Dynamic imaging of cytosolic zinc in Arabidopsis roots combining FRET sensors and RootChip technology



  • Zinc plays a central role in all living cells as a cofactor for enzymes and as a structural element enabling the adequate folding of proteins. In eukaryotic cells, metals are highly compartmentalized and chelated. Although essential to characterize the mechanisms of Zn2+ homeostasis, the measurement of free metal concentrations in living cells has proved challenging and the dynamics are difficult to determine.
  • Our work combines the use of genetically encoded Förster resonance energy transfer (FRET) sensors and a novel microfluidic technology, the RootChip, to monitor the dynamics of cytosolic Zn2+ concentrations in Arabidopsis root cells.
  • Our experiments provide estimates of cytosolic free Zn2+ concentrations in Arabidopsis root cells grown under sufficient (0.4 nM) and excess (2 nM) Zn2+ supply. In addition, monitoring the dynamics of cytosolic [Zn2+] in response to external supply suggests the involvement of high- and low-affinity uptake systems as well as release from internal stores.
  • In this study, we demonstrate that the combination of genetically encoded FRET sensors and microfluidics provides an attractive tool to monitor the dynamics of cellular metal ion concentrations over a wide concentration range in root cells.


Zinc (Zn) is the second most abundant transition metal in plants (Marschner, 2012). Owing to its Lewis acid properties, Zn2+ is an excellent structural and functional element. Bioinformatic analyses of Zn2+ binding sites covering 57 representative organisms predict that c. 6% of the proteins in prokaryotes and 9% of proteins in eukaryotes bind Zn2+. During evolution, Zn2+ involvement has expanded from playing a catalytic role in biochemical reactions present in all living organisms to specific functions as a cofactor and potentially a regulator of DNA transcription so far found only in eukaryotes (Andreini et al., 2009; Andreini & Bertini, 2012).

Even though Zn2+ has a single oxidation state and is not able to enter the Haber Weiss reaction that generates harmful reactive oxygen species, the ability of Zn2+ to compete efficiently with other transition metals and to strongly bind sulfhydryl groups renders free cytosolic Zn2+ potentially harmful to cells. It is commonly accepted that Zn2+ concentration and localization are tightly controlled through transport and binding to proteins and low-molecular-weight ligands. Gauging metal concentration in cells has proved to be challenging. In Escherichia coli and mammalian cells, total Zn2+ concentration ([Zn2+]) was determined to be in the range of hundreds of micromolars, whereas cytosolic ‘free’ or ‘loosely bound’ [Zn2+] was estimated from the pico- to nanomolar range (Outten & O'Halloran, 2001; Dittmer et al., 2009; Bellomo et al., 2011; Qin et al., 2013). Because the cytosolic buffering capacity for Zn2+ is not known, it is not clear how external zinc supply affects the free cytosolic zinc concentration. Do steady-state concentrations depend on external zinc supply? Even though Zn2+ is highly buffered in the cytoplasm, transient increases in free Zn2+ concentration can be measured upon stimulation and were shown to mediate signaling (Maret, 2011, 2012). Cellular signals, including NO, glucose and glutamate, trigger Zn2+ release from internal stores (Chang et al., 2004; Dittmer et al., 2009; Bellomo et al., 2011; Qin et al., 2011). As in the case of calcium, knowledge of both the total and free cytosolic Zn2+ concentrations is thus required to address Zn2+ cellular functions. In multicellular organisms, such as plants, potential differences in Zn2+ concentrations in tissues and cell types further increase the complexity of Zn2+ probing, because spatial resolution is needed.

In Arabidopsis thaliana, many players in the Zn2+ homeostatic network have been identified: the main protein families in charge of Zn2+ uptake, compartmentalization and root-to-shoot transport are ZIP transporters, cation diffusion facilitator (CDF) transporters and metal pumping P1B-type ATPases (Grotz et al., 1998; Hussain et al., 2004; Kobae et al., 2004; Desbrosses-Fonrouge et al., 2005; Arrivault et al., 2006). Vacuoles are a major sequestration site for Zn2+. AtMTP1 and AtMTP3, two members of the CDF family, are involved in Zn2+ vacuolar loading (Kobae et al., 2004; Desbrosses-Fonrouge et al., 2005; Arrivault et al., 2006). Recently, the characterization of a nicotianamine transporter (ZIF1) indicated that Zn2+ could also be sequestrated in the vacuole as a complex with nicotianamine (Haydon et al., 2012). In addition, AtHMA2 and AtHMA4, P1B-ATPases, are important for root-to-shoot transport of Zn2+ in Arabidopsis (Hussain et al., 2004). Activation of AtMTP1/3 or AtHMA2/4 would be expected to decrease free cytosolic [Zn2+], whereas activation of AtZIP4 would lead to an increase (Assuncao et al., 2010). However, only a few studies have addressed the role of these transporters in Zn2+ distribution (Sinclair et al., 2007; Hanikenne et al., 2008; Song et al., 2010; Haydon et al., 2012).

Diverse techniques are available to investigate Zn2+ cellular distribution in plant cells and tissues. Local elemental analysis methods such as energy-dispersive X-ray analysis or electron-spectroscopic imaging allow the imaging of metals with high resolution, but their sensitivity is low (c. 1000 μg g−1)(Lanquar et al., 2005). Elemental mapping using synchrotron radiation-induced X-ray fluorescence and particle-induced X-ray emission enable the visualization of metal distribution at the level of tissues and have been successfully applied in Arabidopsis embryos (with a sensitivity in the range of μg g−1) (Kim et al., 2006; Punshon et al., 2009; Schnell Ramos et al., 2013). It is also possible to measure metal concentration by inductively coupled plasma atomic emission spectroscopy (ICP-AES) after subcellular fractionation (with a sensitivity in the range of fg g−1) (Seigneurin-Berny et al., 2006; Lanquar et al., 2010). In any case, these approaches only provide static information on total [Zn2+] with different thresholds in their spatial resolution or sensitivity. Real-time dynamic visualization of free [Zn2+] is not possible.

Chemical probes, such as Zinpyr-1, provide an alternative to these techniques (Sinclair et al., 2007). However, problems of cellular penetration can be encountered, intracellular localization of the probe cannot be controlled and long-term imaging (i.e. over days or weeks) cannot be achieved. The probe concentration necessary for the analysis can be destructive and could possibly disturb the cellular equilibrium (Qin et al., 2013). Genetically encoded Förster resonance energy transfer (FRET) sensors provide an innovative solution to obtain a better understanding of metal homeostasis and to monitor in vivo the dynamics of metals in plants (Bermejo et al., 2011; Jones et al., 2013). FRET sensors generally consist of a substrate-specific binding domain that is flanked by two, spectrally overlapping fluorescent proteins. A conformational change of the binding domain as a result of substrate binding that causes a change in distance or orientation of the two fluorophores translates into a measurable change in energy transfer. Moreover, these chimeric proteins may be targeted to a specific cellular compartment. FRET sensors with a high dynamic range have recently been developed for a large spectrum of substrates (Vogel et al., 2006; Bermejo et al., 2011; Okumoto et al., 2012; Jones et al., 2013). Different sets of FRET-based Zn2+ sensors were developed and used to measure zinc in yeast and animal cells (Pearce et al., 2000; Bozym et al., 2006; Qiao et al., 2006; Dittmer et al., 2009; Vinkenborg et al., 2009, 2010; Miranda et al., 2012; Lindenburg et al., 2013). The robust response and the wide range of Zn2+ affinities covered by the eCALWY set of FRET sensors render them particularly attractive to accurately determine intracellular Zn2+ concentrations (Vinkenborg et al., 2009). Such sensors, combined with microscopic imaging, are powerful tools to monitor in vivo cellular and subcellular changes in free metal concentrations in plant cells. A recently developed microfluidic imaging platform, the RootChip, allows root growth in a controlled environment on the microscope and pulsed perfusion with test solutions (Grossmann et al., 2011). These features make the RootChip suitable for monitoring changing intracellular substrate concentrations in roots using FRET sensors.

In the present study, we have expressed a suite of eCALWY FRET-based Zn2+ sensors in Arabidopsis thaliana. This enabled us to estimate free [Zn2+] concentrations in Arabidopsis root cells grown under different Zn2+ supplies. Moreover, we also analyzed the dynamics of free Zn2+, after intracellular Zn2+ chelation by N,N,N′,N′-tetrakis(2-pyridinylmethyl)-1,2-ethanediamine (TPEN) and upon root exposure to different Zn2+ concentrations. This work thus demonstrates the feasibility of using genetically encoded Zn2+ FRET sensors to study Zn2+ homeostasis and signaling in plant cells.

Materials and Methods

Material preparation

eCALWY-1 to six cDNAs were excised from peCALWY vectors (Vinkenborg et al., 2009) using NdeI/NotI sites, blunted and subcloned into the vector pRT101 at the SmaI site (Truong et al., 2001). CaMV 35S promoter::eCALWY fragments were excised from pRT101 by HindIII restriction and introduced into the pPZP312 vector using the PstI site (Hajdukiewicz et al., 1994). The binary constructs were introduced into Arabidopsis thaliana (L.) Heynh. Col0 and rdr6 through Agrobacterium-mediated transformation (Peragine et al., 2004; Deuschle et al., 2006). Transgenic seeds were selected for Basta resistance on soil. Basta-resistant plants were then analyzed for fluorescence and six to 11 lines were selected and propagated for each construct. Analyses were performed on T3 generation seedlings (lines: eCALWY-1 1A, eCALWY-1 3D; eCALWY-2 4C, eCALWY-2 11A; eCALWY-3 4B, eCALWY-3 5A; eCALWY-4 2A, eCALWY-4 15B; eCALWY-6 2A, eCALWY-6 5A).

Growth media

Seedlings were grown on a sugar-free modified Hoagland media (MHM) containing 0.28 mM KH2PO4, 1.25 mM KNO3, 0.75 mM MgSO4, 1.5 mM Ca(NO3)2, 25 μM H3BO3, 50 μM KCl, 0.1 μM Na2MoO4, 0.5 μM CuSO4, 10 μM FeHBED, and 3 mM MES-KOH, pH 5.5. Fe was provided as Fe3+ chelated to HBED (N,N′-di(2-hydroxybenzyl) ethylene diamine-N,N′-diacetic acid monochloride hydrate; Strem Chemicals) and FeHBED was prepared as previously described (Lanquar et al., 2005). Seedlings were grown in a climate chamber under the following conditions: 16 h light; light intensity 90 μmol m2 s−1; constant temperature of 21°C; 60% relative humidity.

As the sensors displayed a low dynamic range in the MHM growth medium, possibly as a result of the presence of micronutrients (metal ions), the growth medium was exchanged for a buffered macronutrient medium (BMM) before measurements (10 mM MES, pH5.5, 0.28 mM KH2PO4, 1.25 mM KNO3, 0.75 mM MgSO4, 1.5 mM Ca(NO3)2). TPEN and pyrithione were freshly prepared as 50 mM stock solutions in dimethyl sulfoxide and diluted as indicated in each experiment. All fluorescence imaging was performed with BMM. The ratio was not affected by the exchange of medium (Supporting Information, Fig. S1).

For sensor occupancy (SO) measurements, seeds were surface-sterilized and sown on sugar-free MHM with 1% (w/v) washed Agar M (Sigma-Aldrich). Washed Agar M was prepared as described in Haydon et al. (2012). ZnSO4 (0, 5 or 30 μM) was supplied as indicated. Four days after germination, seedlings were mounted between a 24 × 60 mm cover glass and a 0.8% agar pad (c. 1 mm thick) affixed to a 22 × 22 mm cover glass (Gutierrez et al., 2009). The 0.8% agar pad held BMM supplemented with 0, 5 or 30 μM ZnSO4 as starting conditions, or 500 μM TPEN or 10 μM pyrithione for treatment. The roots were transferred from one agar pad to the other. The same roots were imaged before and after treatments. During the 1 h incubation times, the mounted seedling/coverglass was maintained in a humid chamber.

For RootChip experiments, seedling growth and RootChip mounting were performed as described in Grossmann et al. (2011, 2012). Briefly, after 2 d stratification, seeds were germinated on sugar-free half-strength MHM, with or without 5 μM of ZnSO4 solidified with 1% agar, in cut pipette tips. Five days after germination, seedlings were transferred onto the chip where they grew for a further 2 d under constant perfusion with MHM, supplemented with 5 μM ZnSO4 (Zn2+-sufficient) or no Zn2+ (Zn2+-deficient). During measurements, seedlings were perfused with BMM media supplemented with TPEN, pyrithione, Zn2+ or Zn2+ pyrithione (ZP) as indicated.

Total zinc concentrations

For ICP-AES measurements, seeds were surface-sterilized and sown on sugar-free MHM with 1% (w/v) washed Agar M (Sigma-Aldrich) containing either 5 μM ZnSO4 (Zn2+-sufficient) or no ZnSO4 (Zn2+-deficient). Four days after germination, seedlings were harvested and washed. The DW of the samples was measured after drying at 60°C for 3 d. The dried samples were digested in a mixture of 1 ml of 65% nitric acid and 0.5 ml ultrapure H2O in a CEM MARS5Xpress microwave (CEM GmbH, Kamp-Lintford, Germany). After dilution in trace metal-free water, the metal content of the samples was determined by inductively coupled plasma optical emission spectroscopy using a Thermo Scientific ICAP 6300 Duo View Spectrometer (Thermo Scientific, Franklin, MA, USA). Measurements were performed at The Environmental Measurements Facility, Stanford University.

Root imaging

Roots were imaged on an inverted epifluorescence microscope (Leica DM IRE2, Buffalo Grove, IL, USA), equipped with a motorized stage (Scan IM 127x83; Marzhauser Wetzlar, Germany), a tunable Polychrome V monochromator light source (TILL Photonics, Gräfelfing, Germany), an excitation filter (Chroma 59017x), a beam splitter (Chroma 89002bs), a DualView beam splitter (Photometrics, Tucson, AZ, USA), and an electron multiplying charge-coupled device camera (QuantEM:512SC; Photometrics). For FRET measurements, cerulean was excited at 430 nm. As a control, citrine was excited at 500 nm. The DualView, carrying an ET470/24 m and an ET535/30 m filter setup, enabled simultaneous capture fluorescence emissions from cyan and citrine fluorophores. The objective used was ×10, na 0.4, HC PL APO. Leica Imaging data were acquired using SlideBook 5.0 software (Intelligent Imaging Innovations, Denver, CO, USA).

Sensor occupancy calculation

Image processing and analysis were performed in SlideBook 5.0. For each treatment, the background was subtracted from all measured fluorescent intensities; donor (ID) and acceptor (IA) intensities under donor excitation were acquired and ratios of ID/IA were calculated. To calculate [Zn2+] cytosolic free zinc concentration [Zn2+], SO vs sensor affinity curves were fitted to the equation math formula (where h is the Hill coefficient, SOmax is the maximal value of SO, the y variable, and Kd is the binding constant of the different eCALWYs, the x variable) using the Prism software with a nonlinear regression. Note that the Hill slope was not fixed but the fit yielded values close to 1 when seedlings were grown on sufficient or excess Zn2+.

RootChip FRET measurements, image processing

Image processing and analysis were performed in FIJI (http://fiji.sc/). To reduce movement of regions of interest as a result of root growth, images were registered using the StackReg plug-in (Thevenaz et al., 1998), modified by Brad Busse (Stanford University).

Mean gray values of regions of interest of the root tip area and c. 200 μm up in the elongation zone were calculated as follows: background was subtracted from all measured fluorescent intensities; ID and IA intensities were corrected against citrine emission upon citrine excitation to correct for intensity fluctuation caused by focus drift, root movement, or variations in the amounts of sensor protein during long-term measurements. Ratios of ID/IA were calculated. Data were normalized against the baseline. To calculate rates of [Zn2+] variation, fluorescence ratios were translated into concentrations using the equation math formula, where Kd is the binding constant of the sensor used, R is the variable fluorescence ratio, Rapo is the maximal fluorescence ratio in the presence of 0.5 mM TPEN and Rbound is the minimal fluorescence ratio in the presence of Zn2+ pyrithione. The slope was calculated by linear regression using the Excel SLOPE function.


Evaluation of cytosolic Zn2+ concentration

We investigated free cytosolic [Zn2+] in root cells using a suite of high-affinity FRET Zn2+ sensors, called eCALWY (Vinkenborg et al., 2009). In this work, we define free cytosolic [Zn2+] as the sum of Zn2+ in aqueous solution plus Zn2+ that can be easily disassociated from weak binding sites in the cytosol. eCALWY sensors consists of two metal binding domains, ATOX1 and WD4, linked via a flexible linker and flanked by self-associating variants of cerulean and citrine as donor and acceptor fluorescent domains. In the absence of Zn2+, the two fluorescent domains interact, which gives rise to high energy transfer. Binding of Zn2+ in between the two metal binding domains disrupts the weak intramolecular interaction between the fluorescent domains, which results in a substantial decrease in FRET. Zn2+ affinities ranged from 2 pM for eCALWY-1 to 3 nM for eCALWY-6. The purified proteins display a high dynamic range with a ratio change of c. 2.4 between apo and bound forms of the sensors.

Arabidopsis thaliana Col-0 ecotype was transformed with five eCALWY affinity variants (eCALWY-1, Kd = 1.8 ± 0.5 pM; eCALWY-2, Kd = 9 ± 3 pM; eCALWY-3, Kd = 45 ± 11 pM; eCALWY-4, Kd = 630 ± 160 pM; eCALWY-6, Kd = 2900 ± 1000 pM). However, all the transformants (c. 10 per construct) showed low or no fluorescence and thus could not be used for reliable Zn2+ measurements. Genetically encoded sensors are prone to transgene-induced silencing in Arabidopsis. We therefore expressed the eCALWY variants in the rdr6 genetic background, which is deficient in transgene-induced silencing (Peragine et al., 2004; Deuschle et al., 2006). Transgenic plant lines were screened for high intensities of fluorescence and homogenous fluorescence patterns for cerulean and citrine (c. five lines per construct) (Fig. 1a). In the case of FRET-based sugar sensors, the highest intensities of fluorescence, as shown by the FRET-independent citrine emission, were detected in the root tip and the root vasculature (Chaudhuri et al., 2008). The eCALWY lines showed some variation in the fluorescence pattern (Fig. 1a). We used three homozygous T3 lines with similar expression levels as judged from fluorescence intensity for each sensor and found that the responses from the lines were equivalent. Arabidopsis lines expressing eCALWY sensors did not display any obvious morphological alteration or change in bulk Zn concentration compared with the rdr6 or wildtype backgrounds (Table S1).

Figure 1.

Estimation of free cytosolic Zn2+ concentration in Arabidopsis thaliana roots. (a) Expression pattern of eCALWY sensors in A. thaliana roots. Citrine emission channel after citrine excitation is shown. Bars, 50 μm. (b) Illustration of the agar pad method to determine sensor occupancy (SO). (c) Sensor occupancy plotted against Kd for eCALWY- 1, -2, -3, -4 and -6. SO values were measured in roots of 4-d-old seedlings grown in the absence of Zn2+ (0 Zn2+), with 5 μM Zn2+ or 30 μM Zn2+. No Zn2+, = 62; 5 μM Zn2+, = 87; 30 μM Zn2+, = 73. The data points (n) were collected from three independent experiments in which three roots were analyzed. Error bars, ± SE. Free cytosolic Zn2+ concentration ([Zn2+]cyt) was calculated by fitting the SO values to the Hill equation.

The apparent cytosolic zinc concentration was assessed by measuring the sensor Zn2+ occupancy for each sensor as a function of the sensor Kd. Our calculations are based on the assumption that the affinity of the sensors is unchanged in vivo. Lines carrying eCALWY variants 1, 2, 3, 4 and 6 were germinated and grown for 4 d under three zinc external conditions: no zinc (EDTA-washed agar), 5 μM Zn2+ or 30 μM Zn2+. For this purpose, roots were transferred from plates to agar pads (Fig. 1b). In three independent experiments for each condition, we measured fluorescence signals of three to four roots in the first 500 μm of the root tip zone extending from the tip to the differentiation zone. For analyses, each root was divided in seven to 10 squares of 50 × 50 μm (region of interest, ROI). For each ROI, cerulean and citrine emission after cerulean excitation was acquired and the ‘resting ratio’ R of citrine/cerulean was calculated (or IA/ID). The sensors were subsequently calibrated using TPEN, a cell permeant Zn2+ chelator with femtomolar affinity (10−15), to chelate all Zn2+ from the sensors, and the zinc ionophore pyrithione in the presence of 1 mM zinc (ZP) to saturate all the sensors (Fig. 1b). For each ROI, emission ratios were calculated after addition of TPEN (Rapo) and ZP (Rbound). The sensor occupancy, SO = ((RapoR)/(RapoRbound)) × 100, was calculated. In 70% of cases, we observed an SO ranging between 0 and 110% and a Δ ratio between Rapo and Rbound states of c. 2, consistent with that observed for the purified proteins and for measurements in HEK293 cells (Vinkenborg et al., 2009). The remaining aberrant SO values (SO < 0 or SO > 110%) were attributed to nonhomogenous distribution of TPEN or Zn2+ pyrithione and were excluded from calculations. Each SO calculated was plotted against its own Kd, for the three zinc growth conditions, and fitted to the Hill equation (Fig. 1c). For plants grown under control conditions (5 μM Zn2+), the apparent cytosolic [Zn2+] was 420 ± 200 pM, while plants grown at excess Zn2+ (30 μM) had an apparent cytosolic [Zn2+] of 2000 ± 600 pM. In seedlings grown under Zn2+-depleted conditions, the total [Zn2+] was approximately three times lower than in seedlings grown in sufficient condition (Table S1). In Zn2+-starved root cells, the SO values were similar to those obtained with 30 μM Zn2+ (Figs 1c, S2), albeit with a shallower curve impairing a correct fit and thereby preventing accurate determination of the free cytosolic [Zn2+]. This result may reflect a decrease in the cytosolic Zn2+ buffer capacity, leading to an elevated pool of loosely bound Zn2+ under Zn2+ deficiency. Nevertheless, the SO values obtained indicate that free cytosolic [Zn2+] in Zn2+-starved root cells is in the same range as in plants grown under sufficient and excess Zn2+. Under sufficient and excess conditions, plant cells maintain a pM to nM free [Zn2+] in the presence of μM concentrations in the medium (i.e. a gradient > 104). This indicates a tight control of free Zn2+ concentration by buffering and compartmentalization. Nevertheless, seedlings showed early symptoms of Zn2+ deficiency or Zn2+ toxicity at 0 and 30 μM Zn2+, respectively (pale or red leaves, shorter roots compared with Zn2+ sufficient seedlings), indicating that Zn homeostasis cannot be maintained in these extreme conditions.

Mapping of intracellular Zn2+ concentrations

To analyze cytosolic free Zn2+ concentration in a minimally invasive system, we examined eCALWY plant lines using a microfluidic device called the RootChip (Grossmann et al., 2011, 2012). The RootChip enables growth of seedlings in a highly controlled microenvironment and direct microscopic observation of growing roots. The chip perfusion system provides growth medium during seedling development and features micromechanical valves that allow pulsed treatments of roots. The RootChip allows the medium to be changed while keeping the roots untouched during imaging.

The eCALWY-3 SO for seedlings grown in 5 μM Zn2+ was c. 70% on agar pads. In the RootChip, eCALWY-3 lines gave the highest Δ ratio (RapoRbound). Therefore, we evaluated the spatial distribution of free Zn2+ by imaging 7-d-old eCALWY-3-expressing roots, grown in Zn2+-sufficient conditions. The R ratios were measured for 25 × 25 μm ROIs and the sensor was calibrated by successive perfusion with 500 μM TPEN and 1 mM Zn2+ pyrithione. Sensor occupancies ranged from 11 to 40% along the root, translating to Zn2+ concentration between 4.1 ± 0.9 and 126 ± 12 pM depending on the root and the root zone (Fig. 2, Tables 1, S2). Most of the variation in free cytosolic [Zn2+] was observed between different roots (Tables 1, S2). By contrast, free cytosolic [Zn2+] displayed low spatial heterogeneity within a given root (Fig. 2).

Table 1. Sensor occupancy (SO) and [Zn2+] concentration in Arabidopsis thaliana roots expressing eCALWY-3 grown for 7 d in Zn2+-sufficient conditions
 Whole root (mean ± SE)Elongation zone (mean ± SE)Root tip (mean ± SE)
Root 1SO (%)35.88 ± 0.9639.21 ± 0.9932.06 ± 1.30
[Zn2+] pM25.85 ± 1.0729.44 ± 1.1821.72 ± 1.42
Root 2SO (%)11.78 ± 0.9312.90 ± 1.048.14 ± 1.59
[Zn2+] pM6.22 ± 0.546.87 ± 0.614.11 ± 0.91
Root 3SO (%)33.92 ± 3.0021.56 ± 1.3370.23 ± 3.53
[Zn2+] pM41.74 ± 6.9313.15 ± 1.02125.71 ± 11.95
Figure 2.

Mapping of eCALWY-3 sensor occupancy in a root. Arabidopsis thaliana seedlings were grown in Zn2+-sufficient conditions and measurements were performed on the RootChip (a). Sensor occupancy values calculated for each region of interest; bar, 100 μm. (b) Expression pattern of the sensor in each channel for the same root. From left to right: cerulean emission upon cerulean excitation (Cerulean); citrine emission upon cerulean excitation (Cerulean ex. Citrine em.); and citrine emission upon citrine excitation (Citrine).

Dynamics of intracellular Zn2+ concentrations

To analyze the dynamics of free cytosolic [Zn2+] in roots, eCALWY-3 lines were grown in the RootChip for 5 d in the presence of 5 μM Zn2+. Once the roots had reached the observation chamber, seedlings were perfused with Zn2+-free growth medium for a further 2 d. Before starting the imaging, growth medium was exchanged for Zn2+-free BMM (Figs S1, 3). Perfusion with 500 μM TPEN led to an increase of the ratio IA/ID, which reached a plateau after c. 10 min, suggesting that at that point, most of the sensors were in the apo form. Subsequently, TPEN was removed by perfusion with zinc-free BMM. Coincident with the medium change, the ratio slowly decreased, indicating that Zn2+ became available for binding to the sensors, even though Zn2+ had been omitted from the perfusion medium (Fig. 3). We estimated the rate of free cytosolic [Zn2+] decrease at 0.68 ± 0.25 pM min−1 (= 3). The rate of change in free cytosolic [Zn2+] is determined by Zn2+ fluxes and the Zn2+ buffering capacity of the cell. We first suspected that BMM was contaminated by zinc traces. Based on inductively coupled MS measurements, we estimated that the measure medium contains a contaminating Zn2+ trace concentration in the 100 nM range (0.01 ppm). A high-affinity transporter would be able to take up Zn2+ at this concentration. An alternative hypothesis, which could explain the ratio decrease, is that the addition of TPEN triggers Zn2+ starvation by chelating the pool of cytosolic Zn2+. Cells could compensate by releasing Zn2+ from intracellular stores. Vacuoles, which contain c. 100 μM total Zn2+ based on our previous measurements, provide a large Zn2+ reservoir (Lanquar et al., 2010). Finally, perfusion of roots with 1 mM Zn2+ led to a 100 times faster rate of decrease in the ratio (corresponding to 64.3 ± 18 pM min−1; = 3).

Figure 3.

Response of the eCALWY-3 sensor in root tips using the RootChip. Arabidopsis thaliana seedlings were grown for 5 d in the presence of 5 μM Zn2+ followed by 2 d of starvation. Ratios (IA/ID) normalized to time 0 are shown during perfusion of buffered macroelement medium (BMM), 500 μM N,N,N′,N′-tetrakis(2-pyridinylmethyl)-1,2-ethanediamine (TPEN) and 1 mM Zn2+. IA, acceptor emission intensity; ID, donor emission intensity.

To test whether the source of cytosolic Zn2+ accumulation after removal of TPEN is the Zn2+ taken up from trace amounts present in the medium, we compared the responses of high- (eCALWY-1, Kd = 2 pM) and medium-affinity (eCALWY-3, Kd = 45 pM) sensors in Zn2+-sufficient and -deficient conditions. Low amounts of TPEN (10 μM) were added to chelate the nanomolar traces of zinc in the medium (Fig. 4a,b). Both eCALWY-1 and eCALWY-3 roots responded to 500 μM TPEN treatments with an increased ratio consistent with a depletion of Zn2+ irrespective of the Zn2+ nutrition status of the seedlings. Subsequent perfusion with 10 μM TPEN led to a slow decrease in ratio only in the case of eCALWY-1 roots of Zn2+-deficient seedlings, with a reassociation rate estimated at 0.31 pM min−1 (Fig. 4b). This result indicates that release from internal stores in conjunction with a decreased Zn2+ buffering capacity accounts in part for the reassociation of Zn2+ with eCALWY-1 in Zn2+-starved cells. By contrast, the ratio remained stable for the seedlings grown in Zn2+ sufficient conditions (Fig. 4c,d). This result argues in favor of uptake of Zn2+ traces from the medium into root cells. Our data are thus consistent with the activation of Zn2+ release from intracellular stores upon treatment with 500 μM TPEN as well as import of Zn2+ traces present in the medium.

Figure 4.

eCALWY-1 and eCALWY-3 sensor responses to measure media supplied with 10 μM N,N′,N′-tetrakis(2-pyridinylmethyl)-1,2-ethanediamine (TPEN). Ratios (IA/ID) normalized to time 0 are shown. Measurements were performed on the RootChip. (a, b) Arabidopsis thaliana seedlings expressing eCALWY-1 (a) or eCALWY-3 (b) were grown for 5 d in the presence of 5 μM Zn2+ followed by 2 d of starvation. (c, d). Seedlings expressing eCALWY-1 (c) or eCALWY-3 (d) were grown for 7 d with 5 μM Zn2+. IA, acceptor emission intensity; ID, donor emission intensity.

Zinc uptake in root cells

Supplying 10 μM TPEN in the extracellular buffer is sufficient to prevent Zn2+ reassociation with eCALWY-3 and maintain the sensor in the apo form. Thus, in the subsequent experiments, we measured the effect of Zn2+ influx on free cytosolic [Zn2+] in the presence of 10 μM TPEN. We supplied 10 μM, 100 μM or 1 mM of Zn2+ to roots of seedlings expressing eCALWY-3 grown in Zn2+-sufficient conditions and estimated the rate of free [Zn2+] increase in the cytosol (Fig. 5a,b; Table 2). When 10 or 100 μM was supplied, intracellular [Zn2+] increased at a rate at c. 1 pM min−1 (= 2) (Fig. 5a). When 1 mM Zn2+ was supplied, the accumulation rate increased to 84 pM min−1 (= 1) (Fig. 5b). A similar rate of 71 ± 29 pmol min−1 (= 4) (Table 2) was measured in response to 1 mM of external Zn2+ using eCALWY-4 seedlings (lower Kd = 630 pM). To obtain an estimate of the maximal accumulation rate, the zinc ionophore pyrithione was added in the presence of 1 mM Zn2+. The free cytosolic [Zn2+] increase rate measured was 373 ± 124 pM min−1 (= 4) (Fig. 5a, Table 3). This indicates that the rate measured in response to 1 mM of external Zn2+ is not limited by diffusion.

Table 2. Rates of free cytososlic [Zn2+] variation in individual root tips expressing eCALWY-3 and eCALWY-4
 10 μM Zn2+100 μM Zn2+1 mM Zn2+
  1. Arabidopsis thaliana seedlings were grown for 7 d in the presence of 5 μM Zn2+. Perfusions were 10 μM, 100 μM or 1 mM of Zn2+.

  2. nd, not determined.

eCALWY-3Root 10.94 pM min−10.81 pM min−1nd
Root 20.54 pM min−10.2 pM min−1nd
Root 3ndnd84 pM min−1
eCALWY-4Root 1ndnd49 pM min−1
Root 2ndnd90 pM min−1
Root 3ndnd101 pM min−1
Root 4ndnd44 pM min−1
Table 3. Calculation of rates of free cytosolic [Zn2+] increase in individual root tips expressing eCALWY-3
 ZPMean ± SE
  1. Arabidopsis thaliana seedlings were grown for 7 d in the presence of 5 μM Zn2+. 10 μM of pyrithione and 1 mM of Zn2+ were perfused.

eCALWY-3Root 1464 pM min−1373 ± 124 pM min−1
Root 2492 pM min−1
Root 3237 pM min−1
Root 4300 pM min−1
Figure 5.

eCALWY-3 sensor responses to extracellular Zn2+ perfusion. Arabidopsis thaliana seedlings expressing eCALWY-3 sensors were grown for 7 d with 5 μM Zn2+. Ratios (IA/ID) normalized to time 0 are shown. Measurements were performed on the RootChip in the presence of 10 μM N,N′,N′-tetrakis(2-pyridinylmethyl)-1,2-ethanediamine (TPEN). (a) Perfusion of 10 μM Zn2+, 100 μM Zn2+, 1 mM Zn2+ plus 50 μM pyrithione (ZP). (b) Perfusion of 1 mM Zn2+. Rates of cytosolic [Zn2+] increase are indicated on the graphs. IA, acceptor emission intensity; ID, donor emission intensity; BMM, buffered macronutrient medium.

The observation of the concentration-dependent decrease in ratio indicates that cells expressing eCALWYs are a suitable system to dynamically monitor the response of cytosolic free [Zn2+] in response to a change in extracellular Zn2+ in vivo. Our measurements suggest the involvement of several Zn2+ uptake mechanisms and cytosolic Zn2+ buffering.


The goal of this study was to establish a nondisruptive and dynamic method to measure apparent cytosolic free Zn2+ concentrations with cellular resolution in live tissues. By expressing a series of FRET sensors in plants growing in a minimally invasive environment, the RootChip, we found that cytosolic Zn2+ concentrations are homogenous through the root but vary between the pM to low nM range, depending on external supply. In addition, our results provide evidence of two uptake systems in Arabidopsis root cells, a high-affinity, low-capacity uptake system and a low-affinity, high-capacity uptake system, as well as of a mechanism allowing Zn2+ release from internal stores.

Cytosolic [Zn2+] in plants grown in the presence of different zinc concentrations

Plants have to acclimate to very different nutritional conditions, specifically in the context of metals, which limit growth when scarce and are toxic if accumulated to high concentrations. While the mechanisms of iron uptake, storage and signaling are well studied, less is known about the important cofactor Zn2+.

To examine how cytosolic [Zn2+] in root cells responds to external Zn2+ supply, we exposed seedlings to excess or lack of Zn2+. In optimal growth conditions (5 μM Zn2+), the apparent free cytosolic concentration was estimated at 420 ± 200 pM. This value is close to the Kd of Zinpyr (700 pM) and further validates the usefulness of this fluorophore to probe Zn2+ in plant cells (Sinclair et al., 2007). Assuming that a plant cell is a 20 μm cube and that cytosol represents c. 10% of its volume, a Zn2+ concentration of 420 pM means that only c. 100 Zn2+ ions are free or loosely bound per cytosol. Based on our previous measurements of total Zn2+ in leaf protoplasts (Lanquar et al., 2010), only 1 Zn2+ ion per million is available. The cytosolic [Zn2+] we measured is comparable to those found in diverse mammalian cell types using FRET sensors (from 80 to 400 pM) (Vinkenborg et al., 2009; Qin et al., 2011, 2013). This is also consistent with data reporting that Zn2+ fluctuates from 100 pM to 1 nM in cells depending on the Zn2+ binding protein demand (Maret, 2011). Our measurements in Zn2+ excess conditions showed an increase in the cytosolic [Zn2+] to 2 nM. Root cells are thus able to maintain an extremely high Zn2+ gradient across the plasma membrane, even under Zn2+ excess. However, under excess, the buffering mechanisms are probably overcome, as cytosolic Zn2+ is elevated and toxicity symptoms are visible. Buffering against external fluctuations may be achieved by several mechanisms: limiting the entry of the metals by down-regulating uptake systems; chelation of Zn2+ by low-molecular-weight ligands or metallothioneins; sequestration into the vacuole or other organelles by transporters such as the Cation Diffusion Facilitator/Metal Tolerance Protein, MTP1 and 3; and extrusion from the cytosol possibly involving the cysteine-rich domain protein, Plant Cadmium Resistance 2 (PCR2) and Heavy Metal transporting P1B type ATPases (HMA2 and 4). HMA2 is localized at the plasma membrane and has a Kd for Zn2+ in the nM range that would be compatible with this function (Eren et al., 2007; Zimmermann et al., 2009).

From the agar transfer technology to the use of RootChips for monitoring zinc

The analyses of free Zn2+ in the cytosol using agar pads were characterized by a high standard error (SE). One factor contributing to the error is the c. 30% SE of the Kd value of eCALWY-6 (2.9 ± 1.0 pM). Another likely source of error is the handling of the plants and the efficiency of treatments. Medium change using the agar pad technique can lead to induction of stress responses that often induce significant increases in endogenous fluorescence in plant cells. Moreover, this technique does not guarantee homogenous distribution of the chelators and ionophores applied to the roots, leading to uncertainties in the calibration. This technical limitation probably accounts for the high variability in SO values measured using the agar pad method (Fig. S2). Finally, biological factors may account for an even higher SE of Zn2+-starved seedlings (1.5 ± 1.5 nM): it is likely that a decrease of the cytosolic Zn2+ buffer capacity triggered by Zn2+ deficiency prevents accurate determination of free cytosolic [Zn2+].

To overcome errors associated with root handling or uneven root access of test solutions and image roots in optimal conditions, our laboratory recently developed a microfluidic platform, the RootChip, integrating seedling growth and root imaging (Grossmann et al., 2011). We calculated SO in 7-d-old eCALWY-3 roots grown in the same conditions. Compared with the agar pad measurements, the SO values obtained were lower. This difference could be a result of the developmental stage: seedlings used for agar pad experiments were 4 d old, while seedlings in the RootChip were 3 d older. In addition, the RootChip is less disruptive and the chemicals are homogenously delivered to the roots. Note that using a single sensor to convert SO into [Zn2+] is not as accurate as when using a set of sensors having different affinities, but allows comparison between experiments.

Dynamics of free cytosolic [Zn2+] in response to extracellular Zn2+

After the seedlings were perfused with 500 μM TPEN, reassociation of Zn2+ to the sensors was observed upon perfusion with a nominally Zn2+-free buffer. In the case of the seedlings grown with sufficient Zn2+ supply (5 μM for 7 d), adding 10 μM of TPEN to chelate trace Zn2+ in the external medium allowed a stable ratio to be maintained. This supports the hypothesis that root cells possess a high-affinity mechanism, which allows efficient uptake of Zn2+ traces in the 100 nM range. The case of plants starved for zinc is more complex: the fact that, in eCALWY-1, seedling cytosolic [Zn2+] still decreases suggests that Zn2+ is released from an internal compartment, such as the vacuole. This possibly indicates that starvation induces the expression of transporters able to remobilize Zn2+ from internal stores. AtNRAMP4 is a candidate for Zn2+ release from the vacuole (Lanquar et al., 2004). A decrease in cytosolic Zn2+ buffering capacity in Zn2+-deficient seedlings may also enhance the reassociation rate between Zn2+ and eCALWY-1. Note that the Zn2+ released could be detected only by a high-affinity sensor, such as eCALWY-1 (Kd 1.8 ± 0.5 pM) but not with the lower-affinity sensor eCALWY-3. This suggests that the release mechanism operates in a range of low cytosolic [Zn2+].

Förster resonance energy transfer sensors have been used to characterize sugar fluxes and identify new transporters (Chaudhuri et al., 2008; Chen et al., 2010). We exposed root cells of seedlings grown in the RootChip under a Zn2+-sufficient condition to three external Zn2+ concentrations. When perfused with 10 and 100 μM Zn2+, similar rates were recorded (c. 1 pmol min−1). This suggests the presence of a high-affinity, low-capacity uptake system, active at 10 μM Zn2+. Note that in the presence of 10 μM TPEN and 10 μM Zn2+, the free Zn2+ concentration is estimated to be only 100 pM. Free cytosolic [Zn2+] variation rates did not further increase when 100 μM Zn2+ was perfused. This may indicate that the high-affinity uptake system is saturated or that cellular Zn2+ buffering mechanisms efficiently limit the rate of free cytosolic [Zn2+] increase in this concentration range. When cells were exposed to 1 mM Zn2+, rates increased 100 times (c. 100 pmol min−1, Fig. 5). Seedlings grown in absence of Zn2+ showed a similar rate when exposed 1 mM of external Zn2+ (64.3 ± 18 pM min−1; = 3), suggesting the involvement of a low-affinity, high-capacity transport system in both growth conditions.

The identity of these transporters remains elusive. A high-affinity, low-capacity uptake system could be mediated by members of the ZIP transporter family (Colangelo & Guerinot, 2006; Assuncao et al., 2010). In plants, the best-characterized members are IRT1, IRT3 and ZIP4. IRT1, described as the high-affinity Fe transporter, is also able to mediate Zn2+ uptake (Korshunova et al., 1999; Rogers et al., 2000; Vert et al., 2002). When expressed in yeast, IRT1 displays a Km for Zn2+ of 2.8 ± 0.6 μM, which may be consistent with the data we obtained in this study (Korshunova et al., 1999). IRT3 shares 91% identity with IRT1 and is also a potential candidate for involvement in Zn2+ uptake from the soil (Lin et al., 2009). ZIP4 which is induced under Zn2+ deficiency may display higher affinity for Zn2+ (Assuncao et al., 2010). The molecular identity of the low-affinity, high-capacity transport system remains to be discovered. It is likely that, at mM concentrations, Zn2+ enters nonspecifically through transporters for other cations, such as calcium, for example.

Future use of Zn2+ sensors

In this study we have established the application of genetically encoded FRET sensors to measure intracellular Zn2+ concentrations in plants cells. These sensors allow measuring steady state Zn2+ concentrations, but can also be used to measure Zn2+ intracellular changes in response to external cues. Silencing of eCALWY expression is a bottleneck for the use of the Zn2+ nanosensors in plants. Silencing can be alleviated by placing the eCALWY under the control of a weaker promoter such as UBI10 (Grefen et al., 2010) or under the control of inducible or cell-specific promoters. As silencing is often associated with tandem sequence repetition, altering the codon usage for one of the fluorescent proteins of the eCALWY coding sequence may also reduce silencing. Such improvements, combined with the background work presented in this study, will enable the implementation of the FRET sensor technology in Zn2+ signaling and Zn2+ transport mutants and help to dissect the role of individual genes in the Zn2+ homeostasis network.


We thank Heather Cartwright and Erika Valle-Smith for valuable help. This study was supported by grants. This work was made possible by grants from a Marie Curie International Outgoing Fellowship to V.L. (PIOF-GA-2008-221482), EMBO Long Term Fellowship to G.G. and the Department of Energy to W.B.F. (DE-FG02-04ER15542).