Ubiquitination is a type of protein modification in which the protein ubiquitin is attached to a target protein. In eukaryotes (organisms that include fungi, plants and animals), the addition of a ubiquitin tag can act as a signal for various cellular processes. A prime example is the destruction of ubiquitinated proteins by a eukaryotic protein complex called the proteasome. The ubiquitination process is also the target of many bacterial pathogens, which have developed techniques to hijack it for their own benefit. In papers in Nature, Akturk et al.1, Dong et al.2 and Kalayil et al.3 describe the X-ray crystal structure of the bacterial enzyme SdeA, which catalyses ubiquitination. And, writing in Cell, Wang et al.4 report the structure of a bacterial enzyme called SidE from the same protein family as SdeA.
The eukaryotic ubiquitination pathway requires a three-enzyme cascade5. An enzyme called E1 activates ubiquitin using a molecule of ATP and a magnesium ion (Mg2+) to covalently bind the ubiquitin through a type of linkage called a thioester bond. The activated ubiquitin is then transferred to downstream enzymes, which attach ubiquitin through an isopeptide bond to a lysine amino-acid residue in the target protein. The discovery6 of the SidE family of ubiquitin ligase enzymes in the bacterial pathogen Legionella pneumophila revealed a ubiquitination pathway with striking differences from the eukaryotic system. Not only can SidE ligases carry out the complete process without the aid of other enzymes, but this pathway also generates a different form of ubiquitin, termed phosphoribosylated ubiquitin (PR-Ub), in which a phosphoribose-sugar linkage attaches ubiquitin to the target protein7.
The bacterial ubiquitination pathway requires a molecule of NAD+ instead of the ATP and Mg2+ used by eukaryotes. In the first step, the mono-ADP-ribosyltransferase (mART) domain of SdeA uses ubiquitin and NAD+ to covalently attach an adenosine diphosphate ribose (ADPR) molecule to an arginine residue (Arg42) of ubiquitin6, producing an ADPR-Ub molecule. The phosphodiesterase (PDE) domain of SdeA then cleaves this ADPR-Ub to release the molecule AMP, generating PR-Ub, and this group forms a bond with a serine residue in the target protein7. However, the molecular details of how ubiquitination is catalysed were a mystery until now.
The four papers1–4 provide a detailed picture of the bacterial reaction pathway, with complementary insights into the catalytic mechanisms. Akturk, Dong and Kalayil, and their respective colleagues, report atomic structures of SdeA’s catalytic core, which consists of the PDE and mART domains.
Dong et al. imaged the largest fragment of SdeA, which includes part of the protein’s carboxy-terminal domain. They observed that the C-terminal domain is required to anchor the PDE and mART domains, stabilizing the enzyme in an active conformation. The structure of SidE reported by Wang and colleagues reveals that its catalytic domains are similar to those of SdeA, but the authors conclude that the C-terminal domain mediates SidE dimerization. This disparity might result from differences in the experimental conditions used by the two groups, or might reflect specialized functions of the individual SidE family members.
The generation of ADPR-Ub from ubiquitin and NAD+ in the first step of the reaction is revealed in a structure presented by Dong et al. of the mART domain in complex with ubiquitin and the molecule NADH, which is similar to NAD+ but can inhibit catalysis by the enzyme. This revealed that the Arg42 residue in ubiquitin that becomes modified is located too far away from the ribose group of NADH for modification to occur directly. By contrast, another of ubiquitin’s arginine residues, Arg72, which was previously shown to be important in SdeA-mediated ubiquitination7, is located much closer to the enzyme-bound NADH. The authors used computer simulations of the complex, called molecular dynamics, to show that Arg72 and one other arginine residue (Arg74) anchor ubiquitin to mART. Once the nicotinamide group from NADH is released from the enzyme, a conformational change can occur, allowing Arg42 to replace Arg72 in the active site. This model explains why ADPR attaches selectively to Arg42 and not to other arginines in ubiquitin, but further study is warranted to fully understand the process.
As the reaction proceeds, ADPR-Ub is processed by the PDE domain and PR-Ub is attached to a serine residue on a substrate protein. In addition to their studies of SdeA, Akturk et al. present the structure of ADPR-Ub in complex with SdeD, a member of the SidE family that contains only a PDE domain. Kalayil et al. used mass spectrometry techniques to study the SdeA catalytic intermediates at this stage. Both groups propose a two-step reaction mechanism for SdeA on the basis of studies of SdeA or SdeD.
First, the Glu340 amino-acid residue of SdeA binds ADPR-Ub. The His277 residue of SdeA interacts with a phosphate group on ADPR-Ub, resulting in the release of a molecule of AMP. Second, His407 activates the hydroxyl group of a serine residue on the target protein, which enables the attachment of PR-Ub to the serine. Using a mutated version of SdeA in which the histidine residue at position 407 was replaced with asparagine to trap a catalytic intermediate, Kalayil et al. captured PR-Ub bound to His277 of SdeA (Fig. 1), confirming the catalytic mechanism. Wang et al. report the structures of related complexes of ADPR and ubiquitin with SidE.
If a water molecule enters the PDE domain’s active site instead of a serine amino-acid residue, the reaction product released is unbound PR-Ub. PR-Ub can inhibit host E1-dependent ubiquitination because the PR modification prevents this form of ubiquitin from being a substrate for eukaryotic ubiquitination enzymes7. Kalayil et al. answered the question of whether the pathogenicity associated with SdeA arises from the generation of unbound PR-Ub or from the ubiquitination of host proteins. The authors tested bacterial mutants lacking SidE proteins that were engineered to express either wild-type SdeA or a mutant version of SdeA that generates only unbound PR-Ub. The authors observed that the bacteria that express mutant SdeA were unable to grow in host cells, indicating that the enzyme’s key role is ubiquitination of host proteins.
The role of PR-Ub is an emerging topic in the field of ubiquitin research. These structures of SidE family members now pave the way for more questions to be answered. For example, how is ADPR-Ub shuffled between the PDE and mART domains? The active sites of the PDE and mART domains are far apart (55 ångströms) and do not face each other. There is conflicting evidence as to whether SidE proteins exist as monomers or dimers, and, as a result, there are different models of how the gap between the domains might be bridged.
And what range of functions does the enzyme’s C-terminal domain have? The C-terminal domain stabilizes the catalytic core in SdeA but mediates protein dimerization in SidE. Dong et al. observed that ubiquitin molecules bind to the C-terminal domain of SdeA and induce a large conformational change in the enzyme, which suggests a possible regulatory role for this domain.
How many host proteins are ubiquitinated by SidE-family ligases? So far, only a few SdeA substrates have been identified6,8,9; these include the GTPase enzymes Rab and Rag, as well as the protein RTN4. From analysis of the ubiquitination sites in host proteins, Kalayil et al. and Wang et al. propose that the ligase enzyme specifically targets serine residues in disordered protein regions.
Finally, perhaps the most exciting question still to be answered is this: do enzymes that mediate this type of ubiquitination process also exist in eukaryotes?
Nature 557, 644-645 (2018)