PU.1 Regulates the CXCR1 Promoter*

The interleukin-8 receptors (CXCR1 and CXCR2) are specifically expressed at high levels in cells of the neutrophil lineage. In this work we identify promoter elements of the CXCR1 gene and the ets family transcription factor PU.1 as a major regulator for activation of the CXCR1 promoter. We first showed that the upstream sequence of CXCR1 (−800 to +86 base pairs (bp)) directs myeloid-specific expression of reporter gene constructs. Second, we showed the presence of negative elements in the sequence from −800 to −128 bp and positive elements from −128 to +50 bp. Third, we demonstrated that the fragment −22 to +14 bp binds PU.1. Fourth, we showed that PU.1 transactivates the CXCR1 promoter. These data are the first demonstration of PU.1-mediated transcriptional regulation of a neutrophil chemoattractant G protein-coupled receptor.

pression in myeloid cells, the transcriptional mechanisms regulating myeloid-specific expression of the IL-8 receptor genes are unknown.
Previous studies have shown that CXCR1 and CXCR2 expression is under transcriptional control (14). The genomic organization and promoter regions of CXCR1 and CXCR2 have been previously mapped (15,16). The CXCR1 gene consists of two exons interrupted by an intron of ϳ1.7 kilobases. The entire open reading frame is encoded in exon 2. Sprenger et al. (15) detected promoter activity in the T lymphoma cell line Jurkat when transfected with the chloramphenicol acetyltransferase reporter gene driven by 5Ј flanking sequences of the CXCR1 gene (Ϫ800 to ϩ21 bp). However, identification of regulatory elements and transcription factors regulating the myeloid lineage-specific expression of the CXCR1 gene have been precluded by the lack of suitable myeloid cell lines expressing high levels of CXCR1 or CXCR2. In this work we delineate cis-acting elements that direct myeloid-specific expression of the CXCR1 promoter and identify the ets family transcription factor PU.1 as a major regulator of the CXCR1 promoter.

EXPERIMENTAL PROCEDURES
Cell Culture-Cell culture supplies were purchased from Life Technologies, Inc. 32Dcl3 cells were a gift from Dr. Joel Greenberger (University of Pittsburgh Medical School, Pittsburg, PA). NIH3T3, RAW264.7, and CEMT4 cell lines were purchased from the American Type Culture Collection (Rockville, MD). 32Dcl3 cells were cultured in RPMI 1640 medium supplemented with 15% WEHI-3B conditioned media, 15% fetal bovine serum, and 50 IU/ml penicillin-50 g/ml streptomycin. RAW264.7 cells were cultured in Dulbecco's modified Eagles's medium supplemented with 10% fetal bovine serum and 50 IU/ml penicillin-50 g/ml streptomycin. CEMT4 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum and 50 IU/ml penicillin-50 g/ml streptomycin. NIH3T3 cells were cultured in Dulbecco's modified Eagles's medium (1 g/l D-glucose) supplemented with 10% bovine serum and 50 IU/ml penicillin-50 g/ml streptomycin. All cell lines were grown at 37°C and 5% CO 2 .
Plasmid Construction-Peripheral blood was collected from healthy human donors and placed in a 1% sodium citrate (anticoagulant) solution. Genomic DNA was extracted from leukocytes as described (17). A forward primer corresponding to Ϫ800 to Ϫ777 bp of CXCR1 and containing an XhoI site (5Ј-CGGCTCGAGGCTAACCAGCCAGACTCT-GGGAGT-3Ј), and a reverse primer corresponding to ϩ63 to ϩ86 bp of CXCR1 and containing a HindIII site (5Ј-GGACACACCTAAGCACCG-GCCAGGTGTGTC-3Ј) were used to amplify an 889-bp 5Ј-flanking region of CXCR1 (15 72°C for 1 min 30 s, and a final extension of 72°C for 10 min. The amplified PCR product was gel-purified and digested with XhoI and HindIII, and cloned into SalI-HindIII sites upstream of a luciferase cDNA in a promoterless pGEM-3 vector referred to as pLUC. Deletion constructs were made by digestion with unique restriction sites within the CXCR1 sequence (A2 construct, NdeI; A3 construct, AccI; and A4 construct, PstI). The A2 mutant construct was generated by cassette mutagenesis. The sequence (Ϫ58 to ϩ50 bp) was removed at AccI-PstI sites and replaced by a synthetic double-stranded oligonucleotide in which guanine nucleotides at positions Ϫ13, Ϫ12, Ϫ3, and Ϫ4 bp in the wild-type sequence were replaced with thymine nucleotides. The pGEM-3 vector containing the SV40 promoter and enhancer upstream of a luciferase cDNA (pSV40/LUC) was used as a positive control. The ␤-galactosidase reporter gene driven by the cytomegalovirus promoter (pCMV/␤-gal) was used as an internal control to correct for differences in transfection efficiency between experiments. The vectors pSV40/LUC and pCMV/␤-gal were provided by Dr. Allan Braiser (University of Texas Medical Branch, Galveston, TX).
Plasmids were purified by cesium chloride gradient centrifugation as described by Ausubel et al. (18), and the sequences were verified by dideoxy sequencing using a Sequenase kit (United States Biochemical Co., Cleveland, OH).
Reporter Gene Assays-Luciferase and ␤-galactosidase assays were performed in duplicate with cell lysates (10 l) from each transfection. Luciferase assays were performed as described in Promega Protocols and Applications Guide (19). Lysates were incubated with a luciferase assay reagent composed of 20 mM Tricine, 1.07 mM (MgCO 3 ) 4 MgOH 2 , 2.67 mM MgSO 4 , 0.1 mM EDTA, 270 M coenzyme A, 470 M luciferin, 530 M ATP, and 33.3 mM dithiothreitol. (Coenzyme A, luciferin, ATP, and dithiothreitol were added immediately before each assay.) Purified luciferase was used as a standard. Emitted light was recorded as relative light units using an Analytical Luminescence Laboratory Monolight 2010 luminometer. ␤-Galactosidase assays were performed as described by Sambrook et al. (20).
Electrophoretic Mobility Shift Assays (EMSAs)-Peripheral blood was collected from healthy human donors and placed in a 1% sodium citrate (anticoagulant) solution. Neutrophils and monocytes were separated from other cell types by centrifugation with mono-poly resolving medium (ICN Pharmaceuticals, Inc., Costa Mesa, CA). Nuclear extracts were prepared by cell lysis with the detergent Nonidet P-40 as described by Johnson et al. (21). All extracts were prepared in the presence of 0.5 mM phenylmethylsulfonyl flouride. The protease inhibitors leupeptin (1 M) and aprotinin (0.3 M) were also used in the preparation of extracts from neutrophils and monocytes. Protein concentrations were determined using the Bradford assay (22). The sense strand sequences of the probes are as follows: Double-stranded oligonucleotides were 5Ј end-labeled by polynucleotide kinase using 15 Ci of [␥-32 P]ATP (DuPont NEN). Unincorportated [␥-32 P]ATP was removed from the labeled probe with nick spin columns-Sephadex G-50 (Amersham Pharmacia Biotech).
DNA probes (0.5 ng) were incubated with 5 g of nuclear extracts and 2 g of poly(dI⅐dC) in binding buffer (20 mM Tris-HCl, pH 7.5, 2 mM EDTA, 2 mM dithiothreitol, 100 mM NaCl, 10% glycerol) for 30 min at room temperature. A 100-fold molar excess of unlabeled probe was used for competition experiments. For supershift analyses, 2 g of a rabbit polyclonal PU.1 antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA) or 2 l of rabbit preimmune sera were incubated with nuclear extracts for 15 min at 4°C before addition of probe. Binding reactions were electrophoresed on 6% acrylamide gels with 0.5 ϫ Tris borate-EDTA (4.45 mM Tris base, 4.45 mM boric acid, 1 mM EDTA, pH 8.0) as a running buffer. Gels were dried and exposed with intensifying screens to x-ray film for 24 -48 h at Ϫ70°C.

IL-8 Receptors Are Expressed in Myeloid Precursor
Cells-CXCR1 and CXCR2 are highly expressed in neutrophils and readily detected by Northern blot analysis. The short life span of neutrophils and the difficulties of transfecting these cells have precluded the use of neutrophils as a cellular system to elucidate the transcriptional regulation of neutrophil-specific genes. On the other hand, most cell lines of hematopoietic origin express trace amounts of CXCR1 or CXCR2 (23). Recently, we have identified a myeloid cell line, 32Dcl3 (32D), that increases intracellular calcium in response to IL-8 (12). Northern blot analysis of murine neutrophils and 32D, NIH3T3 (3T3), CEMT4 (CEM), and RAW264.7 (RAW) cell lines revealed high levels of murine IL-8 receptor mRNA in 32D cells and neutrophils (Fig. 1). The 32D cell line is a murine interleukin-3-dependent myeloid precursor cell line that differentiates into neutrophils in the presence of G-CSF (24). Recently we found that IL-8 and the related chemokine melanoma growth-stimulating activity suppress the proliferation of 32D cells (12). Although previously it was argued that mouse and rat only express the human homolog of CXCR2, two reports have identified the human homolog of CXCR1 in rat and mouse (25,26). These observations indicate that the murine 32D cell line is a suitable cellular system to identify the regulatory elements and transcription factors regulating the myeloid lineage-specific expression of the human CXCR1 gene.
Myeloid Lineage Specificity of the CXCR1 Promoter-Previous studies showed that CXCR1 promoter constructs containing the first 800 bp of 5Ј-flanking DNA directed activation of the reporter gene chloramphenicol acetyltransferase in hematopoietic cell lines expressing negligible levels of IL-8 receptors (15,16). To determine the specificity of the CXCR1 promoter for the myeloid lineage, myeloid (32D and RAW) and nonmyeloid (3T3 and CEM) cell lines were transfected with the Ϫ800 to ϩ86 bp/luc promoter construct (A1). As shown in Fig. 2, the A1 promoter construct exhibited the greatest reporter activities in 32D and RAW cell lines, 200-and 400-fold higher than the promoterless luciferase vector, respectively. In the nonmyeloid cells CEM and 3T3, the promoter activity of construct A1 was significantly lower than the activity observed in myeloid cells. Deletion of the sequence Ϫ800 to Ϫ126 bp (construct A2) increased reporter activity 2-fold relative to construct A1. This finding suggests that the sequence Ϫ800 to Ϫ126 bp contains negative promoter elements. Further deletion of the sequence Ϫ800 to Ϫ58 bp (construct A3) still exhibited high levels of promoter activity in the myeloid cells 32D and RAW. Surprisingly, the A3 construct showed higher levels of promoter activity in 3T3 cells than the A1 and A2 constructs, suggesting the presence of negative elements between Ϫ126 and Ϫ58 bp. Deletion of the Ϫ800to ϩ51-bp region (construct A4) abolished promoter activity. These data indicate the presence of cisacting elements from Ϫ126 to ϩ51 bp that direct positive promoter activity in the myeloid cells 32D and RAW and elements from Ϫ126 to Ϫ58 bp that direct negative promoter activity in the nonhematopoietic cell line 3T3.
The Sequence Ϫ22 to ϩ14 bp Binds Nuclear Proteins Present in 32D and RAW Cell Lines-EMSAs were performed to identify transcription factors binding to the CXCR1 promoter sequences in a myeloid-specific fashion. EMSAs were performed with 32 P end-labeled DNA fragments encompassing the Ϫ58to ϩ50-bp sequence, because the majority of myeloid-specific promoter activity is contained in this region. Only a fragment corresponding to the sequence Ϫ22 to ϩ14 bp (A2 probe) produced a fast-migrating, myeloid-specific complex (Fig. 3). An excess of unlabeled A2 probe displaced the binding of the labeled A2 probe to the myeloid-specific factor (Fig. 3, lanes 3 and  7); however, an excess of unlabeled nonspecific probe did not displace the binding of A2 probe (Fig. 3, lanes 11 and 15). A similar myeloid-specific complex was observed with the A2 probe and nuclear extracts from human neutrophils and human monocytes (data not shown). These results suggest that myeloid-specific transcription factors bind to the Ϫ22to ϩ14-bp fragment.
The Ϫ22to ϩ14-bp Fragment Binds to the Hematopoietic Transcription Factor PU.1-Analysis of the Ϫ22 to ϩ14 bp sequence revealed consensus binding sequences for transcrip-tion factors of the ets family. PU.1 is a member of the ets family that is expressed in myeloid cells, including neutrophils and macrophages, and B cells (27). To determine whether PU.1 binds to the Ϫ22to ϩ14-bp sequence (probe A2), the two putative PU.1 binding sites were first mutated. Disruption of the two putative PU.1 sites (M3) or the PU.1 site proximal to the transcription start site (M2) did not displace the A2 probe bound to the myeloid-specific protein, and M3 and M2 probes did not generate the myeloid-specific complex (Fig. 4). By contrast, disruption of the PU.1 site distal to the transcription start site (M1) effectively displaced the A2 probe bound to the myeloid-specific protein, and the M1 probe generated the myeloid-specific complex (Fig. 4). These findings strongly suggest that the core PU.1 binding motif is located Ϫ7 to Ϫ4 bp. Second, a fragment corresponding to the PU.1 binding site of the CD11b promoter was an effective competitor for formation of the myeloid-specific complex (Fig. 4, lanes 2-5) but not for nonmyeloid DNA-protein complexes formed with myeloid (Fig. 4, lanes 2-5) and nonmyeloid extracts (Figs. 4, lanes 6 -9, and 5A). Third, the PU.1 fragment produced a similar myeloid-specific complex as the Ϫ22to ϩ14-bp fragment A2 with extracts from 32D (Fig.  5A, lane 11) and RAW (Fig. 5A, lane 13), and A2 was an effective competitor for the formation of the PU.1 complex (Fig.  5A, lanes 12 and 14). Fourth, EMSA in the presence of antibodies that specifically recognize human and murine PU.1 produced supershifts of the myeloid-specific complexes generated with the PU.1 fragment (Fig. 5A, lanes 3 and 6) and the Ϫ22 to ϩ14 bp sequence (probe A2) (Fig. 5B, lanes 10 and 12). Fifth, Northern blot analysis revealed that PU.1 mRNA is expressed specifically in the myeloid cells 32D and RAW but not in nonmyeloid cells CEM and 3T3 (Fig. 6). These results demonstrate that the myeloid-specific complex is generated by PU.1 binding to the Ϫ22 to ϩ14 bp fragment.
PU.1 Binding Site Is Essential for IL-8 Receptor A Promoter Activity-To determine whether the PU.1 binding site in the Ϫ22to ϩ14-bp sequence is functional, myeloid and nonmyeloid cell lines were transfected with the Ϫ126 to ϩ86 bp/luc construct mutated at this site. Disruption of the PU.1 binding site abolished the promoter activity of the Ϫ126 to ϩ86 bp/luc construct (A2) (Fig. 7). This result indicates that the PU.1 site is essential for promoter activity and that compensatory elements are not present in this construct to drive the expression of the reporter gene. To directly demonstrate that PU.1 binds and activates the IL-8RA promoter, the nonmyeloid cell lines that do not express PU.1, CEM, and 3T3 were cotransfected with the Ϫ126 to ϩ86 bp/luc construct (A2) and an expression vector encoding PU.1. Cotransfections with PU.1 cDNA increased promoter activity 4-fold in CEM cells (Fig. 8A) and Ͼ16-fold in 3T3 cells (Fig. 8B) compared with vector alone. These findings strongly suggest that the myeloid-specific expression of the CXCR1 gene is activated by PU.1 interacting with promoter sequences adjacent to the transcription start site.

DISCUSSION
Chemokines are major regulators of the proliferation of myeloid precursor cells (12,13,28). However, little is currently known about the mechanisms regulating the expression of chemokine receptors during the commitment and differentiation of progenitor cells toward myeloid lineages and neutrophil development, in particular. The restrictive expression of IL-8 receptors in myeloid precursor cells and neutrophils provides a system to identify the regulatory elements and transcription factors that may regulate the commitment of progenitor cells to the neutrophil lineage. In this study, the promoter activity of the proximal Ϫ800 bp of the CXCR1 gene was analyzed. This sequence was found to contain the regulatory elements that direct CXCR1 promoter activity in a myeloid-specific fashion. High levels of promoter activity were detected specifically in 32D (myeloid precursor cells) and RAW (macrophages). Sequences Ϫ800 to Ϫ126 bp were found to contain negative regulatory elements. This finding is in agreement with that of Sprenger et al. (15), who suggested the presence of silencer elements between positions Ϫ841 and Ϫ280 bp on the basis of transfection studies with chloramphenicol acetyltransferase reporter genes in nonmyeloid cell lines. Because most of the promoter activity is localized within the Ϫ56to ϩ50-bp sequence, myeloid-specific proteins binding in this region were identified. On the basis of EMSA, the transcription factor PU.1 was shown to bind the Ϫ22to ϩ14-bp fragment with the common GGAA binding motif at position Ϫ7 to Ϫ4 bp. Disruption of the PU.1 binding site abolished the myeloid-specific transcriptional activity of the CXCR1 promoter. Transfection of nonmyeloid cell lines CEM (T cells) and 3T3 (fibroblasts) with cDNA encoding PU.1 increased the promoter activity of the Ϫ126 to ϩ86 bp/luc construct. Because PU.1 expression in nonmyeloid cell lines produced high levels of promoter activity, this suggests that PU.1 does not require other myeloid-specific factors for activation of the promoter construct. These data show for the first time the transcriptional regulation of a chemoattractant G protein-coupled receptor by the myeloid transcription factor PU.1.
Our studies reveal that the regulatory sequences analyzed in this study do not direct cell-specific expression, because high levels of promoter activity are demonstrated in both 32D cells, which express IL-8 receptors, and RAW cells, which do not express IL-8 receptors. Similar findings were observed with the promoter of the eosinophil-specific IL-5 receptor, in which high levels of promoter activity were obtained in both myeloid and eosinophilic cell lines (36). The neutrophil lineage-specific expression of CXCR1 transcripts could possibly be attributable to post-transcriptional mechanisms or transcription factor regulatory sites located further upstream or downstream of the sequences analyzed in this study. Further experiments will be focused on mapping additional functional elements of the CXCR1 promoter and identifying the transcription factors that bind to these elements to elucidate the mechanisms regulating the neutrophil lineage-specific expression of CXCR1.