"Dividing cells pass through a regular sequence of cell growth and division, known as the cell cycle", said Helena Curtis in her excellent text book, Biology, in the 1985 edition. I have always liked that modest description of what we work on, partly because it is completely apt, and partly because, to read it, you would scarcely think there was anything to research at all! In fact, making sure that cells either grow, or divide, under the right circumstances, is the whole point of cell cycle control. Premature division, before DNA replication has been completed, is a fatal disaster for cells.
During the past year, we have made some good progress towards understanding how the process of cell division is initiated and terminated. In the Annual Report for last year, we said that we had discovered a phosphatase activity that reversed the action of mitotic cyclin-dependent protein kinases, whose activity was high in interphase (when the cells grow) and low in mitosis (when the cells divide), but we did not know its identity. Now we do, thanks to the investigations of Satoru Mochida, who simply asked which protein phosphatase activity was responsible. He raised antibodies against the catalytic subunits of the protein phosphatases present in frog egg extracts, and asked which of them removed 'phosphatase X' activity. It turned out that the antibodies against protein phosphatase 2A (PP2A) did not work very well for this purpose, but Gernot Walter and his colleague Ralf Ruediger kindly provided a monoclonal antibody against the A (scaffolding) subunit of PP2A, which works well in immunodepletion. This showed that phosphatase X was some form of PP2A. PP2A also contains a B subunit, of which several different varieties are known, so Satoru raised antibodies against all those he could find in the frog egg, and together with a summer student, Satoshi Ikeo, found that depletion of B-δ removed the majority of phosphatase X.
What happens when this form of PP2A is removed from an extract that is running throiugh the cell cycle? The answer, as can be seen in Figure 1, is quite dramatic. First, entry into mitosis is greatly accelerated, and subsequently, the extracts are stuck permanently in mitosis. To confirm that this really was due to the loss of PP2A-Bδ, Satoru had to make an active preparation of the enzyme by expressing recombinant subunits in bacteria and insect cells, and then show that adding the synthetic enzyme back to the depleted extracts restored their normal cell cycle entry and exit. It took some time to achieve this end, and here we were greatly helped by Yigong Shi and his colleagues in Princeton who recently solved the structure of this form of PP2A. They generously provided clones and protocols for expression and assembly of trimeric PP2A complexes. Sure enough, when recombinant PP2A-Bδ is added back to the depleted extracts, they behaved exactly as normal undepleted extracts (Figure 1). Adding back too much PP2A-Bδ either slowed down entry into mitosis, or prevented it entirely (data not shown). These results clearly demonstrate that the level of PP2A-Bδ activity regulates the timing of entry into mitosis, and that this particular phosphatase is also necessary for exit from mitosis. These are two somewhat different things, and at the moment we do not know exactly how to interpret the results. We can also see from more detailed analysis of various parameters in the extracts that some subtle effects occur in extracts with too much or too little PP2A-Bδ. For example, entry into mitosis is marked by loss of tyrosine-15 phosphorylation of CDK1, in turn reflecting the balance of inhibitory Wee1/Myt1 kinase and its antagonist, the Cdc25 phosphatase. Probably, dephosphorylation of this pair of enzymes keeps CDK1 inhibited, whereas their phosphorylation tends to activate CDK1, regardless of any effects on the substrates of CDK1. Thus, depletion of PP2A-Bδ probably helps to activate CDK1. and then makes it quicker and easier for CDK1 to work (because of the absence of the antagonising phosphatase). It is hard to disentangle these effects. Similarly, we note that the activation of the APC/C and subsequent degradation of cyclin B occur on time in relation to the activation of CDK1 in the PP2A-Bδ depleted extracts, but the inactivation of APC/C to allow the re-accumulation of cyclins to drive the next cell cycle is not quite right. At higher levels of added PP2A-Bδ, both activation and inactivation of the APC/C are seriously affected.
Figure 1. Three concentrated extracts of activated frog eggs were set up. The first was subjected to 'mock' depletion with protein G beads, while the second and third reactions contained purified polyclonal antibody to the Bδ subunit of Xenopus PP2A. To the third reaction, purified PP2A-Bδ was added after the immunodepletion procedure to restore the normal concentration of the phosphatase. Reactions were incubated at 23°C and aliquots analyzed by 1-dimensional SDS-PAGE followed by immunoblotting with anti-Apc3, or cyclin B2, or anti-phospho-SP antibodies. Entry into mitosis is marked by a shift in the mobility of Apc3 and the appearance of phosphorylated CDK targets. The Bδ-depleted extract entered mitosis at 14 minutes rather than the 28 minutes of the control, and this extract never exited from mitosis.
This is going to be very complicated to work out, and our main task now is to discover the molecular basis for the mitotic inactivation (and post-mitotic reactivation) of PP2A-Bδ. We now have the tools to pursue this, and plenty of ideas about what might be happening, but no data yet. Meanwhile, there has been progress on other projects. For example, Tohru Takaki, who was a postdoc in this laboratory before he joined Mark Petronczki's lab, made a beautifully pure preparation of cyclin D3/CDK4 that enabled our friends in Oxford, Martin Noble and Jane Endicott and their colleagues, to obtain crystals and a structure for the complex. Although the preparation was largely phosphorylated and active as a protein kinase, it was the unphosphorylated form of the enzyme that crystallised. After dephosphosphorylation with λ-phosphatase, all residual activity is lost, but can be restored by CDk7/cyclin H. This is the first available structure for cyclin D, but we are slightly disappointed that the structure is of an inactive form of the enzyme.
Julian Gannon has been pursuing his studies of the activation of the APC/C, focussing on the role of phosphorylation of its activator, Cdc20. He has raised a number of phosphate specific antibodies corresponding to sites in its N-terminus that are thought to play regulatory roles. Some of these sites (e.g. S50 and T79) show reversible phosphorylation when Xenopus egg extracts enter mitosis, and lose the phosphate upon return to interphase, but S114 is always phosphorylated. It is still extremely difficult to make active recombinant Cdc20 (except by translation of mRNA in vitro), which hampers our efforts to understand the regulation and role of this important activator of the APC/C. Alessia Errico continues her work on Tipin and its role in DNA replication. She and members of Vincenzo Costanzo's laboratory learned how to 'comb' DNA to visualise replication forks in single DNA molecules.
For a list of refereed research papers, see Publications (in navigation on left).
