Quantification of Aluminum-Induced Changes in Wheat Root Architecture by X-ray Microcomputed Tomography

ABSTRACT Root architectural traits are of fundamental importance for plant performance, especially under unfavorable soil conditions. This study examined the effect of aluminum (Al) toxicity in different growing media (nutrient solutions and soil) on root architecture of two wheat (Triticum aestivum L.) cultivars with different Al tolerances. Seedlings were grown in acidic and limed soil and in two contrasting nutrient solutions. Root systems of soil-grown plants were scanned using x-ray microcomputed tomography (µCT) while that of nutrient solution–grown plants were assesses using WinRhizo, 3 and 5 days after planting (DAP), respectively. Aluminum caused significant reduction of all examined root traits (number of seminal roots, root length, length of the longest seminal root, root surface area, and root volume). Growth in acidic soil caused significant reduction in root length, length of the longest seminal root, and root surface area at 5 DAP. Soil-grown plants produced a larger root system compared to plants grown in nutrient solutions. Aluminum toxicity–induced differences of root traits were also found between different nutrient solutions. Beside the well-known reduction of root length, Al toxicity had a profound effect on other root architectural traits. X-ray µCT has revealed root architectural changes under specific conditions of acidic, Al-toxic soil. Differences obtained in Al-induced effects on root architecture between different nutrient solutions as well as between different growing systems emphasize the need for further study of root architecture, especially under specific conditions of Al toxicity in acidic soils.


Introduction
Aluminum (Al) is a major limiting factor of crop productivity in acidic soils (Kochian 1995). Acidic (pH < 5.5) soils exhibiting Al toxicity comprise up to 30-40% of the world's arable land, and it is estimated that more than 50% of world's potentially arable land is acidic (Von Uexküll and Mutert 1995). Solubilization of Al oxides and hydroxides is enhanced by low pH, and the predominant form of Al in the acidic soils (pH <5.0) is Al 3+ (Delhaize and Ryan 1995). The most easily recognized symptom of Al 3+ toxicity is the inhibition of root growth (Delhaize and Ryan 1995). Therefore, measurement of the root growth in solution culture assays has been used for screening Al-tolerant genotypes (Samac and Tesfaye 2003). Nevertheless, in only a few cases has Al tolerance observed in solution cultures been correlated with Al tolerance in acidic soils (Samac and Tesfaye 2003). Discrepancies in genotype rankings regarding Al tolerance have been attributed to different factors that affect effective Al concentration in nutrient solutions and in addition can reduce repeatability of the results. Typically researchers used simple nutrient solutions with low ionic strength and a wide range of Al concentrations. However, Gregory and Hinsinger (1999) highlighted that research on roots needs to involve complex growth medium such as soil, opposed to commonly used hydroponics, gels, and sand culture. Furthermore, most research performed in nutrient solutions has focused on the root apex, which is the most sensitive site of root to Al toxicity, while the whole root architecture has gained less attention.
Although Al tolerance in wheat appears to be controlled by a single dominant gene (Delhaize et al. 1993a;Riede and Anderson 1996), many root traits are under polygenic control and expression of these genes is influenced by mutual interactions of roots with the abiotic and biotic soil environment (McCully 1999). The importance of root architecture for plant growth and performance, especially under environmental stress has recently gained more attention (e.g., Lynch 1995;López-Bucio, Cruz-Ramıŕez, and Herrera-Estrella 2003). Noninvasive techniques such as x-ray microcomputed tomography (µCT) provide an opportunity to examine three-dimensional (3-D) root architecture (Tracy et al. 2010) nondestructively in the opaque matrix of soil.
The aim of this study was to quantify Al-induced changes in root architecture of two wheat cultivars that differ in Al tolerance (Al-tolerant Sivka and Al-sensitive Scout 66) grown in different growing systems (nutrient solutions and soil) and to compare the usefulness of two methods (WinRhizo and x-ray µCT) for assessing Al-induced changes in the root architecture.

Plant material and growing conditions
Seeds of Al-tolerant wheat cultivar Sivka were obtained from the University of Zagreb, Faculty of Agriculture, Department of Plant Breeding, Genetics, and Biometrics (Zagreb, Croatia), and Alsensitive cultivar Scout 66 from the Crop Research Institute, Gene Bank Department (Prague, Czech Republic). Seeds were surface sterilized in 2.5% sodium hypochlorite, thoroughly rinsed with distilled water, and soaked for 6 h in distilled water. All seeds were germinated for 64 h on filter paper soaked with 0.2 mM CaCl 2 at 23/18°C with a 16/8 h, day/night regime.

Soil-based experiment
Soil samples (silty loam, luvisol) were collected from the Ap horizon of an arable field near Gospić, Croatia (44°32ʹ 45ʹʹ N, 15°22ʹ 28ʹʹ E). Soil samples were air dried and sieved to <2 mm diameter. Selected physical and chemical characteristics of the soil are shown in Table 1. To get soils with different pH values, half of the soil samples were limed using 1.0 g calcium carbonate (CaCO 3 ) kg −1 . Prepared soil samples were moistened to field capacity and incubated for 2 months at room temperature. After the incubation period, soil pH was 5.8 and Al saturation was 3.32%. Before planting, soil samples were sieved through <1.0-mm diameter mesh and were placed into 50-mm-  diameter and 100-mm-high plastic columns to achieve a bulk density of 1.0 g cm −3 . The soil was watered, maintained at a volumetric water content of 15%, and kept in growth chambers during the seed germination period (64 h). Four uniformly developed seedlings per cultivar were selected for growth (one plant per column). Germinated seeds were placed in 1-mm-diameter, 2-mm-deep holes drilled in the soil columns. The seeds were placed in the hole with the radical downwards before being covered with soil. Plants were grown in a growth chamber with 16/8 h, 23/18°C day/night regime and 75% relative humidity.

Nutrient solution experiment
The experiment was prepared as a randomized block design with ten replicate plants of each cultivar per treatment. Plants were grown on an opaque plastic mesh in two different nutrient solutions, which were previously used in experiments related to Al toxicity. The first nutrient solution (NSR) was used previously by Jurkić (1992, 1993) and the second nutrient solution (NSD) was used by Delhaize et al. (1993aDelhaize et al. ( , 1993b. Treatments were represented as control nutrient solutions, pH 4.0, without aluminum (NSR0 and NSD0, respectively), and nutrient solutions with aluminum (supplied as AlCl 3 ), pH 4.0 (NSR1, NSD1, respectively). Ionic activities and Al speciation in nutrient solutions were calculated by GEOCHEM-EZ  and are shown in Table 2. Based on the calculations, the free activities of Al 3+ were 0.0 (in NSR0 and NSD0) and 72.0 µM L −1 (in NSR1 and NSD1). Nutrient solutions were continuously aerated and replenished daily, and the pH was adjusted with 0.1 M HCl. Plants were grown in a growth chamber with 16/8 h, 23/18°C day/night regime and 75% relative humidity.

Root imaging
For the x-ray µCT scanning, the columns with live plants were scanned on the third and fifth day after planting (DAP) using a Phoenix Nanotom (GE Measurement & Control Solutions, Wunstorf, Germany) x-ray µCT scanner set at 100 kV and 210 µA, with a 0.2-mm copper filter, and voxel resolution was set at 50 µm. For each column, 1200 image projections were collected over a 30-min period. Image slices were reconstructed into 3-D volumes using software Datos|x with beam-hardening reduction algorithms applied and then visualized and analyzed in VGStudioMax 2.0 (Volume Graphics GmbH, Heidelberg, Germany). Roots were segmented from the obtained images using the Region Growing selection tool following the method of Tracy et al. (2012). Segmented root systems were used for quantitative determination of number of seminal roots, root length, length of the longest seminal root, root surface area, and root volume. After the final µCT scan at 5 DAP, roots were extracted from the soil and carefully washed and scanned using Epson Perfection V700 photo scanner and WinRhizo software (WinRhizo 2009 Reg., Regent Instruments Canada Inc.). Root measurements of the plants grown in nutrient solutions were conducted at 3 and 5 DAP, using Epson Perfection V700 photo scanner and WinRhizo software.

Statistical analysis
Data were analysed using the SAS 9.2 statistical package (SAS Institutes, Cary, NC). For the comparison of the scanning techniques (x-ray µCT versus WinRhizo), results of the root traits (number of seminal roots, root length, the length of the longest seminal root, root surface area and root volume) of soil-grown plants that were obtained at 5 DAP were compared using analysis of variance (ANOVA), followed by the use of Tukey's honestly significant difference (HSD) test. For comparisons of different nutrient solutions (NSR0, NSR1, NSD0, and NSD1), soil treatments (acidic versus limed soil), and growing systems, results of the root traits were analyzed using repeated measures (Mixed Model Repeated Measures, Littell et al. 1996).

Effects of aluminum toxicity and soil acidity on root traits
Root traits of Al-tolerant (Sivka) and Al-sensitive (Scout 66) wheat cultivars grown in different nutrient solutions with toxic concentrations of Al and in control solutions (without Al) and in acidic and limed soil are shown in Figures 1,2,3,4, and 5.
The number of seminal roots was consistently larger for Scout 66 compared to Sivka across nutrient solutions (P < 0.001) and soil treatments (P < 0.05). In both nutrient solutions (NSD and NSR) and at both measurement times (3 and 5 DAP) Al treatments reduced (P < 0.01) the number of seminal roots (from 5.01 in Al-treatment solutions to 4.61 in control solutions) ( Figure 1A). In soil the number of seminal roots increased over time, from 3.5 (3 DAP) to 4.88 (5 DAP) (P < 0.05) ( Figure 1B). . For plants grown in soil roots were scanned by X-ray µCT and measured by (VGStudioMax), and for plants grown in nutrient solutions roots were scanned and measured by WinRhizo, at 3 DAP and 5 DAP, respectively. Error bars associated with the histograms are ±1 standard error of the mean. The vertical bars represent standard error of the difference (SED) for (1) day, (2) nutrient solution, (3) nutrient solution treatment, (4) cultivars; (1') day, (2') soil treatment, (3') cultivars. Below the vertical bars (SED) ANOVA for the main effects is presented as: *Significant at the 0.05 probability level; **Significant at the 0.01 probability level; ***Significant at the 0.001 probability level; and ns = not significant. For figure A: means with the same letter are not significantly different between nutrient solution treatments within each nutrient solution type at each measurement time; for Scout 66 (capital in NSD and capital with apostrophe in NSR, respectively) and for Sivka (small in NSD and small with apostrophe in NSR, respectively). For figure B: means with the same letter are not significantly different between soil treatments at each measurement time for Scout 66 (capital) and for Sivka (small).
In nutrient solutions Al treatments reduced (P < 0.001) root length of both cultivars (Scout 66 and Sivka), grown in both nutrient solutions (NSD and NSR) and at both measurement times (3 DAP and 5 DAP). In addition for all cultivar × nutrient solution × treatment combinations, root length increased with time (P < 0.05) except for Scout 66 grown in NSR1 (P > 0.05) (70.64 mm and 85.65 mm, 3 and 5 DAP, respectively) and in NSD1 (P > 0.05) (55.6 mm and 68.47 mm, 3 DAP and 5 DAP, respectively) ( Figure 2A). Root length of plants grown in soil was affected by cultivar × treatment × measurement time interaction (P < 0.05). Reduction of root length of Scout 66 grown in acidic soil was evident at both measurement time (i.e., 87.5 mm vs. 146.23 mm at 3 DAP, P < 0.05 and 454.37, mm vs. 194.7 mm 5 DAP, P < 0.001, in acidic vs. limed soil, respectively). A significant reduction in root length of Sivka grown in acidic soil was recorded at 5 DAP (376.86 mm in acidic vs. 453.24 mm in limed soil, P < 0.05) ( Figure 2B).  . For plants grown in soil roots were scanned by X-ray µCT and measured by (VGStudioMax), and for plants grown in nutrient solutions roots were scanned and measured by WinRhizo, at 3 DAP and 5 DAP, respectively. Error bars associated with the histograms are ±1 standard error of the mean. The vertical bars represent standard error of the difference (SED) for (1) day, (2) nutrient solution, (3) nutrient solution treatment, (4) cultivars; (1') day, (2') soil treatment, (3') cultivars. Below the vertical bars (SED) ANOVA for the main effects is presented as: *Significant at the 0.05 probability level; **Significant at the 0.01 probability level; ***Significant at the 0.001 probability level; and ns = not significant. For figure A: means with the same letter are not significantly different between nutrient solution treatments within each nutrient solution type at each measurement time; for Scout 66 (capital in NSD and capital with apostrophe in NSR, respectively) and for Sivka (small in NSD and small with apostrophe in NSR, respectively). For figure B: means with the same letter are not significantly different between soil treatments at each measurement time for Scout 66 (capital) and for Sivka (small).
In nutrient solutions Al treatments reduced (P < 0.001) length of the longest seminal root of both cultivars (Scout 66 and Sivka), grown in both nutrient solutions (NSD and NSR) and at both measurement times (3 DAP and 5 DAP). In addition, there was a significant interaction of cultivar × nutrient solution × treatment (P < 0.01). No significant difference in length of the longest seminal root was obtained for Sivka grown in NSR0 (P > 0.05) (39.77 mm) and NSD0 (34.68 mm), while cultivar Scout 66 produced longer seminal root in NSR0 (P < 0.001) (51.25 mm) compared to NSD0 (31.78 mm). The opposite was obtained in Al treatment solutions where no significant difference (P > 0.05) was found between NSR1 (21.40 mm) and NSD1 (16.37 mm) grown Scout 66, while significantly (P < 0.05) longer seminal roots were obtained for NSR1 (23.24 mm) compared to NSD1 (17.55 mm) grown Sivka ( Figure 3A). In soil, length of the longest seminal root was affected by measurement time (P < 0.01), by cultivar (P < 0.01) with average length of 55.97 mm for Scout 66 Figure 4: Comparison of root surface area of wheat cultivars Scout 66 and Sivka grown in Al treatment solutions (NSR1 and NSD1) and control nutrient solutions (NSR0 and NSD0) (A) and in acid and limed soil (B). For plants grown in soil roots were scanned by X-ray µCT and measured by (VGStudioMax), and for plants grown in nutrient solutions roots were scanned and measured by WinRhizo, at 3 DAP and 5 DAP, respectively. Error bars associated with the histograms are ±1 standard error of the mean. The vertical bars represent standard error of the difference (SED) for (1) day, (2) nutrient solution, (3) nutrient solution treatment, (4) cultivars; (1') day, (2') soil treatment, (3') cultivars. Below the vertical bars (SED) ANOVA for the main effects is presented as: *Significant at the 0.05 probability level; **Significant at the 0.01 probability level; ***Significant at the 0.001 probability level; and ns = not significant. For figure A: means with the same letter are not significantly different between nutrient solution treatments within each nutrient solution type at each measurement time; for Scout 66 (capital in NSD and capital with apostrophe in NSR, respectively) and for Sivka (small in NSD and small with apostrophe in NSR, respectively).For figure B: means with the same letter are not significantly different between soil treatments at each measurement time for Scout 66 (capital) and for Sivka (small). . For plants grown in soil roots were scanned by X-ray µCT and measured by (VGStudioMax), and for plants grown in nutrient solutions roots were scanned and measured by WinRhizo, at 3 DAP and 5 DAP, respectively. Error bars associated with the histograms are ±1 standard error of the mean. The vertical bars represent standard error of the difference (SED) for (1) day, (2) nutrient solution, (3) nutrient solution treatment, (4) cultivars; (1') day, (2') soil treatment, (3') cultivars. Below the vertical bars (SED) ANOVA for the main effects is presented as: *Significant at the 0.05 probability level; **Significant at the 0.01 probability level; ***Significant at the 0.001 probability level; and ns = not significant. For figure A: means with the same letter are not significantly different between nutrient solution treatments within each nutrient solution type at each measurement time; for Scout 66 (capital in NSD and capital with apostrophe in NSR, respectively) and for Sivka (small in NSD and small with apostrophe in NSR, respectively). For figure B: means with the same letter are not significantly different between soil treatments at each measurement time for Scout 66 (capital) and for Sivka (small). compared to 74.08 mm for Sivka, and by treatment (P < 0.01) with average length 55.35 mm in acidic soil compared to 74.71 mm in limed soil ( Figure 3B).
In nutrient solutions, root surface area was affected by nutrient solution × treatment (P < 0.01) and cultivar × treatment (P < 0.001) interaction. Aluminum treatments reduced root surface area in both nutrient solutions, as well as for both cultivars. However, this reduction was more pronounced in NSR (344.51 mm 2 in NSR0 vs. 165.35 mm 2 in NSR1) compared to NSD (258.65 mm 2 in NSD0 vs. 142.57 mm 2 in NSD1) and for cultivar Scout 66 (332.99 mm 2 in control solutions vs. 142.1 mm 2 in Al-treatment solutions) compared to Sivka (270.17 mm 2 and in control solutions vs. 165.82 mm 2 in Altreatment solutions) ( Figure 4A). When grown in soil, the largest mean root surface area was obtained for Sivka (634.41 mm 2 ) compared to Scout 66 (475.14 mm 2 ) (P < 0.05), and the interaction of treatment × measurement time was significant (P < 0.05). No significant differences (P > 0.05) in root surface area were found between plants grown in acidic (422.1 mm 2 ) and limed (417.27 mm 2 ) soil at 3 DAP, while at 5 DAP plants grown in limed soil produced root systems with bigger (P < 0.01) surface area (788.59 mm 2 ) compared to those grown in acidic soil (591.16 mm 2 ) ( Figure 4B).
In nutrient solutions, root volume was affected by interactions of cultivar × treatment (P < 0.001) and cultivar × nutrient solution × measurement time (P < 0.05). Aluminum treatments reduced root volume of cv. Scout 66 at both measurements and in both nutrient solutions. On the other hand, significant reduction of root volume of cv. Sivka was found only at 5 DAP in NSR (P < 0.01) (35.9 mm 3 in NSR1 compared to 51.6 mm 3 in NSR0) ( Figure 5A). In soil, the interaction of cultivar × treatment (P < 0.05) for root volume was significant. Root volume of cultivar Sivka was greater (P < 0.05) in acidic (74.22 mm 3 ) compared to limed soil (56.11 mm 3 ), while no significant differences (P > 0.05) were found for Scout 66 grown in acidic (58.14 mm 3 ) and limed soil (62.63 mm 3 ) ( Figure 5B).

Comparison of the scanning techniques: X-ray µCT versus WinRhizo
A comparison of the root traits (root length, length of the longest seminal root, root surface area, and root volume) measured by VGStudioMax after x-ray µCT scanning and by WinRhizo (after washing soil from roots) at 5 DAP are shown in Figure 6. Although all measured root traits were slightly larger when measured by WinRhizo compared to VGStudioMax, there were no significant difference in root length (P > 0.05), root surface area (P > 0.05), and the length of the longest seminal root (P > 0.05) when these two techniques were compared. However, a significantly larger (P < 0.05) root volume was obtained by WinRhizo (117.96 mm 3 ) compared to VGStudioMax (89.44 mm 3 ), which can be attributed to the former capturing more of the finer roots (Figure 7).

Discussion
Although all measured root traits were larger when measured by WinRhizo compared to VGStudioMax, comparison of the results of root traits obtained by these two scanning techniques showed that they did not differ significantly, except for root volume. Relatively poor correlation between root volumes measured destructively by WinRhizo and nondestructively, after x-ray µCT scan, was already described by Tracy et al. (2012). Tracy et al. (2012) have attributed these discrepancies to the better contrast between roots and their surroundings that can be obtained using WinRhizo and on the other hand to the image resolution limitation gained by x-ray µCT. This could also be the truth for our results (Figure 7). An additional technical disadvantage of the x-ray µCT scanning technique is the limited soil volume that can be used for growing plants, which prevents this technique from being used to study older plants with more complex root architecture. Namely, all roots of the plants used in this study reached bottom and/or side walls of the columns by 5 DAP (Figure 7). However, results of this study showed that the x-ray µCT scanning technique provide reliable and good quality 3-D scans of roots in the soil, and despite its current limitations, new developments of this technique, such as automated root segmentation, and bigger, faster, and more precise x-ray CT scanners with greater resolution would give the opportunity to study older, more complex root systems (for the review, see Mooney et al. 2012).
Aluminum toxicity reduced all examined root traits in the experiment with nutrient solutions while in soil-based experiments it caused reduction of root length, length of the longest seminal root, and root surface area. Aluminum-induced reduction of root size is most likely the primary cause of commonly described symptoms of Al toxicity, such as impairment of nutrient and water acquisition. Aluminum toxicity, both in acidic soil and in Al-treatment nutrient solutions, caused a more pronounced reduction of all examined root traits for Al-sensitive cv. Scout 66 compared to Altolerant cv. Sivka (Figures 1, 2, 3, 4, and 5). Differences in root traits determined between cv. Scout 66 and cv. Sivka are in accordance to their tolerance to Al. It is well known that there is significant genetic variability in Al tolerance among wheat cultivars, and cv. Scout 66 was used as a model of an Al-sensitive cultivar in previous studies related to Al toxicity (e.g. Rengel and Jurkić 1992;Ryan, Shaff, and Kochian 1992); on the other hand cv. Sivka was evaluated as moderately tolerant cultivar in a screening for Al tolerance among Yugoslavian wheat cultivars (Rengel and Jurkić 1992).
The first and most easily recognized symptom of Al toxicity is the inhibition of root growth (Delhaize and Ryan 1995). Barceló and Poschenrieder (2002) stated that sensitive plants exhibit statistically significant inhibition of root elongation after approximately 30 min to 2 h exposure. Our results show that Al toxicity caused slower reduction of root growth in acidic soil compared to those that were obtained in experiments with nutrient solutions. For example, reduction of root length and root surface area for plants grown in Al treatment solutions was evident at 3 DAP while reduction of root length for acidic soil-grown cv. Sivka and reduction of root surface area for both acidic soilgrown cultivars was evident only at 5 DAP. These delayed response to Al toxicity observed for acidic soil-grown plants could be explained as a lag phase. Barceló and Poschenrieder (2002) described the lag phase as the time or concentration required for Al to interfere with key processes in root growth. It was estimated (Delhaize et al. 1993a) that significant Al inhibition of root growth in wheat occurs at root tip Al concentrations around 1000 µg Al g −1 . Therefore, these results indicate that acidic soilgrown plants, especially cv. Sivka, can tolerate a longer period of exposure to toxic Al concentrations. Although there are some reports about the Al induced inhibition of lateral roots in sensitive genotypes of rice (Oryza sativa L.) (Famoso et al. 2010), soybean (Glycine max L.) (Vilagarcia et al. 2001;Silva et al. 2001), and maize (Zea mays L.) (Clark et al. 2013), there is lack of data about the effect of Al toxicity on other root traits, especially under real acidic soil conditions. Villagarcia et al. (2001) developed a sand-based screening technique, which simulated growth in acidic soil. In their experiments, they compared hydroponic and sand-based experiments by measurements of different root traits of soybean. These authors reported Al toxicity (18 days of exposure to 450 µM Al L −1 ) in sand-based experiments did not greatly affect the tap root length, while it caused significant reduction of root surface area (by 58%) compared to the control, probably due to reduction in length of basal roots and branches. In our experiments, Al toxicity induced reduction of early stage root volume for both cultivars grown in Al-treatment nutrient solutions ( Figure 5A), while soil acidity did not affect root volume of cv. Scout 66 and that of cv. Sivka was greater when grown in acidic compared to limed soil ( Figure 5B). Aluminum-injured roots are often described as stubby and brittle, with thickened lateral roots (Foy, Chaney, and White 1978). Possible explanations of equal root volume (limed and acidic soil grown cv. Scout 66) or increased root volume in acidic soil- grown cv. Sivka could be the Al-induced increase in viscous and elastic extensibility of cell wall of the root apices (Ma et al. 2004) or Al-induced reduction of cell length accompanied by radial cell expansion, which was found on Al-treated rice roots (Alvarez et al. 2012).
Plants grown in acidic soil produced larger root system (root length, length of the longest seminal root, root surface area, and root volume) compared to plants grown in Al-treatment solutions. These results could be explained by greater activities of toxic Al in Al-treatment solutions (Table 2), as well as possible mitigating effect of soil compounds like plant nutrients and organic matter on Al toxicity. Despite the high Al saturation percentage of soil used in this experiment (Table 1), Delhaize and Ryan (1995) found that exchangeable Al in soil is a poor indicator of Al toxicity. In sand-based experiments, Villagarcia et al. (2001) reported that an approximate 100-fold increase in Al concentration was required to inhibit root growth to a comparable degree to hydroponic-based experiments. However, high concentrations of toxic Al are not the only reason for decreased root size in nutrient solutions. This statement is supported by the fact that acidic soil-grown plants produced a larger root system compared to plants grown in the control nutrient solutions. Reduced root growth of plants grown in nutrient solutions could be explained by stress caused by transfer of young seedlings to hydroponics (Tamas et al. 2006). Another possible explanation could be the more efficient detoxification of Al in soil due to slower diffusion rates of organic acids (malate) away from root surface and Al toward root surface. Kinraide, Parker, and Zobel (2005) proposed biphasic diffusion hypothesis of Al detoxification, which suggests that majority of Al detoxification occurs just beneath the root epidermis. Our observed increase in root volume in acidic soil-grown plants possibly caused by radial expansion of epidermal and cortex cells may represent the evidence for such detoxification.
Despite equal concentrations of free Al in both NSD1 and NSR1 solutions (Table 2), Al toxicity caused more pronounced reduction of root growth in NSD compared to NSR. A possible explanation may lay in the different concentration of nutrients in these two nutrient solutions, especially those of calcium and magnesium, and differences in ionic strength of the solutions (Table 2). With the increasing ionic strength of the nutrient solution increases the competition between Al 3+ and other cations for negatively charged sites within the root cell wall and plasma membrane. Because of the complex chemistry of Al and its multiple interactions with different nutrients in solution, in previous studies of Al toxicity researchers used simple nutrient solutions with low ionic strength and wide range of Al concentrations (from 5 to 200 µM L −1 ) (Wang et al. 2006), often avoiding usage of different plant nutrients, such as sulfur and phosphorus (Samac and Tesfaye 2003). However, it has been well documented that different concentrations of nutrients such as nitrate, phosphate, sulfate, and iron can lead to alterations in root growth and architecture (for a review, see López-Bucio, Cruz-Ramıŕez, and Herrera-Estrella 2003).
Results of this study indicate that beside the well-known reduction of root length Al toxicity also has a profound effect on other root traits; for example, in nutrient solutions Al toxicity reduced the number of seminal roots, the length of the longest seminal root, the root surface area, and root volume. In addition, differences obtained in Al-induced effects on root architecture between different nutrient solutions (NSD and NSR) and even more profound differences found between two growing systems (soil and nutrient solutions) emphasize the need for further investigation of wheat root architecture under specific conditions of Al toxicity. In previous experiments Al toxicity was studied under simplified conditions. X-ray µCT provides the opportunity to nondestructively study 3-D root system development in their natural environment of soil. With the further development of this technique, it will be possible to examine larger number of samples, to monitor root development over a more prolonged period across the growth cycle of a plant, and to include different environmental factors or plant microbial interactions that could have significant effect on Al toxicity. For example, it would be useful to investigate Al-induced root architecture changes across specific soil pH ranges (pH 4.0-6.0) in which Al toxicity occurs in arable soils. Furthermore, considering that in many arable soils Al toxicity occurs in acidic subsoil layer, further research should focus on larger number of genotypes and on root architectures of mature more established plants.