From literature [9] and our own experiments, we know that the folded OmpA TM domain does not unfold at all at 50°C. Increasing the temperature further from 50°C to 99°C, the OmpA TM domain unfolds and the intact fusion (HMW band) shifts to its

expected molecular weight of 49 kDa. These results demonstrate that the OmpA TM domain MLN0128 remains heat-modifiable and therefore is correctly assembled into the OM when mCherry is fused to its C-terminus. With increasing exposure to heat, the initially faint LMW (degradation) band also increased in intensity, and displays the exact same heat-modifiability behavior as the intact fusion between the OmpA β-barrel and mCherry. Because we know that mCherry does not exhibit heat-modifiability, the degradation band must see more consist of the OmpA β-barrel with (based on a MW of 28 kDa and assuming C-terminal degradation) the N-terminal part of mCherry (~55 residues), which appears to contain the epitope recognized by the monoclonal antibody. We conclude that cells expressing OmpA-177-SA-1-mCherry contain a mixture of intact fusion assembled

in the OM, and OmpA-177-SA-1 with a C-terminal part of mCherry proteolytically removed. Assuming C-terminal degradation, the removed part then contains the chromophore [30], and therefore this would represent a dark sub-population of OmpA TM domain in the OM. For the full-length OmpA-mCherry fusion (pGI10), we already knew that the full-length OmpA with C -terminal linker, but without mCherry (pGI9), was inserted properly in the OM [10]. Therefore, we only checked that the mCherry fluorescence was associated with Adavosertib mouse the PG/OM layer by fluorescence microscopy of plasmolyzed cells (Figure 2) [31]. This was indeed the case. FRAP results on cytoplasmic mCherry To maximize the likelihood of observing OmpA mobility, we avoided the cell poles (poles contain

inert PG and retain some OM proteins [7]) and performed the FRAP experiments in the cylindrical part of elongated cells. To create elongated cells (filaments) we grew the cells in the presence of the antibiotic cephalexin which blocks cell division but allows further elongation [11, 12]. The effect of cephalexin on bacterial cells is well-known: it binds with high affinity to PBP3, interfering with its ability to function in cell division. In addition, it has recently been shown that PBP3 ALOX15 only interacts with PBP2 (part of the protein complex responsible for elongation) during division at mid-cell [32]. We expect therefore that the structure of the cell wall in filaments will be highly similar to that of normal length cells. We tested our setup by starting with cells expressing cytoplasmic mCherry, which should give a recovery rate similar to that observed for cytoplasmic GFP, for which diffusion constants of 6–9 μm2/s are reported [11, 12]. The average length scale that corresponds with such a diffusion constant is = 2–3 μm when t = 0.5 s.