The laboratory continues to focus its effort in mapping, monitoring and intervening in the network of signalling pathways linked to the aberrant phenotypic properties of transformed cells. Specifically the phenotypes under study relate to migration/ invasion and cell division. Much of this work is focused on members of the protein kinase C (PKC) superfamily and the growth factor pathways that they control. Nevertheless aspects of our derived insights, particularly around spatial organisation and the properties of specific players in these pathways, are instructive in understanding the principles and behaviour of quite distinct parts of this complex signalling network.
Cell migration
The spatial distribution of signalling processes is increasingly recognised as a dominating influence on cell behaviour. This is perhaps manifest most clearly in the polarised behaviour of cells in a differentiated tissue epithelium. It has been established in numerous models that atypical PKC isoforms (aPKCs) play a role in the establishment and maintenance of epithelial polarity. More recently it has emerged that aPKCs also contribute to the polarised, directional, behaviour of migrating cells. Work in the laboratory has continued to dissect those pathways involved in aPKC control of migration. This has focused on the finding that aPKCs and the Exocyst complex, mutually control their cellular traffic to the leading edge of migrating cells. In part, the purpose of this traffic is to control at the leading edge the activation of two MAPkinases, ERK and JNK. In turn these contribute to promoting the turnover of focal adhesions at the leading edge, enabling efficient directional migration.
A second migratory model that is influenced by PKC action is the response to HGF through its receptor c-Met. We continue to work in a collaboration with Dr Stephanie Kermorgant (Bart¿s Hospital) on understanding the signalling pathways in relation to migratory behaviour and the PKC inputs into its (spatial) properties. The demonstration that PKC-dependent c-Met traffic to a perinuclear compartment is required for at least one of the nuclear bound signals (STAT3) esvidences the spatial constraints on signalling behaviour. Further mapping of HGF/c-Met triggered plasma membrane events is ongoing. As part of our efforts to dissect these migratory models we have completed three siRNA library screens. One is directed at HGF-dependent 2-D migration and a second at a transwell migration model. A third, whole genome library screen (collaboration with Dr Mike Howell in the High Throughout Screening Laboratory), has focused around a specific invasion-associated PKCdependent property. Hits from these screens have been validated by multiple means and the actions of a number of these hits are being assessed across the various models.
Cell division
We have been working to determine the basis of the requirement for PKCε in controlling exit from cytokinesis in specific cellular models. This has developed into multiple strands of work embracing the requirements for recruitment to the furrow, the potential actions at the furrow and more broadly, the nature of the trigger that engages this particular pathway.
In dividing cells, as the furrow invaginates and the presumptive daughter cells become apparent, PKCε accumulates at the furrow specifically when it is specifically inhibited. This implies that there is an activity dependent exit from an otherwise dynamic association with the furrow. The nature of this retention and its activity dependence are under investigation. One possibility is that apposition of the two presumptive daughter cells creates a cell-cell 'contact' that facilitates this recruitment/retention under PKCε inhibited conditions. Various approaches are being employed to address this and other candidate mechanisms, including the use of optical traps to investigate cell-cell contact recruitment. This is part of a collaboration with Prof. David Klug (Imperial College) under the umbrella of the Chemical Biology Consortium.
Intriguingly, not all cell types require the activity of PKCε to successfully complete cytokinesis, which begs the question as to what particular circumstances lead to its engagement. To address this issue we are screening a series of proximal cell cycle events to uncover a process that requires PKCε activity only under certain conditions.
PKC assembly and the behaviour of protein kinases
Optimal catalytic activity of PKC isoforms requires the modification of the catalytic domain on multiple priming sites (3 sites for classical and novel isoforms, and 2 sites for the atypical isoforms and PKC-related/PKN proteins). Elucidation of the mechanisms involved in driving, retaining and losing these priming modifications have been a long-standing effort in the laboratory. Recent evidence on the behaviour of PKCα and PKCε added a significant insight into the mechanisms involved in these processes. Studies with kinase domains mutated at the lysine residue normally contacting the α-β phosphates on ATP, have demonstrated that these mutants are poorly phosphorylated in their priming sites. However when the nucleotide pocket is occupied by an inhibitor phosphorylation occurs rapidly.
The implication from this and other observations is that physiologically, ATP binding induces a conformation in the catalytic domain that is the favoured substrate for the upstream kinases and the disfavoured substrate for the competing protein phosphatases. These kinases are thus ATP-driven conformational switches, which become locked in their active state through priming phosphorylation. This switch model for ATP binding is likely shared by a number of protein kinases providing an explanation for inhibitor induced phosphorylation of these kinases. From an evolutionary perspective these observations yield a simple, compelling general model of induced function onto which is layered the evolving action of upstream kinases.
Clinically, there are important implications of this behaviour associated with the use of these priming sites as biomarkers for intervention (they will be uninformative in many cases), as well as for the potential drug liabilities associated with pseudokinase and kinase scaffold action (as non-catalytic ATP-switches). Conversely however, there is also the prediction of opportunities to target pseudokinases where such conformational switches are indicated.
Figure 1. The nucleotide dependence of kinase modification. The scheme illustrates a primary translation product (1) representing an inactive, unmodified, generic kinase. This inactive protein samples conformational space and can take on a metastable active conformation (2) that is competent to bind ATP, leading to the bound conformer (3). In this state the kinase can become phosphorylated, stabilising the conformation of the kinase (4). Nucleotide loss from the kinase (exchange or during the process of catalysis i.e. ADP dissociation), leads to a phosphorylated apo form (5) that may then revert to an inactive, modified conformer (6) that is readily dephosphorylated.
For a list of refereed research papers, see Publications (in navigation on left).
