Phytoextraction by high-Cd-accumulating rice to reduce Cd in wheat grains grown in Cd-polluted fields

Abstract

Phytoextraction using indica rice plants (Indian Rice Oryza ssp.) is a promising technique for remediating cadmium (Cd)-polluted paddy fields. Because this technique has only been established for paddy fields, we decided to examine phytoextraction in upland fields that have been converted from paddy fields. Although “CHOKOUKOKU” shows a shattering habit and lodging, its Cd uptake was significantly higher than that of other indica rice varieties. On the other hand, “IR8”, which was able to accumulate only moderate levels of Cd in its shoots, has a lodging tolerance, making it an optimal variety for southwest Japan, which experiences several typhoons each year. Therefore, both “CHOKOUKOKU” and “IR8” were useful in estimating practical phytoextraction in upland fields. A practical phytoextraction examination showed that the total Cd uptake of “CHOKOUKOKU” and “IR8” was 822 and 545 g ha⁻¹, respectively, after a 4-year period. After phytoextraction by planting high Cd-accumulating rice plants, the Cd concentration of the plot soil decreased by approximately 35%, compared to the initial Cd concentration. To evaluate phytoextraction efficiency in the upland field, wheat (Triticum aestivum) was subsequently grown in the remediation field. The Cd concentration in the grains of “SHIROGANEKOMUGI” grown in the phytoextraction plot was lower than that grown in a non-phytoextraction plot; regrettably, it exceeded the Codex Alimentarius Commission standard, whereas the grain Cd concentrations of “CHUGOKU165” planted in phytoextraction plots complied with the Codex standard for wheat grain. These results suggest that phytoextraction using high Cd-accumulating rice varieties is a practical remediation system for low Cd-polluted upland fields. Moreover, we determined the end point of the phytoextraction process, which occurs when the soil Cd concentration of the phytoextraction fields is less than 0.6 mg kg⁻¹ if “CHUGOKU165” is cultivated in this area.

Key words:

• phytoextraction
• cadmium
• upland field
• indica rice variety
• wheat

INTRODUCTION

Agricultural soils in some regions of Japan have been polluted with cadmium (Cd) because of industrialization during the 1960s (Makino et al. 2012). Cd is one of the most hazardous environmental toxicants of human health; its long-term intake from contaminated food or water is known to cause the Itai-itai disease. The 28th (2005) and the 29th (2006) session of the Codex Alimentarius Commission adopted the maximum permissible concentration of Cd in staple crops, allowing for less than 0.2 mg kg⁻¹ in wheat (Triticum aestivum) grain, and less than 0.4 mg kg⁻¹ in polished rice (Oryza sativa ssp.), respectively. In response to this, the Japanese government revised the Food Sanitation Act and set a new maximum Cd level for rice grains, equal to that of the Codex standard. To comply with this standard, water management based on flooding in the later growth stage was efficient for brown rice (Inahara et al. 2007). The Cd divalent cation forms insoluble cadmium sulfide in strongly reduced conditions, resulting in a low availability of Cd in paddy rice. However, there is a trade-off relationship between Cd and arsenic (As), which is another environmental toxicant of human health (Honma et al. 2012). Indeed, flooding for 3 weeks before and after heading was effective in reducing Cd concentrations in rice grains, whereas As concentrations were markedly increased (Arao et al. 2009). Therefore, to produce safe rice grains, it is necessary to decrease the soil Cd concentration.

Top-dressing of non-polluted soil to Cd-contaminated fields has been the most popular remediation technique in Japan (Ono 2003). In addition, an on-site soil washing technology using ferric iron chloride has been developed and demonstrated (Makino et al. 2007). Although these techniques are effective for high or moderate Cd-contaminated fields, they are not practical for low Cd-contaminated fields because of their high cost (USD$ 300,000 to 500,000 ha⁻¹; Murakami et al. 2007). Thus, alternative low-cost remediation methods are needed for low Cd-contaminated fields.

Phytoextraction has been proposed as a cost-effective and environmentally friendly technique to remove heavy metals from fields using plants (Ebbs et al. 1997). The estimated cost of phytoextraction using four crops was approximately USD$ 120,000 ha⁻¹ (Azuma et al. 2012); it was lower than that of soil dressing. For successful phytoextraction, it is important to select a phytoextraction plant with high Cd-accumulating ability (Ishikawa et al. 2006). Some indica rice varieties can accumulate relatively high Cd concentrations in their shoots (Arao and Ae, 2003). As the cultivation practices of rice are well established, rice is convenient as a phytoextractor plant for Japanese paddy fields. Indica rice varieties “IR8” (Honma et al. 2009; Azuma et al. 2012) and “CHOKOUKOKU” (Murakami et al. 2009) were reported to be useful phytoextractor varieties for Cd-contaminated paddy fields. In a previous study (2009), we found that phytoextraction using “IR8” with early drainage of irrigation water was a practical phytoextraction system. Currently, indica rice is grown in some Cd-contaminated Japanese paddy fields as phytoextractor plants. On the other hand, the area of the Japanese paddy field and upland field is 2.47 and 1.17 M ha, respectively (Statistics Bureau, Ministry of Internal Affairs and Communications 2012). Therefore, phytoextraction in the upland field is required. It is well known that Thlaspi caerulescens (Hammer and Keller 2003) can absorb a large amount of Cd in upland conditions. Kurihara et al. (2005) reported that a significant amount of Cd was taken up from Cd-polluted paddy fields converted to upland fields through phytoextraction using kenaf (Hibiscus cannabinus). Regrettably, a cultivation system of these weeds has not been established, making it difficult to evaluate weeds for practical phytoextraction. Therefore, it is important to select a crop plant for phytoextraction, such as the indica rice plant. However, studies evaluating the abilities of crop plants in upland fields are limited.

Assessment of phytoextraction efficiency has to be based on specific goals. Because Japan has no reference values for Cd in the soil, most studies were unable to directly show the obvious objective of phytoextraction in contaminated fields. Nishiyama et al. (2005) suggested that a 50% decrease in the total soil Cd concentration should be met in order to confirm that phytoextraction is indeed successful. However, no scientific evidence has been identified for this target. To ensure decontamination of Cd in paddy fields, Azuma et al. (2012) planted rice grains in some parts of the phytoextraction field. By detecting a decrease in the Cd concentration of rice grains planted in the previous year, they then judged the success of phytoextraction. However, soil Cd concentration as a goal of phytoextraction has not been declared.

The objective of this study was therefore to select a high Cd-accumulating rice variety for upland fields and to establish a practical phytoextraction system. Furthermore, we investigated the effect of phytoextraction on Cd concentration in subsequently cultured wheat, and also attempted to show the objective goal of phytoextraction in upland fields in this area.

MATERIALS AND METHODS

Field experiment

Cd-polluted upland fields converted from paddy fields (A, B) located in southwest Japan were used in this study. Field A was used to determine the most promising rice variety for phytoextraction in upland fields and field B was used to evaluate the practical method of phytoextraction. These were fields that had been converted to upland at least 10 years ago. The source of Cd pollution in the field was fallout dust from nearby zinc-refining plants. The soil in these fields is generally classified as grey lowland and the texture as sandy clay loam. The 0.1 M hydrochloric acid (HCl) extractable soil Cd concentration of field A and field B was approximately 1.7 and 1.0 mg kg⁻¹, respectively. Initial soil layer and bulk density of field B were 0.17 m and 1.3 kg L⁻¹, respectively. Other soil characteristics are shown in Table 1.

Table 1 Soil properties of fields A and B

Field	pH	Cd concentration (0.1 M HCl extractable) mg kg^−1	T-N (g kg^−1)	T-C (g kg^−1)	Clay (g kg^−1)
A	5.43	1.70	1.10	15.2	225
B	5.30	0.98	1.06	11.0	203

Cd, cadmium; HCl, hydrochloric acid: T-N, total nitrogen; T-C, total carbon.

Selecting promising rice varieties for phytoextraction in the upland field that was converted from paddy field

We used “IR8”, “CHOKOUKOKU”, “MORETSU” and “MINAMIYUTAKA” indica rice varieties, which have been previously reported to accumulate high amounts of Cd (Azuma et al. 2003; Ibaraki et al. 2009; Murakami et al. 2009). The experimental plots (5 × 5 m) were located in field A. Four plots were prepared for each variety (four replicates). A basal slow-release fertilizer was applied at the rates of 110, 77, and 83 kg ha⁻¹ for nitrogen (N), phosphorus (P) and potassium (K), respectively (173 and 100 kg ha⁻¹ for diphosphorus pentoxide (P₂O₅) and potassium oxide (K₂O), respectively). Rice seeds were sown at a row spacing of 0.25 m on June 2, 2006. Rice was cultured under rain-fed conditions throughout the growth period. Weeds were fundamentally controlled with herbicide. When weeds could not be controlled, weeds were picked by human hands. Rice shoots were harvested on October 11, 2006, from 2 m of rows per plot and mixed. Soils were also collected from each point.

Phytoextraction using high Cd-accumulating rice varieties in upland fields

Two rice varieties, “IR8” and “CHOKOUKOKU”, were used in this experiment. Because it had been earlier established that rice plants accumulate injuries when continuously cultured in upland conditions (Yahiro and Tanaka, 1989), we located a crop rotation plot. This plot was alternately cultured with “CHOKOUKOKU” and sorghum (Sorghum bicolor (L.)). Sorghum has the capability to accumulate moderate levels of Cd in its shoots (Ito et al. 2003). The experimental plots (3.3 × 3.3 m) were located in field B. Three plots were prepared for each variety (three replicates). The phytoextraction plants were cultured for 4 years (2008 to 2011). The same basal fertilizer as previously described was applied. A basal slow-release fertilizer for “IR8” was applied at the rates of 220, 154 and 166 kg ha⁻¹ for N, P and K, respectively (346 and 200 kg ha⁻¹ for P₂O₅ and K₂O, respectively). Because “CHOKOUKOKU” tends to lodge, the amount of fertilizer for this variety was half of that of “IR8”. Because the yield of sorghum is generally very high, the amount of fertilizer applied to this crop was twice that for “IR8”. The phytoextraction plant seeds were sown at a row spacing of 0.3 m in late May and harvested in mid-October each year. The plant shoots were taken from 2 m of rows per plot and mixed. Soil was also collected from each point.

Effect of phytoextraction on the Cd concentrations in wheat grains subsequently grown in the remediation field

To evaluate the effect of phytoextraction using indica rice varieties in upland fields, the wheat cultivars “SHIROGANEKOMUGI” and “CHUGOKU165” were subsequently grown in the remediated field. An experimental plot (width: 1.4 m × length: 2 m) was located in the remediated field. The soil pH of the plots was adjusted to 6.0 with magnesium oxide after phytoextraction. A basal fertilizer for wheat was applied at the rates of 60, 60 and 60 kg ha⁻¹ for N, P and K, respectively (135 and 72 kg ha⁻¹ for P₂O₅ and K₂O, respectively). The same amounts of fertilizer were applied on January 25, 2012, and one-third these amounts were applied on May 7, 2012. Wheat varieties were sown in five rows on December 6, 2011, and harvested on June 1, 2012. The plant shoots were taken from 2 m of rows per plot and mixed. Soils were also collected at the time of harvest.

Analyses of the phytoextraction plant, wheat grain and soil

The shoots of the phytoextraction plants were dried at 80°C for 72 h to measure the dry weight. These shoots were divided into ears and other parts (leaves and stems). Each part was weighed and analyzed individually. Each sample was ground to a powder. Two grams of the powdered sample was digested in 40 mL of nitric acid (HNO₃)- perchloric acid (HClO₄)- sulfuric acid (H₂SO₄) (5:1:1) mixed solution (Ibaraki et al. 2009). The Cd concentration of this digested solution was determined using a flame atomic absorption spectrophotometer (FAAS; AA-6800, Shimadzu, Kyoto, Japan). The Cd concentration of the shoots was calculated from the concentrations of the ears and the other parts, taking into account the weight ratio. The wheat grain was air-dried to 15% moisture. Five grams of wheat grain was measured into a 100-mL tall beaker. It was heated to 350°C for 2 h and to 500°C for 10 h. Ten milliliters of HNO₃ and 5 mL of hydrogen peroxide (H₂O₂) were added to the beaker and the grain ash was digested completely (Honma et al. 2009). The Cd concentration of the digested solution was determined using inductively coupled plasma-mass spectrometry (ICP-MS; 7700 ×, Agilent Technologies, California, USA). The soil samples were air-dried and passed through a 2-mm sieve. The soil Cd was extracted with 0.1 mol L⁻¹ HCl [soil: solution = 1:5 weight/volume (w/v), 30°C, 1 h] (Murakami et al. 2007). The Cd concentration of the soils was determined by FAAS. The total carbon (T-C) concentration of soil was determined by Turin’s method. Soil samples were also digested using Kjeldahl’s method, and the total nitrogen (T-N) concentration was determined by steam distillation.

RESULTS

Selecting promising rice varieties for phytoextraction in the upland field that was converted from paddy field

The rice varieties all grew favorably through the middle of October (Table 2). The maturity time of “CHOKOUKOKU” was approximately late September and that of “MORETSU” and “MINAMIYUTAKA” was mid-October. The “IR8” was immature on harvest day (October 11, 2012). The plant height of “IR8” was short, whereas that of the other varieties was tall (113–125 cm). The “CHOKOUKOKU” variety was lodging significantly, and “MORETSU” and “MINAMIYUTAKA” showed signs of lodging because of Typhoon 13, which hit the area on September 17, 2006. Because of the short height of the “IR8”, no lodging was observed. The grains of “CHOKOUKOKU” and “MORETSU” were markedly shattered, whereas the “MINAMIYUTAKA” variety, which was bred to improve this habit of “MORETSU” (Yoshioka et al. 2006), was not shattered. The dry weight of these varieties was not significantly different. The Cd concentration of the “CHOKOUKOKU” shoot was significantly higher than that of the other varieties. Accordingly, the Cd uptake of “CHOKOUKOKU” was highest among all the varieties used in this experiment (P < 0.05, Tukey’s multiple-comparison test). The Cd uptake of “IR8” was half that of “CHOKOUKOKU”.

Table 2 Plant or growth characteristics and cadmium (Cd) absorption of indica rice (Indian Rice Oryza ssp.) varieties at harvest time (2007)

Rice variety	Plant height (cm)	Culm length (cm)	Lodging	Shattering habit	Weight (Mg kg^−1)	Cd concentration (mg kg^−1)	Cd uptake (g ha^−1)
“IR8”	62.4 b	–	–	±	8.16	30.4 b	248 ab
“CHOKOUKOKU”	125.4 a	77.1	++	++	10.32	50.1 a	517 a
“MORETSU”	113.0 a	70.8	+	++	11.39	18.4 b	205 b
“MINAMIYUTAKA”	113.2 a	68.6	+	–	10.69	21.4 b	221 b

The same letter indicates no significant difference at P < 0.05 based on Tukey’s multiple-comparison test.

Phytoextraction using high Cd-accumulating rice varieties in upland fields

The shoot dry weight of “IR8” was higher than that of “CHOKOUKOKU”, with the exception of 2009 (Fig. 1). The yearly change in dry weight of “IR8” and “CHOKOUKOKU” was small (relative standard deviation (RSD) was 2.5 and 9.9%, respectively), with the average value over 4 years being 15.3 and 11.7 Mg ha⁻¹, respectively. No differences in the dry weight were observed among the “CHOKOUKOKU” varieties planted in the “CHOKOUKOKU” plot and those planted in the rotation plot on the third planting (2010); i.e., injury by continuous “CHOKOUKOKU” cropping was not observed. The Cd concentration of the “CHOKOUKOKU” shoot was 2.5 times higher than that of “IR8” in the first year (Fig. 2). However, no statistical differences were observed between the Cd concentration of “CHOKOUKOKU” and that of “IR8” in the following years. Because the soil Cd concentration decreased each year (Table 3), the shoot Cd concentrations of these rice cultivars also decreased annually. The Cd uptake of phytoextraction plants depends on the combination of dry weight and Cd concentration of the shoot. Because the Cd concentration of “CHOKOUKOKU” shoot was higher than that of “IR8”, the Cd uptake of “CHOKOUKOKU” was higher than that of “IR8” (Fig. 3). The total Cd uptake of the “IR8” and “CHOKOUKOKU” for 4 years was 545 and 822 g ha⁻¹, respectively. The soil Cd concentration of the phytoextraction field in May 2008 was approximately 1.0 mg kg⁻¹ (Table 3). The soil Cd concentrations of plots in which the phytoextraction plants were planted gradually decreased as a consequence of the phytoextraction plants assimilating the soil Cd. The soil Cd concentration of the “CHOKOUKOKU” plot was lower than that of the “IR8” plot. The Cd concentrations of phytoextraction plot soils ultimately reached 0.64 mg kg⁻¹ (“IR8”) and 0.61 mg kg⁻¹ (“CHOKOUKOKU”).

Figure 1 Dry weight of shoot of phytoextraction plant. Error bars indicate standard deviation of three replicates. The same letters are not significantly different in the year at P < 0.05 based on Tukey’s multiple-comparison test.

Figure 2 Cadmium (Cd) concentration of phytoextraction plant. Error bars indicate standard deviation of three replicates. The same letters are not significantly different in the year at P < 0.05 based on Tukey’s multiple-comparison test.

The dry weight of sorghum was apparently higher than that of the rice varieties. Regrettably, as the Cd concentration of the shoot was low (3–4 mg kg⁻¹), the Cd uptake of sorghum was moderate. No differences in the Cd uptake of sorghum were observed at the second planting (2009) and fourth planting (2011). The total Cd uptake of the rotation plots for 4 years was 779 g ha⁻¹.

Table 3 Changes in soil cadmium (Cd) concentration in the field soil

	2008		2009		2010		2011
	May	October	May	October	May	October	May	October
Phytoextraction plot	(before sawing)	(after planting)	(before sawing)	(after planting)	(before sawing)	(after planting)	(before sawing)	(after planting)
control	0.97 ± 0.09	0.96 ± 0.12	0.95 ± 0.11	0.92 ± 0.12	0.94 ± 0.10	0.91 ± 0.12	0.97 ± 0.15	0.95 ± 0.10
“IR8”	0.95 ± 0.10	0.84 ± 0.09	0.87 ± 0.07	0.73 ± 0.08	0.76 ± 0.07	0.72 ± 0.08	0.70 ± 0.04	0.64 ± 0.05**
“CHOKOUKOKU”	1.01 ± 0.09	0.78 ± 0.05	0.80 ± 0.03	0.68 ± 0.08	0.70 ± 0.02	0.61 ± 0.09	0.66 ± 0.07	0.61 ± 0.08**
Rotation	1.05 ± 0.03	0.77 ± 0.05	0.79 ± 0.05	0.73 ± 0.16	0.75 ± 0.07	0.63 ± 0.08	0.70 ± 0.06	0.63 ± 0.09**

Data are the mean ± standard deviation (n = 3). **; P < 0.01 t-test of the soil Cd concentration before (before sawing at 2008) and after phytoextraction (after planting at 2012).

Effect of phytoextraction on the Cd concentration of wheat grains subsequently grown in the remediation field

The soil Cd concentration of control plots was 0.98 mg kg⁻¹ and that of the phytoextraction plots was 0.63–0.68 mg kg⁻¹ during wheat planting (Table 4). The yield of wheat grain of “SHIROGANEKOMUGI” and “CHUGOKU165” was 488–540 Mg ha⁻¹ and 552–622 Mg ha⁻¹, respectively. The yields were higher than the average for “SHIROGANEKOMUGI” cultured in this area (data not shown). The grain Cd concentrations of “SHIROGANEKOMUGI” in the phytoextraction plots were lower than those in the non-phytoextraction plots. However, they exceeded the Codex standard for wheat grain (0.2 mg kg⁻¹). On the other hand, the grain Cd concentration of “CHUGOKU165” was lower than that of “SHIROGANEKOMUGI”. The mean Cd concentrations of the “CHUGOKU165” grains planted in the phytoextraction plots complied with the Codex standard for wheat grain.

Table 4 Soil properties during wheat cultivation and cadmium (Cd) concentration in wheat (Triticum aestivum) grain of each cultivar

	Soil		Wheat cultivar
Phytoextraction plot	pH	Cd Concentrationmg kg^−1	“SHIROGANEKOMUGI”mg kg^−1	“CHUGOKU165”mg kg^−1
Control	5.84	0.98	0.79	0.28
“IR8”	6.00	0.68 **	0.40 **	0.17 **
“CHOKOUKOKU”	5.92	0.68 **	0.43 **	0.17 **
Rotation	5.97	0.63 **	0.35 **	0.13 **

**; Significantly different at P < 0.01 based on Dunnet’s multiple-comparison test (n = 3).

DISCUSSION

Selection of a rice variety is important for practical phytoextraction in upland fields. Although “CHOKOUKOKU” had some disadvantages, it was useful as a phytoextractor plant because of its high Cd uptake. While “IR8” is known to accumulate high levels of Cd in its shoots (Azuma et al. 2003), in this study, the amount was lower than that of the “CHOKOUKOKU” (Table 2, Fig. 3). However, the plant height of “IR8” was short, and it was assumed to have lodging tolerance (Ibaraki et al. 2009). Because several typhoons pass through southwest Japan each year, lodging tolerance is an important characteristic for phytoextractor plants grown in this area. The indications are that “CHOKOUKOKU” and “IR8” varieties are promising phytoextraction rice varieties for upland fields converted from paddy fields in southwest Japan. Because the “CHOKOUKOKU” has a shattering habit, it will be essential to use a non-shattering variety such as “MJ3” or “MA22” in the future (Abe et al. 2013). On the other hand, the Cd uptake of sorghum was higher than that of rice varieties at the fourth planting. As the yearly change in the Cd uptake of sorghum was very small, it seems that sorghum can be also used in phytoextraction of upland fields.

Figure 3 Cadmium (Cd) uptake of phytoextraction plant. Error bars indicate standard deviation of three replicates. The same letters are not significantly different in the year at P < 0.05 based on Tukey’s multiple-comparison test.

Murakami et al. (2009) showed that “CHOKOUKOKU” cultured for 2 years could remove Cd from paddy fields and reduce the soil Cd concentration by 38%. In our upland field examination, “CHOKOUKOKU” cultured for 2 years (2008–2009) reduced the soil Cd concentration by 32% (Table 3). Although this percentage was lower than Murakami’s experimental value, comparing these exact values is impractical because planting conditions were different. The Cd uptake of the rice plants decreased annually. However, this phenomenon may have been influenced by the decrease of the soil Cd concentration each year. As a result, the rate of degradation of the soil Cd concentration was reduced. The soil Cd concentration levels of the phytoextraction plot also decreased during the summer period, when the phytoextraction plants were cultivated. However, the levels increased slightly during other seasons. Honma et al. (2009) observed this phenomenon and attributed this finding to the decay of roots or stumps of phytoextraction plants left in the soil. Perronnet et al. (2000) demonstrated that Cd associated to the leaves of the hyperaccumulator T. caerulescens was very mobile after incorporation into the soil. Because the Cd concentrations of the roots and stumps were higher than that of the other parts (data not shown), it seems that their decay released Cd into the soil, resulting in a slight increase in its concentration level. These findings thus suggest that long-term cultivation of phytoextraction plants may be necessary to generate non-Cd-contaminated soil.

Table 5 shows the mass balances of soil Cd amounts in the phytoextraction plot. The initial Cd amount of field B as calculated from soil Cd concentration, soil layer and bulk density would be estimated from 2100 to 2321 g ha⁻¹. The final Cd amount in the soil was estimated as 1555 g ha⁻¹ for the “IR8” plot, 1410 g ha⁻¹ for the “CHOKOUKOKU” plot and 1542 g ha⁻¹ for the rotation plot. If these values were converted into the soil Cd concentrations, it would be 0.70, 0.64 and 0.70 mg kg⁻¹ for “IR8”, “CHOKOUKOKU” and the rotation plot, respectively. These estimated values are slightly higher than the measured values (i.e., 0.64 mg kg⁻¹ for the “IR8” plot; 0.61 mg kg⁻¹ for “CHOKOUKOKU” and the rotation plot). These differences may be attributable to Cd that has remained in the roots and stumps. Thus, the soil Cd concentration estimated from plant shoots was higher than the measured soil Cd values. As the roots and stumps of sorghum were thicker and firmer than those of the rice plant (data not shown), the decay of the roots and stumps of sorghum was slower than that of the rice plant. This characteristic may have also been responsible for the differences between the estimated value and measured value of the rotation plot, which were larger than that of other rice plots. As previously described, Cd in the roots and stumps is ultimately released into the soil; therefore, phytoextraction efficiency should be evaluated using the amount of Cd absorbed by the plant.

Table 5 Mass balances of soil cadmium (Cd) in the phytoextraction plots

	Initial soil	Soil	Bulk	Initial soil	Cd uptake	Final soil	Final soil	Final soil
	Cd conc.	layer	density	Cd amount	of the plant	Cd amount	Cd conc.	Cd conc.
	(A)	(B)	(C)	(estimated)	(D)	(estimated)	(estimated)	(measured)
Phytoextraction plot	mg kg^−1	m	kg L^−1	g ha^−1	g ha^−1	g ha^−1	mg kg^−1	mg kg^−1
“IR8”	0.95	0.17	1.3	2100	545	1555	0.70	0.64
“CHOKOUKOKU”	1.01	0.17	1.3	2232	822	1410	0.64	0.61
Rotation	1.05	0.17	1.3	2321	779	1542	0.70	0.61

Initial soil Cd amount; A × (B × 10,000) × C, final soil Cd amount; (initial soil Cd amount)-D, final soil Cd concentration; (final soil Cd amount)/(B × 10,000)/C.

Wheat varieties bred in Hokkaido tend to have a low Cd concentration compared to those bred in central and southwest Japan (Kubo et al. 2008). “CHUGOKU165” was recently bred to introduce this trait from the Hokkaido genotypes, including its ability to lower Cd concentration. The grain Cd concentration of “CHUGOKU165” was lower than that of “SHIROGANEKOMUGI” (Table 4). The grain Cd concentration of “CHUGOKU165”, which was cultured in the phytoextraction field, was reduced to the level of the Codex standard value (less than 0.2 mg kg⁻¹). “CHUGOKU165” is therefore the most effective wheat variety for the remediation of Cd-contaminated upland fields. On the other hand, the reduction ratios of the Cd concentrations in the wheat grains after phytoextraction to those of the control were higher for “SHIROGANEKOMUGI” than for “CHUGOKU165”. Unfortunately, the mechanism behind this discrepancy could not be established.

To maintain the Cd concentration of wheat grains within the Codex standard value, it is necessary to reduce soil Cd concentration. Fig. 4 shows the relationship between the Cd concentration of field soil and the Cd concentration of wheat grain. The grain Cd concentration was strongly and positively correlated with the soil Cd concentration. Regrettably, the grain Cd concentration of “SHIROGANEKOMUGI” was not reduced to the level of the Codex standard value; there was no intersection between the Codex standard value of wheat grain and the regression line. The intersection of the Codex standard value for wheat grain with the regression line and with the upper limit of the 95% confidence interval of “CHUGOKU165” corresponded to approximately 0.8 and 0.6 mg kg⁻¹ of the soil Cd concentration, respectively. This indicates that the Cd concentration of 95% of the “CHUGOKU165” grains would satisfy the Codex standard when the soil Cd concentration was less than 0.6 mg kg⁻¹. On the other hand, it is important to determine the process length of phytoextraction that would result in remediation of Cd-contaminated soil. Hammer and Keller (2003) suggested that the assessment of phytoextraction efficiency must be based on objective goals, including legislative requirements. However, as the Cd uptake in plants does not necessarily correlate with the soil Cd concentration, the Japanese government has not established regulatory limits for Cd concentrations in farmland soils. The Cd concentration in the rice grains could be reduced by keeping paddy fields irrigated until harvest time even when the soil Cd concentration of the paddy fields is high (Ono, 2003). Therefore, it is difficult to define the exact goal of phytoextraction in paddy fields. The upland field crops are grown under oxidative conditions and, thus, the grain Cd concentration of upland crops is affected by the soil Cd concentration. Therefore, the soil Cd concentration of upland fields should be maintained at low levels. In this study, the intersection between the Codex standard line and the upper limit of the 95% confidence interval of “CHUGOKU165” corresponded to 0.6 mg kg⁻¹ of the soil Cd concentration (Fig. 4). Thus, the terminal point of phytoextraction was considered to be when the soil Cd concentration of the field was decreased to 0.6 mg kg⁻¹. In the present study, indica rice plants were grown in specific conditions (i.e., soil classification, a grey lowland soil; texture, sandy clay loam; 0.1 M HCl extractable soil Cd concentration, 1.0 mg kg⁻¹; bulk density, 1.3 kg L⁻¹). Because soil properties seem to affect the Cd uptake of wheat, to assess the terminal point of phytoextraction in other conditions, it may be necessary to perform a similar experiment.

Figure 4 Relationship between the soil cadmium (Cd) concentration extracted with 0.1 M hydrochloric acid (HCl) and grain Cd concentration (2012). □: “SHIROGANEKOMUGI”, ●: “CHUGOKU165”. Straight lines and broken lines represent the regression line and confidence interval (0.95), respectively.

Based on the results of this study, we conclude that the indica rice varieties of “CHOKOUKOKU” and “IR8” are effective for practical phytoextraction in upland fields converted from paddy fields. In addition, we find that the terminal point of phytoextraction is when the Cd concentration of field soil is less than 0.6 mg kg⁻¹.

ACKNOWLEDGMENTS

This study was supported by a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan (Research project for hazardous chemicals to table HC-1110 and ensuring food safety from farm to table AC-1223). We appreciate M. Murakami, S. Ishikawa and T. Abe for their valuable suggestions. We thank A. Taniguchi, J. Aoki and H. Shinozaki for their assistance.