Histone hyperacetylated domains across the Ifng gene region in natural killer cells and T cells

Abstract

Local histone acetylation of promoters precedes transcription of many genes. Extended histone hyperacetylation at great distances from coding regions of genes also occurs during active transcription of gene families or individual genes and may reflect developmental processes that mark genes destined for cell-specific transcription, nuclear signaling processes that are required for active transcription, or both. To distinguish between these, we compared long-range histone acetylation patterns across the Ifng gene region in natural killer (NK) cells and T cells that were or were not actively transcribing the Ifng gene. In T cells, long-range histone acetylation depended on stimulation that drives both T helper (Th) 1 differentiation and active transcription, and it depended completely or partially on the presence of Stat4 or T-bet, respectively, two transcription factors that are required for Th1 lineage commitment. In contrast, in the absence of stimulation and active transcription, similar histone hyperacetylated domains were found in NK cells. Additional proximal domains were hyperacetylated after stimulation of transcription. We hypothesize that formation of extended histone hyperacetylated domains across the Ifng gene region represents a developmental mechanism that marks this gene for cell- or stimulus-specific transcription.

Mechanisms that permit cell-type-specific transcription of genes are incompletely understood. For example, T lymphocytes and natural killer (NK) cells develop from a common bone marrow-derived hematopoietic progenitor (1). T cells emerge from additional developmental programs in the thymus and migrate to peripheral lymphoid organs. Once in the periphery, they must endure further differentiation before they are competent to transcribe key effector cytokine genes, such as the Ifng gene, necessary for their function in the immune system (2-6). In contrast, NK cells fully mature in the bone marrow and are preprogrammed to rapidly transcribe the Ifng gene after appropriate stimulation in the periphery (7). The molecular bases for these developmental differences are incompletely understood.

Broad, nonuniform histone hyperacetylation patterns that span >100 kb of genomic DNA are formed in chromatin surrounding gene families, such as the growth hormone or β-globin gene locus during tissue-specific transcription (8-11). Long-range, nonuniform histone hyperacetylation patterns are also formed across the Ifng gene region, a gene not known to be a member of a gene family, during the differentiation processes that lead to tissue-specific transcription in CD8⁺ T cells (12). One question raised by these studies is whether establishment of long-range histone hyperacetylation patterns across the Ifng gene region reflects a developmental process that is necessary for active gene transcription, a nuclear signaling process that is required for efficient transcription, or both. A second question is whether key transcriptional regulators of T helper (Th) 1 lineage commitment, Stat4 and T-bet, are required to establish long-range histone hyperacetylation. To distinguish between developmental and transcriptional processes, we compared long-range histone hyperacetylation patterns across the Ifng gene region in Th1 and Th2 effector cells and in NK cells in the resting state and after induction of Ifng gene transcription by stimulation with IL-12 and IL-18. In T cells, we also determined whether establishment of long-range histone acetylation patterns was Stat4- or T-bet-dependent. We performed quantitative chromatin immunoprecipitation (ChIP) analysis, using anti-acetyl histone H4 antibodies to assess levels of long-range histone acetylation. Establishment of long-range histone hyperacetylation patterns in T cells depends on the transcription factors, Stat4 and T-bet, and numerous sites across this region bind Stat4 and T-bet under Th1 conditions in vivo. NK cells exhibit a preexisting long-range pattern of histone hyperacetylation across the Ifng gene region before significant transcription. Induction of transcription changes this pattern. We hypothesize that broad, nonuniform histone acetylation across the Ifng gene region may mark this region for tissue- or stimulus-specific gene expression and may serve to recruit protein complexes to the region to facilitate transcription.

Materials and Methods

Mice. C57BL/6, BALB/c, C57BL/6.Stat4^-/-, and BALB/c.T-bet^-/- mice were obtained from The Jackson Laboratory, bred in Vanderbilt University's animal facilities, and used between 4 and 5 wk of age. Research using mice complied with all relevant institutional and federal guidelines and policies.

Cell Purification and Cultures. NK cells were purified from spleen by negative selection using two different methods. First, NK cells were purified from spleen by using a NK purification system (Miltenyi Biotec, Auburn, CA) according to the manufacturer's protocol. Second, spleen cells were labeled with anti-CD4, anti-CD8, and anti-Ia monoclonal antibodies. NK cells were purified by negative selection with magnetic beads and expanded with human IL-2 (1,000 units/ml). Cultures were performed in RPMI medium 1640 supplemented with 10% FBS/penicillin/streptomycin/glutamine/5 × 10^-5 M 2-mercaptoethamol at 37°C in an incubator containing 5% CO₂ in air. Purity of both freshly isolated and IL-2-expanded NK cells was >95% as determined by flow cytometry. NK cells, in the presence of IL-2, were stimulated with IL-12 and/or IL-18 (5 ng/ml each) for varying periods of time as outlined in Results. Splenic CD4⁺ T cells were purified by negative selection using specific monoclonal antibodies essentially as described in refs. 16 and 17. CD4⁺ T cells (1 × 10⁶ cells per ml) were stimulated with immobilized anti-CD3 monoclonal antibody and irradiated spleen cells (1 × 10⁶ cells). Anti-CD3-coated plates were prepared by treating tissue culture plates with a 10 μg/ml coating solution overnight at 0-4°C (2C11 monoclonal antibody, American Type Culture Collection). T cells were stimulated under neutral (no cytokine or other monoclonal antibody additions), Th1 (10 μg/ml anti-IL-4 and 5 ng/ml IL-12), or Th2 (10 μg/ml anti-IFN-γ and 5 ng/ml IL-4) conditions, essentially as described in refs. 16 and 17. IFN-γ assays were performed by ELISA with mAbs recommended by BD Pharmingen.

ChIP Assays. NK cells or T cells were processed for the ChIP assays. Briefly, cells (≈1 × 10⁷) were fixed with 1% paraformaldehyde and lysed in 1% SDS/10 mM EDTA/50 mM Tris·HCl, pH 8.1 (600-μl total volume). Lysates were sonicated to achieve an average length of genomic DNA of ≈500 bp. Protein-DNA complexes (100 μl of the 600-μl total volume) were immunoprecipitated overnight at 4°C with anti-acetyl histone H4 antibody (Upstate Biotechnology, Lake Placid, NY), anti-Stat4, anti-T-bet (both from Santa Cruz Biotechnology), or normal rabbit IgG as control. Antibody-histone-DNA complexes were purified with protein A agarose. Aliquots (40 μl of the 600-μl total volume) were processed without immunoprecipitation as input controls. DNA was purified after reversal of cross-links at 65°C and suspended to a total volume of 100 μl, and 2 μl was used for PCR. PCR primers for the ChIP assay spanned nonrepetitive conserved (mouse and human) genomic sequences from ≈70 kb upstream to ≈70 kb downstream of the Ifng gene. Sequence conservation was identified by using the UCSC Genome Browser web site (18-20). Real-time PCR was used to determine quantitative histone acetylation (Q-HAc) levels (for histone H4) at the different sites across the ≈140-kb region by using SYBR green. Dissociation curves after amplification showed that primer pairs generated single products, and this observation was confirmed by analysis on agarose gels. Relative levels of histone H4 acetylation at the different positions were obtained by determining the amount of immunoprecipitated DNA from a standard curve generated by serial dilution of input DNA. Standard curves produced from input DNA were identical among the different mouse strains used in the experiments. Actb was used as an internal standard to control for interexperimental variability. Quantitative ChIP assays were performed a minimum of three times.

Results

Stat4, T-bet, and Establishment of Long-Range Histone Acetylation Across the Ifng Region in Th1 Cells. We wanted to determine whether establishment of long-range histone hyperacetylation across the Ifng region depended on two key transcription factors that are required for Th1 lineage commitment, Stat4 and T-bet. We also wanted to use quantitative ChIP assays and additional primer pairs to gain a greater appreciation of differences in levels and extent of subdomains of histone hyperacetylation. CD4⁺ T cells from C57BL/6, BALB/c, C57BL/6.Stat4^-/-, and BALB/c.T-bet^-/- mice were stimulated for 3 d under neutral, Th1, or Th2 differentiation conditions. CD4⁺ T cells from C57BL/6 mice were also stimulated in the presence of the specific inhibitor of histone deacetylases (HDACs), trichostatin A (TSA) (100 nM). Cells were harvested and processed for ChIP histone acetylation assays, and culture fluids were harvested and analyzed for levels of IFN-γ by ELISA. High levels of IFN-γ were found in Th1 cultures but not in Th2 cultures from WT cells. Levels of IFN-γ were markedly reduced in Th1 cultures from Stat4^-/- and T-bet^-/- mice compared with WT mice (Fig. 1). Addition of TSA induced IFN-γ production under neutral conditions but not under Th2 conditions. Somewhat less IFN-γ was found in BALB/c Th1 cultures than in C57BL/6 Th1 cultures.

Fig. 1.

Stat4- and T-bet-dependent Th1 differentiation. Purified CD4⁺ T cells from C57BL/6, C57BL/6.Stat4^-/-, BALB/c, or BALB/c.T-bet^-/- mice were stimulated under neutral, Th1, or Th2 conditions in the presence or absence of TSA. After 3 d, cultures were harvested and analyzed for levels of IFN-γ by ELISA. Results are expressed as the mean level of IFN-γ in ng/ml ± SD (total of three experiments).

We compared the degree of DNA sequence conservation among species to Q-HAc at the Ifng gene and distal regions across ≈140 kb of genomic DNA. This region is devoid of other known genes. The Ifng gene contains four exons and three introns that span ≈5 kb (Fig. 2A). Three distal enhancers are known at ≈-6, -3.5, and +17-19 kb. Known DNase hypersensitivity sites are localized within distal enhancers, the promoter, and the first and third introns. We used the UCSC Genome Browser web site to derive levels of DNA sequence conservation across this region. Promoter elements of genes usually show some of the strongest DNA sequence conservation across species. Therefore, we were interested to see whether other regions besides the promoter had similar levels of DNA sequence conservation to the Ifng promoter (Fig. 2B). A region at ≈-35 to -36 kb 5′ of the Ifng start site contained the highest level of DNA sequence conservation across this region, which was slightly higher than the promoter and a region in the last exon of the Ifng gene. Other regions both 5′ and 3′ of the Ifng gene also possessed high levels of DNA sequence conservation. Next, we compared Q-HAc across this region in naïve CD4⁺ T cells and in T cells cultured under neutral (±TSA), Th1, or Th2 conditions (Fig. 2C). Histone acetylation was undetectable in naïve CD4⁺ T cells (not shown) and almost completely blunted in neutral and Th2 cells. In contrast, this region was heavily histone hyperacetylated in Th1 cells with seven distinct subdomains located at -70 to -50, -30 to -20, ≈-6, -0.4 to +2, ≈+20, ≈+30, and +50 to +60 kb. The pattern of histone hyperacetylation was similar in Th1 cells and T cells cultured under neutral conditions with TSA. TSA did not induce histone hyperacetylation in Th2 cells (not shown).

Fig. 2.

Comparison of DNA sequence conservation among species and long-range Q-HAc in T cells. (A) Schematic of the murine Ifng gene, including positions of known DNase hypersensitivity sites (arrows) and distal enhancer regions (rectangles) (from refs. 27-29). (B) Sites of DNA sequence conservation were obtained from the UCSC Genome Browser web site and are expressed as logarithm-of-odds (LOD) scores derived from a phylogenetic Markov model produced by the phastcons program. (C) Histone acetylation or Q-HAc was determined in chromatin that was isolated from naïve CD4⁺ T cells or T cells cultured under neutral, Th1, or Th2 conditions for 3 d in the presence or absence of TSA as indicated. Results are expressed in mean relative units ± SD.

We considered that regions of high, long-range Q-HAc might display the greatest level of DNA sequence conservation. However, this hypothesis proved not to be completely the case. For example, the region between -35 and -36 kb from the Ifng start site displayed the greatest level of sequence conservation but relatively weak Q-HAc. Similarly, the region at -22 kb showed strong Q-HAc but relatively weak sequence conservation. The promoter at -0.4 kb and the last exon of the Ifng gene at +4.2 kb also showed strong sequence conservation compared with internal portions of the gene at +0.4, +1, and +2 kb. However, the internal regions displayed the greatest level of Q-HAc. In contrast, both sequence conservation and Q-HAc were high in the region around +20 and +30 kb. In the region between +50 and +70 kb, strong Q-HAc did not correlate with strong DNA sequence conservation. Thus, high levels of sequence conservation do not necessarily correlate with the highest levels of Q-HAc.

Next, we wanted to determine whether formation of long-range histone hyperacetylation during Th1 differentiation required Stat4 or T-bet. To do so, we compared long-range Q-HAc in CD4⁺ T cells from WT and Stat4^-/- or T-bet^-/- mice after culture under Th1 conditions. We examined long-range Q-HAc in T cells from WT C57BL/6 and BALB/c mice because the Stat4^-/- mice are backcrossed to C57BL/6 mice and the T-bet^-/- mice are backcrossed to BALB/c mice. Each subdomain was histone acetylated in WT T cells from both strains (Fig. 3A). Next, we examined levels of Q-HAc in Stat4^-/- T cells cultured under Th1 conditions. Long-range histone hyperacetylation was greatly diminished along the entire Ifng gene region in Stat4^-/- T cells (Fig. 3B). We also examined levels of long-range Q-HAc in T-bet-deficient T cells. We found that the three subdomains 5′ of the Ifng gene, and the Ifng gene itself, exhibited similar histone hyperacetylation levels in WT and T-bet^-/- T cells. In contrast, Q-HAc levels in the subdomains 3′ of the Ifng gene were greatly diminished in T-bet^-/- T cells (Fig. 3C). TSA partially restored long-range histone hyperacetylation patterns in both Stat4 and T-bet^-/- T cells cultured under Th1 conditions and partially restored IFN-γ levels (not shown). Taken together, these data demonstrate that both Stat4 and T-bet play key roles in establishing long-range histone hyperacetylation across the Ifng gene region during Th1 differentiation.

Fig. 3.

Long-range Q-HAc depends on the transcription factors Stat4 and T-bet, which are required for Th1 lineage commitment. (A) Long-range Q-HAc is similar in Th1 cells from C57BL/6 and BALB/c mice. Chromatin was isolated from C57BL/6 and BALB/c T cells cultured under Th1 conditions for 3 d, and Q-HAc was determined as in Fig. 2C. (B) Long-range Q-HAc depends on Stat4. Chromatin was isolated from T cells from C57BL/6 and C57BL/6.Stat4^-/- mice cultured under Th1 conditions for 3 d. Q-HAc was determined as in Fig. 2C. (C) Dependence of long-range Q-HAc on T-bet. Chromatin was isolated from BALB/c and BALB/c.T-bet^-/- T cells after culture for 3 d under Th1 conditions. Q-HAc was determined as in Fig. 2C. (D) Stat4 and T-bet bind to multiple sites along the Ifng gene region in Th1 cells. ChIP assays were performed essentially as described, except that anti-Stat4 or anti-T-bet antibodies were used for immunoprecipitation.

We determined whether T-bet and Stat4 were bound to chromatin at these hyperacetylated sites across the Ifng gene region by using ChIP analysis, and we identified numerous T-bet- and Stat4-binding sites. Stat4 binding to chromatin was clearly evident at -0.4, 0.4, 20, and 40 kb, and T-bet binding to chromatin was clearly evident at -0.4, 10, 20, 30, and 40 kb from the Ifng gene in Th1 cells (Fig. 3D) but not in Th2 cells (not shown). We suspect that other weak binding sites (-70, -63, etc.) may be significant and that we may need to perform a more detailed analysis to pinpoint the exact positions of these binding sites across this genomic region.

Stable, Long-Range Histone Acetylation in NK Cells. In contrast to T cells, once NK cells are in the periphery, they do not have to endure further differentiation programs before they are able to rapidly produce large amounts of IFN-γ after appropriate stimulation (21, 22). Therefore, we examined long-range Q-HAc in NK cells that were or were not actively transcribing the Ifng gene. We compared IFN-γ production and Q-HAc in both freshly isolated NK cells and NK cells that were expanded by culturing in IL-2 to assess changes in long-range Q-HAc and IFN-γ production after cell division. During the expansion period, we monitored levels of IFN-γ, and culture fluids contained undetectable levels of IFN-γ (<0.03 ng/ml). Upon stimulation with IL-12, IL-18, or IL-12 and IL-18, in the presence of a constant amount of IL-2, NK cells rapidly produced large amounts of IFN-γ (Fig. 4A). Levels of IFN-γ in culture fluids after stimulation of NK cells did not vary as a function of days after isolation and culture. They were comparable to those produced by differentiated Th1 or T cytotoxic (Tc) 1 cells on a per-cell basis (Fig. 4B). Increases in mRNA levels of Ifng preceded the increase in IFN-γ in culture fluids but essentially mirrored the rate of change of IFN-γ levels under the different stimulation conditions (data not shown).

Fig. 4.

NK cell production of IFN-γ after cytokine stimulation. (A) NK cells were harvested from tissue culture at the indicated times after isolation and stimulated with IL-12, IL-18, or IL-12 and IL-18. Cultures were harvested 24 h after stimulation, and levels of IFN-γ (ng/ml ± SD) were determined by ELISA. (B) NK cells, naïve T cells, effector Th1 cells, and effector Tc1 cells were stimulated with IL-12 and IL-18. Cultures were harvested at the indicated times. IFN-γ levels (ng/ml ± SD) were determined by ELISA.

We performed ChIP analysis using anti-acetyl histone H4 antibodies to assess Q-HAc across the ≈140-kb Ifng gene region in freshly isolated NK cells and in NK cells that had been cultured with IL-2 only and were not producing IFN-γ. ChIP analysis revealed that both populations of NK cells exhibited similar patterns of Q-HAc with as many as five separate domains of Q-HAc across the Ifng gene region under these conditions (Fig. 5), including 5′ domains at -70 to -50 and -30 to -20 kb from the Ifng gene, a domain that included the Ifng promoter (-0.4 kb), and 3′ domains at +20 and +30 to +40 kb from the Ifng gene. Regions in between these domains had low levels of Q-HAc. Identical domains exhibited high levels of Q-HAc in Th1 cells. In contrast, domains at -6, +0.4 to +2, and +50 to +60 kb that were histone acetylated in Th1 cells were not histone acetylated in NK cells.

Fig. 5.

Resting NK cells have a preexisting long-range histone acetylation pattern across the Ifng gene region. NK cells were freshly isolated from spleen or harvested after 3 wk of culture and processed for ChIP assay. Positions of the different primer pairs relative to the Ifng gene are shown. Q-HAc was determined as in Fig. 2C.

Stimulus-Specific Histone Acetylation. Two regions that were not histone acetylated in resting NK cells but were histone acetylated in Th1 cells included the region at -6 kb and the Ifng gene (+0.4 to +2 kb). Therefore, we determined whether stimulation of Ifng transcription in NK cells by IL-12, IL-18, or IL-12 and IL-18 altered long-range histone acetylation patterns in NK cells. Stimulation of NK cells with IL-12 resulted in the selective increase of histone H4 acetylation at -6 kb 5′ of the Ifng gene (Fig. 6) but not at other regions of the Ifng gene (not shown). The level of Q-HAc was similar to that observed in Th1 cells (compare with Fig. 2C). In contrast, stimulation of NK cells with IL-18 did not cause detectable alterations in Q-HAc. Stimulation of NK cells with IL-12 and IL-18 caused a marked increase in Q-HAc at two sites in the first intron of the Ifng gene (Fig. 7). Levels of Q-HAc at these sites were similar in stimulated NK cells and Th1 cells. Increases in histone acetylation after IL-12 and IL-18 stimulation were restricted to the first intron of the Ifng gene. We did not detect other alterations in histone acetylation across the extended Ifng gene region. Thus, distinct signaling paths that stimulate Ifng gene transcription cause histone acetylation of different regions of the Ifng gene.

Fig. 6.

IL-12 stimulates histone acetylation at a specific site at -6kb5′ of the Ifng gene. NK cells were unstimulated or stimulated with IL-12 or IL-18 for 6 h. Samples were processed for ChIP assay and analyzed as outlined in Materials and Methods. Results of Q-HAc analysis at -6kb5′ of the Ifng gene are shown.

Fig. 7.

Stimulation of NK cells with the combination of IL-12 and IL-18 leads to selective histone acetylation of the first intron of the Ifng gene. NK cells received no stimuli or were stimulated with IL-12, IL-18, or IL-12 and IL-18 for 6 h. Samples were processed for ChIP assay and analyzed as described in the legend of Fig. 6. Results of Q-HAc analysis at positions -0.4, +0.4, and +1.1 kb from the Ifng start site are shown.

Discussion

Histone posttranslational modifications in promoter regions of genes, such as acetylation, methylation, and phosphorylation, play key roles in regulation of promoter activity and transcription (23-26). As examples, promoters of both the Ifng and Il4 genes are histone hyperacetylated during active transcription by Th1 and Th2 cells, respectively (27, 28). We have investigated alterations in histone hyperacetylation along the Ifng gene region in CD8⁺ T cells as they differentiate into effector Tc1 cells that actively transcribe the Ifng gene and effector Tc2 cells that do not actively transcribe the Ifng gene (12). The Ifng gene region acquires extensive histone acetylation across the ≈140-kb region both 5′ and 3′ of the ≈5-kb Ifng gene during Tc1 differentiation and active gene transcription.

The purpose of this study was to determine whether (i) Th1 differentiation results in similar changes in long-range histone acetylation, (ii) key transcriptional regulators of Th1 differentiation, Stat4 and T-bet, are required to establish long-range histone acetylation, and (iii) NK cells exhibit preexisting long-range histone acetylation patterns or acquire them after stimulation. We used quantitative methods to gain a greater appreciation of the level of histone acetylation of separate subdomains in different cell types and to produce a higher resolution map. We identified seven subdomains of histone acetylation at ≈-70 to -55, -28 to -14, -6, -0.4 to +2 (the Ifng gene), ≈+20, +30 to +46, and +50 to +60 kb. Generally, the overlap between regions of highest Q-HAc and highest DNA sequence conservation among species is imperfect. Many regions of highest sequence conservation do not correspond to regions of highest histone acetylation, whereas, in other regions, the overlap is greater. A higher resolution map will be necessary to gain a greater appreciation of the relationship between long-range histone acetylation and DNA sequence conservation across this region.

The transcription factors Stat4 and T-bet are key regulators of Th1 differentiation and Ifng transcription (13-15). Precise mechanisms by which these transcription factors induce Th1 development are incompletely understood. Here, we show that establishment of long-range histone acetylation across the entire 140-kb Ifng gene region is almost entirely dependent on the presence of Stat4. In contrast to this, absence of T-bet results in loss of histone acetylation at the distal 3′ subdomains at ≈+20, +30 to +46, and +50 to +60 kb. This result essentially confirms that histone acetylation at ≈+20 kb is T-bet-dependent (29) and extends the positions of T-bet-dependent histone acetylation domains to include ones at +30 to +46 and +50 to +60 kb. In contrast to these 3′ subdomains, histone acetylation of subdomains 5′ of the Ifng gene is relatively T-bet-independent. Stat4 and T-bet bind to numerous sites across the ≈140-kb Ifng gene region under Th1 culture conditions in vivo.

A model consistent with our results is that the state of histone acetylation across the Ifng gene region reflects the balance of histone acetyltransferases (HATs) and HDACs recruited to this region in T cells. Under neutral conditions, the balance is shifted toward HDAC recruitment, resulting in the overall absence of histone hyperacetylation. Inhibition of HDACs by TSA shifts this balance to favor histone hyperacetylation and Ifng transcription under neutral culture conditions. Under Th1 culture conditions, association of Stat4 and T-bet with multiple DNA sites along the Ifng gene region favors recruitment of HATs to the region and shifts the balance to histone hyperacetylation and Ifng transcription. In the absence of Stat4 or T-bet, histone hyperacetylation and Ifng transcription can be partially restored by inhibition of HDACs, indicating that the level of histone hyperacetylation and Ifng transcription is still a reflection of the balance of recruitment of HATs and HDACs to the region in Th1 cells. It is likely that additional transcription factors and additional DNA elements contribute to establishing the level of histone hyperacetylation across this region. Inhibition of HDACs alone is not sufficient to completely restore Ifng transcription to Th1 levels in the absence of T-bet or Stat4, arguing that T-bet and Stat4 contribute to total Ifng transcription in Th1 cells by additional mechanisms.

Because both cellular differentiation and gene transcription occur simultaneously in the Th1 differentiation model, it is not possible to determine whether long-range histone acetylation is associated with differentiation or transcription or both. To address this distinction, we examined histone acetylation across the Ifng gene region in NK cells in the resting state in the absence of active Ifng gene transcription and after IL-12 and/or IL-18 stimulation to induce Ifng transcription. Based on this analysis, we find that histone acetylation can be segregated into several forms (Figs. 5, 6, 7). First, regions very distal to the Ifng gene at -70 to -53, -30 to -23, and +30 to +40 kb, as well as the promoter (-0.4 kb), are histone acetylated in resting NK cells. Second, more proximal regions at -6 kb and the first intron (+0.4 and +1.1 kb) are histone acetylated in response to IL-12/IL-18 signaling. Several subdomains (-70 to -53, -30 to -23, +20, and +30 kb) are histone acetylated to similar degrees in Th1 and resting NK cells. Other domains (-6 kb and the Ifng gene) are histone acetylated to similar degrees in Th1- and IL-12/IL-18-stimulated NK cells. In contrast, histone acetylation at +40 kb is limited to NK cells, and histone acetylation at +50 to +60 kb is limited to Th1 cells. Thus, there are overall similarities in the histone acetylation pattern in Th1 and NK cells. Discrete differences also exist, suggesting that molecular mechanisms that give rise to long-range histone acetylation patterns in Th1 and NK cells may not be identical.

Both -5- to -6-kb and +18- to +19-kb regions exhibit strong DNA sequence conservation and possess transcriptional enhancer activity when linked to reporter genes (29, 30). The -5- to -6-kb region contains consensus T-bet- and NFAT1 (nuclear factor of activated T cells 1)-binding sites and exhibits Th1-specific binding of these transcription factors (29, 30). The +18- to +19-kb enhancer also contains T-bet-binding sites (30). An additional distal enhancer also exists at ≈-3.5 kb from the Ifng transcription start site (murine gene) (31). Similar to what we have reported here, this site is histone hyperacetylated in Th1 but not Th2 cells and in freshly isolated NK cells in the absence of active Ifng gene transcription. The level of histone hyperacetylation at this site is increased after stimulation of NK cells with IL-2. Additional transcriptional enhancers may also exist across this 140-kb region that coincide with regions of DNA sequence conservation and extensive histone hyperacetylation.

These regions may communicate with each other by employing a looping-like mechanism to bring distal regions in close proximity to the Ifng gene (32). Evidence indicates that both the Ifng gene region and the Th2 cytokine gene locus (Il4, Il5, and Il13 genes) exist in specific three-dimensional chromosome conformations in T cells (32, 33). Core conformations of these genomic regions are present in both Th1 and Th2 cells, suggesting that their formation is not dependent on transcription of either Ifng or Th2 cytokine genes. Therefore, it has been argued that formation of these chromatin conformations can provide a general mechanism to deliver distal positive and negative regulatory signals to a given gene or gene family. Additional lineage-specific alterations in conformation may serve, for example, to loop in a specific enhancer element or loop out a specific silencer element in a cell that expresses a given gene. For example, the Ifng gene region exists in a chromatin conformation that is relatively similar in both Th1 and Th2 cells, with one distal region looped out in Th1 cells. We hypothesize that the looped-out region may contain a silencer element that may suppress Ifng gene expression in T cells that do not efficiently express the Ifng gene (32).

The conformation of the Ifng gene region may also contribute to formation of extended patterns of histone acetylation. One model (10) is that local recruitment of HATs by DNA-bound transcription factor complexes may recruit bromodomain-containing HATs that bind acetylated histones independent of transcription factor binding and catalyze histone acetylation at neighboring sites. This process could spread in an iterative fashion to establish a histone acetylated domain. Multiple domains could result if chromatin existed in a preexisting conformation that brought distal genomic regions into close proximity to the initiating HAT-nucleoprotein complex.

A general view is that acetylation of histones in the promoter during active transcription opens the chromatin structure, allowing access to transcriptional and other regulatory factors (34). However, structural evidence argues that posttranslational modifications do not result in alterations of basic chromatin structure (35), which implies that posttranslational histone modifications may serve other functions beyond opening chromatin structure. We hypothesize that the long-range pattern of histone acetylation along the Ifng region in NK cells is established during lineage development and maintained during cell division independent of active Ifng transcription. These studies and others (31, 36) support the view that NK cells have a preexisting pattern of histone and DNA modifications that facilitate rapid gene transcription in response to external stimuli. We hypothesize that the function of preexisting histone modifications is to serve as signposts to direct nuclear communication between the chromatin region surrounding Ifng and nuclear signaling pathways that are required for rapid and efficient transcription in response to appropriate extracellular stimuli. Thus, intrinsic signaling pathways in NK cells must be sufficient to sustain this long-range pattern of histone acetylation. This notion is not entirely new (37). Recent evidence demonstrates that the mouse λ5-VpreB1 locus is marked by histone acetylation and methylation at a discrete intergenic site in embryonic stem cells and spreads toward the VpreB1 and λ5 genes at later stages of B cell development, resulting in a large, active chromatin domain in pre-B cells that coincides with active transcription.

Although our results focus on a single gene, we hypothesize that long-range histone acetylation may represent a general mechanism to mark genes for tissue-specific transcription. This type of mechanism may play a key role in the transcription of cell-specific genes, may enable a cell to rapidly transcribe a given gene in response to external stimuli, or both. Investigation of other genes that exhibit complex cell- and stimulus-specific induction of transcription will be necessary to determine the general applicability of this model.