How does the clustering of phosphatidylinositol phosphates assist in pleckstrin homology domain binding of membranes?

The cells which make up our bodies consist at the simplest level of a lipid membrane containing a water-filled space controlled from the centre by the cell nucleus (where DNA is stored and which has its own secondary membrane for protection). Cells have to signal to one another and possess systems of communication in order to regulate the basic processes of life: DNA replication, cell division, cell growth and, by cell-cell interaction, the formation of tissues and organs. The major signalling routes employed are based on proteins which bind to one another and to other molecules like DNA and lipids. This means that which proteins a cell expresses and at what levels will determine how it behaves and the interactions it enjoys with its surroundings.

To enable the efficient evolution of cell signalling natural selection relies on protein "modules" or "domains". Individual modules crop up in multiple forms, and each kind shares a common basic structure adapted in each case to a different activity. This means that rather than interactions between two molecules having evolved repeatedly from scratch, over the course of evolution existing interaction interfaces have been adapted for new roles.

This grant proposal concerns one important module type, the pleckstrin homology (or PH) domain. PH domains are associated with protein-membrane interactions but also get involved in protein-protein interfaces. When binding membranes, PH domains often interact with special lipid molecules which have been modified by the chemical addition of phosphate groups, in particular lipids called phosphatidyl inositol phosphates (PIPs). Although some PH domains have been shown to bind tightly and very specifically to one kind of PIP (they have high affinity), others seem to have a low affinity - maybe too low to be relevant in the complex environment of a cell - and not to be very choosy about binding partners. We recently found that one PH domain seems to bind not to a single PIP but to several PIPs clustered together and that this increases the tightness of its binding about 300x. We are proposing that this mechanism whereby a PH domain relies on strength in numbers to target specific membranes could be relevant to a number of other PH domains whose measured affinity for their putative PIP target is low.

Our starting point will be to look at the PH domains of three related human proteins, called kindlin-1, kindlin-2 and kindlin-3. Having looked at kindlin-3 already, we want to see if the other two members of this small protein family have similar properties and recognise clusters of PIPs with increased affinity. All the data we have collected to date suggest that they probably will.

Kindlins are an interesting group of proteins to study because they have very imporant roles at several different points in the cell, and so understanding how it is that they bind to one membrane in one part of the cell or move to another part of the cell is something that matters very much. For example, the kindlins are well known to activate mechanisms by which cells stick to one another, but they also control things like growth signals and move to the cell nucleus to take up completely different activities, controlling gene expression.

We will compare the power of PIP clustering to enhance kindlin PH domain binding, to other proteins which we already know bind tightly to a single PIP molecule. Then we will test whether more proteins considered to have low PIP affinity make use of "strength in numbers" to tighten their binding.

The proteins we are going to study are known to be important in disease, including a large number of different cancers, and also heart disease and stroke. This means that investigating the possibility of a new kind of mechanism which underlies their biological effect is something that can help us understand better factors influencing the roles they play - not only in normal cell function but also when things go wrong.

Pleckstrin homology (PH) domains are important signalling modules evolutionarily deployed into a host of different signalling protein types. Some PH domains have been well characterised binding with nM affinity to phosphatidylinsotiol phosphates (PtdInsPs). Others, however, have had only 0.01-1mM affinity measured - despite the fact they seem to confer membrane localisation on the proteins they are part of. It has been suggested that the dimerisation of PH domains may assist in enhancing the tightness of PtdInsP binding, and it has also been found that auxiliary lipids alongside the PtdInsPs play a role. Here we consider the existence of a complementary mechanism.

In studies on the mammalian cell adhesion protein kindlin-3, we have found that the affinity of its PH domain for PtdIns(3,4,5)P3 lipids is enhanced 300x when clusters of five lipids can form. Kindlin-3 is one of three mammalian isoforms with important roles not just in cell adhesion but also in the control of cell surface receptor expression and cytoplasmic signalling and in gene expression control in the nucleus; consequently, kindlins not only control cell adhesion (and kindlin-3 is necessary for blood clotting) but aberrant kindlin expression is associated with a large number of cancer types.

We wish to use a combination of biophysical techniques to investigate the hypothesis that PtdInsP clustering is a general mechanism deployed by a subset of PH domains, and that PH domain oligomerisation is also an imporant factor in affinity. We will study the kindlin PH domains, a set of high-affinity 1:1 binding PH domains, and a set of other low-affinity domains. We will complement our biophysical measurements with molecular dynamics simulations of PH domains binding to membranes, image PH domain-lipid interactions directly to obtain information on the motional properties of PH domains on membranes and lipid cluster formation, and use live cell imaging to determine effects on PH sub-cellular localisation.

We are aiming to demonstrate the existence of a previously-undescribed mechanism for the recruitment of proteins containing PH domains to biological membranes. Localisation within the cell is critical for protein function, as exemplified by the proteins on which we are focusing. Understanding the basis on which PH domain-containing proteins get localised to one place or another has value for commercial private sector scientists seeking to target them pharmacologically. Our data will have the capacity to shape new proposals within R&D efforts and may identify new and more effective strategies for drug design, or for understanding why variant forms or varying concentrations of PH proteins have the health effects they do.

Policy makers at a national or international level should benefit because we will be addressing a fundamental mechanism in cell biology which is likely to be of general relevance and which therefore can shape research funding strategies going forward. In addition, demonstration of the clustering mechanism for PtdInsP binding will in itself pose a set of questions which researchers will wish to address in the future and to which funding agencies will react. The involvement of the proteins studied in a variety of public health issues (several types of cancer as well as blood clot related pathology) will be another potential funding strategy impact of our work and may shape the assessment of health provision risk and its management. All these points mean that our work will assist in the effectiveness with which public resources are deployed because it will give a clarified understanding of biological mechanisms in action during membrane adhesion and cell signalling.

In the third sector, charities promoting research towards improved therapies against a variety of diseases will also benefit because we will be providing tools for understanding in more detail how they might realistically be ameliorated. For example, one of the proteins we will study, PKB/Akt, is well-known as a cancer-related signalling protein and also is thought to be involved in the aetiology of Type II diabetes. Understanding how membrane binding by the PKB/Akt PH domain relates to lipid bilayer structure is highly relevant in both cases.

In the wider public, we expect that a detailed description of how basic molecular processes can be understood to underlie high-interest conditions such as the ones relevant to this proposal will be of significant benefit in helping an understanding of the role of basic scientists and their contribution to the health and wellbeing of the population at large.

This proposal represents a ground-breaking area of science where the team of complementary researchers assembled have the capacity to deliver a novel understanding of the basis of cell localisation and cell signalling with respect to PH domain-containing proteins. We believe it represents a major opportunity for BBSRC-funded science to take the lead in establishing a new level of understanding in membrane biology. Those involved in this multidisciplinary project will gain new insights and skills for future deployment and develop new techniques for protein-membrane interaction analysis. The post-doctoral appointee, in particular, would certainly gain new skills via his participation in the project, given its multidisciplinary nature.

The timescale for all the benefits we envisage is potentially relatively short; while pharmacological innovations would necessarily be some way off, the benefit otherwise in shaping thinking and policy is, we believe, within a 3-5 year range from the start of the grant.