Introduction

Nanotechnology is an interdisciplinary field at the forefront of scientific research, encompassing chemistry, fundamental physics, biology, medicine, and material science1,2,3. The concept of nanoscience was first presented by Nobel prize-winning physicist Richard Feynman to construct different materials at the nanoscale level in 19593,4. Nanomaterials comprise various categories of materials if those have at least one dimension smaller than 100 nm3,4,5. Among them, magnetic nanoparticles (MNPs) are an intriguing group of nanoparticles that have been considerably utilized in diverse fields because of their magnetic properties such as targeted drug delivery, sensing technologies, magnetic labeling, engineering, magnetic separation, and environment applications6,7,8,9,10,11,12. Among MNPs, iron oxide nanoparticles (IONPs) are promising nanoparticles due to their superior biocompatibility for biomedical applications such as magnetic hyperthermia6,13,14. The comprehensive investigation efforts are underway to commercialize IONPs for progressive biomedical applications15. The superparamagnetic magnetite (Fe3O4) is the most common iron oxide compared to α-Fe2O3 (hematite) and γ-Fe2O3 (maghemite) due to its superior magnetic properties, higher biocompatibility, and lower cytotoxicity6,16. Several techniques for preparing nanoparticles have been described, including physical, chemical, and biological methods3. The biological synthesis of nanoparticles is being performed by various macro and microscopic organisms such as bacteria, fungi, viruses, yeast, and plant extracts3,17. The biosynthesis of nanoparticles has received considerable attention in recent years due to their environment-friendly behavior, lower toxicity, more biocompatibility, and more potential for pharmaceutical and biomedical applications than conventional synthesis including chemical and physical methods18,19. Biogenic nanoparticles are remarkably stable and can be made mono-dispersed by changing various factors such as temperature, shaking speed, pH, and incubation duration20. Furthermore, magnetic nanoparticles synthesized by biological methods are pharmacologically better than those produced with physical or chemical techniques21. However, microbial-assisted Synthesis is slow and needs long stages such as culturing, nanoparticle extraction, and purification20. The efficacy of biosynthesis techniques is yet limited to the laboratory step because biosynthesis methods are not still capable of producing large-scale homogeneous nanoparticles19. Thus, there is a critical necessity for stable, efficient, scalable, and eco-friendly fabrication procedures19. Although there have been significant advancements in the fabrication of nanoparticles using various methods such as hydrothermal reactions, laser pyrolysis, co-precipitation, sol–gel method, electrochemistry, flow injection syntheses, as summarized by Laurent et al13. developing an ideal synthetic method for IONP still poses a challenge. This optimal synthesis should meet certain criteria, such as high production yield and avoiding the use of harmful chemicals22. Design of experiments (DoE) is described as a mathematical model that can be employed to design and investigate experiments23. In addition, this method significantly reduces the number of experiments required to screen a large number of variables (more than three variables) simultaneously24. In a DoE experiment, the various factors are discretized into a collection of values that are presented as levels. These levels experiment in various assortments, and a model predicting the response of the system is created based on input data and data collected via multiple iterations of investigation25. Fractional Factorial Design (FFD) supplies details on potential high-order interactions, something that is difficult utilizing a full factorial approach – measuring one factor at a time24. Another advantage of FFD over full factorial is fewer experimental runs due to decreasing the time needed to obtain data and reducing experiment costs24,26. In this study, we carried on our previous microbial synthesis approach using the thermophilic strain ASFS1‏, producing intracellular superparamagnetic Fe3O4 nanoparticles. Utilizing fractional factorial design (FFD), as a collection of mathematical and statistical techniques, we determined the optimal medium composition and ideal productive conditions for cultivating the strain ASFS. Additionally, we employed the dry weight of biomass and iron in the cell plate to monitor the corresponding biomineralization process.

Material and method

Materials and reagents

Tryptone, ferrous sulfate (FeSO4), yeast extract, NaCl, ethylenediaminetetraacetic acid (EDTA), and ferric chloride (FeCl3) were provided by Merck Chemicals (Darmstadt, Germany).

Bacterial strain

Bacillus tequilensis strain ASFS1 (GenBank accession number is MZ669973.1), which magnetic behavior and synthesis of superparamagnetic Fe3O4 nanoparticles had proven by Satarzadeh et al27. in 2024. This strain was used in this study to optimize the growth of bacteria and the iron uptake.

Biosynthesis of superparamagnetic Fe3O4 nanoparticles

To produce magnetite nanoparticles by strain ASFS1, 2 mL of fresh inoculum (equal to 0.5 McFarland) was added to 500 mL Erlenmeyer flasks containing 200 mL of Luria–Bertani medium (containing yeast extract 0.5 g, tryptone 1 g, and NaCl 1 g in 100 ml distilled water) supplemented with FeSO4 (200 μg mL−1) and FeCl3 (400 μg mL−1) and incubated at 37 °C and shaken for 72 h (120 rpm) and then 24 h not shaken according to previous report. After incubation, the bacterial cells were removed from the culture medium by centrifugation (12,800 g for 5 min). The cell pellets were transmitted to a mortar and disrupted by freeze–thaw cycles (5 cycles). Magnetic nanoparticles were separated using a magnetic and reproducible method previously described. In summary, magnetite nanoparticles were separated from disrupted cells using passing via a magnetic separation with a syringe placed near a neodymium-iron-boron magnet and set on solid phase extraction (SPE) Manifolds. Subsequently, the syringe was washed with 1.5 M Tris/HCl buffer (pH 8.3), 1% SDS, and deionized water respectively. Finally, the magnet was separated from the syringe to elute magnetite nanoparticles with deionized water. The obtained magnetite nanoparticles were studied by field-emission scanning electron microscopy (FESEM, Sigma VP model, ZEISS, Germany). This study was conducted to optimize the components of the culture medium to increase the number of cells and subsequently increase the amount of magnetite nanoparticles. The amount of magnetite nanoparticles was analyzed using an indirect method as previously described28. In summary, one ml aliquots (96 h incubated) were centrifuged (9600 g for 5 min) and the supernatant was discarded. The cell pellets were washed with sterile 0.9% NaCl solution mixed with 69% nitric acid (1 ml) and incubated at 98 °C for 2 h. The amount of iron in the cell pellets were measured using an atomic absorption spectrometer (AA670, Shimadzu, Japan). To measure the biomass dry weight, one ml of the samples was centrifuged and then dried and weighed.

Experimental design

A 25−1 FFD, was employed to screen the most effective variables in superparamagnetic Fe3O4 nanoparticles production by strain ASFS1 and for developed a model for optimal conditions. Luria–Bertani medium was used as the basal medium, and then FeSO4, FeCl3, and ethylenediaminetetraacetate disodium salt dihydrate (Na2-EDTA) were added to the basal media. The parameters were selected yeast amount (X1), tryptone amount (X2), FeSO4 concentration (X3), Na2-EDTA concentration (X4), and FeCl3 concentration (X5). The range and the levels of the factors showed in Table 1.

Table1 Utilized Variables for the experimental domain.
Full size table

A 25−1 FFD with three center points consisting of 19 factorial runs were designed by Design Expert version 6.0.4 statistical software (Table 2). The coding of the factors was accomplished according to the following equation:

Table 2 Experimental matrix for the 25−1 FFD and responses for iron uptake and growth.
Full size table
$${x}_{i}=\frac{{X}_{i}-{X}_{i;0}}{\Delta \text{X}} i=\text{1,2},3,\dots ., k$$
(1)

In this study, \({x}_{i}\) refers to the coded value of an independent variable, \({X}_{i}\) represents the real value of the independent variable, \({X}_{i;0}\) denotes the real value of the independent variable at the center point, and \(\Delta {X}_{i}\) represents the step change value. The dry weight of biomass (one ml of bacterial culture) and iron content in biomass (one ml of bacterial culture) were measured as the dependent variable or response. The runs were conducted randomly in an attempt to minimize the influence of extraneous variables. The selected factors can be linked to the response through a model, which is presented in the following equation:

$$Y={b}_{0}+{\sum }_{i=1}^{n}{b}_{i}{x}_{i}+{\sum }_{i=1}^{n}{b}_{ii}{x}_{i}^{2}+{\sum }_{i=1}^{n-1}{\sum }_{j=i+1}^{n}{b}_{ij}{x}_{i}{x}_{i}+e$$
(2)

In this equation, \({x}_{i}\) (i = 1–5) represents the variables of the experiment. Y denotes the response function, which in this study is Fe content in biomass (Response 1) and dried biomass (Response 2). The variable e represents the experiment's error. \({b}_{0}\) showed the constant coefficient, while bi (i = 1–5) refers to the linear coefficient. \({b}_{ij}\) (i ≠ j) represents the second-order coefficient, and \({b}_{ii}\) (i = 1–4) signifies the second-order interaction coefficient. The independent variables denoted by x. In the present study, the independent variables were coded as A, B, C, D, and E. Eventually, the predicted model was confirmed for variables used in current design.

Statistical analysis

The statistical analysis of the outcomes was conducted by Design Expert@ version 7.0.0. The qualities of fitted models assessed based on the coefficient of determination (R2). To assess the significance of each term, the ANOVA combined with the F-test was employed to analyze the iron concentration in the cell plate and dry weight of biomass, with a significance level of p ≤ 0.05. The location of the optimum variables was estimated using solving the fitted model.

Results and discussion

Biosynthesis of nanoparticles

As demonstrated in Fig. 1A (the FESEM image of the purified superparamagnetic Fe3O4 nanoparticles), biogenic magnetite nanoparticles well dispersed and spherical shape. The particle size distribution of biogenic magnetite nanoparticles based on the FESEM image is demonstrated in Fig. 1B. The particle size histogram of magnetite nanoparticles indicates that the particle size ranges from 10 to 110 nm which may be due to the magnetic property that they attach. The frequency distribution revealed that almost 50% of the particles are in the range of 10–40 nm.

Fig.1
figure 1

(A). FESEM image of Fe3O4 nanoparticles synthesized by strain ASFS1;(B ). Particle size distribution of the biogenic magnetite nanoparticles obtained from FESEM images.

Full size image

Optimization of biogenic magnetite nanoparticles production

The first step towards optimization is to identify the components of the medium that have a notable effect on magnetite nanoparticles biosynthesized by strain ASFS1. Five medium components were evaluated at specified levels (Table. 1). The color change of the culture medium to a brownish-black hue after 96 h showed in 19 runs, which indicates the production of magnetite nanoparticles (Fig. 2). The Fe content in biomass and dried biomass results in an augmented 25−1 FFD with the levels in coded units described in Table 2. The response range were from 25.6 to 160.8 μg mL−1 for Fe content in biomass (response 1) and 1.0–3.3 mg mL−1 for dried biomass (response 2). To the analysis of the response effects and assess the significance of the model terms, the ANOVA tables were designed (Table 3 for response 1 and response 2).

Fig. 2
figure 2

Changing the color of the culture medium in 19 different runs designed with different components of the culture medium. The change in the color of the culture medium to brownish-black indicates an increase iron uptake by strain ASFS1. The result of increase for iron uptake by this strain is to increase the synthesis of magnetic nanoparticles. The result of increase for iron uptake by this strain is to increase the synthesis of magnetic nanoparticles.

Full size image
Table 3 ANOVA for dependent variable: Fe content in biomass.
Full size table

The normality of the data for both responses examined using plotting a normal probability plot of the residuals. According to Fig. 3, the data points on the plot fall fairly near to the straight line, which indicates the data follows a normal distribution with a relatively consistent variance throughout the responses range.

Fig. 3
figure 3

The normal probability plot of the residuals, (A). for Response 1 and (B). for response 2.

Full size image

In response 1, X4, X5, X1 X2, X3 X4 are significant model terms. The p-value of Na2-EDTA (X4) was less than 0.05 (Table 3), thus this parameter had a significant effect on superparamagnetic Fe3O4 nanoparticles production strain ASFS1. Alphandéry et al29. reported addition of iron-chelating agents (specially EDTA) to the bacterial growth medium enables the enhance of the growth of Magnetospirillum magneticum strain AMB-1, an increase production yield of the magnetosomes size and chain length and yielded improved magnetosome heating properties. The kinetics of magnetosome synthesis in Magnetospirillum magneticum strain AMB-1 probably happens after the bacteria have reached the stationary phase29,30. Adding iron-chelating agents to the growth medium improves the rate of production of magnetic nanoparticles strain AMB-129. Nevertheless, in the strain MSR-1, magnetosomes are produced before the bacteria have reached the stationary phase, possibly due to various mechanisms of magnetosome formation or iron uptake29,31. Thus, the effects of the presence of iron-chelating agents on the rate of magnetosome formation may be less noticeable for strain MSR-1 than strain AMB-129,31. Therefore, EDTA as an iron-chelating factor has a positive effect on the growth and iron uptake in the strain ASFS1 of this study, similar to the AMB-1strain29, due to probably the mechanism of formation of magnetic nanoparticles in both strains being similar and different from the MSR1 strain31. Similar to our study, Heyen et al31. and Alphandéry et al29. optimized the components of the culture medium for strains MSR-1 and AMB-1 respectively, unlike the present study, none of them used the experimental design method. The use of statistical methods such as Design Expert to design optimization experiments enhanced the efficiency and precision, and it provides investigation of several factors simultaneously. Finally, it improved the comprehensiveness of optimization results in bioprocessing applications25. Iron is a significant factor influencing the production of magnetic nanoparticles using magnetotactic bacteria32,33. Fe (III) deposited directly inside the bacterial cells comprising magnetosomes32. However, magnetotactic bacteria can utilize different iron sources, but the association between growth bacteria and iron uptake is not so clear33. Analysis of various concentrations of FeCl3 (X5) exhibited that this factor had considerable effect on magnetite nanoparticles production by strain ASFS1 (p < 0.05), while different concentration of FeSO4 had not significant effects (p < 0.05). Similar to our study, kabary et al33. reported that FeSO4 were less significant effects for bacterial growth and iron uptake by Pseudomonas aeruginosa Kb1 (confirmed as magnetotactic bacteria). Some studies have used FeSO4 as a single source of iron34,35. Berny et al22. used FeSO4 and FeCl3 as iron sources for the production of magnetic nanoparticles by magnetotactic bacteria, similar to our study. Although, various concentrations of FeSO4 had not significant effects, the interaction effect of X3 X4 is significant (p < 0.05) in the present study. Table 3 revealed that the yeast concentration (X1) and tryptone concentration (X2) not significant separately, while the interaction of X1X2 is significant. Yang et al36. showed yeast extract is not significant in magnetic nanoparticle production by bacteria. Their results are similar to our study, whereas peptone only improves the final bacterial cell density. Furthermore, adding only yeast extract and polypeptide as a culture medium produces a small number of magnetosomes37, similar to the current study.

Figure 4A, B exhibited one factor affecting Fe content in biomass within their experimental range while holding four factors constant. Furthermore, Fig. 4C, D illustrate the interaction of two factors, which the other three factors are kept constant. In Fig. 4A, tryptone (10 g L-1), yeast extract (5 g L-1), FeSO4 (50 μg mL−1), and FeCl3 (100 μg mL−1) are constant, and with increasing Na2-EDTA concentration, the amount of iron in the biomass significantly increased. Analysis of response at the different levels of FeCl3 when tryptone (10 g L-1), yeast extract (5 g L-1), FeSO4 (50 μg mL−1), and Na2-EDTA (100 μg mL−1) are constant, indicating that FeCl3 considerably increased iron in the biomass (Fig. 4B). The interaction of tryptone and yeast extract on iron in the biomass (Fig. 4C) exhibited that by raising tryptone, the amount of iron in biomass not significantly raised, while using increasing yeast extract, the amount of iron in biomass considerably increased (constant factors: 50 μg mL−1 of FeSO4, 100 μg mL−1 of Na2-EDTA and 100 μg mL−1 of FeCl3). Thus, yeast extract alone has no effect on the amount of iron in biomass, but its interaction with tryptone has a significant effect on increasing the amount of iron in biomass. In Fig. 4D, the interaction of FeSO4 and Na2-EDTA shows that the amount of iron in biomass does not increase with the increase level of FeSO4, while increase level of Na2-EDTA lead to the amount of iron in biomass increases significantly).

Fig. 4
figure 4

Interaction plot within the ANOVA for factors affecting on Fe content in biomass (of 1 cc culture medium); A) Na2-EDTA, B) FeCl3, C) Tryptone- yeast extract interaction and D) Na2-EDTA and FeSO4 interaction.

Full size image

In response 2, X1, X5, X1 X2 are significant model terms. The p-value of yeast extract (X1) and FeCl3 (X5) was less than 0.05 (Table 4). Therefore, both parameters have significant effects on magnetite nanoparticles production strain ASFS1. Contrary to Yang et al36. study that polypeptone increases the biomass dry weight and yeast extract has no significant, the yeast extract increases biomass dry weight in the present study. Also, the interaction effects of yeast extract and tryptone (X1X2) is significant (p < 0.05).

Table 4 ANOVA for dependent variable: dried biomass.
Full size table

Figure 5A, B displayed one factor affecting dried biomass within their testing range and keeping four factors’ constants. In addition, Fig. 5C revealed the interaction of two factors, which the further three factors are held constant. The results show that increasing the level of yeast extract and FeCl3 separately, has led to a significant increase in dry biomass (Fig. 5A, B, respectively). In Fig. 5C, the interaction of yeast extract and tryptone displays that dry biomass does not significantly rise with the raised level of tryptone, while the enhanced level of yeast extract leads to dry biomass enhancing considerably (constant factors: 100 μg mL−1 of FeCl3, 100 μg mL−1 of Na2-EDTA and 50 μg mL−1 of FeSO4).

Fig. 5
figure 5

Interaction plot within the ANOVA for factors affecting on dried biomass (of 1 cc culture medium); A) yeast extract B) FeCl3 and C) Tryptone- yeast extract interaction.

Full size image

The fitted model (in the terms of coded values) for estimating Fe content in biomass (3) and dried biomass from 1 cc culture (4) by the selected strain were, respectively:

$$\begin{aligned} {\text{Fe}}\,{\text{content}}\,{\text{in}}\,{\text{Biomassfrom}}\,{1}\;{\text{cc}}\,{\text{ culture}} = & + { 82}.{23 } + { 4}.{19}X_{1} - { 5}.{6}0X_{2} + { 9}.{29}X_{3} \\ & + {11}.0{6}X_{4} + { 25}.0{9}X_{5} - { 1}0.{81}X_{1} X_{2} + {11}.{88}X_{3} X_{4} \\ \end{aligned}$$
(3)
$${\text{Dried}}\,{\text{Biomass}}\,{\text{from}}\,{1}\,{\text{cc}}\,{\text{culture }} = \, + { 1}.{94 } + \, 0.{35}X_{1} + \, 0.{15}X_{2} + \, 0.{28}X_{5} - \, 0.{19}X_{1} X_{2}$$
(4)

The effects of X1, X3, X4, X5, and X3 X4 were positive (synergistic), while X2 and X1 X2 had negative (antagonism) effects on Fe content in biomass. In response 2, the effects of X1, X2, X5 were synergistic and only X1 X2 had antagonism effects on Dried Biomass.

The Model F-value of 8.42 and 8.57 for response 1 and response 2 respectively, implies the models are significant. Furthermore, values of "Prob > F" less than 0.0500 for both responses indicate model terms are significant. There is only a 0.16% and a 0.13% chance that a "Model F-value" this large could occur due to noise for responses 1 and 2, respectively. The relatively high value of R2 = 0.8549 (response 1) and R2 = 0.7251 (response 2) indicated that the models were significantly fitted. The higher the R2 value (closer to the number one) indicates better the model fits, as it explains more of the data's variance. In this study, R2 = 0.8549 (response 1) and R2 = 0.7251 (response 2) indicate a good quality model that explains 85% and 72% of the data variance, respectively.

The location of optimum levels by the selected model for Fe content in biomass (158 μg mL−1) and dried biomass (2.59 mg mL−1) were predicted to be 7.95 g L−1 of yeast extract, 5 g L−1 of tryptone, 75 μg mL−1 of FeSO4, 192.3 μg mL−1 of Na2-EDTA and 150 μg mL−1 of FeCl3.

To confirm the model adequacies for predicting Fe content in biomass and dried biomass, one additional experiment (repeated three times) by the same medium composition were performed. In the event that the average of the validation outcomes was within the limits of the confidence interval (CI), then the significant parameters and suitable levels for acquiring the desired outcomes were correctly selected38. The outcomes of three validation experiments demonstrated that there are good agreements between the predicted (The Fe content in biomass and dried biomass were 158 μg mL−1 and 2.59 mg mL−1, respectively) and experimental (The Fe content in biomass and dried biomass were 139 ± 13 μg mL−1 and 2.2 ± 0.2 mg mL−1, respectively) outcomes that revealed the chosen model was able to navigate within the design space.

Conclusion

The present study was aimed to optimize culture media components for strain ASFS1 growth and magnetite nanoparticles formation by fractional factorial designs. Yeast extract, tryptone, and FeSO4 were confirmed not to be very significant factors considering to Fe content in biomass assisted by strain ASFS1, while Na2-EDTA, FeCl3, interaction yeast extract-tryptone, and interaction FeSO4- Na2-EDTA were illustrated the significant effect on Fe content in biomass. Tryptone was proved not significant parameter considering to dried biomass, while yeast extract, FeCl3, and interaction yeast extract-tryptone were demonstrated a significant effect on dried biomass of strain ASFS1. In both responses 1 and 2, the factors of FeCl3 and interaction effects yeast extract-tryptone are significant (p < 0.05). The maximum fe content in biomass, along with the maximum dried biomass, were achieved at 7.95 g L−1 of yeast extract, 5 g L−1 of tryptone, 75 μg mL−1 of FeSO4, 192.3 μg mL−1 of Na2-EDTA and 150 μg mL−1 of FeCl3. Nevertheless, further studies must be accomplished to find about the related pathway involved in magnetite nanoparticles biosynthesis and the potential application of this strain for large-scale production of superparamagnetic Fe3O4 nanoparticles.