The superoxide anion (O2−)-generating system is an important mechanism of innate immune response against microbial infection in phagocytes and is involved in signal transduction mediated by various physiological and pathological signals in phagocytes and other cells, including B lymphocytes. The O2−-generating system is composed of five specific proteins: p22-phox, gp91-phox, p40-phox, p47-phox, p67-phox, and a small G protein, Rac. Little is known regarding epigenetic regulation of the genes constituting the O2−-generating system. In this study, by analyzing the GCN5 (one of most important histone acetyltransferases)-deficient DT40 cell line, we show that GCN5 deficiency causes loss of the O2−-generating activity. Interestingly, transcription of the gp91-phox gene was drastically downregulated (to ∼4%) in GCN5-deficient cells. To further study the involvement of GCN5 in transcriptional regulation of gp91-phox, we used in vitro differentiation system of U937 cells. When human monoblastic U937 cells were cultured in the presence of IFN-γ, transcription of gp91-phox was remarkably upregulated, and the cells were differentiated to macrophage-like cells that can produce O2−. Chromatin immunoprecipitation assay using the U937 cells during cultivation with IFN-γ revealed not only that association of GCN5 with the gp91-phox gene promoter was significantly accelerated, but also that GCN5 preferentially elevated acetylation levels of H2BK16 and H3K9 surrounding the promoter. These results suggested that GCN5 regulates the O2−-generating system in leukocytes via controlling the gp91-phox gene expression as a supervisor. Our findings obtained in this study should be useful in understanding the molecular mechanisms involved in epigenetic regulation of the O2−-generating system in leukocytes.
Generation of superoxide anion (O2−) is an important function in phagocytes (1) and B lymphocytes (2). On stimulation, the O2−-generating system in leukocytes carries an electron from NADPH to molecular oxygen and generates O2−, which is released either outside the cells or inside phagosomes. Five specific proteins are essential for the O2−-generating system: namely, large (gp91-phox) and small (p22-phox) subunits of cytochrome b558 in membranes (3–6) and cytosolic p40-phox, p47-phox, and p67-phox proteins (7–10). The cytochrome is dormant in resting leukocytes but becomes activated during phagocytosis to generate O2−, a precursor of microbicidal oxidants. Activation of the cytochrome requires stimulus-induced membrane translocation of these cytosolic proteins and small G protein Rac (11, 12). The importance of the system is emphasized by a genetic disorder known as chronic granulomatous disease (CGD), in which phagocytes and B lymphocytes cannot generate O2− upon stimulation (13). Above all, ∼70% of CGD are the result of defective gp91-phox gene (14). The gp91-phox gene is exclusively expressed in phagocytes and B lymphocytes, and its expression is regulated by cell/tissue-type- and/or development stage-specific mechanisms in hematopoietic cells (15). It is known that the gp91-phox gene expression is regulated via activation of its promoter by several transcriptional factors [i.e., CCAAT-displacement protein (16–19), CCAAT-binding protein-1 (16, 18, 20–22), YY1 (23), IFN regulatory factor-1, IFN regulatory factor-2 (18, 20, 21), hematopoietic-associated factor-1 (24–26), PU.1 (25–29), HOXA9 (30), and NF-κB (31)]. However, the involvement of epigenetic regulation including acetylation levels of core histones in the gp91-phox gene expression remains to be resolved.
N-terminal tails of core histones are exposed outside of the chromatin structure. Histone acetyltransferases (HATs) catalyze transfer of the acetyl group from acetyl-CoA to ε-amino groups of conserved Lys residues of core histones. Acetylation helps to neutralize positive charges of core histones, resulting in the reduced affinity of acetylated N-terminal tails of core histones to DNA (32, 33), and the modified marks provide an effective signal that may mediate interactions with various protein factors (i.e., histone modifying enzymes, chromatin remodeling factors, and histone chaperones). HATs are grouped into two types A and B, based on their intracellular localization and substrate specificity (34). GCN5 is a member of the A-type HATs, as well as the GCN5-related N-acetyltransferase superfamily, and shows cell/tissue type-specific expression characteristics. Therefore, it is expected to play distinct and important roles in transcriptional regulations through its intrinsic HAT activity.
GCN5 was first identified as a global coactivator, and transcription-related HAT (34, 35) and GCN5 deficiency in mice led to early embryonic lethality with increased apoptosis in mesodermal lineages (36–38). Recently, it was reported that loss of GCN5 in mouse embryonic stem cells invoked a cell-autonomous pathway of cell death (39). However, its functions remain poorly defined in vertebrate cells.
Using gene-targeting techniques (40), in contrast, we generated homozygous chicken DT40 mutant cell line, GCN5−/−, devoid of two GCN5 alleles (41). The GCN5 deficiency not only caused delayed growth rate and suppressed cell cycle progression at G1/S phase transition, but also caused down- or upregulation of various G1/S phase transition-related genes. Further, GCN5 was shown to induce premature B cell apoptosis by collaboration with BCR signaling (42). In this study, we show that the depletion of GCN5 drastically inhibits the O2−-generating activity in DT40. In addition, to further examine the role of GCN5 in the O2−-generating system, monoblastic leukemia U937 cell line was used as a model. Based on findings obtained in this study, we clarified the participation of GCN5 in epigenetic regulation of the O2−-generating system in leukocytes via controlling gp91-phox gene expression through elevation in acetylation of H2BK16 and H3K9 surrounding its promoter region.
Materials and Methods
PMA (Calbiochem), human IFN-γ (Roche), Diogenes (National Diagnostics), normal rabbit serum (Vector Laboratories), PMSF (Wako), bovine aprotinin, luminol, normal rabbit IgG, and superoxide dismutase (SOD) (Sigma-Aldrich) were obtained. Anti-CREB–binding protein (CBP), anti-GCN5, anti-p300, and anti-p300/CBP-associated factor (PCAF) Abs (Santa Cruz Biotechnology), all anti-acetylated histone Abs (Millipore), monoclonal anti–gp91-phox Ab (BD Bioscience), anti-β-actin Ab (Abcam), HRP-conjugated goat anti-rabbit Ig, and HRP-conjugated rabbit anti-mouse Ig (DakoCytomation) were used.
Cell cultures and monocytic differentiation of U937 cells with IFN-γ
Generation of GCN5−/− was described in our previous report (41). DT40 cells and all subclones were cultured essentially as described (41, 42). Human monoblastic U937 cells were grown in RPMI 1640 culture medium as described (43, 44). U937 cells (1.5 × 106) in 10 ml culture medium were treated with 100 U/ml human IFN-γ at 37°C for 48 h.
Assay of superoxide generation
O2− was quantified by measuring SOD-sensitive chemiluminescence (CL) using Diogenes-luminol CL probes (45). Cells in PBS containing 1 mM MgCl2, 0.5 mM CaCl2, 5 mM glucose, and 0.03% BSA were stimulated with 200 ng/ml PMA at 37°C, and the O2− generation was measured by an automatic luminescence analyzer LB9505C (Berthold) or TD-20/20 luminometer (Promega).
Total RNAs were isolated from DT40 and its subclones and undifferentiated and differentiated U937 cells. Reverse transcription was performed with a first-strand DNA synthesis kit (Toyobo) at 42°C for 20 min, followed by heating at 99°C for 5 min. PCRs were carried out at 96°C for 20 s, 55°C for 30 s, and 72°C for 30 s for 20–40 cycles using sense and antisense primers, which were synthesized according to the EST data deposited in GenBank for appropriate genes and are listed in Supplemental Table I. Chicken and human GAPDH genes were used as internal controls. PCR products were subjected to 1.5% agarose gel electrophoresis. Data obtained by semiquantitative RT-PCR before reaching the plateau were analyzed by Image Gauge software Profile mode (densitometrical analysis mode) using a luminescent image analyzer, LAS-1000plus (Fujifilm). Nucleotide sequences of all amplified RT-PCR products were confirmed by the PCR sequencing method as described (41).
Cells (1 × 107) were collected by centrifugation and sonicated in 100 μl 50 mM Tris-HCl buffer (pH 7.5) containing 0.25 M sucrose, 2 mM EDTA, and 1 mM PMSF. Cell lysates were centrifuged, and supernatants obtained (cytosolic fractions) were treated with 10% trichloroacetic acid, collected by centrifugation, dissolved in 0.5 M Tris-HCl (pH 6.8) containing 2.5% SDS, 10% glycerol, and 5% 2-ME, and heated at 100°C for 5 min. Precipitates of cell lysates (membrane fractions) were suspended in 50 μl solubilizing solution containing 9 M urea, 2% Triton X-100, and 5% 2-ME. Fifty microliters loading buffer (0.5 M Tris-HCl [pH 6.8] containing 5% SDS and 20% glycerol) was added to the solubilized membrane fractions. Immunoblotting was performed as described (41–43, 45). Data analyses were carried out using a luminescent image analyzer, LAS-1000plus (Fujifilm). Human β-actin was used as a control.
Chromatin immunoprecipitation assay
Chromatin immunoprecipitation (ChIP) assay was performed using a Chromatin Immunoprecipitation Assay Kit (Millipore) as follows. U937 cells were cultured with or without 100 U/ml IFN-γ for 48 h. Cells (1 × 106) were incubated with formaldehyde and lysed in the presence of 1 mM PMSF and 100 μg/ml aporotinin. Lysates were sonicated with Biorupter UCD-250 (Cosmo Bio) to shear DNA to lengths between 200 and 1000 bp and centrifuged for 10 min at 15,000 × g at 4°C. Supernatants were diluted with 10 volumes dilution buffer, precleared with Protein A Agarose/Salmon Sperm DNA slurry for 1 h at 4°C, and subjected to immunoprecipitation with 4 μg anti-human HAT Abs (CBP, GCN5, p300, or PCAF) (or 4 μg normal rabbit IgG as negative control) or 2 μl antiacetylated histone antisera (or 2 μl normal rabbit serum as negative control) with rotation overnight at 4°C. A 1/50 volume of the diluted lysates was removed after preclearing (for input). Ab–chromatin complexes were recovered by incubation with Protein A Agarose/Salmon Sperm DNA slurry at 4°C for 1 h. Precipitates were washed according to the protocol, and Ab–chromatin complexes were removed from the beads by incubating in elution buffer for 30 min at room temperature. Cross-linking of immunoprecipitated and input DNAs was reversed by incubation for 8 h at 65°C. Immunoprecipitated and input DNAs were recovered by ethanol precipitation. DNAs were analyzed to confirm the presence of human gp91-phox promoter sequences containing critical cis-element (and also Hox/Pbx consensus-like cis-element) and inverted PU.1 binding site (25, 30, 46) (Supplemental Fig. 1) by PCR. PCRs were carried out at 96°C for 20 s, 55°C for 30 s, and 72°C for 30 s, for 30–40 cycles, using sense (forward) primer 5′-TCAGTTGACCAATGATTATTAGCCAATT-3′ and antisense (reverse) primer 5′-CTATGCTTCTTCTTCCAATGACCAAAT-3′ (Supplemental Fig. 1). PCR was stopped before reaching the plateau. PCR products were subjected to 1.5% agarose gel electrophoresis and analyzed by Image Gauge software Profile mode (densitometrical analysis mode) using a luminescent image analyzer, LAS-1000plus (Fujifilm).
Loss of superoxide-generating activity and drastic downregulation of gp91-phox gene expression in GCN5−/− cells
We examined effects of GCN5 deficiency on the O2−-generating activity (Fig. 1A). As expected, DT40 cells generated CL when stimulated with PMA, and it was completely abolished by SOD (100 μg/ml). These results suggested that CL generated by PMA-stimulated DT40 cells was preferentially derived from O2−. In contrast, PMA-stimulation for GCN5−/− (three independent clones tested) showed a negligible level of O2− generation, indicating that depletion of GCN5 completely prevented the O2−-generating system. Next, to know the effects of GCN5 deficiency on transcriptions of four genes responsible for CGD (p22-phox, gp91-phox, p47-phox, and p67-phox), we carried out semiquantitative RT-PCR on total mRNAs prepared from DT40 and three independent GCN5−/− clones (Fig. 1B). Interestingly, transcription of the gp91-phox gene was drastically downregulated in GCN5−/− (to ∼4%). Transcription levels of p22-phox and p67-phox genes were decreased (to ∼65 and ∼80%), whereas p47-phox mRNA level was increased (to ∼170%). These results revealed that the drastic effect of GCN5 deficiency on the gene expression of gp91-phox probably resulted in the loss of O2−-generating activity. Next, to confirm the role of GCN5 in the gene expression of gp91-phox, re-expression study was carried out using GCN5 expression vector (Supplemental Fig. 2). Unfortunately, re-expression of GCN5 could not complement both the decreases in the gp91-phox gene expression and the loss of O2−-generating activity (Supplemental Fig. 2B). The gene expression of PCAF that was extremely increased in GCN5−/− (41) was further upregulated by re-expression of GCN5 (to ∼200% of GCN5−/−) (Supplemental Fig. 2C). These results suggested that PCAF could play, in part, complementary roles in chromatin dynamics linked to the acetylation state of core histones in absence of GCN5 (41), and the altered epigenetic state caused by GCN5 deficiency may not be restored by re-expression of GCN5. Its molecular mechanism remains to be resolved. In addition, we examined effects of PCAF deficiency on the O2−-generating activity using PCAF−/− [homozygous DT40 mutant cell line devoid of two PCAF alleles (41)] (Supplemental Fig. 3A). PMA stimulation for PCAF−/− (three independent clones tested) induced about a half level of O2− generation, indicating that the depletion of PCAF partly repressed the O2−-generating system. To determine effects of PCAF deficiency on transcriptions of p22-phox, gp91-phox, p47-phox, and p67-phox genes, we carried out semiquantitative RT-PCR on total mRNAs prepared from DT40 and three independent PCAF−/− clones (Supplemental Fig. 3B). Transcriptions of the gp91-phox and p22-phox genes were downregulated (to ∼45 and to ∼80%), and the decreases were less when compared with those in GCN5−/− clones. The deficiency showed insignificant influence on transcription of the p47-phox gene, whereas the transcription level of the p67-phox gene was considerably increased (to ∼150%). These results suggested that in regulation of gp91-phox gene expression, GCN5 acts as a supervisor, and PCAF plays a supporting role.
Effects of IFN-γ treatment on gene expression of gp91-phox and GCN5 in U937 cells
To clarify molecular mechanisms for regulation of the O2−-generating system via controlling the gp91-phox gene expression by GCN5, human monoblastic U937 cells were used as an in vitro model for monocytic differentiation (43, 44). When U937 cells are cultured in the presence of various agents, they differentiate to macrophage-like cells that can produce O2−. IFN-γ, one of the effective inducers for monocytic differentiation, induces the gp91-phox gene expression, and the induction mechanisms mediated by IFN-γ have been well studied (20, 21, 24, 27, 28, 30, 46, 47).
Undifferentiated U937 cells had no activity of O2− generation when stimulated with PMA. After 48 h incubation with IFN-γ, they acquired the O2−-generating activity as judged by CL assay (43, 44) (Fig. 2A). Remarkable increases in gp91-phox gene expression (to ∼850%) and also in protein level (to ∼1200%) were observed in parallel with induction of the O2−-generating activity, whereas insignificant changes in the GCN5 gene expression and also in protein level were detected (Fig. 2B, 2C). In addition, the IFN-γ–induced O2−-generating activity and increased gene expression and protein level of gp91-phox in U937 cells were completely inhibited by addition of 50 μM GCN5 inhibitor CPTH2 (48) (Fig. 3, Supplemental Fig. 4). Interestingly, CPTH2 remarkably downregulated protein levels of p47-phox and p67-phox (Supplemental Fig. 4B), although it showed insignificant effects on gene expressions of these two factors (Fig. 3B, Supplemental Fig. 4A). We also examined inhibition properties of CPTH2 (Supplemental Fig. 5). CPTH2 inhibited the growth of U937 cells (Supplemental Fig. 5A), induction of O2−-generating activity (Supplemental Fig. 5B), and gene expression of gp91-phox (Supplemental Fig. 5C) in a dose-dependent manner, whereas it showed an insignificant effect on cell viability (data not shown). In addition, CPTH2 showed a high specificity against GCN5 family, especially GCN5 (Supplemental Fig. 5D). These data, together with the findings obtained by the gene targeting study using DT40 (Fig. 1), supported the important participation of GCN5 in regulating the O2−-generating activity via controlling gp91-phox gene expression.
IFN-γ enhances interaction of GCN5 with gp91-phox gene promoter and acetylation levels of H2BK16 and H3K9 residues surrounding the promoter
Many transcription factors are involved in regulation of the gp91-phox gene expression through their interaction with its promoter region (16–31). As is well known, HATs play critical roles in modulation of chromatin topology and thereby regulation of gene expression through acetylation of core histones. Such acetylation neutralizes positive charges to enhance hydrophobicity of core histones, resulting in the reduced affinity of acetylated N-terminal tails of core histones to DNA (32, 33). The attenuated histone–DNA interaction is believed to change chromatin configuration for transcription activation through promoting transcription factor–DNA interaction, but molecular mechanisms of the acceleration of the gp91-phox gene expression by any HATs remain to be resolved. Because GCN5 deficiency caused a drastic decrease in the gp91-phox gene expression in DT40 cells resulting in loss of the O2−-generating activity (Fig. 1), GCN5 is expected to be the most important HAT in controlling the gene expression. To clarify molecular mechanisms for the gp91-phox gene expression regulated by GCN5, we investigated interaction of GCN5 and other major HATs (PCAF, CBP, and p300) with gp91-phox promoter by ChIP assay in U937 cells (Fig. 4). Cross-linked chromatins were coprecipitated from the lysates of U937 cells (48 h treated or untreated with IFN-γ) with Abs specific for GCN5, PCAF, CBP, and p300 or irrelevant normal IgG as a negative control. Precipitated chromatins were amplified by PCR using primers flanking the critical cis-element (Hox/Pbx consensus-like sequence) and inverted PU.1 binding site in the proximal gp91-phox gene promoter (Supplemental Fig. 1) (25, 30, 46). The data obtained showed not only that GCN5 and CBP associated with the gp91-phox gene promoter (Fig. 4A), but also that association of GCN5 or CBP with the promoter was remarkably increased or unchanged in U937 cells by IFN-γ treatment (Fig. 4A). In contrast, associations of PCAF and p300 with the gp91-phox gene promoter could not be detected by ChIP assay using several available Abs during cultivation with IFN-γ (0, 12, 24, and 48 h) (data not shown). Time-course study also showed that in IFN-γ–treated U937 cells, the interaction of GCN5 with the gp91-phox gene promoter was dramatically upregulated (to ∼300% by 48 h), although IFN-γ treatment showed an insignificant effect on the interaction of CBP with the promoter (Fig. 4B). Further, we investigated histone acetylation levels surrounding the gp91-phox gene promoter during cultivation of U937 cells with IFN-γ by ChIP assay using various anti-acetylated histone Abs (Fig. 5). Cross-linked chromatins were coprecipitated from the lysates of U937 cells (48 h treated or untreated with IFN-γ) with various antisera specific for acetylated Lys residues of histones or irrelevant normal serum as a negative control. PCR was carried out as described above. The IFN-γ treatment showed significant influences on acetylation levels of Lys16 of H2B (H2BK16) and Lys9 of H3 (H3K9) among the Lys residues of core histones H2A, H2B, H3, and H4 tested (Fig. 5A, 5B). Acetylations of H2BK16 and H3K9 surrounding the gp91-phox gene promoter were remarkably increased to ∼220 and ∼350% by 48 h, respectively, in a time-course manner during cultivation of U937 cells with IFN-γ (Fig. 5C), whereas IFN-γ treatment showed insignificant effects on acetylations of other Lys residues tested by 24 (data not shown) and 48 h (Fig. 5A). In addition, the increases in acetylation levels of both H2BK16 and H3K9 by IFN-γ were completely inhibited by addition of CPTH2 (Fig. 5D). We also examined interaction of PU.1 with the gp91-phox gene promoter during cultivation with IFN-γ (Supplemental Fig. 6). ChIP assay using anti-PU.1 Ab revealed that interaction of PU.1 with the promoter was upregulated during cultivation with IFN-γ (Supplemental Fig. 6A, 6B), and the increase was completely inhibited by addition of CPTH2 (Supplemental Fig. 6C). These results indicated not only that the association of GCN5 with the gp91-phox gene promoter was significantly accelerated in U937 cells with IFN-γ, but also that GCN5 preferentially acetylated H2BK16 and H3K9 surrounding the promoter, resulting in the activation of gp91-phox gene expression.
We demonstrated that GCN5, one of the most important HATs, regulates the O2−-generating system in leukocytes via controlling the gp91-phox gene expression. Recently, by analyzing the GCN5-deficient DT40 cells, GCN5−/−, we showed not only that GCN5 is preferentially involved as a supervisor in the normal cell-cycle progression through comprehensive control of expressions of various cell cycle-related genes (41), but also that it controls BCR-mediated apoptosis in immature B lymphocytes through regulation of several apoptosis-related genes (42). First, we studied the participation of GCN5 in induction of the O2−-generating activity by exposing GCN5−/− to PMA. Among various NOX proteins (NOX1∼5 and DUOX1, 2) belonging to the NOX family in vertebrates, gp91-phox (NOX2) and NOX5 are expressed in B cells and lymphoid tissues, respectively (15, 49). Although we carried out RT-PCR using chicken NOX5 primers, the gene expression of NOX5 could not be detected in both DT40 and GCN5−/− (data not shown), indicating that the PMA-induced O2−-generating activity in DT40 mostly depends on gp91-phox. Therefore, we focused on gp91-phox in the NOX family in this study. Interestingly, GCN5 deficiency completely inhibited the O2−-generating activity (Fig. 1A) and reversely caused a drastic decrease in mRNA levels of gp91-phox (to ∼4%), whereas it had minor influences on those of p47-phox, p22-phox, and p67-phox (i.e., the elevation [to ∼170%] for the former and the slight suppression [to ∼65 and 82%] for the latter two) (Fig. 1B). In contrast, the deficiency of PCAF belonging to the GCN5 family slightly suppressed both the O2−-generating activity (to ∼50%) and gp91-phox gene expression (to ∼45%) compared with those in GCN5−/− (Supplemental Fig. 3). These results suggested that GCN5 mainly regulates gene expression of gp91-phox as a supervisor, and PCAF plays a supporting role in regulation of the gene expression.
To confirm this hypothesis, we used an in vitro differentiation system of U937 cells. The gp91-phox gene is a major gene regulated by IFN-γ (46, 47). When U937 cells were treated with IFN-γ, the gp91-phox gene was remarkably expressed (to ∼800%), whereas the expression of GCN5 gene was not affected (Fig. 2B). Many groups have reported that GCN5 catalyzes acetylation of several specific Lys residues and causes transcriptional activation (50–57). In this study, we clarified that CPTH2, an inhibitor of GCN5 (48), strongly inhibited IFN-γ–induced O2−-generating activity (Fig. 3A), gene expression of gp91-phox (Fig. 3B), and acetylation of H2BK16 and H3K9 (Fig. 5D) in U937 cells. Further, it is worth noting that CPTH2 can inhibit GCN5 strongly and PCAF moderately (Supplemental Fig. 5D). The binding of GCN5 to gp91-phox gene promoter was remarkably upregulated during cultivation with IFN-γ (Fig. 4A), whereas the association of PCAF with the promoter could not be detected by ChIP assay using several available anti-PCAF Abs (data not shown). Therefore, our data suggested that the major target of CPTH2 could be GCN5. After all, ChIP assay revealed not only that the binding of GCN5 to gp91-phox gene promoter was remarkably upregulated without its increased transcription, but also that GCN5 promoted acetylation of H2BK16 and H3K9 surrounding the promoter, resulting in recruitment of various transcription factors (e.g., PU.1) during cultivation of U937 cells with IFN-γ (Figs. 4, 5, Supplemental Fig. 6) (25, 30, 46). These data, together with the findings obtained for the GCN5-deficient DT40 cells (Fig. 1), suggested the participation of GCN5 in regulating the O2−-generating activity via controlling the gp91-phox gene expression. In contrast, interestingly, GCN5 deficiency in DT40 cells also led to decreased acetylation levels of only two Lys residues in core histones: H2BK16 and H3K9 (41). It has been reported that the acetylation of H3K9 catalyzed by GCN5 participates in the activation of transcription (51–55). In addition, GCN5 is reported to be required for the acetylation of H2BK16, resulting in transcriptional activation in yeast (52). Therefore, acceleration of acetylation of H3K9 (and probably also H2BK16) may cause a remarkable increase in gp91-phox gene expression during cultivation of U937 cells with IFN-γ, although acetylation levels of other possible GCN5-catalyzed acetylation sites in core histones—H3K14, H3K18, H3K23, H3K27, H3K56, H4K8, and H4K16 (50–57)—remained unchanged (Fig. 5A). Previous studies showed that transcription factors may be newly expressed or modified, leading to interaction with DNA followed by gp91-phox gene expression (16). For IFN-γ–induced expression of the gp91-phox gene, PU.1 is thought to first bind to the promoter and participate in recruitment of other factors (28). Our data in this study, together with previous findings (28, 53, 54, 56), suggested that GCN5 (probably as HAT complex including GCN5) catalyzes the acetylation of H3K9 (and also H2BK16) residue surrounding the gp91-phox gene promoter; the increased histone acetylation accelerates binding of PU.1 to the promoter sequence; PU.1 recruits other activators to the region and finally promotes the gp91-phox gene expression during cultivation with IFN-γ (Supplemental Fig. 6).
Epigenetic mechanisms can define alterations in cellular phenotypes without altering genotypes (58). According to this theory, epigenetic control of transcriptional activation or inactivation is mostly influenced by the intricately and timely modifications of chromatin-bound histones mediated by acetylation, methylation, phosphorylation, ADP-ribosylation, and/or ubiquitination known as histone code (58). It is well known that acetylation and deacetylation catalyzed reversibly by HATs and histone deacetylases play critical roles in the modulation of chromatin topology and thereby regulation of gene expression. In conclusion, our results obtained in this study indicate not only that GCN5 probably takes part in transcriptional regulation of the gp91-phox gene through alterations in the chromatin structure surrounding its promoter region, but also that as a result, GCN5 plays a key role in mechanisms of epigenetic regulation of the O2−-generating system. We also revealed the moderate participation of PCAF in regulating the O2−-generating activity via controlling gp91-phox gene expression (Supplemental Fig. 3); its molecular mechanisms remain to be resolved. In addition, the epigenetic regulation within U937 cells may be different from that in normal monocyte/macrophage because U937 cells are lymphoma. Therefore, the participation of GCN5 in the O2−-generating system in development of normal cells should be elucidated in the future. Anyhow, our results, together with enormous previous data, may significantly help in the understanding of CGD and epigenetic regulation of leukocyte differentiation.
We thank M. Nakayama for technical support and H.K. Barman, H. Madhyastha, and K.S. Radha for editorial reading of the manuscript.
This work was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
The online version of this article contains supplemental material.
Abbreviations used in this article:
chronic granulomatous disease
Lys16 of H2B
Lys9 of H3
p300/CREB-binding protein-associated factor
The authors have no financial conflicts of interest.