Hox genes encode proteins that contain the DNA-binding homeobox motif and control early patterning of segments in the embryo as well as later events in development (reviewed in Rezsohazy et al. 2015). Mammals have 39 Hox genes arrayed in 4 linear clusters, with each cluster containing 9 to 11 genes. Based on homologies, the genes have been assigned to 13 paralogous groups. The nomenclature of Hox genes uses a letter to indicate the cluster and a number to indicate the paralog group. For example, HOXA4 is the gene in cluster A that is most similar with genes of paralog group 4 from other clusters.
One of the most striking aspects of mammalian Hox gene function is the mechanism of their activation during embryogenesis: the order of genes in a cluster correlates with the timing and location of their activation such that genes at the 3' end of a cluster are activated first and genes at the 5' end of a cluster are activated last. (5' and 3' refer to the transcriptional orientation of the genes in the cluster.) Because development of segments of the embryo proceeds from anterior to posterior this means that the anterior boundaries of expression of 3' genes are more anterior (rostral) and the anterior boundaries of expression of 5' genes are more posterior (caudal). Expression of HOX genes initiates in the posterior primitive streak at the beginning of gastrulation at approximately E7.5 in mouse. As gastrulation proceeds, further 5' genes are sequentially activated and they too undergo the same chromatin changes and migration. After formation of the axis of the embryo, similar waves of activation of HOXA and HOXD clusters occur in developing limbs beginning at about E9. Retinoids, especially all trans retinoic acid (atRA), participate in initiating the process via retinoid receptors. Other factors such as FGFs and Wnt, also regulate Hox expression. After activation, Hox genes participate in maintaining their own expression (autoregulation), activating later, 5' Hox genes, and repressing prior, 3' Hox genes (crossregulation). Differentiation of embryonal carcinoma cells and embryonic stem cells in response to retinoic acid is used to model the process in vitro (reviewed in Gudas et al. 2013).
Activation of Hox genes is accompanied by a change from bivalent chromatin to euchromatin (reviewed in Soshnikova and Duboule 2009). Bivalent chromatin has extensive methylation of lysine-9 on histone H3 (H3K9me3), a repressive mark, with interspersed punctate regions of methylation of lysine-4 on histone H3 (H3K4me2, H3K4me3), an activating mark. Euchromatization initiates at the 3' ends of clusters and proceeds towards the 5' ends, with the euchromatin migrating to an active region of the nucleus (reviewed in Montavon and Duboule 2013). This change in chromatin reflects a loss of H3K27me3 and a gain of H3K4me2,3. Polycomb repressive complexes bind H3K27me3 and are responsible for maintenance of repression, KDM6A and KDM6B histone demethylases remove H3K27me3, and members of the trithorax family of histone methylases (KMT2A, KMT2C, KMT2D) methylate H3K4.