A Reactive and Specific Sensor for Activity-Based ¹⁹F-MRI Sensing of Zn²⁺

Abstract

The rapid fluctuations of metal ion levels in biological systems are faster than the time needed to map fluorinated sensors designed for the ¹⁹F-MRI of cations. An attractive modular solution might come from the activity-based sensing approach. Here, we propose a highly reactive but still ultimately specific synthetic fluorinated sensor for ¹⁹F-MRI mapping of labile Zn²⁺. The sensor comprises a dipicolylamine scaffold for Zn²⁺ recognition conjugated to a fluorophenyl acetate entity. Upon binding to Zn²⁺, the synthetic sensor is readily hydrolyzed, and the frequency of its ¹⁹F-functional group in ¹⁹F-NMR is shifted by 12 ppm, allowing the display of the Zn²⁺ distribution as an artificial MRI-colored map highlighting its specificity compared to other metal ions. The irreversible Zn²⁺-induced hydrolysis results in a “turn-on” ¹⁹F-MRI, potentially detecting the cation even upon a transient elevation of its levels. We envision that additional metal-ion sensors can be developed based on the principles demonstrated in this work, expanding the molecular toolbox currently used for ¹⁹F-MRI.

The vast majority of imaging probes developed for metal ion sensing comprise two entities: a multidentate organic ligand for cation recognition and binding, and an imageable entity that generates a readable signal.^1,2 Adopting this design principle, responsive paramagnetic contrast agents were developed to monitor labile cations with magnetic resonance imaging (MRI),³⁻⁸ overcoming the restrictions of fluorescent imaging and enabling in vivo investigation of metal ions in deep tissues. In these agents, the ion recognition entity, which participates in the paramagnetic metal coordination, binds the cation of interest, freeing the paramagnetic center to bind water molecules, thus inducing paramagnetic relaxation enhancement and a change in the MRI contrast.⁹ Optimization of the structural and binding properties of such MRI-responsive agents allowed spatial mapping of dynamic changes in Ca²⁺ levels in the brain,¹⁰⁻¹³ Cu²⁺ in the liver,¹⁴ and Zn²⁺ in pancreatic¹⁵⁻¹⁷ and prostate^18,19 tissues in vivo. While these designs benefit from high sensitivity, relying on two different molecular entities, one for cation binding and one for MRI signal amplification, limits the flexibility of the sensor architecture. In addition, the large ¹H-MRI signal of the surrounding tissue may complicate the interpretation of results and their quantification when using paramagnetic contrast agents.

Fluorinated ligands were proposed as an alternative type of cation sensor where the ion-binding entity also serves the signal generator, aiming to overcome some of these limitations.²⁰ With this type of sensor, upon binding the cation of interest, a significant chemical shift offset in the ¹⁹F-NMR spectrum is obtained with a unique MR fingerprint for each ion, allowing the monitoring of multiple ions simultaneously with the aid of a single sensor. Extending this approach for ¹⁹F-MRI studies, combined with the chemical exchange saturation transfer (CEST) contrast mechanism^21,22 has shown promise for noninvasive in vivo mapping of labile Zn²⁺ in prostate²³ and brain²⁴ tissues. Although fluorinated agents provide background-free maps and quantifiable information, which do not apply to paramagnetic MRI-responsive agents, the time needed to acquire their signal is expected to be longer than the changes in metal ion levels in biological systems, thus calling for advances for this group of sensors.

An emerging concept for sensing, which relies on the reactivity of the sensor rather than on molecular recognition and reversible binding, may overcome some of the remaining challenges for the ¹⁹F-MRI of metal ions. In this approach, activity-based sensing (ABS),²⁵ the analyte identification occurs via an analyte-initiated reaction, after which a detectable signal is generated.²⁶⁻²⁹ Applying the ABS principles to ¹⁹F-MRI has shown potential for mapping the activity of several key enzymes³⁰⁻³² and biologically relevant redox conditions.^33,34 The irreversible rapid change in the ¹⁹F-MR properties of the sensor upon conversion creates “turn-on” MRI signals that can be repeatedly acquired, thus overcoming the relatively low sensitivity of the ¹⁹F-MRI approach, which frequently requires long acquisition times. Recognizing the need for advanced and specific MRI-responsive agents for metal ions with biological relevance, we propose a highly reactive but still very specific sensor for the ¹⁹F-MRI mapping of labile Zn²⁺. Upon binding to Zn²⁺, the synthetic sensor is readily hydrolyzed, and the resonance of its ¹⁹F-functional group is shifted by 12 ppm, allowing the display of the Zn²⁺ distribution as an artificial MRI-colored map highlighting its ultimate specificity.

Aiming to obtain a sensor with specific ¹⁹F-NMR properties that shows a different ¹⁹F-NMR chemical shift (Δω) upon Zn²⁺-catalyzed hydrolysis, a fluorinated molecular probe was proposed (Figure 1a,b). To this end, a dipicolylamine (DPA) scaffold was used as a zinc-recognition entity conjugated to a fluorinated phenyl acetate moiety. Based on previous designs, we hypothesized that upon binding to a DPA entity,²⁶ the intramolecular interactions between the bound Zn²⁺ and the acetate’s carbonyl group, along with the Lewis acidity of the ion, would catalyze the hydrolysis of the ester. This hydrolysis is expected to result in a much more stable phenolic complex of Zn²⁺ and induce a change in the ¹⁹F-NMR chemical shift of the probe. Two sensors were obtained, each with an acetate functional group at the phenyl ring and a fluorine substitution at either the meta-position (1-OAc) or the para-position (2-OAc), relative to the acetate group. Then, aqueous solutions of the two sensors were subjected to ¹⁹F-NMR measurements without or with the addition of Zn²⁺. Upon Zn²⁺-induced hydrolysis of the acetate group in 1-OAc, only a small change in the ¹⁹F-NMR chemical shift (Δω = 0.3 ppm) was observed from the resulting 1-OH-Zn²⁺ complex (Figure 1c). In contrast, a significantly large Δω of 12 ppm was detected in the ¹⁹F-NMR spectrum after the hydrolysis of 2-OAc to 2-OH in the presence of Zn²⁺ (Figure 1d). This large Δω, which was applicable for ¹⁹F-MR sensing of Zn²⁺ based on reversible binding,³⁵ allows a clear separation between the two forms of the sensor before and after Zn²⁺-induced hydrolysis.

Figure 1.

Zn²⁺ sensor synthesis and mode of action. (a) The synthetic route for synthesizing 1-OAc and 2-OAc. (b) Illustration of the sensor hydrolysis and its mode of action. (c) ¹⁹F-NMR spectra of 1-OAc (labeled green) and the obtained 1-OH-Zn²⁺ complex (labeled red) upon hydrolysis, with the chemical shift offset between the two forms, Δω = 0.3 ppm(d) ¹⁹F-NMR spectra of 2-OAc (labeled green) and the obtained 2-OH-Zn²⁺ complex (labeled red), Δω = 12 ppm.

The white-transparent crystals of the Zn²⁺-complexes were obtained by crystallizing either 1-OH or 2-OH in the presence of Zn(ClO₄)·6H₂O in methanol at ambient temperature, and the solid-state structures of the complexes of 1-OH and 2-OH with Zn²⁺ were obtained and studied (Figures 2, S1, and S2 and Table S1).

Figure 2.

Crystal structures of complexes of Zn²⁺ with 1-OH and 2-OH. The ORTEP schemes for the dimers obtained are shown in (a) 1-OH-Zn²⁺ and (b) 2-OH-Zn²⁺. Displacement ellipsoids are shown at the 50% probability level, and hydrogen atoms have been omitted for the sake of clarity. Color codes: F^– (purple), O (red), N (blue), and Zn²⁺ (pink).

X-ray crystallography revealed a dimeric structure for both the 1-OH-Zn²⁺ and 2-OH-Zn²⁺ complexes. In 2-OH-Zn²⁺, the preferable molecule for ¹⁹F-MRI (Figure 1d), the Zn²⁺ center is coordinated to three nitrogen atoms from the dicopylamine and two bridging oxygen atoms of fluoro-phenols. The Zn–N bond lengths range between 2.054(2) Å, 2.056(2) Å, and 2.165(2) Å, while the Zn–O bond lengths are 1.995(2) Å, comparable to those found in a previously reported Zn²⁺ fluorescent sensor,³⁶ demonstrating the strong coordination ability of 2-OH’s to the Lewis acidic Zn²⁺ ions in 2-OH-Zn²⁺.

The specificity of the 2-OAc for Zn²⁺ sensing was examined with potentially competitive cations (Figure S3), and the ¹⁹F-NMR spectra of the 2-OAc solution in the presence of Ca²⁺, Mg²⁺, Cu²⁺, Fe²⁺, Fe³⁺, Mn²⁺, Ni²⁺, Co²⁺, Na⁺, or K⁺ were compared to its spectrum in the presence of Zn²⁺. Notably, only the Zn²⁺-containing solution yielded a characteristic ¹⁹F-NMR peak at −129 ppm, which is assigned to the 2-OH-Zn²⁺ complex. No significant ¹⁹F-NMR peak could be assigned to a 2-OH-M⁺ complex in the solution for all of the other studied cations. For the paramagnetic cations examined, either a small shift or a line-broadening could be detected for the ¹⁹F-NMR peak of the fluorinated sensor.

Then, the kinetic properties of the 2-OAc hydrolysis in the presence of Zn²⁺ were compared to that of Ca²⁺, Cu^2+, and Fe²⁺ (Figure 3 and Figure S4). To this end, buffered (HEPES buffer, pH 7.2) solutions of 3 mM 2-OAc in equimolar cation concentrations were prepared, and consecutive ¹⁹F-NMR spectra were measured. Clearly, 10 min after the Zn²⁺ addition, when the acquisition of the first ¹⁹F-NMR was completed, a peak at −129 ppm, a characteristic resonance of the 2-OH-Zn²⁺ complex, was already obtained (Figure 3a and Figure S4a). Over time, the intensity of the ¹⁹F-NMR peak of 2-OAc (−117 ppm) was reduced until its elimination, while the intensity peak of the 2-OH-Zn²⁺ complex was elevated as expected from continuous hydrolysis of 2-OAc by the labile-free Zn²⁺ in the solution. This observation was not detected in the presence of the other cations studied (Figure 3b–d, Figure S4b–d). Even for ions expecting to bind a dipicolylamine scaffold, such as Fe²⁺ and Cu²⁺,³⁷ although some paramagnetic line-broadening and reduction in the intensity of the 2-OAc peak were obtained over time, no evidence of 2-OH complexes in the presence of these cations was shown by ¹⁹F-NMR (Figure 3c,d and Figure S4). Still, traces of such complexes were identified by mass spectroscopy measurements (Figures S5–S7). Importantly, even at a much lower Zn²⁺ concentration, a pronounced peak at the resonance of the 2-OH-Zn²⁺ complex (−129 ppm) could be detected in addition to the nonhydrolyzed 2-OAc compound with its characteristic peak at −117 ppm (Figure S8). Notably, in this case, when the concentration of Zn²⁺ was lower than that of the 2-OAc compound, an additional peak of the hydrolyzed 2-OH compound without Zn²⁺ bound to it was detected at −126 ppm. This observation could be explained by the reversible binding of Zn²⁺ to 2-OH.

Figure 3.

Fluorinated sensor specificity. Real-time ¹⁹F NMR spectra of 2-OAc in the presence of (a) Zn²⁺, (b) Ca²⁺, (c) Cu²⁺, and (d) Fe²⁺. The green rectangle marks the chemical shift of 2-OAc (−117 ppm), and the red rectangle marks the chemical shift of 2-OH-Zn²⁺ (−129 ppm).

The kinetic profile of Zn²⁺-induced hydrolysis of 2-OAc was studied by ¹⁹F-NMR (Figure 4). The half lifetime (t_1/2) of 2-OAc in the presence of Zn²⁺ was then evaluated by plotting the integrals of the two obtaining peaks at the ¹⁹F-NMR spectra, −117 ppm before Zn²⁺-induced hydrolysis and −129 ppm after the hydrolysis. The t_1/2 of 2-OAc in the presence of Zn²⁺ at 25 °C was 9 min (Figure 4a,d) while that at 37 °C was 4 min (Figure 4b,d and Figure S9). Importantly, with the addition of 3 mM Ca²⁺, the t_1/2 of 2-OAc was found to be longer than 30 h, reflecting the sensor stability in an aqueous solution and its specificity for Zn²⁺ (Figure 4c and Figure S10). At acidic conditions (pH = 6.5), although slower than at physiological pH, the t_1/2 of 2-OAc in the presence of Zn²⁺ (at 37 °C) was still relatively short (t_1/2 = 9 min, Figure S11). As expected, the Zn²⁺-induced hydrolysis was faster in alkaline conditions at physiological temperature, with a 100% conversion of 2-OAc to 2-OH before the acquisition of the first ¹⁹F-NMR spectrum was completed (Figure S11). Prominently, under strong basic conditions (pH = 8.5), no hydrolysis of 2-OAc occurred without Zn²⁺, showing the cruciality of the ion in the reaction and the stability of 2-OAc even at an elevated pH (Figure S12).

Figure 4.

Kinetic studies of the 2-OAc activity. ¹⁹F-NMR signal intensity representing the conversion of 2-OAc (green dots, ¹⁹F-NMR peak at −117 ppm) to 2-OH (red dots, ¹⁹F-NMR peak at −129 ppm) in the presence of Zn²⁺ at (a) 25 or (b) 37 °C; or in the presence of (c) Ca²⁺ at 25 °C. (d) The t_1/2 values of 2-OAc as evaluated from the plots in panels a–c.

Similar observations were obtained when the same assay was performed in a cell culture medium (DMEM) containing physiologically relevant concentrations of nucleophilic metabolites (sugars, amino acids, and salts) and 10% fetal bovine serum albumin (FBS). Remarkably, when such a medium contained Zn²⁺, 2-OAc was hydrolyzed entirely before the first ¹⁹F-NMR spectrum was acquired (Figure S13). This contrasts with the exact solution of 2-OAc, to which Zn²⁺ was not added, where even at 37 °C, only a slight hydrolysis of 2-OAc could be detected after 1 h (Figure S14). This observation that serum albumin content accelerates the Zn²⁺-induced hydrolysis of 2-OAc, should be further studied. Nevertheless, it could be attributed to the serum albumin’s ability to stabilize Zn²⁺-bound DPA scaffolds, as previously reported.⁴ This stabilization might strengthen the intramolecular interactions between the bound Zn²⁺ and the acetate’s carbonyl group, thereby accelerating the ester hydrolysis due to the proximity of the Lewis acid (bound Zn²⁺) in the formed 2-OAc-Zn²⁺ complex. Overall, these results show the stability of 2-OAc in the absence of Zn²⁺, regardless of the solution conditions with no apparent hydrolysis without the ion at elevated temperature, alkaline conditions, and the presence of high concentration of nucleophiles and serum protein assuring that the proposed sensor will be stable in a biological environment.

Finally, we set out to examine the ability to map the presence of 2-OAc with ¹⁹F-MRI and present its Zn²⁺-sensing capabilities in a multiplex manner based on the large chemical shifts of the sensor before and after its hydrolysis. For that purpose, a phantom of six tubes containing 3 mM 2-OAc and an equimolar of a different cation (one tube was used as a negative control without cation addition) was set and studied. Figure 5a shows the ¹⁹F-NMR spectra of five tubes containing 2-OAc and a cation in HEPES buffered solution (pH = 7.2). As clearly shown, only for the solution that contained Zn²⁺ a ¹⁹F-NMR peak was observed at −129 ppm, as expected from a 2-OH-Zn²⁺ complex, reflecting the “turn-on” ¹⁹F-MR feature of the Zn²⁺ sensor. A single and clear peak is obtained for all other tubes with a resonance of −117 ppm, reflecting intact, nonhydrolyzed 2-OAc in the studied solution. ¹H-MRI of the studied phantom showed no difference between the tubes (Figure 5b).

Figure 5.

Multiplexed ¹⁹F-MRI of Zn²⁺ sensing. (a) ¹⁹F NMR spectra of 3 mM 2-OAc in the presence of 3 mM Zn²⁺, Ca²⁺, K⁺, Mg²⁺ and Na⁺ at 37 °C. (b) ¹H-MRI (left) of the studied phantom composed of six tubes containing 3 mM 2-OAc and 3 mM cation, i.e., K⁺ (#1), Ca²⁺(#2), Zn²⁺(#3), Na⁺ (#5), and Mg²⁺ (#6). Tube #4 contained only 2-OAc. ¹⁹F-MRI map as obtained with the O₁ set to −117 ppm (middle panel, green) and or −129 ppm (right panel, red).

When ¹⁹F-MRI data was acquired with the center frequency (O₁) set to the resonance of 2-OAc, i.e., 117 ppm, a clear signal was obtained from 5 out of the 6 studied tubes. In contrast, setting up the O₁ of the ¹⁹F-MRI acquisition protocol to 129 ppm, a ¹⁹F-MR signal was obtained only from the tube containing Zn²⁺, a clear indication of the complex of the cation with 2-OH. In contrast to the ¹⁹F-CEST approach, which requires the acquisition of at least two ¹⁹F-MR images for Zn²⁺ mapping (“on-resonance” and “off-resonance”),²⁴ the approach presented here can provide ¹⁹F-MR map of Zn²⁺ distribution by setting up the O₁ of the ¹⁹F-MRI acquisition protocol to −129 ppm and acquiring a single ¹⁹F-MR image.

In addition to this advantage, the large Δω of 12 ppm between in the ¹⁹F-MR resonances of nonhydrolyzed 2-OAc and the Zn²⁺ complex of the sensor after its hydrolysis to 2-OH allowed us to present the results in artificial MRI colors, with green representing the ¹⁹F-MRI map at −117 ppm and red representing the ¹⁹F-MRI map at −129 ppm. This frequency encoding feature resulting in multicolor representations of different species is unique to ¹⁹F-MRI studies and is useful for mapping multiple targets in the same region of interest,³⁸ but it can also be used for other applications.³⁹⁻⁴¹ To summarize this part, we showed that 2-OAc could be used as a sensor for activity-based ¹⁹F-MRI mapping of Zn²⁺ with the capability to present the existence of the cation in artificially colored MRI maps capitalizing on the distinctive chemical shift of the resulted 2-OH-Zn²⁺ complex.

In conclusion, we demonstrated a conceptually novel approach for MRI sensing of labile Zn²⁺, which relies on ABS principles. Specifically, having designed a fluorine-modified phenyl acetate moiety attached to a dipicolylamine motif, a Zn²⁺ sensitive ¹⁹F-MRI sensor (2-OAc) was obtained. We showed that upon Zn²⁺ recognition and binding, 2-OAc readily undergoes hydrolysis to result in the 2-OH-Zn²⁺ complex. The large ¹⁹F-NMR Δω difference between the resonances of 2-OAc and that of 2-OH-Zn²⁺ allows one to spectrally resolve them toward their presentation in a dual-color ¹⁹F-MRI fashion, obtaining a “turn-on” ¹⁹F-MRI sensor for Zn²⁺. The high specificity and reactivity of 2-OAc only in the presence of Zn²⁺ to obtain, nonreversibly, 2-OH makes the proposed strategy advantageous for ¹⁹F-MRI, which frequently requires acquisition times much longer than the time dynamic biological processes occur. We envision that the principles shown in this work for ABS ¹⁹F-MRI sensing of Zn²⁺ could be generalized for imaging other metal ions.

Data Availability Statement

CCDC 2366644–2366645 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_requests/cif, by emailing data_request@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: + 44 1223 336033.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acssensors.4c01895.

• Sensors synthesis and characterizations are provided, along with additional experimental methods and supporting figures (PDF)

Author Contributions

The manuscript was written with contributions from all authors. All authors have given approval to the final version of the manuscript.

This project was funded from the European Research Council under the European Union’s Horizon 2020 research and innovation program (grant No. 101086836) and the Israel Science Foundation (grant No. 1329/20). L.M.L thanks the Zuckerman Postdoctoral Scholars Program.

The authors declare no competing financial interest.