Biomimetic membrane for desalination and water reuse 3

Chapter 4 Aquaporin-embedded Mixed-Matrix Membrane: A Layer-by-Layer Self-assembly Approach 4.1 Introduction Among different water purification techniques, reverse osmosis and nanofilt

Chapter 4 Aquaporin-embedded Mixed-Matrix Membrane:

A Layer-by-Layer Self-assembly Approach

4.1 Introduction

Among different water purification techniques, reverse osmosis and nanofiltration

become more and more popular for drinking water production because of their

effectiveness in removing low molecular weight impurities such as small organic

compounds and ions [116] However, reverse osmosis and nanofiltration are still

energy-intensive processes, and it is crucial to develop high-performance membranes

with new materials AQPs in biological membranes are water channel proteins that

are precisely engineered by nature [36, 106] The exceptional selectivity and

permeability that AQP demonstrates towards water molecules make it a potential

candidate in designing high-performance membranes for water purification Among

the AQP family, AqpZ has attracted particular attention in biomimetic membrane

research [88, 106, 117, 118] The functional reconstitution of AqpZ in both lipid

bilayer and block copolymer membranes has been demonstrated previously [37, 38]

Kumar et al have reported that the water permeability of an AqpZ incorporated

biomimetic membrane could possibly reach 80 times higher than current commercial

reverse osmosis membranes [38] Motivated by this finding, several attempts have

been made to develop AqpZ embedded biomimetic membranes for different

applications such as forward osmosis[86], reverse osmosis [88], or nanofiltration [87,

119] However, even through some of previous works have demonstrated water

transport ability of AqpZ in their membranes, the membrane fabrication in a larger

scale remains a challenge, especially when the membrane stability and integrity have

to be maintained for long-term operations

In this work, a practical design of the AqpZ embedded mixed-matrix membrane (MMM) was introduced using the layer-by-layer (LbL) adsorption approach The LbL matrix has offered a stable and compatible environment for AqpZ-reconstituted vesicles and this design may become one of the potential solutions for fabricating the AqpZ-assisted water purification membranes The LbL film is formed by alternatively depositing polycations and polyanions onto a charged substrate It has been studied in a wide variety of research topics As a versatile and convenient strategy, previous research works have demonstrated the functionalization of LbL films via the incorporation of various compounds of interest such as DNA [120], proteins and enzymes [121, 122], nanoparticles [123], and even liposomes [124, 125] The embedding of liposomes into the polyelectrolyte multilayers has been investigated for the development of drug release or enzymatic nano-reactor systems [124-126] Polyelectrolyte encapsulated liposomes have demonstrated better chemical [127] and mechanical stabilities [128] than intact liposomes Gentle polyelectrolytes such as poly-L-lysine (PLL) have to be used for liposome encapsulation because strong polyelectrolytes could disrupt the lipid membranes during the adsorption process Previous works have shown that PLL-covered liposomes are able to be absorbed onto various polyanion films [126] In our current work, the AqpZ reconstituted proteoliposomes are first stabilized with PLL and then embedded within the LbL film that is formed on a porous membrane for nanofiltration studies The adsorption of PLL onto the proteoliposome surface may not damage the incorporated protein because PLL is a biocompatible material composed of amino acids [129]

To fabricate the LbL film, polyallylamine hydrochloride (PAH) was used to form the

polycation layer, while a blend of polyacrylic acid (PAA) and polystyrene sulfonate

(PSS) was used to form the polyanion layer Since PAA is a weak polyanion but has

good hydrophilicity, while PSS is a strong but rather hydrophobic polyanion, the

blended PSS/PAA system provides both hydrophilicity and relatively strong

interactions with polycation layer [130] It has also been reported that LbL films

formed by a combination of PAH-PSS/PAA have a lower surface roughness [131]

and a better ion rejection [132] than films formed by a single polyanion (PAA or

PSS) paired with PAH The positively charged PLL-covered liposomes will be

encapsulated within the LbL film to form a nano-composite MMM, as shown in

Figure 4.1 The polyelectrolyte layers act as a continuous matrix phase and the AqpZ

incorporated liposomes function as the highly permeable dispersed phase The aim in

this paper is to explore the feasibility of devising the AqpZ-embedded MMM for

pressure-driven water purification, and also to demonstrate the functionality of AqpZ

in the MMM

Figure 4.1 The schematic presentation of the formation procedures for the

liposome-embedded LbL membrane

Poly-L-lysine encapsulated liposome with water channel proteins Hydrolyzed PAN membrane

4.2 Materials and methods

4.2.1 Materials

gly-cero-3-phosphocholin (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt) (POPG), cholesterol were

purchased from Avanti Polar Lipids Hydrochloric acid (HCl), sodium chloride (NaCl), magnesium chloride (MgCl2), sodium hydroxide (NaOH), n-methyl-2-pyrrolidone (NMP), PLL, PAA, PSS, PAH, and glutathione (C10H17N3O6S, MW 307.33) were products of Sigma-Aldrich (USA) Bio-Beads SM-2 absorbents and tris(hydroxymethyl)aminomethane (Tris) were purchased from BIO-RAD (USA) Polyacrylonitrile (PAN) for substrate preparation was obtained from Tong-Hua Synthesis Fiber Co Ltd (Taiwan) Grade V1 mica was purchased from SPI Supplies (USA) 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

4.2.2 Vesicle preparation and characterization

POPC and POPG were mixed for vesicle formation at certain molar ratios together with 5% (w/w) cholesterol (Chol) A multilamellar vesicle suspension in a Tris buffer was prepared using the film rehydration method Small unilamellar vesicles (SUV) with a uniform pore size were produced by extruding the suspension through a polycarbonate Nuclepore track-etch membrane (Whatman, UK) that had a pore size

of 100 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 completely The suspension was protected with high purity argon throughout the experiment

The intact POPC/POPG/Chol liposomes were then stabilized with PLL according to

the following procedures A 1 mg ml-1 liposome solution was added at equal volumes

to a 0.5 mg ml-1 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

Lipo8, lipo15 and lipo30 refer to samples containing 8, 15 and 30 percent of POPG in

the liposomes, respectively Similarly, PLL-lipo8, PLL-lipo15 and PLL-lipo30

represent PLL-covered liposomes with their respective POPG content A Zetasizer

Nano ZS instrument (Malvern, UK) was employed to characterize the vesicle size

distribution and zeta potential

Field emission transmission electron microscopy (FETEM: JEOL, JEM-2100F,

Japan) was used to image POPC/POPG/Chol liposomes before and after PLL

adsorption Before imaging, the liposome solution were diluted to 0.5 mg ml-1 with

Tris buffer and dropped on ultrathin carbon film coated copper grids for 15 min and

then rinsed with ultrapure water dropwise The samples were air dried for 30 min

before the FETEM imaging

4.2.3 Vesicle permeability measurement using stopped-flow

Water permeabilities of both the intact liposomes and the PLL-stabilized liposomes at

different AqpZ-to-lipid ratios were investigated using a stopped-flow spectrometer

(Applied Photophysics, Chirascan, UK) By rapidly mixing the liposome solution

with a hypertonic buffer solution containing 0.6 mol L-1 sucrose, water would diffuse

from the vesicles into the buffer, causing the vesicles to experience a sudden

shrinkage To improve the signal to noise ratio, all the experiments were conducted at

a temperature of 8°C Data have been fitted to Equation 3 to estimate the rate constant

k The final osmotic permeability (P f) of the vesicles was calculated by Equation 4

4.2.4 Liposome-embedded LbL on a mica surface

Three sets of PAH-PAA/PSS were prepared on newly cleaved mica surfaces by depositing solutions containing 1 g L-1 of PAH and 1 g L-1 of PSS/PAA (mixed at a weight ratio of 1:1) alternately onto the mica surface and rinsing with pure water after each deposition step Each deposition step lasted for 15 min PLL-covered liposomes containing 15% or 30% POPG were then deposited on top of the third PSS/PAA layer and incubated for 2 hours After the incubation, the mica was rinsed with buffer to remove unbound vesicles and then covered with one more layer of PSS/PAA The films were imaged by an AFM that was operated in an acoustic alternating current mode The samples were also scanned by OLYMPUS cantilevers (OMCL-TR400PSA, resonance frequency of 11 kHz and a typical force constant of 0.02 N m

-1) in aqueous solutions (prepared with Milli-Q water) at room temperature (23 ± 1°C)

4.2.5 Liposome-embedded LbL membranes for nanofiltration

Flat sheet PAN substrates were prepared by casting a 12 wt% PAN/NMP solution directly on glass plates with a 150 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 50°C to generate negative charges on the membrane surface After hydrolysis, the membrane was washed with excess volumes of ultrapure water and used within one week

To prepare the nanofiltration membrane, a PAH-PSS/PAA bilayer was first deposited

onto the surface of the hydrolyzed PAN for 15 min The membrane was washed with

ultrapure water for 1 min after each layer was deposited The PLL-liposome solution

was incubated on top of the PSS/PAA layer for 2 hours and gently spray-washed with

ultrapure water Finally, one more layer of PSS/PAA was deposited on top of the

PLL-liposomes for stabilizing the liposomes A membrane area of 78.5mm2 was used

for nanofiltration

4.2.6 Nanofiltration studies

The nanofiltration tests were conducted using dead-end permeation cells The

measurements for the pure water permeability (PWP) were conducted at 23 ± 1°C in

terms of L m-2 h-1 bar-1, with the selective layers facing the feed at a trans-membrane

pressure of 4 bar The PWP was calculated from Equation 5 The rejections of

200ppm MgCl2 and glutathione solutions in ultrapure water were tested using a

surface mixing speed of 700 rpm at a trans-membrane pressure of 4 bar The salt

rejection was calculated using the Equation 6 Each reported data was an average of

at least three different samples The solute permeability (B) was estimated using the

following equation [133]:

! = ! ∆! − ∆! 1 − !

! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(7) where A is the intrinsic water permeability or PWP, ΔP is the trans-membrane

hydraulic pressure and Δπ is the osmotic pressure from the feed solution

On the basis of the dissociation constants of glutathione, i.e., the pK a values, the

fraction of glutathione present in different ionization states at different pH values can

be expressed by the following the Henderson-Hasselbalch equation shown below

[134] The concentration of glutathione was measured by a total organic carbon analyzer (Shimazu, TOC ASI-5000A, Japan)

pH = p!!+ log !([!"#$#%!!""#$%&']

[!"#$#%!!"#"$] )!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(8)

4.3 Results and discussion

4.3.1 PLL adsorption on liposomes

To study the vesicle properties at different POPG content, the mean vesicle diameters and the zeta potentials of both the intact and the PLL-stabilized liposomes were measured using a Zetasizer apparatus, as shown in Table 4.1 For the intensity-weighted mean diameters of all the samples, the polydispersity indexes are always less than 0.15, indicating that the particle sizes follows a narrow mono-dispersed distribution Upon mixing the intact liposomes with PLL, the mean vesicle sizes increase because PLL is adsorbed onto the negatively charged lipid bilayer The zeta potential becomes more negative as the content of POPG is increased in the liposome because of the accumulation of a higher charge density as POPG is a negatively charged lipid Consequently, as the POPG content is increased, more PLL molecules could be adsorbed onto the liposome surface, producing PLL-covered liposomes with

a higher zeta potential However, despite the fact that more PLL molecules are adsorbed onto the liposomes that possess higher POPG content, a stronger electrostatic interaction induces a more extended and a more compact molecular structure of PLL Thus, the thickness of the PLL layer is reduced with an increasing amount of POPG in the liposomes POPC/POPG/Chol liposomes before and after PLL adsorption were imaged with FETEM and shown in Figure 4.2 In accordance with the dynamic light scattering results, bolder vesicle walls can be observed from the PLL-liposome images because of the PLL adhesion outside the lipid bilayer

Figure 4.2 FETEM images of (a) intact liposomes, and (b) PLL-covered

liposomes The liposomes content 15% POPG

Table 4.1 The zeta potential and the intensity-weighted mean hydrodynamic

diameter of the intact liposomes and the PLL-covered liposomes The thickness

of the adsorbed PLL layer was calculated from the diameters obtained Three

different liposome samples were used to obtain the mean and the standard

deviation

Sample Mean Diameter* (nm) Zeta potential* (mV) PLL thickness** (nm)

* Three different liposome samples were used to obtain the mean and the standard deviation

** The thickness of the adsorbed PLL layer was calculated from the diameters obtained

200#nm#

200#nm#

200#nm#

100#nm#

4.3.2 Water permeability measurement by stopped-flow

Stopped-flow light scattering was used to study the water permeability of POPC/POPG/Chol vesicles at different AqpZ incorporation ratios The measurements were conducted at 8°C to reduce the noise level Only the results from the liposomes containing 30% POPG are presented here because the POPG content did not influence the function of AqpZ An increase in the light scattering signal corresponds

to a reduction in the vesicle size after hypertonic osmotic shock Moreover, the increased rate that is represented by the rate constant k is directly related to the osmotic permeability of the vesicles (Equation 4) The normalized signals of the intact liposomes with different AqpZ-to-lipid ratios are compared in Figure 4.3(a) With the incorporation of AqpZ, a rapid increment in light scattering signal can be observed in the first 40 ms However, the control liposomes that were prepared without AqpZ took approximately 1.5 seconds to reach steady state, which was a much slower response in comparison with the AqpZ-incorporated vesicles The osmotic permeabilities of the intact liposomes and the PLL-covered liposomes are shown in Figure 4.3(b) From the stopped-flow results, the PLL adsorption does not appear to have much effect on lipid membrane permeability or AqpZ functionality

As the AqpZ-to-lipid ratio was increased from 0 to 1:50, the permeability increased approximately linearly from 13 µm/s to 853 µm/s However, a further increase in the vesicle permeability was not observed at higher AqpZ incorporation ratios This finding is similar to the results that were presented in the previous work [37], and the possible reason could be the interference of a large amount of detergent in the AqpZ incorporation process [38]