Tools and techniques for illuminating the cell biology of zinc

Abstract

Zinc (Zn²⁺) is an essential micronutrient that is required for a wide variety of cellular processes. Tools and methods have been instrumental in revealing the myriad roles of Zn²⁺ in cells. This review highlights recent developments fluorescent sensors to measure the labile Zn²⁺ pool, chelators to manipulate Zn²⁺ availability, and fluorescent tools and proteomics approaches for monitoring Zn²⁺-binding proteins in cells. Finally, we close with some highlights on the role of Zn²⁺ in regulating cell function and in cell signaling.

INTRODUCTION

Zn²⁺ is required by thousands of cellular proteins where it has structural, catalytic and regulatory functions [1]. Zn²⁺ regulates a wide-range of molecular processes that can profoundly affect cellular and organismal biology [2, 3]. The total concentration of Zn²⁺ within mammalian cells is on the order of hundreds of micromolar [4], while readily exchangeable or labile Zn²⁺ is maintained on the order of hundreds of picomolar [5]. Multiple studies have demonstrated that in a wide range of mammalian cell types cytosolic Zn²⁺ levels are near 100 pM [5-7] and the labile Zn²⁺ pool can fluctuate in response to extracellular stimuli [8-10]. Tools and techniques that are capable of differentiating the labile and bound Zn²⁺ pools, manipulating Zn²⁺ levels in media and in cells, and proteomic approaches for profiling the Zn²⁺ proteome are critical to our understanding of Zn²⁺ biology.

ILLUMINATING Zn²⁺ IN LIVING CELLS WITH FLUORESCENT SENSORS

In contrast with elemental mapping techniques which can provide valuable information about total Zn²⁺ distribution in cells [11], fluorescent Zn²⁺ sensors can specifically report on the labile Zn²⁺ pool. Fluorescent Zn²⁺ sensors come in two main flavors: genetically encoded and small-molecule sensors. Both classes contain a metal-binding group and at least one fluorophore that absorbs and emits light in the visible region of the electromagnetic spectrum. Zn²⁺ binding results in a change in the overall structure or electronic configuration of the sensor, resulting in a change in fluorescence that can be measured using a fluorescence microscope.

When selecting a Zn²⁺ sensor, one must consider characteristics inherent to the fluorophore, such as absorption and emission properties, photostability, pH-sensitivity and brightness, as well as metal-binding properties, like selectivity, kinetics and affinity. Fluorescent sensors are most sensitive to changes in Zn²⁺ concentrations that are near their apparent dissociation constant for Zn²⁺ binding (K_d’). The K_d’ is defined as the labile Zn²⁺ concentration at which the sensor is half-saturated and can be determined by performing an in vitro or in situ Zn²⁺ titration [12]. Another important characteristic of fluorescent sensors is the dynamic range, which is determined by the fluorescence signal in the Zn²⁺-bound state versus the fluorescence signal in the Zn²⁺-unbound state. In general, a sensor with a high dynamic range can provide both sensitivity and accuracy for estimating Zn²⁺ levels [13]. Ultimately, selecting a sensor will depend on the intended application. For instance, genetically encoded Zn²⁺ sensors can be selectively targeted to subcellular compartments as a means to measure Zn²⁺ within organelles or detect Zn²⁺ release at the cell surface [14]. Table 1 shows the in situ dynamic range, K_d’ and targeted cellular compartments of the current repertoire of genetically encoded Zn²⁺ sensors. In contrast, small-molecule sensors are difficult to direct to a particular compartment, but they often offer a large dynamic range making them particularly sensitive to Zn²⁺ dynamics. Hybrid, or chemigenetic, sensors are a promising new set of tools that leverage small-molecule Zn²⁺ sensors and take advantage of a genetically encoded component to target organelles [15]. In this review, we cover recent advances in the development and application of fluorescent Zn²⁺ sensors. We also refer readers to excellent reviews that cover earlier work in the field [4, 14, 16, 17].

Table 1.

Genetically-encoded Zn²⁺ sensors

Sensor	Sensor Type a	DR b-d	Kd’ e	Targeted probes f	Ref
ZapCY1	FRET (ECFP/Cit)	2.66 b	17 pM	Cytosol, ER, Golgi, MM	[5, 13, 21]
ZapCmR1.1	FRET (Clover/mR2)	1.5 b	ND	Cytosol, nucleus	[127]
ZapCY2	FRET (ECFP/Cit)	1.4 b	811 pM	Cytosol	[5]
ZapCV2	FRET (ECFP/cpV)	2.1 b	5.3 nM	Cytosol	[22, 119]
ZapCmR2	FRET (Clover/mR2)	1.4 b	ND	Cytosol, nucleus	[127]
ZapCV5	FRET (ECFP/Cit)	1.5 b	300 nM	Cytosol	[22]
eCALWY-1	FRET (Cer/Cit)	1.6 b	2 pM	Cytosol, SV	[7]
redCALWY-1	FRET (mOr/mCh)	1.1 c	12.3 pM	Cytosol	[128]
eCALWY-2	FRET (Cer/Cit)	1.6 b	9 pM	Cytosol	[7]
eCALWY-3	FRET (Cer/Cit)	1.6 b	45 pM	Cytosol	[7]
eCALWY-4	FRET (Cer/Cit)	1.83 b	164 pM	Cytosol, SV, MM, ER	[7, 19, 21]
redCALWY-4	FRET (mOr/mCh)	1.2 c	234 pM	Cytosol	[128]
eCALWY-5	FRET (Cer/Cit)	1.3 b	1.8 nM	Cytosol	[7]
eCALWY-6	FRET (Cer/Cit)	1.3 b	2.9 nM	Cytosol, ER	[7, 19]
eZinCh-1	FRET (Cer/Cit)	ND	8.2 μM	SV	[7]
eZinCh-2	FRET (Cer/Cit)	2.06 b	103 pM	Cytosol, ER, MM, SV	[20, 21]
ZnT72R	FRET (mNG/mR2)	1.4 c	56 μM	Cytosol, SV	[23]
BLCALWY-1	FRET/BRET (NL/Cer/Cit)	ND	4.2 pM		[27]
BLZinCh-1	FRET/BRET (NL/Cer/Cit)	1.25 c	160 pM	Cytosol	[27]
BLZinCh-2	FRET/BRET (NL/Cer/Cit)	1.29 c	117 pM	Cytosol	[27]
BLZinCh-3	FRET/BRET (NL/Cer/Cit)	1.67 c	15.6 pM	Cytosol	[27]
GZnP1	Single FP (cpGFP)	2.6 b	58 pM	Cytosol, MM, PM	[31]
GZnP2	Single FP (cpGFP)	4.5 b	352 pM		[32]
mCherry-GZnP2	Ratiometric (cpGFP/mCh)	ND	ND	Cytosol, IMS, MM	[32]
GZnP3	Single FP (cpGFP)	11 b	1.3 nM	Cytosol, SV, endosome, lysosome	[32]
ZnGreen1	Single FP (mTFP1)	2.5 c	633 nM	Cytosol, cell surface	[34]
ZnGreen2	Single FP (cpTFP1)	ND	20 μM		[34]
ZnRed	Single FP (mApple)	2.5 c	166 nM 20 μM	Cytosol, nucleus	[34]
ZIBG1	Single FP (GFP)	1.6 d	2.81 μM	Cell surface	[36]
ZIBG2	Single FP (cpGFP)	1.8 d	282 nM	Cell surface	[36]

^aFluorescent proteins: Cit = Citrine; mR2 = mRuby2; cpV = cpVenus173; Cer = Cerulean; mOr = mOrange; mCh = mCherry; mNG = mNeonGreen; NL = NanoLuc

^bReported in situ dynamic range (DR) in cytosol

^cEstimated in situ DR in cytosol

^dEstimated in situ DR at cell surface

^eIn vitro K_d’ determined at pH 7.4

^fMM = mitochondrial matrix; IMS = mitochondrial intermembrane space; SV = secretory vesicle

GENETICALLY ENCODED Zn²⁺ SENSORS

Genetically encoded Zn²⁺ sensors tether fluorescent proteins to peptides derived from naturally-occurring Zn²⁺-binding proteins. The metal-sensing moiety tunes the sensor fluorescence upon Zn²⁺ binding. Single fluorescent protein sensors are generally intensiometric because Zn²⁺ binding alters fluorescence intensity of the sensor. In contrast, sensors that contain two fluorescent proteins often exploit Forster resonance energy transfer (FRET), in which Zn²⁺ binding alters the orientation and/or distance between a donor and acceptor fluorescent protein, thereby altering the FRET efficiency. FRET-based Zn²⁺ sensors are ratiometric because they use the ratio of acceptor and donor fluorescence to measure changes in Zn²⁺. One major benefit of ratiometric sensors is that they avoid potential cellular artefacts, such as differences in sample thickness, cellular movement and protein concentration because measurements are independent of sensor concentration [12]. Figure 1 illustrates the distinct mechanisms of different sensor platforms.

Figure 1: Structure and mechanism of fluorescent Zn²⁺ sensor platforms.

A) The FRET sensors ZapCY and ZnT72R and dual BRET/FRET sensors BLZinCh and BLCALWY are shown. The Zn²⁺-sensing mechanisms of BLZinCh and BLCALWY are identical to the FRET sensors eZinCh and eCALWY, respectively. The dual BRET/FRET sensors BLZinCh and BLCALWY contain NanoLuc (NL), whereas eZinCh and eCALWY do not. B) The single fluorescent protein (FP)-based Zn²⁺ sensors GZnP (top example), ZnGreen1 and ZIBG1 couple ZF1 and ZF2 derived from Zap1 with cpGFP, mTFP1 and GFP, respectively. In response to Zn²⁺, GZnP1 and ZIBG1 are fluorescence turn-on and ZnGreen1 is fluorescence turn-off. The single FP-based Zn²⁺ sensors ZnGreen2 and ZIBG2 (middle example) couple Zn²⁺ hook peptides derived from Rad50 with cpTFP1 and cpGFP, respectively. In response to Zn²⁺, ZIBG2 is fluorescence turn-on and ZnGreen2 is fluorescence turn-off. ZnRed (bottom example) couples Zap1’s ZF1 and ZF2 with mApple and is fluorescence turn-on. C) The ratiometric Zn²⁺ sensor mCherry-GZnP2 couples mCherry with GZnP2. GZnP2 fluorescence is directly modulated by Zn²⁺, and mCherry fluorescence is independent of Zn²⁺ changes. D) A chemigenetic Zn²⁺ sensor based on addition of a chloroalkane moiety (yellow box) to the small-molecule Zn²⁺ sensor ZP1 (ZP1-12Cl). ZP1-12Cl can interact with HaloTag (ZP1-HaloTag) or HaloTag-mCherry (ZP1-HaloTag-mCherry) to create targetable intensiometric or ratiometric versions, respectively. E) Multiplexing a Zn²⁺-dependent DNAzyme with genetically encoded fluorescent proteins was achieved by selectively-targeting mRNA encoding mClover2. In the presence of Zn²⁺, the Zn²⁺ DNAzyme binds and degrades mClover2 mRNA, but not mRNA encoding mRuby2, leading to an increased mRuby2/mClover2 ratio with increasing concentrations of Zn²⁺.

FRET-based Zn²⁺ sensors

The first genetically encoded Zn²⁺ sensors developed were FRET-based. Early work in the field led to the development of three families of FRET-based Zn²⁺ sensors (ZapCY, eCALWY, eZinCh), each with a unique Zn²⁺-sensing mechanism (Figure 1) [5, 7, 18]. All three sensor platforms have been used to measure labile Zn²⁺ levels in the cytosol and within organelles, including the ER and mitochondria [5-7, 13, 19, 20]. The cytosolic sensors are robust and have a high dynamic range (> 2-fold), and all three sensors report cytosolic Zn²⁺ concentrations in the low hundreds of picomolar range for a variety of cell types (Table 1). Because small-molecule sensors often exist in multiple cellular compartments (see below), genetically encoded sensors are the platform of choice for cytosolic measurements. However, application of these platforms to organelles has led to inconsistences in reported concentrations, highlighting the need for rigorous characterization and a wider range of organelle-targeted options. For example, early estimations of labile Zn²⁺ levels in the ER were highly-variable, with ER-ZapCY1 reporting that labile Zn²⁺ in the ER is lower than the cytosol [5] and ER-eCALWY-4 reporting that labile Zn²⁺ in the ER is higher than the cytosol [7]. In general, FRET-based sensors suffer from a lower dynamic range than intensiometric sensors, and this problem is enhanced when targeted to organelles [5, 13, 19, 20]. In a recent study, Carter et al optimized the conditions used for performing in situ sensor calibrations and in situ Zn²⁺ titrations to determine the dynamic range and K_d’, respectively [21]. Using optimized conditions, they found that the dynamic range of ZapCY1 far exceeded that of both eCALWY-4 and eZinCh-2 in both the cytosol and ER (Table 1). Further, these optimized calibration conditions revealed that both ER-ZapCY1 and ER-eCALWY-4 report that labile Zn²⁺ is lower in the ER than the cytosol.

The study by Carter et al also sought to better understand variability in ER-targeted measurements using a microfluidic platform to assess the functionality of sensors in a large population of cells [22]. For all three sensor platforms the distribution of FRET ratios was more heterogeneous in the ER than the cytosol, suggesting that the ER environment consistently altered sensor function. A side-by-side comparison of sensor localization by fluorescence microscopy revealed that while ER-ZapCY1 and ER-eCALWY-4 reliably localize to the ER, 25% of the cells expressing ER-eZinCh-2 form bright puncta, suggesting significant distortion of ER structure, and likely function. Furthermore, ER-eZinCh-2 exhibited the greatest propensity to form higher-order oligomers, which may negatively impact sensor function. Overall, this study emphasized the importance of carefully optimizing FRET sensor calibration conditions and establishing metrics for FRET sensor performance in organelles. It also revealed the need for more diverse platforms targeted to the ER, as even the best sensor (ER-ZapCY1) showed some perturbation in the ER environment.

A new platform for FRET-based Zn²⁺ sensing was recently reported that incorporates a Zn²⁺-binding RING motif from the protein TRIM72 that is flanked by the green and red fluorescent proteins mNeonGreen and mRuby2, respectively (Figure 1) [23]. The newly-developed sensor, referred to as ZnT72R, has a modest in situ dynamic range (Table 1). The K_d’ is 7.8 μM (pH 7.4) when measured in vitro; however, the in situ K_d’ is 56 μM, suggesting that perhaps in situ Zn²⁺ titration conditions may need to be further optimized. Researchers are cautioned against using the sensor for rigorous quantification of Zn²⁺ as low dynamic range sensors have been shown to lead to inaccurate overestimates of Zn²⁺ concentrations [13]. However, the sensor could be useful for determining relative levels and monitoring increases and decreases in Zn²⁺ in the micromolar range. The micromolar affinity of ZnT72R was ideal for detecting Zn²⁺ in the Zn²⁺-rich secretory granules of pancreatic β-cells [24]. To accomplish this, the authors fused ZnT72R with the granular cargo protein neuropeptide Y (NPY) to generate NPY-ZnT72R. By calibrating NPY-ZnT72R, the authors found that granular Zn²⁺ levels were significantly reduced in diabetic mice (28 μM) compared with healthy controls (41 μM). Although this result disagrees in magnitude with Merkx and colleagues who used VAMP2-eZinCh-2 and found that granular Zn²⁺ levels are ~120 nM in INS-1 cells [20], both studies suggest that insulin granules contain higher levels of Zn²⁺ than the cytosol and hence represent a pool of labile Zn²⁺. This study further illustrates the need for additional development of organelle-targeted sensors with high dynamic range and a range of Zn²⁺ binding affinities.

BRET-based Zn²⁺ sensors

Bioluminescence resonance energy transfer (BRET)-based sensors are a close relative of FRET-based sensors. BRET-based sensors share the same design principle as FRET sensors with one key difference. Rather than incorporating a fluorescent protein as the donor, BRET-based sensors use a donor luciferase that produces light from a chemical reaction. This bioluminescence then excites an acceptor fluorescent protein (Figure 1). BRET-based sensors require a chemical substrate, such as furimazine in the case of the luciferase NanoLuc [25]. Replacing the donor fluorescent protein with a luciferase affords BRET-based sensors the advantage of reduced phototoxicity, autofluorescence and light scattering [26]. Merkx and colleagues generated a handful of BRET-based Zn²⁺ sensors based on their eCALWY and eZinCh platforms [27]. However, rather than removing the donor fluorescent protein, the authors chose to fuse the bright luciferase NanoLuc to their full-length FRET sensor platforms; thereby, generating dual readout BRET/FRET sensors (Figure 1). The sensors BLCALWY-1, BLZinCh-1 and BLZinCh-2 demonstrated a low in situ and in vitro BRET dynamic range (Table 1) compared with the in vitro FRET dynamic range of ~1.5. Interestingly, the authors demonstrated that by incorporating a chromophore-silencing mutation into the donor fluorescent protein (BLZinCh-3), which prevents FRET but maintains BRET from NanoLuc to the acceptor fluorescent protein, the in situ and in vitro BRET dynamic range is improved (Table 1). This suggests that the donor fluorescent protein could have been replaced with the luciferase NanoLuc. Using two different bioluminescent imaging platforms (plate reader and microscopy), the authors were able to measure cytosolic Zn²⁺ using BLZinCh-3. When directly compared with fluorescence measurements, the bioluminescent measurements had lower background and were less heterogeneous. With a K_d’ of 15.6 pM, BLZinCh-3 was saturated in the cytosol but could be used in the future to detect Zn²⁺ in organelles, such as the ER and mitochondria.

Single fluorescent protein-based Zn²⁺ sensors

Single fluorescent protein-based sensors for Ca²⁺, including GCaMP and GECO families, have been around for many years [28, 29], but sensors for Zn²⁺ are relatively new. Figure 1 illustrates the general design principle, which involves attachment of metal-binding domains to the N- and C-termini of a fluorescent protein that is often circularly permuted. Single fluorescent protein-based Ca²⁺ sensors have shown that metal-binding alters the chromophore environment of circular permuted GFP (cpGFP) and changes sensor brightness [30]. Since single fluorescent-protein based sensors are usually not ratiometric, FRET-based sensors are preferred for quantitative measurements. However, single fluorescent protein-based sensors offer several other advantages, including greater dynamic range and the ability to be multiplexed with other fluorescent sensors.

The first single fluorescent protein-based Zn²⁺ sensor (GZnP) was engineered by attaching two Zn²⁺ fingers (ZF1 and ZF2) derived from the yeast transcription factor Zap1 onto the N- and C-termini of cpGFP (Figure 1) [31]. Rational protein engineering guided by computational modeling was used to develop a sensor library consisting of 43 sensor variants. The library was screened for both baseline intensity and dynamic range. GZnP1 was identified as a fluorescent turn-on sensor with an in situ dynamic range comparable with FRET-based sensors that use the same Zn²⁺-binding domain (Table 1) [5]. Furthermore, the dynamic range remained stable when targeted to the plasma membrane and mitochondria.

Qin and colleagues developed a second-generation sensor called GZnP2 with an improved in situ dynamic range (Table 1). GZnP2 was identified by screening a library of GZnP1 variants with mutations in the linker regions connecting ZF1 and ZF2 with cpGFP [32]. GZnP2 has a lower affinity than GZnP1 (Table 1). In order to use GZnP2 to quantify cellular Zn²⁺ levels, the authors attached the spectrally-distinct fluorescent protein mCherry to generate a ratiometric sensor (Figure 1). The authors subsequently targeted mCherry-GZnP2 to both the mitochondrial matrix and mitochondrial intermembrane space (IMS) of four cell types. Overall, they found that while matrix Zn²⁺ levels are less than 1 pM, as previously reported [13], Zn²⁺ levels in the cytosol (75-115 pM) and IMS (65-100 pM) are comparable. A third-generation Zn²⁺ sensor GZnP3 was recently developed with an even better in situ dynamic range (Table 1) [33]. Its nanomolar-affinity makes GZnP3 unsuitable for Zn²⁺ measurements in the cytosol and many organelles. However, the authors reasoned that GZnP3 could instead be used to detect local Zn²⁺ elevations in the cytosol in response to activation of the non-selective cation channel TRPML. Indeed, when targeted to the cytosolic face of lysosomes and endosomes of primary cultured rat hippocampal neurons, GZnP3 detected Zn²⁺ release in response to TRPML activation. A promising new direction will be to convert GZnP3 into a ratiometric sensor, like GZnP2, and quantify labile Zn²⁺ levels near sites of TRPML-mediated Zn²⁺ release.

Another family of single fluorescent protein-based Zn²⁺ sensors was developed by Ai and colleagues [34]. ZnGreen1 and ZnRed were created by inserting ZF1 and ZF2 of Zap1 into the fluorescent proteins mTFP1 and mApple, respectively (Figure 1). A third Zn²⁺ sensor called ZnGreen2 was also developed by attaching two Zn²⁺ hook peptides derived from the Pyrococcus furiosus protein Rad50 to the termini of cpTFP1 (Figure 1). Thus, while ZnGreen1 and ZnRed share a Zn²⁺-binding domain with GZnP, the sensor design of ZnGreen2 more closely-matches GZnP. While the in vitro dynamic range of these sensors is high (ZnGreen1 = 26.3, ZnGreen = 8.7, ZnRed = 3.8), the in situ dynamic range is much lower than GZnP3 (Table 1). Furthermore, their low Zn²⁺ affinity suggests that all three sensors may be suboptimal for most cellular Zn²⁺ measurements (Table 1). Like the low-affinity sensor GZnP3 though, they may be applicable for measuring high local Zn²⁺ fluctuations. Zn²⁺ is enriched in secretory granules of pancreatic β-cells [24] and is released in response to glucose stimulation [35]. In this study, the authors demonstrate that when targeted to the outer-leaflet of the plasma membrane using pDisplay, ZnGreen1 detects a rapid release of Zn²⁺ in response to glucose stimulation in INS-1 cells.

Ai and colleagues recently engineered two new single fluorescent protein-based Zn²⁺ sensors called ZIBG1 and ZIBG2 [36]. The overall design of ZIBG1 and ZIBG2 resemble ZnGreen1 and ZnGreen2, respectively; however, the fluorescent proteins mTFP1/cpTFP1 have been replaced with GFP/cpGFP from G-GECO1 (Figure 1). Like their predecessors, ZIBG1 and ZIBG2 are low-affinity Zn²⁺ sensors but exhibit lower in situ dynamic range (Table 1). Despite this, the authors performed dual-color imaging of MIN6 cells expressing pDisplay-ZIBG2 and the red cytosolic Ca²⁺ indicator R-GECO1 and simultaneously detected a glucose-induced release of Zn²⁺ and transient rise in cytosolic Ca²⁺ levels. Importantly, they also showed that pDisplay-ZIBG2 can detect Zn²⁺ release from mouse and human primary pancreatic β-cells in response to different stimulations.

SMALL-MOLECULE Zn²⁺ SENSORS

The small-molecule Zn²⁺ sensor 6-methoxy-8-p-toluenesulfonamido-quinoline (TSQ) was developed over thirty years ago to detect Zn²⁺ in brain tissue slices [37]. Since that time, many small-molecule Zn²⁺ sensors have been developed that exhibit a wide-range of properties [14]. While there are few examples of small-molecule Zn²⁺ sensors being used for rigorous Zn²⁺ quantification, they have been used to detect Zn²⁺ dynamics [38-40], identify transporters responsible for Zn²⁺ regulation [24, 41, 42], and compare healthy and diseased cells [43]. The general design principle is an organic fluorophore coupled with an electron-rich Zn²⁺-chelating group, such as dipicolylamine (DPA). While the majority of small-molecule Zn²⁺ sensors include the fluorophore fluorescein, including the ZinPyr (ZP) family [44-50], other sensors use coumarin [51], BODIPY [52], bensoresorufrin [38], rhodamine [53] or rhodol [54]. In the absence of Zn²⁺, the fluorophore moiety is quenched by the metal-binding group through photoinduced electron transfer (PET). PET is an excited state process that involves transfer of an electron from a donor to an acceptor. Upon Zn²⁺ binding, the PET pathway is disrupted, leading to fluorescence turn-on (Figure 1). Since most small-molecule Zn²⁺ sensors operate through this mechanism, they are intensiometric. Owing to their brightness in the Zn²⁺-bound state and low signal in the Zn²⁺-unbound state, small-molecule Zn²⁺ sensors generally have a much greater dynamic range than genetically encoded Zn²⁺ sensors. For example, the small-molecule Zn²⁺ sensor FluoZin-3 has a reported dynamic range of 200 [50] compared with the genetically encoded Zn²⁺ sensor ZnGreen, which has an in vitro dynamic range of 26.3 [34].

While the cellular expression of genetically encoded Zn²⁺ sensors can be finely-tuned, it remains a challenge to limit and quantify the concentration of small-molecule sensors inside cells. Two major concerns with elevated probe concentrations are perturbation of endogenous Zn²⁺ levels and accessing Zn²⁺ outside of the labile pool [55]. Qin et al reported that labile Zn²⁺ levels decrease with increasing concentrations of the small-molecule Zn²⁺ sensor FluoZin-3, but not the genetically encoded Zn²⁺ sensor ZapCY2 [6], suggesting that excess probe indeed exceeds the Zn²⁺ buffering capacity of the cell. There is also convincing evidence that the small-molecule Zn²⁺ sensors TSQ, Zinquin and FluoZin-3 detect Zn²⁺ that is associated with low-molecular-weight ligands and proteins, rather than labile Zn²⁺ [56-58]. Thus, it would appear advantageous to use lower probe concentrations. Lippard and colleagues used acetylated versions of the Zn²⁺ sensors ZBR1 (Ac-ZBR1) and ZBR3 (Ac-ZBR3) at 40- to 50-fold lower concentrations than the unacetylated parent sensors [59], and an acetylated version of ZP1 (DA-ZP1) in brain tissue slices at 10-fold lower concentrations than the parent sensor. Likewise, Que and colleagues developed a novel small-molecule Zn²⁺ sensor called ZincBY-1 that could detect Zn²⁺ fluxes in mammalian eggs at nanomolar probe concentrations [60]. However, it is important to keep in mind that the intracellular concentration of sensor is what matters when it comes to minimizing perturbation of Zn²⁺, not the concentration applied to cells. Caution should be used with acetylated sensors because they can accumulate within cells at a concentration 100-times higher than the concentration applied to cells [55].

Subcellular targeting of small-molecule Zn²⁺ sensors

Another general concern with the cellular application of small-molecule Zn²⁺ sensors is spontaneous localization to organelles. For example, both FluoZin-3 and ZP1 spontaneously localize to the Golgi among other cellular compartments [6, 61]. In some cases, an unpredicted organelle localization can be leveraged to provide Zn²⁺ measurements in organelles. For example, the benzoresorufrin-based family of Zn²⁺ probes ZBR1-3 spontaneously localized to the ER [38, 59] and could be used to detect ER Zn²⁺ dynamics . Furthermore, Han et al demonstrated that SpiroZin-2, which was originally reported to accumulate in acidic cellular compartments [41], can be used to track lysosomal Zn²⁺ [40]. Exploiting the overall charge of a molecule is a more direct approach for targeting small-molecule Zn²⁺ sensors to organelles. For example, addition of a morpholine moiety, which remains neutral in the cytosol but becomes positively-charged and trapped in the lysosome, can be used to generate a lysosomal Zn²⁺ sensor [62]. In an attempt to generate a mitochondria-localized Zn²⁺ sensor, Lippard and colleagues functionalized the Zn²⁺ sensor ZP1 with the lipophilic cation triphenylphosphine (TPP), which utilizes the negative mitochondrial membrane potential to be imported into the mitochondria [43]. Surprisingly, ZP1-TPP instead accumulated in endosomes and lysosomes, suggesting that ZP1-TPP was not sufficiently charged for this approach to be successful. To neutralize the negatively-charged carboxyl group on fluorescein, the authors generated a diacetylated sensor called DA-ZP1-TPP. DA-ZP1-TPP not only localized to the mitochondria but exhibited an in vitro dynamic range of 140, which far exceeds its predecessor [63].

Addition of hydrophobic alkyl side chains to a sensor scaffold is a proven approach to non-selectively target small-molecule Zn²⁺ sensors to the cell surface. Lippard and colleagues functionalized the Zn²⁺ sensors ZP1 and Zinquin with N-terminal palmitoyl groups to generate plasma membrane-localized versions (Palm-ZP1 and Palm-ZQ) [64]. Both Palm-ZP1 and Palm-ZQ were responsive to extracellular Zn²⁺, suggesting that both sensors were present on the cell surface. Similarly, Li and colleagues developed a plasma membrane-localized Zn²⁺ sensor (ZIMIR) by incorporating two dodecyl side chains [39]. Using ZIMIR, the authors could detect Zn²⁺ that is co-released with insulin from pancreatic β-cells. However, they observed that the ZIMIR probe was gradually internalized by cells during long incubations. To overcome this limitation and selectively label pancreatic β-cells, Li et al substituted the alkyl side chains of ZIMIR with Ex4(9-39), a truncated ligand of a pancreatic β-cell-specific cell surface receptor [65]. The newly-developed sensor (ZIMIR-Ex4) remained on the pancreatic β-cell surface over long incubations, suggesting that ZIMIR-Ex4 may be a more reliable indicator of Zn²⁺ release.

Chemigenetic Zn²⁺ sensors

Chemigenetic sensors combine small-molecule fluorescent sensors with a genetically encoded component, facilitating localization to specific compartments. For example, HaloTag consists of a genetically encodable modified version of the enzyme haloalkane dehydrogenase that covalently interacts with haloalkane moieties [66]. In order to promote interaction between HaloTag and a small-molecule sensor, the sensor can be functionalized with a chloroalkane ligand (Figure 1) [67]. ZIMIR-HaloTag was the first demonstration of the HaloTag platform being used in conjunction with a small-molecule Zn²⁺ sensor [68]. In pancreatic β-cells expressing a plasma membrane-targeted version of HaloTag, ZIMIR-HaloTag accumulated on the cell surface. Using ZIMIR-HaloTag and the genetically encoded cytosolic Ca²⁺ sensor R-GECO, the authors were also able to simultaneously monitor Zn²⁺ release and cytosolic Ca²⁺ dynamics in pancreatic β-cells following membrane depolarization.

Zastrow et al developed a chloroalkane-modified version of ZP1 (ZP1-12Cl) with a binding-affinity appropriate for detecting cytosolic and nuclear Zn²⁺ (K_d’ = 340 pM) [15]. The authors targeted ZP1-HaloTag to the nucleus and cytosolic face of the ER and mitochondria, and were able to detect nitric oxide (NO)-induced Zn²⁺ dynamics at all three cellular locations. In the absence of HaloTag expression, ZP1-12Cl spontaneously localized to the Golgi and did not display any fluorescent change in response to NO. To generate a ratiometric chemigenetic sensor, Zastrow et al attached the fluorescent protein mCherry to the N-terminus of the HaloTag (ZP1-HaloTag-mCherry) (Figure 1). Like ZP1-HaloTag, ZP1-HaloTag-mCherry detected NO-induced Zn²⁺ dynamics in the nucleus. ZP1-12Cl exhibited an in vitro dynamic range of 6, while HaloTag- or HaloTag-mCherry-bound ZP1-12Cl demonstrated an in vitro dynamic range of 2.7 and 1.9, respectively. The in situ dynamic range of ZP1-HaloTag and ZP1-HaloTag-mCherry was 1.7-1.8 and 1.6-2.1, which is comparable with most genetically-encoded Zn²⁺ sensors.

DNAzyme-based Zn²⁺ sensors

DNAzymes are catalytic DNA molecules capable of performing a chemical reaction. DNAzymes have recently been exploited as metal ion sensors by combining a DNAzyme that detects a metal ion with organic fluorophores or fluorescent proteins. In this platform, the DNA molecule bind specific RNA sequences and cleaves the RNA in a metal-dependent fashion. Hwang and colleagues pioneered the development of DNAzyme-based fluorescent sensors of several metal ions, including Mg²⁺ and Zn²⁺ [69]. While most DNAzymes incorporate organic fluorophores, Xiong et al recently developed a novel system that multiplexes both Mg²⁺ and Zn²⁺ DNAzymes with genetically encoded fluorescent proteins [70] (Figure 1). The DNAzymes are designed so that they specifically cleave mRNA encoding the green fluorescent protein mClover2 but not mRNA encoding the red fluorescent protein mRuby2 (Figure 1). Thus, increased metal ion levels will lead to decreased mClover2 expression and fluorescent signal, without affecting the level of mRuby2 expression and signal. To test the multiplexed system, the authors transfected live cells with either a Mg²⁺ or Zn²⁺ DNAzyme and plasmids encoding mClover2 and mRuby2 for 6 hours, treated cells with 20-40 mM Mg²⁺ or 30 μM Zn²⁺ for an additional 18 hours and quantified the mRuby2/mClover2 ratio using fluorescence microscopy or flow cytometry. They found that the mClover2 fluorescent signal decreased with increasing concentrations of Mg²⁺ or Zn²⁺ while the mRuby2 signal remained relatively constant, leading to an increased mRuby2/mClover2 ratio.

TOOLS TO MANIPULATE CELLULAR Zn²⁺ STATUS

The tools for sensing labile Zn²⁺ described above permit measurement of Zn²⁺ in a variety of organismal, cellular, and subcellular systems. Equally important are molecules and methods for manipulating Zn²⁺, and these manipulation tools can be used with fluorescent sensors to provide insight into zinc biology. Metal chelators have long been used to control and perturb metal concentrations in media and cells. Chelex 100 resin is a valuable resource for removing transition metals from serum or media through binding transition metals and subsequent filtration of chelex out of solution. One disadvantage is that Chelex 100 has poor metal selectivity, removing significant amounts of iron, copper, magnesium, calcium, and zinc [71]. In studying specific metal ions, it is important to remember that metal chelators have a wide range of sensitivities and specificities for target cations, and inductively coupled plasma mass spectrometry (ICP-MS) studies are useful to define elemental metal composition of experimental media before and after chelation. At times supplementation will be necessary to ensure that only the metal of interest is being manipulated. A promising new Zn²⁺-selective chelator, A12-resin, has been shown to remove 99% of Zn²⁺ from media in a metal-specific manner by using the purified S100A12 Zn²⁺-scavenging protein immobilized on an agarose support. The resin may be washed and re-used up to four times while maintaining its Zn²⁺-binding capability [72]. The A12 resin has since been used to develop an assay for Zn²⁺ uptake into cells by selective Zn²⁺ depletion of cell media and resupply of stable isotope ⁷⁰Zn(II) [73]. A12 resin will also be useful for experiments where Zn²⁺ depletion is desired while preserving the integrity of culture media.

Intracellular Zn²⁺ chelators are small molecules that enter cells and bind labile Zn²⁺ with high affinity and specificity. They are used to induce Zn²⁺-deficient states in cells (e.g. in situ calibrations of fluorescent Zn²⁺ sensors) and to ensure that residual Zn²⁺ in the buffer is sequestered. To accomplish this task, much of the field has turned to N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN). TPEN is a powerful Zn²⁺ chelator because it has an K_d’ for Zn²⁺ of <1 fM [74]. TPEN has been widely used to lower cellular Zn²⁺; however, its high affinity is both a strength and weakness. While it readily binds available Zn²⁺ in a cell, TPEN has also been shown to remove Zn²⁺ bound to the proteome [56] and cause cell death independent of its role in Zn²⁺ chelation [75, 76]. Furthermore, when used in conjunction with genetically encoded fluorescent Zn²⁺ sensors, TPEN demonstrates relatively slow kinetics of Zn²⁺ removal from biosensors, which makes sensor calibration difficult [5, 13]. Fortunately, tris(2-pyridylmethyl)amine (TPA) has been demonstrated as an effective Zn²⁺ chelator with lower Zn²⁺ affinity, lower cytotoxicity, and faster kinetics than TPEN [77]. While TPEN is still widely used within the community, TPA may be a more valuable chelator for manipulating Zn²⁺ in biological contexts.

While intracellular Zn²⁺ chelators are used to manipulate cellular labile Zn²⁺, extracellular chelators have proven particularly useful for studying autocrine and paracrine Zn²⁺ signaling. Many early extracellular Zn²⁺ chelators were non-specific for divalent metal cations, making it difficult to study Zn²⁺ signals in systems that also experience fluctuations in other cations [78]. For example, CaEDTA has slow Zn²⁺ binding kinetics, whereas tricine has a low Zn²⁺ affinity with a K_D equal to 10 μM, which demands the use of high concentrations of chelator to effectively chelate Zn²⁺ [79]. Extracellular Zn²⁺ chelators DPESA and TPESA were developed using the TPEN backbone. Their binding affinities are tighter than tricine, but kinetics are still relatively slow, making them less practical for neuronal applications, where action potentials occur in a fraction of a second [80]. The most recent extracellular Zn²⁺ chelator, ZX1, was designed for neuronal applications, with the ability to chelate Zn²⁺ with reasonably fast kinetics (0.027 s⁻¹, ~10-fold faster than CaEDTA) [81], higher affinity (K_D = 1.0 nM), and selectivity for Zn²⁺ [82]. ZX1 has been used to investigate the role of Zn²⁺ in neurotransmitter receptor modulation [83, 84], synaptic plasticity [82, 85], and higher level processes such as auditory modulation [86]. For a more detailed look at Zn²⁺ chelators, Radford and Lippard have an excellent review [87].

Finally, a commonly used tool for manipulating Zn²⁺ in cells is to use the ionophore pyrithione to facilitate transport of Zn²⁺ across cellular membranes. While this is an effective approach for elevating Zn²⁺, it is important to note that pyrithione is not an innocent ligand. Pyrithione can perturb copper and iron homeostasis [88] and activate stress and cell death pathways [89, 90].

TOOLS FOR PROBING OCCUPANCY OF Zn²⁺-BINDING PROTEINS IN CELLS

The current suite of fluorescent Zn²⁺ sensors quantifies labile Zn²⁺ and detects Zn²⁺ dynamics in the cytosol and organelles (see above), but there are still few tools to study protein-bound Zn²⁺ in live cells. Such tools would be useful for studying how proteins acquire Zn²⁺ ions and whether fluctuations in labile Zn²⁺ cause changes in metalation of the Zn²⁺ binding proteome. The first step to develop sensors of Zn²⁺-bound proteins is finding or creating molecules that can recognize the difference between the apo and bound form of a protein. Medicinal chemistry seeks to do this with probes that bind the Zn²⁺-bound active site of a variety of enzymes including carbonic anhydrase, alcohol dehydrogenase, and matrix metalloproteinases [91]. An alternate option is to design probes that react with the metal-free state of the protein, which has been demonstrated by the Weerapana group to detect changes in Zn²⁺-bound cysteines via proteomics (see below) [92].

Recent work from the Que lab has demonstrated the utility of small molecule binders to fluorescently tag Zn²⁺-bound proteins within a cell [93]. The Zn²⁺-bound holoenzyme form of carbonic anhydrase (CA) was detected using a sulfonamide group that recognizes the Zn²⁺-bound active site [94] linked to the hydrophobic-dequenching dye N,N-dimethylaminopthalimide (DMAP) [95]. Binding of the sensor to Zn²⁺-bound holo-CA leads to a shift in sensor environment to more hydrophobic and dequenching of DMAP fluorescence, leading to 12-fold fluorescence induction. The dye can detect Zn²⁺-bound carbonic anhydrase in cell lysate via native SDS-PAGE in-gel fluorescence and in live cells using fluorescence microscopy. The authors further demonstrate that probe fluorescence in live cells is insensitive to both Zn²⁺ chelation by TPEN and supplementation with Zn²⁺ pyrithione, suggesting that carbonic anhydrase may be insensitive to changing Zn²⁺ levels of this magnitude or that the probe prevents Zn²⁺ exchange [93]. This study demonstrates the potential for metalation probes in live cell studies, and the prospective use of this and future tools to study how proteins acquire their metal cofactors and whether they are sensitive to fluctuations of labile metal concentrations.

PROTEOMICS FOR INTERROGATING THE Zn²⁺ PROTEOME

In addition to the fluorescent tools described above, proteomics has become a valuable technique for probing zinc’s role inside the cell. Quantitative proteomics, either through the addition of chemical (activity) based probes or using protein abundance, allows researchers to gain valuable insight into how the Zn²⁺ proteome reacts to perturbations of the Zn²⁺ pool. Additionally, chemical proteomics can be used to identify novel Zn²⁺ binding proteins and to identify cellular compartments where Zn²⁺ may be stored for times of need.

Probing cysteine-reactivity of Zn²⁺ proteome

The Weerapana group was the first group to use chemical proteomics for studying the Zn²⁺ proteome [92]. Prior to this, most attempts to identify novel Zn²⁺ binding proteins relied on bioinformatic approaches to mine sequence or structure databases [96-98]. However, these techniques are limited to well-characterized Zn²⁺ binding motifs (as in the case of Zn²⁺ finger proteins) or high-resolution structures. Additionally, these techniques are more attuned to finding the most stable Zn²⁺-protein interactions rather than transient ones that are still important in protein regulation. Unbiased chemical proteomics approaches, therefore, provide a route to discover novel Zn²⁺ binding proteins that may be missed using the above approaches. Cysteine residues are excellent nucleophiles and can be alkylated by electrophilic probes such as iodoacetamide. Weerapana and coworkers used iodoacetamide and isotopic labeling to establish a hierarchy of cysteine reactivity of the whole proteome [99]. Cysteine reactivity is directly affected by its local environment and therefore it is probable that coordination by a metal such as Zn²⁺ would severely reduce its reactivity.

Using an in-gel fluorescence assay where either iodoacetamide or the peptide-based alkylating probe NJP14 were reacted with cell lysates and then conjugated with a fluorophore, they identified several protein bands which showed reduced NJP14 labeling, and therefore fluorescence, in high Zn²⁺. This was not seen with other metals (e.g. Mg²⁺, Ca²⁺, Mn²⁺) and the labeling was rescued with EDTA, suggesting that the labeled cysteines coordinate Zn²⁺. Streptavidin enrichment and quantitative proteomics of the NJP14 labeled peptides revealed that the proteins most sensitive to this particular probe were sorbitol dehydrogenase (SORD) and glutathione S-transferase omega-1 (GSTO1). SORD uses a catalytic Zn²⁺ ion that is coordinated with C44, H69, E70, and an activated water molecule to convert sorbitol to fructose. Treatment of the isolated protein with EDTA completely abolished the catalytic activity of the enzyme. Interestingly, treatment with NJP14 only reduced enzymatic activity by approximately 50%, suggesting that NJP14 cannot displace bound Zn²⁺ ions and only reacts with the Zn²⁺-free population of proteins. This was confirmed by pretreatment with excess Zn²⁺ to saturate all Zn²⁺ binding sites followed by labeling with NJP14, which saw no reduction in enzymatic activity. Like SORD, GSTO1 saw reduced NJP14 labeling in the presence of high Zn²⁺. However, its catalytic cysteine (C32) had not been shown to bind Zn²⁺. Mutation of this residue (C32A) completely abolished ability of GSTO1 to oxidize NADPH, suggesting that C32 is required for enzyme activity. Treatment of wild type GSTO1 with low concentrations of Zn²⁺ had no effect on enzymatic activity but at 50 and 100 μM Zn²⁺, GSTO1 activity is reduced to approximately 40% and 5% of baseline activity, respectively. This suggests that the activity of proteins like GSTO1 can be regulated in a Zn²⁺ dependent manner, though the precise mechanism is still unknown.

To more broadly characterize reactive cysteines that coordinate Zn²⁺, Pace and Weerapana revisited their quantitative chemical proteomics platform. Iodoacetamide-alkyne reacts promiscuously with thousands of cysteines in the proteome, and its alkyne moiety allows for click-chemistry attachment of either an isotopically labeled biotin linker (for proteomic enrichment) or a fluorophore (for gel-based assays) (Figure 2A). Isotopic labelling allows for quantification between two conditions (e.g. high Zn²⁺ vs. control, and EDTA vs. control) using the ratio of light:heavy peptide abundance. This screen identified more than 700 cysteines, some of which were known to bind Zn²⁺. Refining the data to those with light:heavy ratios > 1.5 upon EDTA treatment and light:heavy ratios < 0.66 upon Zn²⁺ treatment yielded a subset of cysteines (48 in total) that bind Zn²⁺. Some of these were known Zn²⁺ binding proteins, such as alcohol dehydrogenase 5 (ADH5), ribosomal subunits (RPS27, RPS3, RPS11, RPL23), and tubulin (TUBB2c, TUBA4a). Others had not yet been shown to have any Zn²⁺ related function, such as the serine/threonine kinase Nek9 and the histone chaperone ASF1B, showing that this is an unbiased platform for uncovering novel Zn²⁺-protein interactions.

Figure 2: Proteomic techniques to profile Zn²⁺ biochemistry of proteins.

A) Alkylating probes, such as iodoacetamide and N-ethylmaleimide, react with free thiols but not those complexed with Zn²⁺. Protein samples are then subjected to gel-based and/or proteomic analyses. B) AIZin permeates cell membranes and its electrophilicity is enhanced upon Zn²⁺ binding. When in proximity to a nucleophilic amino acid, the nucleophile attacks AIZin and the acyl group is transferred to the protein along with its fluorescein cargo. Researchers can then perform subsequent enrichment (via immunoprecipitation) and proteomic analysis.

In addition to profiling the entire Zn²⁺ proteome, iodoacetamide has been used to identify novel Zn²⁺ binding cysteine residues and their involvement in protein and cell function. The Canada and Pérez-Sala groups found that one cysteine residue, C328, is responsible for vimentin organization in cells [100, 101]. Iodoacetamide and other electrophilic reagents readily alkylated C328 and though alkylation did not block vimentin polymerization, it severely altered the morphology of the resulting filaments. C328 did not form a disulfide bond between adjacent proteins; therefore, the authors hypothesized that it may be involved in Zn²⁺ coordination instead. Indeed, incubation of vimentin in the presence of 500 μM ZnCl₂ severely inhibited labeling of C328 by iodoacetamide, an effect that was not seen with CaCl₂ or MgCl₂. Additionally, vimentin polymerization in vitro was enhanced by incubation with ZnCl₂ and could be reversed through the addition of TPEN. Addition of TPEN to SW13 cells (a vimentin-null cell line) stably expressing wildtype vimentin, but not the C328S mutant, caused rapid disassembly of vimentin filaments into visible puncta, an effect that could be reversed with ZnCl₂ addition. Thus, Zn²⁺ coordination by C328 is critical for vimentin polymerization and construction of the cytoskeleton.

The Eide group recently examined the Zn²⁺ proteome in Saccharomyces cerevisiae in the context of “Zn²⁺ sparing”, or the repression of Zn²⁺ protein expression to reduce the cellular requirement for Zn²⁺ [102]. Through bioinformatics, they identified 582 putative Zn²⁺ binding proteins in S. cerevisiae, or approximately 8% of all proteins in the cell. Using the “Proteomic Ruler” technique [103], they found that in Zn²⁺ replete cells there are approximately 9 x 10⁶ Zn²⁺ binding sites. Interestingly, the 20 most abundant Zn²⁺ binding proteins (alcohol dehydrogenase 1 (Adh1), aldolase (Fba1), superoxide dismutase (Sod1), among others) accounted for nearly 90% of the total Zn²⁺ requirement in the cell. Upon Zn²⁺ withdrawal, they found that many downregulated proteins were related to general protein synthesis, resulting in approximately 4.6 x 10⁶ Zn²⁺ bind sites per cell under Zn²⁺ deficient conditions, a 45% reduction compared with Zn²⁺ replete conditions. Interestingly, ICP-MS analysis showed that Zn²⁺ deficient cells contained 1.6 x 10⁶ atoms per cell, or approximately 30% of the estimated number of Zn²⁺ binding sites under these conditions. This suggested that many of the Zn²⁺ binding proteins are in fact apoproteins in Zn²⁺ deficient conditions.

Eide and colleagues identified aldolase (Fba1) as the second most abundant Zn²⁺ protein in replete yeast cells and the most abundant Zn²⁺ protein in deficient cells. To probe the metal status of Fba1, they used the reagent N-ethylmaleimide (NEM), a cell permeable cysteine thiol alkylating probe whose reactivity towards thiols is severely hampered by metal binding. Despite the lack of direct involvement of Fba1’s five cysteine residues in Zn²⁺ binding, the authors found that Fab1 was more susceptible to NEM labeling in Zn²⁺ deficiency. This suggests that a conformational change occurs upon Zn²⁺ binding that causes at least one cysteine to become more susceptible to NEM labeling. Methionine synthetase (Met6), is another Zn²⁺-binding protein, which has three cysteine residues, two of which are thought to coordinate Zn²⁺. In Zn²⁺ replete conditions, the majority of Met6 proteins had all three cysteines protected from NEM labeling while in Zn²⁺ deficient conditions, only one cysteine was protected. Similar to Fba1, mutation of a non-cysteine Zn²⁺ binding residue results in constitutive NEM labeling. Taken together, these findings suggest that Fba1 and Met6, and likely other Zn²⁺ binding proteins, exist as apoproteins within Zn²⁺ deficient cells.

Interrogating the proteome with a Zn²⁺-responsive chemical probe

The Hamachi group developed a cell permeable chemical probe, AIZin, that responds to labile Zn²⁺ to become active and label neighboring proteins [104]. AIZin uses acyl imidazole (AI) as the electrophilic group and DPA as the Zn²⁺ binding site (Figure 2B). When Zn²⁺ binds to the DPA moiety, the electron density of the imidazole ring is reduced. This increases the reactivity of the carbamate unit of AI through conjugation and makes it more susceptible to attack by nucleophilic sidechains. Additionally, AIZin is conjugated to fluorescein to allow for visualization in live cell microscopy experiments or for immunoprecipitation (IP) enrichment for proteomic analysis. AIZin has a Zn²⁺ K_D of 0.7 nM and is unaffected by millimolar concentrations of other divalent cations such as Ca²⁺ or Mg²⁺. Some other transition metal cations such as Co²⁺, Cu²⁺, Cd²⁺, and Pd²⁺ facilitated hydrolysis (40-80% reactivity relative to Zn²⁺), but these are substantially less labile than Zn²⁺ in live cells. Additionally, addition of the Zn²⁺ chelator TPEN abolishes all labeling.

AIZin is capable of labeling many cellular proteins and is a powerful tool for studying processes that involve release of Zn²⁺, such as in response to nitric oxide (NO) production. Metallothioneins (MTs) are thought to be a major source of labile Zn²⁺ inside of cells, and others have shown that NO-mediated S-nitrosylation of MTs rapidly increases the labile Zn²⁺ pool [105]. This excess Zn²⁺ may be sequestered in vesicles to reduce potential cytotoxic effects. However, the proteins that mediate this sequestration and their relationship to other cellular components are largely unknown. Therefore, Miki and colleagues used AIZin to conditionally label proteins upon NO induction. The NO donor S-nitrosocysteine (SNOC) induced rapid fluorescence signal throughout the cell and at later time points the AIZin signal localized to Zn²⁺-rich (via FluoZin-3) vesicles. Subsequent differential centrifugation of whole cell lysate into organelle and cytosolic compartments revealed that AIZin labeled proteins were detected in both fractions at early time points but migrated to the organelle fraction over time . IP enrichment using an anti-fluorescein antibody and tandem mass tagging to multiplex samples allowed for quantitative proteomic analysis of the AIZIn labeled proteome. In agreement with their imaging data, NO treatment resulted in AIZin enrichment for vesicular and ER proteins. This enrichment was abolished with TPEN treatment, confirming that this is indeed dependent on Zn²⁺ being concentrated to vesicles.

AIZin has also been used to probe mobile Zn²⁺ dynamics in Arabidopsis thaliana. The Yoshimoto group recently published work detailing the role of autophagy in Zn²⁺ deficient plants [106]. Autophagy is an essential process that aids in nutrient recycling under nitrogen and carbon starvation. Yoshimoto’s group suspected this would also be essential in other deficiencies as well, such as Zn²⁺ deficiency. They generated autophagy-defective plants (atg mutants) and found that Zn²⁺ deficiency enhanced the effects of the atg mutations, with seedlings having significantly shorter roots and lower levels of chloroplastic proteins such as Rubisco. Using AIZin and SDS-PAGE, they found that Zn²⁺ deficiency induced mobilization of Zn²⁺ ions in plants but that the atg mutants showed less Zn²⁺ mobilization, as measured by fluorescein intensity. They suggest that under Zn²⁺ deficient conditions autophagy assists in mobilizing Zn²⁺ by degrading unnecessary proteins to ensure that Cu/Zn superoxide dismutases (SODs) have sufficient Zn²⁺ to function. In the atg mutants, Cu/Zn SOD activity was lower than in wild type plants under Zn²⁺ deficiency, while adequate Zn²⁺ yielded comparable activity between the two plants. This suggests that during Zn²⁺ deficiency, autophagy helps to ensure the pathways involved in oxidative stress retain function to protect against ROS.

APPLICATIONS OF TOOLS TO PROBE CELL BIOLOGY OF Zn²⁺

Zn²⁺ in cellular regulation

It is widely known that Zn²⁺ is necessary for cell proliferation, as Zn²⁺ deficient cells fail to divide and proliferate [107] and one of the symptoms of Zn²⁺ deficiency is stunted growth [108]. However, the molecular mechanisms of how Zn²⁺ deficiency leads to cell cycle arrest have remained elusive. Recently, new tools to study both Zn²⁺ and the cell cycle have begun to provide insight into the role of Zn²⁺. Cell cycle studies typically rely on serum starvation of cells to synchronize the cell cycle phases across a population. While serum starvation removes essential growth factors such as mitogens, it also removes essential vitamins and minerals, including Zn²⁺ [109]. Furthermore, cell synchronization can induce stress response pathways, making it difficult to correlate findings to naturally cycling cells [110, 111]. Advances in fluorescent reporters, high throughput microscopy, and quantitative image analysis have made it possible to monitor cell cycle phases in naturally cycling cells to characterize cells’ temporary exit from the cell cycle, termed quiescence [112].

In an asynchronously-dividing population of cells with chelex-treated serum, Lo and colleagues were able to track cells throughout the cell cycle and determine which aspects of the cell cycle are dependent on Zn²⁺ [113]. Cells were grown in media with chelex-treated serum to simulate a minimal media condition (1.46 μM Zn²⁺, measured by ICP-MS), and experimental conditions were either supplemented with 30 μM Zn²⁺ (Zn²⁺ replete) or further Zn²⁺-restricted by the addition of 2-3 μM TPA. The metal in cells was then quantified using the genetically encoded sensor ZapCV2 [22], which demonstrated that media Zn²⁺ and TPA manipulation resulted in changes in labile cellular Zn²⁺, and that these manipulations are not toxic to cells for the duration of the experiment [113]. The Zn²⁺ deficient condition was found to either induce cellular quiescence or cell cycle stall in S-phase, with insufficient DNA replication. The cellular response to Zn²⁺ deficiency was shown to be dependent on when in the cell cycle Zn²⁺ deficiency was induced. Interestingly, DNA damage was increased in cycling Zn²⁺-deficient cells, but not quiescent cells, suggesting that while Zn²⁺ is necessary for DNA replication and repair, its role in the proliferation / quiescence decision is independent of DNA damage. Furthermore, resupply of Zn²⁺ promotes cell cycle re-entry, but the mechanism of which proteins and signaling pathways mediate cell cycle re-entry has not been identified [113].

Zn²⁺ in cell signaling

Another major push in the field of Zn²⁺ biology is understanding systems in which Zn²⁺ can act as a cellular signal. A variety of cell systems have been shown to exhibit dynamic Zn²⁺ behavior, including neurons and oocytes. It has long been known that specific regions of the brain are Zn²⁺-rich [114] and that the Zn²⁺ transporter, ZnT3, imports Zn²⁺ into synaptic vesicles [115]. Along with ZnT3 knockout mice [116], the extracellular Zn²⁺ chelator ZX1 [82] has led to insights about the role of vesicular Zn²⁺ in neurons. Zn²⁺ has been shown to modulate signaling through neurotransmitter receptors, including the inhibition of NMDA glutamate receptors [82, 83, 117]. New electrophysiology studies show that synaptically-released Zn²⁺ from hippocampal mossy fiber neurons contributes to long term potentiation, or strengthening, of neuronal synaptic connections [118]. Recently, Sanford and colleagues quantified Zn²⁺ dynamics upon stimulation by KCl in cultured hippocampal neurons and demonstrated that over 900 genes exhibit changes in expression in a Zn²⁺-dependent manner in response to subnanomolar fluctuations of Zn²⁺. Many of these transcriptional changes involve synaptic plasticity pathways [119]. Furthermore, in vivo animal studies have demonstrated that Zn²⁺ may play a role in fear conditioning, long-term and spatial memory, and audition [86, 120-122]. These in vivo phenotypes are subtle, suggesting that synaptic Zn²⁺ may play more of a role in fine-tuning neuronal connections or that there are compensatory mechanisms during development that mask synaptic Zn²⁺ deficiency phenotypes.

While it has long been recognized that Zn²⁺ is concentrated in mossy fiber neurons in the hippocampus and released with synaptic activity through studies in brain slices and animals, cellular models of Zn²⁺ dynamics are far less clear. In dissociated hippocampal neuron culture, stimulation with glutamate/glycine [123] or KCl [124] has been shown to increase intracellular Zn²⁺, and this Zn²⁺ signal has important downstream signaling consequences. While it was routinely assumed that this intracellular Zn²⁺ derived from synaptic release, more recent studies suggest that this Zn²⁺ may arise from an intracellular source. In particular, it was suggested that neuronal acidification upon glutamate treatment was responsible for Zn²⁺ mobilization [123]. Sanford and Palmer recently quantified Zn²⁺, Ca²⁺, and pH changes using a series of fluorescent sensors during stimulation of dissociated hippocampal neurons in culture. Both KCl and glutamate stimulation led to increases in cytosolic Zn²⁺, and the signal was comparable in the presence of ZX1, suggesting that in dissociated neuron culture Zn²⁺ was released from intracellular stores. Although the pH decreased upon neuronal stimulation, the magnitude and timing of the changes in pH did not correlate with the magnitude and timing of Zn²⁺ changes, suggesting that Zn²⁺ release from intracellular stores might be due to Ca²⁺ dynamics or reactive oxygen species (ROS) production [124].

Another cellular system that experiences Zn²⁺ signals is the developing oocyte. Recently, Zn²⁺ accumulation during oocyte maturation and release at fertilization were rigorously quantified to better understand the source of the “Zn²⁺ spark” at fertilization. Que and colleagues, used the fluorescent Zn²⁺ probe ZincBY-1, elemental mapping, and extracellular FluoZin-3 Zn²⁺ dye to show that oocytes contain thousands of vesicles loaded with approximately 1 million Zn²⁺ ions each, and that these vesicles undergo exocytosis when the oocyte becomes fertilized, resulting in the extracellular Zn²⁺ spark [60]. Subsequent research demonstrated that the Zn²⁺ released from oocytes upon fertilization is linked to the hardening of the zona pellucida, the oocyte’s glycoprotein extracellular matrix, which prevents subsequent fertilizations and maintains the viability of the newly formed zygote [125]. Furthermore, larger Zn²⁺ sparks have been shown to be indicative of future embryo quality, suggesting that Zn²⁺ could potentially be used as a biomarker for fertility treatments [126]. These recent discoveries in oocyte biology suggest that Zn²⁺ regulation and dynamics can play multiple roles in cell development and physiology.

Table 2:

Commonly used Zn²⁺ chelators

Chelator	Zn2+ affinity	Rate of Zn2+ binding or depletion	Selectivity	Application	Advantages	Disadvantages	Ref
Chelex 100	Undefined	30 minute incubation at 4°C led to 10-fold zinc depletion	Binds most transition metals and divalent cations	Removal of metals from media or serum	Simple Inexpensive Literature precedent	Non-selective Should confirm metal depletion via ICP-MS and supplement as needed	[71]
A12 resin	Low or sub-nanomolar	4 hr incubation led to >99% Zn2+ depletion	No detectable Ca2+, Mg2+, Co2+, or Cu2+ depletion from media, although S100A12 binds Ca2+	Selective removal of zinc from media or serum	Selective for zinc May be washed and re-used four times	Not commercially available Relatively new reagent so not a lot of examples of use	[72, 129]
TPEN	<1 fM	4.24x10−6 M−1s−1 (Huang) or 0.016 s−1 (Pan)	Binds Ca2+ 1011 more weakly than Zn2+ Binds Cu2+ more tightly than zinc	Chelation of zinc in cells and from some cellular proteins	Somewhat selective for zinc High affinity Literature precedent	May strip zinc from cellular proteins Cytotoxicity	[74, 77, 82]
TPA	10 pM	5.81x10−6 M−1s−1	Similar to TPEN: Binds Cu2+ more tightly than Zn2+, but only weakly binds Ca2+	Chelation of zinc in cells	Rapid Somewhat selective for zinc Lower cytotoxicity than TPEN		[74, 77]
ZX1	1.0 nM	0.027 s−1	Does not show appreciable Ca2+ or Mg2+ binding	Rapid chelation of extracellular zinc	Rapid Selective for zinc	Relatively new reagent so not a lot of examples of use	[82]