When It Comes to Cell Cycle, Tail Wags Dog

Ubiquitin Tail Formation Orders Disposal of Key Cell Cycle Regulator Proteins

The cell cycle, like many complex biological systems, depends on the activation and deactivation of a multitude of regulatory proteins. The anaphase-promoting complex (APC), for example, controls key cell cycle events, such as the separation of sister chromatids, by coordinating the sequential degradation of a set of regulatory proteins. But one major question has puzzled scientists since the complex was discovered in Marc Kirschner’s lab in 1994: how does the APC manage to have proteins destroyed in precisely the correct order? In the Jan. 13 Cell, Kirschner, the Carl W. Walter professor of systems biology, and colleagues provide an answer. They report that the APC has built-in proofreading capability that helps it discern the order in which proteins are ubiquitinated, then sent to the proteasome for degradation.

Photo by Graham Ramsay

A process of “kinetic proofreading” explains how the anaphase-promoting complex coordinates the ordered degradation of cell cycle regulators, according to new research from Michael Rape (center), Sashank Reddy (right), and Marc Kirschner.

In one fell swoop, Kirschner and colleagues not only explain how the APC coordinates temporally ordered degradation, but they also offer reasons for the existence of multiple ubiquitinations and of de-ubiquitinating enzymes.

Timing Is Everything
Ubiquitin monomers are attached sequentially to protein targets to make ubiquitin chains. These must usually be four or more units in length before the fate of the protein is sealed and a proteasome steps in to degrade it. But postdoctoral fellow Michael Rape (pronounced RA-pa), with the help of MD–PhD student Sashank Reddy, found that not all APC substrates accumulate polyubiquitin chains at the same rate. Cyclin A, for example, is only poorly ubiquitinated by the APC in the G1 phase of the cell cycle, while geminin and securin—proteins that are degraded earlier—are ubiquitinated much faster.

To investigate this further, Rape and colleagues measured the affinity of each of the substrates for the APC. To their surprise, they found that all three bound equally well to the complex. In fact, in experiments designed to limit ubiquitination to the addition of single monomers, they found that cyclin A, securin, and geminin were mono-ubiquitinated at the same rate.

Why would substrates that are mono-ubiquitinated at the same rate be polyubiquitinated at vastly different rates? Kirschner and colleagues reasoned that this can be explained by “processivity,” a term that refers to the number of ubiquitin molecules that are attached to substrates during a single “visit” to the APC. If cyclin A just received one ubiquitin before it dissociated from the complex, but securin and geminin bound and received two, three, or even four ubiquitin molecules, then the differences in ubiquitination rates could be reconciled.

A Tail Order
To test this theory, Rape and colleagues carried out a simple competition experiment to limit substrates to a single trip to the APC. They prebound radiolabeled substrate to a purified complex, then added ATP, ubiquitin, the ubiquitin-conjugating enzymes, and APC to drive the ubiquitination step. The key to this experiment is that they also added a vast excess of unlabeled substrate to inhibit the radiolabeled proteins from rebinding to the APC complex once they dissociated. Rape found that while addition of unlabeled cyclin limited ubiquitination of radiolabeled cyclin A to a single monomer, unlabeled substrate hardly made any difference to the amount of ubiquitin that bound to radiolabeled securin or geminin, indicating that most of the ubiquitin gets added to these two substrates during a single binding event with the APC. In other words, processivity for securin and geminin is very high, which is in keeping with the proteins’ early degradation in the G1 phase of the cell cycle.

Kirschner and colleagues confirmed this result by using ubiquitin-associated domain proteins (UBAs), which bind to short ubiquitin chains, to capture ubiquitinated substrates as they came off the APC. They found that while UBA-captured securin and geminin were always fully ubiquitinated, UBA-captured cyclin A was only partially ubiquitinated—another strong indication that the cyclin must shuttle on and off the APC before it gets completely tagged with a ubiquitin chain.

The authors used similar experiments to test the processivity of other APC substrates. Toward the end of mitosis, for example, APC regulates the degradation of Cdc20, Aurora A, Plk1, and UbcH10, in that order. Rape found that processivity correlated exactly with the timing of the cell cycle. Cdc20 was most processive, followed by Aurora A and Plk1, and last, UbcH10.

The findings settle an old debate about whether substrate ordering, as it is known, is an intrinsic property of the APC and its substrates or whether it depends on extrinsic factors that prime substrates for degradation in the correct sequence. Kirschner and colleagues found that processivity depends on the sequence of the “D box” domain found in APC substrates. By simply mutating the cyclin A D box so it exactly matched the securin D box, they converted cyclin A from a “distributive” substrate (one that gains ubiquitin molecules in multiple visits) to a processive one.

Kinetic proofreading and the cell cycle. By tagging substrates (S) with chains of ubiquitin (U), the anaphase-promoting complex (APC) regulates the destruction of many cell cycle regulators. Processive substrates are tagged multiple times without ever leaving the APC (black arrows), and once they have accumulated four or more ubiquitins, they are degraded by the proteasome. But because distributive substrates may dissociate (red arrows) from the complex and encounter de-ubiquitinating enzymes (DUBs), kinetics dictates that the processive substrates are degraded first.

“The use of processivity to control ubiquitination is very similar to the concept of kinetic proofreading introduced by John Hopfield to explain the remarkable fidelity of protein synthesis,” said Kirschner. Like substrate ordering by the APC, the correct incorporation of amino acids into nascent peptide chains cannot be explained by affinity alone, because the relative affinities of correct and incorrect aminoacyl tRNAs differ by only about 100-fold, while the chances of incorporating an incorrect amino acid are less than one in 10,000. The high fidelity is achieved because kinetics favors the irreversible dissociation of incorrect tRNAs. Similarly, the kinetics of the APC favors dissociation of distributive substrates before they have been fully ubiquitinated, ensuring that they are not degraded prematurely by the proteasome.

Thinking Inside the Box
The need for processivity to control degradation of cell cycle components also provides a rationale for multi-ubiquitination. Scientists have struggled to find a reason why proteins must be tagged with several ubiquitin monomers before being recognized by the proteasome, but this work suggests that multiple steps provide a means to fine-tune the degradation pathway. Processivity may also explain why de-ubiquitinating enzymes (DUBs) have evolved.

“There are over 100 DUBs in the human genome, many associated with disease, and no one knows their function. But they may add specificity to systems,” suggested Kirschner. It is unknown, for example, if mono-ubiquitinated cyclin A can be de-ubiquitinated once it dissociates from the APC, but for UbcH10 that certainly seems to be the case. Rape found that ubiquitin already bound to UbcH10 is rapidly lost if a more processive substrate, such as securin, is added to the mix, indicating that DUBs do play a role in preventing UbcH10 from being degraded, while more processive substrates can easily obtain a full complement of ubiquitins. This makes sense since UbcH10 is part of the APC and is only degraded once all the other substrates have been dispatched to the proteasome.

—Tom Fagan

top