This is a preprint of a chapterpublished in Chemical Signals in Vertebrates IX (Marchlewska-Koj, A., Lepri,J.J., and Muller-Schwarze, D, eds.) pp. 85-91, Kluwer Academic/PlenumPublishers, New York.

 

 

 

 

 

 

 

Spatialrepresentations of odorant chemistry in

themain olfactory bulb of the rat

 

 

 

Brett A. Johnson and Michael Leon

 

Department of Neurobiology and Behavior,

University of California, 2205 BioSci II,

Irvine, CA, USA 92697-4550

 

 

 

1. Systematicstudies of odorant-evoked spatial activity patterns in the olfactory bulb

 

        Odorantsof very different chemical structure that yield the perception of very differ­entodors have long been known to evoke distinct patterns of neural activity in therat olfactory bulb. This observation suggested that at least one step in thecoding of odor information might involve spatial patterns of bulbar activity.Until recently, however, it was not understood how different spatial patternsmight arise from the differences in odorant chemical structure.

        Ourgeneral approach to odor coding in the rat olfactory bulb has involved lookingfor systematic changes in spatial activity patterns that occur with smallchanges in odorant chemistry. Through this approach, elements of the spatialpatterns can be related to particular molecular features and/or chemicalproperties of the odorant stimuli. To compare the patterns of activity evokedby different odorants systematically, we have developed a proce­dure to mapuptake of [14C]2-deoxyglucose (2-DG) across the entire glomerularlayer (Johnson et al.,1999). By using radial grids to dictate positions of measurements in coronalsections, and by equalizing rostral-caudal positions in relation to anatomicallandmarks, we generate standardized arrays of data that allow patterns to becompared quantitatively across multiple animals exposed to distinctodorants. 

        Wereview here our findings concerning oxygen-containing, aliphatic odorants thatdiffer in carbon number, functional groups, and hydrocarbon structure. Takentogether, our results demonstrate several fundamental principles underlyingspatial representations of odorants in the olfactory bulb. Each of theseprinciples will be discussed in turn below.

 

 

2. Modular representations of odorant chemistry

 

        Figure1 shows glomerular activity patterns averaged across three to six rats exposedto each of 13 distinct odorants chosen to illustrate results we have obtainedby using a larger number of compounds (Johnson et al., 1998, 1999; Johnson and Leon, 2000a, b). The pat­ternsshown involve straight-chained aliphatic acids differing incrementally incarbon number (top row), five-carbon aliphatic acids differing in hydrocarbonstructure (middle row), and aliphatic compounds sharing a four-carbon,straight-chained portion, but differ­ing in the functional group present atthe fifth carbon (bottom row). Each row represents a distinct experiment.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.  Average activity patterns evoked byodorants differing in carbon number (7.2 ppm, top row), hydrocarbon structure(8 ppm, middle row), and oxygen-containing functional groups (bottom row,lowest concentrations evoking a pattern as indicated in Fig. 2). Structures areshown as ball-and-stick diagrams. Circles represent carbon atoms. Hydrocarbonhydrogen atoms are omitted. Rostral is to the left. All other orientation is asshown in upper right inset (dor, dorsal; lat, lateral; vent, ventral; med,medial). Bottom right inset shows a selection of glomerular modules activatedby these odorants. Large arrows indicate modules a and A. Small arrows indicatemodules e and E. Arrows appear in the same location on each chart.

        Asis readily apparent in Figure 1, even very closely related odorant chemicalsevoked distinct, but overlapping patterns of glomerular activity. There are twotypes of differences among the odorant-evoked patterns. One type of differenceinvolves the stimulation by one odorant of a part of the bulb (typically acluster of adjacent glomeruli, which we will refer to as a glomerular module)that is not stimulated by another odorant. For example, at the lowestconcentrations evoking patterns, valeric acid, methyl valerate, and pentanolactivated rostral modules that were not activated by 2-hex­anone orpentanal (Fig. 1, bottom row, large arrows). Another type of differenceinvolves small changes in the position of overlapping modules. For example,rostral mod­ules shifted in position with increasing carbon number ofstraight-chained aliphatic acid odorants (Fig. 1, top row, large arrows). Thisdifference will be discussed below in the Òchemotopic organizationÓ section.

        Inour functional group study, higher concentrations of some of the odorantsactivated glomerular modules that were not activated at lower concentrations(Johnson and Leon, 2000a). The full set of modules in that study is diagrammedin Figure 2, which also shows the relative amounts of 2-DG uptake evoked ineach module for each concentration of each odorant. Each module was stimulatedby at least two of the five odorants in the study. Therefore, the coding of theodors of these compounds must be combinatorial. That is, a comparison of therelative levels of activity in various modules is needed to identify a givenodorant as unique. Also, each odorant evoked activity in multiple modules (Fig.2). As shown in Figure 1, even simple molecules such as propionic acidstimulated uptake in at least four modules distributed across both lateral andmedial aspects of the bulb (Johnson et al.,1999). Two of the modules were located in the rostral part of the bulb, and atleast two in the caudal part of the bulb. This parallel and spatiallydistributed representation of propionic acid likely explains its continueddetection by rats with large experimental ablations of the olfactory bulb(Slotnick et al., 1997).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.2-DG uptake in glomerular modules as a function of the concentration ofodorants differing in oxygen-containing functional groups. Locations of modulesare shown in the left panel. Diameters of circles (right panel) indicate uptakestandardized for each odorant to the largest value obtained at anyconcentration. Asterisks denote the lowest concentrations judged to evoke apattern (Johnson and Leon, 2000a).

        Thespecificity of a given glomerulus probably reflects the specificity of anindividual odorant receptor protein, because individual glomeruli appear toreceive convergent pro­jections from sensory neurons that express the same,single odorant receptor gene (Ressler et al., 1994). Because most receptor proteins actually detect molecular fea­turespresent in multiple ligand molecules, the activation of an individualglomerulus likely indicates the presence of a particular odorant molecularfeature (Johnson et al., 1998).By comparing the odorants we have investigated so far, we can begin tohypothesize the molecular features that are necessary for the stimulation ofsome of the modules we have identified. For example, modules a and A (Fig. 2)responded to low concentrations of aliphatic acids largely independently ofhydrocarbon structure or carbon number (Fig. 1, top and middle rows, largearrows). Much higher concentrations of pentanol were needed to activate thesemodules, and even higher concentrations of methyl valerate and pentanal wererequired (Fig. 2). 2-Hexanone did not activate these modules. They wereactivated by ethyl esters, but not by isoamyl esters (Johnson et al., 1998). These findings are consistent with therecognition of a hydrogen bond acceptor by the receptors associated withmodules a and A. Modules e and E were activated by all compounds that shared afour-carbon, straight-chained hydrocarbon structure in the functional groupstudy (Fig. 2). However, 2-methyl­butyric acid stimulated these modulesrobustly, while the very closely related structural isomer, isovaleric(3-methylbutyric) acid, caused only low activity in these modules whenpresented at the same concentration (Fig. 1, middle row, small arrows). Thisdifference suggests that modules e and E may recognize some steric featurepresent in many aliphatic odorant molecules (i.e., a specific geometricarrangement of hydrocarbon hydrogens) (Johnson and Leon, 2000b).

        Byidentifying glomerular modules activated by one odorant that are not activatedby other odorants differing only slightly in chemical structure, we thereforecan generate specific hypotheses concerning the specificity of individualmodules. These hypotheses then can be tested through the use of another odorantset. This systematic approach to odor coding should accelerate ourunderstanding of what odorant chemical features are com­partmentalizedwithin the olfactory bulb. Furthermore, by studying a wider range of func­tionalgroups and hydrocarbon structures, we should be able to approach a morecomplete stimulus map of bulbar activity.

       

 

3. Chemotopic organization of glomeruli withinmodules

 

        Mostof the modules we have identified are comprised of multiple, adjacentglomeruli. For example, at a concentration of 7.2 parts per million, valericacid stimulated about 30 glomeruli each within modules a and A (Johnson etal., 1999). The 2-DGtechnique is capable of resolving the activation of as few as one or twoglomeruli (Johnson et al.,1998). Therefore, the activation of larger numbers of adjacent glomeruli by asingle odorant likely indicates that these glomeruli are all activated directlyby the odorant, and that the specificities of these glomeruli may be closelyrelated.

        Toexplain tuning of individual bulbar projection neurons to aliphatic acids of apar­ticular carbon number, Mori and coworkers suggested that adjacentglomeruli in the rostral, dorsomedial bulb may respond optimally, but broadly,to acids of slightly distinct carbon number (Yokoi et al., 1995). Lateral inhibition betweenneighboring glomeruli or between projection neurons via granule cellinterneurons was proposed to accomplish the observed tuning. To test thishypothesis, we analyzed centroids of 2-DG uptake within module A. The centroidsdiffered significantly across different straight-chained aliphatic acids(Johnson et al.,1999), and the dorsal-ventral position of the module was correlated with carbonnumber (Fig. 3A). Carbon number in straight-chained, saturated compounds isexactly correlated with hydrophobicity, molecular volume, and molecular length,three properties that could affect an odorant ligandÕs interaction withreceptors. To determine if only one of these properties dictated thedorsal-ventral position of module A, we calcu­lated centroids of 2-DGuptake for five- and six-carbon aliphatic acid odorants possessing differenthydrocarbon structures (Johnson and Leon, 2000b).  The straight-chained, branched, cyclic, and double-bondedmolecules we chose differed independently in hydrophobicity, volume, andlength. Indeed, centroids within module A differed across these odorants, andthe dorsal-ventral position of the module was significantly correlated onlywith molecular length (Fig. 3B). Therefore, longer aliphatic acid odorantsappear to activate glomeruli located progressively more ventrally within themodule.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3. Dorsal-ventral positions of module A aresignificantly correlated with carbon number in straight-chained, saturatedaliphatic acids (A:r = 0.91, P < 0.05), and with molecular length in five- and six-carbonaliphatic acids of different hydrocarbon structure (B: r = 0.85, P <0.005). Lines are the results of linear regression. The y-axes are notcomparable between the two plots due to differences in the boundaries of themodules used for centroid determination. The odorants in B were 2-methybutyricacid (6),cyclobutane-carboxylic acid (q),isovaleric acid (l), tert-butylacetic acid(t),cyclopentanecarboxylic acid (à), isocaproic acid (u), valeric acid (Æ),trans-2-pentenoic acid (:), trans-3-hexenoic acid(¤), and caproic acid (#).

 

 

        Thespatial organization of glomeruli within module A is an example of chemotopicorganization, wherein glomeruli that are nearest neighbors respond to the mostclosely re­lated stimuli. Such arrangements may typify the organization ofglomeruli within and be­tween functional modules of the olfactory bulb. Forexample, module a of the lateral bulb shifts rostrally with increasing carbonnumber for aliphatic acids, and centroids within the caudo-lateral andcaudo-medial aspects of the bulb shift both rostrally and ventrally withincreasing carbon number for both acids and esters (Johnson et al., 1998; 1999). Also, the modules thatdistinguish low concentrations of odorants with related functional groups areclustered together in the caudal bulb (Fig. 1, bottom row). Lateral inhibitionmay enhance contrast between projection neurons associated with these nearbyglomeruli, thereby allowing them to discriminate subtly distinct chemicalproperties that may not be distin­guished adequately by an individualodorant receptor.

 

 


4. Parallelrepresentations of odorants within the lateral and medial aspects of theolfactory bulb

 

        Forevery module in the lateral aspect of the bulb, there appears to be a module ofsimilar specificity in the medial aspect of the bulb (Johnson et al., 1998, 1999; Johnson and Leon, 2000a, b). Thisphenomenon is perhaps best illustrated in Figure 2, where lateral modules areassigned lower case letters and the corresponding medial modules are assignedupper case letters. The similarities in the relative amounts of 2-DG uptakebetween lateral and medial modules across both odorant concentration andodorant functional groups results in a similar pattern of circle size betweenthe left and right sides of the figure. Therefore, there appear to be twosimilar representations of aliphatic odorants within each olfactory bulb, onein the lateral aspect and one in the medial aspect.

        Themedial modules are located about 1.7 mm more caudally than the corre­spondinglateral modules, and the medial modules also are situated more ventrally. Thereare similar spatial relationships between lateral and medial glomeruli thatreceive projec­tions from sensory neurons expressing the same odorantreceptor gene (Ressler et al., 1994). Itis likely that our paired modules reflect this sensory neuron projectionpattern (Johnson et al., 1998,1999). Possible reasons for two representations in each bulb include re­dundancyto insure odor perception after damage to part of the bulb or epithelium,coinci­dence detection to resolve stimulus-evoked activity from spontaneousactivity, and/or sepa­rate cortical projections underlying differentodor-influenced behaviors (Johnson et al., 1999).

 

 

5. different representations at different odorantconcentrations

 

        Ifbulbar spatial activity patterns are involved in the coding of perceived odorquality, then different odors should be associated with different activitypatterns. Indeed, odorants that differ only slightly in chemical structureevoke both different perceived odors and different patterns of 2-DG uptake(Fig. 1). Certain odorants also evoke different perceived odors at differentconcentrations. In our functional group study, we included two odorants thathumans report to change in odor quality with concentration (pentanal and2-hexanone) and three odorants with constant odors (valeric acid, methylvalerate, and pentanol). We found that the patterns evoked by pentanal and2-hexanone were significantly different at different concentrations (Johnsonand Leon, 2000a). New glomerular modules were evoked at higher concentrations,and these modules were located far away from those evoked at lowerconcentrations (Fig. 2). Quantitative comparisons indicated that the patternsevoked by certain concentrations of pentanal or 2-hexanone were more similar tothe pattern evoked by methyl valerate than to patterns evoked by different con­centrationsof the same odorant (Johnson and Leon, 2000a). Patterns evoked by valeric acid,methyl valerate, and pentanol did not differ significantly across differentodorant con­centrations. Thus, our data are consistent with a relationshipbetween olfactory bulb spatial activity patterns and odor quality perception.The data further suggest that rats may better dis­criminate betweendifferent concentrations of pentanal and 2-hexanone than between certainconcentrations of these odorants and methyl valerate.

 

 

6. conclusions

 

        Bymapping activity across the entire glomerular layer in rats exposed tosystematically different odorant chemicals, we find that aspects of odorant chemistryare represented spatially in the olfactory bulb. Responses to particularodorant molecular features are compartmentalized into glomerular modules, whereresponses may be tuned by using chemotopic glomerular arrangements and lateralinhibition. Further research, using odorants that differ in other aspects ofchemical structure, as well as behavioral studies correlating differences inodor perception with quantitative differences in spatial patterns of bulbaractivity, should greatly increase our understanding of olfactory bulb function.

 

 


references

 

Johnson,B.A., and Leon, M., 2000a, Modular representations of odorants in theglomerular layer of the rat olfactory bulb and the effects of stimulusconcentration, J. Comp. Neurol.409:495-509.

 

Johnson,B.A., and Leon, M., 2000b, Odorant molecular length: one aspect of theolfactory code, J. Comp. Neurol.,426:330-338.

 

Johnson,B.A., Woo, C.C., and Leon, M., 1998, Spatial coding of odorant features in theglomerular layer of the rat olfactory bulb, J. Comp. Neurol. 393:457-471.

 

Johnson, B.A., Woo,C.C., Hingco, E.E., Pham, K.L., and Leon, M., 1999, Multidimensional chemotopicresponses to n-aliphatic acid odorants in the rat olfactory bulb, J. Comp.Neurol. 409:529-548.

 

Ressler,K.J., Sullivan, S.L., and Buck, L.B., 1994, Information coding in the olfactorysystem: Evidence for a stereotyped and highly organized epitope map in theolfactory bulb, Cell 79:1245-1255.

 

Slotnick,B.M., Bell, G.A., Panhuber, H., and Laing, D.G., 1997, Detection and discriminationof propionic acid after removal of its 2-DG identified major focus in theolfactory bulb: a psychophysical analysis, Brain Res. 762:89-96.

 

Vassar, R.,Chao, S.K., Sitcheran, R., Nu–ez, J.M., Vosshall, L.B., and Axel, R., 1994,Topographic organization of sensory projections to the olfactory bulb, Cell 79:981-991.

 

Yokoi,M., Mori, K., and Nakanishi, S., 1995, Refinement of odor molecule tuning bydendrodendritic synaptic inhibition in the olfactory bulb, Proc. Natl. Acad.Sci. USA 92:3371-3375.