Inducible Covalent Posttranslational Modification of Histone H3

See allHide authors and affiliations

Science's STKE  26 Apr 2005:
Vol. 2005, Issue 281, pp. re4
DOI: 10.1126/stke.2812005re4


The physiological state of a eukaryotic cell is determined by endogenous and exogenous signals, and often the endpoint of the pathways that transmit these signals is DNA. DNA is organized into chromatin, a nucleoprotein complex, which not only facilitates the packaging of DNA within the nucleus but also serves as an important factor in the regulation of gene function. The nucleosome is the basic unit of chromatin and generally consists of approximately two turns of DNA wrapped around an octamer of core histone proteins. Each histone also contains an accessible N-terminal tail that extends outside the chromatin complex and is subject to posttranslational modifications that are crucial in the regulation of gene expression. Two distinct categories of histone posttranslational modification have been observed: (i) inducible or stimulation-dependent and (ii) mitosis-dependent. Stimulation by mitogens or stress leads to rapid transient posttranslational modifications of histones, in particular histone H3, which are mechanistically and temporarily distinct from modifications associated with mitosis. This Review focuses mainly on the inducible phosphorylation of histone H3 brought about by different stimuli, such as epidermal growth factor, 12-O-tetradecanoylphorbol-13-acetate, arsenite, or ultraviolet radiation. We examine the most recent, and at times controversial, research data concerning the identity of the histone H3 kinases responsible for this phosphorylation. In addition, the interdependence of phosphorylation and acetylation will be discussed in light of data showing patterns of inducible modification at specific genes.

General Description of Chromatin and Histone H3

Eukaryote genomes contain multiple chromosomes, and each chromosome contains a linear molecule of DNA. The DNA is tightly bound to small basic proteins called histones that package the DNA into an orderly complex referred to as chromatin. The small histone proteins contain a high proportion of the positively charged basic amino acids arginine and lysine, which increases their binding affinity to the negatively charged DNA. The histone proteins are found only in eukaryotic cells, and their total mass is about equal to the total amount of cellular DNA. Five major types of histone proteins exist and include H1, H2A, H2B, H3, and H4 (1). Histones H2A, H2B, H3, and H4 are referred to as the core histones because they wrap the DNA to form the nucleosome core of chromatin (2, 3). The core histone sequences are the most highly conserved between species. Variants of each histone class also exist and may be associated with differences in chromatin structure that appear to be tissue-specific (4).

The nucleosome is the basic unit of chromatin and consists of approximately two turns of DNA wrapped around an octamer of core histone proteins (Fig. 1) (5, 6). Each octamer is composed of one molecule each of H2A and H2B, forming two dimers on each side of a tetramer consisting of two molecules each of H3 and H4 (7). Histone H1 completes the package and is bound to the DNA as it enters and exits each nucleosome. Nucleosomes are connected to each other by 20 to 60 base pairs (bp) of linker DNA (8).

Fig. 1.

Posttranslational modifications of the histone H3 tail. Ac, acetylation site; Me, methylation site; P; phosphorylation site.

Each histone also contains an unstructured N-terminal tail domain ranging from 16 (in H2A) to 44 (in H3) amino acids in length, which extends out from the nucleosome core. Histone tails are essential for the higher-order folding of chromatin fibers and provide binding sites for nonhistone regulatory proteins (9). They are the target of extensive covalent posttranslational modifications that include the acetylation, methylation, and ubiquitination of specific lysine amino acids; the phosphorylation of serines and threonines; and poly(ADP-ribosylation) (ADP, adenosine diphosphate) (1016). Acetylation and methylation of different lysine and arginine residues in histones H3 and H4 are linked to either transcriptionally active or transcriptionally repressed states of gene expression (17). Initially, phosphorylation of H3 was linked to chromosome condensation that occurs with mitosis (18, 19). These modifications have a major influence on chromatin structure by affecting the local environment (for example, by altering charge, which allows decondensation), facilitating the binding of transcription factors that regulate gene expression, or allowing the interaction of various chromatin-remodeling factors (2022). Chromatin is often described by its state of condensation. Condensed chromatin is referred to as heterochromatin and is generally not transcriptionally active, whereas euchromatin is extended and is more accessible to transcriptional regulators. Genes that are being actively transcribed are found in the more accessible decondensed chromatin, and alteration in the organization of chromatin is believed to give the transcriptional machinery access to genes.

Histones are now clearly regarded as integral and dynamic components of the machinery that is responsible for regulating gene transcription (2326). The same is true for other DNA-related processes, such as replication, repair, recombination, and chromosome segregation (27, 28). Many types of cancer and other diseases are associated with translocations or mutations in genes encoding chromatin-modifying enzymes and regulatory proteins (2933). For example, Coffin-Lowry syndrome (CLS) (34, 35) is associated with mutations in the ribosomal S6-kinase 2 (RSK2), which is a kinase that phosphorylates histone H3 on Ser10. Furthermore, increased histone kinase activity of aurora B has been associated with colorectal cancer (36).

Brief overview of histone H3 acetylation

Acetylation, the most studied posttranslational modification of histone H3, was very early suggested to be related to increased transcriptional activity (37). Acetylation occurs on specific lysine residues on the H3 tail (Fig. 1) and has been proposed as the hallmark of active chromatin (38, 39). Acetylation of histone H3 is believed to decondense the chromatin structure, allowing access to DNA by transcription factors (22). Indeed, acetylation of histones at specific gene loci correlates with transcriptional activity (40). The acetyl groups can be removed from histones by histone deacetylases (HDACs). Deacetylation is associated with the repression of gene expression. Deacetylation presumably restores the positive charge to the specific lysine residue, thereby increasing the interaction of histones with negatively charged DNA. Overall or global acetylation levels of histones result from a balance between histone acetylation and deacetylation catalyzed by acetyltransferases (HATs) and HDACs. However, chromatin is not acetylated evenly. Hyperacetylation, a higher concentration of acetylated histones as compared to the concentration over all the total chromatin, has been observed with decondensed and transcriptionally active euchromatin, whereas the more condensed heterochromatin has been observed to be hypoacetylated and transcriptionally inactive (17, 21). In particular, the formation of heterochromatin and gene silencing have been associated with hypoacetylation of histones H3 and H4 (41).

Abnormal acetylation or deacetylation is associated with cancer development. For example, chromatin inactivated by deacetylation of histone H3, H4, or both is associated with the silencing of several cancer-related genes, including CDKN2A, COX2, BRCA1, CDKN1C, DAPK, CHFR, and CDH1 (4248). Acetylation and deacetylation of histones H3 and H4 may contribute to MYO18B gene expression and silencing in lung carcinogenesis (49). MYO18B is believed to function as a tumor-suppressor gene and is inactivated in about 50% of human lung cancers by various mechanisms, including promoter methylation (49, 50). In addition, Tani et al. (49) also reported that six of seven cell lines with reduced or silenced MYO18B expression exhibited hypoacetylated histone H3 and H4 in the MYO18B promoter region, suggesting another mechanism for MYO18B inactivation.

Brief overview of histone H3 methylation

Histone methylation was discovered over 40 years ago (51), and until recently (52), unlike the more dynamic modifications of acetylation and phosphorylation, was considered to be a stable, more permanent modification. Both lysine and arginine residues can be methylated. Lysine residues can be both acetylated and methylated (Fig. 1), and dual modification has been observed in vivo (5355). Mono-, di-, and tri-methylation have also been observed (56).

Methylation of Lys4 of histone H3 has been linked to active transcription (5760), whereas methylation of Lys9 of H3 has been primarily associated with condensed chromatin and gene silencing (61), including abnormal silencing in some cancer cell lines (62). Several methyltransferases have been described, but only recently has the first histone demethylase, lysine-specific demethylase 1 (LSD1), been described (52). LSD1 is a homolog of amine oxidases and was discovered to have a role in the repression of transcription in vivo, apparently by functioning as a histone H3 Lys4 demethylase (52). LSD1 demethylase activity appears to be specific for mono- or di-methylated H3 (Lys4), with no activity against tri-methylated H3 (Lys4) or H3 (Lys9, Lys36, or Lys79). The discovery of a specific demethylase has implications for discerning yet another level of complexity for the dynamic regulation of histones and gene transcription. Indeed, the existence of demethylases changes the whole perspective regarding the dynamic nature of methylation in vivo (55).

Compared to lysine methylation, less has been reported regarding histone arginine methylation. However, in vivo methylation of Arg17 of histone H3 was recently demonstrated and suggested to be an activating step in mammalian gene transcription (63). Methylation of Arg17 of histone H3 was increased in breast cancer MCF7 cells when the estrogen receptor–regulated pS2 gene was activated by treatment of the cells with 12-O-tetradecanoylphorbol-13-acetate (TPA) and estradiol, which synergistically activate pS2 mRNA expression, suggesting that methylation is inducible. Methylation coincided with the presence of the CARM1 methyltransferase at the pS2 promoter (63).

Inducible Versus Mitotic Posttranslational Phosphorylation

Chromatin is subjected to an array of posttranslational modifications, and some of these modifications are associated with chromosome condensation that occurs with mitosis and meiosis. Alternatively, inducible posttranslational modifications of chromatin can result when a cell is stimulated by stress, such as exposure to ultraviolet (UV) radiation, to the mitochondrial poison arsenite, or to mitogens, such as growth factors and cytokines (Fig. 2). Phosphorylation of histone H3 is one of the posttranslational modifications associated with both mitosis and meiosis and with the response to stress or mitogenic stimuli. Despite targeted attempts to identify the H3 kinases responsible, much debate continues to be centered on which kinase is the most important in mediating inducible histone H3 phosphorylation (35, 64, 65).

Fig. 2.

General schematic of mitosis-associated and inducible phosphorylation. During mitosis, there is global phosphorylation of H3 on Ser10 and Ser28 (S10 and S28). Stress or mitogens trigger partial phosphorylation of S10 and S28, which allows the expression of IEGs.

Histone H3 phosphorylation (Fig. 1) is clearly known to play important regulatory signaling roles in the apparently contradictory processes of chromosome condensation during mitosis and the chromatin extension involved in transcription after external stimulation of gene expression (53, 6670). During mitosis, chromatin is condensed and each Ser10 of every histone H3 is phosphorylated (19, 67, 71), and this phosphorylation state has traditionally been regarded as a marker of mitosis (19, 72). Wei et al. (28) later confirmed that phosphorylation of H3 at Ser10 is tightly linked to chromosome condensation and segregation during mitosis. Stimulation of mammalian cells with growth factors also leads to phosphorylation of histone H3 on Ser10; however, this phosphorylation occurs rapidly after stimulation, is transient, and does not include every histone H3, which is distinctly different from the Ser10 phosphorylation that is associated with the chromosome condensation observed during mitosis (73).

In 1991, Mahadevan et al. (74) first described the "nucleosome response" as a rapid phosphorylation of histone H3 that coincided with the activation of two immediate-early genes (IEGs), c-fos and c-jun. These and other IEGs are rapidly and transiently expressed in response to extracellular stimuli. The expression of IEGs occurs in response to the activation of cell-surface receptors that activate cellular signaling pathways and include genes, which encode proteins that trigger a cell to leave the G0 phase of the cell cycle and enter G1. No new protein synthesis is required for the expression of these genes, and their activation is dependent only on modifications of molecules already present within the cell. The IEG response has been implicated in differentiation, cell division, and diseases, such as inflammation and cancer (75, 76). Phosphorylation of a number of transcription factors and histone H3 is intricately associated with IEG induction (74, 7779). In the classic study by Mahadevan et al. (74), resting C3H 10T1/2 mouse fibroblasts were treated with epidermal growth factor (EGF), TPA, the phosphatase inhibitor okadaic acid, the protein synthesis inhibitor cycloheximide or anisomycin, or combinations of the compounds to elicit early chromatin-associated signaling events. The results confirmed the link between the phosphorylation of H3 and transcriptional activation of the IEGs c-fos and c-jun (74). Since that report, other activator protein–1 (AP-1) family genes, namely fosB, junB, and junD, have also been shown to be differentially induced (80), as well as the oncogene c-myc (81, 82). Additional data verified that certain stimuli trigger a rapid and transient phosphorylation of histone H3 at Ser10 that targets only a small population of histone H3 molecules, which is distinct from the histone H3 phosphorylation that is observed in mitosis (73, 79, 81, 83, 84).

Although less well characterized than phosphorylation of Ser10, phosphorylation of histone H3 Ser28 is also induced by stress, such as UV irradiation (8587). Mitosis-specific phosphorylation of H3 also occurs at Ser28 (88), Thr11 (89), and Thr3 (9092). The kinase Aurora B phosphorylates Ser10 and Ser28 (67, 9395), Dlk (death-associated protein-like kinase, also known as ZIP) phosphorylates Thr11 (89), and an as-yet-uncharacterized kinase is responsible for phosphorylating Thr3 (9092). However, whether these phosphorylation events are causally linked to each other is not known and whether phosphorylation of Thr3 and Thr11 is inducible is also not known. The histone H3 tail also has three other potential threonine phosphorylation sites, Thr6, Thr22, and Thr32, but as yet no reports indicate that these sites are phosphorylated.

Several kinases have been reported to be involved in the phosphorylation of histone H3 at Ser10 after mitogenic stimulation or stress (Table 1). These kinases include the mitogen-stimulated protein kinases (MAPKs), extracellular signal–regulated protein kinases (ERKs), p38 kinase, and c-Jun N-terminal kinases (JNKs) (73, 9698); p90 RSK2 (35, 99, 100); mitogen- and-stress–induced kinases 1 and 2 (MSK1/2) (65); cAMP-dependent protein kinase (PKA) (84); IκB kinase α (IKKα) (101103); protein kinase C (PKC) (104); mixed-lineage triple kinase alpha (MLTK-α) (105); p21-activated kinase-1 (PAK-1) (106); and Fyn, a Src family tyrosine protein kinase (107). Phosphorylation of H3 at Ser28 also appears to be inducible and mediated by MAPKs, MSK1/2, or MLTK-α (85, 86, 105) (Table 2). Thus, many kinases appear to be involved in histone H3 phosphorylation in response to external stimuli, and it is likely that more may remain to be discovered. Not surprisingly, the response may be stimulus- and cell type–dependent.

Table 1. Kinases involved in H3 phosphorylation of Ser10. Those marked with an asterisk also exhibit kinase activity toward H3 in an in vitro kinase assay. DN, dominant-negative; MEFs, mouse embryo fibroblasts.

Table 2. Kinases involved in H3 phosphorylation of Ser28. Those marked with an asterisk also exhibit kinase activity toward H3 in an in vitro kinase assay.

Inducible H3 Kinases


The involvement of the MAPK cascades, which are activated by a multitude of extracellular stimuli, in H3 phosphorylation is well characterized. MAPK cascades are crucial in the regulation of numerous cellular functions, including proliferation, differentiation, and apoptosis. The ERKs, JNKs, and the p38 kinases are the three most thoroughly characterized subgroups of the MAPK family. All three groups of kinases are activated by dual phosphorylation on threonine and tyrosine at T-X-Y motifs within the activation loop. One or more MAPK kinases (MAPKKs) catalyze this phosphorylation, and they are in turn activated by MAPKK kinases (MAPKKKs). Activated MAPKs can translocate to the nucleus, where they phosphorylate target molecules, including various transcription factors and histone H3. Although not mutually exclusive in function, ERKs are activated and play a critical role in transmitting signals initiated mainly by growth factors, whereas various forms of stress or inflammatory signals activate JNKs and p38 kinases. Constitutive activation of MAPK-related oncogenes, such as H-ras, increases the level of phosphorylated H3 in mouse fibroblasts (81, 83). Several reports have confirmed that the MAPK cascades are key players in the nucleosomal response resulting in the phosphorylation of Ser10 of histone H3 and the activation of IEGs (35, 53, 54, 73, 81, 108, 109). As is well known, the IE cellular response is induced by growth factors, cytokines, stress, and pharmacological compounds, which all differentially activate MAPK pathways.

EGF, a growth factor, and TPA, a tumor promoter, were first reported to induce H3 phosphorylation at Ser10 (35, 53) and at Ser10 and Ser28 (65), respectively, through the ERK pathway. The DNA alkylating agent cisplatin also induced phosphorylation of histone H3 at Ser10 and increased the acetylation of histone H4 in HeLa cells through a pathway involving p38 kinase (110). Cisplatin also weakly induced phosphorylation at histone H3 Ser28. Cisplatin treatment caused similar phosphorylation effects in MCF-7 breast cancer cells and Ntera II testicular cancer cells (110). Thus, diverse stimuli, such as growth factors, phorbol esters, and DNA damage, may induce similar histone H3 phosphorylation in various cell types but with a resultant induction of distinct genes and cellular responses.

Arsenite, which is a carcinogen and a tumor promoter, is also a potent inducer of H3 phosphorylation mediated through the MAPK pathways involving ERKs and p38 kinase (100, 109). Both c-fos and c-jun expression are induced rapidly in response to arsenite (111113). In WI-38 human diploid fibroblasts, arsenite dramatically induced the phosphorylation of histone H3 at Ser10 and the acetylation of histone H3 at Lys14 (114). These modifications coincided with increased c-fos and c-jun expression that was mediated through an ERK-dependent pathway (114). Similar to earlier studies, the phosphorylation and acetylation appeared to be confined to a subset of H3 molecules. No phosphorylation of H1 or H2A was observed, indicating a specific effect of arsenite on the induction of histone H3 phosphorylation. The kinetics of H3 phosphorylation and MSK1 activation were similar, suggesting the possibility that MSK1 may be involved in altering the posttranslational modifications associated with histone H3 after arsenite treatment (114). Overexpression of a N- or C-terminal mutant MSK1 or exposure of the cells to H89 (N-[2-(p-bromocinnamylamino)-ethyl]-5-isoquinolinesulfonamide), which is a protein kinase inhibitor known to inhibit MSK1 (among other kinases), had no effect on arsenite-induced phosphorylation of histone H3 at Ser10 in JB6 C141 cells. However, arsenite-induced histone H3 phosphorylation in cells deficient in RSK2 was absent. These results suggested that RSK2, but not MSK1, contributed to arsenite-induced phosphorylation of histone H3 at Ser10 (100). In the case of arsenite-induced kinase activation, there are contradictory results as well, with Li et al. (114) reporting that MSK1 was involved. However, Li et al. (114) used a high concentration of arsenite as compared to that used by He et al. (100). Thus, arsenic, like mitogenic stimuli, appears to induce the phosphorylation of histone H3, which is associated with the IEG response mediated by MAPK pathways.

UV is another potent inducer of phosphorylation of both Ser10 and Ser28 of histone H3 (86, 97). All three spectra of UV light—UVA, UVB, and UVC—can differentially induce the three MAPK pathways (115). Using various MAPK inhibitors, dominant-negative mutants, and knockout cells, we have shown that both ERKs and p38 kinase, but not JNKs, mediate UVB-induced phosphorylation of histone H3 at Ser10 in JB6 Cl41 mouse epidermal skin cells (97). In contrast, all three MAPK cascades can, to varying degrees, lead to the phosphorylation of histone H3 at Ser28 in response to UVB (86). Active ERK2 or p38 kinase directly phosphorylated histone H3 at Ser10 in vitro (97), and ERK1 and -2 (ERK1/2), p38 kinase, or JNK1 and -2 phosphorylated histone H3 at Ser28 in vitro (86). Thus, multiple MAPK pathways play key roles in the phosphorylation of histone H3 at Ser10 and Ser28 after stimulation by stressful and nonstressful stimuli.


We reported previously that MLTK-α is a kinase that can regulate the activity of oncogenic transcription factors such as ATF2, c-Jun, Elk-1, and c-Myc through MAPK signal cascades, and thereby plays a key role in neoplastic cell transformation and cancer development (116, 117). The mixed-lineage kinases (MLKs) are a family of serine/threonine protein kinases that function in a phosphorelay module to control the activity of specific ERK1/2, p38 kinase, JNKs, and ERK5 (118). On the basis of domain arrangement and sequence similarity within their catalytic domains, the MLKs are categorized into three subgroups, including the MLKs, the dual-leucine-zipper–bearing kinases (DLKs), and the zipper sterile-α–motif kinase (ZAK) (118). MLTK-α has two putative nuclear export signal (NES) sequences and accumulates in the nucleus when cells are treated with leptomycin, an inhibitor of the NES receptor (119). UVB or EGF induced MLTK-α accumulation in the nucleus, which suggests that UVB or EGF influences the translocation of MLTK-α. When overexpressed in cells, MLTK-α activated four MAPK kinase pathways: ERK, JNK, p38, and ERK5 (119). Overexpression of MLTK-α in JB6 Cl41 cells enhanced the phosphorylation of histone H3 at Ser28, but not Ser10, induced by UVB or EGF and, moreover, overcame the general inhibition of the MAPK pathway by the MAPK inhibitors PD 98059 (MAPKK inhibitor) and SB 202190 (p38 kinase inhibitor) (105). MLTK-β could not phosphorylate histone H3, suggesting that the difference in kinase activity resulted mainly from two domains, the SAM domain (residues 337 to 408), which is only present in MLTK-α, and the serine/threonine kinase domain (residues 16 to 277).

Although RSK2 (35) and MSK1 (65, 73, 120) phosphorylate histone H3 at Ser10 in response to EGF or TPA, UVB- or EGF-induced phosphorylation of histone H3 at Ser28 was blocked in the MLTK-α knockdown cells as compared with control groups, suggesting that neither MSK1 nor RSK2 was involved in histone H3 Ser28 phosphorylation under these conditions (105).

RSK2 and MSK1/2

Besides the direct involvement of MAPKs in the phosphorylation of histone H3, RSK2 (Tables 1 and 2) and MSK1/2 are also candidates for mediating EGF-induced H3 phosphorylation (35, 65). In addition to the phosphorylation of histones, growth factor–stimulated expression of IEGs also involves the activation of various transcription factors, including CREB [cyclic adenosine monophosphate (AMP) responsive element–binding protein], Elk-1, ATF1, and serum response factor (SRF). MSK1/2 and RSK2 have both been implicated in the stimulation of IEG expression and in the phosphorylation of both histones and transcription factors involved in IEG stimulation; however, their exact roles remain controversial and have been the subject of much debate (64).

CLS patients exhibit deletion, nonsense, or missense mutations in the rsk2 gene (34). RSK2 is a member of the p90RSK family of broadly expressed serine/threonine kinases that are activated in response to many growth factors, peptide hormones, and neurotransmitters, and they are activated by phosphorylation (121, 122). RSK2 is activated by ERK1/2 and protein-dependent kinase 1 (123, 124). Although RSK2 is required for the stimulation of expression of c-fos and activation of the transcription factors Elk-1 and the serum response factor by EGF, it is not required for the activation of CREB in response to stimulation by platelet-derived growth factor (PDGF) or insulin-like growth factor (IGF) (125). In normal cells in response to EGF, RSK2 phosphorylates Ser133 of CREB (126) and phosphorylates Ser10 of histone H3 (35). De Cesare et al. (127) reported that fibroblasts from CLS patients demonstrate a marked decrease in Ser133 phosphorylation of CREB in response to EGF treatment. Sassone-Corsi et al. (35) reported that EGF-induced phosphorylation of histone H3 was absent in RSK2-deficient cells from CLS patients but that mitotic phosphorylation of condensed chromosomes appeared normal, suggesting that histone H3 phosphorylation is linked to RSK2-dependent and -independent pathways, depending on the state of the cell (growth factor–stimulated or mitotic). The reintroduction of RSK2 restored the EGF response of histone H3 phosphorylation in CLS cells, and disruption of the rsk2 gene in mouse stem cells abolished the EGF-stimulated H3 phosphorylation. Thus, RSK2 appears to be required for EGF-stimulated phosphorylation of H3 in vivo but is distinct from mitotic phosphorylation, which was readily detectable in both wild-type and RSK2-deficient mouse and human cells (35).

MSK1 was also reported to function as an H3 kinase (Table 1) at Ser10 (73, 120). In comparing MSK1/2 wild-type and knockout cells treated with TPA or anisomycin, Soloaga et al. (65) observed that phosphorylation of histone H3 at Ser10 and Ser28 and HMG-4 (also known as HMG-N1) at Ser6 was deficient, even though RSK2 activity was previously reported to be normal in these cells (128). This apparently contradicts the work of Sassone-Corsi et al. (35) by presenting evidence that MSK1/2, but not RSK2, is essential for histone H3 phosphorylation, which led to a debate as to which is the most important kinase for histone H3 that regulates the AP-1 family of genes (98). Knockout of MSK2 was more effective in blocking the H3 phosphorylation as compared with MSK1. However, as with the RSK2-deficient cells, mitotic phosphorylation of H3 at Ser10 was normal in both wild-type and MSK1/2 knockout cells. In apparent contradiction to Sassone-Corsi (35), Soloaga et al. (65) reported that histone H3 Ser10 phosphorylation in CLS cells lacking RSK2 was normal after exposure of the cells to TPA or anisomycin. Furthermore, in contrast to De Cesare et al. (127), Soloaga et al. (65) observed normal phosphorylation of CREB and ATF1 in these RSK2-deficient cells. However, both Sassone-Corsi et al. (35) and De Cesare et al. (127) stimulated CLS cells with EGF, whereas Soloaga et al. (65) used anisomycin or a rather high concentration of TPA, all of which have distinct cellular targets and are not necessarily stimulating identical pathways.

In a separate study, Strelkov et al. (83) reported that TPA or EGF induced activation of MSK1, but not RSK2, to phosphorylate histone H3 at Ser10 in 10T1/2 and Ciras-3 mouse fibroblasts. They also observed that 5 nM Ro318220, an inhibitor of both RSK2 and MSK1, had no effect on TPA-induced phosphorylation of histone H3. The median inhibitory concentration of Ro318220 is 3 nM and 8 nM for RSK2 and MSK1, respectively, which would be consistent with RSK2 not being the TPA-induced kinase in these cells. Furthermore, RSK2 phosphorylated histone H2B, which is associated with chromatin condensation and apoptosis (129, 130), rather than H3.

Results from cells exposed to the environmental human carcinogen arsenite further support the idea that RSK2 is one of several histone H3 kinases activated by this agent in JB6 C141 mouse skin epidermal cells (100). In addition to upstream kinases, p53 may be involved in RSK2 phosphorylation of histone H3 at Ser10 (99). EGF- or UVB-induced phosphorylation of histone H3 at Ser10 occurred in p53 wild-type mouse fibroblasts and in NCI 460 cells, a cancer cell line expressing wild-type p53, whereas no induction was observed in p53-deficient mouse fibroblasts or HT1080 cells, a p53-deficient human tumor cell line (99). Reintroduction of wild-type p53 into p53-deficient fibroblasts effectively restored the EGF- and UVB-induced phosphorylation of histone H3 at Ser10. This may have some bearing on the histone phosphorylation defects reported for cells from CLS patients, because compared to normal cells, RSK2-deficient cells from CLS patients exhibited lower phosphorylation of p53 at Ser15 after exposure to UVB. RSK2 interacted directly with p53, and p53 was required for optimal UVB-induced RSK2-mediated phosphorylation of histone H3 at Ser10 (99). On the basis of these results, we hypothesize that p53 serves as a bridge between the kinases and histone H3 and that this is required for EGF- and UVB-induced phosphorylation of histone H3 at Ser10.

The UV-inducible phosphorylation of histone H3 involves not only RSK2 but also MAPKs (97) and MSK1. UVB activates endogenous MSK1 (131, 132) and contributes to the phosphorylation of histone H3 on Ser28 (85). Treatment of JB6 cells with H89 inhibited TPA-induced phosphorylation of H3 on Ser10 mediated by MSK1 (73) and also inhibited UVB-induced phosphorylation of histone H3 at Ser28 (85). H89 blocked MSK1 activity but did not prevent the UVB-induced activation of ERKs, p38, JNKs, Akt, RSKs, or p70S6K. UVB had no effect on PKA phosphorylation in JB6 cells, suggesting that even though PKA is a known target of H89 (133), it is not involved in the inhibitory effect of UVB-induced H3 phosphorylation. Transfection experiments showed that an N-terminal mutant MSK1 or a C-terminal mutant MSK1 markedly blocked MSK1 activity and strongly inhibited UVB-induced phosphorylation of histone H3 at Ser28 in vivo. These data illustrate that MSK1 is likely to be involved in the UVB-induced phosphorylation of histone H3 at Ser28 (85).

However, the involvement of MSK1 in H3 phosphorylation is not unchallenged. Zhang et al. (134) reported that MSK1 actually inhibited transcription and that its target was instead histone H2A at Ser1. Mutating Ser1 on H2A blocked MSK1 inhibition of transcription, as did the acetylation of histone H3 by p300- and CREB-binding protein (CBP)–associated factor (PCAF). However, this study was done in an artificial VP16 system on reconstituted chromatin templates. Even so, histone H3 may not be the only relevant target for the inducible posttranslational modification of chromatin. However, one must keep in mind that most of the cited studies used either free H3 or a peptide to provide evidence that the kinase of interest was an H3 kinase. Most important, the in situ substrate is H3 in nucleosomes. Work by Bustin’s group (135) showed that MKS1 and RSK2 phosphorylated nucleosomal H3. Clearly, the regulation of chromatin is complex, and the debate as to which kinase is the most important seems to have become a moot point, because multiple kinases appear to be involved in inducible H3 phosphorylation, depending on the type of stimulus, cell, and experimental conditions.


Akt, which is also known as protein kinase B, is best known for its role in the phosphoinositide 3-kinase pathway activated by growth factors and other stimuli. However, the exposure of cells to arsenite stimulates phosphorylation of Akt1 at Ser473, which increases the kinase activity of Akt (100). A dominant-negative mutant of Akt1 inhibited arsenite-induced phosphorylation of histone H3 at Ser10. Additionally, Akt1 phosphorylated histone H3 at Ser10 in vitro. The arsenite-induced phosphorylation of histone H3 at Ser10 was almost completely blocked by a dominant-negative mutant of ERK2 or PD98059, a MEK inhibitor. Thus, Akt may be one of the targets of the MAPK signaling activated by tumor-promoting agents that is relevant for the effect these drugs have on transcription. Akt1 is also involved in the UVB-induced phosphorylation of histone H3 at Ser10, and in this case it is activated by another kinase, Fyn (107).


Recently, two groups simultaneously reported that IKKα, but not IKKβ, directly phosphorylates histone H3 (Table 1) in vivo and in vitro (101, 103). Yamamoto et al. (101) reported that IKKα is crucial for cytokine induction of nuclear factor kappa B (NF-κB)–responsive genes in the nucleus. They observed that IKKα was directly responsible for the cytokine-induced phosphorylation of histone H3 at Ser10 and the subsequent acetylation of histone H3 at Lys14. Treatment of cells with tumor necrosis factor–α (TNF-α) and subsequent chromatin immunoprecipitation assays suggested that IKKα and CBP binding to the IκBα promoter were required for the phosphorylation and acetylation of histone H3 (101). Arnest et al. (103) demonstrated that IKKα accumulated in the nucleus after cytokine exposure and was associated with NF-κB–responsive genes after TNF-α treatment. They also showed that IKKα could phosphorylate histone H3 in vitro. Both groups found that the IKKα phosphorylation of histone H3 was distinct from the role of IKKα in IκBα degradation. IKKα does not appear to be limited to cytokine-regulated gene expression; a role for IKKα in controlling EGF-induced c-fos gene expression (102) has also been reported, with IKKα required for EGF-histone H3 phosphorylation at the c-fos promoter and c-fos gene expression (102).


UVB induces the phosphorylation of histone H3 at Ser10 and Ser28 through MAPKs and MSK1 (86, 97). However, much less is known about the upstream mediators in the signal transduction pathway induced by UVB irradiation. The cellular response to UVB includes a dramatic increase in the activity of Fyn, a kinase of the Src family of nonreceptor tyrosine kinases (107). Dominant-negative Fyn inhibited the increase in ERK and Akt1 phosphorylation and activity triggered by UVB. The activity of Fyn and phosphorylation of histone H3 at Ser10 were suppressed by the Fyn inhibitors 4-amino-5-(4-chlorophenyl)-7(t-butyl)pyrazol(3,4-d)pyramide (PP2) and leflunomide (a Src kinase inhibitor), and by expression of dominant-negative Fyn or small interfering RNA (siRNA) to knockdown Fyn. Fyn is yet another kinase implicated in the inducible phosphorylation of histone H3 at Ser10, possibly through regulation of Akt1 and ERK.


PKCs, a family of kinases activated by many extracellular and intracellular signals and involved in a multitude of physiologic functions in the cell (136, 137), have also been implicated in histone H3 Ser10 phosphorylation (104). Huang et al. (104) analyzed histone modifications that occurred at the gene encoding the low-density lipoprotein (LDL) receptor in response to TPA in human hepatoma HepG2 cells. TPA activates PKC and stimulates LDL receptor gene expression in these cells (138). Huang et al. (104) observed an increase in the phosphorylation of histone H3 at Ser10, which correlated with LDL receptor induction by TPA but was not related to MAPK activity. Active PKCβ and PKCε directly phosphorylated histone H3 at Ser10 in vitro; however, histone H3 and PKC were not detected in a complex in HepG2 cells even after TPA stimulation. Thus, these findings suggest that there is yet another kinase that regulates histone H3 Ser10 phosphorylation in the IEG response induced by TPA but appears to be independent of MAPK activation. However, caution must be used when interpreting these in vitro data, especially when a complex of histone H3 and PKC could not be detected in vivo.


Several years ago, PKA was observed to phosphorylate various chromatin substrates in vitro (139). Treatment of rodent ovarian granulosa cells with follicle-stimulating hormone (FSH) resulted in rapid H3 phosphorylation that was dependent on PKA (140). Phosphorylation of histone H3 at Ser10 and acetylation of histone H3 at Lys14 occurred after stimulation of granulosa cells with FSH (84). Besides histone H3 phosphorylation, FSH induced the phosphorylation of RSK2 and ERKs. However, only inhibitors of PKA decreased the phosphorylation of histone H3 Ser10, suggesting that the phosphorylation stimulated by FSH was independent of that produced by other kinases known to be involved in histone phosphorylation in response to other stimuli, such as MAPKs, RSK1, MSK1, and PKC (84).

Taken together, the results summarized in this section suggest that the inducible phosphorylation of histone H3 that contributes to the regulation of gene expression involves a multitude of kinases and that the specific kinases are dependent on cell type, stimulus, and what genes are activated. Whether these kinases act combinatorially, synergistically, or individually is not yet clear.

Coupling Phosphorylation and Acetylation

Another area of controversy concerns the proposed interdependence of H3 phosphorylation and acetylation. Allis et al. (21, 141) have proposed that the pattern of histone modification at specific genes can affect the binding of other molecules, which subsequently mediates the specificity of transcription to dictate the final biological outcome. This concept is referred to as the "histone code" and has gained a great deal of acceptance, at least partially because of numerous research findings confirming that histone posttranslational modifications are clearly implicated in the regulation of gene expression (142).

Notably, phosphorylated histone H3 has been repeatedly reported to be associated with the promoters of many mammalian IEGs after stimulation. For example, light pulses induce prominent phosphorylation of histone H3 at Ser10 in the suprachiasmatic nucleus in rats, which corresponds with the change in the transcriptional profile of c-fos and the circadian gene Per1 (108). Neuronal activation by various drugs also produced changes in Ser10 histone H3 phosphorylation in hippocampal neurons and was associated with increased c-fos transcription (143). FSH stimulated the phosphorylation of histone H3 at Ser10 and the acetylation of histone H3 at Lys14 at FSH-responsive promoters, including serum glucocorticoid kinase, inhibin, and c-fos (84).

Several years ago, Barratt et al. (79) observed that the minute fraction of IEGs that were phosphorylated at histone H3 were also susceptible to a higher level of acetylation than that observed for other genes. This led to the hypothesis that acetylation and phosphorylation might be coupled during transcriptional activation and possibly associated with the MAPK signaling pathway (96), because IEGs are the genetic targets of MAPK signaling. Many researchers have provided evidence that H3 phosphorylation may have a role in the regulation of transcription by acting as a signal for subsequent acetylation of lysines and, in particular, histone H3 Lys14 (53, 66, 143145).

Most data appear to indicate that H3 phosphorylation precedes acetylation, resulting in multiple modifications of histone H3 (54). In fact, various HATs, including the yeast Gcn5 HAT (53, 54, 145), seem to exhibit a preference for H3 phosphorylated at Ser10 as a substrate (53, 54, 145, 146). Furthermore, cytokine-induced phosphorylation of histone H3 at Ser10 by IKKα appears to be required in the subsequent acetylation of Lys14 by CBP (101). Another example of the coupling of phosphorylation (at histone H3 Ser10) and acetylation (at histone H3 Lys14) is in the PKA-mediated transcriptional activity induced by FSH (84). Cheung et al. (53) suggested that phosphorylation and acetylation are "synergistically coupled" and that phosphorylation of the histone H3 tail occurs first and acetylation follows as a consequence. In cells stimulated by EGF, a rapid and sequential phosphorylation and acetylation of histone H3 was observed, and the phosphoacetylated histone H3 occurred preferentially at the c-fos promoter (53).

Another set of sites where phosphorylation and acetylation may be coupled is Lys9 and Ser28 of histone H3. Trichostatin A (TSA), a histone deacetylase inhibitor that increases acetylation of the N-terminal tails of histone H3, induces both acetylation at Lys9 and phosphorylation at Ser28 of histone H3 in JB6 Cl41 mouse epidermal skin cells (87). TSA treatment before UVB irradiation, which is known to induce phosphorylation at Ser28, prevented any increase in the phosphorylation of histone H3 in response to UVB. In contrast, TSA markedly increased phosphorylation and acetylation of histone H3 in cells exposed to UVB before TSA treatment. TSA treatment also stimulated MAPK activity and the MAPK inhibitors PD98059 and SB202190, and dominant-negative mutants of ERK2 or p38 suppressed TSA-induced H3 phosphorylation but had no effect on acetylation. Thus, although histone H3 phosphorylation and acetylation may be coupled, they may not always be interdependent.

The interdependence of phosphorylation and acetylation is not universally observed. In some cases, H3 acetylation occurred independently from histone H3 phosphorylation at the IEGs (54, 147), and phosphorylation of histone H3 was not required. This is explained by the model proposing that phosphorylation is a MAPK-inducible event, but acetylation and deacetylation are continuously occurring processes mediated by the constitutive action of HAT and HDACs. This hypothesis is supported by observations using MSK-knockout cells, in which the phosphorylation of histone H3 at Ser10 was defective, but there was no detectable loss of H3 acetylation at IEGs (65). In addition, inducible histone H3 phosphorylation at Ser10 independent of H3 acetylation has been linked to the induction of retinoic acid receptor expression by retinoids (144). The chromatin at the retinoic acid receptor gene appears to be constitutively acetylated at histone H3 Lys9 and Lys14 even in the absence of phosphorylation (144). Phosphorylation and acetylation at particular nucleosomes may occur through parallel but either dependent or independent pathways.

Considerable progress has been made in determining the enzymes that modify histones and in determining the induced patterns of nucleosomal modifications at a given gene. For example, two recent reports have specifically described the distinct pattern of modifications that occur during the induction of the collagenase (148) and cyclooxygenase 2 (cox-2) (149) genes. Park et al. (149) demonstrated that H3 phosphorylation and acetylation at the cox-2 promoter have a major influence on the expression of this gene in mouse RAW 264.7 macrophages. Sodium butyrate (NaBT), a short-chain fatty acid that possesses histone deacetylase–inhibiting activity, in combination with endotoxin (a lipopolysaccharide), induced increased COX-2 protein and gene expression. Treatment with NaBT induced increased global phosphorylation and acetylation of chromatin; however, histone H3 phosphorylation at Ser10 at the cox-2 promoter was specifically increased within 15 min without any change in acetylation at Lys14. Stimulating cells with NaBT and endotoxin synergistically increased both phosphorylation and acetylation at the cox-2 promoter, and the induced histone H3 modifications were blocked by MAPK inhibitors, suggesting that in this case the modifications were dependent on MAPK signaling.

In a somewhat more detailed study, Martens et al. (148) used the glioblastoma cell line T98G to systematically examine the sequence of events that occur at the human collagenase I promoter upon induction. TPA, or the addition of serum after a period of withdrawal (or both), stimulated collagenase I expression, and the expression coincided with increased c-Fos/c-Jun and p300 protein expression; and they were associated with the collagenase I promoter 30 and 60 min, respectively, after stimulation. The first modification observed was histone H3 Lys4 dimethylation, followed by phosphorylation and acetylation. The histone H3 Lys4 methyltransferase involved was SET9, and trimethylation was also observed. Both RSK2 and p300 were simultaneously recruited to the collagenase promoter. The phosphorylation and acetylation of histone H3 at Ser10 were suggested to coincide with the binding of RSK2 and p300, respectively, at the collagenase promoter. These types of studies reveal that gene expression involves a dynamic recruitment of different modification factors and a diverse but highly orchestrated and orderly set of covalent modifications that are tightly linked to cell type and to the type (and maybe the duration) of the extracellular stimulus.

Summary and Conclusions

The regulation of gene expression is obviously an extremely complex and intricate process involving a multitude of recognized and, most likely, many still elusive regulatory factors. Posttranslational modifications of histones, and especially of histone H3, clearly have a major impact on the cellular response to various stimuli. In spite of the fact that most of what we know regarding the role of histones has been obtained indirectly from in vivo or in vitro experimental data, we are beginning to unravel the complexities of gene expression mediated by histone H3 modifications, which are induced by a host of diverse stimuli. A particular stimulus may induce a certain level of phosphorylation or dephosphorylation, acetylation or deacetylation, and methylation or demethylation, so some combination of these seems to be dependent on multiple extracellular and intracellular factors. Dissimilar cells respond differentially to distinct stimuli, and the induction of gene expression is clearly dependent on the type of stimuli, duration and strength of stimuli, state of the cell (such as normal or cancerous), and, of course, specific cell type. In divergent cell types, UV irradiation, for example, induces numerous histone modifications in a distinctive wavelength-, dose-, and time-dependent manner. This strongly indicates that the regulation of histone modifications and the resultant gene expression is not just one- or two-dimensional but multidimensional, encompassing an array of factors. The continued discovery of new kinases, which play key roles in the differentially inducible histone H3 phosphorylation, suggests that the identification of the most important histone H3 kinase is no longer worthy of debate.

Second, the outcome of the various modifications under diverse conditions is just beginning to be elucidated. Much of the work with gene expression and chromatin condensation has been performed in yeast, and functional assays have not been as well developed in mammalian cells. More work is clearly needed to specifically determine the end result or function of histone H3 modification. Such functional assays might include the assessment of proliferation, cell transformation, or apoptosis under conditions using wild-type and mutant histone H3. Are histone H3 or other histones primarily the target of modifications that lead to gene expression and chromatin condensation, or do they have a more dynamic role? Perhaps in vivo studies, in which histone H3 or its numerous individual variants (H3.3A and H3.3B, among others) are systematically knocked out or knocked down, could be used to determine functional or phenotypic changes. Proteomic approaches or array analysis might be developed to detect real-time modifications or kinases that are present at specific gene promoters in vivo. This type of technology could theoretically provide the means for systematic investigations of chromatin changes during different physiological and pathological states, and clearly could reveal much more relevant information. Better understanding of the posttranslational modifications of histones and their role in gene expression, cell growth, differentiation, apoptosis, and pathophysiologic processes, such as cancer, will provide new tools to prevent or control diseases.


  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.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
  105. 105.
  106. 106.
  107. 107.
  108. 108.
  109. 109.
  110. 110.
  111. 111.
  112. 112.
  113. 113.
  114. 114.
  115. 115.
  116. 116.
  117. 117.
  118. 118.
  119. 119.
  120. 120.
  121. 121.
  122. 122.
  123. 123.
  124. 124.
  125. 125.
  126. 126.
  127. 127.
  128. 128.
  129. 129.
  130. 130.
  131. 131.
  132. 132.
  133. 133.
  134. 134.
  135. 135.
  136. 136.
  137. 137.
  138. 138.
  139. 139.
  140. 140.
  141. 141.
  142. 142.
  143. 143.
  144. 144.
  145. 145.
  146. 146.
  147. 147.
  148. 148.
  149. 149.
  150. 150.
View Abstract

Stay Connected to Science Signaling

Navigate This Article