Supramolecular optimization of the visual contrast for colorimetric indicator assays that release resorufin dye

Abstract

A tetralactam macrocycle acts as a novel supramolecular adjuvant to capture a released resorufin dye and create a higher contrasting yellow/blue color change for enhanced naked eye interpretation of a colorimetric indicator assay.

Graphical Abstract

A tetralactam macrocycle captures released resorufin and creates a pair of complementary indicator assay colors that are more easily perceived by the naked eye.

There is a growing demand for robust diagnostic technologies that allow point-of-care medical treatment or environmental monitoring to be conducted without electrical power. Technologies that meet the criteria of affordable, sensitive, specific, user-friendly, rapid and robust, equipment free and deliverable (ASSURED) are a high priority.^1,2 Many diagnostic assays have been designed to produce a change in a sample’s absorption and/or fluorescence spectrum.^3,4 While fluorescence assays are inherently more sensitive than colorimetric, they require a powered excitation source which is a practical limitation. In contrast, colorimetric indicator assays, especially those based on test strip technology, do not need external power if the results can be detected using the naked eye. The low sensitivity of a colorimetric assay can be circumvented if the chemical event that produces the color change is a catalytic reaction. For example, many enzyme assays utilize an enzyme substrate whose chemical structure releases a colored dye after enzyme action.³ Similarly, many chemical indicators are reactive molecules that release a colored dye when they are cleaved catalytically by a specific analyte.⁴ In these cases, the color change that is caused by the dye release is a detectable signal that is amplified by catalyst turnover.

The ideal colorimetric indicator assay is a signal transduction process that induces a colorless sample to become intensely colored and in this case the assay performance is governed by standard optical parameters such as molar absorptivity of the released dye, color stability, color perception of the human eye, and color homogeneity if the sample is immobilized on a solid phase.⁵ In practice, most colorimetric indicators transduce a change between two colors (two-color colorimetric assay). In this case, the assay development process should also aim to optimize the contrast of the “before” and “after” colors for easier perception by the eye.^‡ Inspection of the color wheel in Scheme 1 provides systematic guidance for contrast optimization in a two-color colorimetric assay. Colors that are opposite each other on the color wheel are complementary colors because they produce high visual contrast. Thus, one important experimental goal for optimizing two-color contrast for naked eye interpretation of a colorimetric indicator assay is to ensure that the “before” and “after” colors are complementary colors.⁶ In principle, one way to pursue this goal is to modify the substrate structure by chemical synthesis. However, this approach is time consuming because it requires an iterative cycle of modified substrate synthesis and testing to determine if there is an improvement in color contrast while maintaining appropriate reactivity. Here, we describe a novel alternative approach; that is, a supramolecular strategy that adds a host molecule to the assay solution to bind the released dye and enhance the visual contrast for naked eye detection.

Scheme 1.

Colors that are opposite on the color wheel are complementary and produce the highest visual contrast.

Many fluorescent dyes undergo a substantial change in brightness when captured by a supramolecular host, and various host molecules have been studied extensively as adjuvants to enhance the performance of fluorescent chemosensor or enzyme assays.⁷ In contrast, there are very few examples of colored dyes that strongly change absorption profile upon supramolecular capture.^7,8 The focus here is on colorimetric indicators that release the well-known phenoxazine dye, resorufin (RF) whose structure is shown in Scheme 2. There are several dozen publications describing a wide range of different enzyme substrates and reactive chemosensors that release RF (see ESI Table S1 for a summary).^3,4,9-14 These yellow-colored compounds are weakly fluorescent. The cleavage reaction is highly fluorogenic due to the appearance of pink-colored RF which has very bright emission properties, and the vast majority of published studies have focused on fluorescence assays. Some reports have suggested that the sample color change from pink to yellow has value as a colorimetric assay, but as indicated by the color wheel in Scheme 1, the change is hard to perceive with the naked eye. Unfortunately, this means that the large collection of known RF-releasing chemosensors and substrates is currently not very useful for colorimetric analysis. Thus, the general project goal was to find a suitable supramolecular host that could enhance the two-color contrast of a RF-releasing indicator assay.

Scheme 2.

Chemical structures.

The choice of a supramolecular host for released RF was based on our previous experience with tetralactam macrocycles containing aromatic sidewalls, which are versatile synthetic receptors for a wide range of guest molecules.¹⁵ They are especially effective as supramolecular hosts for squaraine dyes where dye complexation produces a noticeable red shift in dye absorption.^16,17 Recent work by Mateo-Alonso and co-workers has shown that tetralactam macrocycles with anthracene sidewalls can capture acene guests by a combination of guest hydrogen bonding with the internally directed macrocycle NH residues and aromatic stacking with anthracene sidewalls.¹⁸ With this knowledge in mind, we wondered if tetralactam macrocycles could be used to capture RF and enhance the effectiveness of RF-releasing substrates and chemosensors in colorimetric assays. Here, we report RF complexation studies using the organic and water soluble tetralactam hosts, M1 and M2, and our finding that RF encapsulation induces a remarkable red-shift effect that produces a highly desired blue colored complex. In addition, we show that supramolecular RF encapsulation can change the before/after colors of a two-color colorimetric assay from a yellow/pink pair to a higher contrasting yellow/blue pair which is easier to perceive by the naked eye.

Early experimental evidence that a tetralactam macrocycle can encapsulate RF was gained by conducting host/guest association studies in chloroform. Absorption and fluorescence experiments found that addition of organic-soluble macrocycle M1 to RF (tetrabutylammonium salt) induced a red-shift in RF absorption maxima band and quenching of RF fluorescence (Figure S5). ¹H NMR titration experiments in CDCl₃ produced spectra exhibiting exchange-induced line broadening, a common occurrence in host/chemistry (Figure S4),¹⁹ but there was still strong proof that the RF was encapsulated inside M1. The spectral comparison in Figure 1 highlights the complexation-induced changes in macrocycle chemical shift upon formation of M1⊃RF. The large down-field changes in chemical shift for macrocycle NH residues and internally directed protons B are strong indicators of hydrogen bonding with RF heteroatoms.

Figure 1.

Partial ¹H NMR spectra (500 MHz, CDCl₃, 25° C) of: (top) RF (0.5 mM), (center) M1⊃RF (0.5 mM), and (bottom) M1 (0.5 mM). For clarity, the tetrabutylammonium cation peaks are not shown

Shown in Figure 2 is a molecular model that is highly consistent with the spectral data and the previous literature on acene encapsulation within macrocycles like M1.^15,18 The modeling indicates that the macrocycle NH residues form bifurcated hydrogen bonds with the internal oxygen and nitrogen atoms on the anionic RF guest which is consistent with literature computations showing that these atoms are reasonable hydrogen bond acceptors.^20,21 RF encapsulation is also favored by the large amount of aromatic surface area that is engaged in favorable π-π stacking interactions.

Figure 2:

Two views of a molecular model of M1⊃RF (macrocycle lacks t-butyl groups) that was computed using the PM7 method.

RF complexation studies in water used the known tetralactam macrocycle M2, with two appended dendritic tricarboxylates as water solubilizing units.^22,23 Absorption and fluorescence experiments found that addition of water-soluble macrocycle M2 to RF created a blue colored complex with absorption maxima band at 624 nm (Figure 3a). In addition, there was substantial complexation-induced quenching of RF fluorescence which enabled fluorescence titration experiments. Shown in Figure 3b is a representative fluorescence titration curve that was fitted, using a standard non-linear computer algorithm, to a 1:1 binding model. The average of three independent experiments determined K_a to be (3.2 ± 0.6) × 10⁵ M⁻¹ which is more than 100 times higher than the RF affinity reported for β-cyclodextrin.²⁴ Control experiments that acquired the absorption profile of RF in different solvents showed that the absorption maxima band is weakly red-shifted by a change from water to aprotic organic solvent (Figures 3a and S3). But in no solvent is the absorption maxima wavelength higher than 600 nm. Similarly, RF absorption is still below 600 nm when it is encapsulated within micelles or β-cyclodextrin.^24,25,26 Thus, the remarkably large complexation-induced red-shift of RF absorption maxima wavelength to 624 nm is not simply due to a less polar microenvironment inside the cavity of M2. More likely, intermolecular π-π interactions between the encapsulated RF and the anthracene sidewalls of M2 induces a Davydov splitting effect that lowers the RF excited state energy,^27,28 and also promotes quenching of RF fluorescence by photoinduced electron transfer.²⁹

Figure 3.

(a) Partial absorption (a) and emission (ex: 540 nm) (b) spectra in water (25°C) of 10 μM RF (sodium salt) (red line) and 10 μM M2⊃RF (blue line), with the latter obtained by mixing 10 μM RF and 200 μM M2. Insert: representative titration isotherm showing decrease in fluorescence (ex: 540 nm, em: 585 nm) for RF (10 μM) with increasing equivalents of M2 in water, red line shows curve fitting to a 1:1 binding model.

To demonstrate potential utility as a colorimetric indicator, we investigated two diagnostic applications that are relevant to the ongoing coronavirus pandemic. The first study examined the effect of M2 on the colorimetric performance of RT-1 (Scheme 1), a known RF-releasing chemosensor for fluorescence detection of hypochlorous acid.³⁰ We reasoned that RT-1 could be repurposed as a colorimetric indicator of hypochlorite (denoted here as HOCl/OCl⁻ since the pKa is 7.5), a reactive chemical that is employed world-wide as a bleaching agent or as a common disinfectant. A manifestation of the COVID-19 pandemic is increased environmental exposure to water samples containing HOCl/OCl⁻. Thus, effective indicator tests are needed to rapidly determine if water samples contain HOCl/OCl⁻ which can cause health problems such as skin irritation or inflammation of the esophagus. In addition, there is a need to know if a procured disinfectant solution is sufficiently concentrated for useful operation. The literature includes several recent reports of chemosensor candidates for colorimetric HOCl/OCl⁻ detection.^31-37 Operationally, RT-1 is an attractive choice as an indicator of HOCl/OCl⁻ because it is chemically quite selective, the cleavage reaction is complete within a few minutes, and there is a linear relationship in the low micromolar range between concentration of HOCl/OCl⁻ and released RF.³⁰ As shown in Figure 4, addition of HOCl/OCl⁻ (5 μM) induces rapid chemical fragmentation of yellow RT-1 to produce a pink solution due to the released RF. Subsequent addition of macrocycle M2 to the sample instantly produces a blue color due to the formation of M2⊃RF. Thus, addition of M2 enhances the performance of RT-1 as a two-color indicator of HOCl/OCl⁻ by converting the yellow/pink color change to a yellow/blue color change that is more easily perceived by the naked eye. A second proof of concept study assessed the potential of M2 to enhance an enzyme assay that released RF. We chose NHPO as a known substrate for the enzyme N-Acetyl-β-D-glucosaminidase (NAG) which is a clinically relevant urinary biomarker of kidney dysfunction.³ A relatively high level of NAG in the urine is thought to reflect chronic kidney disease in disorders such as diabetes,³⁸ and may indicate acute kidney injury in patients with severe COVID-19.³⁹

Figure 4.

Absorbance spectra and photographs of a sample, initially containing RT-1 (50 μM, black line), and 3 minutes after addition of HOCl/OCl⁻ (5 μM, red line), or 3 minutes after a two-step addition sequence of HOCl/OCl⁻ (5 μM) and then M2 (500 μM) (blue line). In 200 mM PBS, pH 7.4 at 25°C.

Thus, there is a need for simple and robust urinalysis tests that allow a patient to identify early stages of kidney dysfunction or track kidney health during treatment. Recently, NHPO probe was shown to be an effective fluorescent substrate for NAG;³ thus, our goal was to ascertain if macrocycle M2 could enhance the performance of NHPO as a colorimetric indicator of NAG enzyme. As shown in Figure 5, addition of NAG leads to enzyme catalysed fragmentation of yellow NHPO to produce a pink solution of RF. Adding an aliquot of M2 to the assay solution leads to a deep-purple color due to complexation of the released RF by the M2. The assay mixture included BSA as an enzyme stabilizer which competes for the RF,⁴⁰ and slightly weakens the supramolecular color enhancement effect (there is no evidence for NAG or BSA interaction with M2, Figure S9). Nonetheless, the presence of M2 still converts the before/after colors of the colorimetric NAG assay using NHPO from yellow/pink to yellow/deep-purple, a color change that is more easily perceived by the naked eye.

Figure 5.

Absorbance spectra and photographs of a sample, initially containing NHPO (50 μM), and 30 minutes after addition of NAG enzyme (0.9 μg/mL), red line), or 45 minutes after a two-step addition of NAG (0.9 μg/mL) and M2 (500 μM) (blue line). In 100 mM PBS + 100 μM BSA pH 7.4 at 25°C.

In summary, the vast majority of known colorimetric chemical indicators and enzyme substrates are “two-color” assays. Thus, indicator development projects should aim to design the “before” and “after” colors to be complementary colors because they produce the highest visual contrast for naked eye perception (Scheme 1).⁶ RF-releasing chemosensors and enzyme substrates are ubiquitous but unfortunately the inherent yellow/pink color change is not very useful for colorimetric detection by the naked eye. We find that tetralactam M2 can be employed as a supramolecular adjuvant to capture the released RF and create a higher contrasting yellow/blue color change. This simple assay modification should be broadly useful because there are several dozens of different enzyme substrates and reactive chemosensors that release RF. From a broader perspective, this study provides proof of concept for a new and potentially generalizable supramolecular chemistry strategy for fine-tuning the visual contrast of two-color indicator assays for naked eye detection. The next step in the research is development of indicator test strips where we will build on our previous success using a surface-immobilized tetralactam host for dye capture.⁴¹ We are grateful for funding support from the NIH (R01GM059078, R35GM136212 and T32GM075762) and NSF (CHE1708240) and we acknowledge D. -H. Li for conducting the molecular modelling.