The H ox 2.1 gene forms part of a cluster of homeo-box -containing genes on mouse chromosome 11. Analysis of Hox2.1 cDNAs isolated from an -day p.c. mouse embryo library predicts that the gene encodes a 269 amino acid protein (Mr, 29432). This deduced protein contains a homeobox 15 amino acids from the carboxy terminus and is very rich in serine and proline. A second partially conserved region present in several other genes containing homeo-boxes, the hexapeptide De-Phe-Pro-Trp-Met-Arg, is located 12 amino acids upstream of the homeodomain and is encoded by a separate exon. Analysis of Hox 2.1 gene expression reveals a complex and tissue-specific series of RNA transcripts in a broad range of feta] tissues (lung, spinal cord, kidney, gut, spleen, liver and visceral yolk sac). Comparison of the temporal patterns of gene expression during development and in the adult suggests that Hox2.1 is regulated independently in different tissues. Evidence is also presented that transcripts from other loci have extensive homology to the Hox2.1 gene in sequences out-side of the homeobox. In situ hybridization shows that Hox2.1 transcripts are regionally localized in the spinal cord in an apparent anterior–posterior gradient extending from the hind brain. The distribution of RNA also displays a cell-type specificity in the. lung, where mesodermal cells surrounding the branching epithelial cell layer accumulate high levels of Hox2.1 transcripts.

Current molecular approaches towards understanding of the control of mammalian development are attempting to expand upon classic embryological or genetic studies. However, the cloning and characterization of genes affecting development has proved difficult because of the lack of closely linked markers and the large chromosomal regions involved. In the mouse, insertional mutagenesis via retroviral infection of embryos (Jaenisch, Harbers, Schnieke, Lohler, Chumakov, Jahner, Grotkopp & Hoffman, 1983) or microinjection of fertilized eggs (Woychick, Sterwart, Davis, D’Eustachio & Leder, 1985) represent one way of homing in on genes that influence developmental processes. An alternative approach is based on the recent discovery that highly conserved sequences in genes implicated in the control of pattern formation during the embryonic development of Drosophila (McGinnis, Levine, Hafen, Kuroiwa & Gehring, 1984a; Scott & Weiner, 1984) have a broad phylogenetic distribution (McGinnis, Garber, Wirz, Kuroiwa & Gehring, 1984b; McGinnis, 1985; Holland & Hogan, 1986). Genetic analysis in Drosophila has characterized two complexes, Antennapedia (ANT-C) (Kaufman, Lewis & Wakimoto, 1980) and Bithorax (BX-C) (Lewis, 1978), which contain genes involved in control of the segmental body plan. Cloning of these complexes revealed eight different genes which contain a common DNA element, termed the ‘homeobox’ (Garber, Kuroiwa & Gehring, 1984; McGinnis et al. 1984b; Scott & Weiner, 1984; Regulski, Harding, Kostriken, Karch, Levine & McGinnis, 1985; Laughon et al. 1985). These sequences share greater than 80 % homology and are classified as Antennepedi-like or Class I homeoboxes. The homeobox itself is a 183bp sequence capable of coding for a highly conserved protein domain (for review see Gehring, 1985) which has sequence similarity to the DNA-binding domains of yeast and bacterial regulatory proteins (Shepherd, McGinnis, Carrasco, De Robertis & Gehring, 1984; Laughon & Scott, 1984; Johnson & Herskowitz, 1985; Whiteway & Szostak, 1985). The Drosophila proteins containing homeoboxes have been localized in the nucleus (White & Wilcox, 1984; Beachy, Halfand & Hogness, 1985; Carroll & Scott, 1985; Carroll, Laymen, McCutcheon, Riley & Scott, 1986) and one gene product, engrailed, exhibits a sequence-specific DNA-binding activity (Desplan, Theis & O’Farrell, 1985). The Drosophila homeobox probes have provided a means of directly isolating a family of genes that might regulate early stages of mammalian embryogenesis. It is therefore important to determine the properties of the homeobox-containing genes in mammals in order to establish whether these genes are involved in developmental control processes.

Genes containing homeoboxes have been isolated from sea urchins (Dolecki, Wannakrairoj, Lum, Wang, Riley, Carlos, Wang & Humphreys, 1986), frogs (Carrasco, McGinnis, Gehring & De Robertis, 1984; Muller, Carrasco & De Robertis, 1984; Harvey, Tabin & Melton, 1986), mice (Colberg-Poley, Voss, Chowdhury & Gruss, 1985a; McGinnis, Hart, Gehring & Ruddle, 1984c; Jackson, Schofield & Hogan, 1985; Joyner, Kornberg, Coleman, Cox & Martin, 1985; Hart, Awgulewitsch, Fainsod, McGinnis & Ruddle, 1985; Rabin, Hart, Ferguson-Smith, McGinnis, Levine & Ruddle, 1985; Hauser, Joyner, Klein, Learned, Martin & Tjian, 1985; Breier, Bucan, Francke, Colberg-Poley & Gruss, 1986; Awgule-witsch, Utset, Hart, McGinnis & Ruddle, 1986; Dubuole, Baron, Mahl & Galliot, 1986) and humans (Levine, Rubin & Tjian, 1984; Hauser et al. 1985; Boncinelli, Simeone, La Volpe, Faiella, Fidanza, Acampora & Scotto, 1985), with the genes often organized in clusters. Initial experiments on the patterns of gene expression supported a potential role in development. Mammalian homeobox genes are expressed in teratocarcinoma cells (Colberg-Poley et al. 1985a; Hauser et al. 1985), early embryos (Jackson et al. 1985; Colberg-Poley et al. 1985a,b; Hart et al. 1985; Hauser et al. 1985) and some adult stages (Jackson et al. 1985; Colberg-Poley et al. 1985b; Wolgemuth, Engelmyer, Duggal, Gizang-Ginsberg, Mutter, Ponzetto, Viviano & Zakeri, 1986), in a tissue-specific (Jackson et al. 1985; Simeone, Mavilio, Bottero, Giampaolo, Russo, Faiella, Boncinelli & Peschle, 1986) and spatially restricted (Awgulewitsch et al. 1986) manner. However, as yet there is no direct evidence that establishes the function or regulatory role of these genes in mammals. As a step towards this goal we have isolated genomic and cDNA clones for a mouse homeobox gene, Hox2.1, to analyse potential gene products and their patterns of expression. In this study we report the complete protein sequence predicted from the cDNA clones for the Hox2.1 gene and describe several interesting features of its structure. A second region of homology upstream of the homeobox was found in homeobox-containing genes of several species. We also describe a detailed analysis of the tissue-specific and temporal expression of this gene via Northern analysis, which reveals that it is differentially regulated during development and in the adult. The presence of multiple RNA species suggests that there is a complex transcription pattern for the gene which could involve differential processing and multiple gene products. In situ hybridization experiments examine the spatial distribution and cell-type specificity of Hox2.1 in -day mouse embryos.

Clone isolation and sequencing

A cDNA library, prepared from poly(A)+RNA of C57BL -day p.c. embryos, by inserting double-stranded cDNA’s into the λtlO vector using oligo adapters (Farhner, Hogan & Flavell, 1987), was screened using a labelled 600bp EcoRI-PvuII fragment (H 24.1 probe A, Jackson et al. 1985) which contains the Hox2.1 homeobox and 3’ region of the gene. Nylon filters (Pall Biodyne) of phage were prehybridized and hybridized in 50% formamide, 5 X SSC, 0·1% bovine serum albumin, 0·1% Ficoll 400, 0·1% polyvinylpyrrolidone, 50mM-sodium phosphate (pH6-8), 0·1% SDS, and 200μgml-1 denatured sonicated salmon sperm DNA at 42°C. Filters were washed in 0·1 x SSC, 0·1% SDS at 65°C. Positively hybridizing clones were plaque-purified and the inserts subcloned as EcoRI fragments into pGEMl or pGEM2 (Promega Biotec). Probes from these clones were then used to rescreen the library for additional cDNAs to maximize the area cloned. Both strands of the recombinant clones were then sequenced in the GEMINI vectors by the procedures of Maxam & Gilbert (1980) or using primers for the T7 and SP6 promoters (Promega Biotec), coupled with chain termination sequencing methods (Sanger in full, 1977).

RNA isolation

Tissues were dissected from CBA or C57BI/10 embryos in Dulbecco’s modified Eagle’s medium with 10 % fetal bovine serum (Hogan, Costantini & Lacy, 1986) or taken directly from adult animals and rapidly frozen in liquid nitrogen. The RNA was isolated from the frozen tissue by a modification of the method of Auffray & Rougenon (1979). Briefly, the tissue was homogenized in 5-10 ml of 3 m-LiCl, 6m-urea per gram of tissue for 2 min on ice, using an ultraturrax or similar tissue disrupter. The homogenate was stored overnight at 4°C, then centrifuged at × 5000g for 10 min at 0°C and the supernatant poured off. A half volume of cold 3M-LiCl, 6m-urea was added, the sample vortexed and recentrifuged, discarding the supernatant.

The pellet was redissolved in 10mm-Tris-HCl (pH 7·6), 1 mM-EDTA, 0·5% SDS using 5 ml per gram of original tissue. The sample was extracted with an equal volume of phenol:chloroform:isoamyl alcohol (24:24:1) and the aqueous phase separated by centrifugation at 5000g for 5 min. The RNA was harvested by ethanol precipitation and poly(A)+RNA isolated by oligo dT cellulose chromatography (Aviv & Leder, 1972). The A260 and a 3H-poly-U binding assay (Bishop, Rosbash & Evans, 1974; Rosbash & Ford, 1974) were used to quantify the amount of poly(A)+ mRNA.

Northern blots

Poly(A)+RNA, quantified by the poly-U assay, were electrophoresed for 6h at 60mV in a 6% formaldehyde-1 % agarose gel in 1 x MOPS buffer (pH 7·0; 20 mM-morpboline propanesulphonic acid, 5mM-sodium acetate, 1 mM EDTA) after denaturation at 60°C for 10min in 70% formamide-6% formaldehyde-1 x MOPS. Following electrophoresis, a marker poly(A) lane was cut from the gel and stained to provide size markers (28S−4·7kb, 18S−2·2kb), the re-mainder of the gel was soaked successively in 50mM-NaOH, 100 mM-NaCl (20min); 100 mM-Tris pH 7·6 (20min) and 2 x SSC (20min), then transferred to a Genescreen (Dupont) membrane in 20 x SSC overnight. The filter was rinsed in 6 x SSC, exposed to 600μwattscm-2 (254 nm u.v. light source) for 5 min to crosslink the RNA to the filter, and then baked for 2 h at 80°C. Synthesis of anti-sense SP6 and T7 RNA probes using 32P-UTP (600 Ci mmole-1; NEN-Dupont) was performed according to polymerase suppliers (Promega Biotec) and as previously described (Melton, Krieg, Rebagliati, Maniatis, Zinn & Green, 1984; Krumlauf, Hammer, Tilghman & Brinster, 1985). The Hox2.1 probe was constructed by cloning the BamIII-HindIII fragment of cDNA 2.1A (Fig. A probe 2) into pGEM2 and linearizing the plasmid with EcoRI for T7 transcription. Filters were hybridized with 107ctsmin-1 ml-1 in 60% formamide, 5 x SSC, 0·1% bonne serum albumin, 0·1% Ficoll 400, 0·1% polyvinylpyrolodone, 20mM-sodium phosphate pH6-8, 1% SDS, 7% dextran sulphate, 100 μg ml-1 denatured sonicated salmon sperm DNA, 100μgml-1 tRNA (bakers yeast), Wpgml-1 poly A at 65°C for 16-24 h. Filters were washed in 2 x SSC, 1% SDS at 50°C for 3x 15min, then in 0·2 x SSC, 1 % SDS at 80 °C for 1-2 h, and exposed to Kodak XAR-5 film with intensifying screens (lightning plus, Dupont) at -70°C for various times. In some cases the filters were given a further wash with RNase A to remove unmatched hybrids. The membrane was soaked in 2 x SSC plus 20 μg ml 1 RNase A (Sigma) at room temperature for 30min, then washed in 2 x SSC, 0·5 % SDS at 50°C for 2x30 min. Filters that had been stripped of Hox 2.1 probe by washing at 70°C with 75% formamide, 0·1% SDS, or duplicate filters were hybridized with an SP6 anti-sense mouse (Lactin RNA probe as a control to test the loading and quality of RNA on the filters. Relative levels of RNA were determined by scanning various exposures of the autoradiographs with a densitometer, using the RNA from -day embryos as the standard.

In situ hybridization

Embryos from CBA mice, days p.c., were fixed with paraformaldehyde, embedded, sectioned using a cryostat and hybridized as previously described (Hogan etal. 1986). The Hox2.1 anti-sense (-) and sense (+) strand probes were synthesized from the EcoRI fragment of cDNA 2.1 A (Fig. 1, probe 1), after subcloning into the vector pGEMl in both orientations. The vectors were linearized with SalI and labelled with 35S-UTP (NEN-Dupont) to a specific activity of 2x109 dissints min−1μg-1 using T7 RNA polymerase, as described by Melton et al. (1984). The probes were hydrolysed to an average size of 100 nt in a controlled reaction and used at a final concentration of 0·035-0·075 ngμl-1. Post-hybridization RNase A washes and autoradiography are according to Hogan et al (1986). Exposure times were 6 to 9 days.

Fig. 1.

Structure of Hox2.1 cDNA clones and location of predicted protein coding region. The thin lines labelled c2.1A, c2.1B and c2.1C represent the overlapping cDNA clones isolated from the 812-day p.c. cDNA library and the thick lines denote subcloned regions of the c2.1A cDNA used for in situ hybridization, probe 1 (EcoRI fragment), or Northern analysis, probe 2 (BomHI-Hindlll fragment). The boxed region on the top line shows the open reading frame in the cDNA which corresponds to the 269 amino acid protein predicted from the sequence in Fig. 2. The large shaded area is the homeodomain and the small shaded area is the conserved hexapeptide region (see text).

Fig. 1.

Structure of Hox2.1 cDNA clones and location of predicted protein coding region. The thin lines labelled c2.1A, c2.1B and c2.1C represent the overlapping cDNA clones isolated from the 812-day p.c. cDNA library and the thick lines denote subcloned regions of the c2.1A cDNA used for in situ hybridization, probe 1 (EcoRI fragment), or Northern analysis, probe 2 (BomHI-Hindlll fragment). The boxed region on the top line shows the open reading frame in the cDNA which corresponds to the 269 amino acid protein predicted from the sequence in Fig. 2. The large shaded area is the homeodomain and the small shaded area is the conserved hexapeptide region (see text).

Isolation of cDNA clones

A mouse homeobox genomic clone isolated by Jackson et al. (1985), originally termed H24.1, is expressed in mouse embryos and is located on chromosome 11 band 11D (Munke, Cox, Jackson, Hogan & Francke, 1986). Based on restriction site mapping and sequence comparisons it is identical to the Hox2.1 (Hart ei al. 1985) or Mu-] gene (Hauser et al. 1985), which forms pan of a complex of at least four and possibly six or more mouse homeobox-containing genes on chromosome 11 (Hart et al. 1985; Krumlauf, unpublished data). To obtain cDNA clones of Hox2.1 for expression and protein analysis, we screened a cDNA library prepared from -day p.c. mouse embryo RNA in the vector λgt10 (Farhner et al. 1987) with a subcloned Hox2.1 genomic probe (H24.1 probe A, Jackson et al. 1985) at high stringency. Sequences from the first positive clones were used to rescreen the library for additional clones. The structure and sequence of three overlapping cDNA clones (2.1A, B and C) spanning the largest area, is shown in Figs 1A, 2. We have confirmed that these cDNA clones are derived from the Hox2.1 gene, by matching them exactly to sequences within genomic clones isolated from this gene (data not shown).

Northern analysis of RNA from -day mouse embryos showed a major Hox2.1 transcript of 2·2 kb (see below). The three cDNAs (Fig. 2) span 1 ·9 kb of sequence which suggests that approximately 300 bp of lhe mRNA are missing from our cDNA sequence. The 3’ end of the mRNA is contained in clone 2.1C, including the polyadenylation signal (AATAAA) and the first few bases of the poly(A) tail. Therefore, the 300 bp, which are predicted to be missing, probably represent some sequences from the 5’ end of the RNA and the length of the poly(A) tail. Some of lhe cDNA clones terminated in an A-rich region between 1236-1251 (Fig. 2). However, there is no adjacent polyadenylation signal and these clones probably represent internal priming from the run of A’s rather than an alternative site of poly(A) addition.

Fig. 2.

The nucleotide sequence of Hox2.1 cDNAs and the predicted amino acid sequence of the protein. The nucleotide sequence is a composite produced from the sequence of the three cDNA clones (c2.1A, B and C) in Fig. 1. The numbering starts from the most 5’ base in clone c2.1A. The polyadenylation signal (AATAAA) is underlined. The boxed regions indicated the location of the homeodomain (large box, 682-864 nt) and the conserved peptide (small box, 628-645 nt). The solid triangle denotes the location of an intron and splice site, deduced by sequence comparison of cDNA and genomic clones.

Fig. 2.

The nucleotide sequence of Hox2.1 cDNAs and the predicted amino acid sequence of the protein. The nucleotide sequence is a composite produced from the sequence of the three cDNA clones (c2.1A, B and C) in Fig. 1. The numbering starts from the most 5’ base in clone c2.1A. The polyadenylation signal (AATAAA) is underlined. The boxed regions indicated the location of the homeodomain (large box, 682-864 nt) and the conserved peptide (small box, 628-645 nt). The solid triangle denotes the location of an intron and splice site, deduced by sequence comparison of cDNA and genomic clones.

The predicted Hox 2.1 protein

The protein predicted by the cDNA sequence is shown in Fig. 2. The longest open reading frame in the sequence is in-frame with the homeodomain, and we have denoted the first AUG, 72bp after an inframe stop codon, the methionine-initiation codon. The predicted protein contains 269 amino acids, and has a calculated Mr = 29432. The 61 amino acids of the homeobox and the following 15 amino acids up to the carboxy terminus are identical to those reported for the mouse (Hauser et al. 1985; Jackson et al. 1985) and human (Boncinelli et al. 1985; Hauser et al. 1985; Simeone, Mavilio, Bottero, Grampaolo, Russo, Faiella, Boncinelli & Peschle, 1986) Hox2.I gene. The protein-coding information is located in lhe 5’ portion of the sequence and there is a long 995bp 3’ untranslated region. The human Hox2.1 cDNA (clO; Boncinelli et al. 1985) also has a long 3’ untranslated region (962 bp) and comparison with the mouse sequence shows that these regions are also remarkably conserved (84 %).

Analysis of the predicted Hox2.1 amino acid sequence using the secondary structure program of Chou & Fasman (1981) shows that the homeodomain contains alternating areas of α-helix, similar to the helix-turn-helix motif common to homeobox genes in the Drosophila Antennapedia class (Laughon & Scott, 1984). These regions are believed to bind and interact with DNA based on their homology to yeast mating-type genes, such as mata2, which have been shown to encode DNA-binding proteins (Johnson & Herskowitz, 1985). The protein sequences outside of the homeobox are unusual in that they are very rich in serine (22%) and proline (9%). Prediction of secondary structure for this region of the protein shows that no α-helices can be formed as a consequence of the widely distributed helix-incompatible residues, such as proline and glycine. It is interesting to note that the Xenopus Xhoxl-A (Harvey, Tabin & Melton, 1986), Drosophila fushi tarazu (Laughon & Scott, 1984) and Antennapedia (Schneuwly, Kuroiwa, Baumgartner & Gehring, 1986) proteins also contain approximately 10% prolines outside of the homeo-box. This may therefore represent a general feature of homeobox proteins.

A comparison of the Hox2.1 protein with other homeobox-containing proteins reveals that there is a second region of homology located at a short but variable distance upstream of the homeobox. The hexapeptide (lle-Phe-Pro-Trp-Met-Arg) starting 18 amino acids upstream of the Hox2.1 homeobox (Fig. 2) is partially conserved in several homeobox proteins from other species (Fig. 3). The mouse Hoxl.l (P. Gruss, personal communication) and Hoxl.3 (W. Odenwald, personal communication) proteins have peptides which differ from the 2.1 sequence by only one amino acid, and the Xenopus Xhoxl-A (Harvey et al. 1985), Drosophila Antp (Schneuwly et al. 1986) and Dfd (Laughon et al. 1985) proteins have peptides which differ at additional positions. The caudal (cad) (Mlodzik, Fjose & Gehring, 1985) protein appears to be the most diverged member of the family, whilst the fushi tarazu (Laughon & Scott, 1984) and engrailed (Poole, Kauvar, Drees & Kornberg, 1985) proteins do not have a recognizably homologous peptide. These differences are summarized in Fig. 3 and a consensus sequence for this peptide region presented. In all of the examples shown in Fig. 3, the conserved peptide is located between 5 and 16 amino acids upstream of the homeobox domain and is encoded by a different exon to the homeobox domain. The degree of homology and location of this peptide suggest that it represents a second conserved region associated with the homeodomain and it will be important to establish w hether this region functionally interacts with the homeobox domain.

Fig. 3.

Conserved hexapeptide region in homeobox genes of different species. The Hox2.1, Hox 1.1 and Hox 1.3 genes are from mouse, Xhoxl-A from Xenopus, and Amp, Dfd and Cad from Drosophila. A consensus sequence is indicated below the boxed conserved amino acids. The 5-18 amino acids above the dotted arrow indicate the distance upstream from the homeobox and the arrow below indicates that the region is located on a different exon than the homeobox. References for sequences: (Hoxl.l (Gruss, persona) communication), Hoxl.3 (Odenwald, personal communication), Xhoxl-A (Harvey et al. 1986), Amp (Schenuwly ei al. 1986), Dfd (Laughon et al. 1985), Cad (Mlodzik etal. 1985).

Fig. 3.

Conserved hexapeptide region in homeobox genes of different species. The Hox2.1, Hox 1.1 and Hox 1.3 genes are from mouse, Xhoxl-A from Xenopus, and Amp, Dfd and Cad from Drosophila. A consensus sequence is indicated below the boxed conserved amino acids. The 5-18 amino acids above the dotted arrow indicate the distance upstream from the homeobox and the arrow below indicates that the region is located on a different exon than the homeobox. References for sequences: (Hoxl.l (Gruss, persona) communication), Hoxl.3 (Odenwald, personal communication), Xhoxl-A (Harvey et al. 1986), Amp (Schenuwly ei al. 1986), Dfd (Laughon et al. 1985), Cad (Mlodzik etal. 1985).

Developmental expression

Jackson er of. (1985) have shown, using RNase protection experiments, that the Hox2.1 gene is active in mouse embryos as early as 7) daysp.c. and transcripts were enriched in fetal spinal cord and adult kidney. However, these experiments gave no indication of either transcript complexity or temporal patterns of gene expression. Therefore we have examined the pattern and timing of Hox2.1 expression by Northern blot analysis of RNA extracted from a series of embryonic and adult tissues. For this analysis we subcloned a fragment (BamHI-Wiridlll; Fig. 1, probe B) of the 2.1 A cDNA clone, which does not contain homeobox sequences, into the vector pGEMl to provide single-stranded anti-sense RNA probes. Using this probe, a complex pattern of transcripts is observed in both mouse embryo and adult kidney poly(A)+mRNA, as shown in Fig. 4A. A major 2·2 kb transcript and a series of higher molecular weight minor RNA species, including bands at 3·8, 6, 10 and 12 kb are found in - and -day embryonic RNA. Most of these RNA species and an additional 1·9 kb transcript are also detected in kidney RNA, but the relative ratio of each species is very different from that of the embryo. In particular, the 3·8 kb RNA is as abundant as the 2·2 kb RNA in kidney. A similar complex pattern is shown in Fig. 4B, where we have extended this analysis to include a wider range of embryonic and adult tissues. Transcripts were detected in a large majority of the tissues and Fig. 4B illustrates the surprisingly wide range in number, size and intensity of the transcripts observed. It is important to note that these RNA samples were analysed under conditions of high stringency in both the hybridization and washing steps (see Methods). These conditions are usually sufficient to detect only highly homologous or identical sequences and in control experiments using a mouse β-actin probe we only detected a single RNA band. These bands do not represent nonspecific association of the probe with ribosomal RNAs, as no signal is found in lanes containing large amounts (20μg) of poly(A)-RNA from day mouse embryos (Fig. 5B).

Fig. 4.

Northern blot analysis of Hox2.l expression in poly(A)+RNA extracted from fetal and adult mouse tissues. Samples of RNA isolated from each tissue were electrophoresed on denaturing agarose gels, transferred to membranes and hybridized with a Hox2.1 cDNA probe (probe 2, Fig. 1). (A) Lane 1, -day; Lane 2, 1212-day embryo and Lane 3, adult kidney po)y(A) + RNA, 5pg per lane. (B) The fetal or adult origin of the poly(A)+ is indicated above each lane. In all samples 5 μg of RNA was loaded per lane, except for fetal lung, adult lung, adult spinal cord and visceral yolk sac (VYS) where 1 pg of RNA was applied. (C) The same filter as B treated with RNase A (see Methods). The relative mobilities of the 18 S (2·2 kb) and 28 S (4·7 kb) rRNAs are shown as size markers.

Fig. 4.

Northern blot analysis of Hox2.l expression in poly(A)+RNA extracted from fetal and adult mouse tissues. Samples of RNA isolated from each tissue were electrophoresed on denaturing agarose gels, transferred to membranes and hybridized with a Hox2.1 cDNA probe (probe 2, Fig. 1). (A) Lane 1, -day; Lane 2, 1212-day embryo and Lane 3, adult kidney po)y(A) + RNA, 5pg per lane. (B) The fetal or adult origin of the poly(A)+ is indicated above each lane. In all samples 5 μg of RNA was loaded per lane, except for fetal lung, adult lung, adult spinal cord and visceral yolk sac (VYS) where 1 pg of RNA was applied. (C) The same filter as B treated with RNase A (see Methods). The relative mobilities of the 18 S (2·2 kb) and 28 S (4·7 kb) rRNAs are shown as size markers.

Fig. 5.

Tissue-specific and temporal expression of Hox2.1 in poly(A)+RNA from mouse tissues. Above each lane is marked the tissue and stage in development when RNA was isolated. (A) Time course of Hox2.1 expression in the kidney and spinal cord. The amount of poly(A)+RNA loaded per lane was 5μg, except in the adult spinal cord (1·5 μg), the 141-day fetal kidney (2·5μg), and the 1212-day total embryo (2·5 μg). (B) Hox2.1 expression in liver and other weakly expressing tissues. Liver samples have 10μg of poly(A)+ per lane, the visceral yolk sac (VYS) and placenta (PLAC) I fig, the fetal gut and 1212-day embryo 2μg and a control with 20μg of poly(A)”RNA from 1212-day mouse embryos. Both A and B were treated with RNase A. The relative mobilities of the 18 S (2·2 kb) and 28 S (4·7kb) rRNAs are indicated as size markers.

Fig. 5.

Tissue-specific and temporal expression of Hox2.1 in poly(A)+RNA from mouse tissues. Above each lane is marked the tissue and stage in development when RNA was isolated. (A) Time course of Hox2.1 expression in the kidney and spinal cord. The amount of poly(A)+RNA loaded per lane was 5μg, except in the adult spinal cord (1·5 μg), the 141-day fetal kidney (2·5μg), and the 1212-day total embryo (2·5 μg). (B) Hox2.1 expression in liver and other weakly expressing tissues. Liver samples have 10μg of poly(A)+ per lane, the visceral yolk sac (VYS) and placenta (PLAC) I fig, the fetal gut and 1212-day embryo 2μg and a control with 20μg of poly(A)”RNA from 1212-day mouse embryos. Both A and B were treated with RNase A. The relative mobilities of the 18 S (2·2 kb) and 28 S (4·7kb) rRNAs are indicated as size markers.

This array of multiple RNA transcripts could be derived from the Hox2.1 gene or might represent extensive homology to transcripts from other genes. As a means of examining this problem, the filters in Fig. 4B were treated with ribonuclease (RNase A) to remove any parts of the labelled probe that were not completely matched with other RNAs. This treatment made a dramatic difference in the results (Figs 4C, 5). Some tissues, such as testes, which previously revealed at least seven bands, showed no appreciable signal after this treatment. Patterns in other tissues (spinal cord and kidney for example) did not vary, while the adult lung had two bands that were removed and two (hat remained unchanged. We have obtained similar results using probes from sequences which are 3’ of the Hox2.1 homeobox. The results, showing that many bands are removed by RNase A treatment, suggest that transcripts from other loci have extensive homology to the Hox2.1 gene in sequences both 5’ and 3’ of the homeobox. The series of bands resistant to RNase A demonstrate that multiple transcripts are also derived from the gene.

The results of Northern blot analysis, using the RNase A treatment, show that the 2·2 kb transcript is still the major species found in expressing tissues. The relative levels of the 2·2 kb transcript in embryonic and adult tissues analysed by Northern blotting experiments were quantified by densitometry, as de-scribed in Fig. 6. The fetal lung, spinal cord, gut, kidney, spleen, liver and visceral yolk sac contain the 2·2kb RNA, and the relative levels of expression in each of these tissues varies over a 1000-fold range, as shown in Fig. 6. Negative tissues include fetal brain, heart, muscle, placenta, amnion and day parietal yolk sac. The temporal expression of the gene during development and in the adult mouse is also shown in Figs 4, 5 and 6. There is a small decrease in spinal cord expression between and days in the fetus, and the sizes of the two transcripts in adult lung are slightly smaller than in fetal lung, which may reflect different poly(.A) lengths or different transcripts. However, in these two tissues there is essentially no change in the RNA levels between the fetal and adult stages. Levels of RNA decrease fourfold in the adult kidney as compared to fetal and neonatal kidney. In the gut, spleen and liver, RNAs are not detectable in the adult, and the Hox2.1 gene therefore appears to be turned off after birth in these tissues. The gut RNA levels display the largest decrease between the fetal and adult stages. However, this result is complicated by the fact that the dissected -day gut may also contain some spleen and pancreas. It is difficult to assess the relative contribution of these tissues to the level of expression at days in the gut and the difference may therefore be smaller than we have observed. However, there remains a large decrease in RNA levels between day gut, which is free of these contaminants, and adult intestine. The results summarized in Fig. 6 clearly show that the Hox2.1 gene is modulated during development and that there are tissue-specific differences which control the level and timing of expression.

Fig. 6.

Relative expression of the Hox2.1 gene in embryonic and adult mouse tissues. The poly(A) + RNA from the various tissues was analysed by Northern blotting in several experiments, including Figs 4, 5, and all filters were treated with RNase A (see Methods). The relative level of the 2·2 kb transcript for each tissue was determined by densitometer scanning of autoradiographs from multiple-timed exposures of each filter and compensated for by the amount of RNA loaded in each lane. The levels of RNA in the 1212-day embryo samples were used as an internal control in all scans and the level in the fetal lung was arbitrarily set as 100 %.

Fig. 6.

Relative expression of the Hox2.1 gene in embryonic and adult mouse tissues. The poly(A) + RNA from the various tissues was analysed by Northern blotting in several experiments, including Figs 4, 5, and all filters were treated with RNase A (see Methods). The relative level of the 2·2 kb transcript for each tissue was determined by densitometer scanning of autoradiographs from multiple-timed exposures of each filter and compensated for by the amount of RNA loaded in each lane. The levels of RNA in the 1212-day embryo samples were used as an internal control in all scans and the level in the fetal lung was arbitrarily set as 100 %.

The complex RNA patterns could represent precursors, multiple transcription starts of differentially processed transcripts from the Hox2.1 gene. At present we cannot distinguish between these alternatives, but two points must be made. First, the spinal cord and kidney characteristically have four different transcripts larger than the 2·2 kb RNA (3·8, 6,10 and 12 kb) which hybridize to the Hox 2.1 probe. The lung also has the same species but they are in very low abundance. The level of each of these larger transcripts does not correlate with the level of the major 2·2 kb RNA, in that the ratio of larger species to the 2·2 kb RNA is different in each tissue. Second, the developmental time course for spinal cord and kidney (Figs 4, 5) shows that despite the different ratios of larger transcripts to the 2·2 kb RNA in these tissues, the relative ratio in a given tissue does not change during development. Therefore, regardless of how these RNA species are derived from the Hox2.1 gene, they are coordinately regulated with the major Hox2.1 2·2 kb transcript in a tissue-specific manner during development.

In situ hybridization

The Northern analysis has allowed us to identify general and temporal patterns of Hox2.1 expression in the mouse embryo, but the relative levels of RNA from dissected tissue (Fig. 6) does not take into account that these tissues are composed of many cell types. Understanding any functional role for a homeobox-containing gene, however, requires more detailed information on the spatial and cell-type specificity of the gene expression. Toward this goal in situ hybridization was used to investigate the distribution of Hox2.1 transcripts in the -day mouse embryo. A part of the Hox 2.1 DNA was cloned into pGEM (Fig. 1, probe 2) to provide single-stranded RNA anti-sense (-) probes for mRNA and sense (+) control probes.

Results of in situ hybridization on frozen sections are shown in Fig. 7. Very strong hybridization, specific to the (—) strand probe, was seen in the embryonic spinal cord and lung. Hybridization in the spinal cord is most intense directly posterior to the hind brain, but it does extend into the hind brain region. There is less hybridization occurring more caudally, suggesting an apparent gradient from the anterior to posterior regions of the spinal cord. No signal above background is detected in the brain. Grain density was higher than background over dense aggregates of red blood cells (e.g. in dorsal aorta). This result, however, was observed with both ( + ) and (—) strand probes, and was therefore considered to be nonspecific. The hybridization pattern in the embryonic lung is particularly striking in a slightly more lateral section, which passes through three branches of the lung (Fig. 7C,D). Comparison of bright-field and dark-ground photomicrographs reveals that the hybridization grains are most intense in the mesenchymal cells that envelop the branching epithelial cell layer. There appears to be little or no expression in this lung epithelial layer.

Fig. 7.

Expression of Hox2.1 sequences in sections of 1212-day mouse embryos revealed by in situ hybridization. Sagittal cryostat sections from frozen1212-day embryos were hybridized with anti-sense (-) and sense (+) single-stranded Hox 2.1 RNA probes (probe 1, Fig. 1). (A) Whole embryo bright field, low’ power, (-) probe; (B) whole embryo bright-field, low-power, (+) probe; (C) lung bnght-field, high-power, (-) probe; (D) lung dark-field, high power, (—) probe. Bar in A represents 800μm, and in C 100μm.

Fig. 7.

Expression of Hox2.1 sequences in sections of 1212-day mouse embryos revealed by in situ hybridization. Sagittal cryostat sections from frozen1212-day embryos were hybridized with anti-sense (-) and sense (+) single-stranded Hox 2.1 RNA probes (probe 1, Fig. 1). (A) Whole embryo bright field, low’ power, (-) probe; (B) whole embryo bright-field, low-power, (+) probe; (C) lung bnght-field, high-power, (-) probe; (D) lung dark-field, high power, (—) probe. Bar in A represents 800μm, and in C 100μm.

The striking degree of conservation observed in homeobox sequences from many species, including mammals, has led to the speculation that, by analogy to Drosophila, the mammalian homeobox counterparts may play an important role in regulation of development. Drosophila homeobox genes are involved in control of the segmental body plan by specifying the number, polarity and identity of segments in the embryo (reviewed by Gehring, 1985; Scott, 1985). However, it is difficult to make direct functional comparisons with mammals, as segmentation in vertebrates appears to have evolved independently of Arthropods, and it is still not known whether lineage restricted compartments are an important feature of mammalian development (for review, Hogan, Holland & Schofield, 1985). Genetic analysis of these developmental strategies in mouse or other mammals is much more difficult than in Drosophila and relatively few loci that affect morphogenesis are known. There is no clear evidence for mouse homeotic genes that affect unique somite identity (Hogan et al. 1985) and none of the mouse homeobox-containing genes isolated thus far have been shown to be allelic with mapped developmental mutants. Indeed, the only support for a role of homeobox genes in mouse development comes from studies on their expression. Reports have demonstrated that mouse homeobox genes are expressed in different tissues at both embryonic and adult stages (Colberg-Poley et al. 1985a,b;Jackson et al. 1985; Ruddle, Hart, Awgulewitsch, Fainsod, Utset, Dalton, Kerk, Rabin, Ferguson-Smith, Fienberg & McGinnis, 1985), and that, in the case of Hox3, the localization of the transcripts is spatially restricted in the central nervous system (Awgulewitsch et al. 1986). Therefore, to address the possible function of the mouse Hox 2.1 homeobox gene we have isolated cDNA sequences to analyse the structure of the predicted protein product and examined the tissue-specific, temporal and spatial patterns of its expression as a step towards linking gene activity with known developmental processes.

Jackson et al. (1985) have demonstrated that the Hox2.1 gene is expressed in mouse embryos as early as day p.c. during the late primitive-streak-stage and it is enriched in fetal spinal cord and adult kidney. The results from our analysis show that the Hox2.1 gene is expressed in a broad spectrum of embryonic tissues, including fetal lung, spinal cord, gut, kidney, spleen, liver and visceral yolk sac, but the levels of mRNA in each of these tissues varies over a considerable range (see Fig. 6). Negative tissues include fetal brain, heart, muscle, placenta, amnion and -day parietal yolk sac. Tbe fetal lung, spinal cord and gut have the highest levels of transcripts, and the level in fetal liver is the lowest that we are able to detect. The timing of Hox2.1 expression during development also varies and displays tissue-specific differences. In general, the highest level of expression is observed in the earliest fetal stage dissected for each tissue, with the possible exception of lung. We have never detected transcripts in an adult tissue if not present in its embryonic or fetal counterpart. Due to the difficulty of dissecting sufficient material for Northern RNA hybridization, we have not been able to extend this analysis to time points earlier than days p.c.

Therefore, we do not know whether Hox2.1 transcription begins during the formation of organ primordia, in the early stages of organogenesis or at later times. Several distinct temporal patterns are illustrated in Fig. 6. Fetal and adult levels of expression are roughly equal in the lung and spinal cord, suggesting that once the gene is activated in these tissues it remains on throughout adult life. This agrees with the finding that Hox2.1 transcripts are observed in the adult mouse central nervous system (Ruddle et al. 1985) and that the human Hox2.1 homologue is expressed in the human spinal cord in early embryonic stages (Simeone et al. 1986). How-ever, in the kidney, adult RNA levels are fourfold lower than fetal levels and the remaining tissues (gut, spleen and liver) do not appear to express the gene at all in adult stages. The decrease in levels of Hox2.1 expression in some of these tissues (kidney, spleen and liver) occurs after birth, while in the gut this transition is detected between days of fetal development.

The overall tissue distribution and temporal pattern of Hox2.1 gene expression is different to that observed for other mouse homeobox genes (Colberg-Poley et al. 1985b; Awgulewitsch et al. 1986; Rubin, Toth, Patel, D’Eustachio & Nuguyen-Huu, 1986; Wolgemuth et al. 1986). However, some tissues, such as spinal cord and kidney, contain transcripts from several homeobox genes. These results imply that the individual mouse homeobox genes respond to different tissue-specific and developmental factors which regulate their expression. There is no obvious common origin or feature of these tissues that could account for the large tissue-specific variations in RNA levels and the timing and extent of temporal changes in Hox2.1 gene expression observed. The pattern of embryonic expression reported in this study may therefore result from the action of multiple control elements or factors active during organogenesis in later stage embryos and adults. By contrast, expression in the earlier embryo may be linked to a different functional role. It will be essential, there-fore, to investigate Hox2.1 expression in to day p.c. embryos via in situ hybridization to observe the patterns of expression in the stages when the germ layers and the body axes are first established.

If the activity of mouse homeobox genes provides a framework for positional information in the embryo it is important to examine the cell-type specificity and spatial localization of transcripts during development. In situ hybridization experiments with the Hox2.1 gene (Fig. 7) in day p.c. embryos clearly demonstrates that transcripts are localized in the spinal cord. The highest levels of RNA appear in the posterior region of the hind brain and extend in an apparent decreasing gradient to posterior regions of the spinal cord. The highest signal appears over the central portion of the cord and at the edges, but cells completely spanning the dorsal-ventral axis of the cord are also expressing Hox2.1. Many of these cells have very different fates and are no longer dividing, so there is no strict correlation of levels of expression with rates of cell division or fate in this region of the spinal cord. The Drosophila homeobox-containing genes of the ANT-C and BX-C complexes also specify transcripts that appear to accumulate in discrete regions of the mature embryonic central nervous system (Harding, Wedeen, McGinnis & Levine, 1985; Levine, Hafen, Garber & Gehring, 1983). The spatial localization of Hox 2.1 in the central nervous system is similar to that observed with the Hox3 gene (Awgulewitsch et al. 1986). However, the Hox2.1 transcripts appear to be expressed in a more anterior portion of the spinal cord. Therefore, these two genes have similar overlapping but not identical regional localizations which may be a feature of other mam-malian homeobox genes as in Drosophila.

The in situ hybridization results (Fig. 7C) also reveal that Hox2.1 is not uniformly expressed in all cells of the day p.c. embryonic lung. The mes-enchymal cells accumulate high levels of Hox 2.1 RNA, in marked contrast to the epithelial cell layer that they surround. Essentially no signal is detected over the epithelial cells. The high level of expression in the lung mesoderm cells is not a general property of ah mesoderm cells, as many other mesodermally derived tissues in the day embryo show no signal with the Hox 2.1 probe (Fig. 7A,B). During this stage the lung epithelia have been induced to branch and it is important to examine whether Hox 2.1 expression is involved in, or marks, this process in the lung.

One of the characteristics of Drosophila homeobox genes, like Antennapedia (Antp), is that they have multiple promoters, which generate a series of different RNA transcripts by alternative RNA splicing, and these transcripts are able to encode similar but different proteins (Scott, Weiner, Polisky, Hazelrigg, Pirotta, Scalengne & Kaufman, 1983; Carroll et al. 1986; Schneuwly et al. 1986). The transcriptional analysis of the Hox2.1 gene (Figs 4, 5) also revealed a complex pattern of expression involving multiple transcripts. This raises an important point concerning any Hox2.1 gene product(s). We have assumed that the Hox2.1 protein is synthesized in parallel with the accumulation of the mRNA, but if multiple proteins are possible a detailed analysis on the distribution of specific products is required to address a functional role. We are presently using the predicted amino acid sequence from our cDNA clones (Fig. 2) to generate Hox 2.1-specific antibodies for this purpose.

Analysis of the predicted Hox 2.1 amino acid sequence shows that the protein contains high levels of serine (22%) and proline (9%) in the domains outside of the homeobox, and that a hexapeptide region upstream of the homeodomain is conserved in homeobox genes from several species (Fig. 3). A consensus squence for this peptide shows an aliphatic amino acid (Ile/Leu/Val) in position one, Phe or Tyr in position two, followed by the tripeptide Pro-Trp-Met, ending in the basic residue either Lys or Arg. The conservation of the sequence and its location relative to the homeodomain raises the possibility that it is functionally related to the homeodomain. The Hox2.1 gene must have additional regions of homology with other mouse loci. The Northern analysis experiments (Fig. 4) carried out under high stringency conditions with and without RNase A, clearly demonstrate that there are transcripts in many tissues with extensive homology to the Hox2.1 gene probe. If the clusters of mouse homeobox genes (Colberg-Poley, 1985a,b; Hart et al. 1985; Dubuole et al. 1986) arose via duplication and divergence, then some members of these gene clusters should be very homologous. The degree of Hox2.1 homology with other family members can only be determined when further sequences are available. Under conditions of reduced stringency we have observed (unpublished) that probe 2 recognizes Hox 2.1 homologous sequences in frog, chicken, cow, cat and human genomic DNA but not in Drosophila. Therefore, the Hox2.1 gene appears to be conserved in vertebrate evolution and hence comparisons of the Hox2.1 protein products from several species could provide clues to important functional domains.

The results presented here provide a detailed descriptive picture of the diversity in tissue-specific, temporal, cell type and spatial expression of the Hox2.1 gene. As similar information on other vertebrate homeobox-containing genes arises, there will be a basis for examining functional relationships between the different genes. However, direct confirmation of an embryonic functional role is still required. The production of transgenic mice via introduction of modified genes can test for dominant effects in embryogenesis. Coupled with the use of antibody probes to investigate protein expression in the early embryo, it may be possible to resolve many of the outstanding questions.

We thank Dr Bob Hill for help in screening the cDNA library, Drs Peter Gruss and Ward Odenwald for unpublished sequence information, Raechel Kenney for help in sequencing, Sarah Harper for technical help on in situ hybridization, Dr Colin Hetherington for animal assistance and Lydia Pearson for preparing the manuscript. We also thank Drs Mike Snow, David Murphy, Nick Hastie, Robin Lovell-Badge, Jack Price, Dennis Summerbell and Ian Jackson for valuable discussion.

Auffray
,
C.
&
Rougeon
,
F.
(
1980
).
Purification of mouse immunoglobulin heavy-chain messenger RNAs from total myeloma tumor RNA
.
Eur. J. Biochem
.
107
,
303
324
.
Aviv
,
H.
&
Leder
,
P.
(
1972
).
Purification of biologically active globin messenger RNA by chromatography on oligothymidylic acid-cellulose
.
Proc. natn. Acad. Sci. U.S.A
.
69
,
1408
1412
.
Awgulewitsch
,
A.
,
Utset
,
M. F.
,
Hart
,
C. F.
,
Mcginnis
,
W.
&
Ruddle
,
F. H.
(
1986
).
Spatial restriction in expression of a mouse homeobox locus within the central nervous system
.
Nature, Lond
.
320
,
328
335
.
Beachy
,
P. A.
,
Halfand
,
S. L.
&
Hogness
,
D. S.
(
1985
).
Segmental distribution of bithorax complex proteins during Drosophila development
.
Nature, Lond
.
313
,
545
551
.
Bishop
,
J. O.
,
Rosbash
,
M.
&
Evans
,
D.
(
1974
).
Polynucleotide sequences in eukaryotic DNA and RNA that form ribonuclease-resistant complexes with polyuridylic acid
.
J. molec. Biol
.
85
,
75
86
.
Bonccnelli
,
E.
,
Simeone
,
A.
,
La Volpe
,
A.
,
Faiella
,
A.
,
Fidanza
,
V.
,
Acampora
,
D.
&
Scorro
,
L.
(
1985
).
Human cDNA clones containing homeobox sequences
.
Cold Spring Harbor Symp. Quant. Biol
.
50
,
301
306
.
Breir
,
G.
,
Bucan
,
M.
,
Francke
,
U.
,
Colberg-Poley
,
A. M.
&
Gruss
,
P.
(
1986
).
Sequential expression of murine homeobox genes during F9 EC cell differentiation
.
EMBO J
.
5
,
2209
2215
.
Carrasco
,
A. E.
,
Mcginnis
,
W.
,
Gehring
,
W. J.
&
De Robertis
,
E. M.
(
1984
).
Cloning of an X. laevis gene expressed during early embryogenesis coding for a peptide region homologous to Drosophila homeotic genes
.
Cell
37
,
409
414
.
Carroll
,
S. B.
,
Laymon
,
R. A.
,
Mccutcheon
,
M. A.
,
Riley
,
P. D.
&
Scott
,
M. P.
(
1986
).
The localization and regulation of Antennapedia protein expression in Drosophila embryos
.
Cell
47
,
113
122
.
Carroll
,
S. B.
&
Scott
,
M. P.
(
1985
).
Localization of the fushi tarazu protein during Drosophila embryogenesis
.
Cell
43
,
44
57
.
Chou
,
P. Y.
&
Fasman
,
G. D.
(
1978
).
Prediction of the secondary structure of proteins from their amino acid sequence
.
Adv. Enzym
.
47
,
415
419
.
Colberg-Poley
,
A. M.
,
Voss
,
S. D.
,
Chowdhury
,
K.
&
Gruss
,
P.
(
1985a
).
Structural analysis of murine genes containing homeobox sequences and their expression in embryonal carcinoma cells
.
Nature, Lond
.
314
,
713
718
.
Colberg-Poley
,
A. M.
,
Voss
,
S. D.
,
Chowdhury
,
K.
,
Stewart
,
C. L.
,
Wagner
,
E. F.
&
Gruss
,
P.
(
1985b
).
Clustered homeoboxes are differentially expressed during murine development
.
Cell
43
,
39
45
.
Desplan
,
C.
,
Theis
,
J.
&
O’Farrell
,
P. H.
(
1985
).
The Drosophila developmental gene, engrailed, encodes a sequence-specific DNA binding activity
.
Nature, Lond
.
318
,
630
635
.
Dolecki
,
G. J.
,
Wannakratroj
,
S.
,
Lum
,
R.
,
Wang
,
G.
,
Riley
,
H. D.
,
Carlos
,
R.
,
Wang
,
A.
&
Humphreys
,
T.
(
1986
).
Stage-specific expression of a homeobox-containing gene in the non-segmented sea urchin embryo
.
EMBO J
.
5
,
925
930
.
Dubuole
,
D.
,
Baron
,
A.
,
Mahl
,
P.
&
Galuot
,
B.
(
1986
).
A new homeo-box is present in overlapping cosmid clones which define the mouse Hox-1 locus
.
EMBO J
.
5
,
1973
1980
.
Farnher
,
K.
,
Hogan
,
B. L. M.
&
Flavell
,
R.
(
1987
).
Transcription of H-2 and Qa genes in embryonic and adult mice
.
EMBO J. (in press)
.
Garber
,
R. L.
,
Kuroiwa
,
K.
&
Gehring
,
W. J.
(
1983
).
Genomic and cDNA clones of the homeotic gene Antennapedia Drosophila
.
EMBO J
.
2
,
2027
2034
.
Gehring
,
W. J.
(
1985
).
The homeobox: a key to the understanding of development?
Cell
40
,
3
5
.
Harding
,
K.
,
Wedeen
,
C.
,
Mcginnis
,
W.
&
Levine
,
M.
(
1985
).
Spatially regulated expression of homeotic genes in Drosophila
.
Science
229
,
1236
1242
.
Hart
,
C. P.
,
Awgulewitsch
,
A.
,
Fainsod
,
A.
,
Mcginnis
,
W.
&
Ruddle
,
F. H.
(
1985
).
Horneo box gene complex on mouse chromosome 11: molecular cloning, expression in embryogenesis and homology to a human horneo box locus
.
Cell
43
,
9
18
.
Harvey
,
R. P.
,
Tabin
,
C. J.
&
Melton
,
D. A.
(
1986
).
Embryonic expression and nuclear localization of Xenopus homeobox (Xhox) gene products
.
EMBO J
.
5
,
1237
1244
.
Hauser
,
C. A.
,
Joyner
,
A. L.
,
Klein
,
R. D.
,
Learned
,
T. K.
,
Martin
,
G. R.
&
Tjian
,
R.
(
1985
).
Expression of homologous horneo box containing genes in differentiated human teratocarcinoma cells and mouse embryos
.
Cell
43
,
19
28
.
Hogan
,
B. L. M.
,
Costantini
,
F.
&
Lacy
,
E.
(
1986
).
Manipulating the Mouse Embryo. A Laboratory Manual
.
Cold Spring Harbor, New York
:
Cold Spring Harbor Laboratory
.
Hogan
,
B. L. M.
,
Holland
,
P. W. H.
&
Schofield
,
P.
(
1985
).
How is the Mouse Segmented?
Trends in Genetics
1
,
67
74
.
Holland
,
P. W. H.
&
Hogan
,
B. L. M.
(
1986
).
Phylogenetic distribution of Antennapedia-like homeoboxes
.
Nature, Lond
.
321
,
251
253
.
Jackson
,
I. J.
,
Schofield
,
P
&
Hogan
,
B. L. M.
(
1985
).
A mouse homeobox gene is expressed during embryogenesis and in adult kidney
.
Nature, Lond
.
317
,
745
748
.
Jaenisch
,
R.
,
Harbers
,
K.
,
Schnieke
,
A.
,
Lohler
,
J.
,
Chumakov
,
I.
,
Jahner
,
D.
,
Grotkopp
,
D.
&
Hoffman
,
E.
(
1985
).
Germline integration of Moloney murine leukemia virus at the MOV13 locus leads to recessive lethal mutation and early embryonic death
.
Cell
32
,
209
216
.
Johnson
,
A. D.
&
Herskowttz
,
L.
(
1985
).
A repressor (MAT a2 product) and its operator control expression of a set of cel) type-specific genes in yeast
.
Cell
42
,
237
247
.
Joyner
,
A. L.
,
Kornberg
,
T.
,
Coleman
,
K. G.
,
Cox
,
D. R.
&
Martin
,
G. R.
(
1985
).
Expression during embryogenesis of a mouse gene with sequence homology to the Drosophila engrailed gene
.
Cell
43
,
29
37
.
Kaufman
,
T. C.
,
Lewis
,
R.
&
Wakimoto
,
B.
(
1980
).
Cytogenetic analysis of chromosome 3 in Drosophila melanogaster. the homeotic gene complex in polytene chromosome interval 84A-B
.
Genetics
94
,
115
133
.
Krumlauf
,
R.
,
Hammer
,
R.
,
Th.Ghman
,
S. M.
&
Brinster
,
R.
(
1985
).
Developmental regulation of alpha-fetoprotein in transgenic mice
.
Mol. cell. Biol
.
5
,
163
168
.
Kuroiwa
,
A.
,
Kloter
,
U.
,
Baumgartner
,
P.
&
Gehring
,
W. J.
(
1985
).
Cloning of the homeotic sex combs reduced gene in Drosophila. An in situ localization of its transcripts
.
EMBO J
.
4
,
3757
3764
.
Laughon
,
A.
,
Carroll
,
S. B.
,
Storfer
,
F. A.
,
Riley
,
P. D.
&
Scott
,
M. A.
(
1985
).
Common properties of proteins encoded by the Antennapedia complex genes of Drosophilia melanogaster
.
Cold Spring Harbor Symp. Quant. Biol
.
50
,
253
262
.
Laughon
,
A.
&
Scott
,
M. P.
(
1984
).
Sequence of a Drosophila segmentation gene: protein structure homology with DNA-binding proteins
.
Nature, Lond
.
310
,
25
31
.
Levine
,
M.
,
Hafen
,
E.
,
Garber
.
R. L.
&
Gehring
,
W.J.
(
1983
).
Spatial distribution of Antennapedia transcripts during Drosophila development
.
EMBO J
.
2
,
2037
2046
.
Levine
,
M.
,
Rubin
,
G. N.
&
Than
,
R.
(
1984
).
Human DNA sequences homologous to a protein coding region conserved between homeotic genes of Drosophila
.
Cell
38
,
667
673
.
Lewis
,
E. B.
(
1978
).
A gene complex controlling segmentation in Drosophila
.
Nature, Lond
.
276
,
565
570
.
Maxam
,
A. M.
&
Gilbert
,
W.
(
1980
).
Sequencing end-labeled DNA with base-specific chemical cleavage
.
Meth. Enzymol
.
65
,
499
560
.
Mcginnis
,
W.
(
1985
).
Homeobox sequences of the Antennapedia class are conserved only in higher animal genomes
.
Cold Spring Harbor Symp. Quant. Biol
.
50
,
263
270
.
Mcginnis
,
W.
,
Levine
,
M. L.
,
Hafen
,
E.
,
Kuroiwa
,
A.
&
Gehring
,
W. J.
(
1984a
).
A conserved DNA sequence in homeotic genes of the Drosophila, Antennapedia and Bithorax complexes
.
Nature, Lond
.
308
,
428
433
.
Mcginnis
,
W.
,
Garber
,
R. L.
,
Wirz
,
J.
,
Kuroiwa
,
A.
&
Gehring
,
W. J.
(
1984b
).
A homologous protein-coding sequence in Drosophila genes and its conservation in other metazoans
.
Cell
37
,
403
408
.
Mcginnis
,
W.
,
Hart
,
C. P.
,
Gehring
,
W. J.
&
Ruddle
,
F. H.
(
1984C
).
Molecular cloning and chromosome mapping of a mouse DNA sequence homologous to homeotic genes of Drosophila
.
Cell
38
,
675
680
.
Melton
,
D.
,
Krieg
,
P.
,
Rebaguati
,
M.
,
Maniatis
,
T.
,
Zinn
,
K.
&
Green
,
M.
(
1984
).
Efficient in vitro synthesis of biologically active RNA and RNA probes from plasmids containing a bacteriophage SP6 promoter
.
Nucl. Acids Res
.
12
,
7035
7056
.
Mlodzik
,
M.
,
Fjose
,
A.
&
Gehring
,
W. J.
(
1985
).
Isolation of caudal, a Drosophila homeobox-containing gene with maternal expression, whose transcripts form a concentration gradient at the pre-blastoderm stage
.
EMBO J
.
4
,
2961
2969
.
Muller
,
M. M.
,
Carrasco
,
A. E.
&
De Robertis
,
E. M.
(
1984
).
A homeo-box-containing gene expressed during oogenesis in Xenopus
.
Cell
39
,
157
162
.
Munke
,
M.
,
Cox
,
D. R.
,
Jackson
,
I. J.
,
Hogan
,
B. L. M.
&
Francke
,
U.
(
1986
).
The murine Hox-2 cluster of homeobox containing genes maps distal on chromosome 11 near the fail-short (Ts) locus
.
Cy tosen. Cell Genets
.
42
,
236
240
.
Poole
,
P. W. J.
,
Kauvar
,
L. M.
,
Drees
,
B.
&
Kornberg
,
T.
(
1985
).
The engrailed locus of Drosophila: structural analysis of an embryonic transcript
.
Cell
40
,
37
43
.
Rabin
,
M.
,
Hart
,
C. P.
,
Ferguson-Smith
,
A.
,
Mcginnis
,
W.
,
Levine
,
M.
&
Ruddle
,
F. H.
(
1985
).
Two horneo box loci mapped in evolutionarily related mouse and human chromosomes
.
Nature, Lond
.
314
,
175
177
.
Reguuski
,
M.
,
Harding
,
K.
,
Kostriken
,
R.
,
Karch
,
F.
,
Levine
,
M.
&
Mcginnis
,
W.
(
1985
).
Horneo box genes of the Antennapedia and Bithorax complexes of Drosophila
.
Cell
43
,
71
80
.
Rosbash
,
M.
&
Ford
,
P. J.
(
1974
).
Polyadenylic acid-containing RNA in Xenopus laevis oocytes
.
J. molec. Biol
.
85
,
87
101
.
Rubin
,
M. R.
,
Toth
,
L. E.
,
Patel
,
M. D.
,
D’Eustachio
,
P.
&
Nuguyen-Huu
,
M. C.
(
1986
).
A mouse homeobox gene is expressed in spermatocytes and embryos
.
Science
233
,
663
667
.
Ruddle
,
F. H.
,
Hart
,
C. P.
,
Awgulewitsch
,
A.
,
Fainsod
,
A.
,
Ljtset
,
M.
.
Dalton
,
D.
,
Kerk
,
N.
,
Rabin
,
M.
,
Ferguson-Smith
,
A.
,
Fienberg
,
A.
&
Mcginnis
,
W.
(
1985
).
Mammalian homeobox genes
.
Cold Spring Harbor Symp. Quant. Biol
.
50
,
277
284
.
Sanger
,
F.
,
Nicklen
,
S.
&
Coulson
,
A. R.
(
1977
).
DNA sequencing with chain-terminating inhibitors
.
Proc, natn. Acad. Sci. U.S.A
.
74
,
5463
5467
.
Schneuwly
,
S.
,
Kuroiwa
,
A.
,
Baumgartner
,
P.
&
Gehring
,
W. J.
(
1986
).
Structural organization and sequence of the homeotic gene Antennapedia of Drosophila melanogaster
.
EMBO J
.
5
,
733
739
.
Scott
,
M. P.
(
1985
).
Molecules and puzzles from the Antennapedia homeotic gene complex of Drosophila
.
Trends in Genetics
1
,
74
80
.
Scott
,
M. P.
&
Weiner
,
A. J.
(
1984
).
Structural relationships among genes that control development: sequence homology between the Antennapedia, Ultrabithorax, and fushi tarazu loci of Drosophila
.
Proc, natn. Acad. Sci. U.S.A
.
81
,
4115
4119
.
Scott
,
M. P.
,
Weiner
,
A. J.
,
Polisky
,
B. A.
,
Hazelrigg
,
T. I.
,
Pirrotta
,
V.
,
Scalengne
,
F.
&
Kaufman
,
T. C.
(
1983
).
The molecular organization of the Antennapedia complex of Drosophila
.
Cell
35
,
763
776
.
Shepherd
,
J. C. W.
,
Mcginnis
,
W.
,
Carrasco
,
A. E.
,
De Robertis
,
E. M.
&
Gehring
,
W. J.
(
1984
).
Fly and frog horneo domains show homologies with yeast mating type regulatory proteins
.
Nature, Lond
.
310
,
70
71
.
Simeone
,
A.
,
Malvilio
,
F.
,
Bottero
,
L.
,
Giampaolo
,
A.
,
Russo
,
G.
,
Faiella
,
A.
,
Boncinelli
,
E.
&
Peschle
,
C.
(
1986
).
A human horneo box gene specifically expressed in spinal cord during embryonic development
.
Nature, Lond
.
320
,
763
765
.
White
,
R. A. H.
&
Wilcox
,
M.
(
1984
).
Protein products of the Bithorax complex in Drosophila
.
Cell
39
,
163
171
.
Whiteway
,
M.
&
Szostak
,
J. W.
(
1985
).
The ARDI gene of yeast functions in the switch between the mitotic cell cycle and alternative developmental pathways
.
Cell
43
,
483
492
.
Wolgemuth
,
D. J.
,
Engelmyer
,
E.
,
Duggal
,
R. N.
,
Gizang-Ginsberg
,
E.
,
Mutter
,
G. L.
,
Ponzetto
,
C.
,
Viviano
,
C.
&
Zakeri
,
Z. F.
(
1986
).
Isolation of a mouse cDNA coding for a developmentally regulated testis-specific transcript containing horneo box homology
.
EMBO J
.
5
,
1229
1235
.
Woychick
,
R. P.
,
Stewart
,
T. A.
,
Davis
,
R.
,
D’Eustachio
,
P.
&
Leder
,
P.
(
1985
).
An inherited limb deformity created by insertional mutagenesis in transgenic mice
.
Nature, Lond
.
318
,
36
40
.