Rear view of an enzyme

The enzyme Ubc9 mediates attachment of the small modifier protein SUMO to target proteins. It emerges that for optimal functioning — and for proper meiotic cell division — Ubc9 itself must be modified by SUMO.

Chemical modification can change a protein's fate and behaviour. One such modification is SUMOylation, which involves covalent attachment of the small protein SUMO to target proteins. SUMOylated proteins can carry individual SUMO molecules or chains, with each modification resulting in a different fate. Typically, the addition of SUMO chains through the process of polySUMOylation marks the target for further tagging with chains of the SUMO-related protein ubiquitin and so for degradation1. Writing in Molecular Cell, Klug et al.2 show that enhanced polySUMOylation is essential for meiotic cell division. The authors also describe how the balance is tipped between the addition of a single SUMO and a SUMO chainFootnote 1.

The attachment of SUMO to other proteins involves the sequential action of E1 and E2 enzymes1. The E1 enzyme Aos1/Uba2 binds to and activates SUMO, before establishing a thioester linkage between SUMO and the E2 enzyme Ubc9. This latter enzyme catalyses the formation of a stable isopeptide bond between the carboxy terminus of SUMO and lysine amino-acid residues in the target protein. This step frequently involves a third enzyme, E3 ligase.

Amino acids adjacent to the target protein's acceptor lysine are often important for Ubc9 activity, particularly when E3 enzymes are not involved. In fact, many target proteins carry a 'consensus' sequence that is preferentially used for SUMOylation in this situation. There are several such consensus sequences in the flexible amino-terminal domain of the SUMO protein found in the budding yeast Saccharomyces cerevisiae, and the lysines in them act as the main sites for building SUMO chains3.

Klug et al. find that, remarkably, the activity of S. cerevisiae Ubc9 is controlled by its own SUMOylation. Previous work4,5 had shown that yeast Ubc9 can be SUMOylated at two lysine residues near its own carboxy terminus, and that these modifications negatively regulate the ability of Ubc9 to conjugate SUMO to target proteins. The present paper demonstrates that although such SUMOylated Ubc9 (Ubc9*SUMO) is catalytically inactive as an E2 enzyme, it nevertheless promotes polySUMOylation. This may seem paradoxical, but Klug and colleagues show that enhanced polySUMOylation reflects the ability of Ubc9*SUMO to act as a scaffold (Fig. 1).

Figure 1: Forging a link in the SUMO chain.

a, Attachment of the carboxy terminus of a SUMO protein to the catalytic site of Ubc9 through a thioester bond potentiates enzymatic activity of Ubc9. By contrast, SUMO conjugation close to the C terminus of Ubc9 through an isopeptide bond (Ubc9*SUMO) blocks its enzymatic activity. Klug et al.2 find that Ubc9*SUMO promotes chain formation by acting as a scaffold. The SUMO moiety of Ubc9*SUMO binds to the rear of the active Ubc9 (b). The catalytic site of Ubc9*SUMO then recognizes a consensus sequence at the amino terminus of the SUMO attached to the active Ubc9, placing this SUMO in an optimal configuration for transfer to a target protein (c). The active Ubc9 recognizes an incoming SUMO (d) and catalyses formation of an isopeptide bond between the thioester-linked SUMO and a lysine with the N terminus of the incoming SUMO (e). Chain release follows (f).

An earlier study has shown6 that Ubc9 can interact non-covalently with SUMO at a surface that is both spatially and functionally distinct from its catalytic site. Notably, this rear surface must be intact for Ubc9 to mediate polySUMOylation efficiently. Klug and co-authors show that, in this Ubc9*SUMO interaction, the SUMO moiety binds to the rear surface of an active Ubc9 (which the authors call the enzyme's backside), which is linked to another SUMO through a thioester bond. At the same time, the catalytic site of Ubc9*SUMO binds the consensus motifs of the thioester-linked SUMO, to place it in an optimal configuration for isopeptide-bond formation with the next SUMO protein as the latter is recruited by the catalytically active Ucb9.

This intricate mechanism shares some features with mechanisms reported earlier, in which E2 and E2-like proteins function not as enzymes but as scaffolds during the conjugation of target proteins. For example, one type of ubiquitin chain is assembled by an E2 enzyme in association with an inactive E2-like accessory protein7. The accessory protein binds an incoming ubiquitin so that the appropriate lysine residue of this ubiquitin is correctly oriented with respect to the catalytic site of the active E2 enzyme, ensuring that this lysine is used to form the next link of the chain.

Additionally, mammalian Ubc9 forms an extremely stable complex with RanBP2, a nuclear-pore protein8. This complex acts as a composite E3 enzyme: the RanBP2-bound Ubc9 has a structural role, with a second Ubc9 catalysing isopeptide-bond formation between SUMO and its target protein. Intriguingly, the rear surface of Ubc9 that is stably bound to RanBP2 does not seem to be exposed, and so the orientation of the two Ubc9 molecules associated with RanBP2 probably differs from the arrangement of paired Ubc9 molecules described by Klug and colleagues. Collectively, these mechanisms suggest that E2 enzymes can be adapted for a variety of structural roles in conjugation pathways, and they hint that many more variations on this theme will be found.

Klug and co-authors also investigated the significance of Ubc9 SUMOylation during meiosis — the process by which a parental cell containing two sets of chromosomes eventually divides into four daughter cells, each carrying a single chromosome set. Yeast cells containing a Ubc9 mutant that could not be SUMOylated failed to form a synaptonemal complex, a meiotic structure that promotes both genetic exchange between chromosome pairs and correct chromosome segregation to daughter cells. In these mutants, individual components of the synaptonemal complex showed altered concentrations as well as faulty incorporation into the complex. These findings build on earlier results9 to indicate that the formation of SUMO chains is necessary for the synaptonemal complex to function, although many details of the meiotic defect remain unclear.

The same Ubc9 mutants grew normally under all the non-meiotic conditions that Klug et al. tested, and showed no increased sensitivity to stress conditions. This may imply either that normal growth and stress responses do not require polySUMOylation levels as high as those needed for meiosis, or that other mechanisms for achieving polySUMOylation can be invoked and are sufficient outside meiosis. The researchers note that wild-type Ubc9 becomes increasingly SUMOylated during meiosis, and it will be interesting to discover how this 'autoSUMOylation' is triggered to promote the polySUMOylation of downstream targets.

Ubc9 is the sole E2 enzyme in the SUMO pathway and so can act as a key regulatory point under circumstances in which the overall pattern of SUMOylation fluctuates in response to varying cellular conditions10,11. Although different consequences of autoSUMOylation on different domains of mammalian Ubc9 have been described5, it would be useful to know whether Klug and colleagues' mechanism also applies in human cells. More generally, the authors' work suggests that there is much to learn about the feedback regulation of the SUMO pathway through SUMOylation and the many circumstances in which these mechanisms contribute to cellular function.


  1. 1.

    *This News & Views article was published online on 22 May 2013.


  1. 1

    Gareau, J. R. & Lima, C. D. Nature Rev. Mol. Cell Biol. 11, 861–871 (2010).

    CAS  Article  Google Scholar 

  2. 2

    Klug, H. et al. Mol. Cell (2013).

  3. 3

    Bylebyl, G. R., Belichenko, I. & Johnson, E. S. J. Biol. Chem. 278, 44113–44120 (2003).

    CAS  Article  Google Scholar 

  4. 4

    Ho, C.-W., Chen, H.-T. & Hwang, J. J. Biol. Chem. 286, 21826–21834 (2011).

    CAS  Article  Google Scholar 

  5. 5

    Knipscheer, P. et al. Mol. Cell 31, 371–382 (2008).

    CAS  Article  Google Scholar 

  6. 6

    Knipscheer, P., van Dijk, W. J., Olsen, J. V., Mann, M. & Sixma, T. K. EMBO J. 26, 2797–2807 (2007).

    CAS  Article  Google Scholar 

  7. 7

    Eddins, M. J., Carlile, C. M., Gomez, K. M., Pickart, C. M. & Wolberger, C. Nature Struct. Mol. Biol. 13, 915–920 (2006).

    CAS  Article  Google Scholar 

  8. 8

    Werner, A., Flotho, A. & Melchior, F. Mol. Cell 46, 287–298 (2012).

    CAS  Article  Google Scholar 

  9. 9

    Watts, F. Z. & Hoffmann, E. BioEssays 33, 529–537 (2011).

    CAS  Article  Google Scholar 

  10. 10

    Hsieh, Y.-L. et al. EMBO J. 32, 791–804 (2013).

    CAS  Article  Google Scholar 

  11. 11

    Kelley, J. B. et al. Mol. Cell. Biol. 31, 3378–3395 (2011).

    ADS  CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Mary Dasso.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Dasso, M. Rear view of an enzyme. Nature 497, 576–577 (2013).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.