Electrophoresis. 2017 Mar;38(5):738-746. doi: 10.1002/elps.201600381.

Phospholipid bilayer affinities and solvation characteristics by electrokinetic chromatography with a nanodisc pseudostationary phase.

PubMed link



Since the first publication on the use of nanodiscs in electrokinetic chromatography (EKC) our research group has focused on understanding the role that copolymer belt chemistry and the lipid bilayer chemistry play in the solvation characteristics and selectivity of the nanodiscs. These studies were conducted to investigate to what extent the interactions, solvation characteristics, and selectivity of the nanodiscs are a result of interactions with the lipid bilayer or of interactions with the copolymer belt which stabilizes the bilayer. In order to answer these questions nanodiscs with different copolymer belt to lipid ratios, different belt chemistries, and different lipid compositions were synthesized and investigated using EKC. The retention factors (a direct measure of partition coefficient) for a wide range of solutes were measured on nanodiscs of varied composition and were compared to octanol-water partition coefficients (Po/w) and between nanodiscs.


Table 1. Comparison of nanodiscs based on copolymer belt chemistry and lipid composition


Changes in the copolymer belt to lipid ratio were determined to have only minor effects on performance and selectivity. When the Xiran 30010 copolymer belt to lipid ratio was increased from 0.85:1.00 to 2.00:1.00 there was only a slight increase in the r2 of the linear relationship between the logarithm of the retention factors, log k, and logarithm of the Po/w, log Po/w. For 38 solute probes the r2 increased from 0.881 to 0.904 as seen in Table 1. Plots of log Po/w vs. log k for three belt to lipid ratios appear very similar (Figure 1). In order to determine if there were changes in selectivity of nanodisc interactions, the log k values analyzed using 1.00:1.00 and 2.00:1.00 Xiran 30010 copolymer belt to lipid ratios were graphed against each other. As presented in Table 2, the resulting r2 for all solutes was 0.997, for hydrogen bond donating solutes was 0.997 and for non hydrogen bond donors was 0.998. This demonstrates that there are no changes in the nanodisc-solute interactions when the copolymer to lipid ratio is increased. These results strongly suggest that interactions between solutes and the copolymer are not significant. If there were strong interactions between the solute probes and the copolymer portion of the nanodisc one would expect that the r2 value for the log k vs log Po/w plots would have decreased with larger copolymer belt to lipid ratio, and the correlation between log k values would decrease as the difference in belt to lipid ratio increases.


Table 2. Comparison of nanodisc interaction selectivity using correlation coefficients (r2) for plots of log k on nanodiscs with different Xiran 30010 belt to lipid ratios (w:w)


The retention, selectivity and solvation characteristics of nanodiscs with copolymer belts of two different chemistries were also studied. Xiran 25010 copolymer belts were introduced and compared with Xiran 30010 copolymer belts. Xiran 25010 is 10.0 kD and contains a 3:1 styrene: maleic acid anhydride ratio while Xiran 30010 is 6.0 kD and contains a 2:1 styrene to maleic acid anhydride ratio. Log k values on nanodiscs with 25010 copolymer belt were plotted vs. log Po/w values, resulting in an r2 of 0.941 as compared to an r2 value of 0.904 when using 30010 copolymer (Table 1). In order to determine if there also was a change in nanodisc selectivity the log k values for the Xiran 25010 and the Xiran 30010 to lipid ratio were graphed against each other and the r2 values are presented Table 3. The r2 for all solutes was 0.979, for hydrogen bond donating solutes was 0.937 and for non hydrogen bond donors was 0.998. These results show that the chemical composition of the belt plays a more important role in nanodisc-solute interactions, particularly for hydrogen bond donors, than the copolymer belt to lipid ratio. It remains unclear whether this indicates a change in lipid bilayer structure with the different belts or is evidence of interactions of hydrogen bond donors with the copolymer belt.


Table 3. Comparison of nanodisc interaction selectivity using correlation coefficients (r2) for plots of log k on nanodiscs with different Xiran copolymers



Lastly, the selectivity and solvation characteristics of four nanodiscs synthesized with different lipid compositions were compared to determine how changes in head and tail chemistry of the lipids affect nanodisc-solute interactions. Three of the nanodiscs contained uniform bilayers and one contained a mixed lipid bilayer. The lipids used were 1,2-dimyristoyl-sn-glycero-3-phosphocholine (14:0 PC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (16:0-18:1 PC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (16:0 PC), Sphingomyelin (16:0 SM), and 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine (14:0 PE). All nanodiscs contained a 2.5:1 Xiran 25010 copolymer belt to lipid ratio. The log k values were graphed against log Po/w for each of the nanodiscs and the results are presented in Table 1. 16:0-18:1 PC resulted in the closest representation of the Po/w values with an r2 value of 0.962. This could be because 16:0-18:1 PC contains a double bond and as a result is a more disordered bilayer, which maybe a better representation of octanol when compared to 14:0 PC which is a more ordered bilayer. An example of the separations with different nanodisc lipid chemistries can be seen in Figure 2, which illustrates minor changes in interaction selectivity as a function of lipid composition. Figure 2 also demonstrates that nanodiscs are capable of producing good peak symmetry and high separation efficiency, with an average efficiency from the 4 runs of 197,000±36,100 theoretical plates.


The analysis of nanodisc parameters show that variations in the ratio of copolymer belt to lipid used in the synthesis does not lead to significant changes in the r2 value for the indirect measurement of Po/w. Considering the concentration of copolymer was increased by more than 2x one would expect a large decrease in the r2 value for the indirect measurement of Po/w. This would suggest that the predominant interaction between the nanodiscs and solutes is with the lipid bilayer. However changes in belt chemistry between Xiran 30010 and Xiran 25010 did lead to changes in selectivity for hydrogen bond donors, suggesting that there are interactions between the copolymer belt and hydrogen bond donors or that changes in copolymer chemistry produce structural changes in the bilayer which lead to more favorable bilayer-hydrogen bond interactions. Lipid chemistry was also able to produce changes in the r2 values for the indirect measurement of Po/w providing further evidence that the interactions between nanodiscs and solutes is a result of bilayer interactions.



Figure 1. Plot of log Po/w vs log k for nanodiscs with different copolymer belt: lipid ratios


Figure 2. Separation of five solutes: (1) Benzonitrile, (2) p-Cresol, and (3) 4-Chloroaniline, (4) Butyrophenone, and (5) 3-Bromophenol. Separation parameters: 5 mM lipid nanodiscs with 2.50:1.00 (w:w) copolymer belt to lipid ratio, in 25 mM phosphate pH 7.0. Capillary dimensions: 48.5 cm x 50μm I.D. the top electropherogram utilized a 150μm extended cell pathlength. The injection was made with 35 mbar of pressure for 5 seconds. The operating voltage was 15 kV with detection at 225 nm.