Initial source of antero-posterior organization for early embryo is so called maternal gradient of morphogens. In Drosophila egg this morphogen is well-studied bicoid protein. It is known that gradients of morphogenetically active factors are used widely by organisms to co-ordinate the initial events of embryogenesis. In the modern developmental biology the system of reading of primary (maternal) gradients is called primary morphogenetic field.
Hence we need to understand the limits on how much gradients alone can specify, since this will govern the use developing organisms are able to make of them (Lacalli & Harrison, 1991, Seminars in Developmental Biology, 2, 107-117). Why is position specified in a complex combinatorial manner via cascades of the segmentation genes, while two opposite concentration gradients of peptides-morphogens appears to fulfill the same requirement in a simpler way?
It is apparent that the gradient contains positional information in an analogue form. Analogue information is less stable and hence more error-prone. The problem exists regardless of the capabilities of the reading device used to discriminate concentration difference. It is known that the genetic response, which involves the mobilization of transcriptional machinery, is a highly co-operative, non-linear phenomenon. However, the stochastic nature of molecular motion is non-eliminated source of errors.
Reading a series of thresholds off a single gradient has some advantages (Lacalli & Harrison, 1991). But if uncorrected, the errors accumulate with each step in the hierarchy, from primary morphogenetic gradients down to final members of segmentation gene cascade. But while each step in the control hierarchy is a potential point for increasing error, each also offers an opportunity for error reduction through interactions involving the genes of that particular hierarchical level.
For even the simplest models for gradient reading, there are strategies that provide a means of minimizing error. From the Lacalli & Harrison (1991) analysis, steep gradients are much better than shallow ones. It is thus better to use a shallow gradient to cue sharper, secondary gradients, rather than using the former directly to specify structures in detail. These conclusions are in agreement with finding of Fujioka et all (1995) that adds to the previous discoveries of embryo-length primary morphogenetic gradients, the more localized gradients of segmentation gene products, and shows that Drosophila uses the same principle again to further refine pattern information. We observe there as the primary morphogenetic field is subdivided into the secondary morphogenetic fields.
As Baumgartner and Noll (1990) wrote, if position needs to be specified more stable, it is important to convert the initial analogue positional information into a digital form. This conversion appears to be mediated by the genes from segmentation network. However, the transition from analogue positional information to a purely digital code specifying position along the antero-posterior axis does not occur in a single step. The distributions of the products of the segmentation genes along the antero-posterior axis are not rectangular but rather exhibit leading and trailing edges, indicating a re-establishing components of analogue positional information.
Early Drosophila body plan is under control of as minimum as three interacting regulatory systems: anterior, posterior and terminal. But what is more, the early Drosophila embryo, especially posterior morphogenetic system is self-regulating. In conditions of experimental manipulations the posterior system governs symmetrical double-abdomen phenotype. This is classical feature of regulations in development. The anterior morphogenetic system does not self-regulate to the same degree, but in principle it could size-regulate through a competitive interaction with the posterior system, and this is enough for a corrective response to changes in egg length.
The key point here is that self-regulating mechanisms can be introduced at any of the early hierarchical levels to correct errors carried over from previous levels. The implication is that control systems dynamically structured to minimize error may be quite complex (Lacalli & Harrison, 1991). Known gene-gene interactions in limits of the segmentation net include cross-inhibition and autoregulation loops. These are sufficient to sharpen the concentration peaks, turning their graded slopes into step functions. Turing-like reaction-diffusion mechanisms are important in this respect because, though they respond to the cues provided by existing prepatterns, the new pattern they impose over these has its own intrinsic wavelength, independent of previous errors.
See also
Cues and Turing Models of Drosophila Segmentation Mechanisms
Get your own Free Homepage