The mechanistic details of the attachment of a small protein, ubiquitin, to other proteins are unclear. Crystal structures of the complexes formed by the E2–ubiquitin and RING E3 enzymes offer new insights. See Article p.115
Cells use molecular tags to modulate the fates and functions of proteins. One such tag is ubiquitin, a small protein that regulates nearly every facet of cellular function in eukaryotes (organisms such as animals, plants and fungi). Tagging a protein with ubiquitin requires the sequential action of three types of enzyme: E1 activating enzymes attach ubiquitin to a cysteine amino-acid residue on E2 conjugating enzymes, and E3 ligases stimulate ubiquitin transfer from E2–ubiquitin onto a lysine residue of the substrate protein. How E3 enzymes — the most common of which belong to the RING family1 — carry out the final step has been a long-standing mystery. Now Plechanovová et al.2 (page 115 of this issue) and Dou et al.3 (writing in Nature Structural & Molecular Biology) illuminate this mechanism at high resolution, by describing the structures of RING E3 ligases engaged with E2–ubiquitin. Their results suggest a mode of action that could apply to other E3 enzymes.
More than 600 human genes encode RING or RING-like E3 ligases, underscoring their biological importance1. Canonical RING proteins contain a zinc-binding region that is rich in cysteine and histidine residues and that, on its own, can bind to E2–ubiquitin and promote ubiquitin transfer1. Previous crystal structures revealed some of the interactions between E2 and E3 enzymes, but none of them had captured the elusive association of an E2–ubiquitin intermediate and a RING E3 ligase. This was largely because the link between E2 and ubiquitin is a labile thioester bond.
Plechanovová et al. and Dou et al. cleverly overcame this challenge by using engineered E2 proteins that were linked to ubiquitin through more-stable bond types (peptide and oxyester bonds, respectively). Both groups of researchers mixed their engineered E2–ubiquitin with an E3 RING ligase, and determined the crystal structures of the resulting RING–E2–ubiquitin protein complexes. For the E3 ligase, Dou et al. used a dimeric BIRC7, whereas Plechanovová et al. used a tandem protein fusion (RNF4–RNF4) to mimic the RNF4 dimer.
Earlier studies showed that, in the absence of an E3 partner, E2–ubiquitin can adopt many inactive ('open') configurations4, which presumably prevent the transfer of the molecular ubiquitin tag to a substrate protein (Fig. 1a). The structures determined by Plechanovová et al. and Dou et al. reveal that RING E3 ligases lock E2–ubiquitin into an activated, closed conformation that is poised for ubiquitin transfer; such a form has also been described in concurrent studies of similar proteins using nuclear magnetic resonance3,5.
In the RING–E2–ubiquitin crystal structures, certain amino-acid residues of one of the two RING monomers interact with both ubiquitin and the E2 protein. Of note, an arginine side chain of one RING monomer bridges the E2 protein and the carboxy-terminal tail of ubiquitin. The opposite RING subunit also contacts ubiquitin through, for example, a highly evolutionarily conserved tyrosine or phenylalanine residue. Moreover, a zinc-bound histidine (which is characteristically found in canonical RING proteins) interacts with ubiquitin through a hydrogen bond.
The crystal structures also show an extensive network of interactions between the E2 protein and its linked ubiquitin. In particular, Plechanovová et al. describe a hydrogen bond between a carbonyl oxygen of ubiquitin's C-terminal tail and a highly conserved asparagine side chain of the E2 protein; this asparagine is known6 to be required for efficient ubiquitin transfer. In addition, an aspartate residue of the E2 protein, which had previously been shown to have a role in activating the substrate protein's lysine7, is reconfigured in the RING–E2–ubiquitin complexes.
The findings support a model by which RING binding reduces the conformational heterogeneity of E2–ubiquitin and constrains ubiquitin's C-terminal tail in a shallow cleft within the E2 protein (Fig. 1a). As a result, the thioester bond becomes suitably positioned for attack by the substrate protein's lysine, and several residues of the E2 protein are rearranged to promote the transfer reaction. Both groups of authors validated the model through careful biochemical studies. For example, ubiquitin transfer was diminished when the authors made amino-acid changes in the E3 ligase that were predicted to impair its interactions with ubiquitin or with the E2 protein8. Moreover, Plechanovová and colleagues describe that their E2–ubiquitin is a competitive inhibitor of E3-mediated ubiquitin transfer to substrate proteins. This result confirms that E2–ubiquitin (in which the two proteins are linked through a peptide bond instead of a thioester) is structurally similar to natural E2–ubiquitin.
Does the model hold for other E2 and E3 proteins? An earlier study9 showed that a non-RING E3 ligase (RanBP2) interacts with E2–SUMO in such a way that both E2 and SUMO (a ubiquitin-like protein) are optimally positioned for the transfer reaction to take place. And the RING–E2–ubiquitin structures show striking similarities to that of the protein complex formed by RanBP2, an E2 protein and a SUMO-tagged protein substrate9 (Fig. 1b). Furthermore, Plechanovová et al. show that CHIP, an E3 ligase belonging to the RING-like U-box family, also stimulates ubiquitin transfer by rearranging E2–ubiquitin into a closed configuration. Moreover, computer modelling10,11 and nuclear magnetic resonance data5 have indicated that some monomeric RING, or RING-related (SP-RING), E3 ligases contain elements that could lock E2–ubiquitin or E2–SUMO into a closed conformation.
However, there is evidence that, for some E2 proteins, E2–ubiquitin can adopt a closed configuration in the absence of E3 ligases12,13,14. And it is unclear whether some other types of E3 ligase, which transfer ubiquitin through mechanisms different from those used by RING proteins, will follow the model described by the authors. For example, for E3 ligases of the HECT and RBR families, ubiquitin is transferred from an E2 protein onto a cysteine in the E3 enzyme, before being attached to the protein substrate. Although details of the second step await elucidation, it has been reported15 that HECT binding to E2–ubiquitin promotes tag transfer without stimulating E2–ubiquitin thioester reactivity, in contrast to RING, SP-RING and some other E3 ligases.
In summary, a unified model emerges for those E3 ligases that activate the reactivity of the thioester bond. The binding of an E3 enzyme restricts the conformations available for E2–ubiquitin, which is then forced to adopt a configuration that optimally aligns the thioester for attack by the substrate's lysine. Future studies are required, however, to address how E3–E2–ubiquitin complexes interact with their protein substrates.
Deshaies, R. J. & Joazeiro, C. A. Annu. Rev. Biochem. 78, 399–434 (2009).
Plechanovová, A., Jaffray, E. G., Tatham, M. H., Naismith, J. H. & Hay, R. T. Nature 489, 115–120 (2012).
Dou, H., Buetow, L., Sibbet, G. J., Cameron, K. & Huang, D. T. Nature Struct. Mol. Biol. http://dx.doi.org/10.1038/nsmb.2379 (2012).
Pruneda, J. N., Stoll, K. E., Bolton, L. J., Brzovic, P. S. & Klevit, R. E. Biochemistry 50, 1624–1633 (2011).
Pruneda, J. N. et al. Mol. Cell http://dx.doi.org/10.1016/j.molcel.2012.07.001 (2012).
Wu, P. Y. et al. EMBO J. 22, 5241–5250 (2003).
Yunus, A. A. & Lima, C. D. Nature Struct. Mol. Biol. 13, 491–499 (2006).
Plechanovová, A . et al. Nature Struct. Mol. Biol. 18, 1052–1059 (2011).
Reverter, D. & Lima, C. D. Nature 435, 687–692 (2005).
Yunus, A. A. & Lima, C. D. Mol. Cell 35, 669–682 (2009).
Dou, H. et al. Nature Struct. Mol. Biol. 19, 184–192 (2012).
Hamilton, K. S. et al. Structure 9, 897–904 (2001).
Wickliffe, K. E., Lorenz, S., Wemmer, D. E., Kuriyan, J. & Rape, M. Cell 144, 769–781 (2011).
Saha, A., Lewis, S., Kleiger, G., Kuhlman, B. & Deshaies, R. J. Mol. Cell 42, 75–83 (2011).
Kamadurai, H. B. et al. Mol. Cell 36, 1095–1102 (2009).
About this article
Cell Reports (2016)
Cell Reports (2015)
Nature Structural & Molecular Biology (2015)
PLoS ONE (2014)
Journal of Proteomics (2013)