Abstract
The adsorption ability of a native Jonesia quinghaiensis strain ZFSY-01, a microorganism isolated from uranium tailing wastewater, to U(VI) in wastewater under different conditions was studied in this work. The results showed that 391.5 mg U/g and 78.3% of adsorption capacity and efficiency were achieved under an optimum adsorption condition, respectively. Especially, the adsorption capacity of this strain reached the maximum (Q=788.9 mg U/g) under 100 mg/L of strain dosage. Simultaneously, the linear regression coefficients for the used isothermal sorption model indicate that the biosorption process is compatible with the Freundlich isotherm, the Temkin isotherm and the Halsey isotherm model. Based on the fitted kinetic parameters, the data from the experiments fit well with models of pseudo-second-order kinetics and intraparticle diffusion, suggesting that the strain ZFSY-01 immobilized U(VI) by physical and chemical adsorption. In addition, thermodynamic parameters demonstrated that the sequestration of U(VI) by the strain is spontaneous and endothermic. Based on the above analysis, strain ZFSY-01 can effectively remove U(VI) ions from high- or low-concentration uranium-containing wastewater and is expected to become a promising biological adsorbent.
HIGHLIGHTS
Specific microbes in special habitats have potential efficacy.
Adsorption capacity of strain ZFSY-01 reached 788.9 mg U/g under 100 mg/L of dosage.
Strain ZFSY-01 immobilized U(VI) by physical and chemical adsorption.
It is expected to be a green and efficient biosorbent material.
INTRODUCTION
The increasing application of nuclear energy has resulted in the emission of large quantities of radioactive wastes containing uranium (U) (Banala et al. 2021a; Banerjee et al. 2022). Among these radioactive wastes, U is mainly present as U(IV) and U(VI) in two fugitive forms, where U(IV) exists as insoluble sediment and U(VI) is mainly in the form of uranyl (
), with higher radiological and chemical toxicity (Maolin et al. 2022; Sourav et al. 2022). It should be noted that uranium is a radioactive heavy metal; small amounts of uranium ingested by plants and animals will destroy the genetic information of cells leading to reproductive disorders, while excessive input of uranium can cause organ failure and even death in the human body (Song et al. 2019; Banala et al. 2021a). If the wastewater is unreasonably remediated, the uranium ions in the wastewater may infiltrate into the soil and groundwater, causing serious ecological contamination as well as eventually affecting people's health by means of the food chain (Nie et al. 2022). Therefore, the treatment of uranium-containing wastewater is an issue of concern.
Nowadays, the commonly used treatment methods for uranium-containing wastewater include chemical sedimentation, redox, ion exchange and adsorption (Gupta et al. 2018; Pudza & Abidin 2020). However, these traditional physicochemical approaches have their own limitations, such as higher costs, generation of secondary pollutants as well as cumbersome treatment processes, which do not conform to the goals of sustainable development (Lakaniemi et al. 2019; Silas et al. 2022). Whereas the bioremediation method just meets this demand. Compared with traditional methods, it has the advantages of low material cost, high adsorption efficiency and excellent selectivity. It is well known that, among all wastewater bioremediation technologies, microbial adsorption is considered a promising yet environmentally friendly way of recovering uranium or other heavy metals (Yu et al. 2021; Sivashankar et al. 2022; Smjecanin et al. 2022).
According to previous reports, bacteria have been found to have an excellent binding capacity for metallic uranium ions as a biosorbent. For example, the Kocuria sp. strain studied by Wang et al. immobilized U(VI) with a biosorption capacity of 104 mg U/g through phosphate on the bacterial surface (Wang et al. 2018). Subsequently, Sánchez-Castro et al. studied a strain of Stenotrophomonas Br8 with an adsorption capacity of 373 mg U/g, which showed high phosphatase activity (Sánchez-Castro et al. 2020), and Yu et al. isolated the Pseudomonas stutzeri strain with an adsorption capacity of 308.72 mg U/g from seawater that is expected to be an adsorbent material for the extraction of uranium from seawater (Yu et al. 2022). Some of them can immobilize heavy metals using the diversity and abundance of functional groups on the cell surface, or accumulate heavy metals in the cell utilizing cellular metabolism (Beni & Esmaeili 2020). However, due to the structural and functional variability of different microorganisms, their ability to bind and tolerance to uranium varies. Moreover, a large amount of literature showed that indigenous microorganisms isolated from uranium-contaminated areas have a high tolerance to the heavy metal uranium and other toxic substances (Islam & Sar 2011; Banala et al. 2021b). Based on the foregoing description, it seems of interest to study indigenous strains isolated from uranium contamination. In this work, a strain was isolated from a uranium mine in north-western China using a screening medium with different uranium concentrations (100, 300, and 500 mg/L) that showed a high tolerance to uranium, named ZFSY-01. The strain was identified as Jonesia quinghaiensis based on the 16S rRNA gene sequence (GenBank Accession No. NR029030.1). To our knowledge, there are no reported studies on the immobilization of uranium for J. quinghaiensis strain ZFSY-01. Thus, the primary purpose of the study was to investigate the removal ability of J. quinghaiensis strain ZFSY-01 to 
in wastewater. First, the effects of different adsorption conditions (such as reaction time, temperature, strain dosage, system initial pH, and initial uranium concentration) on the adsorption performance of strain ZFSY-01 were studied to determine the optimal adsorption conditions. Secondly, the experimental data of uranium ion adsorption by strain ZFSY-01 were analyzed using various biosorption kinetics, thermodynamics as well as adsorption equilibrium models, and the fitting effects of various models were compared. Finally, the adsorption mechanism of uranium ion by strain ZFSY-01 was preliminarily explored.
MATERIALS AND METHODS
Experimental reagents
In this study, Luria Broth (LB) medium was used to culture the strain, in which tryptone and yeast extract were procured from Qingyao Bioengineering Co., Ltd (Qingdao, China), while sodium chloride (NaCl) was provided by Chengdu Kelong Chemical Co., Ltd (Chengdu, China). All experimental chemical substances were of analytical grade and did not need to be purified, and the solutions were prepared with sterile deionized water. Moreover, the uranium solvent used in the experiments was obtained by diluting uranyl nitrate standard stock solution (1,000 mg/L), which was purchased from ANPEL Scientific Instrument Co., Ltd (Shanghai, China).
Source and collection of experimental strain ZFSY-01
The strain ZFSY-01 was isolated from an evaporation pond of a uranium mine waste stream in Northwest China and cultured using an LB medium, as detailed in the Supplementary Information. In order to obtain a single colony, the target strain was purified several times through serial culture and then identified by 16S rRNA gene sequencing. The specific molecular biological identification of strain ZFSY-01 was completed by Sangon Biotech (Shanghai) Co., Ltd. In addition, the strain ZFSY-01 was incubated in a medium at 30 °C with shaking of 180 rpm for 12 h and further collected for subsequent experiments. The bacterial cells were collected by centrifugation at 6,000 rpm for 10 min at 4 °C and washed 2–3 times with sterile water.
Uranium tolerance of strain ZFSY-01
The tolerance of the strain ZFSY-01 for uranium was determined. The bacteria were inoculated in an LB liquid medium with a uranium concentration gradient of 0, 50, 100, 200 and 300 mg/L and incubated with shaking (30 °C, 180 rpm). Then, samples were collected at 0, 2, 4, 6, 8, 14, 24, 36 and 48 h intervals to determine the optical density (OD) values and plot the growth curves of the strains so that the effect of different uranium concentrations on the growth of the strains could be studied. The species distribution of uranium was modeled by Visual MINTEQ version 3.0 at 200 mg/L of initial uranium concentration and T = 30 °C between pH 2 and 10.
Experiments on the immobilization of U(VI) by strain ZFSY-01 under different factors
A batch experimental method was used in this work to evaluate the adsorption properties of strain ZFSY-01 for U(VI) in solution. Under laboratory conditions, the effects of reaction time, temperature, strain dosage, system initial pH and initial uranium concentration on the immobilization of uranium by the strain were investigated sequentially. First, wet strain cells of 20 mg of dry weight were added to 50 ml of uranium-containing an LB liquid medium with a uranium concentration of 200 mg/L and incubated in a constant temperature oscillating incubator (30 °C, 180 rpm). Subsequently, the samples were removed and centrifuged at defined time intervals to separate the supernatant, in which U(VI) concentration was determined using the uranium estimation method (Arsenazo-III) (Banala et al. 2021a). Without adding biomass as a control group, the influence of container wall and LB medium on uranium sorption was eliminated.
and adsorption capacity 
of uranium by bacteria were computed by Equations (1) and (2) (Silas et al. 2022):
and 
were the initial and equilibrium concentrations of the uranium solution (mg/L), respectively. V is the volume of the liquid (ml) and m is the dry weight of the strain (g).Experimental data analysis
All experiments were performed in triplicate, the experimental data were statistically analyzed using Microsoft Excel 2019, and the data were graphically plotted using Origin Pro 2021 software.
RESULT AND DISCUSSIONS
Uranium tolerance of strain ZFCW-01
The growth curve of strain ZFSY-01 at different uranium concentrations.
Effect analysis for experimental factors on the immobilized U(VI) of strain ZFSY-01
Reaction time
), the adsorption efficiency of the bacteria was basically unchanged, which could be considered to be in a dynamic adsorption equilibrium. From the above analysis, it could be concluded that the optimal equilibration time for U(VI) adsorption by strain ZFSY-01 was 4 h, which was used for subsequent adsorption studies. 
(a) Effect of reaction time on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (initial uranium concentration: 200 mg/L, V: 50 ml, T: 30 °C, strain dosage: 400 mg/L, pH: 6.5). (b) Effect of strain dosage on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (initial uranium concentration: 200 mg/L, V: 50 ml, T: 30 °C, reaction time: 4 h, pH: 6.5).
(a) Effect of reaction time on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (initial uranium concentration: 200 mg/L, V: 50 ml, T: 30 °C, strain dosage: 400 mg/L, pH: 6.5). (b) Effect of strain dosage on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (initial uranium concentration: 200 mg/L, V: 50 ml, T: 30 °C, reaction time: 4 h, pH: 6.5).
Strain dosage
The dosage of the strain affects the efficiency of the adsorbent used, so it is very important to explore the optimal dosage of bacteria. In this study, the effect of different bacterial biomass on U(VI) adsorption was investigated in order to determine the optimal solid–liquid ratio for U(VI) adsorption by strain ZFSY-01, as indicated in Figure 2(b). Different concentrations of bacterial biomass were added to the LB liquid medium with an initial U(VI) concentration of 200 mg/L. The concentration range of added bacterial biomass was 100–1,000 mg/L (dry weight). As could be seen from Figure 1(b), with the increase in the dosage of bacteria, the adsorption efficiency of bacteria for uranium in LB liquid medium increased, while the adsorption capacity was the opposite. That is, the adsorption capacity of bacteria reached the maximum (Q = 788.9 mg/g) when the number of bacteria was 100 mg/L. Compared with other reported strains such as Chryseomonas MGF-48 (198 mg/g) (Malekzadeh et al. 2002), Streptomyces levoris (9.04 mg/g) (Tsuruta 2004), Geobacillus thermoleovorans subsp. (11 mg/g) (Özdemir & Kilinc 2012), Bacillus thuringiensis ZYR3 (355 mg/g) (Pan et al. 2015), Bacillus subtilis (alginate–chitosan microcapsules) (376.64 mg/g) (Tong 2017), Bacillus amyloliquefaciens (179.5 mg/g) (Liu et al. 2019) and Candida albicans (41.15 mg/g) (Liu et al. 2021), strain ZFSY-01 showed significantly higher adsorption capacity. When the dosage of the bacteria reached 400 mg/L, the adsorption efficiency of uranium by the strain increased to 60.3%, and the adsorption capacity of bacteria decreased to 312.49 mg U/g. After that, with the increase in the dosage of bacteria, the adsorption efficiency gradually tended to stabilize, reaching a maximum of 64.1%. This might be due to a ‘shielding’ influence on the cell face, which shielded the usable binding sites, thus reducing the ability of the strain to immobilize uranium (Malkoc & Nuhoglu 2005; Wang et al. 2010). Therefore, the optimal dosage of strain ZFSY-01 was considered to be 400 mg/L, considering the best adsorption efficiency of the bacteria on U(VI).
Initial pH
in LB medium under strongly acidic conditions, which competed with 
. However, with the increase of pH in the solution, the 
ions in the solution decreased, with the result that 
could be more involved in binding to the negatively charged active sites on the bacteria surface, thus raising the adsorption capacity of strain ZFSY-01 for U(VI) (Hu et al. 2018). Meanwhile, at pH 5.0, the maximum adsorption efficiency and the adsorption capacity of strain ZFSY-01 for U(VI) was observed simultaneously, i.e., the maximum adsorption efficiency was 78.3%, and the 
was 391.5 mg U/g, indicating that strain ZFSY-01 had the best adsorption efficiency for U(VI) in LB liquid medium solution at pH 5.0. Notably, the adsorption capacity and the adsorption efficiency of uranium by bacteria showed a rapid decrease with increasing pH values between 5.0 and 9.0, and then began to stabilize. This could be explained by the increase of negatively charged uranium ions in solution with pH > 5, which repelled negatively charged active sites on the bacterial surface and inhibited the adsorption of U(VI) by bacterial cells (Salome et al. 2017; Li et al. 2019). In addition, Visual MINTEQ 3.0 software was applied to calculate the distribution of U(VI) species in solution at different pH conditions and the detailed results are reported in Table S2 of the Supplementary Information. The chemical speciation of uranium predicted under experimental conditions is shown in Figure 3(b), the pH of the solution was between 3.0 and 6.0, and the species of U(VI) in the solution were primarily present as 
, 
and 
ions. Whereas the negatively charged uranium ions in solution, such as 
and 
, mainly appeared at pH > 7 and gradually became prevalent (Ding et al. 2014). 
(a) Effect of pH on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (initial uranium concentration: 200 mg/L, V: 50 ml, T: 30 °C, strain dosage: 400 mg/L, reaction time: 4 h). (b) The species of uranium at different pH conditions was simulated with Visual MINEQL 3.0 (initial uranium concentration: 200 mg/L, T: 30 °C).
(a) Effect of pH on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (initial uranium concentration: 200 mg/L, V: 50 ml, T: 30 °C, strain dosage: 400 mg/L, reaction time: 4 h). (b) The species of uranium at different pH conditions was simulated with Visual MINEQL 3.0 (initial uranium concentration: 200 mg/L, T: 30 °C).
Initial uranium concentration
Effect of initial uranium concentration on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (V: 50 ml, T: 30 °C, strain dosage: 400 mg/L, reaction time: 4 h, pH: 6.5).
Effect of initial uranium concentration on U(VI) adsorption capacity and adsorption efficiency of strain ZFSY-01 (V: 50 ml, T: 30 °C, strain dosage: 400 mg/L, reaction time: 4 h, pH: 6.5).
Biosorption equilibrium analysis of uranium adsorption by strain ZFSY-01
(mg/g) and 
(mg/L), respectively. 
(mg/g) is the theoretical maximum saturation adsorption of the Langmuir isotherm, and 
(mg/L) is the balance constant of sorption. Significantly, the non-dimensional separating factor (
) is defined by the Langmuir isotherm and is represented by Equation (4) (Guo & Wang 2019):
(mg/L) is the initial uranium concentration, and the 
(mg/L) value is the equilibrium constant in the Langmuir isotherm. Furthermore, when 0 < 
< 1, it indicated that the biosorption properties were favorable (Esmaeili & Beni 2015). Freundlich and Temkin isothermal models were determined by Equations (5) and (6), respectively (Ezzati 2020):
((mg/g), (L/mg)1/n) and 
(g/L) are the adsorption equilibrium constants of the Freundlich and Langmuir isothermal models, respectively. n is a constant, indicating a better interaction between the adsorbent and the adsorbate when 
was 0–1. R is the molar gas constant (
), 
(K) is the absolute temperature and b is the isothermal constant. The Halsey model is formulated using Equation (7) (Ezzati 2020):
and 
are the equilibrium constants and exponents of the Halsey isothermal model.
), whereas it showed a good agreement with Freundlich and Temkin isotherm models over a wide range of sorption concentrations (
). This indicated that the adsorption of U(VI) ions by the strain ZFSY-01 was consistent with the inhomogeneous surface multilayer sorption under non-ideal conditions. Significantly, the fitting results of the Halsey isothermal model were also excellent (
), and this model was also a multilayer adsorption model. The Freundlich isotherm model, 
indicated that the strain was favorable to the adsorption of U(VI), and there was a strong interaction between sorbent and adsorbent. Isothermal model parameters for uranium adsorption by strain ZFSY-01
| Experimental value | Langmuir | Freundlich | Temkin | Halsey | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
 |  |  |  |  |  |  |  |  |  |  |  |  |
| 384.15 | 346.02 | 0.014–0.124 | 0.80 | 83.25 | 0.29 | 0.97 | 57.70 | 2.18 | 0.96 | 0.29 | 4.19 × 106 | 0.97 |
| Experimental value | Langmuir | Freundlich | Temkin | Halsey | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| | | | | | | | | | | | |
| 384.15 | 346.02 | 0.014–0.124 | 0.80 | 83.25 | 0.29 | 0.97 | 57.70 | 2.18 | 0.96 | 0.29 | 4.19 × 106 | 0.97 |
Isotherm model for uranium adsorption by strain ZFSY-01. (a) Langmuir isotherm model. (b) Freundlich isotherm model. (c) Temkin isotherm model. (d) Halsey isotherm model (initial uranium concentration: 200 mg/L, T: 30 °C, strain dosage: 400 mg/L, reaction time: 4 h, pH: 6.5).
Isotherm model for uranium adsorption by strain ZFSY-01. (a) Langmuir isotherm model. (b) Freundlich isotherm model. (c) Temkin isotherm model. (d) Halsey isotherm model (initial uranium concentration: 200 mg/L, T: 30 °C, strain dosage: 400 mg/L, reaction time: 4 h, pH: 6.5).
Kinetic and thermodynamic analysis of uranium removal by strain ZFSY-01
The 
(mg/g) and 
(mg/g) are the adsorption capacity at the moment of t (min) and the moment of equilibrium in the adsorption process, respectively. 
(min−1) is the pseudo-first-order kinetic adsorption rate constant and 
(g/mg·min) is the pseudo-second-order kinetic adsorption rate permanent. 
(mg/g⋅min1/2) is the model rate coefficient for intraparticle diffusion, and C (mg/g) is a constant in the boundary layer thickness function.
, 
) indicated that the experimental data are a good match for the pseudo-second-order kinetic model. Therefore, it was assumed that the process of uranium adsorption for this strain is mainly chemical adsorption. Meanwhile, the theoretical value 
mg U/g was estimated by the pseudo-secondary model, which was basically consistent with the experimental value. Interestingly, in most previous studies, the pseudo-secondary kinetic model showed a good correlation with a fit coefficient 
(Tuzen et al. 2020). In addition, the intraparticle diffusion model also showed good agreement (
). Considering the complexity of the adsorption mechanism, this model is also employed to describe the mechanism of adsorption. To sum up, in the first stage, U(VI) was adsorbed by bacteria through the diffusion of pores inside and outside the membrane. Subsequently, U(VI) could be immobilized on the surface or inside of the bacteria through chemical reactions with organic functional groups on the surface of the bacteria. Kinetic parameters of uranium adsorption by strain ZFSY-01
| Experimental value | Pseudo-first-order | Pseudo-second-order | Intraparticle diffusion | ||||||
|---|---|---|---|---|---|---|---|---|---|
 (mg/g) |  (mg/g) |  (/min) |  |  (mg/g) |  (/min) |  |  |  |  |
| 314.75 | 146.35 | 0.0243 | 0.79 | 301.20 | 0.00043 | 0.99 | 7.770 | 194.444 | 0.88 |
| Experimental value | Pseudo-first-order | Pseudo-second-order | Intraparticle diffusion | ||||||
|---|---|---|---|---|---|---|---|---|---|
| (mg/g) | (mg/g) | (/min) | | (mg/g) | (/min) | | | | |
| 314.75 | 146.35 | 0.0243 | 0.79 | 301.20 | 0.00043 | 0.99 | 7.770 | 194.444 | 0.88 |
Kinetic and thermodynamic models for the adsorption on uranium from strain ZFSY-01. (a) Pseudo-first-order kinetic model. (b) Pseudo-second-order kinetic model. (c) Intraparticle diffusion model (initial uranium concentration: 200 mg/L, strain dosage: 400 mg/L, T: 30 °C, pH: 6.5, V: 50 ml). (d) Thermodynamic parameter model (initial uranium concentration: 200 mg/L, V: 50 ml, T: 288, 298, 303, and 308 K, strain dosage: 400 mg/L, pH: 6.5).
Kinetic and thermodynamic models for the adsorption on uranium from strain ZFSY-01. (a) Pseudo-first-order kinetic model. (b) Pseudo-second-order kinetic model. (c) Intraparticle diffusion model (initial uranium concentration: 200 mg/L, strain dosage: 400 mg/L, T: 30 °C, pH: 6.5, V: 50 ml). (d) Thermodynamic parameter model (initial uranium concentration: 200 mg/L, V: 50 ml, T: 288, 298, 303, and 308 K, strain dosage: 400 mg/L, pH: 6.5).
) (kJ/mol), (
) (kJ/mol), and (
) (J/mol·K). The relationships between the above three parameters are represented by the following equations (Smjecanin et al. 2022):
and R are the balance constant as well as the constant of the gas, respectively. When 
, it meant that the reaction could proceed spontaneously, and vice versa. The reaction was endothermic as 
and exothermic as 
. If 
kJ/mol, it meant that the system reaction was a physical process. Moreover, if 
, it suggested an increase in the disorder of the adsorption system (Sahmoune 2019).The thermodynamic parameters obtained from the experiments are shown in Table 3 and Figure 6(d). The enthalpy change (
) demonstrated that the reaction process was physical and endothermic. Also, the entropy change (
) as well as Gibbs free energy (
) suggested that the U(VI) absorption by strain ZFSY-01 was an entropy-increasing process that could proceed spontaneously.
Thermodynamic characteristic parameters of uranium adsorption by strain ZFSY-01
| T (K) |  (kJ/mol) |  (J/(mol·K)) |  (kJ/mol) |
|---|---|---|---|
| 288 | −1.21 | 105.51 | 29.17 |
| 298 | −2.27 | ||
| 303 | −2.79 | ||
| 308 | −3.32 |
| T (K) | (kJ/mol) | (J/(mol·K)) | (kJ/mol) |
|---|---|---|---|
| 288 | −1.21 | 105.51 | 29.17 |
| 298 | −2.27 | ||
| 303 | −2.79 | ||
| 308 | −3.32 |
CONCLUSION
This research examined the bioabsorption of U(VI) by J. quinghaiensis strain ZFSY-01, which was isolated from the effluent of a uranium mine in northwest China. According to the aforementioned investigations, this strain was found to be highly tolerant to uranium, maintaining high cellular activity even at uranium concentrations of 300 mg/L and resulting in an adsorption capacity and efficiency of 391.5 mg U/g and 78.3%, respectively, under the ideal adsorption circumstances (initial uranium concentration: 200 mg/L, pH: 5, reaction time: 4 h, strain dosage: 400 mg/L, T: 30 °C). Significantly, it exhibited better adsorption properties compared to other biosorbent materials, especially since the adsorption capacity of this strain reached the maximum (
mg U/g) under 100 mg/L of strain dosage. The obtained data were consistent with the Freundlich isothermal adsorption model, the Temkin isothermal adsorption model, the Halsey isothermal adsorption model, and the pseudo-second-order kinetic model with regression coefficients 
or more. Meanwhile, thermodynamic parameters indicate that the adsorption of U(VI) by strain ZFSY-01 is a spontaneous endothermic process. As a result, it was postulated that the mechanism for U(VI) immobilization by this strain involved a combination of chemical and physical adsorption. Based on the above results, the strain ZFSY-01 can effectively remove U(VI) ions from high- or low-concentration uranium-containing wastewater and is a promising biosorbent. However, further research is needed to achieve industrial application.
ACKNOWLEDGEMENTS
This work was supported by the National Natural Science Foundation of China (No. 32060292). In addition, the authors would like to thank the anonymous reviewers for their valuable comments.
DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.
CONFLICT OF INTEREST
The authors declare there is no conflict.
%20cropped.png?versionId=7144)