PerspectiveCell Signaling

Stat Acetylation--A Key Facet of Cytokine Signaling?

See allHide authors and affiliations

Science  14 Jan 2005:
Vol. 307, Issue 5707, pp. 217-218
DOI: 10.1126/science.1108164

DNA binding proteins called Stats (signal transducers and activators of transcription) carry signals from cytokines and other extracellular stimuli into the cell nucleus (1, 2). Stats each carry a Src homology 2 (SH2) domain that recruits Stats to cytokine receptors at the cell surface that have been stimulated by ligand and phosphorylated. Here, Stat dimers form through association of their two phosphorylated SH2 domains, and then move to the nucleus where they bind to the DNA and switch on the expression of specific target genes. Some Stats are also phosphorylated on a conserved serine amino acid residue in the transcriptional-activation domain of their carboxyl terminus (3). Although phosphorlation is a crucial posttranslational mechanism that regulates the activities of numerous proteins, there are many others including methylation, ubiquitination, sumoylation, isgylation, and acetylation. Indeed, different Stat proteins undergo different modifications, yielding a variety of consequences for the transcription of target genes. On page 269 of this issue, Yuan et al. (4) report that Stats undergo acetylation of a single amino acid residue, lysine 685, in response to cytokine stimulation. They conclude that this modification is essential for Stats to form stable dimers and to activate transcription.

Using an antibody that detects lysine acetylation, Yuan et al. demonstrated that cytokine stimulation induced acetylation of Stat3 in the cytosol of cultured cells. When cells overexpressed p300 or CBP—histone acetyltransferase (HAT) enzymes that add acetyl groups to amino acids—Stat3 became acetylated. Meanwhile, overexpression of histone deacetylase (HDAC) 1, 2, or 3 (which remove acetyl groups from amino acids) reduced the acetylation of Stat3. Cytokine stimulation of cultured cells promoted the association of p300 with Stat3, whereas reducing p300 levels attenuated activation of transcription from a Stat reporter construct. In addition, overexpression of histone deacetylase impaired Stat-mediated activation of transcription from the reporter. Histone acetylases associated with Stat3 after ligand stimulation, but the association of histone deacetylases with Stat3 was variable and only partially modulated by cytokine stimulation.

By mutating lysine residues in Stat3, the investigators found that lysine 685 (an amino acid conserved in several Stats) was required for Stat3 acetylation. Mutant Stat3 carrying an arginine instead of a lysine residue at position 685 could be phosphorylated on tyrosine and serine and did move to the nucleus, but its ability to form dimers and to bind to DNA was impaired. Functionally, this mutant Stat did not mediate cytokine-dependent induction of gene expression or enable cellular proliferation. The authors conclude that acetylation of Stats is a cytokine-induced posttranslational modification that, like tyrosine phosphorylation and serine phosphorylation, is critical for Stats to attain their full transcriptional potential. This provocative finding provides fresh insights into how an important family of DNA binding proteins might regulate gene expression. A number of questions, however, remain to be answered, including how to reconcile this notion with previous findings.

Activating Stat3. (Bottom)

Crystal structure of Stat3 showing the location of lysine 685 (red) in the tyrosine-phosphorylated Stat3 dimer. The four recognized domains in each Stat monomer are the coiled-coil domain (green), the DNA binding domain (red), the linker domain (orange), and the SH2 domain (cyan). A disordered loop (black dotted curve) connects the SH2 domain on the left to Tail 1 (magenta). Two strands of DNA are shown in white and black. The other lysine 685 is on the back side of the SH2 domain on the right. (Top) Detailed view of the dimer interface of the two tyrosine-phosphorylated SH2 domains (cyan) of Stat3. Two tail segments are shown in magenta and yellow, and the dotted curve is a disordered loop (residues 689 to 701) connecting the SH2 domain on the left to its tail segment. Lysine 685 is depicted in red and tyrosine 705 is shown in magenta. [Adapted from (14) using PyMOL]

First, based on the crystal structure of Stat3, it is hard to conceive how the putative acetylated lysine in Stat3 could be directly involved in the formation of Stat dimers. Lysine 685 in Stat3 is situated on the external surface of the SH2 domain with its side chain exposed (see the figure). Although lysine 685 is near the phospho-Stat dimer interface, it seems to have no structural role in mediating dimer formation. An important concern of the Yuan et al. study is that mutation of lysine 685 to an arginine could have disrupted the structure of Stat in a manner unrelated to acetylation.

Second, the involvement of histone acetylases and deacetylases in the regulation of cytokine signaling is far from simple. Recent data point to a positive role for histone deacetylases in cytokine-dependent gene regulation, not the negative role implied here. Previous reports have been ambiguous concerning the effect of Stat acetylation on transcriptional activity (57). Yet more recent data have shown that Stat responses are blocked, not enhanced, by inhibition or reduced expression of histone deacetylase (811). A caveat of the new work is that it relies heavily on overexpression to establish most of its major conclusions (4). Furthermore, experiments using reporter constructs must be viewed with caution, given the very variable effects produced by overexpression of histone acetyltransferase and histone deacetylase. Indeed, the Yuan et al. findings are open to alternative explanations. For example, the effects of histone acetyltransferase, especially in reporter assays, may not be due to Stat3 acetylation per se but instead may be due to effects on other processes.

The data pertaining to the interactions of Stat3 with histone acetyltransferase or histone deacetylase are also puzzling. The authors demonstrate the association of p300 with Stats and the dissociation of HDAC3 from Stat3 after ligand stimulation of cultured cells. Because histone acetyltransferase and histone deacetylase proteins are usually found in the cell nucleus in distinct multimeric protein complexes, it is difficult to know whether all of these interactions are physiologically meaningful. Also, it is difficult to envision the intracellular compartments that house interactions between Stat3 and either histone acetyltransferase or histone deacetylase, and how these interactions are regulated by cytokines. What is the signal and how does the engaged receptor activate this process? Are the Jaks (Janus kinases) involved?

The nuclear transcription factor NF-κB (RelA) is also acetylated by p300 and deacetylated by HDAC3 in a signal-dependent manner (12, 13). In this case, however, the details seem clearer as deacetylation is specifically mediated by HDAC3, not HDAC1. Acetylation of NF-κB prevents this transcription factor from interacting with its inhibitor IκB, resulting in lifting of NF-κB repression and prolonged expression of target genes. Meanwhile, deacetylation promotes assembly of NF-κB with IκB, resulting in decreased expression of target genes. The signal dependence of acetylation appears to be linked to nuclear translocation. Shuttling of NF-κB into the nucleus brings this transcription factor close to nuclear acetyltransferases, thus facilitating its acetylation, providing a mechanistic basic for the effect of cytokines. A similar understanding of Stat3 modification is difficult to reconcile with the finding that Stat3 acetylation and deacetylation occur in the cytoplasm as well as the nucleus and can be mediated by multiple enzymes.

Clearly, much work remains to be done before we can understand the cell biology and biochemistry of Stat acetylation. Defining the mechanisms by which cytokine stimulation controls Stat acetylation and how acetylation regulates Stat function will be key to this understanding. Perhaps some of the murkiness surrounding the new findings is due to the complexity of protein modifications. For example, posttranslational modifications may be mutually exclusive, such that lysine acetylation could prevent an alternative modification that shuts down Stat activity, resulting in a net positive effect. One very important piece of information currently missing is the stoichiometry of these modifications. Although tyrosine phosphorylation of Stats tracks absolutely with dimer-induced activity, similar data for acetylation are lacking. Defining how histone acetyltransferases, histone deacetylases, and other modifying enzymes are involved in cytokine responses and sorting out their substrates will require considerable effort. Reconciling the Yuan et al. results with recent reports of the positive effects of histone deacetylases on transcriptional activity will be key to clarifying this important area of cytokine signaling.


  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.
View Abstract

Stay Connected to Science

Navigate This Article