This protocol presents a workflow for sub-mm 2D visualization of multiple labile inorganic nutrient and contaminant solute species using diffusive gradients in thin films (DGT) combined with mass spectrometry imaging. Solute sampling and high-resolution chemical analysis are described in detail for quantitative mapping of solutes in the rhizosphere of terrestrial plants.
We describe a method for two-dimensional (2D) visualization and quantification of the distribution of labile (i.e., reversibly adsorbed) inorganic nutrient (e.g., P, Fe, Mn) and contaminant (e.g., As, Cd, Pb) solute species in the soil adjacent to plant roots (the ‘rhizosphere’) at sub-millimeter (~100 µm) spatial resolution. The method combines sink-based solute sampling by the diffusive gradients in thin films (DGT) technique with spatially resolved chemical analysis by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The DGT technique is based on thin hydrogels with homogeneously distributed analyte-selective binding phases. The variety of available binding phases allows for the preparation of different DGT gel types following simple gel fabrication procedures. For DGT gel deployment in the rhizosphere, plants are grown in flat, transparent growth containers (rhizotrons), which enable minimal invasive access to a soil-grown root system. After a pre-growth period, DGT gels are applied to selected regions of interest for in situ solute sampling in the rhizosphere. Afterwards, DGT gels are retrieved and prepared for subsequent chemical analysis of the bound solutes using LA-ICP-MS line-scan imaging. Application of internal normalization using 13C and external calibration using matrix-matched gel standards further allows for the quantification of the 2D solute fluxes. This method is unique in its capability to generate quantitative, sub-mm scale 2D images of multi-element solute fluxes in soil-plant environments, exceeding the achievable spatial resolution of other methods for measuring solute gradients in the rhizosphere substantially. We present the application and evaluation of the method for imaging multiple cationic and anionic solute species in the rhizosphere of terrestrial plants and highlight the possibility of combining this method with complementary solute imaging techniques.
Nutrient acquisition by crop plants is a key factor in determining crop productivity. The processes governing efficient uptake of nutrients by crops have been studied intensely, especially the mechanisms controlling nutrient availability to and nutrient internalization by plant roots at the soil-root interface, the rhizosphere, are recognized for their role in crop nutrient acquisition. Important processes for plant nutrient uptake include: nutrient transport towards the root; dynamic sorption equilibria between species dissolved in the soil porewater and species bound to solid soil surfaces; microbial competition for nutrients; microbial mineralization of nutrients that are contained in soil organic matter; and nutrient internalization into the root symplasm. The uptake of inorganic trace metal(oid) contaminants is largely controlled by the same mechanisms.
Depending on nutrient and contaminant availability, plant demand and diffusivity in soil, differential nutrient patterns in the rhizosphere can be observed. For strongly sorbing elements with comparatively high internalization rates (e.g., P, Fe, Mn, Zn, As, Cd, Pb), depletion of the labile (i.e., reversibly adsorbed) element fraction compared to the bulk soil is found, with depletion zone widths often being ≤1 mm, while for more mobile nutrients such as NO3–, depletion zones can extend up to several centimeters1. Moreover, accumulation of elements such as Al and Cd has been observed when availability exceeds plant uptake rates2,3.
Given the importance of rhizosphere processes in nutrient and contaminant cycling, several techniques for measuring the plant-available element fraction at high spatial resolution have been developed4,5. However, measuring small-scale labile solute distributions has proven to be challenging for several reasons. A major difficulty is to sample very small (low µL range) volumes of soil and/or porewater at defined positions adjacent to living plant roots to resolve the steep nutrient gradients in the rhizosphere. One approach to address this problem is to use micro-suction cups for the extraction of porewater samples6. With this method, A. Göttlein, A. Heim and E. Matzner7 measured soil porewater nutrient concentrations in the vicinity of Quercus robur L. roots at a spatial resolution of ~1 cm. A difficulty of analyzing µL volumes of soil or soil solution is, that these small sample volumes, in combination with the low concentrations of all but the major nutrient species, require highly sensitive chemical analysis techniques.
An alternative system, capable of resolving nutrient gradients at a resolution down to ~0.5 mm, is to grow a root mat on the surface of a soil block, with a thin hydrophilic membrane layer separating soil from the roots8,9. In this configuration, solutes can pass through the membrane and roots can take up nutrients and contaminants from the soil while root exudates can diffuse into the soil. After the establishment of a dense root layer, the soil block can be sampled and sliced to obtained soil samples for subsequent extraction of element fractions. In this way, one-dimensional nutrient, and contaminant gradients, averaged across a relatively large area (~100 cm2) can be analyzed.
A further challenge is to obtain samples of the labile, plant-available element fraction, since most chemical soil extraction techniques operate very differently compared to the mechanisms by which plants take up nutrients and contaminants. In many soil-extraction protocols, soil is mixed with an extractant solution with the aim to establish a (pseudo-)equilibrium between dissolved and sorbed element fraction. However, plants continuously internalize nutrients and, therefore, often progressively deplete the rhizosphere soil. Although equilibrium extraction protocols have been widely adopted as soil tests as they are easy to implement, the extracted nutrient fraction often does not represent the plant-available nutrient fraction well10,11,12,13. Sink methods which continuously deplete the sampled soil for nutrients have been proposed as advantageous methods and may better resemble the underlying nutrient uptake mechanism by mimicking the root uptake processes10,11,14,15.
In addition to the methods described above, genuine imaging applications, capable of measuring continuous parameter maps with resolutions ≤100 µm across fields of view of several cm2 have been developed for specific elements and soil (bio)chemical parameters5. Autoradiography can be used to image the element distribution in the rhizosphere provided that suitable radioisotopes are available16. Planar optodes enable visualization of important soil chemical parameters such as pH and pO217,18,19, and enzyme activity or total protein distributions can be mapped using fluorescent indicator imaging techniques such as soil zymography20,21,22,23 and/or root blotting methods24. While zymography and autoradiography are limited to the measurement of a single parameter at a time, pH and pO2 imaging using planar optodes can be done concurrently. The more traditional root mat techniques provide 1D information only, while micro suction cups provide point measurements or low resolution 2D information, however both approaches allow for multi-element analysis. More recently, P. D. Ilhardt, et al.25 presented a novel approach using laser induced breakdown spectroscopy (LIBS) to map 2D total multi-element distributions at a resolution of ~100 µm in soil-root core samples where the natural element distribution was preserved by careful sample preparation.
The only technique capable of targeted 2D sampling of multiple nutrient and contaminant solutes at high spatial resolution is the diffusive gradients in thin films (DGT) technique, a sink-based sampling method that immobilizes labile trace metal(loid) species in situ on a binding material embedded in a hydrogel layer26,27. DGT was introduced as a chemical speciation technique for measuring labile solutes in sediments and waters, and was soon adopted for its use in soils28. It enables sub-mm scale multi-element solute imaging, which was initially demonstrated in a river sediment29, and has been developed further for its application in plant rhizospheres30,31,32,33.
For DGT sampling, a gel sheet of a size of approximately 3 cm x 5 cm is applied onto a single plant root that is growing in the surface layer of a soil block, with a hydrophilic membrane separating the gel from the soil. During the contact time, labile nutrients and/or contaminants diffuse towards the gel and are bound immediately by the binding material incorporated in the gel. In this way, a concentration gradient, and thus a continuous net flux towards the gel is established and prevailed during the sampling time. After sampling, the hydrogel can be removed and analyzed using an analytical chemical technique allowing for spatially resolved analysis. A highly specialized and frequently used technique for this purpose is laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). In some early studies, micro particle induced X-ray emission (PIXE) was also used29. DGT sampling combined with LA-ICP-MS analysis allows for multi-element chemical imaging at a spatial resolution of ~100 µm. If highly sensitive ICP-MS techniques (e.g., sector field ICP-MS) are employed, exceptionally low limits of detection can be achieved. In a study on the effect of liming on Zn and Cd uptake by maize15, we were able to map labile Cd in the maize rhizosphere in uncontaminated soil with a limit of detection of 38 pg cm-2 of Cd per gel area. DGT, planar optodes, and zymography rely on diffusion of the target element from soil into a gel layer, which can be exploited for combined application of these methods in order to simultaneously, or consecutively, image a large number of parameters relevant for plant nutrient and contaminant uptake. Detailed information on analytical chemical aspects of DGT imaging, on the potential of combining DGT and other imaging methods, and on its applications is comprehensively reviewed in ref.34,35.
In this article we describe how to carry out a solute imaging experiment using the DGT technique on roots of terrestrial plants in a unsaturated soil environment, including plant cultivation, gel fabrication, gel application, gel analysis and image generation. All steps are elaborated in detail, including notes on critical steps and experimental alternatives.
The solute imaging protocol presented here is a versatile method to visualize and quantify 2D nutrient and contaminant fluxes in soil-plant environments. It is unique in its capability to generate sub-mm scale multi-element images of labile solute species at the soil-root interface, exceeding the achievable spatial resolution of alternative methods for measuring solute gradients in the rhizosphere substantially4. The targeted in situ sampling approach of DGT, in combination with a highly sensitive chemical analysis method such as LA-ICP-MS, facilitates the detailed investigation of solute flux dynamics around individual plant roots grown in soil or similar substrates. Due to the sink-based sampling process, the obtained images reflect the lability of the visualized solutes, and therefore are an estimation of their plant-availability10. Although the method-inherent measurement of solute fluxes bears considerable advantages like the interpretability as plant-available nutrient fractions, flux measurements are much less straight-forward to understand than porewater concentration measurements. The standard DGT sampling geometry in bulk soil applications (specifically the 0.8 mm-thick diffusion gels used in that setup) allows for comparing the actual porewater concentration, csoln, and a time-averaged porewater concentration estimate by a bulk DGT measurement, cDGT, and for the interpretation of these parameters regarding the resupply dynamics of a solute species. However, such comparison cannot be done based on imaging DGT application with very thin diffusion layers, as the derived cDGT values are unrealistically small34. DGT imaging results are therefore not always simple and quick to interpret and are often not directly comparable to more conventional porewater concentration measurements.
When applying the method, a few critical steps need to be carefully considered, mainly related to filling and watering the rhizotron growth containers. During filling the soil into the rhizotron, it is very important to avoid compacting the soil too much, as the plant roots cannot penetrate strongly compacted soil and root growth will be inhibited. We have observed roots avoiding strongly compacted soil and growing along the inner edges of the rhizotron growth container, where the soil is usually less compacted. In this case, individual roots located in the center of the rhizotrons, where DGT gels can be applied conveniently, may not develop at all, effectively inhibiting successful gel application. In our laboratory, experience showed that dry soil bulk densities of 1.0-1.4 g cm-3 allow unimpeded root development. Moreover, excessive soil compaction is also a potential source of artefacts regarding the solubility of redox-sensitive elements and biogeochemically associated species. As the total pore volume is reduced and the pore diameter distribution is shifted towards lower diameters in highly compacted soil, less air-filled larger-diameter pore volume is available, which may lead to reductive conditions locally. Consequently, MnIII/IV– and FeIII-oxides may be reduced, leading to increased Mn2+ and Fe2+ fluxes. The dissolution of Fe-oxides, which are important sorption sites e.g., for phosphate and micronutrients, may liberate sorbed and/or co-precipitated species and thereby cause artificially elevated fluxes of the biogeochemically associated species. A similar issue may arise if the growth containers are watered too much. Evaporation via the small soil surface area at the top of the growth container is low and the soil may remain water-saturated for up to several weeks after planting, which may also cause redox artefacts.
Another important consideration is the chemical functionality of the fabricated HR-DGT binding gel. By following the protocol, thin gels with a homogeneous distribution of binding phases are obtained. If the gels have areas of inhomogeneous material distribution (e.g., holes in the gel or aggregates of binding phases) these areas need to be removed or, if too extensive, the gel fabrication protocol needs to be repeated. If prepared correctly, the gel must be able to bind the target solute species that diffuse into the gel immediately and quantitatively27, which is determined by the analyte-specific gel binding capacity. While exceeding the gel capacity is less problematic in uncontaminated soils, it should be considered in metal-contaminated soils and saline soil environments. Saturation of the gel binding phases will not only impair quantitative solute sampling, but also result in lateral diffusion of solutes between binding phases in the gel, leading to an indefinite localization of small-scale solute flux features. Thus, if very high quantities of labile nutrient/contaminant species are expected in the target soil environment, preliminary tests should be performed. For estimating expected DGT loadings, bulk soil DGT piston sampling followed by gel elution and wet-chemical analysis can be applied15,49. If necessary, DGT deployment times may be adjusted to reduce the gel contact time and thus avoid gel saturation above capacity thresholds. Conversely, preliminary tests can also be helpful to identify required gel contact times and/or LA-ICP-MS sensitivities if very low solute loadings are expected, which may be important for mapping trace element solutes at natural soil background levels15. Besides, correct DGT gel functioning should be verified before its experimental application via controlled loading of gels in the preparation of DGT LA-ICP-MS calibration standards. The gel standard provides a matrix-matched reference gel analyte loading that can be used to assess if the sample gel loading determined by LA-ICP-MS is within the expected range. If unable to obtain a signal which is different from the gas and method blank background noise, the operator must ensure that laboratory procedures for trace element analysis were implemented and all protocol steps were performed correctly. Sometimes, the DGT gel is accidentally flipped after solute sampling with the soil-exposed, loaded side facing towards the glass plate rather than the laser beam, resulting in low signal intensities and erroneously flipped features in the final solute flux images.
During the LA-ICP-MS analysis, a large quantity of data is generated, which takes considerable time to evaluate. In our lab, we use in-house data evaluation scripts tailored for our target data output format using standard spreadsheet software. After semi-automated sorting and calibration, image plotting is conducted using open source, open access image analysis tools (ImageJ, Fiji50). This approach allows full control over data sorting, evaluation, and presentation, which is essential because the collected data correspond to rectangular, and not quadratic pixels, which needs to be properly displayed in the generated solute maps. Moreover, during data processing, any pixel interpolation should be carefully avoided. Pixel interpolation leads to smoothed gradients in the chemical images, resulting in softened, often circular element distribution features and is therefore an undesirable alteration of the original data. Pixel interpolation is a standard procedure in re-scaling and re-formatting operations in many image processing software products but can be deselected usually.
In conclusion, the described method is a significant advancement for understanding nutrient and contaminant dynamics in natural soil-rhizosphere-plant systems. In addition to DGT-only applications, the method can be combined with other, diffusion-based imaging techniques like planar optodes3,33,42,43,48,51 and zymography20,21,22,23,24, and may be developed further for including additional elements and soil parameters.
The authors have nothing to disclose.
This study was co-funded by the Austrian Science Fund (FWF): P30085-N28 (Thomas Prohaska) and the Austrian Science Fund (FWF) and the Federal State of Lower Austria: P27571-BBL (Jakob Santner).
(NH4)2S2O8 (ammonium persulfate; APS) | VWR | 21300.260 | ≥98.0%, analytical reagent |
2-(N-morpholino)-ethanesulfonic acid (MES) | Sigma-Aldrich | M8250-100G | ≥99.5% |
Acrylamide solution | Sigma-Aldrich | A4058-100ML | 40%, for electrophoresis |
Analyte salts | n/a | n/a | Use water soluble analyte salts of analytical grade or higher |
Buechner funnel | VWR | 511-0065 | 13 cm plate diameter |
Chemical equilibrium modelling software | KTH Sweden | n/a | Visual MINTEQ |
Clamp | Local warehouse | n/a | |
Desktop publishing software | Adobe Inc. | n/a | InDesign CS6 |
DGT cross-linker | DGT Research Ltd | n/a | 2%, agarose derivative |
DGT piston sampler | DGT Research Ltd | n/a | 2 cm diameter exposure window |
Digital single-lens reflex (DSLR) camera | Canon Inc. | n/a | Canon EOS 1000D |
Dispersion device | IKA | 3737000 | Ultra-Turrax T10 Basic |
Double-sided adhesive tape | Tesa | 56171 | |
Ethanol | Sigma-Aldrich | 34923 | Puriss. p.a., absolute, ≥99.8% |
Gel blotting paper | Whatman | 10426981 | Blotting Papers, Grade GB005, 20 × 20 cm, 1.5 mm thickness |
Gel drier | UniEquip | n/a | UNIGELDRYER 3545 |
High-pressure microwave system | Anton Paar | n/a | Multiwave 3000 |
HNO3 | VWR | 1.00456.2500P | 65%, ISO for analysis |
Horizontal shaker | GFL | 305 | |
HydroMed D4 | AdvanSource Biomaterials Corp. | n/a | Ether-based hydrophilic urethane |
ICP-MS software | Perkin Elmer | n/a | Syngistix |
Image analysis software | National Institutes of Health (NIH) | n/a | ImageJ Fiji, freely available at https://fiji.sc/ |
Knife-coating device | BYK | 5561 | Single Bar 6″, 0.5 mils |
LA software | Elemental Scientific Lasers | n/a | ActiveView |
LA system | Elemental Scientific Lasers | n/a | NWR193 |
Laminar flow bench | Telstar Laboratory Equipment B.V. | n/a | Class II biological safety cabinet |
Magnetic stirrer | IKA | 0003582400 | C-MAG MS 7 |
Moisture-retaining film | Bemis Company, Inc. | PM999 | Parafilm M, 4" x 250' |
N,N,N’,N’-tetramethylethylenediamine (TEMED) | Sigma-Aldrich | T9281-50ML | BioReagent, suitable for electrophoresis, ~99% |
NaNO3 | Sigma-Aldrich | 229938-10G | 99.995% trace metals basis |
NaOH | Sigma-Aldrich | 1064980500 | Pellets for analysis |
Overhead shaker | GFL | 3040 | |
Perfluoroalkoxy alkane (PFA) vials | Savillex | 200-015-20 | 15 mL Standard Vial, Rounded Interior |
pH meter | Thermo Scientific | 13-644-928 | Orion 3-Star Benchtop pH Meter |
pH probe | Thermo Scientific | 8157BNUMD | Orion ROSS Ultra pH/ATC Triode |
Plastic cutter | DGT Research Ltd | n/a | Use empty cross-linker vials from DGT research Ltd |
Plastic tweezers | Semadeni | 602 | |
Plasticine | Local stationary shop | n/a | non-drying plastic modelling mass based on paraffin wax and bulking agents |
Polycarbonate membrane discs | Whatman | 110606 | Nuclepore Hydrophilic Membrane, 25 mm diameter, 0.2 µm pore size, 10 µm thickness |
Polycarbonate membrane sheet | Whatman | 113506 | Nuclepore Hydrophilic Membrane, 8 × 10 in, 0.2 µm pore size, 10 µm thickness |
Polyethersulfone membrane discs | Pall Corporation | 60172 | Supor 450 Membrane Disc Filters, 25 mm diameter, 0.45 µm pore size, 0.14 mm thickness |
Polyethersulfone membrane sheet | Pall Corporation | 60179 | Supor 450 Membrane Disc Filters, 293 mm diameter, 0.45 µm pore size, 0.14 mm thickness |
PTFE foil | Haberkorn | n/a | 50 µm thickness |
PTFE spacer | Haberkorn | n/a | Variable thicknesses available |
PTFE-coated razor blades | Personna GEM | 62-0178 | Stainless steel single edge blades (coated) |
PTFE-coated Tygon tubing | S-prep GmbH | SP8180 | 0.32 cm inner diameter |
Quadrupole ICP-MS | Perkin Elmer | N8150044 | NexION 2000B |
Quantitative filter paper, 454 | VWR | 516-0854 | Particle retention 12-15 µm |
Spreadsheet software | Microsoft Corporation | n/a | Microsoft Excel 2016 (v16.0) |
Stainless-steel cutter | Local locksmithery | n/a | 2.5 cm diameter |
Suspended particulate reagent-iminodiacetate (SPR-IDA) | Teledyne CETAC Technologies | n/a | 10 µm diameter polystyrene beads, 10 % (w/v) bead suspension |
Transistor-transistor logic (TTL) cable | n/a | n/a | Consult ICP-MS technician to identify a suitable TTL cable for a specific instrument |
Two-volume cell | Elemental Scientific Lasers | n/a | Two-volume cell 1 |
Vinyl electrical tape | 3M | n/a | Scotch Super 33+ |
Water purification system | Termo Electron LED GmbH | n/a | TKA-GenPure |
ZrOCl2 × 8H2O | Alfa Aesar | 86108.30 | 99.9 %, metals basis |