Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

5-Lipoxygenase (5-LO; arachidonate: oxygen 5-oxido-reductase, E.C. 1.13.11.34, 5-LO) is a calcium-, adenosine triphosphate-, non-heme iron-requiring enzyme that catalyzes the two-step lipoxygenation of arachidonic acid to form leukotriene (LT) A₄. The first 5-LO catalyzed reaction converts arachidonic acid to 5-hydroperoxyeicosatetraenoic acid; the second 5-LO catalyzed reaction yields LTA₄. The enzyme product LTA₄ is an unstable intermediate that degrades rapidly unless it serves as a substrate for one of two enzymes, namely LTA₄ epoxide hydrolase or LTC₄ synthase. In neutrophils, LTA₄ serves predominantly as a substrate for LTA₄ epoxide hydrolase, which catalyzes the formation of LTB₄. Although a variety of biological actions have been ascribed to LTB₄, it is predominantly a chemotactic moiety for eosinophils and neutrophils (1, 2). In other cells, including eosinophils and mast cells, LTA₄ serves predominantly as a substrate for LTC₄ synthase, which adducts glutathione to the C-6 position of LTA₄ to form LTC₄. LTC₄ and its del glutamic acid and del glycine/ glutamic acid congeners, known as LTD₄ and LTE₄, respectively, constitute the biologic activity that had previously been termed slow-reacting substance of anaphylaxis, and these compounds are collectively referred to as the cysteinyl LTs (1, 3, 4). Since their identification in 1979 and 1980, the cysteinyl LTs have been demonstrated to be among the most potent bronchoconstrictor substances ever identified (5).

Several lines of evidence suggest that alterations in the transcriptional regulation of 5-LO are important for functional expression and may have clinical relevance. Leukotriene synthesis occurs in leukocytes after cell differentiation, and increases above basal levels following inflammatory or immunologic stimulation (8). In response to differentiation and activation by phorbol 12-myristate 13-acetate, dimethyl sulfoxide, okadaic acid, transforming growth factor-, interleukin-3, oxidized low-density lipoprotein, or Ca⁺⁺ ionophore, several leukocyte cell lines increase steady-state 5-LO messenger RNA (mRNA) (9). In some studies, at least part of this increase is due to augmented transcription as demonstrated by nuclear runoff analysis (12). The mechanisms by which an inflammatory microenvironment augments transcription have not been established.

The human 5-LO gene has been cloned and shares considerable homology with other members of the LO enzyme family. It is located on chromosome 10; contains 14 exons, 13 introns, and a single transcriptional start site; and spans > 82 kb in length. The 5' upstream region possesses promoter activity and is notable for the absence of TATA or CCAAT sequences (13, 14). The 5-LO gene promoter contains numerous consensus binding sites for many known transcription factors, including zinc-finger transcription factors Sp1, Sp3, early growth response protein (Egr)-1, Egr-2, glucocorticoid receptors, NF-B, GATA, Myb, and AP family members; however, the functional consequences of transcription factor interactions with these binding sites have not yet been described. Hoshiko and colleagues have identified a G+C-rich sequence between 212 and 88 base pairs (bp) relative to the translational start site as necessary for basal promoter-reporter construct activity in HeLa and HL-60 cells (15). This sequence is unique among sequences reported in the Genbank database in that it contains five Sp1/Egr-1 binding motifs in tandem. Recently, a potential mechanism of 5-LO regulation at the level of gene structure within this region of the promoter has been described. In particular, a series of naturally occurring mutations in which there are deletions of one or two, or addition of one, Sp1/Egr-1 binding motifs (i.e., GGGCGG) have been found within the G+C-rich promoter sequence between 176 and 147 bp (16). Although it has been established that oligonucleotides corresponding to the G+C-rich region of the 5-LO promoter sequence can bind both Sp1 and Egr-1, and that binding events are altered by these mutations, it is unknown whether differential regulation of 5-LO transcription by Sp1 or Egr-1 occurs.

Using a minimally active drosophila SL2 cell line we now show that Sp1 or Egr-1 overexpression increases 5-LO promoter-reporter construct activity. Analysis by 5'-deletion localizes part of this trans-activation potential to the G+C-rich sequence where the naturally occurring promoter mutations reside. Moreover, the ability of Sp1 and Egr-1 to activate the 5-LO promoter-reporter constructs is directly proportional to the number of Sp1/Egr-1 consensus binding sites within the G+C-rich sequence. Because it is now well established that interactions of Sp1, Egr-1, and other transcription factors may regulate inducible gene expression, it is possible that basal and cytokine-induced 5-LO gene transcription are partially mediated by an interplay of Sp1, Egr-1, and other as yet unidentified transcription factors within the G+C-rich promoter sequence. It is also possible that the naturally occurring promoter mutations found in the G+C-rich region alter inducible 5-LO gene transcription proportional to the number of Sp1/Egr-1 consensus binding sites through an Egr-1- or Sp1-mediated mechanism.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Recombinant Proteins

Recombinant Egr-1 (zinc-finger region) was generously provided by Dr. F. J. Rauscher, III (Wistar Institute, Philadelphia, PA). Human recombinant Sp1 was purchased from Promega. Samples were stored at 80°C in dilutions containing 25 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (Hepes)-KOH (pH 7.5), 100 mM KCl, 10 µM ZnSO₄, 0.1% Nonidet P-40, 1 mM dithiothreitol (DTT), and 50% glycerol.

In Vitro Deoxyribonuclease I Footprinting

The 5-LOp chloramphenicol acetyltransferase basic wild-type (WT)(5) (5LOpCATbasic-WT[5]) construct (discussed below) was cut with HindIII and end-labeled with T4 polynucleotide kinase. An asymmetrically labeled 5-LO promoter oligonucleotide was cut from the pCAT ^®-basic backbone by XbaI and gel-purified. Binding reactions consisted of binding buffer (20 mM Hepes [pH 7.9], 2 mM MgCl₂, 50 mM NaCl, 1 mM DTT, and 20% glycerol), 20 ng of oligonucleotide (50,000 cpm), and increasing amounts of Sp1 or Egr-1 (0.1-10 µg) in a total volume of 100 µl. After incubation on ice for 30 min, 1 µl deoxyribonuclease (DNase) I solution (20 ng/µl) (Promega, Madison, WI) was added to each sample and incubated at 20°C for 1 min. The digestion was terminated by adding 100 µl stop buffer (50 mM Tris [pH 8.0], 2% sodium dodecyl sulfate, 10 mM ethylenediamenetetraacetic acid, 0.4 mg/ml proteinase K, and 100 µg/ml glycogen) and incubating at 37°C for 30 min. After phenol:chloroform extraction and ethanol precipitation, the sample was run on a standard 10% sequencing gel. "G ladders" were generated by methylation of the probe with dimethyl sulfate and piperidine cleavage and run adjacent to footprint reactions.

Construction of Reporter Plasmids

5LOpCATbasic reporter plasmids were constructed by placing approximately 280 bp of the G+C-rich WT(5) or naturally occurring mutant (M) promoter regions (regions 3, 4, and 6) upstream of the CAT gene in pCAT^®-basic vector (Promega). The 5-LO WT(5) promoter contains five tandem, overlapping Sp1/Egr-1 consensus sites. The mutant promoter designated 5-LO M(6) contains an additional Sp1/Egr-1 binding site, whereas the mutant promoters designated 5-LO M(4) and 5-LO M(3) have lost one or two Sp1/Egr-1 binding sites; these promoters therefore contain six, four, and three Sp1/Egr-1 sites, respectively. Promoter regions were enzymatically cut from the p5LO-CAT constructs described in detail by In and associates (16) using HindIII and XbaI, and ligated into the pCAT^®-basic polylinker using standard methods. This removed the enhancer element, resulting in lower basal CAT levels that were more responsive in cotransfection studies. Correct orientation was confirmed by restriction digest, utilizing sites within the polylinker and the insert, to release the appropriate size fragments. All plasmids were sequenced using a standard dideoxy method (17). G. Suske (Institut fur Molekular Biologie und Tumorforschung, Philipps-Universitat Marburg, Marburg, Germany) provided pPACUSp1, and pPACUEgr-1 was constructed by placing the Egr-1 complementary DNA from cytomegalovirus-Egr-1 (provided by V. P. Sukhatme, Beth Israel Hospital, Boston, MA) into the pPACU vector in frame. Plasmids 5LO5900CAT, 5LO229CAT, 5LO129CAT, and 5LO9CAT were constructed as described in detail by Hoshiko and coworkers (15).

Cell Culture

Drosophila SL2 cells (Schneider cells) were grown in Schneider's Drosophila Medium (Life Technologies, Gaithersburg, MD) containing 50 µg/ml streptomycin, 50 IU/ml penicillin, and 10% fetal calf serum at 27°C and passaged at confluence approximately every 4 d.

Transient Transfection/CAT Assays

Drosophila SL2 cells were transfected at 30% confluence with 10 µg of cesium chloride-purified plasmid using a standard calcium phosphate precipitation protocol (18). The cells were generally cotransfected with 2 µg of pTKGH (Nichols Institute Diagnostics, San Juan Capistrano, CA) to correct for transfection efficiency. After the transfection, the cells were incubated for 48 h at 27°C. At the time of harvest, the conditioned medium was sampled and assayed for human growth hormone by radioimmunoassay (Nichols). CAT assays were performed using a standard scintillation diffusion method (18).

Statistics

Computations were performed using the JMP 3.1.5 (SAS Institute, Cary, NC) statistical package. A transformed log-to-log linear regression model was used to establish a significant correlation between Sp1 or Egr-1 overexpression and 5LOpCATbasic-WT(5) activity. A linear regression model was used to establish a significant correlation between the number of Sp1/Egr-1 consensus sites and reporter activity. A Tukey-Kramer honestly significant difference multiple-comparison test was used to assess differences between plasmid groups. Results are expressed as means ± standard deviation and are considered statistically significant at the P < 0.05 level.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Recombinant Egr-1 and Sp1 Bind the G+C-rich 5-LO Promoter Region by In Vitro DNase I Footprinting

Previous studies identified a series of mutations in the G+C-rich core promoter of the 5-LO gene (16). To determine whether Sp1 and Egr-1 bind specifically to this region of the promoter, DNase I footprinting studies were carried out with recombinant proteins. The binding activity of each transcription factor was determined by titration experiments. Increasing amounts of Sp1 (Figure 1A) or Egr-1 (Figure 1B) protect the region of the promoter containing the overlapping Sp1/Egr-1 binding sites from DNase I digestion. Increasing binding of the transcription factors is associated with the formation of hypersensitive sites at the boundaries of the protected regions. No significant binding cooperativity was observed with either of the proteins. There were other regions of the promoter protected by Sp1, both upstream and downstream of the G+C-rich sequence, that could be explained by known Sp1 consensus sites at 224 to 218 bp (not shown) and 117 to 109 bp. No other obvious Egr-1 binding sites were detected in these experiments.

View larger version (83K): [in this window] [in a new window]   View larger version (86K): [in this window] [in a new window]   Figure 1.   In vitro DNase I footprinting using a 280-bp WT 5-LO promoter fragment spanning the G+C-rich region. Increasing concentrations of recombinant Sp1 (A) or Egr-1 (B) protect the five tandem Sp1/Egr-1 binding sites from digestion, resulting in a footprint. Additional Sp1 footprints were noted corresponding to Sp1 consensus binding sites at 224 to 218 bp (not shown) and 117 to 109 bp. Hypersensitive sites are indicated by an asterisk.

Egr-1 and Sp1 Activate the 5-LO Promoter

Transient transfection analysis using 5-LO promoter- reporter constructs was used to determine the effect of increased Egr-1 or Sp1 on 5-LO promoter-reporter activity. We chose drosophila SL2 cells for all functional studies because, unlike other eukaryotic cell lines, they do not express Sp1 or Egr-1 but contain the transcriptional components capable of interacting with these factors and initiating transcription (19). When transfected with 5-LO promoter- reporter constructs, basal levels of activity were very low in drosophila SL2 cells even though transfection efficiencies were high (data not shown). Overexpressing Sp1 or Egr-1 by adding their respective actin promoter-based expression constructs, pPACSp1 (Figure 2A) or pPACEgr-1 (Figure 2B), resulted in a dose-response activation of the 5LOpCATbasic-WT(5) reporter construct. These experiments suggest that both Sp1 and Egr-1 activate the 5-LO promoter.

View larger version (15K): [in this window] [in a new window]   View larger version (14K): [in this window] [in a new window]   Figure 2.   Cotransfection study in drosophila SL2 cells demonstrating Sp1 and Egr-1 activation of the WT 5-LO promoter. Cells were cotransfected with a reporter construct (10 µg) and increasing amounts of an expression construct using a calcium phosphate precipitation protocol, and harvested 24 h later for CAT assays. Increasing pPACSp1 (A) or pPACEgr-1 (B) augments 5LOpCATbasic-WT(5) transcription in a dose-dependent manner, suggesting that Sp1 and Egr-1 are transcriptional activators of the 5-LO promoter. Relative CAT activity and standard deviations are shown as a function of transfected expression construct (0, 25, 100, and 1,000 ng). A transformed log-to-log linear regression model was used to determine significance (P = 0.02 for Sp1 overexpression, and P = 0.03 for Egr-1 overexpression).

The G+C-rich 5-LO Promoter Sequence Is Essential for Reporter Activation

To identify the region(s) of the 5-LO promoter responsible for mediating Egr-1 and Sp1 trans-activation, we performed cotransfection studies using reporter constructs containing 5'-deletions of the 5-LO promoter placed upstream of the CAT gene in pUCAT. Constructs 5LO5900CAT, 5LO229CAT, 5LO129CAT and 5LO9CAT contain 5,900, 229, 129, and 9 bp upstream of the translation initiation site, respectively. The G+C-rich sequence containing the naturally occurring mutations is located between bp 176 and 147 and is contained within construct 5LO229CAT and the larger construct, but not within 5LO129CAT and the smaller construct. Overexpression of Egr-1, or both Egr-1 and Sp1, resulted in a large increase in 5LO229CAT activity relative to 5LO229CAT transfected alone (Figures 3A and 3B, respectively). The same response occurred with the other larger 5-LO reporter construct containing the G+C-rich sequence. In contrast, construct 5LO129CAT and the smaller construct were minimally responsive or inhibited by cotransfection. Fold induction was dramatically increased when Egr-1 and Sp1 were overexpressed together, compared with Egr-1 alone. These studies suggest that the G+C-rich sequence is at least partially responsible for 5-LO promoter activation by Egr-1, and Egr-1 and Sp1. When Sp1 was overexpressed without Egr-1 the results were significantly different. The Sp1 consensus binding site between 117 to 109 bp, contained within construct 5LO129CAT and absent from construct 5LO9CAT, appeared most important for induction. The additional Sp1 sites within the G+C-rich sequence did not increase fold induction.

View larger version (17K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Figure 3.   5'-deletion analysis demonstrates that the G+C-rich sequence is involved in Egr-1-, and Egr-1 and Sp1-mediated trans-activation. Cells were transfected with 10 µg of reporter construct alone or 10 µg of reporter construct and 100 ng of expression constructs. Fold CAT activity was calculated by dividing expression construct-induced activity with basal reporter activity. Constructs 5LO5900 and 5LO229CAT contain the G+C-rich promoter sequence (hatched area) and showed a much greater induction when Egr-1 (A), and Egr-1 and Sp1 (B) were added, whereas constructs 5LO129CAT and 5LO9CAT, missing the G+C-rich sequence, are minimally inducible or inhibited. The combination of Egr-1 and Sp1 resulted in much greater fold induction than Egr-1 alone. In contrast, studies involving the overexpression of Sp1 without Egr-1 suggest that the 117 to 109 Sp1 consensus binding site in construct 5LO129CAT is most important for Sp1-mediated induction (C ). Fold CAT activities, standard deviations, and statistical significance are shown.

Trans-activation of the WT and Naturally Occurring 5-LO Promoter Mutations Is Proportional to the Number of Sp1/Egr-1 Consensus Binding Sites

Cotransfection studies using reporter constructs containing the WT(5) or one of the naturally occurring promoter mutations (M[3], M[4], or M[6]) were used to compare their responsiveness to Sp1 (Figure 4A) or Egr-1 (Figure 4B). Drosophila SL2 cells were transfected with 10 µg of each reporter plasmid and 500 ng of pPACSp1 or pPACEgr-1 expression construct. For either transcription factor, relative CAT activity was directly proportional to the number of Sp1/Egr-1 consensus binding sites. Construct 5-LO M(6) was most responsive, whereas constructs 5-LO WT(5), 5-LO M(4), and 5-LO M(3) showed progressively less inducibility. The difference was more dramatic with Egr-1 overexpression, which resulted in an approximate 50% fold induction per binding site, compared with a 20% fold induction per site for Sp1.

View larger version (27K): [in this window] [in a new window]   View larger version (24K): [in this window] [in a new window]   Figure 4.   Comparison of the WT and M 5-LO promoters for Sp1 (A) and Egr-1 (B) inducibility. A linear regression model was used to establish significant correlations. CAT activity was directly proportional to the number of Sp1/Egr-1 consensus binding sites. Construct 5LOpCATbasic-M(6) was most responsive, whereas constructs -WT(5), -M(4) and -M(3) showed progressively less inducibility (P < 0.001 for Sp1 overexpression, and P < 0.001 for Egr-1 overexpression). Relative CAT activities and standard deviations are shown. Egr-1 overexpression resulted in greater differences compared with Sp1 (approximately 50% increase per Sp1/ Egr-1 site versus 20% increase).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

5-LO is the first committed enzyme in the metabolic pathway leading to the intracellular synthesis of the LTs and as such, is essential for the formation of all LTs. For reasons that are not understood, the expression of 5-LO is not ubiquitous but limited to cells of primarily myeloid lineage. 5-LO expression and LT synthesis occurs in leukocytes after cell differentiation, and increases above basal levels following inflammatory or immunologic stimulation (8). The mechanisms responsible for the regulation of 5-LO functional expression during these events are manifold and poorly understood. Multiple levels of control appear to be involved and include alterations in gene transcription, RNA translation, enzyme translocation, enzyme inactivation, and substrate availability; the relative importance of each level is unknown.

We have previously identified a family of naturally occurring mutations within the G+C-rich core-promoter region of the 5-LO promoter (16). In the family of mutants, the G+C-rich sequence contains three to six overlapping Sp1 and Egr-1 consensus binding sites, with the number depending on the specific gene allele present. In the study by In and coworkers (16), it was shown by electrophoretic mobility shift assays (EMSA) that nuclear Egr-1 and Sp1 bind oligonucleotides consisting of these G+C-rich promoter sequences, and that binding intensities were proportional to the number of Sp1/Egr-1 consensus binding sites, with 5-LO M(6) producing the most intense bands and 5-LO WT(5), 5-LO M(4), and 5-LO M(3) producing progressively less intense bands. However, the effects of these specific binding events on gene transcription were not determined.

On the basis of these data and consistent with In and associates' initial observations (16), we used in vitro DNase I footprinting in our present study to demonstrate that Egr-1 and Sp1 bind specifically to the WT G+C-rich sequence (179 to 147 bp) in the context of a much larger region of authentic 5-LO promoter. In the case of Sp1, other binding sites were identified just upstream (224 to 218 bp) and downstream (117 to 109 bp) of the G+C-rich sequence. In addition, we have shown that Egr-1 and Sp1 can activate the 5-LO promoter in a dose-response fashion when expression constructs are cotransfected with reporter constructs in a minimally active drosophila SL2 cell culture system. Transient transfection analysis using 5-LO promoter- reporter constructs indicates that the G+C-rich sequence containing the naturally occurring mutations was at least partially responsible for Egr-1-, and Egr-1 and Sp1-mediated trans-activation. However, when Sp1 was expressed without Egr-1 the 117 to 109 region appeared to be most critical for induction. Finally, when the naturally occurring mutations were directly compared, our data indicate that trans-activation was proportional to the number of Sp1/Egr-1 consensus binding sites. The greater the number of tandem Sp1/Egr-1 consensus binding sites, the more CAT activity was detected after transfection with any given amount of Sp1 or Egr-1 expression construct. These data suggest that the correlation between EMSA binding intensity and promoter genotype has direct transcriptional activating consequences.

It is important to note that our promoter-reporter constructs did not function this way in all cell systems tested. For example, in the study by In and coworkers, promoter M(6) was less active than the WT(5) promoter in HeLa cells (16). This contrasts with our findings in a minimally active Schneider cell system where M(6) was more active than WT(5). It is likely that the complement of trans-activators between Schneider cells and HeLa cells varies significantly and may account for the differences in reporter activity. We speculate that 5-LO induction during leukocyte differentiation and activation is transcriptionally mediated by increased Sp1, Egr-1, or other inducible transcription factors, and the naturally occurring 5-LO promoter mutations alter 5-LO functional expression by changing the number of Eg-1 or Sp1 binding events within the G+C-rich promoter region. Because many other cell types in addition to leukocytes express Sp1 and Egr-1, these interactions are unlikely to be responsible for the cell specificity of 5-LO expression.

Many genes have a high G+C content within their promoter sequences, for unclear reasons. Several genes including 5-LO, platelet-derived growth factor A-chain, tissue factor, and urinary plasminogen activator have similar overlapping Sp1 and Egr-1 consensus binding sites; however, 5-LO is the only gene in the Genbank database with five tandem sites. This region is also unique among promoter sequences in that mutations have been identified in which the number of these Sp1/Egr-1 binding sites varies among the mutants from three to six (16). It is unknown whether similar mutations exist within the G+C-rich promoter regions of other genes. Sp1 is a ubiquitous zinc-finger protein expressed in nearly all cell types and is required for the expression of many essential genes. Sp1 binds to G+C-rich sites containing the consensus sequence GGGCGG. In general, the highest levels of Sp1 expression are found in cells undergoing differentiation, and these high levels of Sp1 may be required for the subsequent induction of tissue-specific genes (20). Indeed, it is interesting to speculate that increasing Sp1 may be partially responsible for initiating 5-LO transcription during leukocyte differentiation. Why multiple Sp1 sites are present in the promoter regions of so many genes is unclear. It has been suggested that multiple Sp1 sites may interact synergistically to augment gene transcription (20).

Egr-1 and related family members Egr-2 and Egr-3 are also zinc-finger transcription factors that bind to similar G+C-rich sequences containing the consensus sequence GCG(T/G)GGGCG (21). Egr family members are examples of "immediate-early response" proteins, and are rapidly and transiently induced by a large number of growth factors, cytokines, and injurious stimuli. Egr-1 can displace Sp1 and other transcription factors from promoter regions of several genes and increase transcription above basal levels (19, 22). Egr-1 and Egr-2 levels increase during leukocyte differentiation and after cytokine activation at a time when levels of 5-LO mRNA increase, suggesting that Egr-1 may play a role in 5-LO transcription regulation (23). Egr-1 may stimulate transcription by interacting directly with components of the basal transcription apparatus or by recruitment of a coactivator. Tandem Egr-1 sites may permit multiple simultaneous contacts between the activators and the basal transcription factors, thereby increasing the level of transcription (26). Alternatively, multiple Egr-1 elements may increase the efficiency of coactivator recruitment by increasing the number of protein- protein contacts between the activator and the coactivator. The naturally occurring mutants of the 5-LO promoter could alter the activator's ability to increase transcription by affecting either of these mechanisms.