Subcellular localised small molecule fluorescent probes to image mobile Zn2+

Zn2+, as the second most abundant d-block metal in the human body, plays an important role in a wide range of biological processes, and the dysfunction of its homeostasis is related to many diseases, including Type 2 diabetes, Alzheimer's disease and prostate and breast cancers. Small molecule fluorescent probes, as effective tools for real-time imaging, have been widely used to study Zn2+ related processes. However, the failure to control their localisation in cells has limited their utility somewhat, as they are generally incapable of studying individual processes in a specific cellular location. This perspective presents an overview of the recent developments in specific organelle localised small molecule fluorescent Zn2+ probes and their application in biological milieu, which could help to extend our understanding of the mechanisms that cells use to respond to dysfunction of zinc homeostasis and its roles in disease initiation and development.


Introduction
Zn 2+ is the second most abundant d-block metal ion in the human body, with a total content of about 2 g, most of which is bound to proteins. It is estimated to exist in over three thousand proteins and is widely used for catalytic, regulatory, and structural roles. 1,2 Zinc plays an important part in a wide range of biological processes, such as brain function and pathology, immune function, gene transcription, and mammalian reproduction. Due to this, it is unsurprising that problems with zinc homeostasis are associated with many diseases, including Alzheimer's disease, 3 prostate cancer, 4 type 2 diabetes, 5 and immune dysfunction and infection. 6 Whilst the majority of zinc is found in bound forms, there exists a pool of 'mobile' or 'free' zinc, which initiates transient signals that stimulate various physiological processes, though its concentration in the cytosol is only in the picomolar range. 7 The thiol-rich metallothioneins (MTs) and zinc transporters are normally involved in the processes to maintain cellular zinc homeostasis. 7 MTs acting as Zn 2+ buffers, can bind a large amount of Zn 2+ and release it under conditions of oxidative stress due to the antioxidant role it plays. There are two main families of zinc transporters, ZIP and ZNT, which control the import and export of cytosolic zinc to intracellular organelles or extracellular space (Fig. 1). 8 Mobile Zn 2+ is associated with the regulation of gene expression, insulin secretion, and is also considered as a signalling ion for intra-and intercellular communication, such as neurotransmission. 9 Therefore, effective methods are required to image mobile Zn 2+ at the cellular or subcellular level in order to understand these processes. 10 Since Zn 2+ has a d 10 conguration, it is redox inert in biology and is rendered spectroscopically silent for most of the commonly used photo-spectroscopic techniques. 11 As a result, the development of uorescent chemosensors has become popular for noninvasive real-time Zn 2+ imaging in which a change in the uorescence intensity or wavelength occurs due to analyte binding, which leads to signal output. The most commonly used probes utilise a switch on uorescent mechanism in which photon- induced electron transfer (PET) [12][13][14] is prevented upon zinc binding (Scheme 1). Typically, such uorescent chemosensors contain a uorophore (the signal source), a short spacer unit and a receptor (the recognition site). In addition, uorescent probes utilising intermolecular charge transfer (ICT) 15,16 and Förster Resonance Energy Transfer (FRET) 17,18 mechanisms have also been widely applied in biology. In the biological environment, uorescent chemosensors are generally involved in a competitive exchange equilibrium with endogenous ligands and proteins and thus detect changes of mobile Zn 2+ pools rather than total cellular zinc levels. Therefore probes displaying high sensitivity with low detection limit, high specicity and selectivity to distinguish Zn 2+ from competing metal ions are preferred.
There are three main types of uorescent sensors for biological zinc imaging: small molecule probes, protein-based biosensors and peptide targeting uorescent sensors. Proteinbased biosensors and peptide targeting uorescent sensors can localise to specic organelles by introducing genetically encoded site-directing proteins and selecting specic amino acid sequences, which have high affinity for the particular targets. 10 However, they are not without some limitations. The genetically encoded protein-based sensors cannot be transfected into all cell lines, have a small range of excitation and emission wavelengths available, as well as low photochemical stability and brightness. The peptide-based sensors are sensitive to proteases in vivo and cell internalization can be difficult, except for specic peptide sequences. In contrast, small molecule uorescent probes can display high sensitivity and selectivity, low toxicity, and good photophysical properties. However, the failure to control the small molecule probes' cellular or subcellular location can limit their utility somewhat.
In the last two decades, there have been considerable efforts to overcome this problem. At the subcellular level, organelles require zinc for their normal function and the dysfunction of zinc homeostasis results in pathological processes, such as cellular stress, and these can induce cell apoptosis. Therefore, the development of Zn 2+ probes capable of predictable and reliable subcellular localisation is required for biological applications. Reviews exist that summarize a range of subcellularly localised uorescent probes, for cations, anions, or small molecules, [19][20][21][22] however, a summary of the recent developments in subcellularly localised Zn 2+ probes has, to our knowledge, not yet appeared. In this perspective, an overview of the recent development in specic organelle localised small molecule uorescent Zn 2+ probes is presented together with their applications in biological systems.

Plasma membrane targeting
The plasma membrane, also known as the cell membrane, is mainly formed by a lipid bilayer to protect cells from their external environment. The membrane controls the import and export of substances and the selective uptake of ions and organic molecules. Membrane transporters, a class of membrane protein, play an important role in these processes. As shown in Fig. 1, ZIP 1-6, ZIP8, ZIP10 and ZIP14 are zinc transporters that control cellular zinc uptake, while ZNT1, ZNT5B and ZNT10 are responsible for its efflux through the membrane. Some data have suggested that when zinc is present in sufficiently high concentration it can act as a stabilizer of the cell membrane. 23 It has also been reported that zinc deciency in membranes causes a defect in calcium channels, which also impairs the uptake of Ca 2+ . 24 There are relatively few small molecule uorescent probes reported to image Zn 2+ in the cell membrane. Due to the phospholipid bilayer nature of the membrane, a highly hydrophobic group is normally used as a membrane targeting unit. In 2011 Yamamoto et al. reported that cholesterol could be applied as a cell membrane targeting unit. 25 With uorescein as the uorophore and an o-aminophenol-N,N,O-triacetic acid-based zincchelating moiety, probe 1 (LF-Chol) showed good cell membrane localisation and uorescence response when Zn 2+ was added or removed (Fig. 2). More recently, cholesterol was also applied by You et al. in a deep-red uorescent probe 2 (JJ, Fig. 3) to image Zn 2+ . 26 Probe 2 displayed good photophysical properties and tight Zn 2+ binding with a low dissociation constant (4 pM). Co-localisation experiments in HeLa cells revealed that 2 was rapidly internalised to intracellular spaces, including the lysosome and endoplasmic reticulum (ER), but the localisation in the membrane was very low (the Pearson's coefficient was reported to be 0.12). Meanwhile, the authors found that probe 2 facilitated the inux of exogenous Zn 2+ in the absence of the pyrithione ionophore, while a control probe without cholesterol could not, thus suggesting that the membrane targeting cholesterol unit permeabilised the cell membrane (Table 1).
Besides cholesterol, long alkyl chains have also been widely used for membrane targeting. Rutter, Li et al. reported the uorescent probe 3 (ZIMIR) in which a pair of dodecyl alkyl chains were used as the membrane targeting unit as a Zn 2+ indicator to image dynamic insulin release. 27 The probe was quenched based on the PET mechanism and displayed a robust uorescence increase aer binding with Zn 2+ and showed low toxicity and membrane labelling in a wide range of cell lines. Using this probe, the authors demonstrated exocytotic activity at subcellular resolution from pancreatic b cells in intact islets and found that the sites of Zn 2+ /insulin release are mainly in small groups of adjacent b cells. Similar results were observed by Watkinson et al. with the plasma membrane targeting Zn 2+ probe 4. 28 The di-dodecylamide motif as the membrane targeting unit was introduced through a one-pot modular 'click-S N Arclick' approach, which was shown to be more efficient in the synthesis. Probe 4 showed a signicant switch on uorescence response to Zn 2+ due to aggregation phenomena, and in cellulo experiments in mouse pancreatic islets demonstrated its localisation to the exterior of the plasma membrane. This probe was subsequently used to measure dynamic insulin secretion by Hodson et al. since zinc is co-released from insulin-containing granules. 29 Similarly, Cho et al. synthesized the two-photon (TP) probe, 5 (FZn-mem), to image near-membrane Zn 2+ by introducing a long alkyl chain to 2-amino-7-(3-oxo-1-dodecen-1 0 -yl)-9.9-dimethyluorene (ADF), a two-photon uorophore with a large TP cross section. 30 With the same metal-binding motif as 3, probe 5 was shown to be highly selective for Zn 2+ over competing cations, and its dissociation constant K d was determined to be 20 AE 0.4 nM and 19 AE 0.2 nM by one-photon and two-photon spectroscopy, respectively. Aer TP excitation by 820 nm femtosecond laser pulses, the uorescence of 5 could be collected in the emission wavelength range of 450-600 nm, and near-membrane Zn 2+ could be detected in living cells and tissue at a depth of 110 mm.

Mitochondria targeting
The mitochondrion is an important organelle in eukaryotic organisms. It generates most of the chemical energy supply of adenosine triphosphate (ATP) in all cells, and takes part in   Table 1 The photophysical properties of probes 1-34 The results were obtained in solution EtOH : MOPS buffer (1 : This journal is © The Royal Society of Chemistry 2020 Chem. Sci., 2020, 11, 11366-11379 | 11369 Perspective Chemical Science many biological processes, such as signalling, the cell cycle, cell growth and cell death. 31 Mitochondria are involved in intracellular Zn 2+ storage and as a co-factor for a wide range of enzymatic reactions and Zn 2+ is closely related to the mitochondrial respiratory chain. 32 It was also reported that zinc can reduce mitochondrial damage under stress conditions, which can protect cells from oxidatively-induced apoptosis. 33 In contrast, mitochondrial dysfunction may cause rapid Zn 2+ entry, which is one of the major contributors to neuronal injury. 34 Therefore, to understand these processes better, mitochondria localised zinc imaging is in demand. The rst example of a mitochondrial uorescent Zn 2+ probe was reported by Gee et al. in 2003. 35 Probe 6 (RhodZin-3, Fig. 4) displayed a 75-fold uorescence increase aer binding Zn 2+ and good selectivity over competing cations. The ester form was loaded into cells due to its membrane permeability and it showed co-localisation with a mitochondrial marker Mito-Tracker Green and Zn 2+ response in cortical neurons.
In 2006, Lippard et al. synthesized a series of probes and found 7 (ZS5), displaying green uorescence, could localise to mitochondria 36 although this was the only probe to display mitochondrial localisation within a larger series of structurally similar probes synthesized. Another probe 8 (ZBR4), a red uorescent probe, was subsequently found to spontaneously accumulate in mitochondria in 2014 by the same group. 37 However, the localisation was again apparently serendipitous as this was the only probe in another large family of probes that failed to localise to the endoplasmic reticulum (ER).
Delocalised lipophilic cations, such as the triphenylphosphonium salt (TPP), are known as effective mitochondrial targeting groups, since they can cross hydrophobic membranes and accumulate in the mitochondria in living cells. By introducing this group Kim and Cho et al. obtained a mitochondriatargeting two-photon probe 9 (SZn-Mito) for Zn 2+ imaging (Fig. 5). 38 The probe showed a 7-fold increase of two-photon excited uorescence aer being bound with Zn 2+ and was able to detect zinc in living tissues to a depth of 100-200 mm through two-photon microscopy. Just one year later, the same group developed a similar mitochondria targeting probe 10 (SZn2-Mito) which had a 70-fold two-photon excited uorescence increase in response to Zn 2+ and displayed similar properties. 39 Besides these, another two-photon Zn 2+ probe 11 (FZn-Mito) was synthesized with a different uorophore, but it had similar mitochondria targeting behaviour and two-photon emission response to Zn 2+ . 40 In 2012 Guo et al. synthesized the ratiometric probe 12 (Mito-ST) which was capable of imaging zinc ions in the mitochondria. 41 With Zn 2+ associated, the probe showed a blue shi in both excitation and emission wavelength. Interestingly, they observed the concentration of Zn 2+ in mitochondria started to increase immediately upon H 2 O 2 (10 mM) stimulation and increased to 1.6 nM (from 0.6 nM) with the ratiometric imaging result based on the method reported by Tsien et al. 42 However, the Zn 2+ release stimulated by S-nitrosocysteine (SNOC, 10 mM), a precursor of the NO radical, experienced a lag phase, and the concentration was much higher (76 nM) compared to that stimulated by H 2 O 2 . In the same year Jiang et al. reported the quinoline-based ratiometric probe 13 (DQZn2) with similar properties. 43 The emission of the probe showed a 46 nm blue-shi, with about a 5-fold change of the emission intensity ratio aer binding with Zn 2+ , which allowed the concentration of mitochondrial free Zn 2+ to be quantitatively measured in NIH3T3 cells.
The mitochondria localised Zn 2+ probe 14 utilising a FRET mechanism was reported by Zhu et al. 44 The FRET donor uorophore displayed a bathochromic shi of emission when coordinated with Zn 2+ . The spectral overlap between the emission of the donor and the absorption of the acceptor increased upon Zn 2+ binding, which enhanced the FRET efficiency. In HeLa cells, when the probe was excited at 405 nm, red uorescence with emission wavelength 580-680 nm could be observed aer addition of ZnCl 2 .
In 2014, a reaction-based uorescent probe 15 (DA-ZP1-TPP) was reported to investigate mobile Zn 2+ in mitochondria. 45 As shown in Fig. 6, the phenolic oxygen atoms of the xanthene ring in the uorescein moiety was protected with an acetyl group, which rendered it non-uorescent in metal-free media. In the presence of Zn 2+ , the ester group was hydrolysed, and the PET mechanism quenched by Zn 2+ association, giving a strong uorescence response (more than a 140-fold increase). Interestingly, using this probe, the authors found that tumorigenic epithelial prostate cells could not accumulate mobile Zn 2+ in their mitochondria, compared to that of healthy epithelial prostate cells, which could.

Lysosome targeting
The lysosome acts as the waste disposal system in cells by digesting unwanted materials in the cytoplasm. Zn 2+ homeostasis in the lysosome is regulated by zinc transporters, ZIP3, ZIP8 and ZNT2, and its dysregulation is associated with a wide range of pathological processes at the cellular level. For example, oxidative stress induced by H 2 O 2 causes Zn 2+ accumulation in the lysosome in hippocampal neurons, which eventually undergo lysosomal membrane permeabilization (LMP) and this may be the mechanism of oxidative neuronal death. 46 It was also reported that lysosome-related organelles in intestinal cells of C. elegans (gut granules) are the major site of zinc storage, which promotes detoxication and subsequent mobilization, regulating cellular and organismal zinc metabolism. 47 Therefore, it is vital to develop uorescent probes to image Zn 2+ in the lysosomal space to help understand these kinds of biological processes.
In 2009 Guo et al. synthesized the intramolecular charge transfer (ICT) based probe 16 (NBD-TPEA, Fig. 7). 48 The probe demonstrated a good selectivity for Zn 2+ , a large Stokes shi, and was also applied for in vivo Zn 2+ imaging in zebrash larva. However, subcellular experiments showed it to not only accumulate in the lysosome but also in the Golgi apparatus, which may perhaps limit its application in lysosomal Zn 2+ imaging.
The basic ethylenediamine group has been reported to accumulate in the lysosome since it can be protonated by the acid environment in lysosomal space (pH ¼ 4.5-5.5). The basic dimethylethylamino moiety has therefore been introduced as a lysosome targeting group in the ratiometric probe 17 (DQZn4) based on quinoline and showed good targeting behaviour. 49 It performed well in the acidic pH environment, showing  Chemical Science a signicant turn-on uorescence response and 47 nm blueshi of emission wavelength on binding Zn 2+ . With these desirable properties in vitro, it was successfully utilised for imaging lysosomal Zn 2+ changes in NIH 3T3 cells.
Another group that has been frequently used for targeting the lysosome is the morpholine unit. One example is the twophoton probe 18 based on a naphthalimide dye with an N,Ndi-(2-picolyl)ethylenediamine (DPEN) Zn 2+ ligand, which can image intracellular Zn 2+ in the lysosome and mouse brain tissues under two-photon excited microscopy. 50 Unlike the other subcellular targeting probes, in this case organelle differentiation relied on a concentration gradient of the probe intracellularly. Probe 18 showed a strong uorescence switch on response to Zn 2+ in the lysosomal pH range (pH ¼ 4.5-5.5) but the intensity increase was much smaller at cytosolic pH range (pH ¼ 7.2-7.4), therefore it was able to detect lysosomal Zn 2+ specically.
Using the same targeting unit Peng et al. reported a ratiometric probe 19 (LysoZn-1) to image lysosomal Zn 2+ (Fig. 8). 51 The authors introduced an electron donor 4-ethoxylphenyl to the meso-position of a styryl-Bodipy-DPA scaffold, which distinguished Zn 2+ from Cd 2+ very well. They explained this using Hard-So Acid-Base theory, 52 which was supported by theoretical calculations. Probe 19 exhibited a signicant uorescence increase and ratiometric (F 578 nm /F 680 nm ) changes upon Zn 2+ binding and had a good response to Zn 2+ in the lysosomal pH range. Using 19 it was observed that lysosomal Zn 2+ concentration increased upon H 2 O 2 stimulation in neuronal stem cells and a similar phenomenon was observed by Uvdal et al. with a PET based lysosome localised probe 20 (Lys-NBD-TPEA) in which morpholine was again utilised as the targeting unit. 53 Based on a FRET mechanism, two uorescent Zn 2+ probes 21 were prepared by Zhu et al. with different aliphatic amino groups as the lysosome targeting units. 54 A Zn 2+ -sensitive arylvinylbipyridyl uorophore was selected as the FRET donor, and through the efficient intramolecular FRET process, its broad emission band was transformed into a strong, narrow emission band of the acceptor Bodipy, which is preferable for multicolour imaging. With a 2 : 1 stoichiometry between 21 and Zn 2+ , the molar absorptivity of the donor was increased upon  Zn 2+ coordination, leading to the enhancement of acceptor emission. The lysosome localisation was conrmed by confocal microscopy and with the high resolution of structured illumination microscopy (SIM), they found 21b localised to the interior of lysosomes in HeLa cells, rather than anchoring at the lysosomal membranes (Fig. 9). Sessler et al. synthesized a series of probes 22 (LysoDPP-C2-C4), which were designed based on AND logic to detect both Zn 2+ and the acidic pH of the lysosome. 55 The morpholine moiety served not only as a lysosome targeting unit but also as a pH-responsive marker, since the nitrogen in the morpholine moiety was protonated at low pH and the PET was quenched, which increased the uorescence of the diketopyrrolopyrrole (DPP) uorophore in the same way as the PET quenching upon Zn 2+ binding. The experiments in cellulo showed 22c was the most effective probe in terms of the initial uorescence and the uorescence response to Zn 2+ . When incubated with 22c, the cancerous prostate cell lines PC3 and DU145 showed no change in uorescence intensity, while the normal human prostate epithelial cell line RWPE1 displayed a signicant increase on the addition of exogenous Zn 2+ . The authors also demonstrated that the probe was capable of imaging the prostate in vivo nude mice models and discriminating between cancerous and normal prostate tissue through histological studies.

Endoplasmic reticulum (ER) targeting
The ER, an organelle in eukaryotic cells, serves a number of important cellular roles, such as protein synthesis and transport, protein folding, carbohydrate metabolism, lipid and steroid synthesis. 56 Proteins synthesized in the ER are normally properly folded and transported to the Golgi apparatus, however, when there are changes to ER function, resulting from factors such as ageing, genetic mutations, or the environment, unfolded or misfolded proteins are synthesized and accumulate in the ER, causing ER stress, which activates the unfolded protein response (UPR). 57 It is known the ER acts as an intracellular store for biological mediators, including zinc, which it requires for normal function. For example, it has been found that zinc can be released from thapsigargin-and inositol 1,4,5-trisphosphate (IP3)sensitive ER storage in cortical neurons. 58 ZIP7, ZIP9, ZIP13 and ZNT5-7 are the transporters to regulate Zn 2+ inside the ER and the depletion of zinc transporters and zinc deciency can cause ER stress and upregulate the UPR, 59,60 which can result in inammation 61 and a wide range of diseases, such as diabetes 62 and neurodegenerative disorders, including Parkinson's and Alzheimer's diseases. 63 However, the role of 'mobile' zinc in these processes is little understood due to the lack of suitable molecular tools to image this subcellular region that exist.
Reports of ER localised small molecule Zn 2+ probes are very limited, and all early reports were of systems found to accumulate in the ER adventitiously. In 2013 Lippard et al. reported a series of benzoresorun based red-emitting uorescent probes 23 (the ZBR family, Fig. 10) for labile Zn 2+ . 64 The probes displayed a broad absorption band and a bathochromic shi aer binding with Zn 2+ , while the emission showed up to 8.4fold increase with addition of Zn 2+ . In cellulo studies revealed all probes accumulated in the ER in a variety of cell lines, as the Pearson's correlation coefficient with ER tracker was much higher the other organelle tracker dyes (Fig. 11). Interestingly, with 23a, the authors observed the depletion of labile zinc in the ER of neural stem cells under ER stress induced by peroxynitrite.
The conveniently prepared small molecular uorescent probe 24 included benzothiazole as the uorophore and an ohydroxyl Schiff base as the Zn 2+ receptor. 65 Upon binding Zn 2+ , the PET process was blocked, and the uorescence response had a 65-fold increase with a slight red shi, with a Job's plot showing a 2 : 1 binding stoichiometry between 24 and Zn 2+ . In human liver hepatocellular carcinoma cells, 24 was able to  Given the ongoing need to develop an effective and reliable strategy for targeting the ER, a number of targeting units for the organelle have been explored. Glibenclamide functions as the ER targeting group in the commercial ER tracker red and ER tracker green, and using a sulfonyl urea analogue, we developed two probes 25 and 26 as ER targeting probes. 66 It was shown that both probes accumulated in the ER in a number of cell lines, and displayed a good switch on uorescence response to Zn 2+ . Probe 25 was used to demonstrate that a decrease of mobile Zn 2+ concentration in the ER occurs under conditions of ER stress induced by tunicamycin and thapsigargin.
In addition to this cyclohexyl sulfonylurea moiety, the recently reported methyl phenyl sulfonamide was also incorporated as an ER targeting group through an alternative modular 'click-S N Ar-click' approach. 67 The probe obtained, 27, was also found to localise to the ER compared to the other organelles tested.

Golgi apparatus targeting
The Golgi apparatus works as a central station in cells, receiving secretory cargoes exported from the ER packing proteins into membrane-bound vesicles and sending them to their intra-and extra-cellular destinations. Zn 2+ is integral to these processes for a variety of proteins, functioning in catalytic, regulatory, and structural roles. For example, it was found that zinc takes part in the interaction between the two main Golgi proteins GRASP55 and Golgin45, maintaining the normal morphology of the Golgi apparatus; 68 Zn 2+ also coordinates with insulin monomers in the trans-Golgi network to package it into secretory granules, which are then released from pancreatic b-cells. 69 To regulate Zn 2+ homeostasis in the Golgi apparatus, transporters ZNT4-7 and ZIP7, ZIP9, ZIP11, ZIP13 are responsible for Zn 2+ import and export, respectively. The breakdown of Zn 2+ homeostasis in the Golgi apparatus is likely to be associated with a range of human disorders, such as cancer and neuronal, liver, kidney and eye diseases 70 and the development of effective methods for its detection and monitoring are required.
The rst Zn 2+ probe to localise in the Golgi apparatus was reported in 2000 by Lippard, Tsien et al. 71 With uorescein as the signalling unit, probe 28 (Zinpyr-1, Fig. 12) has a large extinction coefficient, high quantum yield, and good membrane permeability. Experiments in Cos-7 cells showed it colocalised well with a galactosyl transferase-enhanced cyan uorescent protein fusion (GT-ECFP), conrming 28 stains the Golgi apparatus. Some years later Lippard et al. presented two cell- Fig. 12 The structures of Golgi apparatus localised fluorescent probes 28-32. trappable uorescent Zn 2+ probes: a carboxylic ester probe 29a (QZ2E) and its carboxylic acid analogue 29b (QZ2A). 72 The probes were poorly emissive but had a signicant increase in emission aer binding with Zn 2+ (120-fold for 29a, 30-fold for 29b). Interestingly, the authors found the 29b was cell membrane-impermeable due to its carboxylic acid moieties, but 29a was membrane permeable and mainly localised to the Golgi apparatus; aer 18 hours' incubation time, the ester was hydrolysed to produce 29b, which was trapped inside cells.
Guo et al. modied the 8-sulfonamidoquinoline based commercial probe Zinquin to produce a new uorescent probe 30 (BPSQ) to image Zn 2+ with a 1 : 1 binding stoichiometry, which was better at discriminating mobile Zn 2+ . 73 Colocalisation experiments showed it to accumulate preferentially in the Golgi apparatus.
Following the previous experience on mitochondria targeting Zn 2+ probes, 38,39 Kim et al. reported a Golgi-localised twophoton probe 32a (SZnC), which had a strong two-photon excited uorescence enhancement in response to Zn 2+ for real-time monitoring of Golgi Zn 2+ changes (Fig. 13). 74 According to theoretical predictions for Golgi apparatus staining, the lipophilicity value (log P value) should be within the range of 3-5. 75 The log P values were calculated through measuring the probe's partitioning ratio between n-octanol and buffer, and the value of 32a was found to be 2.9 AE 0.1, which is reasonably matched to the theoretical range whereas it was determined to be 2.5 AE 0.1 for the control compound 31 (SZn), which experiment proved to be non-Golgi targeting. However, the log P values of another control probe 32b was 3.1 AE 0.1 and this was spread over the entire cell except the nucleus. The authors attributed the Golgi localisation of 32a to the weakly basic pyridyl group, which 32b does not possess. Therefore, both the lipophilicity and pyridyl moiety of 32a appear to contribute to its accumulation in the Golgi apparatus and the probe was applied for imaging Zn 2+ in rat hippocampal tissue and could be observed at a depth of more than 100 mm with two-photon microscopy. In addition to the log P value, it may be that other factors, such as amphiphilicity, electric charge and pK a values, molecular and ionic weights and conjugated bond number of the probes could also be considered to predict their location more accurately. 75 It appears that all of the probes described above were found to reside in the Golgi apparatus adventitiously, therefore a Golgi targeting unit is still required in order to develop reliable Golgi apparatus localised Zn 2+ probes. Ceramide may be an option for Golgi apparatus targeting since it has been employed in commercial stains for the Golgi apparatus such as NBD-ceramide and Bodipy-ceramide. However, it requires multi-step synthesis and also co-localises to the plasma membrane, which may limit future applications to some extent. More recently a phenylsulfonamide-based uorescent probe was reported to localise to the Golgi apparatus for H 2 S imaging, 76 which may provide a possible Golgi apparatus targeting unit for Zn 2+ probes. Furthermore, cysteine has been reported as a Golgi apparatus targeting unit in other systems, 77,78 since galactosyltransferase and protein kinase D were found to anchor to the Golgi region via their cysteine residues or cysteine rich domains 79,80 and this may also nd utility. However, it has been noted that cysteine is too hydrophilic, 81 and consequently not membrane permeable in some cases. We therefore incorporated an S-trityl protected cysteine moiety into a new Zn 2+ probe through 'click' chemistry, and conrmed the probe to be membrane permeable. 82 Interestingly, the trityl group appeared to be removed in cellulo within 24 hours, and the deprotected probe accumulated in the Golgi apparatus.

Nucleus targeting
The nucleus, as the control centre of the cell, maintains the integrity of genes and controls the activities of the cell by regulating gene expression. In eukaryotic cells, there is a double membrane which encloses the organelle from the cytoplasm. This nuclear membrane is impermeable to large molecules, and the nuclear pores are the channels for large molecules, which must be actively transported by carrier proteins, while also allowing free movement of small molecules and ions. Among these, ZNT9 and ZIP7 were found to be responsible for zinc inux and efflux. About 30-40% of total intracellular zinc is found in the nucleus, which plays an important role in the regulation of cell proliferation. It stabilises the structures of DNA, RNA and the ribosome, and is involved in DNA and protein synthesis. Nuclear hormone receptors are regulated by zinc nger domains, and zinc deciency impairs their responsiveness, which also has effects on metabolic processes related to growth. 83 Some zinc-dependent proteins have been found in the nucleus functioning as transporters. 84 It is therefore necessary to image Zn 2+ in the nucleus to better understand zinc inux and distribution and its related biological processes.
Nucleus localised Zn 2+ probes have been realised by genetically encoded sensors 85 or small molecule fusion protein tags, 86 Fig. 13 The colocalisation of probe 32a (SZnC) ( Fig. 14), reported by Guo et al. 89 displayed a 4-fold uorescence enhancement and blue shi of the ICT absorption band on Zn 2+ binding. The intracellular distribution of the probe was studied in HeLa and HepG2 cells and when the cells were stained with 33, only cytosolic uorescence was observed. However, when exogenous Zn 2+ was added, the entire cell became uorescent, including the nucleus, which was conrmed with the Hoechst stain, a commercial DNA dye (Fig. 14). It demonstrated the ability of the probe to penetrate the nuclear envelope, but the nuclear uorescence response was small due to the low nuclear concentration of labile Zn 2+ . The authors speculated that incorporation of the 4-amino-1,8-naphthalimide moiety to be the reason for its nuclear envelope penetrability through positive and negative control experiments. Furthermore, in vivo Zn 2+ imaging in zebrash larva was also performed. In 2018, we developed the biotinylated probe 34 based on 1,8-naphthalimide moiety for imaging Zn 2+ in breast cancer MCF-7 cells and found it has nuclear envelope penetrability. 90 However, it suffers from the same issue as 33 in that it was distributed over the entire cell and its sensitivity was insufficient to detect endogenous nuclear Zn 2+ with the nucleus uorescence only being switched on aer the addition of exogenous Zn 2+ . Nonetheless, the challenge remains to develop small molecule probes to image Zn 2+ in the nucleus specically, where the concentration of mobile Zn 2+ in this organelle is also very low. In this case, targeting groups that bind to DNA may be possibilities, such as the non-uorescent pyrrole-imidazole polyamides. 91,92 Summary and outlook Efforts in the development of small molecule uorescent probes in last two decades have provided a wide range of tools to study Zn 2+ related processes at the subcellular level, which are already beginning to be utilised to provide answers to questions around the role of this important cation in biology. However, its role, especially its function in individual cellular processes, remains far from being fully understood, and a number of challenges and requirements endure for the further development of small molecule Zn 2+ probes.
Reliable and effective small molecule probes to study mobile Zn 2+ , as well as a number of other analytes, in some specic organelles, such as the Golgi apparatus and nucleus, are still needed. It should be noted that many factors need to be considered to achieve organelle targeting, including their pH, 93 ionic strength and temperature of the medium, concentration of the probe and incubation time. 75 It is also desirable to demonstrate the organelle targeting efficacy of new probes in a wide range of cell lines in case of false positive results in individual cell lines if they are to nd widespread utility in biological applications.
There remains a signicant demanded to develop tools to accurately quantify the endogenous Zn 2+ in specic locations by developing ratiometric uorescent probes, since most of the probes discussed herein are based on the typical PET mechanism and are not able to image Zn 2+ quantitively due to either no or a small shi in emission wavelength on analyte binding. One viable route towards the development of ratiometric PET probes has been reported by Radford et al. in which a second zinc-insensitive uorophore was appended to a zinc-sensitive PET based probe using a polyproline helix as linker to provide sufficient rigidity. 94 This approach may open possibilities for other ratiometric organelle targeted probes, although this would further complicate their synthesis and it is possible that such derivatisation may affect cell uptake and targeting properties. As an alternative to PET probes, small molecule ICT probes utilising similar organelle targeting vectors may come to the fore, but again their addition to existing systems adds to synthetic complexity and may adversely affect their photochemical properties.
Probes with longer wavelength excitation and emission proles, such as in the near-infrared range (650-900 nm), which are nding increasing application in vivo, are also required and could be utilised to study Zn 2+ in deeper tissue with fewer undesirable effects from autouorescence of surrounding tissues and less harm to biological samples. More importantly, with the rapid development of high spatiotemporal resolution uorescent microscopic techniques, especially super-resolution microscopy, a number of new challenges are emerging. Besides the basic requirements for biological applications, probes with excellent photostability under the high photon intensity that these techniques utilise are required, whilst also displaying selective organelle targeting ability and a capability to provide real-time information about the location, dynamics and interactions of Zn 2+ in living cells. In this case, small molecule markers are preferable as they have shown higher photostability compared to uorescent proteins. The cyanine, BODIPYS, and especially the rhodamines and their derivatives have been successfully applied in super-resolution imaging, 95,96 which could have potential as uorophores for organelle specic Zn 2+ imaging at the nanometre level. In this case, it may also be necessary to develop routinely performed experiments for the assessment of photostability as part of the characterisation of newly reported probes so that they can be compared more easily.
Additionally, the extension of the application of fusion tags, such as SNAP-tag 97 and HaloTag, 86 or bioorthogonal reactions could be considered to localise Zn 2+ probes to an individual protein, to study spatiotemporal zinc uctuations in a specic process, or its response to intra-or extracellular environmental changes, which may help better understand the mechanism of disease initiation and development due to the dysfunction of zinc homeostasis.
Thus, whilst many exciting developments in small molecule uorescent probes have occurred in recent years and many signicant advances have been made there is still much work to be done and synthetic chemists, collaborating with biologists and biomedical scientists, are likely to be at the vanguard of future developments in the area.

Conflicts of interest
There are no conicts to declare.