March 7, 2005. Biology 104. Albert Harris   Background on some popular "models" = categories of theories Quote from a Feb 2002 paper from Eric Wieschaus' Lab: "The morphogen gradient model is the most widely accepted model describing how cells in early embryos acquire their positional information" (Nature vol 405, pages 798-803) Which cites J. Theoretical Biology vol 25, pages 1-47 (1969) During periods of progress in a science, competing sets of ideas ('"models") often co-exist, like different religions or sects Hypothesis = Theory ~= Model ~= Ideology Over the last 30 years, "Positional Information" has come to dominate many people's thinking in developmental biology. It was invented by Lewis Wolpert. "Positional Information" is a combination of related hypotheses: (Remember that each of the following is a guess) a) perpendicular diffusion gradients of chemicals form inside all developing embryos, with "sources" (= locations of synthesis or secretion of each chemical * at the extreme front and/or rear; ** at the extreme top and/or bottom; and ***either/or at the middle, or along one edge) b) Embryonic cells decide what cell type to differentiate into by measuring the local concentrations of these 3 + chemicals. Here is what one cell might think "The concentration of the anterior-posterior chemical is 20% of maximum; the concentration of the top-bottom chemical is 95% maximum; & the concentration of the medio-lateral chemical is 92% of max." And this combination of concentrations would somehow activate the genes that cause that cell to differentiate into nerve cell of the rear of the spinal cord. Chemical concentrations serve as the X, Y, Z coordinates of graph paper; or as measurements of longitude, latitude and altitude. A key difference from other "models" is that anatomical geometry is generated by the combination of responses of cells to chemicals The chemical gradients do not tell cells what to differentiate into; they tell them where they are relative to the rest of the embryo. c) When an embryo is cut in two (for example by Driesch!) then these chemical gradients adjust their slopes, by forming a new "source" maximum along one side and a new "sink" minimum along the other side. Adjustment of gradient slopes is how Positional Information "explains'" embryonic regulation, such as by separated cells of early sea urchin embryos. Wolpert re-defined the word "Morphogen" (that Turing invented and used with a very different meaning) Many people now say that "a true morphogen" must be able to cause differentiation of 2 or more different cell types, depending on the local concentration of the chemical where the cells are. Positional information claims that anatomical patterns result from how embryonic cells respond to monotonic gradients of chemical concentrations. In other words, they claim there is no geometric pattern in the molecular signals, which just provide the equivalent of latitude and longitude. That is very different from the Turing's hypotheses about The operation of reaction-diffusion system generates alternating regions of high and low concentrations of "morphogens" Then (the idea is) that cells differentiate into one cell type where the morphogen concentration is high, and another where it is low. The geometric pattern of the future anatomy is generated first as waves of chemical concentrations. Such a chemical pattern is called a "Pre-Pattern" Pre-patterns are a very different idea from Positional Information, because the geometry is in spatial distribution of signals. The defining idea is that anatomical geometry IS GENERATED by the different cells' responses to the gradient chemicals . A related difference is whether evolutionary changes in shapes and positions of organs results from changes in the spatial patterns of the signalling molecules, or are such changes produced by changed responses to the same signals patterns. For example, how would evolution cause an organ to form at a more anterior location? The PI explanation: Genes that used to be "turned on" by lower concentrations of the Ant-Post morphogen now become turned on by a higher concentration of that chemical (Maybe at 80% maximum instead of 70% maximum) The prepattern explanation: Some change in the geometric distribution of signalling molecules, perhaps with a concentration peak shifting to a more anterior position. Originally the claim was that (Driesch's) regulation would be impossible unless the embryos of a species used P.I. Drosophila development is very mosaic, however, but when long-range diffusion gradients were found to control early development the P.I. people said: "See, we were right all along!" Many biologists (for example, those who teach Biology 52) use the term "positional information" very loosely, as if it meant any molecular signals that control where an organ will develop. They think it means the same thing as pre-pattern; but some also credit Wolpert with the idea that something controls the positions at which cells will differentiate into each organ. They are wrong. Many others had proposed gradient theories as early as the 1800s. Wolpert's innovation was the claim that regulative development requires control by adjustable gradients. But the best examples of diffusion gradients were found in flies, which are mosaic! Researchers coming from molecular biology tend to like P.I. Researchers trained in embryology tend to hate it rabidly. Wolpert is the roadrunner, Embryologists are the Coyote . On the other hand, it must be admitted that "bicoid", "nanos", "dorsal" and other proteins really do form gradients in developing fly embryos; Mutations in the genes for these and related proteins really do cause abnormal embryos, with reversed axis development. The roadrunner zooms across empty space by not looking down! A Feb 2002 paper in Nature, by Houchmandzadeh & Wieschaus & Leibler (vol. 415: p798 reports measures of bicoid protein distributions: "Here we show that the Bcd gradient displays a high embryo-to embryo variability, but that this noise in the positional information is strongly decreased ('filtered') at the level of hunchback gene expression. In contrast to the Bcd gradient, the hb expression pattern already includes the information about the scale of the embryo." For those not fluent in Wolpert-Speak, this can be translated as: Concentrations of the bicoid protein vary much more irregularly than do concentrations of the hunchback protein (that bicoid supposedly controls). Instead of doubting that bicoid really is the A-P "morphogen" (or, heaven forbid, doubting P.I. dogma!) we prefer the idea that something "filters" the irregularities. I am trying to invent a really good analogy to parody this thinking and I would appreciate suggestions from members of the class. The pattern of X causes the pattern of Y. But then evidence shows that X varies more irregularly than Y. So we conclude that Y filters out the parts of X that are wrong. (but then isn't the filter the true cause of the Y pattern?) Other possible explanations are: A) That bicoid isn't really the controlling molecule (most insects turn out not to have any bicoid, anyway) B) That several kinds of molecules combine their effects. (& bicoid has been discovered, but not the others) C) Some feedback loop, with negative feedback, damps ("filters out") variations in concentrations of any one "morphogen". D) Can you invent one or more alternatives? Drosophila (fly) embryology Two Nobel prizes were given for embryological discoveries: 1) To Spemann for the discovery of induction (in salamander gastrulation). Awarded in 1935. 2) To Nusslein-Volhard, Wieschaus, and E. B. Lewis, for discovering genes that control early development in Drosophila. Awarded in 1995. Fly embryos have two special properties, very different from vertebrates: A) Their axes of asymmetry are already decided, before fertilization, by molecules from the surrounding cells. B) For the first twelve mitotic cells, the nuclei do not become separated into different cells. The embryos remain syncytial! At the 13th division cycle, there is a single mass cleavage. Until then, even proteins are free to diffuse from one end of the embryo to the other; or from the top to the bottom. Regarding #A) There are some species of wasps whose eggs inside caterpillars, which develop into multicellular masses that bud off variable numbers of wasp embryos. This phenomenon is called polyembryony. Can the mechanisms that control axis determination be utterly different in these wasps? Regarding #B) From books I have read, some species of insects do not remain syncytial like this during the early mitoses; but many other arthropods do have syncytial embryos (crayfish). I am puzzled to find out the taxonomic and evolutionary boundaries of this phenomenon. Nusslein-Volhard and Wieschaus did a "genetic screen", seeking mutations that caused embryos to become structurally abnormal and/or stop development at early stages. Later, they did "in situ" labeling of normal embryos with "probe" nucleic acids whose base sequences were complementary to the normal versions of the genes they had discovered in their screen for early extreme birth defects. (Nature vol 287: 795-801) Experiments were done introducing extra copies of genes, injecting proteins, etc. Among other genes, they discovered one "Bicoid" * whose m-RNA is localized at the anterior end, ** and whose protein forms an anterior-posterior diffusion gradient. Another gene's proteins vary in behavior along the dorso-ventral axis "Dorsal"

 

back to syllabus

back to index page