The Hox genes are probably the most widely recognized family of genes that regulate animal developmental processes.
Wendy Bickmore, Jeremy S. Dasen, Jacqueline Deschamps, Denis Duboule (Organiser), Walter Gehring (1939-2014) (Organiser), Yacine Graba, Bradley Hersh, François Karch, Tomonori Katsuyama, Marie Kmita (Organiser), Robb Krumlauf, Shoichiro Kurata, Ulrike Löhr, Moisés Mallo-Perez, Richard Mann, Olivier Pourquié, Filippo Rijli, Ernesto Sanchez-Herrero
by Bradley Hersh
8 – 13 October 2007
The Hox genes are probably the most widely recognized family of genes that regulate animal developmental processes. Mutations in the Hox genes can result in spectacular transformations of one structure into another, such as conversion of antennae to legs in the fruit fly, Drosophila, or transformation of cervical to thoracic vertebrae in mice. At evolutionary time scales, changes in the function or expression of the Hox genes are correlated with changing identities of segmental structures along the animal anterior-posterior axis. This workshop explored research on how Hox gene clusters are organized within animal genomes, how their complex expression patterns are generated, how they select target genes and generate diverse transcriptional outcomes, and how they ultimately regulate the development of animal morphology.
Clustering of Hox genes
One of the most striking features of Hox genes is their clustered arrangement and the remarkable colinearity of expression pattern along the anterior-posterior axis and position within a cluster. Drosophila possesses 8 Hox genes, each of which is expressed in a distinct domain along the head-tail body axis that corresponds to gene order on the chromosome. Vertebrates have a greater number of Hox genes, having expanded the posterior group gene complement and duplicated entire clusters. In addition, vertebrate clusters exhibit temporal colinearity—the Hox genes are activated in a progressive sequence, with anterior genes expressed earlier than posterior genes. Denis Duboule (University of Geneva) and Walter Gehring (University of Basel) presented contrasting models for the evolution of the clustered arrangements. Duboule suggested that the compact clusters of vertebrates are derived from ancestral clusters that were less compact and less well organized. Because such a consolidation seems counter-intuitive, Duboule proposed that the recruitment of global control regions outside of the cluster would selectively favor compaction. Walter Gehring argued, instead, that the original clusters must have been organized, based on observed conservation of colinearity, and that unequal crossing over expanded the initial simple cluster. He also suggested the developmental ground state of a tissue corresponds to the fate of the tissue that expresses the original member of the cluster, or the “UrHox” gene.
Regulation of Hox gene expression
The complex expression patterns of Hox genes are regulated by a variety of methods, ranging from global alterations in chromatin accessibility to specific, local cis-regulatory elements (CREs). Francois Karch (University of Geneva) described mechanisms acting within the iab regulatory domains that regulate parasegment-specific posterior Hox gene expression in Drosophila. He suggested that boundary elements in the Bithorax Complex act to tether chromatin loops, each harboring an iab domain, to the Abdominal-B promoter. As the boundary elements release in sequential parasegments, initiator sequences in the untethered iab domains are freed to activate expression. Wendy Bickmore (MRC Human Genetics Unit) presented beautiful fluorescence in situ hybridization experiments that demonstrated in vivo the looping out of Hox clusters from compact chromosome territories, and chromatin decondensation allowing resolution of loci within a cluster. Tomonori Katsuyama (ETH Zurich) demonstrated physical interactions of Polycomb group (PcG) components with the regulatory elements of Abd-B in the inactive state, and replacement of the PcG factors with Trithorax protein and Pleiohomeotic protein in the active state. Shoichiro Kurata (Tohoku University) demonstrated that Winged Eye(WGE), a chromatin-associated protein, binds to all the sites on polytene chromosomes that are bound by Posterior Sex Combs of the PcG. Additional genetic experiments suggested WGE may act both in complexes that maintain repressed and active states of chromatin.
Because chromatin conformation alone cannot fully account for spatial and temporal colinearity of Hox gene expression, there must be additional cis-acting controls to generate the observed patterns. Duboule described a global control region (GCR) lying 200kb outside the HoxD cluster that facilitates protein-promoter interactions and affects positional and quantitative colinearity in the developing vertebrate limb. Robb Krumlauf (Stowers Institute for Medical Research) described the modular CREs controlling expression of Hox genes in specific rhombomeres within the vertebrate hindbrain, and how cross-talk between Hox genes establishes and then maintains expression.
Exploring Hox target specificity
Understanding how Hox expression is translated into functional effects on target genes presents challenges in both specificity and diversity: since Hox proteins appear to have very similar DNA binding sequence specificities, how do these proteins act on distinct target gene sets; second, how does a single Hox protein modulate different functions during development? One mechanism to generate a diversity of responses is to produce an assortment of protein isoforms from a single Hox gene. Ernesto Sanchez-Herrero (Universidad Autónoma de Madrid) tested the ability of different Ultrabithorax (UBX) isoforms to rescue defects in a Ubx mutant background. In addition, he showed how the posterior group gene Abdominal-B M and R isoforms perform different roles in the regulation of target genes abdominal-A and Distalless in the genital disc. The different effects of these isoforms may involve modification of protein-protein interactions rather than modulation of DNA binding, based on a model for Hox functional diversity described by Yacine Graba (IBDML, Marseille). Graba found that the short UbdA motif in UBX can mediate an interaction with the well-characterized Hox co-factor, Extradenticle (EXD), and suggested that the mode of Hox/EXD interaction depends on the particular target gene, thereby providing a potential structural basis for distinct activities. The EXD co-factor can enhance the DNA binding specificity of Hox proteins, and Richard Mann (Columbia University) described the crystal structure of a SCR/EXD/DNA complex that suggested an important role for DNA electrostatic potential in the recognition of specific DNA targets by Hox proteins. Bradley Hersh (Clemson University) presented analyses of a UBX-responsive CREs to identify additional DNA sequences beyond the UBX core binding site (TAAT) important for regulation by UBX. Jeremy Dasen (New York University) described the generation of vertebrate spinal motor neuron diversity through the combinatorial action of Hox proteins and context-specific transcriptional co-factors, such as FoxP1.
Hox regulation of morphogenesis
Given their dramatic effects on animal shape, Hox genes must ultimately regulate processes of morphogenesis, whether directly or indirectly. Ulrike Löhr (Max-Planck Institute for Biophysical Chemistry) shared experiments that manipulated the gradient of bcd, a homeobox gene within the Hox cluster but without Hox function, to determine the effect on positional information in the early Drosophila embryo. Olivier Pourquié (Stowers Institute for Medical Research) described both the role of Hox genes as regulators of cell ingression during vertebrate gastrulation and the signaling mechanisms that establish final Hox expression boundaries in the somitic mesoderm. Jacqueline Deschamps (Hubrecht Institute) demonstrated that knockout combinations of Cdx genes, members of the ParaHox cluster, resulted in posterior truncations of the axial skeleton and loss of expression for multiple Hox genes. Moisés Mallo (Instituto Gulbenkian de Ciência) manipulated Hox expression in the pre-somitic mesoderm and found that Hoxa10 misexpression could eliminate ribs, whereas Hoxb6 misexpression could trigger rib formation throughout the axial skeleton. Hox genes also modulate FGF and Sonic Hedgehog signaling during morphogenesis of the vertebrate limb. Marie Kmita (Institut de Recherches Cliniques de Montréal) showed that Shh expression is lost in the limb bud of animals with HoxA and HoxD clusters conditionally deleted, and that the reduced limb buds in these animals is due to an increase in apoptotic cell death. Filippo Rijli (IGBMC, Strasbourg) extended the morphogenetic role of Hox genes in the brain in his presentation by demonstrating tthat Hoxa2 is important for proper migration of sensory neurons from the periphery to their targets in the vertebrate hindbrain.
The broad and varied presentations at this meeting demonstrated that Hox genes control fundamental cellular processes, including proliferation, apoptosis, migration, and cell specification. These detailed accounts, generated through sophisticated use of novel cellular, molecular, and computational techniques, expand our mechanistic understanding of Hox gene action. Though the variety of ways in which Hox genes are themselves regulated and by which they, in turn, regulate morphogenetic processes suggests many layers of complexity yet to be unraveled, the talks presented at Les Treilles demonstrate that we are making exciting progress.
A modified version of this summary was published as “Hox en Provence” Developmental Cell (2007), doi:10.1016/j.devcel.2007.11.013