Tests of the chromatographic theory of olfaction with highly soluble odors: a combined electro-olfactogram and computational fluid dynamics study in the mouse

Abstract

The idea that the vertebrate nasal cavity operates like a gas chromatograph to separate and discriminate odors, referred to herein as the 'chromatographic theory' (CT), has a long and interesting history. Though the last decade has seen renewed interest in the notion, its validity remains in question. Here we examine a necessary condition of the theory: a correlation between nasal odor deposition patterns based on mucus solubility and the distribution of olfactory sensory neuron odotypes. Our recent work in the mouse failed to find such a relationship even across large sorption gradients within the olfactory epithelium (OE). However, these studies did not test extremely soluble odorants or low odor concentrations, factors that could explain our inability to find supporting evidence for the CT. The current study combined computational fluid dynamics (CFD) simulations of odor sorption patterns and electro-olfactogram (EOG) measurements of olfactory sensory neuron responses. The odorants tested were at the extremes of mucus solubility and at a range of concentrations. Results showed no relationship between local odor sorption patterns and EOG response maps. Together, results again failed to support a necessary condition of the CT casting further doubt on the viability of this classical odor coding mechanism.

Keywords: Maps; Odor discrimination; Olfactory receptors; Sorption.

Fig. 1.

Recording locations and receptor zone boundaries. (A) Midsagittal drawing of endoturbinates (rostral to the left). The primary EOG recording locations on the dorsal branch of endoturbinate two (II_d) are designated with numerically. Roman numerals designate endoturbinates using standard nomenclature (Coppola et al., 2017). (B) Micrograph of midsagittal view of whole mount showing endoturbinates with NQO1 immunolabeling revealing boundaries of dorsal-central (Zone 1; n=6 mice). Note brown label margins (arrowheads) and compare to recording locations in A. (C) Immunolabeling for NQO1 of coronal sections through a portion of the nasal cavity (n=5 mice). Arrowheads show limited distribution of label in dorsal meatus (top) and dorsal portion of turbinate II_d. Arrows show approximate rostrocaudal location of sections using the whole mount as a reference. S, nasal septum; NT, nasal turbinates; endoturbinates are labeled with Roman numerals by convention (Coppola et al., 2017). Scale bars: 1 mm (top), 0.5 mm (bottom).

Fig. 2.

CFD simulations and EOG results. (A_i,B_i,C_i) CFD simulations of airflow paths and odor deposition (flux) patterns are illustrated in the olfactory recess for three odors chosen because they have contrasting EOG response profiles (see Results). Flow patterns are illustrated with streamlines calculated from the CFD solution. In these medial views note that a region (black line) of the septum has been digitally removed to reveal the underlying endoturbinates. The EOG recording locations on endoturbinate II_d are shown as black-outlined white circles. Location one is caudal-most (right) and location six is rostral-most (left). (A_ii,B_ii,C_ii) Sample raw EOG traces from individual animals at each of the standard recording locations in response to 0.1% concentration of stimulus. The thick horizontal segments above the traces show when the stimulus was turned on. (A_iii,B_iii,C_iii) Mean (±s.e.m.; n=>9 mice per odor) EOG amplitudes at different recording locations are shown with red lines and symbols for three odorants [see odors and recording locations in A_i,B_i,C_i. For comparison to the EOG responses, odorant flux values were extracted from the CFD simulations of odorant deposition at the recording locations and are plotted in blue (right vertical axes)]. Note that ordinates have different scales to account for the different odor intensities so as to highlight EOG and CFD relationships. Pearson-r correlations and P-values are shown for each graph.

Fig. 3.

The effects of mucus solubility, functional group and concentration on EOG gradients. (A) For the highest odor concentrations only (see Results and below for justification), the relationship between the slope of the mean EOG response-gradient across six recording locations and air-mucus partition coefficients (Log₁₀) is plotted. Pearson-r correlation and P-values are shown. (B) Mean (±s.e.m.; n=9 mice) EOG response amplitudes across six recording locations using half-log dilution series of octanoic acid. (C) Mean (±s.e.m.; n=12 mice) EOG response amplitudes across six recording locations using half-log dilution series of acetophenone. Note for B and C that concentration has little effect on the slope or shape of the response gradient across endoturbinate II_d. (D) Mean (±s.e.m.; n=9–18 mice depending on odor) response gradients across six recording locations for four fatty acids (see text for explanation of stimulus concentrations and slope statistics).