A model for reading morphogenetic gradients: autocatalysis and competition at the gene level.

How are morphogenetic gradients interpreted in terms of embryonic gene transcription patterns within a syncytium such as the Drosophila blastoderm? We propose a hypothetical model based on recent findings in the molecular biology of transcription factors. The model postulates a morphogen which is itself a spatially distributed transcription factor M or which generates a distribution of such a factor. We posit the existence of an additional, zygotically transcribed "vernier" factor V. M and V form all possible dimers: MM, MV, and VV. These are differentially translocated to the nuclei and bind with various affinities to responsive elements in the V promoter, thereby contributing to activation/inactivation of V transcription. We find four generic regimes. In order of complexity, they are as follows: (i) MM activates V; the M gradient gives rise to a sharp transcriptional boundary for V and to a secondary gradient in the concentration of protein V; (ii) MV activates V; a sharp boundary in transcription and distribution of V arises; (iii) MM and MV compete for binding; a stationary stripe of active V transcription is generated; (iv) MM and VV are in competition; a stripe of V transcription moves from one end of the embryo toward the other and may stop and/or dwindle at an intermediate position. Tentative interpretations in terms of Drosophila genes such as bicoid and hunchback are presented.