Since the publication of the fluid-mosaic membrane theory by Singer and Nicolson in 1972 generations of scientists have adopted this fascinating concept for all biological membranes. Assuming the membrane as a fluid implies that the components embedded in the lipid bilayer can freely diffuse like swimmers in a water body. During the detailed biochemical analysis of the thylakoid protein components of chloroplasts from higher plants and algae, in the '80 s and '90 s it became clear that photosynthetic membranes are not homogeneous either in the vertical or the lateral directions. The lateral heterogeneity became obvious by the differentiation of grana and stroma thylakoids, but also the margins have been identified with a highly specific protein pattern. Further refinement of the fluid mosaic model was needed to take into account the presence of non-bilayer lipids, which are the most abundant lipids in all energy-converting membranes, and the polymorphism of lipid phases, which has also been documented in thylakoid membranes. These observations lead to the question, how mobile the components are in the lipid phase and how this ordering is made and maintained and how these features might be correlated with the non-bilayer propensity of the membrane lipids. Assuming instead of free diffusion, a "controlled neighborhood" replaced the model of fluidity by the model of a "mixed crystal structure". In this review we describe why basic photosynthetic regulation mechanisms depend on arrays of crystal-like lipid-protein macro-assemblies. The mechanisms which define the ordering in macrodomains are still not completely clear, but some recent experiments give an idea how this fascinating order is produced, adopted and maintained. We use the operation of the xanthophyll cycle as a rather well understood model challenging and complementing the standard Singer-Nicolson model via assigning special roles to non-bilayer lipids and non-lamellar lipid phases in the structure and function of thylakoid membranes.
