Imaging Zinc: Old and New Tools

See allHide authors and affiliations

Science's STKE  13 May 2003:
Vol. 2003, Issue 182, pp. pe18
DOI: 10.1126/stke.2003.182.pe18


As recently as 20 years ago, all zinc in biological systems was believed to be tightly bound to proteins, and the idea of imaging zinc was considered heretical. Beginning with Maske's research with dithizonate staining of the hippocampus in the 1950s, however, zinc-sensitive dyes have indicated that, in mammalian cells, free zinc can exist in at least three separate pools. These pools include vesicular zinc sequestered in presynaptic vesicles and secretory granules, zinc released from these vesicles into the extracellular space after physiological stimulation, and transient increases in zinc in cells in the regions where extracellular release of zinc has occurred. This Perspective covers the zinc-imaging tools, from dithizonate to the newest FRET-based sensors, that have galvanized biomedical science.

Dark Ages

As recently as 20 years ago, the very idea of "imaging zinc" was heresy among the cognoscenti. This was because dogma held that all zinc in biological systems was tightly bound to the proteins where it functioned (or to zinc-chaperone molecules) and would therefore be impossible to image. The very idea of "free" zinc that could be imaged was, as late as the 1980s, simply not acceptable.

Free Zn2+ Is a Signal

Times have changed. It is now absolutely clear that at least a dozen types of mammalian cells sequester rather startling amounts of "free" zinc in their secretory granules and secrete that zinc in a regulated fashion, under the precise control of action potentials and secretagogues (1,2). Furthermore, it is becoming more and more clear that the secreted zinc is not just some epiphenomenon that accompanies secretion of the "real" messengers from these cells. Indeed, to paraphrase McLuhan, it appears that "the cationic medium is the message." Specifically, the brief "puffs" of ionic (rapidly exchangeable) Zn2+ that cells deliver into their immediate microenvironment are now recognized as pivotal and essential modulators and mediators of cell-to-cell signaling (3,4). Even more intriguing, cases are emerging in which the Zn2+ released by one cell travels through specific, gated, zinc-permeable channels in adjacent cells to enter the latter. This makes the Zn2+ ion an orthograde, transcellular, transmembrane signal, a completely novel signal type (5,6).

Zinc imaging tools were the agents that freed Zn2+ from proteins. Thus, as early as 1955, Maske [reviewed in (7)] published the odd result that the colorimetric zinc-binding reagent, dithizone, when injected intravitally, produced a rather stunning band of bright-red, apparent zinc:dithizonate staining in the hippocampal formation of the brain (Fig. 1). Similar vivid staining filled the islets of Langerhans in the pancreas and (unbeknownst to Maske) also stained odd places such as the secretory cells of the submandibular salivary gland (2), the prostatic fluid ducts (8), and other scattered types of secretory cells, all with apparent zinc:dithizonate. This result implied that loosely bound ("free") Zn2+ was richly abundant in select cell types and cell organelles in brain and other tissues.

Fig. 1.

Zn2+ in the hippocampal formation of the brain. (A) The hilus (H) of the dentate gyrus is shown stained for Zn2+ with dithizone, as discovered by Maske. (B) Two individual neurons from the hilus of the dentate gyrus stained for Zn2+ with ZP4 and imaged with confocal microscopy. Pseudocolor image shows that after excitotoxic injury (seizures, in this case), neurons fill with Zn2+, which is rich in the nucleus as well as the perinuclear cytosol. [Courtesy of Cathy Frederickson, NeuroBioTex, Inc.] (C) The Timm-Danscher method shows individual silver labeling of Zn2+ in the vesicles of a single, giant mossy fiber bouton of the hippocampus. [Courtesy of J. Perez-Clausell] Two such giant boutons are shown labeled with a fluorescent Zn2+ probe (ZP1) in the inset in the bottom row. ZP1 [courtesy S. Lippard, NeuroBioTex] stains all of the presynaptic Zn2+ in the various zinc-positive regions of hippocampal neuropil of a live slice. (D) The entire hippocampus with all of the different zinc-containing cytoarchitectonic regions fluorescing. (E) A confocal image centered on the stratum lucidum (SL), the chevrons illustrating the corresponding locations in the two views. (Inset) Two enlarged individual giant mossy fiber boutons (about 10 μM diameter) fluorescing because of the high Zn2+ content of those boutons. SO, stratum oriens; SR, stratum radiatum; SL, stratum lacunosum-moleculare; S, dendritic spine. [Courtesy of S. Sensi and J. Weiss]

The definitive identification of Maske's bright red band as a pool of rapidly exchangeable Zn2+ that was selectively localized in neuronal presynaptic vesicles took another 25 years. Two new, complementary methods of imaging Zn2+ were required. First, Timm and his followers developed the beautiful methods of silver-enhancing weakly bound metals in tissue. In the hands of Finn-Mogen Haug (9) and Gorm Danscher (10,11), these silver-enhancement methods were refined and fine-tuned until [in the current Danscher variants (10)] they now quite selectively label only transition metals (Cu, Zn, Fe, etc.) that are weakly bound in situ in their normal, physiological milieu. When used this way, the Timm-Danscher method stains essentially nothing in the mammalian cerebrum except presynaptic vesicles of certain (now called "zinc-containing") neurons (Fig. 1). The confirmation that the labeled metal in the vesicles is Zn2+ was provided by the second method, by using 6-methoxy-8-quinolyl-p-toluenesulfonamide (TSQ), which produces bright fluorescence when in a complex with Zn2+, but not when in a complex with other Timm-positive biologically relevant transition metals (Fe, Cu, etc.) (12). Moreover, because prior exposure to Timm's reagents (S2-) blocks all TSQ staining, there is essentially no doubt that (i) Timm-Danscher and TSQ stain the same pool of metal, (ii) this metal is Zn2+, and (iii) the stained Zn2+ is located exclusively in the vesicles of neurons in the cerebrum.

This combined approach, using silver staining and inspection at the electron-microscopic level together with fluorescent labeling (to verify the cationic species), remains the definitive method for establishing the nature and fine-structural location of metal deposits in tissue (for instance, Figs. 1 and 2).

Three Separate Zn2+ Pools (and Counting?)

In the brain, we now know that there are three separate "pools" of "free" Zn2+ that can be (indeed must be) imaged separately.

The first pool consists of the Zn2+ in the presynaptic vesicles. This Zn2+ can be imaged by Danscher-type silver staining (Fig. 1) in post mortem histologic sections. For imaging intravitally (or in brain slices or slice cultures), one must use a probe that can penetrate both the plasma membrane and the vesicular membrane. The original TSQ and its congeners (Zinquin, TFL-Zn) will penetrate into the vesicles, but they have unstable partition coefficients, and bind Zn2+ very tightly (Kd ~ 10–12) and so are of limited use for in vitro Zn2+ studies (13).

A better choice for imaging the vesicular pool of Zn2+ is the new probe Zinpyr-1 (ZP1) (14,15), which is stable in physiological conditions, remains in vesicles (probably in the membranes), and undergoes a sufficiently vigorous change in emission intensity upon binding Zn2+ that individual synaptic boutons can be resolved in acute brain slices with routine confocal microscopy (Fig. 1). Similar images can be obtained with the new Zn2+-sensing probe of Kikuchi (16). The probe ZP1 has low-nanomolar affinity for the Zn2+ in vesicles, but the concentration of Zn2+ in the vesicles is believed to be as much as 1 million fold higher than that, around 1 mM or so. Thus, the ideal probe for work on the biophysics of Zn2+ release, reuptake, and general "life cycle" in neuronal vesicles would be a small, lipophilic molecule that binds Zn2+ in the low millimolar to high micromolar range, such as the one recently described by Burdette and Lippard (17).

The situation concerning the Zn2+ in other secretory cell granules seems to be about the same as in neuronal secretory vesicles. Thus, the silver methods show the exquisitely precise labeling of zinc-filled secretory vesicles in the pancreas (18), for example (Fig. 2), an organ that has often been imaged by using the vesicle-permeating TSQ and Zinquin stains (19).

Fig. 2.

Zinc signals: Imaging the release and translocation of Zn2+. (A) Zinc release from pancreatic beta cells imaged in green with membrane-impermeable dye from Molecular Probes (FluoZin-3); the sequential time-lapse pictures cover 66 ms from left to right. The image is of a single exocytotic event and therefore corresponds to attomole quantities being released over a few milliseconds. [Courtesy of K. Gee and R. Kennedy] (B) Graph at left depicts the rise in extracellular Zn2+ during electrical stimulation of dentate granule neurons. The corresponding diagram at right shows in pseudocolor the release of Zn2+ into the extracellular fluid of a brain slice during electrical stimulation of the Zn2+-rich axons of the dentate granule neurons (circles) (measured with membrane impermeant Newport Green). The released Zn2+ could be reliably detected within 33 ms (after 4 stimulus pulses); Zn2+ concentration during peak release was about 20 μM. T, time in seconds. [Courtesy of Y. Li and J. Sarvey] (C) This pseudocolor image shows that a resting cortical neuron in culture displays virtually no Zn2+ fluorescence (faint blue cell body in lower left inset), but a brief stimulation with kainate opens zinc-permeable Ca-A/K channels, causing the middle and distal dendrites to rapidly fill with Zn2+ (white). FluoZin-3 probe from Molecular Probes. [Courtesy of S. Sensi and John Weiss] (D) The Timm-Danscher method of silver staining was used to label the Zn2+ in pancreatic beta cell secretory granules. Degranulation causes the Zn2+ release shown in the top row. [Courtesy of G. Danscher]

The second pool of free Zn2+ that is of interest for both clinical and basic science is the Zn2+ that is released from secretory vesicles into the extracellular milieu. In the brain, this synaptically released Zn2+ has been successfully imaged with a number of methods, all involving the use of membrane-impermeable fluorescent Zn2+ probes. Synaptic Zn2+ release was first visualized by using a protein biosensor derived from apocarbonic anhydrase (apoCA) coupled to a fluorescent reporter that changes emission when apoCA binds Zn2+ (20). The ratiometric, quantitative nature of this apoCA system has recently been harnessed to image released Zn2+ collected from microdialysis probes placed in the living brain (20). Optical fiber probes for Zn2+ (and Cu2+) are now in testing phase (21), and these probes should allow real-time monitoring of Zn2+ release in vivo.

Another method of monitoring released Zn2+ optically has evolved that uses single-reagent, membrane impermeable dyes, such as Newport Green (5,6) and newer compounds (for example, FluoZin-3) from Molecular Probes (Eugene, OR) (22). Although these methods are not ratiometric, they can be used for semi-quantitative studies of the timing and magnitude of Zn2+ translocation within neurons (Fig. 2) and Zn2+ release from other sources, such as pancreatic beta cells (Fig. 2).

In addition to the Zn2+ sequestered in secretory vesicles and the just-secreted Zn2+ in the interstices, there is a third "pool" of rapidly exchangeable Zn2+ that is transiently present in some cells under some conditions. This is the Zn2+ that is found in the perikaryal and dendritic cytoplasm of neurons after Zn2+ currents have been allowed, under special circumstances, to flow across the cell membrane (Fig. 1). This occurs in neurons in the brain whenever (i) there is a release of Zn2+ into the synaptic space around the target neuron and, simultaneously, (ii) zinc-permeable channels in postsynaptic neuron are opened by a ligand or by depolarization (23). This situation is illustrated beautifully by the image of Sensi and Weiss (Fig. 2), which shows the virtual absence of any Zn2+ staining in a normal cultured neuron at rest (inset) and the vivid Zn2+ signal visualized throughout several dendrites after the neuron was exposed briefly to a channel-opening ligand (kainate) and elevated extracellular Zn2+ (Fig. 2). It is noteworthy that the zinc-permeable channels operating in this case (the Ca-A/K channels, which are sensitive to α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and kainate) are known to be preferentially located along the more distal dendrites, where the Zn2+ influx has occurred (24).

These transitory intracellular (and transmembrane) Zn2+ fluxes are reminiscent of similar intracellular (and transmembrane) Ca2+ fluxes. We recently adduced evidence that this flux of synaptically released Zn2+ into neuronal dendrites is necessary and (with glutamate) sufficient for the induction of a lasting potentiation of synaptic strength at one type of zinc-modulated synapse (5,6).

Whereas the Zn2+ fluxes across the plasma membrane and the resulting intracellular Zn2+ currents are likely involved in normal, physiological cell signaling, they are fairly certainly involved in the injury of neurons in so-called excitotoxicity. For in excitotoxicity caused by ischemia, seizure, or head trauma, neurons that are injured (eosinophilic, acidophilic neurons) routinely show vivid Zn2+ staining in the perikaryal cytoplasm and proximal dendrites, and (sometimes) in the nucleus as well (Fig. 1). This injury-associated accumulation of free Zn2+ in the cytosol can be imaged in living tissue by membrane trappable dyes such as Newport Green and related compounds. In post mortem harvested tissue sections, the best method for imaging this pathological Zn2+ is to use a membrane-impermeable probe so that the perikaryal staining will not be overshadowed by the background staining of the neuronal presynaptic vesicles. Thus, the new probe ZP4 (14), which is membrane-impermeable, will hardly stain vesicles in frozen sections at all, but will easily stain the cut-open somata of 20- to 30-μm diameter neurons in 10- to 15-μm histologic sections (Fig. 1).

On the Horizon: in Vitro and in Vivo Imaging

Biosensors are coming in increasingly sophisticated forms. For example, Wolfgang Maret and colleagues have developed a metallothionein (MT)-based sensor that capitalizes on the fast and selective binding of Zn2+ by metallothionein (25). By attaching two halves of a fluorescence resonance energy transfer (FRET) pair to the two ends of the MT molecule, he has produced a sensor that will give a report proportional to Zn2+ signals occurring in the cytoplasm.

Another biosensor is under development by the Thompson-Fierke team, who has developed several exquisitely sensitive Zn2+ and copper sensors (26-28). In this ingenious model, one naturally fluorescent protein is attached to the carbonic anhydrase molecule, and another lipophilic small fluorescent molecule is added, which will coordinate with the holoprotein but not the apoprotein. Because the two proteins form a FRET pair, Zn2+ detection can be performed by excitation wavelength ratiometry, a powerful quantitative method.

Even further out on the horizon are possibilities for noninvasive imaging of Zn2+ deep in tissue. We have, for example, preliminary data that T1 and T2 weighted imaging of Zn2+ by in vivo magnetic resonance imaging (MRI) may be possible (29), and others (30) have proposed other MRI imaging methods for Zn2+.

The Wish List

For intravital imaging studies of Zn2+ in living systems, one needs a ratiometric probe with a strong fluorescent signal that has (i) the right affinity (Kd roughly in the middle of the range of expected signals), (ii) the right on and off rates (fast enough to see the expected signals), (iii) the right permeability (able to permeate and be trapped in the cellular compartment under study), and (iv) lack of toxicity to cells or tissues. Zn2+ probe development is still in the bootstrapping phase, because, for example, neither the basal extracellular level of Zn2+ nor the basal level of intracellular Zn2+ is known for mammalian cells or tissues. Various indirect evidence indicates that both of these are in the range of 10–13 to 10–9 M in normal healthy tissue (31), but direct quantitative determinations are still absent. Transient Zn2+ signals, in contrast, have been measured in the 10–7 to 10–5 M range within signaling microdomains (5,6). Like the absolute concentrations, the absolute signal speeds are not fully understood, but it is clear that signals in the single-digit millisecond range must be detectable.

From its modest beginnings buried within proteins, the Zn2+ ion has now emerged as an intracellular and intercellular signal ion that rivals calcium in the ubiquity and essentiality of its functions. As the tool kit for Zn2+ research grows, biomedical science will surely be galvanized.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
View Abstract

Stay Connected to Science Signaling

Navigate This Article