Evidences for the augmented Cd(II) biosorption by Cd(II) resistant strain Candida tropicalis XTA1874 from contaminated aqueous medium

Cadmium is one of the most dreadful heavy metals and is becoming a major toxicant in ground water with increasing concentration above the WHO Guidelines in drinking water (0.003 mg/L). The potential sources of cadmium include sewage sludge, phosphate fertilizers and ingredients like Ni–Cd batteries, pigments, plating and plastics. Cadmium levels are increased in water owing to the use and disposal of cadmium containing ingredients. Water draining from a landfill may contain higher cadmium levels. The authors have tried to evaluate the optimized nutritional conditions for the optimal growth and Cd(II) remediation capacity for a developed Cd(II) resistant yeast strain named Candida tropicalis XTA 1874 isolated from contaminated water-body in West Bengal. By analyzing the optimization conditions, a synthetic medium was developed and the composition has been given in the main text. The strain showed much better Cd(II) adsorption capacity under the optimized nutritional conditions (Mean removal = 88.077 ± 0.097%).

In order to determine the amount adsorbed and intracellular accumulation the obtained pellet was undergone some treatments. The pellet thus obtained was washed three times with deionized water and the pellet was treated with 0.1 M EDTA solution for 10 min. The adsorbed Cd(II) over the biomass surface was recovered as EDTA washable fraction and was measured again by Flame Atomic Absorption Spectroscopy (Shimadzu AA-7000, Japan) 37 . The amount that was mobilized intracellularly was enumerated by acid digestion (0.2 N H 2 SO 4 and HNO 3 ) of the biomass and Cd(II) was measured in the lysate by Flame Atomic Absorption Spectroscopy (Shimadzu AA-7000, Japan) at 228.8 nm 38 (2.2

. Q2).
Modeling of biosorption isotherms and kinetics. Langmuir and Freundlich isotherm models have been used to describe the experimental data. Langmuir model describes the formation of monolayer over the adsorbent surface and assumes continuous adsorption energy regardless of the degree of coverage [39][40][41] .
The Langmuir model is described by the following equation: where q max signifies maximum adsorption capacity (mg/g d.w.), K L is the Langmuir constant (L/mg) and C e is the equilibrium Cd(II) concentration in the solution in mg/L. The reciprocal form of the equation is (1) q e = V(C i − C e )/m (2) Removal (%) = C i − C e / C i × 100 (3) q e = q max K L C e /1 + K L C e sion 13, Stat-Ease Inc, Minneapolis, MN, USA) has been used to fit quadratic model to the experimental data and to determine the best combination of parameters that resulted in the optimum response value. Optimization of Cd(II) biosorption by Candida tropicalis XTA 1874 was determined by Central Composite Design (CCD) under Response Surface Methodology (RSM). RSM constitutes a group of empirical techniques evaluating the relationship between clusters of independent variables and the measured responses. Since empirically determining the effects of single factors at a time is time consuming, RSM boosts up the operational conditions as well as save the economy of the process by reducing experimental runs 42 . The modern study depicts the impact of various nutrients influencing the growth and metal bioremediation capability of the resistant strain Candida tropicalis XTA 1874. Sixteen independent variables for the current study were: Glucose concentration (%), Urea Concentration (%), K 2 HPO 4 concentration (%), KH 2 Table 1 and Supplementary Table 1 where xi, x 2 i , x 2 j ,….,x 2 k , x i x j , x i x k and x j x k denote the linear, quadratic and the interaction effects of the variables respectively. The terms β0, βi, βii, and βij are the regression coefficients for the constant, linear, quadratic and interaction terms respectively, the random error is ε and the response variables are indicated by Y. The experimental design along with the alteration in the trend in the variables is shown in Table 2 and Supplementary  Table 2. The initial Cd(II) concentration used in the study was 500 ppm Cd(II).

Instrumental analysis. FT-IR analysis.
FT-IR spectroscopy analyses the electromagnetic radiation absorbed by the sample. The yeast biomass of both before and after optimization was centrifuged (10000 rpm, 10 min, + 4 °C, Cooling Centrifuge, Remi c24BL) and separated from the broth culture. The cell pellets of the biomass of both biosorbents were washed thrice with deionized water to remove the growth media residu-(4) 1/q e = 1/q max + 1/q max K L × (1/C e ) (5) R L = 1/1 + K L C o (6) q e = K f × (C e ) 1/n (7) Log q e = Log K f + 1/n Log C e (8) Ln q e − q t = Ln q e − k 1 t (9) t/q t = t/q e + 1/k 2 q 2 e (10) Z = Z 0 − Z c /�Z  Tables 1-3 and Supplementary Tables 1-3). The F and p value are considered to be important in determining the significance of each of the variables. The Model F-value of 21.68 implies the model is significant. There is only a 0.01% chance that an F-value this large could occur due to noise. It has been confirmed by the regression analysis the linear model term (Q), the interactive model terms (AK), (BF), (BG), (CO), (DL), (EH), (FH), (LN), (LQ) and the quadratic terms (F 2 ) and (Q 2 ) were significant (p < 0.05) ( Table 3,  Supplementary Table 3). The estimation of the quadratic model design matrix was done by using p-values. The (12) M desorbed /M sorbed × 100%  Table 4). Table 2, and  Supplementary Tables 2 and 5 and the plot in Fig. 1 and Supplementary Fig. 1 both shows that the actual and predicted values are very close to each other and distribution of the data is close to the fitted line. This indicated that the experimental model is suitable in describing the experimental data. According to the analysis the small probability value of the model is confirm to reject the null hypothesis and the data followed a normal distribution. The equation obtained from the coefficient terms of the coded factors (Supplementary Table 6) for the response variable has been shown in Eq. (14).
Interaction effects of the variables on Cd(II) removal by the strain and selection of synthetic media. The nutrient composition plays significant roles in the growth and removal capacities by the organisms apart from the physical parameters. The interaction effects of glucose concentration and dry cell mass on Cd(II) biosorption efficiency (%) has been assessed in Fig. 2 and Supplementary Fig. 2. The contour plots show that the dry cell mass has much significant effect compared to the glucose concentration on Cd(II) biosorption efficiency (%) by the strain. According to the ANOVA results, the p-value is more than 0.05 proving the interaction effect on  Table 5, Supplementary  Tables 7, 8). The coded values have been calculated using Eq. (10). From the contour and 3D plots (Fig. 2, Supplementary Table 2) it is evident that Cd(II) biosorption efficiency (%) significantly increased with the increase with the increasing amount of both the carbon and nitrogen sources (Glucose and Urea respectively) along with increasing dry cell mass. Similar increase has also been witnessed with the increasing amounts of the trace elements used in the study. But in all cases the effect of dry cell mass is quite prominent. It suggests that the amount of dry cell mass has considerable impact on metal ion biosorption and it too has been found in our recent study. Regarding the ANOVA results, the interaction between Glucose concentration and CaCO 3    www.nature.com/scientificreports/ Candida tropicalis CBL-1 strain has been reported to remove 70% Cd(II) in lab scale and maximum 60% from industrial wastewater 37 . The data showed that the most important constituents for organisimal growth and optimal biosorption are carbon (Glucose) and nitrogen (Urea) sources. Divalent ions such as Zn 2+ , Fe 2+ and Ca 2+ are mimicked by Cd 2+ and sometimes competitively obstruct its adsorption 46,47  Validation of the model. Optimized conditions were maintained have been maintained for checking the suitability of the model for response value prediction. Optimized Cd(II) biosorption was validated under optimized experimental conditions. The response value at optimized nutritional conditions was 95.028%. On the other hand, experimental value under optimized conditions was 95.972 ± 0.0001% using 500 ppm of initial Cd(II) concentration. Experimental response value was well in agreement with the predicted response value (Table 5,  Supplementary Table 8). Based on the above observation the synthetic media selected for optimum growth was shown in (Table 6, Supplementary Table 9).
Biosorption kinetics. Biosorption kinetics determines the rate of adsorption of dissolved adsorbates on the surface of biological adsorbents. Thus kinetic analysis aids to determine the biosorbent's ability to use as an effective Cd(II) adsorbent. Among the most profoundly used kinetic models described in the literature, those that uses order of chemical reactions are well considered. These models are the Pseudo First Order (Lagargren) and Pseudo Second Order (Mckay and Ho) kinetic models 41,49,50 . Cd(II) biosorption kinetics by the developed resistant strain Candida tropicalis XTA 1874 was performed in the before and after optimized conditions. Usually adsorption kinetics involves two phases: a rapid removal stage (first 60 min) from the aqueous solution followed by a slower removal stage before reaching the equilibrium (150 min) (Supplementary Tables 10-13, [15][16][17][18]. The detailed kinetic analysis before and after optimization conditions has been described in the supplementary files (Supplementary Tables 10-13, [15][16][17][18] respectively. Intracellular accumulation of Cd(II) in the due course of removal has also been estimated during kinetic analysis and shown in the above mentioned tables. The kinetic data for Cd(II) biosorption by the strain before and after optimized conditions has been shown in (Tables 7, 8, Supplementary Tables 14, 19) along with the linear plots (Figs. 3, 4, Supplementary Figs. 3, 4). Considering the correlation coefficient obtained by linear plotting of pseudo first and second order equations it can be concluded that Cd(II) adsorption by the biomass of Candida tropicalis   Supplementary Figs. 3, 4). From the above observation it can be concluded that the rate limiting step of the Cd(II) biosorption was chemisorption 41,51 . The calculated q e values obtained before and after optimization from the pseudo second order model were closer to that obtained by the experiment (Tables 7, 8, Supplementary Tables 14, 19). The rate constant k 2 increased with increase in initial Cd(II) concentration indicating the presence of more than one mechanism influencing Cd(II) binding to the yeast cell biomass surface in the culture 41 .    (Supplementary Tables 20, 22). Langmuir and Freundlich isotherms are most frequently used isotherms describing solid-liquid adsorption 52 .
The mathematical analysis of Cd(II) sorption at equilibrium by the strain can be best described by the Langmuir equation. According to the Langmuir's theory adsorption occurs at homogeneous sites on the adsorbent surface by monolayer sorption 53 . After analyzing the data presented in (Tables 9, 10, Supplementary Tables 21,  22) it can be concluded that both Langmuir and Freundlich models can describe the experimental data. The linear plots of Langmuir and Freundlich model both before and after optimization are depicted in Fig. 5 and Supplementary Fig. 5. The higher R 2 values obtained from the Langmuir model better describes the relationship between the amount of Cd(II) sorption at equilibrium. The values of the separation factor R L lies between 0 and 1 (Tables 9, 10, Supplementary Table 21, 22) describing favorable adsorption both before and after optimization.   (Tables 9, 10, Supplementary Tables 21, 22). The higher value of K F (11.721 ± 0.002) after optimization indicates increase in the affinity of the adsorbent towards the toxicant after optimization. The parameter 'n' from Freundlich isotherm indicates the intensity of Cd(II) adsorption. The values (Tables 9, 10, Supplementary Tables 21, 22) lies between in the range 1 < n < 10, confirms the efficiency of the adsorption process. The inverse parameter (1/n), an irrational fraction, informs us about the degree of diversity of the adsorption sites. Its values lies between 0 < 1/n < 1 confirms significant homogeneity of the yeast cell surface 41 . The strain showed mean Cd(II) removal of 88.077 ± 0.097% which is significantly higher than the mean Cd(II) removal before optimized conditions 75.007 ± 0.002% (Tables 9, 10, Supplementary Tables 21, 22). From the statistical analysis using Student's T-test (paired two tail) [sample size (n) = 6] (Table 11, Supplementary Table 24) it can be seen that there are significant differences in both the maximal surface adsorption capacity (q max ) and mean removal (%) (p < 0.05) before and after optimizing conditions. The results were compared with those published by Ref. 24 where it was showed increase in Cd(II) biosorption capacity after optimization using Turbinaria ornata biomass. The immobilized biomass showed an increase in the q max value compared to the free biomass and biosorption better fitted the Langmuir 54 . The equilibrium adsorption also followed Langmuir model. As from the previous study using Candida tropicalis CBL-1 strain it can remove 70% Cd(II) has been reported 37 . The developed resistant strain Candida tropicalis has the capability to remove 85.55% Cd(II) under optimized conditions using the synthetic media developed from statistical optimization using response surface methodology.

Instrumental evidences.
FT-IR analysis. The FT-IR spectra (Fig. 6, Supplementary Fig. 6) between 4000 and 400 cm −1 show the number of peaks indicating the presence of several functional groups in the control (C) and the Cd 2+ resistant strain before (BO-LC), and after (AO-LC) optimisation. The peaks at 3368 cm −1 (C), and 3390 cm −1 (BO-LC) arose due to the stretching of the N-H bond of the amino groups indicating the presence of bonded -OH group 55 . The change in the peak position from 3368 cm −1 , and 3390 cm −1 to 3400 cm −1 (AO-LC) indicates the binding of Cd 2+ ions with N-H and -OH groups. The broader peaks at 2923 cm −1 (BO-LC) and 2926 cm −1 (AO-LC) were due to -CH stretching vibrations of -CH 3 and -CH 2 functional groups 55 . The peaks between 1750 cm −1 and 1740 cm −1 were due to the C=O stretching vibration indicating the presence of carboxylic acids or esters 56 . The 1641 cm −1 (C) and 1644 cm −1 (BO-LC) peaks are due to the C=O group stretching from aldehydes and ketones 57 . The shifting of the peaks from 1641 and 1644 cm −1 to 1649 cm −1 (AO-LC) were due to the binding capability of these groups with Cd 2+ ions 55 . The peak at 1570 cm −1 was due to CO, C-O, and O-H groups in the BO-LC spectra 55 . The region between 1410 and 1060 cm −1 was due to OH, C-H stretching vibrations and C-O bending vibrations 55 . The shifting of the bands from 1063 cm −1 (C) and 1066 cm −1 (BO-LC) to 1078 cm −1 (AO-LC) was due to the binding capability of C-O bonds with Cd 2+ ions 55 . The region between 810 and 800 cm −1 was due to C-C, C-O, and C-O-P stretching vibrations of cellular polysaccharides 58 . The shifting

Analysis of the desorption efficiency and reusability of the biomass. Desorption efficiency (η%)
and reusability of the biomass is regarded as one of the most important properties that make waste water treatment as a cost effective process. As can be seen from (Table 12, Fig. 8, Supplementary Table 25, Supplementary Fig. 8) that biomass from the developed resistant strain showed efficient desorption capacity (Q5: Results) (91.648 ± 0.197%) at the first round of the desorption experiment. The adsorbent was reused with slight decrease in the adsorptive removal and desorption efficiency (η%). Desorption analysis was carried out for five cycles after that no change in desorption efficiency was observed. In each reusage cycle of the biomass the surface and intracellularly accumulated amount (mg/g) has been shown which was determined by EDTA chelation and acid digestion respectively (Supplementary Table 25). Kinetic analysis also showed that equilibrium was reached Table 11. Statistical analysis of the significance of the before and after optimization model by student's T-test (n = 6).

Before optimization (q max )
After optimization (q max ) T-test   Table 26). The terms q i and q t signifies the initial amount (mg/g) of surface accumulated Cd(II) and the amount of Cd(II) still remained in the biomass at time (t) after contact with the eluent solution respectively (Supplementary Table 26).
C r , Concentration of Cd(II) in the desorbing solution (ppm), V r , Volume of the desorbing solution, C i , initial Cd(II) concentration at the adsorbing solution (500 ppm), V, volume of the adsorbing solution (0.1L), C e , Cd(II) concentration (ppm) in the adsorbing solution at equilibrium.
Kinetic analysis of desorption was carried out using liner plotting of parabolic diffusion model and Elovichtype model (Table 13, Fig. 9, Supplementary Table 27, Supplementary Fig. 9) 62 . FE SEM image with EDAX analysis of the cells (3.956 ± 1.296 × 3.878 ± 0.097 µm) after desorption have been shown in (Fig. 10, Supplementary  Fig. 10). EDAX analysis showed a little retention of Cd(II) (0.5wt%) even after desorption of Cd(II) from the biomass.
To analyze the best fitting of the models, the coefficient of determination (R 2 ) and standard error of estimate (SE) were calculated by the following formula (15) Desorption efficiency (η %) = C r × V r /(C i − C e )V × 100   (6).
Based on the values of R 2 and SE (Table 13, Supplementary Table 27), it can be demonstrated that desorption kinetics is following the Elovich Kinetic Model where the calculated and experimental values of C a0 are very close. The derived parameter data complied with the Elovich model assumption αβt > > 1 62 . Cd(II) release from soil has been tested by various organic acids where it has been found that parabolic diffusion best fitted Cd(II) desorption kinetics 65 . One the other hand 62 found Elovich type model to be best fitted for desorption kinetic data.

Conclusion
Toxicant removal by microbial biosorption represents an efficient and cost-effective means of environmental remediation. The developed resistant strain Candida tropicalis XTA1874 exhibited high biosorption capacity after optimizing the culture conditions. In this work the contributions of various nutritional factors have been considered to aggravate microbial growth and biosorption capacity. The obtained results indicate significant Cd(II) binding after optimized conditions. The data follows the Langmuir isotherm model and biosorption plowed pseudo second order kinetics. Each of the nutritional factors plays vital role in accelerating microbial growth and toxicant removal process besides the physical parameters. Based on the optimization study a synthetic media has been developed which aids in accelerated microbial growth and bio-removal capacity. The strain was also undergone efficient desorption and showed significant bio-removal capacity as far as six cycles. Based on the above findings it can be concluded that the strain has tremendous bio-removal capacity and can be assumed to be effectively used in Cd(II) removal from polluted water bodies with an efficient and easily doable technique.

Data availability
All data generated or analysed during this study are included in this published article [Supplementary Information files].