Characterisation of Hybrid Polymersome Vesicles Containing the Efflux Pumps NaAtm1 or P-Glycoprotein

Investigative systems for purified membrane transporters are almost exclusively reliant on the use of phospholipid vesicles or liposomes. Liposomes provide an environment to support protein function; however, they also have numerous drawbacks and should not be considered as a “one-size fits all” system. The use of artificial vesicles comprising block co-polymers (polymersomes) offers considerable advantages in terms of structural stability; provision of sufficient lateral pressure; and low passive permeability, which is a particular issue for transport assays using hydrophobic compounds. The present investigation demonstrates strategies to reconstitute ATP binding cassette (ABC) transporters into hybrid vesicles combining phospholipids and the block co-polymer poly (butadiene)-poly (ethylene oxide). Two efflux pumps were chosen; namely the Novosphingobium aromaticivorans Atm1 protein and human P-glycoprotein (Pgp). Polymersomes were generated with one of two lipid partners, either purified palmitoyl-oleoyl-phosphatidylcholine, or a mixture of crude E. coli lipid extract and cholesterol. Hybrid polymersomes were characterised for size, structural homogeneity, stability to detergents, and permeability. Two transporters, NaAtm1 and P-gp, were successfully reconstituted into pre-formed and surfactant-destabilised hybrid polymersomes using a detergent adsorption strategy. Reconstitution of both proteins was confirmed by density gradient centrifugation and the hybrid polymersomes supported substrate dependent ATPase activity of both transporters. The hybrid polymersomes also displayed low passive permeability to a fluorescent probe (calcein acetomethoxyl-ester (C-AM)) and offer the potential for quantitative measurements of transport activity for hydrophobic compounds.

of the absorbance and required concentrations above 1mM TX-100 to achieve complete solubilisation.
These observations demonstrate that inclusion of polymer provides the vesicles with considerable resistance to surfactant mediated solubilisation.
Panels (f-j) demonstrate the effects of switching the lipid component of vesicles to the more complex crude mixture EcCL. Once again, vesicles with 0-25 mol% PBd-PEO displayed biphasic solubilisation profiles; however, where the proportion of polymer was 50 mol% or greater, a monophasic solubilisation profile was observed. Increasing the proportion of PBd-PEO from 50 (Csat = 0.62mM) to 75 (Csat = 0.85mM) and 100mol % (3.04mM) progressively increased the amount of TX-100 to saturate vesicles. Similarly, higher amounts of TX-100 were required to fully solubilise vesicles with 100% PBd-PEO (Csol = 9.7mM), compared to 75% (Csol = 7.4mM) and 50% (Csol = 5.6mM). Finally, an increase in the slope of the solubilisation profiles was observed as the polymer proportion increased.
Consequently, there are only marginal quantitative differences in the interaction between TX-100 and hybrid vesicles containing POPC or EcCL lipids.
The next series of solubilisation profiles ( Figure S2) for the two distinct lipid species in vesicles with PBd-PEO was done with the mild non-ionic alkyl-maltoside detergent DDM.
Panels (a-e) reveal the solubilisation profiles for hybrid PBd-PEO vesicles containing varying amounts of POPC. POPC vesicles with no polymer (panel a) displayed an initial increase in the optical density, which has been suggested to indicate vesicle coalescence.
Inclusion of 25mol% PBd-PEO revealed a complex solubilisation that also contained a segment associated with an increase in optical density. Vesicles with 50mol%, or greater, PBd-PEO displayed complex biphasic solubilisation profiles. The vesicles comprising 100mol% polymer revealed a shallow solubilisation profile that occurred over a broad range of DDM concentrations. Consequently, it was particularly difficult to accurately define the point of vesicle saturation with DDM.
Panels (f-j) displayed the solubilisation profiles for vesicles comprising the crude EcCL lipid mixture in conjunction with PBd-PEO. Vesicles composed solely of EcCL did not display the increase in optical density observed with POPC lipids, which has also previously been demonstrated. Vesicles with 25 and 50mol% PBd-PEO (panels g-h) were characterised by shallow sigmoidal and monophasic profiles. At higher proportions of PBd-PEO (panels i-j), the profiles became biphasic with a decidedly shallow initial phase. Similar to the observations with POPC containing vesicles, precisely assigning a saturating concentration of DDM was difficult.
Overall, the data in Figures S1-S2 demonstrated that the nature of the lipid species did not greatly affect the solubilisation profiles by detergent. In contrast, the detergent species was associated with a significantly greater effect. Clearly, chemical and physical properties of detergent will impact on its interaction with polymersomes.

Solubilisation of two classes of hybrid polymersome by Triton X-100
Vesicles containing varying proportions of PBd-PEO and either POPC (a-e) or crude EcCL (f-j) lipids were solubilised by the detergent Triton X-100 (10 -6 to 10 -1.5 M). Percentage proportions of polymer are indicated on each graph. Absorbance (optical density) was measured at 500nm in a 1cm path length quartz cuvette. Detergent was added and signal monitored continuously. Data were normalised to the optical density in the absence of detergent.

Solubilisation of two classes of hybrid polymersome by dodecyl--maltoside
Vesicles containing varying proportions of PBd-PEO and either POPC (a-e) or crude EcCL (f-j) lipids were solubilised by the detergent DDM (10 -6 to 10 -1.5 M). Percentage proportions of polymer are indicated on each graph. Absorbance (optical density) was measured at 500nm in a 1cm path length quartz cuvette. Detergent was added and signal monitored continuously. Data were normalised to the optical density in the absence of detergent.