J~uRNALOF

NEUROPHYSIOLOGY

Vol. 68, No. 6, December

1992.

Prlntrd’

ing of Odor Molecules by itral/Tufted Olfactory Bulb. I. Aliphatic Compounds KAZUYUKI

IMAMURA,

Department

ofNeuroscience,

SUMMARY

AND

NOBUKO

MATAGA,

AND

Osaka Bioscience Institute,

MORI

Suita, Osaka 56.5, Japan

INTRODUCTION

CONCLUSIONS

1. Recordings of extracellular spike responses were made from single mitral/tufted cells in the main olfactory bulb of urethan-anesthetized rabbits. Olfactory epithelium ipsilateral to the recorded olfactory bulb was stimulated with homologous series of aliphatic compounds using periodic artificial inhalations. 2. In the dorsomedial part of the main olfactory bulb, single mitral/ tufted cells were activated by subsets of y1-fatty acids with similar hydrocarbon chain lengths. Response selectivities of single mitral/tufted cells were examined in detail using a series of y1-fatty acids at five different concentrations. The results indicate that although the range of effective fatty acids is broader at the higher concentrations, the best response at higher concentrations was similar to that determined at lower concentrations. 3. Analysis of single-unit responses to the panel of fatty acids, including those with branched hydrocarbon chains, suggested that the determinants for the response specificities of individual mitral/ tufted cells in the dorsomedial region include the overall size of hydrocarbon chains of the odor ligand molecules. 4. Single mitral/tufted cells in the dorsomedial region tended to be activated not only by fatty acids but also by n-aliphatic aldehydes. For a panel of a homologous series of n-aldehydes at five different concentrations, individual mitral/tufted cells showed response selectivity to subsets of aldehydes with similar hydrocarbon chain lengths. 5. In most cases, normal aliphatic alcohols and alkanes were ineffective in activating mitral/tufted cells in the dorsomedial region. This suggests that carbonyl group (- C = 0) in the odor molecules plays an important role in determining response specificity of these neurons. 6. Examination with an expanded panel of stimulus odor molecules that included ketones and esters indicated that single mitral/ tufted cells sensitive to subsets of fatty acids and n-aliphatic aldehydes were also responsive to subsets of ketones and/ or esters having hydrocarbon chain lengths similar to those of the effective fatty acids and aldehydes. 7. The present results show a clear correlation between the tuning specificity of individual mitral/tufted cells and the stereochemical structure of the odor molecules, with respect to 1) length and/ or structure of hydrocarbon chain, 2) difference in functional group, and 3) position of the functional group within the molecule. 8. A hypothetical diagram suggesting functional convergence of olfactory nerve input to individual glomeruli is proposed to explain the mechanism for selective activation of individual mitral/ tufted cells by a range of odor molecules with similar stereochemical structures.

1986

KENSAKU

Cells in Rabbit

In the visual, somatosensory, and auditory systems, knowledge of the receptive fields and tuning specificities of individual neurons at successivelevels of central pathways has provided the basis for understanding sensory information processing in the CNS (Aitkin et al. 1984; Hubel and Wiesel 1962; Mountcastle 1984). In the olfactory system, however, becauseof the immense number of different odor molecules, it has been technically difficult to characterize the responsespecificities of individual neurons to odor stimulation in a manner similar to that used in the other sensory systems.The most critical information relates to 1) mechanisms of interactions between odor molecules and receptor molecules, 2) functional subsetsof olfactory receptor neurons, and 3) their direct or indirect connections with neurons in the central olfactory system. Olfactory receptor neurons in the nasal epithelium recognize odor molecules and send the information via their axons to the olfactory bulb. Recently, Buck and Axe1 ( 199 1) have found a multigene family with > 100 members, which presumably encodes odor receptor proteins. These proteins are members of G protein-coupled receptors with seven transmembrane domains, suggesting that the basic mechanism of molecular interaction between odor ligands and receptor proteins can be deduced from knowledge of well-characterized members of the G protein-coupled receptor protein family such as,&adrenergic receptor proteins ( O’Dowd et al. 1989) and photoreceptor proteins (Nathans and Hogness 1984; Nathans et al. 1986). It has thus been suggestedthat, as is the casewith other G protein-coupled receptors, odor receptor proteins may recognize stereochemical features of ligand odor molecules (Buck and Axe1 1991; Lancet 1986; Shepherd and Firestein 199 1). However, the odor responsespecificity of individual subtypes of the odor receptor protein is not known yet, nor has the response specificity of individual olfactory receptor neurons been studied systematically in terms of stereochemical properties of ligand odor molecules. Olfactory receptor neurons project their axons to the glomeruli in the olfactory bulb, where they make excitatory synaptic connections with dendrites of mitral, tufted and periglomerular cells (Shepherd 1972). In the rabbit main olfactory bulb (MOB), - 2,000 discrete glomeruli are arranged in the superficial layer (Clark 1957 ) . These glomeruli form structural units with the following characteristics: 1) an individual olfactory receptor neuron sends its

0022-3077192 $2.00 Copyright 0 1992 The American Physiological Society

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR

CODING

axon only to a single glomerulus; and 2) an individual mammalian mitral/tufted cell sends its primary dendrite only to a single glomerulus. Therefore it can be expected that the response specificities of individual mitral/tufted cells to odor molecules reflect the response specificities of a subset of olfactory receptor neurons that project their axons to the glomerulus. The anatomic structure suggests that the glomeruli are functional units as well, a notion which is supported by studies using 2-deoxyglucose mappings (Lancet et al. 1982) and electrophysiological recordings ( Buonviso and Chaput 1990). In the present study, we recorded extracellular single-unit responses of mitral/ tufted cells to examine in detail the specificities of cellular tuning in terms of differences in the stereochemical properties of the odor molecules. In a previous paper (Tori et al. 1992), we reported that there is a group of mitral/ tufted cells sensitive to fatty acids in the dorsomedial region of the olfactory bulb. This focus of fatty acid-sensitive neurons enabled us to sample consistently in different animals from the group of mitral/ tufted cells, many of which are sensitive to fatty acids, and to characterize in detail the response specificities of these neurons. In the present experiments, we have concentrated our recording in or around the dorsomedial region and used a homologous series of aliphatic compounds (fatty acids, aliphatic aldehydes, aliphatic alcohols, alkanes, ketones, and esters) as stimulus odor molecules. METHODS

Surgical preparation Experiments were performed on 19 male adult rabbits (Japanese white, Oriental Bio, Kyoto) weighing 1.9-2.8 kg. Animals were anesthetized with an intravenous injection of 30% urethan ( 1.2 g/kg). Fine polyethylene tube (PE-60, Clay Adams) was inserted into a radial vein for delivery of infusion solution. Tracheotomy was performed for double cannulation (Onoda and Mori 1980). A glass cannula was inserted caudally into the trachea for spontaneous respiratory flow. A flexible polyethylene tube was inserted rostrally into the postnasal cavity through the larynx for control of nasal air flow. Animals were then mounted in a stereotaxic apparatus. Body temperature was maintained at 385°C by a homeothermic heat pad system ( ATB- 1100, Nihon Kohden) . Electroencephalogram ( EEG), electrocardiogram (ECG) , the rate of the respiration, and the heart beat were monitored throughout the experiment. The cerebrospinal fluid was drained at the atlantooccipital membrane to minimize brain pulsation. The bone and dura overlying the left olfactory bulb were widely removed. The exposed surface of the brain was covered with a thin layer of vaseline-minera1 oil mixture to prevent drying. A small hole was drilled in the bone overlying the anterolateral portion of the frontal neocortex for inserting a bipolar stimulating electrode into the lateral olfactory tract (LOT) running at the surface of the anterior piriform cortex. The final position of the stimulating electrode was determined by monitoring the LOT-evoked field potentials in the MOB (Phillips et al. 1963). The stimulating electrode was then anchored to the skull with dental cement. The depth of anesthesia was judged by the EEG pattern and the rate of respiration. Animals were maintained under deep anesthesia throughout the recording session by infusion of lactated Ringer solution (3.3 ml/h) containing urethan ( 125 mg kg-’ h-l). In addition, the infusion solution contained glucose (250 mg/h), l

l

IN

OLFACTORY

1987

BULB

dexamethasone (0.2 mg/ h), essential amino acids (0.1 ml/h, Hypleamine), and riboflavin phosphate (0.05 ml/h, Towa, Bislase).

Artzjicial

inhalation

of odor-containing

air into the nose

The postnasal cannula was connected with an inlet of an artificial respirator to produce negative-pressure pulses periodically for drawing air and odor molecules through the nasal cavity. The rate of the artificial inhalation was held constant at once per 1.5 s. One cycle consisted of a period of negative pressure application (750 ms) and a period of no air flow (750 ms). A pressure transducer (TP-300T, Nihon Kohden) was connected to the postnasal cannula by a T tube to measure the air pressure within the cannula. The voltage output of the transducer was monitored throughout the experiment using a thermal array recorder. A flexible polyethylene tube was inserted into the nostril to keep the profile of air flow constant. Compressed air was deodorized by passage through activated charcoal and then humidified by passage through distilled water. The air was then led at a constant flow rate through a glass funnel to the space in front of the animal’s nose to minimize the contamination with odors present in the experimental room. To remove the odor-containing air used for stimulation, a gentle suction connected to a vacuum pump was placed at a distance of - 15 cm from the animal’s nose.

Stimulation

with odor molecules

Tables 1 and 2 show the compounds that were used for odor stimulation in this series of experiments. In addition, a number of compounds were used occasionally to examine further the response specificity of olfactory bulb neurons; these include ketones, esters, and aromatic compounds. All chemicals were analytical grade. Each chemical was diluted to 5 X 10 -2 ( vol/ vol ) in mineral oil (Sigma). The mineral oil itself produced no response. For examining dose-response relationships, normal fatty acids and aldehydes were diluted in mineral oil to 1 X 1OA3, 3 X 1OV3, 1 X 1Op2, 3 x 10A2, and 1 X 10-l (vol/vol). One milliliter of each diluted solution was stored in a glass test tube ( 10 cm height, 12 mm ID) sealed with a screw cap. For stimulation, the test tube was uncapped and then placed in front of the animal’s nostril tube. The distance between the open top of the test tube and the tip of the animal’s nostril tube was -7- 10 mm. The duration of an odor application was 3-5 s. By comparison with the method using a flow dilution olfactometer (cf. Meredith 1986)) the present method of odor delivery has disadvantages in controlling exact odor intensity and timing of delivery. However, the present method proved to be far more suitable than the olfactometer method for preparing a large number of different odor molecules. Because the purpose of the present study is to examine response specificities of individual mitral/tufted cells using as many different odor molecules as possible, we adopted this method. To check the consistency of the responses of individual cells, the examination of the response to each odor molecule was repeated at least three times. The sequence of application of different odor molecules was changed randomly. Except for a small variation in the number of elicited spikes, we observed highly reliable responses to repetitions of the same stimulus.

Recording ofsingle-unit

activity

Extracellular responses of single units and oscillatory local field potentials (OLFPs in the external plexiform layer (EPL), Freeman and Skarda 1985; Mori et al. 1992) in the olfactory bulb were recorded by glass micropipettes (7-9 MQ) filled with 2% pontamine sky blue in 0.5 M sodium acetate The signals were fed into a conventional amplifier and monitored on an oscilloscope. The

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

K. IMAMURA,

1988 TABLE

N. MATAGA,

AND

MORI

K.

Initial panel of odor molecules

1.

Abbr.

Molecules

Molecules

Structure

Abbr.

Structure

A. Carboxylic acids Molecules Acetic

Abbr.

acid

Molecules

Structure Isobutyric

(2)COOH

Abbr.

acid

\OOH Propionic

acid

(3)COOH

n-Butyric

acid

(4)COOH

Structure

(IB)COOH

OOH Isovaleaic

acid

&OOH

2-Methylbutyric

acid

(MB)COOH

-0OH n-Vale&

acid

Benzoic

(5)COOH

acid

-0OH n-Caproic

acid

Phenylacetic

(6)COOH

acid

-0OH n-Heptylic

acid

Cinnamic

(7)COOH

acid

-0OH n-Caprylic

acid

Crotonic

(8)COOH

acid

COOH n-Pelargonic

acid

(9)COOH

COOH n- Capric

acid

(lO)COOH

COO/\

B. Others Limonene

LIM

Benzaldehyde

(Bz)CHO

a-Ionone

a-1

Abbr., abbreviations. action potentials and OLFPs were differentially amplified and separated using bandpath filters (300 Hz-3 kHz and 1.6- 100 Hz, respectively). They were recorded together with monitors for artificial inhalation and odor stimulation by a thermal array recorder (Nihon Kohden). Field potential responses to supramaximal LOT volleys were monitored to check the location of the tip of the recording micropipettes in relation to the layers in the bulb. Single units were recorded in the external plexiform layer and the mitral cell layer (MCL). Mitral cells were identified by their antidromic spike responses (judged by the short and fixed onset latencies) to LOT stimulation. Single units showing LOT-evoked spike responses typical for granule cells or short-axon cells (Mori and Takagi 1978 ) were excluded from the present study. In the previous report (Mori et al. 1992)) we demonstrated that there is a group of mitral cells sensitive to fatty acid odor molecules in a region at the dorsomedial part of the olfactory bulb, and that large OLFPs are elicited in or around the region after stimulation with the fatty acid odor molecules. To place the recording micropipette in or near the fatty acid-sensitive region, we first made in each experimental animal a rough mapping (300- to 500pm intervals between recording points) of fatty acid-induced OLFPs in the dorsomedial part of the MOB. When a large OLFP was obtained in response to normal fatty acid odors, the recording

point was noted in a sketch in reference to the distribution of blood vessels running at the surface of the bulb. Single-unit recordings were obtained mainly in or around the fatty acid-sensitive region. However, single-unit recordings were obtained also at sites distant from the fatty acid-sensitive region for comparing the characteristics of the response specificities. Recording micropipettes were inserted vertically from the dorsal surface of the bulb. During one penetration, the pipettes usually encountered the MCL twice, once at the dorsal MCL and then at the ventral MCL. In this experiment, neurons in dorsal regions indicate single units recorded in or superficial to the dorsal MCL. Neurons in ventral regions are those encountered in or deeper to the ventral MCL.

Histological

procedures

For the reconstruction of recording tracks, a cathodic DC current ( 10 PA, 5 min) was passed to eject the marker dye at one or two different depths. The animals were then given a lethal dose of urethane and perfused transcardially with saline followed by phosphate-buffered 10% Formalin. The brains were removed and post-fixed overnight. Olfactory bulbs were sectioned serially in the frontal plane at 50 pm using a freezing microtome and stained for

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR TABLE 2.

Normal

aliphatic

compounds

No. of Carbon Atoms

2

CODING

Aliphatic Aldehydes

b\ OH Acetic acid (Z)COOH

acid

0 02 r/ 'H Propylaldehyde (3)CH0 0 WC
In’addition to [ (4) and ( 5 )COOH] , this mitral cell was activated also by fatty acids with a branched hydrocarbon chain [ (IB), (IV), and (MB)COOH]. Among the three iso-fatty acids, the response to [ (MB)COOH] was most intense and showed three burst discharges per each inspiration pulse. The spike response lasted 530 s after removal of the odor test-tube. The mitral cell showed no spike response to aromatic carboxylic acids, nor to crotonic acid (Table ‘QW r (6 1A), nor to three control odor molecules (limonene, benzaldehyde, and cr-ionone, Table 1B). To determine whether tuning specificity depends on the 5 set concentration of the stimulus odor molecules, we exam8 (7)COOH ined dose-response relationships for the n-fatty acid odor molecules. Figure 211 shows spike responses of the same mitral cell as in Fig. 1 to a graded concentration of [ (4)COOHl. The mitral cell showed no spike response to (8 1 X 10 -3 ( vol/ vol ) of [ (4) COOH] . Clear responses were obtained with concentrations of 23 X 1OA3 (vol/ vol ) . Increased concentration of [ (4)COOH] resulted in a larger number of spike discharges, and at a concentration of 1 X FIG. 1. Single-unit responses of a mitral cell to carboxylic acid odors. 10 -I ( vol/ vol) , the responses outlasted the period of odor A : sample traces of spike activity before, during, and after the odor stimulaapplication. tion. Odor molecules used for stimulation are indicated in the abbreviated Spike responses of this mitral cell to five different conform on the left shoulder of each trace (see Table 1 for structure and centrations of all members of the n-fatty acid odors [ (2)nomenclature of the odor molecules). Traces at left show responses to ( lO)COOH] (Table 1) were likewise examined, and the normal fatty acids [ ( 3 ) - ( 8 ) COOH 1, whereas those at right indicate responses to iso-fatty acids [ (IB), (IV), and (MB)COOH]. Monitors for results are summarized in a three-dimensional diagram in artificial inhalation (thin line with periodic downward reflections) and for Fig. 2 B. At the concentration of 3 X 1Oe3, only the the period of odor stimulation (thick bar) are shown under both left and [ (4)COOH] was effective in activating the mitral cell. right spike activity traces. Each downward reflection in the inhalation When the concentration was increased to 3 X 10e2, both monitor indicates an inhalation pulse during which the air in front of the animal’s nostril is drawn into the nose. There is no air flow in the animal’s [ (4)COOH] and [ ( 5 )COOH ] activated the mitral cell, nose during the period between these inhalation pulses. Spike height is -6 whereas the other n-fatty acids were ineflective. With the mV. B: superimposed traces (4 sweeps) of spike response to lateral olfachighest concentration ( 10 -’ ) , four members of the n-fatty tory tract (LOT) volley recorded from the same neuron as that shown in A. acids [(2)-( 5)COOH] with similar hydrocarbon chain This neuron was identified as a mitral cell on the basis of the antidromic lengths activated the mitral cell, but the remaining five spike response. Arrow: time of LOT stimulation. )COOH

)COOH

a$

)COOH

t

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR

CODING

IN OLFACTORY

BULB

1991

A 1O-3 3x1o-3 1O-2 3x1o-2 10-l

(4) COOH 3 set FIG. 2. Concentration-response relationship of a mitral cell to the stimulation with a homologous series of M-fatty acid odors. A : spike responses to increasing concentrations of n-butyric acid [ (4)COOHl. Concentrations (vol/vol) of the (4)COOH are shown on the left shoulder of each trace. A monitor for artificial inhalation is shown under the spike activity traces. Thick horizontal bar at the bottom: period of odor application. Spike height is -6 mV. B: 3-dimensional diagram showing the response specificity for M-fatty acid odor molecules of the mitral cell shown in A. This diagram is constructed from the data obtained by measuring the spike responses to 9 different members of yt -fatty acids (.x-axis, aligned according to the length of the hydrocarbon chain) over 5 different concentrations (y-axis). Each column indicates mean number of odor-elicited spikes. All spikes occurring within the 1st inhalation cycle ( 1.5 s in duration) of 3 repetitions of the stimulus were averaged.

members [ (6)-( lO)COOH] were ineffective. This indicates that, although the range of effective fatty acids was broader at the higher concentrations, the best response of the mitral cell at higher concentrations was similar to that determined at lower concentrations. The above results indicate that for a given mitral/ tufted cell, a molecular specificity graph (cf. Fig. 2 B) for y2-fatty acids can be determined by measuring the number of spikes elicited by the mitral/ tufted cell in response to a series of n-fatty acids with different hydrocarbon chain lengths over different concentrations. Using this method, molecular specificity graphs for n-fatty acids were obtained in 12 mitral/ tufted cells, and all these graphs showed characteristic response selectivity profiles (similar to that shown in Fig. 2B) with selective responses to subsets of the fatty acids with similar hydrocarbon chain lengths. To obtain a complete molecular specificity graph for the n-fatty acids, it is necessary to examine single unit responses to 45 different odor molecule samples (9 different fatty acids X 5 different concentrations) and at least two trials of each sample, which usually takes >45 min. Thus it was technically difficult to obtain the complete molecular specificity graphs for a large percentage of the recorded mitral/ tufted cells. To check rapidly the molecular specificity of individual mitral/ tufted cells, we used a series of YI-fatty acids with a concentration of 5 X 10v2, because the complete molecular specificity graphs obtained from the 12 mitral/ tufted cells indicated that the range of effective chain

length can be obtained with a series of n-fatty acid molecules at a single relatively high concentration. In the dorsomedial part of the MOB, we recorded 105 mitral/tufted cells that responded to at least one member of the normal fatty acids. Figure 3 summarizes the response selectivity of these 105 mitral/ tufted cells tested with a series of n-fatty acids at a concentration of 5 X 1Oe2. It is clearly seen that, except for one case, all neurons showed response selectivity to subsets of y2-fatty acids having hydrocarbon chains of similar length. These results confirm the previous report ( Mori et al. 1992) and further suggest that a homologous series of aliphatic molecules are good tools for showing examples of a rational relationship between response selectivity of individual mitral/ tufted cells and the structure of odor molecules. Responses to branched fatty acids and crotonic acid, a fatty acid with a double bond To examine the response specificity of individual mitral/ tufted cells in relation to structural variations in odor ligands, the panel of stimulus odor molecules (Table IA) contained three branched fatty acids [ (IB), (IV), and (MB)COOH] with relatively short hydrocarbon chains, three aromatic carboxylic acids [ (Bz), (PA), and (Ci)COOH] with a benzene ring, and crotonic acid, a fatty acid with a double bond in the chain. For 154 MCL units, we checked spike responses for all

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

1992

K. IMAMURA, C6

C7

.

CS

c9

N. MATAGA,

Cl0

I(*) 3. Response specificities of individual mitral/ tufted cells to normal fatty acids. Each bar represents the range of fatty acid odors by which the mitral/tufted cell(s) are activated. C2, C3, and C 10 indicate acetic acid, propionic acid, . . ., and n -capric acid, respectively ( see Table 1) . Numbers at each bar indicate the number of cells showing the same response specificity. For example, the bar shown in the 2nd row from the top indicates that we recorded 5 neurons that were activated by (2 ) , ( 3 ) , and (4)COOH but were not activated by other members of the n-fatty acids [ (5)-( 1O)COOH] . Note that except 1 cell shown at bottom (* ), all mitral/tufted cells are activated by subsets of fatty acids with equivalent ranges of hydrocarbon chain lengths. FIG.

l

l

l

the carboxylic acid odor molecules (at the concentration of 5 X 10-2) listed in Table 1A. Figure 4 showsthree examples representing typical types of responsespecificity of individual mitral/ tufted cells in the fatty acid-sensitive region of the MOB. The first type is characteristic in that it was activated not only by a subset of n-fatty acids but also by varying numbers of the branched fatty acids. For example, a mitral cell shown in A was activated by [ (4) and (5)COOH] and [(IB), (IV), and (MB)COOH]. The mitral cell shown in Fig. 1 also exhibited this type of response snecificitv. The second tvbe of neuron. as exemblified bv

AND

K. MORI

the cell in B, showed spike responsesselectively to a subset of n-fatty acids. They were not activated by any member of the branched fatty acids nor by any member of the aromatic carboxylic acids. Mitral cells of the third type (C) responded with spike dischargesto all or a subsetof the three aromatic carboxylic acids in addition to a subset of the normal fatty acids. Because mitral/ tufted cells of the first type showed tuning to both normal (unbranched) fatty acids and branched fatty acids, we have examined the structural relationship among the fatty acid molecules effective in activating individual mitral/ tufted cells. Figure 5 shows the response selectivities of individual mitral/tufted cells ( numbered from 1 to 25) that were activated by at least one member of the branched fatty acids. Figure 5, fefi, shows responsesto the branched fatty acids [ (IB), (IV) ,&and( MB)COOH] . Sixtyfive percent ( 16 out of 25 ) of the mitral/ tufted cells responded to two or all of the three branched fatty acids. The response specificity of the 25 iso-fatty acid sensitive neurons for the normal fatty acids [ (2)-( lO)COOH] is shown at right. It is clearly seen that the majority (92%) of the iso-fatty acid-sensitive cells were activated also by at least one member of the normal fatty acids. In addition, the isofatty acid-sensitive neurons tended to be activated by subsetsof n-fatty acids with overall hydrocarbon chain lengths similar to those of (IB), (IV), and (MB)COOH. Among the 23 neurons that were activated by both the iso-fatty acids and the normal fatty acids, 19 ( 82%) cells were activated by [ (4) and/ or ( 5 )COOH] (Fig. 5 ). Except for one cell ( cell 23 in Fig. 5 ) , M-fatty acids with long hydrocarbon chain lengths [ ( 8)-( lO)COOH] did not elicit spike responsesin the iso-fatty acid-sensitive mitral/tufted cells. These results suggestthat the determinants for the response specificity of individual mitral cells in this region include the overall size of the hydrocarbon chains of the fatty acid odor molecules. Crotonic acid [ (Cr)COOH] has a molecular structure that is quite similar to n-butyric acid [ (4) COOH] except that it contains a double bond within the hydrocarbon chain (Fig. 6). We have recorded seven neurons that were activated by crotonic acid. The response selectivities of these neurons for the n-fatty acids are shown in Fig. 6, right. As expected from the structural similarity, six of the seven mitral/ tufted cells were activated by subsets of y2fatty acids having hydrocarbon chain length similar to crotonic acid or (4)COOH. Responseto normal alipkatic compounds During the experiments using carboxylic acid odor molecules (Table 1)) we noticed that neurons responsive to benzoic acid [ (Bz) COOH] tended to be activated also by benzaldehyde [ (Bz)CHO] , which was used as a reference odor and sharesthe same hydrocarbon skeleton, a benzene ring, with benzoic acid. This suggeststhat single neurons in the dorsomedial region may respond not only to a range of carboxylic acids but also to a range of aldehydes. In addition, it seemedpossible that single neurons in the dorsomedial region might be activated also by certain members of a seriesof normal alcohols or a seriesof alkanes if the shanes

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR Number

A 0

CODING

B

of impulses/inhalation 10 20

Number 0

IN

OLFACTORY

BULB

1993

C

of impulses/inhalation 10 20

(2)COOH

(2)COOH

(2)COOH

(3)COOH

(3)COOH

(4)COOH

(3)COOH (4)COOH

(EWOOH

(6)COOH

(6)COOH

(6)CooH

(6)CooH

(6)cooH

tWOOH (8)COOH

(7)COOI-I (8)CO0H

(7)COOH

(9)COOH (1O)COOH

(9)COOH

(9)COOH (10)COOH

(S)CCKIH

.

.

(MBXOOH (BdCOOH (PA)COOH

of impulses/inhalation 10 20

(4)CoOH

(10)COOH

t-IB>CoOH (IVJCOQH

Number 0

.

(IB)COOH (IVXWOH

(IBXOOH (rv>COOH

CMBNXOH GIz)CoOH

(MB)COOH (Bz)COOH

(PANXKIH

(PA)COOH

(Ci)CoOH

(Ci)COOH

(Ci)CoOH

(Cr)CCKIH

(Cr)COOH

(CWCOOH

FIG. 4. Three representative examples of response specificity of individual mitral cells (A-C) to normal fatty acids [(2)-( lO)COOH], iso-fatty acids [(IB), (IV), and (MB)COOH], aromatic carboxylic acids [(Bz), (PA), and (Ci)COOH] , and crotonic acid [ (Cr)COOH] . Mean number of odor-elicited spikes per 1 inhalation cycle ( 1.5-s period) was counted for each neuron and indicated by the length of the bar. All the units were identified as mitral cells by their antidromic spike responses to the lateral olfactory tract (LOT) volley. The mitral cells showed almost no spontaneous activity.

or lengths of the hydrocarbon chains were similar to those of the effective carboxylic acids. To examine these possibilities, we expanded th .e panel of stimulatin .g odor molecules to include a series of normal aliphatic aldehydes [ ( 3)x?

$3

$4

$5

$6

$7

.Ct3

C9

Cl0

( 11 )CHO] , a series of normal aliphatic alcohols [ ( 3)( lO)OH], and a series of alkanes [( 5)-( lO)CH] (Table 2). All members contain in common normal aliphatic hydrocarbon chains. Figure 7 shows an example of a mitral cell in the dorsomedial region that responded to n-aliphatic aldehydes as well as n-aliphatic acids. The mitral cell showed spontaneous discharges only occasionally. During single-unit recording from this cell, we first examined the response specificity of the mitral cell using the panel of carboxylic acid odors (Table 1) and found that [( 3)COOH] and [(4)COOH] activated the mitral cell, but other members of the carboxylic acid odors were ineffective. The response specificity of the mitral cell was further examined using the normal aliphatic odor molecules listed in Table 2. As shown in Fig. 7, not only the (3)COOH and (4)COOH but also (3)CH0 and (4)CH0 elicited responses. However, the mitral cell C5

C6

1 2 3 4 5 6 7

5. Response specificities of individual iso-fatty acid-sensitive mltral/ tufted cells to normal fatty acids. The mitral/ tufted cells were numbered from 1 to 25 and listed in rows. Filled bars at each row indicate iso-fatty acids (left) and n-fatty acids ( right) by which the mitral/tufted cell was activated. The 2 cells at the bottom (cells 24 and 25) are exceptional in that they were not activated by any member of the n-fatty acids. FIG.

C rotonic

acid

WcoOH

n-Butyrlc

acid

-COOH

.C7

. C8

.C9

. Cl0

FIG. 6. Response specificities of individual crotonic acid-sensitive mitral/ tufted cells to normal fatty acids. Crotonic acid-sensitive mitral / tufted cells were numbered from 1 to 7 and listed in rows. Filled bars at each row indicate crotonic acids (Cr, left) and n-fatty acids (right) by which the mitral/ tufted cell was activated. The molecular formula of the crotonic acid (Cr) and the n-butyric acid (C4) are shown at bottom.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

IS. IMAMURA,

N. MATAGA,

Figure 8 shows the response specificity of a mitral cell for the homologous series of n-aliphatic aldehydes [ ( 3) ( 11 )CHO] over the five different concentrations. At relatively low concentrations ( 1Om3, 3 X 10 -3, and 10 -2), this cell was activated selectively by ( 9)CH0 and ( 1O)CHO. With increasing the concentration to 3 X 10 -2, not only the ( 9) CHO and ( 10) CHO but also ( 8) CHO activated the cell. With the highest concentration ( 10-l ) , [ ( 7)-( 11 )CHO] were effective in activating the mitral cell; ( 1O)CHO was most effective, followed by (9)CHO, ( 8) CHO, ( 11 )CHO, and ( 7 ) CHO. Molecular specificity graphs for the y1-aliphatic aldehydes were obtained for 14 mitral/ tufted cells using the five different concentrations. All these graphs showed the characteristic response selectivity profiles (as in Fig. 8) with specific responses to subsets of the n-aliphatic aldehydes, similar to the response selectivities for the n-fatty acids (cf. Fig. 2 B) . In addition, we determined in 52 mitral/ tufted cells the response selectivity for the homologous series of n-aliphatic aldehydes at a concentration of 5 X 10m2. As shown

t 1o-2

AND K. MORI

.d-

lsec (3

)CHO

(4

)CHO

FIG. 7. Concentration-response relationship of a mitral cell to the stimulation with a homologous series of n-aliphatic aldehydes. Responses to increasing concentrations of propylaldehyde [ ( 3 ) CHO ] and n -butylaldehyde [ (4) CHO] are shown in A and B, respectively. Monitors for artificial inhalations are shown under spike activity traces. Thick horizontal bars: period of odor application. Numbers at Zeft: concentration ( vol/ vol) of the n -aldehyde odors used. Arrows: onset of the inhalation of odor-containing air. Spike height is -2 mV.

was not activated by other members of the n-aliphatic aldehydes. In addition, no member of the n-alcohol series and n-alkane series in Table 2 was effective in generating an excitatory response.

C3 . C4 . C5 . C6 . C7 . a3 . c9 . Cl0 . Cl1 I2 -1 12

I

5

1

I

II”

-1 I I

FIG. 8. Three-dimensional diagram showing concentration-response relationship of spike responses of a mitral/ tufted cell to 7 different members of a homologous series of n-aliphatic aldehydes over 5 different concentrations. This cell showed no spike responses to (3)CH0 or (4)CH0 at any concentration examined. The procedure for data acquisition was similar to that used in Fig. 2 B.

3

I

FIG. 9. Response specificities of individual mitral/ tufted cells to normal aliphatic aldehydes. Each open bar represents the range of n-aliphatic aldehyde odors by which the mitral/ tufted cell(s) is/are activated. Numbers shown near each bar indicate the number of cells showing the same response specificity. For example, the bar shown at fopindicates that we recorded 2 neurons that were activated only by (3)CH0. Note that all mitral/ tufted cells described here were activated by subsets of aliphatic aldehydes with equivalent hydrocarbon chain lengths.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR

CODING

IN OLFACTORY

1995

BULB

rons were compared between n-fatty acids and n-aliphatic aldehydes, and 24 of them were shown in Fig. 10. Excluding the exceptions of units 16 and 24, the tuning profiles were similar between the n-fatty acids and aldehydes series. In most units, the tuning range in the aldehyde series overlapped at least partially with that in the n-fatty acid series in terms of hydrocarbon chain length. Precise correspondences were found in four units ( Ul, U4, U9, and UlO) (Fig. 10). To examine whether the fatty acid-sensitive neurons respond to n-alcohols and n-alkanes, single-unit responses to these odor molecules were examined in 44 MCL cells that were activated by at least one member of the n-fatty acids. Except for two cells ( 5%), none of the mitral/ tufted cells showed facilitatory response to any member of the n-alcohols. In addition, all members of the n-alkanes listed in Table 2 were ineffective in activating any of the 44 MCL cells. These results suggest that the carbonyl group of the odor molecules plays a key role in determining the tuning specificities of neurons in the fatty acid-sensitive region. Because of the low rate of spontaneous spike activities of the mitral/ tufted cells in our experimental conditions, it was difficult to determine whether the unresponsiveness of ine

/

/

OB

“3 O \ -2

-3 -4

10. Comparison between y1-fatty acid series and y1-aldehyde series in the range of members effective in activating individual neurons. Thirtytwo mitral cell layer ( MCL) neurons were found to respond to both n-fatty acids and aldehydes, and 24 representative units were shown here. The range of examined n-fatty acids was [(2)-( lO)COOH], whereas that of examined n -aldehydes was [ ( 3)-( 11) CHO] . C2-C 11 indicate the length of hydrocarbon chain (number of carbon atoms) of the odor molecules. Each horizontal bar (filled: fatty acids; open: aldehydes) indicates the range of compounds to which individual mitral/tufted cells showed facilitatory spike response. The neurons indicated by asterisks were identified as mitral cells by their antidromic responses to the lateral olfactory tract (LOT) volley. FIG.

in Fig. 9, all cells showed tuning to one to four n-aliphatic aldehydes having similar hydrocarbon chain lengths. Both within and outside of the fatty acid-sensitive region in the dorsomedial part of the olfactory bulb, we examined single-unit responses to all members of n-fatty acids and n -aldehydes on 100 MCL units. Fifty-six units showed facilitatory response to at least one of these odor molecules. Out of the 56, 17 units ( 30%) responded selectively to n-aliphatic aldehydes, whereas 7 units ( 13%) responded exclusively to n-aliphatic acids. The remaining 32 units ( 57%) responded to both n-aliphatic acids and aldehydes. In these 32 units, tuning specificities of individual neu-

-5 -5

1

7mm

Lateral FIG. 1 1. Localization of the carboxylic acid-sensitive mitral/tufted cells in the main olfactory bulb (MOB). A dorsal view of the animal’s left MOB, accessory olfactory bulb ( AOB), and frontal pole (FP) of the neocortex. Each filled circle in the MOB indicates the position of the mitral/ tufted cell, which was activated by 2 1 member of the carboxylic acid odor molecules listed in the Table 1A. Open circles: positions of mitral/ tufted cells that were not activated by any member of the carboxylic acid odors. Arrow with a letter (Midline): midline of the brain. Broken line: extent of the mitral cell layer in the MOB. Dotted line: extent of the AOB. Arrows with broken lines and letters “a” and “b”: frontal planes used for reference for combining data from different animals; “a” indicates the most rostra1 plane containing the mitral cell layer, and b indicates the most rostra1 plane containing the AOB.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

1996

K. IMAMURA,

these cells to alcohols and alkanes reflects inhibition of any response.

N. MATAGA,

or lack

Spatial localization of neurons responsive to carboxylic acid odor molecules In seven animals, the positions of the recorded mitral/ tufted cells were located in the serial frontal sections of the olfactory bulbs. Using the section through the most rostra1 part of the mitral cell layer (indicated by an arrow with “a” in Fig. 11) and that through the most rostra1 part of the accessory olfactory bulb (AOB, arrow with “b”) as two reference planes for alignment in the anterior-posterior axis, data on the positions were collected from the seven experiments and reconstructed in a diagram shown in Fig. 11. This diagram shows a dorsal view of the MOB. Each filled circle indicates the position of an individual mitral/ tufted cell showing an excitatory response to at least 1 of the 16 members of the carboxylic acids (Table 1) . Open circles indicate neurons that showed no response to any member of the carboxylic acid odors. The carboxylic acid-sensitive neurons tested here were located mostly in the dorsal part of the MOB. In addition, they tended to be distributed in a region in the dorsomedial part, as shown in Fig. 11. Figure 12 illustrates an example of the spatial distribution of the carboxylic acid-sensitive neurons in a coronal section. The data were obtained from one animal. Neurons showing excitatory responses to > 1 of the 16 carboxylic acids are shown by filled circles and filled squares. It is clearly seen that mitral/ tufted cells sensitive to carboxylic acids are grouped together in the dorsomedial part of the olfactory bulb (Fig. 124. Mitral/tufted cells recorded in the dorsolateral and ventral parts at this coronal level did

AND

K. MORI

not show excitatory responses to any of the 16 carboxylic acid molecules. Figure 12 B illustrates the relationship between positions and tuning specificities of individual mitral/ tufted cells in the dorsomedial region. Neurons U3, U4, u6, and U7 responded selectively to n-fatty acids. In neurons that responded to both n-aliphatic acids and aldehydes, there were overlaps in the tuning ranges. Neurons with different tuning specificities were found to be intermingled within the dorsomedial region. In the ventral part of the MOB, we recorded 53 mitral/ tufted cells and examined the response selectivities for nfatty acids and n -aliphatic aldehydes (Table 2). Excepting one mitral cell, no mitral/ tufted cells in the ventral part responded to any member of the n-fatty acids. In contrast to this, we found that a number of neurons in the ventral region showed facilitator-y responses to n -aliphatic aldehydes. Although the extent of sampling from regions other than the dorsomedial region is relatively limited, the above results suggest that mitral/ tufted cells responsive to the nfatty acids are mainly localized in the dorsomedial region, whereas n -aliphatic aldehyde-sensitive cells are distributed in both the dorsomedial region and the ventral region. This is in good agreement with the results obtained with the mapping of the odor-elicited OLFPs (Mori et al. 1992). Further characterization of tuning specificity using ketones and esters The results presented in previous sections indicate that I the carbonyl group ( - C = 0) of the odor m olecu les plays

B

’

”

)GrLMCLjEPLIGL

ONL

i

50Opm \

,/

IMMedi

FIG. 12. Spatial distribution of neurons showing spike responses to carboxylic acids and M-aldehydes. A : recording tracks in a coronal section through the MOB. Asterisks: positions of blue dye spots made for marking the e - -> were reconstructed electrode positions. Recorded neurons were numbered from 1 to 26. Neurons that showed facilitatory responses to 2 1 of the 16 carboxylic odor molecules listed in Table 1A are shown by filled symbols. Open symbols: neurons showing no spike response to the 16 carboxylic odor molecules. Squares: mitral cells identified by their antidromic spike response to lateral olfactory tract (LOT) volley. Circles: neurons that did not show antidromic spike response. B: dorsomedial region of A is enlarged. Odor molecules that elicited facilitator-y spike responses of the recorded neuron are shown in brackets. Abbreviations of compounds are listed in Table 1 and 2. ONL, olfactory nerve layer; GL, glomerular layer; EPL, external plexiform layer; MCL, mitral cell layer; GrL, granule cell layer; V, ventricle.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR

CODING

IN

OLFACTORY

1997

BULB

joz:o ]o=!oJ o=!. [oxjo ] or~o]w~o ]08jo ]ojo

Esters

El

E2

E3

E4

E5

E6

E7

E8

C2

C3

C4

C5

C6

C7

C8

C9

n-Alkanes

n-Aliphatic

aldehydes

n-Aliphatic

acids

0x0

d

i

I

08

i

Cl0

Cl1

I

I

Ketones FIG. 13. Characterization of response specificity of a mitral cell using odor molecules of n-primary aliphatic compounds (Table 2) and nonprimary compounds (esters and ketones). Filled bars: mean number of spikes per l-inhalation cycle elicited by stimulation with respective odors. No bar: no facilitatory response. One digit of the scale indicates 10 impulses. Molecular formula of compounds are shown on the right side of each set of coordinates. C2-Cl1 indicate the number of carbon atoms in the side chain of the primary compounds. Esters and ketones were arranged according to the length of 1 side chain and tentatively indicated by E 1-E8 and Kl-K9, respectively.

a key role in activating neurons in the dorsomedial part of the bulb. The n-fatty acids and n-aliphatic aldehydes are primary compounds having the carbonyl group at the end of the hydrocarbon chain. To examine further the tuning specificity of individual mitral/ tufted cells in this region,

we expanded the panel of stimulus odor molecules to inelude nonprimary compounds, ketones, and esters that contain the carbonyl group at various positions along the hydrocarbon chain. An example of these results is shown in Fig. 13. When the

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

1998

K. IMAMURA,

N. MATAGA,

AND K. MORI

Esters El

E2

C2

C3

E3

E4

E5

E6

E7

C7

C8

E8

n-Alkanes

n-Aliphatic

acids C4

C5

C6

C9

Cl0

Cl1

K1 K2 K3 K4 K5 K6 K7 K8 K9 14. Another example of response specificity of a mitral cell to primary n -aliphatic compounds, esters, and ketones. The pattern of response is shown in a fashion similar to that in Fig. 8, except that 1 digit of the scale in this figure indicates 3 spikes. FIG.

primary aliphatic compounds were examined, this mitral cell showed excitatory responses to [ ( 3) and (4)COOH] and [ (4) and ( 5)CHO], with the largest response at [ (4)CHO]. This neuron is one of the exceptional cases in that it responded also to [ (4)OH], because in the dorsomedial region we encountered only two neurons that showed excitatory responses to the n-alcohols. It can be seen in Fig.

13 that this mitral cell was activated by subsets of ketones and esters. The subset of ketones effective in activating this mitral cell were ethyl-propyl, methyl-butyl, ethyl-butyl, propyl-butyl, and dibutyl ketones. Other members of the ketone series did not elicit spike responses in this mitral cell. The subset of the effective ketones shares [o = c-c-c-c-c] or [o = c-c-c-c] structures. The subset of esters effective in ac-

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR

CODING

tivating this mitral cell are propionic acid methyl ester, butyric acid methyl ester, propionic acid ethyl ester, and butyric acid ethyl ester. The subset of esters shares [o = c-c-cc] or [o = c-c-c] structures. Another example of the response selectivity of mitral cells examined with the expanded panel of odor molecules is shown in Fig. 14. This cell showed excitatory responses selectively to [ (6) and (7)COOH], [ (6) and (7)CHO], and also to a subset of esters, pentanoic acid methyl ester, hexanoic acid methyl ester, and hexanoic acid butyl ester. Other compounds, including all the ketones shown in this figure, were ineffective in eliciting excitatory responses in this cell. It can be noted in these two examples that all the effective molecules share similar hydrocarbon chain structure and a carbonyl group connected to the hydrocarbon chain. We examined 15 mitral/ tufted cells sensitive to n aliphatic compounds for all members of the ketones and esters. Twelve ( 80%) of them were activated by subsets of the ketones and/or esters. In all these cases, the effective ketones and/ or esters contained hydrocarbon chains similar to those present in the effective n-fatty acids and aldehydes. These results indicate that not only the primary odor molecules (n-fatty acids and n-aliphatic aldehydes) but also nonprimary odor molecules (ketones and esters) are effective in activating mitral/ tufted cells in the dorsomedial region. The above results indicate also that the length of the hydrocarbon chain connecting the carbonyl group in esters and ketones plays a critical role for activation of individual mitral/tufted cells. DISCUSSION

A number of studies have reported that single mitral/ tufted cells in the mammalian olfactory bulb respond differently to different odor molecules (Buonviso and Chaput 1990; Harrison and Scott 1986; Mair 1982; Mathews 1972; Meredith 1986; Tanabe et al. 1975; Wellis et al. 1989; Wilson and Leon 1988 ) . In general, these studies used only a small number of different odor molecules for stimulation and therefore lacked information on the characteristics of the response specificity in terms of the stereochemical structures of odor molecules. In frog olfactory bulb, Daving ( 1966) reported that hydrocarbon chain length and functional groups of odor molecules were of importance in determining response patterns of mitral cells. This study provides a first systematic physiological report for correlations between the response specificity of single mitral/tufted cells, the structural parameters of the stimulating odor molecules, and the spatial localization of the responses in MOB. Tuning specificity of individual

mitral/tufted

cells

The present study has demonstrated that by using homologous series of aliphatic compounds as stimulus odor molecules, it is possible to determine, at least partially, the odor tuning specificities of individual mitral/ tufted cells in the dorsomedial region of the rabbit olfactory bulb. In addition, the results show a clear correlation between the tuning specificity of mitral/tufted cells and the stereochemical structure of the odor molecules with respect to three vari-

IN

OLFACTORY

BULB

1999

ables in the structure: 1) length and/or structure of hydrocarbon chain, 2) difference in functional group (- COOH, -CHO, -OH, and no group), and 3) position of the functional group within the molecule. First, individual mitral/ tufted cells in the region were activated by subsets of n-fatty acids and/ or n -aliphatic aldehydes having similar hydrocarbon chain lengths. The property of the best response to specific chain length did not change despite the change in molecular concentration, although the range of the effective aliphatic compounds increased with increasing concentrations (Figs. 2 and 8). It was also found that cells that were activated by branched fatty acids [ (IB), (IV), and (NB)COOH] tended to be activated by subsets of n-fatty acids having similar hydrocarbon chain lengths (Fig. 5 ). Furthermore, most of the crotonic acid-sensitive mitral/ tufted cells showed excitatory responses to [ (4)COOH], which has a hydrocarbon chain quite similar to crotonic acid. These results indicate that to activate individual mitral/ tufted cells sensitive to the aliphatic compounds in the region, odor molecules should contain hydrocarbon chains having an appropriate length or structure. Second, in the dorsomedial region, the majority of those cells responsive to n-aliphatic acids were also responsive to n -aliphatic aldehydes, whereas n -aliphatic alcohols and alkanes were mostly ineffective in activating these cells. This suggests that in addition to the structure of the hydrocarbon chain, the carbonyl group in odor molecules plays an important role for activation of these mitral/ tufted cells. This is consistent with our previous results obtained using randomly selected odor molecules ( cf. Fig. 3A of Mori et al. 1990), that individual neurons at the dorsomedial region tended to be activated by both caproic acid

c-c-c-c-c-c’

/

\

0

OH

and t-2-hexenal c-c-c---c--c-c’

/

\

0

H.

Finally, we found that ketones and esters that have the carbonyl group in the middle of the molecule are effective in activating the aliphatic compound-sensitive mitral/ tufted cells in the region, provided that they have a specific carbohydrate chain substructure. Again, individual mitral/ tufted cells were activated by subsets of ketones and/ or esters having similar hydrocarbon chain structures. The structures of the odor molecules effective in activating mitral/ tufted cells (e.g., Figs. 13 and 14) share common stereochemical features. Taken together, these results strongly indicate that mitral/ tufted cells in the olfactory bulb display a stereochemical structure-dependent response to odor molecules. The importance of the stereochemical structure of odor molecules in determining the tuning specificity of individual mitral/ tufted cells is found also in other regions of the

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

2000

K. IMAMURA,

N. MATAGA,

olfactory bulb and for other types of odor ligand molecules. For example, we found that some mitral cells in the medial region responded with spike discharges to para-xylene but were not activated by its structural isomers, ortho-xylene and metha-xylene (Mori, Mataga, and Imamura, unpublished observations). Although variables in molecular structure are multiple and relatively complex, the present results indicate that mitral/ tufted cells may be sensitive to specific ranges of odor molecules (receptive molecular ranges) that can be determined experimentally. In analogy with the term “epitope” in immune system, Shepherd ( 1987) proposed the term “odotope” for expressing the idea of the common substructure of odor molecules responsible for activating a subtype of odor receptor protein. This term is equivalent to “olfactophore,” which was suggested by Ham and Jurs ( 1985) . The present finding that the odor molecules effective for activating individual mitral/tufted cells show common stereochemical substructures suggests that the common substructure reflects the odotope and that the tuning specificity may be determined mostly by the molecular interaction between these odor ligand molecules and one or a few types of olfactory receptor proteins that recognize specific odotopes. In the present study, we obtained tuning graphs for individual mitral/ tufted cells by measuring the number of impulseselicited in responseto homologous seriesof aliphatic odor molecules with different hydrocarbon chain lengths and with different concentrations (Figs. 2 and 8). These tuning graphs showed that 1) among the homologous series of n-fatty acids or n-aliphatic aldehydes, it is possible to determine for individual cells one characteristic member (or 2 members with minimal difference in the chain length) that activates the cell at the lowest concentration; and 2) with increasing concentration, other members having similar chain lengths to the characteristic member become effective in activating the cell. These characteristics of the tuning graph are features explainable by the interaction between odor ligand molecules and one or a few type(s) of receptor proteins, as will be discussedin the next section. The tuning graphs of mitral/ tufted cells are comparable to the tuning curves against sound frequency reported for auditory nerve fibers (Goldstein 1980) and the tuning curves of cones in the retina against wave length (Tomita et al. 1967). In the auditory system, the sound frequency most effective in activating the recorded neuron is called its characteristic frequency. In analogy with this, for individual aliphatic compound-sensitive mitral/tufted cells in the dorsomedial region, it is possible to determine a characteristic (e.g., optimal) hydrocarbon chain length or structure. Different mitral/ tufted cells typically showed tuning to different characteristic hydrocarbon chain structures. This suggeststhat there are different groups of mitral/tufted cells in the dorsomedial region, each group carrying information about odor molecules having a specific hydrocarbon chain length or structure.

Functional convergence ofoljbctory c

nerve input to glomeruli

What is the mechanism for selective activation of individual mitral/ tufted cells by a range of odor molecules having

AND

K. MORI

similar stereochemical structures? Because an individual mitral/ tufted cell projects to a single primary dendrite within a single glomerulus in the rabbit MOB, the tuning specificity of the cell may be determined mostly by the combination of the tuning specificities of all olfactory receptor neurons that project their axons to the glomerulus and make synaptic connections with the primary dendrite of the cell. In the rabbit, a single glomerulus receives converging axonal inputs from w 26,000 receptor neurons (Allison and Warwick 1953). Figure 15 shows a hypothetical diagram that we propose for explaining the mechanism for selective activation of individual mitral/ tufted cells by a range of odor molecules having similar stereochemical structures. This diagram contains three major hypotheses: 1) a single receptor cell is assumedto express a single or a small number of different kind(s) of odor receptor proteins; 2) a single receptor protein can be activated by a range of odor molecules having similar stereochemical structures; and 3) axons of receptor neurons that express the same or similar receptor proteins are assumed to project to a functionally specific glomerulus. These are extensions of hypotheses proposed previously by several authors (Boeckh et al. 1990; Kauer 1987, 1991; Lancet 1986; Rhein and Cagan 1980; Shepherd 1991; Shepherd and Firestein 199 1). According to this hypothetical diagram, the tuning selectivity of an individual mitral/ tufted cell (the shaded mitral cell, for example, in Fig. 15) reflects the tuning specificities of the receptor neurons (shaded) that project their axons to that cell. Because the receptor cells express one or a few common odor receptor protein(s) ( shadedreceptor protein in Fig. 15), the tuning specificities of these receptor cells are

FIG. 1% Hypothetical scheme to explain the response selectivities of individual mitral/tufted cells. In this model, individual glomeruli receive converging axonal inputs from a number of olfactory receptor neurons expressing the same or similar receptor protein subtype(s). For simplicity, this scheme shows only 2 subsets of receptor neurons (R, shaded and blank), each subset expressing a specific subtype of receptor protein on the cilia1 membrane. Glom, glomerulus; M, mitral cell.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

MOLECULAR

CODING

IN OLFACTORY

200 1

BULB

the MOB that were activated by subsets of n-aliphatic aldeassumed to reflect the tuning specificities of the receptor hydes and n-aliphatic alcohols. In contrast to the cells in the protein ( s) . Thus the tuning specificity of the mitral/ tufted dorsomedial region, the majority of these cells were not cell may reflect mostly the tuning specificity of the receptor activated by any member of the n-fatty acid odor moleprotein(s) that are commonly expressed by these receptor cules, so far as we have observed (Mori, Mataga, and Imacells. At present, there is no direct evidence to prove or dis- mura unpublished observations). These observations indiprove any of the above three hypotheses. However, several cate that 1) n-aliphatic aldehydes activate neurons located in at least two discrete regions in the olfactory bulb, the studies have reported suggestive evidence supporting them. For hypothesis 1, Buck and Axe1 ( 199 1) estimated that dorsomedial region and the ventral region; and 2) although each receptor neuron contains only 25-30 transcripts de- n-aliphatic aldehydes activate neurons in both regions, the rived from the putative odor receptor gene family, suggest- receptive molecular range differs between neurons in the ing that individual receptor cells express one type or only a dorsomedial region and ventral region, because only the small number of different types of receptor proteins. For dorsomedial cells are activated by n-fatty acids. These observations are consistent with the results rehypothesis 2, it has been shown in a number of the G proported by 2-deoxyglucose studies that stimulation of the tein-coupled receptor proteins, such as adrenergic receptor proteins and dopamine receptor proteins, that a single sub- olfactory epithelium with a single type of odor molecules type of the receptor protein can be activated by a number of causes high 2-deoxyglucose uptake in glomeruli located at different ligand molecules (agonists) having similar stereo- one to several discrete regions in the olfactory bulb (Bell et al. 1987; Jourdan et al. 1980; Stewart et al. 1979; cf. also chemical substructures ( O’Dowd et al. 1989). For hypothesis 3, recent histochemical studies have demonstrated that a Meredith 1986; Wilson and Leon 1988 for electrophysiolognumber of subsets of olfactory receptor axons identified by ical mappings). The difference in the receptive molecular different molecular markers show a tendency to project to range between the dorsomedial region and the ventral redistinct glomeruli (Akeson 1988; Mori et al. 1985, 1990; gion suggests that these two regions receive inputs from Schwab and Gottlieb 1986 ) . In addition, physiological stud- different types of receptor proteins, although both types can be activated by n -aliphatic aldehydes. ies using simultaneous recordings from different mitral The present study provides a first step toward charactercells have suggested that mitral cells projecting to the same izing individual output neurons in the olfactory bulb by glomerulus share similar response properties to odor stimulation (Buonviso and Chaput 1990). Moreover, studies us- systematically determining their response selectivities for odor molecules. To obtain information on signal processing 2-deoxyglucose have demonstrated that increased metabolic activity induced by stimulation with a given type of ing in the olfactory bulb, it will also be necessary to examine cells and odor molecule occurs in glomeruli in unitary fashion, sug- the roles of local interneurons (periglomerular granule cells) in contributing to the response selectivity (cf. gesting glomerular convergence of receptor axons having similar tuning specificities (Lancet et al. 1982; Stewart et al. Wellis and Scott 1990) and the temporal firing patterns of mitral/ tufted cells, and to examine the interaction via local 1979). cells with different reThe functional convergence of projecting axons pro- interneurons among mitral/tufted posed in this hypothetical scheme can be found in several sponse selectivities. In addition, a comparison of the charother systems. For example, in the somatosensory system, acteristics of receptive molecular ranges between olfactory primary afferent fibers from different receptor cell types are bulb neurons and olfactory cortical neurons is an intriguing mixed with each other in the aRerent bundles but are sorted subject for future research, because this type of approach out in the spinal cord to terminate in different laminae in a has provided important clues in understanding information processing in visual, auditory, and somatosensory syssubmodality-specific manner (Brown 198 1; Mountcastle tems. 1984). In the above discussion, it was assumed that the excitatory responses exhibited by mitral/ tufted cells were proWe thank Drs. Y. Yoshihara and K. Katoh for reading the manuscript. This work was supported by grants from the Ministry of Education, duced mostly by inputs from olfactory nerve terminals. Science, and Culture, and from the Science and Technology Agency of However, it is possible that in addition to the olfactory Japan, and from Brain Science Foundation. nerve input, excitatory synaptic interactions within the gloAddress for reprint requests: K. Mori, Dept. of Neuroscience, Osaka merular layer and rebound excitation after hyperpolarizaBioscience Institute, 6-2-4 Furuedai, Suita, Osaka 565, Japan. tion contribute in the excitatory responses. Received

Spatial representation

27 April

1992; accepted

in final

form

9 July

1992.

in the olfactory bulb

The present results, together with the previous study (Mori et al. 1992), suggest that a large group of glomeruli located in the dorsomedial region receive inputs from receptor neurons expressing structurally related odor receptor proteins. In addition, present results have shown that the dorsomedial mitral/tufted cells tend to be activated also by n-aliphatic aldehydes. It should be noted that we observed a number of mitral/tufted cells in the ventral part of

REFERENCES ADRIAN,

E. D. The electrical

activity

of mammalian

olfactory

bulb.

Elec-

troencephalogr. Clin. Neurophtivsiol. 2: 377-388, 1950. ADRIAN,

E. D. The action

of the mammalian

olfactory

organ.

J.

L-aryngol.

Otol. 70: l-l 5, 1956. L. M., IRVINE, D. R. F., AND WEBSTER, W. R. Central neural mechanism of hearing. In: Handbook qf‘Physiology. The Nervous Systern. Sensory Processes. Bethesda, MD: Am. Physiol. SOC., 1984, sect. 1, vol. III, p. 675-737.

AITKIN,

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.

2002

K. IMAMURA,

N. MATAGA,

subclasses. In: Molecular Neuroby F. L. Margolis and T. V. Getchell. New York: Plenum, 1988, p. 297-3 18. ALLISON, A. C. AND WARWICK, R. T. T. Quantitative observations on the olfactory system of the rabbit. Brain 72: 186- 197, 1949. BELL, G. A., LAING, D. G., AND PANHUBER, H. Odor mixture suppression: evidence for a peripheral mechanism in human and rat. Brain Res. 426: 8-18, 1987. BOECKH, J., DISTLER, P., ERNST, IS. D., H&L, M., AND MALUN, D. Olfactory bulb and antenna1 lobe. In: Chemosensory Information Processing, edited by D. Schild. Berlin: Springer-Verlag, 1990, p. 20 l-227. BROWN, A. G. Organization in the Spinal Cord. Berlin: Springer-Verlag, 1981, p. 136-153. BUCK, L. AND AXEL, R. A novel multigene family may encode odorant receptors: a molecular basis for odor recognition. Cell 65: 175-l 87, 1991. BUONVISO, N. AND CHAPUT, M. A. Response similarity to odors in olfactory bulb output cells presumed to be connected to the same glomerulus: electrophysiological study using simultaneous single unit recordings. J. Neurophysiol. 63: 447-454, 1990. CLARK, W. D. LE GROS. Inquiries into the anatomical basis of olfactory discrimination. Proc. R. Sot. Lond. Ser. B. Biol. Sci. 146: 299-3 19, 1957. DC~VING, IS. An electrophysiological study of odour similarities of homologous substances. J. Physiol. Lond. 186: 97- 109, 1966. FREEMAN, W. J. AND SKARDA, C. A. Spatial EEG patterns, non-linear dynamics and perception: the neo-Sherringtonian view. Brain Res. Rev. 10: 147-175, 1985. GOLDSTEIN, M. H., JR. The auditory periphery. In Medical Physiology ( 14th ed.), edited by V. B. Mountcastle. St. Louis, MO: Mosby, 1980, vol. 1, p. 428-456. HAM, C. L. AND JURS, P. C. Structure-activity studies of musk odorants using pattern recognition: monocyclic nitrobenzenes. Chem. Senses 10: 491-505, 1985. HARRISON, T. A. AND SCOTT, J. W. Olfactory bulb responses to odor stimulation: analysis of response pattern and intensity relationships. J. Neurophysiol. 56: 157 I- 1590, 1986. HUBEL, D. H. AND WIESEL, T. N. Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex. J. Physiol. Lond. 160: 106-154, 1962. JOURDAN, F., DUVEAU, A., ASTIC, L., AND HOLLEY, A. Spatial distribution of ( 14C) 2-deoxyglucose uptake in the olfactory bulbs of rats stimulated with two different odours. Brain Res. 188: 139- 154, 1980. KAUER, J. S. Coding in the olfactory system. In: Neurobiology ofTaste and Smell, edited by T. E. Finger and W. L. Silver. New York: Wiley, 1987, p. 205-232. KAUER, J. S. Contributions of topography and parallel processing to odor coding in the vertebrate olfactory pathway. Trends Neurosci. 14: 79-85, 1991. LANCET, D. Vertebrate olfactory reception. Annu. Rev. Neurosci. 9: 329355,1986. LANCET, D., GREER, C. A., KAUER, J. S., AND SHEPHERD, G. M. Mapping of odor-related neuronal activity in the olfactory bulb by high-resolution 2-deoxyglucose. Proc. Natl. Acad. Sci. USA 79: 670-674, 1982. MAIR, R. G. Response properties of rat olfactory bulb neurons. J. Physiol. Lond. 326: 341-359,1982. MATHEWS, D. F. Response patterns of single units in olfactory bulb of the rat to odour. Brain Res. 47: 384-400, 1972. MEREDITH, M. Patterned response to odor in mammalian olfactory bulb: the influence of intensity. J. Neurophysiol. 56: 572-597, 1986. MORI, K. Membrane and synaptic properties of identified neurons in the olfactory bulb. Prog. Neurobiol. 29: 274-320, 1987. MORI, K., FUJITA, S. C., IMAMURA, K., AND OBATA, K. Immunohistochemical study of subclasses of olfactory nerve fibers and their projections to the olfactory bulb in the rabbit. J. Comp. Neural. 242: 2 14-229, 1985. MORI, K., IMAMURA, K., AND MATAGA, N. Differential specificities of single mitral cells in rabbit olfactory bulb for a homologous series of fatty acid odor molecules. J. Neurophysiol. 67: 786-789, 1992. AKESON,

R. A. Primary

olfactory

neuron

biology of the Olfactory System, edited

AND

K. MORI

MORI, K., olfactory 1990, p. MORI, K., displaced

IMAMURA, K., AND ONODA, N. Signal processing in the rabbit bulb. In: ISOT X, edited by K. B. Doving. Oslo: GCS A. S., 134-141. KISHI, K., AND OJIMA, H. Distribution of dendrites of mitral, mitral, tufted, and granule cells in the rabbit olfactory bulb. J. Comp. Neurol. 219: 339-355, 1983. MORI, K. AND TAKAGI, S. F. Inhibition in the olfactory bulb: dendrodendritic interaction and their relation to the induced waves. In: Food Intake and Chemical Senses, edited by K. Katsuki, M. Sato, S. F. Takagi, and Y. Oomura. Tokyo: Univ. of Tokyo Press, 1977, p. 33-43. MORI, K. AND TAKAGI, S. F. An intracellular study of dendrodendritic inhibitory synapses on mitral cells in the rabbit olfactory bulb. J. Physiol. Lond. 279: 569-588, 1978. MOUNTCASTLE, V. B. Central nervous mechanisms in mechanoreceptive sensibility. In: Handbook of Physiology. The Nervous System. Sensory Processes. Bethesda, MD: Am. Physiol. Sot., 1984, sect. 1, vol. III, p. 789-878. NATHANS, J. AND HOGNESS, D. S. Isolation and nucleotide sequence of the gene encoding human rhodopsin. Proc. Nat/. Acad. Sci. USA 76: 44054408,1984. NATHANS, J., THOMAS, D., AND HOGNESS, D. S. Molecular genetics of human color vision: the genes encoding blue, green, and red pigments. Science Wash. DC 232: 193-202, 1986. O’DOWD, B. F., LEFKOWITZ, R. J., AND CARON, M. G. Structure of the adrenergic and related receptors. Annu. Rev. Neurosci. 12: 67-83, 1989. ONODA, N. AND MORI, K. Depth distribution of temporal firing patterns in olfactory bulb related to air intake cycles. J. NeurophysioZ. 44: 29-39, 1980. PHILLIPS, C. G., POWELL, T. P. S., AND SHEPHERD, G. M. Responses of mitral cells to stimulation of the lateral olfactory tract in the rabbit. J. Physiol. Lond. 168: 65-88, 1963. RALL, W. AND SHEPHERD, G. M. Theoretical reconstruction of field potentials and dendrodendritic synaptic interactions in olfactory bulb. J. Neurophysiol. 3 1: 884-9 15, 1968. RHEIN, L. D. AND CAGAN, R. H. Biochemical studies of olfaction: isolation characterization and odorant binding activity of cilia from rainbow trout olfactory rosettes. Proc. Natl. Acad. Sci. USA 77: 44 12-44 16, 1980. SCHWOB, J. E. AND GOTTLIEB, D. I. The primary olfactory projection has two chemically distinct zones. J. Neurosci. 6: 3393-3404, 1986. SHEPHERD, G. M. Synaptic organization of the mammalian olfactory bulb. Physiol. Rev. 52: 864-9 17, 1972. SHEPHERD, G. M. A molecular vocabulary for olfaction. In: Olfaction and Taste IX, edited by J. Atema. New York: NY Acad. Sci., 1987, p. 98103. SHEPHERD, G. M. Computational structure of the olfactory system. In: Olfaction, A Model System for Computational Neuroscience, edited by J. L. Davis and H. Eichenbaum. Cambridge, MA: MIT Press, 199 1, p. 3-41. SHEPHERD, G. M. AND FIRESTEIN, S. Making scents of olfactory transduction. Curr. Biol. 1: 204-206, 199 1. STEWART, W. B., KAUER, J. S., AND SHEPHERD, G. M. Functional organization of rat olfactory bulb analyzed by 2-deoxyglucose method. J. Comp. Neural. 185: 715-734, 1979. TANABE, T., 11~0, M., AND TAKAGI, S. F. Discrimination of odors in olfactory bulb, pyriform-amygdaloid areas, and orbitofrontal cortex of the monkey. J. Neurophysiol. 38: 1284-1296, 1975. TOMITA, T., KANEKO, A., MURAKAMI, M., AND PAUTLER, E. L. Spectral response curves of single cones in the carp. Vision Res. 7: 5 19-53 1, 1967. WILSON, D. A. AND LEON, M. Spatial pattern of olfactory bulb single unit responses to learned olfactory cues in young rats. J. Neurophysiol. 59: 1770-1783,1988. WELLIS, D. P. AND SCOTT, J. W. Intracellular responses of identified rat olfactory bulb interneurons to electrical and odor stimulation. J. Neurophysiol. 64: 932-947, 1990. WELLIS, D. P., SCOTT, J. W., AND HARRISON, T. A. Discrimination among odorants by single neurons of the rat olfactory bulb. J. Neurophysiol. 6 1: 1161-l 178, 1989.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (132.174.254.157) on January 14, 2019.