Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Interleukin (IL)-5 is a 12.8-kD protein synthesized predominantly by T cells and to a lesser extent by mast cells, basophils, and eosinophils. In its native state, it exists as a 20 to 23 kD glycosylated monomer and a 40 to 50 kD disulfide-linked homodimer (1). Unlike most other cytokines that act upon multiple cell types, human IL-5 exerts its influence predominantly on eosinophils. In vitro studies have demonstrated multiple effects of IL-5 on eosinophil biology, including proliferation, differentiation, migration, activation, and survival (2). It is becoming clear that IL-5 and eosinophils play significant roles in the inflammatory component of asthma and contribute to the associated airway hyperresponsiveness observed in this disease (3).

T cells can be classified into several subtypes, depending on the cytokines they produce (4, 5). T helper (Th) 1 cells are associated with cell-mediated immunity and produce interferon (IFN)- and tumor necrosis factor-. Th2 cells participate in humoral immunity and produce IL-4 and IL-5. Th0 cells express both subsets of cytokines. Much is now known about the regulation of cytokine genes at the molecular level. In particular, several regulatory regions have been identified within the 5' promoter of IL-5 that control activation-dependent and cell-specific expression. The conserved lymphokine element 0 (CLE 0; 62 to 42) (6, 7) and GATA (73 to 66, 128 to 123) (8, 9) as well as Oct (244 to 237) (12) regions act as positive transcription regulatory sites. Notably, the transcription factor GATA-3 plays a significant role in IL-5 gene transcription and is preferentially expressed in Th2 cells (9, 10, 13). The region from 123 to 92, known as response element-II (RE-II) (8, 14), containing binding region (BR)5 and part of BR4 (15) also has been shown to be important in inducible expression of IL-5. This region interacts with members of the nuclear factor of activated T cells (NFAT) family of transcription factors, and electromobility shift assays indicate that additional proteins influence this site.

In the present report, we describe the identification and characterization of a repressor protein that works specifically through RE-II to reduce IL-5 gene transcription. Remarkably, this protein is translated from a messenger RNA (mRNA) that is transcribed with an initiation site near or within intron 11 of a much larger gene known as Wolf- Hirschhorn syndrome candidate 1 (WHSC1; also known as multiple myeloma SET domain [MMSET]) (16, 17).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture

Human Th0 SP-B21 cells (18) and murine Th2 D10.G4.1 cells (5) were maintained as described (14). Jurkat T cells were cultured in RPMI with 5% fetal bovine serum and 20 U/ml recombinant human IL-2 (BioSource International, Camarillo, CA). Human CD4⁺ Th2 cells were prepared from peripheral blood mononuclear cells from healthy volunteers by culturing CD45RO⁺ cells with IL-2, recombinant human IL-4 (Schering-Plough, Kenilworth, NJ), anti-IFN- monoclonal antibody (mAb) (R&D Systems, Minneapolis, MN), anti-CD28 mAb (PharMingen, San Diego, CA), and immobilized anti-CD3 mAb (clone UCHT1; PharMingen) as described (19). Cells exhibited a Th2 phenotype after approximately 3 wk and were used for experiments after a resting phase. Human Th1 cells were prepared according to the previously described culture conditions but with anti-IL-4 mAb (R&D) and recombinant human IFN- (R&D) instead of IL-4 and anti-IFN- mAb, and the cultures were used after 2 wk.

Expression Cloning

SP-B21 cells were stimulated for 4 h with 2 µg/ml anti-CD3 mAb immobilized to tissue culture plates. PolyA⁺ mRNA was purified, converted to complementary DNA (cDNA), ligated to appropriate adapters, and directionally cloned into the SalI (5') and NotI (3') sites of pSPORT1 (GIBCO-BRL, Rockville, MD). Escherichia coli (DH10B) transformed with cDNA was screened to identify clones that exhibited RE-II oligonucleotide binding activity as described (20) with modifications. Briefly, 10⁶ transformed cells were replica-plated, and one set of plates was overlaid with nitrocellulose filters (Hybond-C Extra; Amersham Pharmacia Biotech, Inc., Piscataway, NJ) presoaked in 200 mM isopropylthiogalactopyranoside (IPTG). Filters were exposed to chloroform vapors for 25 min and soaked in Tris-buffered saline (TBS; 50 mM Tris-HCl, pH 7.6, 150 mM NaCl) containing 0.1% Tween-20, 400 µg/ml lysozyme, and 10 µg/ml DNase I for 1 h at 25^oC. After washing in TBS plus Tween-20, filters were equilibrated in binding buffer (BB; 25 mM NaCl, 5 mM MgCl₂, 0.5 mM dithiothreitol, 25 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid, pH 7.9, 0.25% nonfat powdered milk), and immobilized proteins were denatured with 6 M guanidine HCl in BB for 10 min at 4°C. Immobilized proteins were renatured by sequential washes with 3, 1.5, 0.625, 0.3125, and 0 M guanidine HCl in BB. Filters were blocked in BB containing 5% nonfat powdered milk and equilibrated in BB containing 0.25% nonfat powdered milk. Probe (2 × 10⁶ cpm/ml) and denatured calf thymus DNA (5 µg/ ml) were added and filters were incubated for 1 h at 25°C. Filters were washed with BB and exposed to film overnight. The probe consisted of concatamers of RE-II containing oligonucleotides prepared by separately forming LU-LL, MU-ML, and RU-RL hybrids followed by ligating the three double-stranded oligonucleotides at 15°C overnight (see Table 1 for sequences). The resulting mixture of fragments contained varying numbers of MU-ML with 5' LU-LL and 3' RU-RL ends. This synthetic DNA was labeled by nick translation. A single clone exhibited DNA binding activity and was sequenced (clone 7 [1,644-2,753]).

Glutathione-S-Transferase-Fusion Protein Expression and Western Blot Analysis

cDNA from clone 7 (1,644-2,753) was subcloned into the SalI and NotI sites of pGEX-6P-1, pGEX-6P-2, and pGEX-6P-3 (Amersham Pharmacia Biotech), and the resulting plasmids were used to transform DH5 and BL21 strains of E. coli. A pGEX-6P-2 clone was chosen that expressed glutathione-S-transferase (GST)-fusion protein of the appropriate size.

For electromobility shift assays, GST-fusion protein was purified from cleared E. coli sonicates by batch chromatography with glutathione-sepharose 4B. Clone 7 protein was cleaved from the GST fusion and released from the chromatography media by treatment overnight at 4°C with PreScission protease (Amersham Pharmacia Biotech). For Western blot analysis, E. coli sonicate containing clone 7 GST-fusion protein was separated by preparative electrophoresis under reducing but not denaturing conditions using a Mini Prep Cell as described by the manufacturer (BioRad Laboratories, Hercules, CA). Fractions from the Mini Prep Cell containing clone 7 GST-fusion protein were further separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) followed by transfer to polyvinylidene difluoride membranes by electroblotting. For the analysis of native protein, nuclear protein extracts from human CD4⁺ Th2 and Jurkat cells were separated by SDS-PAGE and transferred to supported nitrocellulose membranes. Nuclear protein extracts from CD4⁺ Th2 cells were derived as described (14). Nuclear protein extracts from Jurkat T cells were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

Polyclonal antiserum (SP547) was raised in NZW rabbits injected with a keyhole limpet hemocyanin-conjugated peptide corresponding to amino acids 421 to 433 (CLGDRPKTSTTLSS; Covance Research Products, Inc., Denver, PA) of the predicted full length RE-II binding protein (RE-IIBP). This peptide was chosen from a region predicted to be relatively antigenic (MacVector 5.0; Eastman Kodak Co., Rochester, NY) and non-zwitterionic. For immunodetection of SP547-reactive proteins, membranes were blocked with Blotto (5% nonfat powdered milk, 0.1% Tween 20 in TBS or phosphate-buffered saline [PBS]) and incubated with SP547 (1:10,000) followed by antirabbit immunoglobulin (Ig) G-horseradish peroxidase (1:3,000). Washes were with 0.1% Tween 20 in PBS. Antibodies were detected with enhanced chemiluminescence reagents as described by the manufacturer (Amersham Pharmacia Biotech).

Electromobility Shift Assay

Electromobility shift assays were performed as described (14). The activator protein-1 (AP-1) oligonucleotide (Santa Cruz) corresponded to a consensus binding site. All other oligonucleotides were synthesized commercially (Oligos Etc., Inc., Wilsonville, OR; GIBCO-BRL). Double-stranded oligonucleotides contained GGG overhangs to allow efficient incorporation of [³²P]deoxycytidine triphosphate (NEN Life Science Products, Boston, MA) by the Klenow fragment of DNA polymerase I as described (14).

Rapid Amplification of cDNA Ends and Reverse Transcriptase/Polymerase Chain Reaction

A human pancreas cDNA library (Marathon Ready; Clontech Laboratories, Inc., Palo Alto, CA) was used as starting material for 5' rapid amplification of cDNA ends (RACE) with a 5' primer corresponding to vector sequences (adapter primer 1, 5'-CATCCTAATACGACTCACTATAGGGC-3'; Clontech Laboratories) and a 3' antisense primer corresponding to RE-IIBP (2,273'-2,253', 5'-GAGGTCTTTGGTCTATCCCCG-3'). The first round of polymerase chain reaction (PCR) began with a 1-min incubation at 94°C, followed by five cycles of 94°C for 30 s, 72°C for 5 min, and ending with 25 cycles of 94°C for 30 s, 70°C for 5 min. The resulting amplification was separated by electrophoresis through low melting-point agarose, and gel containing DNA from 1.5 to 3 kbp was retained. PCR was repeated with 2 µl of melted gel as template with primers and conditions as described previously. After electrophoresis, a diffuse, approximately 1.5 kbp band was purified and subcloned into pGEM-T Easy (Promega Corp., Madison, WI). Clones were selected by colony hybridization with [³²P]-labeled clone 7 (1,644-2,753) cDNA. A clone with the largest insert was selected for sequence analysis (clone 4.3 [451-2,273]).

Additional 5' sequence information was obtained by reverse transcriptase (RT)-PCR using cDNA from CD4⁺ Th1 cells. Total RNA was treated with DNAse I and reverse-transcribed with an antisense oligonucleotide corresponding to 676' to 657' of RE-IIBP (5'-CCGGCTTCTCACACAGCTGG-3'). PCR with sense primer corresponding to 31 to 50 (5'-GAGTAGCATTGTGGTTATAT-3') and anti-sense primer corresponding to 467' to 451' (5'-GATGCTGCCGTGGAAGCCCG-3') consisted of one cycle at 94°C for 30 s, 40 cycles at 94°C for 15 s, 60°C for 25 s, followed by a final incubation at 72°C for 7 min. The resulting amplicon was purified by electrophoresis through polyacrylamide gel and sequenced directly. All nucleotide numbering refers to the full-length 3,649-bp RE-IIBP gene encoding a protein of 584 amino acids with GenBank accession number AF330040 for nucleotides 31-3,649.

Transient Expression Plasmid Construction

The complete coding sequence and selected deletion mutants of RE-IIBP were subcloned into the eukaryotic expression vector pME18S (21) under the control of a simian virus 40 and retrovirus (SR) promoter. pME18S was prepared by removing a portion of the polylinker contained within the XhoI sites (pME18S-Xho). Full-length RE-IIBP (clone 4.8 [451-3,649]) was reconstructed within the XbaI site of pME18S-Xho from a 5', NotI-NspHI (1,571 bp) fragment from the 5' RACE clone 4.3 (451-2,273) and a 3' NspHI-KpnI (1,665 bp) fragment from expressed sequence tag (EST) AA064672 (I.M.A.G.E. Consortium CloneID no. 5254893 [22]; Genome Systems, Inc., St. Louis, MO). The coding region contained within a SalI-NotI (1,150 bp) fragment from the first pSPORT clone (clone 7 [1,644-2,753]) was subcloned into the SpeI site of pME18S-Xho (clone 1.1 [1,644-2,753]). Clones 2.8 (2,005- 2,753) and 3.3 (1,644-2,005) consisted of clone 7 (1,644-2,753) fragment NspHI-NotI (749 bp) or SalI-NspHI (401 bp) cloned into the XbaI site of pME18S-Xho, respectively. Appropriate restriction sites were added by ligation of oligonucleotide linkers to the ends of each fragment to facilitate cloning. A translation initiation codon (ATG) was added in the correct frame to the 5' linkers for clones 1.1 (1,644-2,753), 2.8 (2,005-2,753), and 3.3 (1,644- 2,005). Clone 4.8 (451-3,649) contained an endogenous translation initiation codon and none was added to the 5' linker. An endogenous translation termination codon (TAG) was located on clones 1.1 (1,644-2,753), 2.8 (2,005-2,753), and 4.8 (451-3,649). In addition, an in-frame translation termination codon (TAG) was added to the 3' linker for all clones. The luciferase expression vector pGL2basic (Promega) containing a concatamer of five RE-II sites immediately 5' of 80 to +42 of the human IL-5 promoter (pGL2-huIL5-5XRE-II[80]) was used as a reporter gene in transient transfection experiments. The five RE-II sites comprising the concatamer were constructed from the oligonucleotides shown in Table 1. All junctions formed during plasmid assembly were verified by DNA sequencing.

Transient Transfections

Preliminary experiments indicated that optimal transfection efficiency was obtained by electroporation for murine D10.G4.1 cells (14) and by a lipid-mediated method using DMRIE-C reagent (Roche Molecular Biochemicals, Indianapolis, IN) for human Jurkat cells (data not shown). These preliminary experiments also established the appropriate amounts of plasmid DNA to use for each transfection experiment. Murine D10.G4.1 cells were transfected by electroporation as described (14) with 50 µg pGL2-huIL5-5XRE-II, 2 µg pCMV-SEAP (secreted placental alkaline phosphatase; Tropix, Bedford, MA), and 50 µg pME18S-Xho, RE-IIBP-pME18S clone, or NFATc1-pME18S (23). Transfected cells were rested at 37°C overnight and half of each transfection was stimulated with 1 µg/ml immobilized antimouse CD3 mAb for 18 to 20 h. Human Jurkat cells were transfected using DMIRIE-C reagent with 5 µg pGL2-huIL5-5XRE-II, 2 µg pCMV-SEAP, and 1.3 pmol (experiment 1) or 1.9 pmol (experiments 2 and 3) RE-IIBP-pME18S clone or NFATc1-pME18S. pME18S-Xho was included for a total of 8 µg of plasmid DNA in each transfection. After resting at 37°C (1 d for experiment 1, 2 d for experiments 2 and 3), half of the transfected cells were stimulated with 1 µg/ml phytohemagglutinin (PHA) and 50 ng/ml 12-O-tetradecanoylphorbol-13-acetate (TPA) for 18 h. Cell lysates were prepared from all transfections and assayed for luciferase activity as described (14). Cell culture supernatants were collected for the determination of SEAP activity derived from the pCMV-SEAP control plasmid with CSPD, a chemiluminescent substrate as described by the manufacturer (Tropix). SEAP activity was induced after stimulation of Jurkat T cells and was constitutive in D10.G4.1 T cells. SEAP activity was used to normalize all samples for transfection efficiency. In stimulated D10.G4.1 T cells the use of three plasmids (pME18S-Xho, pGL2-huIL5-5XRE-II, and pCMV-SEAP) in transfections resulted in 6-fold less luciferase activity compared with that observed in transfections with two plasmids (pGL2-huIL5-5XRE-II and pCMV-SEAP). This effect was not observed in Jurkat T cells. For this reason, comparisons were made only among transfections done with the same total amount of DNA consisting of all three plasmids (pME18S expression plasmid, 5X-RE-II reporter plasmid, and pCMV-SEAP control plasmid). In all experiments, activity for a given expression construct was compared with activity for the expression vector containing no cDNA insert (pME18S-Xho). Endogenous murine IL-5 levels in cell culture supernatants were assayed by enzyme-linked immunosorbent assay (ELISA) with reagents from PharMingen. The significance of differences between experimental groups (P  0.05) was determined by analysis of variance using Fisher's protected least significance test (StatView; Abacus Concepts Inc., Berkeley, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning of RE-IIBP

To identify DNA-binding proteins in addition to NFAT (8, 14) that interacted with the RE-II region of the human IL-5 promoter, a cDNA library was prepared from polyA⁺ RNA from anti-CD3 mAb-activated SP-B21 cells. Under these conditions SP-B21 cells synthesize IL-5 in addition to IFN- and are phenotypically a Th0-like cell line (18). SP-B21 cells are not transformed and exhibit a cytokine expression profile in response to stimuli similar to that seen in primary lymphocytes. Therefore, these cells mimic as closely as possible signaling mechanisms that occur in primary human T cells (14). The library was prepared in pSPORT, a vector that facilitated the synthesis of protein corresponding to insert cDNA when induced with IPTG. Thus, it was possible to transfer recombinant proteins to nitrocellulose membranes and assay the membranes for DNA-binding activity (20). Denaturing and renaturing the proteins in the presence of guanidine HCl before probing were done to facilitate the correct folding of the proteins for optimal activity (20). A [³²P]-labeled probe was generated from a mixture of oligonucleotides that contained concatamers of the RE-II binding site (Table 1). After screening 10⁶ colonies, a single clone was identified (clone 7 [1,644-2,753]; Figure 1) and sequenced. Based on sequence analysis, clone 7 (1,644-2,753) was a partial clone containing the 3' coding portion of a larger gene. It had a large open reading frame with a TAG translation termination codon but no clear translation initiation codon. As with many cDNA libraries, the library from which clone 7 (1,644-2,753) was identified was made by the addition of NotI and SalI linkers followed by restriction endonuclease digestion. Therefore, it was likely that this gene had an internal NotI site. This library and any other library initially primed with oligo-dT and restricted with NotI, could not be used to identify any additional sequence 5' of the NotI site.

View larger version (58K): [in this window] [in a new window]   Figure 1.   Cloning of RE-IIBP. (A) The full-length sequence of RE-IIBP mRNA was determined by expression cloning (clone 7), RT-PCR (clone 4.3 and 5' RT-PCR) and electronic searching of sequence databases (EST AA064672). Coding region is indicated as open bar. An antibody was generated to a peptide corresponding to sequences indicated by the solid bar (SP547). Enzyme sites relevant to subcloning described in the text are indicated and those in parentheses were from adjacent vector sequences. Bars are drawn to scale. (B) The 5' region of RE-IIBP is compared to the known structure of WHSC1 with intron and exon numbering as described by Stec and coworkers (16). Numbers above and below RE-IIBP indicate sizes (bp) of segments and triangles indicate intronic sequences. Vertical dashed line at the 5' end of RE-IIBP depicts a 30-bp region predicted to be the 5' terminus but not verified by sequencing. RT1, PCR1 and PCR2 indicate the location of primers used to verify the 5' region of the gene. The first eight amino acids are shown. Relative positions are correct but are not drawn to scale. (C) DNA sequence of the 5' noncoding region of RE-IIBP and the first 3 amino acids are shown. Underlined sequence was not confirmed by sequencing and predicted transcription initiation start site is indicated in boldface. Primers RT1, PCR1, and PCR2 are highlighted.

Demonstration of DNA-Binding Activity of Clone 7 (1,644-2,753) cDNA-Derived Protein

Clone 7 (1,644-2,753) cDNA was subcloned into a prokaryotic expression vector (pGEX-6P-2) and expressed as a GST-fusion protein. After purification by affinity chromatography and cleavage of GST from the N-terminus, clone 7 (1,644-2,753) protein was used in electromobility shift assays with oligonucleotides that corresponded to sequences within the RE-II region of the human IL-5 promoter (Figure 2). When a probe corresponding to 99 to 128 (H-128[30]) was used, a major slower migrating band (complex I) and minor faster migrating band (complex II) were identified. Both complexes could be competed by unlabeled H-128[30], by a sequence corresponding to 108 to 127 (H-127[20]), and by an NFAT sequence from the murine IL-4 promoter (mIL-4 NFAT) (24). A mutated sequence corresponding to 99 to 127 (H-117[20]mut) partially competed for binding to complex I but not complex II. When H-127(20) was used as a probe, two weaker and less resolved bands were identified, both of which could be competed by unlabeled H-128(30), H-127(20), and mIL-4 NFAT. H-117(20)mut competed for binding to the slower migrating band but not the faster migrating band. When H-117(20)mut, was used as a probe, a weak band corresponding to the slower migrating complex I was detected. This band was competed by unlabeled H-128(30), H-127(20), H-117(20)mut, and mIL-4 NFAT. Although the BR5 region of RE-II contains an AP-1 site, AP-1 complexes do not bind to RE-II (8, 14) and an AP-1 consensus sequence did not compete with labeled H-128(30), H-127 (20), or H-117(20)mut for binding to RE-IIBP. These data were consistent with the assignment of complex I to binding of clone 7 (1,644-2,753) protein to 99 to 107 and to 108 to 122. Complex II corresponded to binding of clone 7 (1,644-2,753) protein to 108 to 122 of the IL-5 promoter. Formation of the faster migrating but weak complex II occurred with the core sequence 5'-GGnnTnnnAAA-3'. Formation of the slower migrating stronger complex I occurred with the core sequence 5'-TTTnnnnT-3' or 5'-GGnnTnnnAAA-3'. The slower migrating complex may represent an alternate conformation or dimer of clone 7 (1,644-2,753) protein.

View larger version (66K): [in this window] [in a new window]   Figure 2.   Protein derived from RE-IIBP (clone 7 [1,644-2,753]) specifically binds RE-II-containing oligonucleotides. (A) Oligonucleotides used in electromobility shift assays are shown aligned with the RE-II region of the human IL-5 promoter. Base pairs different from those in RE-II are shown in lowercase letters. NF-AT (37), GATA (8), and BR5 (15) sites are indicated by boxes, as well as regions that are required for the formation of complex I or complex II. The ability of each oligonucleotide to participate in complex I or complex II is indicated by a plus and was determined by direct binding with radiolabeled oligonucleotide or by competition with unlabeled oligonucleotide from data depicted in B. (B) Electromobility shift assay using clone 7 (1,644-2,753) protein. Labeled probes and competitors are indicated. Complex I and complex II are shown by arrows. A representative experiment of two is shown in which two different preparations of clone 7 (1,644- 2,753) protein were used with similar results.

Identification of Endogenous RE-II BP

Polyclonal antiserum was generated to a peptide corresponding to 13 amino acids predicted from the clone 7 (1,644- 2,753) sequence (Figure 1A). This antibody (SP547) bound to purified fractions of clone 7 (1,644-2,753) GST-fusion protein (78 and 83 kD; Figure 3A). SP547 reactivity with protein in nuclear extracts from nonstimulated CD4⁺ Th2 cells was competed by excess peptide used to generate the antibody, and reactivity was not observed with pre-immune serum (Figure 3B). Endogenous proteins of 62 and 66 kD were detected in nuclear protein extracts from nonstimulated and stimulated human CD4⁺ Th2 cells (Figure 3C). In nonstimulated and stimulated Jurkat cells, a cell line that does not express appreciable amounts of IL-5, the amount of 62-kD protein was dramatically reduced compared with that of the 66-kD protein (Figure 3C). Jurkat cells contained less SP547-reactive protein than did CD4⁺ Th2 cells.

View larger version (59K): [in this window] [in a new window]   Figure 3.   Western blot analysis of fusion protein and endogenous RE-IIBP. Rabbit antihuman RE-IIBP antiserum (SP547) was used to probe Western blots. (A) Fractions from preparative polyacrylamide gel electrophoresis (PAGE) of protein derived from clone 7 (1,644-2,753)-GST (5 µl) were separated by PAGE and probed with SP547. (B) Nuclear protein from unstimulated human CD4+ Th2 cells were separated by PAGE in a single long well. Proteins were transferred to PVDF membrane and the sample was cut into strips containing approximately 5 µg of protein. After incubation with preimmune serum (lane 3), 80 µM of the peptide used to generate the antibody (lane 4), or buffer (lane 5), the strips were incubated with SP547. (C) Nuclear protein extracts (10 µg per lane) were separated by PAGE and probed with SP547 (lane 6, unstimulated CD4+ Th2 cells; lane 7, 2 h anti-CD3 mAb-stimulated CD4+ Th2 cells; lane 8, unstimulated Jurkat cells; lane 9, 4 h TPA-stimulated Jurkat cells). Molecular weight is indicated in kilodaltons.

Identification of the Complete Transcript for RE-IIBP

RACE was used to identify the 5' end of the gene using Marathon-Ready cDNA (Clontech) derived from human pancreas RNA. PCR was performed with primers that corresponded to an internal sequence of clone 7 (1,644-2,753) and sequence from the 5' linker/adapter used to prepare the cDNA. After subcloning the products of two rounds of PCR and screening the resultant clones with a probe made from clone 7 (1,644-2,753), the largest cDNA was chosen and sequenced (clone 4.3 [451-2,273]). This clone provided a potential translation initiation codon, which included adjacent methionines. By electronic searching of genomic databases, a number of ESTs were identified, including AA064672, AA143620, and AA159454, which completed the 3' portion of the gene (Figure 1). Combining sequence information from clones 7 (1,644-2,753) and 4.3 (451- 2,273), and AA064672 resulted in a gene that contained 3,198 bp with a complete 3' terminus. This gene was identical to the 3' end of MMSET (16, 17).

Identification of Putative 5' Translation Initiation Region

When clone 7 (1,644-2,753) cDNA was used to probe multiple tissue Northern blots (Clontech), mRNA species of 8.8, 6.2, and 3.7 kb were identified, consistent with earlier reports for MMSET and WHSC1 (data not shown; 16, 17). The probe used in our experiments and by Stec and coworkers (16) only identified sequences from the 3' end of the gene and could not hybridize to splice variants terminating within the middle of WHSC1 (16, 17). In addition, the proteins detected by Western blot analysis were consistent with a coding region of approximately 1.8 kb from the 3' end of the gene, and the antibody SP547 could not detect protein from the N-terminus of WHSC1. It was possible that our 3.2-kb sequence represented only a partial cDNA and the actual sequence was the full-length WHSC1 containing 6.2 or 8.8 kbp. However, it was more likely that the sequence was short by approximately 400 bp, and that RE-IIBP transcription initiation occurred within an intron of WHSC1 (intron 11). In support of the latter possibility, the predicted protein encoded by the longest open reading frame of the 3.2-kb cDNA sequence described here (from the first methionine) was 64 kD, remarkably similar to the size of the protein detected by SP547 in T-cell nuclear protein extracts. In addition, electronic analysis identified a transcription initiation consensus sequence (ATATAAGA; Gal1 TATA, TATA-3) in genomic DNA within intron 11 that would result in a 3.7-kb mRNA (FindPatterns using tfsites.dat; Wisconsin Package Version 8.0; Genetics Computer Group (GCG), Madison, WI). To test the "intronic" initiation hypothesis, cDNA was synthesized from T cell- derived RNA in a reverse transcription reaction using a primer (RT1) that overlapped the exon 13 to exon 14 border of WHSC1 (16) (Figure 1B). This cDNA was used in a PCR with primers corresponding to sequences in exon 11 (PCR1) and intron 11 (PCR2) of WHSC1. The 5' PCR primer (PCR2) was 22 bp from the end of the TATA site. If the prediction was correct, i.e., that transcription began within the intron, then a product of 436 bp would be detected. If transcription did not occur in this region and the splicing pattern was the same as in WHSC1, then the reaction would not yield a product because the 5' primer (PCR2) would not hybridize to cDNA sequence. The PCR yielded a product of 436 bp (Figure 4) that was sequenced and found to correspond to the sequence from intron 11 to exon 11 of WHSC1 as hypothesized. This sequence was not genomic DNA because cDNA reactions containing no reverse transcriptase yielded no PCR product. Also, the product was unlikely to be a result of detecting nonspliced RNA because reverse transcription with a primer overlapping the exon 13 to exon 14 border would have had to yield cDNA of greater than 11 kbp (Figure 1B) to provide the appropriate template for amplification, a highly unlikely possibility.

View larger version (40K): [in this window] [in a new window]   Figure 4.   RT-PCR with primers corresponding to the 5' end of RE-IIBP. Total RNA treated with DNase I from CD4+ Th1 cells was used to generate cDNA template with a primer overlapping exon 13 and 14 of RE-IIBP (RT1; see Figures 1B and 1C). The resulting cDNA was amplified (lane 1) with primers corresponding to sequences within intron 11 (PCR1) and exon 11 (PCR2). Control PCRs included template made without reverse transcriptase (lane 2) or with water (lane 3). Size markers are 0X174 digested with HaeIII (lane 4) and a 100-bp ladder (lane 5). DNA length is indicated in basepairs.

Demonstration of RE-IIBP Activity on RE-II-Mediated Transcription

To demonstrate the activity of RE-IIBP on transcription mediated by the RE-II region, mouse and human T cells were cotransfected with an RE-II reporter gene, RE-IIBP-expressing constructs, and a transfection control plasmid that expressed SEAP. The amounts of reporter, expression, and control plasmids as well as the total amount of plasmid DNA were held constant for each transfection in a given experiment. The RE-II reporter construct contained five copies of the RE-II region as well as the CLE 0 and 70 GATA sites to provide measurable promoter activity. The presence of truncated RE-IIBP proteins was confirmed in nuclear extracts from representative cotransfections with clones 1.1 (1,644-2,753) and 2.8 (2,005-2,753) by Western blot analysis with SP547 (data not shown). Western blot analysis of nuclear extracts from cotransfections with clone 4.8 (451-3,649) was uninformative because of the presence of endogenous protein, and SP547 cannot detect protein expressed from clone 3.3 (1,644-2,005).

RE-II promoter activity in PHA plus TPA-stimulated Jurkat cells was reduced by as much as 40% when RE-IIBP or various portions of it were coexpressed (Figure 5), although these reductions were not statistically significant. The RE-II promoter contains homology to an NFAT site (Figure 2) and coexpression of NFATc resulted in a 2-fold increase in RE-II promoter activity. Therefore, the observed reductions in promoter activity mediated by RE-IIBP and derivatives were not due to nonspecific effects of protein overexpression.

View larger version (32K): [in this window] [in a new window]   Figure 5.   Transient transfections of human Jurkat T cells and murine D10.G4.1 T cells with RE-IIBP-expressing constructs, an RE-II reporter gene and a control SEAP gene. (A) Jurkat or D10.G4.1 T cells were transfected with the indicated RE-IIBP-containing pME18S constructs, the pGL2-huIL5-5XRE-II luciferase reporter gene, and the pCMV-SEAP transfection control gene by lipid-mediated transfection (Jurkat) or electroporation (D10.G4.1) as described in MATERIALS AND METHODS. After transfection, half of each sample was stimulated with PHA plus TPA (Jurkat) or immobilized anti-CD3 antibody (D10.G4.1). Luciferase activity for each construct was normalized for transformation efficiency by comparison to SEAP levels and presented relative to luciferase activity measured in stimulated cells transfected with pME18S-Xho without insert. After stimulation, luciferase activity in pME18S-Xho-transfected Jurkat cells was 39,603 ± 14,826 U (n = 3), 168-fold more than that measured in unstimulated cells. Luciferase activity in stimulated pME18S-Xho- transfected D10.G4.1 cells was 5,022 ± 2,507 U (n = 3), 9-fold more than that measured in unstimulated cells. Data represents the average (± standard deviation) of two to three experiments. ND, Not done; *P < 0.05. (B) Graphical depiction of the proteins expressed by the pME18S constructs 1.1 (1,644-2,753), 2.8 (2,005- 2,753), 3.3 (1,644-2,005) and 4.8 (451-3,649) used in A. Shaded bar indicates region used to make anti-RE-IIBP antiserum (SP547). PHD zinc finger and SET domains are indicated. A nuclear localization signal is shown as a shaded box.

Although other investigators used human Jurkat cells to study the IL-5 promoter (25) and there was detectable activity from the reporter gene used in the present studies, they generally do not exhibit inducible expression of IL-5. Therefore, to determine the effect of forced expression of RE-IIBP in a cell that produces significant amounts of IL-5, murine D10.G4.1 cells (a Th2 cell line) were transfected. D10.G4.1 T cells were used in the transfection experiments that contributed to the initial identification of RE-II (14). In these cells, coexpression of all RE-IIBP constructs resulted in a 75% inhibition of RE-II-mediated transcription (Figure 5).

The murine IL-5 gene promoter contains an element similar to human RE-II in sequence and relative location (14). Therefore, it was possible to assess the affect of overexpressing RE-IIBP on the endogenous promoter by measuring the production of murine IL-5. Culture media from transfected D10.G4.1 cells described previously were collected and analyzed by ELISA for murine IL-5 content (Table 2). Murine IL-5 production by cells transfected with RE-IIBPs was reduced compared with pME18S-Xho- transfected cells, and this reduction was statistically significant for the longest RE-IIBP (clone 4.8 [451-3,649]).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this report, we identify and describe a DNA-binding protein, RE-IIBP, that interacted with the RE-II region of the human IL-5 gene promoter and repressed gene transcription mediated by this element. In addition, we demonstrate that ectopically expressed RE-IIBP reduced the level of IL-5 produced by D10.G4.1 T cells in response to anti-CD3 stimulation.

The 62/66-kD RE-IIBP is identical to the carboxy-terminal half of a protein encoded by WHSC1. This gene is located on chromosome 4p16.3 at the breakpoint associated with t(4;14) multiple myeloma resulting in fusion with IgH (17). It also is located within the region of chromosome 4, hemizygously deleted in the Wolf-Hirschhorn syndrome (16). There are no known correlations of IL-5 production with Wolf-Hirschhorn syndrome, although there has been little reason to investigate potential associations prior to the present report. However, an early study demonstrated the in vitro production of B-cell growth factor II (IL-5) by three human myeloma cell lines (28). Although IL-6 is thought to play a dominant role, several cytokines, including IL-5, can synergize with IL-6 to support myeloma cell proliferation (29). In addition, plasma cell myeloma has been associated with eosinophilia, but in only one case was patient serum analyzed for cytokines and IL-5 was not detected (30). Therefore, it is difficult at this time to establish a definitive link between IL-5 and myeloma.

Transcription of RE-IIBP mRNA uniquely begins adjacent to an intron of WHSC1 that undergoes complex alternate splicing, resulting in several possible mRNA species, including two polyA⁺ terminated forms (16, 17). Most of the transcripts described before the present report were transcribed from the 5' terminus of WHSC1, yielding transcripts ranging in size from 2.4 to 10 kb (16, 17). However, we and Stec and colleagues (16) detected a 3.5-kb mRNA with a probe to the 3' end of WHSC1. One explanation for this result was that this mRNA was initiated at an alternative transcription initiation site within the WHSC1 gene. In the present report, experimental evidence supporting the "intronic" transcriptional start site for RE-IIBP included (1) an mRNA species identified by Northern blot analysis with a 3' probe that placed the transcription initiation site 3.7 kb from the polyA⁺ signal; (2) the predicted protein product from RE-IIBP was the same size as the protein detected by antiserum to an RE-IIBP peptide; (3) computer analysis of the sequence identified a putative transcription initiation signal that occurred at the correct position (3.7 kb from the polyA⁺ signal); and (4) the predicted 5' sequence of the mRNA was consistent with the results of RT-PCR analysis. Although computer analysis of genomic DNA proximal to the proposed RE-IIBP transcription initiation site revealed potential regulatory element consensus sites (data not shown), it is not known if RE-IIBP transcription is controlled by this potential promoter or by the distal WHSC1 promoter.

RE-IIBP contains plant homeodomain (PHD) and SET motifs that suggest it plays some role in gene silencing. Two PHD-type zinc fingers are present containing the characteristic Cys₄-His-Cys₃ amino acid pattern. This motif has been postulated to be involved in protein-DNA or protein-protein interactions important in chromatin-mediated positive and negative transcriptional regulation (31). RE-IIBP also contains a single SET domain, named after a motif found in su(var)3-9, enhancer-of-zeste, and trithorax. The SET domain interacts with a family of dual-specificity phosphatases (dsPTPases) that dephosphorylates serine-, threonine-, and tyrosine-containing peptides in vitro. SET domain-containing proteins are important in positive and negative developmental regulation of genes at the level of chromatin structure (32).

Several mechanisms have been proposed regarding the negative regulation of cytokine gene transcription (33). The presence of the PHD and SET motifs together with the data in the present report are consistent with a model in which RE-IIBP acts as a repressor protein that binds to the RE-II region of the IL-5 promoter by means of the PHD motifs and suppresses gene transcription. Phosphorylation of RE-IIBP might reduce the binding affinity for RE-II, leaving the promoter accessible to interactions with other transcription factors, such as NFAT, and allowing for the induction of transcription. Subsequently, interaction of the SET domain with recruited dsPTPases could result in dephosphorylation and the recycling of RE-IIBP to a form capable of binding RE-II and again suppressing transcription mediated by this site. The presence of the 62-kD form of RE-IIBP in Th2 cells but not in Jurkat cells (Figure 3C) might indicate that a modified version of the protein permits IL-5 gene transcription to occur. This concept has been proposed as the mechanism by which occupation of the IL-5 CLE 0 site by constitutive binding factor(s) in resting T cells results in no transcription and occupation of the site by inducible factor(s) with loss of the constitutive factor(s) results in gene transcription (34). In addition, the NFAT consensus site in RE-II has been shown to cooperate with the 70 GATA site and support transcription of a reporter gene in response to IgE plus antigen stimulation in murine mast cells (8). This cooperation potentially extends the influence of RE-IIBP to regions adjacent to RE-II.

It is unclear why the 3.3 derivative of RE-IIBP exhibited activity in the reporter experiments given that the full-length protein and its derivatives containing the C-terminal PHD domain were also active. It is possible that both the N-terminal and C-terminal halves of RE-IIBP contain functional domains, each of which is capable of affecting IL-5 gene transcription. This could occur by direct interaction with promoter DNA or by altering the activity or availability of other associated regulatory proteins. Consistent with this concept is the observation that the full-length protein (construct 4.8) inhibited the production of murine IL-5 to a greater extent than any of the truncated forms of the protein (Table 2).

Although the RE-II region positively contributes to activation-dependent IL-5 gene transcription (8, 14, 15), recent reports also have characterized a negative role for this element (26, 35). The RE-II region has been associated with suppression of IL-5 synthesis in studies with prostaglandin E₂ (35) and ionizing radiation (26). Cyclic AMP (cAMP) (35, 36) is generally considered an activator of IL-5 synthesis (11, 36). However, a recent report demonstrated that increasing intracellular cAMP levels by treatment of T-cell receptor-stimulated human T cells with prostaglandin E₂ enhanced the binding of an unknown protein to the RE-II region (35). Under these conditions, IL-5 synthesis was downregulated compared with T cells not treated with prostaglandin E₂. In addition, overexposure to ionizing radiation often results in eosinophilia, suggesting an affect on IL-5 production (26). Transient transfections of Jurkat T cells with IL-5 promoter-CAT constructs uncovered a suppressive role for the RE-II region in radiation-induced transcriptional activity. It is possible that one of the RE-II-specific binding proteins induced by prostaglandin E₂ or ionizing radiation that are associated with suppression of IL-5 transcription is influenced by or is RE-IIBP.

In conclusion, we have identified a DNA-binding protein that acts as a repressor of IL-5 gene transcription. RE-IIBP negatively contributes to T-cell-specific regulation of IL-5 expression and provides an additional mechanism to modulate the synthesis of this cytokine. In addition, transcription of RE-IIBP mRNA begins within the middle of WHSC1 and encodes a protein corresponding to the C-terminal half of the predicted WHSC1 protein. The involvement of WHSC1 in Wolf-Hirschhorn syndrome and in the translocation observed in multiple myeloma might suggest a role for RE-IIBP in these diseases.