Sean P.
Bew
*a,
Sunil V.
Sharma
b and
William H.
Gardiner
c
aSchool of Chemistry, UEA, Norwich Research Park, Norwich, NR4 7TJ, UK. E-mail: s.bew@uea.ac.uk
bSchool of Chemistry, University of St Andrews, Purdie Building, North Haugh, St Andrews, Scotland, UK
cEvotec UK Ltd, Innovation Drive, Milton Park, Abingdon, Oxfordshire, OX14 4RZ, UK
First published on 14th March 2023
A unique, readily synthesised, upper-rim 1,3-difunctionalised calix[4]arene is reported; equipped with a fluorophore and an alkyne it mediates the efficient chemoselective detection of Au3+. Its ability to detect Au3+ is not perturbed by Au+ or excess competing and or contaminating metal e.g. platinum, cadmium, mercury, silver, sodium, magnesium or potassium cations.
Recent years have witnessed a sharp increase in the use of gold in developing new methodology based on, for example, homogeneous and heterogeneous catalysis. It is, however, not the sole preserve of chemists. As noted, its potential has been recognized, explored, and unlocked in biology, medicine, and physics. For example, Au55 clusters ([Au55(Ph2PC6H4SO3)12]12−) are used to visualize cell component's5 and in medicine, Tauredon, Solganal, Sanocrycin, and Auranofin are used to fight asthma, HIV, malaria, lesions, cancer, and arthritis. Thus, although gold is in many respects ‘compatible’ it can have adverse effects on the environment,6 and in biological systems Au3+ metallodrugs can be reduced to Au+.7 For these reasons and due to its increased use, the development of new modalities and protocols for differentiating Au3+/Au+ are urgently required. Ideally, these must be of relatively low cost, easily synthesized, chemoselective and, importantly, they should not be perturbed by other metal contaminants e.g., mercury, silver, platinum, and cadmium. No reaction-based fluorescent probes for gold detection have been developed which proceed via a ‘turn off’ mechanism and use readily synthesized calix[4]arenes derived from 1 (Scheme 1).8 Exploiting our expertise and knowledge of calix[4]arenes and their upper-rim chemical functionalization9 we were intrigued by the possibility of transforming scaffold 1 into a chemospecific Au3+ metal-binder (2) which specifically presents, facilitates and, subsequently, reports its presence via adjacent alkyne and dansyl ‘antenna’ (cf.3, Scheme 1). Attracted by the chemical and operational simplicity of our system, the straightforward synthesis of entities based on 3 uses easily generated starting material 1 and proceeds via a simple, readily observable ‘turn-off’ mechanism (i.e., 2 → 3) when the metal is bound/detected (Scheme 1). With strategies to generate upper-rim functionalized calix[4]arenes in hand, key to our proposed application is the π-aurophilic alkyne which chemoselectively ‘captures’ the Au3+. Substantiating this as a mode of action, Ultimoto and Fukuda reported internal alkynes react readily with a catalytic source of Au3+ in the form of sodium tetrachloroaurate which in aqueous methanol transforms an alkyne into a ketone (17, Scheme 3).10 With the gold ‘captured’ the conformational mobility of the calix[4]arene11 helps facilitate the fluorescent dansyl group to move closer and interact with the metal whilst ‘reporting’ its presence via fluorescent quenching (Scheme 1).
With this rationale in mind we were also convinced of the merits of generating entities based on 2 due to the widespread use of the calix[4]arene motif in alternate metal detection regimes which, when dovetailed with the known ability of gold salts to mediate efficient internal nonradiative relaxation decay, results in fluorescent quenching (FQ).12 Supporting our approach to gold detection are the abundant applications of the calix[4]arene motif in metal recognition but, surprisingly, the lack of any specific calixarene-based Au3+ modality. Thus, although there are many examples of cone-confined calix[4]arenes with lower-rim metal cation sensing properties,13 this fact contrasts sharply with the handful of known metal sensing upper-rim functionalized calix[4]arenes.14
The desired mono-phenylacetylene-derived 5 (Scheme 2) was used ‘as is’ in the next step [an estimated 15% of the corresponding di-phenylacetylene adduct (not shown) was also formed]. A subsequent Suzuki–Miyaura coupling with 3-(dansylamino)phenylboronic acid (6) catalysed by bis(triphenylphosphine)palladium(II) dichloride (5 mol%), tri-tert-butylphosphine (10 mol%) and potassium carbonate afforded, after chromatographic purification, the desired product 7. We propose the detection capabilities of 7 (vide infra) rely on the capacity of the dansyl group and π-aurophilic alkyne to ‘communicate’ with the gold only when all three are in close proximity (cf.3, Scheme 1). Clearly, it was important to verify the presence of the dansyl and alkyne groups, as well as the conformation of the calix[4]arene. Frustrated by its reluctance to afford crystals suitable for X-ray diffraction, we analyzed the 13C-NMR data associated with the bridging methylene's which, Mendoza et al. report, allow the conformation of a calix[4]arene to be established.17 The 13C-NMR signal at 31 ppm confirmed 7 resided in a cone-conformation. Additional physicochemical studies using FT-IR and mass spectrometry verified the dansyl and alkyne groups were present. A fundamental reason for incorporating and using the dansyl group is its strong fluorescence signal and long emission wavelength. Probing the photophysical properties of 7 a 5 μM solution had an excitation absorption maximum at 288 nm (λex) which generated a strong fluorescent emission signal (λem) with a λmax of 522.9 nm. A weaker signal was observed at 353.9 nm (Scheme 2). Confirming the signal at 522.9 nm was broadly like other λem of alternative upper- and lower-rim dansyl-derived calix[4]arenes our value is in general agreement with literature reports of λmax at 538 nm and 520–575 nm respectively.18 Recording additional spectra of 7 every two minutes for 45 minutes we established the emission (λem) intensity remained constant. This validated the fluorophore was photochemically stable.
Scheme 2 Synthesis of upper-rim bis-5,17-functionalised calix[4]arene 7 from bis-5,17-N-urea-derived calix[4]arene 4. |
It was important to assess if 7 was susceptible to non-chemoselective metal mediated FQ. That is, does it generate false positives? Screening individual solutions of Na+, K+, Cs+, Ca2+ and Mg2+ salts for their effect, if any, on the strong λem signal would help answer this question. Adding 1–5 equivalents of aqueous sodium chloride (2.5 μM) afforded almost identical ‘before’ and ‘after’ λem spectra, even after 3 hours (cf. Scheme 2 with Fig. 1 in the ESI†). Similarly, 1–5 equivalents of a mixture of Li+, Na+, K+, Cs+, Ca2+ and Mg2+ salts (all at 2.5 μM) had little effect. There are reports that Group 12 metal cations FQ the dansyl group. For example, Bartsch et al. described a calix[4]arene appended on the lower-rim with two dansyl groups as a fluorogenic Hg2+ selective extractant which underwent FQ.19 Screening individual solutions of Group 12 salts [cadmium acetate, mercury(II) chloride, hydrogen tetrachloroplatinate and silver tetrafluoroborate] had little effect on the photophysical properties of 7.
Indeed, only with >15 equivalents and extended mixing times was partial FQ observed (see ESI†). Further stress testing its photophysical properties 1–5 equivalents of a mixture of Group 2 (Mg2+ and Ca2+), Group 10 (Pt4+) and Group 12 (Cd2+ and Hg2+) salts had minimal effect on 7's λmax, even after 3 hours (Fig. 1).
Fig. 1 The fluorescent spectrum above of 7 outlines how a mixture of Group 2, 10 and 12 metal salts had little FQ effect on 7. |
Au3+ and Au+ are strong Lewis acids, indeed both can activate alkynes to nucleophilic attack with weak nucleophiles e.g., water or simple alcohols.20 Clearly, if 7 was not able to chemoselectively differentiate Au3+ and Au+ its use would be severely limited. With this in mind we were delighted adding 5 equivalents of chloro(triphenylphosphine)gold(I) to 7 had little impact on the λmax, even after 3 hours (Fig. 2).
With a reliable synthesis in hand and robust photophysical data confirming little evidence of FQ with either single or mixtures of metal salts, it was essential to establish if 7 was able to ‘capture, sense and report’. Adding a solution of sodium tetrachloroaurate to a 5 μM solution of 7, an almost instantaneous reduction in the λmax signal was observed (Fig. 3). Increasing the equivalents from 1 to 2, 3 and 4 induced further fluorescent quenching and a reduction in the λmax intensity (Fig. 3). The lack of FQ with 7 in a protic solvent when presented with Group 10 or 12 metal salts contrasts sharply with observations using the gold salt (Group 11); this suggests key to FQ is the formation of significant and favourable interactions between the alkyne on 7 and the highly charged Au3+.
Fig. 3 Efficient fluorescent quenching of 7 appended with an upper-rim dansyl and phenylacetylene group. |
Evidently, bifunctional calix[4]arene 7 readily and efficiently detected the gold salt in the absence of, potentially, competing metal salts. Critical to its wider application however is its ability to retain this property when presented with a mixture of metal salts. So, do externally added salts inhibit or complicate the use of 7? Generating individual solutions of (2.5 μM) sodium chloride, potassium hexafluorophosphate, calcium chloride, magnesium chloride and cesium chloride (see ESI†) as well as a mixture (2.5 μM) of lithium chloride, sodium chloride, potassium hexafluorophosphate, cesium chloride, magnesium chloride and calcium chloride their addition (5 equivalents each) to 7 and, subsequently, sodium tetrachloroaurate (0–5 equivalents) allowed us to compare ‘contaminated’ (Fig. 4) with a ‘non-contaminated’ gold salt solution (Fig. 3). Fig. 4 clearly demonstrates the strong λmax signal at 520.3 nm was quenched in a manner comparable to ‘pure’ sodium tetrachloroaurate. Further probing and testing the robustness of a ‘contaminated’ system employed a mixture of 20 equivalents of cadmium acetate, mercury(II) chloride and hydrogen tetrachloroplatinate. Comparing Fig. 4 with Fig. 5 their similarities validated the efficient FQ by Au3+ to be unaffected by extraneous, potentially, competing salts.
Fig. 5 Fluorescent quenching of 7 in the presence of 20 equivalents each of Cd2+, Hg2+ and Pt4+ salts. |
Thus far, the alkyne was assumed to be an integral component of 7, confirming this was essential. Hydrogenation of 7 afforded the alkyne-free hydrogenated ‘diphenylethane’ adduct (not shown). Acquiring its fluorescent spectrum after 30 equivalents of sodium tetrachloroaurate had been added verified it did not undergo FQ. This result substantiates the key role the alkyne has in, presumably, ‘capturing’ and acting as an ‘aurophile’ bringing the metal into close proximity to the dansyl group generating ‘alkyne-Au3+-dansyl’ interactions (cf.3) which, ultimately, initiate FQ. Seemingly, removing its ability to form a gold-alkyne complex is the first step since, by a process of elimination, the alkyne-free system seems unable to form gold-dansyl interactions which, if formed, would still, presumably, result in FQ.
Although the exact role the macrocycle occupies is not clear, its presence is evidently important when combined with the two upper-rim substituents. Testing for a ‘macrocycle’ effect ‘linear’ macrocycle-free 8 was generated from readily available starting materials. With its urea and dansyl modalities (Fig. 6) matching those of its cyclic ‘cousin’, 8 afforded the perfect opportunity to evaluate the importance of the calix[4]arene. Adding increasing equivalents (up to 20) of sodium tetrachloroaurate a slight reduction in FQ was observed. The propensity of sulfonamide groups to bind cationic metal centers is known.21 Thus, it seems plausible the observed FQ is associated with intermolecular functional group interactions derived from two molecules of 8 combining to form a weak ‘dansyl-Au3+-dansyl’ complex like 9 (Fig. 6). The ability of upper-rim appended urea-derived calix[4]arenes to bind anions is well documented.22 Aqueous solutions of sodium tetrachloroaurate are known to undergo rapid equilibrium with water generating hydroxytrichloroaurate and chloride.23 With its two upper-rim derived urea's 7 has the potential to generate multiple hydrogen-bonds and bind chloride (10) near the alkyne and dansyl gold binding motifs. Unfortunately, quantifying this using NMR and mass spectrometry proved inconclusive. Likewise generating crystals of 10 suitable for X-ray diffraction was unsuccessful. Our thoughts focused instead on the possibility that 7 when reacted with tetraethylammonium tetrafluoroborate affords a hydrogen bonded BF4− complex (cf.11) with the tetraethylammonium cation in close proximity to the tetrafluoroborate anion, a consequence of electrostatic attraction.
Fig. 6 Examples of motifs generated to probe the role of the calix[4]arene (8) and the two urea's (10 and 11). |
We anticipated the formation of 11 would reduce gold π-alkyne co-ordination and, consequently, diminish FQ. Mindful of reports outlining N-tetraalkylammonium cations residing in, largely unsubstituted upper-rim derived calix[4]arene cavities,24 we considered it unlikely the bulky tetraethylammonium would occupy the calix[4]arene cavity of 7; a fact further compounded by its sterically encumbered upper-rim substituents. Stirring 7 with 5 equivalents of tetraethylammonium tetrafluoroborate afforded, presumably, the urea N–H bonded BF4− with its Et4N+ counterion (11, Fig. 6). Although attempts to substantiate the formation of 11via X-ray diffraction were unsuccessful when the gold salt was added sequentially (>30 equivalents), a relatively small drop in the λmax was observed (ESI†). This indicates the metal is largely blocked from interacting with the triple bond and dansyl group of 11. Accordingly, we surmise from 8 and 11 that two urea groups are key parts of the architecture of 7. These preliminary results establish the importance of the alkyne, with this in mind we wanted to probe this further exploring the stereoelectronic effect associated with reducing or increasing alkyne electron-density and, consequently, its ability to interact with the Au3+. Our assumption focused on the premise that reducing alkyne electron density (12) would reduce its interaction with the metal and, as such, afford reduced fluorescent quenching. Of course, on this basis, incorporating electron-rich13 the opposite would be expected; that is, increased fluorescent quenching. The synthesis of para-nitrophenyl-derived 12 and para-methoxyphenyl 13 followed the same general procedure outlined in Scheme 2 in 56% and 68% yields respectively.
Subjecting separate solutions of metal-free 12 and 13 to fluorescent analysis (2.5 μM, λex 288 nm, λmax 520.7 nm) established, as anticipated, the λem of 12 to have reduced fluorescence intensity (Fig. 8) whilst electron-rich 13 exhibited increased intensity (Fig. 9). Sequential addition of 1, 2, 3, 4 and 5 equivalents of aqueous sodium tetrachloroaurate afforded, under otherwise identical conditions, a greater reduction in the fluorescence intensity of electron-rich13. Accordingly, we surmise increasing electron-density on the alkyne resulted, as anticipated, in augmented Au3+ sensitivity.
Fig. 7 Dansyl- and urea-derived cone-calix[4]arenes synthesised and appended with an electron-poor (12), electron-rich (13) or an anthracene-derived alkyne (14). |
Fig. 8 Effect on the metal-mediated fluorescent quenching properties of 12 when a conjugated electron-withdrawing para-nitrophenyl group is appended. |
Fig. 9 Effect on the metal-mediated fluorescent quenching properties of 13 when a conjugated electron-donating para-methoxyphenyl group is appended. |
Preliminary work established 7 did not detect Au+ (Fig. 2). However, it is known that alkyne activation using Au+ or Au3+ co-ordination (π → σ*) is facilitated if the electrophilic, cationic intermediate is stabilised via extensive delocalisation.25 Having synthesised extensively conjugated 9-(4-ethynylphenyl)anthracene-derived 14 (Fig. 7) from 4 in 2 steps (58% overall yield) we wanted to probe its ability to display enhanced sensitivity to Au3+ in the presence of Au+. Mixing solutions of 14 (2.5 μM) and 5 equivalents of chloro(triphenylphosphine)gold(I), the emission spectrum was acquired over 3 hours; little change was observed viz. Fig. 10. To the same solution containing the Au+ was added 1 equivalent of sodium tetrachloroaurate (cf.7). An almost immediate reduction in the λem and rapid, fluorescent quenching was observed (Fig. 10).
(i) The ability of a calix[4]arene bearing an N-sulphonylcarboxamide e.g., 7, 12–14 to interact with Au3+.
(ii) The propensity of the alkyne to act as an aurophile inducing electron-charge transfer from the alkyne to the metal which stimulates alkyne polarization and subsequent reaction with a nucleophile.
Interestingly, although alkyne coordinated linear Au+ complexes have been known since the 1970′s, alkyne π-systems comprising square-planar Au3+ have, only recently, been verified.26 Nevertheless, invoking the findings of Belanzoni et al.27 we propose Au3+-alkyne coordination proceeds via a classically described dominant σ-component coupled with smaller, but significant, π-back-donation (Dewar–Chatt–Duncanson model). Resulting in significant polarization of the π-alkyne electrons the Au3+ induces strong triple bond polarization and increased alkyne reactivity; this facilitates nucleophilic attack by, for example, a protic solvent ( 15 → 16, Scheme 3).28 Supporting the proposed formation of (E)-vinylgold(III) species 16 we note, with interest, the identification via X-ray diffraction of 19, see crystal structure 20 (Scheme 3). Oxyauration of N-(propargyl)benzamide 18 using Au3+ activates the C–C triple bond (Scheme 3) which initiates a 5-exo-trig cyclisation and the formation of 19 with the Au3+ in a square-planar complex.29 As noted earlier, support for ketone formation (17, Scheme 3) was from a 1991 report by Ultimoto and Fukuda who described how a wide variety of structure and function diverse alkynes readily react in the presence of 2 mol% sodium tetrachloroaurate and aqueous methanol affording ketones in excellent yield. Importantly, and with reference to our observed lack of FQ with chloro(triphenylphosphine)gold(I), when Ultimoto incorporated gold(I) potassium dicyanoaurate [KAu(CN)2] the reaction completely failed to mediate any alkyne hydration.10 As the dansyl group on 7 is slightly acidic (pKa 3.8 and 9.9) partial sulphonamide dissociation in methanol would afford a dansyl anion which we propose has the potential to undergo complex formation with the closely located, upper-rim bound Au3+. Thus, overall and consistent with known mechanisms of transition-metal ion-induced fluorescent quenching the ligand-immobilized excited dansyl fluorophore is quenched via a photoinduced electron transfer (PET) mechanism from the excited dansyl group which is in close proximity to the π-[LAu]-bound complex (15 → 16, Scheme 3).
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3qo00021d |
This journal is © the Partner Organisations 2023 |