The subject of sweetness is one that I have covered a number of times. Chemists have managed to overcome some of the problems associated with sugar by creating molecules that replicate sweetness.

Cover image of New Scientist published 19 June 1986The shape of sweeteners to come
This article was written with the help of Professor Leslie Hough whose research student Shashikant Phadnis discovered sucralose, which is now commercially available as Splenda. The intense artificial sweeteners most commonly used are sucralose, aspartame and acesulfame, the last two of these are often used in conjunction because the one intensifies the sweetness of the other so less of each needs to be used. The former most popular sweetener saccharin is also still available.

Anyone who can predict how sweet a molecule is will make a lot of money. Chemists are busy trying to find the sweet solution

Leslie Hough and John Emsley

Published: 19 June 1986

In ages past, when the only natural source of sugar was honey, it was a delicacy much appreciated. People began to cultivate sugar cane in Southern Europe only around AD 800, though it was known in India as long ago as 3000 BC. Sugar beet made its debut as a crop in Europe much later, in around AD 1800.

Today, the average person in the West consumes about 750 grams of sugar a week, mostly in the form of sucrose, the carbohydrate obtained from cane sugar or sugar beet. Honey, by contrast, is a mixture of carbohydrates. Part of its sucrose has split or "inverted" to form the simpler sugars, fructose and glucose. Fructose is even sweeter than sucrose. Golden syrup and treacle, like honey, also contain a proportion of inverted sugar.

Dentists, doctors and dieters often condemn sugar, which they see as the cause of decay, disease and disfigurement. Yet factual evidence for each of these claims is weak, and some scientists such as Vincent Marks, Professor of Biochemistry at Surrey University, deny any connection between sugar and heart disease. Yet the abuse thrown at sugar – "empty calories", "pure, white and deadly" – is partly justified. Although sugar is an important part of many people's diet, we often use this chemical purely for its taste. The trouble is that sugar is not particularly sweet. To offset the bitterness of tea, coffee and cocoa we need to add 5 to 10 grams of sugar per cup. To counterbalance the acidity of fruit and carbonated drinks we need even more.

The average person's weekly consumption of 750 grams of sugar provides about 12,500 kilojoules of energy. This level of intake seems folly if all we seek to do is to indulge our fondness for sweet-tasting food. If that were so, then it would be a simple matter to reduce our energy intake. Sugar is not a very sweet chemical, so all we would need to do is to find one that is much sweeter.

The Romans discovered the first-ever artificial sweetener. They called it sapa, a name that lends itself to contemporary words such as saporous, meaning sweet-tasting. According to Pliny (AD 23–74), people made sapa by boiling grape juice down in lead pans until the juice concentrated into a very sweet syrup. The underlying process involved the grape acids forming lead salts which can taste sweet. (One lead salt, lead ethanoate, Pb(CH3CO2)2.3H2O, used to be called sugar of lead.

The Romans used sapa to sweeten cooked foods and "improve" wines. People still practised the vintners' trick of adding a pellet of lead shot to each bottle of wine until fairly recent times. This method relied on the lead both to sweeten the wine and to prevent microorganisms from souring it. Sapa was also a favourite of Roman prostitutes: they found that it gave them a pale complexion (lead causes anaemia) and acted as an oral contraceptive (lead will cause abortions).

Chemists' more recent discoveries of new sweeteners have happened more by chance than by design, for sweetness is not a property that you can measure on a scientific instrument (see Table below). However, sweetness stems from molecular structure, so chemists have been able to construct molecules which have the necessary components built in to act as synthetic sweeteners. Unfortunately, our search for substitutes for sugar has had to meet demands not only that artificial sweeteners are safer than sugar, but that they are absolutely safe. This application of double standards has bedeviled attempts to introduce sweet-tasting chemicals other than sugar into our diet. Saccharin, however, managed to gain wide acceptance before it came under attack from this quarter.

Table: How other sweeteners measure up to sugar
Sugar 1 Excess calories; health risks
Cyclamate 30 Unproven links with cancer
Acesulfam 130 None
Aspartame 200 Short shelf-life; slight risk to people with phenylketourea
Saccharin 550 Bitter aftertaste; unproven links with cancer
Sucralose 650 None
Hernandulcin 1000 Slightly bitter
Thaumatin 3000 Lingering taste, unstable
*Chemists measure sweetness by dissolving the substance in water and asking a panel of tasters to compare the taste to a 4 per cent solution of cane sugar in distilled water. Sweetness factors vary considerably from person to person. Some people, for example, find saccharine 700 times sweeter than sugar.

The American chemist Ira Remsen and the Russian-born Constantin Fahlberg, at the Johns Hopkins University, discovered saccharin in 1879. It came into use as an artificial sweetener almost immediately and proved invaluable during the years of food rationing earlier this century. Curiously, this heavy human consumption led no one to suspect that it was anything other than harmless.

Admittedly many people find that saccharin has a metallic or bitter after-taste but few thought it dangerous, because the body cannot metabolise it, and excretes it unchanged. It is used in tiny amounts. A 15-milligram tablet has the same effect as 8000 milligrams of sugar, and saccharin comprises less than 10 parts per million of the average diet.

In 1937, Michael Sveda of E. I. DuPont de Nemours put the cigarette he was smoking on the edge of his bench. When he next put it to his lips he was struck by a sweet taste: he had discovered cyclamate. By the mid-1960s, this sweetener dominated the market in the US. In 1970 it was banned in Britain and North America. Rats fed the blend of cyclamate and saccharin most commonly used had shown a higher- than-normal incidence of bladder cancer. Cyclamate, by far the major component, got the blame. Saccharin, however, because of its chemical similarity to cyclamate, also came under suspicion. Further tests on rats showed that those fed on saccharin developed more bladder cancers than a controlled group, just as had happened with cyclamate. But the well-armed saccharin lobby quickly translated the dose fed to the rats – 3000 parts per million of their diet – into an equivalent consumption for humans of 800 cans of cola per day or almost 300 000 cans per year, so demonstrating how unrealistic such high exposure had been.

Supporters of cyclamate were less lucky, although some countries such as Australia found the evidence unconvincing. Some, such as Canada, which had initially followed the American lead, later lifted the ban. European manufacturers make and sell over 2000 tonnes of cyclamate each year, although they sell none to Britain. Researchers have carried out scores of studies on cyclamate. Some of these studies have indicated links to various cancers, but never convincingly. Even in the US there is doubt about the ban, and a committee of the US National Academy of Sciences that reviewed the evidence on behalf of the Food and Drugs Administration concluded that cyclamate by itself was safe, although mixed with saccharin it may cause cancers. Unfortunately, cyclamate is only a tenth as sweet as other artificial sweeteners so that relatively large amounts are needed.

Today's newest sweetener, aspartame, is probably as safe as chemists can make it. Sales of this compound now exceed $1 billion and G. D. Searle & Co. of Stokie Illinois, on the strength of their patent on aspartame, were taken over recently by Monsanto for $2.7 billion. James Schlatter, who was working on anti-ulcer drugs for the company, accidentally discovered the chemical's sweet taste in 1965. The Food and Drugs Administration gave aspartame approval in 1974, then banned it a year later in 1975, and finally lifted the ban in 1981. Despite its high cost it has gone from strength to strength, chiefly as a blend of aspartame and saccharin under the trade name NutraSweet.

Aspartame is not without its drawbacks. Chemically it seems above suspicion. It consists of two amino acids joined together; both of these are easily metabolised and one, phenylalanine, is essential. Curiously it is this very amino-acid that makes it necessary for aspartame to carry a health warning. One person in 15 000 has a genetic condition called phenylketonuria which becomes worse if he or she eats too much phenylalanine. By avoiding foods like meat, eggs, cheese and milk, which are rich in this amino acid, sufferers can limit their intake.

The other drawbacks of aspartame are its instability in water and its methyl group. The instability limits its use to foods that have a quick turnover such as soft drinks and fruit yoghurts. Aspartame decomposes at a rate of 10 per cent a month at room temperature. On heating, this breakdown speeds up, so that aspartame is useless for sweetening cooked foods. The methyl group constitutes no real hazard. Some researchers have pointed out that in the body the methyl group would form methanol-the dangerous alcohol beloved of Italian wine forgers. However, with aspartame, the amounts are so small that they pose no threat to health.

Despite some doubts about aspartame when it first came onto the market, and the disquiet that The Guardian newspaper showed when this sweetener first became available in Britain in 1983, no serious evidence has emerged to suggest that it carries any risk to health. Its use will no doubt grow, especially when its patent runs out in 1987.

The success of aspartame has prompted chemists to search for similar sweeteners based on amino acids. Recently, Murray Goodman of the University of California, San Diego, and William Fuller and Michael Verlander of BioResearch Inc, also of San Diego, came up with a method of making a new class of amino-acid sweeteners in which they reverse the normal way of combining amino acids and obtain a product that is more stable than aspartame (Journal of the American Chemical Society, vol 107, p 5821).

Sweet molecules made of amino acids also occur naturally. One is thaumatin (trade name Talin), which comes from the West African plant ketemfe (Thaumatococcus danielli i). This protein is a large peptide polymer of relative molecular mass 22 000. It is unusual in that it takes time to act but, when it does, it is 3000 times sweeter than sugar and much sweeter than the other artificial sweeteners (Table). Sung-I-Ion Kim and a group from the University of Utrecht in The Netherlands published its polymer structure in the Proceedings of the National Academy of Sciences (USA) (vol 82, p 1406, biochemistry section).

The polymer structure probably explains why thaumatin is of only limited use as a sweetener. It has many binding sites. so it clings to the tongue and has a long lingering effect. This prolonged aftertaste may be all right for chewing gum and unpleasant-tasting medicines, but it precludes this sweetener from many uses in human food. As a flavour enhancer, and in animal foods, however, it is very good. Pet-food manufacturers already use thaumatin as a sweetener. Farmers have found that pigs eat more, and gain up to 10 per cent more weight, if they add thaumatin to the pigs' feed.

A finger-lickin' compound

Chemically, saccharin and cyclamate are quite similar, so perhaps it was not too surprising when another compound containing sulphur and nitrogen also turned out to have an intensely sweet taste. This sweetener, acesulfam K (trade name Sunnett), was discovered quite by chance when Karl Claus of Hoechst licked his finger to pick up a piece of paper. That was in 1967. Researchers have since carried out over 40 trials to test acesulfam K's safety tests. So far these have shown no adverse side effects. The compound now has clearance for use as a general sweetener. Acesulfam K has advantages over aspartame in that it is stable in,water and does not decompose when heated. It is not metabolised by the body and it is excreted easily. So why has it not had the same success as aspartame despite having official approval and being much cheaper? Perhaps because of the quality of its taste.

One of the most curious discoveries last year was that by Cesar Compadre, John Pezzuto, Douglas Kinghorn and Saritri Kamath of the University of Illinois. They followed up a report of 1576 by Francisco Hernandez in his book The Natural History of New Spain. In it, he noted that the Aztecs of Mexico referred to a sweet herb which they called izonpelic xihuitl. This plant is Lippia dulcis, the leaves and flowers of which are indeed sweet. The American scientists extracted the active constituent from the plant, determined its chemical structure, and named it hemandulcin after Francisco Hernandez. They also prepared it by chemical means and tested its sweetness (Table). Unfortunately, despite its intense sweetness it has a bitter component which makes it unacceptable as a substitute for sugar. On the other hand, hernandulcin passed its first tests for side effects and toxicity on rats. Its discoverers hope that, with suitable chemical modifications, chemists might be able to make a derivative of hernandulcin that retains the sweet component but loses the bitter one.

Another approach to artificial sweetness is to modify the sugar molecule. Chemists have managed to identify those atoms responsible for the sweetness of the sugar molecule, sucrose (Box). Even such a minor change as turning the hydroxyl (OH) group on carbon 4 so that it points upwards and not sideways is enough to take all sweet taste away. Past experience has shown that any tampering with sugar destroys its sweetness. Indeed, one of the bitterest compounds known is a derivative of sugar, sucrose octaacetate. The plant Clematis japonica produces this substance naturally. Manufacturers use it to denature industrial alcohol-in other words, to make it too unpleasant to drink.

Yet chemists in the laboratories of Queen Elizabeth College (now part of King's College London) made the remarkable discovery in the 1970s that they could turn sugar into a molecule over a thousand times sweeter. The journal Chemical Society Reviews has recently told the story of this discovery (vol 14, p 357). The project started by replacing sugar's hydroxyl (OH) groups with chlorine atoms (Cl), using the reagent sulphuryl chloride (SO2Cl2). The first group to react is that on atom 6' and this change makes the sugar sweeter. Put on the next chlorine atom at site 6 and all sweetness vanishes. The third chlorine atom attaches itself to carbon atom 4, making the product ten times sweeter. Put in a fourth chlorine atom on carbon l', however, and you have a sugar that is 200 times sweeter than sucrose.

Curiously, these substances might still have been waiting to be tested for sweetness had not Shashikant Phadnis, a graduate researcher, misheard a telephone call requesting samples of the chlorinated sugars for testing. As the call came from a large sugar company it is perhaps understandable that Phadnis thought that the company had requested them for tasting and so he tried them himself. The results are available in UK patents 1 543 167 and 1 543 l68.

Since then, chemists have tried various combinations of chlorine atoms. Four chlorine atoms on carbons 4, l', 4' and 6' turn out to form the sweetest molecule-2200 times sweeter than sugar. Yet four chlorines at positions 2, 6, 1', and 6' create extreme bitterness. The sugar becomes as bitter as quinine.

The most useful derivative of sugar is not one with chlorine atoms on carbons 4, l', 4', 6' as this structure is chemically tricky to make. As this derivative is sweeter than the one with chlorine atoms at carbons 4, 6, l', 6', we can see that the chlorine on atom 6 is responsible for holding back the sweetness. Take this chlorine away so we have only three chlorines, on 4,l' and 6', and the compound is 650 times sweeter than sugar. As it is also relatively easy to make, this is the one that is being tested for safety as an artificial sweetener. Its real name is 4,l',6'-trichloro-4,1',6'-trideoxy-galactosucrose and its trade name is Sucralose. This sweetener scores over saccharin as it has no bitter after-taste and there are no signs that it might cause cancer.

So why is sucralose so sweet? To answer this question brings us to the theory of Robert Shallenberger and Terry Acree who proposed, 20 years ago, that sweet-tasting molecules had pairs of hydrogen-bonding sites in close proximity.

Hydrogen bonding is the most common interaction between molecules. When a hydrogen atom is attached to oxygen or nitrogen it becomes denuded of electrons. Consequently, it is capable of attracting to it other atoms that have surplus electrons known as lone pairs. Nitrogen, oxygen, fluorine and other non-metal atoms are examples of atoms with lone pairs of electrons.

Sweetness is triangle-shaped

Protein molecules are ideally suited to form pairs of hydrogen bonds, as they have amide (N–H) and carbonyl (C=O) sites strategically placed about 0.3 nanometres apart. The carbonyl group (C=O) is seeking a hydrogen on another atom and the amide group (N–H) is seeking an atom rich in electrons. A molecule that has these complementary sites, which we call AH and B respectively, 0.3 nanometres apart, can bind to the protein. If the protein forms part of the taste receptors at the tip of the tongue, the brain receives a signal saying that a sweet taste is present.

So far so good, but sweetness is such a rare effect that other factors must be involved. Lemont Kier, of the Massachusetts College of Pharmacy at Boston, studied sweet molecules. He noted that most shared a third common factor-part of the molecule was hydrophobic, in other words repellent to hydrogen bonds. He called this the X site. He also found that X must be further from B than from AH. The whole molecular arrangement, the triangle of sweetness, is called the "glucophore".

Further support for Kier's triangle theory came from the observation that, whereas L-phenylalanine is bitter to taste, D-phenylalanine is seven times sweeter than sugar. These two molecules are optical (mirror) isomers. In other words, they differ only in that one is the reflection of the other. In all other respects, the molecules are identical. So the receptor site on the taste buds must recognise more than the hydrogen bonding with the AH and B groups. The situation of the third site (X) must also be critical. However, optical isomers are not always different. D- and L-fructose are both sweet, although this may be because the AH and B sites in these molecules are interchangeable, something which is not possible in the amino acids.

If the triangle theory is correct, it should be possible to construct a molecule with the correct distances between AH, B and X and see if it is sweet. This is just what Tetsuo Suami of Keio University in Japan did. He noted that the sugar D-fructopyranose is the sweetest natural carbohydrate, l.8 times sweeter than sugar. He then built a molecule that he called pseudofructopyranose, which had exactly the same triangle in its structure. It was indeed sweet (New Scientist, 31 October 1985, p 22).

Arnold van der Heijden, Henk van der Wel and Hein Peer of the Unilever Research Laboratories, Vlaardingen, The Netherlands, have refined the AH, B and X theory, using a computer program to link the distances between atoms in molecules and their sweetness. Their results, published in Chemical Senses (vol 10, p 57 and p 73) even suggest that a fourth site may be involved. Their prediction that a derivative of nitroaniline should taste sweet, despite a report in the literature that it was tasteless, supported the validity of the theory. When people tasted this substance, they did indeed find that it was very sweet.

Not everyone accepts the triangle theory, however, as some sweet molecules have no X corner. In addition, it is difficult even to find AH or B sites in certain sweet-tasting substances such as the salts of lead or beryllium. (The discoverer of beryllium in 1798, Nicolas-Louis Vauquelin, called this new substance glucinium, meaning sweet.) In these cases, it may be the water molecules clustering around the metal in solution that provide the AH and B sites to activate the receptors.

Mohamed Mathlouthi and Anne-Marie Seuvre, of the University of Dijon, and Gordon Birch, of Reading University, believe that water has an even more important role to play in sweetness than many chemists have so far realised. Water, as a solvent, is highly structured through a framework of hydrogen bonds. Any solute dissolved in water interferes with these bonds, in some cases strengthening the framework in the immediate vicinity of the solute, in others weakening or even disrupting it. Mathlouthi in Food Chemistry (vol 13, p 1), and Birch in Chemical Senses (vol 10, p 325) discuss ways in which the effects of dissolved sugars may affect the water itself. So far, however, neither group has found a simple relationship between the volumes of sugar molecules or the effects on surface tension that sugars cause, and their perceived sweetness.

Recently they have collaborated on a more fundamental discovery and one that might eventually lead to an instrumental way of detecting sweetness. If sweet molecules have an AH centre, and they generally do, then the AH centre should reveal itself by infrared spectroscopy, as this technique makes it easy to observe hydroxyl (OH) and amide (NH) bonds. It is also possible to detect whether these groups are involved in hydrogen bonding.

Using an advanced Fourier Transform Infrared Spectrometer, Mathlouthi, Seuvre and Birch have measured several sweet carbohydrates including Sucralose. They report that they can detect a free hydroxyl (OH) group in these compounds. This observation indicates a very active AH site, and one that will compete very easily for the corresponding site on the taste bud where it will undoubtedly have to displace a resident water molecule. This work, soon to appear in Carbohydrate Research, holds the promise of an objective assessment of sweetness. Find a molecule with a free AH band in its infrared spectrum and there is a fair chance that it could be sweet.

Chemists can certainly satisfy the human desire for sweet-tasting foods with artificial sweeteners, although the quality of sugar's sweetness has proved difficult to imitate, Yet we do need sugar to preserve food, How else could we tum the fruits of summer into home-made jam for winter? Sugar also plays an important role in helping to save the lives of children afflicted with diarrhoea. Over 12 million children die each year of this condition, mainly in poor countries. Most of them could be saved simply by drinking a solution containing an 8:1 ratio of sugar to salt (New Scientist, 3 April, p 27). Nor is this the only use of sugar in medical treatment. Argentinian doctors have successfully used refined sugar, applied to open wounds, to treat diabetic patients with chronic abscesses.

Third World countries could grow more sugarcane, which is probably the cheapest source of sugar, protein and minerals. As the West turns away from sugar, and its price continues to fall, poor countries prone to famine could build their own sugar mountains as insurance against future drought. Sugar will keep indefinitely and without refrigeration. Moreover, it is a superb food if you have little else-it is cheap, rich in energy, easy to digest and tastes delightful. Even if the world's demand for sugar was already met, it would still be worth growing sugar cane as a fuel crop or as a source of raw materials for the chemical industry. Cuba, for example, plans to use sugarcane to make animal feeds, drugs and even furniture (New Scientist, 3 April p 24).

Leslie Hough is a professor and John Emsley a reader in the department of chemistry, King's College, London.