Functionalized Magnetite Nanoparticles: Characterization, Bioeffects, and Role of Reactive Oxygen Species in Unicellular and Enzymatic Systems

The current study evaluates the role of reactive oxygen species (ROS) in bioeffects of magnetite nanoparticles (MNPs), such as bare (Fe3O4), humic acids (Fe3O4-HA), and 3-aminopropyltriethoxysilane (Fe3O4-APTES) modified MNPs. Mössbauer spectroscopy was used to identify the local surrounding for Fe atom/ions and the depth of modification for MNPs. It was found that the Fe3O4-HA MNPs contain the smallest, whereas the Fe3O4-APTES MNPs contain the largest amount of Fe2+ ions. Bioluminescent cellular and enzymatic assays were applied to monitor the toxicity and anti-(pro-)oxidant activity of MNPs. The contents of ROS were determined by a chemiluminescence luminol assay evaluating the correlations with toxicity/anti-(pro-)oxidant coefficients. Toxic effects of modified MNPs were found at higher concentrations (>10−2 g/L); they were related to ROS storage in bacterial suspensions. MNPs stimulated ROS production by the bacteria in a wide concentration range (10−15–1 g/L). Under the conditions of model oxidative stress and higher concentrations of MNPs (>10−4 g/L), the bacterial bioassay revealed prooxidant activity of all three MNP types, with corresponding decay of ROS content. Bioluminescence enzymatic assay did not show any sensitivity to MNPs, with negligible change in ROS content. The results clearly indicate that cell-membrane processes are responsible for the bioeffects and bacterial ROS generation, confirming the ferroptosis phenomenon based on iron-initiated cell-membrane lipid peroxidation.

The possibility to attain functionalized iron-based nanoparticles by alkoxysilanes is of wide scientific interest to researchers [56]. The chemistry of the silica surface is well established, enabling a large number of applications of silica-coated nanoparticles [56,57]. Alkoxysilane can bind to metal oxide by adsorption or covalent linking. A functional alkoxysilane consists of at least two functional groups, the first group being used for attachment to the surface of the MNPs, and the second group providing an adhesion point to the bond of the molecule [1]. Nanoparticles coated with silica or alkoxysilanes are promising reagents in nanobiotechnology due to their biocompatibility and hydrophilic properties. The great variety of alkoxysilanes makes it possible to carry out various types of functionalization of the surface of nanoparticles by introducing charges into the system. Thus, functionalization prevents from the aggregation of nanoparticles in liquids and improves their chemical stability by controlling surface charge [58].
In this study, we chose a luminous marine bacterium as a model microorganism to evaluate and compare the biological effects of MNPs. The bioluminescence bacteria-based assessment is known as a toxicity assay and it has been applied for more than 50 years for toxicity monitoring due to its simplicity and high sensitivity [59][60][61][62][63][64]. The bioassay uses luminescence intensity as a physiological test parameter providing a convenient approach with high rates of analysis (1-10 min), low costs, and ease of the bioassay procedure.
We compare the cellular bioeffects of MNPs in the bioluminescence enzymatic assay system, proposing a relatively new approach in a toxicology practice [65][66][67][68]. The conventional enzymatic bioluminescent assay is based on the bacterial bioluminescent enzyme system, which involves two coupled enzymatic reactions: where, FMN is flavin mononucleotide, NADH is nicotinamide adenine dinucleotide disodium salt-reduced. Similar to cellular bioassay, the enzymatic analysis can assess the general toxicity in test samples allowing the integration of all the interactions of toxic compounds with the bioluminescent system such as redox processes, polar and nonpolar binding, etc. Moreover, the enzymatic bioassay is specific to oxidizers due to its additional kinetic parameter, the induction period, which directly depends on the redox potentials of toxicants in solutions [67]. Previously [69,70], we used the bioluminescent enzymatic assay system to evaluate the toxicities of general and oxidative types in solutions of organic and inorganic oxidizers (quinones and polyvalent metals, respectively). Changes in general toxicity and oxidative toxicity of aqueous media under exposure to humic substances were studied previously in our previous studies [71][72][73]. Later [74][75][76][77][78][79], the toxicity and antioxidant activity of a series of fullerenols (i.e., carbon nanostructures, water-soluble derivatives of fullerenes, and perspective pharmaceutical agents) were evaluated and compared.
In our previous work [80], we used bioluminescence bioassays, cellular and enzymatic, to evaluate the bioeffects of three types of MNPs: bare MNPs, modified by humic acids, and silica (3-aminopropyltriethoxysilane, APTES), i.e., (1) Fe 3 O 4 , (2) Fe 3 O 4 /HA, and (3) Fe 3 O 4 /APTES, respectively. The bioeffects under study were: toxic effects and anti/prooxidant activity of the MNPs. Current work continues the study of three types of MNPs [80]; it evaluates the role of ROS in the bioeffects of these nanoparticles. Contents of reactive oxygen species (ROS) were determined using chemiluminescence luminol assay in the bioluminescence assay systems. Correlations of ROS content with toxicity/anti-(pro-)oxidant coefficients were evaluated, thus contributing to comprehending ferroptosis phenomenon.

Local Environment of Fe Atom/Ions
The composition and morphological properties of MNPs were studied by Mössbauer spectroscopy, and related spectra, along with their description and results of simulation in different state superparamagnetic relaxation models are presented in Table 1 and Supplementary materials ( Figure S1, Table S1). Relying on the presented results, it can be argued that magnitudes of magnetic splittings, obtained even for spectra at low temperatures, are lower than expected for massive samples of magnetite and maghemite [81], which is typical for nanosized materials [82]. Indeed, the sizes of magnetic domains can be estimated in the range from 12.6 (for Fe 3 O 4 -HA) to 15.7 (for Fe 3 O 4 -APTES) nm as reported in Table 1.  Table S1, D-magnetic domain diameter.
Moreover, by analyzing the areas of the subspectra related to iron atoms in different crystallographic sites of magnetite, the value of the non-stoichiometric parameter-δ for nano-magnetite- [15,83], can be estimated by using the following expression: (1) where S i is the relative area of the subspectrum with isomeric shift δ i related to the iron atoms in the B-site, S j is the relative area of the remaining subspectra, δ 2 and δ 3 are the isomeric shifts of iron atoms (+2) and (+3) in the octahedral oxygen environment for given temperature (here δ 2 = 1.16 ± 0.06 и 1.33 ± 0.09 mm/s for 296 and 78 K, respectively, δ 3 = 0.37 ± 0.04 and 0.49 ± 0.04 mm/s for 296 and 78 K, respectively) [84]. Computed results are reported in Table 1, and the data reveal that magnetite samples are quite strongly oxidized indicating a magnetite-maghemite solid solution [15]. At the same time, the Fe 3 O 4 -HA sample contains the smallest amount of Fe 2+ ions, whereas the Fe 3 O 4 -APTES contain the largest amount. It should be noted that, in the case of the studied samples, there is a clear correlation between their particle sizes and nonstoichiometric composition which leads to the following result: the smaller the particle size, the more nonstoichiometric the nanomagnetite composition is (Table 1).

Magnetic Parameters
The most crucial property of MNPs, which allow a variety of applications, especially in biomedicine, is represented by their ferrimagnetism. As a matter of fact, the application of an external magnetic field makes it possible to concentrate magnetic nanoparticles in the target point, thus reducing likely side effects [85][86][87][88]. Some magnetic characteristics for the Fe 3 O 4 MNPs are presented in Table 1. The hysteresis loops for the three different magnetite NMPs results were closed and symmetrical with respect to the origin of the coordinate system as reported in Figure 1.
The most crucial property of MNPs, which allow a variety of applications, especially in biomedicine, is represented by their ferrimagnetism. As a matter of fact, the application of an external magnetic field makes it possible to concentrate magnetic nanoparticles in the target point, thus reducing likely side effects [85][86][87][88]. Some magnetic characteristics for the Fe3O4 MNPs are presented in Table 1. The hysteresis loops for the three different magnetite NMPs results were closed and symmetrical with respect to the origin of the coordinate system as reported in Figure 1. The shape of the loops indicates the ferromagnetic features of the material which are prone to their potential application. The saturation magnetizations of bare Fe3O4, Fe3O4/HA, and Fe3O4/APTES were respectively 68.2, 30.9, and 31.2 emu/g (Table 1), suggesting a 40% (w/w) of HA content in Fe3O4/HA. The attained values of saturation magnetization indicate that NMPs are stabilized by the HA or APTES functionalization exhibiting superparamagnetic properties at room temperature ( Figure 1) and the corresponding variations, respectively of 54.6% for the Fe3O4/H and 54.5% for the Fe3O4/APTES, compared to bare magnetite can be explained by the noncollinearity of surface spins of MNPs and likely by the same thickness of the HA and APTES shell surrounding the nanomagnetite particles. In the absence of a magnetic field, all samples showed a similar low residual magnetism ∼±4-7 emu g −1 (Table 1), due to magnetic viscosity for superparamagnetic materials [89]. The further increase of the coercivity can be related to cumulative anisotropy by phase transformation to maghemite in the case of Fe3O4/HA in terms of concentration compared to Fe3O4/APTES, which correlates with Mössbauer spectroscopy data [90].
In general, bare and functionalized nanoparticles exhibit a saturation magnetization sufficient to control an external magnetic field. In many works, focused on the preparation of materials for catalyzing ferroptosis, the authors have used iron salts as iron ion sources  (Table 1), suggesting a 40% (w/w) of HA content in Fe 3 O 4 /HA. The attained values of saturation magnetization indicate that NMPs are stabilized by the HA or APTES functionalization exhibiting superparamagnetic properties at room temperature ( Figure 1) and the corresponding variations, respectively of 54.6% for the Fe 3 O 4 /H and 54.5% for the Fe 3 O 4 /APTES, compared to bare magnetite can be explained by the noncollinearity of surface spins of MNPs and likely by the same thickness of the HA and APTES shell surrounding the nanomagnetite particles. In the absence of a magnetic field, all samples showed a similar low residual magnetism ±4-7 emu g −1 (Table 1), due to magnetic viscosity for superparamagnetic materials [89]. The further increase of the coercivity can be related to cumulative anisotropy by phase transformation to maghemite in the case of Fe 3 O 4 /HA in terms of concentration compared to Fe 3 O 4 /APTES, which correlates with Mössbauer spectroscopy data [90].
In general, bare and functionalized nanoparticles exhibit a saturation magnetization sufficient to control an external magnetic field. In many works, focused on the preparation of materials for catalyzing ferroptosis, the authors have used iron salts as iron ion sources [29,91,92], however, as reported by Huo et al., 2017, [93] without targeting, only 6.95% of nanoparticles were stored and localised in the 4T1 cancer cell 48 h upon injection.

Iron Ion Release
The study of the release of Fe 2+ ions and Fe 3+ ions is also associated with the different contributions of ions of different redox states to the Fenton reaction [94]. It is known that the rate constant of •OH generation via Fenton reaction catalyzed by Fe 2+ ion [76 (mol/L) −i s −s ] is 10 4 −10 5 times higher than this of the Fe 3+ ion-catalyzed Fenton-like reaction [0.001−0.07 (mol/L) −i s −s ] [94]. In this regard, estimation of the concentration of released ions of different redox states is of interest in order to identify their contribution to the induction of ROS production.
UV-Vis spectroscopy was used to investigate the kinetics of the release of Fe 2+ and Fe 3+ ions from nanoparticles Fe 3 O 4 , Fe 3 O 4 /HA, and Fe 3 O 4 /APTES ( Figure S2). It is noticeable from the spectra, that Fe 3 O 4 suspension reveals a release of both Fe 2+ and Fe 3+ ions, respectively of 6.3 mg/L and 3.5 mg/L, as shown in Figure 2A. The Fe 3+ /Fe 2+ ratio during the first hour was close to the stoichiometric ratio for magnetite, i.e., 1.8-2.3 ( Figure 2B). Upon 24 h, Fe 3+ concentration increases taking the Fe 3+ /Fe 2+ ratio to a value of 3.3 due to the oxidation of Fe 2+ to Fe 3+ ions.
contributions of ions of different redox states to the Fenton reaction [94]. It is known that the rate constant of •OH generation via Fenton reaction catalyzed by Fe 2+ ion [76 (mol/L) −i s −s ] is 10 4 −10 5 times higher than this of the Fe 3+ ion-catalyzed Fenton-like reaction [0.001−0.07 (mol/L) −i s −s ] [94]. In this regard, estimation of the concentration of released ions of different redox states is of interest in order to identify their contribution to the induction of ROS production.
UV-Vis spectroscopy was used to investigate the kinetics of the release of Fe 2+ and Fe 3+ ions from nanoparticles Fe3O4, Fe3O4/HA, and Fe3O4/APTES ( Figure S2). It is noticeable from the spectra, that Fe3O4 suspension reveals a release of both Fe 2+ and Fe 3+ ions, respectively of 6.3 mg/L and 3.5 mg/L, as shown in Figure 2A. The Fe 3+ /Fe 2+ ratio during the first hour was close to the stoichiometric ratio for magnetite, i.e., 1.8-2.3 ( Figure 2В). Upon 24 h, Fe 3+ concentration increases taking the Fe 3+ /Fe 2+ ratio to a value of 3.3 due to the oxidation of Fe 2+ to Fe 3+ ions. Fe3O4/HA and Fe3O4/APTES do not release Fe 2+ ions within 24 h, Figure 2A. In the case of Fe3O4/HA, the absence of Fe 2+ ions within the solution can be explained by the complete oxidation of magnetite Fe3O4 to maghemite Fe2O3, according to the Mössbauer spectroscopy data (Section 2.1.1). However, analyzing the same data as well as results of X-ray diffraction [82], APTES-functionalized nanoparticles contain both Fe 2+ and Fe 3+ ions. The absence of Fe 2+ ions into the Fe3O4/APTES solution can be explained by the very low concentration which can fall beyond the sensitivity limits of the employed analysis technique. Additionally, Fe 2+ would undergo spontaneous chemical oxidation with molecular oxygen or other potential oxidants to Fe 3+ , which eventually precipitate as ferric hydroxides [95,96].
Both functionalized oxides, Fe3O4/HA and Fe3O4/APTES, release Fe 3+ ions, however, the release kinetics differ significantly ( Figure 2A). So, in the case of Fe3O4/HA, the Fe 3+ concentration in the solution increases from 10.6 mg/L to 19.0 mg/L in a half an hour, and  . However, analyzing the same data as well as results of X-ray diffraction [82], APTES-functionalized nanoparticles contain both Fe 2+ and Fe 3+ ions. The absence of Fe 2+ ions into the Fe 3 O 4 /APTES solution can be explained by the very low concentration which can fall beyond the sensitivity limits of the employed analysis technique. Additionally, Fe 2+ would undergo spontaneous chemical oxidation with molecular oxygen or other potential oxidants to Fe 3+ , which eventually precipitate as ferric hydroxides [95,96].
Both functionalized oxides, Fe 3 O 4 /HA and Fe 3 O 4 /APTES, release Fe 3+ ions, however, the release kinetics differ significantly ( Figure 2A). So, in the case of Fe 3 O 4 /HA, the Fe 3+ concentration in the solution increases from 10.6 mg/L to 19.0 mg/L in a half an hour, and then decreases to 9.5 mg/L in 24 h. In the case of Fe 3 O 4 /APTES, on the contrary, a decay of Fe 3+ release (from 58.2 mg/L to 29.6 mg/L in half an hour after suspension) was revealed with a subsequent increase of Fe 3+ concentration almost to the initial value of 51.4 mg/L after 24 h. In general, functionalized MNPs release more Fe 3+ ions compared to bare MNPs, but do not or negligibly release Fe 2+ ions. For this reason, bare MNPs represent an effective source of a pair of Fe 2+ /Fe 3+ ions, whereas the HA and APTES functionalized nanoparticles, under the specified synthesis conditions, are a source of only the Fe 3+ ions.
The difference in the behavior of iron ions released from the different considered samples would need an attempted explanation. In the case of bare Fe 3 O 4 , their occurring aggregation leads to a decrease in the concentration of surface ions and consequently to their hydrolysis forming stable hydrocomplexes as supported by the binding constant increasing for lgK(Fe(OH) 3aq ) and lgK(Fe(OH) 4 ), from 12 and 22, respectively [97]. Due to a high complexing potential of HA, the binding constants (lgK) of Fe 2+ and Fe 3+ ions are, five and four, respectively, according to [98]. Based on this evidence, HA functionalization not only induces stabilization of the magnetite, but also triggers iron ions to be pulled out by HA from the NP's surface leading into the solution. The higher release of iron ions, in the case of Fe 3 O 4 /APTES, is probably associated with a high level of porosity and amorphous structure of silica, according to the results reported by Oliveira, et al., 2019, and Otalvaro, et al., 2019, [99,100].
In 1987, Minotti et al. [101] showed that lipid peroxidation can be induced both by the participation of Fe 2+ or Fe 3+ ions in the formation of hydroxyl radicals •OH as a result of the Fenton reaction, and also by the participation of pairs of Fe 2+ /Fe 3+ ions immediately to form the Fe 3+ -dioxygen-Fe 2+ complex. However, in a physiologically neutral or slightly acidic environment such as a tumor area, the effectiveness of the Fenton reaction is relatively low [102]. Even under acidic pH conditions, the Fenton reaction catalyzed by Fe 2+ has a low reaction rate (~63 M −1 s −1 ), leading to the slow formation of ROS [103]. For these reasons, the development of nanopreparations featured by high catalytic activity and specificity, is highly prospective in a weakly acidic and neutral media (pH 6-7) as a tumor microenvironment.
The ability to generate ROS by MNPs in highly diluted solutions was studied in relation to bacterial and enzymatic systems at a pH value of~6 in Sections 2.2 and 2.3.

Effects of MNPs on Bacterial Cells
The luminescence intensity of bacterial cells was studied in the presence of MNPs considering different concentrations of MNPs. The MNP effects were analyzed at <1 g/L, as higher content was limited by both low water solubility and nanoparticles dispersion stability during the experiments and also by the demand to avoid the "optical filter" effect, which limits the use of luminescence signal detection in solutions with high optical density and/or light-scattering suspensions [104].

Toxic Effects of MNPs in Bacterial Suspension (a) Effects of MNPs on Bacterial Bioluminescence Intensity
All three types of MNPs demonstrated moderate bioluminescence activation (1.0 < I rel < 1.25, p < 0.05) in a lower concentration range (i.e., 10 −8 -10 −2 g/L), as shown in Figure 3. The activation of bacterial response by MNPs might be described in terms of the conventional "hormesis" model [105][106][107][108]. It is known that the model includes, in the broadest case, three stages of the biological dose-dependent response-stress recognition (I), activation (II), and inhibition of organismal functions or toxic effect (III). As a concept, hormesis involves favorable biological responses to low exposures to stressors, in all cases [109,110].
To characterize the toxicity of the MNP samples, the bioluminescence inhibition (I rel < 1.0, p < 0.05) effects due to higher concentrations of MNPs (>10 −2 g/L) were analyzed. The EC 50 values of MNPs (i.e., MNP concentrations at 50% bioluminescence inhibition) were determined and presented in  Figure 2A,B), which actively produce toxic hydroxyl radicals (according to Fenton reaction [101]) and the result in lipid peroxidation in the cellular membrane.  To characterize the toxicity of the MNP samples, the bioluminescence inhibition (I r < 1.0, p < 0.05) effects due to higher concentrations of MNPs (>10 −2 g/L) were analyzed. The EC50 values of MNPs (i.e., MNP concentrations at 50% bioluminescence inhibition) were determined and presented in Table 2. The results show that the values are comparable however, the lowest value of EC50 of bare Fe3O4 is evidence of its higher toxicity compared to surface-modified MNPs (Fe3O4 /APTES and Fe3O4/HA). Probably, this is a result of the Fe 2+ release in suspensions of bare Fe3O4, the participation of Fe 2+ /Fe 3+ pairs (see Section 2.1.3, Figure 2A,B), which actively produce toxic hydroxyl radicals (according to Fenton reaction [101]) and the result in lipid peroxidation in the cellular membrane.     Figure 4A) and (2) MNPs + bacteria ( Figure 4B). Figure 4A shows an ROS increased content in bacteria-free media at higher concentrations of MNPs (10 −4 -10 −1 g/L) and only a negligible rise was observed for bare Fe 3 O 4 . Whereas, a relevant variation was observed for both modified MNPs at >10 −3 g/L. Such difference between ROS formation into MNP suspensions between modified and non-modified surface MNPs could be probably related to free Fe 3+ ions content in the suspension, Figure 2A.
Similar differences in ROS content due to modified and unmodified MNPs were assessed in bacterial suspensions at high nanoparticle concentrations (>10 −3 g/L, refer to Figure 4B). In fact, an increase in the ROS content was found at high concentrations (>10 -3 g/L) of two modified MNPs, but not observed for bare Fe 3 O 4 . Similarly to the bacteria-free media ( Figure 4A), this may be due to a higher concentration of Fe 3+ released by functionalized nanoparticles (Figure 2A).
It should be noted, that a moderated increase in the ROS content (as compared to the control samples, ROS rel > 1.0, p < 0.05) was observed over a wide range of MNP concentrations (10 -15 -10 -3 g/L) in bacterial media ( Figure 4B) compared to the bacteria-free media ( Figure 4A). This clearly indicates the effective ROS generation associated with the bacterium itself within the low-concentration solutions of MNPs. The above-mentioned effect represents an important achievement that should be further studied in detail.  Figure 4A shows an ROS increased content in bacteria-free media at higher concentrations of MNPs (10 −4 -10 −1 g/L) and only a negligible rise was observed for bare Fe3O4. Whereas, a relevant variation was observed for both modified MNPs at >10 −3 g/L. Such difference between ROS formation into MNP suspensions between modified and nonmodified surface MNPs could be probably related to free Fe 3+ ions content in the suspension, Figure 2A.
Similar differences in ROS content due to modified and unmodified MNPs were assessed in bacterial suspensions at high nanoparticle concentrations (>10 −3 g/L, refer to Figure 4B). In fact, an increase in the ROS content was found at high concentrations (>10 -3 g/L) of two modified MNPs, but not observed for bare Fe3O4. Similarly to the bacteria-free media ( Figure 4A), this may be due to a higher concentration of Fe 3+ released by functionalized nanoparticles (Figure 2A).
It should be noted, that a moderated increase in the ROS content (as compared to the control samples, ROS rel > 1.0, p < 0.05) was observed over a wide range of MNP concentrations (10 -15 -10 -3 g/L) in bacterial media ( Figure 4B) compared to the bacteria-free media ( Figure 4A). This clearly indicates the effective ROS generation associated with the Correlations between bioluminescence intensity I rel and ROS rel (Figures 3 and 4B) were analyzed. The correlation coefficients, r, were calculated for three MNP types, as reported in Table 3 Figure 2A) lead to the reasonable assumption of additional occurring mechanisms, such as the production of toxic −OH via Fenton reaction [101] and break of quantitative correlations between ROS and MNP contents (Table 3). Table 3. Correlation coefficients, r, between dependencies of I rel and ROS rel on MNP concentrations. Previously, high positive correlations between I rel and ROS rel were found in suspensions of luminous bacteria and fullerenol at high "toxic" fullerenol' concentrations [79]. Fullerenols, conversely to MNPs, are antioxidants and active trappers of free radicals, and they could lower the ROS content effectively [111][112][113][114][115][116] suppressing the rates of oxidative processes in cells. Hence, the opposite signs of correlations for fullerenols and MNPs could be concerned with their anti-and pro-oxidant properties, respectively, indicating the complexity and multiplicity of processes of different bioactive compounds.

Effects of MNPs in Bacterial Suspension under Model Oxidative Stress
The MNP anti-(pro-)oxidant activity in bacterial suspensions was investigated under specific conditions of model oxidative stress. During this study, toxic concentrations of MNPs (i.e., >10 −2 g/L) were excluded, while bacterial bioluminescence kinetics was monitored and recorded assuming model solutions of an oxidizing agent, 1,4-benzoquinone (Bq), at its EC 50 to suppresses bioluminescence intensity by 50% (EC 50 , Bq = 10 -7 M). The ROS content was monitored by chemiluminescence technique, within the same media after bioluminescence measurements. It has been assumed that restoration of the bioluminescence intensity (I rel Bq > 1) results in evidence of MNP antioxidant activity, whereas the additional bioluminescence suppression (I rel Bq < 1) would indicate a prooxidant attitude of MNPs.
(a) Effects of MNPs on Bacterial Bioluminescence Intensity under Model Oxidative Stress Figure 5A shows the I rel Bq -values at different concentrations of MNPs. All considered MNPs did not activate reliably the bacterial bioluminescence under oxidative stress, hence, it can be reasonably concluded that MNPs do not induce any antioxidant activity. Oppositely, the I rel Bq -values showed a suppression of the bioluminescence intensity (i.e., I rel Bq < 1, p < 0.05) at higher concentrations of MNPs (>5 × 10 −5 g/L), confirming the prooxidant properties of MNPs. The HA-modified MNPs (Fe 3 O 4 /HA) suppressed bioluminescence intensity (i.e., I rel Bq < 1, p < 0.05) reliably in a wider range of concentrations (10 -12 -10 -2 g/L), as compared to Fe 3 O 4 and Fe 3 O 4 /APTES and this can be explained by the partial hydrolysis of the HA coating in Bq solutions which leads, as a result, to an oxidized iron higher content (Fe 3+ , surface and/or free) in Fe 3 O 4 /HA suspensions.

(b) Effects of MNPs on ROS Content in Bacterial Suspension under Model Oxidative Stress
The ROS content was preliminarily measured in the absence of MNPs at the following different conditions: in physiological saline (3% NaCl) solution, in bacteria-free and in bacterial media under oxidative stress model conditions (i.e., at EC 50 of Bq, EC 50, Bq = 10 −7 M). We showed previously, that Bq lower the ROS content by 10% and by 15% in bacteria-free media and bacterial media, respectively. Analogous results were attained by [79], earlier. The ROS content reduction can be explained considering the peroxides, representative of ROS groups, whose tendency to combine to Bq follows Scheme 1 as presented below.

(b) Effects of MNPs on ROS Content in Bacterial Suspension under Model Oxidative Stress
The ROS content was preliminarily measured in the absence of MNPs at the following different conditions: in physiological saline (3% NaCl) solution, in bacteria-free and in bacterial media under oxidative stress model conditions (i. e. at EC50 of Bq, EC50, Bq = 10 −7 M). We showed previously, that Bq lower the ROS content by 10% and by 15% in bacteriafree media and bacterial media, respectively. Analogous results were attained by [79], earlier. The ROS content reduction can be explained considering the peroxides, representative of ROS groups, whose tendency to combine to Bq follows Scheme 1 as presented below.
We aimed to reveal how the MNPs addition changes the ROS content in the bacterial suspension under oxidative stress conditions (i.e., in the presence of Bq). Figure 5B shows We aimed to reveal how the MNPs addition changes the ROS content in the bacterial suspension under oxidative stress conditions (i.e., in the presence of Bq). Figure 5B shows the ROS content in the bacterial system at various MNP concentrations in the Bq solutions. An additional valuable decay in the content of ROS (ROS rel Bq < 1, p < 0.05) was observed in the concentration range of MNPs 10 −4 -10 −2 g/L. The further ROS decay might be explained if the Bq balance with its reduced form-hydroquinone-in the bacterial suspension is hypothesized. In this case, the Fe 3+ addition should shift the balance to the oxidized form, i.e., Bq, according to Scheme 2, and thus decreasing ROS additionally. Correlation coefficients, r, were calculated between the dependencies of I rel Bq ( Figure  5A) and ROS rel Bq ( Figure 5B) for the corresponding MNP concentration range as reported in Table 4. Direct correlations were found, showing direct interrelations between these dependencies. The correlation values confirm interrelations between bioluminescence intensity and ROS under oxidative stress conditions, supporting the envisaged chemical mechanism of free Fe 3+ involvement (Figure 2A) to the prooxidant activity of MNPs as depicted in Schemes 1 and 2.

Effects of MNPs on Enzymatic Reactions
The effect of MNPs on bioluminescence enzymatic reaction is a question of interest. Comparison of the MNP effects on bacterial cells and their enzymatic reactions forms understanding at the cellular and biochemical levels, respectively, identifies processes most sensitive to MNP exposure, and allows supposing a role of cellular membrane-related processes in the MNP bioeffects.
Analogously to luminous marine bacteria (Section 2.2.1), the enzyme system luminescence intensity was studied in the presence of Fe3O4, Fe3O4/HA, and Fe3O4/APTES; at low concentrations (i.e., 10 -15 -10 -1 g/L). Correlation coefficients, r, were calculated between the dependencies of I rel Bq ( Figure 5A) and ROS rel Bq ( Figure 5B) for the corresponding MNP concentration range as reported in Table 4. Direct correlations were found, showing direct interrelations between these dependencies. The correlation values confirm interrelations between bioluminescence intensity and ROS under oxidative stress conditions, supporting the envisaged chemical mechanism of free Fe 3+ involvement (Figure 2A) to the prooxidant activity of MNPs as depicted in Schemes 1 and 2.

Effects of MNPs on Enzymatic Reactions
The effect of MNPs on bioluminescence enzymatic reaction is a question of interest. Comparison of the MNP effects on bacterial cells and their enzymatic reactions forms understanding at the cellular and biochemical levels, respectively, identifies processes most sensitive to MNP exposure, and allows supposing a role of cellular membrane-related processes in the MNP bioeffects.
Analogously to luminous marine bacteria (Section 2.2.1), the enzyme system luminescence intensity was studied in the presence of Fe 3 O 4 , Fe 3 O 4 /HA, and Fe 3 O 4 /APTES; at low concentrations (i.e., 10 -15 -10 -1 g/L). Figure 6 presents the dependence of bioluminescence intensity I rel on MNP concentration in the enzyme system. The plotted data reveal negligible inhibition of the intensity of the bioluminescence.

Effects of MNPs without Oxidative Stress (a) Effects of MNPs on Bioluminescence Intensity of Enzyme Reactions
The difference in MNP effects on the bioluminescence of bacteria and enzymes is evident from the comparison of Figures 3 and 6, respectively. In fact, the MNPs inhibited or moderately activated bacterial bioluminescence, according to Figure 3, but they did not affect the enzymatic bioluminescence as clearly analysing Figure 6. The difference might be related to the complexity of cellular processes, specifically the cell-membrane ones.

(b) Effects of MNPs on ROS Content in Enzyme System
As discussed above for the bacterial system (Section 2. 2.1 b), the ROS content was also investigated in the enzyme system using (1) MNPs in an enzyme-free medium and (2) MNPs in enzyme reactions. MNP concentrations were altered as reported in Figure 6, but none of the ROS changes were found for both types of solutions over the entire concentration range (Figures S3 and S4).

Effects of MNPs without Oxidative Stress
(a) Effects of MNPs on Bioluminescence Intensity of Enzyme Reactions Figure 6 presents the dependence of bioluminescence intensity I rel on MNP concentration in the enzyme system. The plotted data reveal negligible inhibition of the intensity of the bioluminescence. The difference in MNP effects on the bioluminescence of bacteria and enzymes is evident from the comparison of Figures 3 and 6, respectively. In fact, the MNPs inhibited or moderately activated bacterial bioluminescence, according to Figure 3, but they did not affect the enzymatic bioluminescence as clearly analysing Figure 6. The difference might be related to the complexity of cellular processes, specifically the cell-membrane ones.

(b) Effects of MNPs on ROS Content in Enzyme System
As discussed above for the bacterial system (Section 2. 2.1 b), the ROS content was also investigated in the enzyme system using (1) MNPs in an enzyme-free medium and (2) MNPs in enzyme reactions. MNP concentrations were altered as reported in Figure 6, but none of the ROS changes were found for both types of solutions over the entire concentration range (Figures S3 and S4).

Effects of MNPs in Enzyme System under Model Oxidative Stress
To verify the molecular mechanisms of redox MNP effects under oxidative stress conditions in bacteria (Section 2.2.2), the MNPs influence on luminescent enzymatic reactions under similar conditions (i.e., in solutions of model oxidizer Bq, EC50 = 5 × 10 −7 M) was studied. Similar to bacteria, the effects on bioluminescence intensity were explored in the presence of three types of MNPs-bare Fe3O4, Fe3O4/HA, and Fe3O4/APTES. We did not find any reliable deviations in I rel Bq (p < 0.05) from the control samples (without MNPs) ( Figure S5).
The ROS content was measured preliminarily in enzyme-free and enzymatic media, both, under model oxidative stress conditions, i.e., at EC50 of Bq (5 × 10 −7 M). It has been shown that Bq increases the content of ROS in enzyme-free and enzyme solutions by 110% and 80%, respectively. Such an increase is probably associated with the formation of intermediate peroxide with the occurrence of bacterial luciferase bioluminescence reaction [117], which contributes additionally to overall ROS content. Dark processes in the reaction of bacterial luciferase with hydrogen peroxide production can contribute to the ROS

Effects of MNPs in Enzyme System under Model Oxidative Stress
To verify the molecular mechanisms of redox MNP effects under oxidative stress conditions in bacteria (Section 2.2.2), the MNPs influence on luminescent enzymatic reactions under similar conditions (i.e., in solutions of model oxidizer Bq, EC 50 = 5 × 10 −7 M) was studied. Similar to bacteria, the effects on bioluminescence intensity were explored in the presence of three types of MNPs-bare Fe 3 O 4 , Fe 3 O 4 /HA, and Fe 3 O 4 /APTES. We did not find any reliable deviations in I rel Bq (p < 0.05) from the control samples (without MNPs) ( Figure S5).
The ROS content was measured preliminarily in enzyme-free and enzymatic media, both, under model oxidative stress conditions, i.e., at EC 50 of Bq (5 × 10 −7 M). It has been shown that Bq increases the content of ROS in enzyme-free and enzyme solutions by 110% and 80%, respectively. Such an increase is probably associated with the formation of intermediate peroxide with the occurrence of bacterial luciferase bioluminescence reaction [117], which contributes additionally to overall ROS content. Dark processes in the reaction of bacterial luciferase with hydrogen peroxide production can contribute to the ROS concentration, as well. The ROS content was monitored in the same media after bioluminescence measurements in the presence of Bq (5 × 10 −7 M) and MNPs recording unchanged ROS content in enzyme solutions' overall concentration ranges of MNPs (see Figure S6); no changes were found.
Hence, model oxidative stress has not revealed any reliable changes due to MNPs either in bioluminescence enzymatic intensity or in ROS content in the enzyme solutions. As for the case of absent oxidative conditions (Section 2.3.1), the enzymatic system appeared insensitive to MNP content under oxidative stress conditions, too.
Apparently, all the differences in sensitivity of the bacterial cells and their enzymatic reactions to MNPs are concerned with the cell-membrane processes of luminous bacteria. This result is in agreement with the supposed mechanism of the ferroptosis phenomenon [18,19], which considers the iron-initiated cell-membrane lipid peroxidation in the cellular systems.

Preparations of Fe 3 O 4 MNPs and Humic Acids-and Amino-Silica Functionalized Fe 3 O 4 MNPs
Full details regarding the preparation of the samples were previously described [80]. Briefly, magnetite was obtained in accordance with the Elmore reaction. So, FeCl 3 × 6H 2 O salts (Sigma-Aldrich Chemie GmbH, Steinheim, Germany) and FeCl 2 × 4H 2 O (Sigma-Aldrich Chemie GmbH, Steinheim, Germany) were dissolved in deionized H 2 O and added to 25% solution of ammonium hydroxide (Sigma-Aldrich Chemie GmbH, Steinheim, Germany) with stirring in argon at 50 • C. The formed Fe 3 O 4 nanoparticles were washed five times with Millipore water and ethanol and dried at 70 • C in a vacuum.
To prepare the aminosilica-functionalized MNPs, the Fe 3 O 4 surface modification was performed by using 3-aminopropyltriethoxysilane (APTES, (Sigma-Aldrich Chemie GmbH, Steinheim, Germany)). Fe 3 O 4 was dispersed in ethanol/water (volume ratio, 1:1) solution, and APTES was added under an argon atmosphere, at 40 • C for 2 h with a molar ratio of APTES:Fe 3 O 4 as 4:1. After cooling down to a room temperature, Fe 3 O 4 /APTES MNPs were collected with a magnet (Nd, 0.3 T), rinsed three times with ethanol and deionized water and dried in vacuum at 70 • C for 2 h.
Commercial sodium salt of humic acids (HA) (Powhumus, the total acidity of the HA was 5.3 mmol/g of acidic COOH and OH-groups, weight-average molecular weight Mw was 9.9 kD; Humintech, Grevenbroich, Germany) was used for Fe 3 O 4 /HA. For the synthesis of NPs, FeCl 3 × 6H 2 O and FeCl 2 × 4H 2 O were dissolved in 100 mL water. Upon heating at 40 • C, two solutions, namely the ammonium hydroxide (25%), and the Has were added rapidly and sequentially. The mixture was stirred at 1000 rpm at 40 • C for 10 min under an argon atmosphere and then cooled to room temperature. The black precipitation of Fe 3 O 4 /HA NPs was collected by Nd-magnet (0.3 T) and washed to neutral by distilled water (90 • C) before drying under a vacuum at 40 • C.

Characterization of the MNPs
Mössbauer spectra were obtained on an MS1104EM spectrometer (CJSC Kordon, Rostov-on-Don, Russia) with 57 Co/Rh (RITVERC JSC, St. Petersburg, Russia) with an activity of 5 mCi as a source of γ-radiation. The spectra were recorded for each sample both at room temperature (296 K) and at the boiling point of liquid nitrogen (78 K) in an evacuated cryostat. The temperature control accuracy of the samples was ±2 and ±0.5 • C, respectively. The spectra were obtained in high resolution (1024 points) with a noise/signal ratio lower than 2%. Experimental data were processed using SpectrRelax 2.8 and associated software (MSU, Moscow, Russia). The values of chemical shifts in the manuscript are given relative to α-Fe. The magnetic properties of MNP dry powders were haracterized with a Vibrating Sample Magnetometer Lake Shore (Lake Shore Cryotronics, Westerville, OH, USA) at 300 K. To measure the Fe 2+ /Fe 3+ stimuli-responsive release the obtained samples were dispersed in buffer solutions (dispersed water or 3% NaCl to simulate conditions of bioluminescence assay) containing Potassium thiocyanate (50% solution) and HCl (18.25% solution) as an indicator of Fe 3+ ions and o-phenanthroline (1% solution) as an indicator of Fe 2+ ions. At different time points (i.e., 0, 0.5, 1, 3, and 24 h), the mixtures were centrifuged to collect the supernatant. The absorbance in the regions of 490 nm or 690 nm was detected by UV-Vis-NIR spectrophotometry and then analyzed for Fe 3+ and Fe 2+ , respectively. Elemental analysis of the samples was performed using a CHNS/O elemental analyzer Vario Microcube (Elementar GmbH).

Bioluminescence Assay Systems
Effects of MNPs on microbiological and biochemical processes were evaluated using model bioluminescence assay systems, both cellular and enzymatic, i.e., luminous marine bacteria and a system of coupled enzymatic reactions of the marine bacteria, respectively.

Bioluminescence Cellular Assay
The bacterial assay, i.e., intact marine luminous bacteria Photobacterium phosphoreum, strain 1883 IBSO from the Collection of Luminous Bacteria CCIBSO 863, Institute of Biophysics SB RAS [118] was used.
For the cultivation of P. phosphoreum 1883 IBSO, the semisynthetic medium containing: 10 g/L tryptone, 28.5 g/L NaCl, 4.5 g/L MgCl 2 × 6H 2 O, 0.5 g/L CaCl 2 , 0.5 g/L KCl, 3 g/L yeast extract, and 12.5 g/L agar was used. P. phosphoreum was plated on 25 mL of the semisynthetic medium and incubated at 25 • C for 24 h (stationary growth phase corresponding to the maximum bioluminescence) in an incubator (WIS-20R, WiseCube Laboratory Instruments, Wertheim, Germany). Prior to experiments, bacteria were collected by pipetting of 3% NaCl solution directly onto the agar to release bacteria. The 3% NaCl solution was used to imitate the marine environment for bacterial cells and to balance osmotic processes. The bacterial suspension was diluted to Abs 660 = 0.025 and stored at 4 • C for 30 min to allow the bioluminescence stabilization. The reagents for bacterial cultivation were: tryptone and yeast extract from Dia-M, Moscow, Russia; sodium chloride (NaCl) from Khimreactiv, Nizhny Novgorod, Russia; magnesium chloride hexahydrate (MgCl 2 6H 2 O), calcium chloride (CaCl 2 ), and potassium chloride (KCl) from Pancreac AppliChem GmbH, Darmstadt, Germany; agar from Difco Laboratories, Detroit, MI, USA. The reagents were of chemical or analytical grade.
Sodium chloride (NaCl) from Khimreactiv, Nizhny Novgorod, Russia was used to prepare a 3% NaCl solution.

Bioluminescence Enzymatic Assay
The enzymatic assay, i.e., enzymatic preparation was based on the system of coupled enzyme reactions catalyzed by NADH:FMN-oxidoreductase from Vibrio fischeri (0.15 a.u.) and luciferase from Photobacterium leiognathi at 0.5 mg/mL [118]. The enzyme preparation was produced at the Institute of Biophysics SB RAS, Krasnoyarsk, Russia. Enzyme reactions (1) and (2) are presented in the Introduction.
The used chemicals were FMN and tetradecanal from SERVA, Heidelberg, Germany; NADH from ICN Biochemicals, Costa-Mesa, CA, USA; all reagents were of chemical or analytical grade.

Experimental Data Processing
1. The toxic (for bacterial assay) and inhibitory (for enzymatic assay) effects of MNPs on the bioluminescent systems were characterized by the relative bioluminescence intensity, I rel : where I contr and I MNP are the maximum bioluminescence intensities in the absence and presence of MNPs, respectively. Values of I rel were determined at different concentrations of MNPs. The dependence of I rel vs. MNP concentrations was studied and plotted.
Additionally, to characterize the toxic effects of MNPs on the bacteria, their effective concentrations that inhibited the luminescence intensity by 50% (I rel = 0.5), EC 50 , were determined and compared.
2. The anti-(pro-)oxidant activity of MNPs on the bacterial and enzyme systems was assessed under the conditions of oxidative stress. The methods for determining the anti-(pro-)oxidant activity were elaborated on and developed previously by different authors [67,71,73,75,77,80].
To create conditions of the model oxidative stress, we used an organic oxidizer, 1,4benzoquinone (Bq) (Sigma-Aldrich, St. Louis, MO, USA). The Bq was prepared in 3% NaCl solutions and in distilled water for bacterial and enzymatic assays, respectively. The Bq concentrations that inhibited the bioluminescence intensity of bacterial and enzymatic systems by 50% were applied in the experiments (10 −7 M and 5 × 10 −7 M, respectively). To study and compare changes in the toxicity in the Bq solutions with the addition of MNPs, the antioxidant coefficients I rel Bq were determined as follows:

Luminol Chemiluminescence Assay
We used the luminol chemiluminescence method to evaluate the content of reactive oxygen species (ROS) in the experimental bacterial suspensions, enzymatic solutions, and non-biological systems [119,120]. This technique was used to determine the integral content of ROS assuming that there was a dynamic equilibrium between different ROS forms.
Stock luminol solution (10 −2 M) was prepared as follows: luminol powder was dissolved in 5 mL in a 1N solution of KOH and then, 5 mL of distilled water was added. The chemiluminescence luminol reaction was initiated by an alkaline luminol solution; the maximum value of chemiluminescence intensity was determined. The concentrations of the alkaline luminol solution and aquatic solutions of K 3 [Fe(CN) 6 ] in the experimental samples were 10 −4 M and 10 −3 M, respectively. Initially, the dependences of the chemiluminescence intensity on the concentration of H 2 O 2 (Tula Pharmaceutical Factory, Tula, Russia), were determined in distilled water and 3% NaCl solution for the enzymatic and bacterial luminescence systems, respectively; they were used as calibration dependences to evaluate the ROS content in all the experimental samples.
To verify the role of ROS in the biological effects of MNPs, chemiluminescence intensities were measured in the bioluminescence assay systems (bacterial and enzymatic), as well as in bacteria-free/enzyme-free aqueous solutions. The relative values of ROS content (ROS rel ) were calculated as ratios of the ROS content in the experimental solutions to that in the control solutions without MNPs: The ROS rel values were obtained at different concentrations of MNPs (10 −15 -1 g/L) in the absence and presence of Bq at EC 50 (see Section 3.3.3). The chemiluminescence assay was carried out immediately after the bioluminescence measurements in the bacterial and enzymatic samples.
The values of ROS rel were plotted vs. the concentration of MNPs.

Preparation of the MNP Suspensions for the Bioluminescence Analyses
Solid samples were ground in a mortar and dissolved in a 3% aqueous NaCl solution (for bacterial assay or non-biological systems) or distilled water (for enzymatic assays or non-biological systems). To obtain homogeneous suspensions, the stock suspensions were exposed to ultrasound for 10 min. The MNP solutions of different concentrations (10 −15 -10 g/L) were prepared from the stock suspensions, with the additional 5 s ultrasonic treatment at each dissolution stage. Before the measurements, the dispersions were additionally treated ultrasonically for 5-10 s.
To exclude the effect of the optic filter in the bio-(chemi-)luminescence measurements, the suspensions with optical density higher than 0.1 (at 490 nm and 425 nm, respectively) were excluded. Hence, the effect of the "optic filter" [104] did not skew the results of the bio-(chemi-)luminescence measurements.

Equipment
Measurements of the bio-(chemi-)luminescence intensities were carried out with a bioluminometer Luminoskan Ascent (Thermo Electron Corporation, Solon, OH, USA) equipped with an injector system. All the luminescence measurements were taken at 25 • C and without pre-incubation.
Absorption spectra of the bacterial and MNP suspensions were recorded using a UVIKON 943 Double Beam UV/VIS Spectrophotometer (Kontron Instruments, Milan, Italy).

Statistical Processing
All bio-(chemi-)luminescence measurements were conducted in five replicates for all the control and MNP solutions.
To reveal correlations between the bioluminescence intensity and ROS content, the statistical dependence between the rankings of two variables was analyzed [121], and the correlation coefficients r were calculated.
Statistical processing of the results of bioluminescence and chemiluminescence assays was carried out; p-values were calculated with GraphPad Prism 8 using ANOVA. The p-values were assessed by the Kruskal-Wallis test of two independent sample distributions.

Conclusions
In the current paper, we used bioluminescence bioassays, cellular and enzymatic, to evaluate the bioeffects of three types of magnetite nanoparticles (MNPs), namely, the pristine bare Fe 3 O 4 , humic acid, and silica (3-aminopropyltriethoxysilane) modified nanoaparticles. Additionally, the toxic and anti-(pro-)oxidant bioeffects of MNPs were analyzed assessing the role of ROS. Bacterial bioluminescence was applied as a signal physiological parameter and the ROS content was evaluated by chemiluminescence luminol assay. Higher toxicity related to bare NMPs, Fe 3 O 4, was revealed, with a minimum effective concentration of EC 50 at 0.05 g/L. The inhibition ability of modified nanoparticles at higher concentration intervals (i.e., 10 −2 -1 g/L) was associated with ROS accumulation in the bacteria suspensions, but the effect of bare Fe 3 O 4 appeared to be more complex and it should be further investigated. Wide MNP concentration ranges (10 −15 -1 g/L) have revealed a higher ROS content in bacterial suspensions compared to bacteria-free media, and this supports the conclusion that ROS production is enhanced by bacteria in the presence of MNPs. Under model oxidative stress conditions (i.e., in the presence of model organic oxidizer 1,4benzoquinone) and higher concentrations of MNPs (>5 × 10 −5 g/L), the bacterial bioassay has suggested a pro-oxidant activity of all three types of MNPs, with a corresponding decay of ROS content. The ROS decay can be due to shifting the redox balance in the system, such as the reducer-oxidizer model (i.e., 1,4-hydroquinone-1,4-benzoquinone). Bioluminescence enzymatic assay did not show sensitivity to MNPs, and ROS content was unchanged at all applied concentrations of MNPs (<10 −1 g/L), involving model oxidative stress conditions. The result probably indicates that cell-membrane processes are responsible for the MNP bioeffects and ROS generation by the bacteria, according to the supposed mechanism of ferroptosis phenomenon, which considers the iron-initiated cell-membrane lipid peroxidation in connection of ROS production within the cells.
Author Contributions: Data processing, interpretation, data analysis, writing the manuscript, A.G.K.; investigation, writing-review and editing, A.G.K. and E.S.S.; investigation, software, formal analysis, L.S.B., D.A.P. and A.A.D.; data curation, G.I.D. and N.S.T.; conceptualization, general leadership of the work, data analysis, writing-original draft preparation, writing-review and editing M.Z., K.A.K. and N.S.K. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.