The Chemistry Of Flavor - C&EN Global Enterprise (ACS Publications)

---

A C&EN Feature

The Chemistry Of

Flavor

92 C&EN APRIL 3, 1967

DR. IRWIN HORNSTEIN DR.ROYTERANISHI U.S. Department of Agriculture

T

he perfume of a rose, the tang of an ocean breeze, the aroma of a sizzling steak—tastes and smells, two of our senses by which we characterize the world around us. And yet, we can not adequately express, define, or explain our taste and smell sensations. We can record the sounds we hear, we can photograph the sights we see, but we cannot store and retrieve the flavor of a food or the scent of a flower except in and from our mind. We can transmit sounds and sights across continents, but we cannot perceive an odor beyond the range of our nose. We can accurately define and characterize light and sound by physical measurements, but we cannot adequately define flavor either qualitatively or quantitatively. R. W. Moncrieff in "The Chemical Senses/' published in 1951, defines flavor as "a complex sensation comprising taste, odor, roughness or smoothness, hotness or coldness, and pungency or blandness." Of these qualities, odor and taste are the main contributors to flavor. Yet if one's sense of smell is deadened, most of the flavor disappears. Taste remains but, without odor, an apple and an onion taste alike. We all know how flavorless foods become when we suffer from a head cold. Odor, then, greatly affects flavor. Determining what volatile compounds are largely responsible for the flavor of a food is the approach many flavor chemists take to try to define flavor. Using instrumental techniques, flavor chemists have detected and identified trace amounts of volatile constituents of foods at an ever increasing rate.

These amounts can be extremely minute. Our sense of smell is so sensitive, that we can detect skatol (3methylindole) at a 3 X 1 0 - n % weightto-weight concentration in air. Recently it was found, however, that skatol purified by zone refining is odorless. The ability of submicrogram amounts of volatiles to profoundly affect odor and thus flavor is apparent. Nevertheless, the increasing sensitivity of chemical instrumentation during the past decade has put the chemist and his nose in the same ball park. Studying the physiology of taste and odor reception is another approach to studying flavor. There are thousands of taste receptors and millions of olfactory receptors. These are minute in size and complex in structure. To understand the mechanism by which these neurons respond and to unravel the code by which we associate the neural response with flavor perception is indeed an awesome challenge to the physiologist. A great deal is known about taste, far less about odor. Let us first review what is known about taste and odor and then look at the contributions of analytical chemists to our understanding of flavor. Taste and chemical

structure

Today, there is general agreement that we recognize four basic t a s t e s sweet, salty, bitter, and sour. However, tastes have not always been so classified. Linnaeus in 1754 classified tastes into 11 groups—sweet, sour, sharp, salty, bitter, fatty, insipid, as-

tringent, viscous, aqueous, and nauseous. A hundred years later, only sweet and bitter were considered basic tastes. Wundt in 1865, the founder of the first experimental psychology laboratory, listed six categories—salt, sour, bitter, sweet, alkaline, and metallic. Later, he dropped the last two categories. Going beyond classification, von Skramlik in 1922 at the University of Freiburg tried to duplicate natural tastes using compounds that he assumed would give the basic four pure taste sensations—quinine (bitter), sodium chloride (salty), tartaric acid (sour), and glucose (sweet). In some cases, his attempts were successful, but complex tastes could not be duplicated. To state that there are four, six, or 11 basic tastes is perhaps more a matter of convenience than of fact, but undoubtedly sweet, sour, bitter, and salt play the dominant roles. Inevitably, attempts to correlate chemical structure and taste qualities are made. Correlations of taste quality and chemical structure have been established for sour and salty tastes. Success at correlating structure with sweet and bitter taste qualities has been less marked. Sour tasting compounds are acidic. Mineral acids cannot be distinguished from one another by taste and their sourness is proportional to hydrogen ion concentration. The relationship between sourness and pH in weak organic acids is not as clear-cut as in the mineral acids. For example, acetic acid is not as sour as an equimolar concentration of a mineral acid, but acetic acid tastes more sour than the APRIL 3, 1967 C&EN 93

The taste receptors in man and other mammals are located on the tongue and soft palate. Distributed over the tongue are numerous small finger-like projections or papillae. The papillae at the back of the tongue are each surrounded by a trench into which a watery fluid is secreted by the surrounding glands. Along the walls of the trench are located the taste buds. Papillae closer to the anterior of the tongue have their taste buds on top rather than in trenches. Within each taste bud are 10 to 15 single taste cells arranged like segments of an orange. The taste cells have small microvilli about 4 microns in length and 0,2 microns in diameter that are presumed to be the excitable portions of the taste receptors. Taste sensitivity, flavor appreciation, and the total number of taste buds decrease with increasing age. Adults have about 9 0 0 0 taste buds; they are located chiefly on the upper portion of the tongue. There are few taste buds in the center of the tongue. A sweet taste is most readily sensed at the tip of the tongue, a sour taste at the edge of the tongue, and a salty taste along the tip and edge. As a result, if a solution containing sweet and bitter components is placed in the mouth we taste the sweet before the bitter and the bitter lingers after the sweet has vanished. Discrimination between tastes, therefore, is, in part, a recognition of these spatial and temporal patterns. Although taste is simpler than smell in that only a limited number of taste qualities are sensed, there are more nerves associated with taste than smell. Three of the cranial nerves are associated with taste, only one (two if the trigeminal component of an odor is included) with smell. These nerve cables contain touch, temperature, and pain fibers in addition to taste fibers. One of the three cranial nerves associated with taste, the chorda tympani can be readily studied by electrophysiological methods. The electrical response of the chorda tympani has been recorded for a variety of taste stimuli in many animal species including man. The integrated or total response f r o m this nerve is a summation of the responses of all the taste receptors in the anterior two thirds of the tongue. When a 0.01M solution of sodium chloride is arranged to flow continuously over the surface of the tongue the integrated response pattern of these taste receptors shows the following pattern. The time between application of the taste stimulus and the recording of a response is approximately 50 milliseconds. During the next two seconds, there is a rapid rise in output voltage followed by a decline to a steady state. This steady-state voltage is more typical of a given chemical stimulus than is the initial rise. If the constant stim94 C&EN APRIL 3, 1967

^^i^âà

In this thin cross section of a typical papilla from a human tongue, the taste buds are the small structures imbedded in the walls of the trench. Each taste bud contains 10 to 15 taste cells. Small microvilli project from these cells and are presumed to be the sites of the triggering reactions

ulus is maintained, sensory adaption takes place. In man, the response to a steady-state application of sodium chloride decreases in intensity until saltiness disappears. This adaption takes place in a matter of t w o minutes. Integrated nerve responses have shown that the taste response patterns differ in every animal species studied. Thus, rodents have strong taste responses and are more sensitive to sodium chloride than to potassium chloride. Carnivores, however, exhibit low taste sensitivity and respond to potassium chloride more readily than to sodium chloride, and cats show practically no response to sweet. Marked differences in taste responses also exist within a given order. For example, rats respond poorly to sugar, but hamsters and guinea pigs respond well. To answer the question: " A r e there specialized taste buds for specific taste qualities," single nerve fibers have been studied. Pfaffmann, while at Brown University, worked with rats and hamsters and found t h a t a single nerve fiber responds to more than one of the so-called four basic taste stimuli. In a series of single-fiber studies, he found that some nerve fibers exhibited broad sensitivity to sugar, salt, quinine, and acids, whereas other fibers responded to salt and sugar or to salt, sugar, and acid, and the like. A small number of fibers did show relatively high specificity to a single stimulus, however. There was no indication that specific receptors were present for any basic taste qualities. Instead, many different patterns of nerve activity appeared to exist depending upon the nature of the chemical stimulus. The magnitude of the response of single nerve fibers to the same stimulus also varied

f r o m fiber to fiber. The basic taste qualities cannot be produced by specific receptor types, but rather by response patterns that are interpreted as salty, sweet, bitter, and sour, and the like. Single nerve fiber studies have conclusively shown that the taste buds do not exhibit specific responses. However, one nerve fiber innervates (supplies with nerves) several taste cells within a taste bud and the possibility that single taste cells may respond selectively to only one type of chemical stimulus has also been studied. Kimura and Beidler, at Florida State University, used very fine microelectrodes to study the responses of single taste cells in a single taste bud of the rat. The results were similar to those obtained

0.03iV 0.1M HC1 KC1

0.1M 0.01M 1.0M NaCl Qu. Sucrose

0.03AT 0.1M HC1 KC1

f r o m single nerve fiber studies. A single taste cell could respond to more than one taste quality. Marked variation in sensitivity to the same stimulus in neighboring taste cells was also observed. It is also known that the nerve fibers intercommunicate through fine branches near their roots and there is evidence that the resulting neural interaction leads to preliminary processing of information at the level of the papillae. The coding of information concerning taste quality is more complex, therefore, than one might anticipate f r o m a system based on four basic taste sensations. The total pattern of nerve activity coming f r o m many taste cells must determine the taste quality perceived.

0.1M 0.01M 1.0M NaCl Qu. Sucrose

0.03N 0.1M 0.1M 0.01M 1.0M HC1 KC1 NaCl Qu. Sucrose

The bar graphs show the frequency of response of nine different single-fiber preparations (A through I) to standard solutions of hydrochloric acid, potassium chloride, sodium chloride, quinine, and sucrose. The graph superimposed on the center graph of a single-fiber response, E, is the relative magnitude of response of a whole, undissected nerve. Results such as these indicate that receptors are not specific for the basic taste qualities APRIL 3, 1967 C&EN 95

Minor structural changes can m a k e a sweet s u b s t a n c e bitter or tasteless S-Amlno-4iiitropropoxybansana

oCHltri1CH3

mineral acid at the same pH. The greater sourness of acetic acid may be due to the dissociation of un-ionized acid molecules to form additional hydrogen ions as the acetate ion is tied up in some manner by the taste receptors. Organic acids may have tastes other than sour; for example, citric acid is sweet and sour and citraconic acid is bitter and sour. Sodium chloride has a characteristic salty taste. However, not all salts are equally salty; for that matter they do not all produce a salty taste. The tastes of salts are a function of both cation and anion. Examples of salts that taste predominantly salty include NaCl, KC1, NH 4 C1, LiCl, RbCl, NaBr, NH 4 Br, LiBr, Nal, Lil, N a 2 C 0 3 , K N 0 3 , and N a 2 S 0 4 . Salts that elicit simultaneous bitter and salt sensations include KBr and NH 4 Br and salts that are predominantly bitter include CsCl, RbBr, CsBr, KI, Rbl, Csl, and MgS0 4 . Salts of beryllium are usually sweet as are some lead salts. Mg 2+ and N H 4 salts are usually bitter. An absolute correlation between taste and physical or chemical properties of salts has not been established but, in general, low-molecular-weight salts are salty and as the molecular weight increases, the taste shifts to bitter. Sour and salty responses are produced by reasonably well-defined classes of compounds. In contrast, no single class of compounds can account for either the sweet or bitter taste responses. Almost every class of organic compounds contains substances that will evoke either sweet or bitter sensations. The structural relationship between compounds that are characterized as either sweet or bitter is often very close. Minor changes in the structure of a molecule can make the difference between a sweet substance and a bitter substance. For example, 2-amino-4-nitropropoxybenzene is about 4000 times sweeter than sucrose, 4-amino-2-nitropropoxybenzene is tasteless, and 2,4-dinitropropoxybenzene is bitter. Dulcin (p-ethoxyphenylurea) is very sweet, but the corresponding thiourea is bitter and the o-ethoxyphenylurea is tasteless. Similarly, saccharin is very sweet, but the introduction of a methyl group on the imido nitrogen produces a tasteless compound. Stereoisomers and even anomers may have widely different tastes. The anti form of the oxime of 96 C&EN APRIL 3f 1967

Four thousand times tweeter than micro*

4-Amino-2nitropropoxybensana

3,4-Dinitroprapoxybansana

oeH-tCHtCrtj

OCH a CH t CH 3

Taeteiem

Bitter

p-Ethoxyphenyhira* p-Ethoxyphenyltbiouraa o-Kthoxyphenyluraa

II

II Very tweet

Bitter

Taeteiem

Saccharin

JV-Mathylaaocharin

Very tweet

Taeteiem

rAniaaldoziina (anti form)

i-Aniaaldoxima (ayn form)

CH OH

* O"«

Very

anisaldehyde, a-anisaldoxime, is intensely sweet but /?-anisaldoxime (syn) is tasteless, D-glucose is sweet, whereas L-glucose has a slightly salty taste. The anomers of D-mannose have different tastes; a-D-mannose is sweet while ^-D-mannose is bitter, and many, many other examples could be cited. It is interesting to note that both saccharin and cyclamate (sodium cyclohexylsulfamate), the two most widely used synthetic sweeteners, were accidental discoveries. Although no valid theory for predicting which substances will be sweet or bitter exists, Moncrieff has proposed limited generalizations. Among these are the following: •Polyhydroxy and polyhalogenated aliphatic compounds are generally sweet. • a - A m i n o acids are usually sweet, but the ft- or y-amino acids are not. The closer together the amino group and the carboxyl group are, the greater the sweetness. • On ascending a homologous series, the taste frequently changes from sweet to bitter, and taste and water solubility disappear simultaneously.

Taeteiem

•Alkylation of an amino or amido group often gives a sweet tasting compound. • One nitro group in a molecule often gives a sweet taste. Two or three nitro groups produce bitter tasting substances. • Some aldehydes are sweet; ketones are never sweet. The phenylhydrazones of sweet aldehydes are not sweet but the oximes are often sweet. • Free bases, particularly alkaloids, are bitter. • Introduction of a phenyl group frequently causes a bitter taste and may make sweet compounds bitter— glycerol is sweet, but 3-phenoxy-l,2propanediol is bitter. • Ureas may have a sweet taste; however, symmetrical ureas are not sweet. Thus, urea is bitter, but 1,1dimethylurea is very sweet. Taste perception

mechanism

In a neuron the potassium ion concentration is higher within the cell than outside the cell membrane; the opposite situation is true for the sodium ion concentration. Thus, a

resting potential exists across the cell wall. When this membrane is temporarily "punctured," sodium ions move in and potassium ions move out, the resting potential drops, and a momentary depolarization spreads along the neuron. This wave of depolarization flashing along the fiber constitutes the nerve impulse. Once the neuron has fired it takes a certain time (about 1 millisecond) for the membrane potential to be re-established. Energy is stored during this "recharging" period and the triggering of the nerve requires far less energy than is released by the nerve response. Beidler proposed in 1954 that the important first step in taste stimulation is an adsorption of the chemical stimulant on a receptor site. He suggests that there is a unimolecular reaction between the stimulant and some part of the taste cell. The reaction may be written as: A + B —> AB, where A is the stimulant, B an unfilled receptor site, and AB a filled receptor site. Applying the law of mass action and assuming that the magnitude of the response is proportional to the number of sites filled and that the maximum response occurs when all the sites are filled, Beidler derived a fundamental taste equation: R

Rs

KRf,

In this equation C is the concentration of the applied stimulant. R is the magnitude of the nerve response, Rs is the maximum obtainable response for a given substance, and K is the equilibrium constant for the stimulantreceptor reaction. Plots of the ratio of molar concentration of the stimulant to the magnitude of the integrated response of the taste receptors C/R, vs. the molar concentration, C, for a number of salts gave straight lines as would be predicted from Beidler's equation. From the slope 1/RS and the intercept at C = 0, equilibrium constants were calculated. Salt

Equilibrium constant, K

Sodium chloride

9.80

Sodium formate

9.00

Sodium acetate

8.55

Sodium propionate

7.58

Sodium butyrate

7.72

The low values of K agree with the assumption that the initial step is an adsorption of the chemical stimulant on the surface of the taste receptor and not an enzymic reaction between stimulant and receptor. Beidler also found that the magnitude of the response to stimulus did not vary between 20 and 30° C , another indication that probably only a minor molecular rearrangement at the surface of the receptor is involved. Beidler proposed that this slight rearrangement leads to a small and temporary breakdown in the cell membrane. The resulting depolarization triggers the nerve impulses that are transmitted to the taste centers in the brain. These impulses last only a few milliseconds and are sent along the nerve at maximum frequencies of 100 to 200 impulses per second. It is the frequency of these impulses that signals the intensity of the taste stimulus. The energy to initiate the nerve impulse does not arise from the stimulus. Rather, the stimulus creates small changes in the spatial arrangement of the receptor surface molecules that result in large response changes. Dastoli and Price of Monsanto Research Corp. recently confirmed that weak complex formation may be the triggering mechanism for taste stimulation. They isolated from bovine taste buds a protein fraction that complexed sugars and saccharin. As predicted by Beidler's taste equation, a plot of C/R vs. C, where C was the in vitro sugar concentration and R the change in refractive index upon interaction of the active protein fraction with the sugar, gave a straight line. The calculated equilibrium constants were characteristic of weak interactions such as hydrogen bonding rather than of chemical reactions. The strength of the complexes paralleled the relative sweetnesses of the sugar. Enhancement and inhibition Taste enhancement and taste inhibition also are important phenomena in determining flavor. For example, gymnemic acid completely abolishes sweet sensations. After one chews the leaves of the asclepiad Gymnemia sylvestre (a member of the milkweed family), the plant that contains this material, table sugar has no sweet taste, but is only gritty. Bitter sensa-

tions also decline, but salty and sour responses remain untouched. This effect wears off in an hour or two. Miraculous Fruit, the berry of a Nigerian plant, abolishes the reponse to sour substances and as a result, a sour lemon approaches the taste of a sweet orange. These inhibitory compounds may act by being preferentially adsorbed and then being very slowly desorbed from receptor sites that respond strongly to sweet and sour. Monosodium L-glutamate (MSG) in purified form has been commercially available since 1909 as a flavor enhancer for natural food products. Recently, 5'-nucleotides, especially disodium 5'-inosinate ( I M P ) and disodium 5'-guanylate ( D M P ) , have received considerable attention as flavor enhancers. Using electrophysiological methods, it has been found that single nerve fibers of the chorda tympani respond to MSG, IMP, and DMP, indicating that the action of these compounds is at the receptor surface and is probably similar to that of other compounds evoking taste responses. Beidler has pointed out that the nucleotides are also strong chelating compounds. In addition to initiating taste responses, the nucleotides may combine with metal ions in such a manner as to make additional receptor sites more readily available to the normal taste-stimulating compounds present in foods, thus enhancing the food flavor. The inhibition and enhancement of taste in foods has not been adequately explored. The possibility of finding other compounds that can be used to subdue undesirable flavor notes or accentuate desirable flavors deserves increased study.

Odor classification Attempts at classifying odors according to quality have a long history. Linnaeus attempted to p u t order into odor and taste by postulating seven basic odors. Zwaardemaker in 1895 devised a system in which all odors were divided into nine major categories : • Ethereal—fruits, resins, and ethers. • Aromatic—camphor, cloves, lavender, lemon, bitter almonds. •Balsamic or fragrant—flowers, violet, vanilla, coumarin. • Ambrosial—amber, musk. APRIL 3, 1967 C&EN

97

Electrophysiological studies afford one approach to the understanding of olfaction. Although the contribution of the electrophysiologists has been hampered by the small size of the olfactory neurons, instrumental advances are overcoming inherent difficulties. Progress has been substantial during the past decade. An appreciation of the electrophysiologist's problem can be gained by briefly looking at the olfactory system. The olfactory epithelium is yellow to brown and is composed of chromolipid substances. Whether this pigment is involved in the primary olfactory process is still a matter of conjecture. The olfactory cells are primary neurons. Their number is impressive—there are about 10 to 20 million such receptors in man, and about 100 million in the rabbit. In man the total area of the olfactory region is about 10 cm. 2 and is located in the upper respiratory passages and separated f r o m the brain by a thin plate of bone. The olfactory epithelium, as well as the entire surface of the nasal cavity, contains bare nerve fibers f r o m the trigeminal nerve. It is frequently assumed that the epithelium and nasal surface participate in odor detection solely by evoking a pain response. However, nasal trigeminal sensitivity to phenylethyl alcohol is below that of olfactory sensitivity. This is also true for other compounds, indicating that nonolfactory receptors may be participating in odor detection by mechanisms other than those evoking pain. The sensitive endings of the olfactory cell are minute cilia 0.1 to 0.2 micron in diameter. The cilia (6 to 8 per cell in man, 10 to 14 in the rabbit) are kept moist with mucous secreted by adjacent cells. They extend 1 to 2 microns beyond the surface of the epithelium and are about 100 microns long. The surface area of these cilia is enormous—in the rabbit it is estimated that the 100 million olfactory " h a i r s " have a combined surface area of nearly 600 cm 2 . The axons leading f r o m the cell body to the olfactory bulb are only several millimeters in length and range f r o m 0.1 to 0.5 micron in diameter. Many of these axons are beyond the range of the light microscope and are among the smallest and slowest conducting axons in the body. As the axons leave the basal cell layer of the epithelium, they group together in bundles and pass through the thin perforated bone (cribri98 C&EN APRIL 3, 1967

Man's odor sensors are more closely linked to his brain than any other of his body's sensors. This cutaway sketch of a human nasal cavity shows the specialized olfactory epithelium in the roof of the cavity. In man, this olfactory region covers about 10 cm.2 The olfactory region in one's nasal cavity is separated from the olfactory bulb of the brain by a thin plate of bone, the bony cribriform. The axons, a few millimeters long, lead directly from the olfactory cell bodies through the cribriform plate into the olfactory bulb

I

I I

f o r m plate) separating the nasal chamber f r o m the brain cavity, and enter the olfactory bulb. Within the olfactory bulb, the axons terminate in numerous small bodies (glomeruli). Each olfactory neuron therefore leads directly f r o m the point of stimulation to the olfactory bulb and only one nerve junction in the glomerulus separates the initial receptor f r o m the olfactory lobe of the brain. A more direct path f r o m the outside world to the brain does not exist for any of our other senses. Because of the axon's small size, it has been difficult to record neural activity f r o m a single axon; however, Beidler and Tucker have recorded neural activity f r o m small bundles of primary olfactory neurons. They dissected f r o m the olfactory nerve a small bundle of fibers containing a minimum of about 2 5 0 0 fibers. They cut this t w i g where it entered the olfactory bulb and placed it on platinum electrodes under mineral oil. The electrodes served to stimulate the olfactory nerves and to record odor responses. By stimulating the t w i g and mapping the area where the impulses were received at the olfactory epithelium, bundles 10 to 4 0 microns in diameter were found to innervate oval areas roughly 1 mm. 2 Having mapped the olfactory region, the response of small areas of the olfactory region t o chemical stimuli could be studied. Neural activity has also been recorded by placing electrodes directly into

t h e olfactory e p i t h e l i u m . M o u l t o n , while at Florida S t a t e University, recorded neural activity f r o m p e r m a n e n t l y i m p l a n t e d electrodes in t h e olfactory bulb. O t t o s o n , of t h e Karolinska Institute, Stockholm, m a d e t h e first extensive studies of t h e olfactory mucosa response. H e placed electrodes directly on t h e olfactory mucosa and recorded a slow (2 to 3 seconds) negative potential in response to a chemical stimulus directed at t h e s a m e a r e a . H e n a m e d this response t h e electroo l f a c t o g r a m ( E O G ) and considered this response to be the g e n e r a t o r p o t e n t i a l — t h e potential recorded directly f r o m t h e receptors and an indication of t h e p r i m a r y events in olfaction. However, slow potentials can be induced by m a n y f a c t o r s . For e x a m p l e , H a r t m a n and Wilkens, a t Cornell University, have shown t h a t t h e passage of a volatile chemical across a polished w i r e microelect r o d e in contact with a dilute electrolyte can gene r a t e a potential. It is, t h e r e f o r e , not certain t h a t t h e E O G is indeed a g e n e r a t o r potential. A d r i a n , at C a m b r i d g e University, first recorded t h e activity of t h e olfactory bulb and f o u n d t h a t threshold concentrations of different odorants evoked responses in selective units. However, at concentrations above threshold m a n y c h e m i cals produced responses in t h e s a m e cells. T h e r e w a s no indication t h a t either specific receptors w e r e present or t h a t any one chemical w a s m o r e f u n d a m e n t a l t h a n any other in evoking smell sensations. A d r i a n also observed t h a t although specific receptors w e r e not present, one a r e a of t h e olfactory bulb might respond m o r e vigorously to one chemical t h a n to another. H a i n e r and his c o w o r k e r s at A r t h u r D. Little have suggested t h a t t h e r e m a y be a sorting out of information in t h e olfactory bulb. For e x a m p l e , although about 2 0 , 0 0 0 axons e n t e r each glomerulus in a rabbit, only 2 4 mitral cells (secondary olfactory neurons) lead out. It m a y be t h a t t h e glomeruli serve as a center, w h e r e signals f r o m t h e 2 0 , 0 0 0 p r i m a r y neurons are sorted according to t y p e . T h e signals f r o m one t y p e of sensor are t h e n c o m b i n e d and f e d into a single output c h a n n e l — a single mitral cell. Since 2 4 mitral cells leave each glomerulus t h e r e are 2 4 specific types of receptors. A l t h o u g h no experim e n t a l evidence has been offered to support this view, it does point up t h e possibility t h a t studies

on t h e olfactory bulb which involve secondary neurons need not faithfully reproduce processes t a k i n g place a t t h e olfactory e p i t h e l i u m . Evaluation of t h e electrophysiological d a t a obtained f r o m t h e t h r e e general m e t h o d s of experim e n t a t i o n t e n d s to support t h e following conclusions concerning t h e m e c h a n i s m of olfaction: • Different classes of receptor sites m a y be distributed in different ratios over each p r i m a r y olfactory n e u r o n , t h u s providing a limited d e g r e e of specificity. T h i s type of distribution can account for both t h e lack of specificity and t h e diff e r e n c e in response m a g n i t u d e for different molecular species obtained both f r o m t h e olf a c t o r y bulb and f r o m f r o g mucosa by Gesteland and others of M I T . T h e y found t h a t for any given chemical over one half of t h e olfactory neurons used for recording activity responded, but t h a t no t w o responded in t h e s a m e w a y . • Differences in t e m p o r a l patterns of response evoked by different c o m p o u n d s m a y contribute to odor discrimination. T h e t i m e difference of response m a y be related to t h e distance t h e receptor is f r o m t h e nasal e n t r a n c e , t h e a m o u n t of mucus covering t h e receptor, and t h e partition coefficient of t h e c o m p o u n d b e t w e e n t h e vapor phase and t h e aqueous phase (mucus) and bet w e e n t h e aqueous phase a n d t h e lipoid phase (olfactory e p i t h e l i u m ) . • Differences in spatial patterns of response evoked by different c o m p o u n d s m a y contribute t o odor discrimination. Spatial specificity has been observed in t h e olfactory bulb of cats w h e r e t h e posterior dorsal region is particularly sensitive t o t h e odor of decayed m e a t or fish. Lateral and medial differences of response have been obtained f r o m t h e olfactory mucosa of frogs. Responses f r o m t h e olfactory bulb of other animals also show t h a t at low concentrations s o m e water-soluble c o m p o u n d s s t i m u l a t e t h e f r o n t of t h e bulb while water-insoluble hydrocarbons preferentially excite t h e back of t h e bulb. It m a y well be t h a t odor discrimination resembles t a s t e discrimination and is based on t h e d e g r e e of receptor specificity and t h e varying t e m p o r a l and spatial patterns of response. From t h e electrophysiologists viewpoint it is doubtful t h a t highly specific receptors for particular c h e m ical c o m p o u n d s exist or t h a t specificity alone can be t h e basis of odor discrimination. APRIL 3, 1967 C&EN 99

Compounds with very different structures can have similar odors; these compounds, for example, have a typical musk-like odor

(occur* in natural muok)

A M-And«ort«tt t« ol

0h

AsJk/™*

^ 2r^ »-Ao«tyl-l,l.J,J,M.*k«pt*m«thyllnd&n«

• Alliaceous—hydrogen sulfide, arsine, chlorine. • E m p y r e u m a t i c — r o a s t e d coffee, benzene. • Caprylic—cheese, rancid fat. • R e p u l s i v e — d e a d l y nightshade, bedbug. • Nauseating or fetid—carrion, feces. Other odor classification systems have used four, six, 18, and 24 basic categories. Implicit in these classification schemes is the assumption that specific active sites are present on the olfactory receptors. There have been many attempts to correlate chemical structure and odor quality, but the results have been disappointing. Compounds with very different structures may exhibit similar odors while compounds with very similar structures may have quite different odors. To illustrate, six different classes of compounds possessing a musklike odor are known—macrocyclic musks, steroid musks, nitro musks, indane and naphthalene musks, and benzene musks. Structurally, these classes are completely different, but their odor qualities are alike. Stereoisomers, however, may have quite different odors. The odors of the four menthols are a classic ex-

< ~Yr KJKJ\ CH,