Biomimetic membrane for desalination and water reuse 4

Figure 5.1 The schematic presentation of fabrication procedures for the Poly-L-lysine stabilized liposomes with Aquaporin Hydrolyzed PAN membrane polyanion polycation Magnetic nanopart

Chapter 5 Highly Permeable Aquaporin-embedded

Biomimetic Membrane Featured with a Magnetic-aided Approach

5.1 Introduction

Kumar and his co-workers firstly reconstituted AqpZ into the block-copolymer vesicles, which demonstrated great potential of AQP-based biomimetic membranes [38] Since then, several endeavors have been made to design the AQP-incorporated biomimetic membrane for various water purification applications, like forward osmosis[83-86], Nanofiltration [87, 119, 140, 141], and reverse osmosis [88, 142] Several efforts were made to fabricated AQP-reconstituted black lipid/polymer membrane [82-85, 142] or supported lipid bilayer [88, 143], but only limited progress has been made Wang et al covalently bond the AqpZ-embedded block-copolymer film onto a gold-coated track-etched film to form a pore-spanning membrane, and demonstrated the function of AqpZ in a forward osmosis setup for the first time [86] However, the mechanical stability of these thin films was not sufficient under normal

FO or NF testing conditions when there is strong surface shear or high hydraulic pressure In recent works [87, 140], stability was largely improved by cross-linking the AqpZ-incorporated films with UV, and nanofiltration performance was investigated Nevertheless, the membrane selectivity of these membranes was well below expectation probably due to low AqpZ working efficiency in the membrane

The challenge here is how to fabricate a stable and defect free membrane, and maximize the potential of AQP In the previous study, an AqoZ-embedded mixed-matrix membrane has been developed using the layer-by-layer (LbL) polyelectrolyte

adsorption method [141] The AqpZ-incorporated liposomes, which function as

highly water permeable dispersed phase, have been embedded in a polyelectrolyte

LbL matrix The negatively charged liposomes are protected with a PLL shell, which

provides stability to the liposomes especially when they are embedded in the LbL

films We also observed that the number of liposomes adsorbed on the membrane was

closely related to the amount of charged lipids carried in the liposomes The

membrane has been proven to be strong enough to endure 4 bar hydraulic pressure

and surface agitation However, the designed membrane suffered from a low

liposome-embedding efficiency, and the maximum vesicular fraction in the

membrane is only approximately 20%

The aim of this work is to further improve the embedding efficiency of the

AqpZ-incorporated vesicles in the membrane, by encapsulating magnetic nanoparticles

(MNPs) into the vesicles and utilizing the magnetic force to accelerate the adsorption

of the vesicles on the polyelectrolyte film The MNP-encapsulated liposomes, also

named as magnetic liposomes, has been studied in drug delivery [144, 145],

controlled release[146], or as contrast agents for magnetic resonance imaging [147]

The surface modification of MNPs by hydrophilic molecules avoids aggregation and

precipitation of the particles and allows a stable suspension in biological applications

[148-150]

The membrane fabrication process is schematically shown in Figure 5.1 On top of a

negatively charged membrane substrate, PAH was firstly deposited to form a

polycation layer, and a blend of PAA and PSS was then deposited to form a

polyanion layer Afterward, driven by a strong magnet at the bottom of the

membrane, the PLL-encapsulated magnetic liposomes are precipitated onto the polyanion layer The liposome layer is further stabilized with another PSS/PAA layer

to stabilize this mixed-matrix membrane The magnetic force-driven approach is effective and straightforward In this work, magnetic liposome characterization, membrane morphology study and forward osmosis performance of this biomimetic membrane are discussed

Figure 5.1 The schematic presentation of fabrication procedures for the

Poly-L-lysine stabilized liposomes with Aquaporin Hydrolyzed

PAN membrane

polyanion

polycation

Magnetic nanoparticles (MNPs)

from Tong-Hua Synthesis Fiber Co Ltd (Taiwan) A 10-histidine residual tagged

AqpZ used in this work was obtained from Biochemistry department in National

University of Singapore, and the synthesis of AqpZ followed the procedures in

Borgnia et al’s work [37] Ultrapure water was produced by the Millipore Reference

A+ system (Merck Millipore, USA) A 10 mM Tris buffer at pH=7.5 with 15 mM

NaCl was used throughout this study

5.2.2 Magnetic nanoparticle synthesis

Precursors were prepared by dissolving 40mmol iron chloride (FeCl3.6H2O) and

120mmol sodium oleate in a solvent mixture of 80mL ethanol, 60mL DI water and

140mL hexane Heated to 70 °C and kept for 4 hours The complex was then

separated from the solvent The MNPs were synthesized using the thermal

decomposition method 2mmol complex, 2mmol oleic acid and 6mmol oleyl alcohol

were dissolved in 10g diphenyl ether, heated to 250°C and kept for 30min Ligand

exchanging was applied to graft PMAA on the surface of MNP

5.2.3 Magnetic liposome preparation

6.5mg POPC, 3mg POPG, and 0.5mg cholesterol were dissolved in chloroform 0.5

mole% Rhodamine-PE were added when necessary The dry lipid film was made by

removing the chloroform using a rotary-evaporator, followed by overnight vacuum A

100µL MNP solution (40mg/mL) was added to swell the dry film before bulk

rehydration by 2mL Tris Buffer Small unilamellar vesicles (SUV) with a uniform

size were produced by extruding the suspension through a polycarbonate Nuclepore

track-etch membrane (Whatman, UK) that had a pore size of 200 nm For AqpZ

reconstitution experiments, an AqpZ stock solution was added during the film

rehydration step and the mixture was agitated for at least 8 hours Bio-Beads were then added into the mixture stepwise to remove the detergent The suspension was protected with high purity argon throughout the experiment

To remove the unencapsulated MNPs, a magnetic liposome suspension of 1mL was added into a 30mm x 200mm2 chromatography column filled with Sephadex G100 The suspension was eluted with Tris Buffer at a flow rate of 0.8mL/min The eluted sample was passed through a UV spectrometer and detected at the wavelength of 210nm Fractions of 2mL elution which corresponded to the magnetic liposomes were collected, while the rest of the eluted sample was discarded

The collected magnetic liposomes were then stabilized with PLL according to the following procedures A 2mL liposome solution was added at equal volumes to a 0.5 mg/mL PLL solution in Tris buffer dropwise, while the PLL solution was stirred at a speed of 950 rpm The mixing process was completed within 3 minutes The resultant solution was concentrated to 1mL using a 50mL centrifugal filter (100,000MWCO, Ultracel®, Millipore) A Zetasizer Nano ZS instrument (Malvern, UK) was employed

to characterize the vesicle/nanoparticle size distribution

5.2.4 Field emission transmission electron microscopy (FETEM)

Field emission transmission electron microscopy (JEOL, JEM-2100F, Japan) was used to investigate the MNPs and liposomes with and without MNP encapsulation Before imaging, the liposome solution was diluted to 0.5 mg/mL with Tris buffer and dropped on copper grids coated by an ultrathin carbon film for 15 min and then rinsed

by ultrapure water dropwise The samples were air dried for 30 min before the

FETEM imaging

5.2.5 Liposome-embedded LbL membranes formation

Flat sheet PAN substrates were prepared by casting an 18 wt% PAN/NMP solution

directly on a Teflon plate using a 200 µm casting knife The membranes were then

quickly immersed in a water bath to induce phase inversion and then soaked in

deionized water overnight to remove all traces of NMP The PAN membranes were

later hydrolyzed with a solution containing 1 M NaOH for 1.5 hours at 45°C to

generate negative charges on the membrane surface After the hydrolysis, the

membrane was washed with excess volumes of ultrapure water and used within 3

days

To prepare the LbL membrane with magnetic liposomes, a PAH solution (1g/L) was

first deposited onto the surface of the hydrolyzed PAN, followed by a PSS/PAA

blend solution (1g/L) It took 20 min to deposit each layer and the membrane was

rinsed with ultrapure water three times after each deposition Next, the PLL-stabilized

magnetic liposome suspension was deposited on the top surface of the membrane

with a cubic magnet (0.7T, 80kgf) placed beneath the membrane for 1 hour Finally,

one more layer of PSS/PAA was deposited on top of the PLL-liposomes to stabilize

the liposomes A membrane area of 78.5mm2 was prepared and used for further

studies

5.2.6 Vesicle adsorption study by confocal Laser scanning microscope (CLSM)

To study the vesicle adsorption process, a glass coverslip (12×12mm) was immersed consecutively in 2% Hellmanex solution (60°C), 0.01 M SDS solution, and 0.1 M HCl The glass coverslip was rinsed intensively with ultrapure water after 20 min incubation in each solution Three bilayers of PAH-PSS/PAA were deposited onto the glass coverslip using the same method described above Then, the substrate was incubated with magnetic liposome solution for 15, 30, 45, or 60 min in presence or absence of the magnetic field The glass coverslip was rinsed with Tris buffer to remove the unadsorbed liposomes before imaging

The CLSM images were taken with a Nikon A1 confocal scanning system, equipped with 60× oil immersion lens The excitation laser was set at wavelength of 488 nm and emission was collected from 590 to 650 nm The obtained images were processed and analyzed using ImageJ 1.46r

5.2.7 Forward osmosis measurement

The prepared membrane was placed in a FO testing cell with a draw solution chamber and a feed solution chamber at each side of the membrane The active layer of the membrane was facing the draw solution chamber in all the FO experiments Solutions

of both sides flow co-currently through the cell at a flow rate of 30mL/min in the

cross-flow mode The water flux J w was calculated using the equation shown below

!! =∆!!!!

!!∆!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(10)

where V f is the volume of the feed and ∆C f is the change of salt concentration in the

feed

The membrane was tested by using 0.3M sucrose as draw solution and 200 ppm NaCl

as feed solution The conductivity of the draw solution is very close to DI water,

which can be ignored Therefore, the membrane selectivity can be estimated with the

following equation

!"#"$%&'&%( = !(∆!!− !!!!!!)/!!

!!!!!!/!! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(11)

where C d1 is the final salt concentration of the draw solution in g mL-1; V d1 is the final

volume of draw solution and initial volume of the feed, respectively, in mL The ΔW d

is the weight increase of the draw solution, in g M w and M s are the molecular weights

of water (18.02 g/mol) and NaCl (58.44 g/mol) , respectively

5.3 Results and discussion

5.3.1 Characterization of magnetic liposomes

Figure 5.2 The FETEM images of (a) free MNPs, (b) MNP-encapsulated

liposomes and (c) intact liposomes Scale bar: 100nm The insert in (a) is the

MNP size distribution measured by dynamic light scattering

MNPs were synthesized using the thermal decomposition method and grafted with poly(methacrylic acid) (PMAA) The image taken by field emission transmission electron microscopy (FETEM) shows that the particle size ranges from 2 to 5nm (Figure 5.2a), and a similar result is observed by dynamic light scattering (the Insert

in Figure 5.2a) To produce the magnetic liposomes, a dry lipid film consisting of

a

c b

POPC, POPG and 5 mol% cholesterol were swelled with the MNP solution before

conducting film rehydration Unilamellar vesicles with a uniform size were formed

by extruding the magnetic liposome suspension through a 200nm polycarbonate

Nuclepore track-etch membrane The encapsulated MNPs in the liposomes can be

clearly observed from the FETEM images (Figure 5.2b), while the intact liposomes

were shown as hollow spheres (Figure 5.2c)

The osmotic water permeability of liposomes was studied using a stopped-flow light

scattering apparatus Since an increase in light scattering signal corresponds to a

reduction in vesicle size after the hypertonic osmotic shock, the osmotic permeability

of the vesicles is directly related to the rate of increase in light scattering signal

Figure 5.3a displays the normalized signals of magnetic liposomes with different

AqpZ-to-lipid ratios The AqpZ incorporated vesicles have a much rapid increment in

light scattering signal as compared to the control one In addition, the higher AqpZ

content, the faster the signal response Similar to the intact liposomes (i.e., without

the MNP encapsulation), the permeability of magnetic liposomes reaches the

maximum at an AqpZ incorporation ratio of 2% (or at AqpZ-to-lipid weight ratio of

1:50) but declines at a higher incorporation ratio (Figure 5.3b) The AqpZ

incorporated liposomes comprising MNPs exhibit 60-70% permeability of those

intact liposomes There are two possible reasons One of them might be the formation

of the MNP/AqpZ complex in the liposome suspension The other might be due to the

presence of MNPs within liposomes The former arises from the fact that the PMAA

grafted on the particle surface may entangle with AqpZ and prohibit its effective

reconstitution As a result, the water permeability is reduced Similarly, the latter

might delay the response of vesicles during the osmotic shock

Figure 5.3 Stopped-flow light scattering result (a) The normalized light

scattering signal of magnetic liposomes with different AqpZ-to-lipid weight

ratios, (b) the comparison of the calculated permeability P f of intact liposomes

and MNP-encapsulated liposomes at different AqpZ-to-lipid weight ratios

Therefore, we studied the activation energies of water transport calculated from the Arrhenius plot to confirm our hypotheses The Arrhenius activation energies of water transport across both the intact and magnetic liposomes with and without AqpZ

incorporation were compared in Figure 5.4 The activation energy E a can be

estimated from the plot of the water flow kinetics k against temperature T using the

following equation

! = !!!! ! /!"

or ln ! = ! − !!/!"!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(11) Without AqpZ incorporation, the activation energy of water transport across magnetic

liposomes is 13.9 kcal/mol, while the value drops significantly to 3.85 kcal/mol for

magnetic liposomes containing 1% AqpZ The activation energy of AqpZ-incorprated

magnetic liposomes is slightly higher than that of intact liposomes (3.5 kcal/mol),

probably due to the presence of MNPs in the magnetic liposomes

Figure 5.4 The Arrhenius plots of water flow across liposomes The kinetics of

water transport were studied in intact liposomes (blue) without AqpZ (regression

equation: ln k = 25.76-6.99/T) and intact proteoliposomes (AqpZ-to-lipid weight

ratio: 1:100, regression equation: ln k = 10.63-1.94/T), and in magnetic

liposomes (red) without AqpZ (regression equation: ln k = 25.96-7.06/T) and

with AqpZ incorporation (AqpZ-to-lipid weight ratio: 1:100, regression

equation: ln k = 10.91-1.76/T).

The unencapsulated MNPs were removed from the magnetic liposomes using a gel

chromatography column (GPC, Sephadex G100) The eluted magnetic liposomes

were then stabilized with PLL using the method described in previous work The PLL

layer could protect liposomes from disruption in the LbL deposition process Without this layer, adsorption of POPG/POPC/Chol liposomes directly onto the opposite charged polyelectrolyte film would result in vesicle rupture (Figure 5.5)

Figure 5.5 The FESEM images of ruptured liposomes on membrane surface

The hydrolyzed PAN membrane was firstly covered with a positively charged PAH layer, and then incubated with negatively charged POPC/POPG/Chol liposomes for two hours The liposomes were adsorbed onto the oppositely charged surface, but the electrostatic force induced the rupture of vesicles

Figure 5.6 Study the liposome adsorption process by CLSM Representative

CLSM images of liposome adsorbed LbL film at different liposome deposition

time, (a) in presence, or (b) in absence of the magnetic driven force (c) The

amount of adsorbed vesicles estimated by counting the number of bright dots in

the CLSM images is plotted against the deposition time

5.3.2 Characterization of the LbL membrane

The adsorption process for magnetic liposomes was analyzed using a confocal laser

scanning microscope (CLSM) Three bilayers of PAH-PSS/PAA were deposited on a

glass coverslip, and then the PLL-stabilized magnetic liposome suspension was

incubated on the polyelectrolyte film in presence or absence of magnetic field for

certain time The coverslip was rinsed with Tris buffer before imaging Since the

liposomes were doped with 0.5 mol% Rhodamine-PE, they were displayed as tiny

and bright dots in CLSM images (Figure 5.6a and b) The coverage of the dots

Deposition time (min)

With magnetic field without magnetic field

30min

b

c

increases as the liposome incubation time is increased The number of dots are counted and plotted against the deposition time With presence of the magnetic field, more dots were observed under CLSM, and the surface saturation time is approximately 45 minutes, shorter than the process without the magnetic field (Figure 5.6c) In the absence of the magnetic field, the liposome adsorption is solely based on electrostatic interaction, such that it is less efficient than the magnetic-aided liposome adsorption process

Figure 5.7 The elution profile of magnetic liposome suspension separated by

the cross-linked dextran gel Sephadex G100 and detected by a UV spectrometer

In the magnetic-aided vesicle adsorption process, the unencapsulated MNPs must be completely removed Because under the magnetic field, PMAA modified MNPs could rapidly form complexes with PLL and precipitate onto the PSS/PAA surface as

a thin layer, the PLL-stabilized liposomes can no longer be adsorbed Therefore, the pre-condition to have a successful deposition of magnetic-aided liposomes is to obtain two distinctive peaks in the GPC runs (one for liposomes, one for MNPs) (Figure 5.7) by avoiding particle aggregation of MNPs

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Figure 5.8 The FESEM images of (a) the top surface and (b) the bottom

surface of a blank 18% PAN membrane, and (c) the top surface and (d) the

cross-section of the PAN membrane with a deposition of a PAH-PSS/PAA

bilayer

Table 5.1 Pure water permeability (PWP), molecular weight cut-off (MWCO)

and surface roughness of the hydrolyzed PAN membrane substrate

Hydrolyzed PAN Membrane

PWP 25.2±2.2 Lm-2h-1bar-1

Figure 5.9 The FESEM image of the membrane surface deposited with magnetic

liposomes

The substrate membrane was prepared by casting an 18% PAN/NMP dope solution

on Teflon plates with a 200 µm casting knife The as-cast membrane has a relatively dense top surface and a very porous bottom surface (Figure 5.8a, b) The membrane was hydrolyzed with a NaOH solution to generate negative charges on the membrane surface The hydrolyzed PAN membrane had the pure water permeability of 25.2 Lm-

2h-1bar-1 and the surface roughness of approximately 2.33 nm (Table 5.1) To fabricate the magnetic liposome embedded biomimetic membrane, one bilayer of PAH-PSS/PAA was deposited on the top surface of the hydrolyzed membrane, which formed a selective layer of approximately 100nm (Figure 5.8c, d) Then, the vesicle suspension was deposited onto the top surface while a strong magnet was placed underneath the membrane to facilitate the precipitation of magnetic liposomes onto the membrane surface To stabilize the liposome layer, another layer of PSS/PAA was deposited on the top Figure 5.9 shows the membrane morphology by FESEM In comparison with our previous work (see Figure 4.7), the liposome coverage on the