Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

5-lipoxygenase (ALOX5) (arachidonate 5-lipoxygenase: oxygen 5-oxido-reductase, E.C. 1.13.11.34) is a calcium-, adenosine triphosphate (ATP)-, iron-requiring enzyme that catalyzes the first committed steps in the metabolic pathway leading to the synthesis of all leukotrienes (LT) (1, 2). The human ALOX5 gene spans > 82 Kbp, contains 14 exons, and is located on chromosome 10q11 (3). The gene is expressed in a tissue-specific fashion, primarily in cells of myeloid lineage, including neutrophils, mast cells, eosinophils, monocytes/macrophages, and lymphocytes (6). In response to cytokines and other inflammatory stimuli (e.g., transforming growth factor-, granulocyte-monocyte colony-stimulating factor, interleukin-3, oxidized low-density lipoprotein, phorbol esters, Ca²⁺ ionophores), some leukocyte cell lines increase ALOX5 mRNA, protein, and catalytic activity (7). This regulation is complex and involves control mechanisms at multiple levels, including gene transcription (10), enzyme translocation (11), cofactor interactions (14), and substrate availability (15, 16).

The human ALOX5 promoter has been cloned (GENBANK #M38191) and characterized in a variety of cell types (17). The sequence is G+C-rich, lacks a TATA box, and has multiple transcription start sites. Using promoter-reporter constructs in the context of transient-transfection analysis, two important cis-elements have been comprehensively examined: (i) five tandem SP1/early growth-response protein 1 (EGR-1) consensus-binding sites, between 145 to 179 base pairs (bp) relative to the ATG translational start site¹, that function as positive regulatory elements (17, 18); and (ii) a c-Myb consensus site, located between 1,775 to 1,787 bp, that functions as a negative regulatory element (19). In addition, Hoshiko and colleagues have identified two positive regulatory regions between 3,700 and 5,900 and between854 and 931, and two negative regulatory regions between 1,557 and 3,400, and between 292 and 727. Although the promoter contains numerous consensus-binding sites for transcription factors, including NF-B, glucocorticoid receptors, GATA, and AP1 family members, their functional significance remains unknown.

Direct comparisons of human and mouse genomic sequences have been used to locate important regulatory elements in promoters (20, 21). Although the mouse Alox5 cDNA was cloned (22), there was no promoter sequence available in the literature or public databases to compare with the human sequence. We now describe and report a functional analysis of the mouse Alox5 promoter sequence cloned from a 129/SvJ BAC library. Direct comparison of the mouse promoter sequence with the human sequence facilitated a functional analysis of cell-specific expression in RAW 264.7 cells by the identification of one major regulatory element.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Promoter Cloning

A bacterial artificial chromosome (BAC) 129/SvJ mouse genomic DNA filter (ES BAC Release II; Incyte Genomics, Inc., Palo Alto, CA) was screened by hybridization with overlap radiolabeled oligonucleotide probes corresponding to a portion of Exon I (sense 5'-TTCGCGGGCACCGACGACTACAT-3', antisense 5'-ATGAGGCTCAGGTAGATGTAGTC-3') of the mouse Alox5 cDNA (GENBANK #L42198) (22). Probe radiolabeling and hybridization were performed by standard techniques (23). A strong hybridization signal was detected for eight BAC clones (269F6, 281M7, 352M7, 375M7, 253L20, 280N19, 336B23, 408A11). Clone fragments were generated with standard techniques by EcoRI digest, screened for the presence of Alox5 sequence by Southern blot hybridization, and subcloned into the LITMUS 28 vector (New England Biolabs, Inc., Beverly, MA) for sequencing (23). The promoter sequence was confirmed by polymerase chain reaction with mouse genomic DNA and checked for consistency with GENBANK genomic and EST databases.

Cell Culture

RAW 264.7 (mouse monocyte-macrophage), 3T3-Swiss albino (mouse embryo fibroblasts), and Drosophila SL2 cells were obtained from the American Type Culture Collection (Rockville, MD) and grown in RPMI, Dulbecco's modified Eagle's medium, or Schneider's Drosophila Medium, respectively (Life Technologies, Inc., Gaithersburg, MD). Medium contained 50 µg/ml streptomycin, 50 IU/ml penicillin, and 10% fetal calf serum. RAW and 3T3 cells were grown in 60 mm petri dishes at 37°C, 5% CO₂, and passaged at confluence every 4 d. SL2 cells were grown at 27°C and passaged at confluence every 3 d.

RNA Processing, RT-PCR, and Primer Extension

Total RNA was extracted with Trizol reagent in accordance with the manufacturer's instructions (Life Technologies, Inc.). Samples were aliquoted and stored at 80°C.

Reverse transcription (RT)-PCR was performed with 2 µg total RNA, 1 µg random hexamers (Life Technologies, Inc.), 50 nmoles dNTP, 25 U rRNasin RNase inhibitor (Promega, Madison, WI), and 200 U M-MLV RT (Promega) in a 25 µl total volume consisting of 50 mM Tris-HCL (pH 8.3), 75 mM KCl, 3 mM MgCl₂ and 10 mM dithiothreitol at 37°C for 60 min. PCR reactions were performed with 4 µl 1:5 diluted RT reactions, 40 nmoles dNTP, 10 pmoles primers, and 2 U Taq polymerase (Promega) in a 50 µl total volume consisting of 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% Triton×-100, and 1.5 mM MgCl₂. Alox5 primers were (sense) 5'-CACGGGGACTACATCGAGTT and (antisense) 5'-AAC CTCACATGGGCTACCAG, and p53 primers were (sense) 5'-CCGAGGCCGGCTCTGAGTATACCACCATCC and (antisense) 5'-CTCATTCAGCTCCCGGAACATCTCGAAGCG. Annealing temperature was 60°C with 30 PCR cycles. PCR products were separated on a 1.5-2% agarose gel and photographed.

Primer extension was performed as described by Mason and coworkers (24). Transcript primer was 5'-ATGAGGCTCAGGTAG ATGTAGTC and negative control primer was 5'-CAGCCCAC CAACTTCTTGAC, which is located upstream of the primary transcription start site. Hybridization was performed with 30 µg of total RNA at 65°C for 90 min. Extension reactions were run on an 8% sequencing gel adjacent to the promoter sequence (T7 Sequenase 2.0 sequencing kit, Amersham Life Science, Inc., Cleveland, Ohio).

Construction of Reporter Plasmids and Transient Transfection Analysis

Alox5 promoter-reporter plasmids were constructed by placing PCR-generated promoter fragments into the KpnI and HindIII sites of the pGL3-Basic vector (Promega) by standard techniques. The antisense primer for all constructs was 5'-AAGCTTACTC CGGGCAAGTGAGTG. Site-directed mutagenesis was performed with a QuikChange Site-Directed Mutagenesis Kit according to the manufacturer's instructions (Stratagene, La Jolla, CA). Mutant oligos were Pu.1Mut 5'-GGTAGAGATGGGG CAGAGAGACAAGGGCTTCAGC and Sp1Mut 5'-GACTG GCCAGAGACAGGTTCGGAGCCAGCGCCTGAAG. Correct orientation was confirmed by restriction digest, using sites within the polylinker and the insert, to release the appropriate size fragments. All plasmids were purified with a Qiagen Plasmid Mega Kit (Qiagen Inc., Valencia, CA) and sequenced by standard dideoxy methods to determine accuracy. The details of plasmids pPAC-Sp1, pcDNA1-PU.1, and pcDNA-MutPU.1 (PU.1 dominant negative) are described elsewhere (18, 41).

Cells were transient transfected with SuperFect Transfection Reagent (Qiagen Inc.), according to the manufacturer's instructions, and incubated for 24 h at 37°C or 27°C in growth medium. Luciferase assays were performed with Reporter Lysis Buffer and Assay Reagent (Promega) according to the manufacturer's instructions. Luciferase data were normalized to the core promoter construct 292 arbitrarily set at 100. Transfection efficiency was monitored by cotransfection with pSV--Galactosidase (Promega) according to manufactures instructions. Comparisons of promoter activity between RAW and 3T3 cells required normalization for transfection efficiencies. Cotransfection studies were not normalized because overexpression of PU.1 enhanced pSV--Galactosidase expression. Transfection efficiencies were ~ 10% for RAW cells and 20% for 3T3 cells as determined by cytochemical staining for -Galactosidase activity (23).

Nuclear Protein, EMSA, and Western Blot Analysis

Nuclear proteins were prepared as described by Dignam and coworkers and stored at -80°C in small aliquots (25).

In vitro binding reactions between oligonucleotides and nuclear extract were done in a total volume of 20 µl containing: 2 µl of 10× binding buffer (0.1 M Tris-HCL, pH 7.5, 50% glycerol, 10 mM ethylenediaminetetraacetic acid, 10 mM dithiothreitol), 1 µl of 1 µg/ul poly(dI · dC) poly(dI · dC) (Sigma, St. Louis, MO), 1 µl of 1 µg/µl salmon sperm DNA (Sigma), 2-4 µl of nuclear extract (normalized to ~ 8 µg/µl of total protein), and 1 µl klenow-radiolabeled oligonucleotide (specific activity 25-50,000 cpm/µl). The reaction was allowed to proceed for 30 min at 22°C before the addition of 2 µl of nondenaturing loading buffer (0.2% bromophenol blue, 0.2% xylene cyanol, 20% glycerol). Samples were electrophoresed on 1.5 mm-thick 4-8% polyacrylamide gels with Tris-Boric acid-EDTA running buffer at 200V for ~ 1-2.5 h. Gels were dried under vacuum and autoradiographed overnight with Kodak X-Omat-AR film. Oligonucleotides Sp1/Sp3 (5'-CAGAGACA GGGGCGGAGCCAGCGC), GATA/GGAGA (5'-GGCTTCA GCCTATCAGAGATCACT), and PU.1/Spi-1 (5'-GATGGGGCA GAGGAACAAGGGCTT) were synthesized by Gene Link, Inc. (Hawthorne, NY). Supershift and competition studies were performed as above, except that 1 µl (2 µg) of antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was added 10 min before the addition of radiolabeled oligonucleotide.

Western blots were performed as previously described in detail (26). The membranes were probed sequentially with polyclonal antibodies to Pu.1 and Sp1 (2 µg/µl) (Santa Cruz Biotechnology) at dilutions of 1:10,000, followed by enhanced chemiluminescent detection (Amersham, Arlington Heights, IL) with 1:10,000 horseradish peroxidase-linked secondary antiserum. Blots were exposed to x-ray film for 1-5 min and stained with Coomassie blue.

Statistical Analysis

Relative luciferase activity of the various constructs was compared by analysis of variance or linear regression analysis with StatView software (Abacus Concepts, Inc., Berkeley, CA). When significant differences were noted among groups, the Bonferroni/ Dunn procedure was used for post hoc testing. Tests were conducted at the 5% significance level, and results are expressed as mean ± standard error of the mean.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of the Promoter Region and Comparison with the Human Promoter

Sequencing of 129/SvJ BAC fragments yielded more than 2,000 bp of a novel mouse Alox5 genomic sequence located upstream of the ATG start site. The full-length sequence has been deposited in GENBANK under accession number AF393814. Figure 1 shows ~ 600 bp of promoter sequence with essential features annotated. A direct comparison of the mouse Alox5 and human ALOX5 5'-flanking sequences using Blast 2 Sequences (27), PipMaker (28), and Sequencher 4.0 (Gene Codes Corp. Ann Arbor, MI) failed to align the sequences or find any gross similarities. Both mouse and human 5'-flanking sequences are G+C-rich, with the core promoters (first 300 bp) having GC ratios of 66% and 79%, respectively. Both promoters lack obvious TATA and CCAAT sequences. Screening of both sequences for transcription factor consensus-binding sites using TRANSFACT 5.0/MatInspector 2.2 (29) and TFSearch 1.3 (30) revealed many common consensus-binding sites for transcription factors. However, only consensus binding sites for SP1/SP3, GATA/GGAGA, and ETS/PU.1/ SPI-1 (Figure 1) are found in similar locations in both promoters as discussed in detail below.

View larger version (28K): [in this window] [in a new window]   Figure 1.   Mouse Alox5 promoter sequence. 599 base pairs of the mouse Alox5 promoter cloned from a 129/SvJ BAC library are shown with corresponding numbers relative to the translational start site. Translated sequence is underlined. Intron I flanking sequence is shown in lower case letters. The 5' untranslated region in RAW cells is shown in bold, and the major transciption start site is indicated by the large arrow. As with the human ALOX5 gene, multiple transcriptional start sites were detected. A minor site is indicated by the small arrow. Core sequences with strong homology to the human promoter are outlined along with their contained consensus-binding sites to known transcription factors. Additional promoter sequence can be found in GENBANK AF393814. Of note, the human and mouse intron I location is perfectly conserved.

Identification of the Major Transcriptional Start Site in RAW Cells

RAW 264.7 cells are known to express Alox5 constitutively unlike most other mouse cell lines (2, 31). To define the transcription initiation site and the beginning of the 5'-untranslated region of Alox5 in RAW cells, we used primer extension analysis. Figure 2 shows ~ 4 bands generated with an extension primer corresponding to a sequence in exon 1. The major transcription initiation site in RAW cells is located at 52 bp. Three minor initiation sites appear be located 5' to the major site but the intensity was low and their precise location difficult to determine. In contrast, a control primer beginning 5' to the major initiation site did not produce extension bands. Analysis of the dbEST database revealed one transcript (Id: 1373589) with a 5' end corresponding to -130 bp, indicating that other major initiation sites are possible as suggested by our minor bands.

View larger version (61K): [in this window] [in a new window]   Figure 2.   Major transcription start site by primer extension. RNA from RAW cells was annealed with a primer complementary to a region of exon I, reverse transcribed, and resolved next to a sequencing reaction on an 8% denaturing gel. The location of the major transcriptional start site corresponds to G at location -52 bp. Control primer complementary to the sequence beginning 5' of start site did not produce bands. A few minor transcription start sites are evident, but their precise location could not be elucidated.

Deletion Analysis of the Alox5 Promoter Defines the Core Promoter

To study promoter function, a set of luciferase reporter plasmids containing 5'-deletions were constructed for transient-transfection analysis in RAW 264.7 cells. Figure 3 shows normalized luciferase activity for each plasmid. The largest promoter-reporter construct (2,099) contained 2,099 bp of promoter and resulted in a 27-fold luciferase induction relative to pGL3-Basic backbone (0). Variation in promoter activity as sequential promoter regions were removed suggested the presence of enhancer or silencer elements within 2,099 bp of the ATG start site. The first 292 bp of the 5'-flanking sequence conferred relatively high constitutive activity and was defined as the core promoter. Additional 5'-deletions resulted in a dramatic fall in promoter activity, with virtually all activity eliminated when only 91 bp of promoter remained. Thus, the region 292 to 91 contains regulatory elements essential for maximal constitutive promoter activity. Negative regulatory elements may be contained within the promoter sequence between bp 2099 to 1012 and 484 to 445.

View larger version (21K): [in this window] [in a new window]   Figure 3.   5'-deletion analysis defines location of core promoter. RAW cells were transfected with 10 µg of reporter constructs containing 5'-deletions of the Alox5 promoter. The promoter length is indicated by the plasmid number, and data are expressed as normalized luciferase activity (mean ± SEM). Transfection efficiency was monitored by cotransfection with pSV--Galactosidase Control (Promega). Constitutive activity varies as a function of promoter length, with luciferase levels falling dramatically after -292 bp. A construct containing -91 bp has negligible activity. P values between core promoter constructs are provided (n = 4-7 for all experiments).

Alox5 Promoter Mediates Cell Specific Expression

Unlike RAW cells, 3T3 fibroblasts do not express Alox5 mRNA by RT-PCR (see Figure 4 insert). To assess the Alox5 promoter's ability to mediate cell-specific transcription, the deletion series was compared in these two cell lines (Figure 4). When normalized for differences in transfection efficiency according to -Gal activity, the core promoter (292) yielded 12-fold higher luciferase activity in RAW cells than in 3T3 cells. Constructs 251 and 152 also produced higher luciferase levels in RAW cells than in 3T3 cells, with fold inductions of 21 and 15, respectively. Relative induction was lost with construct -91, but luciferase levels were exceedingly small. These data suggest that cis-elements within the core promoter mediate cell-specific Alox5 expression.

View larger version (34K): [in this window] [in a new window]   Figure 4.   Alox5 promoter has greater activity in RAW versus 3T3 cells. Insert shows agarose gel demonstrating the presence of Alox5 mRNA in RAW cells but not 3T3 cells by RT-PCR. There are no differences in p53 expression shown as a control. To compare promoter activity in the two cell types, each tissue culture plate was transfected with 10 µg of reporter constructs with luciferase activity measured 24 h later. The promoter length is indicated by the plasmid number, and data are expressed as normalized luciferase activity (mean ± SEM). Constructs -292, 251, and -152 were 12- to 21-fold more active in RAW cells than 3T3 cells. Activity was negligible in both cell types with only 91 bp of promoter. P values between cell types are provided.

Pu.1/Spi-1, Sp1, and Sp3 Bind to Cis Elements in the Core Promoter

Direct comparison of mouse and human promoter sequences may reveal important regulatory regions. Thus, we aligned the human ALOX5 core promoter with the mouse Alox5 core promoter beginning from the ATG start site and extending 5'. Three limited areas of strong sequence homology were identified: (i) an Sp1/Sp3 consensus binding site (GGGCGG) located at bp 184 to 189 in the mouse promoter that corresponds to 5 tandem Sp1/Sp3 sites located 145 to 179 in the human promoter; (ii) a Gata/GGAGA consensus binding site located at bp 261 to 269 in the mouse promoter and bp 263 to 273 in the human promoter; and (iii) a Ets-like/Pu.1/Spi1 consensus binding site (AGGAACA) located at bp 289 to 293 in the mouse promoter and bp 284 to 290 in the human promoter (see Figure 1).

Electrophoretic mobility shift assays (EMSAs) were performed to test the hypothesis that these transcription factors are present in RAW cells and capable of binding specifically to these regions of the Alox5 promoter. Incubation of nuclear extracts from RAW cells with a radiolabeled oligonucleotide corresponding to the putative Sp1/Sp3 promoter site (175 to 198) produced three primary complexes (Figure 5). The slowest migrating complex supershifted with antibodies specific for Sp1. The intermediate complex supershifted with antibodies specific for Sp3. Antibodies to Egr-1, Gata-1, and c-jun did not alter the migration pattern of either complex. The identity of the fastest migrating complex is unknown. Alteration of the consensus site by two point mutations eliminated all binding events. Thus, RAW cells contain Sp1 and Sp3 nuclear protein capable of specifically binding to the Alox5 promoter Sp1/ Sp3 consensus site. Egr-1 binding could not be detected.

View larger version (63K): [in this window] [in a new window]   Figure 5.   EMSA demonstrating Sp1 and Sp3 binding to the Alox5 core promoter. RAW nuclear extracts were incubated with radiolabeled Oligo Sp1/Sp3 (WT) corresponding to bp -175 to -198 of the Alox5 promoter and resolved on a nondenaturing gel. Three gel shifts are produced. Coincubation with antibodies against Sp1 and Sp3 supershift low migration (top) and middle complex. Antibodies against Egr-1, Gata1, and c-jun had no effect. Two point mutations (MutSp1) in the Sp1 and Sp3 consensus site eliminate Sp1/Sp3 binding.

Figure 6 shows a similar EMSA using an Alox5 radiolabeled oligonucleotide corresponding to the putative Ets/ Pu.1/Spi-1 site (276 to 299). Incubation with RAW cell nuclear extract produces two gel shifts. The slowest migrating complex supershifts with antibodies specific for Pu.1/Spi-1 and not with an equal concentration of control antibodies. Again, elimination of the consensus-binding site by two point mutations eliminated specific binding. Thus, RAW cells contain Pu.1/Spi-1 nuclear protein capable of specifically binding to the Alox5 promoter Ets/PU.1/ Spi-1 consensus site. We were unable to detect Gata(1-6) binding by EMSA (data not shown).

View larger version (59K): [in this window] [in a new window]   Figure 6.   EMSA demonstrating Pu.1/Spi-1 binding to the Alox5 core promoter. RAW nuclear extracts were incubated with radiolabeled Oligo Pu.1/Spi-1 (WT) corresponding to bp -276 to -299 of the Alox5 promoter and resolved on a nondenaturing gel. Two gel shifts are produced. Coincubation with an antibody against Pu.1/Spi-1 supershifts the low migration (top) band. Antibodies against Sp1, Gata2, and Usf-1 have no effect. Two point mutations (MutPu.1/ Spi-1) in the Pu.1/Spi-1 consensus site eliminate Pu.1/Spi-1 binding.

RAW Cells Express Pu.1/Spi-1, Whereas 3T3 Cells Do Not

We speculated that differences in the expression of Pu.1/ Spi-1 between RAW and 3T3 cells may account for the significantly greater Alox5 promoter activity in RAW cells. To examine this hypothesis we used Western blot analysis and EMSA. Figure 7A shows a Western blot with antibodies to Pu.1/Spi-1 and Sp1. RAW cells contains Pu.1/Spi-1 protein, whereas 3T3 cells do not. In contrast, there were no significant differences in Sp1 or Sp3 between the two cell lines. EMSA confirmed these observations with a complex corresponding to Pu.1/Spi-1 when RAW nuclear extracts but not 3T3 nuclear extracts were used (Figure 7B).

View larger version (46K): [in this window] [in a new window]   Figure 7.   Pu.1/Spi-1 is expressed in RAW cells but not 3T3 cells. (A) Western blot analysis. Nuclear extracts from RAW and 3T3 cells were run on an 8% polyacrylamide gel, transferred to a nitrocellulose membrane, and incubated with anti-Pu.1/Spi-1 antibody. Only RAW cells contain Pu.1/Spi-1, whereas both cells contain Sp1 and Sp3 (data not shown). (B) EMSA, shows Pu.1/Spi-1 band with RAW nuclear extracts but not 3T3 nuclear extracts.

The Sp1/Sp3 Consensus Site is Essential for Constitutive Alox5 Promoter Activity

To examine the importance of these binding events, we performed site-directed mutagenesis on the -292 core promoter such that either the Sp1/Sp3 or Pu.1/Spi-1 sites were ablated. Figure 8 shows transfection results comparing the wild-type genotype (WT) and two mutated promoters. Mutation of the Sp1/Sp3 site has profound effects on promoter function, with a 12-fold reduction in luciferase levels compared with the wild-type promoter (P < 0.0001). In contrast, mutation of the Pu.1/Spi-1 results in only a slight, but statistically significant, decrease (~ 20%) in basal promoter activity (P < 0.02).

View larger version (19K): [in this window] [in a new window]   Figure 8.   Mutation of Sp1/Sp3 or Pu.1/Spi-1 site diminishes Alox5 promoter activity. RAW cells were transiently transfected with 10 µg of WT or mutated -292 luciferase and harvested at 24 h. Mutation of Sp1/Sp3 site decreased basal promoter activity to 8% of WT activity. Mutation of Pu.1/Spi-1 led to a slight but significant reduction (20%) in basal promoter activity (P values are shown; n = 5-8).

Consistent with these observations, overexpression of Sp1 (pPACSp1) in SL2 cells resulted in a dramatic increase in luciferase activity of the 292 promoter in a dose-response pattern (P < 0.01) (Figure 9), whereas overexpression of Pu.1/Spi-1 (pRc/RSV-PU.1) in 3T3 cells, or a dominant negative form of Pu.1/Spi-1 (pRc/RSV-MutPU.1) in RAW cells, had little effect on promoter activity (Figure 9).

View larger version (17K): [in this window] [in a new window]   Figure 9.   Overexpression of Sp1 increases Alox5 promoter activity. SL2, RAW, and 3T3 cells were cotransfected with 9 µg WT 292 luciferase and increasing amounts of Sp1 (pPAC-Sp1), PU.1(pRc.RSV-PU.1), and dominant negative PU.1(pRc/RSV-MutPU.1) expression constructs, respectively. Sp1 overexpression increased luciferase activity in a dose response pattern. PU.1 or MutPU.1 had little effect on Alox5 core promoter activity (P values are shown; n = 3-5).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

ALOX5 catalyzes the two-step conversion of arachidonic acid to LTA₄ and is essential for the synthesis of all LTs (1). The enzyme is tightly regulated in a cell-specific and inflammation-inducible manner through a variety of mechanisms, including transcriptional controls (2). To identify essential regulatory elements in the promoter that controls gene transcription, we cloned the mouse sequence for direct comparison with the human sequence. This strategy has been used to identify important gene regulatory regions, as they are often conserved across species (20, 21, 32).

Alignment of the human and mouse promoters (2,000 bp from ATG start site) revealed several interesting observations. First, there are dramatic differences in the two sequences such that three algorithm-based computer programs failed to find large regions of homology. These findings initially led us to question the authenticity of the cloned promoter sequence; however, subsequent experiments using PCR with multiple primer sets confirmed that the cloned sequence was indeed just 5' of the Alox5 open reading frame. Furthermore, we have shown that the cloned fragment has leukocyte-specific constitutive promoter activity, in the context of transient-transfection analysis with reporter constructs, that is consistent with the tissue expression profile of the gene (6, 33). The second interesting observation is that both sequences lack classic TATA motifs, TFIIB recognition elements, initiator sequences, and downstream core promoter elements (34) and have multiple transcriptional start sites in similar regions of the promoter. The major start site of the human gene is at 65 bp (17) compared with 52 bp in the mouse (Figure 2). The third important observation is that deletion analysis reveals core promoter regions of ~ 300 bp in both genes (17) (Figure 3). We define core promoter as the minimal promoter sequence required for sustained levels of luciferase activity, such that removal of an additional 5' sequence, dramatically decreases activity. Finally, three regions of strong homology (SP1/SP3, PU.1/SPI-1, and GATA/GGAGA) were detected in this core promoter region and contain consensus-binding sites for transcription factors implicated in ALOX5 regulation or leukocyte-specific gene expression. No other obvious sequence element (e.g., inverted repeats, repeated sequence, homology box, Myb, AP-2, NF-B site) listed by Hoshiko and colleagues as a potentially important cis element was identified at a similar location in the mouse promoter (17).

SP1 and SP3 are highly conserved zinc-finger transcription factors expressed in nearly all differentiated cells (35). While SP1 is a transcriptional activator of many promoters, SP3 is generally regarded as a suppressor of SP1-mediated transcription (36). Both SP1 and SP3 have DNA-binding domains that recognize GC boxes (GGGCGG) found in the human and mouse ALOX5 promoters. The one Sp1/Sp3 site located at bp 184 to 189 in the mouse promoter is orthologous to the five tandem SP1/SP3/ EGR-1 sites located at bp 145 to 179 in the human promoter. Hoshiko and coworkers have demonstrated that SP1 binds to this region and that deletion of this region decreases activity of the human promoter by ~ 70% (17). We now show that Sp1 and Sp3 bind specifically to this region of the mouse promoter and that mutation of the consensus-binding site dramatically decreases constitutive promoter activity in RAW cells. Moreover, overexpression of Sp1 in SL2 cells, known to be deficient in Sp1, resulted in a dramatic increase in core promoter activity. These data suggest that this is an important regulatory sequence in the mouse promoter. The five tandem SP1/SP3/ EGR-1 binding sites in the human promoter are of particular interest because of the existence of common polymorphisms in this region of the gene that result in a variable number of these consensus-binding sites (two to seven) (37, 38). The polymorphisms alter affinity for SP1/SP3/ EGR-1 and promoter response to overexpression of these transcription factors (18). Moreover, these sites have asthma pharmacogenetic implications for the treatment of asthma because there is a correlation between these polymorphisms and response to LT inhibitors (39). It is unknown if this site is polymorphic in the mouse. Our data provide additional evidence to suggest that these polymorphisms are important in humans because the Sp1 site is structurally and functionally conserved across human and mouse species. In other words, if a cis element is functionally important in both species, polymorphisms affecting the number of cis elements are more likely to be significant. Furthermore, the fact that the mouse promoter lacks an Egr-1 consensus-binding site in this region and fails to respond to Egr-1 overexpression (data not shown) suggests that SP1 and SP3 are more important in the regulation of ALOX5 than Egr-1.

PU.1, also known as SPI-1, is a member of the ETS family of transcription factors that function as activators or repressors of gene expression in a variety of biologic contexts. Members of this large and divergent group of transcription factors share a conserved winged helix-turn-helix DNA-binding domain that recognizes the core motif GGAA/T (40, 41). This consensus-binding site is found at multiple locations in the human and mouse ALOX5 promoters. Similar to ALOX5, PU.1/SPI-1 is expressed in cells of primarily hematopoietic lineage, including macrophages, mast cells, neutrophils, and B cells (42, 43), and has been shown to regulate the expression of a variety of genes important in inflammation, adaptive and innate immunity, and cell differentiation and proliferation (44). Little is known about the mechanisms responsible for modulation of PU.1/SPI-1's activity, but posttranscriptional modifications, such as phosphorylation and interactions with other proteins in the context of ternary complexes, are probably most important as opposed to changes in the level of protein (40, 44).

We have shown that the human and mouse Alox5 promoters contain an orthologous Pu.1/Spi-1 binding site in their core promoters at bp 284 to 290 and 289 to 293, respectively, and hypothesized that this site may play a role in tissue-specific expression of Alox5. Nuclear extracts from RAW cells express Pu.1/Spi-1 that can bind specifically to the consensus site in the mouse promoter, and point mutations of this site cause a slight but significant downregulation of promoter activity. The fact that Pu.1/Spi-1 is expressed in RAW cells along with Alox5 and that both are undetectable in 3T3 cells suggests that Pu.1/ Spi-1 plays a role in Alox5 cell-specific transcription. However, overexpression of Pu.1/Spi-1 in 3T3 cells, or a dominant negative form of Pu.1/Spi-1 in RAW cells, had no significant effect on core promoter activity suggesting that this interaction is unlikely to have functional significance. We acknowledge that these experiments are inconclusive because protein phosphorylation and interactions with other potentially essential factors are possible and could not be duplicated in these cells. Moreover, our studies have been restricted to constitutive, cell-specific expression and not cytokine induction. Thus, the role of this site in the upregulation of Alox5 in the context of an inflammatory microenvironment remains unknown.

Many other consensus-binding sites for known transcription factors are present in both promoters. Some of these transcription factor families, for example, GATA (45), USF (46), AP, NFAT (47), C/EBP (48), and GGAGA[A/G] (49), have been implicated in leukocyte-specific gene expression. The GATA and GGAGA consensus-binding sites are of particular interest because they are found in similar locations in the human and mouse core promoters; however, we were unable to detect any DNA-protein interactions by supershift analysis with antibodies to these proteins. The roles of these consensus-binding sites remain unknown. As additional ALOX5 sequences become available from genome sequencing projects involving other organisms, it may be useful to perform additional comparisons to identify new regulatory sequences.