Genetical Surface Display of Silicatein on Yarrowia lipolytica Confers Living and Renewable Biosilica-Yeast Hybrid Materials.

In this work, a biological engineering-based biosilica-yeast hybrid material was developed. It was obtained by the aggregation of Yarrowia lipolytica through biosilicification catalyzed using genetically displayed silicatein on its cell surface. With orthosilicate or seawater as the substrate, the silicatein-displayed yeast could aggregate into flocs with a flocculation efficiency of nearly 100%. The resulting floc was found to be a sheetlike biosilica-yeast hybrid material formed by the biosilica-mediated immobilization of yeast cells via cross-linking and embedding, turning the original hydrophilicity of yeast cells into hydrophobicity. In addition, this material was characterized to be porous with an average pore diameter of approximately 10 μm and porosity of over 70%. Because of these properties, this hybrid material could achieve enhanced removal efficiencies for chromium ions and n-hexadecane, which were both above 99%, as compared to the free cells of Y. lipolytica in aqueous environments. Importantly, this hybrid material could be recultivated to generate new batches of yeast cells that maintain parallel properties to the first generation so that the same hybrid material could be reproduced with unchanged highly efficient removal of chromium and n-hexadecane to those of the first generation, demonstrating that this biosilica-yeast hybrid material was living and renewable. This work presented a novel way of harnessing silicatein and Y. lipolytica to achieve biological synthesis of a living inorganic-organic hybrid material that has potential to be applied in water treatment.

Introduction

Although numerous materials and nanomaterials are flourishing in the modern world, it is still significant to highlight the prospects of biological engineered living materials (ELMs), proposed by Gilbert and Ellis in 2018, for its great potential in creating entirely new and useful biological materials as cutting-edge fields involving microbiology, material science, and synthetic biology and its advantages as eco-friendly materials.[1] However, the occurrence of ELMs is largely insufficient.

Presently, anthropogenic activities are causing increasingly heavy contamination of terrestrial and marine aqueous environments on Earth, which endangers the availability of vital water resources for human beings. Therefore, the removal of toxic industrial heavy metals and accidental petroleum spills has received intensive research attention. Modern bioremediation with micro-organisms, which includes bacteria, archaea, fungi, and yeast, is one of the most attractive and active research fields because it could be an alternative approach for the treatment of these pollutants in a cost-effective, safe, and ecofriendly way.[2−4] Environmentally friendly materials are emerging as another group of promising technologies for the effective treatment of environmental pollutants.[2,5,6] Materials/nanomaterials with a large surface area and mesoporous structures have been synthesized exhibiting considerably high capability for adsorbing heavy-metal ions;[7] superhydrophobic materials, also with a high surface area, high porosity, and nanostructured surface were also prepared for highly efficient oil adsorption.[6,8,9] Lately, a new field of combining material science/nanotechnologies with micro-organisms has emerged for the purpose of combining their respective advantages to produce hybrid materials and improve their capacity of removing heavy metals and oil from water.[8,10]

Yarrowia lipolytica is an oleaginous yeast species capable of growing in hydrophobic environments because of its unique physiological and metabolic features, such as the ability to utilize triglycerides, fatty acids, and hydrocarbons as the carbon sources.[11,12] Moreover, this yeast is able to produce various enzymes (such as proteases, lipases, and esterases) and natural products (such as emulsifiers and surfactants), which guarantees its growth in a large array of different conditions.[12,13] These features indicate great potential for Y. lipolytica in the bioremediation of environmental contamination by various pollutants, notably oils and hydrocarbons.[12,14] Moreover, it is documented that Y. lipolytica is an efficient biosorbent for the removal of heavy metals such as nickel, chromium, and silver ions.[12,15,16] This yeast can also tolerate low temperatures, high salt concentrations, and variable pH, which is significant because these features would enable its use for the bioremediation of seawater and the in situ treatment of organic contaminations such as oil spills.[11,12] Therefore, Y. lipolytica has great potential as a multifunctional bioremediation agent for the simultaneous treatment of heavy metals and hydrocarbons in aqueous environments. Inspired by the ELMs, we intend to make endeavors to create a living Y. lipolytica-based material which intersects the material science and synthetic biology to achieve and improve removal capacity of heavy metals and hydrocarbons.

Different from the abovementioned nanomaterials, the cells of Y. lipolytica must be kept intact to maintain their living status for metabolizing hydrocarbons; thus, cells could not be fabricated into mesoporous structures. However, it is feasible to cross-link whole cells to each other so that micropores could be formed. This process would lead to the flocculation of Y. lipolytica cells. On the other hand, the preparation of silica-coated yeast cells[17] has inspired us to make a coating for the surface of Y. lipolytica cells with certain hydrophobic inorganic substances, with which the preparation of nanomaterials for modifying yeast-cell surfaces could be avoided to simplify the process of surface hydrophobization.[8] In this context, biosilicification on the cell surfaces of micro-organisms could cause the deposition of hydrophobic silica[9] on the surfaces of yeast cells,[18] as well as induce the cross-link of yeast cells;[17] however, whether porous structures were formed during this process has not yet been specified. In these cases, silica was deposited on cell surfaces via chemical sol–gel reaction, which required the use of silane substrates. A unique family of enzymes mainly synthesized using marine sponges, namely, silicateins, are able to catalyze the polymerization of soluble marine orthosilicates into biosilica (biologically formed silica nanostructure) to form their spicules.[19] Moreover, silicateins are capable of catalyzing various other silicates and silanes into silica.[19,20] These suggest the possibility of achieving biosilicification mediated by silicateins with versatile substrates, especially, natural ones like seawater, as compared to that for sol–gel reactions. In a former report, silica encapsulation on the surfaces of microbial cells was catalyzed by surface-displayed silicatein α, which could cause simultaneous cell cross-linking and flocculation.[20]

Inspired by these studies, we displayed silicatein SilA1 from a marine sponge on the surface of Y. lipolytica cells through genetic engineering;[21,22] the resulting silicatein-displayed recombinant strain was used to react with organosilane and seawater as the substrates, expecting that the displayed SilA1 could catalyze biosilica synthesis on Y. lipolytica cells and simultaneously medicate their aggregation to obtain a new biosilica–yeast hybrid material that might be capable of enhanced heavy-metal adsorption and n-alkane degradation.

Results and Discussion

Surface Display of Silicatein on Y. lipolytica and Whole-Cell Catalytic Properties

The utilization of the pINA1317-YlCWP110 vector for the surface display of heterologous enzymes requires screening their activities, using different transformants as whole-cell catalysts to identify the transformant exhibiting the highest catalyzing ability.[21−23] Thus, 100 transformants obtained in this work, which had integrated the 6×His-SilA1-YlCWP110 fragment into their genomes, were screened to determine their silicatein activity as described above. As illustrated in Figure a, transformant strain S10 was distinguished as exhibiting the highest specific silicatein activity of 376.5 ± 3.8 U/mg dry cell weight (DCW) among all transformants under the unoptimized reaction conditions (see the Supporting Information), whereas negative control strain S0, with only YlCWP110 integrated into the genome, had silicatein-like activity of 14.2 ± 0.22 U/DCW, which was ascribed to hydrophobic interaction or unspecific catalysis between cell surfaces of the S0 strain and the (tetraethyl orthosilicate) TEOS-hydrolyzed orthosilicate substrate.[19] Upon selection of the S10 strain, cell growth and time-dependent changes of silicatein-specific activities of this whole-cell catalyst were investigated. This allowed the determination of optimal cultivation time when specific activity reaches its maximum, depending on the growth-time-dependent expression of the SilA1 gene, driven by the recombinant hp4d promoter in the pINA1317-YlCWP110 expression vector.[22,24]Figure S1 shows that the specific activity of S10 cells reached 377.2 ± 24.1 U/mg DCW at a cultivation time of 72 h and that it remained unchanged during subsequent cultivation; cell growth was stationary with an approximate cell density of 2.6 × 107 cells/mL at 72 h. Thus, this whole-cell catalyst, harvested after 72 h cultivation, was used in all subsequent assays.

Figure 1

(a) Specific silicatein activities of recombinant strain S10 and control strain S0; (b) flocculation appearance of S10 and S0 Y. lipolytica strains formed in TEOS hydrolysate as the substrate. (c,d) Immunofluorescence assay: observation of the S0 strain under white light and fluorescence microscope (excitation wavelength: 420–485 nm, emission wavelength: 515 nm), respectively. (e,f) Observation of the S10 strain under white light and the fluorescence microscope of the same condition, respectively.

The demonstration of silicatein activity using the S10 whole-cell catalyst not only indicated that SilA1 was successfully expressed in Y. lipolytica but also implied its effective display on the cell surface. To further demonstrate the surface display of SilA1, S10 cells were first subjected to treatment with proteinase K,[21] which led to a drastic decrease in specific activity to 25.70 ± 3.5 U/mg DCW (Figure S2), confirming that the recombinant silicatein was distributed on the cell surface.[21] Thereafter, immunofluorescence assay was performed using 6 × His monoclonal antibody as the primary antibody and IgG/fluorescein isothiocyanate as the secondary antibody. Results in Figure clearly showed that no fluorescence was observed on S0 cells that had only YlCWP110 integrated into the genome (Figure c,d); yeast cells of the S10 strain, on the other hand, whose genome harbors fusion gene 6His-SilA1-YlCWP110, could emit green fluorescence around the cell walls (Figure e,f). Thus, this immunofluorescence assay demonstrated that recombinant silicatein was indeed displayed on the cell surface of the S10 strain through the production of the 6×His-SilA1-YlCWP110 fusion protein anchored on the cell walls.[22,23] Previously, silicatein α was anchored on the cell surface of Escherichia coli by the expression of fusion of its gene with the outer membrane protein A gene. Here, silicatein was displayed for the first time on the cell surface of Y. lipolytica. The effect of surface-displayed SilA1 on the cell growth of the S10 strain was studied. When compared with original Y. lipolytica host strain Po1h, the S10 strain exhibited no significant difference in cell density at each tested time point during 120 h cultivation (Figure S3), indicating that the surface display of SilA1 had no side effects on the cell growth of the S10 strain.

For the purpose of fully understanding the catalytic conditions of whole yeast cell silicatein with TEOS hydrolysate, optimal reaction temperature, pH, and enzymatic stability against different temperatures, pH values, and metal ions were investigated. As shown in Figure S4a, the optimal reaction temperature for S10 cells was 30 °C, at which the relative silicatein activity was maximal. Moreover, surface-displayed SilA1 was relatively stable up to 40 °C, retaining over 80% of its initial activity. However, it was rapidly inactivated when the temperature was above 50 °C (Figure S4a), indicating that the surface-displayed silicatein did not have very strong thermostability. It was once acknowledged that the temperature optimum for silicatein was in the range of 20–25 °C and that the temperature coefficient decreased 2.5-fold above 25 °C.[25] Herein, the optimal temperature for this surface-displayed silicatein was 5–10 °C higher, and it did not significantly lose its activity around the optimal reaction temperature. This positive effect on thermostability could possibly be attributed to the chimeric expression of SilA1 and CWP110, which could increase the rigidity and the consequent stability of the fusion protein. Furthermore, as shown in Figures S4b and S10, cells exhibited the highest relative activity at pH 7.0 under the optimal temperature of 30 °C, both of which are common conditions that should not be difficult to apply in large-scale water-treatment industries. Normally, free or surface-displayed silicateins require a nearly neutral pH value to achieve enzyme-driven silica polycondensation.[25] Thus, the optimal reaction pH value of 7.0 for the S10 whole-cell catalyst is similar to those from previous studies. However, few studies have taken note of the pH stability of silicatein. In this study, we found that the S10 whole-cell catalyst stayed relatively stable within pH values ranging from 4 to 9 at 30 °C (Figure S4b), under which condition, its relative activity was maintained at over 70%. This pH stability profile reflects broad pH adaptability for the S10 whole-cell catalyst which may facilitate silica polycondensation catalyzed using surface-displayed silicatein in versatile aqueous environments.

Finally, the effect of various metal ions on the enzymatic activity of the S10 whole-cell catalyst was evaluated. Table S1 lists the effect of 17 kinds of metal ions on the activity of displayed silicatein. It could easily be observed that the majority of these ions were unable to inhibit the activity of the S10 catalyst, with the exception of Fe3+, which significantly reduced its activity to approximately 59% (p < 0.05); some metal ions, such as Mg2+, Mn2+, K+, Co2+, and Ni2+, even slightly promoted silicatein activity. In particular, we first determined in this study that Cr(III) and Cr(VI) ions did not have inhibiting effects on S10 cells, which guaranteed the possibility of using this whole-cell catalyst in the treatment of these toxic metal ions in water.

Under optimal temperature and pH value, S10 whole-cell silicatein exhibited 381.08 ± 2.4 U/mg of specific activity. Notably, visible aggregation in bulk could be clearly observed under this condition (Figure b), accompanied with a significant decrease of cell density in this solution (from 1 × 107 to approximately 5 × 105 cells/mL). No obvious bulk derived from S0 cells was observable after they were subjected to TEOS-hydrolyzed orthosilicate (Figure b), suggesting that unspecific catalysis using S0 could not result in yeast flocculation. Thereby, this phenomenon demonstrated that S10 cells were flocculated by reacting with the orthosilicate. The reason for this might be the biosilica cross-link during silicatein catalytic polycondensation or the embedding of biosilica in the yeast cells. Therefore, yeast flocs were collected and characterized to address this issue.

Yeast Flocculation Characterization

The reaction products derived from the S10 (S10p) and S0 (S0p) strains with the substrate of the TEOS hydrolysate were filtered for collection, freeze-dried, and subjected to powder X-ray diffraction (XRD) analysis. The XRD S10p pattern demonstrated that amorphous silica accounted for the predominant material, while minor crystalline silica species might have coexisted, deduced by the identification of a characteristic quartz peaks at 2θ = 20.89 and 2θ = 26.63 (Figure a), which was in close agreement with the XRD pattern of pure silica (JCPDS ICDD File Card # 00-001-0647), the commercial biosilica,[26] and biosilica isolated from marine sponges.[27] Moreover, the strong signal at 2θ = 31.08 might correspond to the XRD pattern of the natural silica-containing zeolite of gmelinite;[28] however, the reason for the formation of this crystalline structure required further investigations. In contrast, the product of the S0p control sample showed completely amorphous silica without any definite crystalline peak (Figure b). S10p and S0p were tested with elemental analysis (EA). Table shows that C, H, O, N, and Si elements contributed to the constitution of S10p and S0p, which suggested that S10p and S0p were inorganic–organic hybrid materials. Additionally, S10p contained a larger Si content than that of S0p, which could be attributed to the higher silicatein activity of S10 cells so that more biosilica could be synthesized for S10p. Thus, XRD and EA results demonstrated that biosilica was indeed produced in S10p.

Figure 2

XRD patterns of reaction products derived from (a) S10 (S10p) and (b) S0 (S0p) strains with the TEOS hydrolysate substrate (green dots, 2θ degrees of 20.89, 26.63, and 31.08); SEM images of S10p and S0p at (c,d) 2000× and (e,f) 6000× magnification.

Table 1

EA of Reaction Products Derived from S10 (S10p) and S0 (S0p) Strains with TEOS Hydrolysate

elements	S10p	S0p
Si (%)	17.34 ± 0.23	5.53 ± 0.12
C (%)	20.26 ± 0.17	16.80 ± 0.08
H (%)	4.90 ± 0.22	4.23 ± 0.3
O (%)	27.04 ± 0.09	25.63 ± 0.26
N (%)	4.33 ± 0.12	3.82 ± 0.22

To distinguish this structural difference on the microlevel, S10p and S0p were observed with scanning electron microscopy (SEM). SEM under 2000× magnification revealed large sheetlike morphology for S10p (Figure c), whereas much smaller and irregular structures were observed in the S0p control sample (Figure d). At 6000× magnification, massive S10 cells were observed, tightly cross-linked with each other (Figure e). These cells were embedded using biosilica so that thick layers formed (Figure a), but this embedding was not uniform, and the cells on the layer surface were not enclosed. Cross-linked cells could also be found in the S0p sample, but in a very loose morphology, and biosilica embedding was not observed in this sample (Figure f). An early study reported that a whole-cell catalyst composed of E. coli surface-displaying silicatein could mediate the synthesis of titanium phosphates from a Ti-BLADH substrate using a catalysis mechanism similar to biosilica polycondensation, leading to cross-linking and flocculation in bulk for E. coli cells.[29] Nevertheless, layered amorphous titanium phosphates were formed on the bacterial cell surfaces during this process, but whether E. coli cells were embedded by titanium phosphates was not identified.[29] In this work, the SEM observation directly revealed that the sheetlike aggregates in the bulk of S10p were composed of biosilica-embedded and cross-linked S10 yeast cells. Furthermore, S10p was calcined at 1000 °C for 2 h, and XRD and SEM were again used to characterize these calcinated products. SEM (Figure S5a) showed that S10p calcination caused its disintegration and transformation into tiny pieces. The XRD pattern (Figure S5b) of the calcinated S10 product also revealed that more crystalline silica was formed after calcination. Thereby, these results suggested that yeast cells in S10p were not individually encapsulated with biosilica but rather were cross-linked and embedded in mass, thereby leading to the collapse of S10p into small pieces of amorphous–crystalline hybrid silica. The reason for the failure of yeast-cell encapsulation was highly likely because displayed silicateins were not adjacent on the yeast surfaces, as manifested when much green fluorescence was scattered on the yeast-cell walls, distinguished by fluorescence microscopy (Figure ). Therefore, we propose that yeast-cell cross-linking and embedding were achieved by biosilica growth catalyzed using the surface-displayed silicatein on S10 cells. This process, with the collaboration of cell cross-linking and embedding, led to the immobilization of S10 cells, which was similar to the enzyme immobilization catalyzed using silaffin.[30]

SEM images (Figure c,d) showed that multiple pores were distributed on the S10p, and, at higher magnification, pores with an average diameter of 9.99 ± 1.66 μm could be observed. In addition, mercury intrusion porosimetry showed that S10p had high porosity of approximately 71.48% (Table S2). Pore size and porosity measurement indicated that the S10p biosilica–yeast hybrid was indeed porous with microscale pores all over it. However, the formation mechanism of these pores remained largely unclear, and it required further investigation. Herein, we proposed that because of the nonuniform display of silicatein, these pores were formed by the surrounding of randomly assembled yeast surfaces where silicateins were seldom displayed during the process of yeast flocculation.

After the structure of S10p had been specified, S10p hydrophobicity was taken into account and tested using water contact angle (WCA) assay. Figure a illustrates that the WCA value of S10p was 117.35°, whereas S0p had a WCA value of 61.53°, and WCA values for the unreacted yeast cells of S10 and S0 were even smaller, indicating that S10p had an excellent hydrophobic property.[31] Importantly, this result demonstrated that hydrophobic modification to the yeast cells could be achieved by the catalytic synthesis of biosilica mediated by surface-displayed silicatein, and a porous and hydrophobic hybrid material of biosilica-embedded Y. lipolytica was successfully obtained.

Figure 3

WCA for each reaction product derived from (a) S10 strain with TEOS hydrolysate, (b) pure S10 strain as the control, (c) reaction product derived from S0 strain with TEOS hydrolysate, and (d) pure S0 strain as a control.

Maximal Hybrid Material Production

Upon the successful preparation of the biosilica–yeast hybrid material, its producing conditions were optimized in terms of orthosilicate concentration, cell density, and reaction time, which was measured with the parameter of flocculation efficiency under previously identified optimal reaction conditions of 30 °C and pH 7.0. As shown in Figure a, with a cell density of 1 × 107 cells/mL, S10 flocculation efficiency kept increasing with the increase of orthosilicate concentration from 0 to 10 mol/mL; flocculation efficiency reached 95.3 ± 1.9% and stopped increasing when the orthosilicate level was over 10 mmol/mL, which appeared to be the optimal concentration. Furthermore, with this orthosilicate concentration, the use of different S10 cell densities (1 × 104, 1 × 105, and 1 × 106 cells/mL) appeared to result in lower flocculation efficiency than that for cell density at 1 × 107 cells/mL, and flocculation efficiency also dropped significantly, to 17.1 ± 4.7%, when using 1 × 108 cells/mL (Figure b). This indicated that 1 × 107 cells/mL was the optimal cell density. Using these optimal flocculation conditions, reaction time was optimized. The result, illustrated in Figure c, indicated that flocculation efficiency reached its maximum after incubating with substrates for 30 min. Therefore, the optimal flocculation conditions were specified to be 30 °C, pH 7.0, cell density of 1 × 107 cells/mL, orthosilicate concentration of 10 mmol/mL, and reaction time of 30 min, under which flocculation efficiency was as high as 99.93 ± 2.1%.

Figure 4

Effects of (a) orthosilicate concentration, (b) reaction time, and (c) cell density on flocculation efficiency of S10 cells in TEOS hydrolysate, and (d) effect of reaction time on flocculation efficiency of S10 cells in pure seawater.

The orthosilicate substrate used in the above assays was obtained from TEOS hydrolysates. However, seawater appeared to be another convenient orthosilicate source, which is utilized by marine sponges to synthesize their spicules.[32] Significantly, the use of seawater as a natural substrate to produce the biosilica–yeast hybrid material catalyzed by the silicatein surface-displayed S10 cells truly utilized the real subjects in the bioprocess of silica-based skeletons of marine sponges[33] as a biomimetic and bioinspired issue for designing innovative materials,[34] thus making the preparation of this hybrid material with pure seawater of particular significance. In the beginning, no floc could be observed after S10 cells were subjected to seawater. However, tiny floating flocculant particles began to be visible with an incubation time of over 1 h (Figure S6). On the basis of this phenomenon, time-dependent yeast flocculation in seawater was further investigated. As seen in Figure d, yeast-flocculation efficiency reached 19.2 ± 1.2% after 1 h of incubation and increased rapidly to approximately 65.3 ± 3.4% after 2 h. After that time, flocculation efficiency maintained a very slow increase rate until it reached approximately 95.5 ± 3.4% after 4 h incubation (Figure d). For the S0 strain in seawater, yeast flocculation was not always obvious. Subsequently, the reacted products using seawater as the substrate catalyzed using the S10 (S10sp) and S0 (S0sp) strains were subjected to the XRD, SEM, and WCA analyses in order to verify that the use of the seawater substrate could also contribute to the formation of biosilica–yeast hybrid material like that derived from TEOS hydrolysates. As revealed by SEM in Figure S7a, S10sp had sheetlike morphology. Furthermore, S10sp was porous with an average pore diameter of 9.90 ± 0.91 μm (Figure S7a) and porosity of 70.41% (Table S2), which were consistent with reaction products of S10p derived from the TEOS hydrolysate. In addition, the XRD pattern of S10sp also demonstrated a predominantly amorphous profile for this sample, along with silica crystalline peaks found at θ degrees of 20.95, 26.1, and 50.83 (Figure S7b), which was in agreement with the XRD biosilica pattern from a marine sponge.[27] WCA analysis showed that S10sp also possessed a similar hydrophobic property (Figure S8) to that of S10p. Therefore, these analyses verified that S10 cells could catalyze biosilica formation with the natural substrate of seawater like that used with orthosilicates generated from hydrolyzing TEOS. The bulk was composed of biosilica-embedded yeast cells that could also be produced during this process to produce porous and hydrophobic biosilica–yeast hybrid material similar to S10p. Despite that it would take much longer for S10 to flocculate in seawater than in TEOS hydrolysates, seawater could indeed be used to induce the formation of the biosilica–yeast hybrid material via the surface-displayed silicatein. This slow biosilicification process was likely due to the extremely low concentration of natural orthosilicate at the micromolar level in seawater.[35] Nevertheless, the abundance of seawater on the Earth would guarantee that it could serve as a cost-effective source to be used to prepare this hybrid material.

Enhanced Chromium Removal from Water with Biosilica–Yeast Hybrid Material

At present, remediating chromium-contaminated water by adsorption is a universal problem that has not been solved.[36] In this work, following all characterizations specified above, we evaluated the effects of S10p (obtained from the reaction of S10 cells with the substrate of TEOS hydrolysate) and S10sp (obtained from the reaction of S10 cells with the substrate of pure seawater) hybrid material on the removal of highly toxic chromium ions[37,38] from water. This was implemented by determining the removal efficiency of Cr(III) and Cr(VI) ions, which are the prevailing toxic forms of Cr in natural environments,[39] from water by S10p and S10sp, as well as by nonflocculated S10 cells as the control group. According to results presented in Figure a, the removal efficiency of 100 mg/L Cr(III) ions (Cr3+) by both S10p and S10sp and free cells kept ascending with the increase of their amounts in the reaction solution. At a concentration of 1.3 g/L, Cr(III) removal efficiency for each sample reached the maximum, which was 96.7 ± 1.3% for S10p, 96.0 ± 1.5% for S10sp, and 93.1 ± 1.7% for free S10 cells. Next, the time-dependent profiles of Cr(III) removal efficiency for all samples are shown in Figure b. Results demonstrated that the removal efficiency of Cr3+ by S10p and S10sp further increased to 99.5 ± 1.1 and 99.3 ± 2.3%, respectively, after incubation with 100 mg/L of Cr(III) for 150 min, whereas the Cr(III) removal efficiency of S10 free cells was significantly lower at that timepoint (P < 0.05). In addition, incubation for longer than 150 min did not lead to any further increase of Cr(III) removal efficiency for any sample. For Cr(VI) ions (Cr2O72–), which are 100 times more toxic and 1000 times more mutagenic than Cr(III),[40] similar results were obtained concerning the removal efficiency of all four samples. Briefly, with the same Cr(VI) ion initial concentration of 100 mg/L, the use of 1.3 g/L S10p and S10sp achieved optimal removal efficiency, which was 97.3 ± 1.6 and 96.8 ± 2.1%, within 30 min, whereas free S10 cells only exhibited a removal efficiency of 89.1 ± 1.2% for Cr(VI) (Figure c). Consistently, an incubation duration of 150 min resulted in maximal Cr(VI) removal efficiency of 99.7 ± 1.2 and 99.8 ± 1.1% by S10p and S10sp (Figure d), respectively, which was higher than the Cr(VI) removal efficiency of approximately 90% by the free cells of Y. lipolytica isolates under similar conditions.[41] Although a similar removal efficiency of 99.66% could also be obtained using the free cells of Saccharomyces cerevisiae, this required the chemical and thermal pretreatment of this yeast.[42] In contrast, the hybrid material in this study had the advantageous ability of removing Cr(III) and Cr(VI) pollutants from water with high efficiency and without special pretreatment, making it convenient to operate in possible applications. In addition, the adsorption capacities for Cr(III) by S10p and S0sp (in dry weight) were 76.54 ± 0.77 mg/g and 76.38 ± 1.77 mg/g, while the values for Cr(IV) were 76.69 ± 0.92 mg/g and 76.77 ± 0.84 mg/g. These indicated that the Cr(VI) adsorption capacities of as-prepared hybrid materials were much stronger than the heat-treated yeast cells (7 mg/g) but were obviously weaker than Y. lipolytica cells modified with Fe0/Fe3O4 nanoparticles[43] and many other nanomaterials.[44] However, the preparation of these hybrid materials appeared to be simpler, underlining its potential practicability over the other materials. The adsorption capacity of Cr(III) was rarely reported for the biosorbents,[2] except for the hybrid materials in this work, although weaker than that of phosphate mine (97.23 mg/g).[45] Nonetheless, Cr(III) was also harmful heavy metal ions that need to be removed.[2] Significantly, this high removal efficiency using these two types of hybrid material enabled concentrations of unabsorbed Cr(III) and Cr(VI) ions left in the solutions to be as low as 0.5 mg/L, which met the general standard of the upper limit of total Cr into inland surface water.[2]

Figure 5

Optimization of removal efficiencies for Cr(III) (a,b) and Cr(VI) (c,d) against different conditions of material concentration and reaction time using TEOS hydrolysate-derived hybrid material (S10p), pure-seawater-derived hybrid material (S10sp), and S10 free cells. *, data had significant difference.

From the above data, it can be implied that silica embedding did not interrupt but instead enhanced the original adsorbing abilities of Y. lipolytica for Cr(III) and Cr(VI). Previously, the adsorption mechanism for heavy-metal ions using Yarrowia spp. was ascribed to the binding effects between these ions and reactive moieties present on the cell surface, such as carboxyl, hydroxyl, and amino groups.[15,16,41] Therefore, it could be postulated that the adsorption of Cr(III) and Cr(VI) ions using Y. lipolytica strains used in this work was also achieved through this mechanism. Moreover, it is well known that mesoporous silica nanoparticles have an adsorbing ability, although limited, for heavy-metal ions because of their characteristics.[46] However, the biosilica presented in the hybrid materials here was characterized to have no such mesoporous structure. There was limited iron incorporation in the diatom biosilica, and more than 95% of biosilica-attached iron was in the form of iron clusters.[47] Nonetheless, another study suggested that the silica formed through silicatein catalysis was in the structure of trisiloxane rings and higher-membered siloxane rings,[32] but whether these siloxane rings could incorporate Cr(III) and Cr(VI) ions was yet to be specified. Therefore, it still requires extended studies to unravel the mechanism for enhanced removal efficiency of Cr(III) and Cr(VI) ions that emerge after silica embedding.

Simultaneous Enhanced Removal Efficiency of n-Hexadecane from Water using Hybrid Material

Accidental petroleum spills, especially in oceans, have already caused disasters to ecosystems that directly or indirectly endangered many life forms.[4] Bioremediation using micro-organisms is becoming a major approach for tackling such petroleum contaminations. The well-described capability of Y. lipolytica for metabolizing and degrading hydrocarbons has made this yeast a promising candidate for use in the bioremediation of petroleum spills.[4,12] S10p and S10sp were tested for their removal efficiency of n-alkane in water, with n-hexadecane as a representative n-alkane pollutant. Quantification with gas chromatography (Figure S9 and Table S3) indicated that the use of both S10p and S10sp could better consume n-hexadecane, as reflected by a removal efficiency of 85.48 ± 0.28 and 82.60 ± 0.66% (p > 0.05) within 96 h for S10p and S10sp (Figure a), respectively, under the initial conditions (see the Supporting Information), whereas free S10 cells had a lower n-hexadecane removal efficiency of 45.66 ± 0.30% (Figure a). The removal efficiencies for n-hexadecane reached 99.63 ± 1.12 and 99.73 ± 0.72% after the optimization of reaction conditions, including the concentration of each material and reaction time (Figure S10a,b). As previously shown, S10p and S10sp had porous structures and were highly hydrophobic. Thus, the enhanced n-hexadecane removal ability was due to the large surface area of the hybrid materials and the hydrophobic surface interaction between biosilica and n-hexadecane,[48] which improved the adsorption of n-hexadecane for the biosilica–yeast hybrid material. The similar n-hexadecane removal abilities of S10p and S10sp implied that free S10 cells could be directly applied on alkane-contaminated sites in the sea to allow in situ formation of the as-mentioned hybrid material and simultaneous n-alkane degradation. To testify this assumption, free S10 cells were directly added to seawater containing 1% (v/v) n-hexadecane to simulate the proposed simultaneous flocculation and alkane-degradation process. Results, shown in Figure b, demonstrated that approximately 1.3 g/L of S10sp could indeed be obtained within 4 h. At the same time, n-hexadecane degradation could be determined along with the input of S10 cells, and it reached over 99% after 96 h cultivation. Thus, this newly discovered property suggests that free cells of Y. lipolytica with surface-displayed silicatein could be launched into petroleum-spill spots to achieve better degradation of C10–C16 n-alkanes.[4] Nevertheless, this technique obviously requires further exploration to develop specific details for future use.

Figure 6

(a) n-Hexadecane removal efficiency of TEOS hydrolysate-derived hybrid material (S10p), pure-seawater-derived hybrid material (S10sp), and S10 free cells. (b) Simultaneous characterization of flocculation efficiency and n-hexadecane removal efficiency after launch of S10 free cells into seawater containing 1% (v/v) n-hexadecane.

Sustainability of Biosilica–Yeast Hybrid Materials

According to the characterization of the hybrid material from this work, we proposed that the individual yeast cell was not encapsulated using biosilica. This suggested that the yeast cell inside the material might be able to reproduce when conditions are appropriate, thus allowing the reproduction of new batches of yeast cells and the potential sustainability of this hybrid material. To evaluate this issue, the S10p and S10sp hybrid materials were filtered to collect after the treatment of Cr ions or n-hexadecane, followed by re-inoculation in a PPB medium for a second round of cultivation for 72 h. Then, cell density, silicatein activity, flocculation efficiency, Cr ions, and n-hexadecane removal efficiency were tested with this second batch of yeast cells. As illustrated in Figures a and S11a, the second-batch yeast cells exhibited close values of cell density and silicatein activity to those of the first generation. This unchanged silicatein, which was ascribed to the stable integration of the SilA1 gene into the genome of Y. lipolytica Po1h using the pINA1317-YLCWP110 expression system,[22,24] contributed to the same flocculation efficiency of the second-generation yeast cells to the first generation of S10 cells, so that a second batch of hybrid material could also be obtained by reacting these re-produced yeast cells with TEOS hydrolysates and seawater (Figures b and S11b). The re-produced hybrid material also had unchanged Cr ions (Figures c and S11c) and n-hexadecane removal efficiency to those from the first batch (Figures d and S11d). Furthermore, this process was repeated over a long period, and all these properties were tested on the 10th, 25th, and 50th batch. Figures a–d and S11a–d show that the hybrid materials in this work could be re-cultivated to produce new silicatein-displayed yeast cells, and these cells could be used to re-prepare another batch of hybrid material without any degeneration for at least 5 months from the aspects of silicatein activity, Cr ions, and n-hexadecane treatment, indicating the living and renewable properties of this work’s biosilica–yeast hybrid material and highlighting its advantages of sustainability and cost-effectiveness over the abovementioned hybrid nanomaterials as the agents for the treatment of chromium ions and hydrocarbons in water.

Figure 7

Determination of cell density and whole-cell (a) silicatein activity and (b) flocculation efficiency of each recultivated generation from S10 vs those of S10; (c) Cr(III) and Cr(VI) ion and (d) n-hexadecane removal efficiencies of the hybrid material derived from each recultivated generation from S10 with TEOS hydrolysis as the substrate vs those of S10. (a,b) Data with the same mark had no significant difference.

Conclusions

Enlightened by the ELMs, we first showed that a novel living and renewable biosilica–yeast hybrid material can be obtained by catalytic biosilicification with the simple assistance of surface-displayed marine-sponge-derived silicatein on Y. lipolytica. Biosilica cross-linking and embedding to the yeast cells contributed to their hydrophobic modification and led to the porosity of the hybrid material. Significantly, this hybrid material could also be prepared with seawater as the source of orthosilicate substrates. This was a novel bioinspired case for material fabrication by harnessing silicatein from a marine sponge to react with its natural substrate. Benefiting from hydrophobicity and the porous structure, this hybrid material had enhanced capabilities for treating chromium ions and n-hexadecane, achieving almost 100% removal efficiency for both pollutants. It could be reused as the seed for a second cultivation of new yeast cells, and the resulting new batches of hybrid materials had the same removal efficiency for chromium and degradation capacity for n-hexadecane, demonstrating the cost-effectiveness and sustainability of the biosilica–Y. lipolytica hybrid material for practical use in water treatment. These interesting properties of this hybrid material emphasized the potential of applying it in water treatment as a cost-effective, scalable, and efficient agent to detoxify chromium ions in fresh water and to degrade alkanes in seawater, where petroleum spills occur.

Experimental Section

Construction of a Y. Lipolytica Strain with Surface-Displayed Silicatein

Y. lipolytica Po1h[24] was used as the host strain for genetic manipulations aiming at the surface display of silicatein. The gene-encoding silicatein used in this work was derived from the SilA1 gene of the Latrunculia oparinae(49) sponge. The coding sequence of the SilA1 gene, which was optimized according to the codon usage bias of Y. lipolytica, was synthesized by Nanjing Jinsui Co. Ltd. (Nanjing, China), with an added histidine tag (6His), and was cloned into a pUC57 vector with added SfiI (at 5′ terminus) and BamHI (at 3′ terminus) restriction sites. Expression vector pINA1317-YlCWP110[21] was used for constructing a recombinant plasmid for surface display of SilA1 protein pINA1317-6His-SilA1-YlCWP110 by inserting the SilA1 gene between the SfiI and BamHI cloning sites. The pINA1317-6His-SilA1-YlCWP110 plasmid was linearized by NotI digestion. The transformation of Y. lipolytica Po1h, and the selection and verification of the transformant strains were carried out according to our previous studies.[21,22] As the negative control, empty plasmid pINA1317-YlCWP110 was also linearized and used to transform Po1h, generating control strain S0. After obtaining these recombinant strains, immunofluorescence assay was performed to verify the surface display of silicatein as previously described.[21,22]

Catalysis Optimization using Whole-Cell Silicatein

To determine the specific enzymatic activity of recombinant strains with surface-displayed silicatein, 100 different recombinant strains were inoculated in a PPB medium[21] and cultivated for 96 h. The resulting cultures were harvested by centrifugation at 5000g for 5 min, followed by washing twice with Tris-HCl buffer (20 mM, pH 7.25) and resuspension in 1 mL of Tris-HCl buffer. The reaction system was constructed according to the literature,[50] with modifications. In detail, 1 × 107 cells/mL of each strain was reacted with 10 mM of TEOS hydrolysates (obtained by adding 250 mM TEOS to 50 mM HCl and incubating for 30 min at room temperature) in 50 mL of Tris-HCL buffer (20 mM, pH 7.5) at 28 °C for 20 min. Silica quantification was implemented using a silicon molybdenum blue colorimetric method.[51] Specific silicatein activity was defined as the weight of recombinant yeast (DCW) required to produce 1.0 μM of silica per minute, as in a previous report, with some modifications.[19] The strain with the highest specific silicatein activity was named S10. To determine the optimal reaction temperature, 1 × 107 cells/mL of S10 was incubated with 10 mM TEOS hydrolysates in Tris-HCL buffer (20 mM, pH 7.5) for 20 min at different temperatures ranging from 20 to 80 °C, whereas thermostability was implemented with this method by changing incubation time to 24 h. The optimal pH was measured by varying buffer pH from 2.0 to 10.0 and incubating 1 × 107 cells/mL S10 with TEOS hydrolysates at an optimal temperature for 20 min; pH stability was obtained by switching the time to 24 h. Residual activity was determined using the above assay conditions, and relative activity was calculated as the ratio of residual activity to the activity without incubation. The effects of different metal ions on the whole-cell S10 catalyst were assessed following the procedures described in a former study[21] with Zn2+, Mg2+, Ca2+, Na+, Hg2+, Cu2+, Mn2+, Fe3+, Fe2+, Ba2+, K+, Co2+, Ag+, Ni2+, Cr2O72–, and Cr3+. Cell density was measured with a spectrophotometer at an absorption wavelength of 600 nm.

Optimization of Flocculation Conditions

The aggregation abilities of different strains were measured by flocculation efficiency. Initially, 1 × 107 cells/mL of each strain was reacted with 10 mM TEOS hydrolysates in 50 mL of Tris-HCL buffer (20 mM, pH 7.0) at 30 °C for 20 min. Flocculation efficiency was calculated using the following equation: flocculation efficiency = (1 – A/B) × 100%, where A was OD600nm of S10 free cells left in solution without flocculating and B was OD600nm of S10 free cells before reacting with substrates.[52] The apparent flocculation efficiency of the S10 control strain was also determined with the same procedures. To optimize flocculation efficiency, the effect of the concentration of S10 cells was first studied by setting it as 1 × 104, 1 × 105, 1 × 106, 1 × 107, or 1 × 108 cells/mL, and flocculation efficiency was determined as described above. The effect of orthosilicate concentration was investigated by setting its ramp range between 2 and 14 mM with the optimal cell concentration. The reaction time was optimized by measuring flocculation efficiency under various flocculating durations of 0, 5, 15, 20, 25, 30, 35, 40, or 45 min with optimal cell and orthosilicate concentrations. The floc generated using seawater as the substrates were obtained by incubating 1 × 107 of S10 or S0 cells in filtered seawater at 30 °C for 1, 2, 3, 4, 5, or 6 h.

Obtained Floc Characterization

A SEM image of the freeze-dried sample was taken on a Hitachi S-4800 microscope, as described previously.[53] XRD sample characterizations were performed on a Bruker D8 ADVANCE diffractometer. Diffraction patterns were obtained at diffraction angles between 10 and 60° at room temperature. EA for silicium in each sample was tested with Agilent 7700× inductively coupled plasma–mass spectrometry (Agilent Technologies, USA) after dissolving with HF. EA for carbon (C), hydrogen (H), nitrogen (N), and oxygen (O) was implemented using the Elemental Analyzer EA3000 (Euro Vector, Italy). The hydrophilicity or hydrophobicity of each sample was evaluated by surface-contact-angle measurement between thesessile water drop and sample surface. WCA measurement was performed using Attension Theta Lite (Biolin Scientific, Finland). A drop of water (20 μL) was dropped over the sample using an automatic microsyringe, and then, static images for each surface were taken. Mercury intrusion porosimetry was performed with PoreMaster 60GT (Quantachrome Instruments, USA), following an earlier report.[54]

Removal Efficiency of Cr(III) and Cr(VI) Ions

Flocs or free yeast cells were collected by centrifugation at 5000g for 5 min, freeze-dried, and weighed. For adsorbing chromium ions, 1 g/L of each sample was incubated with 50 mL of 100 mg/L Cr3+(Cr(III)) or Cr2O72–(Cr(VI)) ion solution with an agitation rate of 130 rpm at 28 °C and pH 7.0 for 30 min.[42,55] Then, each reacted group was centrifuged, and the concentration of Cr in the supernatant of Cr(III) adsorbing group was determined using the flame atomic absorption spectrometry method[56] with Sanvant AA apparatus (GBC, Australia). The Cr(VI) concentration was spectrophotometrically quantified using the diphenyl carbazide method.[41] The groups without adding freeze-dried flocs were used as controls for testing the removal efficiency of Cr(III) and Cr(VI). The removal efficiency of Cr(III) and Cr(VI) ions was calculated using the following formula: removal efficiency = [(initial ion concentration – residual ion concentration)/initial ion concentration] × 100%. Subsequently, the dosage effects of both flocs described above were studied by varying final floc concentration as 0.13, 0.26, 0.39, 0.52, 0.65, 0.78, 0.91, 1.04, 1.17, 1.3, 1.43, and 1.56 g/L, respectively. With optimal floc concentration, the effect of incubation time was investigated by setting the time at different intervals ranging from 0 to 210 min.

Removal Efficiency Determination of n-Hexadecane

The utilization of n-hexadecane was measured during the cultivation of flocs or yeast cells with n-hexadecane as the only carbon source. In brief, 1 g/L of each freeze-dried sample was incubated with 50 mL (in 250 mL shaker flasks) of modified PPB medium containing 1% n-hexadecane (v/v) as the only carbon source and cultivated with an agitation rate of 180 rpm at 28 °C. For each culture, 200 μL volume was sampled and centrifuged at 5000g for 5 min at each time interval of 8 h over a period of 120 h. The residual n-hexadecane was extracted with n-hexane and quantified by gas chromatography[57] with n-dodecane as the internal standard at the same time point as mentioned above. The removal efficiency of n-hexadecane was defined as [(initial n-hexadecane – residual n-hexadecane)/initial n-hexadecane] × 100%. The effects of sample concentration and incubation time on the removal efficiency of n-hexadecane were investigated with procedures similar to those used for the chromium ions. The concentration of each sample was set as 0.13, 0.26, 0.39, 0.52, 0.65, 0.78, 0.91, 1.04, 1.17, 1.3, 1.43, and 1.56 g/L; time interval was 8 h, and total duration was 120 h.

Statistical Analysis

Statistical analyses were performed using the Design Expert 7 (Stat-Ease, Minneapolis, MN, USA) statistical package. Data were presented as the mean ± standard deviation (SD) for the parametric data (n = 3). ANOVA and a comparison of the means were conducted using the multiple-range comparison least significant difference test. A probability value of p < 0.05 was considered significant.