Functionality of membrane proteins overexpressed and purified from E. coli is highly dependent upon the strain

Overexpression of correctly folded membrane proteins is a fundamental prerequisite for functional and structural studies. One of the most commonly used expression systems for the production of membrane proteins is Escherichia coli. While misfolded proteins typically aggregate and form inclusions bodies, membrane proteins that are addressed to the membrane and extractable by detergents are generally assumed to be properly folded. Accordingly, GFP fusion strategy is often used as a fluorescent proxy to monitor their expression and folding quality. Here we investigated the functionality of two different multidrug ABC transporters, the homodimer BmrA from Bacillus subtilis and the heterodimer PatA/PatB from Streptococcus pneumoniae, when produced in several E. coli strains with T7 expression system. Strikingly, while strong expression in the membrane of several strains could be achieved, we observed drastic differences in the functionality of these proteins. Moreover, we observed a general trend in which mild detergents mainly extract the population of active transporters, whereas a harsher detergent like Fos-choline 12 could solubilize transporters irrespective of their functionality. Our results suggest that the amount of T7 RNA polymerase transcripts may indirectly but notably impact the structure and activity of overexpressed membrane proteins, and advise caution when using GFP fusion strategy.

: purification yield (mg per L of culture) of PatA/PatB and BmrA transporters.

LMNG
FC12 C41(DE3)-PatA/PatB 6.3 12 BL21(DE3)-PatA/PatB 2.1 11.8 C41(DE3)-BmrA 6.0 6.8 T7 express-BmrA 0.9 3.4 Figure S1. Membrane of C41(DE3) strain containing overexpressed PatA/PatB. A Coomassiestained 12% SDS-PAGE was used to resolve the membrane fraction (Mb) of C41(DE3) strain expressing PatA/PatB, or not (empty vector control). On the right, PatA/PatB purified from Niaffinity was loaded to confirm the identity of the bands.    Figure S5. Doxorubicin transport assays with inverted membrane vesicles prepared from the various strains overexpressing BmrA-GFP. For the assays, 200 µg of total proteins were used and transport was monitored at 25 °C. Transport by membranes expressing BmrA-GFP is displayed as black traces, while transport by control membranes expressing the empty vector pET23 is displayed as orange traces. Figure S6. Quantification of GFP fluorescence in the membrane of the various strains. One hundred micrograms of total membranes were analyzed. Membranes expressing BmrA-GFP are displayed as black traces, while control membranes prepared with the empty vector pET23 are displayed as orange traces. Integrated fluorescence intensities were calculated between 500 and 530 nm. Background fluorescence calculated with control membranes represented respectively 13.5%, 11.6% and 10.6% of the total fluorescence in C41(DE3), BL21(DE3) and T7 exp, respectively.

Figure S7. Relative quantification of BmrA-GFP in the membrane of the various strains.
From the two gels, it was estimated that 5 µg of C41 membranes contained similar BmrA-GFP amounts than 3.5 µg of BL21 membranes or 3 µg of T7 express membranes.  Transport displayed by the C41(DE3) membranes was ∼3-4 times higher than the Lemo21(DE3) membranes prepared with 0.5 mM rhamnose. Western blot analysis was performed with 0.2 µg purified BmrA, 1 µg of membrane proteins (except for the lane with the *, which was loaded with 3 fold less membrane proteins to match the effective transport ratio). B, cells were induced in TB medium at 37 °C. Membrane fractions were analyzed by 12% SDS-PAGE. Transport rates were measured with 100 µg total membrane proteins. Transport displayed by the Lemo21(DE3) membranes prepared with 0.5 mM rhamnose was ∼3 times higher than the C41(DE3) membranes.
Western blot analysis was performed with 0.2 µg purified BmrA, 0.3 µg of membrane proteins (except for the lane with the *, which was loaded with 0.1 µg, i.e. 3 fold less membrane proteins to match the effective transport ratio). Note that the intensity of BmrA from C41(DE3) is intermediate between the intensities of the bands in the Lemo21(DE3) at 0.5 mM rhamnose (see the 2 lanes with 0.5 mM), suggesting that BmrA overexpression is ∼1.5 fold higher in this condition as compared to C41(DE3). C, cells were induced in LB medium at 37 °C. Membrane fractions were analyzed by 12% SDS-PAGE. Transport rates were measured with 100 µg total membrane proteins. Transport displayed by the Lemo21(DE3) membranes prepared with 0.5 mM rhamnose was ∼3 times higher than the C41(DE3) membranes. Western blot analysis was performed with 0.2 µg purified BmrA, 0.3 µg of membrane proteins (except for the lane with the *, which was loaded with 3 fold less membrane proteins to match the effective transport ratio).
Figure S10. Elution profile of PatA/PatB after Ni-affinity purification. A, purification in 0.02% LMNG. PatA/PatB were eluted using a gradient of imidazole (fractions shown between 35 and 50 mM, from left to right). B, purification in 0.3% FC12. PatA/PatB were eluted using a gradient of imidazole (fractions shown between 60 and 80 mM, from left to right).