Membrane‐Associated Nucleobase‐Functionalized β‐Peptides (β‐PNAs) Affecting Membrane Support and Lipid Composition

Abstract Protein‐membrane interactions are essential to maintain membrane integrity and control membrane morphology and composition. Cytoskeletal proteins in particular are known to interact to a high degree with lipid bilayers and to line the cytoplasmic side of the plasma membrane with an extensive network structure. In order to gain a better mechanistical understanding of the protein–membrane interplay and possible membrane signaling, we started to develop a model system based on β‐peptide nucleic acids (β‐PNAs). These β‐peptides are known to form stable hydrogen‐bonded aggregates due to their helical secondary structure, which serve to pre‐organize the attached nucleobases. After optimization of the β‐PNA solid‐phase peptide synthesis and validation of helix formation, the ability of the novel β‐PNAs to dimerize and interact with lipid bilayers was investigated by both fluorescence and circular dichroism spectroscopy. It was shown that duplex formation occurs rapidly and with high specificity and could also be detected on the surfaces of the lipid bilayers. Hereby, the potential of a β‐PNA‐based peptide system to mimic membrane‐associated protein networks could be demonstrated.


Reactions
Air or moisture sensitive reactions were carried out under inert argon or nitrogen atmosphere and in dry solvents. All utilized glassware was heated under reduced pressure for drying beforehand and flushed with inert gas prior to use.

HPLC
RP-HPLC for analysis and purification of samples was performed using an HPLC system from JASCO (Tokyo, Japan), consisting of a diode array MD-2010plus, degasser DG-2080-53 and two PU-2080plus pumps. Optionally, a fraction collector CHF122SC from Advantec (Milpitas, USA) was coupled to the HPLC system. As solvents H2O + 0.1 % (solvent A) and acetonitrile + 0.1 % TFA (solvent B) were used with s flow rate of 1 mL/min for analytical runs and 3 mL/min for semipreperative runs. RP-C18-ec Nucleodur columns from Macherey-Nagel (Düren, Germany) were used (particle size: 5 μm, pore size: 100 Å, column length: 250 mm, column diameter: 4.6 mm (analytical), 10 mm (semipreparative)). Peptides were dissolved in H2O/MeCN or HFIP and filtered before HPLC and UV-detection was performed at 215 nm, 245 nm and 280 nm for non-labelled β-peptides. For labeled β-peptides, UVabsorption was recorded at 464 nm for NBD and 540 nm for TAMRA instead of 280 nm.

Mass Spectroscopy
Electrospray ionization (ESI-MS) and high-resolution ESI (HR-MS) spectra were recorded with a maXis or MicroTOF spectrometer from Bruker Daltonik GmbH (Bremen, Germany). The values are given in m/z.

UV-Spectroscopy
Determination of β-peptide concentrations was conducted using a nanodrop ND-2000c spectrophotometer from Thermo Scientific (Munich, Germany) either by using the implemented pedestal method or with a Quartz SUPRASIL QS cuvette of 1.0 cm path length from Hellma Analytics (Müllheim, Germany). Concentrations were calculated from the absorbance using the Lambert-Beer's law.

CD-Spectroscopy
CD spectroscopy was performed with a J-1500 CD spectrometer from Jasco (Tokyo, Japan) and an F250 recirculation cooler from Julabo (Seelbach, Germany). Measurements were performed with a Quartz SUPRASIL QS cuvette with a 0.1 cm path length from Hellma Analytics (Müllheim, Germany). During measurements, the device and sample cell were flushed with nitrogen. The spectra were recorded in a wavelength range from 180 nm to 350 nm with a bandwidth of 1.0 nm, a data pitch of 0.5 nm, a response time of 1.0 s, a scanning speed of 100 nm/min and a CD scale of 200 meg/1.0 dOD in 'continuous mode'. Measurements were performed with the indicated temperatures and in the indicated solvents. An accumulation of five spectra was recorded per sample and background-corrected against pure solvent without β-peptide. Afterwards, the spectra were expressed as molar ellipticity [θ] (deg × cm2 × mol−1) according to GREENFIELD. [1] For temperature-dependent CD measurements, Samples were heated with a heating rate of 1 °C/min from 5 °C to 95 °C. Data points during heating and cooling cycles were recorded with a sampling rate of 0.5 °C, 1 s wait time, D.I.T. of 2 s, CD scale of 200 mdeg/1.0 dOD and bandwidth of 1.0 nm at 273 nm wavelength. Melting temperatures for the dimerizing β-peptides were determined by the peak maximum of the first derivation of the fitted melting curves. Additionally, spectra scans were performed during heating at 5 °C, 20 °C, 40 °C, 60 °C, 80 °C and 95 °C in a wavelength range from 180 nm to 320 nm with a bandwidth of 1.0 nm, a scanning speed of 100 nm/min, a data pitch of 0.5 nm, a D.I.T. of 2 s and CD scale of 200 meg/1.0 dOD in 'continuous mode'. An average of five spectra was recorded per sample and background-corrected against pure solvent without β-peptides. Afterwards, the spectra were expressed as molar ellipticity as described before.

Fluorescence Spectroscopy
Concentration-dependent fluorescence spectra and FRET assays between NBD-labeled and TAMRAlabeled β-peptides were measured with an FP-6200 spectrometer from Jasco (Seelbach, Germany) with an ETC-272T temperature controller from Jasco (Seelbach, Germany) and a WKL26 recirculation cooler from HAAKE (Karlsruhe, Germany). Measurements were conducted at the indicated temperatures as well as solvents and with a Quartz SUPRASIL QS fluorescence cuvette of 1.0 cm path length from Hellma Analytics (Müllheim, Germany). While the TAMRA-labeled β-peptide concentration was varied with the mole fraction ΧA varying from 0.0 to 0.5, the NBD-labeled β-peptide concentration with 4 μM as well as the total β-peptide concentration with 8 μM was kept constant by addition of the corresponding acetylated β-peptide. Emission spectra were recorded in a wavelength range from 470 nm to 650 nm with excitation at 460 nm, bandwidth of 5 nm, response set to 'fast', scanning speed of 125 nm/min, data pitch of 1.0 nm and sensitivity set to 'medium'.
Time-resolved measurements were started with Tris-HCl buffer (5 mM, pH 7.5) which contained either 1 mM DMPC/DHPC (q=2) or 0.3 mM DHPC or no lipids at all in a cuvette with a stirrer at 20 °C. After 120 s the NBD-labeled β-peptide was added to yield a concentration of 0.5 µM. The NBD fluorescence emission was recorded for 180 s before the TAMRA-labeled β-peptide (0.5 µM) or buffer was added and the fluorescence emission was further recorded for 600 s. Excitation was set to 460 nm and the fluorescence intensity was detected at 530 nm with a bandwidth of 5 nm, a data pitch of 10 s, response set to 'fast', measure time set to 900 s and sensitivity set to 'high'.
FRET assays to investigate membrane interaction of the β-peptides were performed with a Clariostar plate reader from BMG Labtech (Ortenberg, Germany) at room temperature in a black pp 96-well Fbottom plate from Greiner Bio-One (Kremsmünster, Austria) with a total sample volume of 200 μL. NBDlabelled β-peptides and Rhodamine-labelled LUV (0.75 %) were mixed right before measurement at a P/L ratio of 1:150 in Tris-HCl buffer (10 mM, pH 7.5). After focal and gain adjustment, emission spectra were recorded with an excitation wavelength of 460 nm in a range of 480-660 nm while the bandwidth was set to 10 nm. [2] LUVs were prepared according to the following protocol modified from MACDONALDS et. al.: [2] The required lipids were dissolved in CHCl3 on ice and the required amount was transferred to small glass test tubes. The solvent was removed with a nitrogen stream yielding a clear lipid film on the inner test tube walls. The lipid films were dried overnight under reduced pressure at 50 °C. Then, the lipid films were rehydrated in up to 1 ml filtrated buffer by 1 h incubation either at 40 °C for DMPC lipids or at rt for DOPC lipids. Subsequent vortexing for 30 s followed by 5 min incubation in three cycles yielded a milky MLV suspensions. The suspensions were loaded into a LiposoFast extruder with gas-tight syringes from Avestin (Ottawa, Canada). The vesicle suspensions were extruded 31 times through polycarbonate membranes with a pore size of 100 nm to yield a clear LUV suspension. [3] DMPC/DHPC bicelles were prepared according to the following protocol modified from DEANGELIS et. al.: [3] The required amounts of DMPC dissolved in CHCl3 were transferred to small glass test tubes. The solvent was removed with a nitrogen stream resulting in clear lipid films on the inner test tube walls. The lipid films were dried overnight at 50 °C and under reduced pressure. The lipid films were dissolved in 20 mM DHPC in Tris-HCl buffer (5 mM, pH 7.5) to yield a DMPC concentration of 40 mM and a qDMPC/DHPC of 2. Subsequent vortexing for 30 s, incubation on ice for 5 min and heating to 42 °C for 10 min in three cycles yielded a clear bicelle solution. The bicelle solution was stored on ice and diluted shortly before measurements to keep it stable.
Cleavage of the β-peptides was performed by addition of the cleavage mixture of TFA/mcresol/tioanisole/EDT/TFMSA (10:1:1:0.5:1 v/v) to the resin. First the cleavage solution was added without TFMSA and incubated for 5 min on ice before TFMSA was added drop-wise to the icecold mixture. After incubation for 1 h on ice and 2 at room temperature on a shaker, the cleavage solution was filtered from the resin and collected. After concentrating the solution in a nitrogen stream, the crude β-peptide was precipitated with ice-cold diethyl ether and centrifuged for 20 min at -15 °C with 9000 rpm.
The supernatant was removed and the pellet was washed three times with ice-cold ether and dried. Afterwards, the crude β-peptides were purified via HPLC.

Analytical Data of the β-Peptides
Peptide 1 HPLC λ in nm: 215,245,280) Figure S1. CD spectra of the indicated β-peptides at 20 °C in 10 mM Tris-HCl puffer at pH 7.5.
Measurements were performed in 10 mM Tris-HCl buffer at pH 7.5 and at 10 °C.  Figure S6. CD spectra of 1 and 3 (a), 5 and 7 (b) as well as 9 and 11 (c) measured at 20 °C separately, their calculated average and measured together with and without annealing at 80 °C. Temperature dependence of the CD spectra at 273 nm of the combinations 9+11, 5+7 and 1+3 measured together and separately (d). Measurements were performed in 10 mM Tris-HCl puffer at pH 7.5 (n=3).

CD Spectra of the β-Peptide Interaction on Lipid Bilayers
a) b) Figure S9. CD spectra of β-PNA interaction on the surface of DMPC/DHPC bilayers (q=2). 13 measured either together with 15 additionally with separate measurements and calculated average (a) or with 11 additionally with separate measurements and calculated average (b). Measurements were performed at 20 °C in 5 mM Tris-HCl buffer at pH 7.5 (n=3). Figure S10. Mean size distribution plot determined by DLS of an LUV suspension with a mean hydrodynamic diameter of 133.8 nm prepared by extrusion through a polycarbonate membrane with 100~nm pore size. Measurements were performed at 25 °C in 10 mM TRIS-HCl buffer (pH 7.5).

Characterization of LUVs and bicelles by DLS
Figure S11. Mean size distribution plot determined by DLS of a bicelle suspension with a mean hydrodynamic diameter of 30.35 nm and 332.2 nm for the different size populations prepared from DMPC and DHPC with a q of 2. Measurements were performed at 25 °C in 10 mM TRIS-HCl buffer (pH 7.5).