Holding such a heretical view famously cost one of France’s top allergy researchers, Jacques Benveniste, his funding, labs and reputation after his findings were discredited in 1988.

Yet a paper is about to be published in the reputable journal Physica A claiming to show that even though they should be identical, the structure of hydrogen bonds in pure water is very different from that in homeopathic dilutions of salt solutions. Could it be time to take the “memory” of water seriously?

The paper’s author, Swiss chemist Louis Rey, is using thermoluminescence to study the structure of solids. The technique involves bathing a chilled sample with radiation. When the sample is warmed up, the stored energy is released as light in a pattern that reflects the atomic structure of the sample.

Twin peaks

When Rey used the method on ice he saw two peaks of light, at temperatures of around 120 K and 170 K. Rey wanted to test the idea, suggested by other researchers, that the 170 K peak reflects the pattern of hydrogen bonds within the ice. In his experiments he used heavy water (which contains the heavy hydrogen isotope deuterium), because it has stronger hydrogen bonds than normal water.

Aware of homeopaths’ claims that patterns of hydrogen bonds can survive successive dilutions, Rey decided to test samples that had been diluted down to a notional 10^-30 grams per cubic centimetre – way beyond the point when any ions of the original substance could remain. “We thought it would be of interest to challenge the theory,” he says.

Each dilution was made according to a strict protocol, and vigorously stirred at each stage, as homeopaths do. When Rey compared the ultra-dilute lithium and sodium chloride solutions with pure water that had been through the same process, the difference in their thermoluminescence peaks compared with pure water was still there (see graph).

“Much to our surprise, the thermoluminescence glows of the three systems were substantially different,” he says. He believes the result proves that the networks of hydrogen bonds in the samples were different.

Phase transition

Martin Chaplin from London’s South Bank University, an expert on water and hydrogen bonding, is not so sure. “Rey’s rationale for water memory seems most unlikely,” he says. “Most hydrogen bonding in liquid water rearranges when it freezes.”

He points out that the two thermoluminescence peaks Rey observed occur around the temperatures where ice is known to undergo transitions between different phases. He suggests that tiny amounts of impurities in the samples, perhaps due to inefficient mixing, could be getting concentrated at the boundaries between different phases in the ice and causing the changes in thermoluminescence.

But thermoluminescence expert Raphael Visocekas from the Denis Diderot University of Paris, who watched Rey carry out some of his experiments, says he is convinced. “The experiments showed a very nice reproducibility,” he told New �Ӱ�ԭ��. “It is trustworthy physics.” He see no reason why patterns of hydrogen bonds in the liquid samples should not survive freezing and affect the molecular arrangement of the ice.

After his own experience, Benveniste advises caution. “This is interesting work, but Rey’s experiments were not blinded and although he says the work is reproducible, he doesn’t say how many experiments he did,” he says. “As I know to my cost, this is such a controversial field, it is mandatory to be as foolproof as possible.”

Polythene is made of chains of ethylene, also called ethene. In the presence
of a catalyst, these molecules join together to form polythene. The longer the
chains, the tougher the plastic. High-density polythene is used for containers
and pipes, while low-density polythene, which is more flexible, is used to make
bags and packaging.

The catalyst developed at Imperial College consists of an atom of iron
bonded to two chlorine atoms and surrounded by nitrogen and
carbon. The aluminium co-catalyst replaces one of the chlorine atoms
with the carbon of a methyl radical. This iron-carbon bond is highly reactive
and the ethylene inserts itself between the iron and the carbon. The new
molecule still has an iron-carbon bond, so another ethylene inserts itself so
the polymer grows outwards from the iron centre.

The carbon atoms around the iron allow the chemists to control the reaction.
“The beauty of these new iron catalysts is that by modifying the groups around
the iron centre, we can control the activity of the catalyst and select the size
of the giant polythene molecules we want,” says Vernon Gibson, who heads the
team of chemists at Imperial College. “This new iron-based catalyst system is
relatively cheap, readily available and more environmentally friendly than some
of the alternatives”.

Gibson is due to give details of his work next week at a Royal Society of
Chemistry meeting in Durham.

There are significant advantages in using local rather than general
anaesthetics in operations. Patients recover more quickly without the “hangover”
that general anaesthetics cause. And the risk of infection is greater with
general anaesthetics because they suppress the immune system. They can even
cause metabolic changes that lead to wasting and clogged arteries.

Many minor operations—for example, dental procedures or varicose vein
removal—are carried out using local anaesthetics such as lignocaine. But
these drugs are not sufficiently long-lasting for major operations taking
several hours. Bupivacaine lasts up to 50 per cent longer than lignocaine, but
its applications are limited because it can cause dangerous side effects, such
as arrhythmia and convulsions.

“This makes administration of bupivacaine a highly-skilled procedure,” says
Jonathan Bannister, an anaesthetist at Ninewells Hospital in Dundee. “To do a
neural block—like an epidural, which deadens the lower part of the
body—we have to get the drug very close to a nerve. As many nerves are
accompanied by blood vessels, it isn’t difficult to miss the nerve and hit a
blood vessel. That means the patient then gets a dose of bupivacaine directly
into the bloodstream, which can cause cardiac complications, such as arrhythmia
and lowered blood pressure and, in extreme cases, death.”

But Chiroscience realised that only one form of the bupivacaine molecule
causes these problems. Like many biologically-active molecules, the drug
exists as right and left-handed isomers. Both are anaesthetics, but one causes
most of the dangerous side effects. Levobupivacaine, the left-handed form, is
the safer isomer.

Bannister has been investigating the potential advantages of levobupivacaine.
He says: “This will give us an added safety margin. It means we can put in more
of the drug for longer periods.”

Chiroscience expects the new drug to be available in Britain within two years
and in the US in three.

Liquid crystals incorporating metal are not new, but so far only a handful
of metals have been used. According to a team of British chemists, the new
molecule opens up the possibility of exploiting as many as 60 metals in
liquid crystals. This is important because many metals form coloured and
magnetic compounds, so it will be possible to make liquid crystals which
are coloured and which can be switched on and off magnetically – both valuable
properties in commercial applications.

Liquid crystals are a state of matter between a crystalline solid and
a liquid. In a solid, the molecules are arranged in ordered lattices, which
means their physical properties – such as refractive index and electrical
conductivity – depend on the direction of measurement. However, the physical
properties of a liquid are the same in all directions because the molecules
are in perpetual random motion.

Liquid crystals are fluids that retain some of the order of solids
because their molecules are either rod-like (cala-mitic) or disc-like (discotic).
For example, in the so-called nematic phase of a calamitic liquid crystal,
the molecules will on average be pointing in the same direction. If, in
addition, there is some order in the positions of the molecules, this constitutes
a so-called smectic phase.

A liquid crystal’s properties are therefore said to be highly directional.
In a calamitic liquid crystal, for example, the refractive index depends
on whether the refraction is measured along or across the direction in which
molecules point, a property called birefringence.

The addition of metals to calamitic liquid crystals can enhance their
usefulness in a number of ways. Metals are highly polarisable – that is,
they have high electron densities that can be easily distorted in an electric
field. Because the stability of liquid crystal phases is thought to depend
on their ability to polarise, incorporating metals into liquid crystals
could increase their stability. The higher polarisability of a metal would
also increase birefringence in the liquid crystal, a property related to
the quality of viewing of crystals in displays.

Also, many metal complexes are coloured; so, if the metal forms a complex
with a calamitic liquid crystal, coloured liquid crystal displays can be
made without costly filters. And, as many metals are paramagnetic – that
is, they have unpaired electrons which respond to external magnetic fields
– it should be possible to switch such metal-containing liquid crystals
magnetically as well as electrically, so that liquid crystals could function
in magnetic memory devices.

However, before liquid crystal technology can reap all these benefits,
there is a snag to be overcome: most metals do not form calamitic liquid
crystal complexes. Metals that do maintain the linear rod-like geometry
needed for a liquid crystal include nickel, palladium and platinum, copper,
silver and gold, rhodium and iridium – a small group. The rest of the 60-odd
metals in the periodic table, many of which form coloured and paramagnetic
complexes, are happier forming tetrahedral or octahedral complexes, both
of which are the wrong shape for rod-like liquid crystal molecules.

Now Duncan Bruce and his team at the University of Sheffield have overcome
the problem. They have grafted a manganese compound onto the side of a rod-like
liquid crystal molecule in a well-defined smectic phase. Though the geometry
around the manganese cation (positively charged ion) is octahedral, the
original liquid crystal molecule is so long and large that the disruption
caused is minimal. The smectic phase of the rod-shaped molecule is exchanged
for a less orderly nematic liquid crystal phase with the manganese complexed
(see Figure).

The work will be published in the March issue of The Journal of The
Chemical Society, Chemical Communications.

The mountaineering metaphor is apt, though. Vitamin B12 is the most
complicated of an essential group of bio-pigments which includes haem, the
red pigment in haemoglobin, and chlorophyll, the green pigment in plants.
Physical danger may be minimal, but the task of unravelling how nature builds
a molecule as complicated as B12 has had a similar pattern of forward surges,
punctuated by long periods at ‘base camps’ waiting for the intellectual
weather to change. Now, developments in molecular biology and genetic engineering
have put an Anglo-French team of B12 ‘climbers’ in sight of the summit.

Since 1926, it has been known that large quantities of partially cooked
liver can cure pernicious anaemia, a disease which seems to be unique to
humans and is often fatal. Among its symptoms are a deficiency in red blood
cells and in the formation of haemoglobin, and severe damage to the central
nervous system. In the 1940s, vitamin B12 was identified as the substance
in liver that acts against pernicious anaemia, and it was isolated in crystalline
form. A decade later, the crystallographer Dorothy Hodgkin resolved B12’s
complex structure using X-ray crystallography. Ever since then, chemists
have been trying to work out the secrets of how this structure is made.

But, as Battersby explains, the reason for rising to this challenge
is not simply ‘because it is there’. ‘Vitamin B12 is the most complex single
molecular species made by any organism, and we want to know how on earth
a little microbe manages to put such an exquisitely complicated structure
together,’ he says. There are wider implications too: ‘Chemistry is the
enabling science. In understanding how microbes make B12 we are learning
new ways of doing chemistry, for example, making carbon-carbon bonds – the
essence of organic chemistry – and large ring compounds. New chemistry leads
to new compounds and new materials, and better ways of making them.’

Vitamin B12 is made only by a few species of microorganisms. Many of
these live symbiotically in the large intestines of animals, which is how
these animals obtain their daily B12 needs. Rabbits periodically eat their
own faeces for this purpose. For the human body to absorb it in the small
intestine, the vitamin has to form a complex with a glycoprotein secreted
by the stomach. If there is no glycoprotein, the vitamin cannot be absorbed,
and the deficiency results in a build-up of a substance called methylmalonyl
coenzyme A. This in turn affects the production of fatty acids, components
of the myelin sheath that surrounds nerve fibres. The myelin degenerates,
resulting in neurological disorders. But the human liver normally stores
enough vitamin B12 to last for years, so true dietary deficiency is extremely
rare.

Vitamin B12’s main role is in the metabolism of amino acids, molecules
which are the building blocks of proteins. Amino acids consist of an amino
group (-NH₂) and a carboxylic acid group (-COOH) separated by
a carbon atom that also carries a hydrogen atom along with another group
of atoms, which varies in each type of amino acid. Vitamin B12 is involved
in the conversion of homocysteine to another sulphur-containing amino acid
called methionine by adding a CH₃. It is also involved in the
breakdown of some branched-chain amino acids. Moreover, B12 acts as a helper
molecule or ‘coenzyme’ for enzymes that transfer a hydrogen atom from one
carbon to an adjacent one, an important process in many biochemical reactions.

B12 is one of a family of bio-pigments called the tetrapyrroles, which
Battersby calls the ‘pigments of life’ (see figure opposite and Box 1).
They include the porphyrin-based ring systems of haem, which is essential
to the biochemistry of molecular oxygen; chlorophyll and bacteriochlorophyll,
which are crucial in trapping solar energy during photosynthesis; and sirohaem,
a substance in organisms such as Escherichia coli which is responsible for
converting sulphite to sulphide and nitrite to ammonia. Then there is coenzyme
F430, which is responsible for methane generation by certain bacteria. Finally,
there is the smaller corrin ring system of vitamin B12. All these molecules
have the same basic structure and it was this realisation which started
chemists on their 50-year journey towards understanding how they were made.

An American chemist, David Shemin, discovered in 1945 that glycine is
the starting point for the biosynthesis of the tetrapyrroles in animals.
He swallowed 66 grams of glycine labelled with the heavy isotope of nitrogen,
nitrogen-15, over three days. During this period he took regular blood samples,
extracted the haem and measured its enrichment with nitrogen-15 using mass
spectroscopy. Shemin found that the level of nitrogen-15 rose sharply,
remained constant for about four months, and then dropped steadily back
to the its natural level. Four months is about the lifetime of a red blood
cell in the body, so he concluded that glycine was being incorporated into
the haem during its biosynthesis. Chemists now know that glycine provides
the four nitrogen atoms found in haem’s porphyrin rings.

Some chemists, including Albert Eschenmoser at the University of Zurich,
think that molecules based on the corrin ring system of vitamin B12 could
be older in evolutionary terms than the porphyrin-based ring systems in
haem and chlorophyll. This is because there are anaerobic bacteria that
produce corrins, but none that generate porphyrins, and anaerobic bacteria
are older in evolutionary terms than aerobic ones. Also, the biological
task of corrin-type cofactors is to speed up biosynthesis, whereas porphyrins
are involved in energy metabolism, a set of reactions that ultimately will
not work without an aerobic atmosphere.

Certainly, the biosynthetic pathways of the plant, animal and bacterial
pigments soon parted company with their older bacterial and pre-bacterial
forbears. The split occurs at the molecule called uro’gen III (see Box 2
overleaf). In animals the path from here onwards involves a series of steps
eventually leading to haem, while in plants it leads to chlorophyll. In
single-celled creatures such as the bacteria Propionibacterium shermanii
and Pseudomonas denitrificans, a different set of processes leads to vitamin
B12.

Organic chemists set about discovering what these processes were in
1968. They began by working out how to construct from simple building blocks
the cobalt corrin ring that lies at the heart of B12. But these routes turned
out to bear little resemblance to the way nature goes about it. Then Battersby,
and others including Ian Scott at Texas A&M and David Shemin in the
US, Gerhard Muller at the University of Stuttgart and Diulio Arigoni in
Switzerland began to unravel the route, step by laborious step. For each
organism, they had to isolate each intermediate and work out both its structure
and the way it is converted to the final product. This meant isolating
large enough quantities of each intermediate.

They managed this trick by stopping the organism partway through the
biosynthetic process. For example, making vitamin B12 involves going from
uro’gen III to a corrin ring system with one less carbon atom and more double
bonds. To get there the process initially goes through three ‘precorrins’
which are numbered according to how many methyl (CH₃) groups
they have gained since the uro’gen III stage. To stop biosynthesis at precorrins
1, 2 and 3, they excluded cobalt from the organism’s diet.

Another way to generate large quantities of the intermediate is to get
at the enzyme that produces it. In the early days, this was not easy. ‘We
would have to grow an organism over a period of three months, harvest it
and, if we were lucky isolate three to four milligrams of the enzyme from
a huge amount of green stuff,’ Battersby recalls.

Once chemists had worked out what they thought was the structure of
the intermediate, they had to chemically synthesise it in a way that labelled
specific atoms. Then they fed the labelled intermediate to the purified
enzyme (or the organism). If the structure was correct, the label appeared
in the final product. In this way, the chemists were able to tell exactly
what manipulations the enzyme or the organism had performed on the intermediate.
‘In the early days, we used radioactive labels, such as carbon-14,’ says
Battersby. ‘This meant chemically cutting up the final compound and counting
the radioactivity in the bits. If we had stayed with that technique, we
could never have made the progress we have with B12.’

Two developments proved to be crucial to further progress. The first
was reliable nuclear magnetic resonance (NMR) spectrometers that allowed
chemists to detect carbon-13. This isotope of carbon is present naturally
in minute quantities – 1.1 per cent – in all living material. Like hydrogen,
carbon-13 has a nuclear spin of one-half, so it can be detected by NMR.
Battersby’s team managed to build in carbon-13 labels during the synthesis
of prospective intermediates to B12. This meant that the chemists no longer
had to break down the final products in order to determine where in the
molecule the labelled atoms were. With this powerful new tool, Battersby
and others made great strides in working out the pathways to the pigments
of life.

Most of the advances in chemists’ knowledge of B12 at this time came
from experiments on the organism P. shermanii. Growing this organism in
a medium free of cobalt gave the three intermediates called precorrins 1,
2 and 3, according to how many methyl groups were added at each step. But
everyone in the field at that time found it difficult to get past the ‘base
camps’ of precorrins-1, 2 and 3. Researchers were held up at this point
for some eight years. Something else was needed.

The answer was genetic engineering. In 1989, French molecular biologists
Joel Crouzet, Francis Blanche and their teams at the Rhone-Poulenc-Rorer’s
research centre outside Paris, cracked the genetic code of another vitamin
B12-producing organism called P. denitrificans. RPR is the world’s leading
producer of B12, which is used to boost meat pro-duction in poultry and
cattle. In 1990, Battersby teamed up with Crouzet and Blanche and their
colleagues. The French team had already detected and sequenced the genes
that code for the various enzymes that synthesise B12 in P. nitrificans.
This allowed them to tinker with the organism’s genetic code so that the
organisms overproduced these enzymes, making them easier to purify and
so produce many new intermediates. ‘What used to take three months and
gave only 3 to 4 milligrams of enzyme, we can now do overnight and produce
much more,’ says Battersby. Used together with spectroscopy, intermediates
beyond precorrin-3 were produced.

In 1990, the French team isolated precorrin-6x, a compound several steps
on from precorrin-3. Using carbon-13 NMR spectroscopy, the French and British
teams worked out the com-pound’s structure. To their surprise they found
that ring shrinking, which chemists had believed happened later on in the
pathway, had already occurred. Later came a series of steps, all triggered
by one enzyme, that led to precorrin 8x. The final step – leading to the
corrin ring system of vitamin B12 – involved the transfer of the methyl
group on carbon atom 11 to its neighbour on carbon atom 12. This left them
with another piece to fit into the puzzle: what had happened to the methyl
group at carbon-20 in pre-corrin-3? It was not there in precorrin-6x. The
answer came earlier this year, when yet another intermediate, precorrin-4,
was isolated by the French team; it had been converted to an acetyl (CH₃CO-)
group attached to carbon 1.

Finally, thanks to the efforts of Muller and Scott; the French team
at RPR; and Battersby’s Cambridge group, chemists now know that the cobalt
atom appears at the centre of the vitamin’s ring at a different stage in
the pathway, depending on whether the particular B12-producing organism
is anaerobic or aerobic. If anaerobic, cobalt appears earlier.

Like any good mountaineer, Battersby relishes all of the milestones
that he has passed in his 25 years of work on B12. Plenty of midnight oil
has been burned en route, and long weekends spent waiting patiently for
microorganisms, enzymes and NMR spectrometers to do their work. ‘What keeps
us all going, even through the lean times, is an intense, single-minded
interest in the problems of B12 biosynthesis,’ he says.

Perhaps the milestone that gave him the most satisfaction, and caused
the greatest sensation, was the French discovery of precorrin-6x. ‘Many
of our ideas about B12 biosynthesis had to be changed. For example, we used
to think that ring contraction to give the corrin system occurred quite
late in the pathway. That had to go. Ring contraction is now seen as an
early step.’ He also cites the steps that give precorrin-4, precorrin-8
and hydrogenobyrnic acid, the precursor to the final corrin ring.

Battersby is at pains to emphasise the importance of the cross-disciplinary
collaboration that has brought all these fresh insights since 1990. ‘One
hand washes the other,’ he explains. ‘Chemists got as far as they could
go with precorrins-1, 2 and 3, which encouraged the biologists. This in
turn has enabled chemists to detect the new biosynthetic intermediates,
such as precorrin-6x.’ Battersby thinks every step in the pathway will be
known before the century is out. He estimates that they have now pinned
down 90 per cent of the links in the B12 pathway: ‘There are now only a
couple of steps to be filled in, notably between precorrin-3 and 4 and between
4 and 6x,’ he says. ‘Our goal now is to finish the climb to the summit.’

Battersby ‘retired’ late last year from his chair at Cambridge, but
is still collaborating with the French biologists. He is already looking
forward to more detailed work on precorrin-4 and the discovery of the elusive
precorrin-5. Like Everest of 40 years ago, the assault on B12 is nearly
over.

Lionel Milgrom is a lecturer in inorganic chemistry at Brunel University
and a freelance science journalist.

* * *

1: Running rings round the pigments of life

The anatomy of all naturally occurring tetrapyrroles is essentially
the same. Haem, for example, consists of four pyrrole-type rings joined
by carbon bridges to form a larger ring system called a macrocycle which
consists of 20 carbon atoms and 4 nitrogen atoms. The easiest way to remember
how the carbons are numbered is to remember the bridging carbons. The topmost
is carbon atom 5 (C-5), the right-hand one is C-10, the bottom one is C-15,
and the left-hand one is C-20. The rest follow.

The way other groups are arranged on the macrocycle shows the family
relationship between the pigments of life. The outermost groups also fine-tune
the properties of macrocycles. For example, the vinyl groups in haem, among
other things, help to stabilise the iron (II) oxidation state by acting
as a repository for the iron’s electrons. In chlorophyll, modification of
one of the side chains helps to stabilise the reduced chlorin ring in the
presence of oxygen (chlo-rins without such rings are easy to oxidise).

The crowding of groups around the shrunken cobalt corrin ring of vitamin
B12 allows nature to have its chemically reactive cake and eat it. The cobalt-carbon
bond of B12 is easily broken to give a radical – a molecule with one unpaired
electron. Such radicals are usually highly reactive, but with high reactivity
usually comes indiscrimination. By fencing off the reactive centre of B12
with lots of groups, the molecule keeps its high reactivity and specificity.

* * *

2: Mapping out the route

All the steps to vitamin B12 are tightly controlled by enzymes, which
are themselves coded in the genes of an organism’s DNA. This control is
the envy of human chemists. To reproduce the range of reactions of the lowliest
microbe would need a chemical works the size of Britain. The route to vitamin
B12 begins with 5-aminolaevulinic acid, ALA. In animals, ALA is synthesised
from the amino acid glycine and succinic acid, that has been made more reactive
by joining to coenzyme A or CoA, for short.

The next three steps are the same for all the pigments of life. Two
molecules of ALA are joined together to give a pyrrole – a five-membered
ring consisting of four carbon atoms and one nitrogen atom – called porphobilinogen
or PBG. This reaction is similar to that used by chemists to make pyrroles
in the laboratory. Next, four molecules of PBG are ‘bolted’ together, to
give an open chain made of four pyrroles, called a hydroxymethylbilane,
in which the four PBG units have been brought together nose to tail.

In the next step the hydroxymethylbilane ‘eats its own tail’ and turns
into a molecule whose name is abbreviated to uro’gen III. Something strange
happens during this step. One of the PBG units is turned round so that the
order of two large groups is reversed. In the same laboratory experiment
done without one of them, all the PBG units are still nose to tail, giving
uro’gen I.

Why should an organism go to the trouble of producing the type III uro’gen?
No one knows, but the chemistry of PBG gives a clue. In acid, four PBGs
join up to give four possible uro’gen isomers, but mostly type III. If the
prebiotic chemical environment of the Earth contained molecules similar
to uro’gens, then statistically half of them would have been type III. The
first biological systems probably took up these structures as ‘proto-vitamins’
or catalysts, and would have found it easier to make do with what was around
than to have to ‘discover’ how to make them on their own. So the most abundant
isomer was probably genetically ‘fixed’, copied biosynthetically, and its
structure gradually improved.

From uro’gen III the path splits into two. In plants and animals, it
leads to chlorophyll and haem. In the other pathway leading to vitamin B12,
at least eight methyl groups are added to uro’gen III, which shrinks to
give a 19-membered corrin ring with an atom of cobalt at its centre.

Two Dutch chemists have made a simple molecular clip that can mimic
the way in which complex enzymes bind to other molecules. The clip can grip
a molecule only with the aid of potassium ions. This means that it acts
in a similar way to large proteins involved in important biological processes,
such as oxygen transport and enzyme activity. These proteins use a metal
ion or some other species to ‘lock’ themselves into the best possible shape
to bind another molecule or ‘substrate’.

A protein molecule that changes its shape in this way is said to be
acting ‘allosterically’. Allosteric effects are responsible for making enzymes
such good catalysts. They ensure that an enzyme can ‘recognise’ and bind
to its substrate, transform it, release the product, and then be able to
repeat the process within a short time, usually a few millionths of a second.

Roeland Nolte and Rint Sijbesma of the University of Nijmegen in the
Netherlands, started with a compound that contained napthalene groups,
then modified it (Journal of the American Chemical Society, vol 113, p 6695).
A peculiarity of their molecule is that it changes constantly between three
different shapes, known as ss, sa and aa.

The aa form is clip-shaped, enabling the compound to bind another molecule
called 1,3-dinitrobenzene, or DNB. This form is sandwiched between its naphthalene
groups. In solution, most of the molecules exist in the sa form and only
a small fraction exist in the clip-shaped aa form.

Nolte and Sijbesma tinkered with the structure of their compound to
make a related compound with an aza crown-ether ring attached to each naphthalene
group. This new compound also exists mainly in the sa form. However, the
aza crown-ether rings are known to strongly bind metal ions such as potassium,
and when Nolte and Sijbesma added potassium ions, they found that the compound
switched from the conformation sa to aa.

They also found that in the presence of potassium ions the modified
compound was six times as good at binding DNB as it was without. The presence
of potassium induced allosteric binding to DNB.

The allosteric effect is not large, but it may point the way to simple
synthetic systems that are able to mimic the productivity of enzymes.

Ultrasound does several things to a solution. First, it agitates it,
mixing up chemical reactants much more efficiently than heating or stirring.
Secondly, ultrasound causes tiny bubbles to form and collapse repeatedly
in the solution. This process causes enormous changes in the pressure and
temperature at ‘hot spots’ in the solution. As the bubbles grow and collapse,
the temperature can change by thousands of millions of degrees per second.
Although the hot spots last for much less than a millionth of a second,
chemical reactions within the hot spots are speeded up. This happens despite
the fact that, overall, the temperature of the solution is little higher
than its surroundings.

When Suslick exposed the iron pentacarbonyl to ultrasound, he found
that the iron atoms became completely stripped of their carbonyl groups.
The changes in temperature caused by the hot spots are so great that the
iron atoms do not have time to order themselves into crystals, but instead
form an amorphous solid called a metallic gas (Nature, 3 October, p 414).

Amorphous iron powder which is created in this way consists of very
small particles, none of which is larger than 10 nanometres across. The
iron powder is such a good catalyst mainly because the very small particles
have a very large total surface area, according to Suslick.

He focused his attention on one particular reaction, known as the Fischer-Tropsch
reaction, which needs iron to work. In the reaction, carbon monoxide and
hydrogen combine over the iron catalyst to produce a mixture of hydrocarbons
which can be processed into petrol.

The Fischer-Tropsch process was invented during the Second World War
by the Germans in order to counter the oil shortage imposed by the Allied
blockade. South Africa has also developed the technology to implement the
reaction in order to defeat the embargo imposed by the rest of the world.
The technology is expensive but as oil stocks dwindle it could become economic,
turning nations with large coal deposits into major manufacturers of petrol.
Any catalyst that makes the Fischer-Tropsch reaction work more efficiently
could, therefore, be very important.

Suslick found that at 200 °C – a lower temperature than normal for
Fischer-Tropsch chemistry – his amorphous iron could convert carbon monoxide
and hydrogen into hydrocarbons 10 times as effectively as the polycrystalline
iron powders usually used. He also found that amorphous iron was 30 times
as active as crystalline iron during the process of converting cyclohexane
to benzene and methane – another industrially important pair of reactions.