ELK1

Figure 1

As depicted in Figure 1, the Elk1 protein is composed of several domains. Localized in the N-terminal region, the A domain is required for the binding of Elk1 to DNA. This region also contains a nuclear localization signal (NLS) and a nuclear export signal (NES), which are responsible for nuclear import and export, respectively. The B domain allows Elk1 to bind to a dimer of its cofactor, serum response factor (SRF). Located adjacent to the B domain, the R domain is involved in suppressing Elk1 transcriptional activity. This domain harbors the lysine residues that are likely to undergo SUMOylation, a post-translational event that strengthens the inhibition function of the R domain. The D domain plays the key role of binding to active Mitogen-activated protein kinases (MAPKs). Located in the C-terminal region of Elk1, the C domain includes the amino acids that actually become phosphorylated by MAPKs. In this region, Serine 383 and 389 are key sites that need to be phosphorylated for Elk1-mediated transcription to occur. Finally, the DEF domain is specific for the interaction of activated extracellular signal-regulated kinase (Erk), a type of MAPK, with Elk1.[6]

Given its role as a transcription factor, Elk1 is expressed in the nuclei of non-neuronal cells. The protein is present in the cytoplasm as well as in the nucleus of mature neurons.[6] In post-mitotic neurons, a variant of Elk1, sElk1, is expressed solely in the nucleus because it lacks the NES site present in the full-length protein.[7] Moreover, while Elk1 is broadly expressed, actual levels vary among tissues. The rat brain, for example, is extremely rich in Elk1, but the protein is exclusively expressed in neurons.[8]

Aside from the full-length protein, the Elk1 gene can yield two shortened versions of Elk1: ∆Elk1 and sElk1. Alternative splicing produces ∆Elk1. This variant lacks part of the DNA-binding domain that allows interaction with SRF.[9] On the other hand, sElk1 has an intact region that binds to SRF, but it lacks the first 54 amino acids that contain the NES. Found only in neurons, sElk1 is created by employing an internal translation start site.[10] Both ∆Elk1 and sElk1, truncated versions of full-length protein, are capable of binding to DNA and inducing various cellular signaling. In fact, sElk1 counteracts Elk1 in neuronal differentiation and the regulation of nerve growth factor/ERK signaling.[8]

Figure 2

The downstream target of Elk1 is the serum response element (SRE) of the c-fos proto-oncogene.[11][12] To produce c-fos, a protein encoded by the Fos gene, Elk1 needs to be phosphorylated by MAPKs at its C-terminus.[13][14] MAPKs are the final effectors of signal transduction pathways that begin at the plasma membrane.[15] Phosphorylation by MAPKs results in a conformational change of Elk1.[16] As seen in Figure 2, Raf kinase acts upstream of MAPKs to activate them by phosphorylating and, thereby activating, MEKs, or MAPK or ERK kinases.[17][18][19][20] Raf itself is activated by Ras, which is linked to growth factor receptors with tyrosine kinase activity via Grb2 and Sos.[21] Grb2 and Sos can stimulate Ras only after the binding of growth factors to their corresponding receptors. However, Raf activation does not exclusively depend on Ras. Protein kinase C, which is activated by phorbol esters, can fulfill the same function as Ras.[22] MEK kinase (MEKK) can also activate MEKs, which then activate MAPKs, making Raf unnecessary at times.[23] Various signal transduction pathways, therefore, funnel through MEKs and MAPKs and lead to the activation of Elk1. After stimulation of Elk1, SRF, which allows Elk1 to bind to the c-fos promoter, must be recruited. The binding of Elk1 to SRF happens due to protein-protein interaction between the B domain of Elk1 and SRF and the protein-DNA interaction via the A domain.[6]

The aforementioned proteins are like recipes for a certain signaling output. If one of these ingredients, such as SRF, is missing, then a different output occurs. In this case, lack of SRF leads to Elk1's activation of another gene.[16] Elk1 can, thus, independently interact with an ETS binding site, as in the case of the lck proto-oncogene in Figure 2.[16] Moreover, the spacing and relative orientation of the Elk1 binding site to the SRE is rather flexible,[24] suggesting that the SRE-regulated early genes other than c-fos could be targets of Elk1. egr-1 is an example of an Elk1 target that depends on SRE interaction.[16] Ultimately, phosphorylation of Elk1 can result in the production of many proteins, depending on the other factors involved and their specific interactions with each other.

When studying signaling pathways, mutations can further highlight the importance of each component used to activate the downstream target. For instance, disruption of the C-terminal domain of Elk1 that MAPK phosphorylates triggers inhibition of c-fos activation.[16] Similarly, dysfunctional SRF, which normally tethers Elk1 to the SRE, leads to Fos not being transcribed.[21] At the same time, without Elk1, SRF cannot induce c-fos transcription after MAPK stimulation.[16] For these reasons, Elk1 represents an essential link between signal transduction pathways and the initiation of gene transcription.

• Overview of all the structural information available in the PDB for UniProt: P19419 (ETS domain-containing protein Elk-1) at the PDBe-KB.