Hox genes encode a family of transcription factors containing a highly conserved DNA binding domain, the homeodomain (Levine and Hoey, 1988). Hox genes are essential for anterior-posterior (AP) body axis patterning, apendicular patterning, and organogenesis (McGinnis and Krumlauf, 1992; Krumlauf, 1994). In vertebrates, Hox genes are organized into four chromosomal clusters, homologous to those of the Drosophila melanogaster homeotic gene complex (HOM-C). Hox genes are turned on sequentially from the hindbrain to the posterior end of the embryo in a nested manner, colinear with their positions in the clusters. Hox genes at the 3′ ends of the chromosome clusters are expressed earlier and in more anterior regions of the embryo, whereas Hox genes at the more 5′ ends of the clusters are expressed at later times in embryogenesis and in more posterior regions of the embryo (Lewis, 1978; Levine and Harding, 1987; Duboule and Morata, 1994; Krumlauf, 1994). Alterations in normal Hox expression patterns lead to homeotic transformations and malformations (McGinnis and Krumlauf, 1992).
Retinoids such as RA play major roles in the regulation of gene expression during development (Means and Gudas, 1995). Most RA target gene transcriptional activation is mediated through retinoic acid receptor proteins, RARs and RXRs, which are members of the steroid/thyroid nuclear receptor superfamily (Leid et al., 1992; Mendelsohn et al., 1994a). The RAR family contains three major isoforms (RARα, RARβ, and RARγ), which are differentially expressed during embryonic development and display evolutionary conservation (Mangelsdorf et al., 1994). Nonetheless, few developmental defects were observed in mouse mutants lacking individual RAR isoforms, suggesting that they mediate overlapping functions (Lohnes et al., 1993; Lufkin et al., 1993; Mendelsohn et al., 1994b; Luo et al., 1995; Ghyselinck et al., 1997). However, compound mutant mice lacking multiple RAR isoforms died before or shortly after birth and displayed severe malformations of many organs and tissues characteristic of the fetal vitamin A-deficient syndrome (VAD) (Lohnes et al., 1994; Mendelsohn et al., 1994a; Ghyselinck et al., 1998). RARs form heterodimers with RXRs and are thought to bind cooperatively to cis-acting RAREs on the RA-responsive genes to activate transcription. High-affinity RAREs consist of direct repeats of GGTTCA sequence motifs separated by five nucleotides (DR5) (Mangelsdorf et al., 1994). A DR2 motif (GGTTCA motifs separated by two nucleotides) can also function as a binding site for RAR:RXR complexes, but these receptors have a lower binding affinity for a DR2 RARE as compared with a DR5 (Mangelsdorf et al., 1994). In contrast, DR4, DR3, and DR1 motifs favor interactions of RXR with thyroid hormone receptors (TRs), vitamin D3 receptors (VDRs), and retinoid X receptors (RXRs), respectively. Coactivators and corepressors are also involved in mediating transcriptional activation and repression by the RARs and RXRs (Robyr et al., 2000).
There is considerable evidence implicating retinoids as key signaling molecules that determine anterior-posterior (A-P) patterning through regulation of Hox gene expression. First, during the retinoic acid (RA) -induced differentiation of teratocarcinoma cells, most Hox genes are activated by RA sequentially, colinear with their position in the clusters (Simeone et al., 1990, 1991). Second, administration of all-trans retinoic acid to pregnant mice results in a more anterior expression of Hox genes and causes axial patterning defects in the central nervous system, axial skeleton, and limbs of the embryos (Holder and Hill, 1991; Kessel and Gruss, 1991; Morriss-Kay et al., 1991; Conlon and Rossant, 1992). Third, targeted inactivation of retinoic acid receptors in mice results in mice that exhibit developmental abnormalities that correspond to those seen in loss-of-function mutations of Hox genes (Lohnes et al., 1994; Mendelsohn et al., 1994b). In addition, targeted disruption of the retinaldehyde dehydrogenase-2 (Raldh2) gene, which encodes an enzyme required for RA synthesis from retinaldehyde, results in the generation of mice that exhibit a developmental block at E10.5, demonstrating that embryonic RA synthesis is essential during early embryogenesis (Niederreither et al., 1999). Finally, studies carried out both in vitro and in vivo have led to the identification of conserved RA-response elements (RAREs) in the flanking regions of several Hox genes, including Hoxa1, Hoxb1, Hoxa4, Hoxb4, and Hoxd4. These RAREs can confer RA inducibility on a lacZ reporter gene in embryonal carcinoma (EC) cells (Langston and Gudas, 1992; Popperl and Featherstone, 1993; Ogura and Evans, 1995a,b; Langston et al., 1997) and RA inducibility and regulated expression in transgenic mouse embryos (Marshall et al., 1994; Studer et al., 1994; Frasch et al., 1995; Morrison et al., 1996; Zhang et al., 1997; Huang et al., 1998; Packer et al., 1998; Zhang and Featherstone, 2000). Mice with a targeted mutation in the DR2 RARE exhibit a reduction in the early neuroectodermal and mesodermal expression of Hoxb1 (Studer et al., 1998). In addition, a targeted mutation in the Hoxa1 DR5 RARE results in a partial phenotype of the Hoxa1 total knockout, with mice exhibiting rhombomere and cranial nerve abnormalities (Dupé et al., 1997). There is evidence that some additional Hox genes, such as Hoxc6, are induced directly by RA during limb development (Oliver et al., 1990). The discovery of retinoic acid-response elements (RAREs) in the regulatory regions of multiple Hox genes suggests that transcriptional activation of at least some Hox genes by RA occurs as a primary response, as has been shown for Hoxa1 in EC cells (LaRosa and Gudas, 1988). These RAREs are, therefore, most likely a key part of the regulatory machinery underlying the temporal and spatial regulation of Hox gene expression in development (Moroni et al., 1993; Gudas, 1994; Conlon, 1995; Doerksen et al., 1996; Gould et al., 1998).
The murine Hoxb-1 gene is located at the 3′ end of the HoxB homeobox gene cluster on mouse chromosome 11 (Wilkinson et al., 1989; Frohman et al., 1990). The Hoxb-1 gene shows two phases of expression in the developing embryo. Hoxb-1 mRNA is first detectable at 7.5 days post coitum (dpc) in the posterior half of the embryo, in the neuroectoderm, the primitive streak, and in the newly formed mesoderm rostral to the node. By the early somite stage at 8.5 dpc, Hoxb-1 mRNA expression has become divided into two domains: the prospective rhombomere 4 in the hindbrain and the posterior half of the embryo, which includes the posterior neural tube, paraxial (somite) mesoderm, and gut (Wilkinson et al., 1989; Frohman et al., 1990; Murphy and Hill, 1991). The anterior expression of Hoxb-1 induced by exogenous RA transforms rhombomere 2 to a rhombomere 4 identity (Marshall et al., 1992). Therefore, the Hoxb-1 gene is involved in specifying the identities of segments in the hindbrain. A knockout of the Hoxb-1 gene results in alterations in rhombomeric identity, and in failure to form a functional facial (VIIth) nerve (Goddard et al., 1996; Studer et al., 1996). Hoxb1 also controls aspects of the sonic hedgehog and Mash1 signaling pathways (Gaufo et al., 2000). Hoxa1/Hoxb1 double mutants exhibit loss of both rhombomeres 4 and 5, loss of the second branchial arch, and loss of second branchial arch–derived tissues (Gavalas et al., 1998; Rossel and Capecchi, 1999). Hoxa1/Hoxb1 double knockout mice also display lung hypoplasia, with from five smaller lobes to only two lobes observed in these mice (Rossel and Capecchi, 1999). Collaborating with Hoxa1, Hoxb1 also functions in the patterning of neural crest–derived craniofacial structures (Gavalas et al., 1998; Barrow et al., 2000).
Analysis of the 3′ flanking region of the Hoxb1 gene in transgenic mice led to the identification of two RAREs in the Hoxb1 3′ flanking DNA sequence. Marshall et al. (1994) identified a functional DR2-type RARE located 1.5 kb 3′ of the Hoxb1 coding region; this DR2 RARE mediates the RA responsiveness of the Hoxb-1 gene early in embryogenesis and acts as a neural enhancer in transgenic animals. We identified an RA-inducible enhancer (RAIDR5) containing a DR5 type RARE located approximately 6.5 kb 3′ of the murine Hoxb-1 coding region. We have shown that this DR5 RARE is required for RA induction of Hoxb-1 transcription in embryonal carcinoma cells (Langston et al., 1997) and for Hoxb1 expression in the foregut of transgenic animals (Huang et al., 1998). The DR5 containing enhancer was shown to be highly conserved between the Hoxa1 and Hoxb1 genes in mouse and human; this enhancer also includes two other blocks of highly conserved sequences called CE1 and CE2 (Thompson et al., 1997). In contrast, the DR2 RARE is apparently not conserved between the Hoxb1 and Hoxa1 genes; no DR2 RARE has been identified in the 3′ flanking region of the Hoxa1 gene. It is not known whether or not the regulatory functions of the Hoxb1 DR5 RARE have been conserved compared with those of the Hoxa1 DR5 RARE. We also do not understand why the Hoxb1 gene has two types of 3′ RAREs, a DR2 and a DR5 RARE.
Thus, we have now examined whether the 3′ Hoxb1 DR2 and DR5 RARE functions are overlapping or independent in the transactivation of RA responses in vivo. We show that the Hoxb1 DR2 and DR5 RAREs play distinct but overlapping roles in the regulation of different region-specific expression of Hoxb1, as assessed by lacZ transgene expression.
Mutagenesis of Hoxb1 RAREs and the Responses of Hoxb1/lacZ Constructs to RA in F9 and P19 Cells
Transgenic animal analysis of several Hoxb1 3′ regulatory regions has previously identified two conserved retinoic acid–inducible enhancers that contain DR2 and DR5 RA responsive elements (RAREs), respectively. The DR2 RARE is located 1.5 kb 3′ of the Hoxb1 coding region and is required for Hoxb1 early expression in the neuroectoderm in headfold stage embryos, and for the RA-induced anterior extension in the neuroectoderm (Marshall et al., 1994; Studer et al., 1998). The DR5 RARE is located approximately 6.5 kb 3′ of the Hoxb1 coding region and is required for later Hoxb1 expression in the foregut region and for its response to exogenous RA in the foregut (Huang et al., 1998). The foregut gives rise to esophagus, stomach, liver, and pancreas, as well as to the lungs. However, it is not clear whether there are any cooperative interactions or functional redundancies between these two types of RAREs in the activation of Hoxb1 regional-specific expression and RA responsiveness in vivo. To address these questions, we generated new transgenic lines with Hoxb1/lacZ reporter constructs containing wild-type DR2 and DR5 RAREs; lines with a mutated DR2 RARE and wild-type DR5 RARE; lines with a wild-type DR2 and a mutated DR5 RARE; and lines with mutations in both the DR2 and DR5 RAREs (Fig. 1). As described previously, the wild-type Hoxb1/lacZ reporter construct (WT) contains 6 kb of genomic sequence 5′ to Hoxb1 coding sequence (Huang et al., 1998). The lacZ reporter gene is ligated into the Hoxb1 coding sequence, creating an in frame fusion at amino acid 34 of Hoxb1 of exon 1. The construct also contains the Hoxb1 intron, exon 2, and 7.5 kb of 3′ genomic sequence, which includes both the DR2 and DR5 RA-responsive enhancers (Huang et al., 1998). The mutant DR2Hoxb1/lacZ reporter construct (designated as DR2m) is identical to the wild-type construct except that a 5 base pair mutation within the RARE direct repeats was created (Fig. 1). This mutation within the DR2 RARE element eliminated RAR/RXR binding to this RARE element (Marshall et al., 1994). The mutant DR5Hoxb1/lacZ reporter construct (DR5m) has been described previously (Huang et al., 1998). The DR2/DR5 RARE double-mutant construct (designated as 2m5m) is identical to the wild-type construct, except for the mutations in both the DR2 and DR5 RAREs.
To determine how each RARE contributes to the RA responsiveness of Hoxb1 and to assess potential functional interactions between the two RAREs, two murine embryonal carcinoma (EC) cell lines, F9 and P19, were transiently transfected with the wild-type Hoxb1/lacZ and three different RARE mutant constructs (Fig. 2). The cells were either untreated or cultured in the presence of RA for 24 hr. F9 stem cells differentiate into primitive endoderm when treated with RA, whereas P19 cells differentiate into cell types related to mesoderm and neuroectoderm lineages when treated with RA (Gudas, 1994). Our results show that in F9 cells, a mutation of the DR2 RARE alone had no effect on the RA-inducible activity of the lacZ reporter gene; the degree of RA inducibility was similar to that of the WT construct (Fig. 2A). The DR5m construct exhibited reduced lacZ activity in response to RA, as shown previously (Huang et al., 1998). Mutation of both the DR2 and DR5 RAREs resulted in a reduction of the RA responsiveness of lacZ reporter gene to 1.3-fold (Fig. 2A). This result indicates that the DR5, but not the DR2 RARE, is required for full RA-induced transactivation of the lacZ reporter gene in F9 cells.
In P19 cells, a mutation in the DR2 RARE resulted in a loss of RA responsiveness of the lacZ reporter, compared with the WT construct, which exhibited a 2.3-fold increase in activity in response to RA (Fig. 2B). Mutation of the DR5 RARE resulted in a slight reduction in the RA responsiveness of the lacZ reporter (1.8-fold vs. 2.3-fold for the WT construct). Mutation of both the DR2 and DR5 RAREs (2m5m) resulted in a complete loss of the RA-associated increase. These results indicate that transactivation of the Hoxb1/lacZ reporter gene by RA in P19 cells is dependent on the presence of the DR2 RARE and, to a lesser extent, on the DR5 RARE.
These results are consistent with our previous observations that the DR2 RARE, in the context of a smaller 2-kb EcoRI-Hind3 fragment upstream of the CAT reporter, is able to confer RA responsiveness on a heterologous promoter in P19 cells but not in F9 cells. In contrast, the DR5 RARE mediates RA responsiveness in both F9 and P19 embryonal carcinoma cell lines, and in the embryonic stem cell (ES) lines tested (Langston et al., 1997). These results are consistent with the observed absence of a DNase I-hypersensitive site at the DR2 RARE in F9 cells (Langston and Gudas, 1994; Langston et al., 1997). Together, these data demonstrate that the DR2 and DR5 RAREs are differentially used for activating the Hoxb1/lacZ reporter gene in two different EC cell lines, which differentiate along different pathways in response to RA treatment. We next examined the roles of these two types of RAREs in gel shift assays.
Similar Nuclear Proteins Bind to Both the Hoxb1 DR2 and DR5 RAREs In Vitro
To determine whether there were any differences in the nuclear proteins that bound to the Hoxb1 DR2 vs. DR5 RARE, oligonucleotides containing either the DR5 (Fig. 3A) or the DR2 RARE (Fig. 3B) were radiolabeled and used in gel mobility shift assays. The nuclear extracts used were prepared from F9 or P19 cells treated with RA for 24 hr (Fig. 3). Several DNA-protein binding complexes were observed, including two RARE-specific binding complexes (c1, c2) and two nonspecific binding complexes (c3, c4). The electrophoretic patterns formed with the nuclear extracts from F9 cells and P19 cells were very similar. To compare the relative binding specificity of the nuclear proteins to the DR2 and the DR5 RARE oligonucleotides, binding competition experiments were performed by incubating various unlabeled duplex oligonucleotides with nuclear extracts for 5 min before the addition of labeled probes.
Figure 3A shows competition experiments by using the Hoxb1 DR5 RARE as a probe. The unlabeled DR5 RARE oligonucleotide (DR5-B1), in 100-fold excess, competed well for the nuclear proteins bound to oligonucleotide probe containing the radiolabeled DR5-B1 RARE (Fig. 3, lane 1). The unlabeled DR2 RARE oligonucleotide (DR2), in 100-fold excess, competed as efficiently as unlabeled oligonucleotide containing the consensus DR5 RARE derived from the RARβ (DR5-β) (Fig. 3A, lanes 3 and 4). Nuclear proteins bound to the radiolabeled DR2 probe migrated similarly in the electrophoresis as compared with those binding to the radiolabeled DR5 RARE probe (Fig. 3A, lanes 11, 12 vs. lanes 1, 6).
Figure 3B shows gel shift experiments using a radiolabeled DR2 RARE probe. The unlabeled oligonucleotide with a mutation within the DR2 RARE (DR2m), in 100-fold excess, did not compete for the specific nuclear proteins (c1 and c2) bound to the radiolabeled wild-type DR2 RARE probe (Fig. 3A, lanes 4 and 9). In contrast, the DR5 RARE from Hoxb1 (DR5-B1), as well as the RARβ DR5 RARE (DR5-β) both competed very efficiently for the specific proteins bound to the radiolabeled WT DR2 probe (Fig. 3B, lanes 2, 3, and 10). Taken together, these results suggest that very similar nuclear proteins bind to both the DR2 and the DR5 RAREs in both F9 and P19 cells. Thus, these data from the in vitro gel shift experiments cannot explain the differences in the transcriptional responses of the Hoxb1 lacZ constructs containing the DR2 vs. the DR5 RARE in the two cell lines (Fig. 2).
DR2 and DR5 RAREs Play Distinct Roles in Hoxb-1 Regional-Specific Expression During Embryogenesis
To examine the functions of the two different Hoxb1 3′ RAREs and their potential cooperative interactions in vivo, constructs identical to those used for the cell culture analysis mentioned above were used to generate transgenic mice (Fig. 1). At least two independent transgenic founder lines were generated for each construct, and very similar expression patterns were obtained for all of the transgenic founder lines for each construct, with only minor variations in the overall staining intensity levels.
The expression patterns of the embryos carrying the wild-type Hoxb1/lacZ transgene compared with embryos carrying each of the different RARE mutant constructs at E7.5 are shown in Figure 4. Expression of the Hoxb1/lacZ wild-type (WT) construct is detected in the primitive streak and in the newly formed mesoderm in the posterior half of the embryos (Fig. 4A). The anterior boundary of the transgene expression extends to a position near the node in the bottom of the embryonic cylinder. The expression pattern observed recapitulates that of endogenous Hoxb1 mRNA expression (Frohman et al., 1990; Murphy and Hill, 1991). A similar pattern of transgene expression is seen in the embryos with either the DR5m or DR2m Hoxb1/lacZ transgene. LacZ staining is observed in the primitive streak and in the mesoderm rostral to the node (Fig. 4C,E). In contrast, in embryos expressing the DR2/DR5 double-RARE mutant (2m5m) transgene (Fig. 4G), lacZ reporter expression is detected at a significantly lower level compared with that of the WT transgene. The anterior boundary of the 2m5m transgene construct expression remains similar to that of the WT transgene, however. These results suggested that both the DR2 and DR5 RAREs contribute to the generation of the normal level of transgene expression in the primitive streak and mesoderm in pre–headfold stage embryos.
In embryos of 8.5–9.5 dpc (Figs. 4B, 5), which express the WT reporter construct, lacZ expression is seen in rhombomere 4 (r4), the foregut, the posterior neural tube, lateral mesoderm, paraxial mesoderm (somites), and notochord (Huang et al., 1998). The expression pattern observed recapitulates that of endogenous Hoxb1 mRNA expression (Frohman et al., 1990). In addition, we also observed WT Hoxb1 transgene expression in the forelimb buds starting from embryonic day 9.0. Expression is only observed within a wedged area in the anterior-proximal region of the forelimb bud (Fig. 5B).
In embryos containing the Hoxb1 DR5 mutant (DR5m) transgene, lacZ staining is seen in regions similar to those observed in the embryos carrying the wild-type construct, except for the lack of expression in the foregut region as previously described (Figs. 4D, 5D) (Huang et al., 1998). In embryos containing the Hoxb1 DR2 mutant (DR2m) transgene, lacZ staining is seen in rhombomere 4 (r4), the foregut, forelimbs, somites, and the notochord at levels similar to those observed in animals carrying the wild-type transgene (Figs. 4F, 5G). No lacZ staining is seen in the neural epithelium, except in r4. Expression of the transgene in the foregut is reduced compared with that in the embryos containing the WT transgene. These results indicate that the DR2 RARE is required for lacZ transgene expression in the neural epithelium. Because the DR5 is intact in this construct, these results also demonstrate that the DR5 RARE cannot substitute for the DR2 RARE in driving Hoxb1 expression in the neural epithelium. This finding is in contrast to the previous reported functional role of the evolutionarily conserved DR5 RARE enhancer of Hoxa1 as a neural enhancer (Frasch et al., 1995; Dupé et al., 1997).
The transgene expression pattern of embryos carrying the Hoxb1 DR2/DR5 RARE double-mutant construct (2m5m) is shown in Figure 4H, 5J. LacZ staining is seen in r4, the somite mesoderm, and the notochord at a significantly lower level compared with that of the WT construct at 8.5 dpc (Figs. 4H, 5J). Transgene expression in both the posterior neural tube and in the foregut is abolished. These results suggest that the DR2 and DR5 RAREs both contribute to generation of the maximal level of transgene expression in r4 and in the somite paraxial mesoderm. These results also indicate that the DR2 and DR5 RAREs are required for normal expression of the transgene in both the posterior neural tube and the foregut region.
In 10.5–13.5 dpc embryos, transgene expression in the wild-type and in the various RARE mutant lines is greatly reduced relative to that seen in earlier stages, but expression of the transgene is still observed in all of the same regions in which expression is seen in younger embryos (Fig. 5). Transgene expression retreats posteriorly and is observed in the posterior neural epithelium and somite mesoderm in the tail. Interestingly, lacZ transgene expression is not observed in the forelimbs in the 2m5m transgenic embryos, suggesting that Hoxb1 3′ DR2 and DR5 RAREs have redundant functions for Hoxb1 lacZ transgene expression in the forelimb.
Retinoic Acid Responses Are Differentially Altered in Hoxb1 Transgene Constructs With DR2 and DR5 RARE Mutations
To examine the effects of administration of exogenous RA on transgenic embryos, wild-type, nontransgenic female mice were mated with males carrying the transgenic lines Wt, DR2m, DR5m, and 2m5m. Pregnant females were treated with 20 mg/kg body weight all-trans retinoic acid by oral gavage on specified times. The embryos were then isolated 4–16 hr after the RA treatment.
The effects of exogenous RA treatment on the WT and the DR2m, DR5m, and 2m5m transgenes at 7.5 dpc are shown in Figure 6. We first treated the pregnant females at 7.5 dpc and harvested the embryos 4 hr after the treatment, because there were no morphologic abnormalities apparent in embryos 4 hr after RA treatment. In contrast, administration of RA to pregnant females with embryos at 6.75 dpc, followed by harvesting 16 hr later at 7.5 dpc, resulted in some teratogenic defects. After 4 hr of RA treatment, embryos harboring the WT transgene construct showed anterior expansion of lacZ transgene staining and a dramatic increase in the levels of the lacZ staining in the primitive streak and in the newly formed mesoderm; ectopic transgene staining in extraembryonic compartments was also observed (Fig. 6A–D). Embryos carrying the DR5m Hoxb1/lacZ transgene showed enhanced transgene staining after RA treatment compared with that seen in the embryos carrying the WT transgene construct, but no anterior expansion and no expression in extraembryonic tissues (Fig. 6C–F). Embryos carrying the DR2m transgene showed similar levels of transgene staining compared with the WT transgene and no anterior expansion, whereas RA treatment induced ectopic expression in extraembryonic tissues such as the allantois and amnion (Fig. 6G–J). Embryos with the 2m5m transgene showed no response to RA treatment. No anterior expansion and no observable changes in the transgene expression levels were detected after RA, and no transgene expression in extraembryonic tissues was seen (Fig. 6K–N). These data suggest that the DR2 and DR5 RAREs both contribute to the production of RA-regulated transgene expression in the primitive streak and mesoderm. The DR5 RARE is required for RA-induced ectopic expression in the extraembryonic compartments (exocolon, amnion mesoderm; parietal, and visceral yolk sac). The DR2 RARE is required for the up-regulation of transgene expression in primitive streak and mesoderm by RA.
The effects of RA treatment on embryos carrying the WT and the various mutant RARE mutant transgenes at 8.5 dpc are shown in Figure 7. The WT transgene recapitulates the RA response of the endogenous Hoxb1 gene (Conlon and Rossant, 1992), showing anterior expansion of lacZ transgene staining into the midbrain and forebrain and anterior extension of the foregut staining, in agreement with our previously data (Huang et al., 1998). In addition, the overall level of lacZ transgene staining after RA treatment is higher than that in the control, untreated embryos containing the WT Hoxb1 transgene (Fig. 7A,B). After RA treatment, DR5m embryos show anterior extension of transgene expression into the midbrain such that the sharp r4 expression becomes more expanded and irregular, and there is still no foregut staining (Fig. 7C,D). No transgene expression is detected in extraembryonic tissues such as the yolk sac and allantois in DR5m RA-treated embryos (Fig. 7D). RA-treated DR2m embryos exhibit anterior expansion of the transgene expression in the foregut. Neural epithelium staining in DR2m embryos displays an anterior shift from r4 after RA treatment (Fig. 7E,F). The allantois and yolk sac are stained for lacZ in the RA-treated DR2m embryos (data not shown). The 2m5m double-mutant embryos show no response to exogenous RA treatment (Fig. 7G,H). The weak r4 staining remains as a defined band. The posterior paraxial mesoderm staining remains at the same level compared with that in untreated embryos. No staining is detected in the extraembryonic tissues in the 2m5m embryos.
These data are consistent with previous data suggesting that the DR2 RARE is required for early anterior expansion of the reporter in response to RA in neural epithelium (Marshall et al., 1994). Moreover, these results also demonstrate that the DR5 RARE cannot substitute for the function of the DR2 RARE as an early neural enhancer and RA responsive enhancer in the neural tube, because the DR5 RARE is intact in this construct. The DR5 RARE is required for RA-induced ectopic expression of the reporter gene in the foregut and in extraembryonic tissues. Together, both the DR2 and DR5 RARE are required for maintaining high levels of transgene expression in the somite mesoderm and posterior tail bud region.
In 9.5 dpc embryos, it has been shown previously that RA treatment causes an anterior expansion of endogenous Hoxb1 and WT transgene expression into the branchial arch units and has no effect on staining in the neuroepithelium, including r4 (Huang et al., 1998). The effect of RA treatment on WT and DR2m embryos at 9.5 dpc is shown in Figure 8. In agreement with what has been previously described, WT transgenic embryos, exposed to RA in utero at 8.75 dpc and examined at 9.5 dpc, show expansion of lacZ transgene expression into the most anterior branchial arches (Fig. 8A,B) (Frohman et al., 1990). RA-treated DR5m embryos (Fig. 8C,D) show no expression of the transgene in the foregut region and in any of the branchial arch units, as reported previously (Huang et al., 1998). RA-treated DR2m embryos (Fig. 8E,F) show an anterior expansion of the transgene expression domain from the branchial arch 3 to branchial arch 1 (Fig. 5), similar to that seen for the WT transgene construct. After RA treatment, no expansion of transgene expression in the neural epithelium, including r4, was seen at 9.5 dpc. Only minimal changes were observed in other regions, such as the somite mesoderm. These results indicate that the exogenous RA-induced Hoxb1 anterior shift of expression in the foregut is dependent on the DR5, but not on the DR2 RARE.
For embryos carrying the 2m5m construct (Fig. 8G,H), administration of exogenous RA in utero at 8.75 dpc had no effect on lacZ expression in either the neural epithelium or the foregut endoderm at 9.5 dpc. No lacZ transgene expression was observed in either of these tissues. In addition, we could not detect any effect of exogenous RA treatment on the expression level and pattern of transgene expression in the forelimbs.
In summary, the transgene response to exogenous RA in the foregut is dependent on the DR5 but not on the DR2 RARE, and the DR2 RARE cannot substitute for the role of the DR5 RARE as a late RA-responsive element in the foregut.
Hoxb-1/lacZ staining is detected in several novel sites during organogenesis
Staining of transgenic embryos at later stages (12.5–15.5 dpc) is shown in Figure 9. It was previously reported that no endogenous Hoxb1 mRNA was detectable after 12.5 dpc (Wilkinson et al., 1989; Frohman et al., 1990), but these authors were assaying tissue sections by in situ hybridization. After 12.5 dpc, we observed that WT lacZ transgene expression was still seen in restricted regions of the embryos (Fig. 9). Many of these stained regions were similar to the stained regions seen at earlier stages, such as r4, somite-derived tissues, and the posterior neural tube.
In addition, transgene expression in the Hoxb1 WT transgenic embryos was seen in the forelimb and in several other tissues in which transgene expression was not detected at early stages (E7.5–E10.5). These tissues include the hernia gut and genital eminence (Fig. 9A–C). Between 12.5 and 14.5 dpc, the hernia gut is a structure formed during the rapid growth of the mouse digestive tract at this stage, which results in part of the intestine looping out into the umbilical cord. The hernia gut withdraws back into the body at 16.5 dpc. Transgene expression in the hernia gut is lost when both RAREs are mutated, but not in the single DR2 or DR5 RARE mutant transgenic lines (data not shown). By 14.5 dpc, lacZ staining of the WT transgene was also seen in the genital eminence (Fig. 9C), and staining was observed in the DR5m animals (Fig. 9D).
The WT Hoxb1 transgene staining in the developing limbs (Fig. 10C,D) generally recapitulated the pattern of expression of the endogenous Hoxb1 gene, as we demonstrated by whole-mount in situ hybridization (Fig. 10A,B). This transgene expression in the limbs showed some functional redundancy between the DR2 and DR5 RAREs. Staining was observed in both of the single RARE mutant lines, DR5m (Fig. 10D) and DR2m, but not in the double RARE mutant (2m5m) lines (data not shown).
The WT Hoxb1 transgene and the DR5m transgene expression patterns in the developing lung at 13.5 dpc are shown in Figure 9E,F. The WT Hoxb1 transgene exhibits an extensive staining pattern in the lung (Fig. 9E) as it undergoes branching (Schuger et al., 1993; Warburton et al., 2000), whereas lung expression in the DR5m transgenic lines is not seen (Fig. 9F). These data indicate that the expression of the Hoxb1 transgene in the lung between 12.5 and 14.5 dpc requires the DR5 RARE and that Hoxb1 plays a role in lung development. These data are consistent with earlier data showing expression of Hoxb1 in the lung bud (Kappen, 1996; Mollard and Dziadek, 1997).
Understanding how Hox genes establish their coordinated and spatially restricted expression patterns along the anteroposterior axis of the mammalian embryo is critical for elucidating their biological roles in development. Several lines of evidence indicate that retinoids are involved in regulating the graded and colinear expression patterns of Hox genes along the anteroposterior body axis. It was shown previously that the Hoxb1 3′ DR2 RARE functions as a neural enhancer and RA responsive element in the neural tube (Marshall et al., 1994), whereas the 3′ DR5 functions as a gut enhancer and RA-responsive element in the gut (Huang et al., 1998). In this study, we have now made specific mutations in each of these two different RARE enhancers and in both the DR2 and DR5 RAREs to assess their potential interactions, functional redundancies, or both, in transgenic mice. We have shown in this report that the DR2 and DR5 RAREs 3′ of the Hoxb1 gene play distinct roles. Normal neural epithelial expression and the anterior shift in expression in response to exogenous RA are dependent on the DR2 RARE, but not the DR5 RARE, whereas normal foregut expression and the anterior shift in transgene expression in the gut in response to exogenous RA require the DR5 but not the DR2 RARE. However, the DR2 and the DR5 RAREs together contribute to the overall level of Hoxb1 transgene expression in other regions such as in the rhombomere r4 and in the somite mesoderm. That the double mutant (2m5m) transgenic lines exhibited very low levels of transgene expression both in the absence and presence of exogenous RA suggests that endogenous retinoids are the major positive transcriptional signals for regulation of the Hoxb1 gene during early development and that the two RAREs are the major DNA elements involved in the Hoxb1 gene RA transcriptional response.
Retinoid Responsive Elements and Activation of the Hox Complexes
A distinguishing characteristic of Hox complexes is the correlation between the physical order of genes along the chromosome and their expression/function along the anteroposterior axis of the embryo, normally referred to as colinearity. Moving along the complexes in a 5′ direction, each successive gene adopts a progressively later, more posterior, and less RA-sensitive pattern of expression (Duboule and Morata, 1994). That temporal expression and sensitivity of Hox genes to RA in cultured cells reflects the colinearity of Hox genes in the embryo suggests that RA is instrumental in the regulation of the colinear expression of Hox genes in vertebrates. Thus, it was critical to investigate the molecular basis of the RA response to identify the DNA elements involved in colinear regulation.
The temporal windows of the competencies of different Hox genes to respond to RA are well known to differ (for review see Means and Gudas, 1995). Our experiments presented here clearly demonstrate that the 3′ DR2 and DR5 RARE enhancers are each able to mediate distinct, stable anterior shifts in neural-restricted and gut-restricted Hoxb1 transgene expression, respectively. The initial RA-responsive anteriorization of Hoxb1 gene expression in the neural epithelium (7.5–8.5dpc) is mediated by the 3′ DR2 RARE. After that, the RA-responsive anteriorization of the Hoxb1 gene expression in the foregut endoderm at and beyond 8.5 dpc is mediated by the DR5 RARE. This finding suggests that the temporal differences in the competence of the Hoxb1 gene to respond to RA in different subregions can be linked to distinct RARE-containing enhancers 3′ of the Hoxb1 gene. That RAREs are required for both the early and late responses suggests that there is a restriction on the abilities of specific Hoxb1 RARE elements to respond to ligand-activated RARs present at different times during development. Additional factors present in the cells at specific times in development may facilitate the response of the RAREs; for instance, the expression of specific coactivators could potentially vary during development (Robyr et al., 2000).
In Drosophila, it has been shown that not only are Hox genes themselves located in order on the chromosome but also that the enhancers which activate individual Hox genes in subregions of their expression domains are in a colinear order. For example, the Drosophila Contrabithorax mutation, which transposes a Ubx enhancer to a more 3′ position in the genome, causes Ubx to be expressed in a more posterior body region (Duncan, 1987). In the murine Hoxb1 gene, the DR5 RARE is located 3′ relative to DR2 RARE. The DR5 enhancer activates Hoxb1 expression in the foregut, a more posterior body position compared with the DR2 RARE-mediated Hoxb1 expression in the neural tube. Therefore, Hoxb1 gene expression in the subregions of its expression domains correlates both temporally or spatially with the order of the locations of the two RARE enhancers in the chromosome.
Differences Between DR2 and DR5 RAREs
Our results from transgenic mice suggest that the DR2 and DR5 RAREs play distinct roles in driving Hoxb1/lacZ transgene expression in different regions of the embryos. We also observed different RA-induced transactivation activities of the DR2 and DR5 RAREs in F9 and P19 EC cells (Fig. 2). This difference in the transactivation activities of the DR2 and DR5 RAREs in F9 and P19 cells may reflect the different capabilities of these cells to differentiate into primitive endoderm or neural epithelial cell lineages, respectively, after RA treatment. This idea is consistent with the requirement of the DR5 RARE for Hoxb1 expression in gut endoderm and in extraembryonic visceral endoderm vs. the requirement of the DR2 RARE for Hoxb1 expression in the neural epithelium in the transgenic animals (Fig. 11).
The gel mobility shift experiments indicated that identical proteins bound to the DR2 and DR5 RAREs (Fig. 3). This finding is consistent with the previous data that recombinant RARs can bind to both DR2 and DR5 RAREs under gel mobility shift conditions (Langston and Gudas, 1992; Marshall et al., 1994). The affinity of RARs for an RARE is affected by both the sequence of the half site and the flanking nucleotides. Notably, the sequences of the Hoxb1 DR5 RARE half sites (5′-GGTTCA-3′) are identical to those of the DR5 RARE element of RARβ promoter, an RARE for which RARs have high affinity (Fig. 12) (de Thé et al., 1990; Sucov et al., 1990). In contrast, the Hoxb1 3′ DR2 RARE sequence closely resembles the characterized DR2 RARE found in the promoter of the mouse cellular retinol binding protein I (CRBP-I). Even though the gel mobility shift data suggested that similar or identical RAR/RXR proteins bound to both the DR5 and DR2 RAREs, we cannot rule out the possibility that there are different RARs bound to these RAREs in vivo. We are currently examining the contribution of each of the different RARs (α, β, γ) to the Hoxb1/lacZ transgene expression in different regions of the embryo by mating individual RAR knockout mice with the Hoxb1/lacZ transgenic mice.
It is also likely that there are different tissue-specific coactivators or corepressors that are recruited to the RAR/RXR complexes bound to the DR2 or DR5 RAREs (Robyr et al., 2000). For example, N-CoR, a corepressor of RAR, was shown to dissociate from RAR/RXR bound to a DR5 RARE, but not from RAR/RXR bound to a DR1 element, in an RA-dependent manner (Kurokawa et al., 1995). Moreover, there have been established links between the coregulators and histone acetylation, which is a crucial event in the transcriptional response within chromatin. It will be very interesting to assess the roles of coregulators such as SMRT, N-CoR, P/CAF, CBP/p300, and Brahma proteins (Robyr et al., 2000) in the DR2 and DR5 RARE-mediated Hoxb1 transactivation in different cell types and tissues in the embryo and in the RA-induced sequential activation of Hox cluster genes. Because the RARs and RXRs are essentially ubiquitous in their expression, the selective expression of coactivators, corepressors, or both, could provide a particularly effective method to restrict RA inducibility of Hox genes to selective cell types and to enable different tissues to respond to RA at the appropriate developmental times.
Alternatively, the functional differences between DR2 and DR5 RAREs may be determined by adjacent regulatory DNA sequences. Previous studies by Ogura and Evans (Ogura and Evans, 1995a,b) suggested that the human HOXB1 3′ DR2 RARE activity was augmented by an adjacent, tissue-specific enhancer. Similarly, in HOXD4 promoter studies, a 5′ flanking sequence of the DR5 RARE was shown to be essential for the RARE enhancer activity in vivo (Morrison et al., 1996). Previously, we showed that the CE1 element in the Hoxa1 RAIDR5 enhancer was required for Hoxa1 early somite expression (Langston et al., 1997; Thompson et al., 1997, 1998). The CE1 and CE2 elements are located close to the Hoxa1 DR5 RARE in the RAIDR5 enhancer. The CE1 and CE2 elements are also found in close proximity to the DR5 RARE in the Hoxb1 RAIDR5 enhancer and may interact with the DR5 RARE in activating Hoxb1 expression in the early mesoderm and primitive streak.
Hoxb-1 vs. Hoxa-1 3′ Enhancer: RA-Responsive Enhancers From Paralogous Hox Genes
One of the findings from this study is that there is a functional difference between the Hoxb1 3′ DR5 enhancer and the highly conserved Hoxa1 3′ DR5 enhancer. The expression patterns of the Hoxa1 and Hoxb1 genes are very similar, if not completely overlapping, except for the rhombomere 4– specific expression of Hoxb1 (Murphy and Hill, 1991). Previously, it was shown that several DNA elements within the murine Hoxb1 3′ DR5 RARE (RAIDR5) enhancer, including the DR5 RARE and conserved elements CE1 and CE2, were conserved in the chicken Hoxb1 as well as in murine and human Hoxa1 (Langston and Gudas, 1992; Langston et al., 1997). Mutation of the DR5 RARE in the Hoxa1/lacZ transgene eliminated all of the neuroectodermal expression as well as the foregut expression (Frasch et al., 1995). In the current studies, we show that mutation of the corresponding conserved Hoxb1 DR5 RARE eliminated only the foregut expression, and that this mutation had no effect on the Hoxb1 transgene expression in the neural tube. Thus, our results show that, unlike the Hoxa1 DR5 RARE, the Hoxb1 DR5 RARE is not required for transgene expression and for response to exogenous RA in the neural tube.
The potential for conservation of regulatory elements between homologous and paralogous Hox genes has been widely discussed (Maconochie et al., 1996; Morrison et al., 1996). The present studies point out another level of complexity for functional conservation of the regulatory elements between paralogous Hox genes. Although the Hoxa1 and Hoxb1 genes share identical DR5 RAREs in their RA-inducible enhancers, our results suggest that these two paralogous Hox genes are controlled by differential usage of these RAREs (Fig. 11 for summary diagram). These results also indicate that the sequence of the RARE alone is not sufficient for encoding the spatial and temporal information required for the Hox gene expression.
Hoxb1/lacZ Reporter Expression in Extraembryonic Tissues and the Lung, Genital Eminence, and Forelimb
We found that multiple transgenic lines carrying the WT Hoxb1 transgene all showed similar patterns of expression in these locations at later times (12.5–14.5) (Figs. 9, 10). Thus, the observed transgene expression is independent of the transgene insertion site in the genome. We found that the transgene expression in these tissues was lost in the embryos carrying the Hoxb1 RARE double mutant (2m5m), suggesting that the transgene expression is regulated by endogenous retinoids. Consistent with our data that the Hoxb1 DR5 and DR2 RAREs are active in these tissues are previous reports that transgenic mice carrying lacZ driven by an RARβ promoter, which contains a DR5 RARE, express the transgene in many of the same tissues at similar developmental stages. Some of these locations, such as the anterior proximal region of limb bud, the umbilical cord, and the genital eminence, are regions where RARβ mRNA is known to be expressed (Mendelsohn et al., 1991; Reynolds et al., 1991). Additionally, transgenic mice carrying a construct with RARE tandem repeats linked to a heterologous promoter/lacZ transgene (Rossant et al., 1991) express the transgene in the umbilical cord and in the hernia gut. Moreover, these sites have been reported to contain localized, high levels of biologically active retinoids (Mata De Urquiza et al., 1999).
Transgene Constructs and Mutagenesis
Some of the Hoxb1/lacZ reporter transgene constructs have been described in detail previously (Huang et al., 1998). Briefly, a 15-kb murine Hoxb1 genomic DNA fragment containing 7.5 kb of the 3′ flanking region was used. To construct the Hoxb1/lacZ reporter, the Escherichia coli LacZ gene was inserted into exon 1, creating an in-frame fusion with the first 34 amino acids of the Hoxb1 polypeptide.
Multiple nucleotides within the DR2 and DR5 RAREs were mutated by PCR site-directed mutagenesis (Prelich, 1993). The DR5 RARE mutation has been described previously (Huang et al., 1998). Point mutations in the direct repeats of the DR2 RARE were chosen based on previous studies (Marshall et al., 1994); these RARE mutations have been shown to result in the loss of RAR binding in electrophoretic mobility shift assays (Langston and Gudas, 1992; Marshall et al., 1994).
To make the DR2 RARE mutation, four oligonucleotide primers were synthesized, as shown below. The first two (inside) primers carry the mutated DR2 RARE with a diagnostic SalI site (mutated nucleotides are in lower case letters and underlined sequence is the direct repeat of the RARE): 5′ -AGAaGTAcAAAtGTCgaCCCAGAATCGTGGTG-3′ (DR2M-F primer); 5′-GtcGACaTTTgTACtTCTAAGCGCCTCTGGGC-3′ (DR2M-R primer). The two outside primers are as follows: 5′-CAGTTCAGTGCCTTTGAG-3′ (BstB1 primer); 5′-GGCAG ATCGAAGCTTGATGGCAATTTCAA-3′ (Hind3 primer). The PCR products containing the mutated DR2 RARE were digested with BstB1 and Hind3 and subsequently put into full-length Hoxb1/lacZ reporter constructs in two steps. First, the BstB1- and Hind3-digested 2.1-kb PCR products were cloned into a vector containing a 10-kb ClaI-SphI fragment (fragment-1) of Hoxb1/lacZ DNA. The DR5 RARE mutation was made in a 2-kb SphI-KpnI fragment (fragment-2) as described before (Huang et al., 1998). Another KpnI-ClaI fragment (fragment-3) from the full-length Hoxb1/lacZ was prepared. The full-length Hoxb1/lacZ transgene construct was made by a three fragment ligation (1+2+3) of the DNA fragments mentioned above; this resulted in two transgene constructs identical to the Hoxb1/lacZ (WT), except for different point mutations in the DR2 RARE, DR5 RARE, or both. The resultant DR2 single mutant (DR2m) contains the DR2 mutation and an intact DR5 RARE. The DR2/DR5 double mutant (2m5m) contains mutations of both the DR2 and DR5 RAREs (Fig. 1). All of these DNA constructs were sequenced by the Cornell DNA Sequencing Facility (Ithaca, NY) to confirm that the mutations were correct and to ensure that there were no other unwanted sequence changes.
Cell Culture and Transient Transfections
F9 embryonal carcinoma cells were cultured in DMEM with 10% calf serum as previously described (Langston et al., 1997). F9 cells were transfected with Hoxb-1/lacZ plasmids by calcium phosphate coprecipitation. β-Galactosidase activities were normalized to reporter activity of a plasmid containing β-actin/CAT. Transient transfections, CAT assays, and β-galactosidase assays were performed as described previously (Langston et al., 1997). Quantitation of CAT assays was carried out by using a PhosphorImager (Molecular Dynamics).
Gel Mobility Shift Assays
Nuclear extracts were prepared from RA-treated F9 or P19 cells according to the protocol of Dignam et al. (Dignam et al., 1983). Protein concentration was quantitated with the Bradford reagent (Bio-Rad Laboratories, Richmond, CA). Gel mobility shift assays were performed as previously described (Thompson et al., 1997). Briefly, 10,000 to 30,000 cpm of γ-32P-labeled duplex oligonucleotide probe and 3–5 μg of cell extract were mixed at room temperature for 15 min in 20 μl of binding buffer containing 10 mM Tris (pH 7.5), 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 1 μg of poly(dI-dC), and 5% glycerol. The binding complex was separated by 4.5% polyacrylamide gel electrophoresis at 4°C in buffer containing 25 mM Tris, 190 mM glycine, and 1 mM EDTA at 25 mA. Gels were then dried and autoradiographed. Competition experiments were performed by incubating unlabeled duplex oligonucleotide in 100-fold excess with nuclear extracts for 5 min before the addition of radiolabeled probe. Duplex oligomers were annealed from complementary strands by heating to 90°C and cooling slowly to room temperature, and then they were gel purified. The oligonucleotides of the forward strand used in the binding assay were as follows (mutated nucleotides are in lower case letters, underlined sequences are the RAREs): 5′-TCCTTCCTGGTTCATAGAGAGTTCAGCTCTGA-3′ Hoxb1 DR5 RARE (DR5-B1); 5′-TGGGTAGGGTTCACCGAAAGTTCACTCGGG-3′ RARβ DR5 RARE (DR5-β); 5′-CTTAGAGGTAAAAAGGTCAGCCCAG-3′ Hoxb1 DR2 RARE (DR2); 5′-CTTAGAaGTAcAAAtGTCgaCCCAG-3′ Hoxb1 DR2 mutant (DR2m).
Generation of Transgenic Mice
DNA was prepared as a linearized NotI fragment with vector sequence removed and microinjected into one-cell mouse embryos (C57B/6 x CBA/J F1) at the Sloan Kettering Cancer Institute/Cornell Transgenic Mouse Facility. To identify the transgenic offspring, tail DNA was digested with EcoRI, transferred to Hybond N nylon membrane (Amersham) and hybridized with a random-primed [32P]dCTP-labeled lacZ fragment as described (Sambrook et al., 1989). Production of transgenic embryos was performed by mating transgenic males with nontransgenic females (C57BL/6) in the evening and monitoring the plugs the following morning; a plug found the following morning was considered 0.5 dpc. Embryos were removed at the appropriate gestational ages and stained for β-galactosidase activity. The expression patterns of the WT, DR5m, DR2m, and 2m5m transgenes were confirmed at all stages of development in at least two independent founder lines of transgenic mice. For each of the constructs, the expression patterns in these independent founder lines were shown to be identical. The transgenic lines of the WT/lacZ (WT) are Tg11, Tg21, and Tg36. The transgenic lines of Hoxb1/lacZ DR5m(DR5m) are Tg49, Tg51, and Tg75. The transgenic lines of Hoxb-1DR2m/lacZ (DR2m) are Tg40, Tg42, and Tg49; the lines of Hoxb-12m5m/lacZ (2m5m) are Tg8 and Tg19. The Hoxb1/lacZ transgenes were carried in the heterozygous state.
Retinoic Acid Treatment
Pregnant females were treated with RA at various times before embryo dissection. RA was administered at a dose of 20 mg/kg of maternal body weight by oral gavage (Conlon and Rossant, 1992). A stock solution of 25 mg/ml of all-trans-RA (Sigma) in dimethyl sulfoxide was dissolved just before use in corn oil so that a 0.2-ml dose of oil contained the requisite amount of RA. Control mice received corn oil alone. Embryos were isolated 12–16 hr later and stained for lacZ expression, unless mentioned specifically in the text.
β-Galactosidase Staining and Embryo Sectioning
Embryos were fixed as previously described (Conlon and Rossant, 1992) in 1% formaldehyde, 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM EDTA, and 0.02% NP-40 in PBS, for 30–90 min at 4°C. The embryos were then washed three times in PBS plus 0.02% NP-40, and stained for 12–24 hr in 1 mg/ml X-gal, 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, 2 mM MgCl2, 0.01% sodium deoxycholate, and 0.02% NP-40 in PBS, at room temperature and in the dark. Stained embryos were washed with several changes of PBS plus 0.02% NP-40. Photomicrographs were taken on a Leica dissecting microscope.
For sectioning, the embryos were dehydrated and paraffin-embedded as previously described (Means and Gudas, 1997). Sections were cut (7 μm) and mounted on glass slides. Slides were deparaffined and counterstained with eosin as described previously.
Whole-Mount In Situ Hybridization
Hoxb-1 plasmid was kindly provided by Dr. Joe Grippo. The plasmid DNA contains a 435-bp EcoRI/HindIII fragment of Hoxb-1 cDNA. Digoxigenin-labeled anti-sense probes were synthesized from linearized plasmid with HindIII digestion and transcribed in vitro by using T7 RNA polymerase as described by the manufacturer (Boehringer Mannheim Biochemicals, Indianapolis, IN). The Hoxb-1 plasmid was linearized with EcoRI, and transcribed with SP6 RNA polymerase to obtain the sense probe.
Whole-mount in situ hybridization was performed as previously described (Conlon and Rossant, 1992) with the following changes. Embryos were treated with 10 μg/ml proteinase K for 8 min at room temperature, and prehybridized for at least 1 hr at 70°C in hybridization buffer (50% formamide, 0.75 M NaCl, 1 mM EDTA, 50 μg/ml tRNA, 0.05% heparin, 1% sodium dodecyl sulfate). The hybridization buffer was replaced, single-stranded RNA probes labeled with digoxigenin were added (1 μg/ml), and embryos were hybridized overnight at 70°C.
We thank Memorial Sloan Kettering Cancer Institute Transgenic Mouse Laboratory and Rockefeller University Transgenic Mouse Facility for the microinjection of the transgenic mouse founder lines. We thank Dr. Willie Mark for helpful discussions on transgenic mice genotyping and characterization. We thank all of the members of the Gudas laboratory for helpful discussions and Taryn Resnick for editorial assistance. All authors received funding from the NIH.