SUMO and Transcriptional Repression: Dynamic Interactions Between ...

[Cell Cycle 2:6, 528-530; November/December 2003]; ©2003 Landes Bioscience

Extra Views

SUMO and Transcriptional Repression Dynamic Interactions Between the MAP Kinase and SUMO Pathways

rib ut io n.

ABSTRACT

Shen-Hsi Yang1 Ellis Jaffray2 Biruntha Senthinathan1 Ron T. Hay2 Andrew D. Sharrocks1,*

Previously published online as a Cell Cycle E-publication at: http://www.landesbioscience.com/journals/cc/tocnew26.php?volume=2&issue=6

KEY WORDS

©

20 03

La

nd

es

rd

fo

B

io

sc

SUMO, Elk-1, TCFs, transcription repression, MAP kinase, ERK

ot

Received 09/11/03; Accepted 09/11/03

N

*Correspondence to: A.D. Sharrocks; School of Biological Sciences; University of Manchester; 2.205 Stopford Building; Oxford Road; Manchester, M13 9PT, UK; Tel.: 0044.161.275.5979; Fax: 0044.161.275.5082; Email: [email protected]

e.

9ST UK.

Post-translational modification of proteins by SUMO is an important regulatory mechanism that impinges on many cellular processes (for reviews see refs. 1-4). However, recent studies by numerous groups have shifted our focus towards a nuclear function for SUMO modification, including a plethora of reports on transcription regulatory proteins (for reviews see refs. 4-5). We have recently added the ETS-domain transcription factor Elk-1 to this rapidly growing list of nuclear SUMO substrates.6 SUMO conjugation has been shown to regulate several different protein functions including protein stability, sub-nuclear localisation and transcriptional activation capacity (for reviews see refs. 3-5). In common with many transcription factors, SUMO negatively affects the transcriptional activating function of Elk-1. Elk-1 acts together with the transcription factor SRF and is rapidly activated by ERK MAP kinase-mediated phosphorylation in response to mitogenic stimulation (for reviews see refs. 7-9). This modification correlates with the movement of cells from the G0 state to the G1 phase of the cell cycle. Cell cycle entry is accompanied by the activation of immediate-early genes, including the Elk-1 targets c-fos and egr-1. Elk-1 phosphorylation is thought to promote transcriptional activation through recruitment of the co-activator Sur-210 or by the activation of preassembled coactivators such as the histone acetyl transferase p300.11 Activation of the MAP kinase pathway also results in the loss of SUMO modification of Elk-1.6 Thus, activation of the ERK pathway leads to the loss of SUMO conjugation at the same time as increased Elk-1 phosphorylation takes place. What then is the role of this desumoylation process? We have demonstrated that even in the absence of mitogenic stimulation, removal of SUMO through either mutation of its sites of conjugation, or by inhibiting the activity of the SUMO conjugation pathway, leads to enhanced Elk-1 transcriptional activity. This in turn implies that the Elk-1 transcriptional activation domain (TAD) has high intrinsic activating capacity, that must be suppressed by SUMO conjugation. Thus, SUMO modification appears to play an important role in dampening down the activity of the TAD. A model therefore emerges whereby Elk-1 exists in a “transcriptionally poised complex” with SRF on immediate-early gene promoters, that can respond rapidly to mitogenic signals. Desumoylation in response to mitogenic signalling through the ERK MAP kinase cascade, unmasks the latent transcriptional activation capacity of the Elk-1 TAD and phosphorylation of Elk-1 further augments this by promoting Sur-2 recruitment and p300 activation (Fig. 1). Following transcriptional activation, immediate-early gene expression and Elk-1 activity are rapidly downregulated. This downregulation is accompanied by recruitment of the mSin3A-HDAC complex and dephosphorylation of Elk-1 (for reviews see refs. 7-8). It is currently unclear what happens to the SUMO modification status of Elk-1 after the initial desumoylation response following mitogenic stimulation, but one possibility is that Elk-1 becomes resumoylated as the promoter becomes reset to a “poised” state (Fig. 1). In mammals, there are two other proteins, SAP-1 and SAP-2, that are highly related to Elk-1 and form the TCF subfamily of ETS-domain transcription factors. These proteins share several domains of high sequence and functional similarity including the ETS

nc

2School of Biology; University of St. Andrews; The North Haugh; St. Andrews, KY16

ie

1School of Biological Sciences; University of Manchester; 2.205 Stopford Building; Oxford Road; Manchester, M13 9PT UK.

is t

SUMO modification of proteins is being increasingly linked with transcriptional repression. We recently demonstrated that SUMO modification also downregulates the transcriptional activity of the ETS-domain transcription factor Elk-1. However, as Elk-1 becomes activated through MAP kinase-mediated phosphorylation, the SUMO modification is lost, providing an elegant molecular switch that promotes the loss of repressive activities at the same time as permitting the recruitment of coactivator proteins. However, the mechanism by which SUMO promotes transcriptional repression remains enigmatic.

528

Cell Cycle

2003; Vol. 2 Issue 6

SUMO AND TRANSCRIPTIONAL REPRESSION

Figure 1. Model for the dynamic regulation of the transcriptional activity of Elk-1. Elk-1 is initially maintained in a “poised” state by sumoylation. Activation through ERK pathway signalling results in the loss of SUMO modification and enhanced phosphorylation of Elk-1. This in turn leads to the recruitment of the coactivator Sur-2, thereby activating Elk-1. Subsequently, Sur-2 is lost, the corepressor Sin3A is recruited and Elk-1 becomes dephosphorylated and inactive. Resumoylation of Elk-1 will recreate the “poised” complex. For simplicity, p300/CBP is omitted from this diagram but this is thought to be continually present in Elk-1 complexes throughout the activation/deactivation cycle.

DNA-binding domain, the B-box SRF binding motif, a MAP kinase docking motif (D-domain) and the C-terminal TAD (for reviews see refs. 7-8). However, while Elk-1 contains a repression domain known as the R-motif,12 neither SAP-1 nor SAP-2 contain homologous motifs, but instead contain alternative repression domains.13-15 As the R-motif is the site for SUMO modification in Elk-1, this suggests that this regulatory event might be specific for the function of Elk-1. However, the NID repression domain in SAP-1 and SAP-2 also contains a putative SUMO modification site that conforms to the consensus site ΨKxE, suggesting that this might also impart a similar regulatory mechanism on these proteins. However, blocking the activity of the SUMO modification pathway does not affect the activity of the SAP-1 NID repression domain, suggesting that this is not the mechanism through which this domain functions (our unpublished data). Our study leaves two important outstanding questions that need resolving. Firstly, how is SUMO modification lost in Elk-1? SUMO modification is a reversible process and several SUMO proteases have been identified in mammalian cells (for review see ref. 3). Thus ERK pathway activation might act to enhance the activity of a SUMO protease. However, such a possibility would probably necessitate prior recruitment of this protease to Elk-1 in an inactive state to the “poised complex”. A similar mechanism would involve phosphorylated Elk-1 actively recruiting a SUMO-specific protease. Alternatively, phosphorylation of Elk-1 might promote a conformational change that exposes the SUMO-modification site to enable more rapid cleavage by constitutive SUMO proteases. Indeed, such a conformational change does occur upon ERK-mediated Elk-1 phosphorylation16 making this an attractive possibility. A further option is that the subcellular location of Elk-1 might be altered to a region that is rich in SUMO proteases, although to date, we have no evidence to support a SUMO-dependent relocalisation. It is possible that these different events are not mutually exclusive and may act together to promote desumoylation of Elk-1. The second major outstanding question is how SUMO modification acts to repress the activity of Elk-1 and, in more general terms, other transcriptional regulatory proteins. There are several possible www.landesbioscience.com

Figure 2. Potential mechansims of transcriptional repression by SUMO modification of Elk-1. The DNA binding domain (DBD), repressive R-motif (R) and transcriptional activation domain (TAD) in Elk-1 are indicated. SUMO conjugation to Elk-1 might act through steric hinderance where interactions with coactivators like Sur-2 are blocked or coactivators are actively excluded from the promoter. Alternatively, SUMO might induce a conformational change that occludes the interaction surface for coactivators. A third mechanism is that SUMO might actively recruit corepressor (CoRep) complexes and associated HDAC activity that oppose the activation properties of Elk-1.

modes by which SUMO might impart repressive properties on target proteins (see Fig. 2). Firstly, it is possible that SUMO recruits target proteins into repressive domains such as PML bodies, although Elk-1 does not appear to be regulated in this manner. In the case of Elk-1, SUMO might act to inhibit the recruitment of coactivator proteins. Alternatively, SUMO might actively recruit corepressor complexes. A third possibility is that a combination of these events might occur. In the first model, SUMO might sterically hinder the activation domain of Elk-1 that in turn stops its interaction with coactivators or the basal machinery. An extension of this model is that SUMO modification might allosterically inhibit coactivator recruitment by inducing a conformational change in the TAD that locks it in an inactive conformation. In the case of Elk-1, this latter model is attractive as phosphorylation of the TAD causes a conformational change in the protein that accompanies its activation. It is easy to imagine that SUMO modification might drive this process in the opposite direction. SUMO can also repress transcription when taken out of its normal context either by fusing the R-motif or, indeed, SUMO itself to an alternative DNA binding domain.6,17 Thus, the SUMO molecule itself appears to have repressive properties. These results also imply that the attachment mechanism of SUMO is relatively unimportant for this activity as either normal peptide bonds or isopeptide linkages can be accommodated. Furthermore, this suggests that in addition to steric hinderance, SUMO might be able to exclude coactivator proteins from promoter regions. However, a more likely mode of action is through recruitment of corepressor complexes. This has been hinted at by the observation that SUMO-modified p300 shows enhanced recruitment of the deacetylase HDAC6 which is a key component of corepressor complexes.18 By recruiting HDACs, SUMO-modified proteins would alter the local balance between acetylases and deacetylases, favouring the formation of deacetylated chromatin that leads to transcriptional repression. As the R-motif of Elk-1 shows significant similarity to the CRD1 domain of p300 that contains the SUMO modification sites,12,18 it is possible that this mode of action also operates in Elk-1. Indeed, the repressive activity of the R-motif in Elk-1 is TSA-dependent12 supporting this hypothesis. Finally, it should be noted that SUMO modification plays an important role in the repressive activity of the so-called synergy control (SC) motifs that are found in several transcription factors.19 SC

Cell Cycle

529

SUMO AND TRANSCRIPTIONAL REPRESSION

motifs represent a subset of SUMO-dependent transcriptional repression motifs and are only functional on promoters containing reiterated binding sites. However, it is currently unclear how this requirement relates to the mechanisms underlying the repressive function of SUMO. In addition to questions related to the specific mode of action of SUMO modification in transcriptional repression through Elk-1, it will also be important to more generally ask how SUMO acts to modify protein functions in the cell. Inspection of the SwissProt database, reveals that around 30% of all human proteins contain at least one site conforming to the minimal consensus ΨKxE. Furthermore, there are an increasing number of sumoylated proteins that are modified at sites that diverge from this consensus sequence. It appears unlikely that all these proteins are actually modified by SUMO and, indeed, that in all cases SUMO will impart repressive properties. Thus, two important questions will be to determine how the specificity of sumoylation is achieved and why SUMO is repressive in some contexts and not others. In terms of specificity, it is likely that both the local sequence context and specific enzymatic activities such as specific E3 ligases may play an important role. Indeed, it appears that a nuclear localisation sequence is an important additional requirement for sumoylation in vivo.20 It is also possible that the local sequence context also plays a role in determining the repressive activities of SUMO modification. It is envisioned that further studies on the mechanism of action of SUMO in the context of Elk-1 and other proteins will shed further light on these outstanding questions.

17. Ross S, Best JL, Zon LI, Gill G. SUMO-1 modification represses Sp3 transcriptional activation and modulates its subnuclear localization. Mol Cell 2002; 10:831-42. 18. Girdwood D, Bumpass D, Vaughan OA, Thain A, Anderson LA, Snowden AW, Garcia-Wilson E, Perkins ND, Hay RT. p300 transcriptional repression is mediated by SUMO modification. Mol Cell 2003; 11:1043-54. 19. Subramanian L, Benson MD, Iniguez-Lluhi JA. A synergy control motif within the attenuator domain of CCAAT/enhancer-binding protein alpha inhibits transcriptional synergy through its PIASy-enhanced modification by SUMO-1 or SUMO-3. J Biol Chem 2003; 278:9134-41. 20. Rodriguez MS, Dargemont C, Hay RT. SUMO-1 conjugation in vivo requires both a consensus modification motif and nuclear targeting. J Biol Chem 2001; 276:12654-9.

Acknowledgements The work in our laboratories was supported by grants from the Wellcome Trust (ADS), the MRC and BBSRC (ADS, RTH) and a Lister Institute of Preventive Medicine Research Fellowship to ADS. References 1. Hay RT. Protein modification by SUMO. Trends Biochem Sci 2001; 26:332-3. 2. Jackson PK. A new RING for SUMO: wrestling transcriptional responses into nuclear bodies with PIAS family E3 SUMO ligases. Genes Dev 2001; 15:3053-8. 3. Muller S, Hoege C, Pyrowolakis G, Jentsch S. SUMO, ubiquitin's mysterious cousin. Nat Rev Mol Cell Biol 2001; 2:202-10. 4. Verger, A., Perdomo J, Crossley, M. (2003) Modification with SUMO. A role in transcriptional regulation. EMBO Rep. 4:137-42. 5. Gill G. Post-translational modification by the small ubiquitin-related modifier SUMO has big effects on transcription factor activity. Curr Opin Genet Dev 2003; 13:108-13. ‘ 6. Yang SH, Jaffray E, Hay RT, Sharrocks AD. Dynamic interplay of the SUMO and ERK pathways in regulating Elk-1 transcriptional activity. Mol Cell 2003; 12:63-74. 7. Sharrocks AD. Complexities in ETS-domain transcription factor function and regulation; lessons from the TCF subfamily. Biochem Soc Trans 2002; 30:1-9. 8. Shaw PE, Saxton J. Ternary complex factors: prime nuclear targets for mitogen-activated protein kinases. Int J Biochem Cell Biol 2003; 35:1210-26. 9. Wasylyk B, Hagman J, Gutierrez-Hartmann A. Ets transcription factors: nuclear effectors of the Ras-MAP-kinase signalling pathway. Trends Biochem Sci 1998; 23:213-6. 10. Stevens JL, Cantin GT, Wang G, Shevchenko A, Shevchenko A, Berk AJ. Transcription control by E1A and MAP kinase pathway via Sur2 mediator subunit. Science 2002; 296; 755-8. 11. Li Q-J, Yang S-H, Maeda Y, Sladek FM, Sharrocks AD, Martins-Green M. MAP kinase phosphorylation-dependent activation of Elk-1 leads to activation of the coactivator p300. EMBO J 2003; 22:281-91. 12. Yang S-H, Bumpass DC, Perkins ND, Sharrocks AD. The ETS-domain transcription factor Elk-1 contains a novel class of repression domain. Mol Cell Biol 2002; 22:5036-46. 13. Criqui-Filipe P, Ducret C, Maira SM, Wasylyk B. Net, a negative Ras-switchable TCF, contains a second inhibition domain, the CID, that mediates repression through interactions with CtBP and de-acetylation. EMBO J 1999; 18:3392-403. 14. Maira SM, Wurtz JM, Wasylyk B. Net (ERP/SAP2) one of the Ras-inducible TCFs, has a novel inhibitory domain with resemblance to the helix-loop-helix motif. EMBO J 1996; 15:5849-65. 15. Stinson J, Inoue T, Yates P, Clancy A, Norton JD, Sharrocks AD. Regulation of TCF ETS-domain transcription factors by helix-loop-helix motifs. Nucleic Acids Res 2003; 31:4717-28. 16. Yang S-H, Shore P, Willingham N, Lakey JH, Sharrocks AD. Mechanism of phosphorylation-inducible activation of the ETS-domain transcription factor Elk-1. EMBO J 1999; 18:5666-74.

530

Cell Cycle

2003; Vol. 2 Issue 6