I am currently too busy to write the posts that are on my to-do list, which includes i) the next part of my soul searching series, ii) a text discussing if bioinformatics should be considered a science or a black art, iii) talking about a very blatant example of bad bioinformatics in a recent Cell paper, and finally two provocative pieces on v) what I don’t like about open access publishing and vi) why I think that most of junk DNA is actually junk. I am looking forward to being publicly ridiculed for the last one.
Today, I just want to mention that the last few months have been a very good time for ubiquitin research. Here are a few papers that have appeared recently, which I consider to be must-reads:
- E3-Independent Monoubiquitination of Ubiquitin-Binding Proteins from Ivan Dikic’s group. This paper probably qualifies as a ‘everything-you-know-is-wrong paper‘ as it breaks a quasi-dogma in protein ubiquitination. Just a brief reminder: ubiquitin gets transfered to proteins in a three-step process. The first step is activation: under consumption of ATP, the C-terminus of ubiquitin is attached to a Cys side chain of an E1 ‘ubiquitin activating enzyme’. In the second step, the ubiquitin gets transferred to a Cys side chain of another protein, the E2 ‘ubiquitin conjugating enzyme’. In the third step, another class of proteins, the E3 ‘ubiquitin ligases’ help to transfer the ubiquitin to a Lys side chain of the target. Most of the substrate specificity is thought to lie in the E3 step, and there exists a myriad of different E3 proteins and protein-complexes for serving the different targets that need to be ubiquitinated.
In the new paper, the authors show that proteins that contain a ubiquitin-binding domain do not require an E3 for becoming ubiquitinated – the E2 alone is sufficient. This finding is interesting in several respects. It has been known for some time that proteins with certain ubiquitin binding domains have a tendency to become mono-ubiquitinated. This phenomenon has usually been explained by the ubiquitin-binding domain interacting with the first attached ubiquitin, thus preventing chain elongation. While this explanation might still be true for explaining the absence of poly-ubiquitin chains, it looks like the ubiquitin binding domains actively help to get the first ubiquitin tag onto the protein. In addition, the large number of ubiquitin binding domains that occur in many proteins have been interpreted as ubiquitin receptors, i.e. they were thought to sense the ubiquitination status of other proteins. The results presented in this paper suggest that at least some of the domains have a different function, focusing on the protein that hosts the domain. It is not clear if this property is shared by all known classes of ubiquitin binding domains. The authors mention “UBA, UIM, UBM, NFZ, and UBZ” (never heard of NFZ, they probably mean NZF), which – if true – would include the majority of known ‘professional ubiquitin receptor domains’.
- Sequential E2s Drive Polyubiquitin Chain Assembly on APC Targets from David Morgan’s group. This paper addresses the interesting question which E3 works together with what E2, using the anaphase promoting complex APC ( a multi-subunit E3) as an example. The result is quite exciting. As I said above, the E2s normally don’t do substrate recognition (and don’t have recognition domains), this is the task of the E3 component. Nevertheless, there are quite a few different E2s in all eukaryotic genomes. The usual explanation for this multiplicity is that the different E2s serve the different E3s. What the authors did was to test all yeast E2s for their ability to partner with the APC complex. They found two different E2s: UBC1 and UBC4 are both important, although for separate tasks. UBC4 supports the rapid mono-ubiquitination of multiple lysines of the APC targets, while UBC1 is required for chain elongation on pre-attached ubiquitin. Again, there are two aspects that I find particularly interesting. First, this system is a nice contrast to another E2/E3 system described in March 2007: The pair UBE2G2 (E2) and gp78 (E3) was found to transfer preformed ubiquitin chains to the substrate. In this latter system, a poly-ubiquitin chain grows on the catalytic Cys residue of the E2 enzyme; the substrate goes from being non-ubiquitinated to poly-ubiquitinated without passing the mono-ubiquitinated stage. There is an interest twist to this Ub-chain preformation on the E2-Cysteine: a paper from Mark Hochstrassers group in April 2007 has shown these Cys-liked poly-Ub chains can also serve as a degradation signal for the E2, adding another layer of regulation to this process. The second interesting aspect concerns the role of UBC1: in yeast, this is the only example of an E2 that has an additional domain besides the canonical UBC catalytic domain. UBC1 and its mammalian counterpart E2-25K carry a ubiquitin binding UBA domain at their C-terminus. Ever since the discovery of the UBA domain 11 years ago, I have been thinking that an E2 with an additional ubiquitin recognition domain should be involved in transferring one ubiquitin onto another one. I am very glad to see this idea confirmed in 2007.