Histone Deacetylases as Transcriptional Activators? Role Reversal in Inducible Gene Regulation

See allHide authors and affiliations

Science's STKE  09 Aug 2005:
Vol. 2005, Issue 296, pp. re11
DOI: 10.1126/stke.2962005re11


Histone deacetylation enzymes have often been associated with the suppression of eukaryotic gene transcription. In contrast, recent studies of inducible gene regulation indicate that protein deacetylation can also be required as a transcriptional activation signal. The concept of protein deacetylation as a requirement for transcription activation seems to contradict earlier conclusions about the function of deacetylation in gene suppression. However, in the context of a more global interpretation, these opposing effects of deacetylation imply its dynamic role in the overall control of gene expression. The exact requirement for deacetylation differs among promoters, depending on their specific architecture and regulation scenario.

Posttranslational Modifications Regulate Protein Functions

The regulation of protein function frequently relies on the addition or removal of posttranslational modifications. Virtually all biological responses in eukaryotic organisms are controlled by the actions of modified proteins. These modifiers include well-studied examples such as phosphorylation (which regulates enzymatic activity, substrate specificity, and signal transduction) and ubiquitination (which regulates protein stability and subcellular localization). Equally important are modifications such as protein acetylation and protein methylation (which control gene expression, cellular differentiation, and development), fatty acylation (which affects protein localization and enzymatic activity), and glycosylation (the attachment of sugars that typically regulates the structure and function of transmembrane receptors and other cellular proteins). Here, we explore new findings regarding reversible protein acetylation, which occurs on both histones and nonhistone proteins and has been widely studied in the context of genetic and epigenetic control of gene expression.

Reversibility of posttranslational modification is essential for the plasticity of typically transient biological signals. The attachment and removal of these reversible modifications is usually mediated by opposing enzymatic actions, allowing rapid and dynamic responses to both intracellular and extracellular stimuli. For example, kinases and phosphatases add or remove phosphate groups from target proteins. Ubiquitin ligases are opposed by deubiquitinating enzymes, and protein methylation is reversed by demethylase enzymes. Histone acetyltransferase (HAT) activity is counteracted by histone deacetylase (HDAC).

HAT proteins are histone-modifying enzymes that can add acetyl groups to lysine residues in the N-terminal tails of histones to regulate transcriptional output (1). HAT proteins are a diverse group of enzymes that contain an acetyltransferase catalytic domain required for biological activity (2). In many cases, HAT enzymes function within the context of high molecular weight multiprotein complexes (3), and distinct HAT complexes vary with respect to subunit composition and substrate specificity. In addition to histones, HATs also modify nonhistone proteins. For example, p300, CBP, and PCAF—all widely expressed mammalian HATs—acetylate nonhistone substrates, such as the tumor suppressor p53, EKLF (erythroid Kruppel-like factor), GATA1 (GATA-binding protein 1), tubulin, importin-α, and others, resulting in altered ability of these factors to interact with DNA or other proteins (411).

HATs and HDACs

Acetyl groups are removed from HAT-modified proteins by HDAC enzymes. The HDAC family consists of a number of proteins that contain catalytic deacetylase domains. Three classes of HDACs have been described to date in mammalian systems on the basis of sequence similarity to the yeast founding members. Class I HDACs are related to Rpd3, the first described yeast deacetylase protein, and include HDAC1, HDAC2, HDAC3, and HDAC8 (12, 13). Class II HDACs are similar to yeast Hda1 and include HDAC4, HDAC5, HDAC6, HDAC7, HDAC9, and HDAC10. On the basis of differences in regulatory domains that influence subcellular distribution, class II can be further subdivided. In addition, a unique deacetylase, HDAC11, has been recently identified (14). Class III HDACs are related to the yeast Sir2, a nicotinamide adenine dinucleotide–dependent deacetylase. These proteins are conserved from bacteria to mammals, and human homologs include hSIRT1 through hSIRT7.

The class I HDAC proteins are best known for their ability to participate in transcriptional regulation and are often described as gene repressors (15). In mammalian cells, transcriptional repression can be achieved through either passive or active mechanisms (16, 17). Passive repressors function by sequestering essential transcriptional activators or by inertly occupying enhancer binding sites on promoters. By definition, such passive repressors generally do not require enzymatic activity for their inhibitory function. HDAC proteins are considered to be active transcriptional repressors, as they directly remove acetyl groups from histones or other substrates, providing a negative transcriptional signal. Indeed, inactivation of the deacetylase catalytic activity by mutation or pharmacological inhibition can eliminate repression activity (18).

Acetylation and Transcription

The association of some HAT-containing complexes with promoter-specific transcriptional activators and the recruitment of some HDAC-containing complexes to promoters by transcriptional repressors has led some researchers to link acetylation with transcriptional activation and deacetylation with transcriptional repression. Increased acetylation of histones in gene promoter regions has been correlated with increased gene expression (19), and decreased acetylation of histones at promoters has been correlated with HDAC recruitment and transcriptional repression (12). For example, p300 and CBP associate with many mammalian transcription factors, including c-Jun (20), MyoD (21), nuclear hormone receptors (22), and signal transducer and activator of transcription 1 (STAT1) (23), in all cases acting as transcriptional coactivators. In contrast, the recruitment of the histone deacetylase Rpd3 to yeast promoters is associated with decreased occupancy of those promoters by TATA binding protein (TBP), the Swi/Snf nucleosome-remodeling complex, and the SAGA histone acetylase complex, resulting in transcriptional repression (24). In vitro studies have shown that DNA accessibility to restriction enzymes and transcription factors is enhanced in response to histone acetylation, presumably because of an increase in the cumulative negative charge on acetylated histone N termini and repulsion of negatively charged DNA, which may destabilize histone-DNA interactions (25, 26). Acetylation has also been associated with transcriptional activation because it can allow docking of coactivator proteins containing bromodomain (an acetyl-lysine binding motif) to modified histones on target promoters (27). Conversely, removing these modifications could decrease coactivator binding and thus repress transcription. The range of protein-protein interactions that might be enhanced or disrupted by acetylation and deacetylation are almost infinite. Moreover, acetylated lysine residues occupy lysines normally available for alternative modification systems such as ubiquitin, SUMO, or methylation. Increased amounts of mRNA accumulate in cells subjected to prolonged (longer than 18 hours) treatment with pharmacological HDAC inhibitors. Deacetylase inhibitors have been recognized as potential cancer therapeutics because they induce cell cycle arrest and apoptosis, effects generally attributed to increased expression of certain proapoptotic or cell cycle–mediating genes (28, 29). Other examples of mRNA accumulation in response to HDAC inhibitor treatments include activation of NFκB-dependent genes (30, 31), p53-independent expression of the cyclin-dependent kinase inhibitor p21WAF1 (also called CIP1) (28, 3234), and increased expression of growth-differentiation factor 11 (Gdf11) (35), among others.

Role Reversal for HATs and HDACs

These examples substantiate the model associating HATs with transcriptional activation and HDACs with transcriptional repression. However, many studies indicate that gene regulation by acetylation is more dynamic and complex, and that HATs can also act as repressors and HDACs can also function as activators. For example, p300 and CBP both contain cell cycle regulatory domain 1 (CRD1), a transcriptional repressor domain that may allow these HATs to act as transcriptional repressors (36). The removal of yeast repressor proteins Rpd3, Sin3, or Hda1 results in decreased transcription of a number of genes and an increase rather than a loss of silencing, pointing to an alternative role of HDACs as transcriptional activators (3739). Furthermore, SIN3, a major component of the HDAC-containing transcriptional repressor complex Sin3p, functions as a transcriptional corepressor for the human progesterone receptor, but is also a transcriptional coactivator for the GAL4 and HAP1 transcription factors (40). Another class I deacetylase from yeast, Hos2, preferentially binds to transcriptionally active promoters and deacetylates histones H3 and H4 (41). Also, Rpd3 has been shown to be required for transcriptional activation of yeast osmoresponsive promoters in response to osmotic stress (42).

The importance of HDACs as activators of gene expression is underscored by gene expression profiling of cells treated with HDAC inhibitors. Only a small percentage of the total genes examined responded to the treatment (about 7% of mRNAs), with similar proportions of genes activated or repressed (43, 44). SRC promoter repression by HDAC inhibition is one recent example of the negative action of HDAC inhibitors (45). Detailed studies of the interferon (IFN)–β promoter have shown that acetylation and deacetylation of histones are both essential for enhanceosome function (46). Specific acetylated lysine residues of histones H3 and H4 are rapidly deacetylated before chromatin remodeling, recruitment of transcription-activating complexes, and transcription initiation. It seems reasonable that temporally regulated transient acetylation and deacetylation results in precise fine-tuning of the chromatin template to provide adhesive surfaces for the binding of additional transcriptional regulators required to maintain signal integrity (47, 48). If the proposed interrelationship between acetylation and other posttranslational modifications that play a role in transcription regulation (such as methylation or ubiquitination) is correct, the addition or removal of acetyl groups from substrate lysine residues would set the stage for downstream signaling events, potentially resulting in diverse influences on protein activity or chromatin accessibility.

Deacetylation in Cytokine-Inducible Gene Regulation

The transcription-activating role of deacetylation is perhaps best exemplified by studies of cytokine-inducible gene regulation. Recent reports indicate that stimulation of gene expression downstream of IFN (either IFN-α/β or IFN-γ, cytokines responsible for immune regulation and antiviral responses) fails after general blockade of HDAC activity (4952).

Normally, IFN binding to its transmembrane receptor rapidly induces transcription through activation of the Janus kinase (JAK)–STAT signal transduction system (53). IFN-α and IFN-β trigger the tyrosine phosphorylation of latent STAT1 and STAT2 proteins. The activated STATs are held together by intermolecular Src homology 2 (SH2)–phosphotyrosine interactions, and, together with the IFN regulatory factor 9 (IRF9) DNA binding subunit, form the transcription factor complex known as IFN-stimulated gene factor 3 (ISGF3). ISGF3 translocates to the nucleus, where it binds to IFN-stimulated gene (ISG) promoters and recruits coactivators required to initiate mRNA synthesis. ISGF3 can associate with several coactivators, including GCN5, CBP, p300, the Brg1 adenosine triphosphatase, and the metazoan Mediator complex (5457). IFN target genes are generally not actively transcribed in cells not exposed to IFN, but ISGF3 initiates a great increase in the ISG transcription rate that results in rapid and transient accumulation of mRNA. The biological outcome of ISGF3 signaling in cells is the expression of genes involved in innate antiviral responses, creating a cellular state that is resistant to virus infection (58).

Because ISGF3 is a potent activator of IFN responses, and because HDACs are associated with suppression of transcription, one might predict that ISG mRNAs would strongly accumulate in cells treated simultaneously with IFN and HDAC inhibitors. In contrast, this treatment regimen completely prevents or greatly delays IFN-induced gene transcription. HDAC inhibitors potently block accumulation of all ISG mRNAs tested, both in individual analysis and in genome-wide microarray analysis, indicating that HDAC activity is generally required for ISG expression (4951). An important consequence of this suppression of ISG by HDAC inhibitors is the prevention of the biological response to IFN. Simultaneous treatment of cells with IFN and HDAC inhibitors prevents IFN from establishing the cellular antiviral state, enabling viruses to replicate unimpeded. Analysis of the signaling pathways upstream of promoter activation provided no evidence for an acetylated intermediate in the IFN-α signaling pathway. The inhibition of ISG expression was not caused by a defect in tyrosine phosphorylation of STAT proteins, or in assembly, nuclear translocation, or DNA binding by ISGF3. Instead, the HDAC-sensitive step was found to lie between promoter occupation by ISGF3 and recruitment of RNA polymerase II.

The requirement for deacetylase activity in the IFN system is clear, but the exact mechanism underlying this effect is yet to be completely resolved by identification of the required deacetylation enzymes and substrates. One plausible explanation for the data is that ISGF3 recruits an HDAC required to deacetylate proteins present on the target promoter before RNA polymerase II recruitment. This model is supported by the observation of localized deacetylation of histone H4 at the ISG54 promoter after treatment of cells with IFN-α, an effect prevented by HDAC inhibitors. Screening of candidate HDACs revealed efficient interaction between HDAC1 and both STAT1 and STAT2. Tests for coactivator activity supported a role for HDAC1 in IFN-responsive transcription. Altering HDAC1 levels in cells produced responses in IFN signaling consistent with a coactivation function. Reduction of HDAC1 expression by RNA interference resulted in decreased IFN-α responsiveness, whereas increased expression of HDAC1 augmented ISGF3 activity. Although the available evidence indicates a role for HDAC1 as an ISGF3 transcriptional coactivator, the effects of inhibiting HDAC1 activity specifically are not as great as the effects of the general chemical HDAC inhibition. Although this difference might simply reflect technical concerns such as inefficient delivery of small interfering RNA into the cells, it remains possible that additional deacetylase proteins besides HDAC1 are required for ISGF3 activity and the totality of IFN biological responses. Such flexibility or redundancy in HDAC recruitment to promoters could be indicative of another level of epigenetic specificity conferred by an "HDAC code."

The finding that HDACs function as coactivators in IFN-induced ISGF3 signaling is consistent with reports of other cytokine systems that require HDAC activity. The IFN-γ pathway, a similar JAK-STAT signaling system that activates STAT1 homodimers to induce transcription of IFN-γ–activated genes, is also inhibited in the absence of deacetylase activity (49, 52, 59). Furthermore, a similar requirement for deacetylation is observed for interleukin-2– and interleukin-3–responsive genes that are expressed in response to STAT5 signaling (60). The gene activation capability of HDACs is also relevant for the increased expression of Bcl2 and control of apoptosis in B cell lymphomas (61).

How can these findings be reconciled with the abundant and credible evidence from many systems that HDACs function as transcriptional corepressors? One attractive hypothesis is that HATs and HDACs act together to establish a balance of power on activated promoters, ultimately resulting in rapid transcriptional initiation in response to a stimulus. This idea that HATs and HDACs may act in concert to activate gene expression is supported by the finding that these enzymes can exist in close spatial proximity in the cell, their interaction being essential for maintaining the delicate balance of acetylation patterns to regulate not only transcription but a variety of other cellular processes (62). Clearly, both HATs and HDACs are required for the complete gene expression profile from inducible promoters. Removal of either one of these essential components would therefore result in the similar outcome of failed transcription.

Balance of Powers: HATs and HDACs on Activated Promoters

Given the importance of HDACs in both activation and repression of genes, it seems peculiar that microarray studies found only a small subset of promoters to be sensitive to HDAC inhibitor treatments. One answer may be that cumulative HDAC inhibition affects multiple deacetylases, each of which could have distinct roles in gene activation or repression as targets of specific signaling pathways. Further delineation of inherent differences in promoter architecture and distinct signal transduction mechanisms leading to promoter activation could provide refinement of the HDAC requirements, which can vary on a case-by-case basis. Analysis of individual genes underscores these promoter-specific differences, as some mRNAs accumulate in response to prolonged HDAC inhibitor treatments, whereas others remain unchanged or exhibit decreased expression under similar conditions (32, 61). However, brief HDAC inhibitor treatments can prevent activation of inducible transcription, even though more prolonged treatment of cells with HDAC inhibitors appears to increase steady-state accumulation of mRNA.

Promoters Respond Differently to HDAC Inhibitor Treatments

To reconcile these apparently paradoxical effects, we posit that extended HDAC inhibition in the cell might derepress promoters being held inactive by the action of HDACs in repression complexes, thereby enabling these promoters to be available for activation. This mechanism would differ for the cytokine-inducible promoters, where passive basal inactivity in the absence of a stimulus such as IFN is transformed into rapid activation by the action of transcription factors and their recruited coactivator complexes, an essential component of which are HATs and HDACs. Several cases of promoter regulation differentially influenced by HDAC inhibitors are illustrated in Fig. 1. At the steady state, genes could be basally active and unaffected by HDACs (and, in turn, HDAC inhibitors). Alternatively, genes can be silenced by either non-HDAC repressors or HDAC-containing repressors. HDAC inhibition will only cause mRNA accumulation from the latter category, and only if the cellular milieu contains the requisite activation machinery to induce promoter activation. If not, the gene will be alleviated of HDAC-dependent repression but will remain silent until the appropriate activators are present.

Fig. 1.

Effects of HDAC inhibitors on gene expression. (A) Comparison of the effects of HDACs on basal, repressed, and inducible promoters. Basally expressed genes that are not regulated by HDACs are insensitive to HDAC inhibitor treatments. Genes silenced by non-HDAC repressors are insensitive to HDAC inhibitors (HDIs). HDAC-repressed genes are relieved of repression by HDAC inhibitors, possibly resulting in mRNA accumulation. Inducible promoters that are simply "off" would not be expected to change, but those "repressed" by HDAC complexes will be derepressed, possibly resulting in mRNA accumulation in the absence of stimulus. These same genes are activated in response to stimulus, but if other HDACs are part of the activating machinery, gene expression may fail in the presence of stimulus plus HDAC inhibitors. (B) Secondary effects of HDAC inhibitors in derepression. HDAC inhibitors can derepress genes that encode a transcription activator. As this activator accumulates, it can activate expression of its normal target promoters, resulting in secondary activation. (C) Secondary effects of HDAC inhibitors in activated transcription. An inducible gene encoding a repressor is inhibited by HDAC inhibitors, allowing secondary derepression of the downstream target gene.

For inducible systems such as IFN-stimulated genes, in the absence of stimulus an individual promoter could be simply silent (because of a lack of activation signal) or could be actively repressed by HDACs. Although expression of nonrepressed genes would be unchanged in response to HDAC inhibitors, effects on the repressed gene could result in mRNA accumulation, provided that appropriate activators are available. In the presence of inducer, the genes become switched on, responding to transcription factors together with HATs and HDACs. However, blocking of the activating HDACs would result in a failure to induce transcription.

Alternative regulatory phenomena could also produce secondary effects downstream of HDAC inhibition (Fig. 1, B and C). If the product of an HDAC-repressed gene encodes a transcriptional activator and accumulates in response to HDAC inhibitors, the activator protein could become available to bind to its own target promoters, potentially inducing their transcription. Conversely, HDAC inhibition may result in the decreased production of a transcriptional repressor, relieving the repressor’s target gene promoter of its negative action (32). In these scenarios, the target promoters could be erroneously identified as being sensitive to HDAC inhibitors, but the direct nature of their regulation could easily be tested with the use of protein synthesis inhibitors such as cycloheximide. These indirect effects have been ruled out for ISG induction by IFN because the HDAC dependence was also observed in the presence of cycloheximide (49).

The field of histone and protein acetylation is rapidly evolving, and the importance of HATs and HDACs in the regulation of numerous cellular responses is becoming more apparent. Although all the roles of acetylation in cellular regulation are yet to be uncovered, the unexpected function of HDACs as transcriptional activators in inducible transcription is captivating, as it will assuredly prove to be a fundamental mechanism for controlling inducible gene regulation.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
View Abstract

Navigate This Article