Conjugated Polymer-Based Fluorescence Turn-On Sensor for Nitric Oxide

Abstract

A turn-on fluorescent sensor for NO (g) in solution was synthesized using a bipyridyl-substituted poly(p-phenylene vinylene) derivative (CP1) as the sensory scaffold. The action of NO (g) upon the CP1–Cu(II) complex reduces it to the CP1–Cu(I) complex with a concomitant 2.8-fold increase in emission intensity. The reagent is selective for NO (g) versus other biological reactive nitrogen species, except for nitroxyl, and has a detection sensitivity limit of 6.3 nM.

The discovery that nitric oxide (NO) is the endothelium-derived relaxing factor astonished the scientific community.¹ Since then, an even wider range of biological roles for NO have been elucidated.² We are interested in NO as a possible signaling agent in the central nervous system,³ where it is proposed to mediate synaptic plasticity.⁴ The biological chemistry of NO has inspired researchers to devise techniques for its bioimaging,⁵ of which turn-on fluorescence is especially attractive because of the demonstrated success of this strategy in real-time monitoring of other cellular signals.⁶ Neurological research in particular has benefited from the development of small-molecule fluorescent sensors specific for Ca^{2+ 6b 7} or Zn²⁺.⁸ Here we report a turn-on sensor for NO with nM sensitivity employing a π-conjugated polymer (CP) as the fluorescent reporter.

The enhanced sensitivity of CPs vs small molecule-based sensors, together with their structural and optoelectronic tunability, renders them intriguing scaffolds for the design and construction of detection systems.⁹ We therefore investigated this class of materials for imaging NO by turn-on fluorescence, a resistivity-based CP sensor for NO having been previously described.¹⁰ The successful realization of CPs for fluorescence-based detection of NO, in the form of a CP–copper complex that exhibits a turn-on response to NO in solution, represents an advantageous new strategy with many possible applications.

Previously, we reported transition metal-based NO sensors that involved displacement of a fluorescent ligand from a mono- or dimetallic center with attendant turn-on of fluorescence, one of which is reversible (Scheme 1a, L = fluorescent ligand).¹¹ More recent work unveiled a related approach,¹² involving NO-induced copper redox chemistry,¹³ with net conversion of fluorophore-labeled, paramagnetic (quenched) Cu(II) complexes to a diamagnetic (fluorescent) Cu(I) state (Scheme 1b). This mechanism has been considered as a possibility for colorimetric NO sensing.¹⁴

Scheme 1.

Strategies for Fluorescent Detection of Nitric Oxide

In further pursuit of the latter strategy, we prepared a series of CPs integrating copper-binding units at defined intervals and screened their relative fluorescence intensities in both CP–Cu(I) and CP–Cu(II) forms, anticipating turn-on NO detection by the process depicted in Scheme 1b.¹⁵ A poly(p-phenylene vinylene) (PPV) derivative incorporating periodic bipyridyl units along the main chain displayed the most suitable properties for further investigation (CP1, Figure 1, Hx = n-hexyl). A structurally related CP is reportedly quenched in the presence of Cu(II), whereas only moderate quenching occurs with added Cu(I).¹⁶ A careful study of the structural and photophysical characteristics of poly(p-phenylene ethynylene)/bipyridyl–Cu(I) complexes has recently been undertaken.¹⁷ This work indicates that stable complexes exist with two bipyridyl units per Cu(I), indicating that only 0.5 equiv of Cu(I) are necessary to attain maximum quenching of polymer luminescence, as independently observed in the present study (Supporting Information).

Figure 1.

Structure of CP1.

CP1 is a bright red-orange solid (λ_max = 462 nm) with strong fluorescence emission centered at 542 nm (Φ = 0.30). Upon addition of 1 equiv of Cu(OTf)₂ to a solution of CP1, the integrated fluorescence intensity was quenched 4-fold, whereas addition of 1 equiv of [Cu(NCMe)₄][BF₄] decreased the fluorescence by ~30% (Figure 2).¹⁵ Adding >0.5 equiv of Cu(II) afforded little additional emission quenching.

Figure 2.

Emission spectra of CP1 (black), CP1–Cu(I) (green), CP1–Cu(II) (red), and CP1–Cu(II) immediately following addition of 300 equiv of NO (g). All measurements were made in 4:1 CH₂Cl₂/EtOH.

Introduction of 300 equiv of NO (g) to the CP1–Cu(II) complex rapidly (<1 min) increased the integrated emission by 2.8-fold, producing a fluorescence spectrum similar to that of the CP1–Cu(I) complex (Figure 2). NO (g) did not alter the fluorescence of CP1 in the absence of Cu(II).

In the proposed mechanism, Scheme 1b, protons are formed that could affect the fluorescence of CP1. Protons alone, however, decrease the emission from bipyridyl-PPVs.¹⁸ Using a handheld UV lamp, a decrease in emission was qualitatively confirmed in the current case following addition of 3 μL of glacial acetic acid to 5 mL of a 2 μM solution of CP1 in 4:1 CH₂Cl₂/EtOH.

Following initial trials with NO (g), we examined the selectivity of the sensor for nitric oxide vs other biologically relevant reactive nitrogen species (RNS). In parallel with ongoing work in our laboratory,¹⁹ we evaluated a nitrosothiol (SNAP, S-nitroso-N-acetylpenicillamine), a nitroxyl (HNO) donor (Angeli’s salt, Na₂N₂O₃),²⁰ and a nitrosonium (NO⁺) source (NOBF₄) for their ability to alter the emission spectra of CP1, CP1–Cu(I), or CP1–Cu(II). These donors were selected on the basis of their commercial availability in high purity and well-studied kinetics of RNS formation.²² None of the donors (50 equiv of SNAP, 50 equiv of NOBF₄, or 16 equiv of Angeli’s salt, Na₂N₂O₃) induced a change in the emission spectra of CP1 or CP1–Cu(I). When 50 equiv of SNAP was added to CP1–Cu(II), a 1.5-fold turn-on of fluorescence occurred slowly over 2 h, in accord with the known ability of cupric ion to catalyze the release of NO from nitrosothiols.^22e The same effect was observed upon reversing the order of Cu(II) and SNAP addition, in which case the expected 4-fold quenching occurred immediately upon Cu(II) addition, followed by a slow turn-on. This control indicates that the presence of excess SNAP does not interfere with Cu(II) binding by CP1. A full 24 h was required following addition of SNAP before a turn-on response (2.1-fold) similar to that evoked by NO (g) was attained. Because of the significantly longer time for SNAP (24 h) vs NO (<1 min) to elicit a fluorescence response, nitrosothiols should not be considered as seriously interfering analytes.

The addition of 50 equiv of NOBF₄ did not affect the fluorescence of solutions of CP1 or of preformed CP1–Cu(I) or CP1–Cu(II). When Cu(II) was added to a solution of CP1 containing 50 equiv of NOBF₄, the emission spectrum was the same as that of CP1–Cu(II). In the presence of EtOH, NOBF₄ will form EtONO. The solution resulting from addition of NOBF₄ to CP1–Cu(I) was therefore expected to contain the same species present in the reaction of CP1–Cu(II) with NO. These two solutions exhibited identical emission spectra.

Although no response was elicited by nitroxyl with CP1–Cu(I), it is noteworthy that an immediate 2.8-fold increase occurred upon reaction of CP1–Cu(II) with 50 equiv of nitroxyl, formed from decomposition of Angeli’s salt. The spectrum produced was nearly identical to that exhibited by CP1–Cu(I), suggesting that nitroxyl may reduce Cu(II) to Cu(I).

The effect of O₂ was also investigated as another possible interfering biological species. Aeration of a cuvette containing a nitrogen-purged CP1–Cu(II) solution slightly decreased the integrated fluorescence (~6%). This response is presumably due the ability of O₂ to serve as a collisional quencher and is within the error of detection.

The sensitivity of the CP1–Cu(II) complex to NO was evaluated by addition of progressively lower concentrations of SNAP to a 630 nM solution of the sensor. On the basis of multiple measurements versus an emission standard, NIST-issued quinine sulfate dihydrate, we determined a 10% increase in integrated emission to be the lowest quantifiable change discernible at our instrumental detection limit. By using this method, we computed a sensitivity of about 6.3 nM for the CP1–Cu(II) system.

Although previous work has provided strong evidence for the mechanism presented in Scheme 1b,¹³ we carried out a number of additional checks to confirm its validity in the present context. The strongest supporting evidence is the near perfect match of emission spectra derived from CP1–Cu(I), CP1–Cu(I)/NO⁺, and CP1–Cu(II)/NO. Exposure of CP1–Cu(II) to NO (g) in the absence of EtOH, which is required to form RONO in the proposed mechanism, did not produce an increase in fluorescence. Finally, the expected diminution of the EPR signal of CP1–Cu(II) occurred upon addition of 1 equiv of NO, confirming reduction to Cu(I). When 1 equiv of Cu(OTf)₂ was added to a solution of CP1 (0.5 equiv bind) followed by 1 equiv of NO (g), the integrated signal decreased by 45%, indicating that only CP1-coordinated Cu(II) is reduced to Cu(I). No bands attributable to stable copper nitrosyls were observed by IR spectroscopy.

Despite the strong evidence and precedence for the proposed mechanism, we note that, in addition to the d⁹ → d¹⁰ transformation that occurs upon reduction, some degree of ligand rearrangement or other structural alteration, to which CP emission is very sensitive,^17,21 could play a role in the observed sensory response. Work in progress with small-molecule model compounds will elucidate such possibilities.

The system reported here represents an early manifestation of a new strategy for the fluorescent detection of NO. To the best of our knowledge, it also signifies the first fluorescence-based sensor for NO employing a conjugated polymer scaffold.¹⁰ Studies have commenced to identify additional transition metal-conjugated polymer complexes for improved sensory response, devise specificity for NO over nitroxyl, and prepare highly fluorescent water-soluble derivatives for biological imaging of nitric oxide.²¹