The icosahedron is the shape nature builds most insistently at very small scales, and the reason is understood rather than mysterious.
Most viruses have tiny genomes and cannot afford to encode many different coat proteins. Crick and Watson pointed out in the 1950s that a small genome forces a shell built from many copies of a few identical subunits, and identical subunits in a closed shell produce high symmetry. Caspar and Klug worked out the consequences in 1962: a closed shell of identical units contacting one another in nearly but not exactly the same way takes icosahedral symmetry, and only certain sizes are possible. Their triangulation number, T = h squared plus hk plus k squared, permits shells of 60, 180, 240, 420 subunits and so on, and high-resolution structures have confirmed the pattern ever since. The shape is an economy measure, not a symbol.
Other genuine occurrences are worth separating from the folklore. Boron forms clusters of twelve atoms at the vertices of an icosahedron, and those clusters are the building blocks of elemental boron and of boron carbide. Some radiolaria, single-celled marine organisms with mineral skeletons, grow strikingly icosahedral shells; Haeckel's nineteenth-century drawings of them are famous and flatter the geometry somewhat. In 1982 Dan Shechtman observed a metal alloy diffracting with fivefold symmetry, believed impossible in a crystal; the results, published in 1984 and awarded the Nobel Prize in Chemistry in 2011, established icosahedral quasicrystals, ordered solids that never repeat.
Against all that, the modern esoteric assignments deserve a plain label. Placing the icosahedron on the sacral chakra, or with the west, or with the moon, comes from twentieth-century New Age writing and from no ancient source; different authors assign it differently, and the classical Indian diagram for the water element is a flat crescent rather than a solid. The water association itself is Plato's and is genuinely that old. The rest is recent.
Sources Physical Principles in the Construction of Regular VirusesD. L. D. Caspar and A. Klug with Metallic Phase with Long-Range Orientational Order and No Translational SymmetryD. Shechtman, I. Blech, D. Gratias and J. W. Cahn