Real-time whole-plant dynamics of heavy metal transport in Arabidopsis halleri and Arabidopsis thaliana by gamma-ray imaging.

Heavy metals such as zinc are essential for plant growth, but toxic at high concentrations. Despite our knowledge of the molecular mechanisms of heavy metal uptake by plants, experimentally addressing the real-time whole-plant dynamics of heavy metal uptake and partitioning has remained a challenge. To overcome this, we applied a high sensitivity gamma-ray imaging system to image uptake and transport of radioactive 65Zn in whole-plant assays of Arabidopsis thaliana and the Zn hyperaccumulator Arabidopsis halleri. We show that our system can be used to quantitatively image and measure uptake and root-to-shoot translocation dynamics of zinc in real time. In the metal hyperaccumulator Arabidopsis halleri, 65Zn uptake and transport from its growth media to the shoot occurs rapidly and on time scales similar to those reported in rice. In transgenic A. halleri plants in which expression of the zinc transporter gene HMA4 is suppressed by RNAi, 65Zn uptake is completely abolished.

Inductively coupled plasma atomic emission spectroscopy

Region of interest

Single photon emission computed tomography

We have used gamma‐ray imaging to visualize the stark differences in real‐time whole‐plant dynamics of zinc root‐to‐shoot transport in heavy metal hyperaccumulating and non‐accumulating Arabidopsis.

INTRODUCTION

Most plants actively prevent the accumulation of high levels of metals in their aboveground biomass through a functional network of metal homeostasis, in order to avert toxicity. However, metal hyperaccumulators have developed rare adaptations to their environments and selectively extract specific metals from the soil and accumulate them in their shoots at very high concentrations without incurring symptoms of toxicity (Baker & Brooks, 1989; Frérot et al., 2010). In their natural habitat, metal hyperaccumulator plant species accumulate one or several metalloid elements in their aboveground biomass at concentrations two to three orders of magnitude higher than in leaves of most species on normal soils, and at least one order of magnitude greater than the usual range found in plants from metalliferous soils. (Pollard, Reeves, & Baker, 2014). Metal hyperaccumulation in plants can have commercial applications, including phytostabilization (revegetation of contaminated soils), phytomining (extraction of metals from plants for their value) or phytoremediation (plant‐based remediation of soils).

Arabidopsis halleri (formerly known as Cardaminopsis halleri) is a metal hyperaccumulator species and a facultative metallophyte, that is, it grows naturally on both metal‐contaminated (metalliferous) toxic soils and pristine (non‐metalliferous) soils. Zinc (Zn) hyperaccumulation (at least 10,000 μg of Zn per g of dry leaf tissue) in A. halleri is species‐wide, whereas cadmium (Cd) hyperaccumulation (at least 300 μg of Cd per g of dry leaf tissue) is geographically confined (Bert, Macnair, de Laguerie, Saumitou‐Laprade, & Petit, 2000; Bert et al., 2003; Frérot et al., 2010; Pauwels, Frérot, Bonnin, & Saumitou‐Laprade, 2006). The molecular mechanisms underlying metal hyperaccumulation in A. halleri are only partly understood. Typically Zn2+ and other poorly soluble transition metal cations are mobilized in the soil through acidification of the rhizosphere and the release of organic chelators by the root. These transition metal ions are then taken up into the root symplasm through ZIP transporters (ZRT, IRT‐like Protein) located in the epidermal plasma membrane (Grotz et al., 1998; Ramesh, Shin, Eide, & Schachtman, 2003). They are then thought to move symplastically from cell to cell and are eventually loaded into the apoplastic xylem through the transporters HEAVY METAL ATPASE 4 (HMA4) and HMA2 (Hussain et al., 2004). A quantitative trait locus for Cd2+ hypertolerance in A. halleri was found to map to a chromosomal region containing the HMA4 gene (Courbot et al., 2007). Transgenic A. halleri lines in which HMA4 was silenced by RNA interference were employed to demonstrate a key role for HMA4 in Zn and Cd hyperaccumulation as well as Zn and Cd hypertolerance (Hanikenne et al., 2008). In comparison to A. thaliana, the strongly enhanced expression of HMA4 in A. halleri results from a combination of modified cis‐regulatory sequences and copy number expansion (Hanikenne et al., 2008). Transfer of an A. halleri HMA4 gene into A. thaliana recapitulates Zn partitioning into xylem vessels and the constitutive transcriptional upregulation of Zn uptake system‐encoding Zn deficiency‐responsive genes including ZIP4 and IRT3 in roots. Other genes associated with metal hypertolerance and/or hyperaccumulation in A. halleri include constitutively highly expressed ICOTIAN MINE YNTHASE (NAS2 and NAS3), ETAL OLERANCE ROTEIN 1 (MTP1) and ATURAL ESISTANCE‐ SSOCIATED ACROPHAGE ROTEINS (NRAMP) genes (Hanikenne & Nouet, 2011; Krämer, Talke, & Hanikenne, 2007).

The functional characterization of candidate metal transporters can reveal the molecular mechanisms by which metal hyperaccumulation occurs. Imaging methods can reveal the spatiotemporal dynamics of metal translocation within and between roots and shoots. According to the present model, hyperaccumulation is primarily a consequence of high rates of metal loading into the root xylem, which depletes the metal in the root symplasm and secondarily induces elevated rates of metal uptake into the root (Hanikenne et al., 2008; Krämer, 2010). Additional contributions were proposed from chelation in the root symplasm (Deinlein et al., 2012), high rates of unloading from the xylem and distribution across leaf tissues where the metal can be stored in mesophyll cells and is present in high amounts in vacuoles (Hanikenne et al., 2008). However, the time scale by which metal uptake into the root is followed by root‐to‐shoot translocation remains unknown. Within the root, mathematical modeling approaches have explored the importance of ZIP regulation, HMA abundance and symplastic transport in the creation of the radial pattern of Zn within primary roots of A. thaliana (Claus, Bohmann, & Chavarría‐Krauser, 2013). Modeling predicted that Zn transport in the symplasm takes place at the same time scale as symplastic transport with Zn carried along the water flow path at the same velocity of water (Claus et al., 2013). Transpiration is thus a key determinant of the rate of this water transport for symplastic, root cell‐to‐cell transport. The metal must then be loaded into the xylem by the HMA4 transporter to undergo apoplastic transport from the root to shoot. In the hyperaccumulator A. halleri, Cd translocation from the external hydroponic medium into the xylem was very rapid: After a 2‐hr exposure to Cd, its concentration in xylem sap was 5‐fold higher than that in the external solution (Ueno, Iwashita, Zhao, & Ma, 2008). This Cd concentration in xylem sap decreased with increasing the concentration of external Zn suggesting that it is a function of the HMA4 transporter. Based on nuclear magnetic resonance (NMR) and computational modeling of metal speciation by Geochem‐PC of xylem sap, Cd was proposed to mainly occur in the free ionic form in the xylem sap.

Invasive methods involving decapitation of plants or detachment of stems have revealed aspects of the dynamics of apoplastic transport once metal is loaded into the xylem. Apoplastic transport of water in xylem of A. thaliana occurs within 30 min (Park et al., 2014), while xylem sap of A. halleri, collected after exposure to high Cd, accumulates Cd within the same time scale of 30 min (Ueno et al., 2008). At the single cell or tissue level, Förster resonance electron transfer (FRET) imaging in root cells revealed a high‐affinity, low‐capacity uptake system, a low‐affinity, high‐capacity uptake system as well as of a mechanism allowing Zn2+ release from internal cell stores in A. thaliana (Lanquar et al., 2014). Synchrotron‐based techniques revealed that high concentrations of Cd were located in the vascular system of both the hyperaccumulator A. halleri and the non‐accumulator, A. lyrata in the vascular system of the mid‐rib and in secondary veins. The two species differ in the extent of their accumulation of Cd in the mesophyll, with A. halleri showing enriched levels relative to A. lyrata (Isaure, Fayard, Sarret, Pairis, & Bourguignon, 2006).

To complement these often spatially narrow or invasive studies, whole plant imaging enables direct large‐scale material transport studies, addressing spatiotemporal transport within the context of the entire intact plant, from root to shoot for a period of up to several days. Here, we use a high‐sensitivity uncollimated detector single photon emission computed tomography (SPECT)‐type imaging system (UCD‐SPI (Walker, Judenhofer, Cherry, & Mitchell, 2015)), to study Zn metal transport in a hyperaccumulator, A. halleri, and in a non‐accumulator A. thaliana. We further examine the influence of the HMA4 transporter on Zn transport in A. halleri by use of an HMA4 knockdown line.

MATERIALS AND METHODS

Plant material and growth conditions

All plants were cultivated in a growth chamber with a 16 hr:8 hr light:dark cycle at 22°C and 50%–75% humidity with a light intensity of 100 μmol m−2 s−1. We used A. thaliana (accession Columbia [Col‐0]), A. halleri (LAN 5, Langelsheim, Germany), and AhHMA4‐RNAi A. halleri (Line 4.2.1, Langelsheim, Germany) (Hanikenne et al., 2008). Seeds of A. thaliana Col‐0 were grown to an age of 18 days on a mix of vermiculite and sand. In this study, 18‐day‐old plants were transferred from vermiculite and sand (3:1. Perlite:sand) to a hydroponic Hoagland nutrient solution, pH5.8, adjusted with KOH (Becher, Talke, Krall, & Krämer, 2004) but omitting ZnSO4. Zn concentration in this low‐Zn medium lacking any added Zn was measured to be 0.12 μM (inductively coupled plasma–atomic emission spectrometry (ICP‐AES), UC Davis Analytical Lab). The hydroponic growth was conducted in Magenta vessels (GA‐7, Sigma Aldrich), each plant in an individual vessel, on a mesh square suspended over 150 ml of growth medium. A. halleri plants were initially grown in Sunshine Mix (3:1 soil:vermiculite). Vegetatively growing A. halleri plants were replicated via cuttings and grown in hydroponics. The cuttings were first allowed to root on 1x SensiGrow, prepared according to the manufacturer's recommended working concentration (Advanced Nutrients, USA) nutrient solution for 14 days, before being transferred to Hoagland nutrient solution without Zn. The hydroponic solutions were changed every 4–5 days. Both A. halleri and A. thaliana plants were grown on Hoagland solution without Zn for 19–21 days prior to imaging to acclimate to Zn deprivation. Both A. thaliana and A. halleri were at mature vegetative growth stage during the imaging experiments.

Imaging protocol and conditions

Zn isotope was obtained from the National Isotope Development Center (Oak Ridge National Laboratory), in the form of ZnCl2 in solution. The half‐life of 65Zn is 243.9 days, and its primary gamma‐ray decay product is of energy 1,116 keV (51% branching ratio). A small positron emission branching ratio of 1.4% also exists.

Plants were transferred from the growth chamber to the imaging laboratory prior to imaging in batches of three where they grew during the time it took to image them. The imaging laboratory light level was 12 μmol m−2 s−1 PAR, with a light:dark cycle of 10 hr:14 hr, and was at temperature of 20°C. Prior to imaging, each plant was transferred from the low‐Zn Hoagland solution into a solution with radioactive Zn and 1 μM ZnSO4. The plants were incubated in a centrifuge tube containing 50 mL of Hoagland solution with 1 μM ZnSO4, to which 200 μCi of the 65Zn had been added. Each plant was held in a plastic funnel to avoid direct contact of the plant shoot with the radioactive solution – roots were in the spout and the shoot in the funnel mouth. The funnel was placed in the centrifuge tube and incubated in the 65Zn solution for 60 min, with the fluid level coming to the full level of the funnel spout to fully immerse the roots. To avoid generating a large amount of radioactive waste all plants were incubated in the same tube. For the first plant, the 65Zn concentration was 7.2 nM and for the final plant the 65Zn concentration was 5.6 nM, corresponding to an activity range of 200 μCi to 150 μCi in the 50 mL tube, as determined by a calibrated well counter (Capintec, Inc., Florham Park, NJ). The uptake into each plant was typically 4 to 5 μCi.

Following incubation, each plant was rinsed through three washes of Hoagland solution (with 1 μM ZnSO4) to remove any (non‐uptaken) radiolabeled media remaining on the root surface. This was performed with the plant still in the funnel and fresh 50 mL centrifuge tubes. Each rinse was of 20 s, during which the plant was gently swirled in the tube. The rinsing tubes were measured for activity rinsed off of the plant roots and only a small fraction was ever present. Final rinsing for each plant was always performed with a fresh solution. Rinsing solutions were one of the larger components of radiological waste which was generated. The imaging was carried out on Hoagland solution with 1 μM ZnSO4.

For gamma ray imaging on the UCD‐SPI system, the plants were carefully removed from the funnel and gently constrained with a holder of two 10 cm × 15 cm pieces of thin transparent plastic, spaced to be separated by 2 cm. The narrow holder constrained the plant leaves to be close to the imaging system and to have the entire plant roots and shoot to be within the system field of view. The holder contained the growth media at a level which completely covered the roots, a few cm below the top opening of the holder. Most of the leaves extended above the holder opening. Only one of the detector heads of the system was used in order to provide better monitoring of the plant condition throughout the imaging process. Plants were imaged continuously for a period of several hours until the gamma ray spatial distribution reached a steady state. Time periods of imaging ranged from 20 hr to 70 hr for the 11 plants.

Data were collected in files recording positions and energies of up to 8 × 106 gamma ray events detected by the system. Given the efficiency and sensitivity of the system, this corresponded to a time extent of 20 min for the larger plants with high Zn uptake, to multiple hours for the somewhat smaller A. thaliana plants. The data files were used as the time points for subsequent analysis. For each plant, two non‐overlapping regions of interest (ROIs) of area 1,280 mm2 (16 by 20 pixels of detector, each of dimension 2 by 2 mm2) were defined on the detector area, corresponding to the roots and to the leaves of the plant. Total counts recorded in each area in a data file were normalized to the time duration of the data file, to result in a value for counts s−1 1,280 mm−2. Due to the lack of spatial resolution of the system, some counts in each ROI may have come from the opposite part of the plant, but given the close geometry and solid angle considerations, the fraction of the total is small. A total of eleven plants were imaged in the UCD‐SPI system, three A. thaliana, and four each of the A. halleri genotypes.

Quantification of Zn remaining in growth media after imaging

At the end of plant imaging, two values were recorded from an electronic pulse counter unit (Tennelec TC512 Dual Counter Timer) for the overall trigger rate of the system: the final rate with the plant in place in the system, and the rate with the plant removed (but the growth media still remaining in the holder). Measurements of the plant components in the well counter were made post‐imaging, but due to the low activity levels (often 2 μCi or lower), the measurements were not stable above background.

Parallel resupply experiments with no radiolabel were done in triplicate to quantify the amount of Zn in the media before and after 24 hr resupply of 1 μM Zn, and hence the ability of the plants to deplete Zn from their growth media. The Zn concentration in the growth media samples were quantified using ICP‐AES analysis (UC Davis Analytical Lab). The resupply experiments without radiolabel were carried out in Magenta boxes in the growth chamber for a 24 hr period.

Quantification of transport from imaging data

Absolute sensitivity of the system is known from measurement with a centrifuge tube containing known 65Zn activity. The calibrated system performance as measured with point gamma ray sources of a variety of energies broadly agrees with expectations from detailed calculations and simulations (Walker, Cherry, & Mitchell, 2014; Walker et al., 2015).

Gamma‐ray image data processing

The quantification measurements for regions of interest (ROIs) provide a count of average gamma‐rays detected s−1 1,280 mm−2 (these units are used since both ROIs are of that area on the face of the detector). High levels of gamma rays were detected during the early time points, especially in A. halleri, and the gamma ray levels dissipated to a local minima around 3 hr. In order to compare the dynamics of the Zn transport from root to shoot across the samples, the shoot ROI measurements were normalized to the local minima by subtraction. To calculate the rate of transport of Zn into the shoot ROI, the initial slopes of gamma ray build up for shoot ROI were calculated with the time points between 3 hr and 24 hr. To carry out a comparison of time points where the first differences in Zn accumulation into shoot could be observed, the data needed to be processed further. The imaging setup collects 8 x 106 gamma‐rays (over a typical time period of 1,800 s) which summed together then constitutes a time point. Due to the difference in rates for each plant imaged, the time points for imaging are not matched sample to sample. In order to compare specific time points between the samples, values for 3 hr, 4 hr, 5 hr, and so on until 12 hr, were calculated using the two adjacent time points. This was done by calculating the slope and intercept between each pair of time points. Zn signal for each interpolated time point (x) was calculated using the formula y = mx + b (where m = slope and b = intercept from the values of two adjacent time points). Statistical analyses and time course plots were done in R version 3.4.2 (R Core Team, 2017) using agricolae (de Mendiburu, 2017), lsmeans (Lenth, 2016) and ggplot2 (Wickham, 2016) packages and wesanderson palette (Ram & Wickham, 2015).

RESULTS

Dynamics of Zn2+ movement in A. halleri vs. A. thaliana

The dynamics of Zn uptake and movement after Zn deprivation were visualized in vivo in intact plants of Arabidopsis halleri wild type (accession Langelsheim), the A. halleri AhHMA4‐RNAi line 4.2.1 (in the Langelsheim accession background), and A. thaliana Col‐0. First, the two species were grown to mature vegetative stage. A. halleri cuttings were grown hydroponically for 14 days to allow them to root. A. thaliana was grown on soil for 18 days, rinsed and transferred to Zn deprivation media (Hoagland media with no added ZnSO4) for hydroponic growth. The plants were allowed to acclimate to Zn deprivation for 19–21 days. ICP‐AES analysis of the Zn deprivation media showed that Zn concentration of the media lacking any added Zn was 0.12 μM and is thus considered low zinc media. Immediately prior to imaging, the plants were transferred to resupply media with 1 μM ZnSO4 (normal concentration of ZnSO4 in Hoagland media and control concentration in (Talke, Hanikenne, & Krämer, 2006)) and with 5.6–7.2 nM radioactive 65Zn. The plants were on the radiolabelled resupply media for 60 min, after which they were rinsed three times with non‐radiolabelled resupply media before placing them on fresh non‐radiolabeled resupply media for imaging with the UCD‐SPI system. The time at which a plant was placed in the imaging system is considered the 0 hr time point (Figures 1 and 2). Zn signal was measured in both the root region of interest (“root ROI”, blue box, Figure 1) and in the shoot region of interest (“shoot ROI”, red box, Figure 1). In A. halleri wild‐type plants, the signal was detected in the root at 0 hr, and then gradually moved up toward the shoot visibly at 12 hr and 24 hr, and subsequently appeared stationary in the upper part of the root from 36 hr to 60 hr (Figure 1a). In A. halleri HMA4‐RNAi plants, the total gamma‐ray signal never moved upwards from the root ROI, and instead moved slightly downwards in the root ROI over time, and the signal weakened considerably between 0 hr and 12 hr (Figure 1b). In A. thaliana wild‐type plants, similarly to A. halleri HMA4‐RNAi plants, no root‐to‐shoot movement of gamma‐ray signal was visible, and the signal dissipated along the time course (Figure 1c). The radioactive signal in both the shoot and the root ROI started at their maximum for all A. halleri samples (Figure 2a, Supporting Information Figure S1), and reduced to a local minimum at approximately 3 hr. This effect is likely attributable to loosely bound 65Zn in the root apoplast, which is gradually desorbed into the external solution, thereby leaving the center of the field of view and reducing the likelihood of an emitted gamma ray being detected in the imaging system. In order to quantify and investigate root‐to‐shoot Zn transport dynamics, the focus was on the timepoints after the local minima, thus measurements were normalized to the time point at which the detected signal was the lowest close to 3 hr. Shoot data corrected for the local minima is shown in Figure 2b.

Figure 1

Visualization of whole‐plant Zn resupply dynamics as detected by UCD‐SPI. Zn dynamics were measured for Arabidopsis halleri wild type (a), A. halleri RNAi (b) and Arabidopsis thaliana Col‐0 (c) plants upon Zn resupply (1 μM ZnSO 4) after 3 week Zn deprivation. Processed gamma ray detection data is overlaid with static pictures of the plants from which the data were collected. The red square represents the shoot ROI, and the blue square represents the root ROI

Figure 2

Quantification of Zn root‐to‐shoot transport dynamics. (a) Examples of raw data points for shoot (green) and root (brown) ROI measurements across the imaging time lapse (average gamma‐rays detected 1,280 mm−2 s−1). For Arabidopsis halleri wild type and the A. halleri RNAi line, the 0 hr time point shows a high level of gamma‐ray detection that dissipates to a local minimum at 3 hr. (b) Gamma‐rays in the shoot normalized to the 3 hr local minima of the shoot ROI. In order to compare the dynamics of the 65Zn transport to the shoot across the samples, the values were normalized to the local minima. The slopes represent the rate of signal intensity change over time transport from 3 hr to 24 hr. The initial slopes are significantly steeper in A. halleri wild type compared to the two other genotypes (ANOVA, TukeyHSD, p = 0.0126). Pink data points: A. halleri wild type. Beige data points: A. halleri . Brown data points: A. thaliana. Different symbol shapes indicate replicates (n = 4 for A. halleri wild type and A. halleri = 3 for Arabidopsis. thaliana). (c) Gamma‐rays visualized in shoots as in (B), displayed at discrete interpolated time points. The gamma‐ray imaging setup collects 8 × 106 gamma‐rays which then constitute a time point. Due to this, the time points for imaging are not matched sample to sample. In order to compare specific time points between the samples, values for 3 hr, 4 hr, 5 hr, and so on until 12 hr, were interpolated using the two adjacent time points. The letters indicate samples that are significantly different between genotypes at each time point (p < 0.05, least‐squares means). Lines represent the means and smoothed areas standard errors for each genotype. (d) Analytic ICP‐AES quantification of Zn in resupply media after 24 hr resupply compared to 0 hr. E: Gamma‐rays detected in the imaging media after plant imaging (average gamma‐rays detected s−1 1,280 mm−2). As there is no radiolabelled 65Zn in the imaging media prior to when the plant is placed for imaging, this represents radiolabeled 65Zn leached from the plant into the media. Letters (in panels d, e) indicate significant differences (p < 0.05, Student's t‐test) and error bars (panels d, e) indicate standard error

At the interpolated time points after 3 hr of resupply of radio‐labeled Zn, a continuous linear relative increase in the amount of 65Zn was observed in the shoot of A. halleri wild‐type plants (Figure 2c). By contrast, in both A. halleri HMA4‐RNAi and A. thaliana shoots, no additional 65Zn accumulated over the time period examined. On the contrary, the A. halleri HMA4‐RNAi shoots lost further 65Zn signal after the initial local minimum, and from 7 hr onwards this was statistically significant (Tukey's HSD test, p < 0.05). The differential net ability of the three genotypes to take up Zn from the growth medium was determined through independent Zn resupply experiments. These experiments were conducted over the identical time period (24 hr) to the gamma ray imaging experiment, but ICP‐AES was used to measure the amount of Zn remaining in the growth media (Figure 2d). These data demonstrated that also at the whole‐plant level, A. halleri wild‐type plants took up significantly more Zn from the growth media than A. halleri HMA4‐RNAi and A. thaliana (Figure 2d).

The root ROIs were also analyzed for change in 65Zn levels (Supporting Information Figure S2). All three genotypes had identically no change in the 65Zn signal across the analyzed time course (Tukey's HSD test).

Silencing of HMA4 completely abolishes root‐to‐shoot transport of Zn and results in Zn extrusion

In our whole‐plant imaging system, the A. halleri HMA4 RNAi line, shoot 65Zn levels in fact decreased over a 24 hr time period (Figure 2b‐c, Tukey's HSD test, p < 0.05). Finally, the amount of 65Zn signal remaining in the growth media after the imaging run is six‐fold higher after imaging of A. halleri HMA4‐RNAi compared to wild‐type A. halleri (Figure 2e), indicating a higher net loss of 65Zn from the A. halleri HMA4 RNAi root into the surrounding media. This 65Zn represents Zn that was transferred into the imaging setup with the plant after the 65Zn treatment and root rinsing.

DISCUSSION

Advantages of gamma‐ray imaging in intact plants

Aspects of metal uptake and homeostasis in plants may be understood well at the molecular level, but understanding of the whole‐plant dynamics has lagged behind due to the limitations of traditional experimental approaches and imaging systems. Radiolabeled molecules are widely used to measure the transport dynamics in biological systems. In experiments with whole plants and radiolabeled molecules, the biodistribution of the radiolabel is most typically analyzed by plant dissection and counting in a well counter, and for β‐emitters, by ashing the plant biomass and counting the radiolabel with a liquid scintillation counter. Several whole plant positron emission tomography (PET) imaging systems have been developed using 11C (Jahnke et al., 2009; Kawachi et al., 2011; Weisenberger et al., 2009), other groups have developed large scanners for β‐imaging using 32P (Kanno et al., 2007), and autoradiography has been used to image radioisotope distribution of whole plants (Page & Feller, 2005). A scanner (Kawachi et al., 2011) has been used to perform PET imaging of 65Zn and 107Cd in rice plants (Fontanili et al., 2016; Suzui, Yin, Ishii, Sekimoto, & Kawachi, 2017). Each of these nuclear imaging methods has drawbacks: for β‐imaging, the range of β‐particles can be too short to escape the plant; for PET, radioisotope lifetimes are short and the range of the positron can be too long for a thin, low density plant, limiting the yield of annihilation gamma‐rays; and autoradiography is invasive and can require exposure times of weeks or months. Therefore, UCD‐SPI for nuclear imaging of gamma‐ray emitters in plants has significant opportunity to contribute in a new way to transport studies. Indeed, the UCD‐SPI system has extremely high sensitivity (~10%) for a single‐photon imaging system, orders of magnitude higher than is found for SPECT systems used in medical imaging (often ~0.01%). This is extremely useful for imaging over long periods of time with high energy gamma rays, as it allows for imaging very low levels of radiotracer and thereby eases the practical issues of radiation safety and shielding and waste generation.

In particular, the uniquely high sensitivity of the system means that very small amounts (nCi) of radioisotopes may be imaged and followed over time. The system used here had the advantage that one entire time course can be recorded on a single plant. In time courses involving destructive sampling or imaging, data for the different time points are from distinct plant individuals, which generates substantial noise. This is particularly disruptive in experiments with non‐model plants such as A. halleri, which show substantially larger variation in plant architecture between independently grown plants even of an identical genotype.

Common PET isotopes of interest for plant studies include several found in organic compounds: 11C (T1/2 = 20 min), 13N (T1/2 = 10 min), and 18F (T1/2 = 109 min). These atoms can be substituted into amino acids or sugars in plants to follow the natural in situ processes (Cherry, Sorenson, & Phelps, 2003). However, due to the short lifetimes of PET radioisotopes, only short biological processes, such as photosynthesis, may be imaged. In contrast, single gamma‐ray emitting radiotracers used in single photon emission computed tomography (SPECT) are typically metals and do not easily label organic molecules. However, many trace element metals are essential to a plant's survival (Clemens, 2001; Williams, Pittman, & Hall, 2000). An active area of plant research is studying hyperaccumulation of metals in plants using a radioisotope of that metal; commonly studied metals include: Cd, Zn, Mn, Co, and Ni (Krämer, 2010; Page & Feller, 2005). Other potential applications for SPECT imaging include: studying plant ion transport in xylem (Macklon, 1970; Ueno et al., 2008), studying metabolic processes such as tracers for phloem transport (Omid, Malter, Peleg, & Wolf, 2008), and studying signaling by labeled exogenous peptides or proteins (Santner & Estelle, 2009). One further advantage of imaging systems based on gamma‐ray detection is the possibility of detecting the interactions of multiple radioisotopes simultaneously as the gamma‐rays that they emit have distinct energies that can be distinguished from each other by the detector. For example, simultaneous imaging of 65Zn (gamma‐ray 1,116 keV) and 109Cd (22 keV) would enable teasing apart the competition dynamics in their uptake.

However, there is an inherent tradeoff in increased sensitivity of the UCD‐SPI system with spatial resolution. Spatial resolution at the mm‐scale could be obtained using a collimator to better define the spatial origin of detected gamma rays, but this would lead to a greatly reduced event rate in the system. For the high‐energy gamma rays of 65Zn, collimation is a particular challenge. Given the hours‐long time scale of the transport studied here, it is possible that the choice of a collimator could have provided improved event positioning while preserving a usable event rate. A possible hybrid approach could have included using an insertable/removable collimator to acquire an alternating combination of two types of images: high sensitivity‐low spatial resolution without the collimator; and low sensitivity‐higher spatial resolution with a collimator. However, the use of a collimator for this system is unexplored thus far.

Zn transport dynamics in A. halleri and A. thaliana

Zn uptake into the symplast in the outer root layers and loading into the apoplastic xylem stream are well understood on molecular level (Moreira, Moraes, & dos Reis, 2018). However, the dynamics of symplastic movement and patterning of the radial transport have thus far only been modeled to elucidate the timescales of these events (Claus et al., 2013). After xylem loading, the mass flow‐mediated movement of Zn into the shoot inside the xylem is expected to occur within 30 min in Arabidopsis, as previously shown for water (Park et al., 2014) and Cd in xylem sap (Ueno et al., 2008). From previous SPECT imaging, we have shown that a pulse of radiolabelled pertechnetate (99mTcO4 −) moving in the xylem stream reaches the shoot apical meristem of a 2 week old sunflower already in 5 min (Walker et al., 2015).

The rate‐limiting step for root‐to‐shoot translocation of Zn was proposed to be xylem loading involving HMA4 transporters in both A. halleri and A. thaliana (Hanikenne et al., 2008; Sinclair, Sherson, Jarvis, Camakaris, & Cobbett, 2007). The dynamics of root‐to‐shoot Zn flux, however, have so far remained unclear in different species and transgenic lines. Estimates of Zn translocation rates from root to shoot were first obtained by spectroscopy methods of ashed shoot tissues. Early work with metal hyperaccumulator Noccaea caerulescens suggested that the speed of root‐to‐shoot Zn transport was between 20 and 60 hr (Lasat, Baker, & Kochian, 1996; Lombi, Zhao, McGrath, Young, & Sacchi, 2001). Recently, positron imaging of Zn uptake estimated the time for Zn root‐to‐panicle transport in dwarfed mature rice to be 5.3 hr (Suzui et al., 2017). Here, we have produced the first Zn root‐to‐shoot imaging data for A. halleri using UCD‐SPI. Zn accumulates within the shoot of A. halleri, consistent with its ability to hyperaccumulate Zn, different from the HMA4 RNAi line. The speed of Zn transport into the shoot in our data as observed with the smoothed standard error show clear shoot accumulation within 5–7 hr, respectively (Figure 2c). These results are in line with previous reports for rice (Suzui et al., 2017). This contrasts strongly with the faster speed of the other xylem‐transported compounds, such as water in A. thaliana (Park et al., 2014), Cd in A. halleri (Ueno et al., 2008) and pertechnetate in sunflower (Walker et al., 2015), all measured to reach the shoot in 30 min. It should be noted, however, that the experiments demonstrating water transport (Park et al., 2014) and Cd transport (Ueno et al., 2008) were carried out using decapitated stems and are thus destructive in nature, but also far more sensitive to small quantities than the method used here. The slower speed of Zn transport indicates that Zn loading into the xylem by HMA4 is slow and under tight control even in the metal hyperaccumulator A. halleri. Modelling the radial transport of Zn uptake has indeed indicated that HMA concentration is one of the key determinants of the uptake dynamics (Claus et al., 2013).

The HMA4 transporter pumps Zn2+ from the root symplasm into the apoplastic xylem sap of A. thaliana (Verret et al., 2004). Strongly elevated expression of A. halleri HMA4 was suggested to be responsible for the increased in root‐to‐shoot translocation of Zn in A. halleri relative to A. thaliana (Hanikenne et al., 2008). This conclusion was drawn based on the quantification of shoot Zn concentrations after long‐term growth in HMA4‐RNAi lines and wild‐type A. halleri and in A. thaliana Col‐0 (Hanikenne et al., 2008). In the same experiment, root Zn concentration was elevated in some A. halleri HMA4 RNAi lines relative to A. halleri wild‐type plants and even relative to A. thaliana (Hanikenne et al., 2008).

HMA4 is critical to the ability of A. halleri to hyperaccumulate Zn. We tested the functional role of HMA4 for A. halleri Zn translocation from root to shoot by imaging the Zn uptake dynamics of A. halleri HMA4‐RNAi line relative to A. halleri. We found that the Zn signal in the shoot of HMA4‐RNAi line did not increase over our 40‐hr imaging period, but conversely, we saw a continuous decrease in shoot Zn signal with significant differences observable at 3 hr (Figure 2c). The lack of an increase in shoot Zn confirms that Zn loading into the xylem is abolished in the HMA4‐RNAi plants (Hanikenne et al., 2008). The continuous decrease in the Zn signal in the shoot ROI seems to reflect bleeding of the strong early Zn signal from the root ROI into the shoot ROI. The dissipating signal through the A. halleri HMA4‐RNAi time course could be due to apoplastic 65Zn adsorbed to the cell walls of outer root layers during the 65Zn pulse (Lasat et al., 1996) and not removed by the triple rinsing with Hoagland solution. This cell wall‐adsorbed 65Zn would be desorbed into the growth medium during the imaging period by diffusion. The influx of Zn into the root symplasm is very tightly and rapidly regulated in Zn‐concentration dependent fashion (Claus et al., 2013; van de Mortel et al., 2006; Talke et al., 2006). Without the loading of Zn into the xylem, Zn builds up in the root symplasm. In the case of A. halleri HMA4‐RNAi, the symplasm could be saturated with Zn at 3 hr after the resupply, leading to prevention of further uptake of the cell wall‐adsorbed 65Zn and thus higher Zn desorption than Zn uptake into the symplasm.

Finally, we compared the dynamics of Zn movement in the Zn hyperaccumulator A. halleri with those in the related species A. thaliana, a non‐metal hyperaccumulator. Based on previous studies comparing Zn‐deficient to Zn‐sufficient plants of A. thaliana and/or A. halleri, the Zn concentrations in our hydroponic solutions can be estimated to result in moderate Zn deficiency (Talke et al., 2006; Sinclair et al., 2018). The net concentration of 65Zn in the resupply media over the 24 hr period showed a net decrease, suggesting that Zn was taken up into the shoot, although these levels are variable. We found that Zn resupply after Zn deprivation in A. thaliana did not lead to detectable uptake or change of Zn in the shoot or the root ROI. It is possible that the small size and flat rosette growth habit of A. thaliana affected our ability to detect Zn dynamics. Also, low abundance of HMA4 transporters in A. thaliana roots may lead to much slower dynamics that we were unable to capture. In the absence of quantification of 65Zn levels in the shoot, it is possible, although unlikely, that Zn was not translocated in A. thaliana.

The heavy metal imaging study presented here is of interest for phytoremediation applications (Kärenlampi et al., 2000; Krämer, 2010; Robinson et al., 1998; Salt, Smith, & Raskin, 1998; Sarma, 2011). Although most plants prevent the accumulation of heavy metals so as to avert toxicity, metal hyperaccumulators selectively extract high concentrations of metals from the soil into their shoots without incurring symptoms of toxicity (Baker & Brooks, 1989; Frérot et al., 2010). By using the heavy metal radiolabel 65Zn and the UCD‐SPI imaging system, we gained a more detailed spatiotemporal understanding of the dynamics of metal movement into plants, which may be a path toward the use and understanding of metal hyperaccumulating plants for such advantageous applications.

CONFLICT OF INTEREST

The authors declare no conflict of interest associated with the work described in this manuscript.

AUTHOR CONTRIBUTIONS

KK, GM, UK, SC, and SB designed the research; KK, KW, and GM performed research; KW and GM contributed new analytic/computational/etc. tools; KK, KW, GM, SC, and SB analyzed data, and KK, GM, and SB wrote the paper.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.