Vanadium Pentoxide Phytotoxicity: Effects of Species Selection and Nutrient Concentration
- First Online:
- Cite this article as:
- Smith, P.G., Boutin, C. & Knopper, L. Arch Environ Contam Toxicol (2013) 64: 87. doi:10.1007/s00244-012-9806-z
- 321 Views
Vanadium concentrations in soil can be increased through anthropogenic inputs and can be harmful to plants. A Petri dish experiment was conducted to assess the effect of vanadium toxicity on the germination and survival of the garden lettuce, Lactuca sativa. A second study was conducted in a greenhouse to investigate the influence of species selection and nutrient concentration on the toxicity of vanadium pentoxide to plants. L. sativa and four non-crop native plant species, two grasses (Elymus virginicus and Panicum virgatum) and two broad-leaved species (Lycopus americanus and Prunella vulgaris) were selected. Artificial soil was used in both experiments, and a geometric progression of five vanadium concentrations plus controls was selected for the soil treatments. Results of the Petri dish experiment showed that seedling survival is a less sensitive end point than above-ground dry weight (DW) as measured in the greenhouse experiment. Nutrient level (100, 10, and 1 kg/ha) was found to strongly influence vanadium toxicity in the greenhouse study. At 100 kg/ha, plant tolerance to vanadium was greatest, as indicated by higher no-observed, lowest-observed, and percentage effect concentration values. Results showed that forbs (L. americanus and P. vulgaris) tended to be more sensitive than both the crop (L. sativa) and grasses (E. virginicus and P. virgatum) at high concentrations of vanadium. Soil concentrations resulting in a 25 % decrease in shoot DW were generally less than the Canadian soil quality guideline for vanadium, suggesting that 130 mg/kg may not be protective of the Canadian native plant species used in this study.
Vanadium is an abundant element in the Earth’s crust at a mean concentration of 150 g/t (Moskalyk and Alfantazi 2003; Anke 2004). Overall, the average amount of vanadium in soils worldwide ranges from 10 to 220 mg/kg and is especially abundant in limestone soils (Ovari et al. 2001; Poledniok and Buhl 2003). In Canada, mean soil concentrations of naturally occurring vanadium ranging from 16 to 220 mg/kg have been reported (Canadian Council of Ministers of the Environment [CCME] 1999; Environment Canada 2010). However, significantly greater concentrations of vanadium (1,510–3,600 mg/kg) have been found in areas directly affected by anthropogenic inputs (Panichev et al. 2006).
One of the primary anthropogenic inputs of vanadium into the environment is the combustion of fossil fuels, such as coal and oil (Moskalyk and Alfantazi 2003; Meija et al. 2007; Environment Canada 2010). Modelled air stack emissions from coal combustion (Lee and Wu 2002) suggest that 90 % of the vanadium released from this source can be in the form of toxic vanadium pentoxide (V2O5). The V2O5 content of the fine ashes suspended in the air varies from 1 to 18 % (World Health Organization 2000). V2O5 is also an important compound for numerous industrial applications, such as a component of metal alloys and as a catalyst for the production of sulfuric acid (Khorfan et al. 2001). Vanadium can be mined, such as in South Africa (Panichev et al. 2006), or extracted as a byproduct of industrial activities as performed in China and Russia (Moskalyk and Alfantazi 2003). Regardless of the source, inputs from anthropogenic activities contribute to increased concentrations of vanadium in soil.
It has been shown that vanadium is accumulated by plants when present in soil (Bache et al. 1991), and some perennial plants and mushrooms have proven to be extremely rich in this element (Anke 2004). Although the biological importance of vanadium as a trace element for plants is still debated (Venkataraman and Sudha 2005), it is widely accepted that increased levels of vanadium in soil can be harmful to a variety of organisms, including plants (Ullah and Gerzabek 1991; Environment Canada 2010). The growth of plants can be stimulated by trace quantities of vanadium (1–10 μg/L), but concentrations >100 μg/L were found to be toxic (Venkataraman and Sudha 2005). Gil et al. (1995) found that HN4VO3 was toxic to lettuce grown in hydroponic solution containing as little as 0.2–1 mg V/kg. Soybean grown in different types of soils expressed maximum toxicity at 30 mg V/kg in fluvo aquic poor clay soil (Wang and Liu 1999). Although a vanadium concentration of 140 mg/kg in soil may be toxic to plants, mustard and spinach did not show any sign of toxicity at concentrations reaching 170 and 250 mg/kg, respectively (Anke 2004). These examples suggest that plant species selection plays an important role in determining vanadium phytotoxicity.
Phytotoxicity can also be influenced by a wide range of environmental factors that can affect vanadium availability. Many studies focus on the effects of pH and soil type; however, Kalsch et al. (2006) showed that phytotoxicity of 2,4,6-trinitrotoluene (TNT) was also influenced by nutrient availability. There is evidence for interactions between vanadium and soil nutrients (magnesium, calcium, phosphorus) negatively affecting soybean yield (Olness et al. 2001a, b). Vanadates (compounds containing an oxanion of vanadium) can be taken up by plants in place of phosphate (Olness et al. 2001a, b), and can be mobilized in nitrogen-poor soils by nitrogen-fixing organisms, such as Azotobacter vinelandii (Bellenger et al. 2008), thus making them available for plant uptake.
To assess the potential risk of vanadium in soil to community-based receptors (e.g., soil invertebrates, terrestrial plants), many countries have developed guidelines that can be used to assess toxicity based on community responses (rather than at the species level). In Canada, a vanadium soil-quality guideline for the protection of environmental health of 130 mg/kg has been provided based on toxicological studies for vascular plants and soil invertebrates (CCME 1999). Similar ecological guidelines have been developed for plants and/or soil invertebrates in the Netherlands (42 mg/kg), the Czech Republic (180 mg/kg), and Slovenia (120 mg/kg) (Cappuyns 2012). Some regions of the United States (United States Environmental Protection Agency Regions 4 and 6) have adopted the Oak Ridge National Laboratory plant screening benchmark for vanadium of 2 mg/kg (Efroymson 1997). For the Canadian guideline, the applicability of the data used in the assessment of toxicological risk to native noncrop plants exposed to vanadium in soil can be questioned because it was based on limited data that used the crop species cabbage, lettuce, and radish as the test organisms (Environment Canada 1999). Although often used as test organisms in laboratory phytotoxicity experiments, these species are not part of the natural plant communities encountered when conducting ecological risk assessments.
The purpose of this study was to investigate the influence of species selection and nutrient level on the toxicity of V2O5 to plants. Focus was placed on noncrop native plant species, which were considered to be representative of flora observed at sites where risk assessments might be conducted in Canada. An additional objective was to test the value of using short-term germination and emergence studies in risk assessment.
Materials and Methods
This article presents results from both a Petri dish vanadium experiment and a greenhouse vanadium experiment. The materials and methods are similar for both experiments and are discussed in the following sections.
Plant species used in this study were selected because seeds were readily available, easy to handle, and simple to germinate. Four of the plant species selected are found in natural environments in Canada and included two grasses (family: Poaceae) and two forbs (family: Lamiaceae). The grasses E. virginicus (lot no. 9-330) and P. virgatum (lot no. 04SD4) were obtained from the Ontario Seed Company (Waterloo, ON). The forbs L. americanus (lot no. PM1013J) and P. vulgaris (lot no. 17200) were obtained from Prairie Moon Nursery (Winona, MN) and Richters (Goodwood, ON), respectively. In addition, lettuce (L. sativa) was selected for use in the experiments presented here as it is frequently used in phytotoxicity studies. Seeds of the lettuce cultivar Buttercrunch were obtained from the Ontario Seed Company (Waterloo, Ontario; lot no. 11-3064).
The native plants are perennial species [http://plants.usda.gov/java/]. The grasses grow equally well in wetland or nonwetland environments; whereas L. americanus almost always occurs naturally in wetlands, and P. vulgaris usually grows in nonwetland conditions [http://plants.usda.gov/java/]. E. virginicus (Canadian wild rye) can be found in all of the Canadian provinces and grows best in soils with pH 5–7. P. virgatum (switch grass) can be found east of the Rocky Mountains from Saskatchewan to Nova Scotia in Canada and can grow in a soil pH range of 4.5–8. L. americanus (water horehound) can be found throughout southern Canada in soil with pH ranging from 5.2 to 7.8. P. vulgaris (selfheal) can be found in all provinces as well as the Yukon and grows best in soil with pH ranging from 5.4 to 8. Given the wide range of these species in Canada, exposure to vanadium deposition from anthropogenic sources is likely.
An artificial soil (AS) was made according to an Environment Canada standard for measuring the emergence and growth of terrestrial plants exposed to contaminants in soil (Environment Canada 2005). Batches of AS consisted of 10 % peat, 20 % clay, and 70 % sand based on a dry weight (DW) basis (Supplemental Data, Table S1). Peat was sieved through a wire mesh of approximately 5 mm, and in most cases calcium carbonate (CaCO3) was added to adjust the pH. The final average soil pH of approximately 4.7 was close to the optimum pH range for most plant species. The AS was stored for a minimum of 3 days in a cool, dark area before commencing the experiments.
Vanadium Solution and Soil Treatments
Vanadium was added to the AS as a solution. A stock vanadium solution was made from V2O5 using the method provided in HydroQual (1995) with some modifications. A mixture of V2O5 in water was heated to 70 °C and the pH adjusted with 1 M sodium hydroxide (NaOH) until it turned from an orange turbid mixture to a clear and colorless solution. The pH was then adjusted to approximately 7 with 1 M hydrochloric acid (HCl). The solution was made within 1 week of commencing the experiment. The final measured concentration of vanadium in the stock solution was 3.7 mg/L.
Measured vanadium soil concentrations at completion of experiments
Vanadium concentration (±SD) in soil (mg/kg)
Grasses and forbs
Sodium hydroxide and HCl were used to make the V2O5 solution. These chemicals were not anticipated to affect germination or survival of plants (HydroQual 1995); however, controls containing NaOH and HCl in the absence of vanadium were included in the experimental designs.
Samples of solution were analyzed by Maxxam Analytics (Ottawa, ON), and total vanadium in the stock solution was determined by inductively coupled plasma-mass spectrometry (ICP-MS). Soil samples were analysed by Maxxam Analytics and Paracel Laboratories Ltd (Ottawa, ON, Canada), and total vanadium was determined by ICP-MS according to an aqua regia digest. Reportable detection limits for solutions ranged from 1 to 10,000 μg V/L and for soil it ranged from 1 to 50 μg V/g.
For the Petri dish experiment, all of the soil from each treatment was submitted for chemical analysis. The greenhouse soil samples were a composite of six randomly selected pots of soil collected for each treatment with the root material removed.
Petri Dish Experimental Design
The Petri dish experiment design was based on the HydroQual (1995) seedling-emergence test. Two ASs were used: one amended with CaCO3 (pH 4.7) and the other without CaCO3 (pH 3.8). Twenty seeds were spread over 30 g of AS in a Petri dish (diameter 90 mm, depth 20 mm). Vanadium solutions were added to the Petri dishes to obtain a final moisture content equivalent to 80 % of the soil water-holding capacity according to the method described by Environment Canada (2005). Six vanadium treatments (Table 1) were tested on each soil type, with three replicates (i.e., three Petri dishes) per vanadium treatment. The CaCO3-amended soil was also used to test the effect of NaOH and HCl on L. sativa germination and survival in the absence of vanadium. Petri dishes were sealed with parafilm, placed in a growth chamber (model PGW36; Conviron, Winnipeg, Canada), and exposed to a photoperiod of 16 h light (photosynthetic active radiation: 274 ± 42 μmol photons/m2/s) and 8 h dark for 5 days. Temperature within the modules averaged 27.6 ± 0.4 °C during the day and 15.3 ± 0.04 °C at night. End points for the Petri dish experiment included percent germination scored on day 3 and percent survival (defined as two open, healthy cotyledons) scored on the day 5.
Greenhouse Experimental Design
A new batch of AS amended with CaCO3 was used in the greenhouse experiment (average pH 4.8). The toxicity tests included five plant species (two grasses, two forbs, and one crop), five vanadium concentrations, four controls, three nutrient levels, and six replicates/vanadium treatment (one plant/pot) for a total of 675 pots (9 × 9 × 10 cm in size). Aliquots of vanadium stock solution were diluted such that 4 L of liquid were mixed with the AS to achieve the final desired soil vanadium concentrations (Table 1).
One control consisted of tap water only, and six replicates were included. The three remaining controls consisted of NaOH and HCl at concentrations that corresponded to the amount of these chemicals that would be added to treatments with the highest vanadium concentrations. Three replicates were included for each of the NaOH and HCl controls.
The three nutrient concentrations were chosen based on the nitrogen, phosphorus, and potassium values (minimum starting soil concentration) for lettuce grown in organic soils provided in Fertilizer Recommendation Tables for Ontario, Canada [www.omafra.gov.on.ca]. The high-nutrient level treatment contained approximately 100 kg/ha 20:20:20 (nitrogen, phosphorus, and potassium, respectively) All Purpose Fertilizer (Plant Products Co. Ltd., ON, Canada), whereas the medium- and low-nutrient level treatments contained approximately 10 and 1 kg/ha of 20:20:20 fertilizer, respectively.
Seeds were germinated in trays containing unamended AS (no CaCO3), and seedlings with fully emerged cotyledons (approximately 7 days) were transplanted into amended AS. The experiment was conducted in the Environment Canada, National Wildlife Research Centre, greenhouse (Ottawa, ON) from May 18, 2011, when seedlings were transferred until June 20, 2011, when plants were harvested. The average minimum and maximum temperatures recorded during this period were 17 and 35 °C, respectively. Nutrients were added to the surface of pots on days 4 and 21 by pipetting 10 mL of the appropriate concentration of 20:20:20 fertilizer onto the soil surface. After fertilization, plants were not watered again for at least 12 h. During the course of the experiment, plants were watered as needed (generally twice a day) from the top with tap water. The experiment was terminated within 4–5 weeks of exposure to vanadium. Shoot material was harvested, stored in paper bags, and placed into a drying oven at approximately 70 °C for at least 3 days before weighing.
End points for the greenhouse experiment included visual rating comparing vanadium treatment plants with control plants (Hamill et al. 1977; Supplemental Data, Table S2), shoot DW, grass shoot length from soil to longest leaf tip, and number of forb and crop leaves.
Normality of untransformed, square root-transformed, and log-transformed data were determined using Shapiro–Wilk test, and homogeneity of variance was determined on the residuals using Levene’s test. When the assumptions of normality and heterogeneity were met, those data were analysed using one-way analysis of variance, followed by a Bonferroni test, when all variables needed to be compared or Dunnett’s test when only comparison with the control treatment was required. When data were found not to meet the assumptions of normality and/or homogeneity of variance, they were analysed using Kruskal–Wallis test followed by Conover–Inman post hoc test. The minimum significant alpha level was 0.05.
Results of statistical analysis on the effect data (e.g., % survival, DW, length, and number of leaves) were used to determine the no-observable (lowest concentration with statistically no effect) and the lowest-observable effect concentrations (lowest concentration with statistically significant effect) for this study.
The effect concentrations at which 25 % (EC25) and 50 % (EC50) of the population exhibited a response to V2O5 are also presented. The EC25 and EC50 values were determined for data meeting the assumptions for normality and homogeneity using nonlinear regression models (e.g., Gompertz, logarithmic) in Systat version 13.00.05 (Systat software, Chicago). For all remaining nonparametric data, the ICp computer program (Norberg-King 1993) was used with 240 resamples. Any statistical comparisons among plant species used the relative effect data compared with the control average (e.g., shoot DW/average control shoot DW).
The greenhouse-measured values (Table 1) for each vanadium concentration represent an average across all nutrient levels. For each nutrient treatment, the average relative SD (RSD) for any given vanadium concentration, was <15 % with the exception of one soil treatment, which had an RSD of 38 %. In addition, there was no significant difference between vanadium soil concentrations at the three different nutrient levels (Kruskal–Wallis, p > 0.05) for any given vanadium treatment.
Petri Dish Experiment
For each soil type, the RSD for each vanadium treatment was calculated for both germination and survival results. The RSDs for both the germination and survival results were <30 % in all cases. Calculated RSDs <30 % are generally considered to indicate good analytical quality assurance/quality control by commercial laboratories; therefore, the data were considered acceptable for further analysis.
Survival end points for the L. sativa Petri dish experiment
Statistical analysis using shoot DW data from the greenhouse experiment indicated that there was no significant effect (p > 0.05) of the chemical control containing NaOH/HCl treatments without vanadium on growth of any of the plant species (data not shown). Therefore, any significant end point effects observed were not attributed to the addition of NaOH and HCl to the soil.
Through visual assessment, the effects of vanadium on plant species were recognized as decreased growth, necrosis, and mild chlorosis (particularly in L. sativa). At the lowest nutrient level, leaf discolouration, abnormal leaf growth (e.g., twisting and wrinkling), and lightened veins were observed in some plants.
Calculated EC25 and EC50 values for shoot DW values of plants from the greenhouse experiment
Effect concentration (mg V/kg soil)
75 (7.3– > 265)
Vanadium, found naturally in the environment and also introduced through anthropogenic activity, can be toxic to plant species. Toxicity levels assessed in this study were lower than concentrations found in soils in Canada and elsewhere (Rasmussen et al. 2001; Ovari et al. 2001; Poledniok and Buhl 2003). Vanadium toxicity to plants can be reflected in an imbalance in other cations within the plant (Kaplan et al. 1990), which may account for the decreased growth and chlorosis observed in the greenhouse study. In addition, vanadium has been reported to disrupt the function of several enzymes (Irwin et al. 1997) and inhibit plasma membrane proton-pumping ATPases (Gallagher and Leonard 1982; Quist and Hokin 1978; Cantley et al. 1977), likely through competitive inhibition of phosphate-binding on the enzymes (Bowman 1983). This mode of action could result in the observed plant necrosis. There is some evidence that vanadium is an essential element required for plant growth, but studies to date are inconclusive (Venkataraman and Sudha 2005; CCME 1999).
The first objective of this study was to investigate the influence of species selection on the toxicity of V2O5 to plants. Part of the difficulty in environmental risk assessment is to base the measures on representative organisms or groups of organisms and correctly consider uncertainties associated with the methodology used. The crop species used as test organisms in commonly toxicity studies do not represent the terrestrial plants generally encountered as part of ecological risk assessments. The native species selected for this greenhouse experiment are commonly found across Canada, may be present in areas of vanadium contamination, and thus are considered more suitable receptor species for toxicity tests that may be used in the derivation of guidelines. Significant differences were found between species, suggesting that testing one or two species is not sufficient for deriving soil guidelines to be used in risk assessment. Yet, in phytotoxicity studies it is widely assumed that crop species can be used as surrogates for native wild species, although there is a paucity of data to confirm this claim (White and Boutin 2007; Carpenter and Boutin 2010). In these two studies, it was found that when plants were grown under similar controlled conditions, there were no significant differences between crop and noncrop species. This is partly in agreement with the present study in which the crop species L. sativa was no more or less sensitive than the native species used in this study but only at doses <90 mgV/kg soil. However, in other work with herbicides, native species were found to be more sensitive than crop species (Clark et al. 2004; Olszyk et al. 2008). In greenhouse experiments with rare earth metals, it was established that the native species Asclepias syriaca (common milkweed), Desmodium canadense (showy tick trefoil), and, to a lesser extent, Panicum virgatum (switchgrass) were consistently more sensitive to lanthanum, yttrium, and cerium than the two crops Solanum lycopersicum (tomato) and Raphanus sativus (radish) (Boutin et al. 2012a, b; Thomas et al. 2012). The number of species to be used in phytotoxicity studies remains unresolved. A large body of literature is available using herbicides, and it has been argued, for practical reasons and with the use of statistical methods, that 6–10 plant species is acceptable for pesticide risk assessment (Aldenberg and Slob 1993; Boutin et al. 2012c), although this is debated (Newman et al. 2000).
A second objective of this work was to investigate the influence of nutrient level on the toxicity of V2O5 to plants. Under more stressful environmental conditions (i.e., lower nutrient levels), all of the plant species were more sensitive to vanadium. This is a very important finding that should be taken into account in risk assessment. Similar results were obtained for Avena sativa (oat) grown in LUFA 2.2 standard soil spiked with TNT (Kalsch et al. 2006). For shoot length and weight, the no-observable and lowest-observable effect concentrations and EC50 value for A. sativa exposed to TNT were always lower in unfertilized soil than in fertilized soil. This suggests that site-specific factors, such as environmental conditions, are important when considering toxic effects of compounds, including vanadium toxicity. It should be noted that nutrient levels in natural soils can be greater than the highest nutrient dose tested in this experiment and are variable (Page-Dumroese et al. 2000).
Signs of vanadium phytotoxicity have been observed to coincide with imbalances in micronutrients, macronutrients, and metal cations within the plant (Kaplan et al. 1990), including iron (Kohno 1986; Ueoka et al. 2001a), molybdenum (Fargašova and Beinrohr 1998), magnesium, potassium (Gil et al. 1995; Ueoka et al. 2001b), manganese, copper, nickel (Fargašova and Beinrohr 1998; Kohno 1986), calcium (Gil et al. 1995; Kaplan et al. 1990), and possibly aluminum (Kaplan et al. 1990). It has also been noted that vanadium availability to plants over time is influenced primarily by binding to soil constituents (Martin and Kaplan 1998), and important sinks include manganese-oxides, organic matter, phosphates, carbonates, and sulfides (Basta et al. 2005), the concentrations of which can also indicate the fertility of the soil.
An additional objective of this work was to test the value of using short-term germination and emergence studies in risk assessment. It was found that the toxicity of vanadium to lettuce was more pronounced in the greenhouse experiment (EC25 4–120 mg V/kg depending on the nutrient level) than in the Petri dish experiment (EC25 162 mg V/kg) in CaCO3-amended soil. In an experiment similar to the Petri dish work (HydroQual 1995), the calculated EC25 (134 mg/kg [range 63–230]) and EC50 (251 mg/kg [range 172–341]) were slightly lower than those reported here; however, results from the Petri dish experiment in the present study fall within the ranges reported by HydroQual (1995). It has been known for some time that germination and emergence variables were not as sensitive end points as growth tests (Ernst and Nelissen 2000). The short-term period of the Petri dish experiments appears to be insufficient for toxicity to be manifested. The results presented here also demonstrate the inadequacy in using germination and emergence as end points in bioassays.
The purpose of this study was to assess the influence of environmental factors and species sensitivity on vanadium toxicity to plants in greenhouse experiments. A further objective was to determine whether short-term emergence tests were appropriate for use in risk assessment. Results showed that toxicity level was both species and nutrient dependent. The two native forbs tested tended to be more sensitive at high doses of vanadium than both the crop and the two grasses. Nutrient level strongly influenced vanadium toxicity with greater sensitivity observed in plants growing under conditions of nutrient stress. In addition, the germination-and-emergence test conducted with the crop species L. sativa showed the lower sensitivity of such a short-term test.
We gratefully acknowledge the in-kind analytical support provided by Maxxam Analytics and Paracel Laboratories Ltd (Ottawa, ON). Thanks are also given to David Carpenter and all participants involved in the maintenance and set-up of the experiments. The present study received financial support from the Natural Sciences and Engineering Research Council of Canada, the Industrial Research and Development Fellowship Program, and the Chemical Management Plan of Environment Canada.