Length-Dependent Formation of Transmembrane Pores by 310-Helical α-Aminoisobutyric Acid Foldamers

The synthetic biology toolbox lacks extendable and conformationally controllable yet easy-to-synthesize building blocks that are long enough to span membranes. To meet this need, an iterative synthesis of α-aminoisobutyric acid (Aib) oligomers was used to create a library of homologous rigid-rod 310-helical foldamers, which have incrementally increasing lengths and functionalizable N- and C-termini. This library was used to probe the inter-relationship of foldamer length, self-association strength, and ionophoric ability, which is poorly understood. Although foldamer self-association in nonpolar chloroform increased with length, with a ∼14-fold increase in dimerization constant from Aib6 to Aib11, ionophoric activity in bilayers showed a stronger length dependence, with the observed rate constant for Aib11 ∼70-fold greater than that of Aib6. The strongest ionophoric activity was observed for foldamers with >10 Aib residues, which have end-to-end distances greater than the hydrophobic width of the bilayers used (∼2.8 nm); X-ray crystallography showed that Aib11 is 2.93 nm long. These studies suggest that being long enough to span the membrane is more important for good ionophoric activity than strong self-association in the bilayer. Planar bilayer conductance measurements showed that Aib11 and Aib13, but not Aib7, could form pores. This pore-forming behavior is strong evidence that Aibm (m ≥ 10) building blocks can span bilayers.


X-Ray Structures for Compounds 3, 4 and 7 S27
Table S1: Crystal data and structure refinement N 3 Aib 7 O(CH 2 ) 2 TMS S27     To a solution of N 3 Aib n O(CH 2 ) 2 TMS (1 eq) in MeOH or EtOH was added Pd/C (10%) and the resultant mixture was stirred under a H 2 atmosphere until completion of the reaction (TLC monitoring). After this time, the reaction mixture was filtered through a pad of Celite® under vacuum and the filtrate collected and the product isolated as an off-white solid.

HCl·HAibO(CH 2 ) 2 TMS (22)
To a solution of N 3 AibOH (1.29 g, 10.0 mmol) and 2-(trimethylsilyl)ethanol (1.72 mL, 12.0 mmol) in CH 2 Cl 2 (50 mL) were successively added EDC·HCl (2.49 g, 13.0 mmol) and DMAP (122 mg, 1.00 mmol). The mixture was stirred at room temperature for 18 h and was diluted with CH 2 Cl 2 (100 mL). The organic layer was successively washed with 5% KHSO 4 (2 × 50 mL), sat. NaHCO 3 (2 × 50 mL) and brine (1 × 50 mL) and dried over MgSO 4 . After evaporation of the solvent under reduced pressure, the crude was dissolved in MeOH (10 mL) and Pd/C (200 mg) was added. The mixture was stirred under H 2 atmosphere until completion of the reaction (TLC monitoring) and was then filtered on Celite® (eluent EtOAc). After evaporation of the solvent under reduced pressure, the residue was dissolved in Et 2 O (10 mL) and 2 M HCl in Et 2 O (5 mL) was added dropwise. After 5 min stirring at room temperature, the solvent was evaporated under reduced pressure and pentane was added.

Plots of Changes in δ(NH) for Different Foldamers upon Dilution in CDCl 3
In the following dilution studies the protons are labelled alphabetically from high to low field.
Typically the N-terminal NH proton, which is not tied into intramolecular hydrogen bonds, has the smallest chemical shift and experiences the greatest change in chemical shift upon dilution. 5 The other NH resonances remain unassigned.

Iterative Curve Fitting of Dilution Data to a Dimerization Model Theory
The spreadsheet macro developed by Sanderson 6 fits data to binding isotherms for the following equilibrium, using the Solver add-in for Excel (GRG Nonlinear Solving Method).
The values obtained from this macro were verified using the same set of equations, but minimizing in SigmaPlot (Levenberg-Marquardt algorithm).
where A is the Aib foldamer, K is the microscopic association constant and n = 2.
The following equation was used to calculate the binding isotherm: where: δ free is the free chemical shift of the host (Aib foldamer) The concentration of the dimeric complex was calculated iteratively by solving the following equations: The concentration of free host was calculated according to the concentration difference: Graphs S7-S12 show the curve fits using the Solver add-in for each studied Aib foldamer (host A in the previous equation). The data fitted are for the NH resonances most affected by S20 dilution, typically H a and H b , which allows a consensus value for K to be reached (reported in

Ionophoric Activity Experimental Section
Instrumentation: Fluorescence spectroscopy was carried out on a Perkin-Elmer LS55 fluorimeter. Temperature control was attained using a Julabo F25-HE water circulator. UVvisible spectroscopy was carried out using a Jasco V-600 spectrophotometer with the temperature controlled by a Jasco EHC-716 Peltier.

Preparation of Large Unilamellar Vesicles
Egg where F 0 = F t at addition of base pulse, F ∞ = F t at saturation after complete leakage, and I n = normalized fluorescence intensity.

Procedure for the Determination of First Order Rate Constants
The normalized fluorescence data (I n ) was iteratively fitted to first order kinetics using an equation of the general form: The baseline rate constant (methanol only, no foldamer) was then subtracted to give the final apparent rate constant value.

Sodium Ion Transport Rates: Concentration Studies in EYPC
For all sodium ion transport experiments, the HPTS assays were repeated several times and showed good experimental reproducibility (Figures S1 to S5). The data for N 3 Aib 8 O(CH 2 ) 2 TMS ( Figure S1) was fitted several times to pseudo-first order kinetics for each concentration, and the goodness of fit assessed for each fit. This process provided an approximation of the errors inherent in the curve fitting process, estimated as ±0.0003 s −1 .

CF Release Experimental Procedure
The extent of release of 5/6-carboxyfluorescein from the phospholipid vesicles was measured by observing the 5/6-CF emission intensity at 517 nm following excitation at 492 nm.
EYPC-cholesterol (4:1) 800 nm vesicles were prepared as described previously except with 1.  To verify that the bilayers were successfully formed during experiments, the membrane capacitance was measured and a short high-voltage spike was applied to rupture the membrane. Lipid bilayers that could be ruptured by voltage-spikes gave consistent capacitance readings in the range of 60-70 pF, before and after multiple rounds of 'zapping' and reforming. Membranes with lower capacitance values that could not be zapped were likely to be multilayers rather than the more fragile lipid bilayers and were not used during the experiments. Furthermore, the addition of α-hemolysin (Sigma-Aldrich) to the ground-wellside of membranes with capacitances in the range of 60-70 pF, were seen to result in the formation of transmembrane channels within minutes. Since α-hemolysin is unable to form transmembrane channels in multilayer membranes, this confirmed that true bilayers of the appropriate thickness were successfully being formed.
An Axopatch 200B amplifier, a Digidata 1322A, and Clampex 10.4 software (Molecular Devices Corp., CA, USA) were used to apply potential differences, and to S53 measure membrane currents and capacitances. Transmembrane currents were recorded over 10 mV voltage steps from −100 to +100 mV. Current recordings were stopped during the addition of compounds to the wells, and were filtered at 5 kHz and sampled at 50 kHz.      Contact between the droplet and the hydrogel results in the formation of a bilayer.

Droplet interface bilayers (DIB) studies
A pair of silver/silver chloride electrodes, one inserted into the device and one into the droplet from above, facilitates electrical access. An electrical protocol is applied to the electrodes by a patch-clamp amplifier, which simultaneously records the current flowing in the system. It exposes the bilayer to a sequentially increasing potential, rising in 5 mV steps at both positive and negative voltages. Potentials were applied for 180 s, after which the potential is switched to 0 mV for a duration 30 s. The purpose of this 'rest period' is to allow the bilayer to relax to its unperturbed state. The protocol is run for up to 8 cycles, until the potential reaches (±) 40 mV. Figure S11: DIB studies, three runs to ±40 mV in (±5 mV increments) applied to a single 3:1 POPE:POPG bilayer containing Aib 6 (2, 1 nM in droplet).

U-tube metal picrate transport experiments
A glass U-tube (10 mm internal diameter) was incubated in a water bath at 25 °C. Addition of the source phase marked the start of the experiment: aliquots (1 mL) were taken from the receiving phase and analyzed for the presence of picrate by UV spectroscopy (at 356 nm). After measurement, the sample was immediately replaced back in the U-tube.
Measurements were then taken every hour for 4 h, with a final sample taken after 20 h.

Circular dichroism (CD) studies of 7 in vesicles:
The peptaibols alamethicin and Aib-rich peptaibol ampullosporin A have both been shown to adopt mixed α/3 10 helical conformations as found in solution -when embedded in a bilayer. 7 It is likely that Aib m foldamers 1-9 also maintain their 3 10 helical conformation found in solution when embedded in the membrane (as shown in Figure S14). Measuring the circular dichroism (CD) spectra for these foldamers, which exist as racemic mixtures of screw-senses, will not give information on this helical conformation in the bilayer unless a helical excess is induced by the bilayer phospholipids.
The influence of phospholipid chirality on the organization of components in bilayers is generally very weak, especially in the fluid phase. 8 However, although not in a bilayer, there is a recent report of induced CD from a racemic Aib 8 foldamer in micelles composed of Ndodecylproline, a surfactant designed to interact with Aib foldamers. 9 This induced CD was consistent with a biased population of 3 10 helices in the surfactant micelle, and gave residue molar ellipticity values (θ R ) of 1200 deg cm 2 dmol -1 (at 208 nm) and 540 deg cm 2 dmol -1 (at 222 nm), around 4-6 times weaker than analogous values in MeOH solution.
Therefore the CD spectrum of Aib 11 7 in sonicated EYPC/cholesterol vesicles (1 mM 7, 5 mol % loading in EYPC/cholesterol SUVs of ca. 25 nm diameter) was measured, but strong background absorption due to scattering from the SUVs was a problem. After subtracting the background signal due to the phospholipid, we only found a weak CD signal at 222 nm above a very noisy background, corresponding to a θ R of <400 deg cm 2 dmol -1 (at 222 nm).
However due to the strong background scatter from the vesicles, we do not feel confident that these data clearly show induction of chirality. To unambiguously show global 3 10 helical conformations are adopted by Aib m foldamers in vesicle bilayers, either through covalent S59 control from one terminus or by induction by the surrounding phospholipids, long wavelength (λ max > 350 nm) CD reporter groups need to added to the termini of these foldamers.
Model of foldamer 7 in a 3 10 helical conformation in a bilayer: 1-Palmitoyl-2-oleoyl-snglycero-3-phosphocholine (POPC) was chosen as the lipid for this idealized bilayer as POPC bilayers have a similar thickness to that of EYPC, which is a mixture of lipids with different chain lengths.