Environ. Sci. Technol. 2005, 39, 2135-2142
Uptake, Metabolism, Accumulation and Toxicity of Cyanide in Willow Trees MORTEN LARSEN, AHMED S. UCISIK, AND STEFAN TRAPP* Environment & Resources DTU, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark
Chemicals taken up into plants may be accumulated so leading to toxic effects. Uptake and phytotoxicity of free cyanide was determined with the willow-tree transpiration test. Willow sets were grown in sand and irrigated with varying levels of cyanide (CN). Toxicity was determined by measuring transpiration. At CN concentrations below 10 mg/ L, no toxic effects were observed. At 20 mg/L, transpiration was reduced to approximately 50% after 96 h. With 30, 40 and 50 mg/L, the transpiration decreased with a similar rate to < 20% of the initial transpiration within 96 h. Accumulation of cyanide in plant tissue was observed at 40 and 50 mg/L. The kinetics of metabolism of cyanide by roots, stems and leaves of willows was determined by the closed-bottle metabolism test. The Michaelis-Menten parameters vmax and KM (maximal metabolic velocity and halfsaturation constant, respectively) were determined by nonlinear regression. Estimates of uptake and metabolism were balanced using a nonlinear mathematical model. The model predicted that at low doses ( 99% of the KCN in aqueous solution dissociates to release free cyanide (HCN and CN-) at pH 7. Four replicates were made for each concentration. Note that concentrations are expressed as mg CN/L or as mg CN/kg and that 1 mg of KCN equals 0.4 mg of CN. Chemical Analysis. The concentration of free and total cyanide in solution and plant, respectively, was analyzed according to the ISO 11262 standard method (9). Free cyanide reacts with chloramine-T to form cyanogen chloride. This reacts with pyridine-4-carboxylic acid and 1,3-dimethylbarbituric acid to form a colored complex, which is determined photometrically at a wavelength of 606 nm. Total cyanide in roots, stem and leaves was released as hydrogen cyanide by reflux distillation of the sample for 2 h at pH < 2. The hydrogen cyanide was collected in a sodium hydroxide scrubber. To avoid interference from sulfate, the sample was tested for sulfate on lead acetate test paper (10) and afterward analyzed as described above. Phytotoxicity Test. Sets of willow trees (Salix viminalis) were cut in January during frosty weather and placed in a -4 °C freezer. Approximately 2 weeks before the start of the experiments, the sets were thawed and cut in lengths of 30 ( 5 cm. The lower 10 cm of the cuttings was placed in an aqueous solution containing 1 mg/L of indolebutyric acid and 1 mg/L of nicotinic acid for 60 min to induce root growth. Afterward the cuttings were placed in buckets with 10 to 15 cm depth of tap water which were placed at the window. The cuttings were pregrown for approximately 2 weeks, until the mass of roots and leaves was considered sufficient for the trees to survive in sand. The outermost third of the roots was cut off, to increase the formation of root hairs, and the weighed cuttings were then transferred to 500 mL Erlenmeyer flasks. The flasks were filled with 700 g of standard sand no. 1 (sieved to 0.40 to 0.90 mm) from Dansand, 8740 Brædstrup, DK. Tap water (150 mL) was added to the flasks, and they were covered with aluminum foil to prevent algal growth. Cork stoppers, with a hole in the middle, were fitted into the necks of the flasks, to avoid evaporation from the system. Further sealing of the flask was done with plasticine around the cutting. Transparent plastic bags were placed over each plant to decrease transpiration and let root hairs develop. The plants were placed in a climate chamber under a rack of 36 W fluorescent lights. The lights were elevated 65 cm from ground level, and the distance between each light was 20 cm. The temperature and humidity in the chamber VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2135
was kept constant at 24 ( 1 °C and 60 ( 5%, respectively; the light was on 24 h per day throughout the experiment. After 3 days the plastic bags were removed from the plants, and they were left for 3 more days to adapt to the new environment. Afterward, the transpiration was determined by daily weighing of the flasks. When the water content in the Erlenmeyer flasks was below 30 mL (20% of the initial supplied amount), the transpiration started to decrease, although no visible signs of suffering were observed. Consequently, cyanide solution was added to the Erlenmeyer flasks when the water content fell to around 30 mL. A solution (100 mL) containing CN at 0, 10, 20, 30, 40 and 50 mg/L was added to the flasks with four replicates for each concentration. The concentration range was chosen because earlier experiments had shown that willow trees irrigated with CN at 50 mg/L died within a few days, though trees irrigated with 20 mg/L survived more than 17 days (8). Controls without trees were included to account for biological, physical and chemical processes such as degradation, volatilization, complexation and sorption. An additional set of experiments where the leaves or the roots were removed from the trees was performed to determine the influence of diffusive and advective uptake of CN into the trees. The trees used in this set of experiments were cut in August and placed directly in water without treatment with indolebutyric and nicotinic acid. Besides, the experiments were performed as described above. From 5 trees, the leaves were cut off. Another 5 trees were removed from the flasks, the roots were cut off, and the trees were replaced in new flasks just before addition of the solution; 5 replicates with complete trees were also included. A solution (100 mL) containing CN at 20 mg/L was added to the flasks. Also 10 controls without trees, 5 with sand and 5 with autoclaved sand, were included. The transpiration of the plants was determined by weighing the flask systems every 24 h. Inhibition of transpiration is a rapid measure for the toxic effect of a chemical or a substrate to willow trees (11). This effect usually occurs at the latest after a few days, followed by other effects, such as necrosis and growth reduction, general decay and even by the death of the plant. To take into consideration that healthy trees grow quickly and thereby transpiration increases, the normalized relative transpiration (NRT) was calculated. The NRT is the change of transpiration of treated trees divided by the change of control trees
∑ ∑
1 ‚ n NRT(C,t) (%) ) 1 ‚ m
n i)1
Ti(C,t)/Ti(C,0)
m j)1
× 100
Tj(0,t)/Tj(0,0)
where C is the concentration (mg/L), t is the time period (h, from 0 to 24 h, from 24 to 48 h, from 48 to 72 h and so on), T is the absolute transpiration (g/h), i is replicate 1, 2,..., n and j is control 1, 2,..., m (11). The NRT of the controls is always 100%, inhibition effects of the treatment are indicated by NRT < 100%. Four days after the addition of cyanide, the experiments were terminated, and each plant part was analyzed for cyanide. Cyanide was extracted from the sand by adding 100 mL of 1 M NaOH, and the cyanide concentration therein was determined according to the ISO 11262 standard method. Closed-Bottle Metabolism Test. Roots and leaves were cut off the stem. The stem was cut into slices not thicker than 2 mm. Roots, stem slices and leaves were dried with a paper tissue, weighed and transferred to 250 mL blue-cap bottles. Cyanide solution (200 mL of 1, 2, 5 or 10 mg/L) at pH 7 was added to the Erlenmeyer flasks containing plant material (2 g). The flasks were placed on a shaking bench at 75 rpm to 2136
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005
FIGURE 1. Normalized relative transpiration (NRT) of willows grown in sand supplied with cyanide solution at different concentrations. ensure mixing. The concentration of free cyanide in solution was measured at relevant intervals. The experiments were stopped when the concentration of cyanide approached zero, or when no further removal of cyanide was measured. The purpose of the experiments was to determine the metabolism/removal rates of cyanide for roots, stem tissue and leaves of willow cuttings (Salix viminalis). From the degradation rates the maximum degradation rates, vmax, and the Michaelis-Menten constants, KM, were determined by nonlinear fit. Modeling. All equations were implemented in an EXCEL spread sheet. The sensitivity and uncertainty analysis was performed using Crystal Ball software. Safety Measures. Of the different routes of exposure and different cyanide compounds, for humans inhalation of hydrogen cyanide results in the most rapid onset of poisoning. Inhalation of 270 ppm (0.27 mg/L air) is immediately fatal (12). When experimenting with cyanide, make sure there is very good ventilation, never work alone, and carefully consider all safety measures!
Results and Discussion Phytotoxicity Test. Concentrations of CN of 30, 40 and 50 mg CN/L in the irrigation water killed the willow plants (Figure 1). At 20 mg/L the trees survived, but the transpiration was lowered by more than 50% after 96 h. At 10 mg/L transpiration was initially lowered but recovered quickly. Cyanide Accumulation in Plants. Concentrations of cyanide in the plant parts at the end of the experiments were related, but not linearly proportional, to the concentration of the supplied cyanide solution. The highest concentrations were found in the roots. The trees that survived (0, 10, 20 mg/L) did not accumulate cyanide to high levels in any of the plant parts (Figure 3). At 40 and 50 mg/L, the trees died, and CN accumulated in the roots (significantly higher in a one-tailed t-test, R at 10% and 5%, respectively, when compared to all other concentrations) and also in the stems. Addition of 30 mg CN/L killed the trees, but no high cyanide accumulation was found. The same trend is seen for the concentration of cyanide in the solution at the end of the experiments. Remaining cyanide in the 40 and 50 mg CN/L replicates were significantly higher (one-tailed t-test, R at 10% and 5%, respectively), compared to all other concentrations. Mass Balance. In the experiments with the trees, the cyanide was lost almost completely after 4 days (Table 1). Only at the highest rates of CN treatment (40 and 50 mg/L) did some cyanide remain in the sand at 4 days. One set of controls was without trees, but closed with cork stoppers having a drilled hole which was also sealed with plasticine. In this control, the loss was 32%; in those bottles closed with gastight blue caps, the loss was reduced to 23%, while the loss in the controls with gastight blue caps
TABLE 1. Cyanide Mass Balance at the End of the Experiment (t ) 4 d)a mass of CN (µg)
Cadded (mg/L)
sol, initial
roots
stem
leaves
sol, final
total
loss (%)
0 10 20 30 40 50
0 (0) 1019 (34) 1941 (44) 2987 (23) 3977 (22) 4991 (17)
0.4 (0.3) 2.0 (3.0) 1.4 (0.5) 1.8 (0.8) 23.6 (10.6) 35.2 (21.9)
0.5 (0.4) 1.3 (1.2) 0.9 (1.8) 1.6 (2.0) 40.5 (-) 23.7 (30.3)
0.3 (0.3) 1.2 (0.9) 0.8 (0.6) 1.5 (0.3) 1.6 (1.8) 1.1 (0.6)
30.8 (10.5) 35.8 (6.0) 52.6 (14.2) 50.6 (33.2) 300.7 (200.2) 483.4 (173.4)
32.0 40.3 55.7 55.5 366.4 543.4
96 97.1 98.1 90.8 89.1
a In brackets: standard deviation. Sol, initial and Sol, final are the initial and final mass of cyanide in the irrigation water; total is the total mass of cyanide recovered from plant and solution.
TABLE 2. Cyanide Mass Balance at the End of the Experiment (t ) 4 d)a mass of CN (µg)
-roots -leaves whole plants
sol, initial
roots
stem
leaves
sol, final
total
loss (%)
1976 (64) 1982 (68) 1982 (42)
9.9 (6.0) 4.1 (1.1) 5.1 (3.8)
0.7 (0.6)
2.9 (2.1) 4.0 (2.3)
355.7 (35.6) 224.2 (116.3) 281.3 (146.0)
366.0 231.2 291.4
81.5 88.3 85.3
1.0 (1.1)
a
In brackets: standard deviation. Sol, initial and Sol, final are the initial and final mass of cyanide in the irrigation water; total is the total mass of cyanide recovered from plant and solution.
FIGURE 2. Cyanide concentrations at the end of the experiment (t ) 4 d). Error bars denote 95% confidence interval (C.I.).
FIGURE 3. Normalized relative transpiration of willow grown in sand supplied with cyanide solution at different concentrations; -roots and -leaves refer to the experiments with detached roots and leaves, respectively. and sterilized sand was 9%. This shows the loss due to volatilization, sorption, biological degradation and handling. Determination of Diffusive and Advective Uptake. In the second set of experiments the transpiration was lowered very fast when the roots or leaves were detached and a cyanide solution (20 mg/L) was added (Figure 3). Also the transpiration of the experiments with complete plants was lowered more than found in the first experiments. The difference in the concentrations in roots (Figure 4) between the experiments with detached leaves and whole
FIGURE 4. Cyanide concentrations at the end of the experiment (t ) 4 d). Error bars denote 95% C.I.; -roots and -leaves refer to the experiments with detached roots and leaves, respectively. plants could be caused by the higher advective uptake in the whole plants; however, the difference was not significant in a one-tailed t-test (at R ) 5%). A significant difference was found in the final amount of cyanide in the solution; the concentration remaining in the experiments with detached roots was significantly higher (at R ) 5%) compared to the experiments with detached leaves and whole plants. This could be explained by a slower uptake of cyanide into willows when the roots are removed. Much higher amounts of cyanide were found in roots and stem of trees in the second set of experiments (Table 2), when compared to the first set of experiments. Also the total loss of cyanide was smaller. Possibly this was due to the smaller tree mass. For this second set of experiments, the cuttings were taken in August after they came into leaf. They grew slower and transpired less water, and the average leaf and root mass was 70-80% smaller. Though the transpiration in the experiments without leaves was much lower than in the experiments with whole plants, the average removal was higher. Thus, the diffusive uptake of cyanide into the roots seemed to play an important role. Closed-Bottle Metabolism Test. All parts of the willows (roots, stem and leaves) were able to remove cyanide from the solution, with leaves showing the fastest removal and stem the slowest (Figure 5a-d). The rates varied with dosage. VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2137
FIGURE 5. Decrease of solution concentration C (mg/L) in the closed-bottle metabolism test with initial CN concentrations of (a) 1 mg/L, (b) 2 mg/L, (c) 5 mg/L and (d) 10 mg/L. Error bars denote the 95% C.I. the transpiration stream, there is also a loss upward, and the mass balance of the roots is
TABLE 3. Michaelis-Menten Parameters Obtained by Nonlinear Regression roots stem leaves a
vmax (mg/kg/h)
KM (mg/L)
R2
n
6.9 ( 1.7 1.6 ( 0.7 9.6 ( 1.1
0.44 ( 0.33 0.01 ( 0.51 0.59 ( 0.20
0.29 1.5 × 10-5 0.73
55 23 84
change of chemical mass in roots mR (mg) with time t (d) ) inflow - outflow dmR ) CW × Q - CXy × Q dt
Mean of 4 to 5 replicates; ( 95% C.I.
The removal half-time was lower at low doses, although the absolute removal was higher at high doses. Kinetics. The enzyme that metabolizes cyanide in vascular plants, β-cyanoalanine synthase, is well described (13). The final metabolite is asparagine (14). Recently, Ebbs et al. (15) confirmed that free cyanide is taken up and incorporated by willows. Enzymatic reactions can be described by MichaelisMenten kinetics (16)
v)
vmax C ×M KM + C
where v (mg/h) is the removal rate of the substrate concentration C (mg/L), vmax (mg/kg/h) is the maximal removal velocity, KM (mg/L) is the half-saturation constant and M (kg) is the mass of the plant compartment. KM and vmax were determined by nonlinear regression of the removal rate to the cyanide concentration (Table 3). The maximal removal velocity, vmax, was highest for leaves and lowest for stems. The fit was uncertain for stems (R2 insignificant); this affected, however, only KM: vmax was mainly determined by the few measurements for the highest dosage. The MichaelisMenten parameters are in agreement with the KM of 0.6 and 0.7 mg/L and vmax of 9.4 and 10.7 mg/kg/h for leaves and roots, respectively, determined by nonlinear regression by Larsen et al. (8). Model for Balancing Uptake and Metabolism. The uptake of a chemical into roots may occur via advection in the transpiration stream and by diffusion. If the uptake is with 2138
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005
where CW and CXy (mg/L) are the concentrations of the chemical in the external solution and in the xylem sap, and Q (L/d) is the flow of water in the xylem (transpiration stream). If the chemical in the translocation stream is in equilibrium with the root tissue, the mass balance is
dmR Q ) CW × Q - C R × dt KRW where KRW (L/kg) is the partition coefficient between root tissue and transpiration water (17). The uptake by diffusion may be described with Fick’s 1st Law and is then
(
CR dmR ) A × P × CW dt KRW
)
where A is the surface area (dm2) and P is the permeability of the roots (dm/d). Substituting chemical mass mR in the roots by concentration CR (mg/kg) ) mR/MR, where MR (kg) is the mass of roots, gives the differential equation for the chemical concentration in roots with both processes
(
)
(
)
dCR CR Q+A×P Q+A×P ) × CW × dt MR MR KRW - k R × CR where kR is the growth rate of the roots (d-1).
To consider enzymatic metabolism, the term for Michaelis-Menten kinetics is added
v)
vmax C KM * + C
× MR
KM,St* + CSt
where v (mg/d) is the removal rate per plant mass (mg/kg/d) and KM* (mg/kg) is the half-saturation constant related to the concentration inside the plant. It is derived by multiplying the experimental KM related to solution concentration (mg/ L) with the partition constant KRW (L/kg). The complete mass balance is then
CR dmR ) (Q + A × P) × CW - (Q + A × P) × dt KRW vmax × CR KM* + CR
)
(
a ) -kSt -
b)
)
KM * + C R
CR × KM,St* × Q MSt × KRW
In the mass-balance for leaves, there is exchange with air but no translocation upward.
change of mass in leaves ) translocation from stem + uptake from air - degradation - loss to air vmax,L × CL dCL g×A Q × CSt + × CAir ) dt ML × KStW ML K *+C M,L
kL × CL -
At steady-state (dCR/dt ) 0), this yields
KM* × (Q + A × P) × CW (Q + A × P) × CW + × CR MR MR KM* × (Q + A × P) Q+A×P × CR × CR2 - KM* × MR × KRW MR × KRW kR × CR - kR × CR2 - vmax × CR ) 0
aC2 + bC + c ) 0
a ) -kL b)
with the two real solutions
-b - xb - 4ac 2a
C2 )
-b + xb - 4ac 2a
(
Q × CSt CAir × g × A - KM,L* × kL - vmax,L + ML × KStW ML KM,L* × g × A × F KLA × ML c)
2
where
b)
g×A×F KLA × ML
2
a ) -kR -
)
Q+A×P MR × KRW
(
)
Q+A×P Q+A×P × CW - KM* + kR - vmax MR MR × KRW c ) K M* ×
(
L
g×A×F × CL KLA × ML
where L is the index for leaves, CAir is the concentration of the chemical in air (mg/m3), g is the conductance (m/d), A is the leaf area (m2), KLA is the dimensionless partition coefficient leaves to air (mg/m3 leaves to mg/m3 air), and F is the density of the leaves (kg/m3). The parameters of the resulting quadratic equation are
which leads to a quadratic equation of the general form
C1 )
Q MSt × KStW
CR × Q KM,St* × Q - kSt × KM,St* - vmax,St MSt × KRW MSt × KStW
× MR
dCR Q+A×P Q+A×P × CW + kR × CR ) dt MR MR × KRW vmax × CR
- kSt × CSt
resulting in the same quadratic equation with the coefficients
c)
or
(
dCSt Q Q × CR × CSt ) dt MSt × KRW MSt × KStW vmax ,St × CSt
)
Q+A×P × CW MR
Only the first solution, C1, gives realistic (i.e. positive) values. The same mass balance can be written for the stem. We neglect diffusion in/out of the stem, because we currently cannot quantify the processes. Then
KM,L* × Q × CSt KM,L* × CAir × g × A + ML × KStW ML
By this, the solutions for the three main parts of a plant, namely roots, stem and leaves have been defined. For the transient case, the differential equations were solved numerically. Parameterization of the Model. The parameters were determined from the uptake and toxicity experiments. The values of KM and vmax were taken from Table 3. Several parameters were variable and changed from experiment to experiment, e.g., plant mass. Therefore, a set of default data, which represents typical values, was defined (Table 4). Hydrogen cyanide is very hydrophilic (log KOW ) - 0.25 (18)), and it was assumed that the dominant resistance for the uptake into root cells is on the side of the biomembranes. The permeability P of the root membranes for HCN was estimated by (19)
log P ) log KOW - 6.7
change of mass in stem ) translocation from roots degradation - translocation to leaves
P ) 1.12 × 10-7 m/s
This gives the differential equation for the chemical concentration in stem (CSt)
Example Model Simulations. The concentrations of free cyanide in roots for increasing cyanide concentration in VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2139
TABLE 4. Model Parameters; Default Values Represent Typical Average Values parameter
default value
log KOW -0.25 KAW 0.0015 root mass MR 0.002 stem mass MSt 0.018 leaf mass ML 0.0035 transpiration Q 0.008 root area A 0.001 partition coeff root 0.87 to water KRW partition coeff stem 0.56 to water KStW partition coeff leaves 0.54 to air KLA/F leaf-air conductance g 8.64 leaf area A 110 concentration in the 0 to 50 flask CW growth rate k 0.01 CAir 0
source ref 18 ref 20 measureda measureda measureda measureda estimated estimatedb
unit m3/m3 m3/m3 kg kg kg L/d m2 L/kg
estimatedc L/kg estimatedd mg/kg: mg/m3 ref 21 m/d measureda cm2/g calculatede mg/L estimatedf 1/d constant mg/m3
a Experimental values change from tree to tree b K RW ) 0.85 + 0.02 × 1.22 × KOW0.77 (21). c KStW ) 0.316 + 0.684 × 10(-0.28+0.668 logKow) water and wood (22). d KLA ) (0.8 + 0.02 × 1.22 × KOW0.95)/KAW (21); F ) 1000 kg/m3. e Determined from the mixing of the nutrient solution in the flask and the added cyanide solution. f No measurable growth was found during the 4-days experiments.
FIGURE 6. Simulation of the cyanide concentration inside the root, CR (mg/kg), for varying CN concentrations in the external solution, CW (mg/L); default data Table 4 were used; measured values taken from experiment 1; error bars denote 95% C. I. external solution (CW) were calculated (Figure 6) using the default values in Table 4. For CW below 10 mg/L, metabolism in roots was as fast as the uptake and neither accumulation in roots nor translocation to stem was predicted. Thus, at these low levels, we would not expect any toxic effect of cyanide, because none or very little would be present in the tissue. At higher doses, vmax in roots is reached, and though the uptake still increases, the metabolism rate remains constant. From this point on, approximately 20 mg/L, the concentration in roots increases linear with the concentration in solution. Measured concentrations from the first set of experiments are given for comparison and show a similar trend. Sensitivity Analysis. To find the sensitive input parameters of the model, all plant input parameters were varied, and 10 000 Monte Carlo runs were done; the normal distribution was chosen for all parameters, with a standard deviation at 10% of the mean (the default values in Table 4). The probability density of the simulated concentration in roots for a constant concentration in external solution of 20 mg of CN/L ranged from 0.90 to > 9.0 mg/L (Figure 7). The highest probability density was around 4.5 mg of CN/L. The high variance in the calculated root concentration was explained by four parameters: root mass MR (40.2%), vmax of roots (39.7%), root surface area A (11.6%) and transpiration Q 2140
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005
FIGURE 7. Probability density of the calculated concentration in root, CR (mg/kg); Monte Carlo simulation, 10 000 runs, variation of plant input parameters with a normally distributed random variable with standard deviation ) 10% of the mean; concentration in external solution was set to 20 mg/L. (8.3%). The parameters growth rate k and KM of the roots had practically no influence on the cyanide concentration in roots; parameters of stem and leaves had no influence, too. The measured concentrations in roots for a nominal concentration of 20 mg/L in the two sets of experiments were therefore at the lower and higher edge of the Monte Carlo simulation (0.72 mg/kg for the first, Figure 2, and 10.22 mg/kg for the second set of experiments, Figure 4). Since root mass has the highest sensitivity on root concentrations, the difference in the two measured results can probably be explained by the difference in root mass (1.73 ( 0.84 mg and 0.41 ( 0.38 mg, respectively, where ( denotes the 95% C.I.) between the first and second set of experiments. Another outcome of the sensitivity analysis is that diffusive uptake (directly proportional to the surface area A) has a similar but higher influence than the advective uptake (directly proportional to the transpiration Q). However, in case of reduced advective uptake, the model would predict increasing diffusive uptake, because the concentration gradient between outside and inside the root would be higher. This model result is supported by the results from the experiment with/without leaves: Without leaves, transpiration was negligible, but concentrations in roots were only 1/4th lower than in intact plants (Figure 4); furthermore, this reduction was not significant. Simulation of Specific Experiments. Simulations were also carried out with parameter sets measured for single experiments, so that the simulations could be compared to the measurements. The transpiration, Q, was varied with time by inserting the measured values, while the concentration in external solution, CW, was varied with time by considering the calculated uptake by plants and the measured loss in controls without trees (32%). For cyanide applied at 10, 20 and 30 mg/L, the simulated concentrations of cyanide inside the roots were in good agreement with the experimental values. For 40 and 50 mg/ L, the modeled values of CR were much lower than the experimentally determined concentrations, though the maximal simulated concentration over the 4 days was very close to the experimental (Table 5). A likely reason is that when the plants suffer due to toxic effects, the metabolic capacity is decreased. The plot of the simulated concentration in roots, CR, for the whole plants in the second set of experiments (20*) shows a different pattern. This can be explained by the lower root mass and transpiration. Thus, the diffusive and advective uptake as well as the metabolism were lower. For this experiment, the maximal simulated concentration was close to the experimentally determined concentration. Large deviations between simulated and measured cyanide concentrations were found for stem and leaves. This might have several reasons. First, there is always a natural background of CN in willow tissue, as indicated by the cyanide found in untreated controls (C ) 0 mg/L). Second, stems
TABLE 5. Comparison of Experimental and Simulated Concentrations (mg/kg) of Cyanide in the Plants at 4 Days experimental
modeled, final
modeled, max
C (mg/L)
Croot
Cstem
Cleaves
Croot
Cstem
Cleaves
Croot
Cstem
Cleaves
0 10 20 20* 30 40 50
0.24 0.23 0.72 10.22 1.09 15.00 17.03
0.04 0.06 0.07 0.29 0.12 2.26 1.87
0.12 0.26 0.24 1.31 0.51 0.63 0.62
0.20 1.08 6.44 2.04 6.18 6.02
3.9 × 10-4 1.3 × 10-4 5.6 × 10-4 1.5 × 10-4 3.8 × 10-4 4.0 × 10-4
4.1 × 10-7 8.5 × 10-7 1.6 × 10-5 7.8 × 10-7 1.6 × 10-6 1.8 × 10-6
0.36 3.12 8.94 7.65 14.58 16.58
5.0 × 10-5 6.3 × 10-4 1.4 × 10-3 1.8 × 10-3 3.4 × 10-3 4.0 × 10-3
4.2 × 10-7 6.8 × 10-6 6.7 × 10-5 2.7 × 10-5 4.5 × 10-5 5.1 × 10-5
FIGURE 8. Simulation of cyanide concentration in roots with different concentrations of cyanide in the irrigation water. The lines are dotted for the trees that died; * denotes the second set of experiments.
FIGURE 9. Simulation of concentrations of cyanide in the stem using (a) measured (Table 3) and (b) fitted Michaelis-Menten parameters. were partly in direct contact with the cyanide solution (inside the sand) and partly exposed to cyanide vapor (inside the flasks). This might have caused an additional uptake, which could not be taken into consideration in the model. Third, the experimentally determined Michaelis-Menten parameters for the stem were in particular uncertain (Table 3). By variation of vmax of the stem, the model could be fitted to the measured data; by lowering vmax of the stem from 1.6 to 0.24 mg/kg plant/h, the maximal modeled concentration (Figure 9) met the experimentally determined values (Table 5). Interpretation of the Results. Cyanide may be a special case: its rapid mode of action causes its toxic effects to become almost immediately obvious, and it does not accumulate in healthy plants. But the methods and the model developed and tested in this study should also be valid for other plant-chemical combinations.
A clear relation between uptake, metabolism, accumulation and toxicity of cyanide in willows could be found. This relation was nonlinear. The nonlinearity stemmed from the metabolism kinetics of the plant tissue, which could be described by Michaelis-Menten kinetics. The MichaelisMenten parameters were determined in a closed-bottle metabolism test with cutoff, but living plant tissue. The experiments were relatively easy to perform, and the results could be used as realistic input parameters for the massbalance model of cyanide in willows. A satisfying agreement between measured and predicted cyanide accumulation would not have been achieved without this nonlinear metabolism kinetics. However, recent plant-uptake models either neglect metabolism by plants or use zero- or firstorder rates (21-25). The toxicity in this study was determined by weighing the whole system (flask, sand, solution and plant). A reduced weight loss was contributed to an inhibited transpiration. The toxicity measured by this way showed a relatively good correlation with the accumulation of cyanide inside the willows. Since this method of determining toxicity is nondestructive, mass balance and toxicity of cyanide could be determined in the same experiment. The validation of the model provided difficulties: Due to the nonlinearity of the processes, some parameters (in particular root mass and maximum metabolic rate) were very sensitive for the final concentration in roots and for the chemicals toxicity, both in the experiments and in the model simulations. A conclusion is that uptake and metabolism tests of nonpersistent chemicals in soil-plant systems need to be done at varying exposure levels and eventually with plants of varying size. Generally, if chemicals are readily taken up by plants but not eliminated at the same rate, residues may accumulate. This may lead to toxic effects on the plants. If chemicals are accumulated but the plants survive, this may lead to contamination of consumers such as humans. A full-scale phytoremediation project with willow trees growing in cyanide-polluted soil has been implemented in Denmark, but the outcome of the project has not yet been evaluated (26). In phytoremediation, the goal is the complete removal of toxins. If the removal is primarily by plants, both a high metabolic capacity and a high biomass can increase it. However, due to phytotoxic effects of chemicals, rapid metabolism and survival of the trees may only be guaranteed below a certain threshold.
Acknowledgments This work was supported by DTU PhD grants for Morten Larsen and Ahmed S. Ucisik, a RECETO grant for Ahmed, and by the European Commission, project BIOTOOL, contract no. 003998 (GOCE). We thank in particular our cyanide team, Helle Christiansen, Ines Koch, Worasiri Jaiplord Pedersen, Jens Schaarup Sørensen, Alessandro Pirandello, Louise Andresen and Joseph Danquah-Boakye. The authors thank also Jens C. Tjell, Hans Mosbæk, K. Ole Kusk, Ulrich VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2141
Karlson and Tim Mansfeldt for discussion and support. Finally we thank Torben Dolin for preparing the graphs.
Literature Cited (1) Burken, J. Uptake and metabolism of organic compounds. In Phytoremediation - Transformation and Control of Contaminants; McCutcheon, S. C., Schnoor, J. L., Eds.; John Wiley & Sons: Hoboken, New Jersey, USA, 2003; pp 59-84. (2) Komossa, D.; Langebartels, C.; Sandermann Jr., H. Metabolic processes for organic chemicals in plants. In Plant Contamination - Modeling and Simulation of Organic Chemical Processes; Trapp, S., McFarlane, C., Eds.; Lewis: Boca Raton, Florida, USA, 1995; pp 69-103. (3) Schwitzgue´bel, J. P.; Vanek, T. Some fundamental advances for xenobiotic chemicals. In Phytoremediation - Transformation and Control of Contaminants; McCutcheon, S. C., Schnoor, J. L., Eds.; John Wiley & Sons: Hoboken, New Jersey, USA, 2003; pp 123-157. (4) Medina, V. F.; Maestri, E.; Marmiroli, M.; Dietz, A. C.; McCutcheon, S. Plant tolerances to contaminants. In Phytoremediation - Transformation and Control of Contaminants; McCutcheon, S. C., Schnoor, J. L., Eds.; John Wiley & Sons: Hoboken, New Jersey, USA, 2003; pp 189-232. (5) Harms, H.; Bokern, M.; Kolb, M.; Bock, C. Transformation of organic contaminants by different plant systems. In Phytoremediation - Transformation and Control of Contaminants; McCutcheon, S. C., Schnoor, J. L., Eds.; John Wiley & Sons: Hoboken, New Jersey, USA, 2003; pp 285-316. (6) Jacobson, M. E.; Chiang, S. Y.; Gueriguian, L.; Westholm, L. R.; Pierson, J.; Zhu, G.; Saunders: F. M. Transformation kinetics of trinitrotoluene conversion in aquatic plants. In Phytoremediation - Transformation and Control of Contaminants; McCutcheon, S. C., Schnoor, J. L., Eds.; John Wiley & Sons: Hoboken, New Jersey, USA, 2003; pp 409-427. (7) Castro, S.; Davis, L. C.; Erickson, L. E. Phytotransformation of benzotriazoles. Int. J. Phytoremediation 2003, 5 (3), 245265. (8) Larsen, M.; Trapp, S.; Pirandello, A. Removal of cyanide by woody plants. Chemosphere 2004, 54, 325-333. (9) ISO 11262:2003. Soil quality - determination of cyanide; International Organisation for Standardisation, Geneva, Switzerland. (10) ISO 14403:2002. Water quality - Determination of total cyanide and free cyanide by continuous flow analysis; International Organisation for Standardisation, Geneva, Switzerland. (11) Trapp, S.; Zambrano, K. C.; Kusk, K. O.; Karlson, U. A phytotoxicity test using transpiration of willows. Arch. Environ. Contam. Toxicol. 2000, 39, 154-160.
2142
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005
(12) Faust, R. A. 1994. Toxicity Summary for Cyanide. Risk Assessment Information System. July 2004 at http://risk.lsd.ornl.gov/tox/ rap_toxp.shtml (13) Hendrickson, H. R.; Conn, E. E. Cyanide metabolism in higher plants. IV. Purification and properties of the β-cyanoalanine synthase of blue lupine. J. Biol. Chem. 1969, 244, 2632-2640. (14) Manning, K. Detoxification of cyanide by plants and hormone action. In: Cyanide compounds in biology; Evered, D., Harnett, S., Eds.; John Wiley & Sons: Chichester, UK, 1988; pp 92-110. (15) Ebbs, S.; Bushey, J.; Poston, S.; Kosma, D.; Samiotakis, M.; Dzombak, D. Transport and metabolism of free cyanide and iron cyanide complexes by willow. Plant Cell Environ. 2003, 26, 1467-1478. (16) Cornish-Bowden, A. Fundamentals of enzyme kinetics; Portland Press: London, UK, 1995. (17) Trapp, S. Dynamic root uptake model for neutral lipophilic organics. Environ. Toxicol. Chem. 2002, 21, 203-206. (18) Kroschwitz, J. I.; Howe-Grant, M.; Kirk, R. E.; Othmer, D. F. Encyclopedia of Chemical Technology. - 7: Composite Materials to Detergency; John Wiley & Sons: New York, USA, 1993. (19) Trapp, S. Plant Uptake and Transport Models for Neutral and Ionic Chemicals. Environ. Sci. Poll. Res. 2004, 11, 33-39. (20) Kjeldsen, P. Behaviour of cyanides in soil and groundwater: a review. Water, Air, Soil Pollut. 1999, 115, 279-307. (21) Trapp, S.; Matthies, M. Generic one-compartment model for uptake of organic chemicals by foliar vegetation. Environ. Sci. Technol. 1995, 29, 2333-2338; erratum 30, 360. (22) Trapp, S.; Miglioranza, K. S. B.; Mosbæk, H. Sorption of lipophilic organic compounds to wood and implications for their environmental fate. Environ. Sci. Technol. 2001, 35, 1561-1566. (23) Trapp, S.; McFarlane, J. C.; Matthies, M. Model for uptake of xenobiotics into plants - validation with bromacil experiments. Environ. Toxicol. Chem. 1994, 13, 413-422. (24) Paterson, S.; Mackay, D.; McFarlane, J. C. A model of organic chemical uptake by plants from soil and the atmosphere. Environ. Sci. Technol. 1994, 28, 2259-2266. (25) Satchivi, N. M.; Stoller, E. W.; Wax, L. M.; Briskin, D. P. A nonlinear dynamic simulation model for xenobiotic transport and whole plant allocation following foliar application I. Conceptual foundation for model development. Pestic. Biochem. Phys. 2000, 68, 67-84. (26) Trapp, S.; Christiansen, H. Phytoremediation of cyanide-polluted soils. In Phytoremediation - Transformation and Control of Contaminants; McCutcheon, S. C., Schnoor, J. L., Eds.; John Wiley & Sons: Hoboken, New Jersey, USA, 2003; pp 829-862.
Received for review August 2, 2004. Revised manuscript received January 4, 2005. Accepted January 12, 2005. ES048799S
