Here's a refinement in the story of early pattern formation in the Drosophila embryo. Just to recap a little bit, I've told you that in flies there is a molecule, bicoid, that is expressed in a gradient and that is a transcription factor that regulates the expression of other genes. In particular, there is a set of gap genes, Hunchback and Krüppel and Knirps and Giant and Tailless that read the gradient of the bicoid morphogen and are turned on in specific bands along the length of the embryo.

All of this is more or less accurate, but I have to tell you now that it is also a great simplification. Development is much, much more complicated than that. From that description, you'd think that the bulk of the work in specifying identities along the longitudinal axis is done once the bicoid gradient is set up—everything from that point could require nothing but passive obedience from the downstream genes to what bicoid tells them to do. There are also a great many interactions between different downstream genes that are important in shaping the distribution of gene expression, however, and one thing developmental biologists can't do is get trapped into simple linear thinking.

The latest issue of Nature has an article by Jaeger et al. that characterizes the dynamic and interactive properties of the gene circuits that influence gap gene expression. The work involved quantitative measurements of the levels of gap gene expression over space and time in developing flies, using fluorescent probes and a confocal microscope. Oooh, pretty pictures:

Dynamical shifts in gap gene domains are reproduced by gap gene circuits. a, b, Drosophila melanogaster blastoderm stage embryos at late cleavage cycle 14A (time class T8) immunostained for Kr and Gt (FlyEx embryo, rge9; a) and Kni and Hb (rb8; b). Anterior is to the left, dorsal is up. Bars indicate the region included in gap gene circuits.

The authors had also constructed models of the regulatory interactions between the different genes, using the measured values in the embryos as parameters, and ran computer simulations that generated diagrams like the one below that showed how the bands of gap gene expression were fine-tuned and shifted over time. The data and web-based versions of their simulations are all available at the Reinitz Lab website.

Output of a model gene circuit illustrating the changing pattern of anterior-posterior expression (horizontal axis) over time (vertical axis) for three different gap genes.

Biology and the computer models align nicely. On the left are the measured values of gap protein concentrations, and on the right are the values produced by their gene circuit model.

c, d, g, h, Gene expression data (c, g) and gap gene circuit model output (d, h) at early (T1; c, d) and late (T8; g, h) cycle 14A. Vertical axes represent relative protein concentrations, horizontal axes represent position along the A-P axis (where 0% is the anterior pole). There are no Tll data for T1 (c). e, f, Gap domain shifts for Kr, kni and gt covering the time between patterns shown in c, d and g, h. Lines indicate the position of maximum concentration for each domain. Coloured areas represent regions in which protein concentration is above the half-maximum value. Positional values for data were obtained by approximation with quadratic splines.

What does it all mean? It may seem like a small point, that bands of gene expression wobble about a bit during development, but it's actually representative of a fairly important principle. Even in cases where we see a strong maternal pre-pattern imposed on the embryo, where it's as if the mother has drawn a blueprint on the egg to simplify the developmental task of zygotic cells, we are misleading ourselves if we think of the process of development as a passive unfolding of a program. As the authors explain,

Our results indicate that maternal Bcd, Hb and Cad alone are not sufficient for positioning of gap gene domains and hence do not qualify as morphogens in a strict sense. As has been pointed out, an active role of target tissue in specifying positional information contradicts the traditional distinction between the instructive role of maternal morphogens and their passive interpretation. The requirement of specific regulatory interactions in the target tissue for proper interpretation of positional information can be interpreted as a requirement for specific tissue competence. In addition, the dynamical nature of positional information, as encoded by expression boundaries, suggests that positional information in the blastoderm embryo can no longer be seen as a static coordinate system imposed on the embryo by maternal morphogens. Rather, it needs to be understood as the dynamic process underlying the positioning of expression domain boundaries, which is based on both external inputs by morphogens and tissue-internal feedback among target genes.

This is an important attribute of developmental systems in evolution, as well. Information is distributed throughout the different steps of development; it isn't the product of a authoritative master pattern that is defined at one instant. The relative autonomy of different molecular modules in the organism and their ability to independently negotiate with one another to set up domains of expression are properties that confer a great flexibility and robustness on the process.

Jaeger J, Surkova S, Blagov M, Janssens H, Kosman D, Kozlov KN, Manu, Myasnikova E, Vanario-Alonso CE, Samsonova M, Sharp DH, Reinitz J (2004) Dynamic control of positional information in the early Drosophila embryo. Nature 430:368-371.