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 . The uptake into each plant was typically 4 to 5 μCi. Following incubation, each plant was rinsed through three washes of Hoagland solution to remove any radio labeled 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.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,sawtooth greenhouse 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 of area 1,280 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.At the end of plant imaging, two values were recorded from an electronic pulse counter unit 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 . Measurements of the plant components in the well counter were made post-imaging, but due to the low activity levels , the measurements were not stable above background. Parallel resupply experiments with no radio label 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 . The resupply experiments without radio label were carried out in Magenta boxes in the growth chamber for a 24 hr period.The quantification measurements for regions of interest provide a count of average gamma-rays detected s−1 1,280 mm−2 . 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 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 was calculated using the formula y = mx + b . Statistical analyses and time course plots were done in R version 3.4.2 using agricolae , lsmeans and ggplot2 packages and wesanderson palette .The dynamics of Zn uptake and movement after Zn deprivation were visualized in vivo in intact plants of Arabidopsis halleri wild type , the A. halleri AhHMA4-RNAi line 4.2.1 , 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 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 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 UCDSPI system. The time at which a plant was placed in the imaging system is considered the 0 hr time point . Zn signal was measured in both the root region of interest and in the shoot region of interest . 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 . 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 . 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 .

The radioactive signal in both the shoot and the root ROI started at their maximum for all A. halleri samples , 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 time points after the local minima,grow lights 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. 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 . 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 . 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 to the gamma ray imaging experiment, but ICP-AES was used to measure the amount of Zn remaining in the growth media . 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 . The root ROIs were also analyzed for change in 65Zn levels . All three genotypes had identically no change in the 65Zn signal across the analyzed time course .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. Radio labeled molecules are widely used to measure the transport dynamics in biological systems. In experiments with whole plants and radio labeled molecules, the bio-distribution of the radio label 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 radio label with a liquid scintillation counter. Several whole plant positron emission tomography imaging systems have been developed using 11C , other groups have developed large scanners for β-imaging using 32P , and autoradiography has been used to image radioisotope distribution of whole plants . A scanner has been used to perform PET imaging of 65Zn and 107Cd in rice plants . 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 for a single-photon imaging system, orders of magnitude higher than is found for SPECT systems used in medical imaging . 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 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 , 13N , and 18F . These atoms can be substituted into amino acids or sugars in plants to follow the natural in situ processes .