Critical Flanking Sequences of PU.1 Binding Sites in Myeloid-specific Promoters*

The myeloid-specific transcription factor PU.1 is essential for expression of p47 phox , a component of the superoxide-forming phagocyte NADPH oxidase. The consensus PU.1 binding sequence (GAGGAA) is located on the non-coding strand from position −40 to −45 relative to the transcriptional start site of thep47phox promoter. A promoter construct extending to −46 was sufficient to drive tissue-specific expression of the luciferase reporter gene, but extension of the promoter from −46 to −48 resulted in a significant increase in reporter expression. Mutations of the nucleotides G at −46 and/or T at −47 reduced both reporter expression and PU.1 binding, whereas mutations at −48 had no effect. The PU.1 binding avidity of these sequences correlated closely with their capacity to dictate reporter gene transcription. In parallel studies on the functional PU.1 site in the promoter of CD18, mutations of nucleotides G and T at positions −76 and −77 (corresponding to −46 and −47, respectively, of the p47phox promoter) reduced PU.1 binding and nearly abolished the contribution of this element to promoter activity. We conclude that the immediate flanking nucleotides of the PU.1 consensus motif have significant effects on PU.1 binding avidity and activity and that this region is the dominantcis element regulating p47phox expression.

During myelopoiesis pluripotent hematopoietic stem cells become committed as myeloid precursor cells and differentiate into morphologically and functionally distinct end-stage neutrophils, eosinophils, and monocytes. Myeloid development is regulated by a combination of hematopoietic growth factors, growth factor receptors, and lineage-restricted transcription factors (1). The myeloid-specific transcription factor PU.1, the most divergent member of the ets family of transcription factors (2,3), is required for terminal myeloid differentiation and gene expression (4). PU.1 contacts DNA with a novel loop-helix-loop architecture, and binds to a cis element with a core sequence 5Ј-GAGGAA-3Ј (5,6). This element is sometimes clustered with binding sites for other transcription factors, which may facilitate interactions among these different trans-acting factors. Examples include the PIP 1 site within the immunoglobulin light chain gene enhancers E 3Ј , E␥ [2][3][4] , and E␥ 3-1 and the FEF site in the c-fes promoter and chicken lysozyme enhancer (7,8). Whereas PU.1 binding is dispensable for FEF-DNA interaction, it is a prerequisite for PIP participation in the PU.1-PIP-DNA complex (7,8).
PU.1 binds a number of myeloid cell-restricted target genes, such as p47phox (9), gp91phox (10), macrophage colony-stimulating factor receptor (M-CSFR) (11), granulocyte colony-stimulating factor receptor (G-CSFR) (12), F c -receptor (13), scavenger receptor (14), and integrin subunits CD11b (15) and CD18 (16). CD18 is the ␤ 2 subunit of the leukocyte integrins, a family of cell surface proteins that mediate leukocyte adhesion and thereby play a critical role in the inflammatory response (17,18). The importance of the leukocyte integrins in inflammation is illustrated by leukocyte adhesion deficiency syndrome, a rare inherited disorder caused by mutations in the CD18 gene and associated with severe and recurrent infections (17,18). The products of the p47phox and gp91phox genes are components of the phagocyte NADPH oxidase, which by generating superoxide anion serves as a pivotal enzyme in the microbicidal function of phagocytic cells (19,20). The importance of the phagocyte oxidase is demonstrated by chronic granulomatous disease (CGD), an inherited disorder in which a functionally defective NADPH oxidase results in severe and recurrent infections (19 -22).
We previously isolated the p47phox promoter and showed that it contains a consensus PU.1 binding sequence on the non-coding strand from base pair Ϫ40 to Ϫ45 relative to the transcriptional start site (9). Mutation of this sequence abolishes PU.1 binding and promoter activity. Although a p47phox promoter-luciferase reporter construct extending to Ϫ46 could dictate tissue-specific expression in myeloid cells, a larger construct (Ϫ86) showed maximal promoter activity, implying a significant contribution of upstream sequences to the expression of p47phox. Since in DNase I footprint analysis, a protected region (Ϫ37 to Ϫ52) was observed to extend beyond the consensus sequence GAGGAA (Ϫ40 to Ϫ45), the current studies were performed to define the functional role of the flanking residues upstream of the PU.1 site.
Luciferase Vectors-Reporter vectors in the pGL3-Basic luciferase plasmid including pGL3-p47-86, pGL3-p47-48, pGL3-p47-46, and pGL3-p47-36 (i.e. the proximal p47phox promoter extending to positions Ϫ86, Ϫ48, Ϫ46, and Ϫ36, respectively, relative to the TSS) were constructed as described previously (9). Briefly, PCR was carried out using the plasmid construct pGL3-p47-1217 as template, a luciferase antisense oligonucleotide (pGL primer 2; 5Ј-CTTTATGTTTTTG-GCGTCTTCC-3Ј) as the reverse primer and oligonucleotides synthesized with an XhoI restriction site linked to the desired 5Ј terminus of the p47phox promoter as the forward primers. For CD18 analysis, CD18(0.9)/luc (the generous gift of Dr. A. Rosmarin, Brown University, Providence, RI) (16) was used as template and the PCR primers were an antisense sequence (5Ј-CCGAAGCTTTTGCTACCAGTCTGCCC-3Ј) and forward primers synthesized with an XhoI restriction site as for p47phox. The PCR-amplified products were digested with XhoI and HindIII and cloned into the corresponding sites of pGL3-Basic. To generate mutated constructs, altered oligonucleotides were used as the forward primers. The inserts of the p47phox constructs all extended downstream to ϩ52 relative to the TSS and used the p47phox translation initiation codon ligated in-frame to the luciferase open reading frame. The CD18 promoter constructs extended downstream to ϩ27 and used the translation initiation codon of the luciferase open reading frame.
Transient Transfections-THP-1 cells were maintained at a density of ϳ5 ϫ 10 5 cells/ml and for transfection were then resuspended in medium containing 20 g of the luciferase reporter constructs and 2 g of the pRL-CMV plasmid (Promega) as a transfection efficiency control. Electroporation was carried out at 960 F and 250 V. After 48 h the cells were washed three times in phosphate-buffered saline, pH 7.4, lysed in 100 l of reporter lysis buffer, microcentrifuged for 5 min, and 20-l aliquots of the supernatant assayed for both luciferase and Renilla activity using the dual-luciferase assay system (Promega) and a Turner TD-20/20 luminometer.
In Vitro Translation-The mouse PU.1 cDNA (generous gift of Dr. M. Klemsz, Indiana University, Indianapolis, IN) (2) was excised by digestion with EcoRI and ligated into the pBluescript plasmid. A clone with the desired orientation was transcribed and translated in vitro using T3 RNA polymerase and the TnT-coupled reticulocyte lysate system (Promega). When the synthesized [ 35 S]methionine-PU.1 was analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography, a predominant band of approximately 38 kDa was observed, consistent with the molecular mass previously reported (23). The control unprogrammed sample (i.e. no cDNA) gave no corresponding band.
Nuclear Extracts-THP-1 cells were disrupted by cavitation using a technique described previously for neutrophils (24). Briefly, the cells were washed twice in phosphate-buffered saline, pH 7.4, resuspended in 10 ml of cold relaxation buffer (100 mM KCl, 3 mM NaCl, 3.5 mM MgCl 2 , 10 mM PIPES, pH 7.3), and 3.5 l diisopropyl fluorophosphate (Sigma) were added. The cells were kept on ice for 10 min, then centrifuged at 400 ϫ g for 5 min. The cell pellet was resuspended in 10 ml of relaxation buffer and pressurized at 350 p.s.i. in N 2 for 20 min in a nitrogen bomb (Parr Instrument Co., Moline, IL) before release into 750 l of a solution containing 20 mM EGTA, 100 mM MgCl 2 , 20 mM dithiothreitol, 4 mM phenylmethylsulfonyl fluoride, and 2 mM sodium orthovanadate. The cavitated cells were centrifuged at 400 ϫ g for 10 min at 4°C and the nucleus-enriched pellet resuspended and further purified on a discontinuous gradient of sucrose (0.3 M/0.88 M). The nuclear fraction was extracted in 100 l of urea extraction buffer (1.1 mM urea, 1% Nonidet P-40, 5% glycerol, 0.5 mM MgCl 2 , 5 mM KCl, 0.05 mM EDTA, 5 mM HEPES, pH 7.9) and microcentrifuged. The supernatant was collected and stored in aliquots at Ϫ70°C. The protein concentration was determined using the Bradford reagent (Bio-Rad).
Electrophoretic Mobility Shift Assay (EMSA)-EMSA was carried out as described previously (9). Briefly, complementary DNA oligonucleotides were annealed by heating in 1ϫ NET at 95°C for 5 min and cooling at ambient temperature. Probes were labeled with [␥-32 P]ATP and T4 polynucleotide kinase. For gel shift assays nuclear extracts (6 g) were incubated for 20 min at ambient temperature with 5 ϫ 10 4 cpm of the labeled DNA probe in 20 l of binding buffer containing 10 mM Tris-HCl, pH 7.6, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 5% glycerol, 0.7 g/l bovine serum albumin, 2 g of poly(dI-dC). For supershift assays 2 g of specific antibody was added and the reaction continued for 15 min. Samples were loaded on 6% nondenaturing polyacrylamide gels and electrophoresis carried out at 200 V in 25 mM Tris, pH 8.5, with 190 mM glycine and 1 mM EDTA. Competition assays were carried out in the same manner, except that the reaction mixtures were preincubated with competitor DNA for 10 min at 4°C before addition of the labeled probe. The relative binding avidities of various DNA probes to PU.1 were determined by comparisons of band intensities in the presence of a series of dilutions of competing constructs.

RESULTS
Functional Role of the Upstream Sequences Flanking the p47phox PU.1 Binding Site-Our previous work on the p47phox 5Ј regulatory region showed that PU.1 is essential for promoter function (9). Using p47phox-based luciferase reporter gene constructs, we observed that the Ϫ46 construct containing the PU.1 binding site produced myeloid-specific expression, but that the Ϫ86 construct gave maximum promoter activity. Analysis of the p47phox Ϫ86 to Ϫ46 promoter region identified consensus sequences for Sp1 between positions Ϫ78 and Ϫ70 and for PEBP2 between positions Ϫ63 and Ϫ58 (9). However, gel shift experiments showed no evidence that either Sp1 or PEBP2 binds to these sites (data not shown). To investigate the role of the sequences immediately flanking the PU.1 site, we made a new construct pGL3-p47-48. In transient transfection assays of THP-1 cells (Fig. 1), pGL3-p47-48 exhibited twice the promoter activity of pGL3-p47-46, indicating the functional importance of the flanking nucleotides at positions Ϫ47 and/or Ϫ48 (T and C, respectively).
Effect of Flanking Sequences on PU.1-DNA Interaction in p47phox-We next used EMSA to investigate the influence of the flanking nucleotides on PU.1 binding. To facilitate the direct correlation of DNA-protein binding with promoter activity, we designed DNA probes containing the PU.1 consensus motif flanked by the same sequences as those present in the corresponding reporter constructs used in transfections, including some vector sequences as required (Fig. 1A, under- lined). We observed several bands formed between the p47-PU.1-48 probe and THP-1 nuclear extracts ( Fig. 2A, lane 2) and demonstrated binding specificity by competition with unlabeled DNA probe ( Fig. 2A, lane 4). These bands contained PU.1 protein since the addition of PU.1-specific antibody resulted in major decreases in band intensity as well as the appearance of new supershifted bands ( Fig. 2A, lane 3). Comparison with our previous studies (9) indicated that the major band contained intact PU.1 protein, whereas the faster-migrating bands were probably formed by PU.1 degradation products.
The slower-migrating bands may contain other proteins in addition to PU.1. However, this appears unlikely because interaction of the p47-PU.1-48 DNA probe with in vitro synthesized PU.1 protein showed similar slow-migrating bands (Fig.  2D, lane 4), raising the possibility that these bands may contain PU.1 multimers. As will be discussed below, complex formation between the p47 DNA probe and in vitro synthesized PU.1 was readily competed by the unlabeled p47 probe, whereas only weak competition was observed with the CD18-PU.1 probe (Fig. 2D). The p47-PU.1-46 DNA probe exhibited the same general binding patterns, as did p47-PU.1-48, including the slow-migrating bands (Fig. 2C, lane 2). However, based on cross-competition studies, PU.1 bound the p47-PU.1-48 probe twice as avidly as it did the p47-PU.1-46 probe (Fig. 2, B and C). These observations suggest that the nucleotides at positions Ϫ47 and/or Ϫ48 significantly influence PU.1 binding ability and thereby affect gene expression.
Effects of Mutations at p47phox Position Ϫ48 -To test the influence of the nucleotide at position Ϫ48 on PU.1 transactivating activity, we mutated the wild-type nucleotide C to G, T, or A in the pGL3-p47-48 luciferase reporter construct (Fig.  3A). Transient transfection of THP-1 cells demonstrated comparable promoter activities of wild-type and each of the three Ϫ48-mutated constructs (Fig. 3B). Correspondingly altered oligonucleotides were studied by EMSA (Fig. 3C). These base substitutions had no effect on PU.1 binding, since each of the three mutated DNA probes was comparable to wild-type in its ability to compete with the labeled wild-type probe. Thus, the nucleotide at position Ϫ48 does not appear to influence either promoter activity or binding of PU.1.
Effects of Mutations at p47phox Position Ϫ47-Similarly, we mutated the wild-type nucleotide T at position Ϫ47 to A, C, or G (refer to Fig. 3A) and tested the effects on promoter activity and PU.1 binding. The mutant reporter constructs pGL3-p47-48-47A, pGL3-p47-48-47C, and pGL3-p47-48-47G exhibited only minimal promoter activity compared with the wild-type construct containing a T at position Ϫ47 (Fig. 4A). When one of these mutations was introduced into the larger construct pGL3-p47-86 to form pGL3-p47-86-47A, we again observed a major decrease in promoter activity (Fig. 4A). When assessed by EMSA, the oligonucleotides mutated at position Ϫ47 all failed to compete with the wild-type probe for PU.1 binding (Fig. 4B). Cross-competition studies suggested that the avidity of binding of the wild-type probe was about 3-9-fold greater than that of the mutants.
Effects of Mutations at p47phox Position Ϫ46 -At position Ϫ46, immediately adjacent to the core PU.1 consensus binding sequence, the wild-type G residue was mutated to each of the other three nucleotides (refer to Fig. 3A). In transfection studies (Fig. 5A), promoter activities of the 46C and 46A mutants were decreased by 50% and 85%, respectively, whereas those of the 46T and combined 46T47A mutants were completely eliminated. EMSA showed corresponding changes (Fig. 5B). Compared with the wild-type 46G probe, the 46C mutant showed moderately reduced binding, whereas the 46A, 46T, and 46T47A mutants each exhibited dramatically reduced binding. The cross-competition studies suggested that the avidity of binding of the wild-type probe was about 2-3-fold greater than that of the 46C mutant and more than 9-fold greater than that of the other three mutants tested.

Effects of Mutations at Sites in the CD18 Promoter Analogous to Positions
Ϫ46 and Ϫ47 of p47phox-We hypothesized that our findings on the important functional roles of the nucleotides flanking the p47phox PU.1 consensus binding site could be extrapolated to other PU.1-regulated myeloid-specific genes. To test this hypothesis, we chose the CD18 promoter as a model. This promoter contains two PU.1 binding motifs: a distal site at position Ϫ70 to Ϫ75 and a proximal site at position Ϫ55 to Ϫ50. We focused on the distal site, which has been shown to have functional activity that was dramatically decreased following mutation of the core PU.1 binding sequence (16). Moreover, the flanking residues at positions Ϫ76 and Ϫ77 are G and T, respectively, exactly as for the comparable positions (Ϫ46 and Ϫ47) in the p47phox PU.1 flanking region.
A luciferase reporter construct, pGL3-CD18 -81, containing the first 81 nucleotides upstream of the transcriptional start site was transiently expressed in THP-1 cells and compared with a construct (pGL3-CD18 -81-76T77A) in which the flanking wild-type residues 76G and 77T were mutated to T and A, respectively (Fig. 6A). The promoter activity of the mutant construct was decreased by 75% relative to the wild-type (Fig.  6B). That this decrease was quantitatively similar to that seen with mutations of the PU.1 core sequence of CD18 (16) suggests that most of the contribution of the distal PU.1 binding site to the promoter activity of this construct was abrogated by mutation of the flanking residues. Next we investigated the distal PU.1 site of the CD18 promoter by EMSA, including comparisons with the p47phox PU.1 site (Fig. 6C). DNA probes CD18-PU.1 and p47-PU.1 formed similar patterns of complexes with THP-1 nuclear extracts, but p47-PU.1 was far more active (by about 20 -30-fold) than CD18-PU.1 in the cross-competition studies. The mutant probe CD18-PU.1-76T77A exhibited little or no ability to bind PU.1 in THP-1 nuclear extracts (Fig. 6D). When in vitro synthesized PU.1 protein was used instead of THP-1 nuclear extract, similar bands were observed (see Fig.  2D), indicating the likelihood that PU.1 is the only nuclear factor included in the complex. This experiment also confirmed the greater ability of the p47-PU.1, compared with the CD18-PU.1 probe to compete for PU.1 binding.
Correlation of Promoter Activity and PU.1 Binding Avidity-Based on functional analyses of wild-type and mutated p47phox reporter constructs and competition EMSA studies, the relative promoter activities and PU.1 binding avidities (estimated by cross-competition studies) were tabulated ( Table  I). The Spearman rank correlation test showed a striking relationship between these parameters (r ϭ Ϫ0.97, p Ͻ 0.0001). DISCUSSION We previously showed that the myeloid-specific transcription factor PU.1 is essential for the function of the p47phox gene promoter. In the present study, we demonstrate that the upstream nucleotides immediately flanking the PU.1 site in the p47phox promoter (3Ј to the GAGGAA sequence, since this is on the non-coding strand) are important for full PU.1 binding and functional activity. Mutations at base pair Ϫ46 or Ϫ47 dramatically reduced the binding avidity and decreased or abolished promoter activity, indicating the critical role of nucleotides G and T, respectively, at these positions. The avidity of binding of these sequences to PU.1 correlated very closely with their capacity to dictate reporter gene transcription. Analogous results were obtained with the functional PU.1 site in the CD18 gene promoter, suggesting a strong relationship between the avidity with which PU.1 binds its cognate sequence on a promoter and the resulting PU.1-mediated enhancer activity.
Based on the presence of potential binding sites, we speculated earlier (9) that the contribution to p47phox promoter activity of the DNA sequence in the Ϫ86 to Ϫ46 interval might result from the binding of other transcription factors such as Sp1 and PEBP/CBP. Alternatively, increased activity might result from a PU.1-dependent effect of the nucleotides immediately flanking the consensus PU.1-binding site, GAGGAA. Our current work provides evidence for the latter mechanism by demonstrating that the nucleotides G and T at positions Ϫ46 and Ϫ47, respectively, are required for maximal PU.1 binding and p47phox promoter activity. However, the contribution of other accessory factors cannot be completely excluded, since the transfection studies in THP-1 cells showed that the promoter activity of the pGL3-p47-48 construct was still only about half that of pGL3-p47-86, which we have shown previously to be sufficient for maximal promoter activity. In our previous studies, however, we also showed that mutation of the active PU.1 site from GAGGAA to CACCAA was equally effective in abolishing the promoter activity in Ϫ86, Ϫ224, and Ϫ2151 pGL3-p47 constructs. Therefore, the function of any additional transcription factors regulating p47phox transcription appears to be dependent on PU.1 binding.
This study shows that for the p47phox promoter, the nucleotides G and T, which flank the 3Ј end of the GAGGAA sequence, are the most active in promoting PU.1 binding and transactivation of the gene. Inspection of a number of functional PU.1 sites from other gene promoters (Table II) showed that these particular nucleotides occur most frequently at the corresponding positions in these other motifs (18G, 7C, 3A, and no T residues at positions analogous to Ϫ46 and 18T, 5G, 3C, and 2A residues at positions analogous to Ϫ47). The functionally least active combination of bases (TA) was not found. These sequence analyses support our claim that the nucleotides mutated and wild-type reporter constructs. Transfection and the determination and expression of luciferase activity were carried out as in   (6). In these studies, the PU.1 ets domain was shown to bind the DNA with novel loop-helix-loop architecture and to form a series of contacts with both the core GGAA sequence and with nucleotides 3Ј to this motif. A number of amino acid residues contact the phosphate backbone of the two nucleotides immediately 3Ј to the GGAA sequence (corresponding to positions Ϫ46 and Ϫ47 of the p47phox PU.1 site). Substitution of some of these amino acids with glycine was sufficient to abolish binding to the DNA (6). Similarly, nucleotides immediately 5Ј to the GAG-GAA sequence were shown to be contacted by amino acid residues of the ets domain and likewise to be important in PU.1-DNA interaction (5). In agreement with these latter data, we have found that specific nucleotides 5Ј to the PU.1 core sequence in the p47phox promoter are also required for full PU.1 binding and transactivation of the gene. 2 Paxton and colleagues (25) demonstrated that specific sequences flanking a site for NF-B are required for cognate transcription factor binding and tumor necrosis factor-␣-mediated induction of intercellular adhesion molecule-1. Mutations of either the 5Ј-or 3Ј-flanking regions abrogated tumor necrosis factor-␣-induced reporter activity, as did mutations of the core NF-B site. A specific DNA-protein complex was formed when wild-type flanking sequences were included in the EMSA probe, but no complex was formed when random flanking sequences were used. Whether there was a quantitative correlation between binding affinity and promoter activity was not directly addressed. To our knowledge, the current paper is the first to provide a systematic quantitative correlation between transcription factor binding avidity and promoter activity of specific nucleotide sequences. As shown in Table I, flanking sequences determine the avidity of PU.1 binding and thereby influence p47phox promoter activity. Nevertheless, the correlation appears to be promoter-dependent. For example, comparing the p47phox promoter and the CD18 promoter, the wild-type PU.1 binding site of the p47phox gene binds to PU.1 30-fold more strongly than does the wild-type CD18 PU.1 site. If the p47phox PU.1 site is mutated to have such a low binding affinity, it completely loses its contribution to promoter activity.
PIP has been reported to bind to the specific flanking nucleotides of PU.1 sites in the presence of bound PU.1 and thereby increase promoter activity (7). On the other hand, FEF-1 binds to the adjacent sequences independently of PU.1, but cooperates with PU.1 to enhance gene transcription (8). Our current findings that the flanking sequences of PU.1 binding sites influence promoter activity and gene expression by altering the binding avidity of PU.1 provide another model for the mechanism by which the flanking sequences affect transcription. We propose that such a mechanism may also operate in other cis element-transcription factor interactions.
The phagocyte NADPH oxidase is an enzyme complex comprised of several protein subunits, principally the membrane components gp91 phox and p22 phox and the cytosolic components p67 phox , p47 phox , and Rac1/2 (19,20). Genetically determined lack of NADPH oxidase activity results from mutations in the genes for gp91 phox or p47 phox or, rarely, the genes for p22 phox or p67 phox , and leads to the clinical disorder CGD (21,22). Most cases result from mutations in the coding sequence of the affected gene. However, variant forms of gp91 phox -deficient CGD have been described in which low levels of the protein are expressed and single base changes are found at positions Ϫ52, Ϫ53, Ϫ55, or Ϫ57 of the active PU.1 site of the gp91phox promoter (10,26,27). The consensus PU.1 motif GAGGAA of the gp91phox gene promoter is located between Ϫ50 to Ϫ55 relative to the transcription start site. Position Ϫ57 corresponds to position Ϫ47 of the p47phox promoter, shown in the current work to be critical for PU.1 binding and promoter activity. In vitro mimicking of the CGD sequences by mutation of the PU.1 core motif of the gp91phox promoter from GAGGAA to GAGGAG or the flanking sequence from GAGGAAAT to GAGGAAAG led to major losses in promoter activity (10).
In the case of p47 phox -deficient CGD patients, a single lesion, a GT deletion at an intron-exon junction is the predominant mutation. The recent identification of a p47phox pseudogene that contains the GT deletion but is otherwise highly homologous to the normal gene suggests that recombination events, for example crossovers with deletions and/or gene conversions, between the p47phox gene and the pseudogene account for most cases of p47 phox deficiency (28). To date, no CGD patients with mutations in the p47phox promoter have been described. In part, this may be due to the rarity of this genetic disorder. Second, it could be a consequence of activation of the p47phox gene by regulatory elements outside of the promoter region that we have studied. Third, coding regions tend to comprise longer stretches of DNA and therefore are more likely to manifest mutations than are promoter elements, which are usually clustered in small regions upstream of coding segments. Fourth, redundancy of promoter elements may allow for at least partial compensation for mutations of core or flanking sequences of a promoter motif, thereby resulting in subtle phenotypes that could be overlooked clinically. Thus, it remains to be seen whether mutations of p47phox PU.1 core or flanking binding residues may account for some clinical forms of CGD, as is the case for the gp91phox gene.