The Hox genes are key regulators of the regional pattern formation along the anteroposterior (AP) and other embryonic axes. These genes are involved in transducing positional information to the precursors of the embryonic axial and paraxial structures. In vertebrates, the restricted localization of each Hox gene product is important for the correct patterning of tissues surrounding the rostral expression boundary and is essential for the proper orchestration of embryonic morphogenesis (Deschamps et al., 1999).
Hox gene expression in various body regions is collinear with their genomic organization in such a way that the genes at the 3′ end of the cluster are expressed more anteriorly and those at the 5′ end are expressed more posteriorly. The physical position of the Hox gene in the cluster corresponds not only to its expression domain along the primary or secondary body axes (Schughart et al., 1988; Duboule and Dolle, 1989; Graham et al., 1989; Deschamps et al., 1999), but also to its time point of activation (Izpisua-Belmonte et al., 1991) in addition to its response to the morphogen retinoic acid (RA) (Simeone et al., 1990). Generally, the Hox genes located at the 3′ end of the cluster are expressed earlier and are more anterior and more sensitive to induction by RA than those in a more 5′ position.
The Hoxc8 expression pattern during mouse embryogenesis has been studied extensively using RNA in situ hybridization (Awgulewitsch et al., 1986; Utset et al., 1987; Breier et al., 1988; Holland and Hogan 1988; Le Mouellic et al., 1988; Erselius et al., 1990). There is general agreement regarding the overall expression domains of Hoxc8 in the mesodermal derivatives of the thoracic region and in the neural tube at the cervicothoracic levels. Because all these expression domain analyses were performed with the embryo section, there has been some difficulty in interpreting the pattern three-dimensionally. Therefore, this study used whole-mount in situ hybridization and compared the results with the protein pattern examined with the monoclonal antibodies against Hoxc8 in early and midgestation embryos (Belting et al., 1998).
MATERIALS AND METHODS
The RNA was isolated from freshly dissected embryos according to the modified guanidium thiocyanate method (Sambrook et al., 1989) and subjected to RT-PCR. Reverse transcription (RT) was performed with 2 μg of the RNA and 2/25 (vol/vol) of the RT reaction was used for PCR amplification using the following conditions: denaturation for 5 min at 94°C, then 30 cycles of 94°C for 30 sec, 55°C for 30 sec, and 72°C for 1 min. The primers for the Hoxc8 gene (sense, 5′-CACGTCCAAGACTTCTTCCACCACGGC; antisense, 5′-CACTTCATCCTTCGATTCTGGAACC) and for the control β-actin (sense, 5′-CATGTTTGAGACCTTCAACACCCC; antisense, 5′-GCCATCTCCTGCTCGAAGTCTAG) were purchased from Bioneer (Taejon, South Korea).
The plasmid pBS:C8 harboring a 726 base pair murine Hoxc8 cDNA was obtained from Dr. P. Gruss at the Max Planck Institute of Biophysical Chemistry (Göttingen, Germany). In order to generate the pGEM:C8, Hoxc8 cDNA was isolated from pBS:C8 after digestion with the restriction endonucleases BamHI and EcoRI and subcloned into the same sites of the vector, pGEM-7zf (Promega, Fitchburg, WI), containing the T7 and SP6 promoters at each end of the multicloning site. The antisense and sense RNA probes were generated by linearizing the pGEM:C8 plasmid with either BamHI or EcoRI enzymes, which was followed by polymerization with the T7 and SP6 RNA polymerases (Boehringer Mannheim, Mannheim, Germany), respectively, in the presence of the nucleotide mixture containing the digoxygenine-labeled CTP (Boehringer Mannheim) for 2 hr at 37°C. The RNAs were precipitated with LiCl and EtOH and used as the probes for the whole-mount as well as the cold in situ hybridization.
Preparation of Embryos
For mating, pairs of ICR mice (Samtako, Osan, South Korea) were caged together in the late afternoon and the females were examined for the presence of a vaginal plug in the following morning, which was defined as 0.5 day postcoitum (dpc). The pregnant females were sacrificed at the required gestational stages by a cervical dislocation and the embryos were dissected free of the maternal and extraembryonic tissue in PBS. For in situ hybridization, the embryos were fixed in 4% paraformaldehyde (PFA)/PBS at 4°C for 1–12 hr, depending on the age of the embryos. After washing with PBT, 25%, 50%, 70% MetOH/PBT, and 100% MetOH, the embryos were stored at −20°C until needed. For the frozen sections, the embryos were transferred to 30% sucrose overnight at 4°C, mounted in tissue tek (Miles, Elkhard, IN), covered with Tissue Freezing Medium (Triangle Biomedical Science, Durham, NC), then stored at −70°C.
Whole-Mount In Situ Hybridization
Whole-mount in situ hybridization was performed according to the method developed by Wilkinson (1992) with a slight modification. The embryos were rehydrated with 75%, 50%, 25% MetOH/PBT and PBT, then bleached with 6% H2O2/PBT for 1 hr. After treating the embryos with 10 μg/ml proteinase K/PBT for 2–15 min, depending on the embryo stages, 2 mg/ml glycine was added to inactivate the proteinases, and 0.2% glutaraldehyde/4% paraformaldehyde/PBT was added for refixation. After washing in PBT, the embryos were incubated in a prehybridization solution (50% formamide, 5 × SSC, pH 4.5, 50 μg/ml yeast tRNA, 1% SDS, 50 μg/ml heparin) at 70°C for 1 hr, then incubated in a hybridization solution containing 1 μg/ml digoxygenine-labeled RNA probe at 70°C overnight. On the following day, a posthybridization wash was carried out twice with solution 1 (50% formamide, 5 × SSC, pH 4.5, 1% SDS) for 30 min each, three times in a mixed solution containing one volume of each solutions 1 and 2 (0.5 M NaCl, 10 mM Tris-Cl, pH 7.5, 0.1% Tween 20) for 10 min each, and three times with solution 2 at 70°C. The embryos were incubated twice in solution 2 containing 100 μg/ml RNaseA for 30 min at 37°C. After washing with TBST, the embryos were preblocked in 10% sheep serum/TBST for 90 min and incubated in the presence of serum and antidigoxygenine antibodies (Boehringer Mannheim) at 4°C for 1 hr. A postantibody wash was performed three times with TBST for 5 min and five times with NTMT for 1 hr, and the embryos were incubated with NTMT containing 4.5 μl NBT (Boehringer Mannheim), 3.5 μl BCIP (Boehringer Mannheim) per ml in the dark. When color was developed to the desired extent, the embryos were washed twice with PBT. For the paraffin section, the embryos were embedded in a paraplast after clearing with benzene and dehydrating with ethanol.
Cold In Situ Hybridization With Frozen Section
The frozen embryos were cut at 6–12 μm and transferred to pretreated slides (Fisher Superfrost Plus, Pittsburgh, PA). The sections were rehydrated in PBS and postfixed in cold 4% PFA/PBS for 10 min. They were then washed with PBS three times for 10 min each. Acetylation (295 ml H2O, 4 ml triethanolamine, 0.525 ml HCl, and 0.75 ml acetic anhydride) was followed by washing with PBS three times. After applying 500 μl of a prehybridization solution into each embryo section, the slides were incubated in a humidified chamber containing 5 × SSC for 2 hr, a hybridization solution containing 200–400 ng/ml RNA probe was applied for 5 min at 80°C, and slides were further incubated in a humidified chamber containing 5 × SSC and 50% formamide at 72°C overnight. After removing the coverslips by submerging the slides in 5 × SSC at 72°C, the samples were incubated in 0.2 × SSC for 3 hr, 0.2 × SSC for 5 min, and buffer B1 (0.1 M Tris, pH 7.5, 0.15 M NaCl) including levamisole for 5 min. An antidigoxygenine antibody (1:5,000 dilution in B1 buffer) was then added to the slides and further incubated in a humidified chamber overnight at 4°C. The embryos were washed with a B1 buffer three times for 5 min each, then equilibrated with a B3 buffer (0.1 M Tris, pH 9.5, 0.1 M NaCl, 50 mM MgCl2). After applying 60–70 μl of a B4 solution (4.5 μl/ml NBT, 3.5 μl/ml BCIP, 0.24 mg/ml levamisole in B3) onto each embryo section, the slides were covered with parafilm and incubated in the dark for approximately 6 hr to 3 days in a humidified chamber. When the color was developed to the desired extent, the reaction was quenched by washing with PBS, and the slides were mounted with a Universal Mount (Huntsville, AL) after air-dry.
Hoxc8 expression was first observed around 8.5 dpc in the posterior ectodermal and mesodermal cells, not in the cephalic head fold area or in the neural fold, and the tube just closed at the level of somites 4–5 (Fig. 1A and B). Hoxc8 expression was detected in the posterior neural plate at the level of the unsegmented presomatic mesoderm (Fig. 1F and G) and more posteriorly in the mesodermal derivatives, including the paraxial, intermediate, and the proximal part of the lateral plate mesoderm (Fig. 1G). However, Hoxc8 expression was not detected in the notochord.
By 9.5 dpc, the anterior boundary of Hoxc8 expression in the case of the neural tube and paraxial mesoderm was restricted to the level of somites 10 and 16, respectively (Fig. 1C and D). In contrast to the overall expression of Hoxc8 in the whole posterior neural plate at 8.5 dpc, the posterior boundary in the paraxial mesoderm as well as the neural tube was established clearly with an exceptional expression at the tail bud (Fig. 1C). The expression domain in the paraxial mesoderm spanned 5–6 somites, including robust expression from somites 16 to 20. In the case of the ectoderm-derived neural tube, strong expression was detected in the region between the 10th and the 13th somite level and spread gradually toward the more posterior levels (Fig. 1C and D). This kind of position-specific pattern of Hoxc8 expression was maintained throughout the later stages of development.
Around 10.5 dpc, the neural epithelium of the neural tube began to differentiate into three distinct layers, the inner ependymal, the outer marginal, and the intermediate mantle layers. Hoxc8 expression was specifically localized in the ventral horn in the mantle layer. No expression was detected in the ependymal layer, the floor plate, or the roof plate (Fig. 1J). Since the morphology of the neural tube begins to change fundamentally with four columns, two dorsal (alar) and two ventral (basal) columns at this stage, two longitudinal lines of expression (Fig. 1I) and ventral expression (Fig. 1H) were clearly detected in the trunk neural tube area. This is in contrast to the previous stage where one longitudinal line (Fig. 1D) but no dorsoventral restriction (Fig. 1C) was observed along the neural tube. The ventrally restricted expression pattern in the neural tube was maintained through 11.5 dpc (Fig. 1K–M). In the case of the mesoderm derivatives, the sclerotome on the way to the notochord and neural tube away from the dermomyotome expressed Hoxc8 at the level of somites 16–20, while the expression was not detected in the neural tube, spinal ganglia, or notochord at these levels. Weak expression was also detected at the intermediate mesoderm derivatives, including the mesonephric duct and the mesonephric tubule (Fig. 1P). By 12.5 dpc, forward progression of the expression was observed: to the anterior boundary of the forelimb in the neural tube, and to the posterior boundary of the somite level, where the forelimb arises in the case of the mesoderm (Fig. 1N and O).
Since a low amount of expression is difficult to detect using whole-mount in situ hybridization, RT-PCR was applied. The 11.5 dpc embryos were sliced into seven parts along the anterior-posterior axis (Fig. 2A) and the total RNA purified from each slice was used as templates for RT-PCR. As shown in Figure 2A, strong Hoxc8 expression was detected in slice 4 containing the forelimb and the upper thoracic region and slice 5 harboring the lower thoracic region. However, the rest of the area including the brain (slices 1 and 2) and cervical area (slice 3) still expressed Hoxc8 in comparatively low amounts. In the case of the temporal expression pattern, Hoxc8 was expressed in most of the stages of embryonic development from 8.5 to 17.5 dpc (Fig. 2B).
Important information concerning the functions of certain homeobox genes during the development of the fruit fly Drosophila has been gained by comparing the distinct expression patterns in mutant and wild-type embryos (Gehring, 1985; Akam, 1987; Scott and Carroll, 1987). An essential outcome of these studies was that the domains with the highest expression levels of a particular homeobox gene usually correspond to the body regions most severely affected when the gene is mutated (Awgulewitsch and Jacobs, 1990). Accordingly, an important step toward determining the functions of the vertebrate homeobox genes is a detailed analysis of their spatial and temporal patterns of expression during development.
In this study, the dynamic expression pattern of Hoxc8 was determined in embryos from 7.5 to 12.5 dpc using whole-mount in situ hybridization. Previous data analyzed by in situ hybridization had some difficulties in analyzing the time-dependent expression pattern (Awgulewitsch et al., 1986; Utset et al., 1987; Breier et al., 1988; Gaunt 1988; Holland and Hogan, 1988; Le Mouellic et al., 1988; Erselius et al., 1990). Therefore, in this study, whole-mount in situ hybridization was performed because it has advantages in examining the spatiotemporal expression pattern of Hoxc8 and determining the anterior boundary of the expression three-dimensionally.
Thus far, several groups have analyzed the expression pattern of Hoxc8 using antisense RNA/oligo or antibodies either with sectioned embryos or with the whole embryos. Most reports demonstrated a high similarity with some differences. In the case of the whole-mount immunohistochemical analysis (Belting et al., 1998), Hoxc8 was first observed in the extreme caudal region of the embryo at 8.0 dpc, whereas in this experiment it was at 8.5 dpc. Because no expression was detected at 7.5 dpc through in situ (data not shown) or PCR, and rather strong expression was observed at 8.5 dpc, Hoxc8 expression must have been initiated between 7.5 and 8.5 dpc. An interesting discrepancy between the reported immunohistochemical data (Belting et al., 1998) and these results is the expression in the limb bud region. While the data of Belting et al. (1998) showed the expression at the limb bud beginning at 9.5 dpc, no expression was detected in this experiment. This can be partially explained by the fact that the two groups used different mouse strains (FVB vs. ICR mice). However, it might also be due to the protein/mRNA stability; the amount of mRNA not detectable in the limb bud in this study might be just sufficient to produce a detectable amount of the protein by being translated repeatedly. A number of Hox genes have been reported to be transcribed from alternative promoters and/or alternative splicing, resulting in transcripts differing in size, stability, as well as tissue distribution (Stroeher et al., 1986; Cianetti et al., 1990; Sham et al., 1992; Ringeisen et al., 1993; Kim et al., 1998; Patel et al., 1999). Therefore, the possibility that the region of RNA probe could possibly generate some differences in the expression pattern cannot be excluded, even though it is very unlikely. Like previous reports (Awgulewitsch et al., 1990; Belting et al., 1998), the ventrolateral expression pattern in the neural tube and sclerotomal expression pattern were also detected in this study. Because transgenic (Yueh et al., 1998) and molecular studies (Shi et al., 1999; Yang et al., 2000) have shown the evidence of Hoxc8 for the regulation of cartilage differentiation, the Hoxc8 expressed in this sclerotome must have played a role in vertebrae development. With the exception of the limb bud, the overall expression pattern in the establishment and maintenance of the anterior expression boundary in addition to the downregulation in the posterior region along the AP axis showed a strong correspondence between the mRNA and protein level. This consistency in the spatiotemporal expression pattern suggests that Hoxc8 expression is controlled at the level of transcription and provides experimental confirmation for the long-held tenet that the protein and mRNA expression for Hox genes generally coincide.