Human homoeobox-containing genes in development

Essay the (unusual) heuristic value of Hox gene clusters; a matter of time?

Ever since their first report in 1984, Antennapedia-type homeobox (Hox) genes have been involved in such a series of interesting observations, in particular due to their conserved clustered organization between vertebrates and arthropods, that one may legitimately wonder about the origin of this heuristic value. In this essay, I first consider different examples where Hox gene clusters have been instrumental in providing conceptual advances, taken from various fields of research and mostly involving vertebrate embryos. These examples touch upon our understanding of genomic evolution, the revisiting of 19th century views on the relationships between development and evolution and the building of a new framework to understand long-range and pleiotropic gene regulation during development. I then discuss whether the high value of the Hox gene family, when considered as an epistemic object, is related to its clustered structure (and the absence thereof in some animal species) and, if so, what is it in such particular genetic oddities that made them so generous in providing the scientific community with interesting information.

Genome-wide DNA methylation profiling using the methylation-dependent restriction enzyme LpnPI

DNA methylation is a well-known epigenetic modification that plays a crucial role in gene regulation, but genome-wide analysis of DNA methylation remains technically challenging and costly. DNA methylation-dependent restriction enzymes can be used to restrict CpG methylation analysis to methylated regions of the genome only, which significantly reduces the required sequencing depth and simplifies subsequent bioinformatics analysis. Unfortunately, this approach has been hampered by complete digestion of DNA in CpG methylation-dense regions, resulting in fragments that are too small for accurate mapping. Here, we show that the activity of DNA methylation-dependent enzyme, LpnPI, is blocked by a fragment size smaller than 32 bp. This unique property prevents complete digestion of methylation-dense DNA and allows accurate genome-wide analysis of CpG methylation at single-nucleotide resolution. Methylated DNA sequencing (MeD-seq) of LpnPI digested fragments revealed highly reproducible genome-wide CpG methylation profiles for >50% of all potentially methylated CpGs, at a sequencing depth less than one-tenth required for whole-genome bisulfite sequencing (WGBS). MeD-seq identified a high number of patient and tissue-specific differential methylated regions (DMRs) and revealed that patient-specific DMRs observed in both blood and buccal samples predict DNA methylation in other tissues and organs. We also observed highly variable DNA methylation at gene promoters on the inactive X Chromosome, indicating tissue-specific and interpatient-specific escape of X Chromosome inactivation. These findings highlight the potential of MeD-seq for high-throughput epigenetic profiling.

Increased methylation and decreased expression of homeobox genes TLX1, HOXA10 and DLX5 in human placenta are associated with trophoblast differentiation

Homeobox genes regulate embryonic and placental development, and are widely expressed in the human placenta, but their regulatory control by DNA methylation is unclear. DNA methylation analysis was performed on human placentae from first, second and third trimesters to determine methylation patterns of homeobox gene promoters across gestation. Most homeobox genes were hypo-methylated throughout gestation, suggesting that DNA methylation is not the primary mechanism involved in regulating HOX genes expression in the placenta. Nevertheless, several genes showed variable methylation patterns across gestation, with a general trend towards an increase in methylation over gestation. Three genes (TLX1, HOXA10 and DLX5) showed inverse gains of methylation with decreasing mRNA expression throughout pregnancy, supporting a role for DNA methylation in their regulation. Proteins encoded by these genes were primarily localised to the syncytiotrophoblast layer, and showed decreased expression later in gestation. siRNA mediated downregulation of DLX5, TLX1 and HOXA10 in primary term villous cytotrophoblast resulted in decreased proliferation and increased expression of differentiation markers, including ERVW-1. Our data suggest that loss of DLX5, TLX1 and HOXA10 expression in late gestation is required for proper placental differentiation and function.

Structure and sequence of the human homeobox gene HOX7

A cosmid containing the human sequence HOX7, homologous to the murine Hox-7 gene, was isolated from a genomic library, and the positions of the coding sequences were determined by hybridization. DNA sequence analysis demonstrated two exons that code for a homeodomain-containing protein of 297 amino acids. The open reading frame is interrupted by a single intron of approximately 1.6 kb, the splice donor and acceptor sites of which conform to known consensus sequences. The human HOX7 coding sequence has a very high degree of identity with the murine Hox-7 cDNA. Within the homeobox, the two sequences share 94% identity at the DNA level, all substitutions being silent. This high level of sequence similarity is not confined to the homeodomain; overall the human and murine HOX7 gene products show 80% identity at the amino acid level. Both the 5' and 3' untranslated regions also show significant similarity to the murine gene, with 79 and 70% sequence identity, respectively. The sequence upstream of the coding sequence of exon 1 contains a GC-rich putative promoter region. There is no TATA box, but a CCAAT and numerous GC boxes are present. The region encompassing the promoter region, exon 1, and the 5' region of exon 2 have a higher than expected frequency of CpG dinucleotides; numerous sites for rare-cutter restriction enzymes are present, a characteristic of HTF islands.

The human HOX gene family

We report the identification of 10 new human homeobox sequences. Altogether, we have isolated and sequenced 30 human homeoboxes clustered in 4 chromosomal regions called HOX loci. HOX1 includes 8 homeoboxes in 90 kb of DNA on chromosome 7. HOX2 includes 9 homeoboxes in 180 kb on chromosome 17. HOX3 contains at least 7 homeoboxes in 160 kb on chromosome 12. Finally, HOX4 includes 6 homeoboxes in 70 kb on chromosome 2. Homeodomains obtained from the conceptual translation of the isolated homeoboxes can be attributed to 13 homology groups on the basis of their primary peptide sequence. Moreover, it is possible to align the 4 HOX loci so that corresponding homeodomains in all loci share the maximal sequence identity. The complex of these observations supports and extends an evolutionary hypothesis concerning the origin of mammalian and fly homeobox gene complexes. We also determined the coding region present in 3 HOX2 cDNA clones corresponding to HOX2G, HOX2H and HOX2I.

Interference with function of a homeobox gene in Xenopus embryos produces malformations of the anterior spinal cord

XIHbox 1 is expressed in a narrow band across the cervical region of Xenopus embryos. The gene produces two related proteins: "long" and "short" XIHbox 1 homeodomain proteins. Injection of antibodies to the long XIHbox 1 protein into 1-cell embryos caused a phenotype in which the anterior spinal cord was morphologically transformed into a hindbrain-like structure. This alteration was restricted to the region normally expressing long XIHbox 1 protein. Injection of long protein mRNA disrupted segmentation and tissue organization without inhibiting cell proliferation. Injection of short protein mRNA into 1-cell embryos produced spinal cord malformations similar, but not identical, to those caused by the antibodies, suggesting antagonistic roles for long and short XIHbox 1 proteins. We immunostained tadpoles carrying extended hindbrains for N-CAM and consistently found defective organization of spinal nerves over the affected region.

Remarkable intron and exon sequence conservation in human and mouse homeobox Hox 1.3 genes

A high degree of conservation exists between the Hox 1.3 homeobox genes of mice and humans. The two genes occupy the same relative positions in their respective Hox 1 gene clusters, they show extensive sequence similarities in their coding and noncoding portions, and both are transcribed into multiple transcripts of similar sizes. The predicted human Hox 1.3 protein differs from its murine counterpart in only 7 of 270 amino acids. The sequence similarity in the 250 base pairs upstream of the initiation codon is 98%, the similarity between the two introns, both 960 base pairs long, is 72%, and the similarity in the 3' noncoding region from termination codon to polyadenylation signal is 90%. Both mouse and human Hox 1.3 introns contain a sequence with homology to a mating-type-controlled cis element of the yeast Ty1 transposon. DNA-binding studies with a recombinant mouse Hox 1.3 protein identified two binding sites in the intron, both of which were within the region of shared homology with this Ty1 cis element.

Remarkable intron and exon sequence conservation in human and mouse homeobox Hox 1.3 genes

A high degree of conservation exists between the Hox 1.3 homeobox genes of mice and humans. The two genes occupy the same relative positions in their respective Hox 1 gene clusters, they show extensive sequence similarities in their coding and noncoding portions, and both are transcribed into multiple transcripts of similar sizes. The predicted human Hox 1.3 protein differs from its murine counterpart in only 7 of 270 amino acids. The sequence similarity in the 250 base pairs upstream of the initiation codon is 98%, the similarity between the two introns, both 960 base pairs long, is 72%, and the similarity in the 3' noncoding region from termination codon to polyadenylation signal is 90%. Both mouse and human Hox 1.3 introns contain a sequence with homology to a mating-type-controlled cis element of the yeast Ty1 transposon. DNA-binding studies with a recombinant mouse Hox 1.3 protein identified two binding sites in the intron, both of which were within the region of shared homology with this Ty1 cis element.

Expression of Hox-2.4 homeobox gene directed by proviral insertion in a myeloid leukemia

The presence of an altered Hox-2.4 gene in the WEHI3B murine myeloid leukemia suggests that homeobox genes may contribute to neoplasia. A survey of 31 leukemia cell lines of the myeloid, lymphoid and erythroid lineages revealed that Hox-2.4 was expressed only in WEHI3B and the pre-B lymphoid line 70Z/3, in which no DNA rearrangement was observed. To clarify the WEHI3B alteration and normal Hox-2.4 structure, we have sequenced near full length cDNA clones from WEHI3B and 70Z/3, and the 5' portion of the normal Hox-2.4 gene. A WEHI3B cDNA clone demonstrates that an intracisternal A-particle (IAP) provirus has inserted within the first exon of the gene and generated a Hox-2.4 mRNA with a 5' sequence derived from the IAP long terminal repeat. A remarkable degree of similarity found between the amino acid sequences of Hox-2.4 and Hox-3.1, which reside on different chromosomes, supports the notion that an ancient homeobox gene cluster has been duplicated and dispersed early in vertebrate evolution.

Vertebrate homeodomain proteins: families of region-specific transcription factors

Vertebrate homeodomain proteins are transcription factors whose genes can be isolated via a conserved DNA-binding domain called the homeobox. We review recent studies suggesting that one function of these genes is the early subdivision of the embryo along the antero-posterior axis into 'fields' of cells with different developmental potential.

Duplicated homeobox genes in Xenopus

Multiple kinds of clones and restriction fragment polymorphisms are frequently encountered when analyzing genes of the tetraploid frog Xenopus laevis. Two types of cDNA clone have been isolated for homeobox gene 2. Analysis of their corresponding genomic clones confirmed the existence of clearly distinct restriction maps; in addition the nearby presence of two additional homeoboxes suggests that this region is homologous to the Hox 2 gene complex of mammals. We asked whether the genetic polymorphism in Xenopus results from increased allelic differences due to tetraploidy or from having duplicated Hox 2 complexes. Using X. laevis/Xenopus borealis interspecies hybrids we show that the two types of X. laevis homeobox gene 2 transcripts result from two different genetic loci. They cannot represent alleles of the same gene because they do not segregate independently in the F1 hybrid progeny. Most other X. laevis homeobox genes studied so far are also found in two versions. Thus X. laevis seems to have two homeobox genes, both of which are expressed, for each one present in mammals or other vertebrates.

At least three human homeoboxes on chromosome 12 belong to the same transcription unit

Mammalian homeoboxes show a clustered chromosomal organization. In the mouse, at least seven homeoboxes on chromosome 6 and at least six on chromosome 11 identify the murine Hox-1 and Hox-2 loci, respectively. A number of homeoboxes on chromosome 7 define the human HOX-1 locus and homeoboxes on chromosome 17 define the human HOX-2 locus. We studied the genomic organization of three homeobox sequences of the HOX-3 locus on chromosome 12 and analyzed transcripts from this region. Structural characterization and sequencing of several cDNA clones reveal that the three homeobox sequences present in this chromosomal region identify a single transcription unit. Primary transcripts are alternatively processed to give mature messengers with a common 5' noncoding exon encoding different proteins containing one of the three homeodomains.