Experimental approaches for gene regulatory network construction: the chick as a model system

Abstract

Setting up the body plan during embryonic development requires the coordinated action of many signals and transcriptional regulators in a precise temporal sequence and spatial pattern. The last decades have seen an explosion of information describing the molecular control of many developmental processes. The next challenge is to integrate this information into logic "wiring diagrams" that visualize gene actions and outputs, have predictive power and point to key control nodes. Here, we provide an experimental workflow on how to construct gene regulatory networks using the chick as model system.

Figure 1

An example of a simplegene regulatory network. a. In this network, signal 1 from tissue A to signals to tissue B via a receptor (≫); this triggers the expression of transcription factors 1 and 2 (gene 2 and 3) and several downstream targets are activated (genes 4–6). These in turn regulate each others’ expression as indicated by the coloured arrows. Positive interactions are depicted as arrows, negative interactions as bars. At a later time, tissue C emits a second signal, which leads to changes in gene expression in cells nearby, while cells far away are not exposed to signal 2. This leads to differential gene expression in tissue B1 and B2 depending on the transcription factors interactions and signalling input. b. The initial stage in network construction is defining the regulatory state: the sum of genes expressed in each tissue at different times and the signals received from neighbouring tissues (see Fig. 2 stage 2). c. Perturbation experiments suggest interactions and hierarchy (see Fig. 2 stage 3). d. Confirmed interactions after enhancer discovery and testing (Fig. 2 final stage).

Figure 2

Experimental workflow for building a gene regulatory network. Details for each step are described in the text. Generating a GRN is an iterative process, in which each perturbation experiment informs about the network architecture; integration of new information into the network points to novel hypotheses that can then be tested experimentally. Bioinformatics approaches are required to predict regulatory interactions and conserved regulatory modules (CRMs). In vivo testing of CRM activity and their regulation feeds back to the network.

Figure 3

Gain- and loss-of-function experiments in chick embryos. a: Exogenous DNA or oligonucleotides are transfected into chick embryos by electroporation. b: eGFP was electroporated into the ectoderm of a primitive streak stage embryo. GFP fluorescence can first be detected about 3–4 hours after electroporation. After overnight culture the neural plate, the non-neural and extraembryonic ectoderm carries the GFP construct. c and d: electroporation of a Pax2-specific morpholino (MO; green) (Mende et al., 2008) at primitive streak stages into otic precursors leads to loss of Pax2 protein (red) in electroporated cells. c and d show the left and right side of the same embryo, respectively. Note: only few cells carry Pax2 MO on the left hand side, while most are electroporated on the left hand side; this leads to a change in placode morphology (Christophorou et al., 2010).

Figure 4

Testing enhancer activity in chick embryos. a. Diagram showing the GFP-reporter construct containing the putative enhancer, a minimal TK promoter and eGFP; RPF is driven by chick β-actin and CMV promoter. Embryos are electroporated at primitive streak stages and cultured until they have reached the stage when enhancer activity is expected. b–e. The embryo was electroporated at primitive streak stages with ubiquitous RFP and GFP driven by an otic Eya1 enhancer (Ishihara et al., 2008). After overnight culture the embryo has reached the 13-somite stage and shows enhancer activity in the otic placode b. bright field image. c. RFP expression is wide spread. d. GFP is specifically expressed in the otic placode. e. Overlay of bright field and GFP image. White circles indicate the otic placode. mb: midbrain; hb: hindbrain, ov: optic vesicle; som: somite.