- Email: [email protected]
Hox cluster polarity in early transcriptional availability: a high order regulatory level of clustered Hox genes in the mouse
Mechanisms of Development 119 (2002) 81–90 www.elsevier.com/locate/modo
Hox cluster polarity in early transcriptional availability: a high order regulatory level of clustered Hox genes in the mouse Bernard A.J. Roelen 1, Wim de Graaff, Sylvie Forlani 2, Jacqueline Deschamps* Hubrecht Laboratory, Netherlands Institute for Developmental Biology, Uppsalalaan 8, 3584 CT, Utrecht, The Netherlands Received 27 June 2002; received in revised form 23 August 2002; accepted 28 August 2002
Abstract The molecular mechanism underlying the 3 0 to 5 0 polarity of induction of mouse Hox genes is still elusive. While relief from a clusterencompassing repression was shown to lead to all Hoxd genes being expressed like the 3 0 most of them, Hoxd1 (Kondo and Duboule, 1999), the molecular basis of initial activation of this 3 0 most gene, is not understood yet. We show that, already before primitive streak formation, prior to initial expression of the first Hox gene, a dramatic transcriptional stimulation of the 3 0 most genes, Hoxb1 and Hoxb2, is observed upon a short pulse of exogenous retinoic acid (RA), whereas it is not in the case for more 5 0 , cluster-internal, RA-responsive Hoxb genes. In contrast, the RA-responding Hoxb1lacZ transgene that faithfully mimics the endogenous gene (Marshall et al., 1994) did not exhibit the sensitivity of Hoxb1 to precocious activation. We conclude that polarity in initial activation of Hoxb genes reflects a greater availability of 3 0 Hox genes for transcription, suggesting a pre-existing (susceptibility to) opening of the chromatin structure at the 3 0 extremity of the cluster. We discuss the data in the context of prevailing models involving differential chromatin opening in the directionality of clustered Hox gene transcription, and regarding the importance of the cluster context for correct timing of initial Hox gene expression.Interestingly, Cdx1 manifested the same early transcriptional availability as Hoxb1. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Mouse Hox genes during gastrulation; Transcriptional regulation of the Hox genes; Retinoid signalling and transcriptional initiation of the Hox genes
1. Introduction Sequential activation of mouse Hox genes begins at the late primitive streak stage and is tightly coupled with the progress of gastrulation and laying down of the future definitive Hox expressing tissues (Duboule, 1992, 1994; Deschamps et al., 1999). A hierarchy of genetic control mechanisms is thought to orchestrate the subsequent establishment of the definitive Hox expression domains that are crucial for correct embryonic patterning. After the first Hox gene is initially expressed, the release of a cluster-encompassing repression has been shown to successively allow 3 0 to 5 0 Hoxd genes to be transcribed (Kondo and Duboule, 1999). This gradual derepression is thought to precede the numerous regulatory steps directing spatially specific subsets of the complex expression patterns, and involving enhancers * Corresponding author. Tel.: 131-30-2121935; fax: 131-30-2516464. E-mail address: [email protected] (J. Deschamps). 1 Present address. Division of Cellular Biochemistry, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. 2 Present address. De´partement de Biologie Mole´culaire du De´veloppement, Institut Pasteur, 25 rue du Dr. Roux, 75124 Paris, France. Abbreviations: E6.2, embryonic day 6.2; RA, retinoic acid; RAR, retinoic acid receptor; RARE, retinoic acid response element; A-P, antero-posterior
often able to function independently of the cluster context. Progressive escape of Hoxd genes from the global repression, exerted from a remote 5 0 sequence, has been proposed to be facilitated by a gradual opening of the chromatin structure from 3 0 to 5 0 (Van der Hoeven et al., 1996; Kondo and Duboule, 1999). Among the cis-acting, gene-specific regulatory sequences that subsequently would become sequentially accessible and operative are Retinoic Acid Response Elements (RAREs). These RAREs, present in 3 0 and more 5 0 neural enhancers, mediate gene stimulation by endogenous retinoic acid (RA) signals, in a progressively later time window (Gould et al., 1998; Zhang et al., 2000). Transcription of the earliest Hoxb gene, Hoxb1, is activated in the posterior part of the primitive streak at the late streak stage (embryonic day 7.2, E7.2) (Murphy et al., 1989; Frohman et al., 1990). Hoxb1 has been shown to respond to RA subsequently to initial transcription (E7.75, late head fold stage, and later) by rostrally extending its anterior expression boundary in the presumptive hindbrain (Marshall et al., 1992; Conlon and Rossant, 1992). Hoxb2 is regulated very similarly to Hoxb1 at early stages. Both genes share some regulatory elements, a Hoxb2 RAresponse being mediated by Hoxb1 (Maconochie et al., 1997). In addition, a Hoxb2 RARE has recently been iden-
0925-4773/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(02)00329-5
82
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
Fig. 1. Comparison between the expression patterns of Hoxb1 and Hoxb1lacZ in early mouse embryos. (A–C) Whole mount in situ hybridisation with a Hoxb1 probe, and (D–F), Xgal staining of Hoxb1/lacZ activity. Developmental stages were early late streak stage (E7.2, A,D), slightly more advanced early allantoic bud stage (E7.5, B,E), and neural plate stage (E7.75, C,F). Hoxb1/lacZ faithfully recapitulates the endogenous Hoxb1 expression pattern. Posterior is to the right in all panels.
tified, which initiates Hoxb2 expression in rhombomere 4 (cited in Manzanares et al., 2001). Hoxb3 shares regulatory regions with both Hoxb2 and Hoxb4 (Kwan et al., 2001). Expression of more 5 0 Hox genes starts in posterior embryonic tissues at progressively later stages. Hoxb4 is activated at the neural plate stage (E7.75) (Wilkinson et al., 1989; Gould et al., 1998), Hoxb6 at the early head fold (Becker et al., 1996) and Hoxb8 at a more advanced head fold stage (E7.75) (Deschamps and Wijgerde, 1993). Hoxb4 was shown to be RA-sensitive from E8 (Gould et al., 1998), and Hoxb6 and Hoxb8 from E9.5 and E10.5, respectively (Oosterveen et al., submitted).
In our experiments we set out to probe the transcriptional availability of 3 0 versus more 5 0 Hoxb genes very early in development, before any Hox gene starts to be transcribed. For this purpose we cultured carefully staged postimplantation embryos for a short time in the presence of RA as an external stimulus. We show that precocious induction of Hoxb1 and Hoxb2 by RA is operational at E6.0 already, before primitive streak formation and long before the gene is transcriptionally initiated. More 5 0 Hoxb genes, among which the later RA-responsive Hoxb4, do not react to this early stimulation. Furthermore, we show that a Hoxb1 transgene perfectly mimicking the endogenous gene in its expres-
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
83
Fig. 2. Hoxb gene expression in cultured pre- and early-streak stage embryos in control medium (A,C,E,G,I,K) and after a 4 h exposure to 10 26 M RA (B,D,F,H,J,L), detected by whole mount in situ hybridisation. A,B, Hoxb1; C,D, Hoxb2; E,F, Hoxb3; G,H, Hoxb4; I,J, Hoxb6 and K,L, Hoxb8. Posterior is to the right in all panels. While Hoxb1 and Hoxb2 are strongly induced by early RA-exposure, no induction of Hoxb3, Hoxb4, Hoxb6 and Hoxb8 is visible after a 4-h RA treatment at the midstreak stage (E 6.5). Internal positive controls of the probes in these hybridisations were obtained by hybridising E8.5 embryos (not shown) and allowing the staining reaction until a good signal was visible. Bar is 100 mm.
sion features and post-initiation RA sensitivity, did not react to the precocious transcriptional stimulation, suggesting that the early transcriptional availability is a cluster property. Finally we demonstrate that one of the Cdx gene family members, a relative of the Hox genes, also exhibits the extreme early transcriptional availability of Hoxb1.
2. Results 2.1. Exogenous RA rapidly and precociously activates Hoxb1 and Hoxb2 prior to primitive streak formation Hoxb1 is normally transcriptionally activated in the posterior part of the primitive streak, at the junction between embryonic and extraembryonic tissues, at the late streak stage (E7.2). Expression is first confined to posteriormost mesoderm and streak ectoderm, and later spreads rostrally to more anterior regions and laterally in the mesodermal wings (Murphy et al., 1989; Frohman et al., 1990) (Fig. 1A–C) We exposed pre-streak (E6.0) and early-streak (E6.2) embryos in culture, thus before any Hox gene was expressed, to 10 26 M all-trans-retinoic acid (RA) for 4 h, and analysed Hox gene expression by whole mount in situ hybridisation. Hoxb1 expression was strongly activated in these conditions (Fig. 2A, B). Hoxb2 expression was induced as well (Fig. 2C, D), whereas the expression of Hoxb3, Hoxb4, Hoxb6 and Hoxb8 was not induced (Fig. 2E–L). For Hoxb4, Hoxb6
and Hoxb8, which start to be expressed later than Hoxb1 we examined gene expression at progressively later time points, until stages closely preceding normal transcriptional initiation, and did not observe precocious induction upon either a 4 h, or a 7 h-RA treatment (not shown). RA used as an exogenous trigger to sense whether clustered Hox genes are available for transcription, thus rapidly and precociously turns on Hoxb1, and the partially Hoxb1-dependent Hoxb2. In contrast, this trigger does not affect Hoxb3, Hoxb4, Hoxb6 and Hoxb8, in spite of the fact the latter have been shown to be RA-sensitive at later stages. 2.2. Precocious expression of Hoxb1 initially follows spatiotemporal dynamics similar to that of the endogenous gene We examined the spatio-temporal dynamics of the precociously induced Hoxb1 expression, to compare it with the normal evolution of the transcription domain of endogenous Hoxb1. Pre-streak (E6.0) embryos were exposed to RA in culture for periods between 30 min and 4 h. Already after a 30-min exposure, Hoxb1 was precociously induced in the presumptive posterior epiblast at the junction between embryonic and extraembryonic ectoderm (Fig. 3A, B). After culture in these conditions for 1 h, Hoxb1 transcripts had expanded towards antero-distal and lateral positions (Fig. 3C). From the localisation, in whole mount and sectioned embryos, of the expression signal in embryos that had formed an early streak, we conclude that the RAinduced transcription of Hoxb1 in the pre-streak embryos
84
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
begins and spreads from posterior to distal and lateral positions. This spreading of the precocious expression domain of Hoxb1 therefore initially follows spatio-temporal dynamics resembling the expansion of endogenously induced Hoxb1 transcription, but more than 24 h in advance of normal expression. From a 2-h RA exposure onwards, the entire epiblast expressed Hoxb1 at high levels (Fig. 3D) in embryos treated at the prestreak stage, while the most antero-distal region remained non-induced in embryos treated at the early streak stage. Precocious expression of Hoxb1 mainly occurred in the epiblast (Fig. 3F, H) whereas
normally induced transcription of the gene takes place at a later stage in the mesoderm and streak epiblast (Fig. 3G). 2.3. A Hoxb1lacZ transgene, which is RA responsive at later stages, does not react to a pre-initiation RA stimulation We investigated whether a Hoxb1lacZ transgene, able to entirely recapitulate endogenous Hoxb1 expression (Fig. 1D–F), and to respond to RA at later stages (Marshall et al., 1994) exhibited the same extreme sensitivity to early RA activation as endogenous Hoxb1. The transgene used
Fig. 3. Timing of precocious RA-induced Hoxb1 expression (A–D). Hoxb1 expression in a pre-streak control embryo (A) and in embryos cultured for 30 min (B), 1 h (C) and 2 h (D) in the presence of RA. Transversal sections from control and RA-treated embryos (E–H) hybridised with a Hoxb1 probe. Control E6.5 (early streak) embryo (E). E6.5 embryo treated with RA for 4 h (F). Control E7.5 (late streak) embryo (G). E7.5 embryo treated with RA for 4 h (H). Posterior is to the right. Bar is 100 mm in A–F, and 200 mm in G,H. m, mesoderm. e, epiblast.
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
85
Fig. 4. A Hoxb1lacZ transgene is not induced by short RA exposure in early streak stage embryos. A,C,E, control embryos, and B,D,F, embryos exposed to RA for 4 h (B,D) or 6 h (F), assayed for b-galactosidase activity (E,F) or lacZ mRNA expression (A–D). Expression of Hoxb1/lacZ is not induced after RAexposure, in contrast to that of endogenous Hoxb1 (A,B). This Hoxb1/lacZ transgene is induced (anteriorly and laterally extended expression domain) upon a later, post-initiation (E7.5, head fold stage) exposure to RA (H compared to G). Bar is 100 mm in A–F, and 200 mm in G,H.
was a 21.6 kb genomic fragment where lacZ is inserted in frame in the first exon of Hoxb1. It contains, besides the proximal 3 0 and 5 0 RARE elements (Marshall et al., 1994; Studer et al., 1994) the more distal 3 0 RARE that mediates RA induction of Hoxb1 in gut endoderm later in development (Huang et al., 1998). We generated transgenic lines stably expressing this Hoxb1lacZ transgene. As previously described (Marshall et al., 1992, 1994) the Hoxb1lacZ transgene was expressed indistinguishably from the endogenous gene (Fig. 1), and responded to post-initiation RA exposure by rostrally and laterally expanding its expression domain (Fig. 4G, H). We compared endogenous Hoxb1 and Hoxb1lacZ expression after exposure of pre- and early streak stage embryos to RA in culture. No induction of Hoxb1lacZ mRNA and protein accumulation was observed, in conditions where endogenous Hoxb1 was strongly induced (Fig. 4C–F compared to Fig. 4A, B). The Hoxb1lacZ transgene responded to post-initiation RA exposure (Fig. 4G, H), demonstrating its sensitivity to RA at later stages. There
was therefore a dramatic early difference (at E6.2) between transgene and endogenous gene in their ability to be transcriptionally activated, although both responded to RA after Hoxb1 had been normally initiated (E7.5). 2.4. Cdx1 also quickly responds to a short RA pulse by a strong and precocious induction prior to primitive streak formation The mouse homologues of Drosophila caudal, the Cdx genes, are involved in axial patterning and are known to regulate Hox gene expression (Subramanian et al., 1995; Charite´ et al., 1998; van den Akker et al., 2002). Cdx genes have been suggested to be ancient relatives of the Hox genes (Brooke et al., 1998). Cdx1 is initially expressed very similarly to Hoxb1 in time and place. Transcription of Cdx1 is normally initiated in the posterior primitive streak epiblast and mesoderm at the late streak stage. Expression of Cdx1 spreads in a Hox-like manner to eventually reach its
86
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
rostral expression at the level of the preotic sulcus at early somite stages, before becoming down-regulated in the rostral spinal cord and sclerotomes (Meyer and Gruss, 1993). Cdx1 has also been shown to be a direct RA target in vivo at time points following transcriptional initiation (Houle et al., 2000; Prinos et al., 2001). We examined whether Cdx1, like Hoxb1, quickly responded to RA exposure by a strong and precocious induction. Treatment of pre(E6.0) and early streak (E6.2) embryos with exogenous RA for 4 h led to strong and precocious Cdx1 transcription (Fig. 5A, B). This extreme early sensitivity of Cdx1 to exogenous RA is not shared by the other two Cdx genes, Cdx2 (James et al., 1994) and Cdx4 (Gamer and Wright, 1993). No difference in the time, location and level of transcription of Cdx2 (Fig. 5C, D) and Cdx4 (Fig. 5E, F) was noticed between RAtreated and untreated embryos. Therefore, Cdx1 appears to be the only Cdx gene to be precociously transcriptionally activated by RA like Hoxb1. 2.5. Cdx1 does not mediate the early inducibility of Hoxb1 to RA Since both Hoxb1 and Cdx1 are strongly RA-inducible at
pre- and early streak stages, and because Cdx1 is a Hox regulator, we tested whether precocious induction of Hoxb1 by exogenous RA was dependent on Cdx1. We assayed Hoxb1 expression upon RA exposure in Cdx1 null mutant embryos (Subramanian et al., 1995). The same strong and precocious induction of Hoxb1 was measured after early exposure of wild type and Cdx1 null embryos to RA in culture (not shown). This indicates that Cdx1 does not mediate the response of Hoxb1 to RA. We also found that initial Hoxb1 and Hoxb8 expression was not altered in Cdx mutants lacking both Cdx1 alleles and one Cdx2 allele (data not shown), suggesting that Cdx gene products are not mediating endogenous Hox gene initiation during gastrulation.
3. Discussion 3.1. Pre-initiation inducibility reveals a difference in permissivity to transcription before Hox expression begins The observation that 3 0 -most Hox genes rapidly responded to a short pulse of RA before primitive streak formation whereas cluster internal- and 5 0 Hox genes only
Fig. 5. The expression of Cdx1, but not that of Cdx2 and Cdx4, is induced precociously by RA at pre/early streak stages. Cdx1 (A,B), Cdx2 (C,D) and Cdx4 (E,F) expression in cultured control embryos (A,C,E) and embryos cultured in the presence of RA (B,D,F). Posterior is on the right. Bar is 100 m m.
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
did so after endogenous expression was initiated, later during gastrulation, suggests that endogenous 3 0 and 5 0 Hoxb genes differ much more in their response to RA exposure at pre-initiation time points than they do after Hoxb gene expression has started. One possible molecular mechanism for this difference in availability for transcription at early stages might involve a pre-existing difference in the degree of chromatin opening at 3 0 versus internal and 5 0 positions in the Hoxb cluster, before the first Hox gene is activated. Exogenous RA, RA receptors and their co-factors might have access to the 3 0 RARE sequences at the early stages, but be denied access to sequences further 5 0 in the cluster. From the moment 3 0 Hox genes are initially transcribed, a further chromatin opening from 3 0 to 5 0 in the cluster would facilitate progressive release from global silencing, and account for proper timing of activation of more 5 0 Hox genes (Van der Hoeven et al., 1996; Kondo and Duboule, 1999, Kmita et al., 2000). Our observations suggest that the polarity of initial expression may be accounted for by a greater availability of 3 0 genes to the transcriptional machinery, long before Hox genes are expressed for the first time in the embryo. Experiments by Bel-Vialar et al. (2000) suggested that Polycomb-group Gene (Pc-G) products, responsible for a gradually more packed chromatin structure from 3 0 to 5 0 in the clusters, could be the system that controls the accessibility to the Hox RAREs, and differentially modulates temporal activation of the 3 0 to 5 0 Hox genes, both in the absence and in the presence of exogenous RA. Our data support the model of Bel-Vialar et al. (2000), and suggest that in the wild type PcG context, a functional difference exists between the degree of accessibility of the most 3 0 compared with more 5 0 RAsensitive Hox transcription start sites, at stages prior to initial Hox transcription. Alternatively, rather than passively getting access to the RAREs, exogenous RA might actively open the chromatin structure on the 3 0 side of the cluster, and anticipate an opening role of endogenous RA in the normal situation. RA/ RARs, with co-factors and recruited histone acetyl transferases are indeed capable of actively opening the chromatin structure (Bhattacharrya et al., 1997). In such a case, hypersensitivity of 3 0 Hoxb genes to RA-induced transcription would imply that the 3 0 side of the Hox cluster is more sensitive to the ‘opening’ activity of RA. Differential sensitivity to chromatin opening could also depend on the Pc-G epigenetic system. 3.2. Cdx1 shares early susceptibility to activation with Hoxb1 Evidence has accumulated to propose that Cdx genes and Hox genes are evolutionary closely related. Duplication of an ancestral gene cluster would have given rise to the Hox and ParaHox clusters, first identified in Amphioxus (Brooke et al., 1998). Cdx1 is expressed initially as early and as anteriorly as Hoxb1, and we proposed (van den Akker et
87
al., 2002) that Cdx genes possess 3 0 Hox structural and regulatory features, and would be paralogous to relatively anterior Hox genes. The present data further strengthen this view since Cdx1, like Hoxb1 seems to be immediately available for transcription before gene expression is normally initiated, as early as in pre- and early-streak stage embryos. The regulatory sequences of Cdx2 and Cdx4 apparently have diverged away from the initial Hox-like features. Cdx2 has acquired an early expression domain in extraembryonic ectoderm (Beck et al., 1995) before the gene is first expressed in a Hox-like manner in the embryo proper. In spite of the amazing similarity in the way all three Cdx genes and Hox genes establish their expression domains (Deschamps et al., 1999), considerable differences have arisen between Cdx 1 and its paralogues, such as the difference in sensitivity to early RA exposure. In spite of its extreme RA-sensitivity, Cdx1 was shown not to mediate the early response of Hoxb1 to RA. In fact, our experiments strongly suggest that Cdx gene products, which positively regulate the Hox genes during the establishment of their expression domains (Subramanian et al., 1995, Charite´ et al., 1998, Van den Akker et al., 2002), are not involved in the earliest initiation of Hox gene transcription.
3.3. Sensitivity of Hoxb1 to early RA stimulation occurs at a level hierarchically higher than the post-initiation RAresponse of individual Hoxb genes Previous work had made it clear that 3 0 to 5 0 Hox genes differentially respond to RA exposure at time points following their initial expression by a local anterior extension of their expression domains (Marshall et al., 1992; Conlon and Rossant, 1992). A gene-specific window of RA sensitivity has been documented for Hoxb1 to start at E7.5 (Marshall et al., 1994), for Hoxb4 at E8.0 (Gould et al., 1998), and for Hoxb6 at E9.5 and Hoxb8 at E10.5 (Oosterveen et al., submitted). All these RA-mediated regulations occur via RAREs present in neural enhancers operative on transgenes isolated from the cluster context (Marshall et al., 1994; Gould et al., 1998; Huang et al., 1998; Oosterveen et al., submitted). We now show that Hoxb1 expression can be induced by RA at developmental stages much earlier than after endogenous activation of the gene, and that this extreme early RA-sensitivity is not mimicked by a Hoxb1lacZ transgene able to recapitulate all aspects of Hoxb1 expression including RA-inducibility after expression has started. We consider it unlikely that precocious activation depends on additional regulatory sequences not present on the transgene, and favour the view that it relies on a particularly good accessibility, in the endogenous context, of the most 3 0 RAREs flanking Hoxb1, or on a higher sensitivity to RA-induced chromatin opening. This would imply the necessity for Hoxb1 to be in the cluster for its expression to be absolutely correctly regulated, even though randomly integrated Hoxb1lacZ transgenes are expressed extremely
88
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
similarly to their endogenous counterpart at all stages after the earliest initiation time point. The different pre-initiation sensitivity of 3 0 and 5 0 Hoxb genes to RA, and the observation that a 3 0 HoxlacZ transgene does not exhibit the extreme sensitivity of its endogenous counterpart, suggest the existence of a level of Hox regulation governing 3 0 directional gene activation hierarchically higher than that acting via the local RARE-containing neural enhancers. 3.4. Precocious 3 0 Hox gene activation anticipates the normal induction of Hoxb1 at later stages Early RA-induced Hoxb1 transcription starts at the posterior junction between embryonic and extraembryonic tissues and spreads anteriorly, similarly to endogenous gene transcription, at the late streak stage. The spreading of the precocious expression domain of Hoxb1 therefore initially follows spatio-temporal dynamics resembling the expansion of endogenously induced Hoxb1 transcription, but more than 24 h in advance of normal initiation of the gene. RA, its receptors and cofactors do not precociously induce Hoxb1 in all epiblast cells at once. Position-specific factors participate in initiating this induction near the junction between extraembryonic and embryonic ectoderm as early as at the pre-streak stage. The mechanism of action of RA/ RARs at these particular locations may be to stimulate transcription in genomic regions where the chromatin is more open, or to actively open the chromatin structure in particularly sensitive regions. This early RA-induction of 3 0 Hoxb genes may anticipate normal 3 0 Hox gene activation later on, in conjunction with position specific factors. Whichever role endogenous RA plays in normally initiating Hox transcription, it certainly is not the exclusive Hox initiator molecule, since Raldh2 null mutant embryos (Niederreither et al., 1999) did not manifest any change in the spatio-temporal features of Hoxb1 initial expression (B.R., Karen Niederreither, Pascal Dolle´ , Pierre Chambon and J. Deschamps, unpublished). Disruption of the RAREs in early neural enhancers of Hoxb1 and Hoxa1 has a more dramatic effect on early Hoxb1 activation (Studer et al., 1998) than the absence of RA anabolism in Raldh2 mutants observed here. This indicates that other aspects of the regulation by RA/RARs are involved in early gene activation, supporting the second hypothesis mentioned above, of an active role of RA/RARs in opening chromatin structure. The early predisposition of 3 0 Hox genes for chromatin opening and transcription may underlie the polarity of first expression of clustered Hox genes. After the expression of the first Hox gene has been turned on, the 3 0 opening of the cluster would progress in the 5 0 direction, making the genes sequentially available for transcription, concomitantly with the release of the cluster-encompassing repression (Van der Hoeven et al., 1996; Kondo and Duboule, 1999; Kmita et al., 2000). Facilitated transcriptional activation would account for the first expression of the 3 0 most Hox gene, and would subsequently
counteract the 5 0 -nested repression progressively withdrawing from 3 0 to 5 0 positions in the cluster. 4. Experimental procedures 4.1. Generation of Hoxb1lacZ transgenic lines The Hoxb1lacZ transgene used was a 21.6 kb Not1–Sal1 genomic fragment containing lacZ in frame with the first exon of Hoxb1 (Marshall et al., 1992, 1994). Transgenic mouse lines were generated by pronuclear injection of C57Bl6 X CBA F2 zygotes, as previously described (Charite´ et al., 1995, 1998). Five founder mice were generated, all expressing the transgene. Two were selected for further breeding, established as lines and crossed to homozygocity. Both express the Hoxb1lacZ transgene identically. Hoxb1/lacZ embryos were collected from crosses between homozygote transgenic males and C57Bl6 X CBA F1 females. 4.2. Embryo dissection and culture Embryos were collected from crosses between C57Bl6 X CBA F1 mice. Hoxb1lacZ embryos were collected from crosses between homozygote transgenic males and F1 females. Staging of embryos was according to Downs and Davies (1993), and K. Lawson (personal communication, see also Davidson et al., website). Cdx1 mutant embryos were derived from intercrossing Cdx1 homozygotes (Subramanian et al., 1995). Cdx1 homozygous mutant mice were crossed with Cdx2 heterozygous mutant mice (Chawengsaksophak et al., 1997) to obtain transheterozygous offspring. These were crossed with Cdx1 null animals to obtain Cdx1 2/2/Cdx2 1/2 embryos. Raldh2 homozygous mutant embryos were obtained from Raldh2 heterozygous intercrosses (Niederreither et al., 1999). The time of fertilization was taken as the midpoint of the dark cycle before the copulation plug was found. For embryo culture, embryos were dissected from the uterus with intact ectoplacental cones, and the embryo was put in HEPES-buffered Dulbecco’s modified Eagle’s medium (DMEM) containing 7.5% foetal calf serum (FCS), after which Reichert’s membrane was removed. Embryonic stages were determined by morphological inspection (size, streak extension, amount of mesoderm), and by measuring the proximal to distal embryonic part. In the first series of experiments, embryos were cultured for up to 4 h in DMEM containing 50% heat-inactivated rat serum that was prepared by centrifugation immediately after withdrawal from the dorsal aorta. In other experiments, 15% FCS was used instead of rat serum. The results were identical in both conditions. Culture took place in 1 ml in a slowly rotating polystyrene tube or in 650 ml in a 4-well plate (Nunclon), with a maximum of two embryos per tube or well, in a 6.3% CO2 (rat serum), or 7.5% CO2 (FCS) humidified atmosphere at 378C.
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
4.3. Whole mount in situ hybridisation Whole mount in situ hybridisation was essentially as described (Wilkinson, 1992) with minor modifications. Digoxigenin-labelled antisense probes were as follows: for Hoxb1, the probe was generated from an 800 bp EcoRI fragment (Wilkinson et al., 1989), the Hoxb2 probe was a 3 kb long EcoR1-Not1 fragment (Rubock et al., 1990). The Hoxb3 probe was a 900 bp Sac1- BamH1 fragment (Sham et al., 1992). Hoxb4 probe was generated from a 1.3 kb EcoRI fragment (Graham et al., 1988), the Hoxb6 probe was derived from an 1.0 kb EcoRI fragment (Schughart et al., 1988). For Hoxb8 a combination of two probes was used: made from a 420 bp SacI fragment from the first exon, and made from a 350 bp SacI-KpnI fragment from the 3 0 untranslated region (Charite´ et al., 1994). Cdx probes were generated from a 849 bp (Meyer and Gruss, 1993), 227 bp (James et al., 1994) and 615 bp (Gamer and Wright, 1993) fragment of Cdx1, Cdx2 and Cdx4, respectively, all corresponding 3 0 untranslated regions. The lacZ probe was a 700 bp Pst1–Rsa1 fragment from the lacZ gene, used in Charite´ et al. (1994). 4.4. X-gal staining of embryos and sectioning X-gal staining was as described (Charite´ et al., 1994). For sectioning of embryos after whole mount in situ hybridization, embryos were first dehydrated through a graded series of ethanol and infiltrated through cold glycol methacrylate (Technovit 8100) overnight at 48C, after which the plastic was polymerised overnight at 48C. Sections were cut at 7 mm and counterstained with 0.1% neutral red for 15 s and mounted in Depex. Acknowledgements We thank F. Meijlink, K. Lawson, R. Zeller and T. Oosterveen for their comments, and D. Duboule for his helpful criticism of the manuscript. We are grateful to R. Krumlauf and H. Marshall for the Hoxb1lacZ construct, to P. Gruss and B. Meyer for the Cdx1 mutants, to F. Beck and K. Chawengsakshophak for the Cdx2 mutant, and to K. Niederreither, P. Dolle´ and P. Chambon for the Raldh2 mutant mice. Probes for Hoxb1, Hoxb2 and Hoxb4 were made available by R. Krumlauf, for Hoxb3 by M. Sham, for Hoxb6 by K. Schughart, for Cdx1 by B. Meyer, for Cdx2 by F. Beck, and for Cdx4 by C. Wright. S. Forlani was supported by a grant from the Dutch NWO (ALW) organisation. References Beck, F., Erler, T., Russel, A., James, R., 1995. Expression of Cdx-2 in the mouse embryo and placenta: possible role in patterning of the extraembryonic membranes. Dev. Dyn. 204, 219–227. Becker, D., Jiang, Z., Knodler, P., Deinard, A.S., Eid, R., Kidd, K.K.,
89
Shashikant, C.S., Ruddle, F.H., Schughart, K., 1996. Conserved regulatory element involved in the early onset of Hoxb6 gene expression. Dev. Dyn. 205, 73–81. Bel-Vialar, S., Core´ , N., Terranova, R., Goudot, V., Boned, A., Djabali, M., 2000. Altered retinoic acid sensitivity and temporal expression of Hox genes in Polycomb-M33-deficient mice. Dev. Biol. 224, 238–249. Bhattacharrya, N., Dey, A., Minucci, S., Zimmer, A., John, S., Hager, G., Ozato, K., 1997. Retinoid-induced chromatin structure alterations in the retinoic acid receptor beta2 promoter. Mol. Cell. Biol. 17, 6481–6490. Brooke, N.M., Garcia-Ferna`ndez, J., Holland, P.W.H., 1998. The ParaHox gene cluster is an evolutionary sister of the Hox gene cluster. Nature 392, 920–922. Charite´ , J., de Graaff, W., Consten, D., Reijnen, M.J., Korving, J., Deschamps, J., 1998. Transducing positional information to the Hox genes: critical interaction of cdx gene products with position-sensitive regulatory elements. Development 125, 4349–4358. Charite´ , J., de Graaff, W., Shen, S.B., Deschamps, J., 1994. Ectopic expression of Hoxb-8 causes duplication of the ZPA in the forelimb and homeotic transformation of axial structures. Cell 78, 589–601. Charite´ , J., de Graaff, W., Vogels, R., Meylink, F., Deschamps, J., 1995. Regulation of the Hoxb-8 gene: synergism between multimerized cisacting elements increases responsiveness to positional information. Dev. Biol. 171, 294–305. Chawengsaksophak, K., James, R., Hammond, V.E., Kontgen, F., Beck, F., 1997. Homeosis and intestinal tumours in Cdx2 mutant mice. Nature 386, 84–87. Conlon, R.A., Rossant, J., 1992. Exogenous retinoic acid rapidly induces anterior ectopic expression of murine Hox-2 genes in vivo. Development 116, 357–368. Deschamps, J., Wijgerde, M., 1993. Two phases in the establishment of HOX expression domains. Dev. Biol. 156, 473–480. Deschamps, J., van den Akker, E., Forlani, S., de Graaff, W., Oosterveen, T., Roelen, B., Roelfsema, J., 1999. Initiation, establishment and maintenance of Hox gene expression patterns in the mouse. Int. J. Dev. Biol. 43, 635–650. Downs, K.M., Davies, T., 1993. Staging of gastrulating mouse embryos by morphological landmarks in the dissecting microscope. Development 118, 1255–1266. 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., 1994. Temporal colinearity and the phylotypic progression: a basis for the stability of a vertebrate Bauplan and the evolution of morphologies through heterochrony. Development (Suppl.), 135–142. Frohman, M.A., Boyle, M., Martin, G.R., 1990. Isolation of the mouse Hox2.9 gene; analysis of embryonic expression suggests that positional information along the anterior–posterior axis is specified by mesoderm. Development 110, 589–607. Gamer, L.W., Wright, C.V.E., 1993. Murine Cdx-4 bears striking similarities to the Drosophila caudal gene in its homeodomain sequence and early expression pattern. Mech. Dev. 43, 71–81. Gould, A., Itasaki, N., Krumlauf, R., 1998. Initiation of rhombomeric Hoxb4 expression requires induction by somitez and a retinoid pathway. Neuron 21, 39–51. Graham, A., Papalopulu, N., Lorimer, J., McVey, J.H., Tuddenham, E.G.D., Krumlauf, R., 1988. Characterization of a murine homeo box gene, Hox-2.6, related to the Drosophila Deformed gene. Genes Dev. 2, 1424–1438. Houle, M., Prinos, P., Iulianella, A., Bouchard, N., Lohnes, D., 2000. Retinoic acid regulation of Cdx1: an indirect mechanism for retinoids and vertebral specification. Mol. Cell. Biol. 20, 6579–6586. Huang, D., Chen, S.W., Langston, A.W., Gudas, L.J., 1998. A conserved retinoic acid responsive element in the murine Hoxb-1 gene is required for expression in the developing gut. Development 125, 3235–3246. James, R., Erler, T., Kazenwadel, J., 1994. Structure of the murine homeobox gene cdx-2. J. Biol. Chem. 269, 15229–15237. Kmita, M., van der Hoeven, F., Za´ ka´ ny, J., Krumlauf, R., Duboule, D.,
90
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
2000. Mechanisms of Hox gene colinearity: transposition of the anterior Hoxb1 gene into the posterior HoxD complex. Genes Dev. 14, 198–211. Kondo, T., Duboule, D., 1999. Breaking colinearity in the mouse HoxD complex. Cell 97, 407–417. Kwan, C.T., Tsang, S.L., Krumlauf, R., Sham, M.H., 2001. Regulatory analysis of the mouse Hoxb3 gene: multiple elements work in concert to direct temporal and spatial patterns of expression. Dev. Biol. 232, 176–190. Maconochie, M.K., Nonchev, S., Studer, M., Chan, S.K., Popperl, H., Sham, M.H., Mann, R.S., Krumlauf, R., 1997. Cross-regulation in the mouse HoxB complex: the expression of Hoxb2 in rhombomere 4 is regulated by Hoxb1. Genes Dev. 11, 1885–1895. Manzanares, M., Bel-Vialar, S., Ariza-McNaughton, L., Ferretti, E., Marshall, H., Maconochie, M.M., Blasi, F., Krumlauf, R., 2001. Independent regulation of initiation and maintenance phases of Hoxa3 expression in the vertebrate hindbrain involve auto- and cross-regulatory mechanisms. Development 128, 3595–3607. Marshall, H., Nonchev, S., Sham, M.H., Muchamore, I., Lumsden, A., Krumlauf, R., 1992. Retinoic acid alters hindbrain Hox code and induces transformation of rhombomeres 2/3 into a 4/5 identity. Nature 360, 737–741. Marshall, H., Studer, M., Po¨ pperl, H., Aparicio, S., Kuroiwa, A., Brenner, S., Krumlauf, R., 1994. A conserved retinoic acid response element required for early expression of the homeobox gene Hoxb-1. Nature 370, 567–571. Meyer, B.I., Gruss, P., 1993. Mouse Cdx-1 expression during gastrulation. Development 117, 191–203. Murphy, P., Davidson, D.R., Hill, R.E., 1989. Segment-specific expression of a homoeobox-containing gene in the mouse hindbrain. Nature 341, 156–159. Niederreither, K., Subbarayan, V., Dolle´ , P., Chambon, P., 1999. Embryonic retinoic acid synthesis is essential for early mouse post-implantation development. Nat. Genet. 21, 444–448. Prinos, P., Joseph, S., Oh, K., Meyer, B.I., Gruss, P., Lohnes, B., 2001. Multiple pathways governing Cdx1 expression during murine development. Dev. Biol. 239, 257–269.
Rubock, M.J., Larin, Z., Cook, M., Papalopulu, N., Krumlauf, R., Lehrach, H., 1990. A yeast artificial chromosome containing the mouse homeobox cluster Hox-2. Proc. Natl. Acad. Sci. USA 87, 4751–4755. Schughart, K., Utset, M.F., Awgulewitsch, A., Ruddle, F.H., 1988. Structure and expression of Hox-2.2, a murine homeobox-containing gene. Proc. Natl. Acad. Sci. USA 85, 5582–5586. Sham, M.H., Hunt, P., Nonchev, S., Papalopulu, N., Graham, A., Boncinelli, E., Krumlauf, R., 1992. Analysis of the murine Hox-2.7 gene: conserved alternative transcripts with differential distribution in the nervous system and the potential for shared regulatory regions. EMBO J. 11, 1825–1836. Studer, M., Gavalas, A., Marshall, H., Ariza-McNaughton, L, Rijli, P.M., Chambon, P., Krumlauf, R., 1998. Genetic interactions between Hoxa1 and Hoxb1 reveal new roles in regulation of early hindbrain patterning. Development 125, 1025–1036. Studer, M., Popperl, H., Marshall, H., Kuroiwa, A., Krumlauf, R., 1994. Role of a conserved retinoic acid response element in rhombomere restriction of Hoxb-1. Science 265, 1728–1732. Subramanian, V., Meyer, B.I., Gruss, P., 1995. Disruption of the murine homeobox gene Cdx1 affects axial skeletal identities by altering the mesodermal expression domains of Hox genes. Cell 83, 641–653. Van den Akker, E., Forlani, S., Chawengsaksophak, K., de Graaff, W., Beck, F., Meyer, B., Deschamps, J., 2002. Cdx1 and Cdx2 have overlapping functions in anteroposterior patterning and posterior axis elongation. Development 129, 2181–2193. Van der Hoeven, F., Za´ ka´ ny, J., Duboule, D., 1996. Gene transpositions in the HoxD complex reveal a hierarchy of regulatory controls. Cell 85, 1025–1035. Wilkinson, D.G. (Ed.), 1992. In Situ Hybridization, a Practical Approach IRL Press, New York, NY. Wilkinson, D.G., Bhatt, S., Cook, M., Boncinelli, E., Krumlauf, R., 1989. Segmental expression of Hox-2 homoeobox-containing genes in the developing mouse hindbrain. Nature 341, 405–409. Zhang, F., Nagy Kovacs, E., Featherstone, M.S., 2000. Murine Hoxd4 expression in the CNS requires multiple elements including a retinoic acid response element. Mech. Dev. 96, 79–89.
Hox cluster polarity in early transcriptional availability: a high order regulatory level of clustered Hox genes in the mouse Bernard A.J. Roelen 1, Wim de Graaff, Sylvie Forlani 2, Jacqueline Deschamps* Hubrecht Laboratory, Netherlands Institute for Developmental Biology, Uppsalalaan 8, 3584 CT, Utrecht, The Netherlands Received 27 June 2002; received in revised form 23 August 2002; accepted 28 August 2002
Abstract The molecular mechanism underlying the 3 0 to 5 0 polarity of induction of mouse Hox genes is still elusive. While relief from a clusterencompassing repression was shown to lead to all Hoxd genes being expressed like the 3 0 most of them, Hoxd1 (Kondo and Duboule, 1999), the molecular basis of initial activation of this 3 0 most gene, is not understood yet. We show that, already before primitive streak formation, prior to initial expression of the first Hox gene, a dramatic transcriptional stimulation of the 3 0 most genes, Hoxb1 and Hoxb2, is observed upon a short pulse of exogenous retinoic acid (RA), whereas it is not in the case for more 5 0 , cluster-internal, RA-responsive Hoxb genes. In contrast, the RA-responding Hoxb1lacZ transgene that faithfully mimics the endogenous gene (Marshall et al., 1994) did not exhibit the sensitivity of Hoxb1 to precocious activation. We conclude that polarity in initial activation of Hoxb genes reflects a greater availability of 3 0 Hox genes for transcription, suggesting a pre-existing (susceptibility to) opening of the chromatin structure at the 3 0 extremity of the cluster. We discuss the data in the context of prevailing models involving differential chromatin opening in the directionality of clustered Hox gene transcription, and regarding the importance of the cluster context for correct timing of initial Hox gene expression.Interestingly, Cdx1 manifested the same early transcriptional availability as Hoxb1. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Mouse Hox genes during gastrulation; Transcriptional regulation of the Hox genes; Retinoid signalling and transcriptional initiation of the Hox genes
1. Introduction Sequential activation of mouse Hox genes begins at the late primitive streak stage and is tightly coupled with the progress of gastrulation and laying down of the future definitive Hox expressing tissues (Duboule, 1992, 1994; Deschamps et al., 1999). A hierarchy of genetic control mechanisms is thought to orchestrate the subsequent establishment of the definitive Hox expression domains that are crucial for correct embryonic patterning. After the first Hox gene is initially expressed, the release of a cluster-encompassing repression has been shown to successively allow 3 0 to 5 0 Hoxd genes to be transcribed (Kondo and Duboule, 1999). This gradual derepression is thought to precede the numerous regulatory steps directing spatially specific subsets of the complex expression patterns, and involving enhancers * Corresponding author. Tel.: 131-30-2121935; fax: 131-30-2516464. E-mail address: [email protected] (J. Deschamps). 1 Present address. Division of Cellular Biochemistry, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. 2 Present address. De´partement de Biologie Mole´culaire du De´veloppement, Institut Pasteur, 25 rue du Dr. Roux, 75124 Paris, France. Abbreviations: E6.2, embryonic day 6.2; RA, retinoic acid; RAR, retinoic acid receptor; RARE, retinoic acid response element; A-P, antero-posterior
often able to function independently of the cluster context. Progressive escape of Hoxd genes from the global repression, exerted from a remote 5 0 sequence, has been proposed to be facilitated by a gradual opening of the chromatin structure from 3 0 to 5 0 (Van der Hoeven et al., 1996; Kondo and Duboule, 1999). Among the cis-acting, gene-specific regulatory sequences that subsequently would become sequentially accessible and operative are Retinoic Acid Response Elements (RAREs). These RAREs, present in 3 0 and more 5 0 neural enhancers, mediate gene stimulation by endogenous retinoic acid (RA) signals, in a progressively later time window (Gould et al., 1998; Zhang et al., 2000). Transcription of the earliest Hoxb gene, Hoxb1, is activated in the posterior part of the primitive streak at the late streak stage (embryonic day 7.2, E7.2) (Murphy et al., 1989; Frohman et al., 1990). Hoxb1 has been shown to respond to RA subsequently to initial transcription (E7.75, late head fold stage, and later) by rostrally extending its anterior expression boundary in the presumptive hindbrain (Marshall et al., 1992; Conlon and Rossant, 1992). Hoxb2 is regulated very similarly to Hoxb1 at early stages. Both genes share some regulatory elements, a Hoxb2 RAresponse being mediated by Hoxb1 (Maconochie et al., 1997). In addition, a Hoxb2 RARE has recently been iden-
0925-4773/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(02)00329-5
82
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
Fig. 1. Comparison between the expression patterns of Hoxb1 and Hoxb1lacZ in early mouse embryos. (A–C) Whole mount in situ hybridisation with a Hoxb1 probe, and (D–F), Xgal staining of Hoxb1/lacZ activity. Developmental stages were early late streak stage (E7.2, A,D), slightly more advanced early allantoic bud stage (E7.5, B,E), and neural plate stage (E7.75, C,F). Hoxb1/lacZ faithfully recapitulates the endogenous Hoxb1 expression pattern. Posterior is to the right in all panels.
tified, which initiates Hoxb2 expression in rhombomere 4 (cited in Manzanares et al., 2001). Hoxb3 shares regulatory regions with both Hoxb2 and Hoxb4 (Kwan et al., 2001). Expression of more 5 0 Hox genes starts in posterior embryonic tissues at progressively later stages. Hoxb4 is activated at the neural plate stage (E7.75) (Wilkinson et al., 1989; Gould et al., 1998), Hoxb6 at the early head fold (Becker et al., 1996) and Hoxb8 at a more advanced head fold stage (E7.75) (Deschamps and Wijgerde, 1993). Hoxb4 was shown to be RA-sensitive from E8 (Gould et al., 1998), and Hoxb6 and Hoxb8 from E9.5 and E10.5, respectively (Oosterveen et al., submitted).
In our experiments we set out to probe the transcriptional availability of 3 0 versus more 5 0 Hoxb genes very early in development, before any Hox gene starts to be transcribed. For this purpose we cultured carefully staged postimplantation embryos for a short time in the presence of RA as an external stimulus. We show that precocious induction of Hoxb1 and Hoxb2 by RA is operational at E6.0 already, before primitive streak formation and long before the gene is transcriptionally initiated. More 5 0 Hoxb genes, among which the later RA-responsive Hoxb4, do not react to this early stimulation. Furthermore, we show that a Hoxb1 transgene perfectly mimicking the endogenous gene in its expres-
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
83
Fig. 2. Hoxb gene expression in cultured pre- and early-streak stage embryos in control medium (A,C,E,G,I,K) and after a 4 h exposure to 10 26 M RA (B,D,F,H,J,L), detected by whole mount in situ hybridisation. A,B, Hoxb1; C,D, Hoxb2; E,F, Hoxb3; G,H, Hoxb4; I,J, Hoxb6 and K,L, Hoxb8. Posterior is to the right in all panels. While Hoxb1 and Hoxb2 are strongly induced by early RA-exposure, no induction of Hoxb3, Hoxb4, Hoxb6 and Hoxb8 is visible after a 4-h RA treatment at the midstreak stage (E 6.5). Internal positive controls of the probes in these hybridisations were obtained by hybridising E8.5 embryos (not shown) and allowing the staining reaction until a good signal was visible. Bar is 100 mm.
sion features and post-initiation RA sensitivity, did not react to the precocious transcriptional stimulation, suggesting that the early transcriptional availability is a cluster property. Finally we demonstrate that one of the Cdx gene family members, a relative of the Hox genes, also exhibits the extreme early transcriptional availability of Hoxb1.
2. Results 2.1. Exogenous RA rapidly and precociously activates Hoxb1 and Hoxb2 prior to primitive streak formation Hoxb1 is normally transcriptionally activated in the posterior part of the primitive streak, at the junction between embryonic and extraembryonic tissues, at the late streak stage (E7.2). Expression is first confined to posteriormost mesoderm and streak ectoderm, and later spreads rostrally to more anterior regions and laterally in the mesodermal wings (Murphy et al., 1989; Frohman et al., 1990) (Fig. 1A–C) We exposed pre-streak (E6.0) and early-streak (E6.2) embryos in culture, thus before any Hox gene was expressed, to 10 26 M all-trans-retinoic acid (RA) for 4 h, and analysed Hox gene expression by whole mount in situ hybridisation. Hoxb1 expression was strongly activated in these conditions (Fig. 2A, B). Hoxb2 expression was induced as well (Fig. 2C, D), whereas the expression of Hoxb3, Hoxb4, Hoxb6 and Hoxb8 was not induced (Fig. 2E–L). For Hoxb4, Hoxb6
and Hoxb8, which start to be expressed later than Hoxb1 we examined gene expression at progressively later time points, until stages closely preceding normal transcriptional initiation, and did not observe precocious induction upon either a 4 h, or a 7 h-RA treatment (not shown). RA used as an exogenous trigger to sense whether clustered Hox genes are available for transcription, thus rapidly and precociously turns on Hoxb1, and the partially Hoxb1-dependent Hoxb2. In contrast, this trigger does not affect Hoxb3, Hoxb4, Hoxb6 and Hoxb8, in spite of the fact the latter have been shown to be RA-sensitive at later stages. 2.2. Precocious expression of Hoxb1 initially follows spatiotemporal dynamics similar to that of the endogenous gene We examined the spatio-temporal dynamics of the precociously induced Hoxb1 expression, to compare it with the normal evolution of the transcription domain of endogenous Hoxb1. Pre-streak (E6.0) embryos were exposed to RA in culture for periods between 30 min and 4 h. Already after a 30-min exposure, Hoxb1 was precociously induced in the presumptive posterior epiblast at the junction between embryonic and extraembryonic ectoderm (Fig. 3A, B). After culture in these conditions for 1 h, Hoxb1 transcripts had expanded towards antero-distal and lateral positions (Fig. 3C). From the localisation, in whole mount and sectioned embryos, of the expression signal in embryos that had formed an early streak, we conclude that the RAinduced transcription of Hoxb1 in the pre-streak embryos
84
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
begins and spreads from posterior to distal and lateral positions. This spreading of the precocious expression domain of Hoxb1 therefore initially follows spatio-temporal dynamics resembling the expansion of endogenously induced Hoxb1 transcription, but more than 24 h in advance of normal expression. From a 2-h RA exposure onwards, the entire epiblast expressed Hoxb1 at high levels (Fig. 3D) in embryos treated at the prestreak stage, while the most antero-distal region remained non-induced in embryos treated at the early streak stage. Precocious expression of Hoxb1 mainly occurred in the epiblast (Fig. 3F, H) whereas
normally induced transcription of the gene takes place at a later stage in the mesoderm and streak epiblast (Fig. 3G). 2.3. A Hoxb1lacZ transgene, which is RA responsive at later stages, does not react to a pre-initiation RA stimulation We investigated whether a Hoxb1lacZ transgene, able to entirely recapitulate endogenous Hoxb1 expression (Fig. 1D–F), and to respond to RA at later stages (Marshall et al., 1994) exhibited the same extreme sensitivity to early RA activation as endogenous Hoxb1. The transgene used
Fig. 3. Timing of precocious RA-induced Hoxb1 expression (A–D). Hoxb1 expression in a pre-streak control embryo (A) and in embryos cultured for 30 min (B), 1 h (C) and 2 h (D) in the presence of RA. Transversal sections from control and RA-treated embryos (E–H) hybridised with a Hoxb1 probe. Control E6.5 (early streak) embryo (E). E6.5 embryo treated with RA for 4 h (F). Control E7.5 (late streak) embryo (G). E7.5 embryo treated with RA for 4 h (H). Posterior is to the right. Bar is 100 mm in A–F, and 200 mm in G,H. m, mesoderm. e, epiblast.
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
85
Fig. 4. A Hoxb1lacZ transgene is not induced by short RA exposure in early streak stage embryos. A,C,E, control embryos, and B,D,F, embryos exposed to RA for 4 h (B,D) or 6 h (F), assayed for b-galactosidase activity (E,F) or lacZ mRNA expression (A–D). Expression of Hoxb1/lacZ is not induced after RAexposure, in contrast to that of endogenous Hoxb1 (A,B). This Hoxb1/lacZ transgene is induced (anteriorly and laterally extended expression domain) upon a later, post-initiation (E7.5, head fold stage) exposure to RA (H compared to G). Bar is 100 mm in A–F, and 200 mm in G,H.
was a 21.6 kb genomic fragment where lacZ is inserted in frame in the first exon of Hoxb1. It contains, besides the proximal 3 0 and 5 0 RARE elements (Marshall et al., 1994; Studer et al., 1994) the more distal 3 0 RARE that mediates RA induction of Hoxb1 in gut endoderm later in development (Huang et al., 1998). We generated transgenic lines stably expressing this Hoxb1lacZ transgene. As previously described (Marshall et al., 1992, 1994) the Hoxb1lacZ transgene was expressed indistinguishably from the endogenous gene (Fig. 1), and responded to post-initiation RA exposure by rostrally and laterally expanding its expression domain (Fig. 4G, H). We compared endogenous Hoxb1 and Hoxb1lacZ expression after exposure of pre- and early streak stage embryos to RA in culture. No induction of Hoxb1lacZ mRNA and protein accumulation was observed, in conditions where endogenous Hoxb1 was strongly induced (Fig. 4C–F compared to Fig. 4A, B). The Hoxb1lacZ transgene responded to post-initiation RA exposure (Fig. 4G, H), demonstrating its sensitivity to RA at later stages. There
was therefore a dramatic early difference (at E6.2) between transgene and endogenous gene in their ability to be transcriptionally activated, although both responded to RA after Hoxb1 had been normally initiated (E7.5). 2.4. Cdx1 also quickly responds to a short RA pulse by a strong and precocious induction prior to primitive streak formation The mouse homologues of Drosophila caudal, the Cdx genes, are involved in axial patterning and are known to regulate Hox gene expression (Subramanian et al., 1995; Charite´ et al., 1998; van den Akker et al., 2002). Cdx genes have been suggested to be ancient relatives of the Hox genes (Brooke et al., 1998). Cdx1 is initially expressed very similarly to Hoxb1 in time and place. Transcription of Cdx1 is normally initiated in the posterior primitive streak epiblast and mesoderm at the late streak stage. Expression of Cdx1 spreads in a Hox-like manner to eventually reach its
86
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
rostral expression at the level of the preotic sulcus at early somite stages, before becoming down-regulated in the rostral spinal cord and sclerotomes (Meyer and Gruss, 1993). Cdx1 has also been shown to be a direct RA target in vivo at time points following transcriptional initiation (Houle et al., 2000; Prinos et al., 2001). We examined whether Cdx1, like Hoxb1, quickly responded to RA exposure by a strong and precocious induction. Treatment of pre(E6.0) and early streak (E6.2) embryos with exogenous RA for 4 h led to strong and precocious Cdx1 transcription (Fig. 5A, B). This extreme early sensitivity of Cdx1 to exogenous RA is not shared by the other two Cdx genes, Cdx2 (James et al., 1994) and Cdx4 (Gamer and Wright, 1993). No difference in the time, location and level of transcription of Cdx2 (Fig. 5C, D) and Cdx4 (Fig. 5E, F) was noticed between RAtreated and untreated embryos. Therefore, Cdx1 appears to be the only Cdx gene to be precociously transcriptionally activated by RA like Hoxb1. 2.5. Cdx1 does not mediate the early inducibility of Hoxb1 to RA Since both Hoxb1 and Cdx1 are strongly RA-inducible at
pre- and early streak stages, and because Cdx1 is a Hox regulator, we tested whether precocious induction of Hoxb1 by exogenous RA was dependent on Cdx1. We assayed Hoxb1 expression upon RA exposure in Cdx1 null mutant embryos (Subramanian et al., 1995). The same strong and precocious induction of Hoxb1 was measured after early exposure of wild type and Cdx1 null embryos to RA in culture (not shown). This indicates that Cdx1 does not mediate the response of Hoxb1 to RA. We also found that initial Hoxb1 and Hoxb8 expression was not altered in Cdx mutants lacking both Cdx1 alleles and one Cdx2 allele (data not shown), suggesting that Cdx gene products are not mediating endogenous Hox gene initiation during gastrulation.
3. Discussion 3.1. Pre-initiation inducibility reveals a difference in permissivity to transcription before Hox expression begins The observation that 3 0 -most Hox genes rapidly responded to a short pulse of RA before primitive streak formation whereas cluster internal- and 5 0 Hox genes only
Fig. 5. The expression of Cdx1, but not that of Cdx2 and Cdx4, is induced precociously by RA at pre/early streak stages. Cdx1 (A,B), Cdx2 (C,D) and Cdx4 (E,F) expression in cultured control embryos (A,C,E) and embryos cultured in the presence of RA (B,D,F). Posterior is on the right. Bar is 100 m m.
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
did so after endogenous expression was initiated, later during gastrulation, suggests that endogenous 3 0 and 5 0 Hoxb genes differ much more in their response to RA exposure at pre-initiation time points than they do after Hoxb gene expression has started. One possible molecular mechanism for this difference in availability for transcription at early stages might involve a pre-existing difference in the degree of chromatin opening at 3 0 versus internal and 5 0 positions in the Hoxb cluster, before the first Hox gene is activated. Exogenous RA, RA receptors and their co-factors might have access to the 3 0 RARE sequences at the early stages, but be denied access to sequences further 5 0 in the cluster. From the moment 3 0 Hox genes are initially transcribed, a further chromatin opening from 3 0 to 5 0 in the cluster would facilitate progressive release from global silencing, and account for proper timing of activation of more 5 0 Hox genes (Van der Hoeven et al., 1996; Kondo and Duboule, 1999, Kmita et al., 2000). Our observations suggest that the polarity of initial expression may be accounted for by a greater availability of 3 0 genes to the transcriptional machinery, long before Hox genes are expressed for the first time in the embryo. Experiments by Bel-Vialar et al. (2000) suggested that Polycomb-group Gene (Pc-G) products, responsible for a gradually more packed chromatin structure from 3 0 to 5 0 in the clusters, could be the system that controls the accessibility to the Hox RAREs, and differentially modulates temporal activation of the 3 0 to 5 0 Hox genes, both in the absence and in the presence of exogenous RA. Our data support the model of Bel-Vialar et al. (2000), and suggest that in the wild type PcG context, a functional difference exists between the degree of accessibility of the most 3 0 compared with more 5 0 RAsensitive Hox transcription start sites, at stages prior to initial Hox transcription. Alternatively, rather than passively getting access to the RAREs, exogenous RA might actively open the chromatin structure on the 3 0 side of the cluster, and anticipate an opening role of endogenous RA in the normal situation. RA/ RARs, with co-factors and recruited histone acetyl transferases are indeed capable of actively opening the chromatin structure (Bhattacharrya et al., 1997). In such a case, hypersensitivity of 3 0 Hoxb genes to RA-induced transcription would imply that the 3 0 side of the Hox cluster is more sensitive to the ‘opening’ activity of RA. Differential sensitivity to chromatin opening could also depend on the Pc-G epigenetic system. 3.2. Cdx1 shares early susceptibility to activation with Hoxb1 Evidence has accumulated to propose that Cdx genes and Hox genes are evolutionary closely related. Duplication of an ancestral gene cluster would have given rise to the Hox and ParaHox clusters, first identified in Amphioxus (Brooke et al., 1998). Cdx1 is expressed initially as early and as anteriorly as Hoxb1, and we proposed (van den Akker et
87
al., 2002) that Cdx genes possess 3 0 Hox structural and regulatory features, and would be paralogous to relatively anterior Hox genes. The present data further strengthen this view since Cdx1, like Hoxb1 seems to be immediately available for transcription before gene expression is normally initiated, as early as in pre- and early-streak stage embryos. The regulatory sequences of Cdx2 and Cdx4 apparently have diverged away from the initial Hox-like features. Cdx2 has acquired an early expression domain in extraembryonic ectoderm (Beck et al., 1995) before the gene is first expressed in a Hox-like manner in the embryo proper. In spite of the amazing similarity in the way all three Cdx genes and Hox genes establish their expression domains (Deschamps et al., 1999), considerable differences have arisen between Cdx 1 and its paralogues, such as the difference in sensitivity to early RA exposure. In spite of its extreme RA-sensitivity, Cdx1 was shown not to mediate the early response of Hoxb1 to RA. In fact, our experiments strongly suggest that Cdx gene products, which positively regulate the Hox genes during the establishment of their expression domains (Subramanian et al., 1995, Charite´ et al., 1998, Van den Akker et al., 2002), are not involved in the earliest initiation of Hox gene transcription.
3.3. Sensitivity of Hoxb1 to early RA stimulation occurs at a level hierarchically higher than the post-initiation RAresponse of individual Hoxb genes Previous work had made it clear that 3 0 to 5 0 Hox genes differentially respond to RA exposure at time points following their initial expression by a local anterior extension of their expression domains (Marshall et al., 1992; Conlon and Rossant, 1992). A gene-specific window of RA sensitivity has been documented for Hoxb1 to start at E7.5 (Marshall et al., 1994), for Hoxb4 at E8.0 (Gould et al., 1998), and for Hoxb6 at E9.5 and Hoxb8 at E10.5 (Oosterveen et al., submitted). All these RA-mediated regulations occur via RAREs present in neural enhancers operative on transgenes isolated from the cluster context (Marshall et al., 1994; Gould et al., 1998; Huang et al., 1998; Oosterveen et al., submitted). We now show that Hoxb1 expression can be induced by RA at developmental stages much earlier than after endogenous activation of the gene, and that this extreme early RA-sensitivity is not mimicked by a Hoxb1lacZ transgene able to recapitulate all aspects of Hoxb1 expression including RA-inducibility after expression has started. We consider it unlikely that precocious activation depends on additional regulatory sequences not present on the transgene, and favour the view that it relies on a particularly good accessibility, in the endogenous context, of the most 3 0 RAREs flanking Hoxb1, or on a higher sensitivity to RA-induced chromatin opening. This would imply the necessity for Hoxb1 to be in the cluster for its expression to be absolutely correctly regulated, even though randomly integrated Hoxb1lacZ transgenes are expressed extremely
88
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
similarly to their endogenous counterpart at all stages after the earliest initiation time point. The different pre-initiation sensitivity of 3 0 and 5 0 Hoxb genes to RA, and the observation that a 3 0 HoxlacZ transgene does not exhibit the extreme sensitivity of its endogenous counterpart, suggest the existence of a level of Hox regulation governing 3 0 directional gene activation hierarchically higher than that acting via the local RARE-containing neural enhancers. 3.4. Precocious 3 0 Hox gene activation anticipates the normal induction of Hoxb1 at later stages Early RA-induced Hoxb1 transcription starts at the posterior junction between embryonic and extraembryonic tissues and spreads anteriorly, similarly to endogenous gene transcription, at the late streak stage. The spreading of the precocious expression domain of Hoxb1 therefore initially follows spatio-temporal dynamics resembling the expansion of endogenously induced Hoxb1 transcription, but more than 24 h in advance of normal initiation of the gene. RA, its receptors and cofactors do not precociously induce Hoxb1 in all epiblast cells at once. Position-specific factors participate in initiating this induction near the junction between extraembryonic and embryonic ectoderm as early as at the pre-streak stage. The mechanism of action of RA/ RARs at these particular locations may be to stimulate transcription in genomic regions where the chromatin is more open, or to actively open the chromatin structure in particularly sensitive regions. This early RA-induction of 3 0 Hoxb genes may anticipate normal 3 0 Hox gene activation later on, in conjunction with position specific factors. Whichever role endogenous RA plays in normally initiating Hox transcription, it certainly is not the exclusive Hox initiator molecule, since Raldh2 null mutant embryos (Niederreither et al., 1999) did not manifest any change in the spatio-temporal features of Hoxb1 initial expression (B.R., Karen Niederreither, Pascal Dolle´ , Pierre Chambon and J. Deschamps, unpublished). Disruption of the RAREs in early neural enhancers of Hoxb1 and Hoxa1 has a more dramatic effect on early Hoxb1 activation (Studer et al., 1998) than the absence of RA anabolism in Raldh2 mutants observed here. This indicates that other aspects of the regulation by RA/RARs are involved in early gene activation, supporting the second hypothesis mentioned above, of an active role of RA/RARs in opening chromatin structure. The early predisposition of 3 0 Hox genes for chromatin opening and transcription may underlie the polarity of first expression of clustered Hox genes. After the expression of the first Hox gene has been turned on, the 3 0 opening of the cluster would progress in the 5 0 direction, making the genes sequentially available for transcription, concomitantly with the release of the cluster-encompassing repression (Van der Hoeven et al., 1996; Kondo and Duboule, 1999; Kmita et al., 2000). Facilitated transcriptional activation would account for the first expression of the 3 0 most Hox gene, and would subsequently
counteract the 5 0 -nested repression progressively withdrawing from 3 0 to 5 0 positions in the cluster. 4. Experimental procedures 4.1. Generation of Hoxb1lacZ transgenic lines The Hoxb1lacZ transgene used was a 21.6 kb Not1–Sal1 genomic fragment containing lacZ in frame with the first exon of Hoxb1 (Marshall et al., 1992, 1994). Transgenic mouse lines were generated by pronuclear injection of C57Bl6 X CBA F2 zygotes, as previously described (Charite´ et al., 1995, 1998). Five founder mice were generated, all expressing the transgene. Two were selected for further breeding, established as lines and crossed to homozygocity. Both express the Hoxb1lacZ transgene identically. Hoxb1/lacZ embryos were collected from crosses between homozygote transgenic males and C57Bl6 X CBA F1 females. 4.2. Embryo dissection and culture Embryos were collected from crosses between C57Bl6 X CBA F1 mice. Hoxb1lacZ embryos were collected from crosses between homozygote transgenic males and F1 females. Staging of embryos was according to Downs and Davies (1993), and K. Lawson (personal communication, see also Davidson et al., website). Cdx1 mutant embryos were derived from intercrossing Cdx1 homozygotes (Subramanian et al., 1995). Cdx1 homozygous mutant mice were crossed with Cdx2 heterozygous mutant mice (Chawengsaksophak et al., 1997) to obtain transheterozygous offspring. These were crossed with Cdx1 null animals to obtain Cdx1 2/2/Cdx2 1/2 embryos. Raldh2 homozygous mutant embryos were obtained from Raldh2 heterozygous intercrosses (Niederreither et al., 1999). The time of fertilization was taken as the midpoint of the dark cycle before the copulation plug was found. For embryo culture, embryos were dissected from the uterus with intact ectoplacental cones, and the embryo was put in HEPES-buffered Dulbecco’s modified Eagle’s medium (DMEM) containing 7.5% foetal calf serum (FCS), after which Reichert’s membrane was removed. Embryonic stages were determined by morphological inspection (size, streak extension, amount of mesoderm), and by measuring the proximal to distal embryonic part. In the first series of experiments, embryos were cultured for up to 4 h in DMEM containing 50% heat-inactivated rat serum that was prepared by centrifugation immediately after withdrawal from the dorsal aorta. In other experiments, 15% FCS was used instead of rat serum. The results were identical in both conditions. Culture took place in 1 ml in a slowly rotating polystyrene tube or in 650 ml in a 4-well plate (Nunclon), with a maximum of two embryos per tube or well, in a 6.3% CO2 (rat serum), or 7.5% CO2 (FCS) humidified atmosphere at 378C.
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
4.3. Whole mount in situ hybridisation Whole mount in situ hybridisation was essentially as described (Wilkinson, 1992) with minor modifications. Digoxigenin-labelled antisense probes were as follows: for Hoxb1, the probe was generated from an 800 bp EcoRI fragment (Wilkinson et al., 1989), the Hoxb2 probe was a 3 kb long EcoR1-Not1 fragment (Rubock et al., 1990). The Hoxb3 probe was a 900 bp Sac1- BamH1 fragment (Sham et al., 1992). Hoxb4 probe was generated from a 1.3 kb EcoRI fragment (Graham et al., 1988), the Hoxb6 probe was derived from an 1.0 kb EcoRI fragment (Schughart et al., 1988). For Hoxb8 a combination of two probes was used: made from a 420 bp SacI fragment from the first exon, and made from a 350 bp SacI-KpnI fragment from the 3 0 untranslated region (Charite´ et al., 1994). Cdx probes were generated from a 849 bp (Meyer and Gruss, 1993), 227 bp (James et al., 1994) and 615 bp (Gamer and Wright, 1993) fragment of Cdx1, Cdx2 and Cdx4, respectively, all corresponding 3 0 untranslated regions. The lacZ probe was a 700 bp Pst1–Rsa1 fragment from the lacZ gene, used in Charite´ et al. (1994). 4.4. X-gal staining of embryos and sectioning X-gal staining was as described (Charite´ et al., 1994). For sectioning of embryos after whole mount in situ hybridization, embryos were first dehydrated through a graded series of ethanol and infiltrated through cold glycol methacrylate (Technovit 8100) overnight at 48C, after which the plastic was polymerised overnight at 48C. Sections were cut at 7 mm and counterstained with 0.1% neutral red for 15 s and mounted in Depex. Acknowledgements We thank F. Meijlink, K. Lawson, R. Zeller and T. Oosterveen for their comments, and D. Duboule for his helpful criticism of the manuscript. We are grateful to R. Krumlauf and H. Marshall for the Hoxb1lacZ construct, to P. Gruss and B. Meyer for the Cdx1 mutants, to F. Beck and K. Chawengsakshophak for the Cdx2 mutant, and to K. Niederreither, P. Dolle´ and P. Chambon for the Raldh2 mutant mice. Probes for Hoxb1, Hoxb2 and Hoxb4 were made available by R. Krumlauf, for Hoxb3 by M. Sham, for Hoxb6 by K. Schughart, for Cdx1 by B. Meyer, for Cdx2 by F. Beck, and for Cdx4 by C. Wright. S. Forlani was supported by a grant from the Dutch NWO (ALW) organisation. References Beck, F., Erler, T., Russel, A., James, R., 1995. Expression of Cdx-2 in the mouse embryo and placenta: possible role in patterning of the extraembryonic membranes. Dev. Dyn. 204, 219–227. Becker, D., Jiang, Z., Knodler, P., Deinard, A.S., Eid, R., Kidd, K.K.,
89
Shashikant, C.S., Ruddle, F.H., Schughart, K., 1996. Conserved regulatory element involved in the early onset of Hoxb6 gene expression. Dev. Dyn. 205, 73–81. Bel-Vialar, S., Core´ , N., Terranova, R., Goudot, V., Boned, A., Djabali, M., 2000. Altered retinoic acid sensitivity and temporal expression of Hox genes in Polycomb-M33-deficient mice. Dev. Biol. 224, 238–249. Bhattacharrya, N., Dey, A., Minucci, S., Zimmer, A., John, S., Hager, G., Ozato, K., 1997. Retinoid-induced chromatin structure alterations in the retinoic acid receptor beta2 promoter. Mol. Cell. Biol. 17, 6481–6490. Brooke, N.M., Garcia-Ferna`ndez, J., Holland, P.W.H., 1998. The ParaHox gene cluster is an evolutionary sister of the Hox gene cluster. Nature 392, 920–922. Charite´ , J., de Graaff, W., Consten, D., Reijnen, M.J., Korving, J., Deschamps, J., 1998. Transducing positional information to the Hox genes: critical interaction of cdx gene products with position-sensitive regulatory elements. Development 125, 4349–4358. Charite´ , J., de Graaff, W., Shen, S.B., Deschamps, J., 1994. Ectopic expression of Hoxb-8 causes duplication of the ZPA in the forelimb and homeotic transformation of axial structures. Cell 78, 589–601. Charite´ , J., de Graaff, W., Vogels, R., Meylink, F., Deschamps, J., 1995. Regulation of the Hoxb-8 gene: synergism between multimerized cisacting elements increases responsiveness to positional information. Dev. Biol. 171, 294–305. Chawengsaksophak, K., James, R., Hammond, V.E., Kontgen, F., Beck, F., 1997. Homeosis and intestinal tumours in Cdx2 mutant mice. Nature 386, 84–87. Conlon, R.A., Rossant, J., 1992. Exogenous retinoic acid rapidly induces anterior ectopic expression of murine Hox-2 genes in vivo. Development 116, 357–368. Deschamps, J., Wijgerde, M., 1993. Two phases in the establishment of HOX expression domains. Dev. Biol. 156, 473–480. Deschamps, J., van den Akker, E., Forlani, S., de Graaff, W., Oosterveen, T., Roelen, B., Roelfsema, J., 1999. Initiation, establishment and maintenance of Hox gene expression patterns in the mouse. Int. J. Dev. Biol. 43, 635–650. Downs, K.M., Davies, T., 1993. Staging of gastrulating mouse embryos by morphological landmarks in the dissecting microscope. Development 118, 1255–1266. 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., 1994. Temporal colinearity and the phylotypic progression: a basis for the stability of a vertebrate Bauplan and the evolution of morphologies through heterochrony. Development (Suppl.), 135–142. Frohman, M.A., Boyle, M., Martin, G.R., 1990. Isolation of the mouse Hox2.9 gene; analysis of embryonic expression suggests that positional information along the anterior–posterior axis is specified by mesoderm. Development 110, 589–607. Gamer, L.W., Wright, C.V.E., 1993. Murine Cdx-4 bears striking similarities to the Drosophila caudal gene in its homeodomain sequence and early expression pattern. Mech. Dev. 43, 71–81. Gould, A., Itasaki, N., Krumlauf, R., 1998. Initiation of rhombomeric Hoxb4 expression requires induction by somitez and a retinoid pathway. Neuron 21, 39–51. Graham, A., Papalopulu, N., Lorimer, J., McVey, J.H., Tuddenham, E.G.D., Krumlauf, R., 1988. Characterization of a murine homeo box gene, Hox-2.6, related to the Drosophila Deformed gene. Genes Dev. 2, 1424–1438. Houle, M., Prinos, P., Iulianella, A., Bouchard, N., Lohnes, D., 2000. Retinoic acid regulation of Cdx1: an indirect mechanism for retinoids and vertebral specification. Mol. Cell. Biol. 20, 6579–6586. Huang, D., Chen, S.W., Langston, A.W., Gudas, L.J., 1998. A conserved retinoic acid responsive element in the murine Hoxb-1 gene is required for expression in the developing gut. Development 125, 3235–3246. James, R., Erler, T., Kazenwadel, J., 1994. Structure of the murine homeobox gene cdx-2. J. Biol. Chem. 269, 15229–15237. Kmita, M., van der Hoeven, F., Za´ ka´ ny, J., Krumlauf, R., Duboule, D.,
90
B.A.J. Roelen et al. / Mechanisms of Development 119 (2002) 81–90
2000. Mechanisms of Hox gene colinearity: transposition of the anterior Hoxb1 gene into the posterior HoxD complex. Genes Dev. 14, 198–211. Kondo, T., Duboule, D., 1999. Breaking colinearity in the mouse HoxD complex. Cell 97, 407–417. Kwan, C.T., Tsang, S.L., Krumlauf, R., Sham, M.H., 2001. Regulatory analysis of the mouse Hoxb3 gene: multiple elements work in concert to direct temporal and spatial patterns of expression. Dev. Biol. 232, 176–190. Maconochie, M.K., Nonchev, S., Studer, M., Chan, S.K., Popperl, H., Sham, M.H., Mann, R.S., Krumlauf, R., 1997. Cross-regulation in the mouse HoxB complex: the expression of Hoxb2 in rhombomere 4 is regulated by Hoxb1. Genes Dev. 11, 1885–1895. Manzanares, M., Bel-Vialar, S., Ariza-McNaughton, L., Ferretti, E., Marshall, H., Maconochie, M.M., Blasi, F., Krumlauf, R., 2001. Independent regulation of initiation and maintenance phases of Hoxa3 expression in the vertebrate hindbrain involve auto- and cross-regulatory mechanisms. Development 128, 3595–3607. Marshall, H., Nonchev, S., Sham, M.H., Muchamore, I., Lumsden, A., Krumlauf, R., 1992. Retinoic acid alters hindbrain Hox code and induces transformation of rhombomeres 2/3 into a 4/5 identity. Nature 360, 737–741. Marshall, H., Studer, M., Po¨ pperl, H., Aparicio, S., Kuroiwa, A., Brenner, S., Krumlauf, R., 1994. A conserved retinoic acid response element required for early expression of the homeobox gene Hoxb-1. Nature 370, 567–571. Meyer, B.I., Gruss, P., 1993. Mouse Cdx-1 expression during gastrulation. Development 117, 191–203. Murphy, P., Davidson, D.R., Hill, R.E., 1989. Segment-specific expression of a homoeobox-containing gene in the mouse hindbrain. Nature 341, 156–159. Niederreither, K., Subbarayan, V., Dolle´ , P., Chambon, P., 1999. Embryonic retinoic acid synthesis is essential for early mouse post-implantation development. Nat. Genet. 21, 444–448. Prinos, P., Joseph, S., Oh, K., Meyer, B.I., Gruss, P., Lohnes, B., 2001. Multiple pathways governing Cdx1 expression during murine development. Dev. Biol. 239, 257–269.
Rubock, M.J., Larin, Z., Cook, M., Papalopulu, N., Krumlauf, R., Lehrach, H., 1990. A yeast artificial chromosome containing the mouse homeobox cluster Hox-2. Proc. Natl. Acad. Sci. USA 87, 4751–4755. Schughart, K., Utset, M.F., Awgulewitsch, A., Ruddle, F.H., 1988. Structure and expression of Hox-2.2, a murine homeobox-containing gene. Proc. Natl. Acad. Sci. USA 85, 5582–5586. Sham, M.H., Hunt, P., Nonchev, S., Papalopulu, N., Graham, A., Boncinelli, E., Krumlauf, R., 1992. Analysis of the murine Hox-2.7 gene: conserved alternative transcripts with differential distribution in the nervous system and the potential for shared regulatory regions. EMBO J. 11, 1825–1836. Studer, M., Gavalas, A., Marshall, H., Ariza-McNaughton, L, Rijli, P.M., Chambon, P., Krumlauf, R., 1998. Genetic interactions between Hoxa1 and Hoxb1 reveal new roles in regulation of early hindbrain patterning. Development 125, 1025–1036. Studer, M., Popperl, H., Marshall, H., Kuroiwa, A., Krumlauf, R., 1994. Role of a conserved retinoic acid response element in rhombomere restriction of Hoxb-1. Science 265, 1728–1732. Subramanian, V., Meyer, B.I., Gruss, P., 1995. Disruption of the murine homeobox gene Cdx1 affects axial skeletal identities by altering the mesodermal expression domains of Hox genes. Cell 83, 641–653. Van den Akker, E., Forlani, S., Chawengsaksophak, K., de Graaff, W., Beck, F., Meyer, B., Deschamps, J., 2002. Cdx1 and Cdx2 have overlapping functions in anteroposterior patterning and posterior axis elongation. Development 129, 2181–2193. Van der Hoeven, F., Za´ ka´ ny, J., Duboule, D., 1996. Gene transpositions in the HoxD complex reveal a hierarchy of regulatory controls. Cell 85, 1025–1035. Wilkinson, D.G. (Ed.), 1992. In Situ Hybridization, a Practical Approach IRL Press, New York, NY. Wilkinson, D.G., Bhatt, S., Cook, M., Boncinelli, E., Krumlauf, R., 1989. Segmental expression of Hox-2 homoeobox-containing genes in the developing mouse hindbrain. Nature 341, 405–409. Zhang, F., Nagy Kovacs, E., Featherstone, M.S., 2000. Murine Hoxd4 expression in the CNS requires multiple elements including a retinoic acid response element. Mech. Dev. 96, 79–89.