The olfactory system provides mammals with the abilities to investigate, communicate and interact with their environment. These functions are achieved through a finely organized circuit starting from the nasal cavity, passing through the olfactory bulb and ending in various cortical areas. We show that the absence of transient axonal glycoprotein-1 (Tag1)/contactin-2 (Cntn2) in mice results in a significant and selective defect in the number of the main projection neurons in the olfactory bulb, namely the mitral cells. A subpopulation of these projection neurons is reduced in Tag1-deficient mice as a result of impaired migration. We demonstrate that the detected alterations in the number of mitral cells are well correlated with diminished odor discrimination ability and social long-term memory formation. Reduced neuronal activation in the olfactory bulb and the corresponding olfactory cortex suggest that Tag1 is crucial for the olfactory circuit formation in mice. Our results underpin the significance of a numerical defect in the mitral cell layer in the processing and integration of odorant information and subsequently in animal behavior.

The olfactory system can discriminate and process a variety of odors through its fine and sophisticated structure (Fig. 1A). Olfactory sensory neurons (OSNs) from the olfactory epithelium (OE) project to the main olfactory bulb (MOB), where they synapse with dendrites of mitral and tufted cells (MCs, TCs), the olfactory projection neurons, and form the glomeruli structures (Whitman and Greer, 2009). Each OSN expresses a unique type of receptor and projects towards topographically restricted glomeruli (Buck and Axel, 1991; Mombaerts et al., 1996; Mombaerts, 2001; Mori et al., 2006; Sakano, 2010). MCs and TCs from the MOB project their axons towards the lateral olfactory track (LOT), and transmit the olfactory signal to divergent areas of the olfactory cortex (piriform and entorhinal cortices, olfactory tubercle, etc.) (Mouret et al., 2009; Leinwand and Chalasani, 2011). The accessory olfactory system (AOS), which is responsible for pheromone perception, is similarly organized, with the primary sensory neurons of the vomeronasal organ projecting to the accessory OB (AOB, dorsal OB region). Subsequently, the MCs and TCs of the AOB send their axons through the LOT to cortical regions (Baum and Kelliher, 2009; de Castro, 2009). Olfactory system dysfunction has been described in aged individuals as well as in patients with neurodegenerative disorders and dementias. Recent studies propose it as an early biomarker for the diagnosis and progression of these diseases (Attems et al., 2014).

MCs are generated in the ventricular zone of the rostral telencephalon during embryonic days (E)9.5-13.5 and migrate radially towards the intermediate zone (IZ), where they acquire a tangential-like morphology (Blanchart et al., 2006; Imamura et al., 2011; Imamura and Greer, 2013). The precise mechanisms underlying MC development are not yet understood. The adhesion molecules neuropilin 1 and 2 (Nrp1/2), L1 (also known as L1cam), Ocam (also known as Ncam2) and transient axonal glycoprotein-1 (Tag1), also known as contactin-2 (Cntn2), are expressed on OB projection neurons during development but their involvement in the integration of projection neurons into the OB circuitry is not yet elucidated (Wolfer et al., 1994; Yoshihara et al., 1995; Treloar et al., 2003; Inaki et al., 2004). Here we analyze the role of Tag1 in OB development, organization and function. Tag1, an immunoglobulin superfamily (IgSF) member, is implicated in key processes such as axonal fasciculation, pathfinding, neurite extension, neuronal migration and in the maintenance of the integrity of myelinated fibers (Furley et al., 1990; Denaxa et al., 2001; Karagogeos, 2003; Traka et al., 2003; Baeriswyl and Stoeckli, 2008; Wolman et al., 2008).

Tag1 is expressed by OB projection neurons and is detected on the axons that occupy the superficial layer of the LOT during development (Inaki et al., 2004). Our analysis revealed reduced numbers of MCs in the MOB of Tag1-deficient mice, attributed to altered distribution resulting from abnormal migration of a specific neuronal subpopulation, born at E11.5. The cellular defects are accompanied by behavioral deficits related to olfaction and reduced neuronal activity of the main olfactory system (MOS). Our data support a crucial role for Tag1 in the correct positioning of MCs, a prerequisite for proper OB function. Our study further indicates that disruption in a subpopulation of MCs can result in severe olfactory deficits.

Tag1 expression in the developing olfactory system

We analyzed the spatiotemporal expression of Tag1 in the developing OB and correlated it with specific markers. Previous reports show expression by MCs and TCs and their axons in the developing OB (Wolfer et al., 1994; Yoshihara et al., 1995; Inaki et al., 2004). Tag1 is detected as early as E12.5 in the OB and the OE (Fig. S1A-D). The expression in the OB is high at E13.5 and E15.5 and decreases towards birth (Fig. S1C,E,G). Tag1 protein and its mRNA are strongly present in the IZ of the OB primordium at E12.5 and E13.5 which contains mainly the premature MCs (Fig. S1A,C). With the gradual maturation of OB structure, Tag1 is specifically expressed by migrating MCs and TCs detected in the mitral cell layer (MCL) of the MOB, in the internal plexiform layer and the periventricular part of the OB (Fig. S1E,G,H). In the OE, Tag1 mRNA expression in the vomeronasal organ reaches high levels around E15.5, mainly at the dorsal part, and is significantly reduced at birth [postnatal day (P)0] (Fig. S1D,F,G). No expression is detected in the AOB (Fig. S1H, asterisks). Tag1 is also detected in the anterior olfactory nucleus, although there is no expression inside the piriform cortex (PC, layers II/III), olfactory tubercle and nucleus accumbens (Fig. S1I-L).

In the OB, Tag1 protein is localized in the cell body until E16.5, although after E17.5 it is mainly detected on the axon (Fig. 1A,C; Fig. S1). In order to track and study Tag1 expression in vivo, we generated a transgenic mouse line that expresses green fluorescent protein (GFP) under the control of the Tag1 gene promoter (Tag1loxP-GFP-loxP-DTA; Fig. 1B). GFP recapitulates Tag1 expression in the OB of E13.5 mice (Fig. 1Ci). It is detected in the IZ at E14.5, on immature MCs and TCs (co-expression with the early marker Tbr1 in Fig. 1Cii), but not on oligodendrocytes or interneurons in newborn OBs (co-expression with Olig2 and GABA, markers of oligodendrocytes and interneurons, respectively; Fig. 1Ciii,iv). In the OE, GFP is detected in OSNs and co-localizes with Tag1 only in the olfactory nerve (Fig. 1Cv, asterisks). Triple fluorescent in-situ hybridisation (ISH) for Tag1, Gap43 (marker of immature OSNs) and Omp (marker of mature OSNs), showed that Tag1 is predominantly expressed by immature neurons in the OE and is rarely detected in mature OSNs (Fig. 1Cvi, asterisk). Subsequent analysis for two of the main subgroups of OSNs revealed the selective expression of Tag1 by Mash1+ (also known as Ascl1) and not by neurogenin1+ (Ngn1, also known as Neurog1) neurons (Fig. 1Cvii,viii).

Fig. 1.

Tag1 expression in the developing olfactory system. (A) Schematic representation of the olfactory system organization in rodents. Blue, main olfactory system (MOS); red, accessory olfactory system (AOS). (B) Design of Tag1loxP-GFP-loxP-DTA mouse expressing GFP under the Tag1 gene promoter. (C) Analysis of Tag1 expression in the OB and OE. (Ci,v) IHC for GFP and Tag1 at E13.5 OB and OE respectively. There is co-localization of GFP and Tag1 in the intermediate zone (IZ) of the OB (i) and in the olfactory nerve (asterisks in v), whereas GFP expression is found in the OSNs inside the OE. GFP co-localizes with Tbr1 in E14.5 OB (ii), but not with Olig2 (iii) or GABA (iv), which are markers of oligodendrocytes and interneurons, respectively, at P0 in OBs. (Cvi-viii) Fluorescent ISH at E16.5 OE. Tag1 is co-expressed with Gap43 and rarely with Omp (asterisk in vi). It also co-localizes with Mash1 (vii), but not with neurogenin 1 (Ngn1, viii). Scale bars: 100 μm in Ci,v; 50 μm in Cii-iv,vi-viii. AMY, amygdaloid nuclei; AOB, accessory olfactory bulb; AON, anterior olfactory nucleus; BST, bed nucleus of the stria terminalis; EC, entorhinal cortex; LOT, lateral olfactory tract; MOB, main olfactory bulb; nAOT, nucleus of the accessory olfactory tract; OB, olfactory bulb; OT, olfactory tubercle; PC, piriform cortex; TT, tenia tecta; SEZ, subependymal zone.

Fig. 1.

Tag1 expression in the developing olfactory system. (A) Schematic representation of the olfactory system organization in rodents. Blue, main olfactory system (MOS); red, accessory olfactory system (AOS). (B) Design of Tag1loxP-GFP-loxP-DTA mouse expressing GFP under the Tag1 gene promoter. (C) Analysis of Tag1 expression in the OB and OE. (Ci,v) IHC for GFP and Tag1 at E13.5 OB and OE respectively. There is co-localization of GFP and Tag1 in the intermediate zone (IZ) of the OB (i) and in the olfactory nerve (asterisks in v), whereas GFP expression is found in the OSNs inside the OE. GFP co-localizes with Tbr1 in E14.5 OB (ii), but not with Olig2 (iii) or GABA (iv), which are markers of oligodendrocytes and interneurons, respectively, at P0 in OBs. (Cvi-viii) Fluorescent ISH at E16.5 OE. Tag1 is co-expressed with Gap43 and rarely with Omp (asterisk in vi). It also co-localizes with Mash1 (vii), but not with neurogenin 1 (Ngn1, viii). Scale bars: 100 μm in Ci,v; 50 μm in Cii-iv,vi-viii. AMY, amygdaloid nuclei; AOB, accessory olfactory bulb; AON, anterior olfactory nucleus; BST, bed nucleus of the stria terminalis; EC, entorhinal cortex; LOT, lateral olfactory tract; MOB, main olfactory bulb; nAOT, nucleus of the accessory olfactory tract; OB, olfactory bulb; OT, olfactory tubercle; PC, piriform cortex; TT, tenia tecta; SEZ, subependymal zone.

OSNs are unaffected but MCs of the MOB are fewer in Tag1−/− mice

We then analyzed the phenotype of Tag1-deficient mice in the olfactory system. As Tag1 is expressed early in OSN development, we tested whether mature OSNs are affected. ISH for Omp revealed no difference in the number of mature OSNs in deficient (Tag1−/−) vs wild-type (WT) controls (Fig. 2A-C). No gross abnormality in the projections of OSNs to the glomeruli in adult Tag1−/− mice was detected by Ocam immunoreactivity (Fig. S2A,B). Previous studies have shown that the genes encoding the adhesion molecules Kirrel2 and Kirrel3 are transcribed in an activity-dependent manner (Serizawa et al., 2006; Kaneko-Goto et al., 2008). We performed immunohistochemistry (IHC) for Kirrel2 combined with olfactory receptors mOR-EG or OR-I7 (also known as Olfr73 and Olfr2, respectively) in order to analyze the glomerular map formation and activity (Fig. S2C-R). We detected no changes between Tag1−/− mice and WT controls.

Fig. 2.

MCs are fewer in Tag1−/− mice, whereas TCs and OSNs are unaffected in number. (A-C) Fluorescent ISH for Omp in E16.5 mouse embryos showed no difference in the number of mature OSNs in Tag1−/− compared with WT embryos (P>0.1, n=3). D-F. IHC for reelin in adult mice showed no defect in the TC population in Tag1−/− animals (P>0.1, n=3). G-L. IHC in adult (G-I) and P0 mice (J-L) for Tbx21. In the adult OB (I), the number of Tbx21+ MCs is significantly reduced in Tag1+/− (14.2%, P=0.0058, n=6) and Tag1−/− mice (28.73%, P<0.001, n=6) compared with age-matched WT animals. Quantification in P0 OBs (L), showed a significant decrease in Tag1−/− mice (24.8%, P=0.0012, n=4). To-Pro-3 (blue) was used for nuclear staining. Error bars represent s.e.m. **P<0.01, ***P<0.001. Scale bars: 150 μm in A,B,G,H; 75 μm in D,E,J,K.

Fig. 2.

MCs are fewer in Tag1−/− mice, whereas TCs and OSNs are unaffected in number. (A-C) Fluorescent ISH for Omp in E16.5 mouse embryos showed no difference in the number of mature OSNs in Tag1−/− compared with WT embryos (P>0.1, n=3). D-F. IHC for reelin in adult mice showed no defect in the TC population in Tag1−/− animals (P>0.1, n=3). G-L. IHC in adult (G-I) and P0 mice (J-L) for Tbx21. In the adult OB (I), the number of Tbx21+ MCs is significantly reduced in Tag1+/− (14.2%, P=0.0058, n=6) and Tag1−/− mice (28.73%, P<0.001, n=6) compared with age-matched WT animals. Quantification in P0 OBs (L), showed a significant decrease in Tag1−/− mice (24.8%, P=0.0012, n=4). To-Pro-3 (blue) was used for nuclear staining. Error bars represent s.e.m. **P<0.01, ***P<0.001. Scale bars: 150 μm in A,B,G,H; 75 μm in D,E,J,K.

We looked closely at MCs and TCs in the OB of Tag1−/− mice. TC density in adult Tag1−/− animals presents no defects by reelin IHC (Fig. 2D-F). By contrast, we detected a gross reduction in the number of MCs distributed in the MCL of adult Tag1−/− mice, using the MC-specific marker Tbx21 (Fig. 2G,H). Quantification revealed a significant dose-dependent reduction of 14.2% in heterozygous mutants (Tag1+/−) and 28.7% in Tag1−/− mice compared with control age-matched WT mice (Fig. 2I). In newborns (P0), a 24.8% reduction in the number of MCs in the MCL of Tag1−/− mice compared with WT is detected, suggesting that the defect occurs during development (Fig. 2J-L).

The main stream of MCs, born at E11.5 are reduced in the MCL of Tag1−/− mice and ectopically located in the developing AOB

The majority of MCs in the OB originate at E11.5, whereas cells born earlier (E9.5 and E10.5) and later (E12.5 and E13.5) give rise to fewer MCs (Hinds, 1968a,b; Imamura et al., 2011). We injected Tag1+/+;GFP pregnant mice intraperitoneally with BrdU at E10.5, 11.5 and 12.5 and sacrificed at P0, when the organization of the OB is almost complete. We show that BrdU+ cells born at E11.5 and E12.5 co-localize with GFP at the MCL, whereas cells born at E10.5 do not (Fig. 3A-C). In order to identify the subpopulation of MCs whose localization is affected by Tag1 loss, we performed a thorough analysis for BrdU (injected at E11.5 and E12.5, respectively) on P0 OB sections from WT and Tag1−/− mice (Fig. 3D,E, Fig. S3). A severe reduction (39.4%) in the number of E11.5-born cells is detected at the MCL of the MOB in Tag1−/− animals (Fig. 3F), whereas this was not the case for E12.5-born MCs (Fig. S3). Our results suggest that MCs of the MOB, born at E11.5, comprise the main population affected by Tag1 deficiency.

Fig. 3.

Fewer E11.5-born cells reach the MCL of the MOB in Tag1−/− animals, through mislocalization. (A-C) OB coronal sections from newborn Tag1loxP-GFP-loxP-DTA mice immunostained for GFP (green) and BrdU (red). BrdU injections were performed at E10.5 (A), 11.5 (B) and 12.5 (C) in pregnant animals. MCs born at E11.5 and 12.5 express Tag1 (arrowheads: double stained cells), in contrast to E10.5-born MCs. (D,E) BrdU immunostaining (green) in OB sections from P0 WT (D) and Tag1−/− (E) mice, BrdU-pulsed at E11.5 shows a reduction in Tag1−/− MCL. (F-I). Quantification of BrdU+ cells in OB sections from P0 (F,G) and E17.5 (H,I) mice (from rostral to caudal levels) showed a reduction in the MOB of Tag1−/− mice compared with WT controls (at P0: 42.3%, P=0.0205, n=6; at E17.5: 19.48%, P=0.0416, n=5), and increase in the AOB (at P0: 19.81%, P=0.0064, n=6; at E17.5: 37.6%, P=0.0149, n=5). (J) The total number of BrdU+ cells was quantified exclusively in caudal OB sections and no difference was detected (P>0.1, n=5) between genotypes. (K) Graphical representation of the dorsoventral distribution of BrdU+ cells in caudal OB cryosections from E17.5 WT and Tag1−/− mice (n=3) shows that whereas WT MCs are evenly distributed, in Tag1−/− mice, significantly more remain in dorsal areas. To-Pro-3 (blue) was used for nuclear staining. Error bars represent s.e.m. Scale bars: 75 μm in A-C; 150 μm in D,E. *P<0.05, **P<0.01.

Fig. 3.

Fewer E11.5-born cells reach the MCL of the MOB in Tag1−/− animals, through mislocalization. (A-C) OB coronal sections from newborn Tag1loxP-GFP-loxP-DTA mice immunostained for GFP (green) and BrdU (red). BrdU injections were performed at E10.5 (A), 11.5 (B) and 12.5 (C) in pregnant animals. MCs born at E11.5 and 12.5 express Tag1 (arrowheads: double stained cells), in contrast to E10.5-born MCs. (D,E) BrdU immunostaining (green) in OB sections from P0 WT (D) and Tag1−/− (E) mice, BrdU-pulsed at E11.5 shows a reduction in Tag1−/− MCL. (F-I). Quantification of BrdU+ cells in OB sections from P0 (F,G) and E17.5 (H,I) mice (from rostral to caudal levels) showed a reduction in the MOB of Tag1−/− mice compared with WT controls (at P0: 42.3%, P=0.0205, n=6; at E17.5: 19.48%, P=0.0416, n=5), and increase in the AOB (at P0: 19.81%, P=0.0064, n=6; at E17.5: 37.6%, P=0.0149, n=5). (J) The total number of BrdU+ cells was quantified exclusively in caudal OB sections and no difference was detected (P>0.1, n=5) between genotypes. (K) Graphical representation of the dorsoventral distribution of BrdU+ cells in caudal OB cryosections from E17.5 WT and Tag1−/− mice (n=3) shows that whereas WT MCs are evenly distributed, in Tag1−/− mice, significantly more remain in dorsal areas. To-Pro-3 (blue) was used for nuclear staining. Error bars represent s.e.m. Scale bars: 75 μm in A-C; 150 μm in D,E. *P<0.05, **P<0.01.

Tag1 loss could affect migration or guidance of MCs, the pool of MC progenitors (proliferation, specification or survival), post-mitotic differentiation, or a combination of the above. We found that the number of immature Tbr1+ projection neurons is unaffected at E14.5 and that limited cell death is detected in both genotypes by Caspase3 IHC at different developmental stages (E14.5-E18.5) (Fig. S4 and data not shown).

In order to test for mislocalization of MCs, we looked for E11.5-born MCs in the whole volume of the OB. Tag1 is not expressed in the AOB, which is located on the dorsolateral part of caudal OB. However, the number of E11.5-born (BrdU-labeled) cells found in this area was increased in Tag1−/− mice at P0 and E17.5 (Fig. 3G,I), although BrdU+ cells were reduced in the MCL of the MOB compared with WT animals (Fig. 3H,I). We quantified the total number of BrdU+ cells in coronal sections from the caudal OB (that contains regions of both the MOB and the AOB) of E17.5 mice from both genotypes and found no difference, suggesting that Tag1-deficient OB projection neurons are redistributed rather than missing (Fig. 3J). The dorsoventral distribution of BrdU+ cells in caudal OB cryosections from E17.5 WT and Tag1−/− mice shows that more Tag1-deficient MCs tend to remain in dorsal areas as opposed to WT cells, which are equally distributed in dorsal and ventral areas (Fig. 3K).

In order to locate E11.5-born projection neurons in the developing OB, we followed them with BrdU IHC (injection at E11.5) through the anteroposterior axis, including the rostral (containing only the MOB) and caudal (containing both MOB and AOB) regions. Coronal sections were subdivided into four quadrants. The percentages of BrdU+/Tbr1+ cells were quantified per OB quadrant and their distribution is shown in Fig. 4. At E14.5, the dorsoventral distribution of E11.5-born cells is not significantly different between Tag1−/− and WT controls (Fig. 4A,A′,E). However, at subsequent ages (E16.5, E17.5 and P0), a dorsolateral preference (where the developing AOB is located) of Tag1−/− cells becomes evident, in contrast to the uniform distribution in WT mice (Fig. 4B-D′). Quantification of the number of BrdU+/Tbr1+ cells in dorsal versus ventral areas confirmed their tendency to remain in dorsal areas in deficient OBs (Fig. 4F-H). Finally, quantification of the cells in sections containing exclusively the dorsal MOB (DMOB) or the AOB, showed that the increase observed in the dorsolateral quadrant in P0 Tag1−/− mice, results from their redistribution in the AOB (Fig. 4I). In accordance to this finding, few GFP+ cells were found mislocalized in the AOB of Tag1/−;GFP mice (Fig. 4J-L). The trajectories of OB projection neurons in the LOT and the PC display no gross differences in Tag1−/− mice compared with WT controls, as indicated by IHC for the cell adhesion molecules L1 and neuropilin 1(NP1, also known as Nrp1) (Fig. S5).

Fig. 4.

Tracking the location of Ε11.5-born MCs in the developing OB. (A-D′) Radial graphs showing the percentages of BrdU+/ Tbr1+ MCs in different quadrants of the entire OB. BrdU was injected at E11.5 in WT (A-D) and Tag1−/− (A′-D′) mice and the analysis was performed at E14.5 (P=0.5037, n=3, A,A′), E16.5 (P=0.0208, n=3, B,B′), E17.5 (P=0.0074, n=3, C,C′) and P0 (P=0.0034, n=4, D,D′). (E-H) Graphical representation of the dorsoventral distribution of E11.5-born MCs at E14.5-P0. (I) Percentage of double positive cells that are found in dorsal MOB (DMOB) in comparison with the AOB area in P0 animals. (J-L) Using Tag1loxP-GFP-loxP-DTA mice, we found that GFP+ cells are increased inside the AOB of Tag1−/− mice. GFP+ cells are indicated on the border (arrowheads) or inside the AOB (arrows). Error bars represent s.e.m. *P<0.05, **P<0.01. DM, dorsomedial; DL, dorsolateral; VM, ventromedial; VL, ventrolateral.

Fig. 4.

Tracking the location of Ε11.5-born MCs in the developing OB. (A-D′) Radial graphs showing the percentages of BrdU+/ Tbr1+ MCs in different quadrants of the entire OB. BrdU was injected at E11.5 in WT (A-D) and Tag1−/− (A′-D′) mice and the analysis was performed at E14.5 (P=0.5037, n=3, A,A′), E16.5 (P=0.0208, n=3, B,B′), E17.5 (P=0.0074, n=3, C,C′) and P0 (P=0.0034, n=4, D,D′). (E-H) Graphical representation of the dorsoventral distribution of E11.5-born MCs at E14.5-P0. (I) Percentage of double positive cells that are found in dorsal MOB (DMOB) in comparison with the AOB area in P0 animals. (J-L) Using Tag1loxP-GFP-loxP-DTA mice, we found that GFP+ cells are increased inside the AOB of Tag1−/− mice. GFP+ cells are indicated on the border (arrowheads) or inside the AOB (arrows). Error bars represent s.e.m. *P<0.05, **P<0.01. DM, dorsomedial; DL, dorsolateral; VM, ventromedial; VL, ventrolateral.

Tag1 is necessary for the tangential migration of E11.5-born projection neurons from the dorsolateral OB

MCs born at E11.5 are mislocalized in Tag1-deficient mice. We hypothesize that through their radial migration, E11.5-born cells occupy all quadrants of the developing OB by E14.5, and that later on, cells from the dorsolateral area need Tag1 to migrate towards the developing MCL of the MOB.

In order to test this hypothesis, we performed organotypic cultures of OB sections at E14.5 to examine the normal migration pattern of projection neurons detected dorsolaterally, and to investigate possible defects in this pattern in Tag1−/− mice. DiI tracing experiments (crystal placed in the dorsal part) showed that in WT slices, cells migrate tangentially towards more rostral regions; however in Tag1−/− slices, only a few cells have left the dorsal part after 2 days in culture, supporting our initial hypothesis (Fig. 5A). We then set up a transplantation system (Fig. 5B), where the transplants originated from the dorsal OB region (from lateral slices) of Tag1−/−;GFP or Tag1+/+;GFP mice and the recipients were WT or Tag1−/− (Fig. 5Da-d). In another set of transplantation experiments, donor cells were labeled by injecting pregnant donors with BrdU at E11.5 (Fig. 5Da′-d′).

Fig. 5.

Tag1+ OB projection neurons migrate tangentially from the dorsal part and their migration is defective in Tag1−/− mice. (A) Organotypic cultures of sagittal sections from E14.5 brains. A DiI crystal was placed in the dorsocaudal part of lateral OB vibratome sections as shown in the scheme, and the sections were cultured for two days. (B) Transplants were generated on sagittal E14.5 sections as shown in the scheme, with donor cells expressing GFP under the Tag1 gene promoter or BrdU-pulsed at E11.5. (C) WT GFP+ and BrdU+ (BrdU injection at E11.5) cells were transplanted into WT recipients and immunostained for GFP and Tbr1 or GFP and BrdU. Migrating cells (GFP+) are projection neurons (co-localization with Tbr1, upper images) and are born at E11.5 (co-localization with BrdU, lower images). (D) All possible combinations of transplants are presented immunostained for GFP or BrdU (shown in green). WT GFP+ or BrdU+ cells migrated on either WT or Tag1−/− recipients, with some cells reaching the anterior and ventral part of the section (arrowheads in a,a′,b,b′), but only a few Tag1−/−;GFP+ cells migrated on either WT or Tag1−/− recipients (# in c,d). Asterisks in a,a′,b,c′,d indicate a migratory stream, occasionally detected. To-Pro-3 (blue) was used for nuclear staining. Dashed white lines represent the transplant limits. (E,F) Quantification of the average number of migrating neurons and the average distance travelled. n=7-9, *P<0.05, **P<0.01, ***P<0.001. Error bars represent s.e.m. Scale bars: 300 μm in A, 150 μm in C,D.

Fig. 5.

Tag1+ OB projection neurons migrate tangentially from the dorsal part and their migration is defective in Tag1−/− mice. (A) Organotypic cultures of sagittal sections from E14.5 brains. A DiI crystal was placed in the dorsocaudal part of lateral OB vibratome sections as shown in the scheme, and the sections were cultured for two days. (B) Transplants were generated on sagittal E14.5 sections as shown in the scheme, with donor cells expressing GFP under the Tag1 gene promoter or BrdU-pulsed at E11.5. (C) WT GFP+ and BrdU+ (BrdU injection at E11.5) cells were transplanted into WT recipients and immunostained for GFP and Tbr1 or GFP and BrdU. Migrating cells (GFP+) are projection neurons (co-localization with Tbr1, upper images) and are born at E11.5 (co-localization with BrdU, lower images). (D) All possible combinations of transplants are presented immunostained for GFP or BrdU (shown in green). WT GFP+ or BrdU+ cells migrated on either WT or Tag1−/− recipients, with some cells reaching the anterior and ventral part of the section (arrowheads in a,a′,b,b′), but only a few Tag1−/−;GFP+ cells migrated on either WT or Tag1−/− recipients (# in c,d). Asterisks in a,a′,b,c′,d indicate a migratory stream, occasionally detected. To-Pro-3 (blue) was used for nuclear staining. Dashed white lines represent the transplant limits. (E,F) Quantification of the average number of migrating neurons and the average distance travelled. n=7-9, *P<0.05, **P<0.01, ***P<0.001. Error bars represent s.e.m. Scale bars: 300 μm in A, 150 μm in C,D.

Migrating cells from dorsolateral OB (control experiment: donor, Tag1+/+;GFP; recipient, WT) are E11.5-born projection neurons expressing Tag1, as shown by co-localization experiments for GFP and Tbr1 or GFP and BrdU (Fig. 5C). All possible combinations of transplants and recipients were performed, cultured for 3 days and subjected to further sectioning and immunofluorescence for GFP (Fig. 5Da-d) or BrdU (Fig. 5Da′-d′). We observed that Tag1+/+ GFP+ or BrdU+ cells migrate normally on WT or Tag1−/− recipients, following a tangential pattern of migration towards rostral and ventral areas (arrowheads in Fig. 5Da,a′,b,b′), suggesting that Tag1 expression in the recipient is not a prerequisite for proper migration. We combined transplants from BrdU-pulsed Tag1−/− or Tag1−/−;GFP mice on WT or Tag1−/− recipients (Fig. 5Dc,d,c′d′). Donor cells in both cases display severe migration defects. Specifically, only a few BrdU+ cells migrate from the transplant area at short distances (quantified in Fig. 5E,F), whereas only a small group of GFP+ cells migrate in the correct tangential direction (# in Fig. 5Dc,d). These neurons might comprise later-born populations, i.e. at E12.5 or E13.5. Therefore, Tag1 expression on projection neurons from dorsolateral OB (born at E11.5) is indispensable for their proper tangential migration and positioning in the developing MCL of the MOB.

Tag1-deficient mice display severe alterations in their olfactory ability and neuronal activation profile

In order to evaluate the olfactory ability of Tag1−/− mice, we used a battery of behavioral tests. The buried food test is widely used for the evaluation of the ability of rodents to smell volatile odors. In this test, Tag1−/− mice performed well, displaying an insignificant increase in their latency to uncover the hidden food pellet, compared with WT animals (Fig. 6A). In the social interaction test (Fig. 6B), WT, Tag1+/− and Tag1−/− male mice were timed smelling an anesthetized conspecific control animal on repeated trials. All groups of animals progressively reduced the time spent exploring the anesthetized conspecific (trials 1-4), indicative of efficient perception and learning of its odor, although exploration time was increased when the anesthetized animal was replaced with a new one on trial 5. The olfactory memory task was performed 24 h later: experimental animals were exposed to two anesthetized mice, the familiar one (used for trials 1-4 of the learning phase) and a new conspecific animal. Olfactory discrimination index (ODI), showed that Tag1−/− mice display a severe defect in long-term social memory (Fig. 6C). The observation of ODI on Tag1+/− animals is important, as it appears significantly lower than WT but greater than Tag1−/− mice (Fig. 6C). These mice display long-term memory formation but perform worse in this social test that WT, suggesting a correlation with the numerical defect on MCs.

Fig. 6.

Tag1-deficient mice display altered olfaction and neuronal activation profile. (A) Time taken for WT and Tag1−/− mice to locate food in the buried food test, n=9. (B) Social interaction times during the olfactory learning task in WT, Tag1+/− and Tag1−/− mice. One-way ANOVA with repeated measurements (trials), n=4. On trial 5, we used a new conspecific animal. (C) Social memory test. The olfactory discrimination index (ODI) was calculated. Tag1−/− mice show ODI close to 0 (P=0.0054, n=4). Tag1+/− animals present an intermediate ODI between WT and Tag1−/− (P=0.0138 compared with WT, n=4; P=0.0407 compared with Tag1−/−, n=4). (D) Habituation/dishabituation test scheme. (E) Habituation/dishabituation test with social and non-social odors reveals an altered ability of Tag1−/− and Tag1+/− mice to detect volatile odors. For each odor tested, one-way ANOVA with repeated measures (trials), trial×genotype interaction, P<0.01, n=9 for WT and Tag1−/−, n=8 for Tag1+/−. Red asterisks represent statistically significant difference between WT and KO, Green asterisks between WT and HET and # between HET and KO. Colored dashed lines and bars denote three trials for a specific odor. S.O., social odor. (F-H) Neuronal activation was measured in WT and Tag1−/− mice, using c-Fos antibody after exposure of the animals to peanut butter. Reduced number of activated neurons was detected in the MCL (30%, P=0.0291, n=4) and the PC (40.10%, P=0.0028, n=4) of adult Tag1−/− mice. No difference was detected in the AOB after nasal exposure to female urine (P=0.5, n=4). Error bars represent s.e.m. *P<0.05, **P<0.01, ***P<0.001.

Fig. 6.

Tag1-deficient mice display altered olfaction and neuronal activation profile. (A) Time taken for WT and Tag1−/− mice to locate food in the buried food test, n=9. (B) Social interaction times during the olfactory learning task in WT, Tag1+/− and Tag1−/− mice. One-way ANOVA with repeated measurements (trials), n=4. On trial 5, we used a new conspecific animal. (C) Social memory test. The olfactory discrimination index (ODI) was calculated. Tag1−/− mice show ODI close to 0 (P=0.0054, n=4). Tag1+/− animals present an intermediate ODI between WT and Tag1−/− (P=0.0138 compared with WT, n=4; P=0.0407 compared with Tag1−/−, n=4). (D) Habituation/dishabituation test scheme. (E) Habituation/dishabituation test with social and non-social odors reveals an altered ability of Tag1−/− and Tag1+/− mice to detect volatile odors. For each odor tested, one-way ANOVA with repeated measures (trials), trial×genotype interaction, P<0.01, n=9 for WT and Tag1−/−, n=8 for Tag1+/−. Red asterisks represent statistically significant difference between WT and KO, Green asterisks between WT and HET and # between HET and KO. Colored dashed lines and bars denote three trials for a specific odor. S.O., social odor. (F-H) Neuronal activation was measured in WT and Tag1−/− mice, using c-Fos antibody after exposure of the animals to peanut butter. Reduced number of activated neurons was detected in the MCL (30%, P=0.0291, n=4) and the PC (40.10%, P=0.0028, n=4) of adult Tag1−/− mice. No difference was detected in the AOB after nasal exposure to female urine (P=0.5, n=4). Error bars represent s.e.m. *P<0.05, **P<0.01, ***P<0.001.

In order to further elucidate the olfactory ability of Tag1-deficient animals we used a more specific odor discrimination test, namely the olfactory ‘habituation/dishabituation’ assay, using different social (female odor and urine) and non-social [H2O, peanut butter (PB) and 2-methylbutyric acid (2MB)] odors (Fig. 6D). The repertoire of odorants contained both attractive (PB, social odors) and repulsive (2MB) cues. In all odorants used, the odor preference of Tag1−/− animals was significantly reduced compared with WT controls (Fig. 6E; Movie 1). The most profound difference occurred in the attractive PB odor and in female social odors. These results reveal the impaired olfactory ability of Tag1−/− mice regarding the recognition and discrimination of odor cues. Tag1+/− mice display defective odor recognition and discrimination but to a lesser extent compared with Tag1−/− mice (Fig. 6E). To exclude the possibility that our experimental Tag1−/− and Tag1+/− animals develop this behavior as a result of anxiety-related defects we performed an elevated plus maze test which revealed no difference between genotypes (Fig. S6).

We next investigated whether the reduction in the number of MCs of the MOB from Tag1−/− mice results in alterations of the neuronal activation profile that could account for these behavioral defects. No difference was detected in the morphology and neuronal numbers of PC between Tag1−/− and WT mice (Fig. S7). Different groups of male WT and Tag1−/− mice were exposed to either PB for the stimulation of the MOS, or female urine for the stimulation mainly of the AOS (Fig. S8A). We used c-Fos (also known as Fos), encoded by an immediate early gene, to monitor for neuronal activity as its expression increases after synaptic stimulation (Datiche et al., 2001; Roullet et al., 2005).

After PB stimulation, the number of c-Fos+ cells was quantified in 4 different areas (dorsal, lateral, ventral and medial; Fig. 6F, Fig. S8B-E) of the MCL (MOB) as well as in the PC of WT and Tag1−/− animals (Fig. 6G, Fig. S8F-G). There is a significant reduction (30% and 40%, respectively) in the number of neurons activated in the MOB and PC of Tag1−/− mice compared with controls (Fig. 6F,G). 74% of MCs that are born at E11.5 are activated after PB stimulation, which represent 44% of total c-Fos+ cells (Fig. S8M). We propose that the olfaction defects of Tag1−/− mice could result from the reduced output of MCs towards the olfactory cortex, after exposure to volatile odor stimuli. No difference between genotypes was found in the c-Fos profile in the OE, suggesting that activity of the initial station of the odorant information route is not affected (Fig. S8J-L). Furthermore, no difference was detected in the activation profile or the total number of GABA+ cells in the granule cell layer in Tag1−/− mice (Fig. S8N,O).

In order to activate the AOB of adult male mice we used a female urine cue, containing non-volatile signals (in addition to volatile molecules), which are detected by the vomeronasal organ. IHC for c-Fos in adult mice showed that there is no significant difference in the number of activated neurons in Tag1−/− animals, suggesting intact AOB responsiveness after stimulation (Fig. 6H, Fig. S8H,I).

Despite progress on the elucidation of the olfactory receptor map (Bozza et al., 2009; Mori et al., 1999; Mori et al., 2006), little is known about the organization, function and circuit integration of the OB projection neurons. The aim of this study is to shed light on the organization of projection neurons inside the OB and show that alterations in this organization affect olfactory function in mice lacking Tag1 expression.

Tag1 is strongly expressed by immature as well as migrating MCs and TCs of the developing OB. Furthermore, it is expressed by a subset of OSNs, although its absence in Tag1−/− mice does not affect the number of projection to the glomerular layer of the OB. We detected a significant reduction in the number of mature MCs in adult Tag1−/− mice, whereas TCs were unaffected. The projections of MCs into the LOT of Tag1−/− mice towards the olfactory cortex led to neither defects in their fasciculation nor a reduction in the thickness of the axonal tract, suggesting that the redistributed MCs are still able to project normally. MCs are classified according to their birth dates as either early-born (at E10 or earlier) and late-born (at E12 or later). These are localized in the dorsomedial and ventrolateral MCL, respectively, whereas MCs born at E11.5 are uniformly distributed inside the MCL (Imamura et al., 2011). We showed that populations of late-born and intermediate MCs born at E11.5 express Tag1. E11.5-born MCs undergo proper radial migration and thus are uniformly distributed by E14.5 in Tag1/− mice, but are unable to change from a radial to a tangential-like route and subsequently migrate towards the anterior and ventral areas of the developing OB. Transplantation experiments suggest a cell autonomous defect.

Our results are in line with a proposed role of Tag1 in facilitating tangential migration in other neuronal systems, such as the precerebellar nuclei (Kyriakopoulou et al., 2002; Denaxa et al., 2005). There, Tag1 mediates migration through chain interaction among Tag1-expressing axons and soma-axonal appositions between Tag1+ cells and Tag1 axonal scaffold. Imamura et al. (2011) showed that E12.5-born MCs migrate tangentially in contact with Ocam+ preexisting axons of the LOT. Tag1 protein is detected on the cell body of migrating MCs until E16.5, when their tangential migration is almost complete, suggesting that this migration might occur through a soma-axonal apposition mechanism between MC somata and preceding axonal tracts.

It is established that Tag1 exerts its function through either homophilic and/or heterophilic interactions, although our transplantation data indicates that a trans homophilic interaction is improbable in our system (Karagogeos, 2003; Traka et al., 2003; Lieberoth et al., 2009; Ma et al., 2008). Candidate molecules for heterophilic interactions could be the adhesion molecules L1 and Ocam, which are also expressed developmentally by the axons of OB projection neurons (Treloar et al., 2003; Inaki et al., 2004).

Do the molecular deficits present in the MOS of Tag1−/− mice account for any behavioral defects related to olfaction? We have shown that homozygous mutants display a series of behavioral alterations such as learning and memory defects as well as motor coordination and balance abnormalities, resulting from the molecular disorganization of myelinated fibers (Traka et al., 2003; Savvaki et al., 2008, 2010). When the juxtaparanodal domain organization of myelinated fibers was rescued in vivo, learning and memory defects that are dependent on hippocampal function were also rescued (Savvaki et al., 2010). Our present analysis reveals that Tag1−/− mice are able to smell volatile and non-volatile odors in the buried food and social interaction tests respectively, but show a profound disruption in odor discrimination and long-term social memory formation. Aged individuals and patients with neurodegenerative diseases have been reported to display hyposmia as well as impaired odor discrimination and identification, with underlying pathophysiological alterations throughout the olfactory system (Card et al., 1988; Attems et al., 2014). In a mouse model of Alzheimer's disease, denervation of the OB resulted in reduced amyloid plaque load in the OB, the neocortex and the hippocampus (Bibari et al., 2013). Inside the OB, amyloid precursor protein is predominantly expressed by MCs (Card et al., 1988). These studies emphasize the importance of the olfactory system and the crucial role that main projection neurons of the OB could play in the affected olfaction of patients with neurodegenerative diseases.

It is well known that the MOS and the AOS participate in social behaviors such as sex recognition of conspecifics, by detecting volatile and non-volatile molecules from urine or other body odorants, respectively (O'Connell and Meredith, 1984; Pankevich et al., 2004). The volatile sex-specific cues detected mainly by the MOS guide the animal towards the odorant source. After direct interaction with the cue, the AOS is activated, subsequently triggering different behavioral and neuroendocrine responses (Humphries et al., 1999; Pankevich et al., 2004; Hurst, 2009).

Olfactory ‘habituation/dishabituation’ odor discrimination assays showed that Tag1−/− animals were unable to discriminate among volatile stimuli (PB, social odor and 2MB). When we used female urine (containing also non-volatile cues), Tag1−/− mice performed better than with the previous stimuli, but still significantly worse than the control animals. This suggests that the AOS is at least partially functional, but the defects in MOS render deficient mice unable to discriminate cues. Tag1+/− mice display altered odor discrimination ability in this test, but they perform better than Tag1−/− animals, supporting the gene dose-dependent effect. Intact AOS activity is further supported by our results on neuronal activation. The social interaction test revealed that Tag1−/− mice are able to investigate social stimuli and develop short-term social memory, but display severe defects in long-term social memory. This deficit can be attributed to both hippocampal and olfactory discrimination defects. Our data from Tag1+/− mice, where myelinated fibers are not affected in the hippocampus and other brain regions (Savvaki et al., 2008) but long-term social memory is altered, support the contribution of defective olfactory discrimination in this genotype. Noack et al. (2010) showed that long-term social recognition in mice includes the neuronal activation of both the MOB and the AOB, contrary to rats, where the AOB is mainly activated. Furthermore, they showed that the non-volatile fraction (received by AOS) is sufficient for the formation of short-term social memory, whereas in the case of long-term social memory, both the volatile and non-volatile fractions are required.

Our results propose that MCs born at E11.5 are of major importance for the proper function of the OB. Our hypothesis is reinforced by a recent study showing that MCs are heterogeneous in terms of their biophysical properties and that even a small imbalance in this population could significantly influence OB function (Kollo et al., 2014). These researchers showed that approximately one third of MCs and TCs are found in a ‘silent’ state, with very low baseline activity. However, these cells, in contrast to ‘highly active’ projection neurons, can vigorously respond to odorant stimulation. Therefore, the 25% reduction of the otherwise uniformly distributed MC population, born at E11.5, could be correlated with the severe defects in olfactory recognition and discrimination that are observed in Tag1−/− mice. Although the gross morphology of the OE as well as the organization of the glomerular map is not altered in Tag1/− mice, subtle alterations in OE function, i.e. in its electrophysiological properties, cannot be excluded.

This is the first study that proposes a molecular cue involved in the organization and development of projection neuron circuitry in the OB. Tag1 is a key molecule that coordinates the integration of E11.5-born MCs inside the developing MC map of the MOB. Its absence results in aberrant decoding of odorant information towards olfactory cortices, ending up in a disturbed behavioral phenotype regarding olfaction (Fig. 7).

Fig. 7.

Model of olfactory bulb organization and output to olfactory cortices in WT and Tag1−/− mice. The absence of Tag1 in Tag1−/− mice results in a reduced number of MCs found in the MCL of the main olfactory bulb which send reduced output to the piriform corte. Subsequently, reduced activation of the piriform cortex significantly affects olfactory behavior. PC, Piriform cortex.

Fig. 7.

Model of olfactory bulb organization and output to olfactory cortices in WT and Tag1−/− mice. The absence of Tag1 in Tag1−/− mice results in a reduced number of MCs found in the MCL of the main olfactory bulb which send reduced output to the piriform corte. Subsequently, reduced activation of the piriform cortex significantly affects olfactory behavior. PC, Piriform cortex.

Animals

The generation of Tag1−/− mice has been described (Fukamauchi et al., 2001; Traka et al., 2003; Savvaki et al., 2008). All mice were kept as heterozygote breeding pairs and the genotypes were confirmed by PCR. Genetically modified Tag1loxP-GFP-loxP-DTA mice were generated using BAC technology. These mice express GFP under the Tag1 promoter without affecting endogenous expression. They were crossed with Tag1+/− and Tag1−/− mice in order to obtain the genotypes of interest (Tag1+/+;GFP, Tag1−/−;GFP). E0.5: day of the vaginal plug. All mice were of C57BL6/SV129 background. Housing and animal procedures used were according to the European Union policy (Directive 86/609/EEC) and institutionally approved protocols.

Immunohistochemistry

Immunohistochemistry (IHC) was performed on cryosections as described (Vidaki et al., 2012) with the exception of adult OB cryosections (14 μm-thick), which were post-fixed in ice-cold acetone for 10 min, blocked (in 5% BSA in 0.1 M PBS) and incubated with primary and secondary antibodies in 5% BSA (Sigma-Aldrich) 0.5% Triton-X in 0.1 M PBS. Antibodies used: rabbit polyclonal anti-Tbr1 (T-box brain 1, cat. no. Ab31940, Abcam, 1:600), rat polyclonal anti-GFP (cat. no. 04404-84, Nacalai Tesque, 1:1000), rabbit polyclonal anti-Tag1 (TG3, 1:1000, Traka et al., 2003), rabbit anti-Tbx21 (T-box transcription factor, kind gift from Y. Yoshihara, 1:10,000) rabbit polyclonal OCAM (kind gift from Y. Yoshihara, 1:1000), guinea pig a-mOR-EG (kind gift fromY. Yoshihara, 1:1000), guinea pig a-OR-I7 (kind gift from Y. Yoshihara, 1:5000), rabbit a-Kirrel2 [protocol described in Serizawa et al. (2006), a kind gift from H. Sakano, 1:1000] and fluorochrome-labeled secondary antibodies Alexa Fluor 488 and 555 (cat. no. A11034, A21429, A21435, Molecular Probes, 1:800). To-Pro 3 iodide (Invitrogen) was used for the visualization of the nuclei. Hematoxylin/Eosin and Nissl staining was used for morphological characterization.

BrdU injection and immunodetection

BrdU injection and immunodetection were performed as previously described (Vidaki et al., 2012).

OB transplants and immunofluorescence

Brains from E14.5 embryos were dissected in L15 medium (Gibco) containing 100 μ/ml pen/strep (Biosera). Cortices with OBs were embedded in 3% low melting agarose (LMA, Lonza) in L15. Sagittal vibratome sections 250 μm thick were placed on BIOPORE membranes (Millipore) in DMEM/ F12A (DMEM/F12, 0.6% glucose, 5% FBS, 100 μ/ml pen/strep, 1× Glutamax, 1% N2 supplement). Sections were incubated at 37°C, 5% CO2 for 1 h. Afterwards, the transplant was excised from the donor OB and carefully placed on the recipient (after removing the appropriate part). The medium was replaced by fresh Neurobasal/B27 (Neurobasal, 0.6% glucose, 100 μ/ml pen/strep, 1× Glutamax, 2% B27 supplement) and sections were incubated for 3 days at 37°C, 5% CO2. For immunofluorescence, sections were fixed for 1 h in 4% PFA and re-sectioned (60 μm) in 3% LMA in 0.1 M PBS, washed in 0.1 M PBS containing 0.1% Triton-X (1× PBT) and blocked in 1% FBS in 1× PBT for 1 h at room temperature. Finally, sections were incubated with primary and the secondary antibodies in blocking solution, washed and subbed in 0.2% (w/v) porcine skin gelatin in 50 mM Tris pH 7.5. Slides were mounted with MOWIOL and observed in a Leica TS2 confocal microscope. Antibodies used: rat polyclonal anti-GFP (cat. no. 04404-84, Nacalai Tesque, 1:1000), rabbit polyclonal anti-GFP (cat. no. 721, Minotech Biotechnology, 1:50,000), rabbit polyclonal anti-Tbr-1 (cat. no. Ab31940, Abcam, 1:600), rat anti-BrdU (cat. no. OBT0030CX, IgG, AbD Serotec, 1:1000) and fluorochrome-labeled secondary antibodies Alexa Fluor 488 and 555 (Molecular Probes, 1:800).

In situ hybridization (ISH) and fluorescent ISH

ISH was performed on coronal cryosections as described (Denaxa et al., 2001; Ishii et al., 2004). Antisense probes for mouse Tag1 (Denaxa et al., 2001), Omp, Gap43, Mash1 (kind gifts from P. Mombaerts) and Ngn1 (kind gift from F. Guillemot) were used.

Behavioral analysis

Buried food test

Two-month old (mo) male WT and Tag1−/− mice (n=6) were assayed for their ability to locate a hidden, highly palatable food piece (cheese chip) underneath the cage bedding as described (Yang and Crawley, 2009) with minor modifications: a cheese chip was hidden ∼1 cm beneath the cage bedding in a random corner of the experimental cage. We consider that the animal had found the cheese chip when it held it in its front paws and began to eat it.

Social interaction test

Social interaction test was performed as previously reported with minor modifications (Kogan et al., 2000). The olfactory discrimination index (ODI) indicates the memory ability of the animals to recall a previously experienced olfactory stimulus (see also supplementary materials and methods).

Habituation/dishabituation test

The test was performed as previously described with minor modifications (Basic Protocol 2) (Yang and Crawley, 2009); see also supplementary materials and methods.

Elevated plus maze

The test was performed as previously described (Stamatakis et al., 2015).

Neuronal activity experiment, c-Fos labeling and analysis

MOS, AOS activation protocol

Two-month old male mice were used. For MOS activation, a stimulus of 10% w/v PB (dissolved in odorless paraffin oil) or H2O was used. Three to four hours before the test initiation, each animal was transferred in a clean cage (with filter cap) without food, H2O or bedding under the fume hood. Afterwards, the PB odor was introduced into the cage through two cotton tip applicators (each containing 150 μl of PB) and mice were left to smell the stimulus for 90 min. Then, experimental animals were deeply anesthetized, perfused with 4% PFA and brains were dissected. For the AOS we followed the protocol from Pierman et al. (2008) and Veyrac et al. (2011) with minor modifications. After the habituation period, 30 μl of female urine (from a pool of female urine) was applied inside the nose of the subjects (or H2O in controls) and mice were then placed back in the filter cage for 90 min. The following steps were the same as in MOS activation protocol.

c-Fos labeling

A modified protocol was used for the IHC of c-Fos protein in adult OB and PC cryosections (Sundquist and Nisenbaum, 2005); see also supplementary materials and methods. For the c-Fos/BrdU co-labeling the c-Fos protocol preceded the BrdU protocol.

Image analysis and statistics

For the MOS activation analysis, serial sections were obtained at 10× and 40× magnification. At 40× magnification representative images from the MOB were acquired with the same settings for all animals. For PC and granule cell analysis, four images from serial sections (20× magnification) were used, spanning anterior to more posterior regions. For AOS activation analysis, 20× magnification confocal images were acquired and used from two regions of the AOB (anterior and posterior). In all cases, quantification was performed with ImageJ (National Institutes of Health).

Statistics/radial statistics analysis

Student's t-test and one-way ANOVA were used for the statistical analysis by GraphPad Prism version 5.00. Regarding radial statistics analysis, for all developmental stages we obtained sections representative of the entire OB tissue through the anteroposterior (AP) axis. At all developmental stages tested, 3-4 regions were used through the AP axis in both genotypes. Each section was divided into four different anatomical quadrants (dorsolateral, ventrolateral, ventromedial and dorsomedial). Then, double positive (Tbr1/BrdU) cells were manually counted and the total number of cells per quadrant as well as the total percentage of cells per quadrant was calculated for each section and for each age. Radial statistics graphs were generated on Grapher 8 (v8.7, Golden Software), using the average percentage of cells per quadrant (in all sections) for all animals of each genotype. Statistical analysis of dorsoventral positioning was performed using the mean percentages of cells found on dorsal (DL, DM) and ventral (VL, VM) quadrants in all sections of the OB for each animal and genotype (unpaired Student's t-test was performed using GraphPad Prism version 5.00).

The authors are grateful to Drs Sonia Garel (IBENS-INSERM-CNRS, Paris, France), Yoshihiro Yoshihara (RIKEN Brain Science Institute, Japan), H. Sakano (The University of Tokyo, Japan) and Francois Guillemot (NIMR-MRC, London) for providing reagents. We would also like to thank Drs Peter Mombaerts and Markella Katidou (Max Planck Research Unit for Neurogenetics, Frankfurt) for training MS in fluorescent ISH and providing reagents. We also thank Drs K. Kleopa, M. Strigini, L. Zoupi for their comments on the manuscript.

Author contributions

S.M. and B.G.G. contributed equally to this work. Both of them designed and performed experiments, analyzed data and wrote the paper. S.M. also supervised the work. S.A. performed some behavioral experiments and contributed to the analysis of the data. V.M. designed and generated the Tag1loxP-GFP-loxP-DTA mouse line. K.D. supervised the work and wrote the paper.

Funding

This work was supported by the European Commission FP7 programme ‘Translational Potential’ [contract number 285948], InnovCrete [316223], ARISTEIA I [Project 593 MyelinTag] and by Institute of Molecular Biology and Biotechnology intramural grants. B.G.G. has been the recipient of the Manasaki Fellowship and Medical School Fellowship of the University of Crete. V.M. is currently a postdoctoral fellow at the Koch Institute, MIT, Cambridge, MA, USA.

Attems
,
J.
,
Walker
,
L.
and
Jellinger
,
K. A.
(
2014
).
Olfactory bulb involvement in neurodegenerative diseases
.
Acta Neuropathol.
127
,
459
-
475
.
Baeriswyl
,
T.
and
Stoeckli
,
E. T.
(
2008
).
Axonin-1/Tag1 is required for pathfinding of granule cell axons in the developing cerebellum
.
Neural Dev.
3
,
7
.
Baum
,
M. J.
and
Kelliher
,
K. R.
(
2009
).
Complementary roles of the main and accessory olfactory systems in mammalian mate recognition
.
Annu. Rev. Physiol.
71
,
141
-
160
.
Bibari
,
O.
,
Lee
,
S.
,
Dickson
,
T. C.
,
Mitew
,
S.
,
Vickers
,
J. C.
and
Chuah
,
M. I.
(
2013
).
Denervation of the olfactory bulb leads to decreased Abeta plaque load in a transgenic mouse model of Alzheimer's disease
.
Curr. Alzheimer Res.
10
,
688
-
696
.
Blanchart
,
A.
,
De Carlos
,
J. A.
and
López-Mascaraque
,
L.
(
2006
).
Time frame of mitral cell development in the mice olfactory bulb
.
J. Comp. Neurol.
496
,
529
-
543
.
Bozza
,
T.
,
Vassalli
,
A.
,
Fuss
,
S.
,
Zhang
,
J.-J.
,
Weiland
,
B.
,
Pacifico
,
R.
,
Feinstein
,
P.
and
Mombaerts
,
P.
(
2009
).
Mapping of class I and class II odorant receptors to glomerular domains by two distinct types of olfactory sensory neurons in the mouse
.
Neuron
61
,
220
-
233
.
Buck
,
L.
and
Axel
,
R.
(
1991
).
A novel multigene family may encode odorant receptors: a molecular basis for odor recognition
.
Cell
65
,
175
-
187
.
Card
,
J. P.
,
Meade
,
R. P.
and
Davis
,
L. G.
(
1988
).
Immunocytochemical localization of the precursor protein for beta-amyloid in the rat central nervous system
.
Neuron
1
,
835
-
846
.
Datiche
,
F.
,
Roullet
,
F.
and
Cattarelli
,
M.
(
2001
).
Expression of Fos in the piriform cortex after acquisition of olfactory learning: an immunohistochemical study in the rat
.
Brain Res. Bull.
55
,
95
-
99
.
de Castro
,
F.
(
2009
).
Wiring olfaction: the cellular and molecular mechanisms that guide the development of synaptic connections from the nose to the cortex
.
Front. Neurosci.
3
,
52
.
Denaxa
,
M.
,
Chan
,
C. H.
,
Schachner
,
M.
,
Parnavelas
,
J. G.
and
Karagogeos
,
D.
(
2001
).
The adhesion molecule Tag1 mediates the migration of cortical interneurons from the ganglionic eminence along the corticofugal fiber system
.
Development
128
,
4635
-
4644
.
Denaxa
,
M.
,
Kyriakopoulou
,
K.
,
Theodorakis
,
K.
,
Trichas
,
G.
,
Vidaki
,
M.
,
Takeda
,
Y.
,
Watanabe
,
K.
and
Karagogeos
,
D.
(
2005
).
The adhesion molecule Tag1 is required for proper migration of the superficial migratory stream in the medulla but not of cortical interneurons
.
Dev. Biol.
288
,
87
-
99
.
Fukamauchi
,
F.
,
Aihara
,
O.
,
Wang
,
Y.-J.
,
Akasaka
,
K.
,
Takeda
,
Y.
,
Horie
,
M.
,
Kawano
,
H.
,
Sudo
,
K.
,
Asano
,
M.
,
Watanabe
,
K.
, et al.
(
2001
).
Tag1-deficient mice have marked elevation of adenosine A1 receptors in the hippocampus
.
Biochem. Biophys. Res. Commun.
281
,
220
-
226
.
Furley
,
A. J.
,
Morton
,
S. B.
,
Manalo
,
D.
,
Karagogeos
,
D.
,
Dodd
,
J.
and
Jessell
,
T. M.
(
1990
).
The axonal glycoprotein Tag1 is an immunoglobulin superfamily member with neurite outgrowth-promoting activity
.
Cell
61
,
157
-
170
.
Hinds
,
J. W.
(
1968a
).
Autoradiographic study of histogenesis in the mouse olfactory bulb. I. Time of origin of neurons and neuroglia
.
J. Comp. Neurol.
134
,
287
-
304
.
Hinds
,
J. W.
(
1968b
).
Autoradiographic study of histogenesis in the mouse olfactory bulb. II. Cell proliferation and migration
.
J. Comp. Neurol.
134
,
305
-
321
.
Humphries
,
R. E.
,
Robertson
,
D. H. L.
,
Beynon
,
R. J.
and
Hurst
,
J. L.
(
1999
).
Unravelling the chemical basis of competitive scent marking in house mice
.
Anim. Behav.
58
,
1177
-
1190
.
Hurst
,
J. L.
(
2009
).
Female recognition and assessment of males through scent
.
Behav. Brain Res.
200
,
295
-
303
.
Imamura
,
F.
and
Greer
,
C. A.
(
2013
).
Pax6 regulates Tbr1 and Tbr2 expressions in olfactory bulb mitral cells
.
Mol. Cell. Neurosci.
54
,
58
-
70
.
Imamura
,
F.
,
Ayoub
,
A. E.
,
Rakic
,
P.
and
Greer
,
C. A.
(
2011
).
Timing of neurogenesis is a determinant of olfactory circuitry
.
Nat. Neurosci.
14
,
331
-
337
.
Inaki
,
K.
,
Nishimura
,
S.
,
Nakashiba
,
T.
,
Itohara
,
S.
and
Yoshihara
,
Y.
(
2004
).
Laminar organization of the developing lateral olfactory tract revealed by differential expression of cell recognition molecules
.
J. Comp. Neurol.
479
,
243
-
256
.
Ishii
,
T.
,
Omura
,
M.
and
Mombaerts
,
P.
(
2004
).
Protocols for two- and three-color fluorescent RNA in situ hybridization of the main and accessory olfactory epithelia in mouse
.
J. Neurocytol.
33
,
657
-
669
.
Kaneko-Goto
,
T.
,
Yoshihara
,
S.-i.
,
Miyazaki
,
H.
and
Yoshihara
,
Y.
(
2008
).
BIG-2 mediates olfactory axon convergence to target glomeruli
.
Neuron
57
,
834
-
846
.
Karagogeos
,
D.
(
2003
).
Neural GPI-anchored cell adhesion molecules
.
Front. Biosci.
8
,
s1304
-
s1320
.
Kogan
,
J. H.
,
Frankland
,
P. W.
and
Silva
,
A. J.
(
2000
).
Long-term memory underlying hippocampus-dependent social recognition in mice
.
Hippocampus
10
,
47
-
56
.
Kollo
,
M.
,
Schmaltz
,
A.
,
Abdelhamid
,
M.
,
Fukunaga
,
I.
and
Schaefer
,
A. T.
(
2014
).
‘Silent’ mitral cells dominate odor responses in the olfactory bulb of awake mice
.
Nat. Neurosci.
17
,
1313
-
1315
.
Kyriakopoulou
,
K.
,
de Diego
,
I.
,
Wassef
,
M.
and
Karagogeos
,
D.
(
2002
).
A combination of chain and neurophilic migration involving the adhesion molecule Tag1 in the caudal medulla
.
Development
129
,
287
-
296
.
Leinwand
,
S. G.
and
Chalasani
,
S. H.
(
2011
).
Olfactory networks: from sensation to perception
.
Curr. Opin. Genet. Dev.
21
,
806
-
811
.
Lieberoth
,
A.
,
Splittstoesser
,
F.
,
Katagihallimath
,
N.
,
Jakovcevski
,
I.
,
Loers
,
G.
,
Ranscht
,
B.
,
Karagogeos
,
D.
,
Schachner
,
M.
and
Kleene
,
R.
(
2009
).
Lewis(x) and alpha2,3-sialyl glycans and their receptors Tag1, Contactin, and L1 mediate CD24-dependent neurite outgrowth
.
J. Neurosci.
29
,
6677
-
6690
.
Ma
,
Q.-H.
,
Futagawa
,
T.
,
Yang
,
W.-L.
,
Jiang
,
X.-D.
,
Zeng
,
L.
,
Takeda
,
Y.
,
Xu
,
R.-X.
,
Bagnard
,
D.
,
Schachner
,
M.
,
Furley
,
A. J.
, et al.
(
2008
).
A TAG1-APP signalling pathway through Fe65 negatively modulates neurogenesis
.
Nat. Cell Biol.
10
,
283
-
294
.
Mombaerts
,
P.
(
2001
).
How smell develops
.
Nat. Neurosci.
4
Suppl.
,
1192
-
1198
.
Mombaerts
,
P.
,
Wang
,
F.
,
Dulac
,
C.
,
Chao
,
S. K.
,
Nemes
,
A.
,
Mendelsohn
,
M.
,
Edmondson
,
J.
and
Axel
,
R.
(
1996
).
Visualizing an olfactory sensory map
.
Cell
87
,
675
-
686
.
Mori
,
K.
,
Nagao
,
H.
and
Yoshihara
,
Y.
(
1999
).
The olfactory bulb: coding and processing of odor molecule information
.
Science
286
,
711
-
715
.
Mori
,
K.
,
Takahashi
,
Y. K.
,
Igarashi
,
K. M.
and
Yamaguchi
,
M.
(
2006
).
Maps of odorant molecular features in the Mammalian olfactory bulb
.
Physiol. Rev.
86
,
409
-
433
.
Mouret
,
A.
,
Murray
,
K.
and
Lledo
,
P.-M.
(
2009
).
Centrifugal drive onto local inhibitory interneurons of the olfactory bulb
.
Ann. N. Y. Acad. Sci.
1170
,
239
-
254
.
Noack
,
J.
,
Richter
,
K.
,
Laube
,
G.
,
Haghgoo
,
H. A.
,
Veh
,
R. W.
and
Engelmann
,
M.
(
2010
).
Different importance of the volatile and non-volatile fractions of an olfactory signature for individual social recognition in rats versus mice and short-term versus long-term memory
.
Neurobiol. Learn. Mem.
94
,
568
-
575
.
O'Connell
,
R. J.
and
Meredith
,
M.
(
1984
).
Effects of volatile and nonvolatile chemical signals on male sex behaviors mediated by the main and accessory olfactory systems
.
Behav. Neurosci.
98
,
1083
-
1093
.
Pankevich
,
D. E.
,
Baum
,
M. J.
and
Cherry
,
J. A.
(
2004
).
Olfactory sex discrimination persists, whereas the preference for urinary odorants from estrous females disappears in male mice after vomeronasal organ removal
.
J. Neurosci.
24
,
9451
-
9457
.
Pierman
,
S.
,
Douhard
,
Q.
and
Bakker
,
J.
(
2008
).
Evidence for a role of early oestrogens in the central processing of sexually relevant olfactory cues in female mice
.
Eur. J. Neurosci.
27
,
423
-
431
.
Roullet
,
F.
,
Datiche
,
F.
,
Liénard
,
F.
and
Cattarelli
,
M.
(
2005
).
Learning-stage dependent Fos expression in the rat brain during acquisition of an olfactory discrimination task
.
Behav. Brain Res.
157
,
127
-
137
.
Sakano
,
H.
(
2010
).
Neural map formation in the mouse olfactory system
.
Neuron
67
,
530
-
542
.
Savvaki
,
M.
,
Panagiotaropoulos
,
T.
,
Stamatakis
,
A.
,
Sargiannidou
,
I.
,
Karatzioula
,
P.
,
Watanabe
,
K.
,
Stylianopoulou
,
F.
,
Karagogeos
,
D.
and
Kleopa
,
K. A.
(
2008
).
Impairment of learning and memory in Tag1 deficient mice associated with shorter CNS internodes and disrupted juxtaparanodes
.
Mol. Cell. Neurosci.
39
,
478
-
490
.
Savvaki
,
M.
,
Theodorakis
,
K.
,
Zoupi
,
L.
,
Stamatakis
,
A.
,
Tivodar
,
S.
,
Kyriacou
,
K.
,
Stylianopoulou
,
F.
and
Karagogeos
,
D.
(
2010
).
The expression of Tag1 in glial cells is sufficient for the formation of the juxtaparanodal complex and the phenotypic rescue of tag-1 homozygous mutants in the CNS
.
J. Neurosci.
30
,
13943
-
13954
.
Serizawa
,
S.
,
Miyamichi
,
K.
,
Takeuchi
,
H.
,
Yamagishi
,
Y.
,
Suzuki
,
M.
and
Sakano
,
H.
(
2006
).
A neuronal identity code for the odorant receptor-specific and activity-dependent axon sorting
.
Cell
127
,
1057
-
1069
.
Stamatakis
,
A.
,
Kalpachidou
,
T.
,
Raftogianni
,
A.
,
Zografou
,
E.
,
Tzanou
,
A.
,
Pondiki
,
S.
and
Stylianopoulou
,
F.
(
2015
).
Rat dams exposed repeatedly to a daily brief separation from the pups exhibit increased maternal behavior, decreased anxiety and altered levels of receptors for estrogens (ERalpha, ERbeta), oxytocin and serotonin (5-HT1A) in their brain
.
Psychoneuroendocrinology
52
,
212
-
228
.
Sundquist
,
S. J.
and
Nisenbaum
,
L. K.
(
2005
).
Fast Fos: rapid protocols for single- and double-labeling c-Fos immunohistochemistry in fresh frozen brain sections
.
J. Neurosci. Methods
141
,
9
-
20
.
Traka
,
M.
,
Goutebroze
,
L.
,
Denisenko
,
N.
,
Bessa
,
M.
,
Nifli
,
A.
,
Havaki
,
S.
,
Iwakura
,
Y.
,
Fukamauchi
,
F.
,
Watanabe
,
K.
,
Soliven
,
B.
, et al.
(
2003
).
Association of Tag1 with Caspr2 is essential for the molecular organization of juxtaparanodal regions of myelinated fibers
.
J. Cell Biol.
162
,
1161
-
1172
.
Treloar
,
H. B.
,
Gabeau
,
D.
,
Yoshihara
,
Y.
,
Mori
,
K.
and
Greer
,
C. A.
(
2003
).
Inverse expression of olfactory cell adhesion molecule in a subset of olfactory axons and a subset of mitral/tufted cells in the developing rat main olfactory bulb
.
J. Comp. Neurol.
458
,
389
-
403
.
Veyrac
,
A.
,
Wang
,
G.
,
Baum
,
M. J.
and
Bakker
,
J.
(
2011
).
The main and accessory olfactory systems of female mice are activated differentially by dominant versus subordinate male urinary odors
.
Brain Res.
1402
,
20
-
29
.
Vidaki
,
M.
,
Tivodar
,
S.
,
Doulgeraki
,
K.
,
Tybulewicz
,
V.
,
Kessaris
,
N.
,
Pachnis
,
V.
and
Karagogeos
,
D.
(
2012
).
Rac1-dependent cell cycle exit of MGE precursors and GABAergic interneuron migration to the cortex
.
Cereb. Cortex
22
,
680
-
692
.
Whitman
,
M. C.
and
Greer
,
C. A.
(
2009
).
Adult neurogenesis and the olfactory system
.
Prog. Neurobiol.
89
,
162
-
175
.
Wolfer
,
D. P.
,
Henehan-Beatty
,
A.
,
Stoeckli
,
E. T.
,
Sonderegger
,
P.
and
Lipp
,
H.-P.
(
1994
).
Distribution of Tag1/axonin-1 in fibre tracts and migratory streams of the developing mouse nervous system
.
J. Comp. Neurol.
345
,
1
-
32
.
Wolman
,
M. A.
,
Sittaramane
,
V.
,
Essner
,
J. J.
,
Yost
,
H. J.
,
Chandrasekhar
,
A.
and
Halloran
,
M. C.
(
2008
).
Transient axonal glycoprotein-1 (Tag1) and laminin-alpha1 regulate dynamic growth cone behaviors and initial axon direction in vivo
.
Neural Dev.
3
,
6
.
Yang
,
M.
and
Crawley
,
J. N.
(
2009
).
Simple behavioral assessment of mouse olfaction
.
Curr. Protoc. Neurosci.
48
,
8.24.1
-
8.24.12
.
Yoshihara
,
Y.
,
Kawasaki
,
M.
,
Tamada
,
A.
,
Nagata
,
S.
,
Kagamiyama
,
H.
and
Mori
,
K.
(
1995
).
Overlapping and differential expression of BIG-2, BIG-1, Tag1, and F3: four members of an axon-associated cell adhesion molecule subgroup of the immunoglobulin superfamily
.
J. Neurobiol.
28
,
51
-
69
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information