Intracellular dialysis disrupts Zn2+ dynamics and enables selective detection of Zn2+ influx in brain slice preparations


Address correspondence and reprint requests to C. W. Shuttleworth, PhD, Department of Neurosciences, University of New Mexico School of Medicine, Albuquerque, NM 87131-0001, USA. E-mail:


We examined the impact of intracellular dialysis on fluorescence detection of neuronal intracellular Zn2+ accumulation. Comparison between two dialysis conditions (standard; 20 min, brief; 2 min) by standard whole-cell clamp revealed a high vulnerability of intracellular Zn2+ buffers to intracellular dialysis. Thus, low concentrations of zinc-pyrithione generated robust responses in neurons with standard dialysis, but signals were smaller in neurons with short dialysis. Release from oxidation-sensitive Zn2+ pools was reduced by standard dialysis, when compared with responses in neurons with brief dialysis. The dialysis effects were partly reversed by inclusion of recombinant metallothionein-3 in the dialysis solution. These findings suggested that extensive dialysis could be exploited for selective detection of transmembrane Zn2+ influx. Different dialysis conditions were then used to probe responses to synaptic stimulation. Under standard dialysis conditions, synaptic stimuli generated significant FluoZin-3 signals in wild-type (WT) preparations, but responses were almost absent in preparations lacking vesicular Zn2+ (ZnT3-KO). In contrast, under brief dialysis conditions, intracellular Zn2+ transients were very similar in WT and ZnT3-KO preparations. This suggests that both intracellular release and transmembrane flux can contribute to intracellular Zn2+ accumulation after synaptic stimulation. These results demonstrate significant confounds and potential use of intracellular dialysis to investigate intracellular Zn2+ accumulation mechanisms.

Abbreviations used

ethylene glycol tetraacetic acid


field excitatory postsynaptic potentials


metallothionein 3




wild type



Zn2+ is an essential ion required for a wide range of functions in mammalian cells. While total intracellular Zn2+ content is quite high (> 200 μM), cytoplasmic free Zn2+ concentrations are maintained at extremely low levels (< 1 nM), because of the fact that much is bound in structural proteins and also the high activities of intracellular buffer and transporter systems (Colvin et al. 2008; West et al. 2008; Sensi et al. 2009). However, cytosolic Zn2+ transients appear to be important for intracellular signaling. For example, elevated cytoplasmic free Zn2+ levels in neurons have been implicated to modulation of neuronal circuit activity, and activation of neurotoxic pathways when intracellular Zn2+ levels become excessively high (Choi and Koh 1998; Frederickson et al. 2005).

In the mammalian brain, Zn2+ is highly concentrated in synaptic vesicles of glutamergic neurons, because of the activity of the vesicular Zn2+ transporter ZnT3 (Cole et al. 1999). Vesicular Zn2+ can be released as a neuromodulator and can directly modify the function of ion channels and receptors via direct interactions. In addition to extracellular actions, released Zn2+ may also translocate into post-synaptic neurons and potentially contribute to plasticity of some synapses (Huang et al. 2008). Glutamate exposures have been widely used to study neuronal intracellular Zn2+ homeostasis (Sensi et al. 2002, 2003; Dineley et al. 2008; Kiedrowski 2011) and activation of NMDA-type glutamate receptors (NMDARs) have been shown to release Zn2+ from intracellular pools. Synaptic Zn2+ release and influx has been reported to contribute to post-synaptic Zn2+ accumulation (Suh 2009); however, there is not yet evidence for liberation from intracellular stores by endogenous glutamate release. The relative contributions of these two Zn2+ sources following synaptic stimulation remain to be clarified, as does the impact of standard electrophysiological recording methods on intracellular Zn2+ signals.

The whole-cell clamp recording technique results in substantial dialysis of the intracellular compartment, because of large differences in pipette and intracellular volumes (> 10−6 vs. 10−10˜12 L). As a result of effective washout of some intracellular components, intracellular dialysis can lead to rundown of Ca2+ currents (Sakmann and Neher 1984) and mask important neurophysiological responses such as long-term potentiation (Malinow and Tsien 1990). Zn2+ signaling to NMDARs has also been shown to be disrupted by extended dialysis in cultured cortical neurons (Manzerra et al. 2001). On the other hand, intracellular dialysis has been also exploited as a valuable method to manipulate intracellular constituents (Blatow et al. 2003; Eggermann and Jonas 2012). Whether whole-cell recording depletes neurons of important Zn2+ buffers and/or otherwise modifies detection of intracellular Zn2+ responses has not been explicitly tested.

In this study, we demonstrate a significant vulnerability of intracellular Zn2+ buffers and/or pools to intracellular dialysis. Although these dialysis methods may be a significant technical confound, we also demonstrate that they can be exploited to evaluate contributions of both synaptic and intracellular Zn2+ release, following synaptic stimulation.


Slice preparation

All procedures using experimental animals were approved by the Institutional Animal Care and Use Committee of the University of New Mexico. Brain slices were prepared from 4 to 10-week-old WT and ZnT3 KO C57BL/6 animals of both sexes. Data in each specific experiment were collected from matched numbers of each sex, within an age range of 2 weeks. ZnT3 KO animals were originally developed by (Cole et al. 1999) and backcrossed onto the C57BL/6 line for at least 13 generations. Both WT and ZnT3 KO homozygote colonies were established and maintained at the University of New Mexico.

Mice were deeply anesthetized with a subcutaneous injection (0.2 mL) of ketamine/xylazine mix (85 mg/mL and 15 mg/mL, respectively) and decapitated. Brains were carefully extracted into ice-cold cutting solution (in mM: 220 sucrose, 1.25 NaH2PO4, 25 NaHCO3, 3 KCl, 10 glucose, 0.2 CaCl2, 6 MgSO4 equilibrated with 95% O2/5% CO2 gas), hemisected, and sliced at 350 μm thickness with a vibratome (vibratome 3000, Ted Pella Inc., Redding, CA, USA). Slices were allowed to recover in artificial cerebrospinal fluid (ACSF, in mM: 124 NaCl, 1.25 NaH2PO4, 25 NaHCO3, 3 KCl, 2 CaCl2, 1 MgSO4 10 glucose equilibrated with 95% O2/5% CO2 mix-gas) for 1 h at 35°C, and were subsequently maintained at 20–22°C in ACSF. Slices were transferred to a recording chamber (RC-27; Warner Instruments, Hamden, CT, USA) and superfused with ACSF at 2 mm/min and 32°C.

Zn2+ indicator loading into single CA1 neurons

Intracellular Zn2+ dynamics were evaluated using the high affinity indicator FluoZin-3 (Life technologies, Carlsbad, CA, USA), loaded via whole cell dialysis into single CA1 pyramidal neurons. Neurons were visually identified, and patch pipettes (3-5 MΩ) contained (in mM): 135 potassium gluconate, 8 NaCl, 1 MgCl2, 2 Na2ATP, 0.3 NaGTP, 10 Hepes, 0.05 EGTA. pH was adjusted to 7.2 with KOH and FluoZin-3 added to the pipette solution.

A central issue in this study was the influence that whole-cell dialysis had on intracellular Zn2+ dynamics. Therefore, the duration of dialysis was carefully monitored, and the electrode withdrawn from the neuron at specified times after the initial establishment of the whole-cell configuration. Thus, membrane rupture was determined as the time when initial access resistances dropped below 30 MΩ, and if this was not completed within 10 s of initial attempts, the neuron was discarded. During intracellular dialysis, neurons were voltage clamped at −65 mV (holding current range between −50 and +50 pA), and the quality of whole-cell configuration was monitored based on the holding current and membrane response to test pulse (−5 mV, 100 ms). Neurons were discarded when holding current exceeded −100 pA for more than 20 s without any sign of recovery, or when series resistance exceeded 30 MΩ. Following intracellular dialysis, the loading pipette was carefully withdrawn. Successful electrode withdrawal was determined by formation of an out-side-out recording configuration, and could be achieved within 20 s. After successful electrode withdrawal, neurons were allowed 20 min recovery, prior to onset of any stimulation.

The concentration of FluoZin-3 added to the pipette solution depended on the duration of dialysis, to approximately match the final FluoZin-3 concentration achieved in neurons (see Fig. 1). The tested concentration ranged from 40 to 500 μM (see 'Slice preparation').

Figure 1.

Experimental approach for FluoZin-3 loading. The high affinity Zn2+ indicator FluoZin-3 was loaded into hippocampal CA1 pyramidal neurons in acute brain slices via patch pipettes. (a) Three different intracellular loading methods are illustrated; Brief dialysis [2 min, top], Standard dialysis [20 min, middle] and Standard dialysis supplemented with 0.5 μM MT3 [20 min, bottom]. Pipette concentrations of FluoZin-3 were adjusted to achieve similar final intracellular indicator concentrations with the different loading durations (500 μM or 40 μM, as indicated). (b) Representative images of a briefly dialyzed neuron, showing FluoZin-3 increases before and after challenge with a saturating concentration of the Zn2+ ionophore complex ZnPyr (100 μM ZnCl2 and 5 μM pyrithione). Scale bar: 40 μm. Similar challenges with ZnPyr were used to estimate basal intracellular Zn2+ concentrations shown in (c). (c) Comparisons of maximum fluorescence signals generated by saturating concentrations of ZnPyr (100 μM ZnCl2 and 5 μM pyrithione) in the three recording conditions. The left-hand axis shows recorded peak FluoZin-3 signals, and the right-hand axis shows estimated basal Zn2+ concentrations (see 'Methods'). No significant differences were seen (p > 0.5, n = 5 each).

In some experiments, recombinant human metallothionein-3 (MT3) was added to the intracellular solution. This was supplied as lyophilized purified recombinant human MT3, present as a mixture of Zn2+-bound forms [approximately 80% Zn7MT3, 10% of Zn6MT3, and 10% Zn8MT3 (Bestenbalt LLC, Tallinn, Estonia)]. A 5 μM MT3 stock solution was prepared as a 10-times concentrated pipette solution lacking ATP/GTP, and including 10% chelex resin (v/v, Chelex 100, Bio-Rad, Hercules, CA, USA). Chelex is an ion exchange resin and was used here to remove weakly bound Zn2+ from metallothionein as demonstrated previously (Krezoski et al. 1988). This stock was stored at −80°C, and MT3 and ATP/GTP were then added to the pipette solution immediately prior to experiments

Fluorescence imaging

FluoZin-3 fluorescence was excited with 495 nm light (120 ms) delivered from monochromator via a dichoric mirror (505 nm long pass). Emission signals were band-passed filtered (535/50 nm) and acquired using a CCD camera (Till Imago, TILL Photonics, Rochester, NY, USA) controlled by Till Vision software (version 4.04, TILL Photonics, Rochester, NY, USA). Intracellular fluorescence signals were calculated after subtracting background neuronal autofluorescence within the same images. Intracellular basal Zn2+ concentrations were estimated from the equation described in (Grynkiewicz et al. 1985): [Zn2+] = Kd (F − Fmin)/(Fmax − F), where Kd = 15 nM (Gee et al. 2002), Fmax was obtained after exposure to a saturating concentration of ZnPyr (see Fig. 1b), and Fmin was determined from TPEN exposures to be zero.

Because of the high signal to noise ratio achieved by single-cell loading with FluoZin-3, intracellular fluorescence values could be approximated by the maximum fluorescence values obtained from low-pass filtering (3 × 3 pixel averaging) images. The kinetics of responses were analyzed as changes relative to the basal fluorescence intensities. In some experiments (e.g., Fig. 1 100 μM ZnPyr exposures), significant tissue swelling occurred and focus adjustment was required.

Synaptic stimulation

Synaptic responses were evoked with a concentric bipolar stimulating electrode placed > 100 μm from the imaging site. Glass recording electrodes filled with ACSF (0.5–1 MΩ) were placed adjacent to neurons being imaged to verify activation of post-synaptic neurons near the recording sites. In each experiment, input–output curve were generated based on field excitatory post-synaptic potentials (fEPSP) evoked with single current pulses (70 μs, 0.1 Hz). 70% maximum stimulation was used for test stimuli. In all experiments, input–output curves were determined at least 10 min after recovery of neurons from indicator loading and removal of the filling electrode.


Unless otherwise noted, all chemicals were from Sigma Aldrich (St Louis, MO, USA). FluoZin-3 and TPEN (N,N,N′,N′-Tetrakis(2-pyridylmethyl)ethylenediamine were obtained from Life Technologies (Carlsbad, CA, USA). As noted above, recombinant MT3 was from Bestenbalt LLC.

Statistical analysis

All statistical tests were performed using Graph Pad Prism software (GraphPad Software, Inc., La Jolla, CA, USA, version 4.03). One-way anova with post hoc Newman–Keuls multiple comparison tests were used throughout. Values are presented as mean ± SEM. n values indicate numbers of cells tested. A p-value < 0.05 was considered statistically significant.


Intracellular dialysis increased detection of intracellular Zn2+ increases

We first examined the hypothesis that intracellular dialysis could reduce intracellular Zn2+ buffering capacity, and make cytosolic Zn2+ increases more readily detectable. The general experimental approach is shown in Fig. 1, where single pyramidal neurons were dialyzed via a conventional whole-cell patch pipette containing the membrane-impermeable Zn2+ indicator FluoZin-3. Two different durations of dialysis were compared (2 min and 20 min), and the concentration of indicator added to the pipette solution adjusted (500 μM and 40 μM, respectively) so that the final neuronal indicator concentrations were approximately matched. This was confirmed by exposing the indicator-loaded neurons with saturating concentration of ZnPyr (100 μM ZnCl2, 5 μM sodium pyrithione, 20 min exposure) and obtaining maximum fluorescence signals (Fig. 1a and b). The approximate intracellular Zn2+ concentrations were estimated using an equation described in Grynkiewicz et al. (1985), and the calculation confirmed extremely low resting intracellular Zn2+ concentrations (Fig. 1c). These initial experiments verified that initial FluoZin-3 concentrations were closely matched in the neuronal populations, despite very different dialysis durations.

Figure  2 shows that the duration of dialysis had a substantial effect on the amplitude of Zn2+ increases detected by FluoZin-3, when neurons were challenged with a low concentration of zinc pyrithione (ZnPyr: 1 μM ZnCl2 and 1 μM sodium pyrithione). Pyrithione serves to facilitate Zn2+ passage across the plasma membrane, and thereby increase intracellular Zn2+ levels independent of active transport mechanisms. As shown in Fig. 2a, the FluoZin-3 response in the briefly dialyzed neurons was barely detectable, whereas large FluoZin-3 increases were detected in all neurons that were first subjected to standard (20 min) dialysis (Fig. 2b). The increased signals with standard dialysis could be because of washout of an endogenous Zn2+ buffer into the dialysis pipette, and/or changes in transport mechanisms involved in accumulation and clearance.

Figure 2.

Intracellular dialysis strongly modifies detection of intracellular Zn2+ following exposure to ZnPyr. FluoZin-3-loaded neurons were exposed to 1 μM ZnPyr (1 μM ZnCl2 and 1 μM sodium pyrithione, 20 min), followed by 20 μM TPEN. (a–c) Plots of responses from five individual neurons, with either brief dialysis (a: 2 min), standard dialysis (b: 20 min) or standard dialysis supplemented with recombinant MT3 (c). (d) Quantitative analysis of peak FluoZin-3 responses. ***p < 0.005.

Experiments in Fig. 2c show that supplementation of the pipette solution with recombinant MT3 (0.5 μM) was sufficient to abolish the large Zn2+ signals seen with standard dialysis. The recombinant MT3 protein was initially supplied as a mixture of Zn2+ bound forms that is expected to retain significant Zn2+ binding capacity, as well as potentially providing a source of Zn2+ (see 'Methods' and 'Discussion'). We estimated relevant intracellular MT concentrations from previous publications (Hidalgo et al. 1994; Colvin et al. 2008) and examined effects of a range of MT3 concentrations (0.1–5 μM) in an initial set of pilot studies. 0.5 μM MT3 was then chosen for subsequent experiments, as this concentration showed significant effects on intracellular Zn2+ responses while having little deleterious effect on the quality of whole-cell recordings.

Although the difference in the amplitudes of response could be affected by initial fluorescence or Zn2+ concentration, our estimates of maximum fluorescence values as well as near zero minimum fluorescence values (after TPEN exposure) suggested this was not the case (see Fig. 1b and c). These observations imply that intracellular dialysis revealed larger FluoZin-3 signals because of increased cytoplasmic Zn2+ concentration available for detection by the indicator. Supplementation with recombinant MT3 is consistent with the possibility that washout of endogenous Zn2+ buffering proteins could underlie the effect; however, it is emphasized that addition of excess endogenous buffer in these studies could mask other contributing mechanisms (see 'Discussion').

Intracellular dialysis also reduced oxidation dependent intracellular Zn2+ release

Previous work has shown that addition of a membrane-permeable oxidant effectively mobilizes Zn2+ from intracellular stores/binding proteins, and leads to increases in Zn2+ that can be detected by cytosolic indicators (Aizenman et al. 2000). We therefore examined whether dialysis leads to depletion of the size of the oxidation-sensitive intracellular Zn2+ pool. After a stable baseline was collected, neurons were exposed to 200 μM 2,2′-dithiodipyridine (DTDP) for 20 min. As shown in Fig. 3, briefly dialyzed neurons showed a robust increase in FluoZin-3 signals. In contrast, FluoZin-3 signal responses were very small in neurons with standard dialysis, suggesting much smaller oxidation-sensitive Zn2+ pools in these preparations. Intermediate Fluozin-3 responses were observed in neurons with standard dialysis supplemented with recombinant MT3. These results suggest that intracellular dialysis may deplete oxidation-sensitive Zn2+ pools, and addition of Zn2+ bound recombinant MT3 can provide a Zn2+ pool in dialyzed neurons.

Figure 3.

Brief dialysis maintains oxidant-sensitive intracellular Zn2+ pool size. FluoZin-3-loaded neurons were challenged with ACSF containing 200 μM DTDP. After significant Zn2+ responses were obtained, neurons were then exposed to 20 μM TPEN. (a–c) shows responses from five individual neurons under the same conditions as described in Fig. 2, and (d) shows a quantitative analysis of peak responses. *p < 0.05, ***p < 0.005.

Dialysis allows dissection of multiple sources of Zn2+ following synaptic stimulation

The results above suggest that differences in dialysis conditions could be used experimentally to manipulate the ability to preferentially detect transmembrane Zn2+ influx (with standard dialysis) and liberation from intracellular binding sites (with short dialysis). We next examined whether these experimental approaches could be exploited to assess the contributions of different Zn2+ sources to intracellular Zn2+ accumulation following synaptic stimulation.

This was done using trains of synaptic stimulation (20 Hz for 10 s), as this was suggested to be a physiologically relevant stimulation intensity in a recent study of tissue metabolism in a similar preparation (Hall et al. 2012). As shown in Fig.  4, these stimuli provided reliable detection of post-synaptic Zn2+ accumulation. For these experiments, the slow Zn2+ chelator CaEDTA (1 mM) was included in recording bath solution to prevent detection of contaminating Zn2+ [see (Qian and Noebels 2005) and 'Discussion']. Based on previous studies, exposure to 1 mM CaEDTA should have little effect on basal intracellular Zn2+ concentration (Lavoie et al. 2007), and leave a significant fraction of rapidly released Zn2+ available at synaptic clefts (Vogt et al. 2000; Pan et al. 2011). Under these stimulation and recording conditions, slow Zn2+ increases were completely abolished by pre-exposure to a cocktail of glutamate receptor antagonists (20 μM DNQX, 5 μM D-AP5, 10 min), in both standard dialysis and brief dialysis conditions (see Figure S1).

Figure 4.

Manipulation of intracellular dialysis can be used to implicate both intracellular release and transmembrane Zn2+ flux to post-synaptic Zn2+ accumulation following synaptic stimulation. FluoZin-3 loaded neurons from WT and ZnT3 KO slices are shown, with three different intracellular dialysis methods. Following recovery slices were challenged with Schaffer collateral synaptic stimulation (20 Hz, 10 s). (a–f) shows individual responses obtained from multiple neurons in each preparation (n = 10 for brief dialysis, n = 5 for all others). Note that the acquisition rate was changed after 1.6 min (from 2 Hz to 0.4 Hz) in each recording. Peak responses were obtained following data-reduced traces and are compared in (g). ***p < 0.005

Figure  4 shows a summary of intracellular Zn2+ responses of post-synaptic neurons, indicator loaded with standard dialysis, brief dialysis and standard dialysis with MT3. To evaluate contributions of synaptic Zn2+ release, experiments were compared between WT and ZnT3 KO preparations. Strong genotypic differences were seen in the standard dialysis preparations (Fig. 4a and b). Thus, WT preparations showed a robust FluoZin-3 signal increase peaked during 1–2 min after stimulation and slowly decayed over next 5 min, while responses was virtually absent in ZnT3 KO preparations (Fig. 4a and b). These data suggest that the responses observed in dialyzed WT preparations were largely contributed to by presynaptic Zn2+ release, and are consistent with the possibility that significant depletion of intracellular Zn2+ buffering by standard dialysis facilitated detection of the response.

A large difference between WT and ZnT3 KO preparations was not seen in briefly dialyzed neurons. Thus, both WT and ZnT3 KO preparation showed intracellular Zn2+ responses in these cells, following synaptic stimulation (Fig. 4c and d). The responses in ZnT3 KO preparations raised the possibility that these responses were generated by liberation from intracellular sources.

Figure 4e and f show experiments to test whether artificial provision of an intracellular Zn2+ source and sink (by inclusion of MT3 in the pipette solution) could reveal additional Zn2+ release signals in neurons that had been extensively dialyzed. MT3 addition had no additional effect in WT neurons, but did reveal Zn2+ increases in ZnT3 KO neurons (compare Fig. 4c and f).

Taken together, these results suggest that synaptic stimulation leads to post-synaptic Zn2+ accumulation from at least two sources, which can be preferentially demonstrated with different dialysis methods. Synaptic release can be readily demonstrated after standard dialysis, where a large portion of the Zn2+ buffering system is lost. In contrast, briefly dialyzed neurons appear to retain a significant source of intracellular Zn2+, which can generate post-synaptic FluoZin-3 signals, even in the absence of synaptically released Zn2+.



This study examined effects of intracellular dialysis on Zn2+ measurements in neurons subjected to whole-cell recording in acute slice preparations. A main finding is that dialysis appears to effectively deplete intracellular Zn2+ buffering and decrease the size of oxidation-sensitive intracellular Zn2+ pools. Such disruption of intracellular Zn2+ homeostasis was shown to significantly modify detection of intracellular Zn2+ responses to a train of synaptic stimulation. Thus, standard whole-cell dialysis facilitated detection of synaptic Zn2+ translocation, whereas in briefly dialyzed preparations intracellular Zn2+ responses seem to be mediated mainly by intracellular Zn2+ liberation. Together, these findings indicate a high vulnerability of intracellular Zn2+ homeostasis to whole-cell dialysis, and demonstrate its potential use for selective detection of intracellular Zn2+ signals arising from different mechanisms.

Dialysis effects

This study compared effects of two different durations of intracellular dialysis; one standard (20 min) and one intentionally very brief (2 min). It is generally understood that intracellular dialysis is one of the most profound confounds of whole-cell clamp recordings. Washout of intracellular constituents and the imposition of a homogenous intracellular ionic composition improves the resolution of electrophysiological recordings; however, dialysis of channel subunits or signaling molecules can prevent recording of significant physiological responses (see ''). The present demonstration of significant disruption of intracellular Zn2+ homeostasis is another example of the significant impact of dialysis. The 20 min dialysis conditions tested here are relatively common for studies of synaptic physiology or pathophysiology. The current results suggest that loss of Zn2+ buffering and/or intracellular release could be a significant variable in a range of whole-cell studies.

One of the most obvious dialysis effects was the response to low concentrations of the Zn2+ carrier Zn-pyrithione. As noted above, pyrithione serves to facilitate Zn2+ passage across the plasma membrane, and thereby increases intracellular Zn2+ levels independent of active transport mechanisms. The fact that standard intracellular dialysis significantly increased intracellular accumulation following Zn-pyrithione could be because of washout of intracellular buffers, or possibly because of some other factors that decrease Zn2+ extrusion rates. Reversal of the dialysis effect with recombinant MT3 is consistent with the possibility that washout of endogenous Zn2+ buffering proteins could underlie the dialysis effect; however, increased endogenous Zn2+ buffer by MT3 inclusion could have masked dialysis effect on the other contributing mechanisms (e.g., decreased transporter/channel activity).

Likewise, the loss of intracellular Zn2+ accumulation following exposure of the oxidant DTDP is consistent with the hypothesis that dialysis washes out an oxidation-sensitive, diffusible Zn2+-binding source, such as MT3. A similar role for metallothionein in intracellular Zn2+ buffering and regulating the oxidation-sensitive pool size has previously been demonstrated with over-expression of metallothionein in astrocytes (Malaiyandi et al. 2004).

It is noteworthy that even in extensively dialyzed neurons, extremely low resting intracellular Zn2+ concentrations were detected by FluoZin-3 (estimated ~ 500 pM), which were not different from cells loaded with brief dialysis (see Fig. 1). This suggests that mechanisms required for maintaining resting Zn2+ concentrations are different from those that prevent excessive intracellular Zn2+ accumulation. Thus, while diffusible Zn2+ binding molecules (such as glutathione, thionein, and metallothioneins) are likely important defense molecules against severe Zn2+ influx (Cho et al. 2003; Krezel and Maret 2006), resting Zn2+ concentrations may not be under control of these molecules. It was recently reported that the functions of membrane Zn2+ transporters ZIP1 and ZIP3 are important in Zn2+ accumulation in CA1 pyramidal neurons (Qian et al. 2011), and those effects were observed in neurons with significant dialysis (30 min) implying that this pathway could remain intact. Thus, mechanisms such as Zn2+ extrusion or sequestration into organelles alone could potentially be sufficient for maintaining extremely low Zn2+ concentrations at rest (Colvin et al. 2008; Sensi et al. 2009).

In addition to depletion of buffer molecules, the concentrations of small signaling molecules such as inositol phosphate can be modified by intracellular dialysis (Hourez et al. 2005). Because we did not replenish these small molecules, intracellular dialysis could have significantly impaired intracellular signaling pathways. For example, it was reported that Zn2+ dependent NMDAR potentiation by Src kinase is abolished by intracellular dialysis in cultured cortical neurons (Manzerra et al. 2001). These effects certainly could have contributed to reduced detection of intracellular Zn2+ release in the dialyzed neurons, and facilitated selective detection of synaptic Zn2+ translocation in the dialyzed neurons.

In the current studies, dialysis was exploited to evaluate Zn2+ signals following synaptic stimulation, including influx from the extracellular space. However, it is recognized that the same dialysis methods will likely influence cytosolic Zn2+ transients arising from other sources that are resistant to dialysis. Such sources could include intracellular compartments such as mitochondria, endoplasmic reticulum, and lysosomes (see below) and Zn2+ transients arising from these sources may also be more readily detectable in extensively dialyzed cells.

Neuronal intracellular Zn2+ buffer systems

This study revealed that inclusion of MT3 alone was sufficient to restore a large portion of intracellular Zn2+ homeostasis. However, this does not necessarily rule out important contributions of other Zn2+ buffers. Glutathione (GSH) provides an additional major cytoplasmic Zn2+ buffer in hippocampus (Sato et al. 1984), but as GSH is less abundant in neurons (1 mM) compared with the glia (10 mM) (Rice and Russo-Menna 1998) this buffer may not play a major role in the neuronal signals examined here. In addition, it is known that GSH concentrations can be severely depleted during brain slice preparation (Rice 1999). These and other factors might have made contributions of MT3 dialysis relatively more detectable in the brain slice preparations studied here.

Zn2+ binding to MT3 can be quite dynamic despite the high affinity of MT3 for Zn2+ leading to the simultaneous detection of differently Zn2+-bound and -saturated forms (Palumaa et al. 2002, 2005). It has also been shown that MT3 contains weak Zn2+ binding sites which may become available in the presence of FluoZin-3 (Krezel and Maret 2007). This study examined effects of recombinant MT3 originally supplied as a mixture of Zn2+-bound forms (see 'Methods'). The fact that partially Zn2 +-saturated MT3 could act as both a sink (Fig. 2) and a source of Zn2+ (Fig. 3), is consistent with the idea that the protein remained only partially saturated with Zn2+ after dialysis, in FluoZin-3 containing conditions. A similar sink/source function of MT was previously demonstrated in astrocytes over-expressing MT2 (Malaiyandi et al. 2004) and has been suggested to explain differential effects of MT3 deletion in different injury models involving Zn2+ toxicity (see Discussion in Sheline et al. 2010).

The incomplete rescue by MT3 addition of Zn2+ responses in dialyzed neurons leaves open the possibility that the other Zn2+ sinks, such as mitochondria, endoplasmic reticulum, Golgi, and lysosomes (Sensi et al. 2009; Lee and Koh 2010), could contribute to shaping the FluoZin-3 transients seen here.

Intracellular Zn2+ responses during synaptic stimulation

Previous studies suggest that synaptic stimulation may elevate intracellular Zn2+ levels by two mechanisms; intracellular Zn2+ release and synaptic Zn2+ translocation (see ''). Bulk loading of populations of CA1 neurons with a low affinity Zn2+ indicator Newport Green (KD = 1–3 μM) (Li et al. 2001; Suh 2009) has shown intracellular Zn2+ increases in post-synaptic neurons (Li et al. 2001; Suh 2009) and the latter study showed that accumulation was abolished in ZnT3 KO tissues and by application of CaEDTA. Although this suggested a major role of synaptic Zn2+ release and translocation, the results of this study suggest that both synaptic release and intracellular release can contribute to postsynaptic Zn2+ accumulation at Schaffer collateral-CA1 synapses. Thus, in briefly dialyzed preparations, post-synaptic Zn2+ responses were observed in both WT and ZnT3 KO preparations. The presence of responses in ZnT3 KO preparations suggest that, in our recording conditions, intracellular Zn2+ release can significantly contribute to the FluoZin-3 signal changes following synaptic stimulation. Conversely with standard dialysis, FluoZin-3 signals were abolished in ZnT3 KO tissues, implying preferential detection of Zn2+ that is released and taken up by post-synaptic neurons. Taken together, the preferential detection of intracellular Zn2+ release in the briefly dialyzed neurons is likely contributed to by the presence of intracellular buffers, which bind with Zn2+ with significantly higher affinity and masked a large part of fluxed Zn2+ from detection by FluoZin-3. In addition, there was evidence for release of Zn2+ from recombinant MT3 by synaptic stimulation (compare Fig. 4b and f). This may explain why inclusion of this MT3 did not prevent detection of synaptic Zn2+ translocation in extensively dialyzed neurons. While it has been previously shown that glutamate can evoke intracellular Zn2+ release in neuronal culture models (Sensi et al. 2003; Dineley et al. 2008; Kiedrowski 2011), this study appears to be the first to suggest intracellular Zn2+ release during physiological synaptic activity.

A number of experimental differences may underlie the differences between this study, and the previous conclusion that synaptic translocation appeared entirely responsible for Zn2+ signals after electrical stimulation (Suh 2009). Relevant differences include the delivery of the higher affinity indicator FluoZin-3 into single neurons, and use of CaEDTA in the superfusate to prevent detection of contaminating Zn2+ (see Discussion in Carter et al. 2011). Post-synaptic Zn2+ transients observed here were much faster than previously reported by Suh 2009, and these experimental conditions appear to favor detection of a combination of intracellular mobilization, as well as trans-synaptic flux.

We also noted that post-synaptic Zn2+ responses were completely abolished when glutamate receptors were blocked (see Figure S1), regardless of the dialysis method used. These results are consistent with prior demonstration of glutamate receptor dependent intracellular release (Sensi et al. 2003; Dineley et al. 2008; Kiedrowski 2012) and Zn2+ influx through glutamate/depolarization gated channels (e.g., voltage gated Ca2+ channel, AMPAR, NMDAR) (Sensi et al. 1997, 1999, 2000; Kerchner et al. 2000; Huang et al. 2008).

Kinetics of Zn2+ responses

One of the remarkable features of intracellular Zn2+ responses following synaptic stimulation is their very slow kinetics. Both in standard dialysis and briefly dialyzed preparations, similar slow monophasic responses were detected. The intracellular Zn2+ responses were much slower than intracellular Ca2+ transients observed with the same stimuli (data not shown). Previous single-cell Zn2+ imaging during different stimuli (exposures to ouabain, oxygen glucose deprivation, and NMDA) have also shown relatively slow changes in intracellular FluoZin-3 signals (Dietz et al. 2008; Medvedeva et al. 2009; Vander Jagt et al. 2009). One possible explanation of the slow kinetics is that high affinity endogenous intracellular Zn2+ buffers limit mobility of Zn2+ ions and contributed to sluggish responses. However, our dialysis studies suggest this may not be a major contributor. In fact, similarly slow responses were also observed in the dialyzed neurons in which a large fraction of intracellular Zn2+ buffer is likely to be significantly depleted. Instead of large buffer molecules, these slow responses could be contributed by interactions with organic anions (e.g., HCO3, PO43−) which are abundantly present in the cytoplasm with high affinity (Rumschik et al. 2009). An interesting question is whether the observed responses may reflect true intracellular Zn2+ dynamics, or whether signals are distorted because of presence of fluorescence probes. Thus, a high affinity Zn2+ binding molecule such as FluoZin-3 could significantly impact the mobility of Zn2+. Because of extremely low intracellular Zn2+ concentrations at rest and even after stimulation, such a confound may be inevitable for imaging studies of intracellular Zn2+.


This study revealed significant effects on intracellular Zn2+ homeostasis by conditions used in standard electrophysiological experiments. The results also suggest that modifying whole cell indicator loading conditions can be valuable tool to help discriminate between different sources of Zn2+ that contribute to intracellular neuronal Zn2+ signals in adult brain slice preparations.


Supported by NIH grants NS051288 (C.W.S) & DK073446 (C.T.S), AHA grant 11PRE4870002 (I.A.). The authors thank Dr LD Partridge for helpful review of the manuscript. The authors declare that there is no conflict of interest.