Imagine mixing all living animals in North America up in a (very large) bag, identifying which species are in the mix, and then making educated guesses about how those animals live.  Do the fishes and birds share a niche?  How often do monarch butterflies and polar bears interact?  We know the answers to these questions are dependent on the North American habitat in which each animal lives.  Sometimes fishes and birds do interact, but only in particular habitats. 

This mental exercise may seem tedious since we are usually able to examine these animals in their natural habitats, and so have no need to try and reconstruct these interactions from a hypothetical mixed bag of animals.  But we are not as lucky in the case of microbes.  Even though we are now reaching a point where we can reliably identify species of microbes using DNA based approaches, these approaches essentially are like tossing all animals in North America together and trying to figure out which ones share a niche.  For example, when bacteria living in the human gut are studied using DNA based approaches, usually the DNA is extracted from feces.  Therefore, we know that the bacteria were in the human gut but we have no idea where in the gut they were, or who was interacting with who.  And the gut is a long, winding place, and probably has more bacterial niches than we can (or would like to) imagine.

But blessed is technology.  A recent paper by research groups in France and Uruguay (Pedron et al: http://mbio.asm.org/content/3/3/e00116-12.full.html) describes experiments where bacteria were plucked from intestinal crypts of several strains of mice, identified, and compared to bacteria found in nearby lumen samples.  (Crypts are the spaces between long finger-like extensions on the surface of the intestines.) The authors found that bacterial communities from the crypts of different mice were more similar to each other than the crypt and lumen sample from the same mouse.  The bacterial genus Acinetobacter was very abundant in the crypts and almost absent in lumina.  This was very exciting, and suggests that intestinal crypts are Acinetobacter's primary habitat within the gastrointestinal tract. 

Pedron et al. mention that Acinetobacter has also been found in the midgut of many invertebrates (e.g., tsetse flies, mosquitoes, and fruit flies).  Perhaps this is a genus that has co-evolved with animals throughout millions of years.  Or perhaps it has been independently acquired as a gut commensal in multiple evolutionary branches.  More research is needed to figure out if it is the same species of Acinetobacter colonizing crypts in mice and fly, and much more research to figure out what Acinetobacter actually doing there.