Feature: Probing pathogenic proteins

When X-ray crystallographer Professor Michael Parker obtained his first postdoctoral appointment at the European Molecular Biology Organisation’s protein-structure laboratory in Hamburg in 1988, he was offered a potentially career-defining choice.

Former EMBO executive director John Tooze suggested the young Australian researcher could study DNA-binding proteins that operate within the cell nucleus or membrane-associated proteins that transact a cell’s business with the external world.

Parker chose to study colicin A, one of a family of membrane-associated proteins that has figured (in an accessory role) in some of world’s worst outbreaks of Escherichia coli food poisoning.

He showed that E. coli bacteria secrete colicin A in a water-soluble form that spontaneously inserts into the lipid bilayer of the host cell membrane, and adopts a different shape and oligomerises with other colicin molecules, forming a pore in the membrane.

Parker left EMBO in 1991, when Melbourne’s St Vincent’s Institute invited him to set up an X-ray crystallography laboratory at the institute. He continues to work on the structures of membrane-associated proteins, but in recent years his work has shifted from microbial proteins involved in pathogenicity to neuronal proteins with roles in brain function.

Parker’s achievements over the past two decades have earned him the Australian Society of Biochemistry and Molecular Biology’s Lemberg Medal, and an invitation to deliver the Lemberg Memorial Lecture at this year’s ComBio 2011 research conference in Cairns in September.

He also received the coveted Ramaciotti Medal for Excellence in Biomedical Research in October in recognition of his work.

Penetrating proteins

In his first two years at St Vincents, Parker focused on a structural analysis of aerolysins in Aeromonas bacteria. Aerolysins serve a similar function to colicins in E. coli.

Aeromonads are members of the microbial community that purifies water in reservoirs, but can cause food poisoning in humans and problem pathogens for aquaculture.

Parker says studies of bacterial toxins like colicins and aerolysins have shown that pore-forming proteins (PFPs) have the intriguing property of being able to exist in a stable, water-soluble state, then to undergo a large change in conformation before assembling into an integral membrane pore.

In their aqueous alter ego, PFPs crystallise readily, making them ready subjects for X-ray structural analysis. The ability to ‘flip’ between hydrophilic and hydrophobic forms turns out to be a general property of pore-forming proteins in eukaryotes.

Parker moved on to characterise the structure of perfringolysin, secreted by Clostridium perfringens, the anaerobic bacterium that causes gas gangrene by punching holes in muscle cell membranes, causing rapid loss of muscle tissue.

Perfringolysin is from a PFP family that includes perforins, secreted by natural killer (NK) and CD-8 cytotoxic T-cells, specialised immune-system assassins that clear the body of virus-infected and pre-cancerous cells.

---PB---

Parker’s PhD student Ruby Law began work on the structure of perforin at St Vincent’s, but took the project with her when she joined Professor James Whisstock’s protein-structure group at Monash University’s Department of Biochemistry and Molecular Biology.

Law was lead author a Nature paper in November 2010 describing the structure and self-assembling behaviour of perforin by Whisstock’s group and Professor Helen Saibil’s team at the Birbeck College Institute of Structural and Molecular Biology in London.

Like its bacterial cousins, perforin is a multi-domain protein that oligomerises in the outer membrane of the target cell, forming a pore through which the attacking NK or CD-8 T-cell pumps granzymes that trigger the cell’s death by apoptosis.

Parker suspects the similarity of structure and function, along with their self-assembling behaviour, mean perforins are ancient bacterial pore-forming toxins whose cell-killing behaviour has been recruited to protect the body against viral infections and cancer.

In membrane-bound form PFPs are difficult to crystallise, but Parker says he has had success using low-resolution (20 angstrom) electron microscopy to study proteins in situ to gather details that can be used to flesh out structural models.

“I became involved in studying membrane-bound forms of proteins purely to understand what they look like, and how they work, but membrane proteins are potential targets for drug discovery,” Parker says.

“Because I was now well known for my work with membrane-associated proteins, and there are many different types of protein receptors in cell membranes, I became involved in collaborations that led me in new directions.”

Collaborations

One of those collaborations was with University of Melbourne colleague Professor Colin Masters, renowned for his work on Alzheimer’s disease, Creutzfeld-Jakob disease and other neurodegenerative disorders.

In 1987, Masters and Professor Konrad Beyreuther, of the University of Cologne, identified Alzheimers Precursor Protein (APP), a protein of nerve-cell membranes, as a potential cause of Alzheimer’s disease. Specifically, a fragment cleaved from APP, amyloid-beta, aggregates into plaques and tangles that clog the brains of patients with Alzheimer’s disease. It became the prime suspect as the cause of Alzheimer’s disease.

“Nobody looked at the rest of the protein because the hypothesis was that it probably has a very important physiological role,” Parker says. “If you designed drugs to target these domains, they might interfere with the normal structure and function of APP. With hindsight, knowledge of the structure and function of these domains might also help with drug discovery.

“APP is a large protein, so we solve the structures of parts of it and identified new functions for the protein that could have direct application to drug discovery. We have experimental evidence of the way the protein assembles in the cell membrane.”

---PB---

Parker says there is growing evidence that APP is half of an identical pair of molecules that dimerise to form a nerve-cell receptor. The enzymes beta secretase and gamma secretase perform successive cleavage reactions on the precursor protein to release the soluble beta-amyloid fragment.

Parker and Masters believe the dimer structure is crucial to understanding how the APP molecule is processed to release beta-amyloid from the cell membrane, and for designing drugs to inhibit its release and aggregation.

“I’ve worked on two other, closely related projects over the past 5-10 years with Michael Waters, of the University of Queensland, on the human growth hormone receptor (hGFR),” says Parker.

“Genentech discovered the receptor and worked with the extracellular domain. They got a structure for it and proposed that hormones and cytokines act by dimerising the receptor, which then transduces the signal across the membrane.”

But Parker and Waters showed that dimerisation alone does not activate the receptor. When they compared the crystal structure of the extracellular domain in the absence of its ligand, hGH, and with the ligand bound to it, they found no significant change in conformation.

“Mike did some labelling work with fluorescent alanine residues that showed the HGH receptor is actually a dimer before the hormone binds,” says Parker. Their findings suggest that subunits of the dimerised receptor rotate relative to each other within the transmembrane domain, activating the receptor independently of hormone binding.

“Genentech got it wrong because they worked on the water-soluble form of the receptor, not the membrane-bound form,” says Parker.

Therapeutic visions

In a collaboration with researchers at the Ludwig Institute for Cancer Research in Parkville, Parker resolved the structure of the granulocyte-macrophage colony stimulating factor (GM-CSF) receptor complexed with its ligand, GM-CSF, which Ludwig researchers Ashley Dunn, Tony Burgess and Nick Gough discovered in 1983.

“The surprising thing is that the receptor forms large oligomers on the surface of the cell, or even complex networks, so when GM-CSF binds to the receptor, it induces a lot of signal,” says Parker. “Problems arise because the GM-CSF receptor is closely related to the IL-3 and IL-5 interleukin receptors – they share subunits.”

Acting through their cognate receptors, the various ligands determine whether haemopoietic cells will survive or apoptose, proliferate, differentiate, migrate, or perform defensive functions like phagocytosis or releasing oxygen radicals to destroy pathogens.

Parker says problems arise when ligands bind the wrong receptors, activating signalling pathways that can cause chronic inflammatory disorders like asthma and rheumatoid arthritis, or dysregulating cell replication and division to cause leukaemia.

---PB---

By resolving details of the structure and activity of each receptor, Parker says it should be possible to design therapies that will selectively activate or block their activity. “This has been my big passion over the past decade or so,” he says.

“From my early interest in toxins, I went to membrane proteins, to receptors that signal across membranes, to structure-based drug discovery, where you use structural information to design small therapeutic molecules or antibodies.”

Parker has been working with Dr Siew Yeen Chai of the Florey Neuroscience Institute to resolve the structure of the Insulin-Regulated Aminopeptidase (IRAP) neuronal receptor, that Siew’s team discovered half a decade ago.

In 2005, Siew’s team discovered two peptides that enhanced short-term memory in maze-running mice, by suppressing the activity of IRAP, a natural inhibitor of short-term memory.

Parker resolved the structure or IRAP and its natural ligands, and Siew’s team used the data to design two synthetic peptides that are at least 10 times more potent than IRAP inhibitors, markedly improving short-term memory in mice.

The IRAP inhibitors are candidate memory-enhancers for improving memory and cognitive function in Alzheimer’s patients, or stroke patients with memory loss. Parker says IRAP inhibitors might also be useful enhancers of memory and cognition in healthy individuals.

In other projects, Parker has worked with the Cooperative Research Centre for Cancer Therapeutics, to develop lead compounds targeting focal adhesion kinase (FAK), a cellular adhesion molecule whose inactivation is a prelude to metastasis.

He has also worked with Biota Holdings, which developed the anti-influenza drug Relenza, looking for conserved structures in the coat proteins of hundreds of rhinovirus serotypes, that cause common colds. Biota has a potential generic ‘plug drug’ in Phase II clinical trials.