Yen-An
Young
,
Huong T. H.
Nguyen
,
Hieu D.
Nguyen
,
Tuhin
Ganguly
,
Yennie H.
Nguyen
and
Loi H.
Do
*
Department of Chemistry, University of Houston, 4800 Calhoun Road, Houston, TX 77204, USA. E-mail: loido@uh.edu
First published on 10th May 2024
A cyclometalated iridium(III) complex bearing a self-immolative quinolinium moiety was developed as a ratiometric substrate for transfer hydrogenation studies. This photoluminescent probe allowed the rapid screening of a variety of Ir catalysts using a microplate reader, offering a convenient method to assess activity using a minimum amount of catalyst sample.
Scheme 1 Various applications of transfer hydrogenation (TH) catalysts (A) and off–on (B) and ratiometric (C) substrates used to study their activity. |
Herein, we introduce a user-friendly photoluminescence-based method to screen for TH catalyst activity using a microplate reader. Our strategy relies on the use of a ratiometric substrate that allows in situ quantification of the reaction progress.18,19 This approach enables simultaneous monitoring of multiple reactions using minimum amounts of sample, which is particularly attractive for studies of catalysts that are costly (e.g., those containing Ir,20 Ru,21 or Os22) or difficult to obtain on large scales.
The design of our ratiometric probe was inspired by a self-immolative umbelliferone-quinolinium substrate (UmQN) developed by Ward and coworkers (Scheme 1B).11,12 It was shown that when this compound was reduced via TH catalysis, release of umbelliferone (Um) led to an emission turn-on. Although this off–on substrate was used to evaluate TH reactions in the periplasm of E. coli, the total amount of starting substrates inside the cells could not be quantified by fluorescence microscopy because they are non-emissive. To further complicate matters, small molecules such as substrates and molecular catalysts can diffuse in and out of cells so obtaining accurate yields in living systems is highly challenging. To overcome these limitations, we sought to develop a fluorophore-quinolinium construct that would exhibit different emission properties in the starting and product forms so that the reaction yields could be determined by monitoring the emission intensity ratio at two different wavelengths. Such ratiometric substrates may allow non-invasive studies of intracellular TH reactions in real time without having to lyse the cells and analyse their contents by methods such as high performance liquid chromatography. After considering a variety of fluorophore candidates, we selected to use a cyclometalated iridium(III) complex as the reporter due to its tunable emission, chemical stability, and ease of preparation (Scheme 1C).23–28 In addition, its phosphorescence may be more easily distinguished from biological self-fluorescence.29
The synthesis of our ratiometric iridium complex (RIC575, where the subscript denotes its emission maximum) proceeded starting from 2,6-dimethylquinoline (Scheme S1†). Monobromination using N-bromosuccinimide (NBS) afforded compound 1 (68% yield), which was then subjected to nucleophilic attack by 2-bromo-4-hydroxypyridine to furnish 2 in 73% yield. Combining 2 in the presence of 2-(tributylstannyl)pyridine under Stille cross-coupling conditions gave the bipyridine species 3 (62% yield). Metalation of 3 was accomplished by treatment with [Ir(ppy)2Cl]2 (ppy = 2-phenylpyridine), anion exchange using AgOTf (OTf− = triflate anion), and then alkylation using methyl triflate to afford RIC575 in 38% yield (based on 3, Scheme 2A).
The cleaved cyclometalated iridium species RIC525 (where the subscript denotes its emission maximum) was also prepared (Scheme S2†). The starting 2-bromo-4-hydroxypyridine was protected using 2-methoxyethoxymethyl (MEM) chloride to give 5, followed by Stille cross-coupling with 2-(tributylstannyl)pyridine to furnish bipyridine 6. Combining 6 with [Ir(ppy)2Cl]2 and then treatment with HCl gave the corresponding anionic iridium complex (RIC525′). To obtain the neutral form, it was stirred in the presence of NaOH, followed by the addition of AgOTf to afford RIC525 with a yield of 79% based on the amount of [Ir(ppy)2Cl]2 used in the previous step.
To obtain structural characterization, X-ray crystallography was used to analyse single crystals of RIC575 and RIC525. The structure of RIC575 showed the expected octahedral iridium center with a quinolinium moiety attached to the bipyridine ligand (Fig. S62†). The presence of two triflate anions supports the dicationic nature of the complex. The structure of RIC525 also reveals an octahedral Ir geometry but the absence of any counterions in the crystallographic unit cell indicates that this species is charge neutral overall (Fig. S63†). Comparison of the bond metrics in RIC575vs.RIC525 suggests that the former contains a 4-alkoxypyridyl donor whereas the latter contains a 4-pyridonate donor coordinated to iridium (Fig. 1A).30–32 For example, in RIC575, the C–C and C–N bond distances of ∼1.33–1.39 Å in the functionalised pyridine ring are typical of an aromatic structure (Fig. 1A).33 Additionally, its C(3)–O(1) bond length of 1.35 Å is expected for a carbon–oxygen single bond (average = ∼1.37 Å). In contrast, the corresponding six-membered ring in RIC525 exhibits contracted or elongated C–C/C–N bonds relative to those in RIC575, suggesting that it is non-aromatic. The only exception is that the C(4)–C(5) bond distance in RIC525 is similar to that in RIC575 (∼1.37 Å), which may be due to the conjugation of its C4 and C5 p-orbitals with the π-system of the adjacent pyridine ring. Lastly, the C(3)–O(1) bond distance of 1.28 Å in RIC525 is consistent with a CO carbonyl group.33
Due to the electronic differences in their ancillary ligands, RIC575 and RIC525 were expected to display different photophysical properties.23 As shown in Fig. 1B, when RIC575 was excited with 350 nm light in MeOH under N2, a broad emission peak was observed with a maximum at 575 nm and shoulders at 490 and 520 nm. Our data indicate that the non-radiative decay rate (knr) was about 160-fold faster than the radiative decay rate (kr), presumably due to the emission quenching effects of its quinolinium moiety.34 For comparison, the iridium precursor 4 bearing an unmethylated quinoline ring exhibited a knr that is only 1.9-fold faster than kr (Table S4†). When RIC525 was irradiated with 350 nm light, its emission spectrum showed a peak maximum at 525 nm. This hypsochromic shift, relative to the λem of RIC575, is consistent with the more electron-rich nature of the RIC525 bipyridine ligand, which destabilizes the LUMO energy and consequently, increases the Ir complex's HOMO–LUMO gap. Our results also indicate that RIC525 has a higher quantum yield (ΦPL = 0.438 vs. 0.006) and longer luminescence lifetime (τ = 1.26 vs. 1.03 μs) than that of RIC575 (Fig. 1B). A chromaticity analysis shows that the photoluminescence of RIC575 appears yellow whereas that of RIC525 appears green (Fig. S4†). Because RIC575 and RIC525 have poor solubility in water, we were unable to perform photophysical measurements in aqueous solutions such as biological buffers or cell culture media.
Prior to conducting reaction studies with RIC575, we created a calibration curve using standard solutions containing different ratios of RIC575/RIC525 (Table S6†). Sodium bicarbonate was also added to the mixtures to ensure that RIC525 remains in the neutral form. We found that measuring the emission intensities at 550 (I550) and 600 (I600) nm and plotting the I550/I600 ratio gave a linear correlation with the amount of RIC525 present (Fig. S6†).19 Due to the greater than 70-fold difference in quantum yield between RIC525 and RIC575, the choice of 550 and 600 nm for reaction monitoring (rather than at their emission maxima of 525 and 575 nm, respectively) ensures that signals from the product do not overwhelm signals from the starting material for accurate quantification.
Next, we proceeded to study the TH activity of a series of previously reported [Cp*Ir(4-R-picolinamidate)Cl] complexes (where Cp* = pentamethylcyclopentadienyl anion, R = various functional groups; Scheme 2C).11,35–37 To establish that RIC575 is a viable substrate for TH,11,12 a small-scale reaction was performed using a 1-dram vial (Fig. S61A†). When Ir4 (R = NMe2), RIC575, HCOONa, and NaHCO3 were dissolved in 1 mL of MeOH and stirred at RT, the reaction mixture showed increasing luminescence over the course of 1 h with a change in color from yellow to green (Fig. S61B†), suggesting that RIC525 had formed. The presence of the RIC525 product was further confirmed by emission spectroscopy and mass spectrometry. Reduction of the quinolinium moiety was supported by showing that a model substrate, 1,2,6-trimethylquinolinium triflate, was converted to 1,2,3,4-tetrahydro-1,2,6-trimethylquinoline in the presence of Ir4, HCOONa, and NaHCO3 (Fig. S59†). These results indicate that the conversion of RIC575 to RIC525 by the Ir catalysts and HCOONa occurs on a timescale that is suitable for quantification using a microplate reader.
Once we identified the optimal reaction conditions for TH, we carried out our catalyst screening experiments in 24-well microplates (Fig. 2A). In these investigations, each well was charged with an Ir catalyst (2.0 μM), HCOONa (0.7 mM), and NaHCO3 (1.25 mM) in MeOH at 28 °C (Fig. 2A). The mixtures were injected with RIC575 (0.2 mM) and then the microplate was inserted into a microplate reader for continuous emission monitoring at 550 and 600 nm. By converting the I550/I600 ratio to the %yield of RIC525 (see Table S7† for example with Ir1), kinetic plots of the reactions were generated (Fig. 2B, left). The kinetic data showed that in all cases, an induction period of ∼10 min was observed, which may be due to catalyst activation (e.g., formation of the initial iridium-hydride species).36 Based on the RIC525 yields after 1 h, the catalyst activity trend follows the order: Ir4 (R = NMe2) > Ir3 (R = pyrrolidinyl) > Ir2 (R = OMe) > Ir1 (R = H) > Ir5 (R = CF3) > Ir6 (R = CN). The poor solubility of Ir7, which bears a nitro group on the pyridyl ring, precluded studies using this complex. The trend observed that electron-rich Ir catalysts are more active than their electron-poor counterparts is fully consistent with previous studies performed in aqueous media.35,36 It has been demonstrated that hydride formation is typically rate-limiting in TH reactions by [Cp*Ir(4-R-picolinamidate)Cl] catalysts and that this step occurs faster with electron-rich than electron-poor catalysts.36 Although a more in-depth analysis of the kinetic data was not pursued in this work, these results validate our ratiometric substrate-based approach for rapid screening of TH catalysts. Because each reaction in our studies requires only 20 nmol (<∼90 μg) of Ir catalyst, a significant amount of material could be saved compared to performing larger scale reactions in standard vials or flasks.
To illustrate the utility of our method, we used our protocol to evaluate the performance of several new TH catalysts. When developing a catalyst for in cell or in vivo applications, it is often desirable to customize its biological properties, such as cellular uptake, interactions with biomolecules, subcellular localization, etc., without diminishing its catalytic activity.13 Because Ir4 was the most active in our series (Fig. 2B, left), we chose to derivatize this complex to determine whether the catalyst structure and activity could be tuned independently of each other. By adapting previously reported procedures (see ESI†), we successfully prepared three iridium variants featuring tert-butoxyphenyl (Ir8), triethylene glycol phenyl (Ir9), and 1,3-benzodioxole (Ir10) groups on the picolinamidate ligand (Scheme 2C). These substituents were chosen to represent groups with varying shapes and sizes. The new Ir catalysts were tested in TH reactions by combining them with HCOONa, NaHCO3, RIC575, and MeOH in 24-well microplates. Analysis of the photoluminescence data acquired by the microplate reader produced the kinetic traces shown in Fig. 2B (right). The results revealed that Ir8–Ir10 exhibit nearly identical kinetic profiles as that of their parent catalyst Ir4. This finding corroborates our previous observations that for [Cp*Ir(4-R-picolinamidate)Cl] complexes ligand modifications at the N-phenyl moiety do not impact their catalyst activity.36
In summary, we have developed a self-immolative substrate for photoluminescence monitoring of TH catalysts using a microplate reader. Because the emission maximum of the cleaved product RIC525 is blue-shifted relative to that of the starting RIC575, ratiometric quantification of the reaction yields was possible by measuring the intensity changes at two different wavelengths. Application of our method to studies of several new [Cp*Ir(4-R-picolinamidate)Cl] variants revealed that functionalization of the N-phenyl group does not negatively impact their catalytic activity. Our photoluminescence-based method is user-friendly because it is rapid, requires minimal catalyst sample, and can be readily used by researchers with access to microplate readers, which are available in many synthetic and biochemistry laboratories.
There are numerous non-aqueous applications that may benefit from this new molecular tool, such as in high throughput automation for chemical synthesis or metallodrug discovery.38,39 To enable TH reaction imaging in biological cells,40 the RIC575 probe could be rendered more water soluble by functionalizing it with sulfonate groups; however, how this change may affect its cell permeability remains to be investigated.41 Since different cellular compartments have different pH values, modifying the probe so that its emission properties are not sensitive to pH changes could broaden its biological range. Although in-cell applications were not achieved in this work, it provides proof-of-concept demonstration that cyclometalated Ir-quinolinium substrates are useful for ratiometric quantification of TH reactions in solution.
Footnote |
† Electronic supplementary information (ESI) available: Experimental details, characterization data, photophysical studies, and reaction data. CCDC 2336124 and 2336125. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4dt00891j |
This journal is © The Royal Society of Chemistry 2024 |