Aquaporins, which allow water to enter or exit a cell, are one example of which type of protein?

4.2.3 Aquaporins (AQPs)

AQPs or biological water channel proteins are one of the best membrane proteins due to their unique structural and functional properties. Almost no solutes can penetrate AQPs and AQPs are suitable for desalination membrane [59]. However, there are several challenges: (1) it is difficult to make free-standing AQPs in a fabricated membrane; (2) spontaneous AQPs are chemically unstable; and (3) the low mechanical strength of AQPs is also one of the big disadvantages [60]. To solve these issues, studies have been conducted and several methods have been developed [60,61].

11.4 Aquaporins

Aquaporins are transmembrane proteins that regulate the flow of water into and out of cells. For many years, it was thought that water movement into and out of the cell was (1) not regulated in any manner and (2) could be accounted for by simple diffusion across the cell membrane. However, the rapid movement of water in aqueous humor formation could not be described by simple diffusion. In the early 1990s, aquaporins were discovered, and it was found that they can selectively control water movement into and out of cells. One of the critical functions of aquaporins is that whereas they allow the passage of water they prevent the passage of ions. If aquaporins allowed ions through their channels, all ion concentration gradients across the cell membrane would approach zero (i.e., all ions would be in equilibrium with themselves) and therefore the cells would not be able to perform many of the critical functions that depend on the concentration gradient of various molecules (e.g., many transport processes that rely on an electrochemical gradient would fail). Furthermore, if ions were allowed to pass through aquaporins, the amount of energy that cells would expend on maintaining the necessary ion concentration gradients across the cell for the cell to function as we know, would exceed the amount of energy produced during cellular respiration (if the cell could even maintain a concentration gradient under these conditions).

Aquaporins are found in a high concentration in the epithelial cells that produce aqueous humor (as well as other epithelial cells that allow water to move readily across their membrane, e.g., epithelial cells in the kidney). These pores allow water molecules through in a single file. To understand how aquaporins regulate water movement into a single file and prevent the movement of other ions, it is important to understand the three-dimensional (3D) structure of the aquaporin protein. An aquaporin channel is composed of six transmembrane α-helices, with both the amino and carboxyl terminal on the cytoplasmic side of the membrane. Two of the five loops that connect the six transmembrane helices are extremely hydrophobic. One of these loops is on the intracellular side of the membrane, and the other is on the extracellular side of the membrane. The two hydrophobic loops contain a three-amino-acid sequence, termed the asparagine–proline–alanine (NPA) motif. The NPA motif folds back into the aquaporin channel created by the six transmembrane helices. In 3D space, the folding back of these two domains resembles an hourglass shape (or a bottle neck for flow). This hourglass constriction restricts water molecules to a single file as they are passing through the channel. In addition, the hydrophobic portion “coats” one side of the channel (the reason for this and the effect this has on ions will be discussed later). The restriction of water most likely occurs because of an electric field created by the charges on the protein structure, inducing the majority of the channel’s core to be hydrophobic. This electric field also dictates the direction of the water molecules as the flow through the channel. As water molecules enter the channel, they typically are oriented with the oxygen atom facing the entrance of the channel. As the molecules enter the NPA motif, the water molecules flip, so that the oxygen atom is facing toward the channel’s exit. It is thought that the orientation of oxygen changes as a result of a hydrogen-bonding event with the two asparagine molecules within the NPA motif. Therefore, because each water molecule must be reoriented to pass through the aquaporin channel and it can be reoriented only by interacting with the two asparagine molecules within the NPA motif, only one water molecule can flow through the channel at a time. Through these two restrictions it has been observed that the permeability of aquaporin channels toward water molecules is on the order of 6E − 14 cm3/s, which allows approximately 109 water molecules through each pore per second.

There is a second constriction of the aquaporin channel, usually toward the extracellular side of the cell membrane, which acts to restrict the movement of other molecules through the channel. This selectivity filter is termed the aromatic/arginine selectivity filter in aquaporin channels. The selectivity filter is a grouping of amino acids that interact only with water molecules and helps them through the narrowing created by this filter. Other molecules that do not interact with the selectivity filter cannot pass through this narrowing. The aromatic ring weakens the hydrogen bonds between water molecules, and then the partial negative charge on the oxygen atom interacts with the positive charge on the arginine. The interaction between the oxygen and the arginine allows water through the channels and prevents the passage of other molecules, especially protons.

All of the restrictions that were described are physically smaller than hydrated ions; remember that all ions that are in the body are hydrated (e.g., a sodium ion, which has a positive charge, will typically traverse through the biological substrates, with four water molecules associated with it; the partial negative charge on the water’s oxygen form weak hydrogen bonds to the sodium ion; a similar analysis can be made for all other ions). However, ions in solution can dehydrate and this is how ions typically pass through biological ions channels. To dehydrate an ion, a significant amount of energy would be needed, and typically this is counterbalanced by other binding events. In the case of the aquaporin channel, the location where the ions would need to become dehydrated to pass through the pore restriction is associated with hydrophobic amino acid regions. A hydrophobic surface cannot provide a temporary bonding event for a hydrophilic ion. Thus an ion would need an energy source to break the water hydrogen bonding events and not create new hydrogen bonding events. This significant amount of energy is not readily available, which effectively prevents ions from moving through aquaporin channels.

Four aquaporin channels associate with each other in the membrane, so that in one location there are four possible passageways for water to move through the cell membrane. Each aquaporin channel can have a slightly different protein structure. There are at least 4 different aquaporins in mammals and upward of 10 aquaporin channels found in plants. The different structures between the aquaporin molecules may allow for the movement of a small quantity of ions or other solutes (such as glycerol and small sugars) through the channels, although most aquaporin channels restrict ion and solute movement. In addition, some aquaporin channels respond to stimuli from external hormones or other paracrine molecules. On stimulation, the rate of water movement (and possibly the direction of movement) can be altered.

It is important to remember that aquaporins do not actively transport water across the cell membrane; instead they facilitate the diffusion of water across the cell membrane. Because of the slow diffusion of water across the lipid bilayer, aquaporins effectively increase the overall rate of water movement across the cell membrane.

2.6 Biomimetic Membrane

Recently, aquaporin-based biomimetic membranes caught attention because of the intrinsically high water permeability and salt rejection of aquaporin. Commercialization of aquaporin incorporated TFC membrane is now being undertaken.

Tang et al. [111] wrote a review of the properties of aquaporins, their preparation, and characterization. Very high permeability and salt rejection membranes can be obtained based on aquaporin protein function [112].

Kumar et al. [113], based on the measured water permeability of Aquaporin Z (AqpZ)-containing proteoliposomes, postulated that AqpZ-based biomimetic membranes can potentially achieve a membrane permeability as high as 167 μm/s bar (i.e., 601 L/m2 h bar), which is about two orders of magnitude more permeable compared to the existing commercially available seawater RO membranes [114].

Zhao et al. [115] successfully fabricated an aquaporin-based biomimetic membrane via interfacial polymerization method. It was noticed that the resulting membrane AMB-wild, with area greater than 200 cm2, had good mechanical stability when tested up to 10 bar under RO conditions. High water permeability (4.0 L/m2 h bar) and good NaCl rejection (around 97%) were observed at an applied pressure of 5 bar. Rejection was further improved at higher pressure. The membrane had superior separation performance compared to commercial RO membranes (BW30 and SW30HR), demonstrating the great potential of interfacially polymerized ABM membranes.

Several design approaches have been pursued in facing the challenge of making the biomimetic membranes as stable, robust, scalable, and cost-effective as their polymeric counterparts in the form of existing technologies such as RO membranes [114]. However, aquaporin membranes at present are on the laboratory scale, but production of these membranes could easily be established on an industrial scale (full-scale applications) in the near future.

7.3.3 Bolaamphiphiles-based biomimetic membranes for water purification

The aquaporin-based biomimetic membranes are potentially attractive for water purification and green energy production [45]. The aquaporins have high water permeability and selectivity because they are water channel proteins found in biological membranes. Positively charged bolaamphiphiles (see Fig. 7.4) may be used because they can easily be adsorbed on the negative surfaces of the membranes [45]. One privileged form of self-assembly of the investigated bolaamphiphiles is micelles. These micelles have surface of significant curvature as compared to larger nanosized vesicles with an aqueous core such as liposomal phospholipids. The authors have shown that aquaporin Z can be incorporated in structures such as micelles that have substantial spontaneous curvature. The potential application of supported biomimetic membranes can be the biosensors and water purification by filtration. The same authors investigated the preparation of a supported bolaamphiphile membrane on two polymeric nanofiltration membranes: NF-270 containing polyamide and carboxylic surface charges and NTR-7450 made of sulfonated polyethersulfone with sulfonic surface charges. In agreement with the previous studies on silica, mica, and gold, the results confirmed that the supported membrane coverage on the polymeric membranes was governed by the double-layer interactions. To obtain these results, a microfluidic device has been used. The comparison of results on two supports has shown that the formation of the biomimetic membrane was more favorable on the sulfonated polyethersulfone than on the polyamide surface in spite of the fact that both surfaces exhibit a similar surface charge density. The reason for the higher coverage of NTR-7450 may be the higher dissociation constant of sulfonic groups. As the final step of the procedure of formation of biomimetic membranes, spinach aquaporins that are transmembrane proteins that facilitate the water transport were incorporated into a supported membrane on NTR-7450. The result of this experiment was an enhanced pressure-driven water transport through the membrane. Although the transport was not selective, these results provide new insights into the formation of biomimetic membrane on water-permeable polymeric substrates, as a generic approach toward biomimetic water filters.

13.4.1 Misfolded proteins

In most cases, the aquaporin channels are more than capable of keeping up with the waste products produced by the brain, but this system can be overwhelmed. One of the most common sources of interference is accumulation of misfolded brain proteins, specifically intracellular hyperphosphorylated tau and beta amyloid (Ringstad et al., 2017; Da et al., 2018). Particularly, traumatic brain injuries (TBIs) have resulted in hyperphosphorylated tau protein. Normally, tau protein maintains the proper separation between the fast axonal transport tracks (Fig. 13.4), analogous to railroad ties separating the tracks. These transport tracks move metabolites, proteins, and even mitochondria down the axons and waste products back to the cell body for processing. However, when the tau proteins misfold, transport fails. The axon “dies back,” and the proteins begin to pile up within the cell body and precipitate. Both the soluble form and the precipitated form are fatally toxic to the cell. They highjack the synthesis mechanism, producing more defective proteins, and they get transported transsynaptically to previously unaffected cells. Thus, the neurodegenerative process spreads to interconnected regions of the brain. Beta amyloid is another misfolded protein that accumulates in Alzheimer disease. Beta amyloid tends to precipitate in the interstitial spaces, impeding flow needed to clear waste products, but by itself has not been shown to be fatally toxic. As a consequence of brain trauma—particularly repetitive trauma—significant accumulation of hyperphosphorylated misfolded tau occurs with similar consequences to neurodegenerative diseases such as Alzheimer’s disease (Da et al., 2018).

7.5.4 Understanding Mechanism of CAP Selectivity

Recently, a novel model based on aquaporin (AQP) which is the verified H2O2 channels on the cytoplasmic membrane has been proposed [146]. Cancer biologists confirm that most cancer tissues tend to express more AQPs on their cytoplasmic membranes than homologous normal tissues [146]. After the CAP treatment, CAP-originated H2O2 diffuses into cancer cells significantly faster than homologous normal cells, causing a significantly higher rise of ROS in cancer cells than normal cells [146]. Such different H2O2 consumption capacity between cancer cells and normal cells may be the foundation of the selective anticancer mechanism of CAP. The AQP-based model not only successfully explains the observed selective rise of ROS in cancer cells but also explains our recent observation that glioblastoma cells tend to consume H2O2 in medium significantly faster than astrocytes. It should be pointed out that the intracellular antioxidant system may also contribute to the selectivity of CAP, though the direct evidence is still lacking. The rise of intracellular ROS should correlate with the rise of extracellularly originated ROS and the resistance of the intracellular antioxidant system [146]. As mentioned above, the intracellular antioxidant systems include a series of enzymes such as catalase, superoxide dismutase, glutathione reductase, glutathione peroxidase, as well as small molecules such as GSH and NADPH [147]. A very recent study demonstrated that the decreased expression and the enhanced expression of Cu, Zn-SOD, or Mn-SOD enhanced and decreased the plasma-induced HeLa cell death, respectively [147]. Expression of exogenous catalase also blocked HeLa cell death [147]. Thus, the expression level of the intracellular antioxidant system also significantly affects the anticancer capacity of CAP. However, different from AQPs, the expression levels of specific intracellular antioxidant components in cancer cells are not always higher than that in homologous cells. Schematically overall model is shown in Fig. 7.39. According to this model the selective anticancer capacity of CAP may be due to the combined effect of multiple cellular factors, such as the enhanced expression of AQPs as well as the decreased expression of specific antioxidant enzymes such as catalase in cancer cells (Fig. 7.40).

In order to validate the proposed AQP-based hypothesis effect of AQP expression, plasma interaction with cancer cells was studied. Recall that the expression of AQPs in tumor tissues has been widely investigated in past decades [148]. It was found that in the case of U87 cell lines the expression of AQP9 is the most abundant [148]. AQP9 is reported to play important roles in the malignant progression of brain astrocytic tumors [148], such as counteracting the glioma-associated lactic acidosis by clearance of glycerol and lactate from the extracellular space. AQP8 and AQP9 were chosen to study AQPs’ role in anti-glioblastoma effect of plasma by using siRNA technology. It was shown that knockdown of AQP8 in U87MG cells can significantly reduce the anti-glioblastoma capacity of plasma [148]. Overall, it was concluded that silencing specific AQP in U87MG cells indeed weakens the anticancer effect of plasma in agreement with above stated model.

5.4 Application and Future Perspective

Dedicated research has recently been devoted to aquaporin-based membranes, including a certain number of commercialization attempts. The creation of biomimetic nanopores using solid-state materials, such as silicon, is an advancing and newly developed research area. Scale-up to practical dimensions for separation present several challenges. The generation of nanopores that use recent advancements in their fabrication, including technologies such as i-beam and e-beam lithography, are still limited to the lab scale and surely need more development to be put into practical use. The implementation of these pores with the help of specialized biological molecules and chemistry could be difficult in large amounts on a practical scale. Doubts regarding the use of ligands in order to functionalize pores are also difficult to address, especially if ion discrimination, like that seen in potassium channels, is expected. However, the application of DNA sequencing has approached commercialization levels with several technologies in a state to initiate a dedicated organization [92].

LMs and ionophore-based membranes are carrier-mediated biometric membranes. LMs have undergone significant research in the recent decades and there have been good developments in the transport process. However, there are still limitations, such as poor stability, and other practical difficulties, that must be overcome before their commercialization in separation can be succesfully achieved. The practical difficulties include the unstable immobilization of LMs in SLMs and inefficiencies in the separation process of the various recovered materials from the emulsion phase of ELMs. Recently, a few pilot scale efforts have been made and further developments are expected in the coming years. ELMs are utilized for removal of zinc, phenol, and cyanide from industrial waste streams [93].

Membranes based on ionophore are used widely in ion-selective electrodes. Ion-selective membranes are the gold standard for this application. However, their application in separation membranes is yet to experience significant progress because of the low transport rates of ions in practical polymeric matrices [94]. The use of plasticizers for providing fluidity to the polymer matrix could not improve the transport to desired levels in separation applications.

Biomimetic membranes based on artificial channels are still on the ground level of research with most of the work being focused on the synthesis and characterization processes. The transport measurements are still at an introductory level in this field [62] and much more dedicated research is required in order to compare their efficiency to that of the membrane protein channels.

Artificial water channels could be a possible solution as they might prove to be novel materials for water purification. The difficulties in the application of CNTs for desalination include insufficient rejection of salts and difficulties in the manufacturing of large-sale aligned CNT membranes [95]. Studies on water channel-based organic nanochannels have experienced more advancement than before. Studies have shown that that there are no suitable principles expalining the design process [73]. Mimicking the natural selective filters is the only semiempirical principle. Studies indicate that commercialized channel membranes have much lower permeability (> 3 orders of magnitude lower than aquaporins) and imperfect solute rejection in the case of channels with larger diameters [75]. From the studies made by Fei et al. [41], extensive hydrogen bonding helps to encapsulate water wires along the channel, but on the other side it decreases the mobility of the water molecules. This may be the probable reason for the channels showing much lower water permeability, sometimes on the order of more than four magnitudes lower than the lipid background permeability, which led to more changes, including the method used to measure the permeability of less permeable channels. A suitable method for the measurement of the permeability of water is needed and required [62]. The next generation of water channels needs to have an improved design for the pore structure, so as to increase the permeability of water and maintain rejection of the salts. Improvement in the geometry of the channels is one possible area. This would help in packing these channels with much higher lipid density/polymer matrix for fabricating membranes. The studies have proven their capabilities in separation applications because of their higher stability, similar characteristics to natural channels, production scalibility, and most importantly, their ability of immobilization in membrane-like supports.

The strategies that have been proposed using bioinspired antifouling techniques for recent membranes are attracting people in this field. A number of proposed approaches, such as advancement in surface modification, appear to be feasible technically. A cost-effective approach and the practical implications thereof could prove crucial in advancing the strategies to a state in which they can be used in practical applications, as a few of the mentioned approaches initially reduce the permeability of membranes.

6.2.12 Biologically Inspired Membranes

Many biological membranes are highly selective and permeable. Aquaporins are protein channels that regulate water flux across cell membranes. Their high selectivity and water permeability makes their use in polymeric membranes an attractive approach to improve membrane performance. Aquaporin-Z from Escherichia coli has been incorporated into amphiphilic triblock polymer vesicles, which exhibit water permeability at least an order of magnitude over the original vesicles with full rejection to glucose, glycerol, salt, and urea [95]. One potential design is to coat aquaporin-incorporated lipid bilayers on commercial nanofiltration membranes. On this front, limited success was achieved [96]. Aligned CNTs have been shown both experimentally and theoretically to provide water permeation much faster than what the Hagene–Poiseuille equation predicts, owing to the atomic smoothness of the nanosized channel, and the one dimensional single file ordering of water molecules while passing through the nanotubes [97,98]. It was predicted that a membrane containing only 0.03% surface area of aligned CNTs will have flux exceeding current commercial seawater RO membranes [99]. However, high rejection for salt and small molecules is challenging for aligned CNT membranes due to the lack of CNTs with uniformly subnanometer diameter. Functional group gating at the nanotube opening has been proposed to enhance the selectivity of aligned CNT membranes [100]. By grafting carboxyl functional groups on sub-2-nm CNT openings, 98% rejection of Fe(CN6)3− was achieved at low ionic strength by Donnan exclusion [35]. However, KCl rejection was only 50% at 0.3 mM and decreased to almost zero at 10 mM. Grafting bulky functional groups at the tube opening could physically exclude salts. However, steric exclusion will significantly reduce membrane permeability [101]. Thus at the current stage, aligned CNT membranes are not capable of desalination. To achieve reliable salt rejection, the CNT diameter must be uniformly smaller than 0.8 nm [102]. A key barrier for both aquaporin and aligned CNT membranes is the scale-up of the nanomaterial production and membrane fabrication. Large-scale production and purification of aquaporins are very challenging. To date, chemical vapor deposition (CVD) is the most common way to make aligned nanotubes. A continuous high-yield CVD prototype has been designed for producing vertically aligned CNT, paving the way for large-scale production [64]. A postmanufacturing alignment method using magnetic field was also developed [103].

Nanocomposite and TFN membranes have good scalability as they can be fabricated using current industrial manufacturing processes. The high water permeability can reduce the applied pressure or required membrane area and consequently cut cost. This strategy may greatly improve the energy efficiency for treatment of waters with low osmosis pressure, but it may have limited advantage in seawater RO, whose energy consumption is already close to the thermodynamic limit [8]. A recent review ranked current membrane nanotechnologies based on their potential performance enhancement and state of commercial readiness [99].

2.2.4 Concept of directed water channels

The concept of directed water channels is associated with discovery of the structure of aquaporin proteins coexisted with the phospholipid bilayer, which honored Nobel Prize winners in 2003. The size of the hourglass-shaped channel at the narrowest point is 0.30 nm [96–98], therefore, water molecules (0.28 nm) can pass through these protein channels selectively, while other molecules or ions cannot. As an application, aquaporin proteins were incorporated into the barrier layer of a membrane that exhibited 80–1000 times higher water flux than conventional membranes [99]. Water channels could also be found in aligned carbon nanotube-based membranes, where 2 nm inner diameter of the nanotube can be used to get rid of contaminated issues from water with twice higher flux than those of commercially available membranes [100]. Other materials containing nanoporous structure, such as liquid crystals with specific phase [101], graphene oxide nanosheets [102], zeolites [103], and synthesized tubular molecules [104], also provide different options to integrate directed water channels with the barrier layer of a membrane.

Unfortunately, all the abovementioned methods suffer great challenges for practical applications, considering processability, robustness, and fabrication cost. For instance, the density of directed water channels from aquaporin protein embedded in a barrier layer of a membrane was very limited, which greatly reduces the water permeability and the durability of the membrane. Also, carbon nanotubes have to be grown in situ on the substrate, followed by aligning and bonding together to prepare a membrane, which must increase the fabrication cost drastically and the risk of generating defects. Moreover, zeolites-based membrane with tortuous pore structure raises high hydraulic resistance; as a result, the water transportation rate will be increased limitedly. Especially, when graphene oxide was employed to prepare a barrier layer of a membrane for a water treatment process, additional immobilization was needed and therefore, the fabrication cost will be inevitably increased. Moreover, the formation conditions of a specific liquid crystal phase are critical and difficult to scale-up. Finally, the tubular molecules with water channels synthesized from multiple steps are cost-effective for the practical production. Therefore, it is expected to create tunable water channels facilely and extensively in the barrier layer of RO/FO membranes that address all of the abovementioned problems [105,106].

Our research team has expanded the concept of directed water channels and demonstrated a breakthrough on high-flux nanofibrous membranes. TFNC membranes with a nanocomposite barrier layer were fabricated, where the barrier layer was integrated with nanofibrous scaffolds and cross-linked polymer matrix. As mentioned earlier, electrospun nanofibers can be embedded into the polymer matrix and form an integrated top barrier layer of the membrane, where the fiber is surrounded by polymer chains [69,77,107]. Directed water channels are naturally formed by phase separation between an nanofibrous scaffold and a polymer matrix, where the gap between two phases can be used as the channels to discriminate water and other contaminant molecules, based on the mechanism of size exclusion [78,106]. The structure of TFNC membrane based on a three-layered structure with the top nanocomposite layer consisting of directed water channels is shown in Fig. 8.

The circular diagram in Fig. 8 illustrates schematically the formation of directed water channels in the barrier layer of the TFNC membrane with the typical three-tier structure. In the diagram, (magnification part) cellulose nanofibers (yellow) work as one continuous phase to form three-dimensional skeleton, and polymer matrix (pink) surrounding the skeleton work as the other continuous phase. There are connected hollow cylindrical gaps between the two phases, which are regarded as directed water channels (blue). It should be noted that water can also transport through molecular cavities in the polymer matrix, as shown in Fig. 8, however, directed water channels are the preferred options due to the low hydraulic resistance. The advantages of directed water channels over molecular cavities of polymer matrix are obvious: (1) directed water channels, theoretically, can be formed between any bicontinuous phases by phase separation in the barrier layer, not limited by nanofibers and polymer matrix; (2) the size of directed water channels could be adjusted flexibly by tuning the phase separation between two phases, therefore, different types of membranes could be achieved for a variety of applications; (3) the surface properties of directed water channels could also be modified by simply decorate the surface of the gap with, for example, hydrophobic or hydrophilic, charged, or neutral species; and (4) the density of directed water channels could be controlled by the incorporation of nanofibrous scaffolds into the polymer matrix, which not only can provide directed paths for water transportation, but also improve the mechanical properties of the membrane and remain high porosity, therefore, the permeation flux should be increased drastically.

The formation of directed water channels in the barrier layer of a membrane are not limited by electrospun nanofibers, any other nanoscale materials, such as carbon nanotubes [105,107] or cellulose nanofibers [106,108–110], can also be used to create directed water channels. This new membrane design is particularly suited for NF, RO, and FO applications. All membranes containing directed water channels could exhibit 2–5 times higher permeation flux than the corresponding membrane without water channels [39].