Lipid Pore-Filled Silica Thin-Film Membranes for ... - ACS Publications

Subscriber access provided by Caltech Library

Article

Lipid Pore-Filled Silica Thin Film Membranes for Biomimetic Recovery of Dilute Carbohydrates Shanshan Zhou, Daniel M Schlipf, Emma Caroline Guilfoil, Stephen E Rankin, and Barbara Lynn Knutson Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b03844 • Publication Date (Web): 13 Nov 2017 Downloaded from http://pubs.acs.org on November 14, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Lipid Pore-Filled Silica Thin Film Membranes for Biomimetic Recovery of Dilute Carbohydrates Shanshan Zhou, Daniel M. Schlipf, Emma Guilfoil, Stephen E. Rankin, Barbara L. Knutson* Department of Chemical and Materials Engineering University of Kentucky, Lexington, KY 40506 Corresponding Author: [email protected]

1 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

KEYWORDS: boronic acid, lipid bilayer, sugar recovery, biomimetic membrane ABSTRACT Selectively permeable biological membranes containing lipophilic barriers inspire the design of biomimetic carrier-mediated membranes for aqueous solute separation. The recovery of glucose, which can reversibly bind to boronic acid carriers, is examined in lipid pore-filled silica thin film composite membranes with accessible mesopores. The successful incorporation of lipids (1, 2dipalmitoyl-sn-glycero-3-phosphocholine, DPPC) and boronic acid carriers ((4-(N-Bocamino)methyl) phenylboronic acid, BAMP-BA) in the pores of a mesoporous silica (~10 nm pore diameter) through evaporation deposition is verified by confocal microscopy and differential scanning calorimetry (DSC). In the absence of the boronic acid carrier, lipids confined inside the pores of silica thin films (~200 nm thick) provide a factor of 14 increase in diffusive transport resistance to glucose relative to traditional supported lipid bilayers formed by vesicle fusion on the porous surface. The addition of lipid-immobilized BAMP-BA (59 mol% in DPPC) facilitates the transport of glucose through the membrane; glucose flux increases from 45×10-8 mol/m2/s to 225×10-8 mol/m2/s in the presence of BAMP-BA. Furthermore, the transport can be improved by environmental factors including pH gradient (to control binding and release of glucose) and temperature (to adjust lipid bilayer fluidity). The successful development of biomimetic nanocomposite membranes demonstrated here is an important step towards the efficient dilute aqueous solute upgrading or separations, such as the processing of carbohydrates from lignocellulose hydrolysates, using engineered carrier/catalyst/support systems.


Page 2 of 35

Page 3 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

INTRODUCTION The recovery of dilute aqueous solutes such as organic acids and solvents, carbohydrates, small molecule drugs, and proteins is critical to the development of biotechnology applications such as fermentation and biorefining. Dilute aqueous separation process are usually costly and energyintensive due to the low concentration feeds and high latent heat of water.1, 2, 3 In the case of relatively low-value products, the limited options for product recovery and upgrading may act as bottlenecks for commercialization. For example, improved separation of carbohydrates from hydrolysate mixtures resulting from the enzymatic deconstruction of lignocellulosic biomass are needed for the economical production of biofuels and chemicals. Sugars with six or five carbon can be directly converted into biofuels through bacterial fermentation,4 but the optimal sugar concentration is typically around 100 g/L.5, 6 The sugars generated by enzymatic hydrolysis of biomass (e.g. in “cellulosic ethanol” production) are produced at significantly lower concentrations (1~20 g/L)

7, 8

resulting in inefficient fermentation. Alternatively, hydrolysate

sugars can be catalytically dehydrated into platform chemicals for further processing into a variety of value-added chemicals,9 but they need to be removed from the aqueous medium for effective catalysis. In industry, chromatography is the most used method for sugar separation and concentration, but is an expensive process with low production efficiency.3, 10 The recovery of aqueous sugars is a model separation that highlights the need for the development of alternative affinity-based separations. The chemical affinity of sugars to some compounds such as zinc porphyrins, resorcinarenes, alkaline earth metals and boronic acid derivatives presents opportunities for selective separation of different sugars.11 Among these sugar receptors, the boronic acid derivatives are of particular interest as the boronate ligand can reversibly interact with the 1,2-diols or 1,3-diols of saccharides to form cyclic esters with five- or 3 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 35

six-member rings (Scheme 1a).12, 13, 14 Furthermore, the association strength is largely depends on the type of sugar (catechol> fructose> glucose),15 which makes this reversible reaction with specific complexation very attractive for sugar sensing16 and selective separations. Consequently, boronic acid derivatives have been used to functionalize nanoparticles for sugar recovery.12, 13 Besides adsorption, boronic acid derivatives are commonly used as carbohydrate carriers in ion exchange membranes17 and liquid membranes10,

18

to facilitate the separation of sugars. The

selective transport of sugar through liposomes by a boronic acid derivative has also been demonstrated.19

Scheme 1. (a) Reversible binding between a boronic acid derivative and sugar; (b) Reversible binding between Alizarin Red S (ARS) and boronic acid (BA) as a fluorescent indicator. The ability of biological membranes to facilitate the selective transport of ions or hydrophilic molecules using carriers inspired the design of carrier-mediated biomimetic membranes for separation of solutes from dilute aqueous solutions in this work. However, naturally occurring lipid bilayers are fragile structures, not amenable to application in separations without support. High surface area mesoporous silica thin films with tunable pore structure are promising separation platforms for the construction of artificial membrane supports. Composite silica thin films with ordered, perpendicularly oriented hexagonally close packed (HCP) pore channels on a macroporous support have recently been synthesized by our laboratory and demonstrated as size4 ACS Paragon Plus Environment

Page 5 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

exclusion membranes.20 The composite thin film membranes are mechanically robust and the hydrophilic surface of the silica allows for easy surface functionalization and deposition of lipid assemblies. The perpendicular orientation of the 10 nm pores, formed by surfactant templating on a sacrificial chemically neutral crosslinked copolymer surface that is removed during calcination, provides for efficient transport. Thus, incorporating the carrier-loaded lipid bilayers with the silica thin film presents opportunities for designing composite membrane for highly selective separation of molecules. Specifically, the recovery of aqueous glucose from a dilute solution generated during enzymatic depolymerization of biomass is chosen as a model separation. The hypothesized mechanism of separation is by complexing glucose with boronic acid carriers on the donor side of the membrane and releasing it in the receiving phase (Scheme 2).

The

incorporation of carrier in the membrane allows for high permeability and high selectivity that can’t be achieved simultaneously in conventional membrane process.11, 21 The lipid carrier has the additional advantage of being environmentally responsive, for example to temperature.

Scheme 2. Sugar transport through lipid pore-filled silica thin film supported by a macroporous support (anodic aluminum oxide) using a boronic acid carrier. Preparing a defect-free membrane to provide high background transport resistance is important for effective carrier-mediated transport.22 Generally, supported lipid bilayers (SLBs) have been 5 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 35

prepared using Langmuir-Blodgett (LB) deposition or vesicle fusion.23 In the LB method, a single layer of lipids is formed by pulling the solid support through a Langmuir trough which has a lipid monolayer spread at the air-water interface. The LB technique is unable to embed a carrier into the bilayer.24 Vesicle fusion, where pre-formed small unilamellar vesicles (SUVs, 20 -100 nm) of lipid spread on a hydrophilic support through van der Waals interaction, is a simpler method to prepare SLBs.25 For mesoporous surfaces, vesicle fusion produces lipid bilayers that span pore sizes up to 20 nm,26, 27, 28 as the SUVs are too large to rupture inside of the pores.26 However, completely defect-free bilayers are difficult to achieve through vesicle fusion. The existence of pinhole defects has been revealed by AFM.29 Factors including composition and concentration of lipid, preparation method of liposomes, temperature, solution conditions (pH, presence of divalent cations) and physicochemical properties of the support strongly affect the interaction between lipid vesicles and the surface, and therefore the final quality of lipid bilayers.29, 30 Here we propose a robust lipid pore-filled system as a barrier for biomimetic separation (Scheme 2). By dissolving lipid molecules in organic solvent to avoid the formation of vesicles larger than mesopores, lipid assemblies are able to infiltrate into the pores.28 After evaporating the organic solvent and rehydration, lipid assemblies then form inside the pores.

Schlipf et al.28 have

successfully demonstrated the confinement of lipid assemblies of 1,2-dipalmitoyl-sn-glycero-3phosphocholine (DPPC) within the pores of mesoporous silica particles (with 5.4 nm and 9.1 pores). Furthermore, it has been reported that the lipid mobility at different locations of lipidfilled particles (at the core, midway between center and surface, or at the surface) is similar, suggesting that lipid self-assemblies are continuous through the pores.

Continuous lipid is

necessary for the incorporation and proper function of small lipophilic carriers and membrane proteins. This suggests the potential application of lipid pore confinement on membrane supports


Page 7 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

to mimic the barrier and selective transport properties of biological membrane for separation and biosensing applications. In this work, the ability of lipid assemblies of DPPC confined in mesopores of thin silica membranes to provide robust barriers to transport is examined and compared to traditional lipidcoated membranes. The enhanced transport in carrier-mediated biomimetic membranes is confirmed by measuring the permeability of glucose through the composite membrane before and after incorporating 4-((N-Boc-amino)methyl)phenylboronic acid (BAMP-BA) into lipid assemblies. The ability of environmental factors to improve glucose permeability are explored including applying a pH gradient (related to carrier-solute association) and temperature (related to the lipid bilayer diffusivity). The resulting system represents a versatile method for carbohydrate recovery. In addition to the boronic acid carrier for sugar binding, the combination of high surface area silica materials and selectively permeable lipid bilayers suggests potential applications for efficient, selective dilute aqueous solute separations if a specific carrier for the solute is incorporated. EXPERIMENTAL SECTIONS Material: Anodic aluminum oxide (AAO) membranes (Whatman, 25 mm in diameter) with pores of approximately 200 nm in diameter, a porosity of 0.25-0.5, and a nominal thickness of 60 µm were purchased from Fisher Scientific and served as the macroporous support for the silica thin films. Tetraethyl orthosilicate (TEOS, 98%), polyethylene oxide (PEO)-polypropylene oxide (PPO)-PEO triblock copolymer (P123, Mn = 5,800), cetyltrimethylammonium bromide (CTAB, ≥99%), glycerol (≥99%), potassium phosphate buffer tablets, chloroform (≥99%), 1,6diisocyanatohexane (DH, 98%), sodium hydroxide (NaOH, ≥98%) and glucose (≥99.5%) were supplied by Sigma Aldrich. Ethanol (anhydrous) was purchased from DLI and hydrogen chloride 7 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 35

(HCl, 6N), acetone (≥99.5%), and potassium chloride (KCl, ≥99.0%) were purchased from Fisher Scientific. Vybrant® DiO cell-labeling solution was purchased from Thermo Fisher scientific. Alizarin

red

S

(ARS,

certifiable

grade)

was

supplied

by

Amresco.

4-((N-Boc-

amino)methyl)phenylboronic acid (BAMP-BA, 97%) was purchased from Frontier scientific. 1,2dipalmitoyl-sn-glycero-3-phosphocholine (DPPC, >99%) was purchased from Avanti Lipid. Synthesis of thin film silica supported by anodic aluminum oxide (AAO) membrane: Thin film silica membranes with perpendicularly oriented pore arrays were synthesized on macroporous AAO supports using a crosslinked chemically neutral sacrificial polymer layer, as previously reported by Wooten et al.

20

and Koganti et al.31 To block the pores of AAO support

for silica thin film deposition, cross-linked P123 was used as a temporary pore-blocking film. A 0.696 mmol/L P123 in acetone was prepared and an equimolar amount of isocyanate crosslinker DH was added with continuous stirring in a nitrogen bag. Three drop of glycerol were then added and stirred for 10 minutes. The solution was used to dip coat AAO supports, or to modify cleaned glass slides to create chemically neutral surfaces for sandwiching the membranes. The dip coated substrates were cured at 100°C for 24 hours. The mesoporous silica coating solution was prepared by addition of a solution of P123 to a prehydrolyzed silica sol following the procedure of Brinker et al.32 TEOS, anhydrous ethanol, deionized ultra-filtered water and HCl were added together in a mole ratio of 1: 3.8: 1: 5 x 10-5. This solution was refluxed at a temperature of 70 °C for 90 min. A final amount of HCl and water were added yielding a concentration of 7.34 mM HCl. The mixture was stirred at room temperature for 15 min. The mixture was then aged at 50 °C for 15 min. A solution of P123 and ethanol with a mole ratio of 0.01:18.7 was then added to the TEOS mixture and stirred for 10 min. The final mole ratios were 1 TEOS: 22 C2H5OH: 5 H2O: 0.004 HCl: 0.01 P123. The copolymer 8 ACS Paragon Plus Environment

Page 9 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

modified AAO supports were dip coated with the final solution at a speed of 7.6 cm/min. The silica thin films coated AAO support were then sandwiched between two chemically neutral modified glass slides. The thin films of silica on AAO supports were aged and dried at 50 °C for 24 h and then at 100 °C for 24 h. Finally, the films were calcined 500 °C for 4 h with after ramping at 1 °C/min to remove the pore template and crosslinked pore blocking polymer. Membrane characterization: A dual-beam focused ion beam/scanning electron microscope (SEM) instrument (FEI Helios Nanolab 660) was used to characterize the membrane surface morphology. The samples (bare AAO support and thin film silica membrane) were prepared by attaching them to double sided carbon tape on 15 mm aluminum mounts. The pore accessibility of the thin film silica membrane was demonstrated by pressure driven ethanol flux through the membrane. The membrane was placed on a vacuum filtration holder (Cole Parmer) and secured to an Erlenmeyer flask. A chemical duty vacuum/pressure pump (Millipore) was used to control the vacuum pressure. The volumes of ethanol passing through the membrane over 5 min under pressure drops of 16, 26, 34 and 40 kPa were recorded. The ethanol flux through the bare AAO support was also measured to compare with the thin film silica membranes. Measurements were conducted in triplicate. Synthesis of spherical mesoporous silica particles: For visualization of the lipid and BAMPBA infiltration into pores, micron-scale SBAS (Santa Barbara Amorphous Sphere) mesoporous silica particles were prepared using the procedure of Schlipf et al.33 Briefly, 0.465 g of CTAB dissolved in 20 mL of deionized water was first added to 3.10 g P123. This solution was placed in a water bath at 30°C and 7.8 mL of 200 proof ethanol and 45.9 mL of 1.5 M HCl were added under vigorous stirring. After the P123 completely dissolved, 10 mL of TEOS was slowly added dropwise. This solution was stirred for 2 h and poured into a Parr 4748 Teflon lined bomb, which 9 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 35

had been acclimated to the hydrothermal aging temperature, 110 °C. The sample was kept at 110 °C in an oven for 3 days. After cooling to room temperature and homogenization in a high speed mixer, the sample was filtered and rinsed with deionized water. After filtration, the surfactants were removed by Soxhlet extraction for 24 h with 200 mL of 200 proof ethanol. Boronic acid (BA) loaded lipid-filled silica membrane/particle preparation through evaporation deposition: For evaporation deposition, 5 mg of DPPC and 2.5 mg BAMP-BA were dissolved in 1 mL CHCl3 and sonicated for 5 minutes. A silica membrane or 10 mg SBAS particles was then immersed in the lipid solution and sonicated for 25 minutes in a shell vial (25 mm × 95 mm). The above mixture was then blow-dried with nitrogen and was further dried under high vacuum for at least 2 h. To form self-assembled structures, the dried lipid inside the silica membrane or particles was rehydrated in 1 mL PBS solution and sonicated at 47 °C for 1 hr. The sample was sonicated for another 15 minutes while cooling to 30 °C. Approximately 60x excess lipid was used (based on the pore volume of the silica thin film). The excess lipid was removed by washing the membrane or particles with PBS three times. Boronic acid (BAMP-BA) loaded lipid-enveloped silica membrane/particle preparation through vesicle fusion: To prepare vesicles, 5 mg of DPPC and 2.5 mg BAMP-BA were first dissolved in 1 mL CHCl3, and the lipid was blow-dried with air for 1 h followed by high vacuum for at least 2 h. The dried lipid was rehydrated in 1 mL PBS solution to form multilamellar vesicles. The resulting mixture was then extruded at 45 °C through a polycarbonate filter with 200 nm diameter pores 21 times, which produces small unilamellar liposomes with a mean diameter of 200 nm. The liposome solution (BAMP-BA loaded liposomes) was then mixed with a silica membrane or 10 mg silica particles for 1 hr to facilitate vesicle fusion. The excess lipid was removed by washing the membrane or particles with PBS three times. 10 ACS Paragon Plus Environment

Page 11 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR): To confirm the existence of lipid assemblies in the pores of lipid-filled membranes (silica and AAO), they were characterized using a ThermoNicolet Nexus 470 spectrometer equipped with a liquid nitrogen cooled mercury cadmium telluride (MCT) detector and ATR with single-reflection 45° ZnSe crystal (VeeMAX III, PIKE Inc., Madison, WI). Each spectrum was collected at a resolution of 4 cm-1 and 250 scans with IR beam incident at 45°. Confocal laser scanning microscopy: To visualize the lipid, 1 mL of 10 mg/mL lipid-coated particles were mixed with 10 µL of 1 mM fluorescent dye DiO. For the boronic acid loaded lipid-coated particles, 1mL of 10mg/mL boronic acid loaded lipid bilayer-coated particles were mixed with 1 mL of 1 mM Alizarin red S (ARS). The particles were incubated with the dyes on a shaker for 30 min. The samples were then washed with PBS three times and imaged within 2 hours of preparation using a Leica TSP SP5 confocal microscope. Experiments were performed at 28 °C using a 63×/1.4 oil immersion objective. DiO was excited at 488 nm with an argon laser at 6% laser power for imaging and emission was observed between 490 nm and 510 nm. The ARS-BA complex was excited at 514 nm for imaging and emission was observed between 580 nm and 620 nm. Differential Scanning Calorimetry: DSC was used to confirm the formation of lipid bilayers on silica particles by measuring the gel to fluid phase transition temperature of DPPC-loaded particles. Following the preparation of DPPC-filled SBAS, the particle suspensions were centrifuged at 1,000×g to form a soft pellet. 10 µL of the pellet was hermetically sealed in a DSC sample pan. Thermograms were collected with a TA Instruments Q600 DSC between 20 ºC and 70 ºC at a ramp rate of 10 ºC/min and returned to 20 ºC at a cooling rate of 10 ºC/min.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 35

Glucose affinity to BAMP-BA at different pH: Alizarin Red S (ARS) was used a fluorescent indicator (Scheme 1b) for measuring the relative glucose affinity to the boronic acid derivative at different pH values.34 Briefly, 1 mL of 5 mg/mL BAMP-BA loaded liposome solution (prepared as described previously) was mixed with 1 mL of 1 mM ARS to yield a fluorescent ARS-BA complex. Liposomes were necessary to solubilize the lipophilic BAMP-BA. The fluorescence emission at 600 nm of the ARS-BA loaded liposome solution was measured using a plate reader (Synergy Mx, BioTek, Winooski, VT) with excitation at 490 nm before and after addition of 92.8 mM glucose at the various values of pH. The desired pH of these solutions (prepared in PBS buffer) was achieved with the addition of HCl or NaOH. The fluorescence intensity of the solution after addition of glucose, normalized by the value without glucose, indicates the degree of ARS displacement due to glucose binding. Measurements were conducted in triplicate. Glucose diffusion through membranes as a function of temperature and pH: Glucose diffusion through the silica thin film membranes was measured in a side-by-side diffusion cell (PermeGear, Hellertown, PA). The membrane was held in place between two chambers, each holding 7 mL of solution, and tightly clamped in place while allowing a 2 cm2 cross sectional area of the membrane to be exposed. The donor side of the cell was loaded with 7 mL of 5.6 mM glucose solution in PBS buffer (0.9 M, pH 7.4 or adjusted to 10 by NaOH). The receptor side was initially loaded with 7 mL of PBS buffer (pH 7.4 or adjusted to 5 by HCl) with no solute. The diffusion cell was then placed on a heated plate with temperature controlled at 23 ºC or 45 ºC. The concentration of solute in the receptor side and donor sides were both sampled as a function of time. At sampling time, a 200 µL sample was taken from each side and the concentration was measured using high performance liquid chromatography with a Bio‐Rad Aminex HPX‐87H column and a Shodex R01 refractive index detector. The mobile phase was degassed 5 mM aqueous H2SO4 fed at a rate of 0.4 mL/min, and the temperature of the column was 50 ºC. 12 ACS Paragon Plus Environment

Page 13 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

RESULTS AND DISCUSSION Synthesis of mesoporous silica thin film on macroporous support as separation platform: Silica thin films with accessible 2-D hexagonally close packed (HCP) pore structure on porous supports were utilized as the separation platform and support for lipid assemblies. The HCP pore structure formed from cylindrical micelles by surfactant templating is considered to be superior for membrane applications compared to interconnected cubic structures due to their continuous channels, which do not offer alternate pathway for diffusion.35 However, HCP pores naturally tend to align parallel to the substrate during surfactant templating.36 This parallel orientation make the pores inaccessible, and not suitable for membrane applications.37 Recently, we have successfully demonstrated the synthesis of silica thin films with perpendicularly oriented HCP pore on macroporous supports and their ability to act as nanofiltration membranes.20 The same technique was employed here to prepare the composite supported lipid membranes. Porous anodic aluminum oxide (AAO) substrates with ~200 nm pores were used to anchor mesoporous silica membranes. To provide a chemically neutral surface for perpendicular pore alignment and simultaneously prevent the penetration of silica into the pores of the support, the AAO was dip coated with cross-linked copolymer.38 After a stable mesostructure was obtained through a two-step aging process at 50 and 100 °C, calcination was performed to remove the pore template and the blocking copolymer layer. To confirm that pore-accessible silica thin film was successfully deposited on the porous support, SEM images of bare AAO (Figure 1a) and a representative mesoporous silica membrane on AAO after calcination (Figure 1b) were compared. The images show that a continuous silica thin film with accessible pores is formed on the top of the AAO anchoring layer, which results in greater electron scattering and decrease in contrast compared to the bare AAO. Characterization of the silica thin film by TEM showed that 13 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 35

the pores of the silica thin film templated by P123 were highly ordered HCP structures with pore dimension of approximate 10 nm.20, 38 Therefore, AAO-supported silica thin films with accessible continuous 10 nm cylindrical channels were successfully synthesized. However, it is hard to determine whether the deposited silica thin film is defect-free by SEM imaging since the electron beam can induce local heating of the thin film and cause defects. Therefore, ethanol flux measurement through the supported mesoporous silica membranes were performed. The ethanol flux vs. pressure drop (Supporting Information Figure S1) is significantly reduced compared to bare AAO, and consistent with prior flux measurements of membranes with 10 nm pores.20

Figure 1. Plane-view SEM images of (a) AAO support and (b) mesoporous silica thin film coated AAO.


Page 15 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Lipid Pore-Filled Silica Thin Film Membranes as a Transport Barrier : In order to examine the location of DPPC lipid assemblies on porous supports, micron-scale SBAS spherical mesoporous silica particles with pore diameter of 11.8 nm (Figure 2a), which is close to that of the silica thin films, was chosen as a model platform to allow direct visualization of lipid bilayers using confocal scanning laser microscopy.28 The lipid enveloped particles prepared by vesicle fusion have a continuous, single layer of lipid bilayer (tagged with green fluorescent dye DiO) surrounding the spherical particle surface (Figure 2b), but no lipid inside of the particles. Alternatively, the confocal images of the lipid pore-filled particles (Figure 2c) suggest that the lipids are confined inside the pores of composites prepared by evaporation deposition. Typically, bilayers of DPPC have a thickness of 3-5 nm.39 This thickness provides a guideline for the pore diameter required to pack lipid inside the pores by this evaporation deposition method; the pores need to be large enough to accommodate lipid assemblies that would be expected to be approximately the size of the lipid bilayer thickness. If the pore size is too small, lipid bilayers can only form on the particle surface, which results in lipid-enveloped particles.28

Figure 2. (a) SEM image of spherical mesoporous silica particles; confocal microscopy images of (b) lipid enveloped silica particle; (c) lipid filled silica particles; (d) lipid-BA enveloped silica particle; (e) lipid-BA filled silica particles. Lipid is tagged with green fluorescent DiO and 15 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 35

boronic acid (BAMP-BA) is complexed with ARS to give red fluorescence. All confocal images are cross-sections of the core of SBAS particles with pore diameter 11.8 nm. Images with lower magnification are provided in the Supporting Information (Figure S2). The ability of supported lipid bilayers to effectively prevent the transport of sugar in the absence of a carrier is critical for providing subsequent selective membrane transport using the lipidembedded boronic acid carrier. To quantify the effect of lipid location (lipid filled pores versus traditional lipid-enveloped pores) on membrane performance, glucose flux through the two types of supported lipid bilayers was measured and compared with the bare support (silica thin film membrane). As shown in Figure 3, the glucose flux of lipid-enveloped membrane prepared by vesicle fusion is similar to that of the silica membrane and AAO support, which indicates that the external lipid bilayer is not an effective barrier for glucose transport. Meanwhile, lipid-filled membranes, where DPPC is confined inside the membrane pores, show a 14-fold decrease in glucose flux, suggesting that lipid-filled mesopores provide a more effective barrier than lipidenveloped pores. The reason that external lipid bilayers are not an effective barrier is that lipid bilayers prepared by vesicle fusion are prone to several known types of defects.29 The high flux of glucose may be due to defects such as pinholes in the supported lipid bilayer, which can potentially be reduced through the optimization of the surface chemistry of the support, lipid composition and deposition methods to make better solute rejecting bilayer on silica thin film surface. For example, the hydraulic water permeability of commercial polyethersulfone nanofiltration membranes with sulfonic surface groups was significantly reduced (by about 12-fold) after vesicle fusion of 1,2dimyristoyl-sn-glycero-3-phosphocholine (DMPC), while the deposition of DMPC vesicles on polyamide membranes with carboxylic surface groups did not alter the membrane permeability 16 ACS Paragon Plus Environment

Page 17 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

significantly.40 Changing the DMPC to DMPC/DMTAP (1, 2-Dimyristoyl-3-trimethylammonium propane) lipid mixture for vesicle fusion on the polyethersulfone membrane further reduced the hydraulic permeability by 2-fold. Alternatively, the transport of glucose, which is expected to be semi-permeable in lipid bilayers, may present a more stringent test of the integrity of the supported lipid bilayer than the transport of ions; Freeman et al.41 reported that the permeability coefficient of ions (K+, Na+, Cl-) through liposomes is around 10-10 to 10-12 cm/s , while that of glucose is about 10-7 cm/s. The lipid enveloped silica thin film membrane may be better suited to applications involving biomimetic ion transport rather than uncharged hydrophilic molecules, for which it does not have significant barrier properties.

Figure 3. Glucose flux through different membrane (5.6 mM initial glucose concentration on the donor side of a side-by-side diffusion cell). The error bars are based on the sampling times at 1 h, 1.5 h, and 3 h of diffusion). In the case of the lipid-filled pores, the thickness of a DPPC bilayer is about 4 nm, and the pore diameter of the silica thin film is about 10 nm, allowing for the inclusion of lipid self-assemblies. This is consistent with the observation of Schlipf et al.,28 where the evaporation deposition method was used for mesoporous silica particles with varying pore sizes. They found that DPPC was confined to the surface of particles with 3.0 nm pores, while it was able to assemble in the 17 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 35

pores of particles with 5.4 nm and 9.1 nm pores. They also reported that the lipid diffusivity remained consistent regardless of lipid location (at the core, midway between the core and surface, or surface) in the particle, suggesting uniform distribution of lipid assemblies inside the pores. Similarly, dissipative particle dynamics simulations of DMPC confined in hydrophilic pores revealed that the morphology of pore confined lipids is dependent on the pore radius (R) relative to the lipid length (L) and lipid concentration in the pores.42 As the lipid concentration increases, the lipid first forms micelles of distinct morphologies, then cylindrical bilayers (tubes).42, 43 The larger the ratio of pore radius to lipid length, the lower the concentration required to form cylindrical bilayers.42

The effect of pore size and concentration has also been studied for

nonionic surfactant bilayers in confined pores. Cylindrical bilayers of n-dodecyl-penta(ethylene glycol) (C12E5, which forms a 4.2 nm thick bilayer on a flat surface) were formed in 8 nm cylindrical nanopores of silica particles when the bulk concentration was above the CMC, as characterized by grazing incidence small-angle neutron scattering (GISANS). Considering the pore size of the silica thin film membrane used here (~10 nm), and the DPPC concentration (6.9 nM, CMC= 0.46 nM), the lipid most likely forms a continuous tube-like cylindrical bilayer that lines the inner pore surface of the silica membrane to act as barrier to aqueous solute. ATR-FITR was also performed to characterize the location of lipid assemblies in the composite membranes (consisting of silica thin film and AAO support). ATR-FITR (Supporting Information Figure S3) shows the existence of C-H stretching bands (2915, 2846 cm-1), PO2- stretching (1222, 1087 cm-1) and C-O-C stretching (1052 cm-1) for DPPC on both bare AAO and a silica-coated membranes after lipid filling, which indicates that the lipid not only exists inside the silica layer but may be able to load inside of the AAO support when excess amount of lipid is used. However, Karp et al. 44 reported NMR evidence that the DMPC or DPPC bilayers formed on the inner 200 nm diameter pores of AAO were aligned with the molecules normal to the pore surface, 18 ACS Paragon Plus Environment

Page 19 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

suggesting that lipid exists as multilayers coating the pores of an AAO support.45 Considering the 200 nm pore diameter of AAO, the lipid multilayers are not sufficient to block the pores, and the glucose flux of a lipid-filled AAO support prepared using the same procedure as the composite membrane was measured to be 184 x 10-8 mol/m2/s, which 4-fold higher than the lipid-filled silica membrane. Therefore the barrier behavior of this composite membrane primarily results from the lipid pore-filled silica thin film.

Incorporation of boronic acid carrier into lipid bilayers: The lipophilicity of the boronic acid derivative is an important factor for immobilizing the carrier into the lipid bilayer to ensure sugar transport efficiency. Westmark et al.19 encapsulated different types of boronic acid derivatives and glucose into liposomes and found that the glucose efflux from the liposome, as facilitated by the boronic acid derivative, was largely dependent on the lipophilicity of the carrier. Thus the lipophilic BAMP-BA (octanol-water partition coefficient, log Po/w=2.53, as reported by the supplier), is chosen for its ability to reside in the lipid bilayer. A mixture of DPPC/BAMP-BA in a molar ratio of 6.9:10 (59 mol% BA) was dissolved in chloroform prior to lipid rehydration to incorporate BAMP-BA into lipid assemblies. To further confirm that BAMP-BA is successfully immobilized into lipid bilayers, the dye Alizarin red S (ARS), which is not fluorescent itself but becomes fluorescent upon complexation with boronic acid, was utilized (Scheme 1b). ARS was added to the boronic acid immobilized lipid system and indicates that BAMP-BA is present in the same locations as DPPC – that is, at the surface of lipid-enveloped SBAS particles (Figure 2d), and distributed throughout the particles in lipid pore-filled SBAS particles (Figure 2e). The colocalization of the BAMP-BA and DPPC in the particles confirms that the boronic acid carrier is confined in the lipid bilayers.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 35

In order to test whether the addition of BAMP-BA alters the structure of DPPC lipid bilayers, the gel to fluid phase transition temperatures were compared of lipid assemblies with and without boronic acid carrier loaded into SBAS particles. This gel to fluid phase transition temperature is associated with disordering of the acyl chains of lipid molecules as the bilayer becomes more fluid with increasing temperature.46 . As measured by differential scanning calorimetry (DSC) , in the absence of BAMP-BA, DPPC in the lipid-filled particles undergoes the gel to fluid phase transition at 41.3 oC (Figure 4), which is the same as in DPPC liposomes.47 For BAMP-BA immobilized lipid-filled particles, the transitions temperature decreases to 39.3 oC, along with a decrease in the enthalpy of the gel to fluid phase transition. This is consistent with the incorporation of BAMP-BA in the DPPC bilayer, increasing the membrane disorder. This phase transition behavior of lipid bilayers with 59 mol% BAMP-BA incorporation is identical in liposomes, lipid filled silica thin films, and lipid filled particles (Supporting Information Figure S4), suggesting the pore confinement does not affect the loading of BAMP-BA into lipid assemblies.

The phase transition temperature and enthalpy change of the phase transition

gradually decrease as the additive concentration increases in a bilayer, potentially disordering the bilayer structure to the point that this transition becomes negligible at a critical concentration which depends on the type of additive and lipid.48, 49, 50 In DPPC, 50 mol% cholesterol was reported to completely eliminate the phase transition.51, 52 Because the bilayer structure could affect the transport properties in the biomimetic membrane, knowledge of the effect of BAMPBA on the gel to fluid phase transition is important to establish that the carrier is present in the membrane but does not completely distort its structure.


Page 21 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 4. Gel to fluid phase behavior of lipid-filled silica SBAS particles (dashed line) and BAMP-BA immobilized lipid-filled particle (solid line) as measured by DSC (59 mol% BAMPBA, pore diameter: 11.8 nm). Glucose transport through supported lipid bilayers with boronic acid carrier: Transport of glucose through the lipid-filled composite silica thin film membrane was compared with and without boronic acid carrier in the lipid phase using a static diffusion cell. Glucose transport with BAMP-BA was only measured in the lipid-filled pores because lipid enveloping did not provide a sufficient barrier to transport. As shown in Figure 5, glucose flux is enhanced by a factor of four after the immobilization of BAMP-BA in the lipid at room temperature (RT), which suggests that BAMP-BA is carrying glucose through the membrane. The transport mechanism of glucose through the lipid filled silica membrane is hypothesized as described in Scheme 2. On the glucose concentrated side, glucose combines with BAMP-BA, which is immobilized in lipid bilayers, and is then transported through the lipid bilayers. Due to the low glucose concentration on the receiving side, the glucose-boronic acid complex dissociates and releases glucose and BAMP-BA. The free boronic acid carrier then can go back to facilitate transport of another glucose molecule. The combination of passive transport and facilitated transport in a single membrane overcomes the trade-off between high permeability and high selectivity that can’t be 21 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 35

achieved simultaneously in conventional membrane processes. At the same time, the incorporation of carrier into the lipid allows its reuse, which is expected to significantly reduce the cost of using a carrier in the aqueous phase, and eliminates the recovery step of carrier from the product solution.

Figure 5. Glucose flux through lipid filled membrane with or without boronic acid carriers as a function of temperature (room temperature versus 45oC) and pH (pH 7.4 versus a pH 10 (donor)/pH 5 (receptor) gradient). The diffusion was conducted with a 5.6 mM initial glucose feed for 59 mol% BAMP-BA in DPPC, and error is based the sampling at 2 h, 3 h, and 4 h of diffusion. Opportunities to control the glucose flux in the boronic acid carrier system include a) tuning the glucose-boronic acid interaction via choice of boronic acid derivative or environmental conditions; b) altering the transport properties in the lipid assemblies; and c) adjusting the relative concentrations of the glucose and the boronic acid carrier. For a 5.6 mM (dilute) glucose solution, 13 mol%, 59 mol% and 74% BAMP-BA in DPPC deposited by evaporation deposition into the pores of oriented mesoporous silica membranes resulted in glucose fluxes of 142, 412 and 294 × 22 ACS Paragon Plus Environment

Page 23 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

10-8 mol/m2/s at 45oC (neutral pH), respectively. At 74 mol% BAMP-BA, boronic acid precipitation was apparent in the lipid mixture, making the carrier ineffective. Increasing the glucose concentration in the donor phase to 28 mM resulted in a flux of 1700×10-8 mol/m2/s through the membrane with 59 mol% BAMP-BA. This glucose concentration was chosen to represent a dilute sugar concentration consistent with biomass hydrolysates (1-20 g/L).7,

8

However, the concept of boronic acid carriers in liquid membranes has only previously been examined in liquid membranes at sugar concentrations of 100 to 500 mM,10 which suggests that the range of operation with respect to glucose operation concentration can be extended in the current lipid-based system. The ability to operate above and below the gel-fluid transition of lipid bilayers suggests a potential temperature dependence of the flux based on differences in mobility of the carrier due to changes in membrane fluidity. With 59 mol% BAMP-BA, the DPPC bilayers still retains a gelfluid phase transition at approximately 39 oC (see above). Glucose diffusion studies conducted above the transition temperature (at 45 oC) after carrier immobilization show an increase in glucose flux by a factor of two - from 225×10-8 mol/m2/s to 412×10-8 mol/m2/s (Figure 5). While the glucose flux in the absence of BAMP-BA also increases upon increasing the temperature to 45 oC (by a factor of 1.4 to be precise), the effect is more significant for the lipid assemblies containing immobilized BAMP-BA (Figure 5). This dramatic temperature effect on solute flux is attributed to a change in the fluidity of the lipid bilayer. Below the transition temperature, the lipid bilayers is in a “solid” gel phase, where the lateral diffusion is very slow so the lipid bilayers is more rigid and less permeable, which limited the mobility of immobilized boronic acid. Above the transition temperature, the bilayers are in a “fluid” liquid phase.46 In this state, the lipid bilayers shows rapid lateral diffusion and more favorable for compound exchange with the environment,53 which allows the carrier to move fast and improve transport efficiency. 23 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 35

In order to test membrane stability when switching below and above the phase transition temperature, the glucose fluxes through the same BAMP-BA immobilized lipid filled membrane were measured when the temperature was switched in the order of RT (room temperature) to 45oC to RT and finally returned to 45oC (Figure 6). The glucose flux at RT was found to reduce 35% after the first cycle from RT to 45oC and back again. However, no BAMP-BA carrier leakage into the aqueous phase was detected as determined by fluorescence spectroscopy during the diffusion study. The reduction of glucose flux may have been caused by the annealing above the phase transition temperature, which allows the lipid bilayers to reconstruct a better barrier,54 and reduces the background flux. In addition to temperature, lipid composition and support surface chemistry are well documented factors that will affect lipid bilayer fluidity.55, 56 Thus, factors that affect lipid bilayer fluidity and permeability can be examined in future studies in order to improve the transport efficiency and selectivity towards solutes.

Figure 6. Relative glucose flux through boronic acid immobilized lipid filled membrane during multiple temperature-cycle measurements from room temperature (RT) to 45oC.

(At each

measurement, the same membrane started with 5.6 mM initial glucose solution on the donor side and sampling at 2 h of diffusion. Flux reported was normalized to first flux measured at 45oC).


Page 25 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

An alternative approach to changing the glucose flux through the membrane is to change the interactions of the solute-carrier system. The formation and dissociation of glucose-boronic acid complexes is pH dependent.57 The competitive binding of ARS dye and glucose with BAMP-BA was used to quantify the interaction of the glucose with the carrier as a function of pH. As the ARS-boronic acid complex is fluorescent, the addition of glucose results in the dissociation of this complex and decrease in fluorescent intensity (Scheme 1b). The relative fluorescence intensity of the ARS-BA complex as a function of pH at constant glucose concentration is shown in Figure 7. Glucose binds to BAMP-BA strongly at basic conditions (reducing fluorescence by displacement of ARS) and minimally at acidic conditions. Based on this observation, glucose flux was measured across the lipid-filled silica thin film membranes while maintaining a pH gradient with the donor side of the diffusion cell at pH 10 (to promote carrier association), and the receptor side at pH 5 (to promote dissociation from the carrier). The lipid-filled silica thin film membrane acts as an ion barrier, allowing this pH gradient to be maintained across the membrane. Applying the pH gradient improves glucose flux at RT from 225×10-8 mol/m2/s to 287×10-8 mol/m2/s relative to neutral pH experiments when facilitated by BAMP-BA (Figure 5). No significant change in glucose diffusion was observed relative to the neutral pH in the absence of BAMP-BA. This observation further supports the mechanism of boronic acid-mediated transport of glucose in the lipid-filled pores. Most importantly, it demonstrates that tuning the strength of the saccharide-boronic acid interaction can be used to alter transport through the lipid-filled pores. This tuning can be achieved through operating conditions (as in the case of pH), or the choice of boronic acid carrier.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 35

Figure 7. Relative fluorescence intensity after adding glucose to ARS-BA complex (fluorescent) to competitively form BA-glucose (non-fluorescent) as a function of pH in BAMP-BA loaded liposomes. The solution fluorescence intensity after adding glucose was normalized to the original ARS-BA complex intensity at the same experimental pH. The reported error is based on three replicates. A boronic acid-based carrier has previously been incorporated into liquid membranes consisting of 2-nitrophenyl octyl ester supported by a flat sheet of porous polypropylene for sugar separations. Duggan, et al.

18

reported a glucose flux of 0.54×10-8 mol/m2/s using 50 mM 2-

(aminomethyl)-phenylboronic acid in the supported liquid membrane at neutral pH, when the feed concentration of glucose was 300 mM. Glucose flux increased more than ten-fold when a pH gradient was applied (pH of 11.3 on donor side and pH of 6.0 on receptor side), which is consistent with the pH effect seen in this work. Compared to the supported liquid membrane, BAMP-BA immobilized in a lipid filled silica membrane provides a much higher glucose flux of 225×10-8 mol/m2/s at neutral pH and room temperature, which is almost 400 times greater than that of liquid membrane. This advance is notable because the baseline experiments are conducted at only 5.6 mM glucose – much closer to the concentration likely to be encountered during biomass hydrolysis operations. 26 ACS Paragon Plus Environment

Page 27 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

In addition to the significant glucose flux difference, liquid membranes generally have stability problems because the organic phase is held in the support pores by capillary force.11 In addition to stability issues with the carrier liquid, Luccioa, et al.

10

reported that glucose flux reduced by

about half after 48 hours of transport due to the boronic acid carrier leaching out of a liquid membrane.10 To address this problem, quaternary ammonium salts can be added to improve the boronic acid derivative’s solubility in the liquid membrane. It was found that the addition of trioctylmethylammonium chloride to the liquid membrane increased the glucose flux by two to ten times, depending on the pH of solution, which was attributed to the better sugar-boronic acid complex solubilization in the liquid membrane due to its stronger association with the salt.58 In the present case, the BAMP-BA carrier used is compatible with DPPC bilayers without requiring chemical additives to the sugar solution. To demonstrate the potential of the boronic acid immobilized lipid filled silica membrane to selectively separate carbohydrates representative of biomass hydrolysates, the flux was compared during competitive diffusion of a mixed system containing glucose (a six-carbon sugar), xylose (a five-carbon sugars) and cellobiose (disaccharide of glucose) (Figure 8). The effect of temperature (room temperature and 45oC) and pH gradient (pH 7.4 and a pH 10 (donor)/pH 5 (receptor) gradient) on the flux were also evaluated. The results show that xylose had the highest flux through the membrane while the flux of cellobiose was the lowest. The observation is consistent with the reported relative carbohydrate association constant (Keq) to phenylboronic acid at neutral pH, which favors the complexation with xylose: xylose > glucose > cellobiose.57 As expected, the flux of all three sugars significantly increased as the temperature increased above the phase transition temperature, while the selectivity of xylose and glucose over cellobiose was reduced slightly. The temperature change increases the overall fluidity of the lipid bilayers, which enhances the solute permeability regardless of their association constant differences, resulting in 27 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 35

lower selectivity. In contrast, applying a pH gradient improved both solute permeability and selectivity. Because the pH gradient directly changes the association constant, which is dependent on the type of carbohydrate, both the permeability and selectivity increased. The results demonstrate that the proposed boronic acid-mediated biomimetic membrane has potential for selective recovery of monosaccharides from biomass hydrolysates. The selectivity for different carbohydrates may be further improved by reducing boronic acid concentration, selecting a more specific boronic carrier, and tuning the lipid composition or covalently linking the lipid to silica membrane to provide a better background barrier for carbohydrates.

Figure 8. Selective separation of components in a carbohydrate mixture (5.6 mM glucose, 5.6 mM xylose, and 5.6 mM cellobiose) through BAMP-BA immobilized lipid filled membrane as a function of temperature (room temperature versus 45oC) and pH (pH 7.4 versus a pH 10 (donor)/pH 5 (receptor) gradient. (Sampling time: 2 h, S=Selectivity relative to cellobiose flux). CONCLUSIONS Lipid pore-filled silica thin film membranes with an immobilized carrier are successfully demonstrated to facilitate the transport of hydrophilic molecules from dilute aqueous solution. These robust lipid-filled mesoporous membranes act as barriers to ions and small hydrophilic


Page 29 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

solutes that can be matched with biomimetic carriers to provide selective transport as a separation and sensing platform. Relative to traditional supported lipid bilayers that reside on the external surface of the porous film, the lipid pore-filled membrane provides a better barrier which is critical for the carrier function in the membrane. A unique aspect of lipid-filled pore membranes is the temperature dependence of flux due to the gel to fluid lipid bilayer phase transition, which can be used as a temperature switch to increase or decrease mobility through the membrane. Opportunities to tune transport through the solute-carrier interactions can be used to manipulate the flux and achieve selective separations or, in a very different context, to allow temperaturecontrolled release of a solute. In the case of carbohydrate-boronic acid mediated transport, the affinity-dependent transport suggests the potential for separating sugars from dilute aqueous mixtures of processed biomass for the improvement of lignocellulose conversion to chemicals. More broadly, biomimetic membranes which combine high surface area silica thin film membranes with selectively permeable lipid bilayers have potential as an efficient aqueous-based separation technology that can be tuned through the selection of lipid-based carrier molecules. ASSOCIATED CONTENT Supporting Information: Ethanol flux through the bare AAO support and silica membrane and ATR-FTIR characterization of the lipid pore-filled silica thin film. This material is available free of charge via the Internet at http://pubs.acs.org. AUTHOR INFORMATION Corresponding Authors * BLK Email: [email protected]; tel: 859-257-5715; fax: 859-323-1929. [email protected]; tel: 859-257-9799; fax: 859-323-1929. 29 ACS Paragon Plus Environment

SER Email:

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 35

ACKNOWLEDGMENT This work was supported by NSF Experimental Program to Stimulate Competitive Research (EPSCoR) grants (award numbers 1355438 and 1632854), and a United States Department of Agriculture Biomass Research and Development Initiative (BRDI) grant (award # 2011-1000630363).

REFERENCES

1.

Kaghazchi, T.; Kargari, A.; Yegani, R.; Zare, A. Emulsion liquid membrane pertraction of

L-lysine from dilute aqueous solutions by D2EHPA mobile carrier. Desalination 2006, 190 (1-3), 161-171. 2.

Zhang, H.; Wang, Y.; Bai, P.; Guo, X.; Ni, X. Adsorptive Separation of Acetic Acid from

Dilute Aqueous Solutions: Adsorption Kinetic, Isotherms, and Thermodynamic Studies. Journal of Chemical & Engineering Data 2016, 61 (1), 213-219. 3.

Tian, M.; Row, K. H. Separation of glucose and bioethanol in biomass with current

methods and sorbents. J Chromatogr Sci 2013, 51 (8), 819-24. 4.

Taherzadeh, M. J.; Karimi, K. Enyme-based hydrolysis process for ethanol from

lignocellulosic materials: a review. BioResources 2007, 2 (4), 707-738. 5.

Kurian, J. K.; Ashok, M. K.; Banerjee, A.; N., K. V. V. Bioconversion of hemicellulose

hydrolysate of sweet sorghum bagasse to ethanol by using pichia stipitis NCIM 3497 and debaryomyces hansenii sp. BioResources 2010, 5 (4), 2404-2416. 6.

Brooks, A. A. Ethanol production potential of local yeast strains isolated from ripe banana

peels. Journal of Biotechnology 2008, 7 (20), 3749-3752.


Page 31 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

7.

Langmuir

Abels, C.; Thimm, K.; Wulfhorst, H.; Spiess, A. C.; Wessling, M. Membrane-based

recovery of glucose from enzymatic hydrolysis of ionic liquid pretreated cellulose. Bioresour Technol 2013, 149, 58-64. 8.

Malmali, M.; Stickel, J. J.; Wickramasinghe, S. R. Sugar concentration and detoxification

of clarified biomass hydrolysate by nanofiltration. Separation and Purification Technology 2014, 132, 655-665. 9.

Assary, R. S.; Kim, T.; Low, J. J.; Greeley, J.; Curtiss, L. A. Glucose and fructose to

platform chemicals: understanding the thermodynamic landscapes of acid-catalysed reactions using high-level ab initio methods. Phys Chem Chem Phys 2012, 14 (48), 16603-11. 10.

Luccioa, M. D.; Smithb, B. D.; T.Kidab; Borgesa, C. P.; Alvesa, T. L. M. Separation of

fructose from a mixture of sugars using supported liquid membranes. Journal of Membrane Science 2000, 174, 217-224. 11.

Ferraz, H. C.; Duarte, L. T.; Alves, M. D. L. T. L. M.; Habert, A. C.; Borges, C. P. Recent

achievements in facilitated transport membranes for separation processes. Brazilian Journal of Chemical Engineering 2007, 24 (1), 101-118. 12.

Zhao, Y. H.; Shantz, D. F. Phenylboronic acid functionalized SBA-15 for sugar capture.

Langmuir 2011, 27 (23), 14554-62. 13.

Mohapatra, S.; Panda, N.; Pramanik, P. Boronic acid functionalized superparamagnetic

iron oxide nanoparticle as a novel tool for adsorption of sugar. Materials Science and Engineering: C 2009, 29 (7), 2254-2260. 14.

Cheema, M. K. Design and Synthesis of FRET-Based Boronic Acid Receptors to Detect

Carbohydrate Clustering and Development of Diacylglycerol-Based Lipid Probesto Investigate Lipid-Protein Binding Interactions. Master, University of Tennessee2009. 15.

Yan, J.; Springsteen, G.; Deeter, S.; Wang, B. The relationship among pKa, pH, and

binding constants in the interactions between boronic acids and diols—it is not as simple as it appears. Tetrahedron 2004, 60 (49), 11205-11209. 16.

Tsukagoshi, K.; Shinkai, S. J. Specific Complexation with Mono- and Disaccharides That

Can Be Detected by Circular Dichroism. J. Org. Chem. 1991, 56, 4089-4091. 17.

Igawa, M.; Sekimoto, T.; Okochi, H. Facilitated transport of carbohydrates via

complexation with borate ion fixed on an anion-exchange membrane. Journal of Membrane Science 1995, 98, 177-180. 31 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

18.

Duggan, P. J.; Houston, T. A.; Kiefel, M. J.; Levonis, S. M.; Smith, B. D.; Szydzik, M. L.

Enhanced fructose, glucose and lactose transport promoted by a lipophilic 2-(aminomethyl)phenylboronic acid. Tetrahedron 2008, 64 (30-31), 7122-7126. 19.

Westmark, P. R.; Gardiner, S. J.; Smith, B. D. Selective Monosaccharide Transport

through Lipid Bilayers Using Boronic Acid Carriers. J. Am. Chem. Soc. 1996, 118, 11093-11100. 20.

Clark Wooten, M. K.; Koganti, V. R.; Zhou, S.; Rankin, S. E.; Knutson, B. L. Synthesis

and Nanofiltration Membrane Performance of Oriented Mesoporous Silica Thin Films on Macroporous Supports. ACS Appl Mater Interfaces 2016, 8 (33), 21806-15. 21.

Li, Y.; Wang, S.; He, G.; Wu, H.; Pan, F.; Jiang, Z. Facilitated transport of small

molecules and ions for energy-efficient membranes. Chem Soc Rev 2015, 44 (1), 103-18. 22.

Koper, I. Insulating tethered bilayer lipid membranes to study membrane proteins. Mol

Biosyst 2007, 3 (10), 651-7. 23.

Sackmann, E. Supported Membranes: Scientific and Practical Applications Science 1996,

271 (5245), 43-48. 24.

Weirich, K. L.; Israelachvili, J. N.; Fygenson, D. K. Bilayer edges catalyze supported

lipid bilayer formation. Biophys J 2010, 98 (1), 85-92. 25.

Brian, A. A.; McConnell, H. M. Allogeneic stimulation of cytotoxic T cells by supported

planar membranes. Proc. Natl. Acad. Sci. 1984, 81, 6159-6153. 26.

Worsfold, O.; Voelcker, N. H.; Nishiya, T. Biosensing Using Lipid Bilayers Suspended

on Porous Silicon. Langmuir 2006, 22, 7078-7083. 27.

Claesson, M.; Frost, R.; Svedhem, S.; Andersson, M. Pore spanning lipid bilayers on

mesoporous silica having varying pore size. Langmuir 2011, 27 (14), 8974-82. 28.

Schlipf, D. M.; Zhou, S.; Khan, M. A.; Rankin, S. E.; Knutson, B. L. Effects of Pore Size

and Tethering on the Diffusivity of Lipids Confined in Mesoporous Silica. Adv. Mater. Interfaces 2017, 1601103. 29.

Richter, R. P.; Bérat, R.; Brisson, A. R. Formation of Solid-Supported Lipid Bilayers: An

Integrated View. Langmuir 2006, 22 (8), 3497-3505. 30.

El Kirat, K.; Morandat, S.; Dufrene, Y. F. Nanoscale analysis of supported lipid bilayers

using atomic force microscopy. Biochim Biophys Acta 2010, 1798 (4), 750-65. 31.

Koganti, V. R.; Dunphy, D.; Gowrishankar, V.; McGehee, M. D.; Li, X. F.; Wang, J.;

Rankin, S. E. Generalized coating route to silica and titania films with orthogonally tilted cylindrical nanopore arrays. Nano Letters 2006, 6 (11), 2567-2570. 32 ACS Paragon Plus Environment

Page 32 of 35

Page 33 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

32.

Langmuir

Brinker, C. J.; Lu, Y. F.; Sellinger, A.; Fan, H. Y. Evaporation-induced self-assembly:

Nanostructures made easy. Advanced Materials 1999, 11 (7), 579-585. 33.

Schlipf, D. M.; Rankin, S. E.; Knutson, B. L. Pore-size dependent protein adsorption and

protection from proteolytic hydrolysis in tailored mesoporous silica particles. ACS Appl Mater Interfaces 2013, 5 (20), 10111-7. 34.

Springsteen, G.; Wang, B. Alizarin Red S. as a general optical reporter for studying the

binding of boronic acids with carbohydrates. Chemical Communications 2001, (17), 1608-1609. 35.

Koganti, V. R. Controlled evaporation driven synthesis and applications of order

nanoporous ceramic films. Ph.D Ph.D. Thesis, University of Kentucky2006. 36.

Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. Nonionic Triblock and Star

Diblock Copolymer and Oligomeric Surfactant Syntheses of Highly Ordered, Hydrothermally Stable, Mesoporous Silica Structures. J. Am. Chem. Soc. 1998, 120 (24), 6024-6036. 37.

Zhao, D.; Yang, P.; Melosh, N.; Feng, J.; Chmelka, B. F.; Stucky, G. D. Continuous

Mesoporous Silica Films with Highly Ordered Large Pore Structures. Adv. Mater. 1998, 16, 1380-1385. 38.

Koganti, V. R.; Rankin, S. E. Synthesis of Surfactant-Templated Silica Films with

Orthogonally Aligned Hexagonal Mesophase. J. Phys. Chem. B 2005, 109, 3279-3283. 39.

Lundbaek, J. A.; Collingwood, S. A.; Ingolfsson, H. I.; Kapoor, R.; Andersen, O. S. Lipid

bilayer regulation of membrane protein function: gramicidin channels as molecular force probes. J R Soc Interface 2010, 7 (44), 373-95. 40.

Kaufman, Y.; Berman, A.; Freger, V. Supported lipid bilayer membranes for water

purification by reverse osmosis. Langmuir 2010, 26 (10), 7388-95. 41.

Scott Freeman, M. H., Joan C. Sharp Biological Science; Pearson Education Canada:

Toronto, 2009. 42.

Arai, N.; Yasuoka, K.; Zeng, X. Phase diagrams of confined solutions of

dimyristoylphosphatidylcholine (DMPC) lipid and cholesterol in nanotubes. Microfluid. Nanofluid. 2013, 14 (6), 995-1010. 43.

Müter, D.; Shin, T.; Demé, B.; Fratzl, P.; Paris, O.; Findenegg, G. H. Surfactant Self-

Assembly in Cylindrical Silica Nanopores. J. Phys. Chem. Lett. 2010, 1 (9), 1442–1446. 44.

Karp, E. S.; Newstadt, J. P.; Chu, S.; Lorigan, G. A. Characterization of lipid bilayer

formation in aligned nanoporous aluminum oxide nanotube arrays. J Magn Reson 2007, 187 (1), 112-9. 33 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

45.

Page 34 of 35

Wattraint, O.; Sarazin, C. Diffusion measurements of water, ubiquinone and lipid bilayer

inside a cylindrical nanoporous support: a stimulated echo pulsed-field gradient MAS-NMR investigation. Biochim Biophys Acta 2005, 1713 (1), 65-72. 46.

Eeman, M.; Deleu, M. From biological membranes to biomimetic model membranes.

Biotechnol. Agron. Soc. Environ. 2010, 14 (4), 719-736. 47.

Kucerka, N.; Nieh, M. P.; Katsaras, J. Fluid phase lipid areas and bilayer thicknesses of

commonly used phosphatidylcholines as a function of temperature. Biochim Biophys Acta 2011, 1808 (11), 2761-71. 48.

Gardikis, K.; Hatziantoniou, S.; Viras, K.; Wagner, M.; Demetzos, C. A DSC and Raman

spectroscopy study on the effect of PAMAM dendrimer on DPPC model lipid membranes. Int J Pharm 2006, 318 (1-2), 118-23. 49.

Barry, J.; Fritz, M.; Brender, J. R.; Smith, P. E. S.; Lee, D.-K.; Ramamoorthy, A.

Determining the Effects of Lipophilic Drugs on Membrane Structure by Solid-State NMR Spectroscopy: The Case of the Antioxidant Curcumin. J. Am. Chem. Soc. 2009, 131 (12), 4490– 4498. 50.

Gardikis, K.; Hatziantoniou, S.; Viras, K.; Demetzos, C. Effect of a bioactive curcumin

derivative on DPPC membrane: A DSC and Raman spectroscopy study. Thermochimica Acta 2006, 447 (1), 1-4. 51.

Fritzsching, K. J.; Kim, J.; Holland, G. P. Probing lipid-cholesterol interactions in

DOPC/eSM/Chol and DOPC/DPPC/Chol model lipid rafts with DSC and (13)C solid-state NMR. Biochim Biophys Acta 2013, 1828 (8), 1889-98. 52.

McMullen, T. P. W.; Lewis, R. N. A. H.; McElhaney, R. N. Differential scanning

calorimetric study of the effect of cholesterol on the thermotropic phase behavior of a homologous series of linear saturated phosphatidylcholines. Biochemistry 2002, 32 (2), 516-522. 53.

Voskuhl, J.; Ravoo, B. J. Molecular recognition of bilayer vesicles. Chem Soc Rev 2009,

38 (2), 495-505. 54.

R, L.; M, K.; SI, C. The formation and annealing of structural defects in lipid bilayer

vesicles. Biochim Biophys Acta. 1976, 443 (3), 313-330. 55.

Pande, A. H.; Qin, S.; Tatulian, S. A. Membrane fluidity is a key modulator of membrane

binding, insertion, and activity of 5-lipoxygenase. Biophys J 2005, 88 (6), 4084-94. 56.

Lande, M.; Donovan, J.; Zeidel, M. The relationship between membrane fluidity and

permeabilities to water, solutes, ammonia, and protons. J Gen Physiol. 1995, 106 (1), 67-84. 34 ACS Paragon Plus Environment

Page 35 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

57.

Langmuir

Springsteen, G.; Wang, B. A detailed examination of boronic acid–diol complexation.

Tetrahedron 2002, 58 (26), 5291-5300. 58.

Paugam, M.-F.; Riggs, J. A.; Smith, B. D. High fructose syrup production using fructose-

selective liquid membranes. Chem. Commun. 1996, 1 (22), 2539-2540.

TABLE OF CONTENTS/ABSTRACT GRAPHIC


Lipid Pore-Filled Silica Thin-Film Membranes for ... - ACS Publications

Recommend Documents