The Hox genes are a class of putative developmental control genes that are thought to be involved in the specification of positional identity along the anteroposterior axis of the vertebrate embryo. It is apparent from their expression pattern that their regulation is dependent upon positional information. In a previous analysis of the Hox-1·1 promoter in transgenic mice, we identified sequences that were sufficient to establish transgene expression in a specific region of the embryo. The construct used, however, did not contain enough regulatory sequences to reproduce all aspects of Hox-1·1 expression. In particular, neither a posterior boundary nor a restriction of expression to prevertebrae was achieved. Here we show correct regulation by Hox-1·1 sequences in transgenic mice and identify the elements responsible for different levels of control. Concomitant with the subdivision of mesodermal cells into different lineages during gastrulation and organogenesis, Hox-1·1 expression is restricted to successively smaller sets of cells. Distinct elements are required at different stages of development to execute this developmental programme. One position-responsive element (130 bp nontranslated leader) was shown to be crucial for the restriction of expression not only along the anteroposterior axis of the embryo, setting the posterior border, but also along the dorsoventral axis of the neural tube and to the lineage giving rise to the prevertebrae. Thus, Hox-1 ·1 expression is established in a specific region of the embryo and in a specific lineage of the mesoderm by restricting the activity of the promoter by the combined effect of several regulatory elements.

During the development of multicellular organisms, groups of cells acquire different fates and form distinct structures. How cells acquire the properties that allow them to follow the different developmental pathways is of fundamental importance to developmental biology.

A molecular analysis of developmental mutations in Drosophila melanogaster has led to the elucidation of the basic molecular mechanisms that establish the anteroposterior and the dorsoventral axes of the larva (Akam, 1987; Ingham, 1988). A similar understanding of mammalian development, however, is still to be achieved. Several families of putative mammalian control genes have been isolated recently (Dressier and Gruss, 1988; Kessel and Gruss, 1990). Analyses of mouse mutants, ectopic expression in transgenic mice and experiments in Xenopus have shown that these genes serve a regulatory function during embryogenesis (Harvey and Melton, 1988; Wright et al. 1989; Balling et al. 1988, 1989; Kessel et al. 1990; Kessel and Gruss, 1990).

The Hox genes are likely candidates for factors assigning identities to different segments in mesoderm and neuroectoderm (Kessel and Gruss, 1990). It is apparent from their expression pattern that they are themselves regulated depending on the interpretation of positional information. A molecular analysis of the regulatory elements responsible for this positionspecific activity will be a first step to unravel the process establishing the axes in the embryo.

During gastrulation, Hox-1·1 expression is established in a specific region of the neuroectoderm and the mesoderm. During organogenesis, expression in the mesoderm becomes progressively restricted to prevertebral cells and organs such as kidney and intestine (Mahon et al. 1988; Dressier and Gruss, 1989; Kessel, personal communication). At day 12 of development, Hox-1·1 is expressed in two overlapping domains in the neural tube, spinal ganglia (C5 to S4) and prevertebrae (T3 to T13).

A previous analysis in transgenic mice identified Hox-1·1 promoter sequences that were sufficient to establish transgene expression in a specific region of the embryo. The construct used, however, did not contain enough information to specify the posterior boundary and to restrict expression to prevertebrae, resulting in a uniform expression extending from the appropriate anterior limit of Hox-l·l expression to the tail of the embryo. Here we show correct regulation by Hox-1 ·1promoter sequences in transgenic mice and identify the elements that are necessary to specify the posterior boundary of expression and to restrict expression to pre vertebrae.

Recombinant DNA

To construct deletions in the promoter of mólacZl, the sequences between the Sall site (position 0, see Puschel et al 1990) and the following restriction sites were deleted (see also Fig. 1): BspMII at position 987 (mólacZIBm), Ball at 1255 (mólacZIBl), Stul at 1449 (mólacZISt), Xbal at 1801 (mólacZlXII), Alul at 2152 (mólacZlA) and Bam HI at 2047 (mólacZlB). To allow the construction of m61acZ4 (Fig. 3), oligonucleotides were synthesized containing the Hox-1·1 sequences (Kessel et al. 1987; Kessel and Gruss, 1988) from position 3639 (Sad site) to 3767 (ATG) and in addition several restriction sites from the pSPT19-polyhnker (sequence. GAGCTC….ATGAGTTCGGTACC). pm61acZ4 was constructed by linking the following fragments in the indicated order: a 3·7 kbp SaB-Sacl fragment from pm6 Δ (Puschel et al. 1990), the 130bp oligonucleotide, a Ikbp Kpnl-Clal fragment from pSPTlacZ-1 (A.W.P., unpublished), a 2kbp Clal-PstI fragment from pUR292 (Ruther and Muller-Hill, 1983), a 1·2kbp Pstl-Bglll fragment from λ m6 (Colberg-Poley et al. 1985; Kessel et al. 1987) containing a part of the first exon (from the Pstl site to the splice site), the intron and the second exon up to the Bglll site. The vector is based on pmólacZl. The cloning scheme results in the reconstitution of the genomic sequence from the Sall site up to the third codon of Hox-1 1 (Fig 3A). The lacZ coding sequences were fused in frame N-terminally to the first three aminoacids of Hox-1 1 and C-terminally to the Wox-7·2-coding sequences downstream of the Pstl site in the first Hox-1 1 exon. All constructs contain the same 1·7 kbp fragment providing a polyadenylation signal as in mólacZl (Puschel et al. 1990). Several transcription-initiation sites of the Hox-11 gene are clustered around the Sad site (A W P, unpublished results). We assume that the same sites are used in the transgenes. pSPTlacZ-1 was generated by cloning the BamPH-lacZ fragment of pMC1871 (Pharmacia) into the Barn HI site of pSPT19 with the Kpnl site of the polyhnker upstream of the 5’end of the lacZ reading frame. “Die plasmids pm61acZ9 and pmólacZll were generated by exchanging a Clal-Nstl (Clal site in lacZ, Nsil site in the second Hox-11 exon) or a Sall-Clal fragment derived from m61acZ4 with the corresponding fragment of mólacZl, respectively (Fig 3A).

Fig. 1.

Analysis of the Hox-1 1 promoter in transgenic mice Expression of the indicated fusion genes was analysed in transgenic mice by transient expression analysis or by analysis of expression in lines. The constructs used are indicated in the upper part. The constructs contain different amounts of promoter sequences, which are indicated by the lines. The table shows the anterior hmit of expression in prevertebrae, neural tube and spinal ganglia at day 12 p.c. for the different constructs. C5, fifth cervical ganglion; T3, third thoracic prevertebrae; no, expression is seen throughout the embryo The deletion analysis identified two elements: A, a 110 bp AluI-BamHI (B) fragment; P, the 1590 bp promoter extending from ßamHI to Sad (Sc) SI, Sall site, Pv, PVMII site; E, EcoRI site. The hatched box signifies lacZ coding sequences, the open box the Hox-1·1 3’ nontranslated sequence The numbers indicate kbp of DNA.

Fig. 1.

Analysis of the Hox-1 1 promoter in transgenic mice Expression of the indicated fusion genes was analysed in transgenic mice by transient expression analysis or by analysis of expression in lines. The constructs used are indicated in the upper part. The constructs contain different amounts of promoter sequences, which are indicated by the lines. The table shows the anterior hmit of expression in prevertebrae, neural tube and spinal ganglia at day 12 p.c. for the different constructs. C5, fifth cervical ganglion; T3, third thoracic prevertebrae; no, expression is seen throughout the embryo The deletion analysis identified two elements: A, a 110 bp AluI-BamHI (B) fragment; P, the 1590 bp promoter extending from ßamHI to Sad (Sc) SI, Sall site, Pv, PVMII site; E, EcoRI site. The hatched box signifies lacZ coding sequences, the open box the Hox-1·1 3’ nontranslated sequence The numbers indicate kbp of DNA.

Fig. 2.

Hox-1·1 promoter analysis. Embryos were analysed for transgene expression by staining for β -Gal activity. Expression of the following constructs is shown: (A) mólacZISt: embryo E367. (B) mólacZlA: embryo E386. (C) mólacZlB: embryo from line L62 (day 10 p.c., expression in later stages was severely reduced) (D)mólacZlB: embryo E254.(E)mólacZlB: embryo E2564 (F)mólacZlB: embryo E373.

Fig. 2.

Hox-1·1 promoter analysis. Embryos were analysed for transgene expression by staining for β -Gal activity. Expression of the following constructs is shown: (A) mólacZISt: embryo E367. (B) mólacZlA: embryo E386. (C) mólacZlB: embryo from line L62 (day 10 p.c., expression in later stages was severely reduced) (D)mólacZlB: embryo E254.(E)mólacZlB: embryo E2564 (F)mólacZlB: embryo E373.

Fig. 3.

Identification of elements specifiying the postenor limit and causing lineage restnction of expression.(A)Schematic representation of the injected fragments.B, BamHI; C, Cla\ site; E, EcoRI site; N, Afcfl site; P, PstI site; Pv, Pvull site, Sc, Sad; SI, Sall site. The hatched box signifies lacZ coding sequences, the open box the Hox-1.1 3’ nontranslated sequence, the black box the Hox-1 1 homeobox The numbers indicate kbp of DNA. (B) Summary of the results The structure of the Hox-1 1 gene is shown in the upper part. The boxes indicate three identified elements Element A (open box) is the promoter (3640 bp containing element A and P from Fig 1), extending from the Sah site (SI) to the Sad site (Sc). Element B (stippled box) contains 130 bp 5’ nontranslated sequence downstream of the Sad site Element C (hatched box) extends from a Psh site in the first exon to the PvuH site in the second exon For abbreviations see Fig. 1. The expression pattern of the various transgenes was evaluated at day 12 p c according to four entena: appropnate antenor boundary in ectoderm and mesoderm, appropnate postenor boundary in ectoderm and mesoderm, correct lineage restriction (expression in cells contnbuting only to prevertebrae) and appropriate expression relative to the dorsoventral axis in the neural tube. + indicates expression like the endogenous Hox-1·1 gene. – indicates inappropnate expression and (+) partially correct expression.

Fig. 3.

Identification of elements specifiying the postenor limit and causing lineage restnction of expression.(A)Schematic representation of the injected fragments.B, BamHI; C, Cla\ site; E, EcoRI site; N, Afcfl site; P, PstI site; Pv, Pvull site, Sc, Sad; SI, Sall site. The hatched box signifies lacZ coding sequences, the open box the Hox-1.1 3’ nontranslated sequence, the black box the Hox-1 1 homeobox The numbers indicate kbp of DNA. (B) Summary of the results The structure of the Hox-1 1 gene is shown in the upper part. The boxes indicate three identified elements Element A (open box) is the promoter (3640 bp containing element A and P from Fig 1), extending from the Sah site (SI) to the Sad site (Sc). Element B (stippled box) contains 130 bp 5’ nontranslated sequence downstream of the Sad site Element C (hatched box) extends from a Psh site in the first exon to the PvuH site in the second exon For abbreviations see Fig. 1. The expression pattern of the various transgenes was evaluated at day 12 p c according to four entena: appropnate antenor boundary in ectoderm and mesoderm, appropnate postenor boundary in ectoderm and mesoderm, correct lineage restriction (expression in cells contnbuting only to prevertebrae) and appropriate expression relative to the dorsoventral axis in the neural tube. + indicates expression like the endogenous Hox-1·1 gene. – indicates inappropnate expression and (+) partially correct expression.

The fusion gene was separated from the vector for embryo injections by ABndlll digestion. Restriction fragments were separated on agarose gels and isolated by electroelution. The fragment of interest was further purified by phenokchloro-form extraction, chloroform extraction, ethanol precipitation, gel filtration over a P30-column (Biorad) and filtration through a 0 22 μ m filter (Schleicher and Schuell).

Generation of transgenic mice

NMRI outbred mice and C57BL/6/DBA/F1 mice were purchased from the Zentralinstitut fur Versuchstierzucht, Hannover (Germany). Transgenic mice were produced essentially as described by Hogan et al (1986) 6-week-old NMRI female mice were superovulated by injecting intraperitoneally (i.p.) 5 I.u. of gonadotropin from pregnant mare serum 48 h prior to injecting 5i.u. of human chorionic gonadotropin Female mice were then mated with C57BL/6/ DBA/F1 male mice The next day, 1-cell embryos were flushed from the oviducts with M2 medium (Hogan et al. 1986) The eggs were freed of cumulus cells by hyaluronidase treatment. The male pronuclei of the fertilized eggs were microinjected with approximately 2 picohters of DNA at a concentration of 2ng μ l1. The injected, fertilized eggs were then transferred to pseudopregnant recipient NMRI female mice. Transgenic mice were identified by Southern blot analysis of DNA extracted from mouse tail biopsies as described by Hogan et al (1986). The expression pattern directed by the various constructs was analysed both in transient expression assays and in transgenic lines. For transient expression experiments, foster mice were killed 12 days after the retransfer and embryos (corresponding to the founder generation) were analysed for transgene activity Placental DNA was used to assay for the presence of the transgene. Embryos were staged according to Theiler (1989).

X-Gal staining and histology

Embryos between day 7 and 12 (the day of detection of the vaginal plug was designated day zero) were fixed in 1 % formaldehyde, 0·2% glutaraldehyde, 0 02% NP40, 1 × PBS for 30 min at 4 °C, followed by two washes in lx PBS for 20 mm at room temperature. Embryos were stained at 30°C in 1 μ gml − 1 X-Gal, 5HIM KjFe(CN)6, 5mM K4Fe(CN)6, 2mM MgCl2, l × PBS overnight For sectioning, stained embryos were fixed overnight in 4 % formaldehyde in PBS, dehydrated and embedded in paraffin Cryosections were air-dryed, fixed at 4°C for 10 min like whole embryos and stained overnight as described above. Expression levels were estimated by the time required to obtain detectable levels of staining. This time varied from one hour to one day (data not shown). No correlation between copy number of the transgene and expression levels were observed

Analysis of the Hox-1·1 promoter

Previously we were able to reproduce the Hox-1·1 expression pattern partially in transgenic mice by using 3·6kbp of promoter and 1·7kbp 3’ sequences (m6lacZl in Püschel et al. 1990). A series of deletions of the 3·6kbp promoter was analysed for expression in transgenic mice to delineate regulatory sequences further Expression was analysed both in transient assays (the embryos were analysed for β -Gal activity at day 12 of gestation) as well as in lines (Fig. 1). Deletion of up to 1·94 kbp did not result in a change of the expression pattern or the frequency of transgenic mice expressing the transgene (Table 1, Fig. 1, Fig 2A,B) Deletion of another 110bp (m6lacZlB) changed both the pattern and the frequency considerably (Table 2, Figs 1, 2D,E,F). Three independent transgenic embryos (E254, E264, E372) obtained in transient assays did not show any boundary of expression either in ectoderm or in mesoderm (Fig. 1). A different result was obtained when transgenic lines were analysed. The frequency of expressing lines was considerably reduced when compared to other constructs (Table 1). The mólacZlB construct was expressed in only 3 out of 17 lines. Expression in one mólacZlB line (L62) was similar to that of the mólacZl lines between days 8 and 10 (Fig. 2C). Expression at day 12 p.c. was considerably reduced relative to earlier stages in terms of the extent of expressing tissues (data not shown). It is not clear if this is due to the construct or to the site of integration. The construct was particularly sensitive to influences of the integration site, as two of the three fines showed expression that was different from any pattern observed so far and was unique to the particular line (ectopic expression). A low number of expressing lines and high frequency of ectopic expression is commonly observed with weak promoters (Allen et al. 1988). As has been discussed previously (Püschel et al. 1990), we define an expression pattern as characteristic for a construct when the same pattern is found in several independent lines. A characteristic expression of a transgene in a line does not exclude additional ectopic expression in the same line. In the following analyses, we will only describe the pattern characteristic for the construct.

Table 1.

Expression in transgenic embryos (transient assays) at day 12 p.c.

Expression in transgenic embryos (transient assays) at day 12 p.c.
Expression in transgenic embryos (transient assays) at day 12 p.c.

Reproduction of Hox-1·1 regulation

In an attempt to reproduce fully the Hox-1 1 regulation, we created a construct that preserved the genomic configuration of the Hox-1·1 gene by retaining as much genomic sequence as possible. We replaced the coding sequence of the first Hox-1·1 exon by lacZ coding sequences (m61acZ4). This resulted in the fusion of the first three codons of Hox-1·1 to the lacZ reading frame. lacZ was N-terminally fused in frame to the remainder of the Hox-1·1 reading frame. The resulting fusion protein contained the homeobox and is targeted to the nucleus in 3T3-cells (data not shown) and in the embryo (Fig. 6H), indicating the presence of a nuclear targeting signal in the homeobox (Hall et al. 1990), whereas ZucZ-activity of the other constructs showed no particular subcellular localisation. In addition to the sequences of mólacZl, the m61acZ4 construct contains 130bp 5’-nontranslated sequence, 96bp of the first Hox-1·1 exon, the Hox-1·1 intron and 146bp of the second exon all of which were deleted in mólacZl (Fig. 3).

One embryo (E297) and three lines (L14, L30, L53) expressing the m61acZ4 construct at moderate to high levels were recovered and analysed in detail Table 2, Figs 3B, 4B). This pattern was identical to that described for the endogenous Hox-1·1. The development of the embryonic expression pattern was analysed in the three transgenic lines (Table 2). The pattern in embryos of all three lines was identical between day 8 and 10. Line L14 did not show expression after day 10 p.c. Aside from some additional ectopic expression, the patterns of lines L30 and L53 were identical between day 10 and 12 p.c.

Mapping of the regulatory elements

The addition of two DNA segments to mólacZl to create m61acZ4 allowed the complete reproduction of Hox-1·1 regulation. We next determined if one or both of these two sequences are responsible for adding the necessary control elements to the móacZl construct. We constructed two plasmids where only one of these sequences was added, the 5’ nontranslated leader (mólacZll: element B) or the intron (m61acZ9: element C), respectively (Fig. 3B). The m61acZ9 fragment was tested by transient expression experiments in transgenic embryos. Three transgenic embryos (E319, E324, E325) expressing the transgene were recovered (Table 2, Fig 4C). All three embryos had the same expression pattern, which was identical to that of mólacZl (Fig. 3B).

Fig. 4.

Identification of two regulatory elements in the Hox-1·1 gene. Transgenic embryos were analysed for β -Gal activity at day 12 of development: (A) mólacZl (embryo from line L4, Püschel et al. 1990). (B)m61acZ4 (embryo E297). (C)m61acZ9 (embryo E325).(D)môlacZll (embryo from line L35).

Fig. 4.

Identification of two regulatory elements in the Hox-1·1 gene. Transgenic embryos were analysed for β -Gal activity at day 12 of development: (A) mólacZl (embryo from line L4, Püschel et al. 1990). (B)m61acZ4 (embryo E297). (C)m61acZ9 (embryo E325).(D)môlacZll (embryo from line L35).

One embryo (E347) and three lines (L17, L35 and L38) expressing the m6lacZl 1 transgene were obtained (Table 2, Figs 3B, 4D). Line L17 showed ectopic expression starting at day 10 (not shown). With this exception, all integrations gave rise to the same expression pattern between days 8 and 12 p.c. The expression pattern was identical to that of m61acZ4 with one exception. Whereas m61acZ4 expression is restricted at day 11 to a subpopulation of sclerotomal cells (Figs 5B, 6G,H) and at day 12 to prevertebrae (Figs 5D, 61), ß-Gal activity directed by mólacZll is seen in all sclerotomal cells within the expression domain at day 11 (Figs 5A,C, 6D,E) and throughout the prevertebral column, including the intervertebral disc anlagen at day 12 p.c. (Fig. 6F and not shown).

Fig. 5.

Lineage restriction by element C. Day 11 (A, B) and day 12 (C, D) embryos were stained for β -Gal activity. Anterior is to the left in A and C and to the right in B and D. White arrowheads delineate the boundaries of one mesodermal segment (sclerotome).(A) Expression in embryos of line L35 is seen throughout the sclerotome in the thoracic (A) and the caudal region (C). Expression directed by the m61acZ4 transgene is seen in a subset of sclerotomal cells in the thoracic (B, embryo of line L30) as well as in the caudal region (D, embryo of line L53). Also expression in the rib primordia is more limited in embryos expressing the m61acZ4 construct.

Fig. 5.

Lineage restriction by element C. Day 11 (A, B) and day 12 (C, D) embryos were stained for β -Gal activity. Anterior is to the left in A and C and to the right in B and D. White arrowheads delineate the boundaries of one mesodermal segment (sclerotome).(A) Expression in embryos of line L35 is seen throughout the sclerotome in the thoracic (A) and the caudal region (C). Expression directed by the m61acZ4 transgene is seen in a subset of sclerotomal cells in the thoracic (B, embryo of line L30) as well as in the caudal region (D, embryo of line L53). Also expression in the rib primordia is more limited in embryos expressing the m61acZ4 construct.

Change of the expression pattern during development

Expression of m61acZ4 and mólacZll was analysed in detail in transgenic lines (mólacZ4: L30 and L53, mólacZll: L35 and L38) between days 8 and 12 p.c. (Fig. 7 and data not shown). The pattern directed by the two constructs was indistinguishable from day 8 to day 10. Expression is first detectable at day 8·5 in the posterior part of the neural plate (Fig. 7A). In contrast to mólacZl no expression is detectable in the allantois. Subsequently transgene activity is detectable both in the presomitic and the intermediate mesoderm (day 9·0). At day 9·5, the transgenes exhibit the same domains of expression as mólacZl in neural tube, neural crest and throughout the mesoderm posterior to the level of the prospective third thoracic prevertebrae (the 14th somite; Fig. 7B). Thus, with the exception of the allantoic expression, the patterns directed by the mólacZl, mólacZ4 and mólacZll transgenes are identical up to this stage. Beginning at day 10·5p.c., the posterior limit of transgene expression is visible in neural tube and in the somitic mesoderm posterior to the hindlimb (Fig. 7C). However, some expression remains in the tail bud. This pattern is maintained as new segments are added during the following two days of development. At day 11, there is a marked difference between the patterns of m61acZ4 and mólacZll. mólacZll activity is found throughout the sclerotome at day 11 and both in prevertebrae and intervertebral disc anlagen at day 12 (Fig. 6D-F). In contrast, m61acZ4 is active at day 11 in approximately the middle third of the sclerotome At day 12, staining is seen only in prevertebrae in E297, L30 and L53 (Fig. 6G-I). Thus mesodermal expression of m61acZ4 is restricted to the lineage giving rise to prevertebrae (Fig. 8). Expression of both transgenes in the neural tube is detectable in the ventral half, whereas expression of mólacZl is expressed almost along the entire height of the tube (Fig. 9), making element B a potential target for inductive signals from the notochord and/or the floorplate which pattern the architecture of the neural tube (Placzek et al. 1990).

Fig. 6.

Lineage restriction in the mesoderm. (A – C) Expression in the mesoderm is not appropriately restricted in embryos expressing transgenes that lack element B (mólacZl: A, C and m61acZ9: B). (A) Cross section of a day 11 embryo of line L4. Expression is seen in dermamyotome, sclerotome and the surrounding mesoderm. (B) Sagittal section of the caudal region (m61acZ9 E325). (C) Sagittal section of a day 12 embryo from line L4. (D – F) Element B restricts expression in the mesoderm to the sclerotome. (D) Cross section of the sacral region (mólacZll E347) (E) Sagittal section of the caudal region (mólacZll: E347). (F) Sagittal section of the thoracic region (day 11 embryo of line L38). Expression of the transgene in the mesoderm is seen only in the sclerotome. As the sclerotome forms dense and loose regions in the developmg vertebral column which will give rise to the intervertebral disc anlagen and the prevertebrae, transgene expression is maintained in both populations. This expression is also seen at day 12 (not shown). (G-I) Elements B and C restrict expression to a subpopulation of sclerotomal cells. (G) Cross section from the caudal region of a day 11 embryo (line L53) (H) Sagittal section from the caudal region of a day 12 embryo (hne L53) (I) Cross section of a day 12 embryo (mólacZ4: E 297). Expression in the sclerotome is restncted to a subpopulation of cells in the middle third of a segment (G, H) Due to the fusion of the ß-Gal protein to the homeobox, nuclear staining is observed. (I) Staining is seen in the prevertebrae. Embryos were fixed and stained after the dissection and prepared for paraffin sectioning as described in Materials and methods. (C) A cryosection that was stained after sectioning. Blue stain indicates transgene activity The sections were stained with neutral red. dm, dermamyotome; ivd, intervertebral disc; isv, intersegmental vessie; nt, neural tube; pv, prevertebrae, sc, sclerotome. The black arrowheads in B, E, H delineate the boundaries of one mesodermal segment.

Fig. 6.

Lineage restriction in the mesoderm. (A – C) Expression in the mesoderm is not appropriately restricted in embryos expressing transgenes that lack element B (mólacZl: A, C and m61acZ9: B). (A) Cross section of a day 11 embryo of line L4. Expression is seen in dermamyotome, sclerotome and the surrounding mesoderm. (B) Sagittal section of the caudal region (m61acZ9 E325). (C) Sagittal section of a day 12 embryo from line L4. (D – F) Element B restricts expression in the mesoderm to the sclerotome. (D) Cross section of the sacral region (mólacZll E347) (E) Sagittal section of the caudal region (mólacZll: E347). (F) Sagittal section of the thoracic region (day 11 embryo of line L38). Expression of the transgene in the mesoderm is seen only in the sclerotome. As the sclerotome forms dense and loose regions in the developmg vertebral column which will give rise to the intervertebral disc anlagen and the prevertebrae, transgene expression is maintained in both populations. This expression is also seen at day 12 (not shown). (G-I) Elements B and C restrict expression to a subpopulation of sclerotomal cells. (G) Cross section from the caudal region of a day 11 embryo (line L53) (H) Sagittal section from the caudal region of a day 12 embryo (hne L53) (I) Cross section of a day 12 embryo (mólacZ4: E 297). Expression in the sclerotome is restncted to a subpopulation of cells in the middle third of a segment (G, H) Due to the fusion of the ß-Gal protein to the homeobox, nuclear staining is observed. (I) Staining is seen in the prevertebrae. Embryos were fixed and stained after the dissection and prepared for paraffin sectioning as described in Materials and methods. (C) A cryosection that was stained after sectioning. Blue stain indicates transgene activity The sections were stained with neutral red. dm, dermamyotome; ivd, intervertebral disc; isv, intersegmental vessie; nt, neural tube; pv, prevertebrae, sc, sclerotome. The black arrowheads in B, E, H delineate the boundaries of one mesodermal segment.

Fig. 7. Development of the Hox-1·1 expression pattern. Transgemc embryos (m61acZ4: line L30) of different stages were stained for ß-Gal activity (A) Day 8·5. (B) day 9·5 (C) day 10 5. (D) day 11·5.

Fig. 8.

Different elements are required at different times of development for lineage restriction in the mesoderm (A) Schematic representation of expression of the endogenous Hox-1·1 gene and different transgenes in the lineage giving nse to prevertebrae Stippled boxes indicate that the gene is expressed. (B) The regulatory elements are required at different times of development. The bars indicate at what time of development elements A+P, B and C (see Fig 10) are necessary to obtain spatially appropriate expression.

Fig. 8.

Different elements are required at different times of development for lineage restriction in the mesoderm (A) Schematic representation of expression of the endogenous Hox-1·1 gene and different transgenes in the lineage giving nse to prevertebrae Stippled boxes indicate that the gene is expressed. (B) The regulatory elements are required at different times of development. The bars indicate at what time of development elements A+P, B and C (see Fig 10) are necessary to obtain spatially appropriate expression.

Fig. 9. Dorsoventral restriction in the neural tube. (A-C) Cross sections of day 11 embryos from the sacral region are shown mólacZl is expressed throughout the neural tube (nt) (A: embryo of line L4). mólacZll (B: hne L38) and mólacZ4 (C: hne L53) are expressed in the ventral half of the neural tube. The more limited staining in (C) is due to the nuclear localisation of the β -Gal-fusion protein. The stained embryos were embedded in paraffin, sectioned and counter-stained with neutral red

m6lacZ4 expression at day 12

m61acZ4 expression in day 12·5 embryos was analysed in detail by serial sectioning of paraffin-embedded embryos. Expression is detectable in spinal ganglia and the adjacent neural tube from C5 to S4. In prevertebrae, a high level of expression is visible in prevertebrae T3 to T13. However, a low level of expression extends to the sacral region (Fig. 3B) and decreases gradually. Expression in T3 is very low, which is consistent with the in situ hybridization data obtained with Hox-1·1 probes (Puschel et al. 1990; Kessel, unpublished results). The expression can be seen in a small subpopulation of sclerotomal cells that contribute to the prevertebrae (Fig. 6G–1). The posterior boundary of transgene expression is more caudal than has been described for Hox-1·1 This may be due to the lower sensitivity of the in situ hybridisation method. To confirm the authenticity of transgene expression in the tail, a RNAase-mapping with a Hox-1·1 homeobox probe was performed with RNA isolated from tails and hindlimbs prepared from day 12 embryos. In both structures, Hox-1·1 RNA was clearly detectable, whereas RNA from heads did not contain Hox-1·1 message (data not shown). In addition the transgene is expressed in three other structures where Hox-1·1 expression was not reported so far and was probably missed by in situ hybridisation with RNA probes. Both the apical ectodermal ridge of the hindlimb and the ventral ridge of the tail express m61acZ4 (Figs 3B, 7D). In addition expression is seen in the dorsal aorta.

Table 2.

Expression in embryos of transgenic lines

Expression in embryos of transgenic lines
Expression in embryos of transgenic lines

Correct regulation does not require the context of the Hox-1 cluster

Using the m61acZ4 construct, we could reproduce fully the embryonic regulation of Hox-1·1 in transgenic mice. The context of the Hox-1 cluster was not necessary to reproduce qualitatively the expression pattern. However, we did not obtain correct regulation in terms of integration-site independence and copy-number dependence of expression levels. For this aspect of regulation, elements such as a locus control region (LCR) might be necessary as is the case for the β-globin cluster and the Thy-1 gene (Behringer et al. 1990; Greaves et al. 1989; Grosveld et al. 1987; for a review see: Orkin, 1990). Competition of genes for interaction with the LCR in the β-globin cluster is a prerequisite for correct regulation in the context of a gene cluster (Enver et al. 1990). The remarkable conservation of gene order between the four Hox clusters suggests that the context of the cluster might be important for this particular regulation of Hox gene expression (Gaunt and Singh, 1990). Furthermore, there is a clear correlation between the position of a gene within a Hox cluster and its anterior boundary of expression (Graham et al 1989; Duboule and Dolle, 1989). The more 5’ a gene is located the more caudal is its anterior limit of expression in neural tube and prevertebrae. The spatial order of expression boundaries reflects the temporal order of activation in 3’-5’ direction during gastrulation Thus, the gene order both in the βglobin cluster and in the Hox clusters reflects the order of activation during development, which might indicate a common regulatory mechanism. Our observation that the temporal and spatial expression pattern can also be attained outside of the cluster clearly indicates that individual Hox genes carry enough control elements in the vicinity of the gene to exert this control. Therefore, similar to the globin cluster control, a LCR may be required but would represent a different class of regulatory elements.

At least four elements are necessary for positionspecific expression

Previously we have shown that 3·6kbp Hox-1·1 promoter sequences can direct expression of a marker gene to a specific region of an embryo with a discrete anterior boundary. Here we show that 1·7 kbp promoter sequences are sufficient to specify the same expression pattern. Deletion of another 110bp (element A in lacZlB, Fig. 10) results in a loss of the region-specific activity and a uniform expression throughout the embryo, indicating that this element is involved in specifying the anterior boundary of Hox-1·1 expression. Furthermore, this element is required for a correct pattern both in meso- and ectoderm. In one transgenic line (L62), the m6lacZlB expression pattern was identical to that of m6lacZl indicating that element A is not the only regulatory element responsible for establishing the anterior boundary. L62 was the only m6lacZlB line that showed this pattern. As the frequency of transgenic lines expressing the m6lacZlB construct was significantly reduced compared to other constructs, we think that element A could function to ensure high levels of expression in addition to being involved in specifying the anterior border of expression. A result similar to the one described above was obtained when the m6lacZlB was used to express lacZ transiently in transgenic zebrafish (Westerfield et al. in preparation). In general, the same elements that are required to restrict expression from the murine Hox-1·1 promoter in mice are also necessary in transgenic fish. In particular, deletion of element A resulted in uniform expression of lacZ activity in the fish embryo (Westerfield et al. in preparation).

Fig. 10.

Three different regulatory elements restrict the activity of the Hox-1 1 promoter. The genomic structure of the Hox-1·1 gene and the location of the identified regulatory elements A, B, C and P are shown in the upper half. Element A (open box), a 110bp Alul (A)-BamHI (B) fragment; P, the 1590bp promoter extending from BûWHI to SocI (Sc), Element B (stippled box), a 130bp fragment extending from the SucI site to the Hox-1·1 start codon (ATG). Element C (hatched box), a 1341bp PstI (Ps)-PvuII (Pv) fragment The numbers indicate kbp of DNA Transcribed sequences are shown as boxes, coding sequences as stippled boxes. Bg, Bglll site; E, EcoRI site, SI, Sall site; St, Stul site, X, Xbal site. The lower half shows a schematic representation of the patterns that are obtained with different transgenes. P, m6lacZlB; PA, m6lacZl; PAB, m6lacZll, PABC, m61acZ4. Stepwise addition of A, B and C restricts expression of the promoter P to ever smaller sets of cells.

Fig. 10.

Three different regulatory elements restrict the activity of the Hox-1 1 promoter. The genomic structure of the Hox-1·1 gene and the location of the identified regulatory elements A, B, C and P are shown in the upper half. Element A (open box), a 110bp Alul (A)-BamHI (B) fragment; P, the 1590bp promoter extending from BûWHI to SocI (Sc), Element B (stippled box), a 130bp fragment extending from the SucI site to the Hox-1·1 start codon (ATG). Element C (hatched box), a 1341bp PstI (Ps)-PvuII (Pv) fragment The numbers indicate kbp of DNA Transcribed sequences are shown as boxes, coding sequences as stippled boxes. Bg, Bglll site; E, EcoRI site, SI, Sall site; St, Stul site, X, Xbal site. The lower half shows a schematic representation of the patterns that are obtained with different transgenes. P, m6lacZlB; PA, m6lacZl; PAB, m6lacZll, PABC, m61acZ4. Stepwise addition of A, B and C restricts expression of the promoter P to ever smaller sets of cells.

Addition of the 5’ nontranslated sequence (element B in Fig. 10) restricts expression driven by the promoter A+P in three aspects. Element B is necessary to set a posterior boundary both in mesoderm and ectoderm (anteroposterior restriction), to confine expression to sclerotomal cells (lineage restriction) and to restrict expression to the ventral half in the neural tube (dorsoventral restriction). At the moment, it cannot be decided if this element acts at the transcriptional level or by influencing mRNA stability. It also contains several conserved short open reading frames for which a function as a translational control element has been suggested (Bürglin et al. 1987; Kessel and Gruss, 1988).

A third element (element C in Fig. 10) is required to confine expression to prevertebrae. Element C contains a 100bp sequence which upon deletion results in a stabilisation of Hox-1·1 mRNA in F9 stem cells (Colberg-Poley et al. 1987; A.W.P., unpublished results). It is not clear if the same sequence is essential for activity of element C. The three elements A, B and C are required at different times during development to restrict expression of the Hox-1·1 promoter P (Figs 8, 10), reflecting different developmental decisions. We would expect that this mechanism of pattern formation is general for the Hox genes.

A preliminary analysis suggests a similar mechanism for regulation of the Hox-3·1 gene (Biebench et al. 1990). The proposed mechanism, however, differs significantly from those postulated for the Hox-1 3 and HOX-5·1 genes (Zakany et al. 1988; Tuggle et al. 1990). These authors suggested that the additive effect of region-specific enhancers might be responsible for the specific expression patterns of these genes. This would be similar to the stripe-specific elements of several Drosophila segmentation genes (for a review see: Pankratz and Jackle, 1990). The transgenes used in these experiments, however, gave rise to only a very limited aspect of the endogenous expression pattern of the corresponding Hox genes. In particular, transgene activity was first seen considerably later than the onset of expression of the endogenous genes. The constructs employed may therefore not contain the major regulatory elements but sequences that are required after establishment of the pattern.

Two elements restrict Hox-1·1 expression to the prevertebral cell lineage

The axial structures of vertebrates are derived from mesodermal cells. After formation of somites from the presomitic mesoderm, cells at the ventromedial edge of the somites disperse and form the sclerotome, whereas cells from the dorsolateral part of the somites give rise to the dermamyotome. The sclerotomal cells will migrate towards the notochord and form prevertebrae and intervertebral discs. The precise cell lineage relationship between sclerotome and prevertebrae is still unknown. After formation of the sclerotome, a further differentiation into a rostral and caudal half can be observed (Verbout, 1985). This observation has been interpreted as a resegmentation process that results in a half-segment phase-shift during vertebral column formation (Remak, 1855). According to this view, the rostral half of one sclerotome and the caudal half of the preceding sclerotome fuse to form one vertebrae, but the experimental evidence addressing this question is contradictory.

Expression of the m61acZ4 transgene is found in a subpopulation of cells in the central part of the sclerotome. It is possible that these are the progenitor cells for the prevertebrae and that the non-expressing cells will give rise to the intervertebral discs. The expression of m61acZ4 in the middle third of a mesodermal segment is consistent with a recent celllineage analysis of the sclerotomal cells. Sclerotomal cells derived from one somite contribute to only one hemivertebra and the adjacent halves of the intervertebral discs in chicken (C. Stem, personal communication). The expression pattern of Hox-11 and m61acZ4 in the sclerotome and the cell-lineage study argue against the occurrence of resegmentation.

The identification of several position-responsive sequences now enables the identification of the transacting factors conferring positional information. So far the analysis of cw-acting elements has shown that several mechanisms act consecutively at different times during development to direct expression of a gene to a specific structure and a distinct region relative to the anteroposterior axis. Further dissection of the promoter should reveal additional elements involved in specifying the anterior boundary. In addition, it will be interesting to try to separate elements acting specifically in ectoderm or mesoderm.

The authors would like to thank Merve Olowson and Kirsten Schaub for excellent technical assistance and Ralf Altschaffel for photographic work. Oligonucleotides were synthesised by Hans-Peter Geithe. We thank all members of the lab and especially Michael Kessel, Martyn Goulding, Vasanta Subramaman and Urban Deutsch for constructive comments on the manuscript. This work was supported by the Max-Planck Gesellschaft A W.P. and part of the work were supported by the Bundesministerium für Forschung und Technologie (BMFT).

Akam
,
M
(
1987
)
The molecular basis for metamenc pattern in the Drosophila embryo
Development
101
,
1
22
Allen
,
N D
,
Cran
,
D G
,
Barton
,
S C
,
Hettle
,
S.
,
Reik
,
W
and
Surani
,
M A
(
1988
)
Transgenes as probes for active chromosomal domains m mouse development
Nature
333
,
852
855
Balling
,
R
,
Deutsch
,
U
and
Gruss
,
P
(
1988
)
Undulated, a mutation affecting the development of the mouse skeleton, has a point mutation in the paired box of Pax 1
Cell
55
,
531
535
Balling
,
R
,
Mutter
,
G
,
Gruss
,
P
and
Kessel
,
M
(
1989
)
Craniofacial abnormalities induced by ectopic expression of the homeobox gene Hox 1 1 in transgenic mice
Cell
58
,
337
347
Behringer
,
R R
,
Ryan
,
T M
,
Palmiter
,
R D
,
Brinster
,
R L
and
Townes
,
T M
(
1990
)
Human y- to /kglobin gene switching in transgenic mice
Genes and Dev
4
,
380
389
Bieberich
,
C J
,
Utset
,
M
,
Awgulewitsch
,
A
and
Ruddle
,
F H
(
1990
)
Evidence for positive and negative regulation of the Hox-3 1 gene
Proc natn Acad Set USA
87
,
8462
8466
Boncinelli
,
E
,
Somma
,
R
,
Acampora
,
D
,
Pannese
,
M.
,
D’esposito
,
M
,
Faiella
,
A.
and
Simeone
,
A
(
1988
)
Organization of human homeobox genes
Human Reproduction
3
,
880
886
Bürglin
,
T. R.
,
Wright
,
C V E
and
Derobertis
,
E M
(
1987
)
Translational control in homeobox mRNAs9
Nature
330
,
701
702
.
Colberg-Poley
,
A M
,
Püschel
,
A W
,
Dony
,
C
,
Voss
,
S D
and
Gruss
,
P
(
1987
)
Post-transcnptional regulation of a munne homeobox gene transcript in F9 embryonal carcinoma cells
Differentiation
35
,
206
211
Colberg-Poley
,
A M
,
Voss
,
S D
,
Chowdhury
,
K
,
Stewart
,
C. L
and
Gruss
,
P
(
1985
)
Clustered horneo boxes are differentially expressed during munne development
Cell
43
,
39
45
Dressler
,
G R
and
Gruss
,
P
(
1988
)
Do multigene families regulate vertebrate development9
Trends Genet
4
,
214
219
Dressler
,
G R
and
Gruss
,
P
(
1989
)
Antenor boundanes of Hox gene expression in mesoderm denved structures correlate with the linear gene order along the chromosome
Differentiation
.
41
,
193
201
Duboule
,
D
and
Dollé
,
P
(
1989
)
The structural and functional organisation of the munne HOX gene family resembles that of Drosophila homeotic genes
EMBO J
8
,
1497
1505
Enver
,
T
,
Raich
,
N
,
Ebens
,
A J
,
Papayannopoulou Costantini
,
F
and
Stamatoyannopoulos
,
G
(
1990
)
Developmental regulation of human fetal-to-adult globin gene switching in transgenic mice
Nature
344
,
309
313
Gaunt
,
S J
and
Singh
,
P. B.
(
1990
)
Homeogene expression patterns and chromosomal imprinting
Trends Genet
6
,
208
212
Gehring
,
W J
and
Hiromi
,
Y
(
1986
)
Homeotic genes and the homeobox
Ann Rev Genet
20
,
147
173
Graham
,
A
,
Papalopulu
,
N
and
Krumlauf
,
R
(
1989
)
The munne and Drosophila homeobox gene complexes have common features of organisation and expression
Cell
57
,
367
378
Greaves
,
D. R
,
Wilson
,
F D
,
Lang
,
G
and
Kioussis
,
D
(
1989
).
Human CD2 3’-flanking sequences confer high-level, T cell-specific, position-independent gene expression in transgenic mice
.
Ceti
56
,
979
986
.
Grosveld
,
F.
,
Van Assendelft
,
G. B.
,
Greaves
,
D. R.
and
Kollias
,
G.
(
1987
).
Position-independent, high-level expression of the human /Fglobin gene in transgenic mice
.
Cell
51
,
975
985
.
Hall
,
M. N.
,
Craik
,
C.
and
Hiraoka
,
Y.
(
1990
).
Homeodomain of yeast a2 contains a nuclear localization signal
.
Proc. natn. Acad. Sei. U.S.A
.
87
,
6954
6958
.
Harvey
,
R. P.
and
Melton
,
D. A.
(
1988
).
Microinjection of synthetic Xhox-1A homeobox mRNA disrupts somite formation in developing Xenopus embryos
.
Cell
53
,
687
697
.
Hogan
,
B.
,
Constantini
,
F.
and
Lacy
,
E.
(
1986
).
Manipulating the Mouse Embryo. Cold Spring Harbor Laboratory
.
Holland
,
P. W. H.
and
Hogan
,
B.
(
1988
).
Expression of horneo box genes during mouse development: a Review
.
Genes and Dev
.
2
,
773
782
.
Ingham
,
P. W.
(
1988
).
The molecular genetics of embryonic pattern formation in Drosophila
.
Nature
335
,
25
34
.
Kessel
,
M.
,
Balling
,
R.
and
Gruss
,
P.
(
1990
).
Variations of cervical vertebrae after expression of a Hox-1.1 transgene in mice
.
Cell
65
,
301
308
.
Kessel
,
M.
and
Gruss
,
P.
(
1988
).
Open reading frames and translational control
.
Nature
332
,
117
118
.
Kessel
,
M.
and
Gruss
,
P.
(
1990
).
Murine developmental control genes
.
Science
249
,
374
379
.
Kessel
,
M.
,
Schulze
,
F.
,
Fibi
,
M.
and
Gruss
,
P.
(
1987
).
Primary structure and nuclear localization of a murine homeodomain protein
.
Proc. natn. Acad. Sei. U.S.A
.
84
,
5306
5310
.
Mahon
,
K. A.
,
Westphal
,
H.
and
Gruss
,
P.
(
1988
).
Expression of homeobox gene Hoxl.l during mouse embryogenesis
.
Development
104
Supplement
,
187
-
195
.
Orkin
,
S. H.
(
1990
).
Globin gene regulation and switching; circa 1990
.
Cell
63
,
665
672
.
Pankratz
,
M. J.
and
JÄckle
,
H.
(
1990
).
Making stripes in the Drosophila embryo
.
Trends Genet
.
6
,
287
292
.
Placzek
,
M.
,
Tessier-Lavigne
,
M.
,
Yamada
,
T.
,
Jessel
,
T.
and
Dodd
,
J.
(
1990
).
Mesodermal control of neural cell identity: floor plate induction by the notochord
.
Science
250
,
985
988
.
Püschel
,
A. W.
,
Balling
,
R.
and
Gruss
,
P.
(
1990
).
Positionspecific activity of the Hox-1.1 promoter in transgenic mice
.
Development
108
,
435
442
.
Remak
,
R.
(
1855
).
Untersuchungen über die Entwicklung der Wirbelthiere (Berlin: Reimer)
.
Rüther
,
U.
and
MüLler-Hill
(
1983
).
Easy identification of cDNA clones
.
EMBO J
.
2
,
1791
1794
.
Ruiz I Altaba
,
A.
and
Melton
,
D.
(
1989
).
Involvement of the Xenopus homeobox gene Xhox3 in pattern formation along the anterior-posterior axis
.
Cell
57
,
317
326
.
Theiler
,
K.
(
1989
).
The House Mouse
(2nd printing). Springer: New York, Berlin, Heidelberg, London, Paris, Tokyo.
Tuggle
,
C. K.
,
Zakany
,
J.
,
Cianetti
,
L.
,
Peschle
,
C.
and
Nguyen-Huu
,
M. C.
(
1990
).
Region-specific enhancers near two mammalian horneo box genes define adjacent rostrocaudal domains in the central nervous system
.
Genes and Dev
.
4
,
180
189
.
Verbout
,
A. J.
(
1985
).
The development of the vertebral column. Advances in Anatomy
,
Embryology and Cell Bioliology
90
,
1
122
.
Wright
,
C. V. E.
,
Cho
,
K. W. Y.
,
Oliver
,
G.
and
De Robertis
,
E. M.
(
1989
).
Vertebrate homeodomain proteins: families of region-specific transcription factors
.
Trends in Biochem. Sei
.
14
,
52
56
.
Zakany
,
J.
,
Tuggle
,
C. K.
,
Patel
,
M. D.
and
Nguyen-Huu
,
M. C.
(
1988
).
Spatial regulation of homeobox gene fusions in the embryonic central nervous system of transgenic mice
.
Neuron
1
,
679
691
.