Odorant Receptor–Derived cAMP Signals Direct Axonal Targeting

See allHide authors and affiliations

Science  27 Oct 2006:
Vol. 314, Issue 5799, pp. 657-661
DOI: 10.1126/science.1131794


In mammals, odorant receptors (ORs) direct the axons of olfactory sensory neurons (OSNs) toward targets in the olfactory bulb. We show that cyclic adenosine monophosphate (cAMP) signals that regulate the expression of axon guidance molecules are essential for the OR-instructed axonal projection. Genetic manipulations of ORs, stimulatory G protein, cAMP-dependent protein kinase, and cAMP response element–binding protein shifted the axonal projection sites along the anteriorposterior axis in the olfactory bulb. Thus, it is the OR-derived cAMP signals, rather than direct action of OR molecules, that determine the target destinations of OSNs.

Each olfactory sensory neuron (OSN) in the mouse expresses only one functional odorant receptor (OR) gene out of ∼1000 members (13). Axons from OSNs expressing a given OR converge onto a specific site, the glomerulus, in the olfactory bulb (46). It has been proposed that OR molecules at axon termini may directly recognize guidance cues on the olfactory bulb and mediate homophilic interactions of like axons (610). OR molecules are heterotrimeric guanine nucleotide–binding protein (G protein)–coupled receptors (GPCRs) that transduce the odorant-binding signals by activating the olfactory-specific G protein (Golf) expressed in mature OSNs. The activation of Golf stimulates adenylyl cyclase type III, generating cAMP, which opens cyclic nucleotide–gated (CNG) channels. Mice deficient for Golf and CNGA2 are anosmic but form a normal glomerular map (1113), which suggests that a G protein other than Golf may aid in targeting OSNs independent of CNG channels.

OR molecules are rhodopsin-like type A GPCRs that contain a conserved tripeptide motif, Asp-Arg-Tyr (DRY), at the cytoplasmic end of transmembrane domain III (fig. S1A), which is required for coupling of GPCRs to the partner G proteins (14, 15). To examine whether the G protein signaling is involved in guidance of OSN axons, we generated a DRY-motif mutant, Arg-Asp-Tyr (RDY), for the rat OR gene I7 (16) and expressed it using a transgenic system (17) (fig. S1B). Axons from OSNs expressing the wild-type I7, I7(WT), converged to a specific site in the olfactory bulb (Fig. 1A, left), whereas those expressing the DRY-motif mutant, I7(RDY), remained in the anterior region of the olfactory bulb, failing to converge onto a specific glomerulus (Fig. 1A, right). The I7(RDY)-expressing axons never penetrated the glomerular layer but stayed within the olfactory nerve layer (Fig. 1B). These axon termini were devoid of synaptotagmin (presynaptic marker) and microtubule-associated protein 2 (dendritic marker) immunoreactivities and thus probably did not form synapses (Fig. 2A, middle, and fig. S2). OSNs expressing a nonfunctional OR gene can activate other OR genes and will fail to converge onto a single glomerulus (8, 18, 19). However, the inability of I7(RDY) axons to converge on a specific glomerulus was not due to the coexpression of other OR genes (fig. S3); OSNs expressing the I7(RDY) transgene expressed no other OR genes. OSNs expressing I7(WT) all showed Ca2+ signals in response to octanal (an agonist of the I7 receptor), whereas those expressing I7(RDY) did not (Fig. 1C). Thus, the I7(RDY) mutant is deficient in both axon targeting and G protein coupling.

Fig. 1.

A DRY-motif mutant of I7 OR. The DRY sequence in the wild-type OR, I7(WT), was changed to RDY in the mutant protein, I7(RDY). (A) Whole-mount fluorescent views of olfactory bulbs at postnatal day 14 (P14). Medial aspects are shown. Dotted yellow lines demarcate olfactory bulbs. A, anterior; D, dorsal; i, IRES. Scale bars, 500 mm. (B) Coronal sections of olfactory bulbs, stained with antibodies to C/YFP (green) and 4′,6′-diamidino-2-phenylindole (DAPI) (blue). Dashed lines demarcate the olfactory nerve layers (ONL) from the glomerular layers (GL). Scale bars, 100 mm. (C) Fura-2 calcium imaging. OSNs responsive to forskolin (an activator of adenylyl cyclase) were analyzed. Cells expressing the I7(WT) all responded to octanal (n = 12 OSNs), whereas those expressing I7(RDY) did not (n =10). Fsk, 50 mM forskolin; Oct, 500 mM octanal. F ratio (340/380), the ratio of fura-2 fluorescence intensities at 510 nm with excitation at 340/380 nm. Insets show the C/YFP fluorescent images of OSNs analyzed.

Fig. 2.

Rescue of the I7(RDY) phenotype in OSN projection. (A) I7(RDY)-caGs. (B) I7(RDY)-caPKA. (C) I7(RDY)-caCREB. Whole-mount fluorescent views are shown for the medial surface of the olfactory bulbs (age P14). The source of cAMP signals is schematically shown in (A) for I7(RDY)-caGs. Coronal sections stained with antibodies to C/YFP (green) and DAPI (blue) are shown below. Synapse formation was examined with antibodies to synaptotagmin (red). Scale bars, 500 mm for the whole-mount bulbs, 100 mm for sections.

Both Go and Gs genes are expressed in immature mouse OSNs (20). Although the Gs knockout mutation is embryonically lethal (21), the Go-deficient mouse shows no obvious anatomical defect in the olfactory system (22). Because the DRY-motif mutant was assumed to be incapable of coupling with G proteins, we examined whether the constitutively active Gs (caGs) mutant would rescue the defective phenotype of I7(RDY) in axonal projection. We inserted the caGs gene into the I7(RDY) construct with an internal ribosome entry site (IRES), generating I7(RDY)-caGs (fig. S1B). In OSNs expressing this construct, cAMP signals should be generated constitutively by caGs in a receptor-independent manner. Axons expressing I7(RDY)-caGs (Fig. 2A, cyan) converged to a specific site in the olfactory bulb, whereas axons expressing I7(RDY) (Fig. 2A, yellow) did not. Yellow fluorescent protein (YFP)–positive and cyan fluorescent protein (CFP)–positive axons did not intermingle or co-converge, which suggests that homophilic interaction of OR molecules is unlikely. Axons expressing I7(RDY)-caGs were found within a glomerular structure and were immunoreactive for synaptotagmin (Fig. 2A, right). Gs stimulates adenylyl cyclase to produce cAMP, which in turn activates cAMP-dependent protein kinase (PKA). A constitutively active PKA rescued the defective phenotype of I7(RDY) in OSN projection and glomerular formation, although a few projection sites were found in the posterior region in the olfactory bulb (Fig. 2B). When the I7(RDY) construct was coexpressed with a constitutively active variant of cAMP response element–binding protein (CREB), a PKA-regulated transcription factor, axon termini were found within glomerular structures, although with incomplete convergence (Fig. 2C). These results confirm the role of G proteins in OSN axon targeting and suggest the involvement of cAMP in transcriptional regulation of axon guidance molecules.

To study cAMP signaling in OSN projection, we examined the effect of caGs on OSNs expressing the wild-type OR. Two transgenic constructs, I7(WT)-Cre and I7(WT)-caGs, were analyzed. The Cre recombinase gene was assumed not to affect the Gs-mediated signaling. Axons from OSNs expressing I7(WT)-Cre (Fig. 3A, yellow) or I7(WT) (Fig. 3A, cyan) converged in similar regions, whereas those expressing I7(WT)-caGs (Fig. 3B, left) projected to more posterior regions. Additional cAMP signals are generated by caGs. In OSNs expressing I7(WT)-caGs, cAMP signals are generated by both the transgenic caGs and endogenous Gs, whereas, in OSNs expressing I7(RDY)-caGs, the generation of cAMP signals by endogenous Gs is blocked. The glomerulus for I7(WT)-caGs (Fig. 3B, yellow) showed a smaller posterior shift from that for I7(RDY)-caGs (Fig. 3B, cyan). Thus, the signaling level of the endogenous Gs appears to be relatively low when coupled with the wild-type OR. We also tested whether decreased levels of cAMP signals would affect the OSN projection. Axons expressing a dominant-negative PKA (dnPKA) with the wild-type OR converged to the anterior part of the olfactory bulb (Fig. 3C). Unlike axons carrying I7(RDY), axons expressing the I7(WT)-dnPKA construct generated glomerular structures. These transgenic experiments indicate that increased or decreased levels of cAMP signals shift the glomerular target of OSNs posteriorly or anteriorly, respectively.

Fig. 3.

Genetic manipulations of cAMP signals. Whole-mount fluorescent views of olfactory bulbs (medial surface) were analyzed for various double transgenic mice (age P14). OSN axons expressing the transgenes were visualized with CFP (cyan) or YFP (yellow). (A) YFP-tagged I7(WT)-Cre /CFP-tagged I7(WT). (B) Left, YFP-tagged I7(WT)-caGs /CFP-tagged I7(WT); right, YFP-tagged I7(WT)-caGs /CFP-tagged I7(RDY)-caGs. (C) YFP-tagged I7(WT)-dnPKA /CFP-tagged I7(WT). (D) Codingregion–replaced transgenic constructs, Cre and caG his. (E) CFP-tagged I7(RDY)-caGs /YFP-taggedcaG his. (F) YFP-tagged caG his /CFP-tagged I7(WT). Higher power confocal views of the boxed area (insets) are also shown. Arrowheads indicate the trajectories of labeled axons. Sources of cAMP signals are schematically shown for the OSNs expressing the respective transgenic constructs. Scale bars, 500 mm for whole-mount bulbs, 100 mmforsections.

To examine the effect of excessive cAMP signals on OSN projection, we generated the transgenic construct, caG his, in which the OR coding sequence has been replaced with the caGs gene (fig. S1B). More caGs was translated from the cap-dependent caG his than from the IRES-mediated I7(RDY)-caGs (8). Although we expected a posteriorly shifted but scattered pattern of projection with caG his, we detected only one or a few glomeruli (Fig. 3D). Projection sites driven by caG his were located posterior to the I7(RDY)-caGs glomeruli (Fig. 3E). In situ hybridization and single-cell reverse transcription polymerase chain reaction (RT-PCR) indicate that OSNs expressing caG his express multiple OR species (fig. S3). In the double transgenic mouse carrying CFP-tagged I7(WT) and YFP-tagged caG his, a few I7(WT)-expressing axons that probably also expressed caG his projected to the caG his glomerulus (Fig. 3F). Thus, the caG his glomerulus represents a heterogeneous population of axons expressing different ORs. It is possible that caG his produces saturated levels of cAMP signals and generates a distinct glomerular structure regardless of the OR species.

In contrast to Golf, Gs is expressed early in OSN differentiation (11). Our experiments suggest the involvement of a PKA-regulated transcription factor, CREB, in OSN projection (Fig. 2C). We used microarray and RT-PCR analyses to screen for genes with expression levels correlated with cAMP signals. cDNA libraries were prepared from single OSNs from four different transgenic mice, and gene expression profiles were compared between caG his and I7(RDY) and between I7(WT) and I7(WT)-dnPKA (Fig. 4, A and B). Among the genes differentially expressed were some encoding axon guidance molecules [for example, neuropilin-1 (Nrp1)]. Nrp1 was expressed in the caG his OSNs (where cAMP signals might be high), but not in the I7(RDY)-expressing OSNs (where cAMP signaling is blocked) (Fig. 4B and fig. S4). Immunostaining demonstrated a gradient of Nrp1 expression, with low expression in the anterior and high expression in the posterior of the olfactory bulb (fig. S5). In the I7(WT) /I7(WT)-dnPKA mouse, the I7(WT) glomerulus (Fig. 4C, cyan, posterior) was Nrp1-positive, and the I7(WT)-dnPKA glomerulus (Fig. 4C, yellow, anterior) was Nrp1-negative. Nrp1 has been implicated in guidance of OSN axons, because the disruption of the Sema3A gene, which encodes a repulsive ligand for Nrp1, alters glomerular arrangements along the anterior-posterior axis (23, 24). We suggest that Gs-mediated cAMP signals regulate the transcription of genes encoding axon guidance molecules, which in turn guide positioning of glomeruli.

Fig. 4.

Glomerular locations and cAMP signal levels. (A) Projection sites on the medial surface of the olfactory bulb (OB) are schematically shown for various transgenic constructs (fig. S6B). (B) RT-PCR analyses of single OSNs. cDNA libraries were prepared from single OSNs of four different transgenic mice: I7(RDY), I7(WT)-dnPKA, I7(WT), and caG his. The genes up-regulated by the cAMP signals were screened with microarray and RT-PCR analyses. Mixtures of 20 single-cell cDNA samples were analyzed by RT-PCR for the expression of isolated genes. Six examples are shown: pcp4l1, plxna3, nrp1, nxph3, ptprn, and ptprf. The gene ebf1 (olf-1) was used as a control. (C) Expression profiles of Nrp1 in the olfactory bulb. Two horizontal olfactory bulb sections (80 mmapart) from the I7(WT)/I7(WT)-dnPKA double transgenic mouse (age P14) were immunostained with antibodies to Nrp1 (red). The posteriorly located I7(WT) glomerulus (cyan, arrowhead) was immunoreactive for Nrp1, whereas the anteriorly located I7(WT)-dnPKA glomerulus (yellow, arrowhead) was not. On the left, the I7(WT) and I7(WT)-dnPKA glomeruli (dotted traces) are compared for the Nrp1 expression. Quantitative analyses of glomeruli for Nrp1 expression are shown in fig. S5.

Our results explain some puzzling observations about OSN targeting. The β2-adrenergic receptor (β2AR), but not a vomeronasal receptor (V1rb2), can substitute for an OR in OR-instructed axonal outgrowth and glomerular formation (8). The explanation for this observation may be that the β2AR can couple to Gs, but the V1rb2 cannot. This explanation is consistent with the idea that the Gs-mediated cAMP levels set by the receptors determine the target sites of OSN axons. Another puzzling observation is that alterations in OR expression levels can affect OSN projection (8). The level of cAMP signals may be affected by both OR identity and the amount of OR protein, which would be a factor of transcription and translation parameters. OR-instructed Gs signals are not dependent on odorants (23), and disruption of Golf or CNGA2 genes did not affect positioning of glomeruli (1113), which suggests that Gs-mediated cAMP signaling is distinct from that mediated by odor-evoked neuronal activity. It has been thought that ORs at axon termini may recognize guidance cues on the olfactory bulb and mediate the homophilic interactions of like axons (610). However, our results favor a model in which cAMP signals regulate the targeting of OSN axons along the anterior-posterior axis (fig. S6A). These results complement previous studies indicating that the dorsal-ventral arrangement of glomeruli is determined by the locations of OSNs within the olfactory epithelium (2527). We propose that a combination of dorsal-ventral patterning, based on anatomical locations of OSNs, and anterior-posterior patterning, based on OR-derived cAMP signals, establishes olfactory bulb topography. After OSN axons reach their approximate destinations in the olfactory bulb, further refinement of the glomerular map may occur through fasciculation and segregation of axon termini in an activity-dependent manner.

Supporting Online Material

Materials and Methods

Figs. S1 to S6

Table S1


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article