Abstract
A homeobox-containing clone has been isolated from an adult mouse kidney cDNA library and shown by DNA sequence analysis to be a new isolate, Hox-6·1†. A genomic clone containing Hox-6.1 has been isolated and found to contain another putative homeobox sequence (Hox-6·2), within 7 kb of Hox-6·1. In situ hybridization of mouse metaphase chromosomes shows this Hox-6 locus to be located on chromosome 14 (14E2).
Hox-6.1 has been studied in detail and the predicted protein sequence of the homeobox is 100 % homologous to the Xenopus Xebl (formally AC1) homeobox and the human c8 homeobox (Carrasco et al. 1984; Boncinelli et al. 1985; Simeone et al. 1987). Southern blotting shows that the DNA sequence encoding Hox-6.1 is single copy.
Expression of Hox-6.1 has been studied in adult tissues and embryos by RNase protection assays, Northern blotting analysis and in situ hybridization. RNase protection assays show that Hox-6.1 transcripts are present in embryos between days and of gestation and in extraembryonic tissues at day . Adult expression is detectable in kidney and testis but not in liver, spleen and brain. One major transcript is detectable on Northern blots of kidney and day- embryo RNA. In kidney, this transcript is 2·7 kb whereas in embryos the major transcript is smaller at 1·9 kb, a much fainter band being visible at 2·7 kb. Localized expression of Hox-6.1 is observed in the spinal cord and prevertebral column of day- embryos, and in the posterior mesoderm and ectoderm of day-8) embryos. An anterior boundary of expression is located just behind the hindbrain whereas the boundary in the mesoderm is located at the level of the 7th prevertebra.
Introduction
Many of the genes that appear to play key roles in the control of embryonic development in Drosophila, such as segmentation and homeotic genes, have been found to share a conserved DNA sequence of 180 nucleotides termed the homeobox (Gehring 1985; McGinnis et al. 1984).
Sequence analysis demonstrated that this sequence is translated into protein and that the predicted homeobox protein sequences show greater homology than the corresponding DNA sequence, implying conservation of a functional domain. This protein domain is believed to be involved in DNA binding and shows discernible homology to several characterized DNA-binding proteins (Shepherd et al. 1984). Two further characteristics of homeobox-containing genes are that the genes occur in clusters in the Antennapedia and Bithorax loci and that homeobox gene expression is observed in embryos (McGinnis et al. 1984). Homeobox sequences have been demonstrated to exist in a wide variety of nondipteran genomes including sea urchin, Xenopus, mouse and man (Dolecki et al. 1986; Harvey et al. 1986; Hart et al. 1985; Boncinelli et al. 1985). In the mouse, at least 15 homeobox sequences have been identified. These sequences are all potentially protein coding and once again the predicted protein sequences show greater conservation than the nucleotide sequences. Further work has shown that mammalian homeobox genes are expressed during embryogenesis and frequently occur in clusters (Hart et al. 1985). At least two loci containing five or more homeoboxes have been described. Hox-1 on chromosome 6 and Hox-2 on chromosome 11 (Hart et al. 1985; Duboule et al. 1986). Thus, the mouse homeobox genes show many parallels with their Drosophila counterparts and must be considered as possible developmental control genes. Here we describe the identification of another mouse homeobox-gene locus, Hox-6, on chromosome 14. One gene of this locus, Hox-6.1, has been studied in detail and appears to be a homologue of the Xenopus Xebl homeobox gene and the human c8 gene (Carrasco et al. 1984; Simeone et al. 1987).
Materials and methods
cDNA library production and screening
Adult F1, male mouse kidney total RNA was extracted by precipitation in guanidine thiocyanate (Chirgwin et al. 1979). Poly(A)+ RNA was extracted by oligo(dT)-cellulose affinity column chromatography (Aviv & Leder, 1972). 10 μg of poly(A)+ RNA were used to construct a cDNA library λgt10 (Watson & Jackson, 1985). The library consisting of 5× 105 recombinant phage was amplified and screened with a nick-translated 220 bp probe derived from the murine Hox-1.5 gene containing 145 bp of homeobox sequence (McGinnis et al. 1984). One of the positive clones isolated after three successive screenings was mapped and restriction fragments subcloned. Hybridizations were carried out under low stringency conditions, lm-NaCl, 1% SDS, 10% dextran sulphate, 10 μg ml-1 salmon sperm DNA at 60°C for 12 h.
Restriction mapping, subcloning and sequencing
1–2fig of phage DNA from a single plaque was digested with restriction endonucleases, separated on a 0·8 % Trisacetate agarose gels and transferred to Gene Screen plus membranes (NEN). Hybridizations with the Hox-1.5 fragment were carried out using the same conditions as for library screening. Probes were nick translated using 32P-labelled dCTP (Amersham). The whole 540 bp EcoRI gel-purified insert of the clone was ligated into M13mp8. Single-stranded templates were prepared and sequenced by the dideoxy method (Sanger et al. 1977). Reverse primer sequencing was used on a double-stranded template to confirm the sequence (Hong, 1981). The 490 bp EcoRI/ B/II fragment of the insert was subcloned into the EcoRI/ BamHI sites of pGEM 4 (Promega Biotech) for production of riboprobes.
RNase protection assays
RNase protection was carried out essentially as described by Zinn et al. (1983). Total RNAs were extracted as described from adult F1 male mouse tissues. RNA was extracted from embryos and extraembryonic tissues obtained from natural matings between male F1 and female CFLP mice. Extraembryonic tissue consisted of allantois and amnion dissected from the embryo and decidua and contained no visible trophoblast. Midday on the day of the copulation plug was designated day of pregnancy, making the assumption that mating had taken place at midnight. 50 μg samples of RNA were dissolved in 30 μl of 80% formamide, 40mm-Pipes buffer pH 6 · 4, 400mm-NaCl and Imm-EDTA. Radioactive RNA probes complementary to Hox-6.1 mRNA were prepared after digestion of pGEM4 plasmid DNA with HindIII followed by in vitro transcription using SP6 RNA polymerase with 32P-UTP (Amersham; 50 μCi, 0 · 2ftg per 20 μl reaction). The specific activity obtained was about 1 · 0 × 108disintsmin-1μg-1. 5 × 105cts min-1 of riboprobe was added, heated at 85°C for 10 min and incubated overnight at 45 °C. 350 μl of 10 mm-Tris-HCl (pH 7 · 5), 5mm-EDTA, 300mm-NaCl containing 40 μg ml-1 RNase A and lOOi.u.ml-1 RNase T1 (Sigma), previously boiled for 5 min and allowed to cool were added and incubated at 30°C for 30 min. 2 · 5 μl of proteinase K (20 mg ml-1 ) and 20 μl 10 % SDS were added and incubated at 37°C for 15 min. Following phenol/chloroform extraction, the RNAs were ethanol precipitated at – 70°C with 5 μg tRNA as carrier. The precipitates were washed in ethanol, dried and resuspended in 4 μl 50% formamide, 10mm-EDTA, 0 · 5 % xylene cyanol and bromophenol blue. After heating at 65°C for 5 min, 2 μl of each sample was loaded onto a 6 % acrylamide/urea sequencing gel and run at 30 V cm-1. The gel was fixed in 10% methanol/10% acetic acid, and dried onto 3MM paper before autoradiography.
Hybridization to mouse genomic DNA
5 μg of mouse DNA was digested with EcoRI, Wind HI or BamHI and electrophoresed on a 0 · 6 % agarose gel. DNA was transferred to Gene Screen plus membranes by the method of Southern (1975). Nick-translated probes derived from EcoRI/BgII fragment of the insert were hybridized under conditions of high stringency (1 m-NaCl, 1 % SDS, 10% poly(ethylene glycol) at 65 °C and washed under similar stringent conditions of 2 × SSC (0 · 3 m-sodium chloride/0 · 03 m-sodium citrate), 1 % SDS at 65°C.
Northern blot analysis
50 μg of total RNA from adult kidney, liver and day- embryos were electrophoresed at 250 V for 2 h in l × MOPS buffer. The gels were blotted onto Hybond N membranes according to the Amersham protocol. Blots were prehybridized in 5 × SSPE, 50 % formamide, 5 × Denhardt’s solution and 0 · 5% SDS containing denatured salmon sperm DNA at 500 μg ml-1 at 42°C for 2h. 32P-labelled anti-sense riboprobe was added and hybridized overnight at 42°C. Blots were washed in O l × SSPE, 0 · 1 % SDS at 65°C for 30 min before incubation with RNase A at 10 μg ml-1 in 2 × SSPE at 37°C for 10min. The RNase A treatment was required to remove all nonspecifically bound probe which was monitored by the control sample of liver RNA which had been shown by RNase protection assays (Fig. 4) not to contain transcripts of Hox-6.1.
A combination of pBr322 digests and rehybridization of blots with rRNA probes was used as molecular weight markers.
Isolation of Hox 6.1 genomic clone
A mouse genomic library in Charon 35 was screened under conditions of high stringency with a probe derived from a nonhomeobox-containing fragment of a Hox-6.1 cDNA clone. Five positive plaques were identified and one of these was mapped by restriction digests of the phage, Southern blotting and probing with the Hox-6.1 fragment. The blots were washed and reprobed with the Hox-1.5- derived fragment to identify other homeobox sequences.
In situ chromosome hybridization
Mitotic preparations were obtained from third subculture mouse embryonic fibroblasts in a growth phase and the chromosomes banded for recognition after in situ hybridization by substitution with BrdU using an adaptation of a method described for human cells (Zabel et al. 1983). Slides were probed with the entire 540 bp insert labelled with tritium using the method described by Lyon et al. (1986), dipped in Ilford Nuclear Emulsion L4 and exposed for between 15 and 28 days before developing in D19 (Kodak). After developing, replication bands were revealed by treating the slides with Hoechst 33258 at a concentration of 10 μg ml-1 in 2 × SSC for 30 min, placing in 2 × SSC for 1 h at a distance of 15 cm away from a long-wave u.v. light and finally staining in 5 – 10 % Giemsa (Merck) in pH 6 · 8 buffer.
The slides were scored under a × 100 oil immersion lens by counting the total number of grains overlying all the chromosomes in a metaphase spread and relating grain positions with respect to the chromosomes and G-bands as identified from the standard nomenclature for the mouse karyotype (Nesbitt & Francke, 1973).
In situ hybridization to embryo sections
35S-labelled probes for use in in situ hybridization experiments were prepared from the 490 bp EcoRI/ Bg/II fragment of Hox-6.1 subcloned in pGEM4. The probe (antisense probe) used for hybridization with Hox-6.1 mRNA was synthesized in an SP6 RNA polymerase reaction after linearization of the plasmid with HindIII. In some experiments, a shorter anti-sense probe of about 200 bp not including the homeobox region was prepared after digestion of plasmid with BstNI. The control probe (sense probe), of opposite sense, was synthesized in a T7 RNA polymerase reaction after linearization of plasmid with EcoRI. Methods used for the production and alkaline hydrolysis of radioactive probes, for the preparation of embryo sections, and for in situ hybridization were all as previously described (Gaunt et al. 1986; Gaunt, 1987).
Interpretation of embryo sections in autoradiograms was made with reference to Holland & Hogan (1988). These authors distinguished between ‘somites’ of the early embryo (such as seen in our -day embryo sections) and the ‘prevertebrae’ of later stages (such as seen in our -day sections). This distinction is important since prevertebra 1 does not develop from somite 1, but instead may develop from somites 5 and 6 (Holland & Hogan, 1988).
Results
Molecular cloning and sequence analysis of Hox 6.1 A cDNA library from adult mouse kidney poly(A)+ RNA was constructed in λ gt10 and screened with a probe derived from the Hox-1.5 (Mo-10) gene containing 145 bp of homeobox sequence (McGinnis et al. 1984). One of the clones isolated was studied in detail and found to contain an insert of 540 bp with a Bg/II site, in the homeobox (Fig. 1). The nucleotide sequence of the insert from the 5’ end to the first inframe termination codon was determined and is shown in Fig. 2.
Nucleotide sequence homology of the Hox-6.1 homeobox with other mouse homeoboxes is greatest at 86% with Hox-1.2 (Colberg-Poley et al. 1985). However, homology at the amino acid level is greatest with a Xenopus homeobox Xebl (previously called AC1) and a human homeobox c8, being complete at 100% within the 60 amino acids of the homeobox (Carrasco et al. 1984; Muller et al. 1984; Boncinelli et al. 1985). The amino acid sequences of Hox-6.1 and c8 are identical over the regions compared, namely from 10 amino acids upsteam of the homeobox to the termination codon (Simeone et al. 1987). Homology with Xebl extends both immediately upstream and downstream of the homeobox (Fig. 2). Published sequence of Xebl is available for seven amino acids upstream and six amino acids downstream of the homeobox. Six of the seven upstream amino acids are conserved with the nonconserved being alanine in Hox-6.1 (GCG) at position – 2 which is serine (TCG) in Xebl. Five of the six downstream amino acids are also conserved, the exception being threonine (ACG) at position +63 in Hox-6.1 and serine (TCG) in Xebl which is a conservative change. No significant homology in these regions is evident with other homeobox genes. Further sequence of Xebl has not yet been published but from a personal communication (Eddy De Robertis), it is clear that the downstream amino acid sequences of Xebl and Hox-6.1 are highly conserved but not identical.
Hox-6.1 has been sequenced as far as the first inframe termination codon which is found 35 amino acids downstream of the homeobox and is followed by at least 220 bp of 3’-untranslated sequence. Long 3’-untranslated sequences are a common feature of many Drosophila homeobox genes and have also been found in at least one other mouse gene, Hox-2.1 (Krumlauf et al. 1987).
Several other clones isolated from the kidney cDNA library were found to be derived from transcripts of Hox-6.1 but none contained any more coding sequence than that shown in Fig. 2. Surprisingly, the majority of Hox-6.1 cDNA clones examined from the kidney library were found to be derived from unspliced RNA transcripts of Hox-6.1 having an intact 3’ splice acceptor consensus sequence at position – 22 (Fig. 2) (unpublished) which is the same position as the splice site in c8 (Simeone et al. 1987). The clone sequenced in Fig. 2 is derived from a spliced transcript of Hox-6.1 with the 5’ limit of the clone ending just 5’ of the position of the 3’ splice site. The possible significance of these unspliced RNA transcripts will be discussed in a later publication.
When a Southern blot of mouse DNA was probed with the insert under conditions of high stringency, a single band was visible with each enzyme used, showing there to be only one copy of the gene (Fig-3).
Expression of Hox-6.1
(a) RNase protection assays
RNase protection assays were used to detect transcripts of the Hox-6.1 gene during embryonic development and in adult tissues. The 490 bp EcoRl/Bg/II fragment from the insert was subcloned into pGEM4 and 32P-labelled anti-sense riboprobes used for the protection assays. Protected fragments were detected in RNA extracted from embryos on days and of gestation and in RNA from day- extraembryonic tissue, adult kidney and testis (Fig. 4). Very faint protected fragments were detected in day- embryo and extraembryonic RNA after prolonged exposure of the gels but these were not visible when photographed and were therefore not included in Fig. 4. No protected fragments were detected with RNA from adult liver, brain or spleen (Fig. 4). Control experiments were also conducted with the sense riboprobes and no protected fragments were observed (not shown).
Although protected fragments were occasionally visible at 490 bp, corresponding to the size of the intact riboprobe, these were faint even after prolonged exposure of the gel. The major protected fragment was consistently 265 bp (Fig. 4).
(b) Northern blot analysis
Northern blots of total RNA from adult kidney, day- embryos and, as a control, adult liver were probed with the same 32P-labelled anti-sense riboprobe used for the RNase protection assays. After stringent washings and RNase treatment (see Methods), bands were visible in the tracks from kidney and embryo RNA but not in the liver track. One major transcript was visible from kidney and embryo RNA but these were of different sizes. In kidney, the single major transcript detectable was 2 · 7 kb. With day- embryo RNA, the major transcript was 1 · 9kb with a faint band also visible at 2 · 7 kb (Fig. 5).
(c) In situ hybridization
The localization of Hox-6.1 transcripts in developing embryos was examined by in situ hybridization. At days (Fig. 6), intense labelling was seen in the spinal cord, but no labelling above background was seen in any part of the brain. The boundary between anterior (unlabelled) and posterior (labelled) parts of the nervous system was sharply defined and was located within the spinal cord at a position just posterior to the hindbrain (Fig. 6A). No reduction in intensity of labelling was detected between anterior parts of the spinal cord (Fig. 6A) and more posterior parts (Fig. 6D). Spatial restriction of Hox-6.1 expression was also evident within the column of prevertebrae (Fig. 6D), but here the boundary between labelled and unlabelled regions was more posteriorly located than in the nervous system. Prevertebrae 1–6 were not labelled above background. Prevertebra 7 was weakly labelled, and labelling was progressively stronger in prevertebrae 8 and 9. Prevertebrae 9–16 were intensely labelled. Labelling was progressively weaker over prevertebrae 17–20 but persisted at a low, apparently uniform level over all more posterior prevertebrae. In addition, labelling was also seen in several other mesodermal derivatives in the posterior part of the body. Most obvious was labelling in the metanephric kidney, the gonad and the mesodermal components (but not the endodermally derived lining epithelia) of the lung (Fig. 6D), the stomach and some regions of the intestine. No specific labelling was detected over the heart or liver (not shown). In situ hybridization experiments using the sense (control) probe produced only background labelling of tissues with no evidence of boundaries in either the nervous system (Fig. 6B) or the prevertebral column (not shown). In situ hybridization using an anti-sense probe that did not include the homeobox (see Materials and methods) produced an identical pattern of labelling to that described for the full-length probe (not shown). This finding, together with the results of the Northern blot analysis, suggests strongly that the Hox-6.1 probe used for in situ hybridization did not cross hybridize with the transcripts of other homeobox genes.
At days gestation, the embryo was at the 6-somite stage (see Materials and methods for distinction between somites and prevertebrae). In situ hybridization at this stage (Fig. 7) showed that Hox-6.1 transcripts were restricted to the ectoderm and mesoderm layers in the posterior part of the embryo. Labelling did not include any of the somites already formed but was seen in the presomitic mesoderm posterior to somite 6. Labelling in the mesoderm layer extended posteriorly into the allantois, but other extraembryonic tissues including decidual tissue and extraembryonic membranes were not labelled above background (Fig. 7). Experiments using sense (control) probe on the -day embryo showed no specific labelling of tissues (not shown).
Chromosomal location of Hox-6.1
The chromosomal location of Hox-6.1 was determined by in situ hybridization with 3H-labelled probe from the 540 bp insert. A total of 133 grains were scored in a sample of 100 mitotic cells. If the grains were randomly distributed and not a positive indication of a successful hybridization, they would be expected to be distributed at random according to unit length of chromosome. Therefore, estimates were made for the number of grains expected to overlie each homologous pair by reference to the total number of grains scored in relation to the chromosome lengths given by Nesbitt & Franke (1973). Apart from chromosome 14, expected numbers were observed to overlie the other chromosomes. Chromosome 14 represents 4·31 % of the haploid genome and would be expected to carry approximately 6 grains (4·31% of 133 = 5·7); however, 23 grains were observed, a fourfold increase of observed over expected. Further, of the 23 grains, 20 overlay the distal segment of the chromosome with a peak distribution over the middle of band 14E2 (Fig. 8). Clearly, this hybridization indicates that Hox-6.1 maps to this region of chromosome 14.
Isolation of a genomic clone of Hox 6.1
In order to begin to map the Hox 6 locus and determine whether other homeobox-containing genes might be located close to Hox-6.1, a genomic clone containing Hox-6.1 was isolated from a library in Charon 35. The probe used was derived from a region of a cDNA clone of Hox-6.1 that contained no homeobox sequence and was shown by low-stringency probing of Southern blots not to cross hybridize with any other mouse sequences. The genomic library was probed under high-stringency conditions and five positive plaques were obtained. One of these five phage has been mapped, with the position of Hox-6.1 being detected by hybridization of digested fragments to the probe used to screen the library. The same blots were reprobed with the probe derived from Hox-1.5 which identified a possible second homeobox sequence within the clone approximately 7 kb downstream of Hox-6.1 and this putative homeobox-con-taining gene has been named Hox-6.2 (Fig. 9).
Discussion
Homeobox-containing genes have been isolated from a variety of vertebrate species including mammals. The conservation of the homeobox region of these genes and between their Drosophila counterparts is striking, suggesting an important role for these genes in development. It also appears that regions outside the homeobox are conserved between different genes from different vertebrate species (Boncinelli et al. 1985). Furthermore, there are now at least three instances of possible homologous homeobox genes where large portions of the amino acid sequences (if not all), are highly conserved in genes isolated from widely different vertebrate species. The mouse Hox-2.1 gene (Jackson et al. 1985) has homologues in the human (Hui) (Hauser et al. 1985) and in Xenopus (XhoxlB) (Harvey et al. 1986). Similarly there is high amino acid homology between the mouse Hox-1.4 (Duboule et al. 1986) Xenopus XhoxlA (Harvey et al. 1985) and human c13 (Boncinelli et al. 1985) genes. Here, we report the isolation of a new mouse homeobox gene which appears to be homologous to the Xenopus Xebl and human c8 genes (Carrasco et al. 1984; Boncinelli et al. 1985). If these highly conserved genes have an important function in development it might be expected that homologous genes would serve similar functions in their respective species.
The in situ hybridization results now obtained for Hox-6.1 can be compared with results already published for several other mouse homeobox genes (Awgulewitsch et al. 1986; Gaunt et al. 1986; Gaunt, 1987; Utset et al. 1987; Krumlauf et al. 1987). In day mouse embryos, all of the homeobox genes studied show transcripts in the posterior, but not anterior, regions of the nervous system. Individual genes may differ, however, in the position of the boundary between labelled and unlabelled parts. Thus, the boundary for Hox-6.1, located just behind the hindbrain, is clearly posterior to the boundary for Hoxl.5 (Gaunt et al. 1986; Gaunt, 1987), but anterior to that of Hox-3.1 (Awgulewitsch et al. 1986; Utset et al. 1987). The Hox-6.1 transcript boundary within somitic mesoderm derivatives is located posterior to the boundary in nervous tissue and is at the position of the 7th prevertebra. In contrast, all prevertebrae are labelled with Hox 1.5 (Gaunt et al. 1986; Gaunt, 1987). Like Hox-6.1, Hox-3.1 (Utset et al. 1987) shows a transcription boundary within somitic mesoderm derivatives posterior to that in the nervous system. The location of Hox-6.1 transcripts now described within the developing lung seems identical to that previously noted for Hox-2.1 (Krumlauf et al. 1987).
The pattern of Hox-6.1 labelling observed in the day embryo shows similarities to that already described for Hox-1.5 (Gaunt et al. 1986; Gaunt, 1987). Thus, labelling is restricted to posterior ectoderm and mesoderm tissues, and extends posteriorly into the allantois. Hox-6.1 transcripts at this stage do not, however, extend as far forward in the embryo as Hox-1.5 transcripts, which are detected in anterior somites (Gaunt, 1987). The observations for Hox-6.1 are consistent with a view, as suggested for Hox 1.5 (Gaunt, 1987), that this gene serves as one of a series of homeobox genes whose expression provides positional cues during the determination of tissues along the body axis.
Since in situ hybridization data are already published for Xebl (Carrasco & Malacinski, 1987), the Xenopus homologue of mouse Hox-6.1, it is now possible for the first time to compare the spatial distribution of homologous homeobox genes during the development of two different vertebrate species. The transcription boundaries within the central nervous system are similar in both species, being located at the junction of the spinal cord and the brain. Also similar is the concentration of transcripts in the posteriormost parts of earlier-stage embryos (gastrula- and neurula-stage Xenopus embryos, and -day late-primitive-streak-stage mouse embryos). The striking pattern of Hox-6.1 transcription seen in the somitic mesoderm derivatives of the mouse does not seem to be repeated in Xenopus. The somitic mesoderm of Xenopus showed no labelling by the Xebl probe, although labelling was noted in the lateral mesoderm around the blastopore of gastrula stages (Carrasco & Malacinski, 1987). It is possible that homeobox gene transcription in the mesoderm of Xenopus is more transient than in mouse so that it does not persist at stages of somite formation.
The full coding sequence of Hox-6.1 has not yet been determined since, in the cDNA clones so far studied, there appears to be a preponderance of clones derived from unspliced transcripts of Hox-6.1 where the 3’ splice acceptor consensus is intact (unpublished data). In view of the complete amino acid homology between Hox-6.1 and c8 in the regions compared (Fig. 2) it seems probable that the upstream region of Hox-6.1 beyond that currently sequenced will be identical to c8 (Simeone et al. 1987). We are currently sequencing a genomic clone of Hox-6.1 to confirm this.
The position of the 3’ splice acceptor consensus sequence of Hox-6.1 is identical to that of the c8 gene (Simeone et al. 1987). The clones isolated that do contain spliced transcripts of Hox-6.1 terminate just upstream of the position of the 3’ splice site (Fig. 2). It seems probable from the Northern blot data that this preponderance of unspliced Hox-6.1 transcripts may be a feature of the adult kidney and may not occur to the same extent in embryos. The predominant message size in embryos is 1·9 kb which is consistent with the size of other homeobox gene transcripts (e.g. Gaunt et al. 1986; Krumlauf et al. 1987). The major transcript size in the kidney is significantly larger at 2·7 kb and this may reflect transcripts containing unspliced intron sequences. A faint band at 2·7 kb in the embryo RNA is visible which may suggest the occurrence of a small proportion of similar unspliced transcripts in embryos.
The RNase protection data show that Hox-6.1 is expressed in adult kidney and testis but not in liver, spleen or brain. The transcripts are also detected in day- embryonic and extraembryonic tissues and day- embryos. The size of the protected fragments seen is around 265 bp which is 225 bp shorter than the size of the intact probe used. We do not, as yet, have an explanation for this but a fragment of 265 bp from the 5’ end would end about 30 bp past the termination codon and it is possible that this may reflect processing of 3’ untranslated sequences.
Transcripts were detected in extraembryonic tissues (allantois and amnion) from day- embryos. Previously examined extraembryonic tissues, placenta (Jackson et al. 1985) and yolk sac plus placenta (Colberg-Poley et al. 1985) showed no evidence of homeobox gene transcription. These findings for Hox-6.1 are readily explained by the in situ hybridization data, which show transcripts in the allantois but not in other extraembryonic tissues of the -day embryo. Similar findings have been obtained for Hox-1.5 transcripts (Gaunt, 1987). We have examined the expression of Hox-6.1 in day- embryos and extraembryonic tissues and found barely detectable levels of expression using RNase protection (data not shown).
The Hox-6 locus has been located on chromosome 14 in the E2 region. This location does not appear to contain any known obvious developmental mutations. In common with the Hox-1 and Hox-2 loci (Hart et al. 1985; Duboule et al. 1986), Hox-6 appears to contain a cluster of homeobox-containing genes. A single genomic clone of 15 kb contains Hox-6.1 and another putative homeobox (Hox-6.2) detected by Southern blotting, within 7 kb. We are currently examining other genomic clones to determine the extent of this cluster.
Between amino acids 68 and 72 of Hox 6.1 are five consecutive glycine residues. Such a region may be devoid of secondary structure and could form a short flexible hinge-type region in the protein linking two spatially distinct domains. A similar putative hingetype region has been identified in the Ubx homeobox gene in Drosophila but no such regions have been identified downstream in the homeobox genes so far described (Beachy et al. 1985). We have recently isolated a pig homeobox gene, however, which has a glycine-rich region between amino acids 70 – 75 downstream of the homeobox.
Acknowledgements
We wish to thank Kenneth Krauter for the gift of the mouse genomic library, Eddy De Robertis for the unpublished cDNA sequence of Xebl, Andres Carrasco for detailed information on the localization of Xebl transcripts and Don Powell for invaluable help and assistance.
References
The locus has been designated according to the nomenclature system accepted by the International Committee for Standardized Genetic Nomenclature and agreed with T. Roderick, Jackson Laboratory.