Encapsulation versus Self-Aggregation toward Highly Selective

Oct 17, 2018 - High-selectivity issues are determinant for the performance of natural protein channels, but they have been not considered as determina...
0 downloads 0 Views 3MB Size
Download PDF
Article Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/accounts

Encapsulation versus Self-Aggregation toward Highly Selective Artificial K+ Channels Published as part of the Accounts of Chemical Research special issue “Supramolecular Chemistry in Confined Space and Organized Assemblies”.

Downloaded via REGIS UNIV on October 17, 2018 at 12:31:12 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Mihail Barboiu* Lehn Institute of Functional Materials, School of Chemistry, Sun Yat-sen University, Guangzhou 510275, China Institut Europeen des Membranes, Adaptive Supramolecular Nanosystems Group, University of Montpellier, ENSCM-CNRS, Place E. Bataillon CC047, Montpellier 34095, France CONSPECTUS: Natural ion-channel proteins allow ion transport across cell membranes at rates very close to those for ionic diffusion in water. Among them, natural KcsA K+ channels present high transport rates and total selectivity for K+ cations, rejecting all other cations. Most of the reported artificial ion channels cannot reach this type of activity because of their low selectivity. Several synthetic channels have been designed to mimic the natural KcSA channels, but those presenting an important K+/Na+ selectivity are limited. High-selectivity issues are determinant for the performance of natural protein channels, but they have been not considered as determinant in controlling the transport activity of the artificial ion channels. This Account discusses the last developments of artificial supramolecular carriers or channels that selectively transport K+ cations against other cations. Mimicking the complex structures of protein channels is an important research area. These studies are related to such adaptive biomimetic systems that can self-select their functions, with a specific emphasis on artificial superstructures enabling K+ transport like in the natural ones. Alternatively, it is more than interesting to synthetically construct only the active key structures of protein filters or gates that give the chemical selectivity or lead us to describe their dynamic role in the ion pumping and translocation along the channel. Several self-assembled macrocyclic channels are presented here. The macrocyclic binding sites may selectively encapsulate the K+ cations or form aggregated H-bonded central pores of self-assembled macrocycles that coordinate the K+ cations as hydrating water molecules in aqueous solution, compensating for the energetic cost of cation dehydration. These macrocyclic channels are responsive in the presence of K+ cations, even when a large excess of Na+ is present. From the mechanistic point of view, these systems express a synergistic dynamic feature: addition of K+ cations drives the selection and emergence of specific ion channels that selectively conduct the K+ cations that promoted the formation of channel superstructures in the first place. These highly permeable and K+-selective artificial channels may be considered as simple primitive biomimetic alternatives of natural KcsA channels that may find interesting applications in chemical separations, selective sensing, and biomedical materials.

■

INTRODUCTION The exchange of ionic metabolites across cell membranes is a prerequisite for most physiological processes, such as muscular motion, nerve influx, and neuronal ion exchanges.1,2 Their dysfunctions lead to a number of important diseases and even to death.3 Transport across the bilayer membrane is regulated by channel proteins, which usually respond to various stimuli such as external osmotic, ionic, and pH gradients.4 The high rates observed for selective exchanges between intra- and extracellular media via natural channel proteins most of the time lead to stable membrane potentials. Despite their importance in describing the transport activity observed for the natural channels, high-selectivity issues have not yet been considered as important for the activity and selectivity of artificial channel systems. The fact that the artificial channels cannot stabilize this type of stable membrane potential that is useful to control other ionic gradients, is mostly due to their low selectivity. © XXXX American Chemical Society

Indeed, biomimetic ion channels with high selectivity and activity that are made of simpler molecular scaffolds than complex proteins, should open new avenues toward the understanding of important transport phenomena related to ionic translocation through bilayer membranes. Inspired by the tremendous performances observed with biological channels and given the increasing demand for competitive separation processes like blood dialysis, desalination and production of ultrapure water for medicine, etc., the biomimetic design of alternative synthetic ion channels with high selectivity is of expanding interest. Among the protein channels, the KcsA K+ channels are known to conduct K+ cations at high rates selectively over Na+ cations, which are totally rejected.5 Several artificial systems have been synthesized in order to mimic the KcSA K+ Received: June 28, 2018

A

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 1. (A). K+ cations in the KcsA K+ channel filter or outside the channel, surrounded by carbonyl groups or water, respectively.31 (B, C) K+ recognition by (B) 18-crown-6 and (C) 222 cryptand. (D, E) Examples in which 18-crown-6 equatorially coordinates the K+ cations while (D) anionic counterions and water molecules or (E) aromatic indoles bind the vacant apical coordination sites.11 Reproduced with permission from refs 11 and 31. Copyright 2005 Royal Society of Chemistry and 2001 Nature, respectively.

channels, but those showing high K+ versus Na+ selectivity are limited. In general, these artificial channels display significant cation transport rates; however, most of the time the selectivity is not as high as usually is determined by the ratio of the energetic penalties for cation dehydration. At this point, two important questions arise: Would it be possible to construct highly selective artificial K+ channels using only molecular components and simple synthetic methods to compete with and replace the natural KcSA K+ ones? May the KcsA K+ channel structure be used to tailor the performances and functions of such synthetic systems? Indeed, if possible, one could consider that the discovery of highly selective synthetic systems that can be form such competitive artificial channels is most of the time empirical. During the last years, ion selectivity has become important for determining the transport activity, and several artificial channels with high K+ selectivity have been synthesized.6−29 Among other popular molecules, crown ethers have been intensively used as components for the construction of selective ion channels. Lariat crown ethers or macrocyclic “hydraphiles”,6 peptide-appended crown ethers,7 barrel-stave motifs,8 G-quadruplexes,9 and bola-amphiphiles10 have been studied as ion channels translocating different artificial membrane systems. Understanding the behaviors of natural protein channels is often one way to bring artificial channels close to natural functions. Transmembrane ion transport can be achieved by using molecular carriers that reversibly encapsulate species of interest or by self-assembly of supramolecular channels via selfaggregation. On the frame of both basic concepts highlighted in this Account, we were interested to directionally self-organize crown ethers via H-bonding in order to improve the performances of the artificial ion channels to make them closer to the natural ones11−19 or to scale up toward biomimetic hybrid materials with ion-channel transport functions.20−26 Cation-file diffusion occurs via successive siteto-site binding/jumping along the aggregated macrocyclic sites, forming a hydrophilic pathway for ion diffusion within the hydrophobic membrane environment. This approach has been used by other groups to obtain light-responsive27 or peptideamphiphilic macrocycles.28 This Account is devoted to macrocyclic K+ carriers and selfassembled K+ channels and is divided into three sections. The first part describes the KcsA K+ channels, emphasizing the most important structural features leading to K+-selective transport. Related artificial biomimetic K+ channel systems, mostly developed by learning from the natural KcsA K+

channels, are then described as simple artificial alternatives to the most complicated biological machinery. The second part describes recent works, mostly developed in our laboratory, on macrocyclic carriers for selective encapsulation of K+ or on selforganized K+ channels, emphasizing the more recent developments of ion recognition and ion-directed self-assembly. Their inclusion as active artificial channels in bilayer membranes will be described with all matters in the last part of the Account. Several inspiring results discovered during the last 10 years in this field are of highest relevance for important biological scenarios. They open the field to novel K+-selective channels, paralleling those in biological systems.

■

KCSA K+ CHANNELS AS A SOURCE OF INSPIRATION The KcsA K+ channel is a well-known example of a protein channel in which K+ cations are considered to permeate along an ion channel formed from multiple protein subunits within transmembrane domains, which are packed to form the selectivity pore, a crucial element for its activity.5 The selectivity of the KcsA K+ channel is driven in the narrowest region of the pore, which is not able to fit an ion with water molecules around, and thus, dehydration is required to enter this region of the channel. The gating filter of the KcsA K+ channel affords closely located carbonyl groups from the protein scaffold for the perfect binding of dehydrated K+ cations.30 The selectivity is mainly determined by the perfect coordination of K+ cations, together with the orientation of the dipolar carbonyl oxygens that surround permeating cations in the pore filter, rather than the size of the pore (Figure 1A). Conformational motion of the channel adaptively determines that there are eight carbonyl oxygens that compensate for the dehydration energy, and there are four such coordination sites along the filter.31 As a general rule, we can observe that conformational changes induce a variety of recognition behaviors that are directly connected to the channel structure and its average occupancy, coordination behaviors, channel affinity for the K+ cations, and the mobility of the translocating cations. The conformational changes of the pore are crucial for switching of the selectivity filter between the “collapsed” and “conductive” conformations, which are related to the concentration of the K+ at the cellular level.5 Interestingly, the KcsA K+ channel depleted of K+ cations can adapt and change its conformation to selectively transport water 20 times faster than one-directional diffusion of bulk water.32 Inspired by the high performaces shown by the KcsA K+ channel, synthetic methods for the design of alternative B

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research biomimetic artificial K+ receptors or K+ channel systems are of tremendous interest. Molecular recognition involving synthetic receptors is governed by the optimal positioning of the coordinating groups of the receptor, replacing the hydration water molecules around the K+ cations, as in the active gate of KcsA K+ channels. The selective recognition of the optimal against the imperfect coordination features is crucial for both recognition and transport functions. Most of the pioneering examples include crown ethers (Figure 1B) and cryptands, the 3D analogues of crown ethers (Figure 1C). Crown ether receptors equatorially bind the cations, but they do not completing all of the hydration sites, so dehydration is not compensated. This allows anionic counterions, water molecules, or aromatic groups to occupy the vacant apical positions (Figure 1D,E).11 Moreover, cation−π interactions are known, and many examples are of particular biological significance.6 On the other hand, the transport performance does not follow a direct dependence on the stability of the cation− receptor complex: strong binding has a negative effect on the translocation rate.12 Therefore, there exists an optimal association stability for obtaining the best transport performance as long as the complex stability constant is optimal and does not exceed certain values. An elucidation of the behaviors and restrictions that determine the binding of K+ by simple specific receptors is important for understanding the transport mechanisms for the construction of optimal carriers and channels.

Figure 2. Heteroditopic ion-pair recognition and H-bonding selfassembly of macrocyclic supramolecular ion channels. Adapted from ref 24. Copyright 2004 American Chemical Society.

the construction of the ion-channel systems described here (Figure 3). Compounds 1−12 contain rigid benzo-15-crown-5 (1−9) or benzo-18-crown-6 (9−12) macrocyclic ion-binding sites and are decorated with linear or branched alkyl tails of different

■

SUPRAMOLECULAR SELF-ASSEMBLED ION-CHANNEL STRUCTURES IN SOLUTION AND THE SOLID STATE During the past decade, our group was strongly involved in the possibility of constructing self-organized channels based on heteroditopic alkylureido crown ethers. They are among the best K+-selective artificial channels, showing clear regular channel activities.13−26 The transmembrane supramolecular channels arise from H-bonded self-assemblies of crown ethers that are aligned around a central pore. From a structural point of view, we combine monomers as carrier-like transporters13 with supramolecular aggregates with potential channel functions14 (Figure 2). Their self-assembly can be directed by structural design, which determines the distribution of several oligomeric channels of variable stability within the bilayer membrane. Several crown ether molecules have been successfully tested as molecular carriers or as supramolecular channels spanning the lipid bilayers.17−19 Their dynamics at the supramolecular level can offer the basis for understanding novel emerging transport properties that can appear in response to external gradients, similar to natural channels operating under stable membrane potentials. Guided by these principles, we anticipated that modification of the main structural functional moieties of the ionophores may induce controlled recognition and improved transport behaviors. Of special interest are structure-controlled functions for the construction of the channels and to build their superstructure from the directional self-assembly of the optimal units: (i) the cationic binding groups are rigid benzo-15crown-5 and benzo-18-crown-6 of flexible 15-crown-5 macrocycles; (ii) the guiding interactions are the urea or bis(acylhydrazone) H-bonds; (iii) the nature of the hydrophobic tail determines the dynamics of the channel superstructures at the interface with the bilayer membranes. Twenty-four macrocyclic ionophores 1−24 were prepared for

Figure 3. Crown ether compounds 1−24 studied as K+-channelforming components. C

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 4. Single-crystal X-ray structures in stick representation of heteroditopic macrocyclic ionophores. (A, B) Side and top views of the packing of compounds (A) 113 and (B) 3.15 (C−F) Structures of ion-pair complexes: (C) 1·NaCl;13 (D) 1·NaNO3;13 (E) 11·KNO3;26 (F) 3·KNO3.26 (G) Structure of the 23·KTf complex.19 (H) Structure of the (15·Na+)2 complex in the CPK representation (left) and details of the coordination of Na+ by 15-crown-5 (right).33 Legend: Na+ or K+ cations, violet spheres; Cl− anions, green spheres; NO3− anions and H2O molecules, CPK representation. Adapted from refs 13, 15, and 19 and with permission from refs 26 and 33. Copyright 2003, 2006, and 2017 American Chemical Society, 2009 National Academy of Sciences, and 2018 John Wiley and Sons, respectively.

lengths.15,18 Linear alkyl groups proved to be important to promote the formation of self-assembled channels, while crowded bulky tails may disturb the binding and self-assembly, favoring the emergence of small aggregates to a certain extent. Flexible 15-crown-5 macrocycles 13−15 with variable alkyl chain lengths were synthesized in a similar manner. They led to adaptive binding of Na+ and K+ cations as a function of their concentration in the membrane.33 Acyl hydrazide-substituted benzo-15-crown-5 ethers 16−18 show low transport rates, in contrast to their stronger self-association via multiple Hbonding.33 We then designed and prepared a series of squalyl (19)17 and cholesteryl (20−21)16 15-crown-5 ethers, that form H-bonded channels. They are stabilized via aggregation of the bulky anchoring arms, resulting in the formation of the preorganized clusters of the macrocycles in the lipid bilayers. Multivalent macrocyclic systems have been designed for the generation of directional ion-transporting channels based on triarylamine pillars (22, 23)19 or a pillar[5]arene central platform (24).34 Important information concerning the self-assembly and ionrecognition behaviors of the ionophores was obtained from single-crystal X-ray analysis. The single-crystal X-ray structure of 1 reveals the heteroditopic receptor with the urea group orthogonally disposed with respect to the plane of the benzocrown ether ring (Figure 4A). This structure may be directed along parallel off-center macrocyclic aggregates. Replacing the phenyl moiety with hexyl tails in 3 causes the self-assembly change to an antiparallel arrangement of the macrocycles, which is favored by tight contacts between the hexyl chains and

the vicinal macrocycles (Figure 4B). Accordingly, columnar arrays of macrocycles are generated, such that the crown ethers are oriented in parallel or antiparallel arrangements with an average spacing between macrocycles of 4.83 Å in 1 and 8.23 Å in 3. The binding of NaCl by 1 reveals the heteroditopic recognition with the formation of parallel dimers of 12·NaCl, with both Na+ and Cl− sandwiched between two macrocycles and two urea moieties, respectively (Figure 4C). The structure of 1·NaNO3 shows the formation of dimers of (1·NaNO3)2 (Figure 4D). The Na+ cation is equatorially coordinated by the fittest benzo-15-crown-5 ring and apically coordinated by the NO3− anion, which is simultaneously H-bonded to the N−H groups of a vicinal molecule of 1, thus acting as a connecting bridge between two 1·Na+ monomers. Similar antiparallel dimers (11·KNO3)2 have been obtained for 11·KNO3, where the K+ cation is equatorially coordinated to the fittest benzo-18-crown-6 ring, while the apical position is occupied by the NO3− anion, which is simultaneously Hbonded to the urea moiety (Figure 4E). Differently, the crystal structure of 3·KNO3 shows the formation of supramolecular polymers (3·KNO3)n. The biggest K+ cation is sandwiched between two benzo-15-crown-5 macrocycles, while the NO3− anion is H-bonded between two urea moieties, resulting in the formation of knotted bridges between 32·NO3− and 32·K+ dimers, respectively (Figure 4F). All of these structures reveal the formation of the ion-channel superstructures in the solid state, alternating alignments of the macrocyclic/urea receptor D

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 5. (A) Lateral view (left) and top view (center) of the double-barreled model and top view of the toroidal model (right) for the organization of 3 in the bilayer.15 (B) Schematics of the cation/proton transport mechanism via dynamic constitutional channels of 3 (see the text for details).18 (C) Ratiometric fluorescence intensity of the pH indicator HPTS as a function of time for transport of alkali ions through membranes containing compound 19.17 (D−F) Pseudo-first-order rate constants (k) for transport of alkali ions through large unilamellar vesicles containing compounds (D) 19,17 (E) 22,19 and (F) 24.34 Adapted with permission from refs 15, 18, 19, and 34 and with permission from ref 17. Copyright 2006, 2016, 2017, and 2017 American Chemical Society and 2016 John Wiley and Sons, respectively.

This structure encapsulates K+−H2O wires and has a fantastic similarity to the K+−H2O wires observed within the active gate filter of the natural KcsA K+ channel and described later in 1998 by MacKinnon.5 The alternating positioning of K+ cations and H2O molecules overcomes the electrostatic repulsion between the cations and allows K+−H2O cotranslocation along the channel.5 Similarly, Na+−H2O translocation has been proven to occur through macrocyclic channels within single crystals of dibenzo-18-crown-6.36 We believe that the structure of 23 reported by our group completes these

layers obtained via ion-pair encapsulation or ion-pair-directed self-assembly. The single-crystal structure of crown ether−triarylamine complex 23·K+ shows that the K+ cations are equatorially surrounded by benzo-18-crown-6 macrocycles, while the apical positions are coordinated by water molecules acting as Hbonded bridges to vicinal crown ethers (Figure 4G).19 This results in the development of the K+·H2O channel superstructure. A very close superstructure showing a similar stack of crown ethers was published in 1982 by Lehn and co-workers.35 E

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Table 1. Pseudo-First-Order Rate Constants (kK+), EC50(K+) Values,a and K+/Na+ Selectivities (SK+/Na+) for the Transport of K+ Cations through Large Unilamellar Vesicles Containing the Indicated Compounds compound

103·kK+/s−1

EC50(K+)/μM

SK+/Na+

ref

3 19 20 octylphenylalanineamidobenzo-15-crown-5-ether 15 24

175 87.5 3.9 − 100 3.2

10 15.8 26.5 6.2 7.2 38.4

3−17 58.3 14 10 1.5−2 5

18 17 16 29 33 34

a

Here, EC50(K+) is the concentration of the compound needed to obtain 50% K+ transport activity.

external pH gradient.18 Moreover, we discovered that such passive K+-induced influx induces a stable transmembrane potential that activates the transport of the cations as observed in natural channels and rarely detected for artificial ion channels.37,38 Figure 5B shows the K+ cation transport mechanism under passive and active conditions: (i) In the presence of K+, the addition of 3 in the bilayer induces the rapid formation of active ion channels, which cannot be observed in the presence of Na+. The K+ transport via such emergent K+-induced superstructures with fast influx rates results in the generation of a potential resulting from the selective concentration of positive charges on the inside face and negative charges on the outside face of the membrane. It induces a fast proton efflux that restores the electrical balance between the membrane interfaces. (ii) As a consequence, a quite unexpected pumping of K+ inside the vesicle occurs. It reach a maximum value and then slowly re-equilibrates to the initial pH value via a slow K+ antiport efflux. (iii) The addition of NaOH increases the external pH to 7.4, and a very fast efflux of H+ is observed, together with a very fast antiport influx of K+. (iv) The final passive transport step is quite similar to (i). The addition of NaOH activates a strong proton efflux together with a K+ influx. They reach a maximal value, followed by slow proton influx to re-equilibrate the pH. We know from structural studies that macrocycle 3 binds the Na+ cation via equatorial coordination by the oxygens of the crown ether. However, the dehydration sphere around the cation is not totally coordinated by the receptor binding sites, as the apical positions remain hydrated.26 In a different manner, the bigger K+ cations are 10-fold-coordinated in a sandwich-type geometry by two 15-crown-5-ethers and thus are completely surrounded by the macrocycle oxygens, replacing the waters coordinating the cation in aqueous solution. The precise transport mechanism is difficult to describe accurately by these X-ray structures and binding behaviors to explain the selective recognition of the cations. It is most probably possible that the translocating cations can diffuse along the channels formed by aligned macrocycles through a series of K+-sandwiching sites. More complex porous structures resulting from crown ether aggregation may also form, positioning the macrocycles in close spatial proximity. The cation-file translocation occurs as successive binding/ jumping along the successive macrocyclic sites, acting as favorable pseudohydrophilic channels in the hydrophobic membrane environment. Voyer and co-workers showed that the optimal and maximal distances are 6 and 11 Å, respectively, as an ion jumps between two vicinal crown ether moieties.37 From our X-ray crystallographic analysis, the distance between two neighboring benzo-15 crown-5 ethers is 4.8 Å, similar to the measured distance of 3.3 Å between two macrocycles in the

discoveries, displaying an amazing structural similarity with both Lehn’s and natural KcsA superstructures and moreover presenting ion-channel activity in bilayer membranes.19 2-(Dodecylureidomethyl)-15-crown-5 15 forms supramolecular dimers (15·Na+)2 that are connected via two apical coordinative bonds between urea carbonyls and Na+ cations (Figure 4H). It forms a Na+-selective pseudocryptand that encapsulates two Na+ cations inside a self-assembled sandwichtype superstructure of two macrocycles. We also observed that the flexibility of the 15-crown-5 in compound 15 favors apical binding of (15-crown-5)·Na+ by urea carbonyls, allowing the sixfold coordination of the Na+ cation. Differently, the sterically constrained benzo-15-crown-5 in 3 is rigid. This has a significant binding effect on K+ cations 10-foldcoordinated by two benzo-15-crown-5 macrocycles, resulting in the formation of a high-coordination configuration, as is found to induce the selective binding of K+ cations in the filter of the KcsA K+ channel.

■

MEMBRANE TRANSPORT ACTIVITY Convergent synthetic steps have been used for the construction of supramolecular channels resulting from the self-assembly of H-bonded aggregates dynamically exchanging in the bilayer membrane and designed to mimic the natural KcsA channel. As a general trend, compounds 1−24 form channels via the self-assembly of supramolecular oligomers, allowing disruption over regular translocation behaviors at low concentrations. At higher concentration, these channels show multiple conductance activity values and lifetimes, reminiscent of the formation of several types of active superstructures showing a basic regularity. The unifying assumption is that stacks of these oligomers would lead to the variable adaptive structures, depending on their concentration. They successfully form macrocyclic aligned pathways that enable efficient transport through bilayer and hybrid solid membranes.20−26 Compounds 3 and 11 show a complex set of ion-conductance activities for Na+ and K+, respectively. On the basis of planar lipid bilayer experiments, we equated the presence two types of channels formed by compounds 3 and 11 in the membrane: possible stacked crown ethers to form double-barreled channels and large toroidal pores (Figure 5A). The transport activity has a related relationship with the lipophilicity of ionophores, with low activity for the shorter tails and optimal activity for octyl-substituted macrocycles, while longer akyl chains lead to constrained, less dynamic macrocyclic relays. In order to get supporting information on ion conduction mechanisms, we assessed the ion selectivity and serendipitously found that hexylureidobenzo-15-crown-5-ether 3 slowly transports smaller Li+ or the fittest Na+ cations, while K+, Rb+, and Cs+ cations, which are dimensionally bigger than the 15crown-5 hole, are transported faster even in the absence of an F

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research K+(15-crown-5)2 sandwich-type complex. The close proximity of the macrocyclic binding sites favors multiple contacts between the oxygens of the crown ether moieties and the transported cations, recovering the hydration sphere of the cation. This mostly determines the selective binding of highcoordinated K+ cations against the low-coordinated Na+ cations. The K+ cation transport through the channels formed by hexylureidobenzo-15-crown-5-ether 3 occurs with a transport rate of kK+ = 175 × 10−3 s−1 (Table 1) and is certainly determined by the mobility of the self-assembled aggregates of 3, which would favor the emergence of dynamic channel superstructures within the bilayer membrane. This is confirmed by the fact that under the same conditions, cholesterylthioureidoethylamido-15-crown-5 2016 and squalylamidobenzo-15-crown-5 1917 (Figure 5C) with lower fluidities within the membrane environment present 1 order of magnitude lower initial K+ transport rates of kK+ = 3.9 × 10−3 s−1 and kK+ = 87.5 × 10−3 s−1, respectively. Hexylureidobenzo-15-crown-5-ether 3 shows clear regular channel activity (EC50(K+) = 10 μM) and presents good selectivity (SK+/Na+= 3−17). Octylphenylalanineamidobenzo15-crown-5-ether, reported by Zeng et al.,29 presents also interesting selectivity and activity values (SK+/Na+ = 10 and EC50(K+) = 6.2 μM). Under the same conditions, squalylamidobenzo-15-crown-5-ether 19 has the highest selectivity ever reported for K+ over Na+ transport (SK+/Na+ = 58.3) with a reasonable activity (EC50(K+) = 15.8 μM).17 We postulated that the entropic cost for cation binding is far larger than the case when the self-assembled channels have a restricted conformational entropic behavior in the bilayer. Unexpectedly, the simple removal of the phenyl ring in the macrocyclic structure of vflexible 2-(hexylureidomethyl)-15crown-5-ether 13 produced a drastic reduction of the transport activity for all of the alkali cations over a large 5−70 μM concentration domain. This is mostly related to its lower hydrophobicity, resulting in a partition of the channels within the membrane. Increasing the hydrophobicity in 2-(octylureidomethyl)-15-crown-5-ether 14 induced a moderate activity and selectivity for the transport of Na+. The most hydrophobic compound in this series, 2-(dodecylureidomethyl)-15-crown-5-ether 15, is the most active toward all of the alkali cations, and while its selectivity is reversed to Na+ over K+ at low concentrations (SNa+/K+ = 3), it is selective for K + over Na+ at higher concentrations (SK+/Na+ = 1.5−2).33 The selective transport of cations that are dimensionally bigger than the hole of the coordinating macrocycle is also confirmed for the benzo-18-crown-6-based compound 22. The optimal formation of the sandwich-type Rb+(18-crown-6)2 complex results in an optimal conduction of Rb+ over K + through a stacked crown sandwich channel (Figure 5D).19 However, for the much bigger Cs+ cation, which is less adapted to form a sandwich-type complex with 18-crown-6, a much lower conductance has been observed compared with K+ and Rb+. In the same way, the 15-crown-5-ether headgroups in bis(benzo-15-crown-5-etherureido)pillar[5] arene 24 contribute to the formation of the selective channels. Compound 24 is highly active toward K+ and Rb+ and slightly active toward Cs+, Na+, and Li+ (Figure 5E).34 While initially designed as the active binding sites in the membrane, the pillar[5]arene moieties play only the role of supporting groups for the macrocyclic groups used as relays for the cation-file diffusion. They self-assemble in channel-type oligomers of optimally

disposed macrocyclic crown ether sites that are sterically disposed on inactive pillar[5]arene relays.

■

CONCLUDING REMARKS The results described in this Account relate to an important number of artificial K+ channels. As in the natural KcsA channel, molecular-scale self-assembly is of crucial relevance for the observed selective supramolecular translocation. Their dynamics is highly related to structural features of the monomers and to the interactional behaviors of their resulting superstructures within the bilayers. The present examples suggest that the self-assembled macrocyclic ribbons hold significant structural information for the development of innovative K+-selective artificial channels. Of high importance is the demonstration that benzo-15-crown-5 can be efficiently used as multivalent sandwich-type binding sites toward K+. The resulting H-bonded superstructures form highly cooperative channels within bilayer membranes with high translocation rates for K+, which induces the passive polarization of the membrane obtained either under or off pH-active gradients. The crown binding sites align along a directional hydrophilic pore and selectively surround the K+ cations, compensating for the energetic cost of cation dehydration. In contrast, the incomplete coordination and dehydration of the Na+ cations make these channels highly responsive to K+ cations even in the presence of an excess of Na+ cations. These systems can be considered as adaptive self-instructed ion channels, where the K+ solute drives the selection and formation of specific superstructures for its own selective transport. The K+-selective artificial channels described here may be regarded as biomimetic alternatives to KcsA channels. Future prospects include the use of these K+ channels in materials, membranes, and sensor devices presenting a greater degree of structural complexity. The facile design and synthesis of each part of this family of supramolecular channels afford in principle a large number of possibilities to finely control and tune the selectivity and transport activity. The fast transport of K+ through the channels should lead to important practical biomedical or nanotechnology applications for chemical separations using conventional membranes, films, or microdevices.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +33-467-149195. Fax: +33-467-149119. E-mail: [email protected]. Notes

The author declares no competing financial interest. Biography Mihail Barboiu graduated from Politehnica University of Bucharest and received his Ph.D. in 1998 from the University of Montpellier before spending 2 years as a postdoctoral researcher at University Louis Pasteur in Strasbourg, France. He is Senior CNRS Research Director at the Institut Europeen des Membranes in Montpellier and a Fellow of the Royal Society of Chemistry. Since 2014 he has also been Extraordinary Professor at Sun-Yat-sen University in Guangzhou, China. A major focus of his research is dynamic constitutional chemistry toward dynamic interactive systems: adaptive biomimetic membranes, delivery devices, and so on. He is the author of more than 260 scientific publications, three books, 20 chapters, and 380 conferences and lectures and received the EURYI Award in Chemistry G

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Columnar Self-Assemblies of Triarylamines as Scaffolds for Artificial Biomimetic Channels for Ion and for Water Transport. J. Am. Chem. Soc. 2017, 139, 3721−3727. (20) Legrand, Y. M.; Barboiu, M. Self-Assembled Supramolecular Channels: Toward Biomimetic Materials for Directional Translocation. Chem. Rec. 2013, 13, 524−538. (21) Barboiu, M. Constitutional hybrid materials-toward selection of functions. Eur. J. Inorg. Chem. 2015, 2015, 1112−1125. (22) Michau, M.; Barboiu, M.; Caraballo, R.; Arnal-Hérault, C.; Perriat, P.; van der Lee, A.; Pasc, A. Ion-Conduction Pathways in SelfOrganized Ureidoarene Heteropolysiloxane Hybrid Membranes. Chem. - Eur. J. 2008, 14, 1776−1783. (23) Michau, M.; Caraballo, R.; Arnal-Hérault, C.; Barboiu, M. Alkali cation-π aromatic conduction pathways in self-organized hybrid membranes. J. Membr. Sci. 2008, 321, 22−30. (24) Barboiu, M.; Cerneaux, S.; Vaughan, G.; van der Lee, A. Iondriven ATP-pump by Self-Organized Hybrid Membrane Materials. J. Am. Chem. Soc. 2004, 126, 3545−3550. (25) Mihai, S.; Cazacu, A.; Arnal-Herault, C.; Nasr, G.; Meffre, A.; van der Lee, A.; Barboiu, M. Supramolecular self-organization in constitutional hybrid materials. New J. Chem. 2009, 33, 2335−2343. (26) Cazacu, A.; Legrand, Y. M.; Pasc, A.; Nasr, G.; van der Lee, A.; Mahon, E.; Barboiu, M. Dynamic hybrid materials for constitutional self-instructed membranes. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 8117−8122. (27) Liu, T.; Bao, C.; Wang, H.; Lin, Y.; Jia, H.; Zhu, L. Lightcontrolled ion channels formed by amphiphilic small molecules regulate ion conduction via cis−trans photoisomerization. Chem. Commun. 2013, 49, 10311−10313. (28) Liu, T.; Bao, C.; Wang, H.; Fei, L.; Yang, R.; Long, Y.; Zhu, L. Self-assembly of crown ether-based amphiphiles for constructing synthetic ion channels: the relationship between structure and transport activity. New J. Chem. 2014, 38, 3507. (29) Ren, C.; Shen, J.; Zeng, H. Combinatorial Evolution of FastConducting Highly Selective K+-Channels via Modularly Tunable Directional Assembly of Crown Ethers. J. Am. Chem. Soc. 2017, 139, 12338−12341. (30) Thomas, M.; Jayatilaka, D.; Corry, B. The Predominant Role of Coordination Number in Potassium Channel Selectivity. Biophys. J. 2007, 93, 2635−2643. (31) Zhou, Y.; Morais-Cabral, J. H.; Kaufman, A.; MacKinnon, R. Chemistry of ion coordination and hydration revealed by a K+ channel−Fab complex at 2.0 Å resolution. Nature 2001, 414, 43−48. (32) Saparov, S. M.; Pohl, P. Beyond the diffusion limit: water flow through the empty bacterial potassium channel. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 4805−4809. (33) Li, Y.-H.; Zheng, S.-P.; Legrand, Y.-M.; Gilles, A.; Van der Lee, A.; Barboiu, M. Structure-driven selection of adaptive transmembrane Na+ carriers or K+ channels. Angew. Chem., Int. Ed. 2018, 57, 10520− 10524. (34) Feng, W.-X.; Sun, Z.; Zhang, Y.; Legrand, Y.-M.; Petit, E.; Su, C.-Y.; Barboiu, M. Bis-15-Crown-5-Ether-Pillar[5]arene K+-Responsive Channels. Org. Lett. 2017, 19, 1438−1441. (35) Behr, J.-P.; Lehn, J.-M.; Dock, A.-C.; Moras, D. Crsytal structure of a polyfunctional macrocylic K+ complex provide a solid state model of a K+ channel. Nature 1982, 295, 526−527. (36) Fromm, K. M.; Bergougnant, R. D. Transport properties of solid state crown ether channel systems. Solid State Sci. 2007, 9, 580− 587. (37) Otis, F.; Racine-Berthiaume, C.; Voyer, N. How Far Can a Sodium Ion Travel within a Lipid Bilayer? J. Am. Chem. Soc. 2011, 133, 6481−6483. (38) Su, G.; Zhang, M.; Si, W.; Li, Z.-T.; Hou, J.-L. Directional Potassium Transport through Unimolecular Peptide Channel. Angew. Chem., Int. Ed. 2016, 55, 14678−14681.

in 2004 and the RSC Surfaces and Interfaces Award in 2015 for the development of artificial water channels.

■

ACKNOWLEDGMENTS This work was conducted within ANR-15-CE29-0009 DYNAFUN and 1000 Talent Plan, WQ20144400255 of SAFEA, China. I thank the past and the present members of my laboratory and the many individuals with whom we have collaborated in these studies: Yuhao Li, Zhanhu Sun, Shaoping Zheng, Wei-Xu Feng, Arnauld Gilles, Arie van der Lee, Adinela Cazacu, Yves Marie Legrand, Eddy Petit, Mathieu Michau, Remi Caraballo, Gihane Nasr, Andreea Pasc-Banu, Carole Arnal-Herault, and Anca Meffre.

■

REFERENCES

(1) Ball, P. Water as an active consitutent in cell biology. Chem. Rev. 2008, 108, 74−108. (2) Hurtley, S. M. Crossing the Bilayer. Science 2005, 310, 1451− 1451. (3) Ashcroft, F. M. From molecule to malady. Nature 2006, 440, 440−447. (4) Hille, B. Ion Channels of Excitable Membranes, 3rd ed.; Sinauer Associates: Sunderland, MA, 2001. (5) Mackinnon, R. Potassium Channels and the Atomic Basis of Selective Ion Conduction. Angew. Chem., Int. Ed. 2004, 43, 4265− 4277. (6) Gokel, G. W.; Mukhopadhyay, A. Synthetic Models of CationConducting Channels. Chem. Soc. Rev. 2001, 30, 274−286. (7) Voyer, N. The development of peptide nanostructures. Top. Curr. Chem. 1996, 184, 1−37. (8) Matile, S. En Route to Supramolecular Functional Plasticity: Synthetic β-Barrels, the Barrel-Stave Motif, and Related Approaches. Chem. Soc. Rev. 2001, 30, 158−167. (9) Davis, J. T.; Spada, G. P. Supramolecular architectures generated by self-assembly of guanosine derivatives. Chem. Soc. Rev. 2007, 36, 296−313. (10) Fyles, T. M. Synthetic Ion Channels in Bilayer Membranes. Chem. Soc. Rev. 2007, 36, 335−347. (11) Arnal-Herault, C.; Barboiu, M.; Petit, E.; Michau, M.; van der Lee, A. Cation-π Interaction: A case for Macrocycle-Cation πInteraction by its Ureidoarene Counteranion. New J. Chem. 2005, 29, 1535−1539. (12) Arnal-Herault, C.; Michau, M.; Barboiu, M. Mixed supramolecular cation-carrier and anion-carrier facilitated transport for selective alkali cations transport. J. Membr. Sci. 2008, 321, 94−99. (13) Barboiu, M.; Vaughan, G.; van der Lee, A. Self-Organised Heteroditopic Macrocyclic Superstructures. Org. Lett. 2003, 5, 3073− 3076. (14) Barboiu, M. Supramolecular Polymeric Macrocyclic Receptors Hybrid Carrier vs. Channel Transporters in Bulk Liquid Membranes. J. Inclusion Phenom. Mol. Recognit. Chem. 2004, 49, 133−137. (15) Cazacu, A.; Tong, C.; van der Lee, A.; Fyles, T. M.; Barboiu, M. Columnar Self-Assembled Ureidocrown ethers − an Example of Ionchannel Organization in Lipid Bilayers. J. Am. Chem. Soc. 2006, 128, 9541−9548. (16) Sun, Z.; Barboiu, M.; Legrand, Y. M.; Petit, E.; Rotaru, A. Highly Selective Artificial Cholesteryl Crown Ether K+-Channels. Angew. Chem., Int. Ed. 2015, 54, 14473−14477. (17) Sun, Z.; Gilles, A.; Kocsis, I.; Legrand, Y. M.; Petit, E.; Barboiu, M. Squalene Crown-Ether Self-Assembled Conjugates−an example of highly selective artificial K+-channels. Chem. - Eur. J. 2016, 22, 2158− 2164. (18) Gilles, A.; Barboiu, M. Highly Selective Artificial K+ Channels: An Example of Selectivity- Induced Transmembrane Potential. J. Am. Chem. Soc. 2016, 138, 426−432. (19) Schneider, S.; Licsandru, E.-D.; Kocsis, I.; Gilles, A.; Dumitru, F.; Moulin, E.; Tan, J. J.; Lehn, J.-M.; Giuseppone, N.; Barboiu, M. H

DOI: 10.1021/acs.accounts.8b00311 Acc. Chem. Res. XXXX, XXX, XXX−XXX