Lurking in a corridor in the Sydney HQ of Australia's Nuclear Science and Technology Organisation (Ansto) is a crocodile called Sweetheart. Thankfully, he doesn't bite. In fact, it's just a model of a five metre-plus beast that once had a penchant for attacking outboard motors, but this one stays nice and still as Keri Hammerton demonstrates the spot on the neck that's best for taking blood.

Hammerton is a croc enthusiast. She might work for a nuclear science organisation, but that's only a bonus, because the nuclear imaging kit that she and colleague Chris Garvey have in their labs is perfect for studying the physical properties of crocodile blood. Croc blood has some unusual, and desirable, attributes. And they hope that, ultimately, their work will lead to a good artificial blood substitute.

Researchers have been looking for a blood substitute for decades. About one in five people who go into hospital need a blood transfusion - but donated red blood cells have a shelf-life of only about 42 days. Then there's the problem of blood group-matching. Shortages of donated blood, and particularly of certain types, are common.

The blood running through your veins is a cocktail of red blood cells, white immune cells, and other ingredients. But as far as a blood substitute is concerned, the vital part is the haemoglobin - the molecule that carries oxygen around your body. Haemoglobin in blood is packaged inside red blood cells. "But it would be incredibly difficult to engineer artificial cells with haemoglobin inside," says Garvey. Take human haemoglobin outside a red blood cell, though, and it breaks up into its constituent units. Even worse, when these sub-units reach the kidney for excretion, they're just the right size to clog it up and cause renal failure.

There are two artificial blood products that are in the final stages of Food and Drug Administration approval in the US. PolyHeme, produced by Northfield Laboratories in Chicago, is based on purified haemoglobin taken from donated blood. Hemopure is made using haemoglobin from cows. Both get round the problem of isolated haemoglobin breaking up by attaching the molecule to a chemical that holds it together. These new products can be given to anyone, regardless of blood type. "But both suffer from shortages and potential contamination with viruses and prions," says John Olson, an expert on blood substitutes at Rice University in Houston, Texas.

He thinks the best option will be to genetically modify bacteria to produce human haemoglobin. "Recombinant haemoglobin represents an unlimited and reliable supply that can be produced on an industrial scale using bioreactor technology."

But if you're going to get bacteria to produce human haemoglobin, why not modify that haemoglobin so it doesn't break up when isolated, or make a product that will more readily give up its oxygen, so you need less to treat a patient?

This is where the crocodiles come in. Crocodiles can stay underwater for an hour at a time, and in the 70s, German researchers discovered part of the reason for this: bicarbonate ions bind to crocodile haemoglobin and force it to give up more of its oxygen than haemoglobin in human blood.

Then, in the early 1990s, researchers at the MRC laboratory of molecular biology in Cambridge made hybrids of crocodile and human haemoglobin to study bicarbonate binding. They found that they didn't need to alter the human form a great deal to achieve this effect.

Now, at Ansto, Hammerton and Garvey are looking more closely not only at the biochemistry, but also the physical properties of crocodile haemoglobin. Every few months, staff at the Crocodylus Park, a crocodile farm in the Northern Territory, draw blood from a saltwater croc. They fly this overnight to Sydney, and Garvey takes it to the lab.

The haemoglobin is far too small to study under a microscope. So Garvey has been using a standard scanning technique called nuclear magnetic resonance imaging, plus another technique called neutron scattering, to study living blood cells.

By looking at what happens to a beam of neutrons or x-rays when they pass through the cells, he can probe their structure and pinpoint the haemoglobin. This technique has also revealed that isolated crocodile haemoglobin doesn't break apart. In fact, the molecules clump together. "It naturally aggregates," says Hammerton. "This would be a great property for blood substitute haemoglobin to have."

The croc haemoglobin aggregates via sulphur groups on the surface of the molecule. "Crocodiles have the highest proportion of these sulphur groups of any vertebrate - and they're actually antioxidants," Hammerton says. She suspects the reason they have so many is that crocodiles spend long periods of time under water, without oxygen. When they surface, the influx of oxygen in their body produces a vast quantity of damaging oxygen free radicals, which the groups mop up.

There are circumstances when an abundance of these groups could be just what a patient needs. During surgery - say on a liver or a heart - the blood supply is cut, and the tissue becomes oxygen deficient. When blood is then infused, the influx of oxygen produces free radicals, and can cause tissue damage. "So if you used a substitute haemoglobin with lots of these sulphur groups, you might reduce tissue damage," Hammerton says.

So, the future of artificial blood could be variants of isolated human haemoglobin, each designed for a specific medical situation. Producing the haemoglobin in bacteria would get around the problem of supply, and of transferring viruses from donor to recipient. The ideal artificial blood would go a stage further.

The ideal would be some sort of freeze-dried powder to which you could just add water. This could easily be stored in countries without good refrigeration, or shipped in bulk to a disaster or war zone. Haemoglobin is a pretty resilient molecule, so this isn't too far-fetched, says Hammerton - although dispersing the haemoglobin properly in the solution could be tricky.

Another potential hurdle is that human blood can't withstand being too concentrated. So Garvey is using the nuclear imaging equipment to examine the blood of camels, which experience varying blood salt concentrations. "It turns out that camel and human red blood cells are quite different - camel red blood cells hold on to water much more strongly," he says.

It's not clear yet whether this will help in the development of an artificial blood powder. But looking at animal blood, and perhaps borrowing some of their adaptations, should help the quest for the ultimate human blood substitute.

Further reading

www.ansto.gov.au
Australian Nuclear Science and Technology Organisation

www.northfieldlabs.com/
polyheme.html Northfield Labs' PolyHeme blood substitute home page and background information

www.biopure.com
Biopure homepage and details on Hemopure

http://wmi.com.au/crocpark
Crocodylus Park research centre