Krumlauf Lab

The goal of our research is to understand Hox transcriptional regulatory networks that control patterning of the mammalian brain, skeleton, and basic body plan during development, disease, and evolution.

Molecular studies and genomic analyses have revealed that a wide range of animals have a surprisingly similar collection of genes in their genomes. The basis for generating and patterning different tissues within an animal and diversity between species therefore largely depends upon mechanisms that control how and when these genes are utilized. For this reason, it is important to understand the “wiring blueprint” or regulatory networks which control the dynamic patterns of gene expression and cell signaling necessary to form tissues and organs in development. This also facilitates approaches for exploring how such gene regulatory networks may vary in contributing to diversity during evolution. Knowledge of these interactive regulatory networks is also relevant to human health because alterations in their activities can be associated with clinical abnormalities and disease.

The molecular focus of our group is centered on the Hox homeotic genes, which encode highly conserved homeodomain transcription factors that play similar regulatory roles in both vertebrates and invertebrates. Genetic and regulatory analyses in many model systems have demonstrated that the Hox family of transcription factors provides a coordinated and highly conserved molecular framework (code) for specifying and patterning distinct regional properties of tissues during development, tissue homeostasis, and adult life. Mutations that alter Hox genes or their regulation can lead to leukemia, metastatic tumorigenesis, and a variety of patterning defects of the brain, craniofacial region, limb, and skeleton. Therefore, our laboratory is interested in understanding the underlying mechanisms of how Hox-dependent transcriptional regulatory networks control the formation and patterning of the basic mammalian body plan during development, how they are associated with or affected in diseases, and the roles of these pathways in generating morphological diversity during vertebrate evolution.

The major biological focus of our research is aimed at understanding hindbrain, craniofacial, and skeletal development. The hindbrain is a complex coordination center for motor activity, breathing rhythms, and many unconscious functions, and is highly conserved in vertebrates. The hindbrain also plays a key role in regulating the formation of head structures, such as bone and connective tissue, via the generation of cranial neural crest cells. The paraxial mesoderm generates somites, which give rise to diverse bones of the axial skeleton. Segmentation plays an important role in the functional organization of these tissues. The cellular processes of segmentation in both the hindbrain and somite systems depend upon establishing ordered patterns of Hox gene expression as a mechanism for generating a combinatorial code that specifies unique identities to the segments and their derivatives (Alexander et al 2009).

Our long-term goal is to understand the regulatory basis of tissue patterning and morphogenesis by generating a deep mechanistic level of knowledge of the Hox-dependent regulatory cascades which underlie their diverse functional roles in development, disease, and evolution. Towards this end, we investigate the upstream regulatory mechanisms which generate ordered and restricted patterns of Hoxexpression and characterize downstream Hox target genes necessary for elaborating their function. From a biochemical standpoint we want to determine the nature of their in vivo binding sites and rules for target site recognition, how they are recruited to these loci as Hox protein complexes, and the basis for specificity of individual Hox proteins. This insight will aid knowledge of how Hox proteins elaborate their functional roles in specifying unique properties to tissues through the control of cell and developmental processes. To address these questions, we use mouse embryos and ES cells as a primary model system in combination with diverse genetic, genomic, biochemical, and imaging approaches. We also exploit the extensive conservation in hindbrain and somite segmentation between vertebrates by utilizing a variety of other model systems (human cells, chick, zebrafish, and lamprey) to investigate similarities and differences in patterning processes during vertebrate evolution. More information about research model organisms and some of the emerging model systems that my lab and other Stowers researchers study to understand biological processes and relevance to diseases can be found in this Stowers Report article. 

Project Summaries:
 
A. Mechanisms regulating activation and expression of Hox genes
In mammals, there are 39 Hox genes organized into four complexes (HoxA, HoxB, HoxC, HoxD) each located on a different chromosome. These clusters arose through duplication and divergence from a common ancestral cluster. An intriguing property of Hox genes is that their linear arrangement within each complex mirrors their sequential activation and function in the embryo. This phenomenon is known as colinearity and it serves to establish the nested combinations of Hox proteins along the A-P axis that form a combinatorial code for specifying unique regional identities in different tissues. Opposing retinoid and Wnt/Fgf signaling centers drive axial growth and extension and the Hox clusters integrate cues from these signaling pathways to pattern growing tissues. Hence, understanding how regulation of Hox genes is coupled to axial growth and segmentation through these signaling pathways is important for building a picture of the fundamental mechanisms that govern elaboration of the basic body plan.

1. Enhancer sharing and long range regulation of Hox genes and lincRNAs
A number of retinoic acid response elements (RAREs) that modulate transcriptional activity through bounding of heterodimeric RAR:RXR nuclear hormone receptors have been identified within the mammalian Hox clusters. Our analyses have revealed that these RAREs can function as cis-regulatory elements to modulate expression of multiple HoxB genes in a cluster through enhancer sharing and long-range regulation (Nolte et al 2013, Ahn et al 2014). Targeted mutagenesis and gene editing data (CRISPR/cas9) of the endogenous cluster reveals that these RAREs play an important functional role in long-range regulation of multiple HoxB genes during embryogenesis and tissue homeostasis. Some of the RAREs embedded within the Hox clusters are also involved in regulating long intergenic non-coding RNAs (lincRNAs) and microRNAs present in the HoxB cluster (DeKumar et al 2015 DeKumar and Krumlauf 2016). This data provides evidence for shared regulation of coding and non-coding transcripts that may contribute to conservation of clustered Hox organization in animals. The positions of several of these RAREs appear to be highly conserved in Hox clusters during chordate evolution suggesting they may be components of the ancient mechanism which sets up nested domains of axial Hox expression.

Ongoing projects are aimed at investigating the regulatory mechanisms and functional roles of RAREs and lincRNAs in developmental, disease and evolutionary processes. We utilize CRISPR/Cas9 to generate mutations in single and multiple RAREs and lincRNAs in the endogenous mouse Hox clusters to probe in vivo function and synergy between elements. Recombineering of bacterial artificial chromosomes (BAC) containing the mouse HoxB or HoxA clusters and gene editing are also being used to label multiple Hox genes with different fluorescent protein reporters (Parrish et al 2014). Multispectral imaging of hindbrain tissue from these transgenic or gene edited mice is then performed to simultaneously monitor the expression of multiple Hox genes over time in lines with wild type or mutated RAREs (Ahn et al 2014). The mice carrying diverse fluorescent reporters also provide a means for isolating and identifying distinct groups of cells based on their combinations of Hox expression which is useful for global expression profiling and genomic analyses. These studies are uncovering novel roles for retinoids, lincRNAs and Hox genes in neural, cardiac, and hematopoietic development and enhance understanding of how embryos establish appropriate programs of gene expression in response to RA, Wnt and Fgf signaling.

2. Chromatin state, topological domains and boundaries
Genes flanking the four Hox clusters have very different expression properties in comparison to Hox genes implying distinct modes of regulation. Since cis-regulatory elements that modulate Hox expression can work over long range we wish to understand how their activity restricted to the Hox genes and do not impact neighboring genes (Nolte et al 2013). Current models suggest that this may be mediated by chromatin looping, boundary, and/or insulator elements. In an ongoing project, through a combination of regulatory analyses, phylogenetic footprinting, and biochemical approaches we identified a novel sequence flanking the HoxB complex. Functional studies indicate this element functions as a boundary or insulator to generate a HoxB regulatory domain or landscape distinct from the neighboring genes. Deletion of this region from the endogenous cluster alters regulation of adjacent genes. Similar conserved elements are positioned adjacent to other Hox clusters. We are currently performing a series of experiments to characterize the properties, components, and mechanisms through which this boundary region establishes a chromosomal domain to facilitate control of HoxB expression and whether similar mechanisms are used by other Hox clusters.

Dynamic patterns of gene expression can depend upon changes in the state or modification of chromatin. We have utilized the programmed differentiation of mouse ES cells into neural fates and dissected tissue from mouse embryos in combination with genomic technologies to investigate links between chromatin state and regulation of Hox genes (Lin et al 2011, DeKumar et al 2015). Epigenetic changes in chromatin modification and the binding of regulatory proteins to Hox clusters reveal dynamic and specific changes in epigenetic marks associated with activated and/or repressed states and transcriptional activity of Hox genes. Transcriptional profiling using RNA from the differentiating ES cells and mouse tissues has uncovered an extensive array of non-coding RNAs from both strands of the four Hox clusters. While some of these correspond to known miRs or regulatory RNAs, many are novel and display dynamic changes over time (DeKumar et al 2015, DeKumar and Krumlauf 2016). Several of the novel non-coding RNAs map to cis-regions which correlate with changes in chromatin state identified in genomic assays or in regulatory modules. Current projects are testing whether these non-coding RNAs directly participate in the activation or regulation of Hox genes or other developmental targets. 

3. Cis-regulatory elements controlling segmentation
Sequence comparisons between vertebrate Hox clusters in combination with transgenic regulatory analyses have provided insight into mechanisms governing Hox expression in hindbrain segments (rhombomeres). A series of cis-regulatory modules underlie the generation of discrete domains of segmental expression in the hindbrain. Expression of any given Hox gene is the result of the combined activity of a series of cis-modules that control specific subsets of the segmental expression pattern and each cis-module is subject to its own regulatory inputs and timing. We have used this information to construct a gene regulatory network for vertebrate hindbrain segmentation (Alexander et al 2009, Tümpel et al 2009, Krumlauf 2016, Parker et al 2016). Current projects apply proteomic and biochemical approaches to expand our knowledge of the upstream transcription factors, cofactors, and signaling pathways that activate Hox expression through these hindbrain cis-regulatory modules. Experiments are also being conducted to investigate the basis for cis-regulation of Hox genes in somites and cranial neural crest cells.

4. Evolution of mechanisms coupling Hox expression to hindbrain segmentation
The hindbrain gene regulatory network we have characterized in jawed vertebrates controls patterning by integrating segmentation with the generation of segmentally-restricted domains of Hox gene expression. We are investigating when this network evolved and when Hox genes became coupled to it. Non-vertebrate chordates display nested domains of axial Hox gene expression but lack hindbrain segmentation and segmental regulatory elements (Manzanares et al 2000, Natale et al 2011). The sea lamprey, a jawless fish, is poised to provide unique insights into vertebrate origins due to its phylogenetic position at the base of the vertebrate tree. We have cloned lamprey Hox genes and investigated their relationship to the hindbrain regulatory network (Smith et al 2013). In collaboration with Marianne Bronner at Caltech, using a novel tool that allows cross-species comparisons of regulatory elements between jawed and jawless vertebrates we have demonstrated that lamprey Hox genes are coupled to hindbrain segmentation in a manner conserved with jawed vertebrates (Parker et al 2014, Parker et al 2016). This work has shown that coupling of Hox gene expression to segmentation of the hindbrain is a defining feature and an ancient trait with an origin at the base of vertebrate evolution. We are expanding this analysis and investigating how it extends to Hox regulation in cranial neural crest networks.

B. Hox targets, binding sites, and cofactors

1. Transcriptional profiling segmental expression in the hindbrain
We are characterizing rhombomere-specific patterns of gene expression which are important in generating individual segments and identifying genes which are dependent upon Hox activity. Towards this end we have performed transcriptional profiling on whole dissected hindbrains and individual hindbrain segments isolated by laser capture microscopy from transgenic lines expressing fluorescent protein reporters in specific segments. This readily enables the identification and isolation of segments in the absence of overt morphological segmentation. Transcriptional profiling by microarrays and RNA-seq on tissue and segments isolated from wild type and Hox mutant embryos help to identify genes dependent upon Hox activity. Projects in this area are uncovering a number of pathways and targets genes downstream of Hox involved in regulating neurogenesis in the hindbrain.

2. Genome-wide identification Hox binding sites
Very little is known about the downstream target loci Hox proteins regulate and how they are recruited to these loci as Hox protein complexes. One challenge is that all 39 mammalian Hox proteins are highly similar. Hence, their individual specificity is likely to be modulated by differences in cofactors, interacting proteins, target binding sites, or other unknown processes. A major current focus involves projects identifying in vivo relevant Hox response elements on a genome-wide basis to gain insight into how Hox proteins potentiate their functional roles. We are using transgenic mice, gene editing, and the programmed differentiation of ES cells carrying epitope-tagged versions of different Hox proteins in combination with ChIP-seq genomic technologies to identify in vivo binding regions for Hox proteins and their associated target genes. In addition to the known co-factors, such as Pbx and Meis, this analysis is revealing that in vivo binding sites for Hox proteins display a range of novel sequence motifs. The emerging data indicates that different Hox proteins can bind to distinct target regions and is providing insight into new potential partners and cofactors for Hox activity. Current projects are aimed at validating the role of these Hox binding sites in regulatory assays and comparing them to genes whose expression is altered in transcriptional profiling analyses. This approach is serving to identify direct down-stream regulatory targets and pathways under control of individual Hox proteins and to illuminate common targets for multiple Hox proteins. In light of the conserved nature and function of Hox proteins, in collaboration with Julia Zeitlinger’s lab at the Stowers Institute, we are also conducting genome-wide analyses of Hox targets in Drosophila. The goal is to compare and contrast Hox targets in mammalian cells and Drosophila to look for shared properties and common target pathways.

3. Proteomic approaches to identification of Hox cofactors and complexes
Hox proteins bind DNA and exert their regulatory activity in part through interactions with cofactors and recruitment of coactivator/corepressor complexes. To begin to identify some of these interacting proteins, we are collaborating with Michael Washburn and Laurence Florens of the Proteomics Center at the Stowers Institute to use MudPIT-based mass spectroscopy approaches. Exploiting ES cells, gene edited mice and transgenic mice carrying epitope-tagged versions of different Hox proteins in combination with immune precipitation of the tagged Hox proteins followed by protein sequencing we are searching for interacting partners. To complement this approach, in collaboration with Ron and Joan Conaway, we are using Hox binding sites identified by genomic approaches as high affinity templates in combination with nuclear extracts to assemble binding complexes and characterize them by protein sequencing.