Save the flowers: would-be scent engineers aim to resurrect lost floral fragrance.

Vince Agnes, as well-appointed as the flowers that he has been selling for more than 60 years in his shop in Silver Spring, Md., remembers when all his roses smelled as good as they looked. When he opened for business in the 1940s, there were only a few varieties: red, white, yellow, and pink, he recalls. "Now, there are thousands," Agnes says, "but only a few have a lot of scent."

No one knows what's responsible for this waning of fragrance by roses and other ornamental-flower varieties, including carnations and chrysanthemums, but scientists who investigate floral scent suspect that the flower breeding that's led to an estimated 18,000 rose cultivars in an ever-widening spectrum has run roughshod over fragrance.

"Pigment compounds are derived from the same biochemical precursors [as scent compounds are], so it makes sense that if you make more of one you get less of the other," notes floral-scent biochemist and geneticist Eran Pichersky of the University of Michigan in Ann Arbor.

Floral scent may be dwindling because breeders for the $30 billion ornamental-flower industry pay scant attention to this most emblematic attribute of flowers. "In order of [commercial] priority, color is number 1 through 10," says Alan Blowers, head of flower biotechnology for Ball Helix, a biotech company in West Chicago, Ill., devoted to the ornamental-plant industry. Beyond color, breeders have been targeting improvements in flower longevity, shape, size, disease resistance, and other traits likely to improve the growers' bottom lines.

Fragrance is different. It's invisible, and its sensory impression is as subjective as taste. And, as it turns out, fragrance is a genetically complex trait that's difficult to manipulate by ordinary breeding methods. Despite those obstacles, Blowers predicts, "fragrance will become important again," as the molecular biology underlying floral scent becomes better understood.

With a nose both for understanding the molecular origins of floral scents and for engineering what could be blockbuster flower varieties, researchers have been teasing out the complex biochemical orchestration underlying one of life's simplest pleasures. They've been uncovering fragrance-related genes, the enzymes encoded by those genes, the in-cell reactions that these enzymes catalyze, and the fragrant performance of all this molecular biology--a vast aromatic harmony of alcohols, aldehydes, fatty acids, terpenoids, benzenoids, and other volatile, and therefore sniffable, chemicals.

In the past few years, flower scientists have assembled enough knowledge and technology to consider resurrecting scents in flowers that have lost them or engineering plants that produce scents never before experienced by a bee, beetle, or gardener.

The researchers "have pushed the envelope in terms of our eventual ability to change floral scent," says Michael Dobres, head of the Philadelphia biotech company NovaFlora, which is developing genetic methods for controlling various traits of ornamental flowers.

DECONSTRUCTING SCENT The plant world perfumes, or sometimes stinks up, the environment with a vast roster of volatile organic chemicals. Scientists have so far identified about 1,000 of these compounds emanating from petals, leaves, and other tissues.

"There could be up to 50, maybe 100, chemicals involved in a particular scent," says Pichersky.

Usually, only a few of the volatile chemicals in a fragrance are obviously noticeable to human noses. One whiff of 2-phenylethanol, for instance, and images of roses come to mind, even though scores of volatile chemicals contribute to the filly detailed scent of roses. Like harmonics that help the ear distinguish a middle C played on a piano from one played on a violin, the minor chemical components of a scent provide the olfactory subtleties that individualize the scent of a particular rose variety.

Pichersky, who grew up on a kibbutz growing flowers and other crops in his native Israel and now lives on a 30-acre farm outside Ann Arbor, has been gardening all his life. He has made it his mission to uncover as much as he can about the biosynthesis of floral scents and the biological roles that these scents play. In 1996, he and his colleagues in Michigan were the first to discover a gene that produces a floral scent.

Not only do the volatiles in botanical scents attract pollinators and delight the human nose, he notes, but they also serve to protect plants from pathogens and pests. For example, when some plants come under attack by munching caterpillars, they emit specific chemicals as clarion calls for parasitic wasps. The wasps alight on the marauding caterpillars and lay eggs, which hatch into larvae that eat the caterpillars alive. "It's a chemical arms race out there," says Pichersky.

As the first step in analyzing the complex biochemical choreography behind a floral scent, Pichersky and his coworkers in 1994 worked out the amino acid sequence of the enzyme linalool synthase from petals of Clarkia breweri, a purple wildflower native to California. They then used that information to identify the enzyme's gene.

Through painstaking biochemical analysis, the researchers discovered that this enzyme converts the substrate geranyl pyrophosphate into linalool, a volatile compound with what Pichersky describes as a "wine-sweet" smell. Geranyl pyrophosphate was already known as an intermediate in the metabolic pathway that produces cholesterol compounds.

Since then, Pichersky's group and others have uncovered about 25 more floral-scent genes. Natalia Dudareva, a former postdoc student of Pichersky who now runs her own floral-scent laboratory at Purdue University in West Lafayette, Ind., estimates that the present list of known scent genes and their associated enzymes can account for the cellular synthesis of no more than 5 percent of the plant volatiles that scientists have identified.

The enzymes encoded by floral-scent genes fall into a few functional categories with names such as synthases, methyl transferases, and carboxymethyltransferases. Enzymes in a given category impose a particular kind of biochemical transformation on cellular chemicals that arise from the basic, or primary, metabolism that all plants share.

The outcome of the transformation differs according to the specific plant. For example, in snapdragons, one particular methyltransferase enzyme adds a methyl group (a central carbon hub bonded to three hydrogen atoms) to benzoic acid, producing methyl benzoate. In C. breweri, the same enzyme instead methylates salicylic acid, producing methyl salicylate. Scientists call such species-specific biochemical products secondary metabolites.

Mixing and matching enzymes and substrates in varying sequences of reactions creates a bazaar of secondary metabolites, Pichersky notes. That's how a lilac or honeysuckle, for example, can produce its own intoxicating cocktail of fragrance compounds.

Discovery in floral-fragrance biochemistry is on a fast track, now that lab devices can identify and analyze the activity and interactions of hundreds to thousands of genes and proteins at once. For example, Pichersky and a large collaboration of researchers working primarily in Israel, a flower-exporting country, compared the genetic activity of Fragrant Cloud, a scented rose cultivar, with that of Golden Gate, an unscented one. From an initial roster of more than 2,000 genes that the researchers identified as active in these two cultivars, the team pinpointed a few genes that appeared to be involved in scent production. This led the search to previously unrecognized enzymes, which the researchers demonstrated were required in the biosynthesis of various rose-scent chemicals, among them geranyl acetate and germacrene D.

Those are just a few chemical pixels in the vast picture of floral scent. Nevertheless, Robert Raguso, a chemical ecologist at the University of South Carolina in Columbia, characterizes the pace of discovery in the field as explosive. "We are in this beautiful growth phase where everything is new ... and worthwhile. Now, the most interesting challenge is putting it together," says Raguso, who was a graduate student in Pichersky's lab in the early 1990s and whose work led to the discovery of the linalool synthase gene.

SCENT AWAY

Even as researchers uncover more of the molecular story behind floral scent, the goal of controlling how flowers smell remains elusive. The genetic and biochemical complexity of fragrance continues to thwart scientists. "Many attempts at [scent] engineering have been done, but so far there hasn't been a lot of success," says Dudareva.

Pichersky says that scent engineering would be useful for more than just pleasing human noses. For starters, he suggests that it will someday empower growers to choose the pollinators that visit specific plants and to replace some chemical pesticides with living pest controllers such as parasitic wasps.

In one approach to manipulating plants' biochemical pathways, a team of scientists in the Netherlands inserted into petunias the C. breweri gene for linalool synthase that Pichersky's team had discovered. The team, led by Harro Bouwmeester of Plant Research International in Wageningen, the Netherlands, confirmed that the transplanted gene was working and that the transgenic petunias were making linalool in their tissues, but the linalool never made it out of the plant.

A related project, led by Alexander Vainstein of Hebrew University in Rehovot, Israel, got a bit further. It produced transgenic carnations that released linalool, but in amounts too small for a person to smell. While Raguso says that the aroma of linalool reminds him of Earl Grey tea, David Clark of the University of Florida in Gainesville describes the smell as "Fruit Loopy." Clark has been approaching scent engineering by manipulating the native genes of a single plant, a petunia, rather than transferring scent genes from one species to another. Using current techniques of genetic analysis and engineering, he intends to first identify genes that might play roles in petunia scent. Then, he'll either deactivate those genes or pump up their activity. His goal is to make the plants produce unusually small or large amounts of the enzymes encoded by the genes.

"We are just now figuring out where all of the pieces are in the pathways," says Clark. He notes that petunia fragrance emerges largely from 8 to 10 volatiles, each one created by the interplay of several enzymes and substrates. Manipulating the plant's fragrance with finesse, therefore, would require commandeering several genes, a dauntingly difficult task since researchers are only sporadically successful at achieving a desired goal when engineering even single genes.

Roman Kaiser, director of the natural-scents research unit for the Geneva-based company Givaudan, favors studying, rather than manipulating, floral fragrances. However, he predicts that would-be fragrance engineers such as Pichersky, Clark, and Dudareva will eventually have their day.

The growing body of knowledge about floral scents is likely to have a bearing on the perfume industry as well as on flower sellers, Kaiser says. "I could imagine that very special fragrance chemicals found in nature, but difficult to synthesize, might be produced by applying such techniques" in a way that improves the yield of these chemicals in the flowers that naturally make them, he says.

If researchers do approach a time when they can engineer flowers to have novel scents, they may discover, as have scientists in the genetically modified food business, that winning public support for such manipulations is the toughest challenge of all. Using genetic techniques to alter floral scent is, in Clark's words, "a double-edged sword."

Opponents of genetically engineered food may add scent-altered flowers to their list of products that could pose dangers. Consider a project in which agriculturally minded genetic researchers alter fragrance genes and the flower attracts different pollinators. "If we end up with a plant that is covered in flies, someone will say, 'This is a freak show,'" Clark predicts. Such a scenario could easily nip scent engineering in the bud, he says.

Even if future scent engineers can win over public opinion, they may have to contend with a host of low-tech factors that Silver Spring, Md.-florist Agnes suspects have a dulling effect on floral scent. He remembers fondly when he could buy all his flowers fresh from local greenhouses. "Now, I get my flowers from California, from Israel, Holland, all over the place," he says.

Flowers lose their scent while they're refrigerated during long journeys on planes and trucks, he says. And that could be a problem, even for high-tech flowers of the future.