In the vertebrate embryo, placodes are specialized regions of the ectoderm that emerge at the interface between the neural folds and the non-neural ectoderm. A subset, termed neurogenic placodes, makes major contributions to the cranial sensory ganglia and the associated sensory receptor neurons (Graham and Begbie, 2000). The bilateral olfactory placodes are the most anterior of the neurogenic placodes, emerging at the rostral edge of the neural plate (Graham and Begbie, 2000). Cells within these paired placodes will differentiate to produce the mature olfactory epithelium (OE; Klein and Graziadei, 1983; Calof and Chikaraishi, 1989). The OE is populated by three major cell types: basal cells, support cells, and olfactory receptor neurons (ORNs; Klein and Graziadei, 1983). The basal cells act as precursor cells for the OE, differentiating to replace lost support cells or ORNs over the lifetime of the animal (Harding et al., 1977; Monti Graziadei and Graziadei, 1979; Caggiano et al., 1994). The support cells serve a secretory role within the OE (Okano and Takagi, 1974), and the ORNs express odorant receptors and are the sensory cells that transduce odor information (Buck and Axel, 1991). The generation of only primary sensory neurons and the regenerative capacity of the OE creates an excellent model system in which to study the process of neural cell fate specification.
Transcription factors that have been implicated in vertebrate neurogenesis include the bHLH and rHLH families (Kageyama and Nakanishi, 1997; Dubois and Vincent, 2001). The vertebrate bHLH family members were identified based on homology to the genes of the Drosophila achaete-scute complex and atonal. In Drosophila, these genes are required for sensory neuron development (Campos-Ortega, 1997). The rHLH protein Olfactory neuronal transcription factor (Olf-1) was identified in rat for its role in regulating olfactory gene expression (Wang et al., 1997). Additional rHLH homologues to Olf-1, the Early B cell factor genes, were identified in mouse (Ebf, Ebf 2, Ebf 3) as well as in Xenopus (Xebf2, Xebf3), Zebrafish (Zcoe2), and Caenorhabditis elegans (unc 3/CeO/E; Garel et al., 1997; Dubois et al., 1998; Bally-Cuif et al., 1998; Prasad et al., 1998; Pozzoli et al., 2001). Overexpression studies in Xenopus have demonstrated that members from both these families can promote neurogenesis (Lee, 1997a; Dubois and Vincent, 2001), and targeted gene disruption in mice has shown that several members of the bHLH family of transcription factors are required for the specification of certain neuronal populations (Kageyama and Nakanishi, 1997; Bertrand et al., 2002). Although significant progress has been made in understanding how these transcription factors function to regulate neurogenesis, their role during placodal neurogenesis is less well understood.
All three murine Ebf genes are strongly expressed in the developing olfactory epithelium (Garel et al., 1997; Malgaretti et al., 1997). These proteins have been shown in vitro to bind the promoters and activate the expression of several olfactory neuron-specific genes, such as olfactory specific adenylyl cyclase III (Ac-III), olfactory marker protein (OMP), and the olfactory cyclic nucleotide gated channel (OcNC) (Wang et al., 1993, 1997). These data argue that the Ebf proteins may directly regulate the genes specific to odorant detection and/or the commitment of olfactory precursor cells to the ORN lineage. Although these results suggest that the Ebf genes are involved in ORN differentiation, their contribution to generating ORN fate has not yet been determined.
Several lines of evidence argue that the bHLH proteins are involved in the differentiation of olfactory receptor neurons (ORNs). For example, during mammalian olfactory neurogenesis Mammalian achaete-scute homologue 1 (Mash1) is expressed within the central region of the developing olfactory placodes (Cau et al., 1997). Likewise, analysis of Mash1 mutant mice reveals a major loss of ORNs within the mutant olfactory epithelium, as well as a loss of subsequent bHLH gene expression, such as Math 4C and NeuroD in the developing olfactory epithelium (Guillemot et al., 1993; Cau et al., 2002). These data argue that, in the mouse, Mash1 is critical in specifying the ORN lineage and that Mash1 may commit cells to the ORN lineage by activating the expression of additional bHLH proteins.
Although Mash1, a vertebrate achaete-scute homologue, is involved in regulating ORN development, the role of atonal-related HLH homologues during olfactory development is less well understood. Experiments in Drosophila have demonstrated that atonal is both necessary and sufficient for the production of a type of olfactory sensilla (Gupta and Rodrigues, 1997). However, whether vertebrate atonal homologues act to regulate ORN fate is not known. In this study, we demonstrate that in the developing Xenopus olfactory placode both bHLH and rHLH transcription factors are sequentially expressed during olfactory development. In addition, we show that a Xenopus atonal homologue, Xath5, is able to promote the differentiation of ORNs within a defined region of the developing embryo.
Sequential Activation of bHLH and rHLH Transcription Factors During Olfactory Development
To determine whether bHLH or rHLH transcription factors are expressed within the presumptive olfactory epithelium in Xenopus laevis embryos, we used in situ hybridization to visualize the spatial and temporal expression patterns of several HLH genes during olfactory placode development. We examined the expression of the bHLH genes X-NgnR-1, Xath3, Xath5, NeuroD, and the rHLH genes Xebf2 and Xebf3, as well as the neuronal marker N-tubulin beginning at early neural plate stages through progressively later stages. The first gene to be expressed in the presumptive olfactory placode was X-NgnR-1 at stage 15 (Fig. 1A). X-NgnR-1 was expressed at the lateral edge of the anterior neural plate, within the region fated to give rise to the olfactory placodes (Eagleson and Harris, 1990). By stage 16, expression of Xebf2 begins in the developing olfactory placodes (Fig. 1C). At stage 17, the developing olfactory placodes are located more medially in the embryo and initial expression of Xath3 is detected (Fig. 1E). Weak expression of Xath5 also occurs at stage 17 (data not shown) and becomes readily apparent by stage 18 (Fig. 1G). Expression of Xebf3 and N-tubulin did not occur in the presumptive olfactory placode until stage 20, the point at which the neural tube has completely formed (Fig. 1I and data not shown). The expression of X-NgnR-1 was maintained within the developing olfactory placodes, becoming restricted to the periphery of the developing OE by stage 24, whereas expression of Xebf2, Xebf3, Xath3, and Xath5 genes was maintained throughout the OE through stage 24, the latest stage tested (data not shown). Together, these results reveal that these HLH transcription factors are expressed sequentially within the developing olfactory placode. The timing of their expression suggests a potential role in regulating olfactory development.
Overexpression of Xath5 Expands the Region of Xenopus Olfactory Marker Protein Expression
Previous analysis revealed that Xath5 expression is limited to the pineal gland, retina, and the developing olfactory placode in Xenopus (Kanekar et al., 1997). In addition, Xath5 overexpression in Xenopus embryos has demonstrated that this bHLH factor can promote ectopic neurogenesis at neural plate stages as well as bias retinal progenitors to a retinal ganglion cell fate (Kanekar et al., 1997); however, Xath5 function in the olfactory lineage had not been examined. To determine whether Xath5 influences ORN generation in the developing olfactory epithelium, we injected mRNA encoding the lineage tracer β-galactosidase (β-gal) alone or together with mRNA for Xath5 into one cell of a two-cell stage embryo. These embryos were collected at stage 36, because this stage is when markers for mature ORNs are expressed, and were stained with X-gal to detect the lineage tracer. The embryos were then probed by whole-mount in situ hybridization for the expression of either Xenopus Olfactory Marker Protein 2 (Xomp2), which labels mature ORNs (Rossler et al., 1998), or N-tubulin, a neuronal marker that specifically labels developing neuroepithelial cells (Chitnis et al., 1995). Xath5-injected embryos analyzed for Xomp2 expression at stage 36 exhibited a striking expansion of Xomp2 expression within the anterior olfactory region on the injected side (72.5%, n = 91, Fig. 2A,C). In contrast, the control β-gal–injected embryos showed symmetric Xomp2 expression between the injected and uninjected sides (Fig. 2D). Xath5 misexpressing embryos had ectopic N-tubulin staining throughout the injected side (100%, n = 27, Fig. 2G), indicating that Xath5 can promote neurogenesis, as previously described (Kanekar et al., 1997). N-tubulin staining was largely restricted to the developing nervous system on the uninjected side (Fig. 2H). Of interest, unlike the widespread ectopic activation of N-tubulin expression caused by Xath5 overexpression, there was no ectopic Xomp2 expression; instead, the domain of normal Xomp2 expression was enlarged (Fig. 2A,C). These results argue that Xath5 promotes olfactory neurogenesis but only within a restricted region in the anterior of the embryo.
It is possible that other bHLH transcription factors may also promote olfactory neurogenesis. To test this, we overexpressed another closely related Xenopus atonal homologue, XNeuroD, in one cell of a two-cell stage embryo, then assayed for changes in Xomp2 expression. Unlike Xath5, misexpression of XNeuroD did not promote expanded Xomp2 expression; however, we did observe disorganization and/or loss of Xomp2 staining (Fig. 2E,F). In addition, we also overexpressed Xebf2 or Xebf3, because the mammalian homologue of these genes has been shown to bind the OMP promoter and initiate OMP transcription in vitro (Wang and Reed, 1993). Surprisingly, overexpressing either of the Xenopus Ebf genes did not ectopically activate or expand Xomp2 expression in embryos at either early or late developmental stages (data not shown).
Because Xath5 overexpression increases the production of ORNs as assayed by Xomp2 expression at later developmental stages, we next determined whether Xath5 influenced the expression of presumptive olfactory placode markers early during development. The rHLH gene, Xebf2, is one of the earliest genes expressed in the developing olfactory placode and is maintained in the placodes as development progresses (Fig. 1C, and data not shown). It, thus, can serve as an early marker for prospective ORNs. In addition, because the onset of Xebf2 expression precedes that of Xath5, it is unlikely that Xath5 directly activates Xebf2 expression during olfactory development. Embryos overexpressing Xath5 were collected at early (stage 15) and late (stage 20) neurula stages and probed for Xebf2 expression by whole-mount in situ hybridization. At stage 15, Xenopus embryos do not yet express Xebf2 in the developing olfactory placode and Xath5 overexpression did not induce premature activation of Xebf2 at this stage (100%, n = 22; data not shown). Xath5-injected embryos assayed for Xebf2 expression at stage 20 had expanded Xebf2 expression on the injected side (85%, n = 21, Fig. 2I), however, no ectopic activation of Xebf2 was apparent outside this region as a result of Xath5 misexpression. In contrast, stage 20 β-gal–injected control embryos showed symmetric Xebf2 expression between the injected and uninjected sides (Fig. 2J). These findings argue that expression of Xath5 is sufficient to drive a defined population of cells to begin the process of olfactory neuron differentiation even at very early stages of olfactory placode development.
Xath5 Promotes ORN Fate Independent of Cell Proliferation in the Developing Olfactory Placode
The ability of Xath5 activity to promote ORN genesis could be a result of increased proliferation of the ORN precursors. To address this, we treated embryos with the DNA synthesis inhibitors hydroxyurea and aphidocolin (HUA), which block cell division but allow normal morphogenetic movements until stage 32 (Harris and Hartenstein, 1991). We injected one cell of two-cell stage embryos with Xath5 and β-gal mRNA. These embryos were transferred to a solution of HUA at stage 10 and allowed to develop until stage 24–26 and assayed for Xebf2 expression, a marker of the early olfactory placode. We used Xebf2 as a marker of olfactory fate because it is an early marker of ORNs and expression can be analyzed well before embryo viability is compromised due to the HUA treatment. Exogenous Xath5 mRNA led to an expansion of the Xebf2-positive region (90%, n = 31; Fig. 3A), even in the presence of HUA (93%, n = 27; Fig. 3C), whereas control embryos expressing β-gal had no effect on Xebf2 expression in untreated (100%, n = 32; Fig. 3B) or HUA-treated embryos (92%, n = 28; Fig. 3D). HUA treatment effectively reduced proliferation as assayed by anti-phosphohistone H3 (HP3) antibody staining, which labels mitotic cells. In the absence of HUA treatment, Xenopus embryos injected with RNA encoding Xath5 and β-gal (Fig. 3E,F) or β-gal alone (Fig. 3I, J) exhibited extensive HP3 labeling throughout the animal, including the region of the olfactory placode (Fig. 3F,J). Conversely, the extent of HP3 labeling in the HUA-treated animals was dramatically reduced (Fig. 3G,H,K,L). This finding confirms that HUA treatment inhibited proliferation. These data argue that the effect of Xath5 activity on ORN fate occurs independent of cellular proliferation and suggest that Xath5 may recruit surrounding ectoderm to promote a spatially restricted increase in olfactory neural fate.
We have characterized the expression of several HLH genes during olfactory placode development in Xenopus laevis. These genes are expressed within the developing olfactory placode in a sequential manner, reminiscent of the HLH expression observed in the Xenopus neural plate during early neuronal differentiation, or primary neurogenesis (Chitnis, 1999). This pattern of rHLH and bHLH gene expression suggests that HLH family members may act to regulate neurogenesis in the nascent olfactory epithelium. In addition, we show that overexpression of the atonal homologue Xath5 can promote expansion of ORN markers within a defined domain at the anterior of the embryo and that this occurs independent of cellular proliferation. This finding suggests that Xath5 can regulate ORN differentiation during olfactory placode development in Xenopus.
Cascades of HLH Gene Expression Regulate Neurogenesis
Cascades of bHLH and rHLH gene activation have been observed in multiple developmental contexts. For example, during Xenopus primary neurogenesis within the developing neural plate, the differentiation of primary neurons is initiated by the expression of the bHLH gene X-NgnR-1 (Ma et al., 1996), which then promotes the expression of Xebf2, Xath3, and XNeuroD (Perron et al., 1999; Pozzoli et al., 2001). XNeuroD subsequently promotes expression of Xebf3 (Pozzoli et al., 2001). This pattern of sequential gene expression during neuronal differentiation has been suggested to progressively restrict a progenitor cell to the neural fate during primary neurogenesis.
We found all of these genes are also expressed during olfactory neurogenesis in a similar sequence. For example, X-NgnR-1 was the first bHLH gene to be expressed within the anterolateral border of the neural plate, the region fated to produce the olfactory epithelium, and was followed by expression of Xebf2 at stage 16 and Xath3 at stage 17. At this point, the expression pattern within the nascent olfactory placode diverged from the pattern observed in primary neurogenesis. Xath5, which is not expressed in the developing neural plate, was expressed concomitantly with Xath3 in the immature olfactory placodes. By stage 20, Xebf3, NeuroD, and N-tubulin were expressed in the nascent placodes, their later expression consistent with that seen during primary neurogenesis. That a similar complement of HLH factors regulates neurogenesis within the neural plate and olfactory placodes suggests that the distinct identities of these different populations of neurons are not strictly determined by these factors. Within the olfactory placode, it is also possible that particular neuronal precursors are heterogeneous with respect to HLH gene expression, as has been shown in the cortex and the retina (Jasoni and Reh, 1996; Nieto et al., 2001).
A cascade of bHLH gene expression also occurs during mouse olfactory neurogenesis. Analysis of the Mash-1 mutant olfactory epithelium revealed that Mash-1 activity is required for the subsequent expression of additional bHLH genes, specifically Math4C/ngn1 and then NeuroD, in mouse ORN progenitors (Cau et al., 1997). This pattern of gene expression is consistent with proneural genes acting in a cascade to progressively restrict cell fate. It suggests, however, that the precise combination of HLH genes used to regulate olfactory neurogenesis is not highly conserved.
Xath5 Can Regulate Olfactory Neurogenesis
We have found that Xath5 is expressed in the presumptive olfactory placodes and that overexpression of Xath5 broadens the domain of Xebf2 expression early (stage 20), and later causes an expansion of Xomp2 expression (stage 36), which marks mature ORNs. These data suggest that Xath5 activity can promote ORN fate in a population of cells competent to respond to signals that specify the olfactory lineage. The Ebf genes have been shown to activate Omp expression in vitro (Wang et al., 1993), suggesting that Xath5 activity may be indirectly activating Xomp2 expression by means of increased Xebf2 expression. However, overexpression of either Xebf2 or Xebf3 in Xenopus embryos does not result in ectopic Xomp2 expression at early or later stages of development (data not shown). This finding argues that, although Xath5 activity increases Xebf2 expression early in the developing olfactory placode, Xebf2 alone is not sufficient to promote ORN fate.
XNeuroD, an atonal homologue closely related to Xath5, does not promote expansion of Xomp2 expression when overexpressed, even though it is very effective at promoting ectopic neurogenesis (Lee et al., 1995). Paradoxically, we have previously shown that overexpression of XNeuroD causes an expansion of Xath5 expression in anterior of the embryo (Kanekar et al., 1997), raising the question of why Xath5 promotes expanded Xomp2 expression when overexpressed by RNA injection, but not when its expression is activated after XNeuroD overexpression. XNeuroD may be activating the expression of additional genes that may interfere with ORN differentiation, such as negative regulatory factors. Alternatively, XNeuroD overexpression may disrupt or alter the expression of other factors necessary for the ORN fate, such as patterning factors or competence factors (see below). Additional experiments will be necessary to distinguish between these possibilities. It remains to be determined whether Xath5 has a unique ability to promote the olfactory neuron fate, because the mouse ath5 homologue Math5 is not expressed in the developing olfactory epithelium (Brown et al., 1998). However, other atonal homologues, including Math4A and Math4C, are expressed in the developing olfactory placodes and could serve an analogous role to Xath5 during mammalian olfactory neurogenesis (Cau et al., 1997).
If Xath5 activity increases the number of cells committed to the ORN lineage, how are the additional neurons being generated? Xath5 overexpression causes expansion of Xebf2 expression in the presence of HUA, which blocks cell division, arguing that surrounding ectodermal cells are being recruited to the ORN fate in response to Xath5 expression. The proneural bHLH factors have been shown to convert ectodermal cells, which would normally differentiate as epidermis, into neurons, when overexpressed in the developing Xenopus embryo (Ferreiro et al., 1994; Turner and Weintraub, 1994; Ma et al., 1996; Lee, 1997b). Indeed, overexpression of Xath5 promotes the formation of ectopic N-tubulin–positive neurons throughout the ectoderm of injected embryos (Kanekar et al., 1997). However, only a subset of these neurons in a very defined region in the anterior of the embryo adopt the ORN fate and express Xomp2, suggesting that recruitment of cells to the ORN fate by Xath5 is spatially limited.
Defining the Region of Olfactory Neurogenesis
Two possibilities exist to explain this constraint on ORN differentiation in response to Xath5 activity. First, negative regulatory factors may exist to limit the region of olfactory neurogenesis to the anterior of the embryo and, thus, define the region capable of generating olfactory neurons. For example, Notch/Delta signaling, which can regulate Xath5 activity in the neural plate and the retina (Schneider et al., 2001) may limit Xath5 function in the developing olfactory epithelium. The Notch ligand, X-Delta-1 is expressed within the early placodal region (Chitnis et al., 1995; Schlosser and Northcutt, 2000) but likely acts through lateral inhibition to regulate the onset or density of neuronal differentiation within the placodes rather than defining the boundaries of olfactory neurogenesis. It is possible that other negative regulators such as hairy-related genes serve to define the placodal domain. In mouse, the hairy/Enhancer of split homologue Hes1 functions as a prepattern gene to regulate the level and pattern of proneural bHLH expression within the olfactory region, and then acts redundantly with Hes5 as a neurogenic gene to regulate the density of neuronal differentiation within the placode (Cau et al., 2000). Similar mechanisms could be operating in Xenopus.
Alternatively, a positive cofactor may be required to generate ORNs in cooperation with Xath5, and this cofactor may be limited to the presumptive placodal region in the anterior of the embryo. This finding could include other HLH factors that we have shown are expressed also in the developing placode. During spinal cord development, the bHLH factors Olig2 and Ngn2 cooperate to regulate motorneuron development (Zhou et al., 2001). Alternatively, Xath5 could be cooperating with other factors, such as homeobox transcription factors, to specify the ORN fate. For example, X-dlx3 has an early and restricted pattern of expression in the presumptive olfactory region, and this, or some other factor, may define a region of competence for ORN differentiation (Franco et al., 2001). In the mammalian retina, bHLH factors alone do not dictate specific retinal neuron fates, rather they appear to cooperate with homeodomain transcription factors to regulate retinal neuron subtype decisions (Hatakeyama et al., 2001; Inoue et al., 2002). Thus, overexpression of Xath5 in Xenopus embryos promotes the differentiation of neurons broadly throughout the non-neural ectoderm, but whether these become ORNs may depend on a spatially restricted complement of regulatory factors that define the domain of olfactory neurogenesis. Further experiments will be necessary to distinguish these possibilities.
In Situ Hybridization
The following constructs were used in probe synthesis: pBS-X-NgnR-1 (Ma et al., 1996), pBS-Xath5 (Kanekar et al., 1997), pBS-Xath 3 (Turner and Weintraub, 1994), pBS-XNeuroD (Lee et al., 1995), pBS-Xebf2, pBS-Xebf3, pBS-N-tubulin (Chitnis et al., 1995), and pBK-CMV-Xomp2 (Rossler et al., 1998; gift from H. Breer). Digoxigenin-labeled antisense RNA probes were synthesized in vitro per manufacturer's instruction with a Maxiscript kit (Ambion). X-Ngnr-1, Xath5, Xath3, and N-tubulin fluorescein-labeled antisense RNA probes were synthesized in vitro per manufacturer's instruction by using the Fluorescein RNA Labeling Mix (Boehringer Mannheim). Embryos were obtained from adult Xenopus animals using hormone-induced egg laying and standard in vitro fertilization. Whole-mount in situ hybridizations were performed on albino embryos by using the procedure previously described (Kanekar et al., 1997).
Microinjection of RNA
The following constructs were used in microinjection experiments: pCS2-Xath5 (Kanekar et al., 1997), pCS2-XNeuroD (Lee et al., 1995), pCS2-Xebf2 (Dubois et al., 1998; Pozzoli et al., 2001), pCS2-Xebf3 (Pozzoli et al., 2001), and pCS2-nuclear (βgal (Chitnis et al., 1995). Capped mRNAs were transcribed in vitro per the manufacturer's instructions by using a Message Machine kit (Ambion). Embryos were collected as described above and dejellied in a 2.5% cysteine solution at pH 8.0. The mRNA was injected into one blastomere of a two-cell stage embryo (albino) at the following concentrations: Xath5 (200 pg), Xebf2 (200 pg), Xebf3 (200 pg), XNeuroD (50 pg), and βgal (100 pg). Embryos were allowed to develop then collected at various stages according to Nieuwkoop and Faber (1994). Embryos were fixed in MEMFA (Harland et al., 1991) at room temperature for less than 1 hr and stained with X-gal by using previously described methods (Turner and Weintraub, 1994), stored in methanol and analyzed by means of in situ hybridization.
Hydroxyurea and Aphidocolin Treatment
Embryos were injected with mRNA for either Xath5 and nβ-gal or nβ-gal alone at the two-cell stage as described above and then transferred at stage 10–10.5 to a 0.1× MMR solution containing 20 mM hydroxyurea (Sigma) and 150 μM aphidocolin (Sigma) as previously described (Harris and Hartenstein, 1991). Sibling embryos were transferred at stage 10–10.5 to 0.1× MMR solution to serve as controls. HUA-treated embryos were staged relative to the control embryos. Embryos were collected at stage 24–26 and fixed for 30–60 min in MEMFA. Embryos were subsequently stained with X-gal as previously described (Turner and Weintraub, 1994) and then processed by in situ hybridization for Xebf2 expression.
HP3 Antibody Staining
To detect mitotic cells, a rabbit anti-H3 antibody (Upstate Biotechnology) was used. HUA-treated and untreated embryos were fixed and X-gal stained (see above) then stored in Dent's fixative (20% dimethyl sulfoxide, 80% methanol) at −20°C overnight. Embryos were rehydrated through a series of methanol/phophate buffered saline (PBS) washes then washed extensively in PBTx (PBS + 0.1% Triton X-100) before blocking for 30 min at room temperature (RT) in PBTx + 10 mg/ml bovine serum albumin (BSA). Embryos were then incubated overnight at 4°C in anti-HP3 diluted 1:500 in block. Samples were washed extensively the following day in PBTx + 2 mg/ml BSA then blocked for 30 min at RT. Embryos were incubated overnight at 4°C in anti-rabbit Alexa Fluor-555 (Molecular Probes) diluted 1:1,000 in block. The following day, samples were washed extensively in PBTx + 2 mg/ml BSA and then refixed for 15 min at RT in MEMFA then washed in PBS.
We thank Sheryl Scott, Gary Schoenwolf, and Mary Lucero for critical reading of the manuscript; Alejandro Sanchez and Gail Burd for valuable technical advice; and Heinz Breer for the Xomp2 construct. M.L.V. received funding from the NIH and the Pew Scholars Program in the Biomedical Sciences sponsored by the Pew Charitable Trusts.