ÐÓ°ÉÔ­´´

THE LEVEL of glucose in the blood normally rises rapidly after eating a meal. If someone drinks an extract of the roots of certain species of mulberry with the meal, however, this somehow suppresses the rise in blood sugar. A substance with this effect might be of use to diabetics, which is why, in 1976, a team of Japanese workers decided to try to identify it. They called it moranoline, after the plant’s Latin name, Morus. The research showed that moranoline was a simple compound with five molecules of carbon and one of nitrogen in a ring which carried one hydroxymethyl (CH₂OH) group and several hydroxy (OH) groups.

Soon afterwards, workers in Belgium and at the Royal Botanic Gardens at Kew found two similar compounds (named DMDP and DMJ) in tropical beans. The group at Kew was interested in how wild plants protect themselves from being eaten by predators. They realised that moranoline, DMDP and DMJ bore a close resemblance to the simple sugars glucose, fructose and mannose, respectively. In other words, they were mimics, or analogues, of sugars in which nitrogen replaces the oxygen of the sugar ring. It occurred to them that the enzymes of predatory animals might mistake these compounds for real sugars, and so inactivate the enzymes. This hunch proved correct: in many organisms, including insects and mammals, all three substances could selectively inhibit enzymes of a type known as glycosidases.

Moranoline, DMDP and DMJ are just three of some 20 nitrogen compounds isolated from plants and microbes, which can inhibit the enzymes that normally break down chains of sugars. Biochemists have nicknamed these compounds the ‘sugar-shaped alkaloids’. As recently as 1985, when New ÐÓ°ÉÔ­´´ first reported on their interesting and varied properties as enzyme inhibitors (‘The sugar-shaped weapons of plants’, 15 August 1985), few could have foreseen that, by 1989, they would be important tools in cancer, immunology and AIDS research.

The term ‘alkaloid’ describes a small molecule in which nitrogen occurs as part of a ring structure. Chemists have isolated thousands of alkaloids from natural sources. Many of them cause changes to the biochemistry of organisms which ingest them. Those that cause beneficial changes in humans are responsible for the success of many medicines. The first alkaloids to be isolated were those with obvious pharmacological effects, beginning with morphine in the early 19th century. Later, chemists discovered that if they added ammonia to aqueous extracts of plants, they could separate the alkaloids by shaking the extracts with a solvent such as chloroform. When it is left to stand, the mixture separates into two layers, water and chloroform, with the alkaloids in the chloroform.

This method, which is quick and cheap, endeared itself to generations of chemists but its popularity is largely to blame for the fact that a group of alkaloids, probably widespread in nature and of considerable ecological significance, remained undetected until recently. These alkaloids carry so many hydrophilic, or water-seeking, groups on the ring that it is not possible to extract them with chloroform. The sugar-shaped alkaloids that carry several hydroxy groups are of this type. Another reason why they were not investigated sooner is that their effects on human physiology can be hidden, such as the effect of moranoline on the rise in blood glucose.

Moranoline is more widely known today under the name deoxynojirimycin, or DNJ for short. Unknown to the investigators who found it in the mulberry, researchers in another Japanese laboratory had already made it in the 1960s by chemically modifying nojirimycin, a mimic of glucose which occurs in a bacterium, a species of Streptomyces. The researchers had isolated this compound using a method known as ion-exchange chromatography. This involves passing a solution containing the alkaloid over beads of resin. The beads are made of sulphonated polystyrene, modified so that they can bind alkaloids and remove them from solution. After washing the unbound extract off the beads, soaking them in a solution of dilute ammonia releases the alkaloids.

From the mid-1970s onwards, the use of ion-exchange methods led to the rapid discovery of several sugar-shaped alkaloids. DNJ itself turned up in bacteria. Researchers also found two compounds in Australian plants which would later prove to be exceptionally useful. In 1979, chemists in Western Australia solved the 100-year-old mystery of what was poisoning sheep that grazed small pea-like plants of the genus Swainsona. They isolated a bicyclic compound, which they called swainsonine, which damages the nervous system. Other investigators found that swainsonine was also responsible for poisoning cattle that ate other pea-like plants, Astragalus and Oxytropis, on the mid-western cattle ranges of the US. Then, in 1981, biochemists in London found an alkaloid with a similar bicyclic ring in the chestnut-like seeds of Castanospermum australe, a tree from the rainforests of Queensland. They called it castanospermine, and wondered whether it might contribute to the digestive upsets reported to follow if humans or animals ate the seeds.

Plant chemists at Kew were particularly interested in the discovery of an alkaloid in Castanospermum. An important part of their work is to use chemical methods to assist botanists in classifying and naming plants. Chemists had always believed, on the basis of the old chloroform-extraction method, that Castanospermum did not contain alkaloids. The discovery of this compound led not only to a reappraisal of the relationship of Castanospermum to other bean species, but also to the isolation of AB1, which is similar to DMDP, from a plant thought to be related to Castanospermum. More recently, the close partnership of botanists and chemists at Kew has led to the discovery of a new source of castanospermine in a group of South American plants, without anyone setting foot outside Kew Gardens. Botanists had noticed in the 1970s, using material stored in the herbarium, that the flowers and pollen of Castanospermum were remarkably similar to those of a group of South American beans of the species Alexa. The geographical separation seemed to rule out any possibility that they might be related. However, Robert Nash at Kew had a hunch that they were. He tried to extract compounds from a pod of a species of Alexa that had languished in the Kew Herbarium for 40 years – and found castanospermine. This discovery has made botanists think again about the relationship between Australasian and South American plants. It may be that the common ancestor of Castanospermum and Alexa straddled the land bridge which geologists believe linked both continents through Antarctica 50 million years ago. ÐÓ°ÉÔ­´´s at Kew recently confirmed that there is a close relationship between the two groups by isolating from both novel alkaloids with another kind of bicyclic structure, such as alexine and diepialexine.

Like DNJ and its relative, swainsonine, AB1, castanospermine and alexine all exert their effects by inhibiting enzymes known as glycosidases. These remove one sugar at a time from the ends of short chains of sugars. The term ‘glycosidase’ includes enzymes that remove molecules of sugars such as glucose, mannose, galactose and so on. The enzymes are respectively named glucosidases, mannosidases, galactosidases, and so on. Castanospermine and DNJ inhibit glucosidases; swainsonine and DMJ have their effect on mannosidases. Simple sugars can join up in chains in two different ways, with a so-called alpha or beta link. Different enzymes split each type of link and individual alkaloids may inhibit the enzymes to different extents.

The problem with pests

Castanospermine powerfully suppresses the alphaglucosidase found in the guts of mammals but it is almost inactive against the equivalent enzyme from the larvae of many insects, including some small beetles (species of Callosobruchus) which attack stores of beans in the tropics. In contrast, DMDP has a strong effect on the beetle’s enzyme, but not on the mammalian one. A compound that inhibits enzymes of crop pests, but not those of humans, might be useful in protecting crops. Unfortunately, it is impossible to predict how an enzyme from a particular organism will respond to a particular alkaloid, or how altering the structure of the alkaloid will affect its properties. For example, AB1 is identical to DMDP except for the loss of one hydroxymethyl group. That loss enables it to inactivate the mammalian alpha-glucosidase strongly. Nevertheless, biochemists are having some success in working out how the differences in structure affect the functions of the molecules. For instance, they have been able to use DMDP to differentiate between two kinds of glucosidase in cells.

The ability of the sugar-shaped alkaloids to inhibit glycosidases involved in the synthesis of some glycoproteins has proved to be particularly important. Glycoproteins are proteins with chains of sugars (known as oligosaccharides) attached. These compounds play a part in many processes that scientists do not fully understand, such as the transformation of normal cells into cancer cells and the infection of cells by viruses. To understand these events more fully, biochemists need to know exactly how the sugar chains contribute to the activity of the glycoprotein. In theory, the way to find out would be to manufacture glycoproteins which have specific changes in the sequence of the sugars in the chains and then compare the activities of the altered molecule with those of the original. In practice, there was no way of doing this until the discovery of the sugar-shaped alkaloids.

An important group of glycoproteins, widespread in nature, are those in which the sugar chains are linked to asparagine residues in the protein. The final sequence of sugars in the chains is unique to the particular molecule, but the sugars all start life in the same way as a preformed ‘starter’ molecule. This is a branched chain which is transferred intact to the protein during its synthesis. The starter molecule appears to be the same or very similar for all glycoproteins of this type. To reach the finished product, enzymes trim the starter down to a smaller ‘stump’ chain, one sugar at a time. These enzymes are a special group of glycosidases (glucosidases and mannosidases) known as processing or trimming glycosidases. When this operation is complete, enzymes called sugar transferases add new sugars to the stump, one at a time, to produce a unique finished chain.

Some sugar-shaped alkaloids can block the action of the trimming enzymes. For instance, DNJ and castanospermine block the removal of the first glucose from one part of the starter and DMJ and swainsonine prevent the removal of mannose from another. When the processing of any branch of the starter is blocked, that branch appears unmodified in the final molecule, changing its properties. So biochemists can use the alkaloids to bring about selective changes in the carbohydrate parts of glycoproteins, making it possible to discover exactly what the function of the carbohydrates is.

One day, the sugar-shaped alkaloids may have a role in the treatment of cancer. Glycosidases released by some tumours help cells to break away from the tumour so that they can invade other body tissues: this process is called metastasis. Experiments are under way to determine whether it is possible to reduce or block metastasis with sugar-shaped alkaloids. One particularly interesting observation is that these compounds can make some cancer cells behave more like normal cells. For example, cat embryo cells, transformed into cancer cells by a strain of feline leukaemia virus, can grow in soft agar, a property of the transformed (cancerous) state. In the presence of castanospermine, they lose this ability. Swainsonine has no effect in this system, but researchers have found that it prevents the growth in soft agar of mouse cells (of a type called fibroblasts) transformed with DNA from a line of human bladder cancer cells. No one yet knows how these changes come about. They probably result from modifications to the glycoproteins on the surface of the cancer cells. This is particularly likely, as researchers have already observed changes in the sequence of sugars on the side chains of glycoproteins on normal cells when they become cancerous.

Recently, American scientists found that if they treated cultured tumour cells from mice with swainsonine before injecting the cells into healthy mice, the cells were less likely to form tumours in the lungs than the untreated cells. In a second experiment, the researchers provided the animals with traces of the alkaloid in their drinking water for only a few days before injecting them with untreated tumour cells. To the surprise of the researchers, there was a similar reduction in the ability of the cancer cells to invade the lungs. As it seemed unlikely that swainsonine could have produced this effect on the cancer cells directly, the researchers investigated the possibility that swainsonine affects the immune system, enhancing the animals’ ability to reject the invading cells. Their studies showed that swainsonine could increase the number of cells of the immune system known as natural killer cells. These observations confirmed an earlier report from Japan that swainsonine could enhance the immune response of mice.

Since 1987, there has been a surge of interest in the potential of the sugar-shaped alkaloids as a treatment for AIDS, even though earlier studies on the ability of these compounds to reduce the infectivity of viruses were discouraging. High levels of castanospermine, swainsonine, DMJ and DNJ all caused changes to the glycoproteins of the viruses that cause influenza, vesicular stomatitis and fowl plague, as well as to those of the Rous sarcoma virus, but without affecting their virulence. In 1987, biochemists at Kew and St Mary’s Hospital in Paddington, London, showed that castanospermine and, to some extent, DNJ and DMDP, all reduced the ability of the human immunodeficiency virus (HIV), which causes AIDS, to infect cultured cells. Researchers in the Netherlands and the US have also reported the effect of castanospermine on HIV. Studies have since shown that when the alkaloids are present, the viruses that infected cells produce are not infectious. This is due, in part at least, to changes to a glycoprotein on the surface of HIV, known as gp120, which carries an unusually large number of sugar side chains and which plays a crucial role in the process by which the virus infects new cells. The alkaloids are also likely to affect the glycoproteins of the host cell, but scientists have been pleasantly surprised to find that it is possible to inhibit the virus with low concentrations of the alkaloid which do not harm cells. Perhaps castanospermine selectively affects a process specific to the virus, such as the way in which gp120 is produced.

By modifying DNJ or castanospermine, chemists in Britain and the US have made even more powerful compounds to inhibit HIV. Because the alkaloids have a different mode of action from the AIDS drug zidovudine, it may be possible to treat patients with both drugs. ÐÓ°ÉÔ­´´s are now working to produce large enough quantities of the alkaloids for studies in patients. A modified form of DNJ, N-butyl DNJ, is now being tested in clinical trials in the US.

Another application for alkaloids, such as DNJ, that can inhibit glucosidases may be in diabetes. ÐÓ°ÉÔ­´´s have tested many natural and chemically modified forms of such compounds. In 1987, American chemists described how to make a molecule that inhibits intestinal alpha-glucosidase. The molecule consisted of a novel alkaloid, homonojirimycin, joined to a molecule of glucose. Although homonojirimycin was not known in nature then, a scientist at Kew recently discovered it in a vine, Omphalea, of the family Euphorbiaceae. This suggests that, once again, nature had a lot of good ideas well ahead of human chemists.

Plants do not suffer from AIDS or diabetes, but they are attacked by insects. ÐÓ°ÉÔ­´´s at Kew showed in 1984 that traces of DMDP in artificial diets for plants would prevent locusts from feeding. Researchers in the US found that castanospermine had a similar effect on aphids. These compounds act directly on the sensilla (sensory organs) on the mouthparts of insects. This means that the insect leaves the plant alone, without ever eating any of it.

Kinder crop protection

Natural, and therefore biodegradable, chemicals which deter insects from feeding without killing them indiscriminately could be ideal agents for protecting crops. Entomologists at Kew and at Birkbeck College in London have been studying the sensory signals that pass from the sensilla to the brain in Spodoptera caterpillars as they make contact with sugar-shaped alkaloids. Sugars such as glucose, fructose and sucrose provoke a strong signal telling the insect to feed. DMDP provokes an equally strong signal telling them not to feed. If the researchers exposed the caterpillars to DMDP before exposing them to the sugars, the strength of the ‘feed’ signal prompted by fructose and, to some extent, by sucrose, was greatly diminished for up to two hours. As DMDP is an analogue of fructose (and sucrose consists of a molecule of fructose linked to one of glucose), it seems likely that DMDP temporarily blocks the insects’ receptor sites for fructose in some way. DMDP in plant leaves may be the perfect means of discouraging insects. DMDP makes predators ‘blind’ to the presence of potential food in the form of fructose or sucrose, and they presumably move away to other species. Pollinating insects needed by the plant, however, are not affected. This simple and ancient survival strategy might point the way to future ways of protecting crops. Meanwhile, DMDP and other sugar-shaped alkaloids are helping entomologists to map the receptors for different chemicals on the sensilla of insects.

The quest for more sources of these alkaloids, in both plants and microorganisms, continues. Chemists at Kew are using insects to help them in their search. For example, only certain species of Lonchocarpus, a group of beans found throughout the tropics, contain DMDP: it is possible to predict which ones they are by observing which insects live on them. Although DMDP deters locusts, many insects feed happily on plants containing high levels of this substance. Close relatives of insects that are tolerant to DMDP usually eat plants that contain DMDP, too. Biochemists have tracked down more species of Lonchocarpus that contain DMDP in this way.

People have previously considered most plants that supply us with sugar-shaped alkaloids to be of little use. These species do not figure prominently in lists of economically important species. This anomaly may give us cause to reflect that wild plants enabled us to evolve into what we are today, and that we still need all of them.

Linda Fellows is a biochemist at the Royal Botanic Gardens, Kew.