cDNA Cloning of a Novel, Developmentally Regulated Immediate Early Gene Activated by Fibroblast Growth Factor and Encoding a Nuclear Protein*

We have utilized the polymerase chain reaction (PCR)-based differential display methodology (Liang, P., and Pardee, A. B. (1992) Science 257, 967–969) to identify a novel transcript whose expression levels increased inXenopus embryo explants during mesoderm induction by fibroblast growth factor. The PCR product was used to clone a 2.3-kilobase pair cDNA representing this transcript, which we have named er1 (early response1). The er1 cDNA contained a single open reading frame predicted to encode a protein of 493 amino acid residues. A data base homology search revealed that the predicted ER1 amino acid sequence contains three regions of similarity to the rat and human proteins encoded by the metastasis-associated gene, mta1, and two regions of similarity to the Caenorhabditis eleganssequence that is similar to mta1. The fibroblast growth factor-induced increase in er1 steady-state levels was not dependent on de novo protein synthesis, demonstrating thater1 is an immediate-early gene. Northern blot analysis revealed a single 2.8-kilobase pair mRNA that was observed predominantly during the initial cleavage and blastula stages ofXenopus development, with little or no detectable mRNA during subsequent development. Quantitative PCR analysis of early developmental stages showed that er1 peaked during late blastula. Computer-assisted analysis of the predicted ER1 amino acid sequence revealed two putative nuclear localization signals, four highly acidic regions clustered at the N terminus and a proline-rich region located near the C terminus. Subcellular localization by immunocytochemistry revealed that the ER1 protein was targeted exclusively to the nucleus. Transactivation assays using various regions of ER1 fused to the DNA binding domain of GAL4 demonstrated that the N-terminal acidic region is a potent transactivator. These data suggest that ER1 may function as a transcription factor.

The family of fibroblast growth factors (FGFs) 1 consists of nine members related by sequence and their ability to bind heparin (1). FGFs are involved in a number of cellular activities, including mitogenesis, cell differentiation, and angiogenesis (reviewed in Ref. 2). In addition, overexpression of FGF in various cell lines leads to phenotypic transformation (3)(4)(5). To define the mechanisms by which a ligand can have such pleiotropic effects, a better understanding of molecular aspects of the various cellular responses is required.
FGF can induce mesoderm differentiation in Xenopus embryonic tissue (6), and many of the initial events in the cellular response during induction are similar to those previously characterized for the FGF-mediated mitogenic response. During mesoderm induction, FGF binds to high affinity cell surface receptors (7), which in turn become phosphorylated on tyrosine (8). The phosphorylated FGF receptor forms a signaling complex by binding a number of intracellular substrates (9), which results in activation of several well characterized signaling pathways. For instance, protein kinase C becomes activated during FGF-induced mesoderm differentiation (8) as does mitogen-activated protein kinase (10).
The ultimate targets of these signal transduction pathways are the immediate-early genes. To date, very few FGF immediate-early genes have been identified (11,12). Accordingly, we have utilized the differential display methodology (13) to isolate cDNAs representing such genes. In this paper, we describe the cloning and characterization of a cDNA representing a novel immediate-early gene, er1, whose steady-state levels increased in response to FGF. We show that the ER1 protein is targeted to the nucleus and that the N-terminal acidic region of ER1 can function as a transcriptional activator.

EXPERIMENTAL PROCEDURES
Embryos and Mesoderm Induction-Xenopus laevis were purchased from Nasco. Embryos were obtained and cultured as in Ref. 14. The recombinant Xenopus basic FGF used for induction was prepared as in Ref. 15. Animal pole explants (animal caps) were induced to form mesoderm as described (9), and animal caps were treated for 30 min prior to RNA extraction. For inhibition of protein synthesis during induction, animal caps were pretreated for 30 min with 5 g/ml cycloheximide (Sigma), cultured with or without FGF for an additional 30 min, and then processed for PCR analysis as described below. Protein synthesis was measured in parallel samples by including 2 uCi/l of [ 35 S]methionine in the culture medium, and 35 S incorporation into trichloroacetic acid precipitable material was determined according to Clemens (16).
Differential Display-RNA was extracted from induced or uninduced animal caps using the NaCl/EDTA/Tris/SDS protocol (17). Reverse transcription (RT) and polymerase chain reaction (PCR) were performed as in Ref. 13 with the following primers: 5Ј-T 11 AC-3Ј and 5Ј-CTGATCCATG-3Ј. PCR products were separated on a 6% polyacrylamide/6 M urea gel; the gel was dried, and the products were visualized * This work was supported by grants from the Medical Research Council of Canada and from the Cancer Research Society (to L. L. G. and G. D. P.). 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM 1 The abbreviations used are: FGF, fibroblast growth factor; NLS, nuclear localization signal; PCR, polymerase chain reaction; by autoradiography. Differentially expressed bands were excised, and the PCR products were eluted from the gel in 100 l of H 2 O.
cDNA Cloning and Sequencing of er1 from Xenopus Embryos-Eluted PCR products were cloned into the pCR II vector using the TA cloning kit (Invitrogen). The sequences for both strands of the initial er1 PCR product and all subsequent cDNA inserts were determined as in Ref. 14. A 2.3-kb er1 cDNA was isolated from a stage 8 Xenopus (ZAP II) cDNA library (14), using primers designed according to the er1 sequence (5Ј-TCCGTTACACCAGGATGTAG-3Ј; 5Ј-GGCTGAAATTC-CAGTT GGTA-3Ј; 5Ј-GCATCAGCTGCAGATCAAGG-3Ј; and 5Ј-GTT-TAAGAAAGGGC-AGTTCG-3Ј) and the ZAP vector sequence (5Ј-GCTCGAAATTAACCCTCACTAAAG-3Ј and 5Ј-GGTACCTAATA CG-ACTCACTATAGGG-3Ј). The cDNA was cloned into pCR II , and the sequence was determined and verified by sequencing several clones on both strands.
Quantitative PCR and Northern Analysis-Quantitative PCR analysis was performed as described in Ref. 18 with the following modifications: RNA was prepared as in Ref. 19; one-eighth of the RT sample was added to a 50-l PCR reaction, and the annealing temperature was 56°C. Histone H4 was used as a control with forward (F) and reverse (R) primers as described (18), and the primer sequences for er1 were as above. The PCR products were analyzed in the linear range for ampli-fication, determined empirically (18) to be 19 cycles for histone H4 and 24 cycles for er1. Quantitation by densitometry was performed as described in Ref. 19 with normalization to histone H4. Northern analysis was carried out as described in Ref. 20 using the 2.3-kb er1 or histone H4 cDNA as a probe.
Immunocytochemistry and Protein Analysis-Anti-Xenopus ER1 antiserum was prepared by immunizing rabbits as in Ref. 9 with a Cterminal synthetic peptide (CIKRQRMDSPGKEST) of the predicted ER1 protein sequence. Coupled in vitro transcription-translation, immunoprecipitation, and SDS-polyacrylamide gel electrophoresis were performed as in Ref. 9. For immunocytochemistry, NIH 3T3 cells were transfected with either pcDNA3 (Invitrogen) or er1-pcDNA3. After 48 h, the cells were processed for immunocytochemistry as in Ref. 19 using a 1:50 dilution of the anti-ER1 antiserum.
Plasmid Construction and Transient Transactivation Assays-NIH 3T3 cells (ATCC) were maintained in Dulbecco's modified Eagle's medium plus 10% calf serum and transfected with LipofectAMINE according to the manufacturer's directions (Life Technologies, Inc.). The expression vectors used in this assay were engineered to contain various portions of ER1 fused to the GAL4 DNA binding domain of the pM plasmid (CLONTECH) and are named according to the amino acids of ER1 that each encodes. Specific primers incorporating 5Ј and 3Ј BglII sites (ER 1-493 and ER 176 -493) or a 5ЈEcoRI and a 3Ј BamHI site (ER 1-175 and ER 1-25) were used to amplify PCR fragments encoding the appropriate amino acids. The digested PCR fragments were inserted into the complementary sites of the pM plasmid, and all plasmids were sequenced to verify the junctions and the er1 sequence and to ensure the proper reading frame. ER 1-98 and ER1-57 were generated by digesting the ER 1-175 construct with PstI or PvuII, respectively, and religating the cut vector. 0.5 g of a CAT reporter plasmid (pG5CAT, CLONTECH) was cotransfected into 3 ϫ 10 5 cells with 1.0 g of either the pM vector alone or one of the pM-er1 fusion constructs. After 48 h, cell extracts were prepared and assayed for CAT enzyme using a CAT enzyme-linked The nucleotide sequence numbers of the er1 cDNA are shown on the left, and the amino acid sequence numbers of the predicted ER1 protein are shown on the right. The TAA termination codon is indicated by an asterisk. Four stretches of predominantly acidic residues are underlined, the proline-rich region is in bold, and two putative NLS are indicated by double underlines; the second NLS conforms to the consensus for a bipartite NLS.

FIG. 2. Amino acid comparison of ER1 to the rat and human MTA1 and the C. elegans similar-to-MTA1 protein.
A, schematic illustrating alignment of the predicted Xenopus ER1 protein sequence with the rat and human MTA1 and the protein from C. elegans that is similar to MTA1. The N termini were aligned, and gaps (black lines) were introduced in the C. elegans and Xenopus proteins to align the regions of similarity (hatched) identified by the BLAST program. White boxes indicate unique regions. B, alignment of the predicted ER1 amino acid sequence with the MTA1 amino acid sequences in the regions of similarity illustrated in A. Identities are indicated by the one-letter amino acid code, conservative changes are indicated by a plus sign (ϩ), and dashes (-) indicate nonconservative changes. The amino acid sequence numbers of the ER1 protein are shown on the right. immunosorbent assay kit (Boehringer Mannheim) according to the manufacturer's directions.

RESULTS AND DISCUSSION
In our efforts to elucidate the molecular mechanisms of FGFinduced mesoderm differentiation in Xenopus, we employed the PCR-based differential display method (13) to identify and characterize genes that are expressed early during the cellular response to FGF. RNA was isolated and reverse-transcribed from five individual sets of 30-min FGF-treated or control animal pole explants (animal caps) from Xenopus blastulae. PCR products from the five sets were separated on a 6% polyacrylamide/urea gel. Only those bands that were differentially expressed in all five sets were chosen for further analysis. A total of eleven differentially expressed bands were identified, and one of these was eluted from the gel, cloned, and sequenced. A search of the data base for similarity to known sequences revealed that this cDNA represented a novel Xenopus gene, which we have named er1 (early response 1).
The sequence of the er1 PCR product was used to obtain a 2.3-kb cDNA from a Xenopus blastula library (14). This cDNA consisted of a single 1497-base pair open reading frame, bracketed by a 214-base pair 5Ј-untranslated region that contained several stop codons in all three frames and a 626-base pair 3Ј-untranslated region (Fig. 1). The ATG initiation codon is predicted to be at nucleotides 233-235, because this site is positioned within a Kozak consensus sequence for the start of translation (21), with a purine in the Ϫ3 position and a guanine in the ϩ4 position. The open reading frame is predicted to encode a protein of 493 amino acids, beginning at nucleotide 233 and ending with an in-frame TAA stop codon at position 1712 (Fig. 1).
Computer-assisted analysis of the deduced amino acid se- For gastrula stages in lanes 5-7, RNA was isolated at stages 10, 10.5, and 12, respectively, according to morphological criteria (29). RT-PCR and analysis were performed as described in the legend to Fig. 3. quence using MOTIFS and PSORT software programs predicts that ER1 does not contain an N-terminal signal sequence for transfer into the endoplasmic reticulum or a hydrophobic domain characteristic of transmembrane proteins. However, ER1 does contain two potential nuclear localization signals (NLS): RRPR and KKSERYDFFAQQTRFGKKK (Fig. 1); the latter conforms to the consensus sequence for a bipartite NLS (22). ER1 also contains a proline-rich sequence near the C terminus that corresponds to the PXXP motif found in all high affinity SH3-domain binding ligands (23). The N terminus of ER1 includes several highly acidic stretches (Fig. 1), characteristic of the acidic activation domains of many transcription factors (24).
A data base homology search using the National Center for Biotechnology Information BLAST Network Service revealed that ER1 contains three regions of similarity to the product of the rat metastasis-associated gene, mta1 (25) (Fig. 2), a gene that was isolated by differential cDNA library screening and whose expression was associated with a metastatic phenotype. mta1 encodes a 703-amino acid, 79-kDa polypeptide of unknown function that contains a putative SH3 binding domain near the C terminus. ER1 also displays similarity to the human MTA1 (accession number U35113) and to the Caenorhabditis elegans MTA1-like sequence (accession number U41264) (Fig.  2). Within the regions of similarity, the percentage of amino acid similarity ranged from 46 to 64%; however, the overall percentage of similarity was only 13.0, 14.0, and 15.6% to the rat, human, and C. elegans sequences, respectively. To investigate whether er1 represents the Xenopus homolog of mta1 or simply a related protein, we screened by RT-PCR a human breast carcinoma cell line, MDA-468 (26), for er1-related sequences. We obtained a partial human cDNA clone spanning sequence inside and outside the regions of similarity shown in Fig. 2; this sequence displays 91% overall similarity to er1 at the amino acid level (data not shown). The existence of a human gene product that is distinct from human mta1 and that shows a high degree of similarity to er1 suggests that er1 and mta1 are not homologs but possibly related members of a family of proteins or simply proteins containing some of the same functional domains.
Verification that the steady-state levels of er1 were increased in response to FGF during mesoderm induction in vitro was performed by quantitative PCR after a 30-min treatment with FGF, using histone H4 as an internal standard (Fig. 3A). In several independent experiments, densitometric analysis revealed that er1 levels ranged from 3-to 4-fold higher in FGFtreated samples, after normalization to histone H4. These data confirm that er1 levels were increased by treatment with FGF and demonstrate that the increase in er1 occurs early during the cellular response to FGF.
The possibility that er1 is an immediate-early gene was investigated further. By definition, transcription of immediateearly genes is a rapid response and is not dependent on de novo protein synthesis. The FGF-induced increase in er1 levels was measured in the presence or the absence of 5 g/ml cycloheximide. Cycloheximide inhibited 90% of [ 35 S]methionine incorporation into trichloroacetic acid-precipitable material (data not shown) but did not prevent the FGF-induced increase in er1 levels (Fig. 3B), demonstrating that er1 is an immediate-early gene.
Northern analysis of the temporal pattern of er1 expression during embryonic development revealed a single er1 mRNA (Fig. 4A). The estimated 2.8-kb size of the message was slightly larger than that of the cDNA clone, but this is probably due to the presence of a poly(A) tail. er1 was detectable during initial cleavage stages prior to the start of zygotic transcription that occurs at mid-blastula transition (27), indicating that er1 is a maternally derived mRNA. Densitometric analysis revealed that steady-state levels of er1 were relatively constant during early cleavage stages (stages 2, 6, and 7; Fig. 4A, lanes 1-3), increased slightly at blastula stage (Fig. 4A, lane 4), and then decreased 6-fold during gastrula, neurula, and tailbud stages (stages 12, 17, and 22; Fig. 4A, lanes 5-7) and remained below detectable levels during subsequent development (stages 30 and 41; Fig. 4A, lanes 8 and 9). Mesoderm induction, a process in which FGF is known to play a pivotal role, takes place during blastula stages. Therefore, we examined er1 levels at 1-h time intervals during blastula and gastrula stages using a quantitative PCR assay (18). er1 expression levels were shown to increase 2-fold from early blastula (stages 7 and 8; Fig. 4B, lanes 1 and 2) to late blastula stages (stages 8 and 9; Fig. 4B,  lanes 3 and 4), followed by a 5-fold decrease at gastrulation (stage 10; Fig. 4B, lane 5).
Our sequence analysis revealed two putative nuclear localization signals, suggesting that ER1 is targeted to the nucleus. We investigated the subcellular localization of the ER1 protein using a polyclonal anti-ER1 antibody to stain transfected NIH 3T3 cells expressing ER1. This antibody, directed against a synthetic C-terminal peptide, recognizes full-length ER1 protein synthesized in vitro (Fig. 5A, lane 3) and specifically stains the nuclei of cells expressing ER1 (Fig. 5B). Cells transfected with the pcDNA3 vector alone (Fig. 5B) as well as pcDNA3-er1 transfected cells stained with preimmune serum (not shown) gave similar patterns and showed no specific nuclear staining.
The facts that ER1 is targeted to the nucleus and that its N terminus contains stretches of acidic residues characteristic of acidic activation domains (25) suggest that ER1 may function as a transcription factor. We investigated this possibility by testing the transactivation potential of various regions of the ER1 protein. Constructs containing different portions of er1 FIG. 6. The N terminus of ER1 functions as a transcriptional activator. NIH 3T3 cells were transiently transfected with various GAL4-ER1 fusion constructs along with a CAT reporter plasmid. After 48 h, CAT enzyme levels were measured as described under "Experimental Procedures." Vector denotes the control pM plasmid, containing only the GAL4 DNA binding domain, whereas the numbers indicate the amino acids of ER1 encoded by each construct. The value for each construct represents the fold activation relative to the pM plasmid, averaged from 3 to 12 independent transfections. fused to the GAL4 DNA binding domain were used along with a CAT reporter plasmid in transient transfections. Assays of CAT enzyme levels revealed that although full-length ER1 did not activate transcription, the N-terminal region (ER 1-175) containing all four acidic stretches (Fig. 1) stimulated transcription 10-fold (Fig. 6). The complementary C-terminal portion, ER 176 -493, on the other hand, had no transactivational activity. It is unclear why full-length ER1 was unable to stimulate transcription, but one possible explanation is that fusion of ER1 to GAL4 may alter the tertiary structure of the ER1 protein, affecting its activity. A similar observation was made with the ETS transcription factor ER81, which when fused to the GAL4 DNA binding domain lost its ability to activate transcription (28).
Interestingly, deletion of the N-terminal region to produce a construct containing only the first three acidic stretches (ER 1-98), resulted in a much more potent transactivator that stimulated transcription 80-fold (Fig. 6). This suggests that a negatively acting domain is located between amino acids 99 -176. Further truncation of the N terminus to generate ER 1-57 and ER 1-25 completely abolished transactivation. These results demonstrate that the ER1 protein contains regions with transcription transactivating activity and that ER1 has the potential to function as a transcription factor.