Getting the Measure of Positional Information

Understanding the mechanisms that underlie pattern formation is one of the major challenges of developmental biology. The complexity and beauty of the patterns on butterfly wings, fish scales, or bird feathers are not only remarkable products of developmental processes but puzzles that tease our intellects. If we are to understand these beautiful products of cellular activity, we need to first investigate simpler patterns, which are more tractable experimentally. A good example is the subdivision of an embryo along its main axis, which can be represented as a polarized subdivision of a cellular field. Over 40 years ago, Lewis Wolpert offered a conceptual solution to this problem in the form of the French Flag model [1]. The central element of the proposal is that spatial gradients of substances called morphogens are the cause of such subdivision (Figure 1, left panel). The idea is simple: specific concentration thresholds in the gradient are detected by cells in the target tissue and lead to the expression of distinct sets of target genes. The crucial ingredient of the argument was a precise and direct correlation between the input (the gradient) and the output (the response of the tissue)—each threshold corresponds precisely to a border of an expression territory. However, the quantitative nature of the argument made it difficult to test for the lack of adequate measurable observables. The discovery of the Bicoid (Bcd) gradient in the early embryo of the fruit fly Drosophila melanogaster in the 1980s (Figure 2A) provided the first direct evidence for the existence of the postulated morphogen gradients [2–4]. Bcd is a transcription factor that—within the syncytium of the early Drosophila embryo—forms a concentration gradient from anterior to posterior. The gradient is necessary for the antero-posterior patterning of the embryo [3]. This finding led to a general revival of interest in spatial patterning and to the discovery of many similar gradients involved in other developmental processes, such as patterning of the wing disc in Drosophila or the neural tube in vertebrates [5,6]. Bcd activates transcription of gap genes such as hunchback (hb) (Figure 2B) in a concentration-dependent manner [7,8]. This observation—together with the more general correlation between the concentration of Bcd and downstream expression patterns [3]—provided support for the hypothesis that Bcd is a morphogen [9,10]. If this were the case, Bcd alone should be capable of precisely positioning boundaries of downstream gene expression as proposed by the French Flag. But is it? To test this requires precise measurements of Bcd and its targets. One major issue, first highlighted by C. H. Waddington, is the heavy reliance of the French Flag model on precise detection of concentration thresholds by cells in the target tissue [9]. Early efforts to model the system indicated that target gene auto-activation [11], or more complex interactions among downstream factors [12,13], would be necessary to achieve the required robustness. However, these studies were hampered by the absence of reliable experimental evidence on the variability of gene expression against which the models could be tested.