Article
Reference
Genetic analysis of a conserved sequence in the HoxD complex:
regulatory redundancy or limitations of the transgenic approach?
BECKERS, Johannes, DUBOULE, Denis
Abstract
Extensive sequencing in the HoxD complex of several vertebrate species has revealed a set of conserved DNA sequences interspersed between neighboring Hox genes. Their high degree of conservation strongly suggested that they are used for regulatory purposes, a hypothesis that was largely confirmed by using "classical transgenesis" or in vivo mutagenesis through the embryonic stem (ES) cell technology. Here, we show that this is not always the case. We report that the deletion of a conserved regulatory sequence located in the HoxD complex gives different results, depending on the transgenic approach that was used. In
"conventional" transgenesis, this sequence was necessary for proper expression in a subdomain of the developing limb. However, a deletion of this sequence in complexo did not confirm this effect, thereby creating an important discrepancy between the classical transgenic and the ES cell-based, targeted mutagenesis. This unexpected observation may show the limitations of the former technology. Alternatively, it could illustrate a redundancy in regulatory circuits and, thus, justify the combination of parallel [...]
BECKERS, Johannes, DUBOULE, Denis. Genetic analysis of a conserved sequence in the HoxD complex: regulatory redundancy or limitations of the transgenic approach?
Developmental Dynamics , 1998, vol. 213, no. 1, p. 1-11
DOI : 10.1002/(SICI)1097-0177(199809)213:1<1::AID-AJA1>3.0.CO;2-L PMID : 9733096
Available at:
http://archive-ouverte.unige.ch/unige:83156
Disclaimer: layout of this document may differ from the published version.
1 / 1
Genetic Analysis of a Conserved Sequence in the HoxD Complex: Regulatory Redundancy or Limitations
of the Transgenic Approach?
JOHANNES BECKERSANDDENIS DUBOULE*
Department of Zoology and Animal Biology, Sciences III, University of Geneva, Geneva, Switzerland ABSTRACT Extensive sequencing in the
HoxD complex of several vertebrate species has revealed a set of conserved DNA sequences inter- spersed between neighboring Hox genes. Their high degree of conservation strongly suggested that they are used for regulatory purposes, a hypothesis that was largely confirmed by using
‘‘classical transgenesis’’ or in vivo mutagenesis through the embryonic stem (ES) cell technology.
Here, we show that this is not always the case. We report that the deletion of a conserved regulatory sequence located in the HoxD complex gives dif- ferent results, depending on the transgenic ap- proach that was used. In ‘‘conventional’’ transgen- esis, this sequence was necessary for proper expression in a subdomain of the developing limb. However, a deletion of this sequence in complexo did not confirm this effect, thereby creating an important discrepancy between the classical transgenic and the ES cell-based, tar- geted mutagenesis. This unexpected observation may show the limitations of the former technol- ogy. Alternatively, it could illustrate a redun- dancy in regulatory circuits and, thus, justify the combination of parallel strategies. Dev. Dyn.
1998;213:1–11. r1998 Wiley-Liss, Inc.
Key words: HoxD complex; embryonic stem cell technology; region X; targeted muta- genesis
INTRODUCTION
Vertebrate Hox genes are required for the proper organization and development of the body plan. Their correct function depends on the establishment in time and space of a precise pattern of partially overlapping expression domains in the embryo. The genetic mecha- nism(s) responsible for the sequential activation of these genes is largely unknown. However, it relies on a specific feature of this gene family, i.e., genes are controlled according to their genomic positions in the different complexes (colinearity; see Krumlauf, 1994, see references in van der Hoeven et al., 1996a,b).
Whatever process underlies colinearity, it is likely that it involves some cis-acting regulatory sequences to mediate either specific interactions, in the case of upstream factors, or less specific types of controls, such
as in the generation of high-order chromatin configura- tions. The characterization of such sequences, thus, may be an important step in our understanding of Hox gene regulation.
It is generally believed that DNA sequences that are highly conserved among different species and located outside transcribed regions serve some regulatory pur- poses (see, e.g., Burglin et al., 1987). Based on this assumption, the validity of which has been verified in many instances, we set out to sequence the ‘‘posterior’’
part of the HoxD complex in mammals, chick, and fish, i.e., the region containing AbdominalB (AbdB)-related genes (Izpisu´ a-Belmonte et al., 1991). Sequence com- parisons led to the definition of several conserved regions that were further analyzed by using a ‘‘classi- cal’’ reporter transgenic approach (Renucci et al., 1992;
Ge´rard et al., 1993; Beckers et al., 1996; He´rault et al., 1998). Deletion or mutation of particular sequences within large reporter transgenes usually modified some aspect of the regulation, thus allowing the attribution of precise function to DNA sequences. In this way, short stretches of DNA were identified to play a role either in the position of the expression boundaries (see, e.g., Ge´rard et al., 1993; Vogels et al., 1993), in germ layer specificity (Za´ka´ny et al., 1988; Kress et al., 1990;
Pu¨ schel et al., 1991), or in expression domains linked to particular structures, such as the limbs (Schughart et al., 1991; Eid et al., 1993) or the hindbrain (Sham et al., 1993; Studer et al., 1994; Po¨pperl et al., 1995).
Although the transgenic approach has proven to be powerful, it also contains intrinsic problems that should not be overlooked. For example, the almost systematic production of multiple copy insertions through DNA injection may produce a transgenic locus that is not directly comparable to the endogenous locus. Also, autoregulation and cross regulation, which have been demonstrated to be important for labial- or Deformed- related genes (Po¨pperl et al., 1995; Chan et al., 1997;
Grant sponsor: Canton de Gene`ve; Grant sponsor: Swiss National Research Fund; Grant sponsor: Human Frontier Program; Grant sponsor: Claraz and Cloetta Foundations.
Johannes Beckers is currently at The Jackson Laboratory, Bar Harbor, ME 04609.
*Correspondence to: Denis Duboule, Department of Zoology and Animal Biology, University of Geneva, Sciences III, Quai Ernest Ansermet 30, 1211 Geneva 4, Switzerland. E-mail: denis.duboule@zoo.
unige.ch
Received 15 April 1998; Accepted 8 May 1998 r1998 WILEY-LISS, INC.
Gould et al., 1997; Maconochie et al., 1997), may interfere with the interpretation of the results in the presence of both the endogenous locus and other inter- acting gene products. For these reasons, it is desirable that the final output of such transgenic approaches be controlled by a targeted approach within the endog- enous locus in vivo. In the case of Hox genes, few cases have been reported so far in which both approaches were used. In each of those cases, however, the results that were obtained by using the targeted approach confirmed partially or entirely the transgenic results (Ge´rard et al., 1996; Dupe´ et al., 1997; Za´ka´ny et al., 1997a,b; Studer et al., 1998; He´rault et al., 1998;
Gavalas et al., 1998).
Posterior Hoxd genes are important determinants of limb development. During limb budding, these genes are expressed with proximodistal and anteroposterior specificities and follow temporal colinearity (Dolle´ et al., 1989). The expression domains of these genes in limbs are complex and are established through several phases of activation, which are potentially controlled by different enhancer sequences (Duboule, 1994; Nelson et al., 1996; He´rault et al., 1998). In our search for regulatory sequences that are active during limb devel- opment, we identified two regions of high sequence similarities between tetrapods. One region was local- ized between Hoxd-13 and Hoxd-12 and was shown by transgenesis to drive expression of a Hoxd-12 reporter transgene in a restricted posterior domain of the grow- ing limb buds. Deletion of this region in vivo confirmed this result and showed that the corresponding limb expression domain of the Hoxd-12 gene was absent in mutant animals (He´rault et al., 1998). The second conserved sequence was isolated between Hoxd-11 and Hoxd-12 and was shown to drive expression of both Hoxd-11 (Beckers et al., 1996) and Hoxd-12 (He´rault et al., 1998) reporter transgenes in a proximal and poste- rior domain of the endogenous Hoxd-11 gene.
To evaluate the function of this sequence precisely in the regulation of posterior Hoxd genes, we set out to delete it from its endogenous context through an embry- onic stem (ES) cell-based strategy. Mice were recovered at the expected Mendelian ratio. Surprisingly, no abnor- mal phenotype was detected in the absence of the neomycin selection cassette. Likewise, expression pat- terns of the surrounding Hoxd genes remained un- changed in homozygous mutants. We conclude that such sequence may drive expression of Hoxd-11 or Hoxd-12 in an as yet unidentified domain (in late fetuses or adults). Alternatively, it is possible that a regulatory redundancy prevents us from easily assess- ing the functional potential of this sequence.
RESULTS Targeted Deletion of Region X
We previously identified a cis-acting regulatory ele- ment, region X (RX), located in the intergenic region between Hoxd-11 and Hoxd-12 and well conserved within tetrapods as well as in zebrafish. A detailed
analysis of this element in Hoxd-11 and Hoxd-12 trans- genes indicated that it was required to direct expres- sion of reporter genes in a proximal posterior domain of the developing forelimb in mouse embryos (Beckers et al., 1996; He´rault et al., 1998). However, the combined loss of function of group 11 genes and Hoxd-12 did not induce any abnormal phenotype in this precise region (Davis et al., 1995; Davis and Capecchi, 1996). It was therefore unclear whether this expression domain re- flected a genuine function of either Hoxd-11, Hoxd-12, or both genes. To reveal the function of RX, we deleted it from its endogenous position in the HoxD complex through homologous recombination in mouse ES cells (Fig. 1A,). Internal as well as external DNA probes (Fig.
1B) were used to identify four clones that had recom- bined the targeting vector at the expected locus. Male germ line chimeras were generated from two clones. In the F2 progeny, wild type, heterozygous (HoxDRXneo/1) and homozygous (HoxDRXneo/RXneo) animals were ob- tained at the expected ratio. Mice homozygous for the deletion of RX appeared outwardly normal. They were fertile and had life spans of at least 1 year.
HoxDRXneoAllele
Analysis of RX in transgenic mice suggested a func- tion in developing limbs (Beckers et al., 1996). We therefore investigated mutant mice for alterations of the limb skeletal pattern. Limb skeletons of homozy- gous mutants (HoxDRXneo/RXneo) were indistinguishable from wild type litter mates (not shown). Because previ- ous compound mutants revealed that AbdB-related Hoxd and Hoxa genes can compensate one another’s function (see, e.g., Davis and Capecchi, 1996; Za´ka´ny and Duboule, 1996), we genetically combined the ab- sence of RX with different Hox loss-of-function alleles.
When the HoxDRXneo allele was added to a Hoxa-13/
Hoxd-13 transheterozygous background, a strong en- hancement of the limb skeletal phenotype was scored.
Autopods (hands and feet) of Hoxa-13/Hoxd-13 trans- heterozygous mice regularly showed a loss of phalange 2 (P2) in digit V (minimus) as well as a fusion of phalanges 1 and 2 in digit I (thumb). In the distal row of the carpus, the d3 and d4 bones, were occasionally fused, and an additional element (d5) could occur (Fromental-Ramain et al., 1996; see Fig. 2A). In triple mutants (Hoxa-131/2, Hoxd-131/2, HoxDRXneo/1), these alterations became fully penetrant. In addition, mice of this genotype displayed an additional posterior digit- like structure (Fig. 2B; V with asterisk) as well as a strong reduction of the fifth metacarpal bone (Fig. 2B).
Likewise, a fusion of the proximal carpal bones, the navicular lunate and triangular bones, that occurred with incomplete penetrance in mice heterozygous for both Hoxa-11 and Hoxd-11 loss of function became fully penetrant in Hoxa-111/2, Hoxd-111/2, HoxDRXneo/1ani- mals (not shown).
In addition, HoxDRXneo/RXneomice displayed an ante- rior shift of the lumbosacral transition in the vertebral column. Whereas wild type mice showed six lumbar
vertebrae, the lumbar region of homozygous mutant mice (HoxDRXneo/RXneo) was reduced to five vertebrae, leading to an anterior shift of the sacral region (Fig.
3A,B). Normally, the third lumbar vertebra (L3) is the most rostral one to bear an accessory process. In homozygous mutant mice, an accessory process was found only down to the second lumbar vertebra, suggest- ing a posterior transformation of L3 into an L4-like morphology. These alterations in the vertebral column correlated with a clear anterior extension of the Hoxd-11 expression domain in mesodermal derivatives (Fig.
3C,D). Posterior transformations also occurred in het- erozygous mutants, although with lower penetrance, indicating that this phenotype was dose-dependent.
In contrast to this ectopic expression of Hoxd-11 in the trunk, the limb skeletal phenotype was associated with a down-regulation of Hox gene expression in developing limb buds. In wholemount in situ hybridiza- tions of homozygous HoxDRXneo/RXneomutant embryos at 11 days postcoitus (11 dpc), we observed a limb staining that was reduced in intensity for both Hoxd-11 and Hoxd-12. Hoxd-13, however, remained unchanged (not shown). This quantitative change in the expression of the two genes flanking the mutagenized sequence likely induced the skeletal alterations that were observed in compound mutants. Unlike the anterior shift of Hoxd-11 expression in the trunk, we did not observed any qualitative change in the distribution of these two gene transcript domains in limbs.
Deletion of RX Does Not Alter Hoxd Gene Expression
We subsequently excised the PGKneo selection cas- sette from the mutagenized locus by using the Cre recombinase in the recombined ES cells in vitro (Gu et al., 1993). This excision process generated the HoxDRX
allele, which was free of selection marker and harbored only a 261 base pair (bp) large deletion, including RX (Fig. 1C). This configuration was controlled by South- ern analysis and was confirmed further at the nucleo- tide level. The sequence of polymerase chain reaction (PCR) fragments that covered this region and that was obtained from homozygous mutant DNA demonstrated both the proper introduction of the mutagenized locus as well as the accurate excision of the PGKneo cassette.
Sequences at and around the deletion of RX were identical to the targeting vector, thus confirming the mutated nature of the locus. The excision of the selec- tion cassette, as expected, left behind a single loxP site.
No additional change in the nucleotide sequence was found in the subcloned genomic region (not shown).
Analysis of HoxDRXmutant mice revealed that nei- ther the limb skeletal alterations found in compound mutants with the HoxDRXneo allele nor the anterior vertebral transformation were present. In particular, limb skeletons of mice carrying the three alleles Hoxa- 131/2, Hoxd-131/2, and HoxDRX/1 were indistinguish- able from those of Hoxa-131/2/Hoxd-131/2 mice (Fig.
2A,C). Also, the vertebral column of HoxDRX/RX mice was normal (not shown). Therefore, the phenotypic alterations observed in association with the HoxDRXneo allele had been induced by the presence of the selection cassette in the locus.
We next assessed the regulatory function of RX by looking at Hoxd gene expression in HoxDRX/RXmutant mice. We initially focused on Hoxd-11 and monitored its expression by using wholemount in situ hybridizations.
We analyzed either 9.5-dpc embryos, a stage at which Hoxd-11 expression is first observed in the tail bud region, or 12.5-dpc embryos, a stage at which the expression pattern in limbs has developed fully. We did not observe any difference in Hoxd-11 expression be-
Fig. 1. Targeted mutagenesis of Region X (RX) in the murineHoxD complex. A: Targeting construct designed to delete RX. Below the targeting vector, a part of the wild typeHoxD complex is shown extending over the regions of homology. B: Structure of the locus after homologous recombination. Open arrows flanking the selection marker (neo) are loxP sites (34 base pairs; bp). External and internal probes for Southern analysis are depicted (neo and SE47 and PstI-HindIII fragment (PH),
respectively). C: Structure of the mutated RX locus after excision of the selection cassette (HoxDRX). After Cre treatment, a single loxP site flanked by a short stretch of polylinker sequences containingHindIII, XbaI, and SmaI restriction sites is left behind. The asterisks indicate the positions of the oligonucleotides used for polymerase chain reaction (PCR) in typing the offspring from established lines. Xh,XhoI; S, SmaI;
Xb,XbaI; E, EcoRI; H, HindIII.
Fig. 2. Phenotypes of the HoxDRXneo and HoxDRX alleles when combined withHoxa-13/Hoxd-13 transheterozygous mutant background.
A: Forelimb autopod of aHoxa-131/2/Hoxd -131/2mutant mouse. In digit I, phalanges 1 and 2 are fused (P1/P2), whereas, in digit V, P2 is missing.
In the distal carpal row, d3 and d4 are fused (d3/d4). Occasionally, an additional carpal element occurred posteriorly (the arrow indicates an incomplete split of the d3/d4 bone). B: Forelimb autopod of aHoxa-131/2/ Hoxd-131/2/HoxDRXneo/1 mutant mouse. In addition to the alterations
observed in the transheterozygous mutant (A), an additional digit-like structure is found posteriorly (V with asterisk), and the fifth metacarpal (mc5) is strongly abnormal. C: Forelimb autopod of aHoxa-131/2/Hoxd- 131/2/HoxDRX/1mutant mouse. The genotype of this mouse is the same as of that in B, except that the PGKneo cassette was removed. The autopod is indistinguishable from that ofHoxa-13/Hoxd-13 transheterozy- gous mutants (A).
Fig. 3. The PGKneo cassette induced posterior transformations in lumbar and sacral vertebrae. A: Wild type vertebral column. The third lumbar vertebra (L3) is the first to show an accessory process (arrow). B:
Vertebral column of a HoxDRXneo/RXneo mouse showing morphological posterior transformations. All vertebrae posterior to L5 adopt a morphology that resembles the posterior next vertebral element. For example, the normal L6 has features of the first sacral element (S1). Thus, the lumbosacral transition is shifted anteriorly. In addition, the second lumbar element now bears an accessory process (arrow). C: Wholemount in situ hybridization of a wild type embryo at 13.5 days postcoitus (dpc) using a Hoxd-11 probe. The arrow indicates the anterior limit of expression in the sclerotome. D: AHoxDRXneo/RXneo litter mate hybridized with the same probe.Hoxd-11 expression extends anteriorly in the sclerotome (arrow) but not in the central nervous system (CNS).
tween HoxDRX/RX mutants and their wild type litter mates. Unexpectedly, the proximal posterior domain of limb expression that had been associated with RX by using conventional transgenesis (LE1 in Beckers et al., 1996) was not lost upon deletion of this element from the endogenous locus (Fig. 4E,F). This indicated that RX is not critical to drive endogenous Hoxd-11 expres- sion into this limb domain.
Because RX was equidistant from both Hoxd-11 and Hoxd-12, we investigated its potential influence either on this latter gene (Fig. 4C,D) or on more distantly located members, such as Hoxd-13 (Fig. 4A,B), Hoxd- 10, and Hoxd-9. Wholemount in situ hybridizations did not reveal any modifications in the expression patterns of these genes in mutant fetuses. In particular, none of the alterations detected in trunk mesoderm or limb buds of HoxDRXneo mice were observed. We concluded that RX may not be required for the expression of AbdB-related Hoxd genes between 9.5 and 12.5 days of embryonic development.
Genitourinary System
Loss-of-function alleles of groups 10, 11, and 13 Hox genes have revealed their particular functions in the morphogenesis of the urogenital system. Uterine de- fects and homeosis of the vas deferens as well as cryptorchidism were associated with the inactivation of
Hoxa-11 and Hoxa-10 (Hsieh-Li et al., 1995; Rijli et al., 1995; Satokata et al., 1995; Gendron et al., 1997).
Notably, compound mutant mice for Hoxa-11 and Hoxd-11 loss-of-function alleles showed enhanced alter- ations in the male reproductive system, suggesting that Hoxd-11 is involved in the morphogenesis of this sys- tem. In addition, mice double homozygous for these two mutant alleles displayed severe renal malformations, leading to perinatal death (Davis et al., 1995). Like- wise, mice mutant for Hoxd-13 displayed abnormal morphogenesis of the prostate (Podlasek et al., 1997).
We investigated the role of RX in Hoxd gene expres- sion in these urogenital structures by using in situ hybridization on sections of 19-dpc mutant embryos.
Hoxd-11 and Hoxd-10 were expressed in the metaneph- ros and ureter of wild type embryos (Fig. 5A,E), and their expression was indistinguishable from that of control mice (Fig. 5B,F). Similarly, Hoxd-10 was ex- pressed in the prostatic primordia of both homozygous mutant and wild type embryos (Fig. 5A,D). In the urethra, Hoxd-11 again was expressed identically in both homozygous mutant and wild type litter mates (Fig. 5G,H). The situation was the same for Hoxd-12 (not shown). Therefore, in situ hybridization did not point to a potential role of RX in Hoxd gene expression in the urogenital system, at least not at these stages.
Fig. 4. Hoxd gene expression in HoxDRX/RXmutants and wild type 11.5-dpc litter mates. Wild type embryos are on the left (A,C,E), and homozygous mutants are on the right (HoxDRX/RX; B,D,F). Embryos are shown from two sides. A,B: Hybridization with aHoxd-13 probe. Expres- sion domains in fore- and hindlimbs are indistinguishable between wild
type and mutant embryos. The same holds true for expression in the genital eminence (arrow). Hybridizations with probes for eitherHoxd-12 (C,D) orHoxd-11 (E,F) revealed no differences between wild type and mutant embryos either in forelimb or hindlimb expression or in the anterior limits in the CNS or sclerotome.
In addition to its function during the morphogenesis of the reproductive tracts, Hoxa-11 expression oscil- lates with the estrous cycle in adult female uteri, where it is required for early stages of gestation (Gendron et al., 1997). We wanted to determine whether this func- tion was shared by paralogous genes by monitoring expression of Hoxd-11 in the uteri of adult females. We took advantage of an established mouse line that has lacZ reporter sequences inserted in the endogenous Hoxd-11 locus in the presence of an intact RX (Za´ka´ny and Duboule, 1996). LacZ expression was analyzed on histological sections of uteri either before pregnancy or at an early implantation stage (6.5 dpc). Similar to Hoxa-11 mRNA distribution, expression was restricted to the stromal cells underlying the luminal epithelium at 6.5 dpc (Fig. 6B). The reporter gene was not ex- pressed in the corresponding cells from nonpregnant females (Fig. 6A). These observations showed that Hoxd-11, like Hoxa-11, is expressed in uteri of adult mice and that its activation is dependent on stages of the estrous cycle.
We checked whether RX was required for expression of Hoxd-11 during pregnancy by producing mice homo- zygous for HoxDRX/RX and heterozygous for a Hoxa-11 loss of function (Small and Potter, 1993). Such animals were mated and produced litters of normal size. Accord- ingly, there was no evidence for an enhanced hypofertil- ity in this genetic combination. Interestingly, compound mutants mice for Hoxa-11 and Hoxd-11 loss-of-function alleles died shortly after birth due to the lack or hypoplasia of kidneys (Davis et al., 1995). To find out whether the addition of the HoxDRXallele to Hoxa-11 mutant mice would induce an increased lethality, we typed 126 mice derived from transheterozygous crosses 2–3 weeks after birth. The distribution of phenotypes was not significantly different from the expected ratios and did not reveal an increased lethality in association with a particular genotype (Table 1). Double-homozy- gous specimens appeared at the expected frequency. In addition, the kidneys of such mice were histologically normal (not shown).
We also examined urogenital tracts from adult double- homozygous mutant males (Hoxa-112/2/HoxDRX/RX). Dis- sected reproductive systems of Hoxa-11 homozygous mutants were indistinguishable from those of double- homozygous mice. They showed comparable degrees of vas deferens malformation and cryptorchidism. The male urogenital system is comprised of a variety of ducts and glands that join in a defined sequence to form the urethra. These structures are tightly packed, hinder- ing the assessment of morphological alterations by gross inspection. Urogenital systems of at least 6-week- old male mice from the four different genotypes (wild type, HoxDRX/RX, Hoxa-112/2/HoxDRX/RX, and Hoxa- 112/2) were examined by histology in serial transverse sections. An as yet unreported alteration was associ- ated consistently with the Hoxa-11 loss of function. In wild type mice, the ducts of the vas deferens joined at the midline before the vesicles of the glans ductus derefentis (or ampullary gland) started to fuse to form two new ducts that enclose the vas deferens. Thus, they appeared as two tubes contained within one another. In Hoxa-112/2mutant animals, the two tubes of the vas deferens did not join medially. Instead, the vesicles of the ampullary glands had already fused to the vas deferens before the two ducts joined at the midline (not shown). This alteration was understood as a premature fusion of the ampullary gland to the vas deferens.
However, this mutant phenotype was neither found in HoxDRX/RX mutants nor enhanced in urogenital tracts from double-homozygous mutants; hence, it is not documented here.
DISCUSSION
Sequence comparison between orthologous loci from different vertebrate species has been a powerful ap- proach to identify regulatory sequences. The regulatory potential of such sequences was most often assayed by conventional transgenesis, i.e., by using a reporter gene construct with or without particular cis-acting DNA sequences. This approach has been widely used in the case of the Hox complexes, in which regions of orthology were easy to define and were rather small. Conse- quently, several control sequences were isolated in this way and were shown to be responsible for important aspects of Hox gene expression (see, e.g., Bieberich et al., 1990; Kress et al., 1990; Pu¨ schel et al., 1990; Tuggle et al., 1990; Whiting et al., 1991; Sham et al., 1992;
Renucci et al., 1992; Zwartkruis et al., 1992; Charite´ et al., 1995; Ge´rard et al., 1993; Vogels et al., 1993;
Marshall et al., 1994; Studer et al., 1994; Aparicio et al., 1995; Morrison et al., 1995, 1997; Po¨pperl et al., 1995;
Bradshaw et al., 1996; Chan et al., 1997; Gould et al., 1997; Maconochie et al., 1997). Yet, the deduced func- tional potential was rarely assayed in the endogenous context, within the Hox complex itself, due to the technical difficulty of the ES cell-based mutagenesis.
However, in those cases in which regulatory sequences were studied by using both conventional and ES cell- based transgenesis, the results were usually compat- TABLE 1. Genotypes of Offspring From Crosses
Between Hoxa-111/2/HoxDRX/1Males and Femalesa Genotypes from
Hoxa-112/1/HoxDRX/1 intercrosses
Expected ratios (%)
Scored ratios (%)
Wild type 6.25 8.7
HoxDRX/1 12.5 16.7
HoxDRX/RX 6.25 5.7
Hoxa-111/2 12.5 8.7
Hoxa-112/2 6.25 7.9
Hoxa-111/2/HoxDRX/1 25.0 23.8
Hoxa-111/2/HoxDRX/RX 12.5 11.9
Hoxa-112/2/HoxDRX/1 12.5 9.5
Hoxa-112/2/HoxDRX/RX 6.25 7.1
aOne hundred twenty-six mice were collected and scored for their genotype at 2 or 3 weeks after birth. The ratio of each genotype is compared to the expected Mendelian ratio. RX, region X in the murine HoxD complex.
ible, i.e., the phenotypic consequences of the modifica- tion in vivo corresponded to some extent to the expectation generated by the transgenic approach, even though in vivo modifications had systematically weaker effects than those predicted (Ge´rard et al., 1996; Dupe´
et al., 1997; Za´ka´ny et al., 1997a,b; Gavalas et al., 1998;
He´rault et al., 1998; Studer et al. 1998).
Our extensive sequencing of the posterior part of the HoxD complex in mice, birds, and fish has revealed a rather low number of DNA sequences significantly conserved outside the RNA coding regions (Renucci et al., 1992; Ge´rard et al., 1993; van der Hoeven et al., 1996a,b; He´rault et al., 1998). These potential control elements were all studied by both conventional trans- genesis, using lacZ as a reporter gene, and ES cell- based mutagenesis. The deletion or fine mutagenesis of these sequences in their endogenous contexts had different consequences. The study of RXI, a Hoxd-12 control region, showed that it was required in develop- ing posterior limbs, as predicted from conventional transgenesis (He´rault et al., 1998). Likewise, mutagen- esis of a nuclear receptor binding site downstream of Hoxd-11 confirmed the repressive function of another of these regions (RIX; Ge´rard et al., 1996). In contrast, the deletion of RVIII did not exactly reproduce the trans- genic analysis. Although this region proved to be neces- sary for the activation of a Hoxd-11/lacZ transgene, its deletion in complexo did not prevent the activation of the endogenous Hoxd-11 gene, even though this activa- tion did not occur properly (Za´ka´ny et al., 1997b).
Similarly, mutagenesis of retinoic acid response ele- ments (RAREs) that are known to control the expres- sion of Hoxb-1 and Hoxa-1 did not induce phenotypes as strong as the transgenic approach may have predicted (Dupe´ et al., 1997; Gavalas et al., 1998; Studer et al., 1998). Here, we report the deletion of RX, a sequence located between Hoxd-11 and Hoxd-12. In this case, a strong discrepancy was observed between the results obtained by using the two approaches.
Phenotypes
Mice deleted for RX and containing the PGKneo selection cassette showed a clear phenotype in the vertebral column in which an anterior vertebral trans- formation correlated with a gain of function of Hoxd-11.
In the limbs, alterations were detected only in the presence of additional null alleles either for Hoxd-13 and Hoxa-13 in the digits or for Hoxa-11 and Hoxd-11 in the carpus. Because a slight reduction in the expression of both Hoxd-11 and Hoxd-12 was scored in RX mutant embryonic limbs, the reinforcement of these pheno- types can be explained by a local and partial loss of function. The absence of phenotype in RX homozygous simple mutants likely results from the functional redun- dancy known to occur amongst Hox genes, in particular during limb development (Davis and Capecchi, 1996;
Za´ka´ny et al., 1997a). Consequently, both gain-of- function and loss-of-function phenotypes were observed after deletion of RX.
However, these various phenotypic traits disap- peared upon excision of the selection cassette. This cassette contained the PGK promoter and was shown in similar cases to induce either gain-of-function or loss-of- function of genes located at the vicinity of the insertion site (see, e.g., Fiering et al., 1995; Hug et al., 1996;
McDevitt et al., 1997; Za´ka´ny et al., 1997b). It is thus clear that the deregulation in the expression of the neighboring Hoxd-11 and Hoxd-12 genes was caused by the presence of the PGK promoter. This raises once again the need for removing extra sequences from targeted loci, particularly when genes occur in tightly clustered configuration (see, e.g., Olson et al., 1996;
Za´ka´ny et al., 1997b). It also suggests that some previously described Hox phenotypes may require fur- ther analysis.
No Phenotype Associated With RX Deletion Mice homozygous for a deletion of RX showed no particular abnormality in their phenotypes, even after all major sites of expression of posterior Hoxd genes were carefully investigated. Surprisingly, the expres- sion domain that had been associated systematically with RX in transgenic mice (proximal limb) was not affected in the deleted animals. Potential explanations for this unexpected observation are not trivial. One possibility is that multiple sequences either contribute to the establishment of this expression domain or are capable on their own of driving expression therein and that only one such sequence was located on the reporter transgene. In such a scheme, the presence of the gene in the complex would allow for a compensation of the regulation due to a sequence located nearby, whereas an externally localized transgene would be silent in this domain upon deletion of this sequence. It is also pos- sible that RX is a target sequence for auto- or cross- regulation, whereas the activation of the gene is depen- dent on another control element located outside the transgene but within the complex. In this case, the absence of RX would suppress the transcription of the transgene in a particular domain, whereas the endog- enous copy would still respond to the activating mecha- nism.
Another explanation is that the deletion of RX in- duced a phenotype that we did not detect. Although we analyzed all of the sites where either Hoxd-11 or Hoxd-12 loss-of-function mutations were reported to have an effect, we cannot rule out the possibility that these Hox genes have important functions in as yet unidentified tissues or organs, in particular during adult life. It is also possible that the functional redun- dancy observed for Hoxd-11, for example, in the fore- arms and kidneys, equally applies to these unidentified expression sites, making their detection even more difficult. This issue will be addressed by using large deficiencies covering the HoxD complex. In such configu- rations, RX might reveal its genuine functional poten- tial.
Despite these tentative explanations, we would like to stress that RX is the largest sequence of such high sequence similarity reported so far in Hox intergenic regions. The conservation between zebrafish and mouse DNA (Beckers et al., 1996) suggested an important function that is shared by all vertebrates. The absence of a clear phenotype linked to its deletion is therefore an unexpected observation. This is especially surpris- ing when one considers that RX is the only conserved sequence identified between Hoxd-12 and Hoxd-11.
Interestingly, none of the control elements located within the AbdB-related Hoxd genes (group 9–13) turned out to be critical for the major expression features of the neighboring genes. Instead, deletion or mutagenesis of these elements lead to alterations in the modulation of Hoxd gene expression, such as slight differences either in the expression timing (Za´ka´ny et al., 1997b), in the position of the AP boundary (Ge´rard et al., 1996), or in the regulation of a restricted and specific expression domain (He´rault et al., 1998). How- ever, none of these modifications seriously challenged the major colinear activation of these genes in time and space. It thus appears that conserved regulatory se-
quences within the posterior HoxD complex are in- volved mainly in the refinement and secondary control of the expression patterns rather than in the original establishment of the transcript domains. This suggests that the control mechanism responsible for the sequen- tial activation of these genes during development may not necessarily rely on classical cis-acting regulatory elements that are detectable by using our strategy.
Transgenic vs. ES Cell-Based Approaches
The deletion of RX from its endogenous context failed to reproduce the effect that had been observed previ- ously on a transgenic context. This discrepancy, rather than pointing to the weakness of one particular system, may illustrate the need for both approaches in combina- tion. On the one hand, the random integration of multiple copies of a partial genomic locus in the pres- ence of a functional, endogenous counterpart is cer- tainly very far from physiological conditions, in particu- lar when clustered genes are analyzed individually. On the other hand, however, the clustering of regulatory sequences bearing some functional redundancy may prevent the dissection of their respective importance if
Fig. 5. A–H:Hoxd-11 and Hoxd-10 expression in wild type (left) and homozygous mutant (HoxDRX/RX; right) em- bryos at 19 dpc. The sections in A–D are hybridized with Hoxd-10, and those in E–H are hybridized with Hoxd-11. Both genes are expressed in the metanephros (kidney; kd) and the ureter (ur) of wild type (A,E) and mutant (B,F) embryos.
Hoxd-10 is expressed in the primordia of the prostate (p) in wild type (A) and mutant (D) embryos. In the urethra (ut), Hoxd-11 transcripts were also detected in both wild type (G) and mutant (H) embryos. In addition, both genes were still transcribed at 19 dpc in the spinal cord (sc) and midgut (mg) of mutants and wild type embryos.
they are analyzed in their endogenous configuration. In this view, a conventional transgenic approach may help to unmask a particular regulatory function by transpos- ing it outside a complex and redundant genomic environ- ment. This illustrates the necessity to carry out both types of analyses and suggests that contradictory re- sults may not always be inconsistent with the molecu- lar physiology of the system.
EXPERIMENTAL PROCEDURES Homologous Recombination and Mutant Lines
The deleted region in the targeting vector was identi- cal to the one in transgene md11D described in Beckers et al. (1996). The 261-bp-large deletion started 24 bp upstream and ended 10 bp downstream of the DNA sequence given in the same publication (see Fig. 3 in Beckers et al., 1996). The PGKneo selection cassette, which was flanked by two loxP sites in the same orientation, was inserted as a blunted BamHI/XhoI fragment, in a reverse orientation with respect to the transcription of Hoxd genes, into a blunted HindIII site
located 303 bp upstream of the deleted RX. The 58 fragment of homology extended over 5.7 kb from the selection marker insertion site up to a BamHI up- stream of Hoxd-12. The 38-flanking sequence covered 3.5 kb from RX to a BsmI site in the Hoxd-11 intron.
ES cells (D3; Doetschman et al., 1985) were cultured on embryonic fibroblasts in presence of LIF. ES cells (107) were electroporated with 40 µg of linearized targeting vector. G418 selection and amplification of the electroporated cells was done as described previ- ously (Joyner, 1993). Excision of the selection cassette from cells that had homologously recombined the target- ing constructs was carried out in vitro by electropora- tion of 10 µg of the pMC-Cre plasmid producing the Cre recombinase of bacteriophage P1 (Gu et al., 1993). ES cells were microinjected into C57Bl/6 blastocysts and transferred to pseudopregnant foster mothers by follow- ing standard procedures.
Recombinant clones were identified by Southern analysis. For an external probe, a PstI HindIII frag- ment (PH; Fig. 1B), in combination with a HindIII digest of the genomic DNA, was used. The excision of the PGKneo cassette was controlled by the combination of an XbaI digest and the use of probe SE47 (Fig. 1B).
Routinely, genomic typing of F2 mice from established lines was performed by PCR. The pair of primers is indicated by asterisks in Figure 1C. It covered the deletion as well as the loxP site left behind. With these primers, the HoxDRXallele was 440 bp, and the wild type was 670 bp large. PCR products were cloned and sequenced from adult mutant mice to finally control the experiment. Compound mutants were produced with previously published alleles, i.e., Hoxd-13 (Dolle´ et al., 1993), Hoxa-13 (Fromental-Ramain et al., 1996; a gift of P. Chambon), Hoxa-11 (Small and Potter, 1993; a gift of S. Potter), and the HoxDdel allele, a triple loss-of- function of Hoxd-13 to Hoxd-11 in cis (Za´ka´ny and Duboule, 1996).
In Situ Hybridization
Fixation, hybridization, and detection of Hoxd gene expression in wholemount embryos was carried out by using a probe concentration of between 10 ng/ml and 50 ng/ml. Detection of digoxigenin-11-UTP labelled ribo- probes was performed by using an alkaline phosphatase- conjugated antidigoxigenin antibody (Boehringer Mann- heim, Indianapolis, IN). The probes used have been published previously: Hoxd-11 (Ge´rard et al., 1996);
Hoxd-13 probe (Dolle´ et al., 1991), and Hoxd-12 (Izpisu´a- Belmonte et al., 1991). The Hoxd-10 riboprobe was transcribed from a 1,138-bp cDNA fragment (Renucci et al., 1992). In situ hybridizations on cryosections were performed by using standard protocols. Briefly, 19.5- dpc embryos were collected, embedded in optimum cutting temperature medium, frozen without fixation, and further processed after typing of extraembryonic membranes. Antisense riboprobes were labelled with (aS35)-CTP, followed by DNAse I digestion and partial alkaline hydrolysis. Riboprobes were purified using
Fig. 6. Detection ofb-galactosidase activity in histological sections of uteri from mice heterozygous for an insertion of thelacZ gene into the Hoxd-11 locus. A: Expression was detected in neither luminal epithelia (le) or stromal cells (asterisk) of nonpregnant uteri. Muscular layers (m) were routinely found positive. B: During early stages of gestation, at 6.5 dpc,lacZ expression was confined to stromal cells (asterisk) underlying the luminal epithelium (le).
quick-spin columns. Embryos were sectioned at 10 µm at (20°C, mounted on gelatin/chrome alum-subbed slides, and stored at (80°C. For hybridization, slides were fixed in 4% formaldehyde and acetylated in 0.25%
acetic anhydride in 0.1 M triethanolamine buffer, pH 8.
Without prehybridization, slides were incubated over- night at 52°C with 20,000–30,000 cpm/µl of denatured probe in hybridization solution containing 50% for- mamide. Slides were then washed several times at 55°C, treated with RNAse A at 37°C, transferred to standard saline citrate buffer, and finally dehydrated.
Slides were processed for emulsion autoradiography (NTB-2; Kodak, Rochester, NY) and developed and fixed after 2–3 weeks (D19 and AL4; Kodak). For histological examination, slides were stained in 0.02%
toluidine blue solution.
Histology and Skeletal Preparations
Male urogenital tracts were collected from typed adult mice, fixed for 1 day in Bouin’s solution, and washed in water for another day, dehydrated, cleared, and embedded in paraffin. Kidneys and testes were sectioned sagitally at 7–10µm, whereas the remaining urogenital tracts were cut transversally. Sections were stained in hematoxylin/eosin. Skeletal staining was carried out as described by Inouye (1976). Typed adult mice were stained for 3–4 days in 2% KOH and 0.1%
alizarin red. Soft tissues were then cleared in 2% KOH.
Uteri were isolated, fixed in 4% paraformaldehyde, and stained with an X-galactosidase cocktail. Stained samples were imbedded in paraffin, sectioned at 7–10 µm, and stained with eosin following standard proce- dures.
ACKNOWLEDGMENTS
We thank M. Friedli for her technical assistance as well as colleagues from the laboratory for their com- ments and for sharing reagents. We also thank Drs. S.
Potter and P. Chambon for making mice available to us.
The laboratory is supported by funds from the Canton de Gene`ve, the Swiss National Research Fund, the Human Frontier Program and the Claraz and Cloetta foundations.
REFERENCES
Aparicio S, Morrison A, Gould A, Gilthorpe J, Chaudhuri C, Rigby P, Krumlauf R, Brenner S. Detecting conserved regulatory elements with the model genome of the Japanese puffer fish, Fugu rubripes.
Proc. Natl. Acad. Sci. U.S.A. 1995;92:1684–1688.
Beckers J, Ge´rard M, Duboule D. Transgenic analysis of a potential Hoxd-11 limb regulatory element present in tetrapods and fish. Dev.
Biol. 1996;180:543–553.
Bieberich CJ, Utset MF, Awgulewitsch A, Ruddle FH. Evidence for positive and negative regulation of the Hox-3.1 gene. Proc. Natl.
Acad. Sci. U.S.A. 1990;87:8462–8466.
Bradshaw MS, Shashikant CS, Belting HG, Bollekens JA, Ruddle FH.
A long-range regulatory element of Hoxc8 identified by using the pClasper vector. Proc. Natl. Acad. Sci. U.S.A. 1996;93:2426–2430.
Burglin TR, Wright CV, De Robertis EM. Translational control in homeobox mRNAs? Nature 1987;330:701–702.
Chan SK, Ryoo HD, Gould A, Krumlauf R, Mann RS. Switching the in
vivo specificity of a minimal Hox-responsive element. Development 1997;124:2007–2014.
Charite´ J, de Graaff W, Vogels R, Meijlink F, Deschamps J. Regulation of the Hoxb-8 gene: Synergism between multimerized cis-acting elements increases responsiveness to positional information. Dev.
Biol. 1995;171:294–305.
Davis AP, Capecchi MR. A mutational analysis of the 5( HoxD genes:
Dissection of genetic interactions during limb development in the mouse. Development 1996;122:1175–1185.
Davis AP, Witte DP, Hsieh-Li HM, Potter SS, Capecchi MR. Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11. Nature 1995;375:791–795.
Doetschman TC, Eistetter H, Katz M, Schmidt W, Kemler R. The in vitro development of blastocyst-derived embryonic stem cell lines:
Formation of the visceral yolk sac, blood islands and myocardium. J.
Embryol. Exp. Morphol. 1985;87:27–45.
Dolle´ P, Izpisua-Belmonte JC, Falkenstein H, Renucci A, Duboule D.
Coordinate expression of the murine Hox-5 complex homeobox- containing genes during limb pattern formation. Nature 1989;342:
767–772.
Dolle´ P, Izpisua-Belmonte JC, Boncinelli E, Duboule D. 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 develop- ment. Mech. Dev. 1991;36:3–13.
Dolle´ P, Dierich A, LeMeur M, Schimmang T, Schuhbaur B, Chambon P, Duboule D. Disruption of the Hoxd-13 gene induces localized heterochrony leading to mice with neotenic limbs. Cell 1993;75:431–
441.
Duboule D. How to make a limb? Science 1994;266:575–576.
Dupe´ V, Davenne M, Brocard J, Dolle´ P, Mark M, Dierich A, Chambon P, Rijli P. In vivo functional analysis of the Hoxa-1 3( retinoic acid response element (3(RARE). Development 1997;124:399–410.
Eid R, Koseki H, Schughart K. Analysis of LacZ reporter genes in transgenic embryos suggests the presence of several cis-acting regulatory elements in the murine Hoxb-6 gene. Dev. Dyn. 1993;196:
205–216.
Fiering S, Epner E, Robinson K, Zhuang Y, Telling A, Hu M, Martin DI, Enver T, Ley TJ, Groudine M. Targeted deletion of 5(HS2 of the murine beta-globin LCR reveals that it is not essential for proper regulation of the beta-globin locus. Genes Dev. 1995;9:2203–2213.
Fromental-Ramain C, Warot X, Messadecq N, LeMeur M, Dolle´ P, Chambon P. Hoxa-13 and Hoxd-13 play a crucial role in patterning of the limb autopod. Development 1996;122:1087–1097.
Gavalas A, Studer M, Lumsden A, Rijli F, Krumlauf R, Chambon P.
Hoxa-1 and Hoxb-1 synergize in patterning the hindbrain, cranial nerves and second branchial arch. Development 1998;125:1123–
1136.
Gendron RL, Paradis H, Hsieh-Li HM, Lee DW, Potter SS, Markoff E.
Abnormal uterine stromal and glandular function associated with maternal reproductive defects in Hoxa-11 null mice. Biol. Reprod.
1997;56:1097–1105.
Ge´rard M, Duboule D, Za´ka´ny J. Structure and activity of regulatory elements involved in the activation of the Hoxd-11 gene during late gastrulation. EMBO J. 1993;12:3539–3550.
Ge´rard M, Chen JY, Gro¨nemeyer H, Chambon P, Duboule D, Za´ka´ny J.
In vivo targeted mutagenesis of a regulatory element required for positioning the Hoxd-11 and Hoxd-10 expression boundaries. Genes Dev. 1996;10:2326–2334.
Gould A, Morrison A, Sproat G, White R, Krumlauf R. Positive cross-regulation and enhancer sharing: Two mechanisms for specify- ing Hox expression patterns. Genes Dev. 1997;11:900–913.
Gu H, Zou YR, Rajewsky K. Independent control of immunoglobulin switch recombination at individual switch regions evidenced through Cre-loxP-mediated gene targeting. Cell 1993;73:1155–1164.
He´rault Y, Beckers J, Fraudeau N, Kondo T, Duboule D. In vivo analysis of a conserved Hoxd-12 regulatory element reveals global versus local modes of controls in the posterior HoxD complex.
Development 1998;125:1669–1677.
Hsieh-Li HM, Witte DP, Weinstein M, Branford W, Li H, Small K, Potter SS. Hoxa 11 structure, extensive antisense transcription, and function in male and female fertility. Development 1995;121:1373–
1385.
Hug BA, Wesselschmidt RL, Fiering S, Bender MA, Epner E, Groudine M, Ley TJ. Analysis of mice containing a targeted deletion of beta-globin locus control region 5( hypersensitive site 3. Mol. Cell Biol. 1996;16:2906–2912.
Inouye M. Differential staining of cartilage and bone in fetal mouse skeleton by alcian blue and alizarine red. Cong. Anom. 1776;16:171–
173.
Izpisua-Belmonte JC, Falkenstein H, Dolle´ P, Renucci A, Duboule D.
Murine genes related to the Drosophila AbdB homeotic genes are sequentially expressed during development of the posterior part of the body. EMBO J. 1991;10:2279–2289.
Joyner AL. Gene Targeting. A Practical Approach. Oxford: IRL Press at Oxford University Press, 1993.
Kress C, Vogels R, De Graaff W, Bonnerot C, Meijlink F, Nicolas JF, Deschamps J. Hox-2.3 upstream sequences mediate lacZ expression in intermediate mesoderm derivatives of transgenic mice. Develop- ment 1990;109:775–786.
Maconochie MK, Nonchev S, Studer M, Chan SK, Po¨pperl H, Sham MH, Mann RS, Krumlauf R. Cross-regulation in the mouse HoxB complex: The expression of Hoxb2 in rhombomere 4 is regulated by Hoxb1. Genes Dev. 1997;11:1885–1895.
Marshall H, Studer M, Po¨pperl H, Aparicio S, Kuroiwa A, Brenner S, Krumlauf R. A conserved retinoic acid response element required for early expression of the homeobox gene Hoxb-1. Nature 1994;370:567–
571.
McDevitt MA, Shivdasani RA, Fujiwara Y, Yang H, Orkin SH. A
‘‘knockdown’’ mutation created by cis-element gene targeting re- veals the dependence of erythroid cell maturation on the level of transcription factor GATA-1. Proc. Natl. Acad. Sci. U.S.A. 1997;94:
6781–6785.
Morrison A, Chaudhuri C, Ariza-McNaughton L, Muchamore I, Kuroiwa A, Krumlauf R. Comparative analysis of chicken Hoxb-4 regulation in transgenic mice. Mech. Dev. 1995;53:47–59.
Morrison A, Ariza McNaughton L, Gould A, Featherstone M, Krumlauf R. HOXD4 and regulation of the group 4 paralog genes. Develop- ment 1997;124:3135–3146.
Nelson EC, Morgan BA, Burke AC, Laufer E, DiMambro E, Murtaugh LC, Gonzales E, Tessarollo LF, Tabin C. Analysis of Hox gene expression in the chick limb bud. Development 1996;122:1449–
1466.
Olson EN, Arnold HH, Rigby PW, Wold BJ. Know your neighbors:
Three phenotypes in null mutants of the myogenic bHLH gene MRF4. Cell 1996;85:1–4.
Podlasek C, Duboule D, Bushman W. Male accessory sex organ morphogenesis is altered by loss of function of Hoxd-13. Dev. Dyn.
1997;208:454–465.
Po¨pperl H, Bienz M, Studer M, Chan SK, Aparicio S, Brenner S, Mann RS, Krumlauf R. Segmental expression of Hoxb-1 is controlled by a highly conserved autoregulatory loop dependent upon exd/pbx. Cell 1995;81:1031–1042.
Pu¨ schel AW, Balling R, Gruss P. Position-specific activity of the Hox1.1 promoter in transgenic mice. Development 1990;108:435–442.
Pu¨ schel AW, Balling R, Gruss P. Separate elements cause lineage restriction and specify boundaries of Hox-1.1 expression. Develop- ment 1991;112:279–287.
Renucci A, Zappavigna V, Za´ka´ny J, Izpisua-Belmonte JC, Burki K, Duboule D. Comparison of mouse and human HOX-4 complexes defines conserved sequences involved in the regulation of Hox-4.4.
EMBO J. 1992;11:1459–1468.
Rijli FM, Matyas R, Pellegrini M, Dierich A, Gruss P, Dolle´ P,
Chambon P. Cryptorchidism and homeotic transformations of spinal nerves and vertebrae in Hoxa-10 mutant mice. Proc. Natl. Acad. Sci.
U.S.A. 1995;92:8185–8189.
Satokata I, Benson G, Maas R. Sexually dimorphic sterility pheno- types in Hoxa10-deficient mice. Nature 1995;374:460–463.
Schughart K, Bieberich CJ, Eid R, Ruddle FH. A regulatory region from the mouse Hox-2.2 promoter directs gene expression into developing limbs. Development 1991;112:807–811.
Sham MH, Hunt P, Nonchev S, Papalopulu N, Graham A, Boncinelli E, Krumlauf R. Analysis of the murine Hox-2.7 gene: Conserved alternative transcripts with differential distributions in the nervous system and the potential for shared regulatory regions. EMBO J.
1992;11:1825–1836.
Sham MH, Vesque C, Nonchev S, Marshall H, Frain M, Gupta RD, Whiting J, Wilkinson D, Charnay P, Krumlauf R. The zinc finger gene Krox20 regulates HoxB2 (Hox2.8) during hindbrain segmenta- tion. Cell 1993;72:183–196.
Small KM, Potter SS. Homeotic transformations and limb defects in Hox A11 mutant mice. Genes Dev. 1993;7:2318–2328.
Studer M, Po¨pperl H, Marshall H, Kuroiwa A, Krumlauf R. Role of a conserved retinoic acid response element in rhombomere restriction of Hoxb-1. Science 1994;265:1728–1732.
Studer M, Gavalas A, Marshall H, Ariza-McNaughton L, Rijli F, Chambon P, Krumlauf R. Genetic interactions between Hoxa-1 and Hoxb-1 reveal new roles in regulation of early hindbrain patterning.
Development 1998;125:1025–1036.
Tuggle CK, Za´ka´ny J, Cianetti L, Peschle C, Nguyen-Huu MC.
Region-specific enhancers near two mammalian homeo box genes define adjacent rostrocaudal domains in the central nervous system.
Genes Dev. 1990;4:180–189.
van der Hoeven F, Sordino P, Fraudeau N, Izpisua-Belmonte JC, Duboule D. Teleost HoxD and HoxA genes: Comparison with tetra- pods and functional evolution of the HOXD complex. Mech. Dev.
1996a;54:9–21.
van der Hoeven F, Za´ka´ny J, Duboule D. Gene transposition in the HoxD complex reveal a hierarchy of regulatory controls. Cell 1996b;85:1025–1035.
Vogels R, Charite J, de Graaff W, Deschamps J. Proximal cis-acting elements cooperate to set Hoxb-7 (Hox-2.3) expression boundaries in transgenic mice. Development 1993;118:71–82.
Whiting J, Marshall H, Cook M, Krumlauf R, Rigby PW, Stott D, Allemann RK. Multiple spatially specific enhancers are required to reconstruct the pattern of Hox-2.6 gene expression. Genes Dev.
1991;5:2048–2059.
Za´ka´ny J, Duboule D. Synpolydactyly in mice with a targeted defi- ciency in the Hoxd complex. Nature 1996;384:69–71.
Za´ka´ny J, Tuggle CK, Patel MD, Nguyen-Huu MC. Spatial regulation of homeobox gene fusions in the embryonic central nervous system of transgenic mice. Neuron 1988;1:679–691.
Za´ka´ny J, Fromental-Ramain C, Warot X, Duboule D. Regulation of number and size of digits by posterior Hox genes: A dose-dependent mechanism with potential evolutionary implications. Proc. Natl.
Acad. Sci. U.S.A. 1997a;94:13695–13700.
Za´ka´ny J, Ge´rard M, Favier B, Duboule D. Deletion of a HoxD enhancer induces transcriptional heterochrony leading to transposi- tion of the sacrum. EMBO J. 1997b;16:4393–4402.
Zwartkruis F, Hoeijmakers T, Deschamps J, Meijlink F. The murine Hox-2.4 promoter contains a functional octamer motif. Nucleic Acids Res. 1992:20:1599–1606.
