A rat homolog of the Drosophila enhancer of split (groucho) locus lacking WD-40 repeats.

In Drosophila, neurogenic loci function in defining cellular fate by interpreting the identity of other cells in the immediate environment. To begin studies of mammalian homologs of these genes, we have isolated two rat homologs of the neurogenic locus Enhancer of split. The protein encoded by the Drosophila Enhancer of split locus is complex and contains five distinct regions based on amino acid composition. One region contains six WD-40 repeats, which were first described in the beta subunit of the heterotrimeric guanine nucleotide-binding protein. One of the rat cDNAs we isolated, R-esp1, encodes a novel form that lacks the WD-40 repeating units. Data are presented demonstrating that the R-esp1 cDNA is a full-length clone encoding an expressed 24-kDa protein. Antibodies raised against this protein stain the nucleus of both PC-12 and GH3 cells. The second clone, R-esp2, encodes a full-length homolog containing WD-40 repeats. The hydrodynamic properties of in vitro translated R-esp1 and R-esp2 proteins indicate that they do not stably self-associate or form heterodimers. A model is presented for the possible role of the R-esp1 protein in the negative regulation of Enhancer of split proteins containing WD-40 repeats.

The fate of cells during development can be affected by the identity of surrounding cells. In the fly, neurogenic gene products play important roles in defining cellular fate by interpreting the identity of other cells in the immediate environment (for review, see Artavanis-Tsakonas et al., 1991). Mutations in each of these loci cause most ventral ectodermal cells to become neuroblasts with little or no formation of epidermis. Hence, the phenotype of mutants is one of neuronal hypertrophy and epidermal hypotrophy. The neuronal hypertrophy does not arise from unregulated growth of the neuroblasts. Instead, cells normally destined to become epidermis are diverted to become neuroblasts. The neurogenic genes not only affect fate decisions in ectodermal cells but also affect mesodermal and endodermal development (Corbin et al., 1991;Hartenstein et al., 1992). At least six different neurogenic genes have been defined by mutations that affect Drosophila embryonic development. These include: Enhancer of split, Notch, Delta, mastermind, big brain and neuralized. Classical genetic approaches in Drosophila were used to identify this group of genes but such studies are The costs of publication of this article were defrayed in part by the * This work was supported by American Cancer Society Grant CD496. payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
T h nucleotide sequence(s) reported in this paper has been submitted (R-espl) and L14463 (R-esp2) to the GenBankmIEMBL Data Bank with accession numbeds) L14462.
Brigham and Women's Hospital, 75 Francis St., Boston, MA02115. %I.: t To whom correspondence should be addressed: Cardiology Division,  difficult to conduct in mammals. The cloning of mammalian homologs of Drosophila genes involved in determination of cell fate may allow identification of mammalian genes that play a similar role.
We have used this approach to study vertebrate homologs of the neurogenic class of Drosophila genes. This paper focuses on identification of mammalian homologs of the Drosophila neurogenic gene Enhancer of split (E(sp1)). In Drosophila the E(sp1) locus is complicated, giving rise to two distinct classes of proteins which are the product of multiple genes (Hartley et al., 1988;Klambt et al., 1989). The E(sp1)-HLH class encodes proteins containing the helix-loop-helix motif characteristic of gene regulatory proteins. The E(sp1)-WD class, also known as m9110 orgroucho, encodes a protein that contains a series of six WD-40 repeating units. The WD-40 repeat typically ends in the sequence tryptophan (W)-aspartate (Dl, and the consensus repeat is approximately 40 amino acids long (for review, see Duronio et al., 1992). This repeating unit was first recognized in the guanine nucleotide-binding protein p subunit, but it has been found in at least 20 other regulatory proteins, including the transcription regulatory factor TUP1 (Williams and Trumbly, 19901, the splicing factor PRP4 (Bjorn et al., 1989), and cleavage stimulation factor which modulates polyadenylation (Takagaki and Manley, 1992). Although the function of the WD-40 repeat is still uncertain, the diversity of proteins that contain this motif suggest that its presence does not necessarily confer participation in signal transduction pathways mediated by G-proteins. In addition to the conserved WD-40 repeating units, full-length E(sp1)-WD proteins contain two other conserved regions. The first spans the amino-terminal 120 amino acids and is rich in glutamine. The second spans the central region of the protein, contains a nuclear localization signal, and is rich in proline, serine, and glycine.
We have isolated two different cDNAs encoding rat E(sp1)-WD homologs which we term R-espl and R-esp2. Both clones show great similarity to both the Drosophila E(spl)-WD gene product (Hartley et al., 1988) and the four human E(sp1)-WD products (Stifani et al., 1992). However, R-espl encodes a novel member of this gene family lacking the WD-40 repeating units.
No equivalent to R-espl has been described in Drosophila. R-esp2 encodes a full-length E(sp1)-WD homolog containing all three domains.

MATERIALS AND METHODS
Cloning Rat E(sp1)-WD Homologs-A rat hippocampal cDNA library in A@-10 (Diyao Zhao and William Chin, Harvard Medical School) was screened with an SspI-XbaI fragment isolated from the Drosophila m9/10 cDNA (C. Delidakis and S. Artavanis-Tsakonas, Yale University) using standard procedures (Maniatis et al., 1982). The fragment used as probe contains 250 nucleotides of 5"noncoding and the first 180 codons of the Drosophila E(sp1)-WD cDNA. Filters were hybridized and washed under reduced stringency conditions as described in Schmidt et al. (1989).
DNA Sequencing and Analysis-DNA sequence was determined using the dideoxy chain termination technique (Sanger et al., 1977) and specific DNA primers. Computer analysis was done on a Vax using the 25681 General Computer Group sequence analysis software (Devereux et al., 1984).
RNA Blot Analysis--Total RNA was isolated from tissue by guanidinium thiocyanate extraction followed by phenolhhloroform extraction (Chomczynski and Sacchi, 1987). RNA was fractionated by electrophoresis through a formaldehyde-agarose gel. The RNAwas transferred to a nitrocellulose filter for hybridization with the full-length R-espl cDNA. DNA was labeled with [32PldCTF' by the method of Feinberg and Vogelstein (1983). Hybridization was carried out at 45 "C in a solution containing 6 x SSC, 50% formamide, 5 x Denhart's solution, 0.2% SDS. ARer hybridization for 18 h, the blot was washed in 0.2 x SSC, 0.5% SDS at 60 "C.
Immunoprecipitation-Rabbits were immunized with peptide corresponding to residues 12-26 of R-espl coupled to keyhole limpet hemocyanin. For metabolic labeling, PC-12 cells were preincubated for 1 h in media lacking methionine. 500 pCi of [36Slmethionine (Amersham Corp.) was added per 100-mm dish, and incubation was continued for 4 h at 37 "C. The cells were then washed three times with 1 x PBS' and then lysed in EBC buffer (50 m~ Tris, pH 8.0, 120 m~ NaCl, 0.5% Nonidet P-40, 1 pg/ml leupeptin, 1 pdml pepstatin, 1 pg/ml soybean and lima bean protease inhibitors). Following addition of EBC, the plates were incubated at 4 "C for 15 min and then the lysate was scraped into a 50-ml Falcon tube. Debris was removed by centrifugation at 2000 x g for 20 mi n, and the supernatant was concentrated to 2 ml by centrifugation through a Centricon 30 microconcentrator. For immunoprecipitation, 450 pl of IB (25 m~ Tris, pH 7.8, 150 m~ NaCl, 4 m~ EDTA, 0.1% SDS, 0.5% deoxycholate, 1% Nonidet P-40 was added to 50 pl of the concentrated supernatant, and the sample was incubated at 95 "C for 2 min. ARer cooling, antibody and a 50% slurry of Sepharose beads suspended in IB was added, and the sample was mixed for 2 h at room temperature. The Sepharose was removed by centrifugation, and protein A-Sepharose was added to the supernatant followed by incubation overnight at 4 "C. Precipitated material was pelleted by centrifugation at 14,000 rpm in a microcentrifuge, and the pellet was then washed once in IB followed by two washes in 25 m~ M s , pH 7.8,150 m~ NaCl, 4 m~ EDTA and two washes in 50 m~ Tris, pH 7.8. The washed pellet was resuspended in loading buffer (Laemmli, 1970) and analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography.
Immunohistochemical Staining-Cells were grown on laminincoated glass coverslips. Cells were fixed in fresh 4% paraformaldehyde prepared in 125 m~ phosphate buffer, pH 7.0, containing 5% sucrose for 3 min followed by two washes with PBS. This fixation and wash procedure was carried out twice followed by permeabilization for 3 min in PBS containing 0.1% Triton X-100. Blocking was done in PBS containing 1% bovine serum albumin, 3% goat serum, 3% horse serum, and 0.05% NaN3. A 1:50 dilution of affinity-purified anti-R-espl antibody in blocking solution was added, and incubation was continued overnight followed by three washes with blocking buffer. Fluoroscein-conjugated goat anti-rabbit IgG (Cappel) prepared in blocking buffer was added and incubation continued for 1 h. Cells were then washed three times with blocking buffer followed by three washes with PBS. Coverslips were then mounted in AFT system mounting media (Hoechst-Japan) and viewed by epifluorescence on a Bio-Rad MRCGOO confocal microscope.
Physical Properties of R-esp Proteins-Stokes radii and sedimentation coefficients were determined as described in Schmidt and Neer (1991). Stokes radii were determined by gel filtration over either Ultrogel AcA 34 or AcA 22 (IBF Inc.) equilibrated with 50 m~ Tris-HC1, pH 7.5,l m~ EDTA, 1 m~ dithiothreitol (TED) and 0.3% Lubrol F' X. Markers used were @galactosidase (Stokes radius = 69 A), fumerase (Stokes radius = 52 A), bovine serum albumin (Stokes radius = 37 A), carbonic anhydrase (Stokes radius = 24 A), and cytochrome c (Stokes radius = 17 A). Fractions from the column that contained in vitro translated R-esp products were identified by SDS-gel electrophoresis followed by autoradiography. Elution profiles were analyzed by scanning densitometry of the autoradiogram, and Stokes radii were determined by comparison with the elution positions of the added markers of known Stokes radius.

Rat E(sp1)-WD
Homologs-We have used the cDNA encoding the Drosophila E(sp1)-WD locus to identify rat homologs from an adult hippocampal cDNA library. Two of these cDNAs have been sequenced revealing that one clone, R-espl, encodes a novel member of the E(sp1)-WD family, whereas the other, R-esp2, encodes a protein that is highly similar to the Drosophila E(sp1)-WD gene product. A sequence comparison of R-espl and R-esp2 with the Drosophila E(sp1)-WD protein is shown in Fig.  1. The Drosophila E(sp1)-WD protein is 720 amino acids long and the six WD repeats are found in the carboxyl-terminal region of the protein. The first clone, Resp-1, encodes a protein of only 197 amino acids which is similar to the amino terminus of the DrosophiZa E(spl)-WD protein. Data presented below indicates that R-espl is a full-length cDNA clone. The second clone, R-esp2, encodes a 741-amino acid protein, which, like Drosophila E(sp1)-WD, contains six WD-40 repeats at the carboxyl terminus (Fig. 2).
Stifani et al. (1992) reported the sequence of four human homologs of the E(sp1)-WD locus which they term transducinlike enhancer of split (TLE11-4. Comparison of the human and Drosophila sequences defined five different regions based on amino acid composition: an amino-terminal region rich in glutamine (Q); a glycindproline-rich region (GP); a CcN region which contains a nuclear localization signal along with potential casein kinase I1 and cdc2 phosphorylation sites; a prolineserine (PS)-rich region; and the six WD-40 repeats which make up the carboxyl-terminal region of the E(sp1)-WD homologs (Fig. 3). R-espl contains only the amino-terminal glutamine and glycine/proline-rich regions. R-esp2 contains all five of these regions.
The five regions defined above were based upon amino acid composition. Another way to compare and classify proteins is to identify regions where the amino acid sequences have been conserved. This is particularly important when comparing sequences that contain regions rich in particular amino acids such as glutamine, proline, glycine, or serine. For example, it is possible for two proteins to be similar merely because they both contain a high glutamine content. However, if the two proteins are homologs, residues other then glutamine would also be conserved. The Pileup and Plotsimilarity programs of the Genetics Computer Group sequence analysis package (Devereux et al., 1987) allows a graphical presentation of the sequence identity between R-espl, R-esp2, D-esp, and the human E(sp1)-WD homologs, TLE 1-4 (Fig. 3). This analysis indicates that these proteins contain three discrete domains based upon the extent of amino acid identity. Domain I extends from the amino terminus to approximately amino acid 120 and corresponds to the Q-rich region. The level of conservation drops in domain I1 which Corresponds to amino acids 120-450 and spans the GP, CcN, and PS regions. Inspection reveals that the high proline, glycine, and serine content is best conserved in domain 11. Many other amino acids have been substituted within this domain. Domain I11 corresponds to the region containing the six WD-40 repeating units. The highest degree of amino acid identity between these proteins is within domains I and 111, suggesting that these two may play an important role in E(sp1)-WD structure or function. Inspection of domain I reveals that the linear sequence of this region, including those amino acids that are not glutamine, are well conserved (Fig. 1, underlined region). This is important to the classification of R-espl as a member of the E(sp1)-WD family, even though it lacks WD-40 repeats. This classification would not be justified if R-espl was related to other E(sp1)-WD proteins solely on the basis of high glutamine content.
Only one E(sp1)-WD gene has been identified in Drosophila, whereas mammalian species appear to contain multiple homologs. Comparison of the amino acid and nucleotide sequences (Tables I and 11) of the rat, Drosophila, and human sequences reveals that the mammalian sequences are more similar to one another than to the Drosophila E(sp1)-WD protein. These observations indicate an amplification of the E(sp1)-WD loci occurred after the divergence of the arthropods and chordates. At the nucleotide level (Table 11) R-espl is most similar to an expressed sequence tag (EST002561 which was identified by randomly selecting and sequencing clones from a human cDNA library (Adams et al., 1991). In addition, R-espl shows greater similarity to the amino-terminal region of the rat and human E(sp1)-WD homologs than to the Drosophila E(sp1)-WD gene product. Apparently, the R-espl gene also arose after the divergence of arthropods from chordates, suggesting that an E(sp1) gene exactly homologous to R-espl may not exist in Drosophila. Given the similarity of R-espl with domain I of the E(spl)-WD gene family, it is likely that R-espl arose in chordates by mutation of a full-length E(sp1)-WD gene. R-esp2 shows 98% amino acid identity with the human E(sp1-WD) homolog termed TLE4 (Table I). This human cDNA is only a partial clone, so it is unclear if this great similarity extends throughout the entire protein.
Northern Blot Analysis of R-espl-To determine if R-espl is expressed in adult tissues, total RNA was extracted from heart, brain, lung, and kidney, electrophoresed through a formaldehyde-agarose gel, and transferred to nitrocellulose filters. The filters were hybridized with [32PldCTP-labeled R-espl and then washed at high stringency. R-espl anneals to a 1.3-kilobase mRNA that is expressed in heart, brain, lung, and kidney (Fig. 4). The length of the mRNA is similar to that of the cDNA (1350 base pairs), suggesting that the cDNA clone is probably full length. This supports the idea that R-espl is a unique clone rather than a truncation artifact of a larger cDNA. The level of & A s varied among tissues. Typically, the R-espl mRNA

FIG. 3. Conservation of E(spl)-WD homologs. The top shows a stick diagram of the R-esp clones indicating the consewed amino acid regions defined by Stifani et al. (1992). Q = glutamine-rich; GP
= glycindproline-rich region; CCN = nuclear localization signal, casein kinaseIYcdc2 kinase site; PS = proline/ serine-rich region; and WD-40, which contains a series of six WD = 40 repeating units. The plot shows the identity score assigned for a comparison of the available rat, human, and Drosophila E(sp1)-WD proteins using a sliding window of 10 amino acids. Three domains, I , II, and III, can be recognized based upon the extent of conservation. Both domains I and I11 contain regions that are conserved perfectly (score = 1) between the rat, human, and Drosophila sequences. Domain I1 is less well conserved. TLE4 is a partial cDNA clone which spans the coding region from the prolindserine-rich region through the six WD-40 repeating units. Harley et al. 1988).

TABLE I1 Comparison of the nucleotide sequence of Efspl)-WD homologs
Sequences were aligned and then the extent of nucleotide identity was determined using the Distances program of the Wisconsin genetics computer group sequence analysis software package. References are given in Table I  level was %fold higher in kidney than lung and 2-fold higher in kidney than heart or brain.
Immunoprecipitation-Antipeptide antibodies have been raised against R-espl and have been used to show that the R-espl cDNA encodes an expressed protein of the size predicted from the nucleotide sequence. The R-espl peptide, spanning amino acids 12-26, was chosen from a region that has only 7 out of 14 residues identical to R-esp2. The R-espl antisera have been tested for their ability to immunoprecipitate protein from either in vitro translated R-espl or from proteins extracted from PC-12 cells that were metabolically labeled with [35S]me-thionine (Fig. 5). In vitro translation of R-espl yields a 24-kDa protein that is immunoprecipitated by the anti-R-espl antisera but not by preimmune sera. In addition, this antiserum does not immunoprecipitate in vitro translated R-esp2 (not shown). One protein is specifically precipitated from the [35Slmethionine-labeled PC-12 cell extract and comigrates with the in vitro translated R-espl. Precipitation of this 24-kDa protein was blocked by the peptide used for immunization and it was not precipitated by preimmune serum. The data show that R-espl encodes an expressed protein of the size predicted from the cDNA.
Immunochemical Localization-The full-length Drosophila and the human E(sp1)-WD homologs contain nuclear localization signals and are found in the nucleus (Hartley et al., 1989;Stifani et al., 1992). However, R-espl lacks a recognizable nuclear localization signal, making it difficult to predict the subcellular location of this protein. To determine where R-espl protein is found in the cell, the anti-R-espl antibodies have been used for immunohistochemical localization in two different rat cell lines: PC-12, a pheochromocytoma line, and GH3, a line established from an anterior pituitary tumor. These cell lines were chosen because they are both derived from ectoderm, a germ layer which is clearly affected by mutations in the Drosophila E(sp1)-WD gene. Furthermore, PC-12 cells can be induced to differentiate into neuronal-type cells with nerve growth factor (NGF). Immunohistochemical staining was visualized using a fluorescein-conjugated goat anti-rabbit IgG as a secondary antibody. As seen in Fig. 6, anti-R-espl stains the nucleus of these cells. This staining is absent when antisera is first incubated with peptide used to raise the antibody (Fig. 6, control), when primary antibody was omitted or when preimmune sera was used (not shown). The staining pattern within the nuclei of the PC-12 cells had a reticulated or sponge-like appearance. Optical sectioning by confocal microscopy indicates that this pattern extends throughout the nucleus. NGFtreated PC-12 cells that had extended neurite processes have a more diffise staining appearance, although regions lacking R-espl Immunoprecipitation PC 12 1 2 3 4 5 6 7 FIG. 5. Immunoprecipitation of R-espl. Samples were precipitated with protein A-Sepharose and then applied to an SDS-polyacrylamide gel and the results determined by autoradiography. Lane 1, in vitro translated I"KSlmethionine-labeled R-espl. R-espl is indicated with an arrowhead. This sample was not immunoprecipitated. Lanes 2-4, protein samples were extracted from [""Slmethionine-labeled PC-12 cells and then precipitated as follows: lane 2, preimmune serum; lane 3, anti-r-espl immune serum; lane 4, anti-R-espl immune serum that was preincubated with peptide that was used as antigen. Lanes 5-7, in vitro translated [35Slmethionine-labeled R-espl. Lane 5, preimmune serum. Lane 6, anti-r-espl immune serum; lane 7, anti-R-espl immune serum that was preincubated with peptide that was used as antigen.
staining are still visible. GH3 cells also exhibit a more diffuse R-espl nuclear staining than the undifferentiated PC-12 cells. Identical fmation and staining protocols were used on these cells. The different pattern seen in the NGF-treated PC-12 cells and GH3 cells argues that the reticulated pattern seen in untreated PC-12 cells is not an artifact of the furation or staining protocol.

Physical Characteristics of in Vitro Danslated R-esp Proteins
" T h e ability of proteins to interact plays a fundamental role in determining the function of the resulting protein complex. To determine whether R-espl or R-esp2 are capable of forming homomultimers or whether they may directly interact to form a heterocomplex, we characterized the physical state of the R-esp proteins. R-espl and R-esp2 proteins were translated in vitro using the rabbit reticulocyte lysate system, and their apparent molecular weight was determined by a combination of gel filtration and sucrose density gradient centrifugation (see Schmidt and Neer, 1991). The results are summarized in Table  111. R-espl behaves as a uniform population of molecules and the physical parameters indicate that in vitro translated Respl is a monomer that forms extended rods. This is based on comparison of the observed sedimentation coefficients and Stokes radii with those of globular proteins. For example, Respl has a s20. u) similar to that of the cytochrome c which has a molecular weight of 13,370. In contrast, the Stokes radius of R-esp 1 is similar to that of glyceraldehyde-3-phosphate dehydrogenase, which has a molecular weight of 145,000. These differences indicate that the in vitro translated proteins do not fold to a globular structure and suggest an extended rod like shape. Immunochemical localization of R-espl in PC-I2 cells, NGF-treated PC-I2 cells, and GH3 cells. Cells were stained with anti-R-espl antibody followed by fluorescein-conjugated goat anti-rabbit IgG. The pattern of staining within the nucleus of the PC-I2 cells appears reticulated, whereas that of NGF-treated PC-I2 and GH3 cells is more diffuse. CONTROL, undifferentiated PC-12 cell control. Antiserum was first incubated with the antigenic peptide before application to the cells. This exposure was three times longer than for the other panels. The peptide blocks the nuclear staining but does not completely block cytoplasmic staining. Nuclear immunostaining of NGF-treated PC-I2 cells and GH3 cells was completely blocked by peptide (not shown).

TABLE 111
Physical characteristics of in vitro translated R-esp proteins The molecular weight was calculated using the following equation, M, = [6?rNn,,,,,,,) J % ,~, A where N is Avogadro's number, nz0,", is the viscosity of water a t 20 "C, and u is the partial specific volume (in this case assumed to be 0.74 mug which is typical of most proteins) pzo.w is the density of water at 20 "C, szo,u, is the sedimentation coefficient a t 20 "C, andA is the Stokes radius in angstroms. The predicted molecular weight is deduced from the cDNAs. 110,000 80,000 a Physical properties of the population that was included in the gel filtration column and sucrose density gradients.

Calculated Predicted
In contrast to R-espl, the in vitro translated R-esp2 separated into two distinct populations on both gel filtration and sucrose density gradient centrifugation (not shown). One population behaved as rod shaped monomers yielding the hydrodynamic data presented in Table I11 for R-esp2. The second population eluted in the excluded volume of the ACA 22 column (exclusion limit = 1.2 x lo6 daltons), and this material rapidly sedimented as a broad peak during sucrose density gradient centrifugation, indicating that it is not a stable complex (not shown). This high molecular weight population represents ei-ther large aggregates of R-esp2 or aggregation of R-esp2 with other proteins present in the rabbit reticulocyte lysate.
To determine if R-espl and R-esp2 can form heterodimers the proteins were translated separately, mixed, and incubated at either 4 or 37 "C for 1 h prior to loading on the gel filtration column or sucrose density gradient. No changes in the physical properties of either R-espl or R-esp2 were detected indicating that the R-esp proteins do not form heterocomplexes under these conditions.

DISCUSSION
The R-espl cDNA encodes a novel homolog of the E(sp1)-WD locus of Drosophila, also known as m9/10 or groucho, which appears to be a truncated form of the previously characterized products of the E(sp1)-WD genes. The data presented here indicate that R-espl is a full-length cDNA clone, because it hybridized with a 1.3-kilobase mRNA in adult rat tissue. R-espl mRNA was detected in kidney, brain, heart, and lung tissue of adult rats. Differences in mRNA level among these tissues suggest that R-espl mRNA is subject to tissue-specific regulation in the adult.
Immunoprecipitation with antibodies raised against a peptide from R-espl detected R-espl protein in PC-12 cells. In vitro translation of R-espl yielded a product that comigrated with a 24-kDa protein that was specifically immunoprecipitated from the PC-12 cells. Finally, immunohistochemical staining revealed this protein localized to the nucleus of both PC-12 and GH3 cells. Taken together, these results argue that the R-espl cDNA is a full-length cDNA encoding a cellular protein.
R-espl may arise from a unique gene or by alternative splicing of a gene that also encodes a protein containing the WD-40 repeating units. Sequence comparison of R-espl and R-esp2 suggest that R-espl cDNA could not be an alternative splice product of a gene also encoding R-esp2. However, R-espl could arise by alternative splicing of some other full-length E(sp1)-WD homolog. Resolution of this question will require isolation of genomic clones encoding these proteins.
When the available nucleotide sequences of the E(sp1)-WD gene family are compared, R-espl shows greatest identity with human cDNA EST00256 (Table 11). EST00256 was originally isolated as an expressed sequence tag and identified as a member of the E(sp1) gene family by partial sequencing of the cDNA (Adams et al., 1991). In their screening of a human brain cDNA library, Stifani et al. (1992) failed to isolate a clone corresponding to EST00256. The probe used to screen their cDNA library corresponded to a region containing WD-40 repeats, a region absent from the R-espl clone. Possibly, EST00256 corresponds to the human homolog of R-espl.
Stifani et al., (1992) recognized that full-length E(spl)-WD homologs contain five discrete regions based upon amino acid composition and sequence. If these regions are important, they are likely to be conserved. Comparison of the amino acid conservation between the two new rat sequences together with the human and Drosophila E(sp1)-WD proteins suggests that three discrete domains (I, 11, and I11 in Fig. 3) are present in the full-length proteins. However, R-espl only contains amino-terminal domain I and the proline/glycine-rich portion of domain I1 while lacking the remaining portion of domain I1 and all of domain 111. R-esp2 encodes a full-length E(sp1)-WD homolog containing all three domains.
The E(sp1)-WD homologs containing WD-40 repeats are nuclear proteins. If R-espl affects the function of full-length E(spl)-WD homologs then R-espl protein may also be found in the nucleus. Although R-espl protein lacks a recognizable nuclear localization signal, immunohistochemical staining clearly shows it in the nucleus of both PC-12 and GH3 cells. The hydrodynamic data suggest that R-espl is a rod-shaped monomer that does not form homodimers. R-espl monomers, with a predicted molecular mass of 24 kDa, would be small enough to pass into or out of the nucleus by diffusion (Paine and Horowitz, 1980). Alternatively, R-espl may associate with another protein which contains a nuclear localization signal. Preliminary results with anti-R-esp2 antibodies indicate that this protein is also found in the nucleus (not shown). Like the human and Drosophila E(spl)-WD homologs, R-esp2 contains a nuclear localization signal.
The pattern of nuclear staining was distinguishable between untreated and NGF-treated PC-12 cells. Staining in untreated cells appeared reticulated with the stained portions forming a network containing many punctate regions lacking stain. It is unclear what confers this reticulated appearance, although one possibility is that R-espl associates with the nuclear matrix in untreated cells. In PC-12 cells that extended processes due to NGF treatment, the fibrillar appearance was less dramatic with staining broadly distributed throughout the nucleus. However, unstained regions are still clearly visible. Differentiation may change the nuclear targets that bind R-espl, resulting in altered immunohistochemical staining.
R-espl contains all of domain I but lacks most of domain I1 and all of domain 111. Possibly, R-espl acts as a negative regulator of full-length E(sp1)-WD proteins by competing for factors that interact with domain I. This model predicts that R-espl and R-esp2 should share a subset of interacting factors. Current work is focused upon evaluating this possibility.