Interrogating Intracellular Zinc Chemistry with a Long Stokes Shift Zinc Probe ZincBY-4

Abstract

Previous work has shown that fluctuations in zinc content and subcellular localization plays key roles in regulating cell cycle progression; however, a deep mechanistic understanding requires the determination of when, where, and how labile zinc pools are concentrated into or released from stores. Labile zinc ions can be difficult to detect with probes that require hydrolysis of toxic protecting groups or application at high concentrations that negatively impact cell function. We previously reported a BODIPY-based zinc probe, ZincBY-1, that can be used at working concentrations that are 20–200-fold lower than concentrations employed with other probes. To better understand how zinc pools can be visualized at such low probe concentrations, we modulated the photophysical properties via changes at the 5-position of the BODIPY core. One of these, ZincBY-4, exhibits an order of magnitude higher affinity for zinc, an 8-fold increase in brightness in response to zinc, and a 100 nm Stokes shift within cells. The larger Stokes shift of ZincBY-4 presents a unique opportunity for simultaneous imaging with GFP or fluorescein sensors upon single excitation. Finally, by creating a proxy for the cellular environment in spectrometer experiments, we show that the ZincBY series are highly effective at 50 nM because they can pass membranes and accumulate in regions of high zinc concentration within a cell. These features of the ZincBY probe class have widespread applications in imaging and for understanding the regulatory roles of zinc fluxes in live cells.

Introduction

Fluctuations in zinc availability plays a wide number of roles across the life cycle of a cell. A good deal of the intracellular Zn(II) is bound to metalloenzymes active sites that can have quite long half-lives for chemical exchange or dissociation. The zinc ions in these sites are described as having ‘catalytic’ or ‘structural’ roles; however, a growing number of studies indicate a third functional category: i.e. a regulatory roles wherein biology uses transient fluctuations in Zn(II) localization to regulate large scale processes including insulin secretion, immune response, neurological signaling, meiotic cell cycle, and fertilization.^1–5 In order to study these fluctuations, a significant number of zinc-responsive fluorescent probes have been developed based on both small molecule and fluorescent protein scaffolds.^6–10 Each of these probes have strengths and weaknesses (see reviews by Sfrzzetto et al., Carter et al., and Que et al.);^{6, 8–9} however, a major limitation of small molecule probes is that the working concentrations required to load the probe into a cell can perturb the biological event under evaluation.

This limitation of zinc probes is particularly apparent in studies of zinc fluctuations that regulate cell cycle progression in the female gamete, i.e. the egg, where addition of an intracellular chelator can induce parthenogenesis.^11–14 A number of studies have shown that subcellular fluctuations in zinc availability during cell cycle progression are essential for the maturation of the egg and, after fertilization, for proper embryo development. Zinc gain and loss is required for the normal transitions of oocyte to egg to embryo.^{11–13, 15} During the egg to embryo transition zinc is lost from the egg in exocytotic bursts termed zinc sparks.^{11, 16–17} Using a new zinc probe, ZincBY-1, Que et al showed that zinc loss occurred from a system of thousands of vesicles at the periphery of the egg that contain high concentrations of labile zinc.¹⁸ Upon fertilization, these cortical vesicles fuse with plasma membrane of the egg, releasing billion zinc ions and temporarily raising the zinc concentration in the egg envelope up to 500 μM.¹⁹ The vesicles containing this large pool of labile zinc correspond to cortical granules, as shown by co-localization of ZincBY-1 with ovastacin, a zinc metalloprotease that is known to be released during cortical granule exocytosis.²⁰

While ZincBY-1 has similar photophysical properties to a variety of zinc probes, it can be used to reveal the localization of labile zinc stores at nanomolar working concentrations.¹⁸ This contrasts with many other probes that require cells to be incubated at micromolar concentrations, which could negatively impact the metal homeostasis of the cell^21–22 It is not clear why ZincBY-1 is able to effectively visualize zinc at such low working concentrations; however, we hypothesized that ZincBY-1 specifically localizes to regions of labile zinc, such as the cortical granules of the mammalian egg^.18 Other zinc probes tested within the mammalian egg were unable to visualize zinc within vesicles at a 50 nM working concentration.¹⁸

One of the challenges in utilizing ZincBY-1 in imaging experiments alongside other probes is its narrow Stokes shift (11 nm).¹⁸ In order to avoid scattered light from the excitation beam from interfering with the emission signal, the emission detection range needs to be selected sufficiently far away from the excitation wavelength. This reduces the strength of the fluorescence signal, leading to a decrease in the signal to noise. Substituents at the 5-position of the boron-dipyrromethane core (BODIPY, see Scheme 1) can modulate the photophysical properties. A red-shift in fluorescence and long Stokes-shift for BODIPY fluorophores have been observed upon the addition of bulky aryl substituents, increase in conjugation through alkynyl groups, or fusion with other aromatic groups.^23–26 In order to optimize the photophysical properties of the ZincBY probe, we switched the methoxy group at the 5-position of the BODIPY core to an ethoxy group with a terminal amide. Alterations of the BODIPY core leads to changes in the hydrophobicity of the molecule providing potential insight into how ZincBY-1 is able to visualize zinc in living cells with only a 50 nM incubation concentration.

Scheme 1.

(A) Structures of ZincBY-1, ZincBY-2, ZincBY-3, and ZincBY-4. Carbons 3 and 5 of the BODIPY core, and nitrogens 3a, 3b, and 4a are indicated on ZincBY-4. (B) Synthetic scheme detailing the synthesis of ZincBY-2, −3, and −4. Full synthetic details are available in Supporting Information

Herein we present the synthesis, photophysical characterization, and utility of three new vital zinc probes obtained via modification of the 5-position of the BODIPY core. By analyzing substituents that possesses either a terminal azide, amine, or acetamide (ZincBY-2, ZincBY-3, and ZincBY-4 respectively) we conclude that these probes avidly concentrate and are electrostatically trapped in compartments of the cell that are otherwise enriched in labile zinc. Furthermore, we find favorable attributes of the ZincBY-4 probe arise from the formation of a hydrogen bond between the amide hydrogen of the pendant chain at the 5-position and one or both of the fluorine atoms of the BF₂ center. This intramolecular hydrogen bond constrains the geometry, alters the basicity of the exocyclic nitrogen, increases the zinc affinity, and increases the Stokes shift. The increase in Stokes shift allows for simultaneous imaging of static zinc pools or zinc fluxes with green fluorescent sensors in live cells, increasing the utility of the ZincBY series.

Results and discussion

Synthesis and Characterization of new ZincBY Probes

The synthetic approach to this series of new zinc probes is shown in Scheme 1. ZincBY-1 (1) was synthesized by the attachment of a trispicen chelator (8) to a 3-chloro-5-meothxy-8-mesityl-BODIPY core according to previously published procedures.¹⁸ ZincBY-2 (2), ZincBY-3 (3), and ZincBY-4 (4) were synthesized by first substituting 2-azidoethanol (6) at the 5 position of 3,5-dichloroBODIPY (5) to create the asymmetric BODIPY, 7. ZincBY-2 was formed by substitution of trispicen (8) for Cl at the 3-position of 7. Staudinger reduction on the azide of ZincBY-2 to a primary amine was accomplished using trimethylphosphine, resulting in ZincBY-3. Finally, capping of the primary amine of ZincBY-3 with acetic anhydride yielded ZincBY-4. Full synthetic and characterization details are available in the Supporting Information (SI) with the ¹H NMR spectra (Figure S1–S4).

To compare the quantum yields and extinction coefficients across the series, we first remeasured these values for ZincBY-1 and found a 10-fold higher brightness for both the apo and holo states compared to the originally reported values (Table S1), however the absorption and emission wavelengths (533 nm and 543 nm respectively) and turn-on ratio (ca. 5-fold) remain the same.¹⁸ The apo form of ZincBY-4 has absorption peaks at 482 and 570 nm and an emission maximum at 585 nm, while the holo form has a single absorption maximum at 482 nm and emits at 537 nm (when excited at 480 nm) (Figure 1).

Figure 1.

Photophysical properties of ZincBY-4. (A) UV/Visible spectrum of holo and apo ZincBY-4 probe (200 nM) in aqueous buffer. (B) Emission spectrum of holo- and apo-ZincBY-4 (200 nM) in aqueous buffer. (C) Determination of the zinc dissociation constant (228.7 ± 18.4 pM) of ZincBY-4 in a buffered EGTA-Zinc solution. (D) Fitted titration curves gave zinc dissociation constants of 3.72 ± 0.87 nM for ZincBY-1 (2.6 nM reported),¹⁸ 18.1 ± 2.6 nM for ZincBY-2 and 97.7 ± 26.1 pM for ZincBY-3.

The quantum yield of ZincBY-4 in the apo form is 0.164, 5% lower compared to ZincBY-1. In the holo form, the quantum yield of ZincBY-4 is 0.574, 10% lower compared to ZincBY-1. In addition, the extinction coefficients are also approximately half that of ZincBY-1 (Table 1). The lower quantum yields and extinction coefficients of ZincBY probes with substituents other than methoxy at the 5-position result in less brightness compared to ZincBY-1; however, there is a greater fluorescence turn-on in the presence of zinc for ZincBY-4 (8.4) compared to ZincBY-1 (5.2) in aqueous buffered solutions. Full characterization details for ZincBY-2 and ZincBY-3 are also provided in the SI (Table S1, and Figures S10–S13) while a comparison of the photophysical properties of ZincBY-1 and ZincBY-4 can be found in Table 1. We speculate that the higher quantum yield of ZincBY-1 relative to the other family members arises from a smaller contribution of charge transfer in excited state of ZincBY-1 (vide infra).

Table 1.

Comparison of the photophysical properties of ZincBY-1 and ZincBY-4.

	Apo-ZincBY-1	Apo-ZincBY-4	Zn-ZincBY-1	Zn-ZincBY-4
λAbs (max)	533 nm	572 nm	533 nm	482 nm
Extinction Coefficient (ε), M−1cm−1	2.42(33) × 104	0.88(4) × 104	4.14(33) × 104	2.12(4) × 104
λEm (max)	560 nm	585 nm	543 nm	537 nm
Quantum Yield (Φ)	0.215(5)	0.164(4)	0.651(2)	0.562(11)
Brightness (ε*Φ/1000), M−1cm−1	5.21(72)	1.44(7) × 104	26.97(2.30) × 104	12.15(34) × 104

Like ZincBY-1, ZincBY-4 responds preferentially to Zn(II) relative to other biologically relevant metal ions (Figure S12). While several other families of zinc probes exhibit significant pH-dependent fluorescence changes, neither the holo or apo forms of ZincBY-1 and ZincBY-4 show significant changes in fluorescence in the physiological pH range (pH 4–8).¹⁸ In addition, the presence of 10 mM glutathione (GSH) protected both holo- and apo-ZincBY-4, leading to a retention of 97% and 40% fluorescence intensity respectively after 12 hours, compared to 83% and 7% fluorescence intensity respectively after 12 hours in the absence of GSH (Figure S12). This suggests that ZincBY-4 is sensitive to oxidative damage; however, the presence of GSH in cells should protect the probe from damage. In addition, the Zn(II) bound form of ZincBY-4 is much more stable compared to the apo-ZincBY-4, thus if zinc pools are being monitored with ZincBY-4 over long periods of time there should be little degradation in fluorescence signal. Intriguingly, the zinc affinity of the probes varied widely depending on the substituent attached to the ethoxy group, which is quite removed from the chelation site. While the affinity of the azide substituent decreased the binding to 18.1 ± 2.6 nM, reduction to the amine increased the binding by three orders of magnitude to 97.7 ± 26.1 pM, and acetylation of the amine to the acetamide reduced this binding to 228.7 ± 18.4 pM (Figure 1).

Mechanistic Analysis of Photophysical Properties

The 180-fold differences in zinc affinities and striking differences in photophysical properties across this series of probes suggests that this pendant arm can exert significant effects on the electronic structure of the BODIPY core. Given that ZincBY-2, ZincBY-3, and ZincBY-4 all share two methylene spacers after the ether linkage at the 5-position in their common BODIPY core, it is unlikely that differences in the terminal groups contribute to any significant inductive effects. One major difference is that the substituents at the 5-position in ZincBY-3 and ZincBY-4 have a terminal amine or amide respectively and can form intramolecular hydrogen bonds to the BODIPY core in contrast to ZincBY-2 with a terminal azide.

We hypothesize that the increase in zinc affinity in ZincBY-3 and ZincBY-4 relative to ZincBY-1 and ZincBY-2 arises because of changes in the electron density of the BODIPY core, stabilized by intramolecular hydrogen bond(s) between the amine or amide hydrogen and a fluorine on the BF₂ center of the BODIPY fluorophore, increasing the basicity of the exocyclic nitrogen 3b (Scheme 1). Such intramolecular hydrogen bonding to the BF₂ center in the BODIPY core has been reported in previous literature and have been demonstrated to increase the Stokes shift in other systems, a phenomenon that we observe in our system as well.^27–28 To probe this idea, we first evaluated the changes in photophysical properties, shielding effects on heteronuclei of the BODIPY core, and in DFT electronic structure models before returning to address the differences in zinc affinity.

Analysis of the excited state of ZincBY-1 and ZincBY-4 through time dependent density functional theory (TD-DFT) calculations using the PBE0 functional^29–31 and def2-TZVP basis set³² indicate that a low energy charge transfer transition is present when the intramolecular hydrogen bond is intact in ZincBY-4; this transition is predicted to be attenuated when no hydrogen bond is present as in ZincBY-1 (Figure 2). The presence of an intramolecular hydrogen bond does not dramatically alter the HOMO or the LUMO of ZincBY-4 (Figure S15). It should be noted that the transitions of the DFT calculations do not accurately match the experimental results. For ZincBY-1, the DFT calculations show a higher energy transition for the holo form; however, experimentally both the apo and holo forms of ZincBY-1 have a transition at the same energy (Figure 2). In addition, despite the presence of a low energy charge transfer transition for apo-ZincBY-4 when a hydrogen bond is present, neither of the transitions completely match the experimental results (Figure 2). Since the PBE0 functional tends to overestimate the absorption energies of BODIPY molecules, we next moved to using a range separated ωB97-XD3 functional;³³ however, in moving to this method, the charge-transfer transition that we saw previously disappeared (Figure S14).

Figure 2.

Computational analysis of ZincBY-1 and ZincBY-4. Comparisons of TD-DFT calculated absorbance spectra utilizing a PBE0 functional compared to the experimentally determined absorbance spectra of (A) ZincBY-1 and (B) ZincBY-4. The calculated spectra for ZincBY-4 were simulated to both enforce (H-bond) and prohibit (no H-bond) hydrogen bonding between the amide and to the BF2 group. (C) Partial charges, indicated in black, of nitrogen 3a, nitrogen 3b, and nitrogen 4a of the BODIPY cores of ZincBY-1, ZincBY-4 without a hydrogen bond (R_H-F = 5.16 Å), and ZincBY-4 with a hydrogen bond (R_H-F = 2.31 Å). Partial charges were calculated using the CHELPG method. ZBY stands for ZincBY.

To test whether intramolecular hydrogen bond formation is possible in ZincBY-4, we carried out heteronuclear NMR experiments in CD3CN, as it is a poor hydrogen bond acceptor and is not expected to disrupt intramolecular hydrogen bond formation. Initially, the ¹H NMR spectrum of ZincBY-4 was fully assigned (Figure S4). Next, we employed ¹⁹F-¹H 2D heteronuclear Overhauser Effect Spectroscopy (HOESY) and found magnetization transfer between the amide proton and the fluorines (Figure S5): this is consistent with close contact of the amide proton with at least one of the fluorine atoms. Further evidence for an intramolecular hydrogen bond can be seen in shielding trends in the ¹⁹F and ¹¹B NMR spectra. For ZincBY-1, the peaks of the BF₂ moiety peaks are centered at −135.1 ppm for ¹⁹F and 1.02 ppm for ¹¹B (Figures S6 and S8). The ¹⁹F resonance for ZincBY-4 shifts upfield to −141.4 ppm and appears as a broad multiplet instead of the canonical quartet, while a modest downfield shift in the ¹¹B spectrum to 1.38 ppm is observed (Figures S7 and S9). A full analysis of the NMR parameters and solutions analyzed can be found in Tables S2 and S3. These results indicate that there is less π-delocalization occurring within ZincBY-4 compared to ZincBY-1, as similar shifts within BODIPYs was noted previously.³⁴ The upfield shift in the ¹⁹F peak is consistent with the fluorine being involved in a hydrogen bond with the amide hydrogen. The upfield shift further indicates that the ¹⁹F nucleus is a better hydrogen bond acceptor. Similar shift has been reported by other groups investigating hydrogen-bonding to the BF₂ center of other BODIPY systems.^35–36 Taken together, these results are consistent with the presence of an intramolecular hydrogen bond between the amide hydrogen and a fluorine on the BF₂ moiety of the BODIPY fluorophore.

Based on analysis of electron delocalization, we propose that the resonance contributions of various resonance forms of the of the BODIPY π system (Scheme 2) are significantly different when this intramolecular hydrogen bond is intact. Alterations to the resonance preference of the BODIPY fluorophore that alter the basicity of the exocyclic nitrogen atom (N 3b) would explain the tighter zinc binding of ZincBY-4 compared to ZincBY-1. For an asymmetric BODIPY such as ZincBY-4, nitrogen 3b is in resonance with the BODIPY core (see structure III in Scheme 2), causing forms II and III to be the major contributors and the formal positive charge shared between nitrogen 3a and nitrogen 3b.³⁷ We hypothesize that the intramolecular hydrogen bonding in ZincBY-4 decreases the electron-withdrawing character of the BF₂ center of the BODIPY, increasing the resonance contribution of form I and decreasing the resonance contribution of III.

Scheme 2.

Resonance forms of ZincBY-4, 4. I and II are the generally accepted resonance forms of BODIPY; however, III is also an expected form as shown by a similar analogue.³⁷

A decrease in the contribution from III would lead to an increase in the electron density on nitrogen 3b, which is key for zinc binding. This hypothesis is supported by a DFT analysis of the electronic structure of ZincBY-4 with and without the hydrogen bond. Electron density at the nitrogens was determined by calculating the partial charges through CHarges from Electrostatic Potentials using a Grid-based method (CHELPG).³⁸ The CHELPG analysis revealed that ZincBY-4 with a hydrogen bond results in a 0.41 more negative charge density on nitrogen 3b compared to ZincBY-4 without a hydrogen bond (Figure 2). The lack of a hydrogen bond in ZincBY-4, also results in a 0.18 more positive charge density compared to ZincBY-1 (Figure 2). Thus, the presence of an intramolecular hydrogen bond in ZincBY-4 leads to a more negative partial charge on nitrogen 3b, i.e. there is more electron density on nitrogen 3b. An increase in electron density on nitrogen 3b explains the strong binding affinity of ZincBY-4 compared to ZincBY-1.

These insights into the effect of hydrogen bonding on the electronic structure of BODIPY core is consistent with the thermodynamics of zinc binding across the series. The probe with a pendant amine (ZincBY-3) will have a strong coulombic contribution and is anticipated to form the strongest intramolecular hydrogen bonds to the BF₂ center. ZincBY-3 has a 38-fold higher affinity for zinc than that of ZincBY-1, while the pendant amide group in ZincBY-4, which has a smaller coulombic contribution, has an affinity that is only 16-fold higher. ZincBY-2, which has a pendant azido group that cannot serve as a hydrogen bond donor, has a five-fold weaker affinity for zinc than that of ZincBY-1. We speculate that the greater bulk of the methylene spacers and azido group in ZincBY-2 relative to the methyl group in ZincBY-1 may partially disrupt hydrogen bonding interactions between the BF₂ unit of ZincBY-2 with solvent.

In Cell Imaging with ZincBY-4

Due to the higher zinc affinity and a change in the emission and absorption properties of ZincBY-4 compared to ZincBY-1, we compared ZincBY-1 and ZincBY-4 in head-to-head studies in the mouse egg. Interestingly, when staining MII eggs with ZincBY-4, the photophysical results differed in the cell compared to the previously measured in vitro properties (Figure 3). We expected to see optimal emission from the holo probes at 495–560 nm with a 488 nm excitation (λ_em,max = 537 nm in buffer); however, significant fluorescence signal was observed in the 575–750 nm region. Only ZincBY-4 was readily excited at 488 nm, whereas both probes were excitable at 552 nm (Figure S16).

Figure 3.

Probing zinc distribution with ZincBY-4 in the MII egg. (A) MII eggs were incubated with ZincBY-4 (50 nM) for 10 minutes and then their fluorescence images were taken upon excitation with a 488 nm laser in and a 2 μm optical section. In contrast to our expectation that the emission of the holo probe would be observed in 495–560 nm, the probe exhibited strong fluorescence emission in 575–750 nm. Addition of TPEN (5 μM) led to a loss of fluorescence signal in the vesicle region. Scale bar is 20 μm. (B-E) To determine if ZincBY-4 interacts with proteins, the emission (B, D) and absorbance (C, E) spectra of ZincBY-4 (200 nM) were measured in the absence (B, D) and presence (C, E) of BSA (10 mg/mL) in buffer. The presence of BSA red-shifted the emission maximum wavelength from 537 nm to 575 nm..

The BODIPY core of the ZincBY series is hydrophobic, and this moiety has a well-established propensity to associate with lipophilic regions of the cells.³⁹ The BODIPY-based probes have been demonstrated to localize to granules in cells and become unresponsive to metal, thus no single probe can effectively image across all organisms.^40–42 When N,N,N’,N’-tetrakis(2-pyridinylmethyl)-1,2-ethanediamine (TPEN) was added to MII eggs stained with ZincBY-4, loss in fluorescence was observed (Figure 3A). It is possible ZincBY-4 is interacting with hydrophobic biopolymers as it binds to zinc, forming a ternary complex. Zinc probes, such as Zinquin, have been shown to form ternary complexes with zinc and proteins and bind to proteins in a non-specific manner in vitro.^43–44 To determine if intracellular proteins affect ZincBY fluorescence in living cells, 10 mg/mL BSA, as a surrogate for the cellular milieu, was added to the buffer used to measure the fluorescence properties of ZincBY-4 in vitro (Figure 3). While the total protein concentration in a cell is approximately 2.5 mM,⁴⁵ we chose 10 mg/mL BSA (0.16 mM) as a conservative number to prevent the potential obscuring of fluorescent signal by a high concentration of protein.

In the presence of BSA, the emission maximum of holo-ZincBY-4, excited at 480 nm, shifted from 537 nm to 575 nm, similar to what is observed in living cells. Emission from the apo form, also shifts from 585 nm to 590 nm (Figure 3B). There is no shift in the absorption maximum of the apo probe; however, the absorbance of holo-ZincBY-4 in the presence of BSA red-shifted giving a broad plateau from 490 nm to 540 nm (Figure 3C). The apo and holo forms both displayed an increase in polarization from 0.03 to 0.37 in the presence of the excess BSA, indicating that ZincBY-4 binds readily to the lipid transport protein. The same interactions can be found for ZincBY-1, ZincBY-2, and ZincBY-3 as well (Figure S17). However, while ZincBY-4 still maintained a fluorescence turn-on in response to zinc in the presence of BSA (ca. two-fold), no fluorescence turn-on was observed for ZincBY-1 in the presence of BSA (Figure S17).

Analogous non-specific interactions between ZincBY-4 and lipophilic regions occur inside cells. To better understand these changes further, insights into the photophysical mechanisms of these probes are required.

Mechanisms of Fluorescence

The ZincBY series of probes utilize a BODIPY fluorophore conjugated to a trispicen chelator. The electron pair of the trispicen nitrogen (electron-donating) attached directly to the BODIPY (electron-accepting) is in conjugation with the fluorophore’s π system. In polar solvent ZincBY fluorescence is quenched via an intramolecular charge transfer (ICT).^46–47 The fluorescence emission of probes undergoing an ICT mechanism are highly dependent on the solvent.^48–49 To test whether the solvent affects the photophysical properties of the ZincBY-4, we varied the ratio of water to acetonitrile and examined how the fluorescence emission and intensity changed as a function of dielectric constant (Figure 4).

Figure 4.

Photophysical properties of ZincBY-4 in an altered dielectric environment. (A) Emission maximum wavelength and (B) Fluorescence intensity of ZincBY-4 as a function of dielectric constant. The fluorescence intensity was normalized to the fluorescence intensity in 100% water of that condition (ε = 78). The dielectric constant of solvent was modulated by using a different ratio of water to acetonitrile. (C and D) Fluorescence spectrum of ZincBY-4 (200 nM) in acetonitrile/water mixture with (C) and without zinc.

Interestingly, as the dielectric constant decreases, the emission maximum wavelength of the apo ZincBY-4 does not change, but its holo form undergoes a red shift from 537 nm to 575 nm (Figure 4A), the same wavelength observed when the probe binds non-specifically to BSA. On the other hand, the fluorescence intensity of the holo form does not change as the dielectric constant changes, but the apo one does change in response to dielectric constant. Hydrogen bonding of the amide to the fluorine (Scheme 2) will become more favorable in a hydrophobic environment due to a lack of interaction with the surrounding solvent. Since the intramolecular hydrogen bond is much more transient in a more polar environment, the Stokes shift is anticipated to be smaller in water compared acetonitrile, which is observed. Other reports of an intramolecular hydrogen bond to a fluorine of the BODIPY fluorophore also observe a long Stokes shift and are consistent with our observations.^27–28

Advantages of ZincBY-4 in live cell imaging studies

Given the large Stokes shift observed under intracellular conditions, we anticipate several advantages when using ZincBY-4 over the parent probe, ZincBY-1, including the ability to simultaneously image multiple reporters in the same cell. We treated live MII eggs with both ZincBY-4 and TubulinTracker Green (a taxol-oregon green 488 conjugate), a vital probe that is used to stain tubulin microtubules in live cells (Figure 5, S18).⁵⁰ MII eggs provide a good proof-of-concept system as the large spatial differences between the zinc vesicles and the tubulin metaphase spindles permits ready visualization of any spectral overlap by the two probes. With careful control of the imaging parameters (e.g. adjusting the detection channel wavelengths for both probes) there is virtually no spectral overlap between the two probes (Figure 5A).

Figure 5.

Simultaneous excitation of ZincBY-4 and a green-emitting fluorophore in the MII mouse egg and C. elegans worm. (A-C) Detection of zinc and tubulin using a single excitation wavelength in a 2 μm optical section of a MII mouse egg. Both ZincBY-4 (50 nM, Red) and TubulinTracker Green (240 nM, Green) are excited at 488 nm. Detection windows are set to allow for minimal overlap of the probe signals. Fluorescence images of a live mouse egg that was stained only with TubulinTracker Green (A), only ZincBY-4 (B), or both ZincBY-4 and TubulinTracker Green (C). Scale bar is 20 μm. (D-F) Detection of Zinc and Tubulin using a single excitation wavelength in C. Elegans worms. EU1067 contains a GFP::Histone and GFP::Tubulin fusion and N2 is the wild type. Both ZincBY-4 (50 μM) and GFP are excited at 488 nm. Detection windows are set to allow for minimal overlap of the probe signals. Fluorescence images show non-stained EU1067 worms (D), ZincBY-4 stained N2 worms (E), and ZincBY-4 stained EU1067 worms (F). The −1 is above the −1 oocyte with the arrow indicating the direction of the earlier oocytes. Scale bar is 50 μm. The * indicates the gut and the surrounding gut granules.

This is evident in samples stained with only a single probe and looking for any signal in the opposite channel. The utility of this probe is not limited to just single cells studies, but can be used in live animal imaging, i.e. in studies of zinc distribution within C. elegans worms (Figure 5B). With the same imaging parameters across all conditions, there is very little overlap of ZincBY-4 in the GFP detection channel and no GFP fluorescence detected in the ZincBY-4 channel, allowing for detection of both probes at the same time with a single excitation wavelength.

We found that worms stained with ZincBY-4 show a pattern of increasing fluorescence intensity as oocytes progress through the gonad, where the most mature oocyte (referred to as the −1 oocyte) showed the highest signal (Figure 5b, S19). The biological significance of more labile zinc in the −1 oocyte is currently unknown and is the subject of further investigation. Additionally, the intense signal observed in the ZincBY-4 stained worms can be attributed to zinc stored in the intestines. C. elegans accumulate zinc within lysosome-related organelles called gut granules and is likely the source of the punctate ZincBY-4 signal.⁵¹

ZincBY-4 can also be effectively utilized in time lapse imaging to evaluate zinc transients in living cells. Extracellular free Zn(II) released upon fertilization of mouse eggs is readily observed using FluoZin-3 acid after parthenogenic activation of MII eggs by treatment with the calcium ionophore, ionomycin, which raises cytoplasmic calcium levels to trigger vesicle release and the resumption of meiosis through calmodulin dependent signaling pathways.¹¹ The transient increase in extracellular free zinc was shown to directly correlate with a decrease in free zinc concentration in vesicles located at the cell surface: these events were followed by loading cells with ZincBY-1 and using another parthenogenic activator, strontium chloride, to stimulate zinc release.¹⁸ When ZincBY-4 is loaded into the cells, we observe the both loss of zinc from the zinc granule region and increase of zinc outside the cell without the need of a second probe (Figure 6, S20, Video S1). The drop of fluorescence intensity in the intracellular zinc granule region occurs upon vesicular fusion with the plasma membrane, which allows release of the vesicle’s zinc cargo. The fluorescence intensity outside the egg increases due to the release of the zinc-probe complex at the time of vesicle exocytosis. These results demonstrate that ZincBY-4 applications are not limited to imaging static pools of zinc, but can also probe dynamic changes in zinc concentration and localization.

Figure 6.

Imaging the zinc vesicle loss in a single MII egg during chemical activation (mimic of fertilization). (A) A cohort of MII eggs were incubated with ZincBY-4 for 10 minutes and then chemically activated with ionomycin and visualized via confocal microscopy with an open pinhole resulting in a 26 μm z-section through the center of the egg. Scale bar is 20 μm. (B) The outside and periphery of the egg were analyzed by creating a region of interested outside of the egg (outlined in red, denoted by X) and the periphery of the egg (outlined in blue, denoted by *). The intensities in these regions were plotted over time and normalized to the intensity of the 0 second time point of those regions. There is an increase in fluorescence intensity outside the egg while fluorescence intensity is lost in the periphery region following chemical activation. Scale bar is 20 μm.

Conclusions

Using a new series of zinc specific probes having variations in the 5-position of the BODIPY core of ZincBY-1, we can now explain the photophysical properties of the ZincBY probes and apply this knowledge to develop better methods for imaging intracellular zinc fluxes in live-cell imaging experiments. Modifications at the 5 position altered the photophysical properties of the BODIPY in a way that facilitates an expanding number of applications. One derivative, ZincBY-4 displayed a large Stokes shift and higher zinc affinity, making it more suitable for detection of zinc within live cells simultaneously with other green fluorescent probes. Through HOESY NMR and computational studies, we showed that the amide proton has significant interactions with the fluorine atoms in the BF₂ moiety at the BODIPY core. These interactions change the resonance stabilization and fluorescence properties. CHELPG analysis of ZincBY-4 reveals that intramolecular hydrogen bond influences the electronic structure of the BODIPY core in a manner that leads to significantly more electron density on the exocyclic nitrogen at the 3b position, thereby increasing the binding affinity for zinc. The large Stokes shift in ZincBY-4 could arise from an intramolecular hydrogen bond in the excited state, which can be broken due to a change in geometry, as is typical for an ICT based probe. A relaxation of the geometry to restore the hydrogen bond results in a loss of energy and increasing Stokes shift.⁵² Modification to the core of the BODIPY that promote intramolecular hydrogen bonds to the boron-center provide a way to alter the photophysical properties of the fluorophore without adding additional steric bulk or increasing the hydrophobicity of the fluorophore provide an interesting avenue to explore for further probe design.

These results provide additional rationale that help explain the differences between the observed photophysical properties of ZincBY-4 observed in cells and in the cuvette. We found that several of the intracellular properties of ZincBY-4 can be replicated in vitro by either providing non-specific protein interactions (i.e. addition of BSA) or by moving to a lower dielectric solvent. Interestingly, we observed similar effects of non-specific binding to BSA for ZincBY-1 as well. The apo and holo spectra of ZincBY-1 are very close in emission maximum wavelength and intensity, indicating that within a hydrophobic environment, the apo probe is potentially indistinguishable from the holo one. One of the striking results from our earlier work suggested that we can achieve detection signal at 50 nM probe concentration due to accumulation of the ZincBY probe in zinc-rich compartments.¹⁸ Our new data corroborates this observation and further suggests the loss of fluorescence upon TPEN treatment removes zinc from the probe in cells, allowing the neutral apo form to diffuse throughout the rest of the cell. Additionally, while the fluorescence of apo-ZincBY-1 matches that of holo-ZincBY-1 in the presence of BSA, holo-ZincBY-4 shows a two-fold increase in fluorescence intensity over its apo form, highlighting another advantage of ZincBY-4 over ZincBY-1 (Figure 3). Unlike ZincBY-1, ZincBY-4 can be excited at 488 nm. Due to the large Stokes shift of ZincBY-4, it can be used in conjunction with fluorescent proteins such as GFP in a variety of experiments that test for co-localization. This property was demonstrated by our ability to excite both GFP and oregon green 488 and ZincBY-4 in C. elegans oocytes and M. musculus eggs with little spectral overlap. Simultaneous detection opens the avenue for greater temporal precision in live cell experiments to examine how zinc interacts with other factors that can be detected by GFP or another green fluorescent sensor.