and morphogenesis. In particular, RA functions during limb bud formation as though it is a morphogen. RA binds to its nuclear receptor (retinoic acid receptor [RAR]), and the ligand-receptor complex presumably recognizes control regions of a set of genes and activates or represses those target genes. Although this hypothesis is widely accepted, the identity of the target genes of RAR is totally unknown. In this view, our finding that the act-3 gene is regulated by RA is extremely interesting. Two important points would emerge from our results. First, the steady state level of ocf3 mRNA is already reduced to 10% 12 hr after RA treatment (Figure lo), suggesting that repression of the o&3 gene takes place very early after the RA treatment. Repression of the act-3 gene should precede commitment of the stem cells, for the following reasons. When P19 cells are exposed to RA for 12 hr and subsequently RA is removed, most of the cells still remain pluripotential (data not shown). Furthermore, D- cell-specific markers such as SSEA-1 and AEC3Al-9 are fully maintained until 48 hr after the RA treatment (Jones-Villeneuve et al., 1983; our unpublished data). The second important point is that the repression is not related to the fate of the differentiating cells but is specific for RA treatment; i.e., RA plus cell aggregation induce P19 stem cells to differentiate mainly into neurons and astroglia, while exposure of the stem cells plated on tissue culture grade dishes to RA eventually leads them to fibroblast-like cells (Rudnicki and McBurney, 1987; our unpublished data). Although the fates of the cells are thus different between two procedures, the ocf-3 gene was repressed in both cases (Figures 9A and 96). In contrast, the act-3 gene was active throughout the DMSO-induced differentiation (Figure 9A). These observations indicate that the repression is not an indirect effect of cell differentiation, but is rather closely linked to RAR.

Cell 470

It is tempting to speculate that the act-3 gene may be a primary target gene of RAR. This possibility should be tested by isolating a genomic ocf-3 gene, identifying its transcriptional control region, and searching for RAR binding sites there. Role of Ott-3 in Embryogenesis While the act-3 gene was actively expressed in P19 stem cells, it was not expressed (at least to a detectable level) in any of the adult organs examined, nor in day 8-18 embryos (Figure 86). These results were somewhat predicted because P19 stem cells are thought to be equivalent to inner cell mass cells (day 3.5 after fertilization). Therefore, it is probably in those early stem cells (at the blastocyst stage) that the o&3 gene is expressed. In this view, it is important to study temporal and spatial expression of o&3 mRNA during mouse development, by in situ hybridization. The structural and functional properties of Ott-3, its expression pattern, and dramatic down-regulation by RA all suggest that it may regulate differentiation of embryonic stem cells by controlling a set of target genes. The behavior of Ott-3 during the RA-induced cell differentiation leads us to speculate that repression of the 01%3 gene may be required for the stem cells to differentiate toward certain cell lineages; disappearance of Ott-3 would result in repression or activation of its target genes. One of the target genes of Ott-3 must be an “early transposon”, from which El and E2 (two embryonic stem cell-specific enhancers used in this study) are derived. Consistent with this, the transposon has been known to be actively expressed in embryonal stem cells but repressed when the cells differentiate (Brulet et al., 1983). However, activation or repression of the transposon would not have a significant effect on stem cell differentiation because its transcript does not seem to encode any functional protein (Sonigo et al., 1987). Presumably, there are other target genes that are functionally relevant to cell differentiation, the identification of which is of paramount importance. It is also important to study the regulation of the ocr-3 gene itself and to identify a gene of a higher hierarchy, if such a gene exists. We believe that these lines of study will reveal a cascade of regulatory genes that operate during early embryogenesis. Experimental

Procedures

cell Culture PI9 cells were cultured as described by Rudnicki and McBurney (1987). They were maintained in minimal essential medium (a modification) supplemented with 10% fetal bovine serum as D- cells. Cell differentiation was induced as follows, unless otherwise mentioned. P19 cells were plated in bacterial-grade dishes in the presence of 0.5 uM RA. The cell aggregates thus formed were plated on tissue culturegrade dishes on day 5 after the induction. The cells were harvested on day 7 or 8, when all the cells differentiated into a variety of cell types, such as neurons, astroglia, and smooth muscles. For establishing cell lines transformed with an act-3 antisense expression vector, P19 cells were cotransfected with pSVneo and a lo-fold excess of pf3Aoct3:AS (Figure 10). After selecting G418-resistant cells, 18 cell lines transformed with pf3AocB:AS were established (detailed characterization of those cell lines will be described elsewhere).

DNA Binding Assays Nuclear extracts were prepared from D- and D+ EC cells essentially as described by Dignam et al. (1983). To avoid protein degradation during extract preparation, four protease inhibitors (1 m M PMSF, 1 &ml leupeptin, 1 w/ml pepstatin A, and 0.3 uglml antipain) were present throughout. DNAase I footprint analysis was performed as described by Jones et al. (1985). The appropriate restriction fragments derived from mouse DNAfragments 052 and 015 (Bhat et al., 1988) containing an EC stem cell-specific enhancer El or E2, were end labeled and used as probes. Gel-shift assays were performed as described by Sen and Baltimore (1988). For the gel-shift assays, two double-stranded oligonucleotides were chemically synthesized. One was designated OCTA28 (GATCAGTACTAATTAGCATTATAAAG); the other was designated EN28 (GATCCCATACTTAAAATTCAACTAGA). Isolation and Characterization of 012.3 cDNA Clones For the screening of the cDNA libraries by hybridization, two oligonucleotides were chemically synthesized: the 89 mer corresponding to the amino acid sequence of the Ott-2 POU-specific subdomain 8, and the 63 mer corresponding to the amino acid sequence of the 06-2 homeobox subdomain. The amino acid sequence of Ott-2 published by Muller et al. (1988) was used for this. The codon usage frequency in mouse was taken into account when the amino acid sequence was reverse translated into nucleotide sequence. The hgtl0 cDNA libraries constructed from D- and D+ PI9 cells were screened with those two probes. Hybridization was done in 6x SSC at 5oOC. Filters were washed in 6x SSC at 50% 3 times. For each library, 5 x lo5 to 1 x lo6 recombinant phages were screened. Positive clones were observed only in the P19 (D-) library. In this library, there were about 20 positive phages in every 5 x IO4 phages screened. More than 20 positive clones were randomly chosen and purified. Their inserts ranged from 0.5 kb to 1.3 kb. These inserts were subcloned in the Bluescript vector. For each insert, the nucleotide sequence was determined from both ends, using the Sequenase kit (United States Biochemical Corporation). The obtained sequence data indicated that all but one clone were perfectly overlapping; therefore, the largest EcoRl insert (1.3 kb) from hC1 was further characterized in detail. The entire nucleotide sequence of the 1Cl insert was determined by constructing a series of deletion mutants, followed by sequencing from the deletion point. The remaining one clone (hH8) turned out to be a cDNA clone for O&l. In Vittu Transcription and Translation of act-3 cDNA The EcoRl insert cloned in 1Cl was subcloned in the Bluescript vector. The resulting plasmid DNA was linearized with BamHI. Sense RNA was transcribed in vitro with T7 RNA polymerase. About 1 ug of RNA was added to a rabbit reticulocyte lysate (Amersham). After the in vitro translation reaction was completed, the reaction mixture (50 ul) was passed through a spun column equilibrated with Dignam buffer D (Dignam, 1983); 5 ul of the column eluate was used for the gel-shift assay. RNA Analyses For preparation of RNA from EC cells, the cells were lysed by homogenization in the presence of 0.5% NP40, and RNA was extracted from the cytoplasmic fraction with phenol and chloroform. For preparation of RNA from mouse embryos and organs, guanidine-hot phenol method (Maniatis et al., 1982) was employed. RNA was passed through an oligo(dT)-cellulose column and poly(A)+ RNA was used for the Northern blots, unless otherwise mentioned. Construction of Expression Vectors and Reporter Plasmids For pBScat, the Bglll-BamHI smaller fragment of pAlOcat was cloned at the BamHl site of Bluescript SK. The cat gene in pBScat is linked to an enhancerless SV40 early promoter. OCTA26 was polymerized and subcloned at the BamHl site of pUC12. For pOCTAcat, six tandem repeats of OCTA26 were excised as the Sacl-Xbal fragment, which was then placed at the corresponding sites of pBScat. The EcoRl insert of 1Cl was subcloned at the EcoRl site of Bluescript KS. act-3 cDNA, cloned in either sense orientation or antisense orientation, was excised at the Sall-BamHI fragment, which was subsequently cloned at the corresponding site of p8APr-1 (Gunning et al., 1987) a cDNA expression vector driven by human p-actin promoter (pf3Aoct3 and

&&Embryonic

Octamer-Binding

Factor

p6Aoct3:AS Figure IO). pbAAN:S and pj3AAN:AS (Figure 10) were constructed in a similar manner using the EcoRl insert of a partial ocr-3 cDNA clone, SBHS. p8Aoctl:S and p6Aoctl:AS were also created by the same strategy using the EcoFtl insert of an Ott-1 cDNA clone, hH8. iH8 contains 1145 bp of an Ott-1 coding sequence (from amino acid residue 352 to 699 of Ott-1, described by Sturm et al., 1988). plFNcat (Hata et al., 1989a), plFNcat-I TRE, and plFNcat-3TRE (Hata et al., 1989b) were kindly provided by Drs. Shigeo Ohno and Akiko Hata. plFNcat contains a human h-interferon gene promoter fragment (-55 to +19) including the TATA box. plFNcat-1 TRE and plFNcat-3 TRE have one copy and three copies, respectively, of a 16 bp oligonucleotide that corresponds to the TRE of a collagenase gene. A c-jun expression vector (p6Ac-jun) was constructed in p8APr-1, using a cDNA clone for rat c-[un (Sakai et al., 1989). For the construction of pgAoct3-jun, the 5’ terminal half of the ICl insert was excised as the Sall-Pstl fragment. The 3’terminal half of the rat c-jun cDNA was obtained as the Pstl-BamHI fragment. Those two fragments were ligated at the Pstl ends and the resulting fragment was subsequently cloned at the Sall-BamHI sites of pj3APr-I. For the construction of ppAAc-jun, the hC1 insert subcloned in Bluescript was digested with Ncol (located at the initiation codon). The Ncol end was filled in by Klenow enzyme (this preserves the initiation codon) and ligated to a Pstl linker. To align the coding sequences in frame, the 12 bp Pstl linker (GCTCTGCAGAGC) was synthesized and used for this. The ligated DNA was then digested with Sall and Pstl and the 80 bp Sall-Pstl fragment containmg the 5’ untranslated region and the initiation codon of act-3 mRNA was isolated. This fragment was ligated to the Pstl-BamHI rat c-jun fragment described above and subsequently cloned at the Sall-BamHI site of p6APr-1. Acknowledgments We thank all the members of our Department for encouragement and comments during the work. We are especially grateful to N. Tanaka for assisting in the footprint assay, Y. Maeda and M. Noda for oligonucleotides, and M. lmagawa for his invaluable advice. We also thank K. Yamamoto for homology search, S. Ohno and A. Hata for plFNcat reporters, and M. Karin for his encouraging advice on harts-activation experiments. This work was supported by grants from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges, This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Received June 26, 1989; revised December

Edwards, M. K. S., Harris, J. F, and McBurney, M. W. (1983). Induced muscle differentiation in an embryonal carcinoma cell line. Mol. Cell. Biol. 3, 2280-2286. Finney, M., Ruvkun, G., and Horvitz, H. R. (1988). The C. elegans cell lineage and differentiation gene uric-86 encodes a protein with a homeodomain and extended similarity to transcription factors. Cell 55, 757-769. Fletcher, C., Heintz, N.. and Roeder. R. G. (1987). Purification and characterization of OTF-1, a transcription factor regulating cell cycle expression of a human histone H2b gene. Cell 57, 773-781. Gardner, R. L.. and Rossant, J. (1979). Investigation of the fate of the 4-5 day post-coitum mouse inner cell mass cells by blastocyst injection. J. Embryol. Exp. Morphol. 52, 141-152. Gorman, C. M., Rigby, f? W. J.. and Lane, D. P (1985). Negative regulation of viral enhancers in undifferenbated embryonic stem cells. Cell 42, 519-526. Gunning, P, Leavitt, J., Muscat, G., Ng, S-Y., and Kedes. L. (1987). A human f3-actin expression vector system directs high-level accumulation of antisense transcripts. Proc. Natl. Acad. Sci. USA 84.4831-4835. Hamada, mosomal

H. (1986a). Activation of an enhancerless integration. Mol. Cell. Biol. 6. 4179-4184,

gene by chro-

Hamada. H. (1986b). Random cloning of gene activator elements from the human genome. Mol. Cell. Biol. 6, 4185-4194. Hata, A., Ohno, S., Akita, Y., and Suzuki, K. (1989a). Tandemly reiterated negative enhancer-like elements regulate transcription of a human gene for the large subunit of calcium-dependent protease. J. Biol. Chem. 264, 6404-6411. Hata, A., Akita, Y., Konno, Y., Suzuku, K., and Ohno, S. (1989b). Direct evidence that the kinase activity of protein kinase C is involved in transcriptional activation through a TPA-responsive element. FEBS Lett. 252, 144-146. He, X., Treaty, M. N., Simmons, D. M., Ingraham, H. A., Swanson, L. W., and Rosenfeld, M. G. (1989). Expression of a large family of PO&domain regulatory genes in mammalian brain development. Nature 340, 35-42. Herr, W., Sturm, R. A., Clerc. R. G., Corcoran, L. M., Balhmore. D., Sharp, t? A., Ingraham, H. A., Rosenfeld, M. G., Finney, M., Ruvkun, G.. and Horvttz, H. R. (1988). The POU domain: a large conserved region in the mammalian pit-l, act-1. act-2 and Caenorhabditis elegans uric-86 gene products. Genes Dev. 2, 1513-1516. Hoey, T., and Levine, M. (1988). Divergent homeo box protems recognize similar DNA sequences in Drosophila. Nature 332, 858-861. Hope, I. A., and Struhl, K. (1985). GCN4 protein, synthesized in vttro, bmds HIS3 regulatory sequences: implications for general control of amino acid biosynthetic genes in yeast. Cell 43, 177-188.

6, 1989.

Bandziults, R. J., Swanson, M. S., and Dreyfuss. G. (1989). RNAbinding proteins as developmental regulators. Genes Dev. 3, 431-437 Bhat, K., McBurney, M. W., and Hamada, H. (1988). Functional clonmg of mouse chromosomal loci specifically active in embryonal carcinoma stem cells. Mol. Cell. Biol. 8, 3251-3259. Bodner, M., Castrillo, J.-L., Theill, L. E., Deerinck, T., Elksman, M., and Karin, M. (1988). The pituitary-specific transcription factor GHF-1 is a homeobox-containing protein. Cell 55, 505-518. Brulet, P., Kaghad. M., Xu, Y. S., Croissant, O., and Jacob, F. (1983). Early differential tissue expression of transposon-like repetitive DNA sequences of the mouse. Proc. Natl. Acad. Sci. USA 80, 5641-5645. Chiu, R., Boyle, W. J., Meek, J., Smeal, T., Hunter, T., and Karin, M. (1988). The c-Fos protein interacts with c-Jun/AP-1 to stimulate transcription of AP-1 responsive genes. Cell 54, 541-552. Courey, A. J., andTjian, R. (1988). Analysis of SPl in vivo reveals multiple transcriptional domains, including a novel glutamine-rich activation motif. Cell 55, 887-898.

Ingraham, H. A., Chen, R., Mangalam, H J.. Elsholtz, H. P. Flynn, S. E., Lin, C. R., Simmons, D. M., Swanson, L., and Rosenfeld, M. G. (1988). A tissue-specific transcription factor containing a homeodomain specifies a pituitary phenotype. Cell 55, 519-529. Jones, K. A., Yamamoto, K. R., and Tjian, R. (1985). Two distmct transcription factors bind to the HSV thymidine kinase promoter in vitro. Cell 42, 559-572. Jones-Villeneuve, E. M. V., Rudnicki, M. A., Hams, J. F., and McBurney, M. W. (1983). Retinoic acid induced neural differentiation of embryonal carcinoma cells. Mol. Cell. Biol. 3, 2271-2279. Ko, H. S., Fast, P, McBride, W., and Staudt, L. M. (1988). A human protern specific for the immunoglobulin octamer DNA motif contains a functional homeobox domain. Cell 55, 135-144. Landschulz. W. H., Johnson, l? F., and McKnight. S. L. (1988a). The leucine zipper: a hypothetical structure common to a new class of DNA binding proteins. Science 240, 1759-1764.

speci-

Landschulz, W. H.. Johnson, P F., Adashi. E. Y.. Graves, B. J.. and McKnight, S. L. (1988b). Isolation of a recombmant copy of the gene encoding CIEBP. Genes Dev. 2, 786-800.

Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983). Accurate transcription mitiation by RNA polymerase II in a soluble extract from Isolated mammalian nuclet. Nucl. Acids Res. II, 1475-1489.

Lenardo, M.. Staudt, L.. Robbins. P, Kuang, A., Milligan, R. C., and Baltimore, D. (1989). Repression of the IgH enhancer In teratocarcrnoma cells associated with a novel octamer factor. Science 243, 544-546.

Desplan, C., Theis. J., and O’Farrell, P H. (1988). Thesequence ficity of homeodomain-DNA interaction. Cell 54, 1081-1090.

Cell 472

Levine, M.. and Hoey, T. (1988). Homeobox proteins specific transcription factors. Cell 55, 537-540.

as sequence-

Maniatis. T., Fritsch. E. F., and Sambrook, J. (1982). Molecular Cloning: A Laboratory Manual (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory). McBurney, M. W., and Rogers, B. J. (1982). Isolation of male embryonal carcinoma cells and their chromosome replication patterns. Dev. Eiol. 89, 503-508. Mermod. N., Cl’Neill, E. A., Kelley, T. J., and Tjian. Ft. (1989). The proline-rich transcriptional activator of CTWNF-I is distinct from the replication and DNA binding domain. Cell 58, 741-753. Mitchell, P. J., and Tjian, Ft. (1989). Transcriptional regulation in mammalian cells by sequence-specific DNA binding proteins. Science 245, 371-378. Muller, M. M., Ruppert, S., Schaffner, W., and Matthias, f? (1988). A cloned octamer transcription factor stimulates transcription from lymphoid-specific promoters in non-B cells. Nature 336, 544-551. Nakabeppu. Y., Ryder, K., and Nathans, D. (1988). DNA binding activities of three murine Jun proteins: stimulation by Fos. Cell 55. 907-915. Prochownik. E. V., and Kukowska, J. (1986). Deregulated expression of c-myc by murine erythroleukemia cells prevents differentiation. Nature 322, 846-850. Rudnicki. M. A., and McBurney, M. W. (1987). Cell culture method and induction of differentiation of embryonal carcinoma cell lines. In Teratocarcinomas and Embryonic Stem Cells: A Practical Approach, E. J. Robertson, ed. (Oxford: IRL Press). Sakai, M., Okuda, A., Hatayama. I., Sato, K.. Nishi, S.. and Muramatsu, M. (1989). Structure and expression of c-jun mRNA: tissue distribution and increase during hepatocarcinogenesis. Cancer Res., in press, Scheidereit, C., Heguy, A., and Roeder, R. G. (1987). Identification and purification of a human lymphoid-specific octamer binding protein (OTF-2) that activates transcription of an immunoglobulin promoter in vitro. Cell 57, 783-793. Scheidereit. C., Cromlish, J. A., Gerster, T., Kawakami, K., Balmaceda, C. G.. Currie, R. A, and Roeder, R. G. (1988). A human lymphoidspecific transcription factor that activates immunoglobulin gene is a homeobox protein. Nature 336, 551-557. Scholer, H. R., Hatzopoulos, A. K., Balling, R., Suzuki, N., and Gruss, P (1989). A family of octamer-specific proteins present during mouse embryogenesis: evidence for germline-specific expression of an Ott factor. EMBO J. 8, 2543-2550. Sen, R., and Baltimore, D. (1986). Multiple nuclear factors interact with the immunoglobulin enhancer sequences. Cell 46, 705-716. Sive, H. L., and Roeder, R. G. (1986). Interaction of a common factor with conserved promoter and enhancer sequences in histone H2B, immunoglobulin and U2 small nuclear RNA (snRNA) genes. Proc. Natl. Acad. Sci. USA 83, 6382-8386. Sonigo, f?, Wain-Hobson. S., Bougueleret, L., Tiollais, P.. Jacob, P., and Brulet, P (1987). Nucleotide sequence and evolution of ETn element, Proc. Natl. Acad. Sci. USA 84, 3766-3771. Sorger, P K., and Pelham, H. R. B. (1988). Yeast heat shock factor is an essential DNA-binding protein that exhibits temperature-dependent phosphorylation. Cell 54, 855-864. Sturm, R. A., and Herr, W. (1988). The POU domain is a bipartite DNAbinding structure. Nature 336, 601-603. Sturm, R. A., Das, G., and Herr, W. (1988). The ubiquitous octamerbinding protein Ott-1 contains a POU domain with a homeo box subdomain. Genes Dev. 2, 1582-1599. GenBank

Accession

The accession J03178.

Numbers

number

for the sequence

reported

in this paper is

A novel octamer binding transcription factor is differentially expressed in mouse embryonic cells.

We have identified a novel octamer binding factor (Oct-3) in P19 embryonal carcinoma cells. Oct-3, which recognizes the typical octamer motif (ATTTGCA...
3MB Sizes 0 Downloads 0 Views