The methods described here allow integration of expression patterns from different genes in different animals, providing the potential to obtain a comprehensive picture of gene expression in every cell. For example, using this methodology it should be possible to describe the patterns of activity of the promoters for all of the transcription factor genes active during embryogenesis. Combined with emerging knowledge of transcription factor binding sites, these expression patterns would begin to reveal the network of regulatory control. Predicted networks could then be tested by quantitative analysis of reporters after RNAi or genetic depletion of predicted regulators. With appropriate controls for integration site and copy number, promoter dissection could allow the identification of the specific DNA sequences required for each regulator's effects.
The ability to trace altered lineages and generate a quantitative readout of cell-type specific reporters extends the phenotypic analysis of mutants and RNAi treatments that perturb development. As the number of cell fate markers increases, the method will become increasingly powerful in phenotypic analysis that will be required to gain a functional understanding of regulatory networks.
Future developments could considerably enhance the already powerful information obtained. Extending automated lineaging to include the last round of embryonic cell division will facilitate describing the expression of those genes only expressed in the terminally differentiated cells. This remains a challenging goal because of the active cell migrations and close packing of nuclei in the late embryo. Another improvement would be the development of an RFP-based lineaging system, which would allow the embryonic expression patterns for existing genomically integrated GFP reporters24-26
to be characterized. Whether the many other non-integrated strains available would be useful for systematic analysis is questionable, because their mosaicism would be expected to necessitate additional replicates to ensure identification of all expressing cells and because a fraction of embryos imaged would not contain the transgene at all.
Our results emphasize that promoter::reporter constructs only partially capture the complexity of regulation, emphasizing the need for more faithful transgenic strategies. We intentionally used transcriptional fusions to histones for this study to maximize sensitivity for weakly expressed genes. Protein fusions could also be generated to reveal subcellular localization of the proteins and to contrast transcriptional and translational controls, at the potential expense of reduced sensitivity for proteins with faster turnover than HIS-24. Improved cloning and transformation methods such as recombineering27
would allow the use of larger genomic segments so that the observed patterns are likely to more faithfully reflect the native pattern. Faster-folding, brighter reporters would ensure that the system detects the earliest expression of even weakly expressed genes. Extension of the system to multiple colors could increase throughput and provide kinetic information about colocalization.
In vivo single-cell analysis of gene expression is an important step towards a comprehensive molecular understanding of development. The transparency of the worm and its invariant lineage make it ideal for such analyses, but with continued progress in non-invasive imaging to track cells and ever-expanding sets of cell fate markers to substitute for the lineage, we envision equivalent analyses for more complex organisms, including mammals.