|Home | About | Journals | Submit | Contact Us | Français|
The coordination of gene expression is critical for cell differentiation and the subsequent establishment of tissue function. We show here that a multiple zinc-finger transcription factor, zfp423/OAZ, is transiently expressed in newly differentiating olfactory receptor neurons (ORNs) and has a key role in coordinating the expression of immature and mature stage-specific genes. OAZ deletion in mice impairs aspects of ORN differentiation, particularly the patterns of axonal projection to the olfactory bulb. OAZ gain-of-function experiments show that sustained OAZ expression throughout ORN maturation arrests ORN development at an immature stage and alters OR gene expression. Importantly, reintroducing OAZ expression in mature ORNs suppresses mature marker expression and reactivates immature-specific markers. Together, theses experiments suggest that OAZ participates in a developmental switch regulating the transition from differentiation to maturation in ORNs.
The olfactory receptor neurons (ORNs) are continuously generated from a population of neuroblast-like stem/progenitor cells located at the basal olfactory epithelium (OE). This lifelong olfactory neurogenesis is under the control of a temporal series of intrinsic transcriptional cascades (Nicolay et al., 2006), some of which are regulated by extrinsic cues (Kawauchi et al., 2004; Kawauchi et al., 2005; Wu et al., 2003). During ORN development, olfactory stem cells undergo infrequent asymmetric cell divisions to generate transient amplifying progenitors (Mash1-positive) and subsequently, Ngn1-positive immediate neuronal precursors (Calof et al., 2002; Cau et al., 2002). The division of Ngn1-positive cells generates daughter cells that differentiate into ORNs and express characteristic markers including neural cell adhesion molecule (NCAM), olfactory maker protein (OMP) and olfactory-specific genes essential for odorant transduction (Bakalyar and Reed, 1990; Brunet et al., 1996; Calof and Chikaraishi, 1989; Jones and Reed, 1989; Kawauchi et al., 2004).
Extrinsic signaling molecules implicated in olfactory neurogenesis include the fibroblast growth factor (FGF) and transforming growth factor-β (TGF-β) superfamilies (Kawauchi et al., 2004; Kawauchi et al., 2005). These signaling molecules, particular the bone morphogenic proteins (BMPs) exert both positive and negative effects on ORN neurogenesis, depending on the type of the signaling molecule, the concentration and the stage at which they act (Kawauchi et al., 2004; Shou et al., 1999; Shou et al., 2000; Wu et al., 2003). Despite considerable progress in defining transcription factors and signaling molecules in OE neurogenesis, much less is known about the key transcriptional regulators mediating the interaction between these two aspects of regulation.
The Zfp423/OAZ protein, containing 30 kruppel-like C2H2 zinc-fingers is an attractive candidate to mediate extrinsic regulation of the transcriptional cascade. OAZ (O/E associated zinc-finger protein) was identified as an inhibitor of O/E proteins, a family of transcriptional factors involved in terminal ORN differentiation (Wang and Reed, 1993; Wang et al., 1993; Wang et al., 1997). OAZ binds to O/E proteins using a subset of zinc-finger domains and abolishes O/E binding to consensus binding sites present in olfactory-specific promoters (Tsai and Reed, 1997; Tsai and Reed, 1998). This negative regulation of O/E activity was proposed to prevent premature differentiation of immature ORNs. Independently, OAZ was also identified as a cofactor of Smad proteins in the BMP-signaling pathway (Hata et al., 2000; Ku et al., 2006; Shim et al., 2002). The zinc-finger domains mediating Smad interaction are distinct from those that engage in O/E proteins (Hata et al., 2000). The presence of BMPs and their receptors in the olfactory system and the proposed role of BMPs in regulation of ORN differentiation (Dewulf et al., 1995; Wu et al., 2003; Zhang et al., 1998) suggest that OAZ might have an important role in integrating BMP signaling and O/E activity at critical stages of ORN differentiation.
In this study, we provide in vivo evidence that OAZ regulates the transition from differentiation to maturation in ORNs. OAZ is transiently expressed in newly generated postmitotic ORNs. OAZ deletion in mice impairs ORN differentiation. Importantly, reintroducing OAZ into mature cells is sufficient to revert mature cells to an immature phenotype. These arrested, immature cells accumulate in OE and are perturbed in normal olfactory receptor (OR) gene selection and expression.
To elucidate the pattern of OAZ expression in OE, we initially utilized an OAZ-lacZ reporter mouse line (OAZlacZ, XB409), derived from a gene-trap screen in ES cells (Stryke et al., 2003). This line contains a β-geo gene-trap vector insertion into the last intron of OAZ gene, resulting in a deletion of nine amino acids from the OAZ C-terminus and expression of an OAZ-lacZ fusion protein. X-gal staining in OAZlacZ/+ mice at E18 revealed a broad OAZ expression in the developing OE (Figure 1A). The expression decreased postnatally and was restricted to the basal OE (Figure 1B, C), the site of continuous olfactory neurogenesis in adults.
During ORN development, progenitors express the molecular markers Mash1 and Ngn1 at successive stages and migrate apically in the OE (Calof et al., 2002; Cau et al., 2002). To identify the stage at which OAZ is expressed, we introduced Mash1GFP/+ and Ngn1GFP/+ reporter knock-in mice, in which a nuclear GFP reporter precisely replaces the respective coding regions and faithfully mimic expression of the endogenous gene (Leung and Reed, unpublished). In OAZlacZ/+Mash1GFP/+ mice, OAZ-expressing cells were identified with X-gal staining (Figure 1D) and Mash1-expressing cells were labeled with an anti-GFP antibody (Figure 1E). OAZ-expressing cells were located apically to Mash1-expressing cells (Figure 1F), and were essentially absent in E18 Mash1-null mice (Figure 1J), suggesting that OAZ is downstream of Mash1. In OAZlacZ/+Ngn1GFP/+ mice, Ngn1 and OAZ-expressing cells were largely located in the same layer (Figures 1G–I). In cofocal images, Ngn1/OAZ double-labeled cells were predominantly located between the more basal Ngn1-expressing cells and more apical OAZ-expressing cells (Figure 1K), suggesting that OAZ expression may be initiated as cells ceased Ngn1 expression. Furthermore, OAZ was confined to postmitotic cells when examined by BrdU labeling (Figure 1L). Together, these observations indicate that OAZ is expressed in newly differentiating ORNs.
In parallel, we generated a reporter-tagged OAZ-null allele by targeted deletion of the OAZ large exon 4 and insertion of IRES-YFP-polyA. This presumptive functional null allele lacks 1180/1292 amino acids of the native OAZ protein including all the known DNA-binding and protein-binding domains and is hereafter referred to as OAZ−. Homozygous OAZ−/− mice exhibited profound cerebellar defects and died by approximately four weeks. The cerebellar phenotype of other OAZ alleles has been described elsewhere (Alcaraz et al., 2006; Cheng and Reed, 2007; Warming et al., 2006). The olfactory turbinates of OAZ−/− mice display normal organization with a slightly smaller nasal cavity and correspondingly reduced OE area. To examine subsequent stages of ORN development, OE sections (PD 20–25) were examined with the ORN lineage marker O/E, mature ORN marker OMP, and differentiation markers GAP43 and NeuroD1. The pattern of O/E immunostaining in the epithelium from wild-type and OAZ−/− mice were indistinguishable (Figure 2A). However, the number of mature cells visualized by OMP immunostaining was fewer in P20 OAZ−/− mice (data not shown) and, when quantified by in situ hybridization (Figure 2B), was reduced 30% (wt, 890±30/mm; OAZ−/−, 635±15/mm, Figure 2F). In contrast, there were no changes in the expression of GAP43 (Figure 2C) and NeuroD1 (Figure S1C in supplemental data). It is noteworthy that Evi3, an OAZ homolog gene that also interacts with O/E protein (Hentges et al., 2005; Warming et al., 2003), is present at low levels in the OE and displays a pattern of expression similar to OAZ (Figure S2), which may obscure a complete analysis of loss-of-function of OAZ in the OE.
The reduction of mature cells could arise from decreased proliferation of progenitors, delayed maturation, and/or reduction of cell lifespan. No significant changes in the number of BrdU-positive cells were observed between wild-type and OAZ−/− mice examined at P2, P7 and P20 (Figure 2D, G). TUNEL staining, however, showed a twofold increase in the number of apoptotic cells in OAZ−/− mice (wt, 7.0±1.9/mm; OAZ−/−, 17.6±10.0/mm, Figure 2E, H) that could account for the decease in mature neurons, but additional effects of OAZ on progression of maturation are possible. The observed OE phenotype was consistent along the anterior to posterior axis and across different dorsal-ventral zones in the OE (Figures S1A, B).
A hallmark of ORN organization is that cells expressing a particular OR gene project their axons to spatially defined glomeruli in the olfactory bulb (OB) (Vassar et al., 1994). The OB in OAZ−/− mice was smaller, a phenotype also observed in other genetic models where primary ORN afferents fail to project or lack odor-evoked activity (Baker et al., 1999; Hirata et al., 2006). We examined ORN projection by introducing the O/E3-tauGFP reporter (Wang et al., 2004). In whole-mount view, ORN axons failed to innervate the dorsal caudal region of the OB in OAZ−/− mice (Figure 3B and Figures S3B–D). In coronal sections, the caudal dorsal OB was essentially devoid of olfactory nerves and glomeruli; instead, additional glomeruli were concentrated on the ventral surface (Figures S3K–M). A similar OB phenotype was observed when projection were visualized with an OMP-taulacZ reporter (Mombaerts et al., 1996) (Figures 3C, D).
We examined the laminar organization of the OB using makers for OB neurons and olfactory axons. In neonatal and adult mice, GAP43 mRNA is detected in mitral cells and periglomerular cells (Verhaagen et al., 1990) (Figure 3E–H). In OAZ−/− mice, the dorsal OB surface was devoid of periglomerular cells and the mitral cell layer was pushed inwardly by the ventrally accumulated glomeruli (Figure 3F, H). Consistently, when olfactory axons were labeled with OMP (Figures 3I–L), the dorsal OB surface lacked an olfactory nerve layer (ONL), but a fibrocellular mass-like structure composed of tangled axons and disorganized glomeruli was observed at the ventral surface (Figure 3L).
We next examined the projection of individual ORNs by introducing reporter-tagged OR alleles (Mombaerts et al., 1996; Zheng et al., 2000). In contrast to two topographically fixed M72 glomeruli in wild-type mice (arrows in Figure 3M, N), multiple M72 projection sites were detected (10/12 OBs), or only wandering axons could be observed (2/12 OBs) in OAZ−/− mice. Interestingly, a subset of M72 neurons projected to a glomerulus located in a similar position as in OAZ wild-type mice (arrows in Figure 3O–R), whereas the remaining majority axons projected to multiple, more ventral glomeruli (arrowheads). The disrupted projection pattern was also observed for the M71-expressing cells using the M71-IRES-TauGFP reporter (control, n=8 mice, OAZ−/−, n=12 mice). ORNs expressing the P2 OR extend much shorter axons and project to a ventral region of the bulb. In OAZ−/− mice, the P2 axons innervate glomeruli that were anteriorly shifted and wandering axons were observed at higher frequency (data not shown).
The loss-of-function experiments described above suggest that OAZ is important for the normal program of ORN differentiation, particularly the projection of olfactory axons. The proposed role of OAZ as an inhibitor of O/E transcription factor function (but not O/E protein expression) led us to ask whether sustained OAZ expression was sufficient to arrest ORN development. We designed a dominant gain-of-function knock-in allele that expresses OAZ from the O/E3 locus (Figure 4A). The O/E3 promoter would drive sustained OAZ expression throughout ORN development. An IRES-driven YFP reporter was included to facilitate identification of the O/E3-OAZ expressing cells.
The LTNL cassette used for ES cell selection affects gene expression from the O/E3 promoter. In the initial O/E3OAZ-LTNL mouse line that still carries LTNL, no YFP expression was detected in the OE. However, after CRE-mediated removal of LTNL, heterozygous O/E3OAZ/+ mice could be readily distinguished by their small size and YFP expression (mean weight at P21: wt, 11.1±0.6g, n=7; O/E3OAZ/+, 7.5±0.8g, n=7, P<0.001). We speculate that LTNL blocks a downstream enhancer of the O/E3 promoter, resulting in the absence of OAZ-YFP expression in cassette-containing mice. A similar phenomenon is seen with other insertions at this locus (unpublished data). Since both the O/E3OAZ-LTNL mice and O/E3OAZ/+ mice contain one deleted copy of the O/E3 gene, analysis of O/E3OAZ-LTNL mice (which display a wild-type phenotype) provides strong evidence that the phenotype we observed in O/E3OAZ/+ mice results from expression of OAZ, rather than simply the loss of one copy of O/E3. This special feature also provides for conditional control of OAZ-YFP expression.
We characterized ORN development in O/E3OAZ/+ mice using mature markers OMP and Adenylyl cyclase 3 (AC3), the general neuronal marker NCAM, and the immature marker GAP43. In O/E3OAZ/+ mice, there was a dramatic reduction in the number of mature ORNs visualized by OMP (Figures 4B, C) and AC3 immunostaining (Figures 4F, G). Intriguingly, there was no overlap between expression of OAZ and the mature cell markers; cells that displayed high levels of OMP or AC3 expression were those that failed to express the OAZ-linked YFP reporter (arrows in Figures 4C–E and G–I). Furthermore, the YFP-positive cells in O/E3OAZ/+ mice expressed NCAM, indicative of a normal neuronal lineage (Figures 4J–M). In situ hybridization revealed that the majority of cells in the epithelium of O/E3OAZ/+ mice were positive for GAP43 (Figures 5C, D, n=6 mice), indicating that the cells were arrested at an immature stage. In contrast, GAP43 expression was restricted to the basal immature cell layer in cassette-containing, phenotypically wild-type mice (Figures 5A, B). Importantly, the number of GAP43 positive cells in O/E3OAZ/+ mice was markedly increased in comparison to wild-type OE.
The reduction of mature cells was also confirmed by using an OMP-taulacZ reporter (Figures 6A–D). There was a 95% reduction of OMP-positive cells in O/E3OAZ/+ mice compared with OMP in situ hybridization of wild-type mice (wt, 865±66/mm; O/E3OAZ/+, 47±28/mm, Figure 6E). In parallel, in situ hybridization with several OR gene probes revealed that OR genes were not expressed at high levels in these immature cells (Figure S4). If these cells are blocked from completing the normal differentiation program by the persistent expression of OAZ, one might expect an increase in apoptosis. We observed an 8-fold increase in TUNEL-positive cells in O/E3OAZ/+ mice (Figures 6F, G and H, wt, 9.4±5.4/mm; O/E3OAZ/+, 76.5±25.2/mm). However, these apoptotic cells still represent only a small fraction of the total immature population in the epithelium. Taken together, these experiments indicate that ORN maturation is arrested at an immature neuronal stage and accumulate in greater numbers inO/E3OAZ/+ mice.
The OAZ-mediated arrest of ORN maturation is consistent with its ability to block O/E-mediated expression from reporter constructs in heterologous cell lines (Tsai and Reed, 1997). We next asked whether OAZ expression is sufficient for an immature phenotype by utilizing the gain-of-function O/E3OAZ-LNL/+ allele to reintroduce OAZ expression in mature ORNs. Specifically, O/E3OAZ-LNL/+ mice were crossed with mice in which the CRE recombinase is under the control of the OMP promoter (OMP-CRE) (Li et al., 2004). As expected, LTNL cassette removal led to appearance of OAZ-YFP in the mature neuronal layer (Figure 5I). In these mice, GAP43 in situ hybridization revealed a second population of GAP43-expressing cells present in the apical region (Figures 5E, F, n=4) in addition to the basal layer of the endogenous GAP43-expressing cells. This pattern contrasts with the basally-restricted GAP43-positive cells observed in the O/E3OAZ-LNL/+ mice (Figure 5B). Subsequent OMP/GAP43 double-labeling revealed that OMP expression was suppressed in these reactivated GAP43-positive cells (Figures 5G, H). Messenger RNA for ORs and other mature ORN markers examined by in situ hybridization were also reduced (data not shown). Together, these experiments indicate that OAZ expression is sufficient to establish an immature ORN phenotype.
We further characterized the OAZ-mediated arrest of ORN maturation and examined ORN projection using the OMP-tauLacZ reporter. In O/E3OAZ/+ mice, the dorsal OB surface was largely devoid of glomeruli; the few projecting axons reaching the dorsal surface migrated along aberrant paths instead of converging (Figures 7A, B). Coronal sections confirmed the lack of an extensive ONL and glomerular layer at the dorsal and lateral OB surface, while some glomeruli remained at the medial surface (Figures 7C–F). Further analysis showed that even though GAP43-expressing immature cells could initiate their axonal growth in O/E3OAZ/+ mice (Figures 7G, H), the majority failed to innervate the bulb (Figure S5). The medial OB surface contained OMP-positive axons and disorganized glomerular structures, but the lateral surface contained only a thin ONL (Figures 7I–K) that presumable derive from the “escaper” cells. The expression of tyrosine hydroxylase, a marker for periglomerular interneurons, was also reduced at the lateral OB surface consistent with the reduced innervation by ORN axons (Figures 7L–N).
We next examined OR-dependent axonal convergence. In O/E3OAZ/+ mice, the projections from both P2 and M72-expressing cells to the OB were greatly reduced. The remaining P2 and M72 projecting axons were unable to converge but spread along the medial surface of the bulb (Figures 8A, B and Figure S6).
Interestingly, whole-mount X-gal staining of taulacZ-tagged M72 receptor revealed a large number of weakly stained cells in the epithelium (arrows in Figure 8D). In coronal sections, these weakly stained cells displayed short dendrites and no obvious axon extensions although the faint signal might prevent visualization of these structures. (Figure 8F). Confocal imaging revealed that most of these cells (94%, n=89) were also expressing OAZ-YFP. The level of M72 gene expression in these immature cells, inferred by the intensity of X-gal staining, was much lower than that observed for axon-projecting M72 cells in wild-type mice (arrowheads in Figure 8E) and could not be detected by in situ hybridization (Figure S4). Remarkably, the total number of M72-taulacZ expressing cells was three-fold greater in the epithelium of O/E3OAZ/+ mice than in control littermates (wt, 1.2±0.3/mm, n=3; O/E3OAZ/+, 4.7±0.7/mm, n=3, P<0.002, Figure 8G). The OAZ-mediated arrest of ORN maturation appears to perturb OR expression and alters the number of cells that have chosen a particular receptor for expression.
First identified as a transcriptional partner of O/E proteins, OAZ was also found to mediate BMP-signaling pathways through independent domains of the protein. The loss-of-function and gain-of function experiments described here suggest that in ORN development, OAZ may coordinately regulate the duration of immature gene expression and the onset of mature gene expression (Figure S7). These two developmental processes may arise through distinct O/E and BMP-mediated pathways.
In the OE, OAZ is transiently expressed in newly generated differentiating ORNs. Consistent with its postmitotic expression, OAZ deletion does not affect proliferation of olfactory progenitors. Instead, we observed decreased mature cells, impaired axonal targeting and increased apoptosis in OAZ−/− mice. Similar but less severe axonal targeting defects were also observed in O/E2 and O/E3-null mice (Wang et al., 2004). Therefore, the O/E-OAZ interaction may regulate (directly or indirectly) ORN axonal growth and/or targeting. The overlapping expression of Evi3 and OAZ in the OE may moderate the phenotype in the loss-of-function OAZ-null mice and obscure a critical role for OAZ in the transition from immature to mature ORNs. However, the series of OAZ knock-in experiments described here provide strong evidence that the presence of OAZ can perturb normal developmental timing and alter immature and mature ORN gene expression patterns. In this scenario, the transient OAZ expression during early differentiation contributes to a critical developmental switch during OE neurogenesis.
Consistent with the ability of OAZ to inhibit O/E-mediated transcription in vitro, ORNs that maintain OAZ expression in vivo are arrested at a differentiating stage precisely when the known targets of O/E (OMP, ORs and other olfactory transduction components) are first expressed. However, the re-expression of GAP43 when OAZ is reintroduced into mature cells (O/E3OAZ-LNL/+OMPCRE/+ mice) suggests that OAZ has additional functions beyond its action as an O/E inhibitor. A potential pathway involved in this transcriptional activation is the BMP-signaling pathway. OAZ is normally expressed at a stage when low-doses of BMP4 promotes survival of newly generated NCAM-positive immature cells in olfactory cultures (Shou et al., 2000). It would be interesting to determine whether BMP4 regulates GAP43 expression. Taken together, our experiments support a model in which OAZ may coordinately regulate immature and mature gene expression by integrating extrinsic microenvironmental cues (BMP-pathway) with an intrinsic transcriptional cascade (O/E pathway).
The molecular mechanisms controlling OR gene choice and expression remains largely unresolved. While a specific genomic element, the “H region”, is suggested to function as a global enhancer promoting OR expression (Lomvardas et al., 2006; Serizawa et al., 2003), the molecular interactions and proteins that participate in this complex process are unknown. These studies and related experiments on feedback mechanisms in OR gene choice suggest that the OR selection takes place within a defined time window during ORN development (Lewcock and Reed, 2003; Lewcock and Reed, 2004; Mombaerts, 2004). The expression of OAZ during this window and the phenotype observed in O/E3OAZ/+ mice suggest that it may participate in OR selection. The increased number of immature M72-expressing cells in O/E3OAZ/+ mice may derive from neuronal precursors that are arrested in development by sustained OAZ expression and accumulate multiple, activated OR genes in each cell. Unfortunately, the low level of reporter expression in these immature cells has precluded direct examination of this interesting question. Alternatively, maintained OAZ expression could promote differential selection of particular OR gene family members. The O/E3OAZ/+ mice may also facilitate the study of OR selection by providing large numbers of genetically-tagged differentiating cells arrested in development.
The dramatic reduction of mature ORNs in O/E3OAZ/+ mice is accompanied by a fraction of cells (~5% - “escapers”) that mature and express OMP and AC3. Double-label experiments reveal that the escaper cells do not express the reporter-tagged OE3-OAZ allele. The arrest phenotype associated with the OE3-OAZ allele is dominant and individual cells that silence this allele epigenetically or by loss-of-heterozygosity (LOH) during mitotic recombination will have a distinct survival advantage. The inability of these escapers to converge to discrete glomeruli may reflect their abnormal developmental history, altered OR expression or the reduced number of cells of each OR type in the epithelium.
The O/E transcription factors are involved in many developmental processes (Corradi et al., 2003; Garel et al., 1999; Garel et al., 2002; Kieslinger et al., 2005; Lin and Grosschedl, 1995). The redundant expression of different O/E family members has limited our ability to study these processes by conventional gene-targeted inactivation. For example, disruption of EBF1, the founding member of this family, leads to profound developmental defects in B cells and striatum tissues that express only this member of the O/E family (Garel et al., 1999; Lin and Grosschedl, 1995), but its loss results in modest, if any, phenotype in the OE where it is highly expressed. Similarly, O/E2 and O/E3 knockout mice as well as O/E2O/E3 double heterozygous mice displayed modest defects of ORN projection suggesting complex interactions and considerable functional redundancy with the additional O/E family members expressed in the OE (Wang et al., 2004). The ability of the OAZ protein to inhibit the function of all O/E family members in vitro (Tsai and Reed, 1997) and the expression of the O/E3-OAZ allele in all olfactory neuronal progenitors provides powerful tools to functionally disrupt all O/E function in this tissue. The abundant, persistent expression of the O/E3 locus in mature olfactory neurons may make these cells particular susceptible to this dominant suppression.
The generation of OAZlacZ and OAZ-null mice, and their phenotype outside of the olfactory system are described elsewhere (Cheng and Reed, 2007; see supplemental methods). The OAZ knock-in was created from the O/E3 knock-out targeting vector (Wang et al., 2004) and modified by inserting the full-length OAZ cDNA 261 base pairs before the initiation codon for O/E3. The OAZ cDNA was followed by an IRES-3NLS-YFP-pA cassette and LoxP-TK(Δ)-Neo-LoxP (LTNL) cassette. The OR reporter mice, OMP-taulacZ and OMP-CRE mice were obtained from Dr. P. Mombaerts. The sequences of primers for genotyping each of the lines are available from the authors.
The X-gal staining (Mombaerts et al., 1996) and immunohistochemistry was performed on PFA-fixed cryostat sections. The primary antibodies were: AC3 (rabbit, 1:1000; Santa Cruz Biotechnology), GAP43 (rabbit, 1:200, Chemicon), GFP (rabbit, 1:1000, Molecular Probes), LacZ (mouse, 1:500, Promega), O/E (rabbit, 1:1000), OMP (goat, 1: 3000; provided by Dr. F. Margolis), Tyrosine hydroxylase (rabbit, 1:500; Chemicon), NCAM (mouse, 1:1000; Sigma). Alexa 488- or Alexa 546-labled secondary antibodies were used at 1:1000 dilution. For double staining of X-gal and GFP, sections were stained with X-gal staining solution overnight, fixed with 4% PFA and then followed by normal antibody staining procedure using the anti-GFP antibody.
In situ hybridization was performed on PFA-fixed sections following a standard protocol (Wang et al., 2004). The probes were: I7 (NM_010983), M4 (NM_146937) and M72 (NM_030553) coding sequence, GAP43 (nucleotides 147 to 860 of NM_008083), NeuroD1 (nucleotides 640 to 1192 of BC018241), Evi3 (nucleotides 1 to 1172 of BC021376) and OMP (NM_011010) partial cDNA. Images were obtained with a Zeiss Axioplan microscope or a Zeiss LSM 510 confocal laser-scanning microscope.
Adult mice or pregnant females were injected intraperitoneally with BrdU (Sigma) 50μg/g body weight 30min before they were sacrificed. PFA-fixed sections were incubated with 3N HCl for 30min before immunostaining with anti-BrdU antibody (rat, 1:100; Abcam). To visualize, either an Alexa 546-labled secondary antibody or ABC kit (vector) and FAST DAB tablet sets (Sigma) were used. For double labeling of BrdU with anti-GFP antibody, sections were incubated with anti-GFP antibody first, then washed and fixed with 4% PFA before proceeding with the acid treatment. The secondary antibody incubations were performed simultaneously. TUNEL staining was performed using Apoptag® in situ apoptosis detection kits (Chemicon) according to the manufacturer’s instruction.
OMP counting was performed on images taken at 20× magnification (4 images at matched locations for each animal, the total length of OE was approximately 5mm). The total number of BrdU, TUNEL and specific OR cells were counted under the microscope from overlapping 20× fields for the entire OE region on matched sections (4–5 sections for each animal, the total length of OE sampled was from 85mm to 160mm). OE length was determined by tracing the outline of the epithelium basal lamina using Axiovision® software. The data is represented as mean±s.d. and two-tailed Student’s t-test was used for statistical analysis.
We thank Dr. P. Mombaerts for the OR-reporter mice, OMP-lacZ and OMP-CRE mice (Rockefeller University); Dr. F. Margolis for the OMP antibody (University of Maryland); Dr. S. Warming for OAZ cDNA (National Cancer Institute). We thank members of the Reed lab and Dr. J. Nathans for helpful discussions. This work was supported by the NIDCD and the Howard Hughes Medical Institute.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.