STRUCTURE ACTIVITY RELATIONSHIPS AND THE SUBJECTIVITY OF ODOUR SENSATION

Dr. Thomas Markert Cognis Deutschland GmbH, Henkelstr 67, D 40551 Duesseldorf, Germany

1 INTRODUCTION

Structure activity relationships (SAR) are one of the most useful sets of tools in both pharmaceutical and fragrance research. Ever since Amoore carried out his studies and formulated his theory of odour recognition, chemists have been looking at the shape of molecules and their associative possibilities to find clues that would explain perceived odour sensations. How difficult it is to go down this research path in finding new chemical entities with interesting odour qualities is clear from the broad variety of odours the human nose is able to detect and identify. I will now attempt to explain how complex the activity side of SAR can be and what the consequences of this complexity are.

In this context I will again follow up the question which Pieter Aarts recently put at the top of an article [1], although he was dealing with a totally different subject: "The Optimal Fragrance - Lucky Shot or Organised Hunt?"

The sense of smell is even able to discriminate between the antipodes of chemical structures like R- and S-carvone or R- and S-p-menthene-8-thiol [2]. When a perfume layman, like a chemist, tries to verify the reported odour descriptions, he becomes aware that the difference between the odours of chemically similar substances is dependent on purity, concentration in your nose, your sniffing technique, the way the air streams through your nose [3], and much more.

As Charles Sell tells us in a remarkable report about structure/odour correlations entitled "The Mechanism of Olfaction and the Design of Novel Fragrance Ingredients" [4], it is sometimes a trace impurity which fundamentally changes the scent of a substance or a mixture of substances.

2 AMOORE'S CONCEPT OF PRIMARY ODOURS

Let us start with John E. Amoore's [5] theory of odour reception (figure 1), which is based on specific anosmia and the concept of primary odours. What I understand about his idea is that he tried to find chemical structures by using the holes in the olfactory epithelium and a negative selection of substances that were reported as resulting in specific anosmia.

In terms of SAR, this would mean he was searching for chemicals with no activity. And from the shape of the molecules he found in this way he tried to reconstruct a receptor site which could in size and shape accept this chemical structure (figure 2). The goal of his studies was a classification of odours by collecting groups of similar molecules, which could fit, specifically into the same receptor. Amoore was limited in his approach to the choice of known substances and he was also dependent on the odour descriptions he was given by the experts. My opinion in this context is that Amoore could never definitely know whether a substance, which would bind to the same specific receptor, would cause the same odour sensations and associations. In other words, he grouped various chemicals together, guided by the similar odour descriptions for those materials.

3 SPECIFIC ANOSMIA AND THE CONCEPT OF PRIMARY ODOURS

I have to admit at this point that I have a problem. My problem is with specific anosmia, which is the basis of Amoore's theory of olfaction. The way Amoore measured specific anosmia demonstrated the usefulness of his approach and proved the reality of this phenomenon. However, the results are not useful to classify scents; they only caused chemists to focus on molecules for which there would probably be a specific receptor in the nasal mucous membrane. When a chemist looks at the structures found by Amoore they are surprised to find four small molecules like trimethylamine and isobutyric aldehyde, alongside two very large molecules like androstenone and pentadecanolide (figure 3).

Those who are able to smell androstenone with 19 carbon atoms describe it as reminding them of stale sweat. Isovaleric acid, a molecule with 5 carbon atoms, is almost officially said to smell sweaty. So, am I to believe that a molecule with 19 carbon atoms is bound to the same specific receptor as a similarly smelling compound containing 5 carbon atoms? The Amoore approach is most interesting because, when you think about it, in the end it doesn't tell you much about the structural side of SAR, nor does it tell you much about the activity on the side of the receptor, but it raises the question of what specific anosmia means. What is the sense of lacking receptors?

When we at Cognis were searching for new sandalwood substances, I noticed that I became anosmic to Sandelice®; first only on Fridays, then later all the time (figure 4).

Then I noticed that my anosmia was a hyperosmia. I was so sensitive to Sandelice® that I had the odour impression for a fraction of a second and then my nose had adapted. So adaptation can also look like anosmia. By contrast, I am truly anosmic to androstenone. True specific anosmics smell the impurities in the compounds. So, although I'm training on our androstenone sample, to me it smells a little bit cedar-woody but not at all like urine or stale sweat. Others nearly had their noses blasted off when they opened the bottle. So I consider that the purity of our androstenone sample is very good.

3.1 What anosmics smell?

When the results of The Smell Survey were published by National Geographic in 1987 [6], I thought how unhappy the 1.2% of people who were suffering from total anosmia must feel. I thought those people would neither smell nor taste anything so delicious as truffles or foie gras. This is by far not the case. I learned from one of my neighbours who lives a few houses away from ours that her bulbus olfactorius had been severed in a car crash. But she is still able to taste and to smell. - Though she might need more cigarettes or beer to have the same activity effect as osmic people. - And I wondered how this could happen without the ability to smell. Then I read [7] about people who, though lacking a sense of smell, were able to cook, detect dry or humid air, and more. At least taste is working well in anosmics.

By thinking about the odour impressions of people lacking olfaction, I found the explanation for some unusual odour descriptions. What do you think a powdery or dusty scent should mean? Would it be a powder or dust that would enter your nose? We were once purifying the essential oil of pinus longifolia and when I smelled the fractions, I immediately imagined smelling powdered bellpepper from a pepper pot. The visual picture of a liquid in a distillation bulb did not fit the odour impression of a powder. Then suddenly I had an idea about what could be the explanation for this curious phenomenon. Like every mucous membrane, the olfactory epithelium is sensitive to touch as well. In other words, your nose does not just smell things, it also feels them (figure 5).

Our sense of taste is based on touch. The only reason we boil our soup or coffee is that we like it hot. Umami (monosodium glutamate) is discussed as a fifth taste quality but it could also work as a transporter of tastemakers (my personal name for taste enhancers) — comparable to the odour binding proteins — by distributing aroma components in your mouth. The result is called mouthfeel. Touch, pain, or trigeminal reception is what anosmics smell, and probably osmics as well.

3.2 The Kallmann syndrome

Kallmann's syndrome is a neuronal migration defect, which also affects olfactory system development. To test the functioning of olfaction with patients suffering from Kallmann's syndrome, doctors use common fragrance materials. In this way it was found that many fragrances have a strong trigeminal component. Anosmic patients were able to assign odour descriptions to fragrances without using olfactorial nerves. So the information must have been transported by a nerve other than the bulbus olfactorius. In his report about "Trigeminal Perception of Odorant Quality in Congenitally Anosmic Subjects" [8], Matthias Laska presents a list of compounds eliciting strong trigeminal responses, which sounds like Amoore's list of primary odours (figure (6).

Again there are four smaller molecules like acetic acid, acetone or ethanol and two larger entities (-)-menthol and 1,8-cineole. The existence of trigeminal sensation, which I would like to call feeling chemicals, is well known from von Skramliks' [9] experiments that were published in 1926. He found a trigeminal component in odorants by monorhinal application. When the test person could detect whether the odorant entered the left or right nostril this was by trigeminal irritation, because olfactorial reception is not able to identify the direction from which the smell has approached the nasal cavity. To test the extent of the trigeminal nerve stimulus within a sniffing process, monorhinal application is recommended.

So what do we learn from those results?

4 THE ACTIVITY SIDE OF SAR REASSESSED

In a report called "Clinical Testing of Olfaction Reassessed" A.J. Pinching [10] speaks of the "poor smell vocabulary of most humans, which was regarded as a barrier to interpretation of olfactory tests. However it has become clear that the great majority of odours have a trigeminal component to their detection." Let us now take a closer look at odour descriptions. In SAR they represent the activity part of the relationship, and the accuracy of this part should be as scientific as the knowledge about the chemical structure. But this is by no means the case. I do not dispute the trigeminal component of odour sensation. What I think is rather that you have a pain sensation in your odour description that does not come from an olfactory reception site. That means that many impressions may stem from stimulating trigeminal nerve endings and you consider them to be your odour reception, not knowing or even wanting to know that you as an individual suffer from specific anosmia regarding this particular scent.

As a perfumer you cannot tell everybody that you are not able to smell floral or musk substances, because you would have all the marketing people crowding round trying to sell you those substances. That was how I learned that a little cedarwood effect could be my reception of androstenone. This is enough to live with, but not enough to detect truffles, which contain markers similar to androstenone.

Methyl dihydrojasmonate (figure 7) is said [11] to smell less intensive as its purity increases. When you have perceived this substance once, you have the impression of blossoming flowers everywhere in nature, especially in springtime. The sensation is not a smell for me but a kind of radiation, which conjures up the picture of a sunbeam sizzling your nose into a springtime feeling. Substances with the same effect, later evaluated by innocent perfumers, were always attributed the quality "smells like paint". So which activity would you propose to search for a receptor for methyl dihydrojasmonate, the paint or the flowery activity?

Being aware of how difficult it is to describe and identify odours, many companies have invented descriptor systems, which they put in a graph showing the intensity for each of 160 descriptors of an aroma (figure 8).

This is a method used to characterise scents more accurately. This is also the way the so-called "electronic noses" work. With their different sensors they adsorb or oxidise vapours of organic material and identify the vapour composition by pattern analysis with neural networks. However, the electronic noses do not give any odour description that could be used in SAR analyses. They are only able to discriminate between headspaces that they have stored.

4.1 Activity in a chemical sense

This is highlighted in an excellent review about "trends in fragrance chemistry" [12] in the German magazine Angewandte Chemie (Applied Chemistry) (figure 9).

What Givaudan researchers make visible and sniffable in this review is a olfactophor model; i.e. known molecules with known scents were put together in one olfactophor, similar molecules with other scents were used to define certain exclusion zones around the olfactophor. This olfactophor model was used to explain scents of new materials that belonged, olfactorily, to the same family. So nobody except the concept users can know what came first, the concept or the molecule ... or the chicken or the egg?

Why don't chemists use the knowledge of, for example, the research results of Hanns Hatt [13] (figure 10) who reports that the first cloned human olfactory receptor "OR 17-40 exhibits a remarkable ability to discriminate structurally closely related molecules like helional and piperonal. Interestingly, to humans, both chemicals smell differently as well." He wonders about the thresholds of some odorants in mammals, especially in humans, which can be as low as a few parts per million. Such high sensitivity is not observed with cloned receptors. Multiple factors may explain the higher sensitivity observed in vivo, including the presence of odorant binding proteins in the nasal mucus. In one figure (figure 11, Diagram B in [13] p.122), the protein encoded by the human OR 17-40 is presented as traversing the plasma membrane seven times).

The finding that the odorant receptors react more sensitively in vivo to odorants than in vitro is analogous to what Amoore found with his dilution method. Some people were extremely sensitive to some molecules, which he recognised as primary odours. The success in isolating and identifying human receptors may mean that special sensitivity to special odorants has nothing to do with receptors but with those multiple factors that may explain higher sensitivity in vivo like a vomeronasal organ.

4.2 The quality of fragrances reassessed

Dietrich Kastner tells us [14] that taste and smell apparently may not be qualities of molecules. This is true in a philosophical sense because it is our mind that makes perception happen by offering the basic conditions for our ability to perceive in time and space. Immanuel Kant therefore created the term a priori. At the point when Hatt was wondering why in his mind structurally related molecules like helional and piperonal smelled different, Kastner knew that from the nearly 20,000 substances which he smelled and characterised, he hadn't found any 2 odorants with a totally identical scent. The conclusion from this observation was, that odour is what we make in our mind out of the reception of odorants. On the other hand, this would mean that the same molecule would also smell different to the same nose at different locations and occasions. This is what I myself am wondering about, since I found in many cases that new synthesised molecules smelled different in Krefeld, where our perfumers work, compared to Holthausen, where I work. This I can explain through what I think is trigeminal nerve stimulation dependency. From small molecules to bigger ones is like switching trigeminal reception to olfactorial perception.

When Günther Ohloff once held a lecture in Düsseldorf about his "triaxial rule of odour sensation in the ambergris odorants family" he was asked during the discussion about the odours of hydrogen cyanide or hydrogen sulphide. His remarkable answer was: "What you perceive from those molecules is not an odour reception" (figure 12, "The nose as spectroscopist" [19]) This is also my theory in explaining the subjectivity of odour reception: Smaller molecules are felt through irritation of trigeminal nerve endings. Examples of these touch sensations are the cooling effect of (-) menthol or the burning sensation of chilli capsaicin or the stinging of acetic acid, the mucous layer membrane wrinkling of acetone, or the pain sensation of carbonic acid. As the studies of impulses with congenitally anosmic subjects have shown, they are well able to receive odour sensation from small molecules and can even make statements about the qualities of the trigger (figure 13).