Using gene targeting, we have created mice with a disruption in the homeobox-containing gene hoxd-11. Homozygous mutants are viable and the only outwardly apparent abnormality is male infertility. Skeletons of mutant mice show a homeotic transformation that repatterns the sacrum such that each vertebra adopts the structure of the next most anterior vertebra. Defects are also seen in the bones of the limb, including regional malformations at the distal end of the forelimb affecting the length and structure of phalanges and metacarpals, inappropriate fusions between wrist bones, and defects at the most distal end in the long bones of the radius and ulna. The phenotypes show both incomplete penetrance and variable expressivity. In contrast to the defects observed in the vertebral column, the phenotypes in the appendicular skeleton do not resemble homeotic transformations, but rather regional malformations in the shapes, length and segmentation of bones. Our results are discussed in the context of two other recent gene targeting studies involving the paralogous gene hoxa-11 and another member of the Hox D locus, hoxd-13. The position of these limb deformities reflects the temporal and structural colinearity of the Hox genes, such that inactivation of 3′ genes has a more proximal phenotypic boundary (affecting both the zeugopod and autopod of the limb) than that of the more 5′ genes (affecting only the autopod). Taken together, these observations suggest an important role for Hox genes in controlling localized growth of those cells that contribute to forming the appendicular skeleton.

In vertebrates there is a matrix of 38 genes that collectively make up the Hox complex. These genes are distributed in four linkage groups, Hox A, Hox B, Hox C and Hox D, which are located on four separate chromosomes. This organization is believed to have arisen during prevertebrate phylogeny by quadruplication of an ancestral complex to create the current linkage groups, making genes in one locus homologous to their corresponding parallel genes in the other linkage groups (Pendleton et al., 1993). Such homologous genes are called paralogues.

A Hox gene product is defined by the presence of a highly conserved 61 amino acid motif called the homeodomain that has DNA-binding properties (Kissinger et al., 1990; Otting et al., 1990). In Drosophila, the homeodomain-encoding genes of the Bithorax and Antennapedia homeotic complexes (HOM-C) are used to pattern the developing embryo (Lewis, 1978; Akam, 1987). The vertebrate Hox genes are similar to those in the HOM-C in three regards. First, sequence similarities in the homeodomains allow for the alignment of mammalian Hox genes with their counterpart Drosophila genes (detailed in McGinnis and Krumlauf, 1992). Second, several mammalian Hox genes have been shown to function and phenocopy their Drosophila HOM-C homologues in vivo when expressed ectopically in transgenic flies (McGinnis et al., 1990). Third, the chromosomal position of the Hox genes reflects the order of the anterior boundaries of expression for each gene along the anteroposterior (A-P) axis. The 3′ genes extend more anteriorly while those on the 5′ end of the complex are limited to more posterior regions (Duboule and Dollé, 1989; Graham et al., 1989). This latter property has been termed structural colinearity.

Metamere formation along the primary axis of the vertebrate embryo is intrinsically coordinated with time since the somites that generate this pattern condense and mature in an anteriorto-posterior direction. Consequently, the order of Hox genes on the chromosome also correlates with their time of activation, with 3′ members being activated before their 5′ neighbors (Izpisúa-Belmonte et al., 1991a). This temporal progression of Hox gene activation may play as important a role as the spatial distribution of Hox gene expression in executing the developmental program for patterning the mammalian embryo.

The structural and temporal colinearity of Hox gene expression suggested that the encoded proteins may determine regional identity along the body axis of the embryo. This hypothesis has been amply confirmed by the analysis of mice with targeted disruptions in Hox genes. These mutant mice show either regional loss of tissues and structures, or apparent localized homeotic transformations along the A-P axis (Chisaka and Capecchi, 1991; Lufkin et al., 1991; Chisaka et al., 1992; LeMouellic et al., 1992; Ramirez-Solis et al., 1993; Carpenter et al., 1993; Mark et al., 1993; Condie and Capecchi, 1993; Jeannotte et al., 1993; Dollé et al., 1993; Small and Potter, 1993; Gendron-Maguire et al., 1993; Rijli et al., 1993; Kostic and Capecchi, 1994).

In addition to patterning the embryo along the A-P axis, a subset of 15 or more Hox genes located at the 5′ end of the Hox A, Hox C and Hox D linkage groups are also likely to be involved in patterning the limb. Interestingly, during the early phases of limb bud growth, the spatial and temporal expression patterns of these Hox genes within the limb also reflect the order of the genes on the chromosome (Dollé et al., 1989, 1991a; Nohno et al., 1991; Izpisúa-Belmonte et al., 1991a,b; Yokouchi et al., 1991; Haack and Gruss, 1993). These 5′ Hox genes express within the developing limb a nested set of transcripts which have been suggested to establish polarity and patterning, as well as influencing the number of digits (Duboule, 1992; Izpisúa-Belmonte and Duboule, 1992; Tabin, 1992).

The vertebrate limb initially appears as a bud of mesenchyme that originates from the lateral plate mesoderm. The outgrowth of the limb bud is controlled by the apical ectodermal ridge (AER), which maintains the underlying mesoderm (the progress zone) in a highly proliferative state. As the limb grows, cells leave the progress zone and participate in forming the prechondrogenic condensations that generate the limb bones (Hinchliffe, 1991). As a result, patterning of the limb is polarized such that the proximal bones (the stylopod and zeugopod) are laid down prior to the distal element (the autopod). An important component of patterning the limb along the A-P axis is a signal emanating from the zone of polarizing activity (ZPA) located at the posterior proximal margin of the limb bud (Tickle et al., 1975; Saunders, 1977). Although the nature of this signal and how it is distributed along the A-P axis of the limb has not been defined, there is current excitement that sonic hedgehog might be the long sought signaling molecule (Echelard et al., 1993; Riddle et al., 1993). Because the anterior expression boundaries of the 5′-Hox D genes vary along the A-P axis of the developing limb bud, it has been postulated that these genes may be involved in interpreting the ZPA signal. Consistent with this hypothesis, ZPA transplantations to the anterior side of the chick limb bud generate mirrorimage duplications of the 5′-Hox D expression patterns that precede and resemble the characteristic mirror-image limb duplications that are induced by the ectopic ZPA (Izpisúa-Belmonte et al., 1991b; Nohno et al., 1991).

Hoxd-11 (formerly hox-4.6) is situated in the center of the 5′-set of Hox D genes. It is expressed both along the primary body axis and in the developing limb. The earliest expression is noted at day E8.75 in lateral plate mesoderm of the most posterior region of the embryo. At E12.5 the expression of hoxd-11 reaches an anterior border along the A-P axis in the sclerotome condensations of prevertebra 25, near the junction between the lumbar and sacral vertebrae (Izpisúa-Belmonte et al., 1991a). Hoxd-11 expression is also seen in the developing urogenital tracts (Dollé et al., 1991b). In the limb, expression of hoxd-11 is first noted at E9.0. With the outgrowth of the limb, hoxd-11 transcripts contribute to the nested set of Hox D transcripts that vary along both the proximodistal (Pr-D) and A-P axes (Dollé et al., 1989, 1991a).

While RNA in situ studies suggest possible sites of gene function, a more direct approach to defining the developmen-tal roles for Hox proteins is to employ the technique of gene targeting to create specific mutations in the mouse (Capecchi, 1989, 1994). Towards this end, we have created mice homozygous for a mutation in the hoxd-11 gene. These mice have a homeotic transformation encompassing the entire sacral region of the axial column as well as regional malformations at the distal end of the forelimb. Our results are compared with the recent targeted analysis of two other Hox genes that also affect patterning of the sacrum and the limb: hoxa-11, a paralogue of hoxd-11 (Small and Potter, 1993); and hoxd-13, another member of the 5′ Hox D locus (Dollé et al., 1993).

Construction of a hoxd-11 targeting vector

A DNA fragment containing the hoxd-11 gene was isolated from a genomic lambda library made from mouse CC1.2 embryo-derived stem (ES) cell DNA, using a probe 5′ to the hoxd-10 locus. The hoxd-11 gene was identified by hybridization to a 45-mer oligonucleotide from the human homologue, HOX4F (Acampora et al., 1989), and the homeobox was sequenced to confirm its identity (Izpisúa-Belmonte et al., 1991a). A replacement type targeting vector (Thomas and Capecchi, 1987; Thomas et al., 1992; Deng et al., 1993) carrying 9.5 kilobases (kb) of genomic DNA containing hoxd-11 and flanked by two herpes simplex virus thymidine kinase genes was constructed in a Bluescript-based plasmid. KT3NP4, a 3.1 kb neomycin (neo) resistance cassette driven by the RNA polymerase II promoter, was inserted into a Bst1107I site in the hoxd-11 homeobox (Fig. 1A). This site corresponds to amino acid 23 of the homeodomain so that the insertion disrupts the coding sequence at the end of the first helix of the DNA-binding motif. Since the KT3NP4 neo cassette contains a very efficient poly(A) addition signal cloned from the hprt gene (Deng et al., 1993), hoxd-11 transcripts 3′ of the neo insertion should not be synthesized. In the numerous other cases in which we have used RT-PCR to detect the presence of such transcripts, none were found. Therefore, this targeting vector should generate a loss-of-function mutation with respect to DNA binding. The potential exists for an NH2-terminal fragment of hoxd-11 to be synthesized and to retain some biological activity aside from DNA binding (Zappavigna et al., 1994). However, such a polypeptide fragment, if synthesized, is likely to be rapidly degraded relative to the intact, normal hoxd-11 protein (Capecchi et al., 1974; Rechsteiner, 1987). Thus, whether the concentration of such a polypeptide fragment within the cells of the embryo could reach the critical concentration necessary to elicit a bio-logical response is questionable.

Fig. 1.

Gene targeting of hoxd-11 and genotype analysis of cell lines and mice. (A) Diagram of the targeting vector and anticipated restriction fragment lengths resulting from gene targeting. The upper line is the targeting vector. The black box indicates the homeobox of hoxd-11, the white box is the disrupting neo cassette and the grey boxes on the ends are the thymidine kinase genes. The thick line underneath is the genomic locus with restriction enzyme sites marked. Ligation of the neo cassette into the homeobox of hoxd-11 destroys an endogenous Bst 1107 site in the gene. The expected restriction fragment lengths (in kb) for the wild-type and the targeted alleles are shown beneath for the two flanking probes D (5′ probe) and I (3′ probe). (B) Southern blot analysis of parental CC1.2 ES cells and the isolated cell line 3f-15. DNA from each cell line was digested with SacI and hybridized with probe D or digested with Bst1107 I and hybridized with probe I to demonstrate a gene targeting event. (C) Southern blot analysis of progeny mice resulting from hoxd-11 heterozygote intercrosses. Tail DNA from each pup was digested with SacI and hybridized with probe D to allow genotyping of the animals (indicated above each lane). B=Bst1107 I, S=Sac I.

Fig. 1.

Gene targeting of hoxd-11 and genotype analysis of cell lines and mice. (A) Diagram of the targeting vector and anticipated restriction fragment lengths resulting from gene targeting. The upper line is the targeting vector. The black box indicates the homeobox of hoxd-11, the white box is the disrupting neo cassette and the grey boxes on the ends are the thymidine kinase genes. The thick line underneath is the genomic locus with restriction enzyme sites marked. Ligation of the neo cassette into the homeobox of hoxd-11 destroys an endogenous Bst 1107 site in the gene. The expected restriction fragment lengths (in kb) for the wild-type and the targeted alleles are shown beneath for the two flanking probes D (5′ probe) and I (3′ probe). (B) Southern blot analysis of parental CC1.2 ES cells and the isolated cell line 3f-15. DNA from each cell line was digested with SacI and hybridized with probe D or digested with Bst1107 I and hybridized with probe I to demonstrate a gene targeting event. (C) Southern blot analysis of progeny mice resulting from hoxd-11 heterozygote intercrosses. Tail DNA from each pup was digested with SacI and hybridized with probe D to allow genotyping of the animals (indicated above each lane). B=Bst1107 I, S=Sac I.

Electroporation and ES cell line analysis

The targeting vector was linearized and electroporated into CC1.2 ES cells (Deng and Capecchi, 1992). The cells were grown in media containing the drugs G418 and FIAU to enrich for cells that had undergone a homologous recombination event (Mansour et al., 1988). Cell colonies that survived selection were picked and their DNA extracted as described (Condie and Capecchi, 1993). A primary screen of the cell lines was performed by Southern blot analysis of SacI-digested genomic DNA, using the 1.0 kb SalI-XhoI 5′ flanking probe D immediately adjacent to the targeting vector DNA (Fig. 1A). The expected shift in band size is from 13.9 kb (wild-type band) to 6.2 kb (mutant band). Targeted cell lines were confirmed by Southern blot analysis using the 3′-flanking probe I (a 1.3 kb XhoI-EcoRV fragment; see Fig. 1A,B) and a neo-specific probe. Six out of 90 cell lines (7%) tested positive for a gene targeting event. The targeted ES cell line 3f-15 was injected into C57BL/6J (B6) blastocysts to produce a chimeric male that, when mated to B6 females, transmitted the mutation through the germline. Sib matings and backcrosses to B6 mice produced animals with largely B6 genetic background. All progeny were genotyped by Southern blot analysis of tail or skin DNA using a SacI digest and probe D, as described above.

Skeleton preparations and whole-mount in situ hybridization

Newborn pups were collected within 48 hours of birth and killed by CO2 asphyxiation. Skin was removed from the carcasses and used for genotyping. Skeleton preparations were made as described (Mansour et al., 1993). Bone lengths were measured under a microscope using a Digital Filar Eyepiece and M/XM processor (LASICO). For Table 2, the bones of three sex- and age-matched hoxd-11/hoxd-11 and hoxd-11+/hoxd-11+ animals were measured individually, averaged and the data expressed as the per cent ratio of the length of the mutant limb to that of the wild-type limb.

Table 2.

Lengths of hoxd-11/hoxd-11 forelimb bones as percent of average wild-type length

Lengths of hoxd-11−/hoxd-11− forelimb bones as percent of average wild-type length
Lengths of hoxd-11−/hoxd-11− forelimb bones as percent of average wild-type length

For whole-mount in situ hybridization, E10.5-E12.5 mouse embryos were fixed and processed as described (Carpenter et al., 1993). The digoxigenin RNA probe was transcribed from a 300 bp fragment containing the hoxd-11 homeobox and 3′ untranslated region (Izpisúa-Belmonte et al., 1991a).

Creating a hoxd-11 mutant mouse

A replacement-type gene targeting vector was con-structed from mouse genomic DNA containing the hoxd-11 gene which was disrupted by insertion of a neo-resistance cassette into the homeodomain encoding exon (Fig. 1A and Materials and Methods for details). The vector was introduced into CC1.2 ES cells which were then cultured under conditions chosen to enrich for a homologous recombi- nation event. DNA isolated from individual ES cell colonies was analyzed by Southern blot hybridization for a gene targeting event using three probes (the 5′ flanking probe D, the 3′ flanking probe I and a neo-specific probe) and three different restriction enzyme digests (Fig. 1A). 7% of the ES cell clones analyzed had the anticipated DNA structure as a result of a hoxd-11 gene targeting event; a representative analysis is shown in Fig. 1B. This targeted ES cell line was injected into B6-derived blastocysts to produce a chimeric male that passed the mutation through the germline. All subsequent progeny were genotyped by analysis of tail or skin DNA.

Disruption of hoxd-11 is not lethal

Mice heterozygous for the hoxd-11 mutation were intercrossed to produce homozygous mutant animals. Homozygosity for this mutation caused neither embryonic nor postnatal lethality as the predicted Mendelian ratio of alleles was observed among the newborn and adult mice. All pups were examined visually for general health and all appeared outwardly normal. However, hoxd-11/hoxd-11 male animals produced no offspring, suggesting that they are infertile (5/5 mice tested). Male mutants were killed and their reproductive tract was examined. Their reproductive organs and penian bone appeared anatomically normal. Sperm collected from the vas deferens appeared normal, were present in normal amounts and were motile. However, no vaginal plugs have been observed in wild-type females mated with males homozygous for the hoxd-11 mutation. The reason for the apparent male infertility has not yet been determined.

Homeosis in the sacral vertebrae

Skeleton preparations of newborn pups were prepared to examine the structure of the vertebral column. A summary of the vertebrae patterns for the hoxd-11/hoxd-11 animals and their wild-type littermates is provided in Table 1. The lower vertebral column of a normal mouse is characterized by six lumbar and four sacral vertebrae (L6:S4 pattern), with the latter fusing together at their transverse processes to form the sacral bone (Figs 2A, 3A). Approximately 10% of wild-type mice show five lumbar vertebrae instead of six (L5:S4 pattern) in the axial skeleton (Green, 1954). While most of our wild-type mice have the L6:S4 pattern, 13% (2/16 mice) display the variant pattern L4:A1:S4, which is similar to L5:S4 except that the most posterior lumbar vertebra is asymmetric (Green, 1954). All hoxd-11 heterozygotes (30/30 mice) examined showed only the wild-type L6:S4 pattern. Hoxd-11 homozygous mice, however, were abnormal in the patterning of these posterior vertebrae. This phenotype showed incomplete penetrance (86%) and variable expressivity.

Table 1.

Distribution of the vertebral phenotypes for hoxd-11 mice

Distribution of the vertebral phenotypes for hoxd-11 mice
Distribution of the vertebral phenotypes for hoxd-11 mice
Fig. 2.

Hoxd-11 mutant mice show an additional lumbar vertebra. Ventral views of the posterior axial columns of newborn skeleton preparations from wild-type (A; genotype +/+) and hoxd-11 mutant mice (B; −/−). The first lumbar vertebra is denoted as L1 and the last lumbar vertebra as either L6 (in A) or as the additional vertebra L7 (in B). Skeletons were stained for bone (with alizarin red) and cartilage (with alcian blue). Scale bar, 1 mm.

Fig. 2.

Hoxd-11 mutant mice show an additional lumbar vertebra. Ventral views of the posterior axial columns of newborn skeleton preparations from wild-type (A; genotype +/+) and hoxd-11 mutant mice (B; −/−). The first lumbar vertebra is denoted as L1 and the last lumbar vertebra as either L6 (in A) or as the additional vertebra L7 (in B). Skeletons were stained for bone (with alizarin red) and cartilage (with alcian blue). Scale bar, 1 mm.

The most abundant phenotypic class of the homozygous mutants (15/28 mice) had one additional lumbar vertebra. Instead of the normal six, a seventh lumbar vertebra is present followed by four sacral vertebrae, creating an L7:S4 pattern (Figs 2B, 3B). There are two possible interpretations to account for this anomaly: (1) the hoxd-11 mutation causes formation of a supernumerary lumbar vertebra, or (2) the mutation causes an anterior homeotic transformation of the entire sacral region in which the first sacral vertebra is transformed into lumbar 7 (S1λL7) and each subsequent sacral vertebra adopts the structure of the next most anterior vertebra.

The second possibility suggests that a deficiency of one vertebra should occur somewhere posterior to the transformation event. Unfortunately, in the remaining caudal region of the mouse the axial skeleton shows a natural variation in the number of vertebra, making such a determination impossible. However, the hoxd-11 mutation shows variable expressivity and this variation in phenotype lends support for the second hypothesis. The next most abundant phenotypic class (7/28 mice) displayed the correct number of six lumbar vertebrae, but an additional sacral vertebra was seen (L6:S5 pattern). The additional sacral vertebra always resembled S1, such that the pattern was S1-S1-S2-S3-S4 (Fig. 3C). In addition, two hoxd-11 homozygous animals clearly showed six normal lumbar vertebrae followed by the intermediate, asymmetric vertebra A2, in which the right half resembled a lumbar vertebra with an unfused transverse process, while the left half was reminiscent of S1 with its process fused to the sacral bone (Green, 1954). In this mutant, A2 was then followed by four sacral vertebrae (L6:A2:S4 pattern). Both of these examples may represent trapped intermediates in which S1 has failed to fully transform into L7 but all of the other sacral vertebrae (and at least one caudal vertebra) have undergone homeosis. To account for these intermediate phenotypes, the first explanation would have to be modified to suggest that this hoxd-11 mutation at times causes an additional lumbar vertebra, or causes an additional sacral vertebra but not both at the same time since we never see an L7:S5 pattern. Instead, we prefer the second hypothesis, i.e., that the mutation causes an anterior homeotic transformation event that affects the entire sacrum and that variable expressivity can account for the range of phenotypes observed.

Fig. 3.

Homeosis of the sacrum in hoxd-11 mutant mice. (A) Wild-type mouse (+/+) with four sacral vertebrae (S1 through S4) following the sixth lumbar vertebra (L6) creating the L6:S4 pattern. (B) Hoxd-11 mutant (−/−) mouse with four sacral vertebrae following the seventh lumbar vertebra (L7) creating the L7:S4 pattern. (C) Hoxd-11 mutant (−/−) mouse demonstrating variable expressivity of the mutation. Here, five sacral vertebrae (S1-S1-S2-S3-S4) are present following the sixth lumbar segment, creating the intermediate L6:S5 pattern. Scale bar, 1 mm.

Fig. 3.

Homeosis of the sacrum in hoxd-11 mutant mice. (A) Wild-type mouse (+/+) with four sacral vertebrae (S1 through S4) following the sixth lumbar vertebra (L6) creating the L6:S4 pattern. (B) Hoxd-11 mutant (−/−) mouse with four sacral vertebrae following the seventh lumbar vertebra (L7) creating the L7:S4 pattern. (C) Hoxd-11 mutant (−/−) mouse demonstrating variable expressivity of the mutation. Here, five sacral vertebrae (S1-S1-S2-S3-S4) are present following the sixth lumbar segment, creating the intermediate L6:S5 pattern. Scale bar, 1 mm.

Abnormal forelimbs in hoxd-11/hoxd-11 mice

Hoxd-11 transcripts have been reported in the mouse limb bud starting at E9.0 (Dollé et al., 1989). From E9.0 to E10.5, hoxd-11 transcripts are restricted to the proximal/posterior region of the developing limb bud (Dollé et al., 1989; and Fig. 4A). However, by E11.5 the pattern of hoxd-11 expression is observed to increase in complexity showing two stripes of expression, a distal and a more proximal one, which extend from the posterior to the anterior margin of the limb bud (Fig. 4B). By E12.5, hoxd-11 expression appears to be restricted to the perichondrium (Fig. 4C). Both in terms of position and intensity, the patterns of hoxd-11 expression are similar in the forelimbs and hindlimbs.

Fig. 4.

Whole-mount in situ hybridization analysis of hoxd-11 expression in the limb bud. Lateral view of fore- and hindlimbs of (A) E10.5, (B) E11.5 and (C) E12.5 embryos. Hoxd-11 expression is initially restricted to the proximal/posterior region of the limb buds (arrows in A) and then increases in complexity, first forming two stripes (arrows in B) and then becoming more restricted to the perichondrium (C).

Fig. 4.

Whole-mount in situ hybridization analysis of hoxd-11 expression in the limb bud. Lateral view of fore- and hindlimbs of (A) E10.5, (B) E11.5 and (C) E12.5 embryos. Hoxd-11 expression is initially restricted to the proximal/posterior region of the limb buds (arrows in A) and then increases in complexity, first forming two stripes (arrows in B) and then becoming more restricted to the perichondrium (C).

Since hoxd-11 transcripts are detected in the mouse limb buds during development, skeleton preparations were examined for defects in both the forelimbs and hindlimbs. A diagram of a wild-type mouse forelimb is shown in Fig. 5. Beneath the metacarpals of each digit, there are four distal carpal bones called d1, d2, d3 and d4. In some strains of mice, d2 is split to produce an additional central carpal bone (called c). Our parental mouse strain B6, which is the predominant genetic background in the hoxd-11 colony, does not show this division of d2 into d2 plus c. Instead, d2 remains as one larger carpal bone (referred to here as d2-c); occasionally, however, d2 has a line demarcating c, but the two bones are never fully separated. Proximal to the distal carpals are the navicular lunate, the triangular and the pisiform bones.

Fig. 5.

Diagram of a dorsal view of the wild-type mouse forelimb skeleton. Digit I has only two phalanges (P2 and P1); all other digits have three. The distal carpal bones are labelled d1 through d4; d2-c represents the fusion between d2 and the central carpal bone c. Distal is up, proximal down, anterior to the left and posterior to the right.

Fig. 5.

Diagram of a dorsal view of the wild-type mouse forelimb skeleton. Digit I has only two phalanges (P2 and P1); all other digits have three. The distal carpal bones are labelled d1 through d4; d2-c represents the fusion between d2 and the central carpal bone c. Distal is up, proximal down, anterior to the left and posterior to the right.

There are two readily apparent forelimb phenotypes observed in hoxd-11 homozygous animals. The length of the metacarpals (especially on digit II) is reduced, making the paw smaller, and a large gap exists between the radius and ulna (Fig. 6). There is, however, no evidence of a homeotic transformation event in which any digit adopts the structure of another. Closer examination of the forelimb revealed that phalange (P) 2 of digits II and V was also noticeably reduced in length (Fig. 7). The lengths of the metacarpals and phalanges were measured, and the results are given in Table 2. Articulations between the phalanges are sometimes malformed, especially those involving P1 and P2 of digit II which are fused in over 40% of the mutant animals (Fig. 7B).

Fig. 6.

Comparison of hoxd-11/hoxd-11 and wild-type forelimbs. Dorsal view of the forelimbs of a wild-type (A; +/+) and hoxd-11 mutant littermate (B; −/−). The mutant has a large gap between the radius and ulna (open arrow in B), and the overall size of the paw is smaller due to the reduction in lengths of the metacarpals (especially digit II, closed arrow in B). Digits are labelled I through V. Notice that none of the digits has undergone any type of a homeotic transformation. Scale bar, 1 mm.

Fig. 6.

Comparison of hoxd-11/hoxd-11 and wild-type forelimbs. Dorsal view of the forelimbs of a wild-type (A; +/+) and hoxd-11 mutant littermate (B; −/−). The mutant has a large gap between the radius and ulna (open arrow in B), and the overall size of the paw is smaller due to the reduction in lengths of the metacarpals (especially digit II, closed arrow in B). Digits are labelled I through V. Notice that none of the digits has undergone any type of a homeotic transformation. Scale bar, 1 mm.

Fig. 7.

Reduction in length of phalanges in hoxd-11 mutants. Digit II of wild-type mouse (A; +/+) and hoxd-11 mutant mouse (B; −/−). In the mutant, phalange 2 (P2) is greatly reduced (brackets) and shows partial fusion to P1 (arrow in B). Digit V of wild-type mouse (C; +/+) and hoxd-11 mutant mouse (D; −/−). Again, P2 in the mutant is severely reduced in length (brackets). Scale bar, 0.5 mm.

Fig. 7.

Reduction in length of phalanges in hoxd-11 mutants. Digit II of wild-type mouse (A; +/+) and hoxd-11 mutant mouse (B; −/−). In the mutant, phalange 2 (P2) is greatly reduced (brackets) and shows partial fusion to P1 (arrow in B). Digit V of wild-type mouse (C; +/+) and hoxd-11 mutant mouse (D; −/−). Again, P2 in the mutant is severely reduced in length (brackets). Scale bar, 0.5 mm.

In the distal carpals of hoxd-11 homozygous mice, d2 was always split into d2 plus the central bone c (Fig. 8A,B) The proximal carpal bones of these animals were also malformed. The navicular lunate, triangular and pisiform were fused together to form one large bone with the pisiform portion misshapen (Fig. 8). As with the axial skeleton, the limb phenotype showed incomplete penetrance (94%) and variable expressivity. 25% of the mutant animals had a detached navicular lunate but the triangular bone was fused to the malformed pisiform. The remaining mutants showed a fusion between the navicular lunate and the triangular while the pisiform remained separated yet malformed (Table 3). Hoxd-11 heterozygotes also displayed a mild phenotype in the wrist. These animals had d2 separated into d2 plus c but no other defects in any of the carpal bones (Table 3).

Table 3.

Carpal bone phenotypes for adult hoxd-11 mice

Carpal bone phenotypes for adult hoxd-11 mice
Carpal bone phenotypes for adult hoxd-11 mice
Fig. 8.

Wrist bone and zeugopod phenotype in hoxd-11 mutant mice. Dorsal view of the carpal region of a wild-type mouse (A; +/+) and hoxd-11 mutant (B; −/−). In the mutant (B), the single carpal bone d2-c is split into two distinct bones called d2 and c; in addition, the navicular lunate and triangular bones which remain separate in the wild-type sib are fused together in the mutant (compare arrows in A and B). In the zeugopod, the distal tip of the radius (r) has a minor excrescence on its dorsal ridge and a thinning on the medial side (open arrowheads, B) while the ulna (u) has a smaller and more rounded styloid process (closed arrowhead, B). Note the large gap between the radius and the ulna in the mutant (B) as compared to the wild-type sib (A). Side view of the wrist in a wild-type mouse (C; +/+) and hoxd-11 mutant (D; −/−) shows the fusion of a misshapen pisiform (p) to the triangular bone beneath Digit V in the mutant (compare arrows in C, D).

Fig. 8.

Wrist bone and zeugopod phenotype in hoxd-11 mutant mice. Dorsal view of the carpal region of a wild-type mouse (A; +/+) and hoxd-11 mutant (B; −/−). In the mutant (B), the single carpal bone d2-c is split into two distinct bones called d2 and c; in addition, the navicular lunate and triangular bones which remain separate in the wild-type sib are fused together in the mutant (compare arrows in A and B). In the zeugopod, the distal tip of the radius (r) has a minor excrescence on its dorsal ridge and a thinning on the medial side (open arrowheads, B) while the ulna (u) has a smaller and more rounded styloid process (closed arrowhead, B). Note the large gap between the radius and the ulna in the mutant (B) as compared to the wild-type sib (A). Side view of the wrist in a wild-type mouse (C; +/+) and hoxd-11 mutant (D; −/−) shows the fusion of a misshapen pisiform (p) to the triangular bone beneath Digit V in the mutant (compare arrows in C, D).

In addition, both the radius and ulna showed minor abnor-malities in hoxd-11/hoxd-11 animals. At the distal head of the ulna, the styloid process was smaller and more rounded than normal, and the radial epiphysis had a minor excrescence on its dorsal ridge and a thinning on the ventral medial side (Fig. 8A,B). Over half of the mutants (9/17 mice) also had a small sesamoid bone located on the ventral side between the radius and ulna at the distal end just below the carpals that was never seen in heterozygotes or wild-type sibs (data not shown). We believe it is a combination of these bone defects together with the fusion of the proximal carpal bones, that lead to the formation of the noticeable gap between the radius and ulna of the mutant limbs (Fig. 6B).

A summary of the forelimb defects observed in hoxd-11−λhomozygous mice is provided in Fig. 9. Interestingly, the set of defects is distributed over the autopod and extends into the zeugopod. This mutation appears to affect limb patterning along both the Pr-D and A-P axes since the lengths of the phalanges and metacarpals as well as the shapes of the wrist bones, radius and ulna are all affected.

Fig. 9.

Summary of the major forelimb phenotypes seen in hoxd-11 mutant mice. Red dots represent the site of limb defects. From proximal to distal, these include: a large gap between the radius and ulna, malformations of both the radial and ulnar epiphysis, an aberrant sesamoid bone between the radius and ulna, two fusions between the three proximal carpal bones, a split of the carpal bone d2-c, the reduction in metacarpal lengths (digits II, III, IV), a fusion between P1 and P2 (digit II) and a strong reduction in the length of P2 (digits II and V).

Fig. 9.

Summary of the major forelimb phenotypes seen in hoxd-11 mutant mice. Red dots represent the site of limb defects. From proximal to distal, these include: a large gap between the radius and ulna, malformations of both the radial and ulnar epiphysis, an aberrant sesamoid bone between the radius and ulna, two fusions between the three proximal carpal bones, a split of the carpal bone d2-c, the reduction in metacarpal lengths (digits II, III, IV), a fusion between P1 and P2 (digit II) and a strong reduction in the length of P2 (digits II and V).

Hindlimb phenotypes

The hindlimbs of hoxd-11 homozygous mice did not have any major abnormalities. However, a sesamoid bone lateral to the tibiale mediale was either absent or greatly reduced in the hoxd-11 heterozygous and homozygous animals (Fig. 10A, B). In addition, as with the distal carpal bones of the forelimb, there often existed a fusion between the tarsal navicular and cuneiforme 3 in wild-type mice while, in most of the heterozygous (6/8 mice) and homozygous mutants (16/17 mice), these two bones were completely separate (Fig. 10C-F). The absence of a severe phenotype in the hindlimbs may be due to compensation by two paralogues of hoxd-11. We might anticipate a greater degree of redundancy of function among the three paralogous genes in forming the hindlimb since hoxc-11 is expressed in the hindlimb but not in the forelimb (S. L. Hostikka and Capecchi, unpublished results).

Fig. 10.

Hindlimb defects in hoxd-11 mutant mice. A sesamoid bone next to the tibiale mediale (tm) in the wild-type mouse (arrow in A; +/+) is absent in the mutant (B; −/−). The normal fusion between the tarsal bones navicular (n) and cuneiforme 3 (c3) in the wild-type mouse (arrow in C; +/+) does not appear in the mutant (arrow in D; −/−). Dissection of the tarsal bones shows how the navicular and cuneiforme 3 remain fused in the wild-type mouse (E) but separate in the mutant (F). Scale bars, (A, B) 0.25 mm; (C, D) 0.5 mm; (E, F) 0.4 mm. The talus is the most proximal tarsal bone and is labelled for orientation.

Fig. 10.

Hindlimb defects in hoxd-11 mutant mice. A sesamoid bone next to the tibiale mediale (tm) in the wild-type mouse (arrow in A; +/+) is absent in the mutant (B; −/−). The normal fusion between the tarsal bones navicular (n) and cuneiforme 3 (c3) in the wild-type mouse (arrow in C; +/+) does not appear in the mutant (arrow in D; −/−). Dissection of the tarsal bones shows how the navicular and cuneiforme 3 remain fused in the wild-type mouse (E) but separate in the mutant (F). Scale bars, (A, B) 0.25 mm; (C, D) 0.5 mm; (E, F) 0.4 mm. The talus is the most proximal tarsal bone and is labelled for orientation.

Hoxd-11 homozygotes are viable and appear outwardly normal. The only apparent abnormality is male infertility. However, the cause of male infertility has not yet been determined. The male reproductive organs and penian bone appear anatomically normal and normal amounts of motile sperm are produced. Examination of the mutant skeletons shows defects in the formation of the vertebral column and the limbs. Thus, structure of the skeleton along the major body axis as well as along the appendicular axis is affected, emphasizing that Hox genes function as a multiaxial patterning system in mammals.

Anterior homeotic transformation of the sacrum

Hoxd-11/hoxd-11 animals have a deviant vertebral pattern in the lumbar/sacral region. Instead of the wild-type pattern of six lumbar and four sacral vertebrae (L6:S4), the homozygous mutants generally have seven lumbar and four sacral vertebrae (L7:S4). This mutant pattern may be the result of an anterior homeotic transformation of the entire sacral region initiating with S1. This conclusion is supported by the intermediate phenotypes produced by variable expressivity of the mutation (see Results). There exists ample precedence that loss-of-function mutations in Hox genes cause apparent homeotic transformations of components of the vertebral column (LeMouellic et al., 1992; Ramirez-Solis et al., 1993; Jeannotte et al., 1993; Kostic and Capecchi, 1994; Condie and Capecchi, 1994). Disruption of hoxd-13, another member of the 5′-Hox D locus, also produces an apparent homeotic transformation in the sacral region (Dollé et al., 1993). This phenotype, however, is restricted to the most posterior portion of the sacrum (i.e., a transformation of S4 to S3). The site of homeosis for hoxd-11 mutants is consistent with its anterior limit of expression in the prevertebrae, thus supporting the posterior prevalence model (Duboule, 1991). However, since the transformation in the hoxd-11 mutant mouse requires each vertebra to adopt the structure of the next most anterior vertebra, as opposed to transformation of the entire region to one vertebral fate, a more complex scenario involving a fine-tuning mechanism must be envisioned.

Interestingly, disruption of hoxa-11, a paralogue of hoxd-11, appears to result in the same homeotic transformation of the sacral vertebrae observed in hoxd-11 mutant mice (Small and Potter, 1993). In this case, however, the anterior limit of expression for hoxa-11 is reported to be at prevertebra 20 rather than at prevertebra 25, the site where homeosis of the sacrum is initiated. Perhaps, the disruption of hoxa-11 alters the expression pattern of hoxd-11, thereby accounting for the identical transformation of the sacrum. Alternatively, hoxa-11 and hoxd-11 may function together to specify the morphology of the sacrum. In this context, it will be of interest to analyze mice defective for both hoxa-11 and hoxd-11.

Forelimb defects in hoxd-11 mutant mice

Hoxd-11/hoxd-11 animals show several malformations in the appendicular skeleton of the forelimb. From proximal to distal, these defects include: a large gap between the radius and ulna, malformations of both the radial and ulnar epiphysis, the presence of an aberrant sesamoid bone between the radius and ulna, a fusion between the three proximal carpal bones (navicular lunate, triangular and a misshapen pisiform), a split of the distal carpal bone d2-c into d2 plus a central bone (called c), reduction in metacarpal length (digit II, III, IV), fusion between phalanges 1 and 2 (digit II), and a strong reduction in length of phalange 2 (digit II, V).

Surprisingly, the hindlimbs of these mutant animals appear fairly normal, except for the failure of a fusion between two tarsal bones and a reduction in the size of a sesamoid bone located next to the tibiale mediale. The absence of more severe hindlimb defects may reflect overlapping functions supplied by a paralogous member of the Hox C linkage group which is expressed in hindlimbs but not in forelimbs. It should be noted that, with respect to the hoxa-11 and hoxd-13 mutant phenotypes, the forelimbs and hindlimbs are equally affected. As might be expected, the degree of overlapping function between paralogous Hox genes varies from gene to gene. The set of digit defects observed in hoxd-11 mutant mice is difficult to interpret in terms of a model where digit identity is specified solely by a gradient of a diffusible morphogen emanating from the zone of polarizing activity (Tabin, 1991). Thus, the set of defects of the digits does not appear to be polarized with respect to the ZPA. The length of phalange 2 of digits II and V is strongly affected with much less effect on the lengths of the same phalanges in digits III and IV. In contrast, the lengths of the metacarpals of digits II, III and IV are reduced without alteration in the metacarpal lengths of digits I and V. Based on the nested set of Hox D gene expression patterns in the limb and the observation that ectopic expression of mouse hoxd-11 in the chick limb sometimes resulted in a homeotic transformation of digit identity, a Hox D code was proposed for specifying digit identity (Morgan et al., 1992). The limb phenotype associated with disruption of hoxd-11 does not support this model. From a comparison of the phenotypes resulting from disruption of hoxa-11, hoxd-11 and hoxd-13, it also appears unlikely that even combining mutations in paralogous ‘limb’Hox genes will reveal homeosis of the digits (see below).

Overlap of phenotypes among hoxd-11, hoxa-11 and hoxd-13 mutant mice

The recent description of mice with targeted disruptions in hoxa-11, a paralogue of hoxd-11 (Small and Potter, 1993), and another member of the Hox D linkage group, hoxd-13 (Dollé et al., 1993), allows us to compare the limb phenotypes of these mutant mice with those of hoxd-11 mutants. Fig. 11 summarizes the forelimb phenotypes for mice carrying disruptions in hoxa-11, hoxd-11 and hoxd-13. Four points are evident. First, there are overlapping defects for hoxd-11 with both hoxa-11 (ulnar epiphysis, sesamoid bone, carpal bone fusions, misshapen pisiform) and hoxd-13 (metacarpal and phalanges) mutants. Secondly, each mutant shows unique phenotypes not present in the other two (for details see Dollé et al., 1993; Small and Potter, 1993). Thirdly, targeted disruption of these Hox genes does not cause homeosis of the digits. Lastly, the position of the most proximal phenotype for all three mice parallels both the chromosome position (structural colinearity) and the activation order of these genes (temporal colinearity). Hoxa-11 and hoxd-11 deformities are initiated in the zeugopod (the radius and ulna long bones) while hoxd-13 mutant animals are limited to defects in the later forming distal autopod.

Fig. 11.

A composite diagram of the mouse forelimb phenotypes from individual gene targeting experiments of hoxd-11 (this report), hoxa-11 (Small and Potter, 1993) and hoxd-13 (Dollé et al., 1993). The colored dots represent the physical position of a mutant phenotype when compared to wild-type sibs. Hoxa-11 is a paralogue of hoxd-11 and hoxd-13 is another member of the Hox D locus. The diagram illustrates several points: (1) the overlapping phenotypes that hoxd-11 mutant mice share with both hoxa-11 and hoxd-13 animals, (2) the unique phenotypes each mutant displays, (3) that hoxd-11 mutants show more proximal phenotypes than hoxd-13 mutants reflecting both structural and temporal colinearity, and (4) that the loss-of-function mutations of Hox genes in the limb do not cause homeosis, but instead result in localized malformation of several bones that do not appear to be polarized with respect to the ZPA.

Fig. 11.

A composite diagram of the mouse forelimb phenotypes from individual gene targeting experiments of hoxd-11 (this report), hoxa-11 (Small and Potter, 1993) and hoxd-13 (Dollé et al., 1993). The colored dots represent the physical position of a mutant phenotype when compared to wild-type sibs. Hoxa-11 is a paralogue of hoxd-11 and hoxd-13 is another member of the Hox D locus. The diagram illustrates several points: (1) the overlapping phenotypes that hoxd-11 mutant mice share with both hoxa-11 and hoxd-13 animals, (2) the unique phenotypes each mutant displays, (3) that hoxd-11 mutants show more proximal phenotypes than hoxd-13 mutants reflecting both structural and temporal colinearity, and (4) that the loss-of-function mutations of Hox genes in the limb do not cause homeosis, but instead result in localized malformation of several bones that do not appear to be polarized with respect to the ZPA.

The overlapping phenotypes seen in hoxa-11 and hoxd-11 mutant mice might have been anticipated, since defects in mice with mutations in paralogous genes appear to be restricted to the same regions of the embryo (Chisaka and Capecchi, 1992; Condie and Capecchi, 1993). We would like to suggest that in the regions of overlap the two paralogous gene products may function synergistically to regulate the growth of localized groups of cells. This suggestion is influenced by our recent analysis of mice doubly mutant for the two paralogous genes, hoxa-3 and hoxd-3 (unpublished data). Even though mice with mutations in one or the other of these genes show no overlap in phenotype, the double mutants demonstrate that these two genes strongly interact. The hoxa-3 mutant phenotype is exacerbated in mice homozygous for both targeted disruptions. Likewise, the hoxd-3 phenotype is exacerbated in the double mutant. The interactions appear to be quantitative since the degree of exacerbation of the hoxd-3 phenotype in (hoxd-3/ hoxd-3; hoxa-3/hoxa-3+) mice is intermediate to that observed in (hoxd-3/hoxd-3; hoxa-3+/hoxa-3+) or (hoxd-3/hoxd-3; hoxa-3/ hoxa-3) mutant mice. The interactions between these genes and the resultant phe-notypes are readily interpretable in terms of models in which these genes synergistically control the rates of proliferation of localized groups of cells. In light of the above discussion, it will be of interest to examine the limbs (and the sacrum) of mice mutant for both hoxa-11 and hoxd-11. Since these genes are on separate chromosomes, double mutants can be generated by intercrosses.

The overlap of forelimb phenotypes observed in hoxd-11 and hoxd-13 mutant mice is also intriguing. Again we can postulate that, in the regions of overlap, the two gene products are cooperating in regulating the formation of bones. Alternatively, some of the defects observed in hoxd-11 mutant mice could be caused by disruption of the temporal sequence of activation of other Hox D genes. For example, phalange 2 of digits II and V in hoxd-11 mice is severely reduced in length; in hoxd-13 mutant animals, however, these elements are entirely missing. Disruption of hoxd-11 may influence hoxd-13 by retarding its activation, and thereby mimicking its phenotype, though with a reduced severity. Mice doubly mutant for hoxd-11 and hoxd-13 might resolve the issue of whether these two gene products interact synergistically to control the proliferation rates of common precursor cells or whether a mutation in hoxd-11 disrupts the temporal progression of the Hox D gene activation during patterning of the limb.

In summary, limb bones are formed from prechondrogenic condensations as the limb bud grows. The pattern of these condensations may be inherently dependent upon the overall limb geometry (Oster et al., 1988). A sequence of segmentations and bifurcations of these condensations occurs in a Pr-D direction to create the long bones (humerus, then radius and ulna). This is followed by condensations progressing in an A-P manner to generate the carpals, metacarpals and digits (Shubin and Alberch, 1986). Disruption of hoxd-11, as described here, affects the morphology of bones arising from both the Pr-D and A-P condensations. In addition, malformations of the same bones are independently affected by disruptions in hoxa-11 and hoxd-13 (see Fig. 11). These observations taken together suggest that these Hox genes do not code for distinct positional information, per se, dictated solely by an A-P morphogenic gradient. Rather, what is emerging is a more complex set of interactions among these Hox genes, in which each is responsible for localized regions of cell growth. Perturbations in the pattern of localized growth rates of the cells contributing to the prechondrogenic condensations could then lead to the loss of structures, fusion between bones and even additional elements. All three phenotypes are seen in mice mutant for hoxa-11, hoxd-11 and hoxd-13.

We want to acknowledge and thank S. L. Hostikka for doing the whole-mount in situ hybridizations. We also want to thank M. Allen, S. Barnett, C. Lenz, E. Nakashima and S. Tamowski for excellent technical assistance. K. Thomas, D. Spyropoulos and T. Tsuzuki provided certain DNA probes and vector cassettes, D. Rancourt provided the phage library of CC1.2 genomic DNA, and B. Condie and T. Musci provided help with histology. We thank C. S. Thummel for the use of the Digital Filar Eyepiece and M/XM processor. L. Oswald helped with the preparation of the manuscript. A. P. D. is the recipient of an NSF predoctoral fellowship and an NIH genetics training grant.

Acampora
,
D.
,
D’Esposito
,
M.
,
Faiella
,
A.
,
Pannese
,
M.
,
Migliaccio
,
E.
,
Morelli
,
F.
,
Stornauiolo
,
A.
,
Nigro
,
V.
,
Simeone
,
A.
and
Boncinelli
,
E.
(
1989
).
The human hox gene family
.
Nucleic Acids Res
.
17
,
10385
10402
.
Akam
,
M. E.
(
1987
).
The molecular basis for metameric pattern in the Drosophila embryo
.
Development
101
,
1
22
.
Capecchi
,
M. R.
,
Capecchi
,
N. E.
,
Hughes
,
S. H.
and
Wahl
,
G. M.
(
1974
).
Selective degradation of abnormal proteins in mammalian tissue culture cells
.
Proc. Natl. Acad. Sci. USA
71
,
4732
4736
.
Capecchi
,
M. R.
(
1989
).
Altering the genome by homologous recombination
.
Science
244
,
1288
1292
.
Capecchi
,
M. R.
(
1994
).
Targeted gene replacement
.
Sci. Am
.
270
,
54
61
.
Carpenter
,
E. M.
,
Goddard
,
J. M.
,
Chisaka
,
O.
,
Manley
,
N. R.
and
Capecchi
,
M. R.
(
1993
).
Loss of Hoxa-1 (Hox-1.6) function results in the reorganization of the murine hindbrain
.
Development
118
,
1063
1075
.
Chisaka
,
O.
and
Capecchi
,
M. R.
(
1991
).
Regionally restricted developmental defects resulting from targeted disruption of the mouse homeobox gene hox-1.5
.
Nature
350
,
473
479
.
Chisaka
,
O.
,
Musci
,
T. S.
and
Capecchi
,
M. R.
(
1992
).
Developmental defects of the ear, cranial nerves and hindbrain resulting from targeted disruption of the mouse homeobox gene Hox-1.6
.
Nature
355
,
516
520
.
Condie
,
B. G.
and
Capecchi
,
M. R.
(
1993
).
Mice homozygous for a targeted disruption of Hoxd-3 (Hox-4.1) exhibit anterior transformations of the first and second cervical vertebrae, the atlas and the axis
.
Development
119
,
579
595
.
Deng
,
C.
and
Capecchi
,
M. R.
(
1992
).
Reexamination of gene targeting frequency as a function of the extent of homology between the targeting vector and the target locus
.
Mol. Cell. Biol
.
12
,
3365
3371
.
Deng
,
C.
,
Thomas
,
K. R.
and
Capecchi
,
M. R.
(
1993
).
Location of crossovers during gene targeting with insertion and replacement vectors
.
Mol. Cell. Biol
.
13
,
2134
2140
.
Dollé
,
P.
,
Dierich
,
A.
,
LeMeur
,
M.
,
Schimmang
,
T.
,
Schuhbaur
,
B.
,
Chambon
,
P.
and
Duboule
,
D.
(
1993
).
Disruption of the Hoxd-13 gene induces localized heterochrony leading to mice with neotenic limbs
.
Cell
75
,
431
441
.
Dollé
,
P.
,
Izpisúa-Belmonte
,
J. C.
,
Boncinelli
,
E.
and
Duboule
,
D.
(
1991a
).
The Hox-4.8 gene is localized at the 5′ extremity of the Hox-4 complex and is expressed in the most posterior parts of the body during development
.
Mech. Dev
.
36
,
3
13
.
Dollé
,
P.
,
Izpisúa-Belmonte
,
J. C.
,
Brown
,
J. M.
,
Tickle
,
C.
and
Duboule
,
D.
(
1991b
).
Hox-4 genes and the morphogenesis of mammalian genitalia
.
Genes Dev
.
5
,
1767
1776
.
Dollé
,
P.
,
Izpisúa-Belmonte
,
J. C.
,
Falkenstein
,
H.
,
Renucci
,
A.
and
Duboule
,
D.
(
1989
).
Coordinate expression of the murine Hox-5 complex homeobox-containing genes during limb pattern formation
.
Nature
342
,
767
772
.
Duboule
,
D.
(
1991
).
Patterning in the vertebral limb
.
Curr. Opin. Genet. Dev
.
1
,
211
216
.
Duboule
,
D.
(
1992
).
The vertebrate limb: a model system to study the Hox/HOM gene network during development and evolution
.
BioEssays
14
,
375
384
.
Duboule
,
D.
and
Dollé
,
P.
(
1989
).
The structural and functional organization of the murine Hox gene family resembles that of Drosophila homeotic genes
.
EMBO J
8
,
1497
1505
.
Echelard
,
Y.
,
Epstein
,
D. J.
,
St-Jacques
,
B.
,
Shen
,
L.
,
Mohler
,
J.
,
McMahon
,
J. A.
and
McMahon
,
A. P. C.
(
1993
).
Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity
.
Cell
75
,
1417
1430
.
Gendron-Maguire
,
M.
,
Mallo
,
M.
,
Zhang
,
M.
and
Gridley
,
T.
(
1993
).
Hoxa2 mutant mice exhibit homeotic transformation of skeletal elements derived from cranial neural crest
.
Cell
75
,
1317
1331
.
Graham
,
A.
,
Papalopulu
,
N.
and
Krumlauf
,
R.
(
1989
).
The murine and Drosophila homeobox gene complexes have common features of organization and expression
.
Cell
57
,
367
378
.
Green
,
E. L.
(
1954
).
Quantitative genetics of skeletal variations in the mouse. I. Crosses between three short-ear strains (P, NB, SEC/2)
.
J. Natl. Cancer Inst
.
15
,
609
627
.
Haack
,
H.
and
Gruss
,
P.
(
1993
).
The establishment of murine Hox-1 expression domains during patterning of the limb
.
Dev. Biol
.
157
,
410
422
.
Hinchliffe
,
R.
(
1991
).
Developmental approaches to the problem of transformation of limb structure in evolution
.
In Developmental Patterning of the Vertebrate Limb
(Eds.
J. R.
Hinchliffe
,
J. M.
Hurle
and
D.
Summerbell
).
New York
:
Plenum Publishing Corp
.
Izpisúa-Belmonte
,
J.-C.
,
Falkenstein
,
H.
,
Dollé
,
P.
,
Renucci
,
A.
and
Duboule
,
D.
(
1991a
).
Murine genes related to the Drosophila AbdB homeotic gene are sequentially expressed during development of the posterior part of the body
.
EMBO J
.
10
,
2279
2289
.
Izpisúa-Belmonte
,
J.-C.
,
Tickle
,
C.
,
Dollé
,
P.
,
Wolpert
,
L.
and
Duboule
,
D.
(
1991b
).
Expression of the homeobox Hox-4 genes and the specification of position in chick wing development
.
Nature
350
,
585
589
.
Izpisúa-Belmonte
,
J.-C.
and
Duboule
,
D.
(
1992
).
Homeobox genes and pattern formation in the vertebrate limb
.
Dev. Biol
.
152
,
26
36
.
Jeannotte
,
L.
,
Lemieux
,
M.
,
Charron
,
J.
,
Poirier
,
F.
and
Robertson
,
E. J.
(
1993
).
Specification of axial identity in the mouse: role of the Hoxa-5 (Hox1.3) gene
.
Genes Dev
.
7
,
2085
2096
.
Kissinger
,
C. R.
,
Liu
,
B.
,
Martin-Blanco
,
E.
,
Kornberg
,
T. B.
and
Pabo
,
C. O.
(
1990
).
Crystal structure of an engrailed homeodomain-DNA complex at 2.8 Å resolution: a framework for understanding homeodomain-DNA interactions
.
Cell
63
,
579
590
.
Kostic
,
D.
and
Capecchi
,
M. R.
(
1994
).
Targeted disruptions of the murine hoxa-4 and hoxa-6 genes result in homeotic transformations of components of the vertebral column
.
Mech. Dev
. In press.
LeMouellic
,
H.
,
Lallemand
,
Y.
and
Brûlet
,
P.
(
1992
).
Homeosis in the mouse induced by a null mutation in the Hox-3.1 gene
.
Cell
69
,
251
264
.
Lewis
,
E. B.
(
1978
).
A gene complex controlling segmentation in Drosophila
.
Nature
276
,
565
570
.
Lufkin
,
T.
,
Dierich
,
A.
,
LeMeur
,
M.
,
Mark
,
M.
and
Chambon
,
P.
(
1991
).
Disruption of the Hox-1.6 homeobox gene results in defects in a region corresponding to its rostral domain of expression
.
Cell
66
,
1105
1119
.
Mansour
,
S. L.
,
Thomas
,
K. R.
and
Capecchi
,
M. R.
(
1988
).
Disruption of the proto-oncogene int-2 in mouse embryo-derived stem cells: A general strategy for targeting mutations to nonselectable genes
.
Nature
336
,
348
352
.
Mansour
,
S. L.
,
Goddard
,
J. M.
and
Capecchi
,
M. R.
(
1993
).
Mice homozygous for a targeted disruption of the proto-oncogene int-2 have developmental defects in the tail and inner ear
.
Development
117
,
13
28
.
Mark
,
M.
,
Lufkin
,
T.
,
Vonesch
,
J. L.
,
Ruberte
,
E.
,
Olivo
,
J.-C.
,
Dollé
,
P.
,
Gorry
,
P.
,
Lumsden
,
A.
and
Chambon
P.
(
1993
).
Two rhombomeres are altered in Hoxa-1 mutant mice
.
Development
119
,
319
338
.
McGinnis
,
N.
,
Kuziora
,
M. A.
and
McGinnis
,
W.
(
1990
).
Human Hox-4.2 and Drosophila Deformed encode similar regulatory specificities in Drosophila embryos and larvae
.
Cell
63
,
969
976
.
McGinnis
,
W.
and
Krumlauf
,
R.
(
1992
).
Homeobox genes and axial patterning
.
Cell
68
,
283
302
.
Morgan
,
B. A.
,
Izpisúa-Belmonte
,
J.-C.
,
Duboule
,
D.
and
Tabin
,
C. J.
(
1992
).
Targeted misexpression of Hox-4.6 in the avian limb bud causes apparent homeotic transformations
.
Nature
358
,
236
239
.
Nohno
,
T.
,
Noji
,
S.
,
Koyama
,
E.
,
Ohyama
,
K.
,
Myokai
,
F.
,
Kuroiwa
,
A.
,
Saito
,
T.
and
Taniguchi
,
S.
(
1991
).
Involvement of the Chox-4 chicken homeobox genes in determination of anteroposterior axial polarity during limb development
.
Cell
64
,
1197
1205
.
Oster
,
G. F.
,
Shubin
,
N.
,
Murray
,
J. D.
and
Alberch
,
P.
(
1988
).
Evolution and morphogenetic rules: the shape of the vertebrate limb in ontogeny and phylogeny
.
Evolution
42
,
862
884
.
Otting
,
G.
,
Qian
,
Y. Q.
,
Billeter
,
M.
,
Muller
,
M.
,
Affolter
,
M.
,
Gehring
,
W. J.
and
Wuthrich
,
K.
(
1990
).
Protein--DNA contacts in the structure of a homeodomain--DNA complex determined by nuclear magnetic resonance spectroscopy in solution
.
EMBO J
.
9
,
3085
3092
.
Pendleton
,
J. W.
,
Nagai
,
B. K.
,
Murtha
,
M. T.
and
Ruddle
,
F. H.
(
1993
).
Expansion of the Hox gene family and the evolution of chordates
.
Proc. Natl. Acad. Sci. USA
90
,
6300
6304
.
Ramirez-Solis
,
R.
,
Zheng
,
H.
,
Whiting
,
J.
,
Krumlauf
,
R.
and
Bradley
,
A.
(
1993
).
Hoxb-4 (Hox-2.6) mutant mice show homeotic transformation of a cervical vertebra and defects in the closure of the sternal rudiments
.
Cell
73
,
279
294
.
Rechsteiner
,
M.
(
1987
).
Ubiquitin-mediated pathways for intracellular proteolysis
.
A. Rev. Cell. Biol
.
3
,
1
30
.
Riddle
,
R.
,
Johnson
,
R. L.
,
Laufer
,
E.
and
Tabin
,
C.
(
1993
).
Sonic hedgehog mediates the polarizing activity of the ZPA
.
Cell
75
,
1401
1416
.
Rijli
,
F. M.
,
Mark
,
M.
,
Lakkaraju
,
S.
,
Dierich
,
A.
,
Dollé
,
P.
and
Chambon
,
P.
(
1993
).
A homeotic transformation is generated in the rostral branchial region of the head by disruption of Hoxa-2, which acts as a selector gene
.
Cell
75
,
1333
1349
.
Saunders
,
J. W.
(
1977
).
The experimental analysis of chick limb bud development
.
In Vertebrate Limb and Somite Morphogenesis
. (Eds.
D. A.
Ede
,
J. R.
Hinchliffe
and
M.
Balls
).
Cambridge
:
Cambridge University Press
, pp.
289
314
.
Shubin
,
N. H.
and
Alberch
,
P.
(
1986
).
A morphogenetic approach to the origin and basic organization of the tetrapod limb
.
Evol. Biol
.
20
,
319
387
.
Small
,
K. M.
and
Potter
,
S. S.
(
1993
).
Homeotic transformations and limb defects in Hoxa-11 mutant mice
.
Genes Dev
.
7
,
2318
2328
.
Tabin
,
C. J.
(
1991
).
Retinoids, homeoboxes and growth factors: toward molecular models for limb development
.
Cell
66
,
199
217
.
Tabin
,
C. J.
(
1992
).
Why we have (only) five fingers per hand: Hox genes and the evolution of paired limbs
.
Development
116
,
289
296
.
Thomas
,
K. R.
and
Capecchi
,
M. R.
(
1987
).
Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells
.
Cell
51
,
503
512
.
Thomas
,
K. R.
,
Deng
,
C.
and
Capecchi
,
M. R.
(
1992
).
High-fidelity gene targeting in embryonic stem cells by using sequence replacement vectors
.
Mol. Cell. Biol
.
12
,
2919
2923
.
Tickle
,
C.
,
Summerbell
,
D.
and
Wolpert
,
L.
(
1975
).
Positional signalling and specification of digits in chick limb morphogenesis
.
Nature
254
,
199
202
.
Yokouchi
,
Y.
,
Sasaki
,
H.
and
Kuroiwa
,
A.
(
1991
).
Homeobox gene expression correlated with the bifurcation process of limb cartilage development
.
Nature
353
,
443
445
.
Zappavigna
,
V.
,
Sartori
,
D.
and
Mavilio
,
F.
(
1994
).
Specificity of HOX protein function depends on DNA-protein and protein-protein interactions, both mediated by the homeo domain
.
Genes Dev
.
8
,
732
744
.