Kylie
Chinner
a,
Niklas
Grabicki
a,
Rei
Hamaguchi
b,
Mitsunori
Ikeguchi
c,
Kazushi
Kinbara
bd,
Sayaka
Toyoda
e,
Kohei
Sato
*be and
Oliver
Dumele
*af
aDepartment of Chemistry and IRIS Adlershof, Humboldt-Universität zu Berlin, Brook-Taylor-Str. 2, Berlin 12489, Germany
bSchool of Life Science and Technology, Tokyo Institute of Technology, Yokohama, Kanagawa 226-8501, Japan
cGraduate School of Medical Life Science, Yokohama City University, Yokohama, Kanagawa 230-0045, Japan
dResearch Center for Autonomous SystemMaterialogy (ASMat), Institute of Innovative Research, Tokyo Institute of Technology, Kanagawa 226-8501, Japan
eDepartment of Chemistry, School of Science, 1 Gakuen Uegahara, Sanda-shi, Hyogo 669-1330, Japan. E-mail: ksato@kwansei.ac.jp; Web: https://www.ksatolab.net
fInstitute of Organic Chemistry, University of Freiburg, Albertstr. 21, Freiburg 79104, Germany. E-mail: oliver.dumele@oc.uni-freiburg.de; Web: https://www.dumelelab.com
First published on 11th September 2024
Nanohoops, an exciting class of fluorophores with supramolecular binding abilities, have the potential to become innovative tools within biological imaging and sensing. Given the biological importance of cell membranes, incorporation of macrocyclic materials with the dual capability of fluorescence emission and supramolecular complexation would be particularly interesting. A series of different-sized nanohoops—ethylene glycol-decorated [n]cyclo-para-pyrenylenes (CPYs) (n = 4–8)—were synthesised via an alternate synthetic route which implements a stannylation-based precursor, producing purer material than the previous borylation approach, enabling the growth of single-crystals of the Pt-macrocycle. Reductive elimination of these single-crystals achieved significantly higher selectivity and yields towards smaller ring-sized nanohoops (n = 4–6). The supramolecular binding capabilities of these CPYs were then explored through host–guest studies with a series of polycyclic (aromatic)hydrocarbons, revealing the importance of molecular size, shape, and CH–π contacts for efficient binding. CPYs were incorporated within the hydrophobic layer of lipid bilayer membranes, as confirmed by microscopic imaging and emission spectroscopy, which also demonstrated the size-preferential incorporation of the five-fold nanohoop. Molecular dynamics simulations revealed the position and orientation within the membrane, as well as the unique non-covalent threading interaction between nanohoop and phospholipid.
Recently, carbon-rich nanohoops as an emerging class of biological fluorophores for sensing and imaging have been of particular interest,27–30 some recent examples being the aqueous-soluble sulfonated [8]cycloparaphenylene which can penetrate live cells27 and [10]cycloparaphenylene assembled into cell-incorporated nanoscale vesicles.29 While nanohoops within membranes have been previously studied for bioimaging and sensing applications, their unique supramolecular complexation capabilities within membranes has yet to be exploited. Additionally, nanohoops have large Stokes shifts (110–250 nm), with no spectral overlap between absorbance and emission.12 This means, in principle, a mixture of ring sizes can all be excited at the same wavelength, resulting in a size-dependent readout, facilitating localisation.
Previously, we reported that amphiphilic molecules can be effectively incorporated into lipid bilayer membranes31–37 to form intriguing supramolecular structures such as reversible calcium-induced assemblies,38 or self-assembled supramolecular transmembrane ion channels.36
Based on those findings, the advantageous fluorescence emission of nanohoops,39 paired with their shape-persistence40 and defined inner void with guest binding capabilities, we envisioned that functionalised cycloparaphenylene-type nanohoops would exhibit unique properties when incorporated into artificial and natural membranes.
While several methods have been established to access the shortest possible repeating units of carbon nanotubes, called cycloparaphenylenes,41–47 applications remain in their infancy due to synthetic challenges imposed by the high energy required to bend an aromatic ring out of plane. Larger nanorings48,49—such as Isobe and co-workers’ phenine nanoring50,51—avoid this, achieving an almost strain-free macrocycle through a considerably wider diameter. While larger nanorings can undergo host–guest binding, prominently with fullerene, efficient binding of non-fullerene guests to nanohoops with smaller diameters is still rare.52,61–66
Recently, we reported the synthesis of a series of highly functionalised nanohoops, [n]cyclo-2,7-(4,5,9,10-tetrahydro)pyrenylenes (CPYs),52 following the platinum-mediated macrocyclisation developed by Yamago and co-workers.54 These nanohoops featured ethylene glycol moieties which decorate the outer rim and confine the inner cavity. The five-fold CPY in particular, possesses high fluorescence emission and an optimal sized cavity for small molecule binding, making it an excellent shape-persistent candidate to study supramolecular host–guest interactions. However, utilising the reversible Pt-complexation often results in complex product mixtures, typically containing a range of ring sizes of reductively eliminated carbon nanorings. 43,44,52 Thus, additional selectivity is desired.
Herein, we report a size-selective route towards CPYs with higher overall yields and ring size selectivity (n = 4–6), with the reductive elimination on isolated single-crystalline Pt-macrocycle as a key step. With this new-found control in synthesis over size to achieve precise molecular entities, we took advantage of the size distribution and identified the five-fold as the most favourable CPY—regarding its high fluorescence emission and optimal central void volume—for encapsulation of non-polar hydrocarbon guests and selective incorporation into phospholipid bilayer membranes (Fig. 1).55
Fig. 1 (a) [n]Cycloparaphenylenes ([n]CPP) and (b) reported [n]CPP derivatives as biological nanohoop fluorophores.27,28 (c) [n]CPYs as biological nanohoop fluorophores for host–guest complexation with polycyclic aromatic hydrocarbon (PAH) guests and incorporation into the hydrophobic layer of phospholipid bilayer membranes. |
Scheme 1 Synthesis of 2 as previously synthesised52 starting from 2,7-bisbromo-pyrene-4,5,9,10-tetrone 5. Reagents and conditions: (Ia) camphorsulfonic acid, ethylene glycol, MeOH, 120 °C, 24 h, 73%; (IIa) 4,4′-di-tert-butyl-2,2′-bipyridine, [Ir(OMe)(COD)]2, bis(pinacolato)diboron, 1,4-dioxane, 120 °C, 18 h, 84%; (IIIa) [PtCl2(COD)], CsF, CH2Cl2, 45 °C, 24 h, 74%. Synthesis of 2 as single-crystals starting from 2,7-bisbromo-pyrene-4,5,9,10-tetrone 5. Reagents and conditions: (Ib) camphorsulfonic acid, ethylene glycol, MeOH, 120 °C, 48 h, 70%; (IIb) Pd(PPh3)4, LiCl, 2,4,6-tri-tert-butylphenol, (Sn(Me3))2, 1,4-dioxane, 110 °C, 2 h, 77%; (IIIb) Pt(COD)Cl2, THF, 60 °C, 24 h, 74%. |
To obtain bisstannane precursor 3, the K-region of pyrene was symmetrically functionalised via a Ru-catalysed oxidation to obtain pyrene-4,5,9,10-tetrone 6 as previously reported (see ESI Scheme S1†).52 Bromination with NBS at the 2 and 7 positions gave 5, followed by an acid-catalysed condensation with ethylene glycol to afford 4 (Scheme 1, bottom). Next, Pd-catalysed stannylation gave the 2,7-bis(trimethylstannyl)pyrene precursor 3, then subsequent transmetalation with an equimolar amount of dichloro(1,5-cyclooctadiene)platinum(II) produced the macrocyclic Pt(II) precursor 2 in 74% yield as an off-white solid that crystallised upon slow evaporation from a saturated solution in CH2Cl2, resulting in crystals suitable for X-ray diffraction (Fig. 2).
Fig. 2 Single-crystal X-ray structure of platinum macrocycle 2 with side view and Yamago and co-workers’ 2014 platinum macrocycle44 for comparison. Ellipsoids are shown at 50% probability; hydrogen atoms and solvent molecules are omitted for clarity. Colour code: carbon, grey; oxygen, red. |
Interestingly, this platinum macrocycle differs in geometry from Yamago and co-worker's platinum macrocycle.44 While both are comprised of a repeating pyrene scaffold, introduction of sterically demanding ethylene glycol motifs distort the nanoring from planarity to a non-planar ‘butterfly’ geometry—as revealed by the CAr–Pt(II)–CAr angles of 86.4°, which are slightly smaller than those of Yamago and co-workers’ 4,5,9,10-tetrahydropyrenylene-containing platinum metallacycle (91.1° and 87.6°, Fig. 2).
Reductive elimination on single-crystals of the platinum metallacycle precursor 2 gave the targeted CPYs 1[n] with unprecedented size-selectivity of 1[4]/1[5]/1[6] 7:1:<1. This is a significant increase in ring size selectivity compared to the reductive elimination from the typical crude Pt-metallacycle mixture, which was previously reported to 1[4]/1[5]/1[6]/1[7]/1[8] 2:2:2:2:1, and an overall improved yield for the preferred smaller ring sizes (from 5% to 21% for 1[4]). Separation of ring sizes by recycling gel permeation chromatography (rGPC) unambiguously illustrates this selectivity, with a significantly larger peak area permeating for the four-fold ring, 1[4], after reductive elimination of single-crystals (Fig. 3, right). The residual larger ring sizes n > 4 originate from either small impurities within the single-crystalline material containing larger platinum metallacycles, or from Pt(PPh3)4–n formation—which is known to insert into strain-activated C–C bonds—and subsequent ligand exchange.59,60 While the more selective synthesis of 1[4] provides valuable insight into the reaction mechanism, this ring size remains of least importance compared to its larger congeners for guest binding due to the very small gate opening and inner cavity. Therefore, we turned to 1[5] to explore new guest binding abilities to expand upon previously reported cationic crown–ether complexes.52,53
Fig. 4 (a) X-ray crystal structure of 1[5]⋯COR (100 K). Ellipsoids are shown at 50% probability; hydrogen atoms and solvent molecules are omitted for clarity. Colour code: carbon, grey; oxygen; red. (b) Calculated geometry of 1[5]⋯pyrene, level of theory: DFT:B3LYP/6-31G(d). (c) Guest molecules used in host–guest studies with 1[5]. Association constants (Ka) obtained via1H NMR titration experiments in benzene-d6 at 298 K (see ESI Fig. S19†). Packing coefficients (PC) obtained from the calculated cavity volumes of X-ray (1[5]) and geometry-optimised (guests) structures using the MS Roll suite implemented in X-Seed (see ESI Section S5†). Diameters (d) were derived from geometry-optimised structures, with the largest distance measured between two hydrogens taken (highlighted in red). |
Through isothermal 1H NMR titrations at slow exchange, corannulene67–70 (9) was found to bind the strongest of all guests screened within this study, with an association constant of 990 M−1 in benzene-d6 at 298 K (see ESI Fig. S19†). The C5 symmetry of 1[5] produces a favourable alignment for corannulene's (9) C5V symmetric bowl-shape and 10 peripheral hydrogen atoms, leading to optimal average CH–π contact distance of dC–π = 3.62(10) Å as satisfyingly revealed by single-crystal X-ray crystallography (see ESI Section S4.2†).
This is particularly intriguing as corannulene (9) is known to bind with smaller sized nanohoops, such as [10]cycloparaphenylene66 (d[10]CPP = 13.8 Å, Ka = 170 ± 50 M−1 in CDCl3 at 293 K) and [4]cyclo-2,8-chrysenylene ([4]CC)64 (d[4]CC = 13.4 Å, Ka = 2940 M−1 in CD2Cl2 at 298 K), whereas the cavity of 1[5] should be too large (15.2 Å) for efficient binding. Nonetheless, binding does occur. Therefore, the highly functionalised rim of 1[5] (gate diameter of 9.8 Å) must be a significant factor in enabling corannulene (9) to complex within the oversized inner cavity of 1[5]. Thus, the unique ethylene glycol-decorated rim of CPYs can enable binding with guests that would typically not occur due unfavourable size difference between host cavity and guest.
Pyrene (10), the second strongest bound hydrocarbon guest of the investigated series, had a significantly smaller association constant of 195 M−1 in benzene-d6 at 298 K (see ESI Fig. S20†). While both corannulene (9) and pyrene (10) possess 10 exposed hydrogen atoms, the hydrogens of pyrene (10) are arranged in an ovoid fashion. This ovaloid shape likely leads to suboptimal contact, and therefore weaker CH–π interactions between pyrene (10) and 1[5]. Additionally, while other strained carbon nanorings can undergo flexible alignment to gain optimal CH–π contacts around guest molecules, as seen with Isobe and co-workers’ [4]cyclo-2,8-chrysenylene,64 the bulky ethylene glycol groups of 1[5] likely prevent this distortion due to steric hindrance.
As the cavity of 1[5] is both confined and restricted from distortion by the highly functionalised rim, size and shape are crucial factors in host–guest complexation. Coronene (12), while possessing 12 hydrogen atoms, increasing the potential for CH–π interactions, is too large and reveals no evidence of binding upon mixing with 1[5] in benzene-d6 at 298 K—as with perylene (13), which is too wide to fit within the confined cavity. Initially, this was surprising as coronene (12) is known to bind with [11]cycloparaphenylene66 (15.2 Å), which shares a similar inner cavity diameter as 1[5] (15.2 Å). However, when considering the smaller gate diameter (9.8 Å) of 1[5], facilitated by its highly functionalised rim, the lack of complexation becomes understandable.
Benzo[1,2-b:3,4-b′:5,6-b″]trithiophene (11), which possesses a similar diameter as corannulene (9), did not bind—possibly lacking the sufficient amount of C–H moieties required for measurable complex formation. We hypothesised that corannulene's (9) ability to undergo bowl inversion72–75 possibly allowed the fluctuating molecule to fit within the cavity despite its curved shape, hence we screened the partially unsaturated and more flexible 2,3,4,5,6,7,8,9-octahydro-1H-trindene (14). However, it did not bind—likely due weaker Csp3H–π interactions and suboptimal contact angles from the less polarised alkane hydrogens which cannot point directly at the π-electrons of 1[5].
Interestingly, changing solvents to chlorinated tetrachloroethane-d2 resulted in a lower association constant for corannulene (9) (622 M−1, compared to 990 M−1 in benzene-d6, both at 298 K) (see ESI Fig. S21†). While, unexpectedly, 1H NMR titrations in CDCl3 displayed no binding. A single-crystal X-ray structure of 1[6] grown from CHCl3 shows competitive solvent interaction between multiple chloroform molecules and the oxygen atoms of the ethylene glycol moieties at the outer rim of the nanohoop's cavity, which reveal an enthalpically favourable solvate complex compared to non-polar hydrocarbon guest binding (see ESI Fig. S13c†).
Fig. 5 1H NMR spectra (600 MHz) for the assignment of the complexation of 1[5] and corannulene (COR) in tetrachloroethane-d2 (TCE-d2) at 298 K. |
For uncomplexed 1[5], two broad 1H NMR signals corresponding to the ethylene glycol moieties (ca. 4.25 and 3.70 ppm) are characteristic, with the equatorial protons shifted further downfield than the axial protons due to hyperconjugation with oxygen.71 The pseudo doublet of the axial protons is typical of the smaller ring sizes of CPYs (1[4] and 1[5]).52 However, upon guest addition, the 1H NMR signals of the ethylene glycol groups in 1[5] began to split (Fig. 6) (see ESI Fig. S21†). As both equatorial and axial inner protons of the ethylene glycol moiety face the inner cavity of 1[5], they can be influenced by the presence of a guest. Consequently, for the complexation of 1[5]⋯COR, the 1H NMR signals of ‘free’ and ‘bound’ ethylene glycol split further into ‘bound’ and ‘free’ equatorial inner protons (Heq. inner) as well as ‘free’ and ‘bound’ axial inner (Hax. inner) (these signals were identified and assigned via1H–1H ROESY and 1H–1H NOESY experiments, see ESI Fig. S22 and S23†). The ethylene glycol protons facing away from the cavity (Heq. outer and Hax. outer) maintain similar chemical shifts to the uncomplexed host, with the same pseudo doublet of the axial protons present. A relative population of free/bound COR in 1[5]⋯COR was found to be 38:62 at 8 equivalents of corannulene at 298 K (see ESI Fig. S21†).
Fig. 6 1H NMR spectra (C2D2Cl4, 600 MHz, 298 K) illustrating the assignment of ethylene glycol groups of 1[5]⋯COR with a schematic complexation of free 1[5], free corannulene, and the complex 1[5]⋯COR. Only one wall fragment of 1[5] in the complex is illustrated for clarity. Solid-convex and transparent-concave corannulene are shown as a representation of corannulene's bowl inversion. Assigned via1H–1H ROESY and 1H–1H NOESY spectra with key NOE cross-signals marked as red arrows (for further details, see ESI Fig. S22 and S23†). |
Additionally, corannulene is known to undergo degenerate bowl-to-bowl inversion72–75 which passes through a planar D5h transition state. As 1[5] can supply a distinct chemical environment through the ‘inner’ and ‘outer’ protons of its ethylene glycol decorated rim, we hypothesised that the unique splitting pattern of our host, paired with the increased stability gained from host–guest complexation, could reveal the bowl inversion of nascent corannulene at low temperatures, similar to that of Exbox4+ revealing the bowl inversion of nascent and ethylcorannulene.75 Unfortunately, despite all efforts, tested solvent systems either froze or were too competitive (see ESI Fig. S24†).
To explore the functions of 1[5], we tested its potential to transport ions across lipid bilayer membranes, expecting its pore to function as a transmembrane channel or ionophore. For this purpose, we prepared LUVs encapsulating 8-hydroxypyrene-1,3,6-trisulfonic acid (HPTS), a pH-sensitive fluorescent dye (see ESI Section S7.4†). We incorporated 1[5] and then created a pH gradient across the membranes so that the transmembrane ion transport could be monitored by an increase in the fluorescence intensity of HPTS. 35,36,55 However, as shown in Fig. S27,† we did not observe any changes in fluorescence intensity. We also performed current recording measurements of 1[5]-containing planar lipid bilayer membranes under the application of voltage, but we did not observe any current signals (Fig. S28†). These results suggest that 1[5] does not possess transmembrane ion transport properties.
To understand the reason for the biased Θ value, we analysed the trajectories during the simulations and unexpectedly found that the alkyl tail of DOPC molecule was incorporated into the cavity of 1[5] (Fig. 8h). Similar to the host–guest complexes of aromatic molecular nanocapsules and hydrocarbons reported by Rebek and coworkers,82,83 the flexible alkyl tail of DOPC adopted a folded conformation to fill the inner void of 1[5]. The biased Θ-value of 1[5] can be understood as the result of this complexation event. The complexation also explains the reason for the inability of 1[5] to transport ions across the membranes, because the inner cavity of 1[5] is fully occupied. This complexation capability of 1[5], along with the enhanced solubility of smaller ring-sized CPYs (n = 4–6), are likely the main contributing factors that drive the size-preferential incorporation of 1[5] into the membrane.
Experimental efforts were then undertaken to investigate the binding ability of 1[5] with DOPC (up to 10 equiv.). Unfortunately, after 1H NMR isothermal titrations of 1[5] with DOPC in benzene-d6 at 298 K, no binding was observed (see ESI Fig. S25†). The association constant of the weak CH–π interactions between DOPC alkyl tail and 1[5] are likely below the detection limit, with the hydrophobic environment of the bilayer membrane likely driving the complexation event to occur. Replicating this environment for 1H NMR titrations remained unachievable as CPYs are insoluble in aqueous media.
These results demonstrate the intriguing function of nanohoops interacting with phospholipids, one of the vital components of life,84 and their potential ability to act as supramolecular sensors or functional modulators of cellular membranes.
Footnote |
† Electronic supplementary information (ESI) available: For ESI and crystallographic data in CIF format. CCDC 2049559 and 2236136. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4sc03408b |
This journal is © The Royal Society of Chemistry 2024 |