The serendipitous identification of a virulence switch in Acinetobacter baumannii
The paper can be found here: https://go.nature.com/2K7w7tT
Every microbiologist is occasionally faced with an “odd” result that leads to the question; do I pursue this finding now or put it on hold for later analysis? Approximately six years ago, I was looking at colonies of Acinetobacter nosocomialis strain M2 on low percentage (0.8%) agar plates under a dissecting microscope. At the time, we were screening for motility mutants, which formed colonies with smooth edges on 0.8% agar, versus irregular colony edges formed by wild-type cells. With this particular microscope, colonies were illuminated from below with a light source where the angle could be adjusted with a mirror. At a certain light angle, two colony types were clearly present on the plate and appeared either opaque or translucent. At first, I assumed the culture was contaminated. However, both colony variants appeared identical under room lighting. Furthermore, when an isolated opaque colony was restreaked, it gave rise to cells forming translucent colonies and isolated translucent colonies gave rise to opaque colonies. A number of subsequent tests confirmed that both opaque and translucent colonies were A. nosocomialis M2 and that these colony variants were capable of rapidly interconverting.
These colony variants provided an explanation for a problematic issue we were having at the time, where a particular mutation that altered surface motility produced variable phenotypes between our lab and another lab studying the same mutant. This variability was subsequently traced to whether an opaque or translucent variant with this mutation was being tested. We immediately realized the importance of this observation for future studies in other pathogenic Acinetobacter, and tested every A. baumannii strain in our collection for these colony variants. To our surprise, virtually every strain formed opaque and translucent colonies. One exception was ATCC 17978, which formed translucent colonies only. Based on these observations, I made the decision to shift part of my lab to study these colony variants. This revealed that multiple phenotypic differences were present between these variants, along with large differences in their global gene expression profiles. A breakthrough occurred while collaborating with my colleague David Weiss, where studies demonstrated that only the opaque variant was virulent in a mouse lung model of infection, as reported in this publication.
Our ability to identify these colony variants was serendipitous for a number of reasons. First, the use of 0.8% agar turned out to be critical, as opaque and translucent variants are very difficult to distinguish on regular (1.5%) agar plates. Second, the frequency of conversion from opaque to translucent colonies is greatly reduced on 1.5% agar versus 0.8% agar. The basis for this is unclear, but likely reflects important information about environmental conditions that influence switching. Third, opaque and translucent variants appear very similar on plates under room lighting conditions and require observation using an oblique light source. While the decision to continue studying these variants was almost put on hold multiple times during the initial stages of this finding, it reinforces the fact that “odd” results can sometimes lead to important findings.