Expression of Hoxa2 in rhombomere 4 is regulated by a conserved cross-regulatory mechanism dependent upon Hoxb1

The Hoxa2 gene is an important component of regulatory events during hindbrain segmentation and head development in vertebrates. In this study we have used sequenced comparisons of the Hoxa2 locus from 12 vertebrate species in combination with detailed regulatory analyses in mouse and chicken embryos to characterize the mechanistic basis for the regulation of Hoxa2 in rhombomere (r) 4. A highly conserved region in the Hoxa2 intron functions as an r4 enhancer. In vitro binding studies demonstrate that within the conserved region three bipartite Hox/Pbx binding sites (PH1-PH3) in combination with a single binding site for Pbx-Prep/Meis (PM) heterodimers co-operate to regulate enhancer activity in r4. Mutational analysis reveals that these sites are required for activity of the enhancer, suggesting that the r4 enhancer from Hoxa2 functions in vivo as a Hox-response module in combination with the Hox cofactors, Pbx and Prep/Meis. Furthermore, this r4 enhancer is capable of mediating a response to ectopic HOXB1 expression in the hindbrain. These findings reveal that Hoxa2 is a target gene of Hoxb1 and permit us to develop a gene regulatory network for r4, whereby Hoxa2, along with Hoxb1, Hoxb2 and Hoxa1, is integrated into a series of auto- and cross-regulatory loops between Hox genes. These data highlight the important role played by direct cross-talk between Hox genes in regulating hindbrain patterning.     

Independent regulation of initiation and maintenance phases of Hoxa3 expression in the vertebrate hindbrain involve auto- and cross-regulatory mechanisms

During development of the vertebrate hindbrain, Hox genes play multiple roles in the segmental processes that regulate anteroposterior (AP) patterning. Paralogous Hox genes, such as Hoxa3, Hoxb3 and Hoxd3, generally have very similar patterns of expression, and gene targeting experiments have shown that members of paralogy group 3 can functionally compensate for each other. Hence, distinct functions for individual members of this family may primarily depend upon differences in their expression domains. The earliest domains of expression of the Hoxa3 and Hoxb3 genes in hindbrain rhombomeric (r) segments are transiently regulated by kreisler, a conserved Maf b-Zip protein, but the mechanisms that maintain expression in later stages are unknown. In this study, we have compared the segmental expression and regulation of Hoxa3 and Hoxb3 in mouse and chick embryos to investigate how they are controlled after initial activation. We found that the patterns of Hoxa3 and Hoxb3 expression in r5 and r6 in later stages during mouse and chick hindbrain development were differentially regulated. Hoxa3 expression was maintained in r5 and r6, while Hoxb3 was downregulated. Regulatory comparisons of cis-elements from the chick and mouse Hoxa3 locus in both transgenic mouse and chick embryos have identified a conserved enhancer that mediates the late phase of Hoxa3 expression through a conserved auto/cross-regulatory loop. This block of similarity is also present in the human and horn shark loci, and contains two bipartite Hox/Pbx-binding sites that are necessary for its in vivo activity in the hindbrain. These HOX/PBC sites are positioned near a conserved kreisler-binding site (KrA) that is involved in activating early expression in r5 and r6, but their activity is independent of kreisler. This work demonstrates that separate elements are involved in initiating and maintaining Hoxa3 expression during hindbrain segmentation, and that it is regulated in a manner different from Hoxb3 in later stages. Together, these findings add further strength to the emerging importance of positive auto- and cross-regulatory interactions between Hox genes as a general mechanism for maintaining their correct spatial patterns in the vertebrate nervous system.     

Zic2 is required for neural crest formation and hindbrain patterning during mouse development

The Zic genes are the vertebrate homologues of the Drosophila pair rule gene odd-paired. It has been proposed that Zic genes play several roles during neural development including mediolateral segmentation of the neural plate, neural crest induction, and inhibition of neurogenesis. Initially during mouse neural development Zic2 is expressed throughout the neural plate while later on expression in the neurectoderm becomes restricted to the lateral region of the neural plate. A hypomorphic allele of Zic2 has demonstrated that in the mouse Zic2 is required for the timing of neurulation. We have isolated a new allele of Zic2 that behaves as a loss of function allele. Analysis of this mutant reveals two further functions for Zic2 during early neural development. Mutation of Zic2 results in a delay of neural crest production and a decrease in the number of neural crest cells that are produced. These defects are independent of mediolateral segmentation of the neurectoderm and of dorsal neurectoderm proliferation, both of which occur normally in the mutant embryos. Additionally Zic2 is required during hindbrain patterning for the normal development of rhombomeres 3 and 5. This work provides the first genetic evidence that the Zic genes are involved in neural crest production and the first demonstration that Zic2 functions during hindbrain patterning.     

Roles of Hoxa1 and Hoxa2 in patterning the early hindbrain of the mouse

Early in its development, the vertebrate hindbrain is transiently subdivided into a series of compartments called rhombomeres. Genes have been identified whose expression patterns distinguish these cellular compartments. Two of these genes, Hoxa1 and Hoxa2, have been shown to be required for proper patterning of the early mouse hindbrain and the associated neural crest. To determine the extent to which these two genes function together to pattern the hindbrain, we generated mice simultaneously mutant at both loci. The hindbrain patterning defects were analyzed in embryos individually mutant for Hoxa1 and Hoxa2 in greater detail and extended to embryos mutant for both genes. From these data a model is proposed to describe how Hoxa1, Hoxa2, Hoxb1, Krox20 (Egr2) and kreisler function together to pattern the early mouse hindbrain. Critical to the model is the demonstration that Hoxa1 activity is required to set the anterior limit of Hoxb1 expression at the presumptive r3/4 rhombomere boundary. Failure to express Hoxb1 to this boundary in Hoxa1 mutant embryos initiates a cascade of gene misexpressions that result in misspecification of the hindbrain compartments from r2 through r5. Subsequent to misspecification of the hindbrain compartments, ectopic induction of apoptosis appears to be used to regulate the aberrant size of the misspecified rhombomeres.     

Cis-regulatory characterization of sequence conservation surrounding the Hox4 genes

Hox genes are key regulators of anterior-posterior axis patterning and have a major role in hindbrain development. The zebrafish Hox4 paralogs have strong overlapping activities in hindbrain rhombomeres 7 and 8, in the spinal cord and in the pharyngeal arches. With the aim to predict enhancers that act on the hoxa4a, hoxb4a, hoxc4a and hoxd4a genes, we used sequence conservation around the Hox4 genes to analyze all fish:human conserved non-coding sequences by reporter assays in stable zebrafish transgenesis. Thirty-four elements were functionally tested in GFP reporter gene constructs and more than 100 F1 lines were analyzed to establish a correlation between sequence conservation and cis-regulatory function, constituting a catalog of Hox4 CNEs. Sixteen tissue-specific enhancers could be identified. Multiple alignments of the CNEs revealed paralogous cis-regulatory sequences, however, the CNE sequence similarities were found not to correlate with tissue specificity. To identify ancestral enhancers that direct Hox4 gene activity, genome sequence alignments of mammals, teleosts, horn shark and the cephalochordate amphioxus, which is the most basal extant chordate possessing a single prototypical Hox cluster, were performed. Three elements were identified and two of them exhibited regulatory activity in transgenic zebrafish, however revealing no specificity. Our data show that the approach to identify cis-regulatory sequences by genome sequence alignments and subsequent testing in zebrafish transgenesis can be used to define enhancers within the Hox clusters and that these have significantly diverged in their function during evolution.     

Synergy between Hoxa1 and Hoxb1: the relationship between arch patterning and the generation of cranial neural crest

Hoxa1 and Hoxb1 have overlapping synergistic roles in patterning the hindbrain and cranial neural crest cells. The combination of an ectoderm-specific regulatory mutation in the Hoxb1 locus and the Hoxa1 mutant genetic background results in an ectoderm-specific double mutation, leaving the other germ layers impaired only in Hoxa1 function. This has allowed us to examine neural crest and arch patterning defects that originate exclusively from the neuroepithelium as a result of the simultaneous loss of Hoxa1 and Hoxb1 in this tissue. Using molecular and lineage analysis in this double mutant background we demonstrate that presumptive rhombomere 4, the major site of origin of the second pharyngeal arch neural crest, is reduced in size and has lost the ability to generate neural crest cells. Grafting experiments using wild-type cells in cultured normal or double mutant mouse embryos demonstrate that this is a cell-autonomous defect, suggesting that the formation or generation of cranial neural crest has been uncoupled from segmental identity in these mutants. Furthermore, we show that loss of the second arch neural crest population does not have any adverse consequences on early patterning of the second arch. Signalling molecules are expressed correctly and pharyngeal pouch and epibranchial placode formation are unaffected. There are no signs of excessive cell death or loss of proliferation in the epithelium of the second arch, suggesting that the neural crest cells are not the source of any indispensable mitogenic or survival signals. These results illustrate that Hox genes are not only necessary for proper axial specification of the neural crest but that they also play a vital role in the generation of this population itself. Furthermore, they demonstrate that early patterning of the separate components of the pharyngeal arches can proceed independently of neural crest cell migration.     

Ectopic Hoxa2 induction after neural crest migration results in homeosis of jaw elements in Xenopus

Hox genes are required to pattern neural crest (NC) derived craniofacial and visceral skeletal structures. However, the temporal requirement of Hox patterning activity is not known. Here, we use an inducible system to establish Hoxa2 activity at distinct NC migratory stages in Xenopus embryos. We uncover stage-specific effects of Hoxa2 gain-of-function suggesting a multistep patterning process for hindbrain NC. Most interestingly, we show that Hoxa2 induction at postmigratory stages results in mirror image homeotic transformation of a subset of jaw elements, normally devoid of Hox expression, towards hyoid morphology. This is the reverse phenotype to that observed in the Hoxa2 knockout. These data demonstrate that the skeletal pattern of rhombomeric mandibular crest is not committed before migration and further implicate Hoxa2 as a true selector of hyoid fate. Moreover, the demonstration that the expression of Hoxa2 alone is sufficient to transform the upper jaw and its joint selectively may have implications for the evolution of jaws.     

Altered neuronal lineages in the facial ganglia of Hoxa2 mutant mice

Neurons of cranial sensory ganglia are derived from the neural crest and ectodermal placodes, but the mechanisms that control the relative contributions of each are not understood. Crest cells of the second branchial arch generate few facial ganglion neurons and no vestibuloacoustic ganglion neurons, but crest cells in other branchial arches generate many sensory neurons. Here we report that the facial ganglia of Hoxa2 mutant mice contain a large population of crest-derived neurons, suggesting that Hoxa2 normally represses the neurogenic potential of second arch crest cells. This may represent an anterior transformation of second arch neural crest cells toward a fate resembling that of first arch neural crest cells, which normally do not express Hoxa2 or any other Hox gene. We additionally found that overexpressing Hoxa2 in cultures of P19 embryonal carcinoma cells reduced the frequency of spontaneous neuronal differentiation, but only in the presence of cotransfected Pbx and Meis Hox cofactors. Finally, expression of Hoxa2 and the cofactors in chick neural crest cells populating the trigeminal ganglion also reduced the frequency of neurogenesis in the intact embryo. These data suggest an unanticipated role for Hox genes in controlling the neurogenic potential of at least some cranial neural crest cells.     

Analysis of Fibroblast growth factor 15 cis-elements reveals two conserved enhancers which are closely related to cardiac outflow tract development

Fibroblast growth factor 15 (Fgf15) is expressed in the developing mouse central nervous system and pharyngeal arches. Fgf15 mutant mice showed defects of the cardiac outflow tract probably because of aberrant behavior of the cardiac neural crest cells. In this study, we examined cis-elements of the Fgf15 gene by transient transgenic analysis using lacZ as a reporter. We identified two enhancers: one directed lacZ expression in the hindbrain/spinal cord and the other in the posterior midbrain (pmb), rhombomere1 (r1) and pharyngeal epithelia. Interestingly, human genomic regions which are highly homologous to these two mouse enhancers showed almost the same enhancer activities as those of mice in transgenic mouse embryos, indicating that the two enhancers are conserved between humans and mice. We also showed that the mouse and human pmb/r1 enhancer can regulate lacZ expression in chick embryos in almost the same way as in mouse embryos. We found that the lacZ expression domain with this enhancer was expanded by ectopic Fgf8b expression, suggesting that this enhancer is regulated by Fgf8 signaling. Moreover, over-expression of Fgf15 resulted in up-regulation of Fgf8 expression in the isthmus/r1. These findings suggest that a reciprocal positive regulation exists between Fgf15 and Fgf8 in the isthmus/r1. Together with cardiac outflow tract defects in Fgf15 mutants, the conservation of enhancers in the hindbrain/spinal cord and pharyngeal epithelia suggests that human FGF19 (ortholog of Fgf15) is involved in early development and the distribution of cardiac neural crest cells and is one of the candidate genes for congenital heart defects.     

Effects of gene regulatory reprogramming on gene expression in human and mouse developing hearts

Lineage-specific regulatory elements underlie adaptation of species and play a role in disease susceptibility. We compared functionally conserved and lineage-specific enhancers by cross-mapping 5042 human and 6564 mouse heart enhancers. Of these, 79 per cent are lineage-specific, lacking a functional orthologue. Heart enhancers tend to cluster and, commonly, there are multiple heart enhancers in a heart locus providing a regulatory stability to the locus. We observed little cross-clustering, however, between lineage-specific and functionally conserved heart enhancers suggesting regulatory function acquisition and development in loci previously lacking heart activity. We also identified 862 human-specific heart enhancers: 417 featuring sequence conservation with mouse (class II) and 445 with neither sequence nor function conservation (class III). Ninety-eight per cent of class III enhancers were deleted from the mouse genome, and we estimated a similar-sized enhancer gain in the human lineage. Human-specific enhancers display no detectable decrease in the negative selection pressure and are strongly associated with genes partaking in the heart regulatory programmes. The loss of a heart enhancer could be compensated by activity of a redundant heart enhancer; however, we observed redundancy in only 15 per cent of class II and III enhancer loci indicating a large-scale reprogramming of the heart regulatory programme in mammals.     

Auto/cross-regulation of Hoxb3 expression in posterior hindbrain and spinal cord

The complex and dynamic pattern of Hoxb3 expression in the developing hindbrain and the associated neural crest of mouse embryos is controlled by three separate cis-regulatory elements: element I (region A), element IIIa, and the r5 enhancer (element IVa). We have examined the cis-regulatory element IIIa by transgenic and mutational analysis to determine the upstream trans-acting factors and mechanisms that are involved in controlling the expression of the mouse Hoxb3 gene in the anterior spinal cord and hindbrain up to the r5/r6 boundary, as well as the associated neural crest which migrate to the third and posterior branchial arches and to the gut. By deletion analysis, we have identified the sequence requirements within a 482-bp element III482. Two Hox binding sites are identified in element III482 and we have shown that in vitro both Hoxb3 and Hoxb4 proteins can interact with these Hox binding sites, suggesting that auto/cross-regulation is required for establishing the expression of Hoxb3 in the neural tube domain. Interestingly, we have identified a novel GCCAGGC sequence motif within element III482, which is also required to direct gene expression to a subset of the expression domains except for rhombomere 6 and the associated neural crest migrating to the third and posterior branchial arches. Element III482 can direct a higher level of reporter gene expression in r6, which led us to investigate whether kreisler is involved in regulating Hoxb3 expression in r6 through this element. However, our transgenic and mutational analysis has demonstrated that, although kreisler binding sites are present, they are not required for the establishment or maintenance of reporter gene expression in r6. Our results have provided evidence that the expression of Hoxb3 in the neural tube up to the r5/r6 boundary is auto/cross-regulated by Hox genes and expression of Hoxb3 in r6 does not require kreisler.     

Differences in Krox20-dependent regulation of Hoxa2 and Hoxb2 during hindbrain development

During hindbrain development, segmental regulation of the paralogous Hoxa2 and Hoxb2 genes in rhombomeres (r) 3 and 5 involves Krox20-dependent enhancers that have been conserved during the duplication of the vertebrate Hox clusters from a common ancestor. Examining these evolutionarily related control regions could provide important insight into the degree to which the basic Krox20-dependent mechanisms, cis-regulatory components, and their organization have been conserved. Toward this goal we have performed a detailed functional analysis of a mouse Hoxa2 enhancer capable of directing reporter expression in r3 and r5. The combined activities of five separate cis-regions, in addition to the conserved Krox20 binding sites, are involved in mediating enhancer function. A CTTT (BoxA) motif adjacent to the Krox20 binding sites is important for r3/r5 activity. The BoxA motif is similar to one (Box1) found in the Hoxb2 enhancer and indicates that the close proximity of these Box motifs to Krox20 sites is a common feature of Krox20 targets in vivo. Two other rhombomeric elements (RE1 and RE3) are essential for r3/r5 activity and share common TCT motifs, indicating that they interact with a similar cofactor(s). TCT motifs are also found in the Hoxb2 enhancer, suggesting that they may be another common feature of Krox20-dependent control regions. The two remaining Hoxa2 cis-elements, RE2 and RE4, are not conserved in the Hoxb2 enhancer and define differences in some of components that can contribute to the Krox20-dependent activities of these enhancers. Furthermore, analysis of regulatory activities of these enhancers in a Krox20 mutant background has uncovered differences in their degree of dependence upon Krox20 for segmental expression. Together, this work has revealed a surprising degree of complexity in the number of cis-elements and regulatory components that contribute to segmental expression mediated by Krox20 and sheds light on the diversity and evolution of Krox20 target sites and Hox regulatory elements in vertebrates.     

Homeotic transformation of branchial arch identity after Hoxa2 overexpression

Overexpression of Hoxa2 in the chick first branchial arch leads to a transformation of first arch cartilages, such as Meckel's and the quadrate, into second arch elements, such as the tongue skeleton. These duplicated elements are fused to the original in a similar manner to that seen in the Hoxa2 knockout, where the reverse transformation of second to first arch morphology is observed. This confirms the role of Hoxa2 as a selector gene specifying second arch fate. When first arch neural crest alone is targeted, first arch elements are lost, but the Hoxa2-expressing crest is unable to develop into second arch elements. This is not due to Hoxa2 preventing differentiation of cartilages. Upregulation of a second arch marker in the first arch, and homeotic transformation of cartilage elements is only produced after global Hoxa2 overexpression in the crest and the surrounding tissue. Thus, although the neural crest appears to contain some patterning information, it needs to read cues from the environment to form a coordinated pattern. Hoxa2 appears to exert its effect during differentiation of the cartilage elements in the branchial arches, rather than during crest migration, implying that pattern is determined quite late in development.