Oxidatively Locked [Co2L3]6+ Cylinders Derived from Bis(bidentate) 2-Pyridyl-1,2,3-triazole “Click” Ligands: Synthesis, Stability, and Antimicrobial Studies

A small family of [Co2(Lpytrz)3]6+ cylinders was synthesised from bis(bidentate) 2-pyridyl-1,2,3-triazole “click” ligands (Lpytrz) through an “assembly-followed-by-oxidation” method. The cylinders were characterised using 1H, 13C, and DOSY NMR, IR, and UV-Vis spectroscopies, along with electrospray ionisation mass spectrometry (ESMS). Stability studies were conducted in dimethyl sulfoxide (DMSO) and D2O. In contrast to similar, previously studied, [Fe2(Lpytrz)3]4+ helicates the more kinetically inert [Co2(Lpytrz)3]6+ systems proved stable (over a period of days) when exposed to DMSO and were even more stable in D2O. The triply stranded [Co2(Lpytrz)3]6+ systems and the corresponding “free” ligands were tested for antimicrobial activity in vitro against both Gram-positive (Staphylococcus aureus) and Gram-negative (Escherichia coli) microorganisms. Agar-based disk diffusion and Mueller–Hinton broth micro-dilution assays showed that the [Co2(Lpytrz)3]6+ cylinders were not active against either strain of bacteria. It is presumed that a high charge of the [Co2(Lpytrz)3]6+ cylinders is preventing them from crossing the bacterial cell membranes, rendering the compounds biologically inactive.

Very recently, Lusby and co-workers reported the synthesis of some kinetically "locked" cobalt(III) tetrahedral cages, which were generated through an "assembly-followed-by-oxidation" process [35,36]. Taking advantage of the different labilities of Co(II) and Co(III), the tetrahedral cages were assembled under thermodynamic control and then "locked" upon oxidation of the Co(II) ions to Co(III). We reasoned that this approach may enable the generation of a family of biologically robust and active [Co 2 (L pytrz ) 3 ] 6+ cylinders. Cobalt is a physiologically active metal that has key functions in many biological processes [37]. Because of this, more and more cobalt complexes are being explored for an extensive range of biological applications such as anti-inflammatory [38], antiviral [39], anticancer [40], and antimicrobial [41,42] agents amongst others [43]. Consequently, development of cobalt complexes could significantly enhance the arsenal of pharmaceutical drugs available to combat a wide variety of harmful diseases and reduce antibiotic resistance.
Herein, we report the synthesis and characterisation of a small series of [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders assembled using bis(2-pyridyl-1,2,3-triazole) ligands. Exploiting the "assembly-followedby-oxidation" process, labile Co(II) ions were used to form helicates, and oxidation to Co(III) yielded the kinetically inert complexes. Additionally, we examined the stability of the [Co 2 (L pytrz ) 3 ] 6+ cylinders in a range of solvents and in the presence of biologically relevant nucleophiles along with their antimicrobial activity against S. aureus and E. coli.

Stability Studies
As previous work had shown that the [Fe 2 (L pytrz ) 3 ](BF 4 ) 4 cylinders were unstable in the presence of DMSO [32], we examined the stability of the new [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders under a variety of conditions before carrying out antimicrobial testing. To allow a direct comparison to the iron(II) systems that were examined previously, we initially studied the stability of the [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders in d 6 -DMSO in the light. The cylinders were dissolved in d 6 -DMSO and 1 H-NMR spectra were obtained over a period of days (Supplementary Materials, Section 3). Like the related iron(II) cylinders, the cobalt(III) systems decompose in d 6 -DMSO. However, due to the more inert nature of d 6 cobalt(III) ions relative to iron(II), the cylinders are observed to decompose over a period of days rather than instantaneously. Interestingly, the family of cobalt cylinders displayed different rates of decomposition. After two days, the [Co 2 L1 3 ] 6+ complex is completely decomposed into "free" L1 and presumably [Co(d 6 -DMSO) 6 ] 3+ . The [Co 2 L2 3 ] 6+ complex is more long lived and complete conversion to free ligand is only observed after four days. The larger [Co 2 L3 3 ] 6+ and [Co 2 L4 3 ] 6+ are more stable again. The [Co 2 L3 3 ] 6+ completely decomposed into "free" L3 after six days and the [Co 2 L4 3 ] 6+ complex is still the major species (>70%) in solution after the same time period. The differences in the rates of decomposition can be attributed to a number of factors. The hexyloxy substituted ligands (L2 and L4) are more electron rich relative to L1 and L3 leading to more stable complexes. The greater stability of the larger [Co 2 L3 3 ] 6+ and [Co 2 L4 3 ] 6+ complexes is presumably due to the lower charge repulsion between the more distant Co(III) cations and the lessening of the lone pair-lone pair repulsion that is present in the smaller [Co 2 L1 3 ] 6+ and [Co 2 L2 3 ] 6+ compounds (Supplementary Materials, Figure S48).
Intriguingly, repeating the d 6 -DMSO stability experiments in the absence of light resulted in a different outcome. The cobalt(III) cylinders are all much more stable in d 6 -DMSO when light is excluded from the reaction mixture. After six days in the dark, only very small peaks consistent with the "free" ligands can be observed in each of the examined systems (Figure 2a and Supplementary Materials, Section 3). The major species present in solution are the intact [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders. This suggests that the decomposition of the complexes in d 6 -DMSO is a light activated process, possibly due to in situ photoreduction of the cobalt(III) ions to the more labile cobalt(II) ions.
As the [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders were soluble in water, we have undertaken similar stability studies in D 2 O in the presence of light (Figure 2b and Supplementary Materials, Section 3). The stability of the complexes in D 2 O was significantly better than that observed in d 6 -DMSO in the presence of light. The cobalt(III) complexes showed no signs of decomposition over a period of a week in D 2 O, and no new signals are observed in any of the 1 H-NMR spectra. Furthermore, no signals due to "free" ligands are observed, and no precipitate is formed during the monitored time period. In some cases, the acidic triazole and methylene protons of the complexes slowly disappear due to hydrogen-deuterium exchange, but the remaining proton resonances do not shift, confirming that the cylinders remain intact in D 2 O. Additional stability tests, in D2O, were also conducted with the [Co2L23](OTf)6 cylinder against the common biological nucleophiles histidine and chloride ions.
[Co2L23](OTf)6 (1 equiv) and tetrabutylammonium chloride (9 equiv) were dissolved in D2O, and the mixture was analysed via 1 H-NMR spectroscopy (Figure 3). The complex proved stable over a day; as seen below, the acidic triazole (Hd) slowly disappears due to hydrogen-deuterium exchange and the slight upfield shifts of all signals due to the increased concentration of negatively charged ions in solution. However, the characteristic AB quartet remains unchanged, confirming the stability of the cylinder under these conditions. 1 H-NMR spectroscopy (D2O) on a mixture of [Co2L23](OTf)6 (1 equiv) and histidine (6 equiv) provided similar results. While hydrogen-deuterium exchange was observed for the protons He and Hf of the complexes, the pyridyl resonances (Ha-d) were unchanged over two days, and no sign of ligand precipitation from the solution was observed, indicating that the cylinder remained intact (Supplementary Materials, Section 3, Figure S47).

Antimicrobial Activity
The resistance of bacteria towards antibiotics is fast becoming a global issue. With pathogens developing a resistance to beta-lactams and other commonly used antibiotics, the need for new antimicrobial medicines is growing. Metal-containing antibacterial drugs have become a global research focus over the past decade and are showing increasing promise [53][54][55]. As the previously Additional stability tests, in D 2 O, were also conducted with the [Co 2 L2 3 ](OTf) 6 cylinder against the common biological nucleophiles histidine and chloride ions. [Co 2 L2 3 ](OTf) 6 (1 equiv) and tetrabutylammonium chloride (9 equiv) were dissolved in D 2 O, and the mixture was analysed via 1 H-NMR spectroscopy (Figure 3). The complex proved stable over a day; as seen below, the acidic triazole (H d ) slowly disappears due to hydrogen-deuterium exchange and the slight upfield shifts of all signals due to the increased concentration of negatively charged ions in solution. However, the characteristic AB quartet remains unchanged, confirming the stability of the cylinder under these conditions. 1 H-NMR spectroscopy (D 2 O) on a mixture of [Co 2 L2 3 ](OTf) 6 (1 equiv) and histidine (6 equiv) provided similar results. While hydrogen-deuterium exchange was observed for the protons H e and H f of the complexes, the pyridyl resonances (H a-d ) were unchanged over two days, and no sign of ligand precipitation from the solution was observed, indicating that the cylinder remained intact (Supplementary Materials, Section 3, Figure S47). Additional stability tests, in D2O, were also conducted with the [Co2L23](OTf)6 cylinder against the common biological nucleophiles histidine and chloride ions.
[Co2L23](OTf)6 (1 equiv) and tetrabutylammonium chloride (9 equiv) were dissolved in D2O, and the mixture was analysed via 1 H-NMR spectroscopy (Figure 3). The complex proved stable over a day; as seen below, the acidic triazole (Hd) slowly disappears due to hydrogen-deuterium exchange and the slight upfield shifts of all signals due to the increased concentration of negatively charged ions in solution. However, the characteristic AB quartet remains unchanged, confirming the stability of the cylinder under these conditions. 1 H-NMR spectroscopy (D2O) on a mixture of [Co2L23](OTf)6 (1 equiv) and histidine (6 equiv) provided similar results. While hydrogen-deuterium exchange was observed for the protons He and Hf of the complexes, the pyridyl resonances (Ha-d) were unchanged over two days, and no sign of ligand precipitation from the solution was observed, indicating that the cylinder remained intact (Supplementary Materials, Section 3, Figure S47).

Antimicrobial Activity
The resistance of bacteria towards antibiotics is fast becoming a global issue. With pathogens developing a resistance to beta-lactams and other commonly used antibiotics, the need for new antimicrobial medicines is growing. Metal-containing antibacterial drugs have become a global research focus over the past decade and are showing increasing promise [53][54][55]. As the previously

Antimicrobial Activity
The resistance of bacteria towards antibiotics is fast becoming a global issue. With pathogens developing a resistance to beta-lactams and other commonly used antibiotics, the need for new antimicrobial medicines is growing. Metal-containing antibacterial drugs have become a global research focus over the past decade and are showing increasing promise [53][54][55]. As the previously investigated, biologically robust, ruthenium(II) triply stranded helicate [Ru 2 (L3) 3 ] 4+ and related dirhenium(I) L pytrz complexes had displayed modest antimicrobial activity [33,56], we screened the new [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders against S. aureus (Gram-positive bacteria) and E. coli (Gram-negative bacteria). The activity of the ligands (L1-L4) and the cobalt(III) complexes was initially examined using Kirby-Bauer disk diffusion assays, but none of the ligands or cylinders displayed any zones of inhibition, indicating that they were not biologically active. As this negative result could potentially be due to the solubility properties of the compounds preventing the diffusion of materials from the disk into the agar, we further investigated the antimicrobial properties using Mueller-Hinton broth micro-dilution assays. Unfortunately, these experiments also indicated that the cobalt(III) cylinders are not active antimicrobial agents, and no inhibition of bacterial growth was observed over a range of concentrations, from 1024 µg·mL −1 to 1 µg·mL −1 . Given the stability of the [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders in aqueous and pseudo-biological conditions and the activity of other triply stranded helicates [13,18], related dinuclear ruthenium(II) compounds [53,57], and cobalt(III) complexes [41,42], the observed lack of activity was initially surprising. However, a series of elegant papers from Keene and co-workers have shown that the charge of the metal complex is an important factor for obtaining active antimicrobial complexes. They have developed a family of highly active dinuclear ruthenium(II) complexes that are overall 4+ cations [53,57]. Subsequently, they generated the analogous dinuclear iridium(III) complexes that are overall 6+ cations and found that these compounds displayed no antimicrobial activity [58]. The have also investigated the corresponding monometallic systems [Ir(Me 4 phen) 3 ] 3+ and [Ru(Me 4 phen) 3 ] 3+ (where Me 4 phen = 3,4,7,8-tetramethyl-1,10-phenanthroline) and shown while the 2+ ruthenium system is highly active the 3+ iridium is not [59]. Based on these findings, Keene and co-workers have suggested that an individual metal centre with a 3+ charge cannot readily pass across the bacterial cell membrane, and our observed results are consistent with that postulate.

Conclusions
A small family of the [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders were synthesised, in good yields, from bis(bidentate) 2-pyridyl-1,2,3-triazole "click" ligands through an "assembly-followed-by-oxidation" method. The cylinders were characterised using 1 H, 13 C, and DOSY NMR, IR, and UV-Vis spectroscopies, along with electrospray ionisation mass spectrometry (ESMS). Stability studies were conducted in DMSO and D 2 O in the presence and absence of light. In contrast to similar, previously studied, [Fe 2 (L pytrz ) 3 ] 4+ helicates, the more kinetically inert [Co 2 (L pytrz ) 3 ] 6+ systems proved stable (over a period of days) when exposed to dimethyl sulfoxide (DMSO). The cobalt(III) cylinders displayed impressive stability in D 2 O and retained their structure, even in the presence of common biological nucleophiles. In vitro agar-based disk diffusion and Mueller-Hinton broth assays with both S. aureus and E. coli showed that the [Co 2 L 3 ] 6+ cylinders were not biologically active against either strain of bacteria. It is presumed that the lack of biological activity is connected to the high charge of the [Co 2 L 3 ] 6+ cylinders, which is preventing them from crossing the bacterial cell membranes. These results suggest that, while the "assembly-followed-by-oxidation" approach would allow the generation of a large family of [Co 2 (L pytrz ) 3 ](OTf) 6 cylinders that are kinetically robust under biological conditions, the high 6+ charge of the complexes will prevent their use as antimicrobial agents.

General
Unless otherwise stated, all reagents were purchased from commercial sources and used without further purification. The ligands L1 [26], L2 [26] and L3 [44] were synthesised using our previously reported procedures. Solvents were laboratory reagent grade. Petroleum ether refers to the fraction of petrol boiling in the range 40-60 • C. Substances (and abbreviations) used in this study include isopropyl alcohol (IPA), methanol (CH 3 OH), dichloromethane (CH 2 Cl 2 ), ethylenediaminetetraacetate (EDTA), tetrahydrofuran (THF), dimethyl sulfoxide (DMSO), and dimethylformamide (DMF). 1 H-, 13  , and Na 2 CO 3 (154 mg, 1.452 mmol), the reaction mixture was stirred at room temperature for 16 hours. It was then added to a 0.1 M EDTA/NH 4 OH aqueous solution (100 mL) and to CH 2 Cl 2 (100 mL) and stirred vigorously for one hour. The organic layer was then washed with water (5 × 100 mL) and brine (100 mL), dried over Na 2 SO 4 , and filtered, and the solvent was removed under vacuum. Column chromatography (silica gel) (CH 2 Cl 2 then 4:1 CH 2 Cl 2 /acetone) afforded the product as a fluffy white solid (155 mg, 0.260 mmol, 72%  (1 mL). An addition of a solution of (NH 4 ) 2 [Ce(NO 3 ) 6 ] (75 mg, 137 µmol) in acetonitrile (0.440 mL) to this solution resulted in the precipitation of a fluffy yellow solid. After centrifugation (13,000 RPM, 5 min), the filtrate was discarded, and the pellet was suspended in acetonitrile and collected by filtration. The solid obtained was dissolved in a minimal amount of water before an addition of a saturated aqueous solution of [NH 4 ]PF 6 (10 mL). The resulting precipitate was collected by filtration, washed with water (2 × 5 mL), dissolved in acetonitrile (5 mL), and filtered through cotton wool. Vapor diffusion of diethyl ether into this solution yielded the product as yellow crystals ( 6 : A solution of [Co(OH 2 ) 6 ](BF 4 ) 2 (2 equiv) in CH 3 CN (2 mL) was added to a suspension of L (3 equiv) in CH 3 CN (2 mL). To this, a solution of (NH 4 ) 2 [Ce(NO 3 ) 6 ] (2.2 equiv) in CH 3 CN (2 mL) was added to form a yellow precipitate. The solid was collected via centrifugation (5 min at 13,000 rpm) and dissolved in DMSO (0.75 mL). [NBu 4 ](OTf) (30 equiv) was added as a solid to the orange solution. The product was collected by precipitation out of DMSO with ethyl acetate (3 mL) and diethyl ether (5 mL