1. Introduction
The role of microorganisms in the biogeochemical cycle of selenium (Se) has been established [
1], although the vast genetic diversity of bacteria makes it difficult to fully elucidate the biological mechanisms behind the biochemistry of one of the most abundant and toxic Se species: the oxyanion selenite (SeO
32−) [
2]. Since the beginning of the 20th century, a variety of microorganisms have been described for their ability to grow in the presence of Se oxyanions and bioprocess them into the less bioavailable elemental form (Se
0) [
3]. In the 1970s, this phenomenon started to be linked to the microbial capability of biosynthesizing Se nanostructures (SeNSs) with defined size and shape [
4]. The simultaneous development of the nanotechnology field in terms of new synthetic procedures, nanomaterial (NM) characterization, and potential applications [
5] resulted in an increased scientific focus on the possibility of using microorganisms as green and inexpensive catalysts to produce SeNSs [
6], reaching its peak in the last 20 years in terms of dedicated research, whose interest was more devoted to investigating the mechanisms behind Se oxyanion bioprocessing than studying potential optimization of NM biosynthesis processes. To date, (1) Painter-type reactions involving thiol (RSH) groups [
7,
8,
9], (2) enzymatic reduction by periplasmic or cytosolic oxidoreductases [
10,
11,
12,
13,
14,
15,
16], (3) inorganic reactions with microbial metabolites [
17], and (4) redox reactions mediated by siderophores [
18] are the four strategies mainly acknowledged as able to achieve microbial processing of SeO
32−. However, no common mechanism has yet been identified for Se oxyanion biotransformation in bacteria, which instead depends more on the bacterial species investigated as well as the diverse physiological state of microorganisms.
The inherent complexity of bacteria also complicates the design of processes to produce SeNSs as controllable and predictable as chemogenic NSs, highlighting the necessity to study the mechanism of their biosynthesis. The choice of the organism to be used as the microbial cell factory, metal(loid) precursor concentration, pH, temperature, bacterial incubation timeframe, cell physiology, and localization of the precursor reduction events are parameters that must be considered when studying the biogenic production of SeNSs [
19,
20,
21], as variations in these conditions can determine the physical-chemical characteristics. Among these features, morphology and size are crucial factors for NM applications [
5], as they directly affect several fundamental properties of material on the nanoscale (e.g., electrical and optical features, potential toxicity or cellular uptake for medical applications) [
5,
22,
23]. A key aspect of biogenically synthesized SeNSs is the presence of an organic material derived from the bacterial systems used, which seems to confer a naturally high degree of thermodynamic stability toward these NMs [
24,
25]. The function(s) and the composition of this organic material, as well as its variation upon changes in bacterial growth conditions, are not completely understood yet, constituting a black hole in the microbial nanotechnology field.
In the present study, we explored how the environmental isolate
Ochrobactrum sp. MPV1 can tolerate high concentrations of SeO
32−. This strain has been previously described for its ability to biosynthesize Se nanoparticles (NPs) and nanorods (NRs) through SeO
32− bioconversion [
26,
27]; thus, it was investigated for the removal of different SeO
32- loads under different conditions to better understand the biomolecular process(es) behind this biotransformation. Metabolically controlled growth conditions were subsequently used to optimize the tuning of SeNS morphology previously observed [
27], and all the recovered biogenic NSs were characterized, focusing on size and shape variations. Finally, the new insights presented in this study regarding composition, physical-chemical features, and role of the organic material enclosing SeNSs recovered from MPV1 cells revealed its paramount importance for the thermodynamic stabilization of biogenic NMs, making their coating with stabilizing agents typically required to prevent the aggregation of those chemically produced unnecessary.
3. Discussion
The investigation conducted to unveil potential mechanism(s) exploited by MPV1 to cope with increasing concentrations of SeO
32− (0.5–10 mM) highlighted the growth and oxyanion removal rates (
Figure 1a,b,
Table 1) comparable to those described for most SeO
32− tolerant bacteria [
3,
29,
30,
31,
32,
33,
34,
35]. Since Se oxyanions exceeding 2.5 mM reappeared in the growth medium upon exposure to 3, 5, and 10 mM SeO
32− (
Figure 1b,
Table 1), 2.5 mM SeO
32− appears to be the threshold concentration biotically processed by MPV1 cells under these experimental conditions, as also observed in the case of
Moraxella bovis [
36]. This evidence indicates that the bioprocess of SeO
32− by MPV1 might involve: (1) the uptake of increasing concentrations Se oxyanions, (2) their bioaccumulation and bioconversion up to 2.5 mM, and (3) a gradual release of exceeding SeO
32− amounts. This last step could be due to either cell lysis events, however unlikely, as similar death events were observed in bacterial cultures incubated with all SeO
32− concentrations (
Figure 1a), or a saturation of the cellular systems responsible for SeO
32− removal, which led to the release of oxyanions to reach a sort of equilibrium between the intra- and extra-cellular environments [
34].
The high level of RSH oxidation measured in MPV1 cells exposed to 0.5 mM SeO
32− (
Figure 1c) indicates a major involvement of these reactive groups for SeO
32− removal. Other cellular systems (i.e., enzymes) seemed to be involved in the bioprocessing of SeO
32− concentrations exceeding 0.5 mM, as suggested by (1) the ability of MPV1 cells to biotically remove ca. 2.5 from 10 mM SeO
32− after 168 h of incubation (
Figure 1b,
Table 1), and yet (2) the RSH levels recover toward later incubation times defining a low level of sustained oxidized RSHs (
Figure 1c), and (3) their minor contribution to the oxyanion conversion as function of SeO
32− concentration, as depicted by the linear relationship observed in
Figure S1. The presence of an inhibitor for glutathione (GSH) synthesis,
S-n-butyl homocysteine sulfoximine (BSO), only slightly affected the biotic removal of 2 mM SeO
32−, revealing only a six-hour delay in the process [
26]. Thus, the key role played by GSHs in MPV1 cells is to bioconvert Se oxyanions, yet ancillary enzymatic mechanism(s) can be induced as function of SeO
32− concentration and time of exposure. Ubiquitous enzymes, like NAD(P)H-dependent thioredoxin reductases and flavin oxidoreductases, sulfate or sulfite reductases, or fumarate reductases, were identified as responsible for the biotic reduction of high concentrations (from 2 to 10 mM) of SeO
32− [
30,
32,
33,
34,
35,
37]. In this regard, NADPH-dependent reduction activity toward high concentrations (5 mM) of SeO
32− was found in the cytoplasmic and, to a minor extent, in the periplasmic fractions of MPV1 cells [
26]. SeO
32− bioprocessing can also be mediated by intracellular SeO
32− reductases [
34,
38], lignin peroxidase [
39], chromate (CrsF), ferric (FerB) and arsenate reductases (ArsH) [
37], or the metalloid-selective channel porin ExtI [
40]. Thus, enzymatic systems might be accountable for the bioconversion of high oxyanion concentrations in MPV1, whereas low amounts of SeO
32− are likely bioprocessed through Painter-type reactions.
Regardless of the initial concentration of SeO
32− precursor, MPV1 biosynthesized SeNPs as the main product of Se oxyanion bioconversion (
Figure 2,
Figure 3,
Figure 4 and
Figure S2). The process behind the formation of NSs relies on a number of parameters (i.e., precursor concentration, reducing agent, reaction time, the concentration of elemental atoms) that influence the rate of growth, morphology, and size of NMs [
41,
42]. Due to the complexity of a biological system, the type of cell factory and the localization of precursor reduction events must be accounted for by NS biosynthesis, as they directly influence the concentration of metal atoms available for NM formation. Previous reports showed that the reduction of SeO
32− occurred in the cytoplasm of MPV1 [
26,
27], leading to the confinement of many Se atoms in the small cellular volume, increasing the chances to exceed the critical level of these atoms to form Se nuclei [
43], which eventually grow as NPs. Thus, the MPV1 intracellular environment can improve the synthesis of SeNSs even at low concentration of Se atoms with respect to chemogenic procedures.
Overall, NMs synthesized by microorganisms generally feature high polydispersity in size [
25], which mostly depends on the uneven distribution of the metal(loid) precursor within the cells during bacterial growth, resulting in the accumulation of different intracellular concentrations of elemental atoms, which can determine diverse NS production rates [
43]. However, despite the different growth conditions tested, the average diameter of biogenic SeNPs was always between 90 and 140 nm (
Table 3), indicating a good monodispersity in size, in line with most studies reported to date [
44]. Although NPs are classically defined as particles having a diameter between 1 and 100 nm, the unique physical-chemical properties of these biogenic Se-structures [
27] and the proximity of their size with the range in question allow them to be considered as NPs, accordingly to some of the definitions coined to date for these NMs [
45,
46]. The monodispersity of biogenic SeNPs may indicate their natural stability within this range size due to the existence of an organic material composed of biomolecules produced by bacterial cells that participate to control NP diameter [
47,
48]. The close association of SeNPs with the organic material was further supported by SEM imaging, which highlighted the presence of a matrix composed of light elements (
Table 4) and enclosing SeNPs (
Figure 2,
Figure 3 and
Figure 4). TEM micrographs revealed the occurrence clusters of NPs in SeNPs
MPV1-0.5_120_e and SeNPs
MPV1-2_120_e (
Figure S2a,b), likely caused by the high bioprocessing rate of low SeO
32− concentrations. Since any significant difference was not observed in the growth profile of MPV1 cells upon exposure to diverse oxyanion concentrations (
Figure 1a), the bacterial incubation with 0.5 and 2 mM SeO
32− corresponded to the highest precursor (SeO
32−)-to-reducing agent (RSHs and enzymatic systems) ratio, which mediated the fastest oxyanion bioprocessing observed (
Figure 1b and
Table 1). This would result in the buildup of a high concentration of Se atoms over a short period of time, causing the rapid formation of SeNPs and their eventual agglomeration [
49,
50] in the intracellular environment, even though their complete aggregation was prevented by the presence of the organic material. The low extent of oxyanion bioprocessing under MPV1 exposure to either 5 or 10 mM (
Figure 1b,
Table 1) led to a decreased amount of Se atoms available for NP synthesis over the time period [
20], preventing the detection of big clusters within TEM micrographs (
Figure S2c,d).
Previous studies concerning the characterization of biogenic SeNSs showed the existence of an organic material playing a key role in their synthesis and stabilization [
6,
44]. Over the past few years, FTIR spectroscopy has been the most-used technique to assess the presence of biomolecules associated with SeNSs, enabling the detection of proteins, carbohydrates, and lipids within most of the extracts analyzed [
20,
32,
37,
47,
51,
52,
53,
54,
55,
56,
57], including those recovered from MPV1 cells grown under optimal conditions [
26]. Here, the detection of light elements attributable to biomolecules co-produced by the bacterial strain alongside Se (
Table 4) highlighted a certain degree of variability among the biogenic NSs, likely due to the exploitation of multiple strategies by MPV1 to remove Se oxyanions [
20,
32]. The detection of N in some cases might be ascribed to the occurrence of proteins or metabolites within the biogenic extracts [
26], whereas the constant presence of S signal may be due to the involvement of RSHs in SeO
32− bioprocessing for MPV1 cultures [
20,
21,
58]. The narrow size distributions of the organic material (
Table 5) suggested that it mostly contained amphiphilic biomolecules able to form nanosized aggregates (e.g., micelles and vesicles) when suspended in aqueous solution [
28,
59]. The low PdI values indicated the ability of these biomolecules to form monodisperse structures [
60]. Since Se does not have a net charge in its elemental state (Se
0), the negative ζ values (
Table 6) may indicate that negatively charged biomolecules were part of the biogenic extracts, whose charges can be attributed to the presence of either carboxyl (–COO
−) or phosphate (–PO
42−) functional groups [
28,
61]. Although similar in elemental composition, the biogenic extracts recovered from MPV1 cells grown under metabolically controlled conditions showed ζ values closed to neutrality (
Table 6), potentially indicating differences in terms of biomolecular composition, depending on the metabolism exploited by MPV1 to cope with Se oxyanion toxicity. The different bacterial physiological states determined morphological changes of SeNSs (
Figure 5), resulting in the production of both NPs and NRs, also observed in the case of
Shewanella sp. HN-41 [
19],
Lysinibacillus sp. ZYM-1 [
20], and
Rhodococcus aetherivorans BCP1 [
28]. This phenomenon can be ascribed to the bivalent nature of Se, as once amorphous NPs are formed, they can spontaneously dissolve and release Se atoms [
62], which might precipitate as nanocrystallinites and grow in one direction to attain a more thermodynamic stable state, allowing NRs to form [
63]. This process is favored by the co-presence of amphiphilic molecules (e.g., surfactants having a bulky structure) that can act as templates to guide the deposition of Se atoms and their growth in one direction [
64]. In this regard, the synthesis of biosurfactants was earlier reported for
Ochrobactrum genus bacterial strains when grown under stress conditions [
65], whereas the shift from SeNPs to SeNRs was previously observed in MPV1-glucose grown cells [
27]. Here, this change in NS morphology was emphasized due to the different MPV1 pre-culturing conditions, and cells also thriving under pyruvate and SeO
32− pressure-produced SeNRs (
Figure 5c,d), suggesting a direct influence of the bacterial physiology on the biosynthesis of different nanomaterial morphologies. Based on both the evidence collected here and previous studies [
26,
27], a putative mechanism illustrating SeO
32− bioprocessing and SeNS production by MPV1 is proposed in
Figure 7.
The biomolecules present in the extracts are also responsible for the thermodynamic stability of biogenic SeNSs, as indicated by the formation of insoluble Se precipitates upon physical removal of the organic material. This conclusion was further supported by the slight effect of the temperature on both surface charge and d
H of SeNPs
MPV1-0.5_120_e, as opposed to
l-cys SeNPs (
Figure 6), whose electrostatic stabilization was completely lost within 15 days. This phenomenon may be due to the overall development of electrostatic (charged moieties) and steric (bulky amphiphilic molecules) interactions between the organic material and the SeNSs within the biogenic extracts, generating the electrosteric stabilization effect [
25,
28,
52,
61], which is used to strongly stabilize chemogenic NMs [
64].