Learning Objectives

In this paper, we present a combination of two complementary approaches to perform in vivo study of the functionality of the Xenopus tadpole olfactory system: i) a method for imaging presynaptic Ca²⁺ changes in the glomeruli of living tadpoles using a fluorescent calcium indicator, and ii) an odor guided behavioral assay that can be used to investigate the response to specific waterborne odorants. Since these approaches have been employed to evaluate the recovery of olfactory processing after injury¹⁰, a simple method for transecting olfactory nerves is also described.

Transection of olfactory pathways in Xenopus tadpoles
There are two ways to certify the validity of the procedure. Both rely on the visualization of olfactory nerves using fluorescent reporters. One method is based on transgenic tadpoles that express fluorescent proteins on their nervous system. Two recommended lines that express GFP under a neuronal β-tubulin promoter are X. laevis tubb2b-GFP and X. tropicalis NBT-GFP (Figure 1, see Table of Materials). If only wild-type animals are available, CM-diI can be used (see step 1.8). Figure 1 shows images of tubb2-GFP X. laevis tadpoles. Images are from four different animals where a single olfactory nerve was transected. The cut should be obvious under the dissection scope. The advantage of using transgenic lines is that reformation of the olfactory nerve can be followed over a period of time. When doing sequential observations, it is recommended to minimize exposure to fluorescent light to prevent photodamage. Sectioning of a single olfactory nerve is useful when an internal control is required, as for example, to compare normally developed vs rewired glomerular units. Sectioning of both olfactory nerves should be applied when the aim is completely suppressing the transmission of information.

Live imaging of presynaptic responses to olfactory stimuli
The correct labeling of ORNs with calcium green-1 dextran can be observed at the level of the olfactory bulb (Figure 2A) using widefield microscopy. Glomeruli are obvious (Figure 2B) and should appear distributed in different layers by moving the focus plane. The morphology of glomerular structures can be better resolved if a confocal microscope is used instead (Figure 2C). The number of labeled glomeruli depends on dye uptake at the level of the olfactory epithelium. Therefore, this procedure does not allow the visualization of all glomerular units. Animals showing a more intense fluorescence staining of the glomerular region should be selected before performing imaging experiments, since they contain more ORNs labeled. This maneuver is highly recommended in order to increase experimental throughput and should be performed under a dissecting scope equipped with a fluorescence lamp. Reject animals that do not show labeled glomerular units or that show fluorescence restricted to particular areas of the glomerular layer. Presynaptic Ca²⁺ responses can be observed as soon as 1 day after dye loading. For carrying out imaging experiments it is desirable to use objectives of high numerical aperture, typically ≥0.9.

Increases of presynaptic calcium levels can be evoked exposing dendritic knobs of ORNs to amino acids. It is important to position the capillary delivering the odorant solution above the nasal capsule. Care should be taken to avoid contact because it could clog the tip of the capillary and/or cause mechanical stimulation of ORNs. Transient increases in presynaptic calcium levels can be observed for stimuli ≥0.1 s (Figure 2D) and are indicative of a correct olfactory transduction. It is also important to visualize basal calcium levels with low camera gain. Presynaptic terminals of ORNs display high fluorescent increases and it is essential to avoid signal saturation. High temporal resolution can be achieved with widefield microscopy. For example, using an electron-multiplied CCD camera, it is possible to achieve frame rates of 50 Hz or higher. Use of confocal microscopy reduces temporal resolution but allows a better definition of glomerular structures.

The high affinity of calcium green for binding calcium (190 nM) is particularly useful to detect small responses. Transient intracellular calcium increases detected in the glomerular layer are variable. Some glomerular regions show changes in ΔF/F₀ ≥0.2, while neighboring processes might not even respond (Figure 2D). The following factors contribute to the variability of the response of glomerular units: i) the overall number of labeled glomeruli, ii) the intracellular concentration of calcium green, and iii) the selectivity of ORNs to detect amino acids. Since a too low number of labeled glomeruli might preclude observing responses, it is absolutely necessary to carry out these experiments with animals containing as many labeled ORNs as possible.

Olfactory-guided behavior
Data analysis
Olfactory-guided behavior is studied using a custom-built system. Figure 3 shows a schematic drawing of equipment used to carry out the assay. The test is based in the ability of tadpoles to detect the presence of amino acids, which act as odorants. A solution combining five different amino acids (methionine, leucine, histidine, arginine and lysine) is used for stimulation. The solution is locally delivered during 30 s to a 35 mm well containing a free-swimming tadpole. The immediate response of tadpoles to the incoming solution is an increase in motility. Random movements occurring during the initial ∼5–10 s of solution application are followed by a direct swim towards the source of odorants. Tadpoles remain for several seconds in the vicinity of the nozzle during delivery of amino acids and gradually recover motility in random directions (see Supplementary Videos 1, 2).

The experimental conditions described allow the normal swim of X. tropicalis tadpoles stages 48-52; however, it must be taken into account that the motility of larger animals might be restricted in 35 mm wells. Tadpole movements are recorded with a CCD camera. Attraction for the odorant solution can be detected as a reduction of the Euclidean distance separating perfusion inlet from tadpole position (Figure 4). Tracking of tadpole head positions within an area of 35 mm x 35 mm (or the equivalent size in pixels) allows obtaining a quantitative analysis of olfactory-guided behavior (Figure 4A). Individual plots of tadpole movements are constructed using X-Y coordinates obtained by image analysis (Figure 4B). The extracted motility plots must faithfully reproduce video images.

There are two possible methods of interpreting olfactory-guided behavior experiments. The first approach is inspired on a previous study using zebrafishes²¹. Measurement of the time spent in the vicinity of the nozzle delivering odorants evidences the presence of a positive tropism. A region of interest of 8.75 mm radius (corresponding to ¼ of the well diameter) centered on the solution inlet it is used to classify the proximity of the animals to the odorant source (Figure 4A, 4C). Binning the time spent by tadpoles in the vicinity of the nozzle during defined periods, i.e., 15 s intervals, allows identifying the ability to detect amino acid solutions (Figure 4D). The overall behavior of a population of tadpoles can be obtained by plotting the distribution of individual data (Figure 5A). A positive tropism can be detected when the solution of methionine, leucine, histidine, arginine and lysine is prepared either at 1 mM or 160 μM (Figure 5A and 5B). Animals do not respond to water application (Figure 5C), thus discarding the participation of mechanosensitive mechanisms. Differences among time intervals defined in each experimental group can be established using nonparametric repeated measures ANOVA with Dunn's multiple comparisons test. The disadvantage of binning data in time intervals of 15 s is a reduced temporal resolution.

A way to increase temporal information of the behavioral response is by making average plots of Euclidean distances from the odor source. Although tadpole movements show an intrinsic variability, the average motility of a population of animals (typically ≥40) shows the olfactory-guided behavior. To carry out this analysis it is necessary to group animal positions before the onset of stimulation. Since tadpoles are found in different locations when the odorant solution enters into the well it is required to set the Euclidean distance to 0 just before stimulation (Figure 6A, see also the inclusion criteria in Figure 7). Negative and positive values therefore indicate an attraction or a repulsion from the odor source, respectively. An attraction for odors is well described by a linear fit with regression coefficients ≥0.9. If water is delivered, the net changes of Euclidean distance are distributed around 0 and it is not possible to fit a line during odorant stimulation, thus indicating the absence of an olfactory-guided behavior (Figure 6B). Comparison of average plots of Euclidean distances for amino acid solutions prepared at 1 mM and 160 µM suggest different delays in the olfactory-guided response (compare Figure 6A and C). The time interval required to initiate the movement towards the source of odorants is shorter when amino acids are applied at a higher concentration. A lack of olfactory-guided behavior is observed in tadpoles with both olfactory nerves transected (Figure 6D).

A limitation of the described olfactory-guided behavioral assay is the establishment of complex fluid plumes. This can be seen if the amino acid solution is substituted by a dye, such as Fast Green, when setting up the system. The use of colored solutions verifies the formation of plumes and shows that waterborne odorants reach any region of the well within 5 s. Turbulences caused by the delivery of the solution are likely detected by the lateral line of tadpoles and probably contribute to the variability observed in animal motility but do not interfere with olfactory guided behavior. Control experiments carried out by using water instead of amino acid solutions reveal that tadpoles are capable to discriminate olfactory from mechanical stimuli. The estimation of time spent in a region of interest (Figure 5) and the average plot of Euclidean distances (Figure 6) are two complementary methods to describe the olfactory-guided response of tadpoles.

Inclusion criteria
Inclusion criteria must also be taken into account for data analysis. Some tadpoles show a resonant movement, which is illustrated by the fitting of the plot of Euclidean distances to a sinusoidal function (Figure 7A). Tadpoles displaying this behavior must be discarded from all analysis.

The exclusion of animals that at the onset of the application of odorant solutions are at a maximum (>30 mm Figure 7B) or a minimum Euclidean distance (<5mm, Figure 7C) from the nozzle allows decreasing the variability of the average plots. The example illustrated in Figure 7B shows a positive tropism for the amino acid solution. The tadpole is located at a maximum distance of the solution inlet at the onset of stimulation. Therefore, this relative position can only reveal an attraction for the odor source. Figure 7C shows the opposite situation. Here, a tadpole is located in the vicinity of the nozzle delivering the amino acid solution. Quantification of the time spent near the odor source shows a response (method used in Figure 5); however, it cannot show a net movement towards the inlet.

In summary, the proposed assay of olfactory-guided behavior defines a binary test. This method can be used to detect the ability of an experimental group of tadpoles to respond to odorants. Further improvements are required if the aim is establishing differences among complex olfactory-guided responses, as for example, determining preferences for given odors.

Figure 1: Transection of olfactory nerves. Representative images of tubb2-GFP X. laevis tadpoles obtained after transection of a single olfactory nerve (arrows). Nerves of tubb2-GFP tadpoles display strong fluorescence. Nerve transection is obvious immediately after surgery (D0). Regrowth of the olfactory nerve is evident 4 days after cut (D4). Eight days after surgery (D8) there is little difference between control and reformed nerves. Tadpoles were anesthetized in 0.02% MS-222 to collect images. Olfactory bulb (O.B.), olfactory nerve (O.N.), nasal capsule (N.C.), tectum (Tec), optic nerve (Op.N.). Arrows indicate the location of the transected nerve. Please click here to view a larger version of this figure.

Figure 2: Labeling of olfactory receptor neurons with calcium green dextran and visualization of presynaptic calcium influx upon stimulation with amino acids. (A) Transmitted light image of a tadpole showing the location of the olfactory epithelium, olfactory nerves and the glomerular layer of the olfactory bulb. (B) Image of an olfactory bulb visualized by widefield (wf) microscopy. Neurons were labeled by application of calcium green-1 dextran at the nasal capsule. The fluorescence observed corresponds to presynaptic terminals from olfactory receptor neurons forming glomeruli. O.N: olfactory nerve; O.B: olfactory bulb. (C) Confocal section located dorsally from the entry of the olfactory nerve to the bulb. The presynaptic component of olfactory glomeruli was labeled using calcium green-1 dextran, as in B). (D) Presynaptic terminals transiently increase their calcium levels upon exposure of the olfactory epithelium to a solution containing five different amino acids. Relative changes in calcium fluorescence obtained before, during and after 1 s stimulation. (E) Distribution of 10 different regions of interest (ROIs) used to quantify ΔF/F₀ changes. ROI11 is outside of the glomerular layer and is used to define background levels of fluorescence. (F) Individual ΔF/F₀ responses for ROIs defined in E). Please click here to view a larger version of this figure.

Figure 3: Olfactory-guided behavior assay. (A) Schematic diagram showing the equipment used in the test. (B) Example of motility tracks recorded over 90 s. Each circle represents a well containing a single animal. Please click here to view a larger version of this figure.

Figure 4: Tracking of olfactory-guided behavior. (A) Example showing a well used for the behavioral assay. Blue dotted lines indicate the location of X,Y coordinates (in mm) used to track tadpole movements (see also Supplementary Video 1). The green ellipse represents the position of the solution inlet. The dotted black line indicates the proximal area to the tube delivering the amino acid solution (see text for details). (B) Motility of the tadpole shown in A) during the behavioral assay. Color-coded traces indicate the position of the animal before (gray) and after olfactory stimulation (violet). Movements during application of the odorant solution are illustrated in a temporal color gradient (30 s, red to blue). (C) Using X,Y tadpole positions it is possible to calculate changes in the Euclidean distance from the tadpole head to the perfusion inlet. Distances shorter than 8.75 mm correspond to the proximal area of the nozzle. (D) Plot of the time spent by tadpoles in the region defined by the dotted line in A). Each dot indicates a 15 s period. The animal is attracted by the odorant solution. Please click here to view a larger version of this figure.

Figure 5: Tadpoles are attracted by amino acids. (A) Time spent by tadpoles in the vicinity of the odor source. Each bin comprises a 15 s period. Box plots represents the median (black horizontal line), 25 to 75% quartiles (boxes), and ranges (whiskers) of data. Delivery of a 1 mM amino acid solution attracted tadpoles to the odor source. (B) Tadpoles were attracted by the odorant solution when the concentration of amino acids was reduced to 160 μM. (C) Delivery of water did not modify animal behavior. Repeated measures ANOVA with Dunn's multiple comparisons test, p < 0.05. Please click here to view a larger version of this figure.

Figure 6: Temporal response of tadpoles to odorants. (A) Plot of the average Euclidean distance to the odor source as a function of time. The Euclidean distance was set to 0 before stimulation in each individual trace. Negative and positive values indicate a decrease and an increase in distance to the odor source, respectively. Attraction to the odor source can be described by a linear fit (r²=0.98). (B) Delivery of water does not modify the distance to the odor source. (C) Tadpoles respond to the application of a 160 μM solution of amino acids as revealed by a linear fit (r²=0.96). (D) Tadpoles with both olfactory nerves transected do not respond to amino acids. Please click here to view a larger version of this figure.

Figure 7: Inclusion criteria for the olfactory-guided behavior assay. (A) Some tadpoles show a resonant movement. This behavior is revealed by successful fit of a sinusoidal function to the plot of the Euclidean distance to the odor source. Tadpoles displaying this activity should be excluded from the test. (B, C) A way to reduce variability in the average temporal response to odorants (Figure 6) is by excluding animals located at a maximum (B) or a minimum (C) Euclidean distance at the onset of stimulation. Red dotted lines indicate the threshold value (30 mm and 5 mm). The Euclidean distance before the onset of stimulation (arrow, left plots) is set to "0" to report attractive or repulsive behaviors as negative or positive values (right plots), respectively. Please click here to view a larger version of this figure.

Supplementary Video 1: Olfactory-guided behavior triggered by delivery of an amino acid solution. The movie shows a tadpole freely swimming on a 35 mm well. The blue ellipse indicates the position of the nozzle delivering the odorant solution. The onset and end of stimulation are indicated by green and red dots, respectively. Figure 4 shows the analysis of the behavior observed. Please click here to view this video. (Right-click to download.)

Supplementary Video 2: Tadpole motility during delivery of water. The movie shows a tadpole freely swimming on a 35 mm well. The blue ellipse indicates the position of the nozzle releasing MQ water. The onset and end of water delivery are indicated by green and red dots, respectively. Please click here to view this video. (Right-click to download.)