Functional analysis of a growth factor-responsive transcription factor complex

Identification of the functional components of the Ras signaling pathway regulating pituitary cell-specific gene expression

Ras, a small GTP-binding protein, is required for functional receptor tyrosine kinase signaling. Ultimately, Ras alters the activity of specific nuclear transcription factors and regulates novel patterns of gene expression. Using a rat prolactin promoter construct in transient transfection experiments, we show that both oncogenic Ras and activated forms of Raf-1 kinase selectively stimulated the cellular rat prolactin promoter in GH4 rat pituitary cells. We also show that the Ras signal is completely blocked by an expression vector encoding a dominant-negative Raf kinase. Additionally, using a molecular genetic approach, we determined that inhibitory forms of p42 mitogen-activated protein kinase and an Ets-2 transcription factor interfere with both the Ras and the Raf activation of the rat prolactin promoter. These findings define a functional requirement for these signaling constituents in the activation of the prolactin gene, a cell-specific gene which marks the lactotroph pituitary cell type. Further, this analysis allowed us to order the components in the Ras signaling pathway as it impinges on regulation of prolactin gene transcription as Ras-->Raf kinase-->mitogen-activated protein kinase-->Ets. In contrast, we show that intact c-Jun expression inhibited the Ras-induced activation of the prolactin promoter, defining it as a negative regulator of this pathway, whereas c-Jun was able to enhance the Ras activation of an AP-1-driven promoter in GH4 cells. These data show that c-Jun is not the nuclear mediator of the Ras signal for the highly specialized, pituitary cell-specific prolactin cellular promoter. Thus, we have defined a model system which provides an ideal paradigm for studying Ras/Raf signaling pathways and their effects on neuroendocrine cell-specific gene regulation.

Control of MAP kinase activation by the mitogen-induced threonine/tyrosine phosphatase PAC1

Intracellular signalling following mitogenic stimulation of quiescent cells involves the initiation of a phosphorylation cascade that leads to the rapid and reversible activation of the mitogen-activated protein (MAP) kinases ERK1 and ERK2. MAP kinase activation is mediated by dual phosphorylation within the motif Thr-Glu-Tyr by MAP kinase kinase (MEK). Following activation, the MAP kinases translocate into the nucleus where they phosphorylate several transduction targets, including transcription factors. We have previously identified PAC1 as an immediate-early mitogen-inducible tyrosine phosphatase in nuclei of T cells. Here we present several lines of evidence indicating that PAC1 is a physiologically relevant MAP kinase phosphatase. Recombinant PAC1 in vitro is a dual-specific Thr/Tyr phosphatase with stringent substrate specificity for MAP kinase. Constitutive expression of PAC1 in vivo leads to inhibition of MAP kinase activity normally stimulated by epidermal growth factor, phorbol myristyl acetate, or T-cell receptor crosslinking. The inactivation of MAP kinase by PAC1 results in inhibition of MAP kinase-regulated reporter gene expression.

Ternary complex factors: growth factor regulated transcriptional activators

Members of a family of Ets domain proteins, the ternary complex factors (TCFs), are recruited to the c-fos serum response element by interaction with the serum response factor. Recent findings indicate that phosphorylation of TCFs occurs in response to activation of the MAP kinase pathway, and that regulation of TCF activity is an important mechanism by which the serum response element responds to growth factor signals.

Real-time DNA binding measurements of the ETS1 recombinant oncoproteins reveal significant kinetic differences between the p42 and p51 isoforms

The sequence-specific DNA binding of recombinant p42 and p51 ETS1 oncoprotein was examined quantitatively to determine whether the loss of the Exon VII phosphorylation domain in p42 ETS1 or the phosphorylation of expressed Exon VII in p51 ETS1 had an effect on DNA binding activity. The kinetics of sequence-specific DNA binding was measured using real-time changes in surface plasmon resonance with BIAcore (registered trademark, Pharmacia Biosensor) technology. The real-time binding of p42 and p51 ETS1 displayed significant differences in kinetic behavior. p51 ETS1 is characterized by a fast initial binding and conversion to a stable complex, whereas p42 ETS1 exhibits a slow initial binding and conversion to a stable complex. All of the p51 ETS1 DNA binding states are characterized by rapid turnover, whereas the p42 ETS1 DNA binding states are 4-20 times more stable. A model describing these kinetic steps is presented. Stoichiometric titrations of either p42 or p51 ETS1 with specific oligonucleotides show 1:1 complex formation. The DNA sequence specificity of the p42 and p51 ETS1 as determined by mutational analysis was similar. The in vitro phosphorylation of p51 ETS1 by CAM kinase II obliterates its binding to specific DNA, suggesting that the regulation of p51 ETS1 sequence-specific DNA binding occurs through phosphorylation by a calcium-dependent second messenger. The p42 ETS1 lacks this regulatory domain (Exon VII), and binding to its specific DNA sequence is not sensitive to calcium signaling.

Role of protein kinase C in ras-mediated fos-expression

HC-11 mouse mammary epithelial cells stably transfected with a glucocorticoid-inducible Ha-ras construct encoding a transforming (val12) p21Ha-ras were cotransfected with a c-fos-CAT construct containing the human c-fos promoter up to position -711 and the CAT reporter gene. Expression of Ha-ras by dexamethasone leads to a transcriptional activation of the fos-CAT construct which was found to be sensitive to the PKC inhibitor ilmofosine (BM41440) and abrogated by PKC depletion following long-term exposure to TPA. The responsiveness to Ha-ras is retained if only the portion of the fos promoter covering the serum response element (SRE) and the adjacent fos AP-1 (FAP) site are put in front of a CAT gene linked to a thymidine kinase (TK) promoter. Further depletion of the FAP-site does not affect the inducibility by Ha-ras. Transcriptional activation of the SRE-FAP-TK-CAT as well as the SRE-TK-CAT constructs by Ha-ras is sensitive to the PKC-inhibitor ilmofosine (BM41440) and blocked by long-term exposure to TPA. Long-term exposure to TPA depletes cells of PKC alpha and significantly reduces the PKC epsilon levels. Long-term exposure in bryostatin 1 selectively depletes PKC alpha. Depletion of PKC alpha by bryostatin 1 does not reduce the transcriptional activation of the SRE-FAP-TK-CAT-construct by Ha-ras. It is concluded that (i) transforming Ha-ras induces c-fos in HC-11 cells via PKC (presumably epsilon), (ii) the signal is mediated to the serum response element (SRE) of the fos promoter and (iii) the fos AP-1 (FAP) site is not required for this mechanism.

Activation of ternary complex factor Elk-1 by MAP kinases

Ternary complex factors (TCFs), one of which is Elk-1, have been implicated in mediation of c-fos induction. They have been shown to be phosphorylated by mitogen-activated protein kinases (MAPKs) in vitro. We demonstrate that recombinant Elk-1 is hyperphosphorylated in vivo upon joint overexpression of MAPKs and constitutively activated Raf-1 kinase, the latter serving as an indirect in vivo activator of MAPKs. This phosphorylation is accompanied by a conformational change and results in an elevated transactivation potential of Elk-1. Mutation of mapped in vivo phosphorylation sites, which are potential targets for MAPKs, reduced Elk-1-mediated transcription. Thus, MAPKs are very probably controlling Elk-1 activity by direct phosphorylation in vivo. Furthermore, Elk-1 was shown to stimulate transcription from both the c-fos serum response element and also from an Ets binding site. While binding of TCFs to the c-fos promoter is dependent on the serum response factor, TCFs can autonomously interact with Ets binding sites. This indicates that TCFs may participate in the transcriptional regulation of two different sets of genes.

LEF-1 contains an activation domain that stimulates transcription only in a specific context of factor-binding sites

Lymphoid enhancer factor 1 (LEF-1) is a member of the high mobility group (HMG) family of proteins and participates in the regulation of the T cell receptor (TCR) alpha enhancer. We have previously shown that DNA binding by the HMG domain of LEF-1 induces a sharp bend in the DNA helix. Together with the dependence of LEF-1 on other factor-binding sites to regulate gene expression, DNA bending induced by the HMG domain suggested an 'architectural' role for LEF-1. In this study, we performed experiments to distinguish between a model in which the HMG domain is the only functional determinant of LEF-1 and a model in which additional domains of LEF-1 are involved in the regulation of gene expression. First, we show that the HMG domain alone is not sufficient to stimulate TCR alpha enhancer function. Second, we replaced the HMG domain of LEF-1 with the DNA-binding domain of the bacterial repressor LexA, which binds a specific nucleotide sequence without inducing a sharp bend in the DNA helix. The chimeric LEF-LexA protein increased the activity of a TCR alpha enhancer in which the LEF-1-binding site had been replaced with a LexA recognition sequence. Transcriptional stimulation by LEF-LexA, however, was less efficient than that observed with endogenous LEF-1. The LEF-LexA-mediated activation of gene expression was dependent upon an amino-terminal region of LEF-1 and a specific context of factor-binding sites in the TCR alpha enhancer. Neither multimerized LexA-binding sites, nor TCR alpha enhancers with altered spatial arrangements of factor-binding sites, were functional for regulation by LEF-LexA. Together, these data suggest that an aminoterminal region in LEF-1 contributes to the context-dependent regulation of the TCR alpha enhancer by LEF-1, presumably by interacting with other enhancer-bound proteins.

Phosphorylation of the c-Fos transrepression domain by mitogen-activated protein kinase and 90-kDa ribosomal S6 kinase

Phosphorylation of the C terminus of c-Fos has been implicated in serum response element-mediated repression of c-fos transcription after its induction by serum growth factors. The growth-regulated enzymes responsible for this phosphorylation in early G1 phase of the cell cycle and the sites of phosphorylation have not been identified. We now provide evidence that two growth-regulated, nucleus- and cytoplasm-localized protein kinases, 90-kDa ribosomal S6 kinase (RSK) and mitogen-activated protein kinase (MAP kinase), contribute to the serum-induced phosphorylation of c-Fos. The major phosphopeptides derived from biosynthetically labeled c-Fos correspond to phosphopeptides generated after phosphorylation of c-Fos in vitro with both RSK and MAP kinase. The phosphorylation sites identified for RSK (Ser-362) and MAP kinase (Ser-374) are in the transrepression domain. Cooperative phosphorylation at these sites by both enzymes was observed in vitro and reflected in vivo by the predominance of the peptide phosphorylated on both sites, as opposed to singly phosphorylated peptides. This study suggests a role for nuclear RSK and MAP kinase in modulating newly synthesized c-Fos phosphorylation and downstream signaling.

A single amino-acid substitution in the Ets domain alters core DNA binding specificity of Ets1 to that of the related transcription factors Elf1 and E74

Ets proteins form a family of sequence specific DNA binding proteins which bind DNA through a 85 aminoacids conserved domain, the Ets domain, whose sequence is unrelated to any other characterized DNA binding domain. Unlike all other known Ets proteins, which bind specific DNA sequences centered over either GGAA or GGAT core motifs, E74 and Elf1 selectively bind to GGAA corecontaining sites. Elf1 and E74 differ from other Ets proteins in three residues located in an otherwise highly conserved region of the Ets domain, referred to as conserved region III (CRIII). We show that a restricted selectivity for GGAA core-containing sites could be conferred to Ets1 upon changing a single lysine residue within CRIII to the threonine found in Elf1 and E74 at this position. Conversely, the reciprocal mutation in Elf1 confers to this protein the ability to bind to GGAT core containing EBS. This, together with the fact that mutation of two invariant arginine residues in CRIII abolishes DNA binding, indicates that CRIII plays a key role in Ets domain recognition of the GGAA/T core motif and lead us to discuss a model of Ets proteins--core motif interaction.

p21ras and calcineurin synergize to regulate the nuclear factor of activated T cells

In T lymphocytes, triggering of the T cell receptor (TCR) induces several signaling cascades which ultimately synergize to induce the activity of the nuclear factor of activated T cells (NFAT), a DNA binding complex critical to the inducibility and T cell specificity of the T cell growth factor interleukin 2. One immediate consequence of T cell activation via the TCR is an increase in cytosolic calcium. Calcium signals are important for NFAT induction, and recent studies have identified calcineurin, a calcium-calmodulin dependent serine-threonine phosphatase, as a prominent component of the calcium signaling pathway in T cells. A second important molecule in TCR signal transduction is the guanine nucleotide binding protein, p21ras, which is coupled to the TCR by a protein tyrosine kinase dependent mechanism. The experiments presented here show that expression by transfection of mutationally activated calcineurin or activated p21ras alone is insufficient for NFAT transactivation. However, coexpression of the activated calcineurin with activated p21ras could mimic TCR signals in NFAT induction. These data identify calcineurin and p21ras as cooperative partners in T cell activation.

A growth factor-induced kinase phosphorylates the serum response factor at a site that regulates its DNA-binding activity

A signaling pathway by which growth factors may induce transcription of the c-fos proto-oncogene has been characterized. Growth factor stimulation of quiescent fibroblasts activates a protein kinase cascade that leads to the rapid and transient phosphorylation of the serum response factor (SRF), a regulator of c-fos transcription. The in vivo kinetics of SRF phosphorylation and dephosphorylation parallel the activation and subsequent repression of c-fos transcription, suggesting that this phosphorylation event plays a critical role in the control of c-fos expression. The ribosomal S6 kinase pp90rsk, a growth factor-inducible kinase, phosphorylates SRF in vitro at serine 103, the site that becomes newly phosphorylated upon growth factor stimulation in vivo. Phosphorylation of serine 103 significantly enhances the affinity and rate with which SRF associates with its binding site, the serum response element, within the c-fos promoter. These results suggest a model in which the growth factor-induced phosphorylation of SRF at serine 103 contributes to the activation of c-fos transcription by facilitating the formation of an active transcription complex at the serum response element.

Conditional transformation of cells and rapid activation of the mitogen-activated protein kinase cascade by an estradiol-dependent human raf-1 protein kinase

We report a strategy for regulating the activity of a cytoplasmic signaling molecule, the protein kinase encoded by raf-1. Retroviruses encoding a gene fusion between an oncogenic form of human p74raf-1 and the hormone-binding domain of the human estrogen receptor (hrafER) were constructed. The fusion protein was nontransforming in the absence of estradiol but could be reversibly activated by the addition or removal of estradiol from the growth media. Activation of hrafER was accompanied in C7 3T3 cells by the rapid, protein synthesis-independent activation of both mitogen-activated protein (MAP) kinase kinase and p42/p44 MAP kinase and by phosphorylation of the resident p74raf-1 protein as demonstrated by decreased electrophoretic mobility. The phosphorylation of p74raf-1 had no effect on the kinase activity of the protein, indicating that mobility shift is an unreliable indicator of p74raf-1 enzymatic activity. Removal of estradiol from the growth media led to a rapid inactivation of the MAP kinase cascade. These results demonstrate that Raf-1 can activate the MAP kinase cascade in vivo, independent of other "upstream" signaling components. Parallel experiments performed with rat1a cells conditionally transformed by hrafER demonstrated activation of MAP kinase kinase in response to estradiol but no subsequent activation of p42/p44 MAP kinases or phosphorylation of p74raf-1. This result suggests that in rat1a cells, p42/p44 MAP kinase activation is not required for Raf-1-mediated oncogenic transformation. Estradiol-dependent activation of p42/p44 MAP kinases and phosphorylation of p74raf-1 was, however, observed in rat1a cells expressing hrafER when the cells were pretreated with okadaic acid. This result suggests that the level of protein phosphatase activity may play a crucial role in the regulation of the MAP kinase cascade. Our results provide the first example of a cytosolic signal transducer being harnessed by fusion to the hormone-binding domain of the estrogen receptor. This conditional system not only will aid the elucidation of the function of Raf-1 but also may be more broadly useful for the construction of conditional forms of other kinases and signaling molecules.

Activation of the inducible orphan receptor gene nur77 by serum growth factors: dissociation of immediate-early and delayed-early responses

We have characterized the genetic elements that mediate the transcriptional activation of nur77, a growth factor-inducible gene encoding a member of the steroid/thyroid hormone receptor superfamily. Although initially identified as a serum-inducible immediate-early gene with expression kinetics similar to those of c-fos, we found that transcriptional activation of nur77 by serum growth factors in fibroblasts is in fact composed of two components: an immediate-early component, which can occur in the absence of de novo protein synthesis, and a delayed-early component, which is dependent on de novo protein synthesis. The expression of nur77 following serum stimulation reflects the superimposition of immediate-early and delayed-early expression. Immediate-early and delayed-early expression can be dissociated from one another by deletion or base substitution mutations of the nur77 promoter. Immediate-early expression of nur77 is mediated primarily by sequences located between nucleotides -86 and -126 upstream of the transcription start site. This region includes a sequence that resembles but differs from the CArG element found in other serum-inducible promoters. Upstream of the CArG-like element is a potential binding site for a transcription factor of the Ets family; the presence of this site is required for significant transcriptional induction. Delayed-early expression of nur77 is mediated by multiple AP-1-like and GC-rich elements, which can interact with products of immediate-early genes such as Fos/Jun and Zif268, respectively. Furthermore, we show that Zif268 can activate transcription of the nur77 promoter, suggesting that it may play a role in the delayed-early expression of nur77.

The protein kinase mos activates MAP kinase kinase in vitro and stimulates the MAP kinase pathway in mammalian somatic cells in vivo

The mos protooncogene encodes a serine/threonine protein kinase that is only expressed at significant levels in germ cells. Recombinant malE-mos protein (Xenopus mos protooncogene fused in frame to the maltose binding protein of E. coli) activates MAP kinase in cell-free extracts prepared from Xenopus oocytes and eggs. Here we show that malE-mos immunoprecipitates from Xenopus extracts phosphorylate and activate MAP kinase kinase in vitro, indicating that mos can function as a MAP kinase kinase kinase. Moreover, ectopic expression of mos in mammalian somatic cells, that lack any endogenous mos protein, triggers the activation of MAP kinase in vivo. These results identify the mos protooncogene as a direct activator of the MAP kinase pathway, with the potential to activate this kinase cascade even in cells where normally there is no expression of mos.

A growth factor-induced kinase phosphorylates the serum response factor at a site that regulates its DNA-binding activity

A signaling pathway by which growth factors may induce transcription of the c-fos proto-oncogene has been characterized. Growth factor stimulation of quiescent fibroblasts activates a protein kinase cascade that leads to the rapid and transient phosphorylation of the serum response factor (SRF), a regulator of c-fos transcription. The in vivo kinetics of SRF phosphorylation and dephosphorylation parallel the activation and subsequent repression of c-fos transcription, suggesting that this phosphorylation event plays a critical role in the control of c-fos expression. The ribosomal S6 kinase pp90rsk, a growth factor-inducible kinase, phosphorylates SRF in vitro at serine 103, the site that becomes newly phosphorylated upon growth factor stimulation in vivo. Phosphorylation of serine 103 significantly enhances the affinity and rate with which SRF associates with its binding site, the serum response element, within the c-fos promoter. These results suggest a model in which the growth factor-induced phosphorylation of SRF at serine 103 contributes to the activation of c-fos transcription by facilitating the formation of an active transcription complex at the serum response element.

Activation of the inducible orphan receptor gene nur77 by serum growth factors: dissociation of immediate-early and delayed-early responses

We have characterized the genetic elements that mediate the transcriptional activation of nur77, a growth factor-inducible gene encoding a member of the steroid/thyroid hormone receptor superfamily. Although initially identified as a serum-inducible immediate-early gene with expression kinetics similar to those of c-fos, we found that transcriptional activation of nur77 by serum growth factors in fibroblasts is in fact composed of two components: an immediate-early component, which can occur in the absence of de novo protein synthesis, and a delayed-early component, which is dependent on de novo protein synthesis. The expression of nur77 following serum stimulation reflects the superimposition of immediate-early and delayed-early expression. Immediate-early and delayed-early expression can be dissociated from one another by deletion or base substitution mutations of the nur77 promoter. Immediate-early expression of nur77 is mediated primarily by sequences located between nucleotides -86 and -126 upstream of the transcription start site. This region includes a sequence that resembles but differs from the CArG element found in other serum-inducible promoters. Upstream of the CArG-like element is a potential binding site for a transcription factor of the Ets family; the presence of this site is required for significant transcriptional induction. Delayed-early expression of nur77 is mediated by multiple AP-1-like and GC-rich elements, which can interact with products of immediate-early genes such as Fos/Jun and Zif268, respectively. Furthermore, we show that Zif268 can activate transcription of the nur77 promoter, suggesting that it may play a role in the delayed-early expression of nur77.

Conditional transformation of cells and rapid activation of the mitogen-activated protein kinase cascade by an estradiol-dependent human raf-1 protein kinase

We report a strategy for regulating the activity of a cytoplasmic signaling molecule, the protein kinase encoded by raf-1. Retroviruses encoding a gene fusion between an oncogenic form of human p74raf-1 and the hormone-binding domain of the human estrogen receptor (hrafER) were constructed. The fusion protein was nontransforming in the absence of estradiol but could be reversibly activated by the addition or removal of estradiol from the growth media. Activation of hrafER was accompanied in C7 3T3 cells by the rapid, protein synthesis-independent activation of both mitogen-activated protein (MAP) kinase kinase and p42/p44 MAP kinase and by phosphorylation of the resident p74raf-1 protein as demonstrated by decreased electrophoretic mobility. The phosphorylation of p74raf-1 had no effect on the kinase activity of the protein, indicating that mobility shift is an unreliable indicator of p74raf-1 enzymatic activity. Removal of estradiol from the growth media led to a rapid inactivation of the MAP kinase cascade. These results demonstrate that Raf-1 can activate the MAP kinase cascade in vivo, independent of other "upstream" signaling components. Parallel experiments performed with rat1a cells conditionally transformed by hrafER demonstrated activation of MAP kinase kinase in response to estradiol but no subsequent activation of p42/p44 MAP kinases or phosphorylation of p74raf-1. This result suggests that in rat1a cells, p42/p44 MAP kinase activation is not required for Raf-1-mediated oncogenic transformation. Estradiol-dependent activation of p42/p44 MAP kinases and phosphorylation of p74raf-1 was, however, observed in rat1a cells expressing hrafER when the cells were pretreated with okadaic acid. This result suggests that the level of protein phosphatase activity may play a crucial role in the regulation of the MAP kinase cascade. Our results provide the first example of a cytosolic signal transducer being harnessed by fusion to the hormone-binding domain of the estrogen receptor. This conditional system not only will aid the elucidation of the function of Raf-1 but also may be more broadly useful for the construction of conditional forms of other kinases and signaling molecules.

Ras-independent growth factor signaling by transcription factor tyrosine phosphorylation

Interferons induce transcriptional activation through tyrosine phosphorylation of the latent, cytoplasmic transcription factor interferon-stimulated gene factor-3 (ISGF-3). Growth factors and cytokines were found to use a similar pathway: The 91-kilodalton subunit of ISGF-3 was activated and tyrosine phosphorylated in response to epidermal growth factor (EGF), platelet-derived growth factor, and colony stimulating factor-1. The tyrosine phosphorylated factor acquired DNA binding activity and accumulated in nuclei. Activation required the major sites for autophosphorylation on the EGF receptor that bind Src homology region 2 domain-containing proteins implicated in Ras activation. However, activation of this factor was independent of the normal functioning of Ras.

Isolation and characterization of SRF accessory proteins

Many genes which are regulated by growth factors contain a common regulatory element, the serum response element (SRE). Activation of transcription by the SRE involves a ternary complex formed between a ubiquitous factor, serum response factor (SRF), and a second protein, p62/TCF. We used a yeast genetic screen to isolate cDNAs encoding a protein, SAP-1, with the DNA binding properties of p62/TCF. The SAP-1 sequence contains three regions of homology to the previously uncharacterized Elk-1 protein, which also acts as an SRF accessory protein. Only two of these regions are required for cooperative interactions with SRF in the ternary complex. The third contains several conserved sites for the MAP kinases, whose activity is regulated in response to growth factor stimulation. We discuss the potential role of these proteins in regulation of the c-fos SRE.

The SRF accessory protein Elk-1 contains a growth factor-regulated transcriptional activation domain

The Elk-1 and SRF transcription factors form a ternary complex at the c-fos serum response element (SRE). Growth factor stimulation rapidly induces a reversible change in the electrophoretic mobility of the ternary complex, accompanied by increased phosphorylation of the Elk-1 C-terminal region and by the activation of a 42 kd cellular Elk-1 kinase. Phosphorylation of Elk-1 in vitro by partially purified p42/p44 MAP kinase induces a similar reduction in ternary complex mobility but has little effect on the efficiency of its formation. In vitro, MAP kinase phosphorylates the Elk-1 C-terminal region at multiple sites, which are also phosphorylated following growth factor stimulation in vivo. The Elk-1 C-terminal region functions as a regulated transcriptional activation domain whose activity in vivo is dependent on the integrity of the MAP kinase sites. These findings directly link transcriptional activation by the SRE to the growth factor-regulated phosphorylation of an SRE-binding protein.