WO2022195059A1 - Filtration of biomolecules using nanoporous membranes functionalized by electrografted polymers mimicking the nuclear pore - Google Patents

Abstract

The present invention concerns nanoporous functionalized membrane comprising a synthetic nanoporous membrane and electrografted polymers, its process of preparation and its uses. In particular, it relates to the use of a nanoporous functionalized membrane according to the invention to improve the filtration of biomacromolecules. It further relates to the use of electrografted polymers to functionalize a synthetic nanoporous membrane.

Description

Filtration of biomolecules using nanoporous membranes functionalized by electrog rafted polymers mimicking the nuclear pore

The present invention concerns a nanoporous functionalized membrane comprising a synthetic nanoporous membrane and electrog rafted polymers, its process of preparation and its uses. In particular, it relates to the use of a nanoporous functionalized membrane according to the invention to improve the filtration of biomacromolecules. It further relates to the use of electrog rafted polymers to functionalize a synthetic nanoporous membrane.

In nature as well as technology, compartment formation is one of the key components of complex and evolved systems. In eukaryotic cells, genetic material is stored inside the nucleus, which is separated from the cytoplasm in order to protect the genetic material inside the nucleus from outer influences. However, there are gateways between the nucleus and the cytoplasm, called nuclear pore complexes (NPCs). These highly selective pores act as passageways to ensure selective and directional transport of a wide range of cargo molecules (Jovanovic-Talisman and Zilman 2017; Kabachinski and Schwartz 2015).

In the context of the invention, the inventors used the NPCs as a model system for selective and directional transport at nanoscales. The first experiments examining the passage of individual molecules through nano-sized pores in membranes had been restricted to naturally occurring nanopores (Kasianowicz et al. 1996). Advances in technology then allowed artificial solid-state nanopores to be fabricated in insulating membranes. By finding new technological ways of mimicking these kind of gateways, nanopore devices emerge as a new powerful class of designable molecule sensors, passageways, and filter systems. Design of bio-inspired nanopores brings the ability to manipulate and control ion and molecule transport inside a confined environment. Thus, they can mimic ionic and molecular transport processes of biological channels (Dekker 2007; Caspi et al. 2008; Hriciga and Lehn 1983).

Synthetic nanopores have many advantages compared to biological ones, such as their stability, tunable dimensions (size and shape), and the possibility to integrate them in nanofluidic systems. By functionalizing the nanopore external and internal surfaces of the nanopore, their physical and chemical properties (e.g. hydrophobicity, selectivity, surface charges, specific molecular recognition) and thus the ionic transport properties are modified (Lepoitevin et al. 2017; Venkatesan et al. 2009; Dekker 2007).

Chemical mimics of the NPCs receptor-mediated transport mostly use modified nano porous membrane filters. As a grafting inside the nanopore several groups first tried polyisopropylacrylamide (pNIPAM), which is well-known for its intramolecular hydrogen bonding and for its hydrophobic-hydrophilic phase transition at physiological temperatures. Below the transition the polymer is soluble, whereas at elevated temperature, a collapsed phase appears. The mobile polymer thus acts as a soluble receptor to usher a macromolecular cargo specifically through the pores (Caspi et al. 2008). But also FG-nups from NPCs as well as individual biomimetic nuclear pore complexes are used to create nano-selectivity. These membranes efficiently pass transport factors and transport-factor- cargo complexes that specifically bind FG nups, while they significantly inhibit the passage of proteins that do not (Jovanovic-Talisman et al. 2009; Kowalczyk et al. 2011 ). Both types of functionalization inside the nanopore form polymer brushes whose conformation is strongly dependent on the solvent and other surrounding conditions. Thus, the grafting has a huge influence on the kinetics of translocation in nanopores.

The inventors of the present invention have designed a nanoporous functionalized membrane comprising a synthetic nanoporous membrane and electrog rafted polymers which, compared to existing technologies, makes it possible to combine geometric selectivity (size of molecules) and physico-chemical while maintaining high mass flows. With this membrane, the method of filtration is faster (hours vs days) than HPLC separating column chromatography and similar in cost to ultrafiltration membranes.

The membrane of the invention allows to maintain the size filtration properties of nanoporous membranes by adding a filtration capacity as a function of physicochemical criteria. Compared to existing methods, this invention makes it possible to maintain high mass fluxes by adding molecular selectivity and for a reduced cost.

As mentioned, the invention relates to a nanoporous functionalized membrane comprising a synthetic nanoporous membrane and electrog rafted polymers.

Said nanoporous functionalized membrane is characterized in that said polymers are grafted within the nanopores of said synthetic nanoporous membrane and in that the contour length of the polymers is greater than once the radius of the nanopores.

Said membrane is able to mimick the filtration properties of the natural pore and more precisely its physico-chemical selectivity. To do this, polymers were electrog rafted within membranes. Once grafted, these polymers form a cohesive network which fills the nanopores of the membrane. This network adds to the original artificial membranes the ability to filter and sort biomacromolecules according to their physicochemical properties. In addition, the thermodynamics properties of these grafted polymers allow this pore to be opened and closed by a change in temperature.

By “nanoporous functionalized membrane” is meant organic or inorganic solid membranes with nanometer-scale pore(s) grafted with polymers. By “synthetic nanoporous membrane” is meant organic or inorganic solid membranes with nanometer-scale pore(s).

By “electrografted polymers” is meant polymers that contain at least one electroactive function that allows the grafting of the polymer on a surface. More particularly, electrografted polymers are chosen from hydrophilic polymers. In some embodiments, hydrophilic polymers of the invention exhibit a lower critical solubility temperature (LCST).

By “lower critical solubility temperature (LCST)” is meant a critical temperature defined for a given polymer below which the polymer is hydrophilic and above which the polymer is hydrophobic.

LCST can be measured by well-known methods in the art. Turbidity can be cited as an example.

As mentioned above, the polymers according to the invention have a contour length greater than once the radius of the nanopores.

By “contour length” is meant the total physical length of the molecule. It can be measured by well-known methods in the art. Mass spectrometry can be cited as an example.

The counter length of the polymers is greater than once the radius of the nanopores, in particular greater than 100 monomers when poly-alkyl oxazoline polymers are contemplated.

The radius of the nanopores can be measured by routine methods of the skilled person. Scanning electron microscopy can be cited as an example of method.

In particular, the polymers according to the invention comprise between 2 and 1000 monomers, more particularly between 10 and 600.

Even more particularly, when radius of the nanopores is comprises between 50 and 200 nm, polymers according to the invention comprise between 100 and 600 momomers.By “hydrophilic polymers” is meant a polymer that is partially or fully soluble in aqueous solution, for example poly-alkyl oxazoline polymers for lower temperature than the LCST. By “partially soluble” is meant that mass parts of aqueous solution required to dissolve 1 mass part of that polymer can ranged from 1 to 1000.

By “hydrophobic polymers” is meant a polymer that is not soluble in aqueous solution, for example poly-alkyl oxazoline polymers for higher temperature than the LCST.

“Poly(2-alkyl-2-oxazoline) polymers” are well known in the art.

In particular by “alkyl” is meant an alkyl chain comprising between 1 and 4 carbons such as methyl, ethyl isopropyl or n-propyl. Poly(2-methyl-2-oxazoline, 2-n-propyl-2- oxazoline and poly(2-ethyl-2-oxazoline), can be cited as examples. In particular, poly(2- alkyl-2-oxazoline) polymers according to the invention comprise between 100 and 1000 monomers of (2-alkyl-2-oxazoline), more particularly between 100 and 600.

In the context of the invention, said polymers comprise at least one electroactive function. By “electro active function” is meant a function that reacts upon electrochemical process to provide a reactive radical. In particular, diazonium electro active probes can be cited.

In particular, in the context of the invention, the polymers of the invention are grafted within the nanopores of said synthetic nanoporous membrane thanks to this electroactive function.

More particularly, the electrografting according to the invention and the type of polymer used makes it possible to graft polymers which are temperature dependent and sparingly soluble with a sufficient density to create the desired network within the nanopores of the membrane.

More particularly, said nanopores mimick the selectivity of nuclear pore complexes in biological cells by selecting molecules on the basis of their physico-chemical parameters and in particular their hydrophobicity.

For example, the polymerization of 2-alkyl-2-oxazolines can be initiated by an electrophile, in particular allyl bromide, (scheme 1 ) (Guillaume Pereira, Cecile Huin, Simona Morariu, Veronique Bennevault-Celton, and Philippe Guegan " Synthesis of Poly (2-methyl- 2-oxazoline) Star Polymers with a beta-Cyclodextrin Core "Aust. J. Chem., 65, 1145-1155 (2012)). The use of a compound possessing a protected aniline function as an initiator has been reported (Waschinski, CJ & Tiller, JC Poly (oxazoline) s with Telechelic Antimicrobial Functions. Biomacromolecules 6, 235-243 (2005)), but the synthesis of the initiator is difficult to reproduce (thesis Dihia Benaoudia, Universite Paris Diderot, defended on November 29, 2018). The functionalization of poly (2-alkyl-2-oxazolines) can be done by reaction of the growing species, oxazolinium, with a nucleophilic compound such as amines (aliphatic, unsaturated or aromatic) (G. Delaitre Telechelic Poly (2oxazoline) s European Polymer Journal, 121 (2019) 109281). Scheme 1 : polymerization of 2-alkyl-2-oxazolines initiated by an electrophile (R can be any aliphatic and aromatic radical) The functionalization of growing poly 2-alkyl-2-oxazolines by a molecule having an aniline function available for subsequent reactions has never been reported (protected aniline function or not).

The invention proposes an efficient method of functionalization of growing poly (2- alkyl-2-oxazoline) in a single step by reaction of the oxazolinium ends with 1 - (4- aminophenyl) piperazine according to the specific reaction of piperazine on the oxazolinium function.

The originality is that it does not require a step of protection/deprotection of the aniline function, the amine function of the piperazine cycle being more reactive than the amine function of the aniline end (scheme 2).

It is therefore a one-step functionalization.

Scheme 2: specific functionalization reaction of the oxazolinium ends carried by a poly (2-alkyl-2-oxazoline) chain growing by 1- (4-aminophenyl) piperazine The poly (2-alkyl-2-oxazolines) obtained have an aminophenyl functionality of between 95 and 100%, and the polymerization is controlled.

The functionalization reactions can be carried out on "growing species", oxazolinium, resulting from the polymerization of 2-methyl-2-oxazoline and 2-n-propyl-2-oxazoline, suggesting that this chemistry is not sensitive to the nature of the alkyl group of the 2-alkyl- 2-oxazoline monomer.

This method of functionalization can thus be extended to other 2-alkyl-2-oxazolines (ethyl, isopropyl, butyl, phenyl, etc.), the polymerization of which can be initiated by initiators of varying nature inducing a counterion other than bromide, as well as polymers resulting from the polymerization of 2-alkyl-2-oxazoline of variable architecture (star, hyperbranched, dendrimer, brush, crosslinked) and of copolymers of variable composition (block, graft, random, gradient, alternating).

As previously mentioned, the present invention also relates to a process of preparation of the nanoporous functionalized membrane according to the invention and described herein, wherein a synthetic nanoporous membrane is electrog rafted of polymers, said polymers being grafted within the nanopores of said synthetic nanoporous membrane and said polymers having a contour length greater than once the radius of the nanopores..

The above definition given for the nanoporous functionalized membrane also apply for the process.

In particular, the grafting is achieved by polarization of a metallic layer present on the surface of the nanopores.

By “metallic layer” is meant a layer of metal atoms sputtered or evaporated on the membrane. The thickness of this layer can be varied from 1 nm to 10 micrometers. In particular, gold and aluminum atoms can be cited.

The polarization can be for example conducted by applying a constant voltage bias to the membrane with a potentiostat.

In one embodiment, the grafting step is repeated more than once, for example between 1 and 100 times, in particular between 1 and 10 times.

As a consequence, the polarization process can be repeated more than once, for example for between 1 to 100 cycles, more particularly for between 1 to 10 cycles.

The present invention further relates to the use of a nanoporous functionalized membrane according to the invention and described herein, to improve the filtration of biomacromolecules. By “biomacromolecule” is meant a molecule of molar masse between 5 000 g/mol and 100000 g/mol and synthesized by a biological process. A typical biomacromolecule is DNA, RNA, polysaccharide, protein, glycoprotein...

In the context of the invention, the term “biomacromolecules” does not cover assemblies of molecules.

In particular, recombinant proteins such as insulin and ARN such as ARNm are contemplated.

The present invention also relates to the use of electrog rafted polymers to functionalize a synthetic nanoporous membrane by grafting the polymers within the nanopores of said synthetic nanoporous membrane, the contour length of the polymers being greater than once the radius of the nanopores.

In particular, the existence of a critical solubility temperature (LCST) associated with the grafted polymers leads to the existence of a phase transition. Below this temperature, the polymer exhibits an extended shape which leads to the formation of a cohesive network in the center of the pore. Above this temperature the polymer collapses on itself and the pore is opened. This reversible mechanism makes it possible to open and close the cohesive network and the nanopores on demand for being able to clean the membrane at the end of the filtration phase.

In particular, said polymers are able to open or close the nanopores of the membrane thanks to a stimuli, i.e. temperature, light, pH, osmotic pressure and more preferably temperature.

More particularly, it allows to clean the filter membranes by opening the nanopores via an increase in temperature.

The invention will be further illustrated by the following figures and examples.

Figures

Figure 1 : Zero mode waveguide setup for nanopore translocation with polymer grafting + pores with different polymer grafting

open, semi-open, close

Figure 2: Transport properties of Dextran: comparison between polymers with contour length shorter and longer than the pore radius

Figure 3: Transport properties of BSA (bovine serum albumin): comparison between polymers with contour length shorter and longer than the pore radius Examples

Example of preparation of polymers according to the invention

1) Polymerization of 2-methyl-2-oxazoline

The 2-methyl-2-oxazoline, allyl bromide and acetonitrile were dried under vacuum separately in the presence of a desiccant, preferably CaH₂. A succession of freezing/thawing cycles were carried out until a vacuum of about 10^-5 - 10^-6 hPa constant, preferably 5 cycles, is obtained. The 2-methyl-2-oxazoline, allyl bromide and acetonitrile were then freeze-distilled separately using a secondary vacuum manifold. Control of the quantity of allyl bromide to be introduced into the reactor was ensured by the preparation of a solution of allyl bromide (AIBr) in acetonitrile (AcN), advantageously 0.1 ml. of AIBr in 0.9 ml. of AcN. The 1- (4-aminophenyl) piperazine being solid was used without further purification. The Xn of polymers was adjusted by the molar ratio of 2-methyl-2-oxazoline to allyl bromide, and was set at about 60 in this case.

The reactor in which the polymerization was carried out is stored at 60°C. It was then evacuated and ignited on the vacuum manifold to remove all traces of moisture from the walls of the reactor. The reactor and all of the cryodistilled reagents were placed in a glove box under an inert atmosphere. The initiator was introduced into the reactor, advantageously 0.17 ml. of the 10% vol/vol solution in acetonitrile, 2-methyl-2-oxazoline, advantageously 1 ml_, then acetonitrile, advantageously 6.9 ml_. The closed reactor was taken out of the glove box and placed at 80°C. during the time of the polymerization, advantageously 24 hours.

At the same time, 175 mg of 1 - (4-aminophenyl) piperazine were dissolved in 5.6 mL of cryodistilled AcN. The preparations were done in a glove box.

At the end of polymerization, the reactor was brought to room temperature, introduced into a glove box and then the solution of 1 -(4-aminophenyl) piperazine in acetonitrile was added to the reaction medium. The amount added corresponded to 5 equivalents of deactivating agent relative to the initiator initially introduced into the reactor, so as to result in complete functionalization of the polymer chains. The reaction mixture was placed at room temperature for 66h. The solvent was then removed by evaporation, then the polymer was dialyzed for 24 hours against methanol using a regenerated cellulose dialysis rod having a cut-off of 1000 g.mol-1 . The methanol from the dialysis is changed 3 times in total over 24 hours. The solvent was then evaporated off, then the polymer placed in an oven at 60°C for 24 hours. The polymer was characterized by 1 H NMR to verify its purity, its degree of functionalization and its molar mass was determined by SEC (eluent DMF) using a PMMA calibration. The molar mass (SEC) of the polymer was 5150 g.mol-1 with a dispersity of 1 .33 and the degree of functionalization is 100%. 2) Polymerization of 2-n-DroDyl-2-oxazoline

The 2-n-propyl-2-oxazoline, allyl bromide and acetonitrile were vacuum dried separately in the presence of a desiccant, preferably CaH₂. A succession of freezing/thawing cycles were carried out until a vacuum of about 10^-5 - 10^-6 hPa constant, preferably 5 cycles, was obtained. The 2-n-propyl-2-oxazoline, allyl bromide and acetonitrile were then cryodistilled separately using a secondary vacuum manifold. Control of the amount of allyl bromide to be introduced into the reactor was ensured by preparing a solution of allyl bromide in acetonitrile, preferably 0.1 ml. of AIBr in 0.9 ml. of AcN. The 1 - (4-aminophenyl) piperazine being solid was used without further purification. The Xn of polymers was adjusted by the molar ratio of 2-n-propyl-2-oxazoline to allyl bromide, and is set at about 60 in this case.

The reactor in which the polymerization was carried out is stored at 60°C. It was then evacuated and ignited on the vacuum manifold to remove all traces of moisture from the walls of the reactor. The reactor and all of the cryodistilled reagents were placed in a glove box under an inert atmosphere. The initiator was introduced into the reactor, advantageously 0.12 ml. of a 10% vol/vol solution in acetonitrile, n-propyloxazoline, advantageously 1 ml. then acetonitrile, advantageously 4.6 ml_. The closed reactor was taken out of the glove box and placed at 80°C. during the time of the polymerization, advantageously 72 hours.

At the same time, 125 mg of 1 - (4-aminophenyl) piperazine were solubilized in 4.5 mL of cryodistilled AcN. The preparation was done in a glove box.

At the end of polymerization, the reactor was brought to room temperature, introduced into a glove box and then the solution of 1 - (4-aminophenyl) piperazine in acetonitrile was added to the reaction medium. The amount added corresponded to an excess of 5 equivalents of deactivating agent relative to the initiator initially introduced into the reactor, so as to result in a total functionalization of the polymer chains. The reaction mixture was placed at room temperature for 66h. The solvent was then removed by evaporation, then the polymer was dialyzed for 24 hours against methanol using a regenerated cellulose dialysis rod having a cut-off of 1000 g.mol-1 . The methanol from the dialysis was changed 3 times in total over 24 hours. The solvent was then evaporated off and the polymer placed in an oven at 60°C for 24 hours. The polymer was characterized by 1 H NMR to verify its purity, its degree of functionalization and its molar mass was determined by SEC (eluent DMF) using a PMMA calibration. The molar mass (SEC) of the polymer is 9000 g.mol-1 with a dispersity of 1 .08 and the degree of functionalization was 98%. The LCST of this polymer was 23°C. Example of preparation and use of a nanoporous functionalized membrane accordina to the invention

MATERIALS AND METHODS

Experimental set-up

Our experimental set-up combined optical detection and pressure control to induce the transport of polymers through a porous membrane following Auger et al. (Auger et al. 2018; Auger et al. 2014) (Figure 1 ).

In brief, the experimental set-up is a classical fluorescence microscopy experiment made of an inverted microscope (Axiovert 200) with a filter box - consisting of several pairs of excitation and emission filters and an associated dichroic mirror - connected to an electron multiplying charge coupled device camera (EMCCD camera, Andor, iXon 897). In direct image capture, it displays a maximum resolution of 512 x 512 pixels in 32 bits and a maximum frame rate of 60 Hz. On top of the microscope, a water objective (ZEISS C- Apochromat) with 63x magnification, 1.2 numerical aperture, and an operating distance of 0.28 mm for a 0.17 mm glass slide is mounted. As excitation source a laser (Cobolt blues 50, Cobolt AB), which emits at 473 nm wavelength was used. The beam was expanded by a telescope and followed an optical path so that the beam was parallel when leaving the lens. The membrane was illuminated from the gold covered side. A shutter was used to interrupt the laser beam when the illumination was not needed. The set-up exhibited a second light source, a fluorescent lamp (Uvico, Optoelectronics GmbH), which was connected directly to the microscope by an optical fiber. The two possible excitation wavelengths are 488 nm and 568 nm. This lamp can be used to get a better overview over the sample and to light the sample with or without heating it up, depending on the chosen excitation wavelength.

The heart of the set-up consists of two chambers connected by a porous membrane: the upper chamber (called the trans chamber) is made of a screw cap (nanion) which consists of a glass bottom with a small hole to which a piece of membrane is glued. This chamber is connected to a pressure control system (MFCS, Fluigent) which allows an application of pressures up to 1 bar with a resolution better than 0.1 mbar. Sealing is ensured by a Teflon gasket inside the cap. The lower chamber is also known as the cis chamber. It consists of a Teflon ring on which a 0.17 mm thick glass strip is glued. The set up furthermore features a smaller Teflon ring that is attached around the screw cap to hold it vertically in the cis chamber.

The translocated polymer - marked with a fluorescent dye - was placed in the trans chamber and by applying a pressure difference between the chambers, the polymer was transported through the membrane into the cis chamber. Since the membrane was illuminated by an extended laser beam on the side where the gold layer in grafted onto, the polymers were observed as they exit the pores. Molecules in the trans chamber and inside the pore were invisible because they were not in the illuminated volume. Eventually, the molecules' fluorescence disappears because of bleaching and optical defocusing as they moved away from the membrane.

This two-chamber construction is located above the lens of the fluorescence microscope. The sample can be moved in all three directions by means of a stage control system with a XY-drive accurate to the pm and a Z-drive accurate to the nm. The stage control system was operated by using the software ImageJ. The observation of the transported molecules was accomplished by a EMCCD camera connected operating software (Andor SOLIS for imaging).

The laser including its optical pathway, the fluorescence microscope, the two-chamber construction, and the EMCCD camera were located on a stable table. The stage control system, the pressure control system, and the shutter control system were placed next to it.

Membrane preparation and polymer grafting

The membranes containing the nanopores were pretreated. These commercially available track-edged membranes (Whatman) consisted of a 6 pm-thick polycarbonate layer and possess cylindric holes (pores) with a diameter of 50 nm and a density of 6 10⁸ pores/cm².

To visualize the translocation of molecules through the pores, we grafted a gold layer with a thickness of 50 nm on top of the membrane.

The gold was evaporated after a low intensity ionic pickling. Evaporation was achieved at a pressure below 10^-6 Pa, gold was deposited at a speed of 0.2 nm/s. The speed and thickness of the deposit were measured by a quartz balance. We chose the thickness of the gold layer to be 50 nm.

The polymer grafting on top of the gold layer was done by means of electro-grafting (connection between the polymers and the gold surface via azo-coupling)

Synthesis of Polymethyloxazoline (PMeOx) :

Table 1

The structure of the porous membrane remains unaffected during the polymer grafting on top of the membrane's gold layer, the only difference was the additional polymer layer on top of the gold coating. The polymers were relatively hydrophobic so in aqueous environment, they are in relative bad solvent. Therefore, the interaction between the polymers is strong and they tend to stay in contact with other polymers. The spatial elongation and other properties of this layer vary depending on the specific type of polymer grafted.

The n-propyl-2-oxazoline polymers exhibit a lower critical solution temperature (LOST). This means that in bulk below the LCST these polymers were fully soluble whereas at a temperature higher than the LCST they accumulate and precipitate. This transition temperature was routinely measured by recording the turbidity of a solution of polymers as a function of temperature. T < LOST: The system is completely miscible in all proportions. For the n-propyl-2- oxazoline polymers this means that they behave in solution like in a good solvent and are free to elongate. In a confined environment (e.g. inside a nanopore), the polymer strands were swelled and form a mesh connecting to each other. This leads to a decrease of the effective diameter of the pore, reducing or blocking the flux through the pore (Figure 1).

T > LOST. Partial liquid miscibility occurs meaning that in bulk, the polymer was collapsed as it tries to minimize its contact surface area with the surrounding solvent. Grafted inside a nanopore, the polymer strands are folded resulting in an increased effective diameter of the nanopore which in turn increases the flux through the pore (compared to T < LOST), the pore is open (Figure 1).

In a confined environment, this behavior is even more pronounced. Energetically speaking, it is very costly to confine polymers in a small tube. For T < LOST the polymers were filling the tube because it was favorable for them to be in interaction with each other since they are in a bad solvent. For DT > LOST the polymers were in a collapsed state since the interaction effects are winning over. Now the energy gained through the collapse was larger than the entropy loss (less degrees of freedom in the collapsed state).

The 2-methyl-2-oxazoline polymers do not exhibit an LOST. Therefore, their conformation inside the nanopore was independent of the surrounding temperature. The polymer strands behave like in an intermediate solvent (called theta-solvent). The effective diameter of the pore is reduced by the grafting but the pore is never fully closed i.e. semi open (Figure 1).

Sample preparation

For all experiments, we used a tris(hydroxymethyl)aminomethane (Tris) and ethylenediaminetetraacetic acid (EDTA) based buffer (TE buffer), made from T ris-HCI (Sigma-Aldrich) and EDTA (Sigma-Aldrich).

We used Lambda-DNA (Invitrogen) fluorescently labeled with YOYO-1 (Life Tech).

For the stationary heating experiments, we introduced 100 mI TE buffer mixed with 2 mI DNA-YOYO mix in the trans chamber and 100 mI TE buffer in the cis chamber. To apply specific temperatures, we made use of a temperature control system (Pecon, TempController 2000-2). This system allows us to heat a metal ring that is located around the objective. As the objective is connected by a drop of water to the cis chamber which in turn is filled with buffer that is in contact with the membrane, we have the possibility to indirectly heat up the membrane. The applied temperature at the level of the heater was chosen to be 50°C. After the temperature control system had reached this temperature, we waited for 60 minutes to make sure that the whole system was heated up to a fairly constant temperature. The temperature of the buffer in the cis chamber was checked to be higher than 30 °C using a thermometer with a long and flexible probe that is placed on the side of the inner edge of the cis chamber so that it is in contact with the liquid inside. Then, we started the experiment while the temperature control system was still heating up the system. For the first 5 minutes, we recorded a video every minute, then one video after 10 minutes and one video after 15 minutes. Thereafter, we switched off the heat supply and let the system cool down while taking one video every minute. With the acquisition of a video we also measured and noted the temperature of the liquid in the cis chamber.

In the scope of the varying pressure experiments, we introduced 100 mI TE buffer mixed with 2 mI DNA-YOYO mix in the trans chamber and 100 mI TE buffer in the cis chamber. The applied pressure was changed between 0 mbar and 150 mbar using the pressure control system. After setting a pressure value at the control system, we waited 3 minutes to let the system adopt. Then, we took four videos in a row at constant pressure. The measurements were either performed at room temperature or while heating the system as explained for the stationary heating experiment, but without changing the heat supply during the experiments.

Data analysis

Single molecule event detection

The analysis of the single molecule experiments was mainly about the interpretation of the recorded videos. To determine the translocation frequency, the translocation events appearing in the video were counted by hand. These were visible as bright spots on the otherwise dark background (the membrane is only slightly auto-fluorescent). The frequency f was then calculated as the number of events N observed during one video sequence divided by the number of pores present in one film NP times the acquisition time t = 15 s. NP can be calculated as the pore density p = 6 10⁸ pores/cm² times the observed area 150 pm x 150 pm:

For the data analysis, all plotting and fitting processes were done by means of MATLAB.

For the analysis of the stationary heating experiments, we plotted the number of translocation events as a function of time but also as a function of the measured temperature. As the temperature was taken at the edge of the cis chamber, the noted temperature differs from the actual temperature of the membrane, but we assumed that these two temperatures are not far from each other. The resulting curves were normalized by setting the starting frequency value to one. Subsequently, the temperature-frequency curves were fitted by an error function. For the varying pressure experiments, we took the mean value of the number of translocation events of the four videos recorded at the same pressure. This value was then plotted as a function of the applied pressure and fitted using the suction model. Since our control variable is the pressure and not the flow rate, we applied the Hagen-Poiseuille law to obtain a relationship between the translocation probability P_t and the ratio of p_cto p. The frequency of translocation f_trans can be written as

/irons = k x exp(-P_c/P) with k being constant, p the applied pressure, and p_c the critical pressure. By fitting this equation to our data points, we were able to obtain values for and p_c. Here, we were mainly interested in p_c, the minimum pressure that is required to transport the molecule through the pore.

RESULTS

Stationary heating experiment In order to examine the effects of temperature on the DNA translocation dynamics in nanopores coated with a polymer grafting, we systematically varied the temperature and recorded the change in translocation frequency through the pore for different polymer graftings. Since two of the polymers we use as a grafting possess a LOST, varying the temperature in a range including the LOST changes the conformation of the polymers in the grafted layer. In turn, this changes the inside of the pore and allowed us to switch between an open and a closed state of the nanopore depending on the temperature of the surrounding medium (Figure 1).

To characterize the kinetics of the system, we used a heating system that warms the whole membrane as described in the materials and methods section. The measurements were conducted for the two n-propyl-2-oxazoline polymer graftings, for one exemplary 2-methyl- 2-oxzoline polymer grafting (methyl-long), and for a membrane without polymer grafting as a negative control.

Due to the way of quantifying the temperature in this measurement, the exact implementation varied from experiment to experiment: if the probe was placed closer to the center of the membrane, the measured temperature was slightly higher than if it was placed further outside and therefore closer to the environment. Anyway, this temperature was never the actual temperature directly at the membrane, but we assumed that this temperature was not very far from the measured one. For the further analysis, we shifted the curves to a common midpoint in order to facilitate the fitting process. Next, we fitted the temperature-frequency curves for the n-propyl-2-oxazoline polymer graftings by an error function / O ⁼ fa ^{" er}f( ^{" ( —} Orans) in order to get values for the mean steepness of the transition m (in °C^_1) and the mean transition temperature 7_frans (in °C). The errors were calculated as the standard deviation of the values of the single experiments divided by the squareroot of the number of performed experiments.

The values resulting from the fitting quantify the transition.

Table 2

The mean steepness of the error function m was found to be or 1 .2 °C ¹ in average for the different length of the polymers. respectively. This means that the transition between the open state of the nanopore and the closed state was extremely sharp. A change of the surrounding temperature of about 1 °C^_1 is already enough to evoke a conformational change of the polymer grafting and with this a change in the openness of the pore starting from an open pore to a completely closed pore (no translocation events). The second fitting parameter was the midpoint position 7_frans that gave the transition temperature. Remarkably, as seen in Table 2, the lower critical temperature observed for polymers with a contour length longer than the pore size (Xn>100) where equal or inferior to 25°C. This property enable to use the temperature switch for the filtration of temperature sensitive proteins.

The translocation frequency was temperature independent for the 2-methyl-2-oxazoline polymer grafting and the membrane without polymer grafting. For both types of graftings the temperature-frequency curves were constant as a function of temperature. The conformation of the polymer grafting was unchanged during the experiment.

With this set of experiments, we were able to get an insight on the statics of the system and to observe the full shape of the temperature-frequency curves for the different polymer graftings by closely studying the cooling down process.

Flux driven DNA translocation experiments

In order to examine the correlation between translocation frequency and the spatial extension of the polymer layer grafted inside the nanopore we performed single molecule ZMW experiments to measure the translocation frequency of long double-strand DNA through the membrane as a function of the applied pressure. Since the critical pressure p_c is proportional to the inverse of the pore radius R to the power of four p_c~R^®, a small change in the pore radius results in a huge difference in the critical pressure. The resulting pressure-frequency curves were fitted with the suction model and the fitting results are then compared to the findings originating from membranes without a polymer grafting (Auger et al. 2014). Doing so allows us in turn draw conclusions about the thickness of the grafted polymer layer and with this the radius of gyration of the grafted polymers.

The measurements were done for n-propyl-2-oxazoline polymer graftings and for the 2- methyl-2-oxzoline polymer graftings. For the DNA translocation experiments using membranes with a 2-methyl-2-oxazoline polymer grafting, the surrounding temperature did not affect the polymers' configuration and therefore has no influence on the translocation frequency since the 2-methyl-2-oxazoline polymers did not exhibit an LOST. Hence, the experiments were performed at room temperature.

For the experiments using membranes with a n-propyl-2-oxazoline polymer grafting, we heated the system up to a temperature higher than 30°C since the grafted polymers propyl- short and propyl-long exhibit an LOST. Heating the system makes the grafted polymers collapse and as a result opened the pore. As a control experiment (data not shown), we tried to perform the same measurements with membranes with a n-propyl-2-oxazoline polymer grafting at room temperature but we recorded translocation events only at very high pressure confirming the assumption that the pores are indeed closed at a low temperature. We then fitted the resulting pressure-frequency curves by the suction model.

Auger et al performed the same experiments for membranes without a polymer grafting (Auger et al. 2014). Comparing our fitting results to the findings originating from membranes without a polymer grafting allowed us to draw conclusions about the radius of gyration of the grafted polymers according to the Poiseuille equation. For the same pores as the ones we used (declared radius: 25 nm, measured radius: 21 nm) but without polymer grafting, Auger et al. found the critical pressure to be 82 ± 4 mbar. Using these values, we could calculate the thickness of the grafted polymer layer R_p as R_p

Table 3 The data presented in the table 3 shows that the translocation of DNA through nanopores was strongly influenced by the grafting inside the pore. We assumed that the additional polymer layer mainly influences the translocation by reducing the diameter of the nanopore. As a result, the DNA molecules to be transported must be spatially restricted even further, and consequently the critical pressure is greater the greater the thickness of the polymer grafting.

The effect of the pore entrance functionalization on the translocation kinetics of l-DNA have been studied by several groups in the past. A simple type of grafting was realized by Auger et al who grafted short polyethylene glycol thiol molecules (mPEG thiol) using a protocol that was already used for other types of grafting by Jovanovic-Talisman et al. (Jovanovic- Talisman et al. 2009).

The polymers we used as a grafting layer are long, hydrophilic, flexible polymers. To verify whether the calculation of the thickness of the polymer layer we did before is too simple, we compared the thickness of the polymer layer extracted from the fitting process to the actual radius of gyration of the grafted polymers. Knowing the structure formula of the polymers, we estimated the size and the mass of one monomer inside the polymer. We determined the molar mass of one monomer by adding up the molar masses of all atoms inside the monomer. Subsequently, we divided the molar mass of the used polymer grafting by the molar mass of one monomer to calculate the number of monomers per polymer. To accomplish this, we took the mean of the two molar mass values (determined by NMR or SEC, respectively). The resulting values for the number of monomers per polymer were rounded up to whole numbers.

In order to calculate the size of one monomer we used the estimation that one C-C bond has a length of approximately 120 pm and one C-N bond a length of approximately 154 pm. As both polymer types exhibited two C-C bonds and one C-N bonds, the length of one monomer was calculated to be about 428 pm. We then used the FJC model to get an estimation of the radius of gyration of the grafted polymer that can be compared to the

T> — JN f thickness of the grafted polymer layer: e '6 with N being the number of monomers and b the monomer size. This polymer model was applicable also to grafted polymers with slight deviations as proved by Halperin in 1988 (Halperin 1988). He showed that there existed a collapse of grafted chains in poor solvent and that the collapse behavior of non overlapping grafted chains was identical to that of free coils but with no phase separation. According to Halperin, there were in fact different types of polymer configuration in a grafted polymer layer depending on the polymer type and the grafting density: For the case of not very densely grafted polymers (as for our work), the radius of free linear chains consisting of N monomers scales as JV^3/5 in a good solvent, as JV^1/2 in an intermediate solvent (Q- solvent) and as JV^1/3 in a poor solvent. So, for the 2-methyl-2-oxazoline polymers (that were assumed to be in an intermediate solvent) we could apply the upper formula with a slight change in the prefactor as the space accessible to the polymer was divided by two as the polymer was grafted onto a membrane (Halperin 1988). Thus, the radius of gyration reduces

Comparing the values for the radii of gyration originating from our data and from the calculation (Table 3) demonstrated that for the 2-methyl-2-oxazoline polymers the values derived from the experimental data were rather close to the theoretically calculated values. In contrast, the values for the n-propyl-2-oxazoline polymers derived from the experiments were significantly smaller than the radii of the gyration calculated using the FJC model. Using pressure to push DNA through nanopores, we primarily measured the free energy barrier of translocation experienced by the DNA due to the grafted polymers. Since both DNA and the polymers are relatively flexible, we might expect their radii of gyration to overlap as they interacted with each other. One would therefore assume that the experimentally determined thickness of the polymer coating was smaller than the theoretically calculated thickness. This was in agreement with our findings presented in table 3 showing that the theoretical data from applying the FJC model concerning the 2- methyl-2-oxazoline polymers were in very close agreement to our experimental findings. The highly hydrophobic 2-methyl-2-oxazoline polymers appeared to form a coating that interacted only weakly with the DNA. Therefore, the thickness of the polymer layer was indeed close to the radius of gyration of the grafted polymer. This implied that we were indeed able to determine the thickness of the grafted polymer layer by measuring the frequency of translocation of Lambda-DNA through nanopores as a function of the applied pressure.

With this in mind, we could also had a closer look to the findings for the n-propyl-2-oxazoline polymers and try to give an explication for the origin of the differing values stemming from our experiments and the theoretical polymer model. For the propyl-short-grafting and the propyl-long-grafting, the value for the thickness of the grafted polymer layer stemming from our experiments was significantly smaller than the value for the radius of gyration calculated with the FJC model. This demonstrated that by performing our experiments at a temperature higher than the LCST of both polymers, the n-propyl-2-oxazoline polymers were indeed brought into a collapsed state. Therefore, the actual thickness of the polymer grafting was smaller than the polymer's radius of gyration, because the FJC model is not applicable to this polymer configuration. This observation was consistent with the assumption of a temperature induced conformational change of polymers having an LCST, mentioned at the beginning of this section. Thus, the radius of gyration of the n-propyl-2-oxazoline polymers could not be calculated with the same formula as the one used before. Here, R_g scales as jV^1/3 since the polymers were in poor solvent. Due to geometric considerations the prefactor

was estimated to be *p , so ³ 4JT , leading to a radius of gyration of 1 .1 nm for the propyl-short and 1.3 nm for the propyl-long which is much closer to the experimentally determined values (Table 3).

Passive transport experiment

We are able to present a thermally controlled gating system for DNA translocation through nanopores. In the two first series of experiments reported in this work deal with the translocation of DNA through nanopores. DNA was huge compared to the size of the pore (its radius of gyration measure about ten times the size of one pore). Therefore, we had to consider entropic and enthalpic effects as the polymers need to be confined to translocate through the pore. In the set of experiments presented in the following, we shall work with molecules that are much smaller than the pore diameter. Due to their lesser size, there was no pressure application necessary to push them through the pores but they translocate through the membrane just by diffusion. Thus, the transport was not driven by a flux but by an osmotic pressure. The translocation of these molecules was not as clearly visible as the DNA translocation which we could examine at a single molecule level using the ZMW effect. This is why we performed the measurements as ensemble experiments. To quantify the diffusive transport we fluorescently labelled the small molecules and recorded the decrease in fluorescent intensity over time. The two chosen molecules were BSA and dextran, some of their important properties are repeatedly listed in table 4.

Table 4: Comparison of the two small biomolecules dextran and BSA used for the passive transport experiments

Dextran experiments Figure 2 show the recorded time evolution of the normalised fluorescent intensity for the different types of grafted polymers for dextran. The measurements were done for n- propyl-2-oxazoline and methyl-oxazoline polymer graftings

The curves depicted in Figure 2show a decrease in intensity for all different types of grafting The decrease in intensity varies for the different types of polymer grafting. Remarkably we observed that the polymers with a contour length larger than the pore radius show a decay kinetics much slower than the polymers with a contour length smaller than the pore radius. Transport properties of hydrophilic molecules through grafted membranes are thus enhanced for long grafting.

BSA experiments

We performed the same type of experiments as for dextran also for BSA as shown in Figure 3 which shows the recorded time evolution of the normalized fluorescent intensity for the different types of grafted polymers for BSA. The measurements were done for n-propyl- 2-oxazoline and and 2-methyl-2-oxazoline polymer graftings. The curves depicted in Figure 3 show a decrease in intensity for all different types of grafting. The decrease in intensity varies for the different types of polymer grafting. Remarkably we observed that the polymers with a contour length larger than the pore radius show a decay kinetics much faster than the polymers with a contour length smaller than the pore radius. Transport properties of hydrophobic molecules through grafted membranes are thus enhanced for long grafting.

Interpretation

The reason for this different transport properties of hyphobic and hydrophilic molecules lied in the interaction with the grafted polymer layer. BSA is charged and hydrophobic. Flence, it has an affinity for the pore, more precisely the layer of hydrophobic polymers grafted inside the pore. Dextran is uncharged and hydrophilic. Our results show that attractive hydropbic affinity between the transported molecule and the grafted membrane can be used to enhance the transport of hydrophobic molecules.

Claims

1. Nanoporous functionalized membrane comprising a synthetic nanoporous membrane and electrografted polymers, wherein said polymers are grafted within the nanopores of said synthetic nanoporous membrane and wherein the contour length of the polymers is greater than once the radius of the nanopores.

2. Nanoporous functionalized membrane according to claim 1 , wherein said polymers are chosen from poly-alkyl oxazoline polymers.

3. Nanoporous functionalized membrane according to claim 1 or 2, wherein said polymers comprise at least one electroactive function.

4. Nanoporous functionalized membrane according to claim 3, wherein said at least one electro active function is chosen from diazonium electro active probes.

5. Nanoporous functionalized membrane according to claim 5, wherein said nanopores mimick the selectivity of nuclear pore complexes in biological cells. by selecting molecules on the basis of their hydrophobicity.

6. Nanoporous functionalized membrane according to any of claims 1 to 5, wherein said polymers are chosen from polymers which exhibit a lower critical solubility temperature (LCST).

7. A process of preparation of the nanoporous functionalized membrane according to any of claims 1 to 6, wherein a synthetic nanoporous membrane is electrografted of polymers, said polymers being grafted within the nanopores of said synthetic nanoporous membrane and said polymers having a contour length greater than once the radius of the nanopores.

8. A process according to claim 7, wherein the grafting is achieved by polarization of a metallic layer present on the surface of the nanopores.

9. Use of a nanoporous functionalized membrane according to any of claims 1 to 6 to improve the filtration of biomacromolecules.

10. Use of electrog rafted polymers to functionalize a synthetic nanoporous membrane, by grafting the polymers within the nanopores of said synthetic nanoporous membrane, the contour length of the polymers being greater than once the radius of the nanopores.

11. Use according to claim 11 , to open or close the nanopores of the membrane thanks to a stimuli, i.e. temperature, light, pH, osmotic pressure and more preferably temperature.

12. Use according to any of claim 11 , wherein said polymers are chosen from polymers which exhibit a lower critical solubility temperature (LOST), in particular poly-alkyl oxazoline polymers.

13. Use according to claim 12 to clean the filter membranes by opening the nanopores via an increase in temperature.

14. Use according to any of claims 11 to 13, wherein said polymers comprise at least one electroactive function, in particular chosen from diazonium electro active probes.