New studies reveal details of the multidrug resistance pump P-glycoprotein
Illustration of a promiscuous multidrug resistance transporter. Created by Graham Johnson of mesoscope.org for Clinical Pharmacology & Therapeutics, Jan 2010. Reproduced with kind permission from G. Johnson.
Collaborative research between groups in the Biochemistry and Chemistry departments in Oxford will help efforts to understand the fundamental characteristics of a protein that is responsible for resistance to many cancer drugs.
The groups describe their studies of the protein, the P-glycoprotein pump, in two papers in PNAS and Proteins (1,2). As well as contributing to research aimed at developing more effective chemotherapies, the work demonstrates the use of an approach that could be applied to the study of other membrane proteins which are a major target for pharmaceutical companies.
Phil Biggin in Biochemistry and Carol Robinson in Chemistry collaborated on the project which brought together the expertise of their groups – in computational modelling and mass spectrometry respectively. Much of the work was carried out by Jerome Ma, Dr Biggin’s DPhil student, and by Professor Robinson’s postdoctoral researcher, Julien Marcoux.
P-glycoprotein (P-gp) is one of the major export pump proteins in the body and plays an important role exporting harmful compounds from the cell and transporting lipids from the inner to the outer leaflet of the cell membrane. It is up-regulated in tumour cells where it protects the cell by pumping out anti-cancer drugs, and can give rise to multidrug resistance. In the brain, P-gp has a neuroprotective role preventing harmful compounds from reaching the brain.
The pump belongs to a group of transporters known as the ATP Binding Cassette (ABC) family which needs ATP to pump compounds out of the cell. The protein is thought to alternate between two different conformations when it functions – the substrate binds to an inward facing conformation, and export proceeds in an ATP-dependent way through conformational changes to an outward facing form.
P-gp is of huge importance in the pharmaceutical industry and all potential drugs must be tested for susceptibility to it. Researchers have only recently determined its molecular structure using X-ray crystallography and many questions remain about how substrate binding and ATP hydrolysis are linked. Studying the details of drug binding has also proved a challenge because the pump’s activity is influenced by the lipids around it.
Model of mouse P-gp in an inward facing conformation used for molecular dynamics simulations. The model was based on the crystal structure PDB ID 3G60.
In the recent work published in PNAS, Dr Marcoux and colleagues turned to mass spectrometry to explore how the pump interacts with its ligands (1). They embedded P-gp in detergent micelles to study intact membrane protein complexes, managing to preserve the interaction between pump and ligand.
Using this approach, the group could follow the binding of a variety of ligands (drugs, lipids and nucleotides) in real time – the first time that this has been done with a membrane protein. They were able to deduce key characteristics of binding: the affinity of ligand for pump, the relative quantities of ligand and pump (stoichiometry), and the synergistic effects of different ligand binding. They combined this analysis with another technique called ion-mobility which provides an insight into the conformational state of the protein.
To test the experimental mass spectrometry data, Dr Marcoux and colleagues turned to Phil Biggin’s group. ‘We used the computational modelling expertise of Phil and Jerome to check if what we thought was happening was reasonable within the context of the membrane,’ explains Dr Marcoux. The molecular dynamics simulations which his Biochemistry colleagues ran supported the experimental data and confirmed the observed stoichiometry and binding sites.
Together, the results give a more detailed picture about ligand binding than it has been possible to resolve before. Their work also suggests that the pump exists in a delicately balanced equilibrium, predominantly in the inward facing form, and that more than one ligand (nucleotide plus drug) must be added together to promote the outward facing conformational state of the pump.
The group also observed synergistic effects of ligand binding, with prior binding of a pump inhibitor favouring subsequent lipid binding. This is consistent with reports that some lipids can affect the transport of different substrates by this family of transporters. Such an effect could be due to the lipids playing more than a structural role and interacting with the transport process itself, for example by stabilising different states of the transport protein.
Molecular docking of two molecules of cyclosporin A, a P-gp inhibitor, showing binding within the large transmembrane hydrophobic cavity of P-gp. (Click to enlarge)
‘It’s always been thought that drugs enter the binding pocket from within the membrane in the first place, so it’s possible that the membrane plays a more active role in transport,’ says Jerome Ma. ‘Our work provides additional evidence for this.’
‘The interpretation has been that you need specific lipids in order to keep the transporter in the proper resting state, in common with what is seen for other membrane receptors,’ adds Dr Biggin. ‘Perhaps our work shifts the view slightly – providing some evidence that lipids are playing a more active role in the transport process itself and doing more than just keeping the protein happy.’
Dr Biggin and Jerome Ma also used a solely computational modelling approach to look further at how ligand binding affects the conformation of P-gp within the context of the membrane.
‘We looked at what happens to the conformation when a transport inhibitor is bound compared with when substrate is bound,’ explains Jerome Ma. ‘Our main finding is that with inhibitor, the structure remains splayed open, whereas with a substrate, it starts closing up. That’s significant because it explains how an inhibitor might act by keeping the structure open and preventing ATP hydrolysis from happening, and why a substrate is more likely to be transported.’
Their research, published in Proteins, ties in with the mass spectrometry and ion-mobility work (2). Both projects indicate that the structure of P-gp is very flexible and that ligand binding can perturb the equilibrium between the two conformational states.
A structural study of another ABC transporter which has been published at the same time in PNAS links with the P-gp work of the two groups. The work is from Structural Genomics Consortium researchers led by Dr Liz Carpenter who have elucidated the first X-ray structure of a human ABC transporter, ABCB10 (3).
Dr Carpenter and colleagues looked at the structure of ABCB10 with nucleotide bound and found that it remained in the resting, inward facing conformation, contrary to what had been seen with other transporters. This work fits neatly with the mass spectrometry/ion-mobility work on P-gp, says Dr Marcoux, hence the papers coming out together.
‘We also find that when P-gp binds just nucleotides, we don’t change the conformation,’ he explains. ‘We see an equilibrium between states, with concomitant binding of nucleotides and drug affecting the proportion of the different states.’
Links to papers: