This is a
preprint of an article published in the Journal of Comparative Neurology, 1999,
409:529-548.
© 1999
Wiley-Liss, Inc
n-Aliphatic Acid Odorants
in the Rat Olfactory Bulb
Brett A. Johnson,*
Cynthia C. Woo, Edna E. Hingco,
Keith L. Pham, and Michael Leon
School of Biological Sciences, University of California, Irvine, CA 92697-4550
Number of text pages: 48
Number of figures: 8
Number of tables: 3
Abbreviated title: Aliphatic acid responses in the olfactory bulb
Associate Editor: Jon H. Kaas
Indexing terms: chemical senses; deoxyglucose; odor; olfactory receptors; olfactory coding
*Correspondence to: Brett A. Johnson, Department of Psychobiology, University of California, 2205 BioSci II, Irvine, CA 92697-4550. Telephone: (949)824-7303. Fax: (949)824-2447. E-mail: bajohnso@uci.edu
Grant sponsor: NIDCD; Grant number DC03545
Abstract
In an effort to understand the means by which similar chemical odorants are encoded in the mammalian brain, we exposed rats to a homologous series of n-aliphatic acids and mapped the response of the entire olfactory bulb glomerular layer using a high-resolution [14C]-2-deoxyglucose uptake technique. We found that these similar odorants evoked spatially clustered, but distinct responses in the bulb that varied systematically with carbon chain length. In addition to these chemotopic responses, different odorants within the series evoked systematic differences along two other dimensions: amount of deoxyglucose uptake and extent of the glomerular layer showing high activity. Increases along these two dimensions also were correlated with increasing carbon number. The focal glomerular responses were mirrored by responses in deeper bulb layers. Decreasing the odorant concentration decreased the deoxyglucose uptake within focal regions. The focal regions of activity occurred in pairs involving both medial and lateral representations in the bulb, perhaps reflecting the paired medial and lateral projections of olfactory sensory neurons expressing specific types of odorant feature receptor proteins. The observed spatial pattern of response also may explain both the failure of some bulb lesions to interfere with behavioral olfactory responses and the success of other lesions in blocking olfactory responses. These data support a model of parallel, distributed processing of odorants along multiple dimensions. They also support the notion that analyses of the spatial relationships among odorant responses in the olfactory bulb can reveal aspects of the mechanism for odor chemical coding.
The mechanism by which chemical cues become encoded into odors in the mammalian nervous system is still not well understood. One approach to this problem has involved the mapping of olfactory bulb activity in response to odorants using [14C]2-deoxyglucose (2‑DG) to determine whether the spatial patterns of neural responses would lend insight into the coding scheme used by the olfactory system. To date, the responses to a limited number of odorants with very different chemical structures have been mapped (Sharp et al., 1977; Stewart et al., 1979; Jourdan et al., 1980; Bell et al., 1987; Royet et al., 1987; Guthrie et al., 1993; Sallaz and Jourdan, 1993; Shepherd, 1994; Johnson and Leon, 1996). These studies provided early evidence for an odor-dependent map by showing that the locations of focal regions of activity differ between very different odorants and that a given odorant produces similar patterns in different animals. Recently, we showed that these odor-evoked patterns of activity may be comprised of unitary responses to distinct molecular features of odorants (Johnson et al., 1998), a finding that supports the hypothesis that each chemical may be identified through a parallel processing of its molecular features (Kauer and Cinelli, 1993; Shepherd, 1994; Axel, 1995; Mori and Yoshihara, 1995; Buck, 1996; Johnson et al., 1998).
Because very closely related chemicals evoke the perception of distinct odors, it remains necessary to address how such similar odorants are coded by the mammalian nervous system. A powerful choice for such studies involves homologous series of odorants differing by single steps of carbon chain length but possessing the same functional chemical group (Dšving, 1966; Mori et al., 1992; Imamura et al., 1992; Sato et al., 1994). Using homologous series of aliphatic acids and other aliphatic, carbonyl-containing compounds, MoriŐs group found that individual mitral/tufted cells in the dorsomedial rabbit olfactory bulb are tuned to odorants possessing a limited range of carbon-chain lengths (Mori et al., 1992; Imamura et al., 1992). The cells responding maximally to one carbon chain length appeared to be located in slightly different positions from those preferring another chain length (Mori et al., 1992). Mori and coworkers proposed that neighboring glomeruli within the dorsomedial bulb may respond broadly to carbonyl-containing odorants of various chain lengths, with a slight preference for a given chain length, and that lateral inhibitory interactions between glomeruli and/or mitral/tufted cells may refine the specificity of an individual mitral/tufted cell to a more limited range of stimuli (Imamura et al., 1992). Consistent with this hypothesis, they found that reduction of lateral inhibition, achieved by local iontophoresis of a receptor antagonist, resulted in a broader tuning of the mitral/tufted cells with respect to carbon chain length (Yokoi et al., 1995).
Spatial clustering of glomeruli to accomplish odorant tuning would suggest that spatial distributions of responses may be used by the olfactory system to encode odor quality. Indeed, spatial clustering of glomeruli responding to related amino acid odorants has been found in the responses of zebrafish olfactory bulbs (Friedrich and Korsching, 1997). In the current study, we have applied a high-resolution, 2-DG mapping procedure to address a number of issues related to odor mapping and its significance for olfaction. Our first goal was to test directly the prediction that neighboring glomeruli in the dorsomedial region of a mammalian olfactory bulb are differentially activated within a homologous series of aliphatic acids.
Many studies have reported a single major focus of glomerular or granule cell activity in the dorsomedial region of the mammalian olfactory bulb in response to propionic acid, a three-carbon aliphatic acid. These neural activity studies have employed recording of olfactory field potentials (Mori et al., 1992), 2-DG uptake (Slotnick et al., 1987; 1989; Sallaz and Jourdan, 1992; 1993), in situ hybridization for c-fos mRNA (Guthrie et al., 1993; Guthrie and Gall, 1995), and immunohistochemistry for Fos-like antigens (Onoda, 1992; Sallaz and Jourdan, 1993). When the portion of the bulb containing this propionic acid-evoked focus is ablated, rats remain capable of detecting propionic acid (Slotnick et al., 1997; Lu and Slotnick, 1994; 1998). Furthermore, rats remain capable of discriminating between odorants following extensive lesions involving all but the caudal 20% of the bulb (Lu and Slotnick, 1998). Even closely related chemicals such as propionic and acetic acid can be discriminated following extensive bulb lesions (Lu and Slotnick, 1994; 1998; Slotnick et al., 1997). These results indicate that lesioned animals are capable of discriminating between the tested odorants using only the spared posteromedial bulb, thus questioning the relevance of focal odorant responses to odor coding. However, because the entire bulb had not been mapped in the prior studies, the second goal of this study was to apply our 2-DG mapping technique to determine whether this posteromedial region contains reliable focal responses to propionic acid that could underlie the retained capacity for odor discrimination.
A single focus of odorant-evoked activity is contrary to the typical projection patterns of homologous sensory neurons. Olfactory sensory neurons expressing the same individual member of the olfactory receptor gene superfamily (Buck and Axel, 1991) project to as few as two of the ~2000 glomeruli present in each rat or mouse olfactory bulb (Ressler et al., 1994; Vassar et al., 1994; Mombaerts et al., 1996). One of the two glomeruli is invariably located on the lateral aspect of the bulb, and the other is located medially, with the lateral glomerulus being situated more rostrally and dorsally than the medial one (Ressler et al., 1994; Vassar et al., 1994; Mombaerts et al., 1996). Indeed, pairs of lateral and medial glomerular fields, possessing the dorsal/ventral and rostral/caudal relationships predicted from the projection patterns of homologous sensory neurons, were observed following high-resolution mapping of the response of the entire olfactory bulb to different aliphatic ester odorants (Johnson et al., 1998). The third goal of our current study was to apply the 2-DG mapping technique to test the prediction that there is at least one additional propionic acid-responsive area in the lateral bulb that would be rostral and dorsal to the dorsomedial focus of activity.
Our results reveal an exquisite order to the spatial representations of these closely related, aliphatic acid odorants. The systematic variations in activity include discrete, chemotopic responses involving paired medial and lateral glomerular foci in both the rostral and caudal parts of the bulb. These data support a relatively simple model of odorant feature processing in the rat olfactory bulb in which spatial coding is used to amplify, segregate, and tune responses to specific chemical attributes of odorant molecules.
Materials and
Methods
Odor exposures
All procedures involving animals were approved by the UC Irvine Institutional Animal Care and Use Committee (IACUC). A total of 55 Wistar rats (postnatal day 19-21) were used in this study. The rats were bred in our own colony and had access to food and water ad libidum. On the day following birth, litters were culled to no more than eight pups. Both males and females were used, with the number of each counter-balanced among odorants. At least 1 hour prior to any exposure, the entire litter with the dam was transferred to a clean cage to reduce carryover of home cage odorants into the test apparatus. Pups were removed from this cage no more than 15 minutes prior to the odor exposure. Up to four rats were used from each litter, but no two rats from the same litter were exposed to the same odorant condition.
In one experiment, rats were exposed to air vehicle (n=6) or to valeric acid (n=6). Air exposures always preceded odorant exposures on a given day to avoid odorant contamination of the air-exposed animals. For exposures to different aliphatic acid odorants (n=6 for each odorant), the order of odorant presentation was explicitly varied across different exposure days to avoid any possible effects of order or systematic cross-contamination between odorants.
As in our prior study of aliphatic esters, we considered it important to compare odorants at the same vapor-phase concentration in order to correlate activity with chemical structures of the odorant molecules. To produce such stimuli, we calculated predicted vapor pressures (Hass and Newton, 1975) and diluted the odorants differentially to achieve a final vapor concentration of 7.2 parts per million (Table 1). Octanoic acid was not sufficiently volatile to achieve this vapor-phase concentration. Therefore, an additional six animals each were exposed to caproic acid or octanoic acid at a vapor phase concentration of 0.8 part per million. One additional rat was exposed to 25 parts per million valeric acid (1/8 dilution of saturated vapor) to explore relationships between patterns in the glomerular layer and deeper bulbar layers. All odorants were acquired at their highest available purity (> 98-99%) (Fisher Scientific, Tustin, CA). Odorants were vaporized by bubbling high-purity nitrogen gas at a flow rate of 230 ml/min through a column of 100 ml undiluted, liquid odorant in a 125‑ml gas washing bottle fitted with a stopper assembly that included a fritted glass cylinder of 40-60 µm pore size. For air-exposed controls, an empty gas washing bottle was used. A portion of the odorized nitrogen was then diverted to a vent, while another portion, regulated using a Gilmont size 1 flowmeter, was mixed with ultra zero grade air to achieve a final flow rate of 2 liters/min of the desired dilution. The apparatus was equilibrated at the desired flow rates for at least 15 min before any exposure. Tubing and connections were made with Teflon, Kynar, brass, or glass to minimize interactions with the odorants. Clean tubing and exposure chambers were used for each odorant condition.
Rats were given a subcutaneous injection of [14C]2-DG (Sigma Chemical Company, St. Louis, MO; 0.16 mCi/kg) immediately prior to odor exposure, which was conducted in 1‑liter Mason jars with odorant entry and exit ports bored into the lid. Odorant entry began after the rat was introduced into the chamber, so that the concentration rose steadily over the course of the exposure. Exposures were for 45 minutes and were terminated by removing the rat from the jar, decapitating it, and freezing its brain in isopentane at ‑45ˇC.
Brains were stored at ‑80ˇC until sectioning, at which time they were warmed to a cryostat temperature of ‑18ˇC for at least 45 minutes prior to cutting. Coronal sections were cut perpendicularly to the long axis of the olfactory bulb. Every sixth 20‑µm section was used for autoradiography. These sections were collected on room-temperature 22 x 22 mm coverglasses, rapidly dehydrated on a slide warmer at 60ˇC, taped to cardboard, and exposed to Kodak SB-5 autoradiography film for 10 days along with 14C-standards (ARC-146A; American Radiolabeled Chemicals, St. Louis, MO) that had been calibrated to tissue equivalents (nCi/g) of isotope. Adjacent sections were taken on cooled, gelatin-subbed microscope slides and were stained using cresyl violet. Slides and autoradiograms were coded prior to analysis.
Anatomical standardizations for measurements of 2-DG uptake
We used further refinements of our previously published methods, which employed polar grids to sample 2-DG uptake across the glomerular layer (Johnson and Leon, 1997; Johnson et al., 1998). In our prior studies, uptake was measured at fixed angles in all bulb sections. Although this strategy was sufficient to compare patterns of activity across odorants, section size increases as one moves caudally through the bulb, so that the resulting arrays of measurements overestimated the representation of the anterior bulb relative to the posterior bulb. In order to achieve measurements that would be more proportional to bulb space, the current study employed a series of 17 grids, which differed in the numbers of angles.
Three anatomical landmarks were used to standardize rostral-caudal distances between bulbs, which vary in length from animal to animal. For each bulb, we determined the first cresyl violet-stained section that possessed an external plexiform layer (the first section to be mapped), the first section that contained an accessory olfactory bulb, and the last section that contained a mitral cell layer on its medial aspect (the last section to be mapped). By comparing values from bulbs in the current study that were used for comparisons between odorants (42 brains total), as well as values from bulbs in 36 other animals of similar age used in other studies, we determined the most commonly encountered (modal) distances between consecutive landmarks. The ŇtypicalÓ bulb was found to possess 25 collected sections (3.0 mm) between the first external plexiform layer-containing section and the first accessory olfactory bulb-containing section, and 19 sections (2.28 mm) after the first accessory olfactory bulb-containing section up to the last mitral cell layer-containing section. Six individual bulbs that represented these modal distances and that yielded intact sections then were used to determine grid angles that would give 120-µm spacing between adjacent measurements in the glomerular layer.
Mapping of glomerular 2-DG uptake
Images of the cresyl violet-stained sections, the autoradiograph sections, and the 14C‑standards were digitized on a light box using a Sony X6-77 CCD camera and NIH IMAGE software (version 1.45) at a magnification giving 108 pixels/mm. Digitized images were saved to disks and were analyzed later using IMAGE (version 1.61). First, the outline, as well as the glomerular layer, of each cresyl violet-stained section was traced onto an acetate sheet positioned on the computer monitor. Boundaries of the subependymal zone in all sections caudal to the first accessory olfactory bulb also were traced. For mapping, the traced outline of the stained section was superimposed on the outline of a pseudocolor-enhanced image of the adjacent autoradiograph section (Fig. 1). The appropriate grid, printed on a transparency, was centered within the traced glomerular layer, with the vertical axis held parallel to the midline between the two bulbs (Fig. 1). For sections caudal to the first accessory olfactory bulb, these grids were centered in the traced subependymal zone (see Fig.1, section C). The appropriate grid for a given section was calculated on the basis of the sectionŐs fractional distance between consecutive landmarks. For a bulb of typical length, grids were changed every section for the first ten sections, where the perimeter of the glomerular layer changed most rapidly, and then every other section for the remainder of the bulb. The number of angles within a grid varied from 19 in section #1 to 80 in the first section containing the accessory olfactory bulb.
Measurements were taken in units of gray scale using a circular sampling area of 120‑µm diameter. Each measurement was taken midway between the traced boundaries of the glomerular layer at an intersection of a gridline, beginning at the dorsal extreme of the section and moving first in the lateral direction and then around the section. For ease in constructing maps, film background measurements were taken for smaller sections before and after measuring from the glomerular layer, so that the 39th measurement in each section represented the ventral-most position within a section, and so that there would be a total of 80 values recorded for every section (Fig. 1). If the section was judged to lack a glomerular layer at a given grid intersection, a background measurement was taken from blank film adjacent to the section. These background readings were less than one-half the value of the lowest glomerular layer measurement and were later deleted. Measurements from a given section were saved in a single file. The files for the different sections in a given bulb then were merged to create an array of values containing 80 rows (measurements) by as many columns as there were sections in that bulb.
Mapping of uptake in deeper bulbar layers
To correlate patterns in the glomerular layer with patterns in deeper bulb layers, uptake of 2‑DG also was measured in the external plexiform layer and in the internal plexiform/superficial granule cell layer of a single animal. For these measurements, the mitral cell layer in the adjacent cresyl violet-stained sections was traced and overlayed on the autoradiograph sections along with the tracings of the other layers. Although the deeper layers were smaller in perimeter than was the glomerular layer, the same grids and sampling tool as described above were used for these analyses to allow for direct comparisons with the glomerular layer. External plexiform layer uptake was measured midway between the inner boundary of the glomerular layer and the mitral cell layer at each grid intersection. Measurements of the internal plexiform/superficial granule cell layer were taken just deep to the mitral cell layer.
In another study, the internal plexiform/superficial granule cell layer was mapped for the rats that were exposed to propionic, valeric, or caproic acids in order to compare the patterns evoked by the different odorants. This analysis used measurements taken at every other grid angle, given the smaller perimeter of the internal plexiform/superficial granule cell layer compared to the glomerular layer. Grids were centered with respect to the mitral cell layer in sections preceding the accessory olfactory bulb. Only the right bulb was mapped for these analyses.
Transformation of the arrays
The arrays of gray scale values were converted to units of nCi/g tissue equivalents using a standard curve derived from measurements of the images of the 14C-standards that were placed on the same films as the sections. The arrays then were expanded or contracted at evenly spaced intervals in order to standardize the number of sections between consecutive landmarks (first accessory bulb = section #25, last mitral cell layer = section #44). Expansion was accomplished by duplicating sections measured using the same grid as would have been appropriate for the ŇmissingÓ section. Contraction was accomplished by replacing pairs of sections measured with the same grid with a single mock section containing the average of the values in the two sections. These adjustments were made independently for the left and right bulbs of each animal. The arrays for the two bulbs of a given animal then were averaged.
Each animalŐs average array of glomerular layer uptake was transformed in two different ways. For most of the analyses, the uptake at each postion in the array was divided by uptake in the subependymal zone. Subependymal zone uptake was calculated by averaging the uptake in this region across five consecutive sections of each bulb (centered at two-thirds the distance between section #1 and the first accessory olfactory bulb). This transformation into glomerular layer/subependymal zone uptake was used to correct for any slight differences in the amount of [14C]2‑DG injected into different animals, and it allowed comparison of both patterns and amounts of uptake across odorant conditions.
The other transformation involved the calculation of a z-score value for each position in the array relative to the average and standard deviation of the values across the whole array of the same bulb, as had been used previously (Kent and Mozell, 1992; Johnson et al., 1998). This intra-bulbar standardization allows for distinguishing differences in the patterns of uptake across different odorants without being biased by differences in the absolute values of uptake. Arrays transformed to z scores were used both to define fields of uptake for statistical analyses and to calculate pattern dissimilarities across odorants.
Contour charts
Arrays were visualized as contour charts constructed using Microsoft Excel 98 software. In these contour charts, individual calculated values are present at the intersections of gridlines. Curves are drawn to represent locations within the array that are projected to have the same level of uptake. The areas between adjacent curves are then assigned colors to represent bins of values. Color coding is chosen such that higher values received warmer colors and lower values received cooler colors. Figure 1 illustrates the relationships between bulbar anatomy and these charts.
The charts have a shape resembling the Ňrolled-out mapsÓ previously used by others to express spatial distributions of bulbar activity (Stewart et al., 1979; Jourdan et al., 1980; Royet et al., 1987), patterns of nerve degeneration in the glomerular layer (Land, 1973), and patterns of antibody staining in the bulb (Schwob and Gottlieb, 1986). Our charts differ from the typical maps in that they open dorsally instead of ventrally and are rotated so that dorsal is located at the top and bottom boundaries of the charts, lateral is located in the upper half, medial is in the lower half, and ventral is at half the height of the chart. Rostral is to the left and caudal to the right. We chose to open our maps dorsally because it is difficult to maintain complete tissue integrity at the dorsal extremity of coronal sections, and this orientation minimizes the impact of occasional missing values on the appearance of individual bulb maps. Both the number of measurements and the amount of space used in the chart increase from rostral to caudal across the anterior two-thirds of the bulb in concert with the increasing section perimeter (Fig. 1). The medial bulb continues further caudally than does the lateral bulb, which is replaced with the lateral olfactory tract and anterior olfactory nucleus in caudal sections. Figure 1 also shows three individual, pseudocolor-enhanced, autoradiography sections containing foci of uptake in a rat exposed to valeric acid vapor. The relationship between the locations of foci in individual sections and the locations of warm colors in the chart provides a further orientation to the areas subject to most consideration in our analyses.
Pattern dissimilarity
To correlate differences in the patterns of odorant-evoked uptake with differences in carbon number between the odorants, indices of pattern dissimilarity were calculated (Kent and Mozell, 1992; Johnson et al., 1998). For these analyses, z score-transformed arrays for animals exposed to the same odorant were averaged together. Each pair of the averaged arrays, representing a comparison of two odorant-evoked patterns of uptake, were subtracted from each other. Because five odorants were investigated, there were ten unique pairs of odorant-evoked patterns that were subtracted. The values in each difference array were squared, and the squared values then were averaged. The square root of this mean square was the index of dissimilarity between the pair of odorant-evoked patterns.
RESULTS
Valeric acid evokes 2-DG uptake in four discrete fields
To determine the areas of the bulb that are likely to respond to aliphatic acids, rats were exposed either to air vehicle or to valeric acid. Figure 2 (top) shows the average patterns of uptake across the rats exposed to the two exposure conditions. When the average array for the air-exposed rats was subtracted from that for the valeric acid-exposed rats, the difference map revealed four circumscribed regions of increased uptake attributable to the odorant exposure (Fig. 2, bottom left). These fields of increased response were numbered from rostral to caudal for easy reference.
Field 1 was located in the dorsal extremity of sections near the rostral pole of the bulb. Field 2 was located in the dorsomedial region, approximately one-half the distance between the first section containing an external plexiform layer and the first section containing an accessory olfactory bulb. This field is in the part of the bulb where focal responses to propionic acid and isovaleric acid have been described previously (Royet et al., 1987; Slotnick et al., 1987; 1989; Sallaz and Jourdan, 1992; 1993). Field 3 was located in the caudal half of the midlateral bulb, and Field 4 was located in the caudal third of the ventromedial bulb.
To assess the animal-to-animal reliability of these areas of increased uptake, separate two-tailed t-tests were performed across rats at each standardized location within the arrays. Spatial clusters of reliable, valeric acid-evoked increases were present within Fields 2, 3, and 4 (Fig. 2, bottom right). Field 1 was not associated with clusters of reliable increases in this analysis: a likely explanation is that individual values contributing to Field 1 were often missing due to loss of tissue along the dorsal edge of small coronal sections from the rostral bulb. Indeed, comparisons among aliphatic acid odorants, including valeric acid, were found to yield significant differences within Field 1 in separate analyses that were less sensitive to missing values (see below). Very few reliable increases or decreases were detected outside of the four fields, and, when present, these changes were rarely clustered (Fig. 2, bottom right).
Changes in the amount of 2-DG uptake evoked by different
aliphatic acids
To evaluate how odorant carbon chain length influences 2-DG uptake in the glomerular layer, we presented six rats each with equal vapor-phase concentrations of five different straight-chain aliphatic acids, varying from two to six carbons in length (Table 1). Arrays of glomerular layer uptake/subependymal zone uptake then were averaged across rats exposed to the same odorants. Contour charts displaying the distribution of values within these averaged arrays are shown in Figure 3A.
In general, all of the aliphatic acids evoked uptake in the same four fields evoked by valeric acid (compare Figure 2, top with Figure 3A). The greatest exception was acetic acid, which did not appear to evoke as much uptake within the posterior fields, 3 and 4. In addition to the four fields seen for the other odorants, caproic acid also stimulated uptake across larger regions of the posterolateral and posteromedial glomerular layer at this concentration (7.2 parts per million).
One of the most obvious effects of increasing carbon chain length was an increase in the amount of 2‑DG taken up within the four fields (Fig. 3A). To analyze this apparent effect statistically, we defined boundaries for the four fields that were unbiased with respect to individual odorants by averaging z score-transformed arrays across all 30 animals in this part of the study. Areas containing contiguous values exceeding 0.5 in this grand average array were taken to represent the extents of the fields for this analysis. These areas are displayed by transitions from green to yellow in the contour chart shown in Figure 3B. (Field 1 wrapped around the dorsal extreme of the sections, so that values contributing to this field correspond to both the top and bottom parts of the chart.)
The average uptake across each field in each animal then was calculated (Fig. 3C). For all four fields, uptake was significantly different across odorants (Table 2), with increasing values of uptake being associated with increasing odorant carbon number (Fig. 3C). Field 1 and Field 2 showed similar effects of increasing carbon length on levels of uptake, with butyric acid giving values intermediate between propionic and valeric acids. For any given odorant, the magnitude of uptake within these two fields was similar. Fields 1 and 2 also had the same rostral-caudal, dorsal-ventral, and lateral-medial relationships to each other that have been described for paired glomeruli receiving projections from homologous olfactory sensory neurons (Ressler et al., 1994; Vassar et al., 1994; Mombaerts et al., 1996). Field 3 and Field 4 also had similar odorant-response profiles, displaying a small increase between acetic and propionic acids and a larger increase between butyric and valeric acids. For any given odorant, these two fields were stimulated to a similar extent. The pair of Fields 3 and 4 also have the predicted anatomical relationships for paired glomeruli receiving homologous olfactory sensory neuron projections.
Odorant-dependent changes in the locations
of 2‑DG
uptake within fields
Although the five aliphatic acids stimulated uptake clustered within each of the four fields, there were odorant-dependent differences in the relative locations of uptake within those fields. To evaluate the statistical significance of these changed locations, centroids of uptake were calculated for each animal within the fields delineated in Figure 3B. The centroids were described by two coordinates, one indicating the rostral-caudal location in units of section number, and one indicating the position along the perimeter of the section in units of measurement number. Centers of ellipses in Figure 3B show the average positions of the centroids in rats exposed to each odorant. The two axes of each ellipse indicate the standard error of the mean of each coordinate of the centroid across animals exposed to the same odorant. The centroids within every field were significantly different across odorants along the perimeter of the section (Table 2), with increasing carbon chain length being correlated with increasingly ventral activity (Fig. 3B). The centroid along the rostral-caudal axis also was significantly different for Fields 1, 3, and 4 (Table 2), with increasing carbon chain length being associated with increasingly rostral activity (Fig. 3B). Field 2 did not show any change in the centroid of activity along the rostral-caudal axis (Fig. 3B, Table 2).
Appearance of foci within the four fields
The changes in amounts and locations of 2‑DG uptake also were apparent upon inspection of individual autoradiography sections from rats exposed to the different aliphatic acid odorants. Figure 4 illustrates foci occurring in Fields 1 and 2 from representative animals exposed to the five odorants. The rostral shift in the location of activity in Field 1 with increasing carbon number is evident in Figure 4A. Acetic, propionic, and butyric acids tended to evoke the greatest uptake in foci located in sections 4 and 5, whereas valeric acid stimulated the most uptake in sections 3 and 4, and caproic acid evoked greater uptake in sections 2 and 3 (Fig. 4A). In some rats, the foci evoked by caproic acid extended even more rostrally and were detected in the section taken prior to section #1 of our analysis (i.e., prior to the section containing the first external plexiform layer). The foci evoked by valeric and caproic acids also extended more consistently into the lateral aspect of the sections than did those evoked by the smaller acids (Fig. 4A), which explains the previously established shift in the centroid along the perimeter of the sections (Fig. 3B). For all odorants, the foci within Field 1 moved to more dorsal locations when followed along the rostral-caudal axis through the bulb (Fig. 4A).
Within Field 2, the shift in foci to more ventral locations with increasing odorant carbon number (Fig. 3B) was readily apparent from individual autoradiography sections (Fig. 4B). Foci in section #15 from caproic acid-stimulated rats actually appeared to be located almost mid-medially (Fig. 4B), and there was some evidence for multiple, spatially segregated foci evoked by caproic acid in this field (double arrows in Fig. 4B). Field 2 was not associated with a changed centroid along the rostral-caudal axis in our statistical analyses (Fig. 3B, Table 2). This lack of a rostral-caudal shift appeared to be due to an expansion in both the rostral and the caudal directions with increasing odorant carbon number (Fig. 4B). For all odorants, there was a tendency for foci in Field 2 to move to more ventral positions when followed from rostral to caudal through the bulb (Fig. 4B).
The foci within Fields 1 and 2 tended to be larger in animals exposed to valeric and caproic acids than in animals exposed to the smaller acids (Fig. 4A,B). This increase in the size of the foci is illustrated further in Figure 5, which shows the peak of activity detected in Field 1 in an animal exposed to propionic acid and in another animal exposed to valeric acid. The focus of uptake evoked by propionic acid appeared to be associated with only one to three glomeruli in a single coronal section (Fig. 5). In contrast, the valeric acid-evoked focus was associated with approximately fifteen glomeruli in a section (Fig. 5).
Foci detected within Fields 3 and 4 are shown in Figure 6. Acetic acid rarely evoked detectable foci within Field 3. When present, the acetic acid-evoked foci tended to be very light and scattered in the caudal portions of the field (Fig. 6A). Propionic acid evoked foci of greater uptake in the caudal portion of Field 3 (Fig. 6A). With further increases in odorant carbon number, the evoked foci became both elevated in amount of uptake and situated in progressively rostral and ventral locations (Fig. 6A), thus leading to the statistically significant shifts in centroids described earlier (Fig. 3B, Table 2). Caproic acid evoked lateral foci that extended rostrally and ventrally beyond the boundaries of Field 3 used in the statistical analyses (Fig. 3B). These rostrally extending, lateral foci can be seen in the same sections as Field 2 (Fig. 4B). To a lesser extent, valeric acid also evoked lateral foci of uptake in the more rostral sections (Fig. 4B).
Similar to the case for Field 3, acetic acid rarely evoked foci of 2‑DG uptake within Field 4, and, when present, these light foci were scattered in caudal portions of the field (Fig. 6B). Propionic acid evoked foci of greater uptke in the caudal parts of the field. With increasing carbon number, the aliphatic acid odorants evoked foci that were more intense and that extended into more rostral and ventral locations (Fig. 6B), consistent with the statistically significant shifts in the calculated centroids (Fig. 3B, Table 2).
Correlations with carbon-chain length
In our previous analyses of aliphatic ester odorants, we noted that molecules of more similar size produced more similar patterns of 2‑DG uptake and that increasing molecular size was associated with increasing levels of pattern complexity. To determine if there were comparable relationships across the aliphatic acid odorants, similar analyses were conducted in the current study. As a measure of pattern complexity, we calculated the standard deviation of values across the arrays of 2‑DG uptake for each animal and expressed this standard deviation as a percent of the mean uptake across the array of the same animal. These values were found to be significantly different across odorants (P < 0.05, F4,25=14.19, single-factor ANOVA) and positively correlated (r=0.98) with odorant carbon number (Fig. 7A). Thus, increasing carbon number results in an increase in the complexity of the pattern of 2‑DG uptake.
To express the degree of difference between patterns evoked by different odorants, we calculated indices of pattern dissimilarity between each possible pair of odorants using averaged, z score-transformed arrays (Table 3). The degree of pattern dissimilarity between odorant pairs was found to be positively correlated (r=0.87) with the difference in the number of carbons between the pair (Fig. 7B). In other words, odorants of more similar size evoked more similar patterns of uptake, whereas odorants of increasingly different size evoked increasingly distinct patterns of uptake. The trend of increasing dissimilarity with increasing disparity in carbon length also was observed for each individual odorant, although caproic acid yielded patterns that generally were more distinct for any given difference in carbon length (Table 3).
Effects of odorant concentration on patterns of 2‑DG
uptake
To determine if odorant concentration affected the patterns of glomerular layer activity evoked by the aliphatic acids, six additional rats were exposed to 0.8 part per million caproic acid, which represents a 9-fold dilution relative to the previously described exposures. Contour charts of 2‑DG uptake averaged across these animals showed the same four fields of activity seen for the other acids (Fig. 8A). The amount of uptake within each of the fields was lower than that observed following exposures to 7.2 parts per million caproic acid, and the activation of large areas of the lateral and medial bulb that was observed for 7.2 parts per million caproic acid was observed to a much lesser extent at the lower concentration (Fig. 8A). The relative locations of the peaks of uptake within the four fields did not appear to be altered by the reduction in odorant concentration (Fig. 8A).
Patterns evoked by octanoic acid
Six rats were exposed to 0.8 part per million octanoic acid, a straight-chain, n-aliphatic acid that possesses two more carbons than does caproic acid. Again, the average patterns of uptake across these six animals included the four fields of response described for the other acids (Fig. 8B, straight arrows). The uptake within Fields 1 and 2 was greater for octanoic acid than for caproic acid when presented at the same concentration (Fig. 8A,B), which extends the finding of increased uptake with increasing carbon length that was established for the other acids at higher concentrations (Fig. 3C). The balance of activity within Field 2 did not appear to shift in location relative to caproic acid, although it expanded along the rostral-caudal axis. The balance of activity within Field 3 continued the trend of shifting to a more ventral position with increasing carbon number. In addition to the four fields of response, there was evidence for two unique foci in rats exposed to octanoic acid, one of which was caudal and ventrolateral, and one of which was more caudal and ventromedial, thus being in relative positions predicted for paired glomeruli receiving projections from homologous olfactory sensory neurons (Fig. 8B, curved arrows). These foci were reliably present in rats exposed to octanoic acid, but were not seen in rats exposed to the other aliphatic acid odorants.
Spatial patterns of uptake are maintained in deeper bulb
layers
The apical dendrites of individual mitral and tufted projection neurons extend into a single glomerulus in the rat (Mori, 1987), which suggests that a simple spatial pattern of glomerular activity should be reflected by a similarly simple spatial pattern of projection neuron activity. To test this prediction, 2‑DG uptake was mapped in the external plexiform layer and the internal plexiform/superficial granule cell layer. The external plexiform layer contains dendrodendritic connections between basal dendrites of projection neurons and dendrites of inhibitory granule cells (Mori, 1987). The internal plexiform/superficial granule cell layer contains axon collaterals of projection neurons (Mori, 1987), and possibly axon termini from tufted cells on the other side of the bulb that form an interbulbar association system (Schoenfeld et al., 1985). Thus, activity in either layer is related to the activity of mitral/tufted cell projection neurons.
Contour charts of the uptake measured in the glomerular, external plexiform, and internal plexiform/superficial granule cell layers in a single bulb of a rat exposed to 25 parts per million valeric acid are presented in Figure 8C. The patterns of uptake measured in the deeper layers very closely mirrored the patterns in the glomerular layer (Fig. 8C). The four fields of activity were readily apparent, with the locations of maxima closely approximating the peaks of uptake in the glomerular layer (Fig. 8C). The uptake associated with Field 1 was more caudal in the internal plexiform/superficial granule cell layer than in the glomerular layer, which is to be expected given that the early sections containing the glomerular foci do not even contain an internal plexiform layer (Fig. 5). The uptake correlated with Field 2 in the internal plexiform/superficial granule cell layer appeared to be somewhat more dorsal than in the glomerular layer, and all fields of uptake were somewhat broader in the deeper layers (Fig. 8C), which is consistent with the approximately 1‑mm extent of the basal dendrites of projection neurons within the external plexiform layer (Mori, 1987). The levels of uptake in the deeper layers were lower than those in the glomerular layer under these exposure conditions.
To determine whether activity in deeper bulbar layers varied across different odorants, the 2‑DG uptake in the internal plexiform/superficial granule cell layer was mapped across single bulbs of the rats exposed to propionic, valeric, or caproic acids at 7.2 parts per million. As shown in Figure 8D, four fields of uptake in this deeper layer were observed for each of the aliphatic acid odorants. The amounts and locations of uptake within the four fields in the internal plexiform/superficial granule cell layer appeared to change across odorants, similar to what was observed for the glomerular layer of these same animals (compare Fig. 8D to Fig. 3A). To assess the statistical significance of these changes, fields in the deeper layer were delineated by averaging z score-transformed arrays and using a cut-off of 0.5, as was described for the analysis of glomerular layer uptake. Amounts of uptake and two coordinates of centroids were compared across odorants for each of the four fields. Results of ANOVAs indicated significant changes (P < 0.05, critical F2,15 = 3.68) in the rostral-caudal location of uptake within Field 1 (F2,15 = 10.94), in both the amount of uptake (F2,15 = 5.04) and the dorsal-ventral location of uptake (F2,15 = 4.27) within Field 2, in the rostral-caudal location of uptake within Field 3 (F2,15 = 9.00), and in both the rostral-caudal (F2,15 = 7.82) and dorsal-ventral (F2,15 = 20.04) locations of uptake within Field 4. Thus, the pattern of uptake differed across aliphatic acid odorants in the internal plexiform/superficial granule cell layer, where uptake is downstream of bulbar projection neurons.
Discussion
Distinct spatial representations of closely related
odorants
Analyses of 2‑DG uptake across the entire glomerular layer revealed that unbranched, n‑aliphatic acids differing by as little as one methylene group produced clustered, but systematically different patterns of activity in the rat olfactory bulb. The information comprising the patterns was multi-dimensional, involving different odorant-specific locations of active glomeruli, different odorant-specific numbers of activated glomeruli, and different odorant-specific amounts of activity. Changes in these parameters were found by analyzing uptake evoked within only four discrete glomerular regions that were distributed with respect to the rostral-caudal axis of the bulb and that were present in parallel in the lateral and medial aspects of the bulb. Our finding of two pairs of fields that contained reliable responses to even small aliphatic acids contrasts with previous descriptions of a major single focus of 2‑DG uptake or olfactory field potentials evoked by propionic acid (Slotnick et al., 1987; 1989; Mori et al., 1992), as well as with other parallel, distributed models suggesting that the entire bulb may be equipotential in the representation of even simple odorants (Kauer and Cinelli, 1993; Cinelli et al., 1995; Slotnick et al., 1997; Lu and Slotnick, 1998).
Clustering of glomeruli with related specificity
One way that spatial coding could be used in olfaction was proposed by Mori and coworkers to explain the fine-tuning of mitral/tufted cell responses to aliphatic acids and aldehydes of particular carbon chain lengths (Imamura et al., 1992). In their model, nearby glomeruli receive input from olfactory sensory neurons differing only slightly in their preference for a given odorant carbon chain length. Because mitral cells activated by neighboring glomeruli are capable of inhibiting one another through reciprocal synapses with shared inhibitory granule cell dendrites, the initially most active mitral cells could quiet their initially somewhat less active neighbors, resulting in a smaller range of effective stimuli for any given projection neuron. To test one prediction of this model, Mori and coworkers applied an antagonist of neurotransmitter receptors presumed to be involved in these inhibitory connections and observed that mitral/tufted cell projection neurons indeed exhibited broader tuning with respect to the carbon chain length of aliphatic aldehyde odorants (Yokoi et al., 1995).
The current data further support this model by demonstrating directly that aliphatic acid odorants of slightly different carbon number stimulate overlapping, but distinct parts of the glomerular layer. Progressions of centroids of activity towards more ventral and/or rostral locations with increasing carbon chain length were identified for all four fields of evoked 2‑DG uptake, suggesting the anatomical dimensions along which this lateral inhibitory tuning network might be laid out. Similar rostral progressions of centroids with increasing carbon number were observed in our previous study of aliphatic ester odorants for posterior fields that overlap with Fields 3 and 4 of the current report (Johnson et al., 1998). Our prior study also showed that glomeruli responding more specifically to isoamyl acetate may be located near glomeruli responding to both isoamyl acetate and isoamyl butyrate, and that glomeruli responding selectively to ethyl butyrate may be near glomeruli responding to both ethyl acetate and ethyl butyrate, which is further evidence for clustering of glomeruli with related specificities (Johnson et al., 1998). The area involved in foci of either 2‑DG uptake (Stewart et al., 1979) or c‑fos expression (Guthrie and Gall, 1995) increases with increasing odorant concentration. These findings are consistent either with a recruitment of neighboring glomeruli that receive projections from olfactory sensory neurons expressing related odorant receptors that differ in their affinities for the odorant or with an increased contribution of inhibitory interneurons whose axons extend to neighboring glomeruli. Clustering of glomeruli with specificities for chemically related amino acid odorants also has been observed through Ca2+-sensitive dye recordings of zebrafish olfactory bulbs, suggesting that the use of spatial relationships to achieve finer odorant tuning might be a mechanism common to many species (Friedrich and Korsching, 1997).
The olfactory receptor proteins expressed by sensory neurons appear to be involved in targeting the axons of the cells to the appropriate region of the mouse olfactory bulb (Singer et al., 1995; Mombaerts et al., 1996; Wang et al., 1998). If receptors with similar specificity also contain a high degree of amino acid sequence homology, and if similar sequences cause guidance to nearby glomerular locations, then axonal guidance by the receptor proteins would be an efficient means to construct clusters of similarly specific glomeruli. The reliability in the locations of uptake across different animals exposed to the same odorant, which must have been present to obtain statistically significant results in our analyses, implies that there is a profoundly rigorous set of parameters used to establish the topography of the projections of homologous sensory neurons.
The specificity of the fields of uptake
With the possible exception of Fields 3 and 4 for acetic acid, all of the aliphatic acids we studied evoked uptake in the same four fields. Caproic acid at the same concentration used for the other aliphatic acids also evoked activity outside of these fields, but the peaks of activity were clearly localized within the fields, and responses to a lower concentration of the same odorant were again confined to the four fields. (We currently are conducting a more detailed study of the effects of odorant concentration on various dimensions of 2‑DG uptake using a larger range of odorants at more concentration steps.) Octanoic acid evoked activity in the four fields, but also stimulated uptake in locations not seen for the other acids, a finding similar to that we obtained for isoamyl butyrate, a nine-carbon aliphatic ester (Johnson et al., 1998).
In the absence of data on compounds differing in additional functional groups and other molecular features, the minimal determinants for 2‑DG uptake in the four fields remains unclear. Mori and coworkers recorded mitral/tufted cell responses to aliphatic acids, aldehydes, ketones, and methyl or ethyl esters within the dorsomedial region of the rabbit olfactory bulb (corresponding to Field 2 of the current study), and they suggested that the minimal determinants may have been related to the carbonyl group shared by these compounds (Imamura et al., 1992). It is possible that further studies with additional odorants will reveal a similar specificity for Fields 1 and 2. Indeed, in our prior study (Johnson et al., 1998), ethyl acetate and ethyl butyrate (75 parts per million) evoked uptake in dorsolateral and dorsomedial fields that may overlap with Fields 1 and 2 of the current study. Aliphatic aldehydes apparently also stimulate mitral/tufted cells in a more ventral position where responses were not observed for aliphatic acids, an observation that suggests a means of discriminating between different molecules sharing the carbonyl group (Imamura et al., 1992).
Our responses in Fields 3 and 4 overlap extensively with fields evoked by a variety of aliphatic esters (Stewart et al., 1979; Royet et al., 1987; Johnson et al., 1998 and unpublished results) and components of peppermint extract (Johnson and Leon, 1997 and unpublished results). Others have reported responses to camphor (Stewart et al., 1979), ethylacetoacetate (Jourdan et al., 1980), cyclohexanone (Jourdan et al., 1980), and limonene (Bell et al., 1987) in the same general regions. In our study of aliphatic esters, we suggested that responses in these areas may represent the activity of receptors responding to some chemical attribute shared by many odorants (e.g., odorant hydrophobicity or size) (Johnson et al., 1998). The relative absence of responses to acetic acid in these regions would be consistent with this hypothesis, given its small, hydrophilic nature. The observation that centroids of uptake within these regions vary with odorant size for both aliphatic acids and aliphatic esters (Johnson et al., 1998) suggests that the regions may contribute to odorant quality coding despite the extensive overlap between individual odorants.
The spatial overlap in responses at the level of the glomerular layer that is observed even for odorants that differ greatly in certain molecular features suggests the possibility that even chemically dissimilar odorants might be confused on a perceptual level. Humans show surprisingly poor initial performance during odorant identification tasks, correctly identifying only about half of the common, familiar odorants presented to them (Cain, 1979; Cain and Potts, 1996). Misidentifications often involve related odorants, but also include dramatic mistakes involving very different odorants (Cain, 1979; Cain and Potts, 1996). This poor initial performance appears to be related, at least in part, to perceptual confusion (Cain and Potts, 1996). Odorant identification becomes virtually perfect with only a limited amount of feedback regarding the correct odorant identifications (Cain, 1979), which suggests that humans might learn to focus on neural information germaine to the correct odorant discrimination or odorant label. Following training, humans can discriminate between very closely related aliphatic acids, although they have greater difficulties when these compounds differ by single steps of carbon-chain length or single changes in branch structure (Laska and Teubner, 1998). Rats can be trained to discriminate between propionic acid and acetic acid, which differ by a single step in carbon-chain length and which are perceived to be very similar by humans (Lu and Slotnick, 1994; 1998; Slotnick et al., 1997). However, it is possible that rats would confuse these odorants (or even more dissimilar ones) without prior discrimination training. Indeed, there is evidence that rats will generalize between a conditioned aliphatic aldehyde odorant and another aldehyde differing by a single carbon (Linster et al., 1998).
Parallel patterns of uptake in the lateral and medial bulb
In agreement with our prior study on aliphatic esters, the current study has revealed that for every identified field of response in the lateral glomerular layer, there is a field in the medial glomerular layer that possesses a similar odorant specificity. These lateral-medial pairs of fields also existed in deeper bulbar layers. The anatomical relationship between these paired fields of activity mirrors both the projection patterns of homologous sensory neurons (Ressler et al., 1994; Vassar et al., 1994; Mombaerts et al., 1996) and the intrabulbar tufted cell associational system wherein tufted cells on one side of the bulb send axons to restricted domains of granule cells on the other side of the bulb (Schoenfeld et al., 1985).
The reason for two maps of odorant chemistry in a single olfactory bulb is not yet known. One possibility is simple redundancy. The paired representations may have arisen to insure one intact pattern following injury or mild nasal infections affecting part of the olfactory epithelium. Another possibility is that the two representations allow coincidence detection to filter out activity in mitral cells that is unrelated to odorant exposure. If similarly odorant-specific mitral cells in the lateral and medial regions of the bulb converge in their projections onto targets in the forebrain, their joint activation may more faithfully convey the presence of an odorant feature than would activity in either bulb region alone. The inhibitory intrabulbar associational system (Schoenfeld et al., 1985) also may be incorporated into this coincidence detection hypothesis. Finally, in lower vertebrates, there are separate lateral and medial olfactory tracts providing connections of the lateral and medial bulb to distinct forebrain targets, and it remains possible that separate projection tracts also are present in mammals (Eisthen, 1997).
Relationships of the current results to prior lesion
studies
The parallel representations of odorant features in the lateral and medial bulb and the distribution of fields in the anterior and posterior aspects of the bulb provide possible explanations for the spared functions reported in studies involving both smaller lesions of a focus evoked by propionic acid and extensive lesions of the anterior bulb. Lesions of the dorsomedial bulb containing foci of 2‑DG uptake and c-fos expression in response to propionic acid (corresponding to our Field 2) did not change the threshold concentration of the odorant necessary for learning a behavioral task (Slotnick et al., 1987; 1997). Nor did such lesions block the recognition of propionic acid (Lu and Slotnick, 1994). Even odorants as similar as propionic and acetic acids still could be discriminated following the lesions (Lu and Slotnick, 1994; 1998; Slotnick et al., 1997). These lesions would have spared the lateral field of similar specificity (Field 1), as well as two other fields responding to propionic acid (Fields 3 and 4). Thus, the results of these relatively discrete lesion studies may simply indicate that the limited redundancy of the representation of propionic acid in the bulb was adequate to support threshold detection and discrimination. It should be stressed that the 2‑DG uptake evoked by propionic acid in Fields 1, 3, and 4 was not diffuse, sporadic, and minor, but rather was focal, reliable across animals, similar in magnitude to the previously described dorsomedial foci within Field 2, and different from that evoked by other aliphatic acids.
Even following larger anterior lesions involving up to 80% of the bulb, including all of the lateral aspect, rats were able to discriminate propionic acid from acetic acid, as well as cineole from propionic acid, amyl acetate, citral, and butanol (Lu and Slotnick, 1998). Given the extremely posterior location of the responses within Field 4 of our current study, it is probable that this field would have been spared in all of these lesion studies. Numerous other odorants appear to stimulate appreciable, focal 2‑DG uptake within the posterior portions of the medial glomerular layer that overlap with Field 4 (see above). If the attribute responsible for responses in this region were distinct for most odorants, and if it resulted in even the subtle spatial differences or differences in the amounts of activity within this posterior field such as those observed in the current study, then one might predict that the posterior, medial bulb indeed could discriminate between the odorants so far tested. Thus, a special involvement of the posteromedial bulb in odor detection and discrimination may explain the negative results of the prior lesion studies. Consistent with a special importance of the posterior, medial bulb is the observation that large lesions sparing this region result in very few deficits in detection and discrimination (Lu and Slotnick, 1998). In rats with lesions that involve this region, deficits are observed, and olfactory performance is positively correlated with the number of remaining glomeruli (Lu and Slotnick, 1998).
Given the large differences in odorant-evoked 2‑DG uptake within foci as compared to the glomerular layer background, it remains reasonable to suggest that the foci have a special significance for the coding of odor quality. If this suggestion is true, it would be predicted that lesions of lateral-medial paired fields of similar specificity would change the quality of the odor, even if it could still be discriminated from other randomly chosen odorants. Large, anterior lesions presumably would remove paired fields, but the possibility of altered perception of odorants has been addressed directly only in studies involving either lateral or medial lesions (Lu and Slotnick, 1994). It does appear, however, that rats previously trained to perform a specific task when propionic acid, amyl acetate, or citral are present need to re-learn associations with these odors following large, anterior bulbar lesions (Lu and Slotnick, 1998). Although interpreted as a retention deficit (Lu and Slotnick, 1998), this observation also is consistent with an altered perception of the previously learned odor after the lesion.
Relatively discrete lesions of the bulb involving either Field 1 or Field 2 of the current study do lead to impairments in the discriminations of mixtures of propionic and acetic acids (Slotnick et al., 1997). Perhaps these more difficult odorant discrimination tasks require all four fields of response.
Relationships of the current results to reports of
low-specificity,
diffusely distributed mitral cell activity
The work of MoriŐs group indicated a well-tuned response of mitral/tufted cells within the dorsomedial olfactory bulb to aliphatic acids and aldehydes in contrast to a low responsiveness of projection neurons in other regions of the rabbit olfactory bulb (Mori et al., 1992; Imamura et al., 1992). Thus, mitral cells in relative proximity to one another were found to have related odorant specificities that could be expressed in terms of the intensity of response to compounds differing systematically in chemical structure (Mori et al., 1992; Imamura et al., 1992; Katoh et al., 1993). These results were in agreement with a study by Buonviso and Chaput (1990) showing that neighboring projection neurons in the rat tended to have more closely related odorant specificities than did projection neurons located further apart. Recently, however, it has been reported that mitral cells distributed widely across the rat olfactory bulb respond with various temporal patterns of excitation and suppression to the majority of odorants tested (Motokizawa, 1996). The same study indicated that there was no greater similarity in response profiles in mitral cells located nearby than in mitral cells located far apart (Motokizawa, 1996). It was argued that the result of Mori and coworkers was likely either an artifact of the depth of anesthesia or an indication of a species difference (Motokizawa, 1996).
By mapping 2‑DG uptake in the external plexiform layer and in the internal plexiform/superficial granule cell layers, we have found spatial representations of aliphatic acid odorants that contain focal areas of response mirroring those detected in the glomerular layer. This activity, which must be secondary to the activation of projection neurons, indicates a distinctive response in the same dorsomedial part of the bulb (Field 2) where Mori et al. found specific responses to aliphatic acids (Mori et al., 1992; Imamura et al., 1992). Furthermore, the 2-DG uptake in the internal plexiform/superficial granule cell layer differed systematically in location across aliphatic acid odorants differing in carbon-chain length, which likely indicates an orderly spatial arrangement of projection neurons exhibiting different optimal specificities. The relative spatial distributions of internal plexiform/superficial granule cell layer responses to propionic acid and caproic acid in Field 2 are consistent with those illustrated by Mori et al. in their electrophysiological studies of mitral/tufted cells in the dorsomedial region of the rabbit olfactory bulb (Mori et al., 1992). Our results arose from studies of rats that were not subjected to anesthesia, and yet we corroborate the results that MoriŐs group obtained using anesthetized rabbits by this entirely independent method. For this reason, it appears very unlikely that the electrophysiological results of Mori and coworkers are artifactual.
In their analysis, Mori and coworkers had guidance from both prior 2‑DG studies of propionic acid-evoked activity and field potential recordings of activity evoked by propionic and caproic acids, and they chose their sites for recording on the basis of the largest responses (Mori et al., 1992). They then pursued systematic studies related to small changes in odorant chemistry. In our analyses, we mapped uptake across the entire bulb in rats exposed to systematically different chemical odorants, and therefore we could easily identify the largest responses. In the analysis of Motokizawa, electrodes were lowered at apparently random locations, and odorants of widely different chemistries were used (Motokizawa, 1996). Thus, no spatially specific selection was made for largest responses, and it is perhaps not surprising that the mitral cell responses usually failed to discriminate between even large differences in odorant chemistry. It seems likely that the primary responses to the odorants were not recorded in that study.
It remains possible that the depth of anesthesia assisted Mori and coworkers in uncovering the largest responses to aliphatic acid odorants. It also is possible that the 2‑DG method exaggerates the highest level of activity. In combination, however, our results together with those of Mori and coworkers indicate that these higher levels of activity can distinguish even small differences in the chemical structure of related odorants, whereas the diffuse and nonspecific responses found elsewhere in the bulb by Motokizawa may be largely without information content and may not distinguish between even chemically unrelated odorants. Thus, all of these results actually lend more support for models involving a special significance of the locations of large responses in the coding of odorant quality than they do for a widely distributed model in which most of the bulb is involved in coding a given odorant.
The question that remains is how the brain filters out the small, diffuse, poorly linked mitral and tufted cell responses from the spatially restricted ones that contain more information. Because this question has not yet been examined directly, there is little data to support a particular hypothesis. Possibilities include a selection for the most intense volleys of projection neuron activity, a selection for impulses linked in time to the early phase of respirations (e.g., by using the phase relationships between sniffing and theta rhythms in olfactory structures (Macrides et al., 1982)), and/or a selection for impulses synchronized either with respect to homologous portions of the lateral and medial aspects of a given bulb (as discussed above) or with respect to homologous portions of the left and right olfactory bulbs. The aforementioned lesion studies cast doubt on the necessity for combined lateral and medial activity (or interbulbar synchronization), although such processes may participate in an intact animal.
The specificity of individual sensory neurons
Given the convergence of projections into the bulb, glomerular activity should reflect homologous sensory neuron activity. Most of our data suggests that this activity possesses a definite selectivity for molecular features of odorants. How, then, can one explain the large number of olfactory sensory neurons that respond to a given odorant, and the large fraction of tested odorants capable of stimulating a given salamander olfactory sensory neuron in vitro (Firestein et al., 1993)? One possible answer may be similar to that for the low specificity mitral/tufted cell responses detected in the olfactory bulb. The study showing broad specificity of sensory neurons used no preselection for cells giving the largest responses. Indeed, in studies that did preselect for mouse olfactory receptor neurons located in a portion of the septum of the epithelium, Sato et al. (1994) found good tuning with respect to the carbon chain length of aliphatic acids and alcohols, especially when the odorants were presented at low concentrations. Similarly, in a study that preselected for mouse sensory neurons sending axons to glomeruli of the dorsomedial olfactory bulb, where responses to aliphatic acids are observed, Bozza and Kauer (1998) found that individual sensory neurons responded to aliphatic acids but not to alcohols of a similar carbon-chain length. Enrichment of a particular rat olfactory receptor gene also resulted in sensory neuron responses that both were very finely tuned with respect to small changes in odorant chemistry and were detectable at low odorant concentrations (Zhao et al., 1998). It also is possible that there is a species difference between salamanders and rodents in the breadth of tuning of individual olfactory sensory neurons for different odorant molecules that could account for the reported differences in response profiles (Sato et al., 1994).
Despite the evidence that individual sensory neurons, glomeruli, and mitral cells are selective for odorants within a particular range of chemical structure, crude counts of glomeruli underlying foci of uptake evoked by high concentrations of valeric acid in the current study would suggest that up to 5% of all glomeruli in the bulb may respond to this odorant to some extent (data not shown), despite the restriction of these glomeruli to four defined fields. High concentrations of caproic acid apparently can stimulate an even greater fraction of the glomeruli (one-third to one-half), which is similar to what we observed previously for relatively high concentrations of isoamyl butyrate (Johnson et al., 1998).
The increases both in the number of active glomeruli and in the average uptake across a given functional field with increasing odorant molecular size that were revealed in the current study are reminiscent of numerous other studies showing increased amplitudes of responses with increases in odorant size and/or hydrophobicity. For example, Sato et al. (1994) found that the number of responsive mouse olfactory sensory neurons increases with increasing carbon number of n-aliphatic acids and alcohols. Ottoson (1958) found that increasing odorant hydrophobicity correlates with an increased response amplitude in frog epithelium as measured using electro-olfactography. Finally, Cometto-Mu–iz et al. (1998) observed an inverse correlation between the carbon-chain length of n-aliphatic acids or aldehydes and the concentrations of these odorants necessary for threshold odor detection in humans.
A simple model of odorant feature processing
The simple spatial patterns of odorant-evoked activity observed when 2‑DG uptake is surveyed across the entire glomerular layer are consistent with previously discussed models of distributed, molecular feature processing of odorants in the rat olfactory bulb (Imamura et al., 1992; Kauer and Cinelli, 1993; Shepherd, 1994; Axel, 1995; Mori and Yoshihara, 1995; Buck, 1996), and our results emphasize the possible use of spatial coding in olfaction. In this view, the spatial convergence of homologous olfactory sensory neuron projections into individual glomeruli may increase the sensitivity at the glomerulus and amplify the response. Glomeruli responding to similar molecular features appear to be clustered together, creating functional fields that can be revealed by studying 2‑DG uptake. Within a given cluster, glomeruli are distributed in such a way that nearest neighbors exhibit the slightest differences in specificity, perhaps related to odorant size or hydrophobicity. This rigorous spatial arrangement insures that lateral inhibition tunes mitral and tufted cells to specific molecular attributes that would be difficult to distinguish without tuning. Thus, the bulb may be viewed as an array of functional fields that serve to amplify, spatially segregate, and separately tune responses to molecular features of odorants.
It should be noted that the odorants studied here were relatively simple chemicals. Odors produced by mixtures of dissimilar chemical odorants or by monomolecular odorants with either a large number of molecular features or a large number of conformations may have more complex spatial representations in the bulb. The coding of such stimuli could involve much of the bulb surface, as seen for high concentrations of caproic acid in the present report and for isoamyl butyrate in our previous study (Johnson et al., 1998). The odor quality of these more complex stimuli may be affected only slightly by lesions involving small parts of the representation.
A map of the chemical specificity of the olfactory bulb
Our findings of well-defined fields of 2‑DG uptake evoked by aliphatic acids further indicate the possibility of rendering an extensive map of odorant chemistry across the glomerular layer of the olfactory bulb (Johnson et al., 1998). Although such an undertaking will be laborious with respect to the number of odorants that will need to be surveyed, it should further elaborate the fundamentals of odor processing by the bulb. Combined with oriented lesion studies and correlations with psychophysical measures of odorant quality, such studies should prove helpful in the design of specific experiments to determine how other regions of the brain decode the chemical feature information that may be segregated and tuned within the olfactory bulb.
Literature Cited
Axel, R. (1995) The molecular logic of smell. Sci. Am. 273:154-159.
Bell, G.A., D.G. Laing, and H. Panhuber (1987) Odour mixture suppression: Evidence for a peripheral mechanism in human and rat. Brain Res. 426:8-18.
Bozza, T.C. and J.S. Kauer (1998) Odorant response properties of convergent olfactory receptor neurons. J. Neurosci. 18:4560-4569.
Buck, L.B. (1996) Information coding in the vertebrate olfactory system. Annu. Rev. Neurosci. 19:517-544.
Buck, L.B. and R. Axel (1991) A novel multigene family may encode odorant receptors: A molecular basis for odor recognition. Cell 65:175-187.
Buonviso, N. and M.A. Chaput (1990) 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.
Cain, W.S. (1979) To know with the nose: Keys to odor identification. Science 203:467-470.
Cain, W.S. and B.C. Potts (1996) Switch and bait: Probing the discriminative basis of odor identification via recognition memory. Chem. Senses 21:35-44.
Cinelli, A.R., K.A. Hamilton, and J.S. Kauer (1995) Salamander olfactory bulb neuronal activity observed by video rate, voltage-sensitive dye imaging. III. Spatial and temporal properties of responses evoked by odorant stimulation. J. Neurophysiol. 73:2053-2071.
Cometto-Mu–iz, J.E., W.S. Cain, and M.H. Abraham (1998) Nasal pungency and odor of homologous aldehydes and carboxylic acids. Exp. Brain Res. 118:180-188.
Dšving, K.B. (1966) An electrophysiological study of odour similarities of homologous substances. J. Physiol. 186:97-109.
Eisthen, H.L. (1997) Evolution of vertebrate olfactory systems. Brain Behav. Evol. 50:222-233.
Firestein, S., C. Picco, and A. Menini (1993) The relation between stimulus and response in olfactory receptor cells of the tiger salamander. J. Physiol. 468:1-10.
Friedrich, R.W. and S.I. Korsching (1997) Combinatorial and chemotopic odorant coding in the zebrafish olfactory bulb visualized by optical imaging. Neuron 18:737-752.
Guthrie, K.M., A.J. Anderson, M. Leon, and C. Gall (1993) Odor-induced increases in c-fos mRNA expression reveal an anatomical ŇunitÓ for odor processing in olfactory bulb. Proc. Natl. Acad. Sci. USA 90:3329-3333.
Guthrie, K.M. and C. Gall (1995) Functional mapping of odor-activated neurons in the olfactory bulb. Chem. Senses 20:271-282.
Hass, H.B. and R.F. Newton (1975) Correction of boiling points to standard pressure. In R.C. Weast (ed): Handbook of Chemistry and Physics. 56th ed. Cleveland, Ohio: CRC Press, pp. D176-D177.
Imamura, K., N. Mataga, and K. Mori (1992) Coding of odor molecules by mitral/tufted cells in rabbit olfactory bulb. I. Aliphatic compounds. J. Neurophysiol. 68:1986-2002.
Johnson, B.A. and M. Leon (1996) Spatial distribution of [14C]2-deoxyglucose uptake in the glomerular layer of the rat olfactory bulb following early olfactory preference learning. J. Comp. Neurol. 376:557-566.
Johnson, B.A., C.C. Woo, and M. Leon (1998) Spatial coding of odorant features in the glomerular layer of the rat olfactory bulb. J. Comp. Neurol. 393:457-471.
Jourdan, F., A. Duveau, L. Astic, and A. Holley (1980) Spatial distribution of [14C]2-deoxyglucose uptake in the olfactory bulbs of rats stimulated with two different odours. Brain Res. 188:139-154.
Katoh, K., H. Koshimoto, A. Tani, and K. Mori (1993) Coding of odor molecules by mitral/tufted cells in rabbit olfactory bulb. II. Aromatic compounds. J. Neurophysiol. 70:2161-2175.
Kauer, J.S. and A.R. Cinelli (1993) Are there structural and functional modules in the vertebrate olfactory bulb? Micr. Res. Tech. 24:157-167.
Kent, P.F. and M.M. Mozell (1992) The recording of odorant-induced mucosal activity patterns with a voltage-sensitive dye. J. Neurophysiol. 68:1804-1819.
Land, L.J. (1973) Localized projection of olfactory nerves to rabbit olfactory bulb. Brain Res. 63:153-166.
Laska, M. and P. Teubner (1998) Odor structure-activity relationships of carboxylic acids correspond between squirrel monkeys and humans. Am. J. Physiol. 274:R1639-R1645.
Linster, C., E. De Rosa, and M.E. Hasselmo (1998) Relation between olfactory bulb activity and behavioral responses to n-aliphatic aldehydes. Soc. Neurosci. Abs. 24:908.
Lu, X.-C. M. and B.M. Slotnick (1994) Recognition of propionic acid vapor after removal of the olfactory bulb area associated with high 2-DG uptake. Brain Res. 639:26-32.
Lu, X.-C. M. and B.M. Slotnick (1998) Olfaction in rats with extensive lesions of the olfactory bulbs: Implications for odor coding. Neuroscience 84:849-866.
Macrides, F., H.B. Eichenbaum, and W.B. Forbes (1982) Temporal relationship between sniffing and the limbic q rhythm during odor discrimination reversal learning. J. Neurosci. 2:1705-1717.
Mombaerts, P., F. Wang, C. Dulac, S.K. Chao, A. Nemes, M. Mendelsohn, J. Edmonson, and R. Axel (1996) Visualizing an olfactory sensory map. Cell 87:675-686.
Mori, K. (1987) Membrane and synaptic properties of identified neurons in the olfactory bulb. Progr. Neurobiol. 29:275-320.
Mori, K. and Y. Yoshihara (1995) Molecular recognition and olfactory processing in the mammalian olfactory system. Progr. Neurobiol. 45:585-619.
Mori, K., N. Mataga, and K. Imamura (1992) Differential specificities of single mitral cells in rabbit olfactory bulb for a homologous series of fatty acid odor molecules. J. Neurophysiol. 67:786-789.
Motokizawa, F. (1996) Odor representation and discrimination in mitral/tufted cells of the rat olfactory bulb. Exp. Brain Res. 112:24-34.
Onoda, N. (1992) Odor-induced fos-like immunoreactivity in the rat olfactory bulb. Neurosci. Lett. 137:157-160.
Ottoson, D. (1958) Studies on the relationship between olfactory stimulating effectiveness and physicochemical properties of odorant compounds. Acta Physiol. Scand. 43:167-181.
Ressler, K.J., S.L. Sullivan, and L.B. Buck (1994) Information coding in the olfactory system: Evidence for a stereotyped and highly organized epitope map in the olfactory bulb. Cell 79:1245-1255.
Royet, J.P., G. Sicard, C. Souchier, and F. Jourdan (1987) Specificity of spatial patterns of glomerular activation in the mouse olfactory bulb: Computer-assisted image analysis of 2-deoxyglucose autoradiograms. Brain Res. 417:1-11.
Sallaz, M. and F. Jourdan (1992) Apomorphine disrupts odour-induced patterns of glomerular activation in the olfactory bulb. NeuroReport 3:833-836.
Sallaz, M. and F. Jourdan (1993) C-fos expression and 2-deoxyglucose uptake in the olfactory bulb of odour-stimulated awake rats. NeuroReport 4:55-58.
Sato, T., J. Hirono, M. Tonoike, and M. Takebayashi (1994) Tuning specificities to aliphatic odorants in mouse olfactory receptor neurons and their local distribution. J. Neurophysiol. 72:2980-2989.
Schoenfeld, T.A., J.E. Marchand, and F. Macrides (1985) Topographic organization of tufted cell axonal projections in the hamster main olfactory bulb: An intrabulbar associational system. J. Comp. Neurol. 235:503-518.
Schwob, J.E. and D.I. Gottlieb (1986) The primary olfactory projection has two chemically distinct zones. J. Neurosci. 6:3393-3404.
Sharp, F.R., J.S. Kauer, and G.M. Shepherd (1977) Laminar analysis of 2-deoxyglucose uptake in olfactory bulb and olfactory cortex of rabbit and rat. J. Neurophysiol. 40:800-813.
Shepherd, G.M. (1994) Discrimination of molecular signals by the olfactory receptor neuron. Neuron 13:771-790.
Singer, M.S., G.M. Shepherd, and C.A. Greer (1995) Olfactory receptors guide axons. Nature 377:19-20.
Slotnick, B.M., S. Graham, D.G. Laing, and G.A. Bell (1987) Detection of propionic acid vapor by rats with lesions of olfactory bulb areas associated with high 2-DG uptake. Brain Res. 417:343-346.
Slotnick, B.M., H. Panhuber, G.A. Bell, and D.G. Laing (1989) Odor-induced metabolic activity in the olfactory bulb of rats trained to detect propionic acid vapor. Brain Res. 500:161-168.
Slotnick, B.M., G.A. Bell, H. Panhuber, and D.G. Laing (1997) Detection and discrimination of propionic acid after removal of its 2-DG identified major focus in the olfactory bulb: A psychophysical analysis. Brain Res. 762:89-96.
Stewart, W.B., J.S. Kauer, and G.M. Shepherd (1979) Functional organization of rat olfactory bulb analysed by the 2-deoxyglucose method. J. Comp. Neurol. 185:715-734.
Vassar, R., S.K. Chao, R. Sitcheran, J.M. Nu–ez, L.B. Vosshall, and R. Axel (1994) Topographic organization of sensory projections to the olfactory bulb. Cell 79:981-991.
Wang, F., A. Nemes, M. Mendelsohn, and R. Axel (1998) Odorant receptors govern the formation of a precise topographic map. Cell 93:47-60.
Yokoi, M., K. Mori, and S. Nakanishi (1995) Refinement of odor molecule tuning by dendrodendritic synaptic inhibition in the olfactory bulb. Proc. Natl. Acad. Sci. USA 92:3371-3375.
Zhao, H., L. Ivic, J.M. Otaki, M. Hashimoto, K. Mikoshiba, and S. Firestein (1998) Functional expression of a mammalian odorant receptor. Science 279:237-242.
Table 1. Odorants used in the study
Number Dilution of saturated Dilution to give
Odorant of carbons Structure vapor to give 7.2 ppm 0.8 ppm
Acetic acid 2 CH3COOH 1/2225 --1
Propionic acid 3 CH3CH2COOH