Cite this article: Zhang C, Dischler A, Glover K, Qin Y. 2022 Neuronal signalling of zinc: from detection and modulation to function. Open

Biol. 12: 220188. https://doi.org/10.1098/rsob.22.0188

Received: 20 June 2022 Accepted: 11 August 2022

Subject Area: neuroscience/biochemistry

Keywords: neuronal, signalling, zinc, sensor

Author for correspondence: Yan Qin e-mail: [email protected]

Department of Biological Sciences, University of Denver, Denver, CO 80210, USA AD, 0000-0002-1798-0469; YQ, 0000-0003-3640-0105

Zinc is an essential trace element that stabilizes protein structures and allosterically modulates a plethora of enzymes, ion channels and neurotransmitter receptors. Labile zinc (Zn2+) acts as an intracellular and intercellular signalling molecule in response to various stimuli, which is especially important in the central nervous system. Zincergic neurons, characterized by Zn2+ deposits in synaptic vesicles and presynaptic Zn2+ release, are found in the cortex, hippocampus, amygdala, olfactory bulb and spinal cord. To provide an overview of synaptic Zn2+ and intracellular Zn2+ signalling in neurons, the present paper summarizes the fluorescent sensors used to detect Zn2+ signals, the cellular mechanisms regulating the generation and buffering of Zn2+ signals, as well as the current perspectives on their pleiotropic effects on phosphorylation signalling, synapse formation, synaptic plasticity, as well as sensory and cognitive function.

As the second most abundant trace element after iron, zinc has the highest concentration in the brain (6–95 µg g−1) [1–3], and is essential for brain function. Prenatal zinc deficiency during the critical period of rapid brain growth in fetuses causes irreversible damage to neural development [4–10], impairs learning and memory [11], and leads to development of autism-like behaviour [12]. Such nutritional zinc deficiency impedes neurogenesis, neuronal migration and synaptogenesis during brain development [13]. Zinc deficiency in adults has been clinically linked to psychological disorders such as depression [14,15], attention-deficit/hyperactivity disorder (ADHD) [16,17], and autism [18], as well as neurodegenerative diseases such as Parkinson’s disease [19–21].

Zinc’s importance in the brain is due to its high protein binding capacity [22], which serves critical roles in the structural stability, catalytic activity and regulatory function of thousands of proteins. In the past few decades, the focus of biological study on zinc has expanded from its structural and catalytical roles to regulatory function and its signalling roles have been extensively investigated, especially in the immune system and central nervous system [23–25]. This review focuses particularly on neurons, the fundamental units that generate electrical and chemical signals in the brain.

Neuronal signalling is partially mediated by free or labile zinc (referred to as Zn2+ ) via intracellular and intercellular signalling pathways. Inside neurons, the majority of zinc is present in a protein-bound state, leaving only a subnanomolar concentration of labile Zn2+ in the cytosol [26–29]. Intracellular Zn2+ homeostasis is efficiently maintained at stable levels with buffering and muffling mechanisms involving Zn2+ transporters, metallothioneins and metal-responsive transcription factor-1 [26,27,30,31]. A high concentration of Zn2+ is stored in the synaptic vesicles of certain neurons [28,29], which can be released into the synaptic cleft and act as intercellular signalling molecules when neurons are activated by physiological stimuli [31,32]. In addition, cytosolic Zn2+, via liberation from intracellular stores, can mediate a series of intracellular signalling events. In this review, we will provide an overview

© 2022 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.

Table 1. List of some sensors used to measure and detect neuronal Zn2+ signals.

localization name dissociation constant (Kd) application references cytosol ZP1 0.7 nM hippocampal slices [35,42–44]

primary neurons ZP3 0.7 nM hippocampal neurons and slices [30,45] ZnAF-2DA 2.7 nM hippocampal slices [46] FluoZin-3 AM 15 nM primary neurons [33,47–51]

hippocampal slices GZnP3 1.3 nM primary neurons [52] mitochondria RhodZin-3 65 nM primary neurons [39,53–57] PC12 cells

Mito-ZapCY1 1.6 pM C. elegans PVD neurons [58,59] lumen of synaptic vesicles

TSQ 155 nM–48 µM hippocampal slices [35,60–66] Zinquin 620 nM primary neurons [60,64–67] SpiroZin-2 3.6 nM hippocampal slices [51,68]

and acid compartments

FluoZin-3 15 nM hippocampal slices [33,69,70] ZP4 0.65 nM hippocampal slices [71–73]

extracellular regions (synaptically released Zn2+)

primary neurons ZnAF-2 2.7 nM hippocampal slices [34,42,46,74–77] NewPort Green DCF 1 μM hippocampal slices [78–81] LZ9 0.57 nM coronal brain slices containing

[82–84]

dorsal cochlear nucleus

2

royalsocietypublishing.org/journal/rsobOpen Biol.12: 220188

of neuronal Zn2+ signalling, both intracellularly and intercellularly. We will first summarize the fluorescent sensors that have been used to illuminate Zn2+ signals within subcellular compartments of neurons and synapses. Next, we will discuss the current understanding about synaptic Zn2+ signalling in the brain. We will then move onto discussing different players that modulate intracellular Zn2+ signals, followed by a discussion of the targets and function of Zn2+ in intracellular signalling pathways.

The prerequisite of a signalling molecule is that its concentration can fluctuate alongside physiological events. Fluorescent sensors, along with time-lapse fluorescence microscopy, have been enormously instrumental in revealing spatio-temporal dynamic changes of Zn2+ signals in neurons and brain tissues. Fluorescent Zn2+ sensors involve a Zn2+binding unit with one or two fluorescent components so that the sensors can display changes in spectral properties in response to the binding of Zn2+ ions. These tools include synthetic small molecule sensors that are constructed by organic chemistry and genetically encoded sensors that are constructed by molecular protein engineering. Cell impermeable small molecule sensors can be used to detect extracellular Zn2+ signals in brain slices, while detection of intracellular Zn2+ can be achieved by the addition of an ester moiety, which allows sensors to cross the plasma membrane into cells, followed by cleavage via intracellular esterases [33,34]. However, incomplete hydrolysis of these esters can cause nonspecific localization of small molecule sensors to

subcellular compartments [35]. Small molecule sensors generally display a very large dynamic range and are easy to use in comparison to genetically encoded sensors. Genetically encoded sensors, on the other hand,aresuperior in spatio-temporal detection because they can be precisely targeted to specific cell types and subcellular compartments [36–38]. Genetically encoded sensors are ideal for long-term imaging due to their ability to be retained in cells for days to weeks [36,39]. Please see these review papers [38,40,41] for a comprehensive overview of different types of Zn2+ sensors. Here we will focus on the sensors that have been successfully used to detect dynamic changes in Zn2+ signals in neuron cultures and brain slices (table 1).

Steady-state Zn2+ concentration in the cytoplasm of mammalian cells is maintained at a subnanomolar range (100 pM– 1 nM) [33,85–87]. To detect cytosolic Zn2+ signals changing from baseline concentrations to high nanomolar concentrations, we need sensors with nanomolar sensitivity which is determined by both binding affinity and dynamic range. A number of cell permeable small molecule sensors with nanomolar affinity have been reported including ZnAF2DA (Kd: 2.7 nM) [88], ZP1 (Zinpyr-1, Kd: 0.7 nM) [89], ZP3 (Kd: 0.7 nM) [45], ZP4 (Kd: 0.65 nM) [71] and FluoZin-3 acetoxymethyl (AM) ester (Kd: 15 nM) [47]. Among these sensors, FluoZin-3 AM is one of the most popular sensors used to detect cytosolic Zn2+ signals because it displays large response to Zn2+ [47]. The fluorescence of FluoZin-3 is not affected by other cations such as Ca2+ and Mg2+, which allows simultaneous recording of Zn2+ and Ca2+ by using FluoZin-3 alongside the Ca2+ dye Fura Red [48], Fura-2FF [49,50] or Fura-6F [50] in primary cultured neurons and hippocampal slices. FluoZin-3 also has low sensitivity to changes

in pH, with its signal remaining unchanged from pH 6 to pH

1. 9 [48,90]. FluoZin-3 has successfully been used with pH sensor pHrodo Red to record dynamic changes in Zn2+ and pH simultaneously in cultured hippocampal neurons [48]. One limitation of using small molecule sensors to detect cytosolic Zn2+ in neurons is that their nonspecific subcellular localization generates variable fluorescence between the cytosol and bright punctate compartments, making it difficult to interpret the subcellular locations of Zn2+ signals. Genetically encoded sensors with more specific subcellular localization allow distinguishing Zn2+ signals coming from various subcel-

lular compartments. For example, GZnP3 (Kd: 1.3 nM) has been used to show that endolysosomal vesicles can release Zn2+ into the cytosol in hippocampal neurons [52].

Measurement of Zn2+ in the mitochondrial matrix by different sensors has determined that mitochondrial Zn2+ concentration is several orders of magnitude lower than the cytosol, approximately 0.2–300 pM [58,85,91–94]. RhodZin-3 (Kd: 65 nM) is the first reported mitochondrial Zn2+ sensor, which has a 75-fold fluorescence change from quenched N,N, N0,N0-tetrakis(2-pyridinylmethyl)-1,2-ethanediamine (TPEN) to saturated Zn2+ levels in vitro [53,54]. The positive charges carried by RhodZin-3 allow it to follow the electrical gradient and accumulate in the mitochondria [54]. RhodZin-3 has been used to record increases in mitochondrial Zn2+ following ischemia [53], N-ethylmaleimide treatment [55] or Aβ42 incubation [39,56] in primary cortical and hippocampal neurons. Mitochondrial Zn2+ was also detected in the neuron-related PC12 cells using RhodZin-3 [57]. However, due to its positive charge, RhodZin-3 signals are reduced in mitochondria when the mitochondrial membrane is depolarized, which limits its proper localization [39,56]. Other small molecule Zn2+ sensors targeted to mitochondriainclude DA-ZP1-TPP[95]andZP1BG which requires co-transfection with mitochondrial-targeted alkylguaninetransferase [96], but these probes have not been widely used in live cells. Three genetically encoded mitochondrial Zn2+ sensors based on fluorescence resonance energy transfer (FRET) have been created with various binding affinities: mito-ZapCY1 (Kd: 1.6 pM, [58]), mito-eCALWY-4 (Kd: 60 pM, [92]), and mito-eZinCh-2 (Kd: 5–10 pM, [91]). These ratiometric probes are useful in quantification of mitochondrial Zn2+ concentrations. For example, mito-ZapCY1 has been used to measure mitochondrial Zn2+ in C. elegans posterior ventral dorsal (PVD) neurons [59]. Mitochondrial intermembrane space (IMS) is separated from mitochondrial matrix by the inner mitochondrial membrane, where oxidative phosphorylation occurs. A single fluorescent protein-based Zn2+ sensor, GZnP2, has been targeted to the mitochondrial matrix and the IMS [85], demonstrating that the concentration of labile Zn2+ in the IMS is 100 pM, revealing differences in Zn2+ concentration across inner mitochondrial membrane by three orders of magnitude [85].

Synaptic vesicles are unique secretory organelles found in neurons, which are categorized by the different neurotransmitters they contain. High concentrations of Zn2+ are concentrated in the synaptic vesicles of certain glutamatergic neurons [29,31,97–100], but also in some glycinergic and GABAergic neurons in neocortex [101,102], hippocampus [100,103–111], amygdala [31,107], auditory brainstem [112,113] and spinal cord [114,115]. The vesicular Zn2+ concentrations are estimated to be in the high nanomolar to low millimolar range [11,29,98,109,116–121]. This pool of Zn2+ inside the synaptic vesicles was visualized by small molecule

sensors such as TSQ (Kd: 155 nM to 48 µM), which is a quinoline-based membrane-permeable sensor [60,61]. TSQ is lipophilic making it ideal for measuring vesicular Zn2+ in brain tissue [35,62,63]. TSQ’s derivative, Zinquin (Kd: 620 nM) was developed to increase cellular retention [64]. However, both TSQ and Zinquin were found to coordinate non-labile Zn2+, that is pre-bound with proteins [60,65] and low molecular weight ligands (glutamic acid, glutathione, histidine and ATP) [66]. Recently, a pH insensitive sensor SpiroZin2 (Kd: 3.6 nM) was used to detect vesicular Zn2+ in hippocampal mossy fibres [68] and lysosomal Zn2+ in lactating mouse mammary epithelial cells [51].

Synapse activity causes Zn2+ to release along with neurotransmitters, subsequently increasing extracellular Zn2+ concentration in the synaptic cleft. Direct quantitative measurement of such brief (within a few milliseconds) [122] and localized Zn2+ signals within the synaptic cleft remains challenging, but a large body of evidence has detected synaptically released Zn2+ and estimated that extracellular Zn2+ can increase from 1–20 nM baseline concentration to high micromolar concentration [34,78,98,123–126]. The cell membrane impermeable version of Zn2+ sensors such as FluoZin-3 (Kd: 15 nM), ZP4 (Kd: 0.65 nM) and ZnAF-2 (Kd: 2.7 nM) have been used to examine Zn2+ release at hippocampal mossy fibre synapses [34,42,46,71–77,101,126,127]. By utilizing different affinities of sensors ZnAF-2 and ZnAF-3 (Kd: 790 nM), it has been demonstrated that membrane depolarization induces differential amounts of Zn2+ release in the hippocampus, with the highest in the dentate gyrus compared to CA1 and CA3 [34]. Cell impermeable Newport Green DCF has low affinity for Zn2+ (Kd: 1 µM), making it ideal for detecting high concentrations of Zn2+, and it has been used to visualize vesicular Zn2+ release in the hippocampal hilus [78,79]. The ratiometric Zn2+ sensor LZ9 (Kd = 0.57 nM) was developed for more precise quantification of extracellular Zn2+ to correct for the varieties in tissue thickness, sensor concentrations and imaging acquisitions [82]. LZ9 was designed by linking a green Zn2+ sensor with lissamine rhodamine B (LRB), which is a Zn2+-insensitive red fluorophore [82,83]. With LZ9, the extracellular Zn2+ concentration in the mice dorsal cochlear nucleus (DCN) was measured and detected under electrical stimulations [82].

3

royalsocietypublishing.org/journal/rsobOpen Biol.12: 220188

The presence of abundant Zn2+ in the synaptic vesicles of zincergic neurons has been confirmed by histochemical staining [29,31,97–100,128], electron microscopy [111], and microscopy imaging using the fluorescent sensors discussed in the previous section. This pool of Zn2+ is concentrated into synaptic vesicles through vesicular transporter ZnT3 [129,130] (figure 1), and co-released with neurotransmitters, such as glutamate, to the synaptic cleft during physiological neuronal excitation [31,101,131]. Synaptically released Zn2+ has been suggested to act via phasic mode, which refers to the situation that free Zn2+ immediately increases in the synaptic cleft, diffuses away from released sites and acts on target cells [118,132–135]. There is evidence that such synaptic Zn2+ can diffuse into extrasynaptic regions during repetitive synaptic stimulations [82]. Synaptically released Zn2+ acts as an intercellular signalling molecule to regulate activity of presynaptic or postsynaptic neurons, astrocytes

2

1

1. 3
2. 4

6

5

7

= Zn2+

8

1. 10
2. 11

9

1. 1 NMDAR ZnT3 ZnT9 MT3 VGCC
2. 2
3. 3
4. 4
5. 5

13

4

royalsocietypublishing.org/journal/rsobOpen Biol.12: 220188

and microglia cells [118,133–135] by targeting a variety of ionotropic and metabotropic receptors located on cells. The neurobiological roles of synaptic Zn2+ have been investigated extensively by three strategies: (a) eliminating Zn2+ inside synaptic vesicles using ZnT3 knockout mice created by Dr Richard Palmiter’s team [130], (b) chelating extracellular Zn2+ with membrane-impermeable chelators such as CaEDTA, Tricine and ZX1[84,136–140], or (c) mutating the Zn2+ binding sites on the receptor proteins such as glycine receptor alpha1 subunit [140] and glutamate receptor GluN2A subunit [98]. In addition to the transient high concentrations of synaptic Zn2+, low concentrations of ambient extracellular Zn2+ (less than 10 nM) might also act as a signalling molecule via tonic mode. Such ambient Zn2+ was suggested to derive from accumulation of synaptically released Zn2+ that is still coordinated with membrane proteins [141], or from efflux of cytoplasmic Zn2+ in postsynaptic cells by Zn2+ transporter ZnT1 [142] (figure 1). Different results were reported regarding whether tonic Zn2+ affects cell excitability, which might be due to the differences in the synapses chosen to study, Mg2+ concentration in the experimental buffers, or chelators used [82,98].

A number of ionotropic receptors can be regulated by Zn2+ including N-methyl-D-aspartate receptors (NMDARs), α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptors (AMPARs) without GluA2 subunit [143], gammaaminobutyric acid receptors (GABARs), glycine receptors, L-type and N-type voltage-gated calcium channels (VGCCs),

voltage-gated sodium channels, voltage-gated potassium channels, P2X purinergic receptors and transient receptor potential ankyrin 1 (TRPA1) channels [144–149] (figure 1). Zn2+ displays differential modulation (inhibition, excitation or activation) of these targets with varying potency depending on target isoform types, Zn2+ concentration and synaptic activity [138]. For example, NMDAR subtypes containing GluN2A subunits possess a nanomolar-affinity Zn2+ binding site, while GluN2B subunits have a micromolar-affinity site for Zn2+ [150–152]. For AMPAR, hundreds of micromolar Zn2+ demonstrates inhibition towards the AMPAR containing GluA2 subunits [153]. By tuning the gating of these ion channels, synaptically released Zn2+ modulates excitatory and inhibitory postsynaptic currents (EPSCs and IPSCs respectively), and hence synaptic activity [120,121,126,140,154]. The effects of synaptic Zn2+ on EPSCs have been studied in brain regions that are rich in glutamatergic Zn2+, including hippocampal mossy fibre-CA3, hippocampal Schaffer collateralCA1 synapses and DCN parallel fibre synapses, where it was discovered that presynaptic stimulation discharges Zn2+ that inhibits NMDAR EPSCs [11,82,98,109] and AMPAR ESPCs [84]. In addition, synaptically released Zn2+ inhibits GABAAR IPSCs in principal neurons in the lateral amygdala [120], but enhances GABAAR IPSCs in somatostatin interneurons in the auditory cortex [155] as well as glycinergic IPSCs in hypoglossal motoneurons of the brainstem [140,156,157].

Emerging studies also revealed the roles of synaptic Zn2+ in regulating metabotropic receptors. Hershfinkel’s group

demonstrated that synaptically released Zn2+ from mossy fibre stimulation interacts with a ‘Zn2+-sensing receptor’ (ZnR/ GPR39) on postsynaptic cells in CA3 of hippocampal slices (figure 1), which triggers the G protein-coupled receptor pathway and subsequently induces the release of Ca2+ from thapsigargin-sensitive intracellular pools [118,158]. Such Zn2+-evoked Ca2+ signals induce a series of MAPK-dependent cascades, promoting interaction between SNAP23 and K+/Cl− cotransporter 2 (KCC2), enhancing the surface expression and activity of KCC2 in hippocampal neurons [158,159]. The increased KCC2-mediated Cl− outward current maintains the hyperpolarizing GABAAR reversal potentials, reducing glutamate excitotoxicity in hippocampal neurons. This has further been supported by the finding that GPR39 knockout mice are more susceptible to kainate-induced seizures compared to wild-type groups [160]. The ZnR was also reported to be present in the DCN, where activation of ZnR by synaptic Zn2+ promotes synthesis of endocannabinoids, resulting in the reduction of presynaptic glutamate release in a retrograde manner [113]. However, ZnR/GPR39 has very low binding affinity for Zn2+ (Kd: 150 µM), raising the question of whether it acts as a physiological receptor of Zn2+. Early studies have suggested that Zn2+ exerts effects via tropomyosin receptor kinase B (TrkB), through either activating metalloprotease to increase mature brain-derived neurotropic factor (BDNF)

1. [161], or transactivating TrkB directly [154]. However, this hypothesis was challenged by inconsistent results in ZnT3 knockout mice. Although several studies showed that there is an increased level of BDNF in ZnT3 knockout mice [162–165], the TrkB protein levels in ZnT3 knockout mice have been reported to be increased [165], downregulated
2. [162,163], or unchanged [164,166]. Through modulating the activity of the vast array of recep-

tors, the signals of synaptic Zn2+ are transduced to impact brain function. For example, synaptically released Zn2+ facilitates long-term potentiation (LTP) at the cortico-amygdala synapses via depressing feedforward GABAergic inhibition of principal neurons [120]. Partially due to the roles of synaptic Zn2+ in amygdala synaptic plasticity, ZnT3 knockout mice showed deficiency in subtle and complex learning of fear [167]. Synaptic Zn2+ in the hippocampus was suggested to enhance LTP in the CA1 region [168] and modulate the mossy fibre LTP in the CA3 regions [11,109]. As hippocampal LTP is the basis of normal cognitive function, synaptic Zn2+ might be involved in learning and memory [77,169]. ZnT3 knockout mice showed mild deficits in long-term memory and spatial memory [170,171]. In addition, synaptically released Zn2+ in the auditory cortex and somatosensory cortex aids in sensory processes as ZnT3 knockout mice have been shown to have deficits in distinguishing different sound frequencies [136,172] and detecting fine textural differences [173].

Dynamic changes in intracellular Zn2+ concentration are tightly regulated in neurons by a multitude of membrane transporters, ion channels and buffering proteins, which control Zn2+ signals both spatially and temporally. The Zn2+ transporters include 10 efflux transporters (ZnTs, SLC30A) and 14 influx transporters (ZIPs, SLC39A). ZnTs extrude

Zn2+ into the extracellular space or intracellular compartments [31,129,130,174,175], while ZIPs allow Zn2+ inflow by or coupled with H+ or HCO3− gradients [176–179]. The expression of Zn2+ transporters has distinct localization and roles in various brain areas and subcellular compartments. ZnT1 (SLC30A1), which is primarily localized on the plasma membrane [180–182], was shown to be closely tied with NMDA receptors at postsynaptic densities [183] (figure 1). ZnT3 (SLC30A3) is localized on the membrane of synaptic vesicles and highly co-localized with synaptic Zn2+ [31,129,130,174,175] (figure 1). ZnT9 (SLC30A9) is a mitochondrial Zn2+ transporter which maintains low Zn2+ inside mitochondria [184] (figure 1). TMEM163 is another Zn2+ transporter expressed in the synaptic vesicles and has been categorized as ZnT11 [185]. Both ZIP1 and ZIP3 are ubiquitous plasma membrane transporters, but they display distinct expressions in the hippocampus [178,186] (figure 1). ZIP1 is enriched in the stratum pyramidale of CA3, while ZIP3 is highly expressed in dentate gyrus granule cells [187]. ZIP4 has high expression in the soma and postsynaptic regions of Purkinje neurons in the cerebellum [188].

Metallothioneins (MTs) are a family of cysteine-rich (30%) proteins, including four human isoforms (MT1, MT2, MT3, MT4), which play critical roles in regulating gene expression, controlling cellular metal metabolism and adjusting cellular adaptation to stress [189]. The MT3 protein consists of two subdomains (α and β domain), and has the capacity of binding up to seven Zn2+ ions [190] (figure 1). MT3 is the brain-specific MT, responsible for buffering cellular Zn2+ in neurons and astrocytes [191–194]. The high amounts of cysteine residues in MTs are accessible to modification by reactive oxygen or nitrogen species, thus liberating Zn2+ to the cytosol [195–197]. Compared to MT1 and MT2, MT3 demonstrated higher reactivity to release more Zn2+ when treated with S-nitrosothiols due to the enhanced nitrosylation of multiple cysteines adjacent to basic and acid residues in MT3 [198]. Accordingly, exogenous nitric oxide increases intracellular Zn2+ in neurons [199]. In addition, cytosolic acidification following Ca2+ influx has been suggested to induce intracellular Zn2+ release in neurons [48,49].

Labile Zn2+ is maintained at nanomolar concentration (approx. 20 nM) in cerebrospinal fluid (CSF) in the normal brain [200], while the concentration of Zn2+ can reach up to 15 000 fold (approx. 300 µM) in the synaptic cleft under both spontaneous and electrically stimulated conditions [109,179,200,201]. It is still unclear whether the synaptically released Zn2+ can flux into neurons, but immediate increases in cellular Zn2+ can be mediated through opening of ion channels on the plasma membrane or lysosomal membrane. Both fluorescence-based imaging techniques and electrophysiology have provided strong evidence that activated voltage-gated Ca2+ channels are permeable to Zn2+ in mammalian neurons [147,149,202–204]. Recent electrophysiology and Zn2+ fluorescent imaging studies show that some TRP subfamilies are Zn2+ permeable [205]. TRPA1 channels, which are mainly located in dorsal root ganglia neurons, are permeable to Zn2+ [146]. Another TRP channel, transient receptor potential canonical type 6 (TRPC6), demonstrates Zn2+ permeability in mouse cortical neurons and contributes to nuclear Zn2+ accumulation [206]. In addition, the lysosomal channel transient receptor potential mucolipin 1 (TRPML1) mediates Zn2+ release from late endosomes in primary rat neuron culture [52] (figure 1).

5

royalsocietypublishing.org/journal/rsobOpen Biol.12: 220188

Increases in cellular Zn2+ signals might act similarly to Ca2+ signals, mediating a cascade of phosphorylation signalling events that involve protein kinases and proteases. The enzymes that can be regulated by intraneuronal Zn2+ should have a half-maximal inhibitory concentration (IC50) within physiological range of Zn2+ concentrations (high picomolar to low nanomolar). Several enzymes demonstrate this high Zn2+ binding affinity in vitro including tyrosine phosphatase beta activity (IC50 = 21 pM) [207], caspase 3 (IC50 <

1. 10 nM) [208], and Ca2+-ATPase (IC50 = 80 pM) [209]. Tyrosine phosphatases are expressed in neurons and play a role in modifying synaptic formation as well as neuronal development [210]. Intraneuronal Zn2+ release, mediated by nitric oxide, was found to activate p38 MAPK, resulting in apoptosis in cortical neurons [199]. Zn2+ reuptake into the hippocampal mossy fibre terminal was suggested to inhibit MAPK tyrosine phosphatase, thereby inducing Erk activation [211]. However, a recent study in HeLa cells and mouse hippocampal neuronal cell line (HT-22) suggests that nanomolar Zn2+ activates Erk and Akt signalling pathways via the upstream molecule Ras, while activation of Erk phosphatase requires a Zn2+ concentration higher than is reached by physiological fluctuations [212]. In addition, members of the MAPK pathway proteins (e.g. MAPK1, MAPK4, Fibroblast growth factor receptor 3, Fibroblast Growth Factor Receptor Substrate 2, Rap guanine nucleotide exchange factor 2, c-Jun) were also upregulated at the transcriptional level by increases in intraneuronal Zn2+ [213].

The Shank proteins are another target of intraneuronal Zn2+ signals and the interaction between Zn2+ and Shank proteins regulates synapse formation and function. The Shank proteins, a family of scaffolding proteins located in the excitatory synapses of neurons, consist of three members (Shank1, Shank2 and Shank3) [214,215] (figure 1). Shank protein is a ‘master’ scaffolding protein, interacting and coordinating with intermediate scaffolding proteins at the postsynaptic regions [214]. Shank proteins indirectly interact with NMDA receptors by linking PSD95 to guanylate kinase (GK)-associated proteins [216,217]. There is a direct interaction between AMPAR’s GluA1 subunit and Shank3’s PDZ domain [218]. In addition, Shank assists synaptic growth and maturation through affecting the internalization of the transmembrane Wnt receptor Frizzled-2 (Fz2) [219]. It has been well studied that internalization and cleavage of Fz2 play an important role in modulating synapse development [220,221]. Among the three members, Shank2 and Shank3 contain a sterile-alpha-motif domain at their C-terminal that can bind Zn2+, while Shank1 is insensitive to Zn2+ [222,223]. Binding of Zn2+ is required for oligomerization and assembly of both Shank2 and Shank3 [223,224], essential for their proper postsynaptic localization during synaptogenesis and synapse maturation [225]. The Zn2+ sensitive Shank proteins are expressed before Shank1 in neurons, resulting in a difference in Zn2+ sensitivity between young neurons and mature neurons [225]. Neuron depolarizationinduced Zn2+ signals interact with Shank2 and Shank3,

recruiting AMPAR GluA2 subunits and removing GluA1 subunits at the glutamatergic synapses, hence enhancing AMPAR synaptic efficacy in young neurons [224,226] (figure 1). In mouse models of Zn2+ deficiency, Shank protein levels were significantly reduced in the striatum, hippocampus, cortex and cerebellum and the mice displayed developmental and behavioural issues mimicking autism spectrum disorders [227].

The surge in development of a full suite of new fluorescent Zn2+ sensors, chelators and genetic mice models allows us to gain a greater understanding about the neuronal signalling of Zn2+ in live primary neuron culture, brain slices and live animals. The dynamic changes in neuronal Zn2+ signals have been evidenced by the detection of extracellular Zn2+ signals during synaptic activity and intracellular Zn2+ signals during influx, neuron excitation and oxidative stress. Simultaneous measurement of intracellular Zn2+ concentrations and signalling molecules in live cells has confirmed that physiologically relevant Zn2+ dynamics regulate Erk signalling events. Behavioural studies in live mice with the application of fast and selective Zn2+ chelators along with genetic knockdown of ZnT3 further elucidated the involvement of synaptic Zn2+ signals in learning, memory, emotion, sensory function and social interaction.

However, it is still not completely unveiled when, where and how physiological changes in intracellular and intercellular Zn2+ signals can regulate their targets. We still lack tools that are sensitive and bright enough to track localized Zn2+ signals within synapses and neurons in whole organisms, preventing us from tackling many of these unanswered questions. Another challenge is that occurrence of Zn2+ signals is accompanied with the changes in other cellular signals such as Ca2+, pH and redox potential, which adds to the complexity of investigating the Zn2+ signalling roles. New sensors are needed to resolve these challenges. In addition, given that there are multiple zincergic neurons and Zn2+ targets, the function of synaptic Zn2+ signals in a specific brain region is hard to clarify from whole-body ZnT3 knockout mice. Conditional knockout of ZnT3 in specific brain regions and neuron types will further elucidate the roles of synaptic Zn2+ in the brain.

6

royalsocietypublishing.org/journal/rsobOpen Biol.12: 220188

Data accessibility. This article has no additional data.

Authors’ contributions. C.Z.: writing—original draft, writing—review and editing; A.D.: writing—original draft, writing—review and editing; K.G.: writing—original draft, writing—review and editing; Y.Q.: conceptualization, funding acquisition, project administration, resources, supervision, validation, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Conflict of interest declaration. We declare we have no competing interests. Funding. This work was funded by the NIH R01NS110590.