Differential antero-posterior expression of two proteins encoded by a homeobox gene in Xenopus and mouse embryos

Molecular and evolutionary basis of limb field specification and limb initiation

Specification of limb field and initiation of limb development involve multiple steps, each of which is tightly regulated both spatially and temporally. Recent developmental analyses on various vertebrates have provided insights into the molecular mechanisms that specify limb field and have revealed several genetic interactions of signals involved in limb initiation processes. Furthermore, new approaches to the study of the developmental mechanisms of the lateral plate mesoderm of amphioxus and lamprey embryos have given us clues to understand the evolutionary scenarios that led to the acquisition of paired appendages during evolution. This review highlights such recent findings and discusses the mechanisms of limb field specification and limb bud initiation during development and evolution.

Development and evolution of the lateral plate mesoderm: comparative analysis of amphioxus and lamprey with implications for the acquisition of paired fins

Possession of paired appendages is regarded as a novelty that defines crown gnathostomes and allows sophisticated behavioral and locomotive patterns. During embryonic development, initiation of limb buds in the lateral plate mesoderm involves several steps. First, the lateral plate mesoderm is regionalized into the cardiac mesoderm (CM) and the posterior lateral plate mesoderm (PLPM). Second, in the PLPM, Hox genes are expressed in a collinear manner to establish positional values along the anterior-posterior axis. The developing PLPM splits into somatic and splanchnic layers. In the presumptive limb field of the somatic layer, expression of limb initiation genes appears. To gain insight into the evolutionary sequence leading to the emergence of paired appendages in ancestral vertebrates, we examined the embryonic development of the ventral mesoderm in the cephalochordate amphioxus Branchiostoma floridae and of the lateral plate mesoderm in the agnathan lamprey Lethenteron japonicum, and studied the expression patterns of cognates of genes known to be expressed in these mesodermal layers during amniote development. We observed that, although the amphioxus ventral mesoderm posterior to the pharynx was not regionalized into CM and posterior ventral mesoderm, the lateral plate mesoderm of lampreys was regionalized into CM and PLPM, as in gnathostomes. We also found nested expression of two Hox genes (LjHox5i and LjHox6w) in the PLPM of lamprey embryos. However, histological examination showed that the PLPM of lampreys was not separated into somatic and splanchnic layers. These findings provide insight into the sequential evolutionary changes that occurred in the ancestral lateral plate mesoderm leading to the emergence of paired appendages.

Reshuffling genomic landscapes to study the regulatory evolution of Hox gene clusters

The emergence of Vertebrata was accompanied by two rounds of whole-genome duplications. This enabled paralogous genes to acquire novel functions with high evolutionary potential, a process suggested to occur mostly by changes in gene regulation, rather than in protein sequences. In the case of Hox gene clusters, such duplications favored the appearance of distinct global regulations. To assess the impact of such "regulatory evolution" upon neo-functionalization, we developed PANTHERE (PAN-genomic Translocation for Heterologous Enhancer RE-shuffling) to bring the entire megabase-scale HoxD regulatory landscape in front of the HoxC gene cluster via a targeted translocation in vivo. At this chimeric locus, Hoxc genes could both interpret this foreign regulation and functionally substitute for their Hoxd counterparts. Our results emphasize the importance of evolving regulatory modules rather than their target genes in the process of neo-functionalization and offer a genetic tool to study the complexity of the vertebrate regulatory genome.

XMeis3 is necessary for mesodermal Hox gene expression and function

Hox transcription factors provide positional information during patterning of the anteroposterior axis. Hox transcription factors can co-operatively bind with PBC-class co-factors, enhancing specificity and affinity for their appropriate binding sites. The nuclear localisation of these co-factors is regulated by the Meis-class of homeodomain proteins. During development of the zebrafish hindbrain, Meis3 has previously been shown to synergise with Hoxb1 in the autoregulation of Hoxb1. In Xenopus XMeis3 posteriorises the embryo upon ectopic expression. Recently, an early temporally collinear expression sequence of Hox genes was detected in Xenopus gastrula mesoderm (see intro. P3). There is evidence that this sequence sets up the embryo's later axial Hox expression pattern by time-space translation. We investigated whether XMeis3 is involved in regulation of this early mesodermal Hox gene expression. Here, we present evidence that XMeis3 is necessary for expression of Hoxd1, Hoxb4 and Hoxc6 in mesoderm during gastrulation. In addition, we show that XMeis3 function is necessary for the progression of gastrulation. Finally, we present evidence for synergy between XMeis3 and Hoxd1 in Hoxd1 autoregulation in mesoderm during gastrulation.

Xwnt8 directly initiates expression of labial Hox genes

Hox transcription factors play an essential role in patterning the anteroposterior axis during embryogenesis and exhibit a complex array of spatial and temporal patterns of expression. Their earliest onset of expression in vertebrates is during gastrulation in a temporally collinear sequence in the presomitic/ventrolateral mesoderm, and it is not clear which upstream signal transduction events initiate this expression. Using Xenopus, we present evidence that Xwnt8 is necessary for initiation of this collinear sequence by activating Hox-1 expression in three Hox clusters: hoxd, hoxa, and hoxb. All three labial genes appear to be direct targets of canonical Wnt signaling through Tcf/Lef. In addition, Xwnt8 loss- and gain-of-function leads to indirect regulation of other Hox genes: Hoxb4, Hoxd4, Hoxa7, Hoxc6, and Hoxc8. These findings shed new light on the early role of Wnt8 as well as of a proposed WNT gradient in patterning the Xenopus central nervous system (Kiecker and Niehrs [2001] Development 128:4189-4201).

Two Hoxc6 transcripts are differentially expressed and regulate primary neurogenesis in Xenopus laevis

Hox genes are key players in defining positional information along the main body axis of vertebrate embryos. In Xenopus laevis, Hoxc6 was the first homeobox gene isolated. It encodes two isoforms. We analyzed in detail their spatial and temporal expression pattern during early development. One major expression domain of both isoforms is the spinal cord portion of the neural tube. Within the spinal cord and its populations of primary neurons, Hox genes have been found to play a crucial role for defining positional information. Here we report that a loss-of-function of either one of the Hoxc6 products does not affect neural induction, the expression of general neural markers is not modified. However, Hoxc6 does widely affect the formation of primary neurons within the developing neural tissue. Manipulations of Hoxc6 expression severly changes the expression of the neuronal markers N-tubulin and Islet-1. Formation of primary neurons and formation of cranial nerves are affected. Hence, Hoxc6 functions are not restricted to the expected role in anterior-posterior pattern formation, but they also regulate N-tubulin, thereby having an effect on the initial formation of primary neurons in Xenopus laevis embryos.

How to build transcriptional network models of mammalian pattern formation

Background

Genetic regulatory networks of sequence specific transcription factors underlie pattern formation in multicellular organisms. Deciphering and representing the mammalian networks is a central problem in development, neurobiology, and regenerative medicine. Transcriptional networks specify intermingled embryonic cell populations during pattern formation in the vertebrate neural tube. Each embryonic population gives rise to a distinct type of adult neuron. The homeodomain transcription factor Lbx1 is expressed in five such populations and loss of Lbx1 leads to distinct respecifications in each of the five populations.

Methodology/Principal Findings

We have purified normal and respecified pools of these five populations from embryos bearing one or two copies of the null Lbx1^GFP allele, respectively. Microarrays were used to show that expression levels of 8% of all transcription factor genes were altered in the respecified pool. These transcription factor genes constitute 20–30% of the active nodes of the transcriptional network that governs neural tube patterning. Half of the 141 regulated nodes were located in the top 150 clusters of ultraconserved non-coding regions. Generally, Lbx1 repressed genes that have expression patterns outside of the Lbx1-expressing domain and activated genes that have expression patterns inside the Lbx1-expressing domain.

Conclusions/Significance

Constraining epistasis analysis of Lbx1 to only those cells that normally express Lbx1 allowed unprecedented sensitivity in identifying Lbx1 network interactions and allowed the interactions to be assigned to a specific set of cell populations. We call this method ANCEA, or active node constrained epistasis analysis, and think that it will be generally useful in discovering and assigning network interactions to specific populations. We discuss how ANCEA, coupled with population partitioning analysis, can greatly facilitate the systematic dissection of transcriptional networks that underlie mammalian patterning.

Evo-devo: variations on ancestral themes

Most animals evolved from a common ancestor, Urbilateria, which already had in place the developmental genetic networks for shaping body plans. Comparative genomics has revealed rather unexpectedly that many of the genes present in bilaterian animal ancestors were lost by individual phyla during evolution. Reconstruction of the archetypal developmental genomic tool-kit present in Urbilateria will help to elucidate the contribution of gene loss and developmental constraints to the evolution of animal body plans.

The developmental bases of limb reduction and body elongation in squamates

Employing an integrative approach to investigate the evolution of morphology can yield novel perspectives not attainable from a single field of study. Studies of limb loss and body elongation in squamates (snakes and lizards) present a good example in which integrating studies of systematics and ecology with genetics and development can provide considerable new insight. In this comment we address several misunderstandings of the developmental genetic literature presented in a paper by Wiens and Slingluff (2001) to counter their criticism of previous work in these disciplines and to clarify the apparently contradictory data from different fields of study. Specifically, we comment on (1) the developmental mechanisms underlying axial regionalization, body elongation, and limb loss; (2) the utility of presacral vertebral counts versus more specific partitioning of the primary body axis; (3) the independent, modular nature of limbs and limb girdles and their utility in diagnosing genetic changes in development; and (4) the causal bases of hind limb reduction in ophidian and nonophidian squamates.

Determination of anterior endoderm in Xenopus embryos

Grafts of anterior endoderm from embryos at stage 28 or later developed according to their fate (i.e., anterior) when transplanted posteriorly. Conversely, grafts from earlier embryos developed according to their new location (i.e., posterior). However, endoderm grafted along with its mesodermal and ectodermal sheath retained its fate regardless of the stage of the donor. We conclude that anterior endoderm in Xenopus embryos is determined at about stage 25 under the influence of mesoderm.

FGF signaling and the anterior neural induction in Xenopus

We previously showed that FGF was capable of inducing Xenopus gastrula ectoderm cells in culture to express position-specific neural markers along the anteroposterior axis in a dose-dependent manner. However, conflicting results have been obtained concerning involvement of FGF signaling in the anterior neural induction in vivo using the same dominant-negative construct of Xenopus FGF receptor type-1 (delta XFGFR-1 or XFD). We explored this issue by employing a similar construct of receptor type-4a (XFGFR-4a) in addition, since expression of XFGFR-4a was seen to peak between gastrula and neurula stages, when the neural induction and patterning take place, whereas expression of XFGFR-1 had not a distinct peak during that period. Further, these two FGFRs are most distantly related in amino acid sequence in the Xenopus FGFR family. When we injected mRNA of a dominant-negative version of XFGFR-4a (delta XFGFR-4a) into eight animal pole blastomeres at 32-cell stage, anterior defects including loss of normal structure in telencephalon and eye regions became prominent as examined morphologically or by in situ hybridization. Overexpression of delta XFGFR-1 appeared far less effective than that of delta XFGFR-4a. Requirement of FGF signaling in ectoderm for anterior neural development was further confirmed in culture: when ectoderm cells that were overexpressing delta XFGFR-4a were cocultured with intact organizer cells from either early or late gastrula embryos, expression of anterior and posterior neural markers was inhibited, respectively. We also showed that autonomous neuralization of the anterior-type observed in ectoderm cells that were subjected to prolonged dissociation was strongly suppressed by delta XFGFR-4a, but not as much by delta XFGFR-1. It is thus indicated that FGF signaling in ectoderm, mainly through XFGFR-4, is required for the anterior neural induction by organizer. We may reconcile our data to the current "neural default model," which features the central roles of BMP4 signaling in ectoderm and BMP4 antagonists from organizer, simply postulating that the neural default pathway in ectoderm includes constitutive FGF signaling step.

Developmental basis of limblessness and axial patterning in snakes

The evolution of snakes involved major changes in vertebrate body plan organization, but the developmental basis of those changes is unknown. The python axial skeleton consists of hundreds of similar vertebrae, forelimbs are absent and hindlimbs are severely reduced. Combined limb loss and trunk elongation is found in many vertebrate taxa, suggesting that these changes may be linked by a common developmental mechanism. Here we show that Hox gene expression domains are expanded along the body axis in python embryos, and that this can account for both the absence of forelimbs and the expansion of thoracic identity in the axial skeleton. Hindlimb buds are initiated, but apical-ridge and polarizing-region signalling pathways that are normally required for limb development are not activated. Leg bud outgrowth and signalling by Sonic hedgehog in pythons can be rescued by application of fibroblast growth factor or by recombination with chick apical ridge. The failure to activate these signalling pathways during normal python development may also stem from changes in Hox gene expression that occurred early in snake evolution.

The HOXC6 homeodomain-containing proteins

The HOXC6 homeodomain-containing proteins act as transcription factors in the genetic control of multiple genes involved in development and cell differentiation. Two HOXC6 polypeptides are encoded by a single homeobox ('HOX') gene described as 'master gene' for the crucial role it plays in the patterning and axial morphogenesis of multiple species. Transcription of the HOXC6 gene is initiated from two promoters and generates two proteins that share the same DNA-binding domain but harbor a distinct N-terminal region. Recent studies have demonstrated that both HOXC6 products can activate or repress transcription, depending on the cellular context. Functional in vivo specificity of HOXC6 proteins may be achieved through combinatorial interactions with other members of the HOX family as well as with co-factors whose identities are largely unknown. Disruption of this 'HOX code' may lead to pathology such as developmental defects.

Growth factors and dexamethasone regulate Hoxb5 protein in cultured murine fetal lungs

Studies on lung morphogenesis have indicated a role of homeobox (Hox) genes in the regulation of lung development. In the present study, we attempted to modulate the synthesis of Hoxb5 protein in cultured murine fetal lungs after mechanical or chemical stimuli. Murine fetuses at gestational day 14 (GD14) were removed from pregnant CD-1 mice, and lungs were excised and cultured for 7 days in BGJb media. The experimental groups were 1) untreated, unligated; 2) tracheal ligation; 3) supplemented media with either epidermal growth factor (EGF; 10 ng/ml), transforming growth factor (TGF)-beta 1 (2 ng/ml), dexamethasone (10 nM), EGF + TGF-beta 1, or EGF + TGF-beta 1 + dexamethasone. After 3 or 7 days, the cultured lungs were compared with in vivo lungs. Immunoblotting signals at 3 days in culture were stronger than those at 7 days. Western blot analyses showed that ligation, EGF, TGF-beta 1, and EGF + TGF-beta 1 downregulated Hoxb5 protein to approximately 20-70% of Hoxb5 protein levels in unligated, untreated cultured lungs. Furthermore, dexamethasone alone or in combination with EGF and TGF-beta 1 downregulated Hoxb5 protein by > 90% (P < 0.05) signal strength, similar to that seen in GD19 or in neonatal lungs. Immunostaining showed that Hoxb5 protein was expressed strongly in the lung mesenchyme at early stages in gestation. However, by GD19 and in neonates, it was present only in specific epithelial cells. A persistent level of Hoxb5 protein in the mesenchyme after EGF or TGF-beta 1 treatments or tracheal ligation was noted. Hoxb5 protein was significantly downregulated by EGF + TGF-beta 1, and it was least in lungs after dexamethasone or EGF + TGF-beta 1 + dexamethasone treatment. The decrease in Hoxb5 protein was significant only in the groups with dexamethasone added to the media. Thus immunostaining results parallel those of immunoblotting. The degree of Hoxb5 downregulation by dexamethasone or EGF + TGF-beta 1 + dexamethasone was similar to that seen in vivo in very late gestation, which correlated to the advancing structural development of the lung.

The Xenopus homeobox gene twin mediates Wnt induction of goosecoid in establishment of Spemann's organizer

We describe the isolation of the Xenopus homeobox gene twin (Xtwn), which was identified in an expression cloning screen for molecules with dorsalizing activities. Injection of synthetic Xtwn mRNA restores a complete dorsal axis in embryos lacking dorsal structures and induces a complete secondary dorsal axis when ectopically expressed in normal embryos. The sequence homology, expression pattern and gain-of-function phenotype of Xtwn is most similar to the previously isolated Xenopus homeobox gene siamois (Xsia) suggesting that Xtwn and Xsia comprise a new subclass of homeobox genes important in dorsal axis specification. We find that Xtwn is able to activate the Spemann organizer-specific gene goosecoid (gsc) via direct binding to a region of the gsc promoter previously shown to mediate Wnt induction. Since Xtwn expression is strongly induced in ectodermal (animal cap) cells in response to overexpression of a dorsalizing Wnt molecule, we examined the possibility that Xtwn might be a direct target of a Wnt signal transduction cascade. First, we demonstrate that purified LEF1 protein can interact, in vitro, with consensus LEF1/TCF3-binding sites found within the Xtwn promoter. Second, these binding sites were shown to be required for Wnt-mediated induction of a Xtwn reporter gene containing these sites. As LEF1/TCF3 family transcription factors have previously been shown to directly mediate Wnt signaling, these results suggest that Xtwn induction by Wnt may be direct. Finally, in UV-hyperventralized embryos, expression of endogenous Xtwn is confined to the vegetal pole and a Xtwn reporter gene is hyperinduced vegetally in a LEF1/TCF3-binding-site-dependent manner. These results suggest that cortical rotation distributes Wnt-like dorsal determinants to the dorsal side of the embryo, including the dorsal marginal zone, and that these determinants may directly establish Spemann's organizer in this region.

The Pax4 gene is essential for differentiation of insulin-producing beta cells in the mammalian pancreas

The mammalian pancreas contains two distinct cell populations: endocrine cells which secrete hormones into the bloodstream, and exocrine cells, which secrete enzymes into the digestive tract. The four endocrine cell types found in the adult pancreas-(alpha, beta, delta and PP-synthesize glucagon, insulin, somatostatin and pancreatic polypeptide, respectively. All of these endocrine cells arise from common multipotent precursors, which coexpress several hormones when they start to differentiate. Expression of some homeobox genes in the early developing pancreas has been reported. The Pax4 gene is expressed in the early pancreas, but is later restricted to beta cells. Inactivation of Pax4 by homologous recombination results in the absence of mature insulin- and somatostatin-producing cells (beta and delta, respectively) in the pancreas of Pax4 homozygous mutant mice, but glucagon-producing alpha cells are present in considerably higher numbers. We propose that the early expression of Pax4 in a subset of endocrine progenitors is essential for the differentiation of the beta and delta cell lineages. A default pathway would explain the elevated number of alpha cells in the absence of Pax4.

Analysis of Hox gene expression in the chick limb bud

The vertebrate Hox genes have been shown to be important for patterning the primary and secondary axes of the developing vertebrate embryo. The function of these genes along the primary axis of the embryo has been generally interpreted in the context of positional specification and homeotic transformation of axial structures. The way in which these genes are expressed and function during the development of the secondary axes, particularly the limb, is less clear. In order to provide a reference for understanding the role of the Hox genes in limb patterning, we isolated clones of 23 Hox genes expressed during limb development, characterized their expression patterns and analyzed their regulation by the signalling centers which pattern the limb. The expression patterns of the Abd-B-related Hoxa and Hoxd genes have previously been partially characterized; however, our study reveals that these genes are expressed in patterns more dynamic and complex than generally appreciated, only transiently approximating simple, concentric, nested domains. Detailed analysis of these patterns suggests that the expression of each of the Hoxa and Hoxd genes is regulated in up to three independent phases. Each of these phases appears to be associated with the specification and patterning of one of the proximodistal segments of the limb (upper arm, lower arm and hand). Interestingly, in the last of these phases, the expression of the Hoxd genes violates the general rule of spatial and temporal colinearity of Hox gene expression with gene order along the chromosome. In contrast to the Abd-B-related Hoxa and Hoxd genes, which are expressed in both the fore and hind limbs, different sets of Hoxc genes are expressed in the two limbs. There is a correlation between the relative position of these genes along the chromosome and the axial level of the limb bud in which they are expressed. The more 3′ genes are expressed in the fore limb bud while the 5′ genes are expressed in the hind limb bud; intermediate genes are transcribed in both limbs. However, there is no clear correlation between the relative position of the genes along the chromosome and their expression domains within the limb. With the exception of Hoxc-11, which is transcribed in a posterior portion of the hind limb, Hoxc gene expression is restricted to the anterior/proximal portion of the limb bud. Importantly, comparison of the distributions of Hoxc-6 RNA and protein products reveals posttranscriptional regulation of this gene, suggesting that caution must be exercised in interpreting the functional significance of the RNA distribution of any of the vertebrate Hox genes. To understand the genesis of the complex patterns of Hox gene expression in the limb bud, we examined the propagation of Hox gene expression relative to cell proliferation. We find that shifts in Hox gene expression cannot be attributed to passive expansion due to cell proliferation. Rather, phase-specific Hox gene expression patterns appear to result from a context-dependent response of the limb mesoderm to Sonic hedgehog. Sonic hedgehog (the patterning signal from the Zone of Polarizing Activity) is known to be able to activate Hoxd gene expression in the limb. Although we find that Sonic hedgehog is capable of initiating and polarizing Hoxd gene expression during both of the latter two phases of Hox gene expression, the specific patterns induced are not determined by the signal, but depend upon the temporal context of the mesoderm receiving the signal. Misexpression of Sonic hedgehog also reveals that Hoxb-9, which is normally excluded from the posterior mesenchyme of the leg, is negatively regulated by Sonic hedgehog and that Hoxc-11, which is expressed in the posterior portion of the leg, is not affected by Sonic hedgehog and hence is not required to pattern the skeletal elements of the lower leg.

Developmentally regulated expression of a cell surface protein, neuropilin, in the mouse nervous system

Neuropilin (previously A5) is a cell surface glycoprotein that was originally identified in Xenopus tadpole nervous tissues. In Xenopus, neuropilin is expressed on both the presynaptic and postsynaptic elements in the visual and general somatic sensory systems, suggesting a role in neuronal cell recognition. In this study, we identified a mouse homologue of neuropilin and examined its expression in developing mouse nervous tissues. cDNA cloning and sequencing revealed that the primary structure of the mouse neuropilin was highly similar to that of Xenopus and that the extracellular segment of the molecule possessed several motifs that were expected to be involved in cell-cell interaction. Immunohistochemistry and in situ hybridization analyses in mice indicated that the expression of neuropilin was restricted to particular neuron circuits. Neuropilin protein was localized on axons but not on the somata of neurons. The expression of neuropilin persisted through the time when axons were actively growing to form neuronal connections. These observations suggest that neuropilin is involved in growth, fasciculation, and targeting for a particular groups of axons.

bFGF as a possible morphogen for the anteroposterior axis of the central nervous system in Xenopus

Vertebrate neural development is initiated during gastrulation by the inductive action of the dorsal mesoderm (Spemann's organizer in amphibians) on neighbouring ectoderm, which eventually gives rise to the central nervous system from forebrain to spinal cord. Here we present evidence that bFGF can mimic the organizer action by inducing Xenopus ectoderm cells in culture to express four position-specific neural markers (XeNK-2, En-2, XIHbox1 and XIHbox6) along the anteroposterior axis. bFGF also induced the expression of a general neural marker NCAM but not the expression of immediate-early mesoderm markers (goosecoid, noggin, Xbra and Xwnt-8), suggesting that bFGF directly neuralized ectoderm cells without forming mesodermal cells. The bFGF dose required to induce the position-specific markers was correlated with the anteroposterior location of their expression in vivo, with lower doses eliciting more anterior markers and higher doses more posterior markers. These data indicate that bFGF or its homologue is a promising candidate for a neural morphogen for anteroposterior patterning in Xenopus. Further, we showed that the ability of ectoderm cells to express the anterior markers in response to bFGF was lost by mid-gastrula, before the organizer mesoderm completely underlies the anterior dorsal ectoderm. Thus, an endogenous FGF-like molecule released from the involuting organizer may initiate the formation of the anteroposterior axis of the central nervous system during the early stages of gastrulation by forming a concentration gradient within the plane of dorsal ectoderm.

Hox genes and the evolution of vertebrate axial morphology

A common form of evolutionary variation between vertebrate taxa is the different numbers of segments that contribute to various regions of the anterior-posterior axis; cervical vertebrae, thoracic vertebrae, etc. The term ‘transposition’ is used to describe this phenomenon. Genetic experiments with homeotic genes in mice have demonstrated that Hox genes are in part responsible for the specification of segmental identity along the anterior-posterior axis, and it has been proposed that an axial Hox code determines the morphology of individual vertebrae (Kessel, M. and Gruss, P. (1990) Science 249, 347–379). This paper presents a comparative study of the developmental patterns of homeobox gene expression and developmental morphology between animals that have homologous regulatory genes but different morphologies. The axial expression boundaries of 23 Hox genes were examined in the paraxial mesoderm of chick, and 16 in mouse embryos by in situ hybridization and immunolocalization techniques. Hox gene anterior expression boundaries were found to be transposed in concert with morphological boundaries. This data contributes a mechanistic level to the assumed homology of these regions in vertebrates. The recognition of mechanistic homology supports the historical homology of basic patterning mechanisms between all organisms that share these genes.

Neural induction and regionalisation by different subpopulations of cells in Hensen's node

Cell lineage analysis has revealed that the amniote organizer, Hensen's node, is subdivided into distinct regions, each containing a characteristic subpopulation of cells with defined fates. Here, we address the question of whether the inducing and regionalising ability of Hensen's node is associated with a specific subpopulation. Quail explants from Hensen's node are grafted into an extraembryonic site in a host chick embryo allowing host- and donor-derived cells to be distinguished. Cell-type- and region-specific markers are used to assess the fates of the mesodermal and neural cells that develop. We find that neural inducing ability is localised in the epiblast layer and the mesendoderm (deep portion) of the medial sector of the node. The deep portion of the posterolateral part of the node does not have neural inducing ability. Neural induction also correlates with the presence of particular prospective cell types in our grafts: chordamesoderm (notochord/head process), definitive (gut) endoderm or neural tissue. However, only grafts that include the epiblast layer of the node induce neural tissue expressing a complete range of anteroposterior characteristics, although prospective prechordal plate cells may also play a role in specification of the forebrain.

Nucleolar localisation of three Hox homeoproteins

Homeoproteins encoded by genes of the Hox family are nuclear proteins believed to act as transcription factors and to participate in the determination of the body plan. Here we show that in several vertebrate cells, they exhibit a subnuclear localisation associated with the nucleolus. We used monoclonal antibodies to study the distribution of three homeoproteins, namely HOXB7, HOXC6 and HOXD4. The immunoreactivity to antibodies against HOXC6 protein in Xenopus laevis embryonic tissues is restricted to one or two spots within the nucleus; this distribution partially overlaps that of fibrillarin, a protein of the fibrillar zone of the nucleoli. Indirect immunofluorescence analysis of the distribution of HOXB7 protein in 3T3 cells, and of HOXD4 protein in human neuroblastoma and Raji lymphoma cell lines and activated lymphocytes, results invariably in a nucleolar localisation. Purified nucleoli from stimulated T lymphocytes, and Raji cells contain an activity capable of binding, in a gel retardation assay, to an oligonucleotide specifically recognised by the HOXD4 homeoprotein. This activity is specifically removed by anti-HOXD4 antibodies and is found associated in southwestern blots with a single band with an apparent M(r) of 30,000, corresponding to that of recombinant HOXD4. The functional significance of the nucleolar localisation of Hox proteins remains to be determined.

Expression of the LIM class homeobox gene Xlim-1 in pronephros and CNS cell lineages of Xenopus embryos is affected by retinoic acid and exogastrulation

The LIM class homeobox gene Xlim-1 is expressed in Xenopus embryos in the lineages leading to (i) the notochord, (ii) the pronephros, and (iii) certain cells of the central nervous system (CNS). In its first expression phase, Xlim-1 mRNA arises in the Spemann organizer region, accumulates in prechordal mesoderm and notochord during gastrulation, and decays in these tissues during neurula stages except that it persists in the posterior tip of the notochord. In the second phase, expression in lateral mesoderm begins at late gastrula, and converges to the pronephros at tailbud stages. Expression in a central location of the neural plate also initiates at late gastrula, expands anteriorly and posteriorly, and becomes established in the lateral regions of the spinal cord and hindbrain at tailbud stages. Thus Xlim-1 expression precedes morphogenesis, suggesting that it may be involved in cell specification in these lineages. Enhancement of Xlim-1 expression by retinoic acid (RA) was first detectable in the dorsal mesoderm at initial gastrula. During gastrulation and early neurulation, RA strongly enhanced Xlim-1 expression in all three lineages and also expanded its expressing domains; this overexpression correlated well with RA phenotypes such as enlarged pronephros and hindbrain-like structure. Exogastrulation reduced Xlim-1 expression in the lateral mesoderm and ectoderm but not in the notochord, suggesting that the second phase of Xlim-1 expression requires mesoderm/ectoderm interactions. RA treatment of exogastrulae did not revert this reduction.

Vertebrate homeobox genes

In the former part of the review the principal available data about Hox genes, their molecular organisation and their expression in vertebrate embryos, with particular emphasis for mammals, are briefly summarized. In the latter part we analysed the expression of four mouse homeobox genes related to two Drosophila genes expressed in the developing head of the fly: Emx1 and Emx2, related to ems, and Otx1 and Otx2, related to otd.

Direct regulation of decapentaplegic by Ultrabithorax and its role in Drosophila midgut morphogenesis

Drosophila homeotic genes encode transcription factors thought to control segmental identity by regulating expression of largely unknown target genes. The formation of the second midgut constriction requires the Ultrabithorax (Ubx) and abdominal-A (abd-A) homeotic genes and decapentaplegic (dpp), a gene encoding a member of the TGF beta family of proteins. We identified a 674 bp enhancer of dpp controlling its expression in the second constriction domain of the visceral mesoderm (parasegment 7). Normal enhancer function requires positive regulation by Ubx and negative regulation by abd-A. This enhancer contains UBX- and ABD-A-binding sites defined in vitro. By generating complementary alterations of the binding sites and the binding specificity of UBX, we show that Ubx directly regulates dpp expression. These regulatory interactions are relevant to normal development, because a transgene made with this enhancer driving a dpp transcription unit rescues the second midgut constriction and larval lethality phenotypes of dpps mutations.

Spatial localisation of transcripts of the Hox-C6 gene

The Hox-C6 gene, in common with its Xenopus and human homologues, is organised as 3 exons with 2 distinct promoters (PRI and PRII) producing 2 different transcripts. The PRII promoter produces a 'typical' Hox transcript by splicing 2 exons separated by a 700 bp intron. The PRI promoter initiates transcription from a 3rd exon, 9 kbp 5' of PRII. This exon is spliced into the 1st exon of the PRII transcript at a point 3' of the translation start codon. This means that the predicted protein product of the PRI protein contains the Hox-C6 homeodomain but is truncated by 82 amino acids compared with the PRII product. In situ hybridisation using probes specific for PRI or PRII showed them to have essentially the same distribution in the developing nervous system and prevertebrae of 12.5 d embryos. However, the distribution of transcripts in the CNS was distinctly different from that reported previously for a probe containing the Hox-C6 homeobox.

Specification of axial identity in the mouse: role of the Hoxa-5 (Hox1.3) gene

Numerous lines of study have suggested that the Hox genes, encoding putative transcription factors, are key genes in the establishment of the body plan of the mammalian embryo. To examine the role of Hoxa-5 (Hox1.3) gene during development, we have used targeted mutagenesis in embryonic stem cells to produce a strain of mice carrying a disrupted Hoxa-5 allele. The viability of homozygous mutant mice is markedly reduced, with 50% of the mutant animals dying at birth or shortly thereafter. Analysis of the skeleton of Hoxa-5 mutants reveals a number of homeotic transformations restricted to the cervical and thoracic regions. Of these, one of the most frequent morphological abnormalities is the posterior transformation of the seventh cervical vertebra into the likeness of a thoracic vertebra complete with a pair of ribs. These results demonstrate that the Hoxa-5 gene has an important role in the establishment of the skeleton during development and contributes to the process whereby the axial structures are determined.

Planar and vertical induction of anteroposterior pattern during the development of the amphibian central nervous system

In amphibians and other vertebrates, neural development is induced in the ectoderm by signals coming from the dorsal mesoderm during gastrulation. Classical embryological results indicated that these signals follow a "vertical" path, from the involuted dorsal mesoderm to the overlying ectoderm. Recent work with the frog Xenopus laevis, however, has revealed the existence of "planar" neural-inducing signals, which pass within the continuous sheet or plane of tissue formed by the dorsal mesoderm and presumptive neurectoderm. Much of this work has made use of Keller explants, in which dorsal mesoderm and ectoderm are cultured in a planar configuration with contact along only a single edge, and vertical contact is prevented. Planar signals can induce the full anteroposterior (A-P) extent of neural pattern, as evidenced in Keller explants by the expression of genes that mark specific positions along the A-P axis. In this review, classical and modern molecular work on vertical and planar induction will be discussed. This will be followed by a discussion of various models for vertical induction and planar induction. It has been proposed that the A-P pattern in the nervous system is derived from a parallel pattern of inducers in the dorsal mesoderm which is "imprinted" vertically onto the overlying ectoderm. Since it is now known that planar signals can also induce A-P neural pattern, this kind of model must be reassessed. The study of planar induction of A-P pattern in Xenopus embryos provides a simple, manipulable, two-dimensional system in which to investigate pattern formation.

Changes in the prevalence of an homeobox gene product during muscle differentiation

We have studied by immunohistofluorescence and confocal microscopy the localization of the XlHbox-1 protein, the product of a Xenopus class 1 homeobox gene corresponding to the human HOX 3C, during the development of Xenopus laevis mesodermal derivatives. The protein, not present at early stages of embryonic development, can first be detected in the neurula where it is weakly expressed in the rostral part of the spinal cord and in the nuclei of the corresponding somites. At later stages of mesodermal development, very high levels of the molecule are present in the nuclei of a small group of myogenic cells in the most dorsal aspect of the myotome, while the nuclei of differentiated muscle fibers within the myotome are either stained weakly or completely negative. A similar transient expression of XlHbox-1 gene product during myogenesis occurs during muscle differentiation in the limb bud and during differentiation of visceral smooth muscles from the lateral plate mesoderm. In both cases the nuclei of precursor cells contain high level of this protein which is rapidly down regulated during further muscle differentiation. In myogenic areas the modulation of XlHbox-1 expression invariably parallels that of the neural cell adhesion molecule N-CAM. These data are the first evidence that a homeobox gene belonging to the Antennapedia-Bithorax complex is transiently expressed in early phases of muscle differentiation. The transient expression of homeobox genes in early phases of embryonic development could act synergistically with the expression of other myogenic transcriptional factors to specify a fine level of differentiation of the muscle cells along the body axis.

Hoxb-4 (Hox-2.6) mutant mice show homeotic transformation of a cervical vertebra and defects in the closure of the sternal rudiments

Two Hoxb-4 (Hox-2.6) mutations were introduced into the mouse germline. The overt phenotype caused by one of the mutations was assayed on two different genetic backgrounds, an inbred 129SvEv and a hybrid 129SvEv-C57BL/6J. The allele hoxb-4' is a disruption of the first exon and causes two obvious skeletal changes: a partial homeotic transformation of the second cervical vertebra from axis to atlas and a defective morphogenesis of the sternum. Both phenotypes have incomplete penetrance and variable expressivity when assayed in the hybrid genetic background, but the sternum defect is completely penetrant in the inbred background. The mutant allele hoxb-4s has a premature stop codon, introduced by the "hit and run" method in the second exon, that disrupts the third helix of the homeodomain. This allele also causes the partial homeotic transformation of axis to atlas, but it does not affect the sternum.

Immunodetection of Xenopus bone morphogenetic protein-4 in early embryos

Specific antibodies to Xenopus laevis bone morphogenetic protein-4 (xBMP-4) were raised by immunizing rabbits with a fusion protein of bacterial beta-galactosidase and xBMP-4. The antibodies were used to detect xBMPs expressed in mammalian cells by Western blotting. The antibodies were found to recognize xBMP-4 specifically and not to cross-react with either xBMP-2 or xBMP-7 which are similar to xBMP-4. In addition, the antibodies recognized dimeric xBMP-4 whereas our previous antibodies recognized the reduced form only. The present antibodies detected an immunoreactive 27 kDa protein in extracts of developing Xenopus embryos from oocyte to tailbud embryo. The xBMP-4 peptide appeared to be monomeric in structure because the molecular weight did not shift upon reduction of disulfide bond(s).

Characterization of the Xenopus Hox 2.4 gene and identification of control elements in its intron

We report on the Xenopus homolog of the Hox 2.4 gene. This gene occupies the next to 5'-most position in the Xenopus Hox2 complex. Hox 2.4 RNA is first detected at the early neurula stage, reaching a peak at the early tailbud stage, and is localized in the middle and posterior portions of the embryos. Antibodies raised against a fusion protein show expression of Hox 2.4 protein in Xenopus embryos in a band located in the mid spinal cord. Thus, the protein is expressed in a narrower domain than that of Hox 2.4 mRNA. The Xenopus Hox 2.4 antibody cross-reacts readily with mouse embryonic tissue, where the protein is detected in migrating neural crest cells, the dorsal portion of the spinal cord, somites, lateral plate mesoderm, and in the forelimb bud. The Xenopus Hox 2.4 intron shares considerable sequence identity with the intron in the mouse homolog. A reporter gene containing an element from this intron which can bind homeodomain proteins is activated following microinjection into Xenopus embryos. The short distance between the end of the Hox 2.4 cDNA and the start site of the neighboring gene in the complex raises the possibility that this transcriptional element might be shared by two Hox genes.

A Hox 3.3-lacZ transgene expressed in developing limbs

We describe transgenic mouse lines that express lacZ under the control of the Hox 3.3 Promoter II. The correct anterior boundary can be fixed by 3.6 kb of promoter DNA (plus 1.6 kb of 5' transcribed sequences), both in tissues of ectodermal and mesodermal origin. The posterior border, however, is not respected, and lacZ expression continues into the tail region. One line has particularly strong graded expression in the anterior proximal limb bud. Other lines, containing a shorter promoter fragment (0.6 kb), have ectopic expression in the head region, including one line that has expression in the anterior half of the retina. Such mouse lines make it possible to molecularly distinguish cells in regions of the embryo that look otherwise identical and may be useful in studying the establishment of molecular differences in the mouse embryo.

Homeotic transformations in the mouse induced by overexpression of a human Hox3.3 transgene

A permanent transgenic mouse line was generated carrying 40 copies of the human Hox3.3 gene. The resulting mice express large amounts of Hox3.3 protein in posterior regions of the embryo where this homeodomain protein is normally not expressed. The transgene causes homeotic transformations of the skeleton, in particular the appearance of an extra pair of ribs in the lumbar region, transformation of the shape of posterior ribs into that of more anterior ones, and the joining of an additional pair of ribs to the sternum. The phenotype of this line resembles that obtained by the targeted loss-of-function mutation of Hox3.1 (Le Mouellic et al., 1992). In transient assays, the human Hox3.3 transgene leads to the formation of additional ribs in more posterior vertebrae as well.

Two distal-less related homeobox-containing genes expressed in regeneration blastemas of the newt

Urodeles, like the newt, are able to replace their limbs and tail following amputation by the formation of a blastema, a mass of proliferating mesenchymal cells originating from the tissue adjacent to the cut surface. As this capacity may involve genetic control, we investigated in adult tissues the expression of genes controlling embryonic development. We screened a newt cDNA library with a redundant oligonucleotide specific to the highly conserved third helix of the DNA-binding domain of homeobox genes. Five classes of cDNA have been isolated. We report the nucleotide sequence and the tissue distribution of two of them, NvHBox-4 and NvHBox-5. The amino acid sequences of both homeodomains are highly homologous (83 and 87% identity) to distal-less, a Drosophila homeobox gene expressed during the development of appendages. NvHBox-4 and NvHBox-5 express respectively 2.8 and 2 kb transcripts. The pattern of expression of both genes is identical in adult tissues of the newt. Polyadenylated transcripts are detectable in the forelimbs, hindlimbs, the tail, flank, and brain as well as in limb and tail blastemas. Analysis of dissected tissue from the hindlimbs indicated that the expression of both genes is restricted to the skin. This work is a first step toward understanding the possible relation between sustained expression of homeobox-containing genes in adult newt tissues and regeneration potential.

Cell-surface changes induced by ectopic expression of the murine homeo☐ gene Hox-3.3

Murine homeobox-containing genes (Hox genes) are postulated as playing key roles in the establishment of the anterior-posterior embryonic body axis, possibly providing cells with positional cues. Little is known, however, concerning how cells might respond to homeobox gene expression to interpret these cues. Since changes in the cell-surface are central to many processes in early development we reasoned that cells expressing different complements of Hox genes might have different surface properties. In order to investigate this we have used the sensitive, non-disruptive technique of multiple two-phase aqueous partition, which is able to detect small differences on the surface of intact cells. Using this technique we have found that ectopic expression of the murine Hox-3.3 gene in cultured cells induces reproducible changes in the cell surface. Changes only occurred above a threshold level of gene expression, but above this level a correlation between surface change and gene expression was seen. The implications for the establishment of a 'Hox' code of homeobox genes acting to specifically change cell-surface properties are discussed.

Planar induction of anteroposterior pattern in the developing central nervous system of Xenopus laevis

It has long been thought that anteroposterior (A-P) pattern in the vertebrate central nervous system is induced in the embryo's dorsal ectoderm exclusively by signals passing vertically from underlying, patterned dorsal mesoderm. Explants from early gastrulae of the frog Xenopus laevis were prepared in which vertical contact between dorsal ectoderm and mesoderm was prevented but planar contact was maintained. In these, four position-specific neural markers (engrailed-2, Krox-20, XlHbox 1, and XlHbox 6) were expressed in the ectoderm in the same A-P order as in the embryo. Thus, planar signals alone, following a path available in the normal embryo, can induce A-P neural pattern.

Mechanism of skin morphogenesis. II. Retinoic acid modulates axis orientation and phenotypes of skin appendages

The factors that determine the axial orientation and phenotypes of skin appendages were analyzed by studying the effect of retinoic acid (RA) on embryonic chicken skin explant cultures. With RA uniformly distributed in the culture media, the feather buds became smaller, were disoriented or were transformed into scale-like structures in a concentration-dependent manner (from 0.05-2.5 microM). With RA distributed as a gradient created by a RA-soaked anion exchange bead, a radial zone of inhibition with a rim of disoriented buds was observed. The new axis of the disoriented buds appeared to be determined by a combination of the original feather axis determining force and a new axial force pointing centrifugally away from the RA source. This observed result can be simulated with a computer model using a vectorial sum of different feather axial determination forces. The size of the inhibited zone is linearly correlated to the RA concentration and may be used to quantify the morphogenetic activity of retinoids. These effects are specific to developmental stages (Hamburg and Hamilton stage 31–34). Both all-trans and 13-cis RA have morphogenetic activity. Retinol has no effect and retinal has a small inhibitory effect but neither phenotypic transformation nor axial disorientation were observed. The antero-posterior gradient of homeoprotein XlHbox 1 in feather buds became diffusive after RA treatment. RA dissolves dermal condensations and the distribution of N-CAM is altered from an anterior localized pattern to a diffusive presence in the bud cores. Endogenous retinoids in developing skins show developmental stage-dependent changes both quantitatively and qualitatively. The results suggest that RA either is or can modulate the endogenous morphogen(s) that determine the orientation and phenotype of skin appendages, and that this morphogenetic pathway involves Hox genes and adhesion molecules.

Retinoic acid induces changes in the localization of homeobox proteins in the antero-posterior axis of Xenopus laevis embryos

We have studied the localization of the proteins of Xeb1 and Xeb2, two homeobox (hbx)-containing genes that are expressed during the early development of Xenopus laevis. Both proteins are expressed in juxtaposed and partially overlapping domains along the antero-posterior axis of Xenopus laevis embryos, with clearly defined anterior boundaries. Xeb2 is predominantly expressed in the caudal region of the hindbrain, whereas the Xeb1 protein is located in the most rostral region of the spinal cord. Furthermore, both proteins are expressed in single cells dispersed in the lateral flanks of the embryo in positions that correlate with the expression domains in the neural tube. We suggest that these cells are migratory neural crest cells that have acquired positional information in the neural tube prior to migration. The Xeb2 protein was also detected in the most posterior branchial arches and the pronephros. In stage 45 embryos, nuclei of the IX-X cranial ganglia, the lung buds and cells spreading into the forelimb rudiment express the Xeb2 antigen. The Xeb1 protein was also detected in the lung buds and the forelimb rudiment. To examine the effect of retinoic acid on expression, gastrula embryos were treated with all-trans retinoic acid (RA). Increasing concentrations of RA caused progressive truncation of anterior structures. The most severely affected embryos lacked eyes, nasal pits, forebrain, midbrain and otic vesicles, and the anterior boundary of the hindbrain seemed to be displaced rostrally. This alteration correlates with a progressive displacement of the anterior boundary of the expression domain of Xeb2. On the other hand, 10(-6) M RA induces an ectopic site of Xeb1 expression at the anterior end of the central nervous system, located just anterior to the extended domain of Xeb2 whereas expression in the spinal cord remains unaffected.

Expression and modification of Hox 2.1 protein in mouse embryos

A polyclonal antibody, alpha Hox 2.1a, has been generated and used to immunolocalize Hox 2.1 protein in mouse embryos. Protein is present in nuclei of all tissues previously shown to express Hox 2.1 RNA. In addition, protein is seen in somites and proximal regions of the limb buds, tissues in which Hox 2.1 RNA expression was not clearly detected previously by in situ hybridization. At the 7 somite stage, protein is detectable in the neural tube up to the level of somite 1, but later retracts to a more posterior position. Immunoblot, in vitro translation, and immunoprecipitation experiments were carried out to characterize the Hox 2.1 protein. The results show that the Hox 2.1 gene produces at least two related phosphorylated proteins present in different proportions in different tissues.

Specific activation of mammalian Hox promoters in mosaic transgenic zebrafish

Homeo box-containing genes (Hox) are expressed in restricted regions of vertebrate embryos and may specify positional information. The organization and expression patterns of these genes are highly conserved among different species, suggesting that their regulation may also have been conserved. We developed a transient expression system, using mosaically transgenic zebrafish, which allows rapid analysis of transgene expression, and examined the activities of two mammalian Hox genes, mouse Hox-1.1 and human HOX-3.3. We found that these Hox promoters are activated in specific regions and tissues of developing zebrafish embryos and that this specificity depends upon the same regulatory elements within the promoters that specify the spatial expression of these genes in mice. Our results suggest that the promoter activities have been remarkably conserved from fish to mammals. To study the regulation of Hox expression in the developing nervous system, we analyzed the promoter activities in spt-1 mutants that have a mesodermal deficiency. Our results suggest that interactions, probably with the paraxial mesoderm, differentially regulate the activities of Hox promoters in the developing nervous system.