Rhomboids give biological heresy a cold shower

Proteases can only perform their high-energy, protein-cleaving reactions in the presence of water. Cell membranes comprise a double layer of lipids that strongly exclude water. The discovery of intra-membrane proteases like rhomboids was biological heresy.

But last October, the international journal Nature published the first crystal structure for a rhomboid, which explained how they spirit water into the hydrophobic membrane.

Dr Matthew Freeman, of the Medical Research Council Laboratory of Molecular Biology in Cambridge, is a speaker at this month's Lorne conference on Protein Structure and Function.

Biophysicists were slow to accept claims of intra-membrane proteolysis as something more than a geneticist's fantasy. How could a hydrophilic enzyme cleave the lipophilic trans-membrane domains of other proteins if they are embedded in a hydrophobic lipid bilayer?

Freeman concedes that rhomboids are "mechanistically weird" enzymes, but didn't accept that lack of water in the cell membrane was an insoluble problem for evolution.

Of the Nature paper, he said, "We now know it's not a fantasy. They're real enzymes."

The frisson surrounding rhomboids' potential as drug targets stems from their apparently ancient role in cell-to-cell signalling.

Freeman says rhomboids are one of four families of intra-membrane proteases with similar function - they cleave proteins and release the resulting fragments from the membrane.

But each family seems to have evolved independently - they are not obviously related by structure, peptide sequence or evolution.

The others include presenilins, which, among other things, are secretases that cleave the trans-membrane domain of the Alzheimer's precursor protein (APP), releasing beta-amyloid fragments that aggregate into neurotoxic plaques in the brain.

Structures have not yet been reported for the other three families of intra-membrane proteases, so we don't know whether they have solved the problem of importing water into the hydrophobic membrane in the same way as rhomboids.

But as they all have trans-membrane domains themselves, Freeman speculates that they all create a hydrophilic micro-environment within which their cleavage reactions occur.

"That's the biochemistry, but in terms of the biology and medicine, the cleavage reaction removes a protein domain that tethers a protein to the membrane, and releases it into the intracellular or extracellular medium, so it can move somewhere else."

Rhomboid proteases

Freeman and his MRC colleagues discovered the first rhomboid protease in Drosophila. "Our approach was initially a genetic one, and this told us that rhomboids were key activators of epidermal growth factor receptor (EGFR) signalling," he says.

All eukaryotes investigated to date appear to have rhomboid proteases, and searches of DNA databanks show that, with a few exceptions, they are also present in prokaryotes - eubacteria and archaea.

"Of the 350 prokaryotic genomes sequenced to date, around 70 per cent have rhomboids. This has the smell that rhomboids were present in the last universal ancestor of life on earth."

After identifying the first rhomboid in Drosophila, the Freeman laboratory searched for others. The only other with a known function was in a bacterium, Providencia. Amazingly, it also appeared to be involved in cell-to-cell signalling.

"That made us wonder if all rhomboids would be involved in intercellular signalling, but the apparently similar functions in Drosophila and Providencia has turned out to be a coincidence," he says.

"So intercellular signalling is not the whole story - it's just that rhomboid-mediated cleavage reactions lend themselves to signalling roles. Rhomboids are also involved in a much wider range of physiologically important control processes."

Cells signal to each other in response to environmental or biochemical cues; the rhomboid system facilitates conditional signalling, an essential process for all life.

While there is no validated medical use for rhomboids yet, Freeman says there are four or five clinical contexts in which they could prove very useful drug targets. "It's early days, but the possibilities are exciting."

One of the first identified potential uses is in novel drugs to control the Apicomplexa parasites that cause malaria, cryptosporidiosis and toxoplasmosis.

All are obligate intracellular parasites, with specialised surface proteins that allow them to attach to host cells. Once attached, parasites then cleave the anchor protein, and Freeman says it now appears that they do so with their own rhomboid protease, allowing them allows them ghost through the membrane of the host cell.

"Final genetic proof is lacking, but it's looking very promising," he says. "Rhomboid inhibitors could be potential therapeutics for Apicomplexa parasitic infections."

The basic attraction of rhomboids is that enzymes make ideal targets for potent, small-molecule drugs. Ordinary protein-protein reactions involve large surface-area contacts that are not easily disrupted with small molecules.

The involvement of rhomboids in signalling by the EGF receptor - a key player in many cancers - suggests rhomboid-targeting drugs could also have a useful role as cancer drugs.

"The coolest potential application I've heard of is in disrupting quorum sensing in bacteria, which is the role of rhomboids in Providencia. But we need to understand a lot more about this before we know whether this is a realistic possibility".

(Quorum sensing is the process by which bacterial cells sense the presence of other bacteria; bacterial cells are typically non-pathogenic in low numbers, but above a threshold density, activate virulence genes.)

Freeman emphasises that all such clinical applications remain speculative, but says the rhomboid research field is rich with "intriguing possibilities".

"We have a very interesting new kind of enzyme mechanism, and biologically, rhomboids do lots of important things: it seems highly likely that they will also have medical significance."