Optimization of a Fungally Bioaugmented Biomixture for Carbofuran Removal in On-Farm Biopurification Systems

  • Karla Ruiz-Hidalgo
  • Juan Salvador Chin-Pampillo
  • Mario Masís-Mora
  • Elizabeth Carazo-Rojas
  • Carlos E. Rodríguez-Rodríguez
Article

Abstract

Biomixtures comprise the active part of biopurification systems (BPS) for the removal of pesticide-containing wastewater from agricultural origin. Considering that biomixtures contain an important amount of lignocellulosic substrates, their bioaugmentation with degrading ligninolytic fungi represents a promising way to improve BPS. The fungus Trametes versicolor was employed for the bioaugmentation of rice husk-compost-soil (GCS) biomixtures in order to optimize the removal of the highly toxic insecticide/nematicide carbofuran (CFN). Composition of biomixtures has not been optimized before, and usually, a volumetric composition of 50:25:25 (lignocellulosic substrate:humic component:soil) is employed. Optimization of the biomixture composition was performed with a central composite design, using the volumetric content of rice husk (pre-colonized by the fungus) and the volumetric ratio compost/soil as design variables. Performance of biomixtures was comprehensively assayed considering CFN removal, the production of toxic transformation products (3-hydroxycarbofuran/3-ketocarbofuran), the ability to mineralize [14C]carbofuran, and the residual toxicity in the matrix. According to the models, the optimal volumetric composition of the GCS biomixture is 30:43:27, which maximizes removal and mineralization rate, and minimizes the accumulation of transformation products. Results support the value of assessing new biomixture formulations according to the target pesticide in order to obtain their optimal performance, before their use in BPS.

Keywords

Biopurification system Pesticides Bioaugmentation Fungi Toxicity Degradation 

1 Introduction

The use of pesticides constitutes a necessary practice in many agricultural activities. Nonetheless, the continuous presence of pesticide residues in environmental compartments results in negative effects on ecosystems and environmental quality (Damalas and Eleftherohorinos 2011). Such pollution with pesticides is mainly produced at point source losses during filling and cleaning of application tanks or inadequate disposal of application residues (De Wilde et al. 2007).

On-farm biopurification systems (BPS) have emerged as a biotechnological tool to address point source contamination. Pesticide-containing wastewaters from agricultural pest control activities are collected and poured onto the biologically active part of the BPS, the biomixture, where accelerated pesticide degradation takes place (Verhagen et al. 2015). Biomixtures are constituted by three components: soil, which acts as the main source of degrading microorganisms; a lignocellulosic substrate, to support the growth and colonization of ligninolytic fungi; and a humic component (usually compost or peat), to favor pesticide retention in the matrix (Castillo et al. 2008; De Wilde et al. 2007). In order to enhance the degradation process that takes place within the BPS, strategies such as biostimulation and bioaugmentation have been adopted. Biostimulation with terpenes and some terpene-rich wastes has shown an increased capacity to remove atrazine (Tortella et al. 2013b, c), and successful removal was demonstrated with the addition of fertilizers to biomixtures in the case of chlorpyrifos (Tortella et al. 2010). The bioaugmentation of biomixtures using pesticide-degrading microorganisms is another interesting approach in BPS design (Verhagen et al. 2013). Given that an important fraction of the biomixture is constituted by a lignocellulosic substrate, the bioaugmentation with ligninolytic fungi is of particular interest to potentially increase the degrading capacity of the biomixture; nonetheless, just a few reports deal with this topic (Bending et al. 2002; Rigas et al. 2007; von Wirén-Lehr et al. 2001). Ligninolytic fungi such as white rot fungi are recognized for their ability to transform a wide range of organic pollutants (Yang et al. 2013) that include pesticides (Bending et al. 2002; Fragoeiro and Magan 2008). The versatility in their degrading capacity has been demonstrated in complex matrices such as soil, sewage sludge, and wastewater (Cruz-Morató et al. 2013; D’Annibale et al. 2005; Novotný et al. 1999; Rodríguez-Rodríguez et al. 2014). The use of white rot fungi in pesticide degradation and their potential application in biopurification systems has been reviewed elsewhere (Rodríguez-Rodríguez et al. 2013; Tortella et al. 2013a).

Efficiency of biomixtures depends on several factors, such as temperature of operation, moisture content, and maturity and composition of the biomixture (Castillo et al. 2008; Tortella et al. 2012). Traditionally, biomixtures employed in BPS contain the volumetric composition of 50:25:25 adopted from the early stages of BPS development, or empiric changes that include the substitution of peat for compost or the use of different lignocellulosic wastes (Coppola et al. 2007; de Roffignac et al. 2008; Karanasios et al. 2010; Kravvariti et al. 2010; Urrutia et al. 2013). It has been postulated that composition of biomixtures should be adapted to availability of local materials (Coppola et al. 2007), which might result in differences in their performance on pesticide degradation. Regardless of the materials employed, formal optimization of the composition of biomixtures has not been performed, in particular, nor in the case of bioaugmented biomixtures.

A recent study by our group revealed the efficiency of biomixtures bioaugmented with the white rot fungus Trametes versicolor to remove the pesticide carbofuran (CFN) (Madrigal-Zúñiga et al. 2015). CFN is a neurotoxic insecticide and nematicide that, despite being banned from agricultural use in some regions, is still extensively used in developing countries (though under the process of being seriously restricted or banned as well).

This work aimed to optimize the composition of a biomixture bioaugmented with the fungus T. versicolor to maximize the degradation of CFN. Optimization considered a multicomponent approach that evaluated the removal of CFN, the accumulation of two of its transformation products, the mineralization rate of radiolabeled CFN, and the residual toxicity of the matrix after the degradation process. A central composite design and response surface methodology were employed to obtain the composition that produces the most desirable effects on the response variables.

2 Materials and Methods

2.1 Chemicals and Reagents

Commercial CFN (Furadan®48 SC, 48 g a.i./100 mL) was purchased from a local market. Analytical standard CFN (2,2-dimethyl-2,3-dihydro-1-benzofuran-7-ylmethylcarbamate, >99 % purity), 3-hydroxycarbofuran (99.5 %), and 3-ketocarbofuran (99.5 %) were obtained from Chemservice (West Chester, PA, USA). Radiolabeled CFN [Ring-U-14C]-Carbofuran (14C-CFN; 2.89 × 109 Bq g−1; radiochemical purity 100 %; chemical purity 99.5 %) was obtained from Izotop (Institute of Isotopes Co., Budapest, Hungary). Carbendazim-d3 (surrogate standard, 99.0 %) and carbofuran-d4 (internal standard, 99.5 %) were purchased from Dr Ehrenstorfer (Augsburg, Germany). Ultima Gold cocktail for liquid scintillation counting was purchased from Perkin Elmer (Waltham, MA, USA). Solvents and extraction chemicals are listed in Ruiz-Hidalgo et al. (2014).

2.2 Fungal Strain and Biomixture Components

T. versicolor (ATCC 42530) was obtained from the American Type Culture Collection and maintained by subculturing every 30 days on dextrose potato agar slants (pH 4.5) at 25 °C. T. versicolor blended mycelial suspension was prepared according to a procedure by Font Segura et al. (1993), modified by the use of Sabouraud broth as culture medium. Clay loam soil (sand 40 %, silt 27 %, clay 33 %) was collected from the upper soil layer (0–20 cm) of an onion field with history of CFN application, in Tierra Blanca, Cartago, Costa Rica. Soil was air-dried and sieved (<2 mm). Rice husk obtained from an agricultural supplier from Cartago, Costa Rica was employed as the lignocellulosic substrate. Garden compost employed as the humic component was collected from a composting station located at Universidad de Costa Rica and sieved as previously described after air-drying.

2.3 Experimental Design, Response Surface Methodology, and Statistical Analysis

A central composite design (CCD) methodology with two factors (k = 2) was applied to optimize the composition of a biomixture composed of rice husk, compost, and soil, to remove CFN. The design variables or factors were the volumetric content (%) of fungal pre-colonized rice husk (RH, A) and the volumetric ratio compost/soil (C/S, B). The effect of these factors was observed on the following response variables: the percentage of CFN removal after 3 days (R1), the residual concentration of the transformation products 3-hydroxycarbofuran and 3-ketocarbofuran after 3 days (R2), the rate of 14C-CFN mineralization over a period of 30 days (R3), and the residual acute toxicity of the matrix after 8 days (R4). CCD employs 2k factorial points representing all combinations of the codified values (±1), 2k axial points at a distance  ± α from the origin, and at least three central points in the origin (encoded as 0, 0). The factor levels were normalized and coded in the range (−α, +α). The α value corresponds to 1.414 (α = F1/4, where F = 2k).

Nine combinations of the design variables were evaluated; to determine the experimental uncertainty, the central point was performed by quintuplicate, resulting in an experimental design that included 13 runs. The experimental design matrix is shown in Table 1 and comprises actual values for the different combinations of factors A and B. The CCD was centered in the point (origin) where A = 50 % and B = 1, which corresponds to the most commonly described volumetric composition of biomixtures employed in BPS, which is lignocellulosic substrate:compost:soil at a ratio of 50:25:25 (Castillo et al. 2008). The volumetric content of pre-colonized rice husk, A, ranged from 20 to 62.4 %, considering that this is the source of T. versicolor as the bioaugmentation agent. The ratio C/S, B, ranged from 0 (which represents absence of compost) to 2; absence of soil was not evaluated as it represents the main source of degrading microbiota.
Table 1

CCD design matrix and response values during the removal of CFN in GCS biomixtures

Run

Actual factors

Corresponding GCS biomixture composition

Responses

A

B

RH (%)

Compost (%)

Soil (%)

R1

3-Hydroxycarbofuran (μg kg−1)

3-Ketocarbofuran (μg kg−1)

R2

R3

R4

RH (% v/v)

Volumetric ratio compost/soil (%/%)

CFN removal after 3 days (%)

3-Hydroxycarbofuran + 3-ketocarbofuran (μg kg−1)

Mineralization rate k (d−1)

Residual toxicity after 8 days (TU)

1

50

1

50.0

25.0

25.0

22.9 ± 2.6

49.2 ± 16.6

648.2 ± 14.3

697.4 ± 21.9

0.00107 ± 0.00008

61

2

50

1

50.0

25.0

25.0

27.4 ± 2.6

47.2 ± 16.7

666.2 ± 14.5

713.4 ± 22.1

0.00250 ± 0.00028

116

3

50

1

50.0

25.0

25.0

26.5 ± 2.6

57.5 ± 16.5

970.3 ± 17.5

1027.8 ± 24.0

0.00156 ± 0.00009

41

4

50

1

50.0

25.0

25.0

28.0 ± 2.5

60.2 ± 16.5

849.8 ± 16.2

910.0 ± 23.1

0.00384 ± 0.00026

43

5

50

1

50.0

25.0

25.0

26.8 ± 2.6

65.4 ± 16.6

1047.6 ± 18.5

1112.9 ± 24.8

0.00152 ± 0.00012

26

6

62.4

1.4

62.4

22.0

15.6

26.6 ± 4.9

87.9 ± 5.7

571.6 ± 61.4

659.5 ± 61.7

0.00186 ± 0.00015

41

7

62.4

0.6

62.4

13.9

23.7

26.1 ± 1.8

73.2 ± 17.3

806.5 ± 64.7

879.7 ± 66.9

0.00334 ± 0.00035

103

8

37.6

0.6

37.6

23.4

39.0

26.6 ± 1.9

65.3 ± 17.0

682.0 ± 249.3

747.3 ± 249.9

0.00266 ± 0.00032

52

9

37.6

1.4

37.6

36.4

26.0

33.2 ± 5.3

54.0 ± 16.7

709.4 ± 163.6

763.3 ± 164.5

0.00396 ± 0.00022

16

10

50

2

50.0

33.3

16.7

46.1 ± 11.9

22.9 ± 14.6

555.2 ± 151.2

578.1 ± 151.9

0.00203 ± 0.00012

48

11

80

1

80.0

10.0

10.0

40.5 ± 4.0

76.7 ± 18.7

1221.2 ± 239.7

1297.9 ± 240.4

0.00121 ± 0.00012

25

12

50

0

50.0

0.0

50.0

35.9 ± 3.0

123.3 ± 35.8

1005.8 ± 59.5

1129.1 ± 69.4

0.00210 ± 0.00029

45

13

20

1

20.0

40.0

40.0

25.1 ± 14.3

37.2 ± 9.4

440.9 ± 115.8

478.1 ± 116.2

0.00580 ± 0.00042

124

Each response variable (Ri) can be fitted to a second order polynomial model (k = 2), according to the equation:
$$ {R}_i={b}_0+{b}_1A+{b}_2B+{b}_{12} AB+{b}_{11}{A}^2 + {b}_{22}{B}^2. $$
(1)

The estimation of the model parameters (bi) and the statistical analysis were performed using the software Design Expert 9.0 (Stat-Ease Inc., Minneapolis, USA). The quality of the fit polynomial model was determined by Fisher’s F test, and model terms were evaluated by the P value with 95 % confidence level; results were completely analyzed by analysis of variance (ANOVA) employing the same software. In addition, a variable called “desirability” (which ranges from 0 to 1) that encompasses the simultaneous effects of the response variables was also determined. Optimization of the biomixture composition was conducted with response surface methodology by the analysis of contour plots and numerical solutions by the software.

2.4 Experimental Procedures

2.4.1 Degradation Experiments

CFN degradation assays were performed in 12 cm × 3.5 cm plastic tubes containing 5 g of each biomixture. First, the proper amount of dry rice husk and distilled water was sterilized at 121 °C during 15 min prior to the inoculation of the blended T. versicolor mycelial suspension (0.35 mL per gram of dry rice husk). After fungal colonization for 10 days at 25 °C, biomixtures were prepared by mixing the pre-colonized rice husk, soil, and compost at the volumetric ratios shown in Table 1 in order to obtain a total of nine biomixtures of different composition, according to the design variables described in Section 2.3. The biomixtures were then spiked with Furadan®48 SC to give a CFN final concentration of ∼15 mg kg−1 and incubated in static conditions at 25 °C during 3 or 8 days. The remaining concentration of the pesticide and its transformation products was determined by sacrificing duplicate unitary systems (quintuplicates for the central point of the CCD) at times 0 h and 3 days, as described in Section 2.5.1. Additional unitary systems (10 g each) obtained after 8 days of treatment (eight per biomixture) were combined as a composite sample for the determination of acute toxicity as described in Section 2.5.3.

2.4.2 Mineralization Studies

The mineralization of 14C-CFN was determined by monitoring the production of 14CO2 in biometric systems containing 14CO2 traps with 10 mL KOH (0.1 M) (Ruiz-Hidalgo et al. 2014). Fifty grams of each biomixture was prepared into triplicate biometric flasks as described in Section 2.4.1 and spiked with commercial CFN (30 mg kg−1) and 14C-CFN (3000 dpm g−1). The systems were incubated in the dark at (25 ± 1)°C during 30 days. The KOH in the flasks was withdrawn at selected times and replaced with the same amount of fresh KOH. Activity of 14C from the trapped 14CO2 was analyzed in the KOH samples as described in Section 2.5.2.

2.5 Analytical Procedures

2.5.1 Extraction and Analysis of CFN and Transformation Products

Extraction was carried out as described in Ruiz-Hidalgo et al. (2014). Carbendazim-d4 and carbofuran-d3 were added as surrogate and internal standard, respectively, to unitary samples (5 g) obtained from degradation experiments. CFN and its transformation products were analyzed by LC-MS/MS using ultra high performance liquid chromatography (UPLC 1290 Infinity LC, Agilent Technologies, CA, USA) coupled to a triple quadrupole mass spectrometer (6460, Agilent Technologies, CA, USA). Chromatographic separation was done at 40 °C by injecting 6 μL samples (2 μL loop) in a Poroshell 120 EC-C18 column (100 mm × 2.1 mm i.d., particle size 2.7 μm; Agilent Technologies, CA, USA) and using acidified water (formic acid 0.1 % v/v, A) and acidified methanol (formic acid 0.1 % v/v, B) as mobile phases. The mobile phase flow was 0.3 mL min−1 at the following conditions: 30 % B for 3 min, followed by a 15-min linear gradient to 100 % B, 4 min at 100 % B, and 0.1 min gradient back to 30 % B, followed by 5 min at initial conditions. Selected transitions for the analytes are shown in Table 2. The mass spectrometer employed a jet stream (electrospray) ionization source operating at the following conditions: gas temperature 300 °C; gas flow 7 L min−1; nebulizer 45 psi; sheath gas temperature 250 °C; sheath gas flow 11 L min−1; capillary voltage 3500 V (for positive and negative); nozzle voltage 500 V (for positive and negative); and heater MS1 and MS2 at 100 °C. Recovery of the method was 91 % for CFN, 98 % for 3-hydroxycarbofuran, and 95 % for 3-ketocarbofuran. Limit of detection (LOD) and limit of quantification (LOQ) were 13 and 26 μg kg−1 for CFN and 3-ketocarbofuran, and 16 and 32 μg kg−1 for 3-hydroxycarbofuran.
Table 2

Selected transitions and other parameters in the detection of CFN and its transformation products using the dynamic MRM method

Compound

Retention time (min)

Transition

Fragmentor (V)

Collision energy (V)

Type of transition

3-Hydroxycarbofuran

3.60 ± 1.00

238 → 163

72

9

Q

238 → 107

33

q

3-Ketocarbofuran

6.10 ± 1.00

236 → 161

82

17

Q

236 → 179

9

q

Carbofuran

7.92 ± 1.00

222 → 165

82

9

Q

222 → 123

21

q

Carbendazim-d4

1.45 ± 1.00

196 → 164

102

17

Q

196 → 136

34

q

Carbofuran-d3

7.92 ± 1.00

225 → 165

86

9

Q

225 → 123

21

q

Q quantification transition, q qualifier transition

2.5.2 Determination of 14CO2 from Mineralization Assays

Scintillant liquid (10 mL) was added to 2 mL aliquots from the withdrawn KOH solution samples, and the 14C activity from 14CO2 was measured by liquid scintillation using a Beckman LS6000SC counter. The total cumulative 14CO2 activity and the initially added 14C-CFN activity were used to calculate the percentage of 14C-pesticide mineralized. Pesticide mineralization was modeled using a first order model.

2.5.3 Acute Toxicity: Daphnia magna Immobilization Test

An acute test based on D. magna immobilization was employed to estimate the residual toxicity in the biomixtures. The test was conducted following the US EPA (2002) and some modifications as described by Ruiz-Hidalgo et al. (2014), using elutriates from the biomixtures at time 0 h (after CFN spiking) and after 8 days of treatment, as analytical matrix. Elutriates were obtained according to the protocol of EPA (2001). EC50, the concentration producing 50 % of immobilization in the daphnids, was determined using the TOXCALC—Toxicity Data Analysis Software from Tidepool Scientific Software. Toxicity results were expressed as toxicity units (TU), calculated according to the following expression: TU = (EC50)−1 · 100.

3 Results and Discussion

3.1 Removal of CFN and Formation of Transformation Products

The use of fungal bioaugmentation to enhance the performance of biomixtures in the removal of pesticides has been barely studied (Bending et al. 2002; von Wirén-Lehr et al. 2001). A work from our group recently demonstrated the improvement in the removal of CFN in rice husk biomixtures bioaugmented with T. versicolor, particularly in the peat-based with respect to the compost-based biomixture (Madrigal-Zúñiga et al. 2015). These biomixtures were elaborated using the conventional volumetric proportion of 50:25:25 (lignocellulosic substrate:humic component, peat or compost:soil), commonly employed in BPS (Castillo et al. 2008). Nonetheless, considering that peat is not an autochthonous material, and therefore its use would increase the cost of the biomixture, it was decided to optimize the composition of the biomixture that employs compost as the humic component, which is a mixture of rice husk-compost-soil (GCS). Moreover, toxicological data has revealed that peat itself and peat-based biomixtures are more toxic than compost and compost-based biomixtures (Chin-Pampillo et al. 2015), and in general, compost is considered as a more environmentally friendly alternative to peat (Fogg et al. 2003). Rice husk was used given that this lignocellulosic substrate permitted the best colonization by T. versicolor, in terms of visual growth and the enzymatic activity of laccase (Madrigal-Zúñiga et al. 2015).

CFN half-life in the bioaugmented GCS biomixture (50:25:25) was 8.1 days (Madrigal-Zúñiga et al. 2015); the removal of the pesticide after a shorter period (3 days) was employed as the response variable to evaluate the performance of the different biomixtures in the CCD, and the results are shown in Table 1. The removal of CFN in 3 days ranged from 22.9 % in one of the 50:25:25 biomixtures to 46.1 % when A = 50 % and B = 2 (that is a 50:33.3:16.7 biomixture). The average removal for the 50:25:25 biomixtures was 26.3 %, significantly below the maximum values achieved with other compositions. Results suggest a more accelerated removal compared to barley straw-based biomixtures bioaugmented with several fungi (T. versicolor, Stereum hirsutum, Hypholoma fasciculare) employed in the elimination of diverse pesticides (Bending et al. 2002).

The correlation between the design variables (RH and C/S) and the CFN removal after 3 days (R1) was analyzed by the response surface methodology, which permitted to fit the data to a second order polynomial model according to Eq. (1) in Table 3. The accuracy of the fit was statistically analyzed by ANOVA, which demonstrated that the model was significant (Pmodel > F = 0.0054), with goodness in the fit (R2 = 0.8108) and adequate precision (8.956, desired over 4). Other statistical parameters of the model are presented in Table 4.
Table 3

Final equations (in terms of actual factors) describing the behavior of response variables on CFN removal in GCS biomixtures

Response variable

Final equation in terms of actual factors

Removal after 3 days (R1)

46.5 − 0.589A − 24.9B + 7.45 × 10− 3A2 + 14.9B2

(1)

3-Hydroxycarbofuran

65.3 + 0.705A − 37.5B

(2)

3-Ketocarbofuran

503 + 9.62A − 202B

(3)

3-Hydroxycarbofuran + 3-ketocarbofuran (R2)

569 + 10.3A − 240B

(4)

Mineralization rate (R3)

5.84 × 10− 3 − 6.43 × 10− 5A − 5.38 × 10− 5B

(5)

Table 4

Analysis of variance (ANOVA) and other statistical parameters for the models employed to fit the response variables during the removal of CFN in GCS biomixtures

Response variable

Source of variations

Sum of squares

df

Mean square

F value

Probability P (>F)

R2

Adjusted R2

Adequate precision

CFN removal in 3 days (R1)

Model

453.60

4

113.40

8.57

0.0054

0.8108

0.7161

8.956

Residual

105.88

8

13.23

     

Lack of fit

89.96

4

22.49

5.65

0.0610

   

Pure error

15.92

4

3.98

     

Total

559.48

12

      

3-Hydroxycarbofuran + 3-ketocarbofuran (R2)

Model

4.093 × 105

2

2.046 × 105

7.24

0.0114

0.5915

0.5098

7.673

Residual

2.827 × 105

10

2.827 × 104

     

Lack of fit

1.454 × 105

6

2.423 × 104

0.71

0.6658

   

Pure error

1.373 × 105

4

3.433 × 105

     

Total

6.920 × 105

12

      

Mineralization (R3)

Model

9.995 × 10−6

3

4.998 × 10−6

4.23

0.0468

0.4580

0.3496

7.386

Residual

1.183 × 10−5

9

1.183 × 10−6

     

Lack of fit

6.951 × 10−6

5

1.158 × 10−6

0.36

0.5460

   

Pure error

4.876 × 10−6

4

1.219 × 10−6

     

Total

2.182 × 10−5

12

      
The coefficient values of the model in terms of coded factors (Table 5) are useful to determine the relative impact of the factors on the response variable. The coefficients related to RH and C/S are both positive, which indicates that higher values of these factors result in increased removal of CFN; in addition, the coefficients have almost the same magnitude, and therefore, their impact on CFN removal is quite similar. The coefficients for the quadratic factors A2 and B2 also had a positive effect on the removal, though higher in the case of B2. Other interactions were not included in the model, as they were not significant. The regression equation was graphically represented by the contour plot (Fig. 1). As observed, the removal of CFN is maximized at two regions: at high values of C/S (regardless of RH values) and at high RH values combined with low C/S values. Maximization of R1 performed in the Expert Design software indicates that optimum conditions for CFN elimination in 3 days take place (within the limits shown in the figure) at A = 30 % and B = 1.6, which corresponds to a 30:43:27 (GCS) biomixture. It is remarkable that the regions of maximum CFN removal do not include the central point that corresponds to the composition 50:25:25, commonly employed in BPS.
Table 5

Regression results from the CCD experiments for the modeling of response variables

Model

Model term

Coefficient (coded factors)

Standard error

F value

P value

CFN removal after 3 days (R1)

Intercept

25.7

1.41

  

A

1.93

0.92

4.42

0.0688

B

1.97

0.90

4.85

0.0587

A2

1.14

0.52

4.89

0.0579

B2

2.38

0.48

24.24

0.0012

3-Hydroxycarbofuran

Intercept

63.1

4.36

  

A

8.74

3.97

4.86

0.0521

B

−15.0

3.87

15.04

0.0031

3-Ketocarbofuran

Intercept

783

46.5

  

A

119

42.3

7.96

0.0181

B

−80.9

41.3

3.84

0.0786

3-Hydroxycarbofuran + 3-ketocarbofuran (R2)

Intercept

846

46.6

  

A

128

42.4

9.11

0.0129

B

−98.9

41.4

5.36

0.0431

Mineralization rate (R3)

Intercept

2.57 × 10−3

3.02 × 10−4

  

A

−7.97 × 10−4

2.74 × 10−4

8.44

0.0157

B

−2.15 × 10−5

2.68 × 10−4

6.46 × 10−3

0.9375

Fig. 1

Surface response (contour plot) of the removal of CFN after 3 days (R1) in rice husk-compost-soil bioaugmented biomixtures of different composition. Standard composition (50:25:25) is indicated with an “×”

The concentration of the transformation products 3-hydroxycarbofuran and 3-ketocarbofuran was analyzed in the different biomixtures after 3 days of treatment. These compounds are the main CFN metabolites detected in several matrices including soil, and their presence is of concern as they produce the same toxic effects on organisms as the parental compound (Gupta 1994). Accumulation of 3-ketocarbofuran in the biomixtures ranged from 440.9 to 1221.2 μg kg−1 and was higher than that of 3-hydroxycarbofuran, whose concentrations ranged from 22.9 to 123.3 μg kg−1 (Table 1). Previous assays with T. versicolor in liquid medium and in rice husk alone have shown the production of 3-hydroxycarbofuran but not 3-ketocarbofuran (Mir-Tutusaus et al. 2014; Ruiz-Hidalgo et al. 2014); however, in the case of biomixtures bioaugmented with the fungus (50:25:25), both metabolites were detected and their complete removal required 20 days in peat-based and 34 days in compost-based biomixtures (Madrigal-Zúñiga et al. 2015). Data corresponding to each transformation product were fitted to linear models as described in the regression Eqs. 2 and 3 in Table 3. From the analysis of coefficients from coded factors (Table 5) and response surface (Fig. 2), the presence of both compounds follows the same trend: high values of RH favor accumulation of metabolites, while high C/S ratios tend to result in lower metabolite accumulation. In the case of 3-hydroxycarbofuran, the most important effect is given by the C/S ratio, contrary to 3-ketocarbofuran for which RH exerts a more important role in the production of the metabolite. These results suggest that higher amounts of fungal biomass might result in undesired formation of toxic transformation products, whose accelerated degradation is potentially favored by the microbial communities present in soil and/or compost.
Fig. 2

Surface response (contour plot) of the accumulation of transformation products from CFN in rice husk-compost-soil bioaugmented biomixtures of different composition. a 3-Hydroxycarbofuran. b 3-Ketocarbofuran. c Summatory (3-hydroxycarbofuran + 3-ketocarbofuran, R2). Standard composition (50:25:25) is indicated with an “×”

Given that both metabolites present similar toxicity, the sum of 3-hydroxycarbofuran and 3-ketocarbofuran concentrations was considered as a response variable for the optimization of the biomixture composition (R2). Values for this variable in the different biomixtures ranged from 478.1 to 1129.1 μg kg−1 (Table 1). As expected, the behavior of R2 followed the same trend as that of the individual metabolites. The data were fitted to a significant linear model (Pmodel > F = 0.0114; lack of fit 0.6658, Table 4), for which the coefficients of the coded factors indicate a higher effect of RH than C/S ratio (128 vs −98.9). As described before, lower amounts of colonized rice husk and higher C/S ratios result in lower accumulation of toxic transformation products (Fig. 2). Minimization of R2 using the Expert Design software resulted in the same optimum composition as for R1, A = 30 % and B = 1.6, corresponding to a 30:43:27 (GCS) biomixture.

3.2 Mineralization of 14C-CFN in the Biomixtures

The mineralization of 14C-CFN was determined in the biomixtures in a period of 30 days, in order to assay their ability to completely oxidize the pesticide, and therefore, to minimize the possible accumulation of toxic transformation products. In fact, removal of the parental compound may occur without mineralization, as it has been observed in compost-based biomixtures used for the elimination of bentazone (Coppola et al. 2011). Mineralization rate constants are shown in Table 1 and ranged from 0.00107 d−1 in one replicate of the CCD central point to 0.00580 d−1 in the biomixture 20:40:40 (GCS). Interestingly, higher mineralization rates do not correlate to higher removal values; this could be ascribed to the fact that part of the 14C was assimilated to microbial biomass instead of being released as 14CO2. On the contrary, the higher mineralization rate did correlate with the lowest accumulation of transformation products (20:40:40 biomixture); this result suggests that further degradation of these metabolites from CFN might favor the accelerated complete oxidation of the pesticide. Mineralization rates were slightly higher than those obtained in ten different non-bioaugmented biomixtures during CFN removal, which ranged from 0.00058 to 0.0025 d−1 (Chin-Pampillo et al. 2015).

The correlation between the design variables and the mineralization rate was fitted to a linear model. Accuracy of the fit by ANOVA indicated that the model was significant (Pmodel > F = 0.0468; lack of fit 0.5460; Table 4). Coefficients of the coded factors reveal that RH exerts the most important effect on the mineralization rate (Table 5); moreover, this coefficient is negative, indicating that higher RH values tend to decrease the mineralization in the biomixture. Similarly, the C/S ratio negatively affects this response variable, though at a lower extent. From the response surface (Fig. 3), it can be seen that low RH values, regardless of the C/S ratio employed, favor the mineralization in the systems. On the contrary, colonized rice husk amounts over 50 % v/v should be avoided in the design of biomixtures for CFN elimination, as they reduce the mineralization of the pesticide. Optimum mineralization rate occurs at RH = 30 % and a C/S ratio = 0.4, which corresponds (within the limits) to a bioaugmented biomixture 30:20:50 (GCS). As it has been observed in the cases of the other response variables, optimum mineralization performance is achieved at compositions significantly different from the conventional 50:25:25 biomixture, particularly at higher proportions of soil; this finding could be ascribed to the effect of larger microbial communities from soil whose presence might accelerate the complete oxidation of intermediate products of the pesticide.
Fig. 3

Surface response (contour plot) of the mineralization of 14C-CFN, represented as mineralization rates obtained during a 30-day period (R3) in rice husk-compost-soil bioaugmented biomixtures of different composition. Standard composition (50:25:25) is indicated with an “×”

3.3 Residual Toxicity in the Biomixtures

Toxicological evaluation of the biomixtures provides a helpful tool to estimate the actual potential application of BPS as eco-friendly devices. In order to compare the different biomixtures, the residual toxicity after 8 days of treatment was assayed using an immobilization test with D. magna. The treatment time to apply the analysis was established in 8 days, as reports of toxicity at longer treatment times (e.g., 60 days) have shown complete elimination in the toxicity of several biomixtures used for CFN elimination (Chin-Pampillo et al. 2015).

Toxicological data could not be modeled in the Expert Design software. Most residual toxicity values ranged from 40 to 60 TU (Table 1). Lower values (below 30 TU) were obtained in three biomixtures, including one of the replicates of the central point and one containing the highest rice husk amount (80 %), which coincides with one region of high CFN removal according to Fig. 1. On the contrary, three biomixtures showed residual toxicity of more than 100 TU; interestingly, in one of these biomixtures (20:40:40, GCS), the lowest concentration of transformation products (3-hydroxycarbofuran and 3-ketocarbofuran) and the highest mineralization rate were achieved. These data suggest that other transformation products of higher toxicity (not considered in this work) might have accumulated at a higher extent in this biomixture. Overall results did not permit to establish a correlation between toxicity and the other response variables used in the work. Previous studies revealed that 48 days were required to completely remove the toxicity in fungal bioaugmented biomixtures used for CFN degradation (Madrigal-Zúñiga et al. 2015), while CFN degradation by T. versicolor in sterile rice husk left a residual toxicity of 11 TU (and incomplete pesticide removal) after 60 days (Ruiz-Hidalgo et al. 2014).

3.4 Global Optimization of the Biomixture Composition

The variable desirability was employed to comprise the effects of the response variables in order to obtain one single optimized biomixture composition. The optimization criteria included the maximization of R1 (CFN removal after 3 days) and R3 (mineralization rate of 14C-CFN in 30 days), and the minimization of R2 (total concentration of transformation products after 3 days). The response surface and the contour plot are shown in Fig. 4. As expected according to the behavior of the individual variables (particularly R1 and R2), the region that results in higher desirability comprises low values of RH and high C/S ratios. The numerical solution for desirability maximization (i.e., optimization) provided by the software was achieved with the values A = 30 % and B = 1.6 (desirability = 0.750; values range from 0 to 1), which corresponds to a volumetric composition of 30:43:27 (GCS), in agreement with the optimization for individual response variables R1 and R2.
Fig. 4

Surface response of desirability (combining the effect of response variables R1, R2, and R3) during the removal of CFN in rice husk-compost-soil bioaugmented biomixtures of different composition. a Three-dimensional response surface. b Contour plot. Standard composition (50:25:25) is indicated with an “×”

Interestingly, the optimized composition differs from the 50:25:25 typical composition, particularly by reducing the amount of lignocellulosic substrate (colonized by T. versicolor in this case) and substituting it with compost. Reduction in the necessary content of lignocellulosic substrate might be explained by the higher amount of fungal biomass due to the pre-colonization of the rice husk. Analogous systems called biopiles, employed in the degradation of pharmaceuticals, UV-filters, and brominated flame retardants, were successful when they were bioaugmented with T. versicolor by previous colonization of a lignocellulosic substrate at 38 % w/w proportions (Rodríguez-Rodríguez et al. 2011, 2012, 2014). Contrary to this work, Coppola et al. (2011) found no difference in the degradation of the pesticides isoproturon and bentazone in non-bioaugmented biomixtures at increasing content of vine branch straw as lignocellulosic substrate (12.5, 25, 50 %).

The optimum amount of soil here determined resembles the 25 % v/v usually employed in BPS; it has been demonstrated that low soil amounts (0.5 %) are enough to provide an appropriate degrading population, however, the lowest the soil inoculum, the largest the lag phase in the pesticide removal (as determined during linuron mineralization by Sniegowski et al. 2012). In this respect, it should be highlighted that most of the degrading populations come from the soil, partly due to the pre-exposure to CFN. Non-bioaugmented GCS biomixtures (50:25:25) have shown CFN half-lives of around 4 days and a mineralization rate of 0.00158 d−1 (Chin-Pampillo et al. 2015), an effect that relies entirely on the native soil microorganisms and those provided by the compost or the coconut fiber. The impact of the bioaugmentation could either produce a collaborative removal of the pesticide by T. versicolor and the indigenous microbiota, or the inhibition of some indigenous (degrading) populations due to fungal activity. With respect to other biomixture components, the evaluation of biomixtures with different compost contents (25 and 50 %) for the degradation of chlorpyrifos showed the best performance at 25 % (Kravvariti et al. 2010), contrary to the optimized value obtained in the present work. On the other hand, Coppola et al. (2011) achieved higher removal of isoproturon and bentazone when they employed 50 % with respect to 87.5 % compost. The contrasting results remark the pertinence of accurate assessment of biomixtures in order to maximize the performance of a BPS.

Bioaugmentation with fungi seems to improve the performance of biomixtures for CFN removal; similar enhanced effect was achieved in the bioaugmentation of a biomixture with the fungus Phanerochaete chrysosporium in the elimination of isoproturon, with respect to non-bioaugmented biomixtures (von Wirén-Lehr et al. 2001). Fungal persistence can be affected due to the competition of indigenous microbiota from the biomixture; therefore, additional studies should be conducted in order to monitor the active biomass of T. versicolor present in the matrix. The persistence of this fungus has been successfully reported in solid-phase systems for periods up to 21 days (Rodríguez-Rodríguez et al. 2012). Further improvement of bioaugmented biomixtures could be potentially obtained by re-inoculation with fungal biomass, a strategy that has been proved successful in the elimination of emerging pollutants in sludge by T. versicolor (Rodríguez-Rodríguez et al. 2014).

4 Conclusions

The optimization of a bioaugmented biomixture for the degradation of CFN using T. versicolor yielded a volumetric composition of 30:43:27 (GCS). Such composition presents a marked difference with respect to the 50:25:25 biomixture that has been commonly applied in BPS, which remarks the necessity to evaluate new biomixture formulations according to specific pesticide application schemes. The use of rice husk (agricultural waste) as the carrier for T. versicolor as the bioaugmentation agent supports the cost-effectiveness of the process. This is the first report known by the authors on the optimization of bioaugmented biomixtures for the degradation of pesticides and provides relevant information on the design of biomixtures for application in BPS.

Notes

Acknowledgments

This work was supported by the Vicerrectoría de Investigación, Universidad de Costa Rica (projects 802-B2-046, 802-B4-503, and 802-B4-609), the Costa Rican Ministry of Science, Technology and Telecommunications, MICITT (project FI-093-13/802-B4-503), and the Joint FAO/IAEA (project TC COS5/029).

References

  1. Bending, G. D., Friloux, M., & Walker, A. (2002). Degradation of contrasting pesticides by white rot fungi and its relationship with ligninolytic potential. FEMS Microbiology Letters, 212, 59–63.CrossRefGoogle Scholar
  2. Castillo, M. P., Torstensson, L., & Stenström, J. (2008). Biobeds for environmental protection from pesticide use—a review. Journal of Agricultural and Food Chemistry, 56, 6206–6219.CrossRefGoogle Scholar
  3. Chin-Pampillo, J. S., Ruiz-Hidalgo, K., Masís-Mora, M., Carazo-Rojas, E., & Rodríguez-Rodríguez, C. E. (2015). Adaptation of biomixtures for carbofuran degradation in on-farm biopurification systems in tropical regions. Environmental Science and Pollution Research, 22, 9839–9848.CrossRefGoogle Scholar
  4. Coppola, L., Castillo, M. P., Monaci, E., & Vischetti, C. (2007). Adaptation of the biobed composition for chlorpyrifos degradation to southern Europe conditions. Journal of Agricultural and Food Chemistry, 55, 396–401.CrossRefGoogle Scholar
  5. Coppola, L., Castillo, M. P., & Vischetti, C. (2011). Degradation of isoproturon and bentazone in peat- and compost-based biomixtures. Pest Management Science, 67, 107–113.CrossRefGoogle Scholar
  6. Cruz-Morató, C., Ferrando-Climent, L., Rodríguez-Mozaz, S., Barceló, D., Marco-Urrea, E., Vicent, T., & Sarrà, M. (2013). Degradation of pharmaceuticals in non-sterile urban wastewater by Trametes versicolor in a fluidized bed bioreactor. Water Research, 47, 5200–5210.CrossRefGoogle Scholar
  7. D’Annibale, A., Ricci, M., Leonardi, V., Quaratino, D., Mincione, E., & Petruccioli, M. (2005). Degradation of aromatic hydrocarbons by white-rot fungi in a historically contaminated soil. Biotechnology & Bioengineering, 90, 723–731.CrossRefGoogle Scholar
  8. Damalas, C. A., & Eleftherohorinos, I. G. (2011). Pesticide exposure, safety issues, and risk assessment indicators. International Journal of Environmental Research and Public Health, 8, 1402–1419.CrossRefGoogle Scholar
  9. de Roffignac, L., Cattan, P., Mailloux, J., Herzog, D., & Le Bellec, F. (2008). Efficiency of a bagasse substrate in a biological bed system for the degradation of glyphosate, malathion and lambda-cyhalothrin under tropical climate conditions. Pest Management Science, 64, 1303–1313.Google Scholar
  10. De Wilde, T., Spanoghe, P., Debaer, C., Ryckeboer, J., Springael, D., & Jaeken, P. (2007). Overview of on-farm bioremediation systems to reduce the occurrence of point source contamination. Pest Management Science, 63, 111–128.CrossRefGoogle Scholar
  11. EPA. (2001). Methods for collection, storage and manipulation of sediments for chemical and toxicological analyses: technical manual (EPA-823-B-01-002). Washington, DC: Office of Water (4305).Google Scholar
  12. EPA. (2002). Methods for measuring the acute toxicity of effluents and receiving waters to freshwater and marine organisms (EPA-821-R-02-012). Washington, DC: Office of Water (4303T).Google Scholar
  13. Fogg, P., Boxall, A. B. A., Walker, A., & Jukes, A. A. (2003). Pesticide degradation in a “biobed” composting substrate. Pest Management Science, 59, 527–537.CrossRefGoogle Scholar
  14. Font Segura, X., Gabarrell Durany, X., Ramos Lozano, D., & Vicent Huguet, T. (1993). Detoxification pretreatment of black liquor derived from non-wood feedstock with white-rot fungi. Environmental Technology, 14, 681–687.CrossRefGoogle Scholar
  15. Fragoeiro, S., & Magan, N. (2008). Impact of Trametes versicolor and Phanerochaete chrysosporium on differential breakdown of pesticide mixtures in soil microcosms at two water potentials and associated respiration and enzyme activity. International Biodeterioration and Biodegradation, 62, 376–383.CrossRefGoogle Scholar
  16. Gupta, R. C. (1994). Carbofuran toxicity. Journal of Toxicology and Environmental Health, 43, 383–418.CrossRefGoogle Scholar
  17. Karanasios, E., Tsiropoulos, N. G., Karpouzas, D. G., & Ehaliotis, C. (2010). Degradation and adsorption of pesticides in compost-based biomixtures as potential substrates for biobeds in southern Europe. Journal of Agricultural and Food Chemistry, 58, 9147–9156.CrossRefGoogle Scholar
  18. Kravvariti, K., Tsiropoulos, N. G., & Karpouzas, D. G. (2010). Degradation and adsorption of terbuthylazine and chlorpyrifos in biobed biomixtures from composted cotton crop residues. Pest Management Science, 66, 1122–1128.CrossRefGoogle Scholar
  19. Madrigal-Zúñiga, K., Ruiz-Hidalgo, K., Chin-Pampillo, J. S., Masís-Mora, M., Castro-Gutiérrez, V., & Rodríguez-Rodríguez, C. E. (2015). Fungal bioaugmentation of two rice husk-based biomixtures for the removal of carbofuran in on-farm biopurification systems. Biology and Fertility of Soils. doi:10.1007/s00374-015-1071-7.Google Scholar
  20. Mir-Tutusaus, J. A., Masís-Mora, M., Corcellas, C., Eljarrat, E., Barceló, D., Sarrà, M., Caminal, G., Vicent, T., & Rodríguez-Rodríguez, C. E. (2014). Degradation of selected agrochemicals by the white-rot fungus Trametes versicolor. Science of the Total Environment, 500–501, 235–242.CrossRefGoogle Scholar
  21. Novotný, Č., Erbanová, P., Šašek, V., Kubátová, A., Cajthaml, T., Lang, E., Krahl, J., & Zadražil, F. (1999). Extracellular oxidative enzyme production and PAH removal in soil by exploratory mycelium of white rot fungi. Biodegradation, 10, 159–168.CrossRefGoogle Scholar
  22. Rigas, F., Papadopoulou, K., Dritsa, V., & Doulia, D. (2007). Bioremediation of a soil contaminated by lindane utilizing the fungus Ganoderma australe via response surface methodology. Journal of Hazardous Materials, 140, 325–332.CrossRefGoogle Scholar
  23. Rodríguez-Rodríguez, C. E., Jelić, A., Llorca, M., Farré, M., Caminal, G., Petrović, M., Barceló, D., & Vicent, T. (2011). Solid-phase treatment with the fungus Trametes versicolor substantially reduces pharmaceutical concentrations and toxicity from sewage sludge. Bioresource Technology, 102, 5602–5608.CrossRefGoogle Scholar
  24. Rodríguez-Rodríguez, C. E., Jelić, A., Pereira, M. A., Sousa, D. Z., Petrović, M., Alves, M. M., Barceló, D., Caminal, G., & Vicent, T. (2012). Bioaugmentation of sewage sludge with Trametes versicolor in solid-phase biopiles produces degradation of pharmaceuticals and affects microbial communities. Environmental Science & Technology, 46, 12012–12020.CrossRefGoogle Scholar
  25. Rodríguez-Rodríguez, C. E., Castro Gutiérrez, V., Chin-Pampillo, J. S., & Ruiz-Hidalgo, K. (2013). On-farm biopurification systems: role of white rot fungi in depuration of pesticide containing wastewaters. FEMS Microbiology Letters, 345, 1–12.CrossRefGoogle Scholar
  26. Rodríguez-Rodríguez, C. E., Lucas, D., Barón, E., Gago-Ferrero, P., Molins-Delgado, D., Rodríguez-Mozaz, S., Eljarrat, E., Díaz-Cruz, M. S., Barceló, D., Caminal, G., & Vicent, T. (2014). Re-inoculation strategies enhance the degradation of emerging pollutants by fungal bioaugmentation in sewage sludge. Bioresource Technology, 168, 180–189.CrossRefGoogle Scholar
  27. Ruiz-Hidalgo, K., Chin-Pampillo, J. S., Masís-Mora, M., Carazo, R. E., & Rodríguez-Rodríguez, C. E. (2014). Degradation of carbofuran by Trametes versicolor in rice husk as a potential lignocellulosic substrate for biomixtures: from mineralization to toxicity reduction. Process Biochemistry, 49, 2266–2271.CrossRefGoogle Scholar
  28. Sniegowski, K., Bers, K., Van Goetem, K., Ryckeboer, J., Jaeken, P., Spanoghe, P., & Springael, D. (2012). Minimal pesticide-primed soil inoculum density to secure maximum pesticide degradation efficiency in on-farm biopurification systems. Chemosphere, 88, 1114–1118.CrossRefGoogle Scholar
  29. Tortella, G. R., Rubilar, O., Cea, M., Wulff, C., Martínez, O., & Diez, M. C. (2010). Biostimulation of agricultural biobeds with NPK fertilizer on chlorpyrifos degradation to avoid soil and water contamination. Journal of Soil Science and Plant Nutrition, 10, 464–475.CrossRefGoogle Scholar
  30. Tortella, G. R., Rubilar, O., Castillo, M. P., Cea, M., Mella-Herrera, R., & Diez, M. C. (2012). Chlorpyrifos degradation in a biomixture of biobed at different maturity stages. Chemosphere, 88, 224–228.CrossRefGoogle Scholar
  31. Tortella, G. R., Durán, N., Rubilar, O., Parada, M., & Diez, M. C. (2013a). Are white-rot fungi a real biotechnological option for the improvement of environmental health? Critical Reviews in Biotechnology, 35, 165–172.CrossRefGoogle Scholar
  32. Tortella, G. R., Rubilar, O., Cea, M., Briceño, G., Quiroz, A., Diez, M. C., & Parra, L. (2013b). Natural wastes rich in terpenes and their relevance in the matrix of an on-farm biopurification system for the degradation of atrazine. International Biodeterioration & Biodegradation, 85, 8–15.CrossRefGoogle Scholar
  33. Tortella, G. R., Rubilar, O., Stenström, J., Cea, M., Briceño, G., Quiroz, A., Diez, M. C., & Parra, L. (2013c). Using volatile organic compounds to enhance atrazine biodegradation in a biobed system. Biodegradation, 24, 711–720.CrossRefGoogle Scholar
  34. Urrutia, C., Rubilar, O., Tortella, G. R., & Diez, M. C. (2013). Degradation of pesticide mixture on modified matrix of a biopurification system with alternatives lignocellulosic wastes. Chemosphere, 92, 1361–1366.CrossRefGoogle Scholar
  35. Verhagen, P., De Gelder, L., & Boon, N. (2013). Inoculation with a mixed degrading culture improves the pesticide removal of an on-farm biopurification system. Current Microbiology, 67, 466–471.CrossRefGoogle Scholar
  36. Verhagen, P., Destino, C., Boon, N., & De Gelder, L. (2015). Spatial heterogeneity in degradation characteristics and microbial community composition of pesticide biopurification systems. Journal of Applied Microbiology, 118, 368–378.CrossRefGoogle Scholar
  37. von Wirén-Lehr, S., Castillo, M. P., Torstensson, L., & Scheunert, I. (2001). Degradation of isoproturon in biobeds. Biology and Fertility of Soils, 33, 535–540.CrossRefGoogle Scholar
  38. Yang, S., Hai, F. I., Nghiem, L. D., Price, W. E., Roddick, F., Moreira, M. T., & Magram, S. F. (2013). Understanding the factors controlling the removal of trace organic contaminants by white-rot fungi and their lignin modifying enzymes: a critical review. Bioresource Technology, 141, 97–108.CrossRefGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  • Karla Ruiz-Hidalgo
    • 1
  • Juan Salvador Chin-Pampillo
    • 1
  • Mario Masís-Mora
    • 1
  • Elizabeth Carazo-Rojas
    • 1
  • Carlos E. Rodríguez-Rodríguez
    • 1
  1. 1.Research Center of Environmental Contamination (CICA)Universidad de Costa RicaSan JoséCosta Rica

Personalised recommendations