**1. Introduction**

Fluorescent gold nanoclusters (AuNCs) embedded in bovine serum albumin (BSA) have been extensively studied in the literature since 2009, following the publication of Xie et al. [1] However, many open questions remain unsolved, especially for the witnessed dependence of the fluorescent properties with the sizes of Au nanostructures. [2–42]. The discussion remains controversial because many reports describe fluorescent AuNCs where the origin of such phenomenon is linked to the small size of the nanoparticles (NPs) (e.g., Burt et al. [37], Zhang and Wang [38], Zheng et al. [39]), while other reports sugges<sup>t</sup> that the key to the fluorescence properties is intimately associated to the oxidation state of the gold cluster, Au(III)- [3] and/or Au(I)-complexes [40]. In Table 1, examples of selected BSA-AuNCs preparations are given, including their synthetic parameters, as reported in literature, in conjunction with the AuNCs size, oxidation state of Au when known, and maximum (maxima) of their fluorescence emissions.

**Citation:** Ostruszka, R.; Zoppellaro, G.; Tomanec, O.; Pinkas, D.; Filimonenko, V.; Šišková, K. Evidence of Au(II) and Au(0) States in Bovine Serum Albumin-Au Nanoclusters Revealed by CW-EPR/LEPR and Peculiarities in HR-TEM/STEM Imaging. *Nanomaterials* **2022**, *12*, 1425. https://doi.org/10.3390/nano 12091425

Academic Editors: Wei Chen and Derong Cao

Received: 28 March 2022 Accepted: 19 April 2022 Published: 22 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).


**Table 1.** Details of selected BSA-AuNCs preparations and properties as reported in literature.

Notes: n.a. = not available, \* = sizes obtained after the reduction performed by addition of NaBH4 to Au(III)-BSA systems which were generated at given pH values, MW = microwave radiation heating, XPS = X-ray photoelectron spectroscopy, FQ = fluorescence quenching, TEM = transmission electron microscopy, HAADF-STEM = high-angle annular dark-field scanning transmission electron microscopy, HR-TEM = high resolution transmission electron microscopy.

From the size perspectives, AuNCs are considered intermediate systems, falling between dimensions of isolated Au atoms (0D, zero dimensional) and Au nanoparticles (3D dimensional). The fluorescent properties emerge and are tuned by the presence of discrete energy levels that mirror the nanoparticle size variance, which approach the Fermi wavelength of free electrons (e.g., [38,43]). More detailed descriptions of different types of luminescent Au nanostructures and the related emission mechanisms, including not only the free-electron model, but also surface ligand effects, are thoroughly discussed in recent reviews (e.g., [39,44–46]).

The oxidation state of Au is undoubtedly an important parameter to unveil for understanding the physical basis of the fluorescence phenomenon. [44] In this context, analysis of the fluorescent properties of Au clusters entrapped by proteins, such as bovine serum albumin (BSA), is particularly debated. Mostly, Au(0) and Au(I) oxidation states are observed and screened by XPS (X-ray photoelectron spectroscopy) in BSA-AuNCs [1,2,4,15,16,30,32,34–36]. The pseudo-polymeric structure, such as SR-(Au-SR)x- (x = 1 or 2) motif, is envisioned to be present, and from XPS results the authors [2] demonstrated that Au-Au and Au-S bonds form in the BSA-AuNCs system. It should be noted that the occurrence of the staple motif arrangemen<sup>t</sup> (-S-Au(I)-S-Au(I)-S-) determined in BSA-AuNCs (as in e.g., [7]) shares similarities with the arrangemen<sup>t</sup> observed in thiol-protected Au25(SR)18 clusters prepared by Zhu M. et al. [47,48]. Based on XRD analysis, a core-shell structure of the thiol-protected Au25 cluster was determined in [47]. The same group of researchers revealed the intrinsic magnetism of these thiolate-protected Au25 super atoms in 2009 [49]. On the contrary, the magnetic properties of BSA-AuNCs have not been investigated so far, to the best of our knowledge. Therefore, the presence of such a property represents one of the aims tackled in this study.

Importantly, the oxidation state of Au in the final AuNCs is closely related to the experimental conditions. There are obvious differences in the presence and/or absence of a strong reduction agen<sup>t</sup> in the course of AuNCs formation as shown in Table 1. In synthetic procedures where NaBH4 is used, BSA represents solely a template (matrix) that prevents the occurrence of coalescence processes in AuNCs at room temperature. On the contrary, when the use of NaBH4 is avoided (as in our previous work [41]), BSA acts as a reducing agen<sup>t</sup> (probably due to tyrosine residues [1] at increased pH above its pKa value) and, simultaneously, as a capping agen<sup>t</sup> for AuNCs. Interestingly, even syntheses exploiting solely BSA as the reducing agen<sup>t</sup> can lead to different fluorescent properties of resulting AuNCs, as shown in Table 1. The reason for such differences in fluorescent characteristics may be either the type of heating (incubator at 37 ◦C vs. MW heating), or the initial concentration of BSA—see Table 1.

Several literature reports that describe organometallic complexes containing Au(I) sites show the existence of strong Au–Au bonds and the term "aurophilicity" was therefore introduced in 1989 to describe the unusual bond properties [50]. Aurophilic interaction is thought to be as strong as a hydrogen bond and shows the presence of bond length as shorter than the sum of two van der Waals radii of gold (3.80 Å). This bond characteristic is suggested to be one of the relevant factors promoting the fluorescence effect of Au(I) complexes in the UV-vis region, when investigated in the solid state [50]. The fluorescent property may then vanish in the solution, but there are scenarios where the luminescence is restored, at least for high concentrations, when a suitable solvent is used [51,52] and/or increases in the solution when AuNCs undergo self-assembly into nanoribbon structures [53]. Importantly, many Au(I) ions, being complexed with thiols and/or compounds containing other functional groups, such as carboxylates and amides, show fluorescent effects when they are partially ordered (from aurophillic interactions) [51,52]. This might also be the case of Au(I) in BSA-AuNCs.

Concerning another oxidation state of Au (besides Au(I) and Au(0)) being evidenced in BSA-AuNCs, however implicitly (i.e., based on additional chemical reduction using NaBH4 and not directly by XPS), Dixon and Egusa [3] reported the presence of a significant amount of Au(III) in 2018. Terminologically, they label the systems as BSA-Au(III) complexes, but, simultaneously, they admit that there is no disproval of the existence of neutral Au(0) NCs in BSA-Au samples [3]. They [3] summarized that the red fluorescence that emerged due to Au(III) cations complexed with BSA at or above the pH value (9.7 +/− 0.2), which was responsible for the conformational change of BSA structure from normal (N) to aged (A) conformation. Therefore, they [3] concluded that upon BSA conformational change, some Cys–Cys bonds were expected to be sufficiently exposed and solvent-accessible and

might interact with Au(III). They also experimentally proved multiple specific Au(III) binding sites in BSA and postulated that the UV-excitable red fluorescence of the BSA-Au(III) complex is due to the internal and potentially cascaded energy transfers among chromophores [3]. Alternatively, they admit strong electron delocalization as a source of the red fluorescence [3].

Furthermore, based on Table 1, it is obvious that sizes of AuNCs, the oxidation state of Au, and consequently their luminescent properties are strongly dependent, among others, on the exact conditions of their synthesis. Moreover, purification (if applied), storage conditions, and time elapsed before the characterization of AuNCs is completed, may represent other factors influencing the fluorescence of the final NCs as demonstrated in our previous work [41]. Obviously (Table 1), visualization of tiny AuNCs is usually performed by TEM and HR-TEM (high-resolution transmission electron microscopy). Particularly, STEM (scanning transmission electron microscopy) using High-Angle Annular Dark Field (HAADF) mode is necessary to be employed in the case of BSA-AuNCs [37]; otherwise, the TEM resolution is obscured by the enveloping protein (BSA) and the accurate size distribution cannot be obtained as stated in [2]. Interestingly, in many research papers (e.g., [3,4,42,54]), the TEM images of AuNCs exceed the size characteristic for their fluorescent properties (maximum of 2 nm in diameter). Besides HR-TEM/STEM, the determination of BSA-AuNCs sizes is often based on mass spectrometry measurements (MS) [55,56]. This is an indirect method where the MS size of BSA is subtracted from the MS size of the BSA-AuNCs system. However, this leads to the determination of the number of Au atoms per BSA. Such estimation does not provide any information about the real size of AuNCs because Au atoms are not localized on a single place in BSA; indeed, they can be spread on different cysteine and other residues of the protein [3,41]. Therefore, it seems that direct visualization of AuNCs as well as determination of individual AuNCs sizes within BSA represent challenging tasks.

As above mentioned, the magnetic properties of BSA-AuNCs have not been investigated so far and, therefore, it is one of the aims of the present study. From the analysis of literature results [57–66], many NPs (mostly prepared using thiolate residues) comprising engineered Au(0) metal clusters give a wide range of magnetic behaviors (diamagnetic, Pauli paramagnetism, antiferro-, and ferromagnetic responses) that depend on the system size (1.8–4.4 nm), the metal's local symmetry, metal coordination environments, and unbalanced/charged states (neutral, positive, and negative metal–organic clusters). For example, the 25 gold (Au(0)) atoms NC stabilized by 18 thiolate ligands prepared by Zhu and co-authors (Au25(SR)18, with R = phenylethyl) shows, in frozen CH2Cl2/toluene solution at *T* = 8 K an EPR (Electron Paramagnetic Resonance) signal characteristic for *S* = 1 2 systems with *g*-values consistent with an orthorhombic symmetry at 2.56(x), 2.36(y), and 1.82(z) [49]. Similarly, Agrachev and co-authors reported the phenylethanethiolate Au complex (neutral), Au25(SC2Ph)18, encoding a radius of 13.2 Å, which also exhibits EPR signatures associated to spin *S* = 1 2 . The EPR spectrum of the complex at *T* = 5 K is rather broad (~200 mT) for Au25(SR)18 and anisotropic, with *g*-tensor components at 2.53 (x), 2.36 (y), and 1.82 (z) [67].

In this report, the authors used CW-EPR (Continuous Wave EPR), light-induced EPR (LEPR), HR-TEM, and STEM to thoroughly investigate the formation of BSA-AuNCs in a solution, which were intentionally prepared by Xie's type of synthesis [1]. So far, EPR has been employed for the investigation of the emerging magnetism of atomically precise AuNCs of different sizes (e.g., Au25(SR)18, Au133(TBBT)52), which were prepared by a multistep synthesis involving NaBH4 as a reduction agen<sup>t</sup> [5,6,49]. The EPR/LEPR study of BSA-AuNCs (further labelled as AuBSA in the present study to be directly distinguished from the other published results for BSA-AuNCs) is more complicated than the previous EPR studies of atomically precise AuNCs [5,6,49] because BSA alone reveals photochemically induced radicals' generation, which results in damages of the BSA structure and oxidation of cysteine and tryptophan residues [68,69]. Several years ago, Lassmann and coworkers [68] analyzed, using the LEPR technique, the formation of C and S radical centers

in BSA by UV-photolysis, especially thiyl (R-CH2-S•) and perthiyl radicals (R-CH2-S-S•), revealing that radiation damages are produced in the amino acid chain in a dose-dependent manner, i.e., the number of radicals increases by increasing the UV light irradiation time. The authors found that in bovine serum albumin (BSA) the thiyl radical (R-CH2-S•) exhibits axial anisotropy, with *g*// = 2.17, *g*⊥ = 2.008, <sup>Δ</sup>*B*pp = 1.8 mT, pH = 7, *T* = 80 K) and the cysteine × HCl radical (phosphate buffer, pH = 3) shows similar features, with *g*// = 2.11, *g*⊥ = 2.011, <sup>Δ</sup>*B*pp = 3.2 mT, *T* = 80 K). The *g*// component was always found to be very broad and weak, to the point that it becomes barely detectable, while the *g*⊥ component gives the dominating, most intense signal. Together with the thiyl radicals, other types of radical species were demonstrated to form as well, which include perthiyl centres (R-CH2-S-S•) that exhibits rhombic character, and likewise, C-radical centers. In BSA (pH = 7, *T* = 80 K), EPR signals for the perthiyl specie show resonances at *g*1 = 2.057, *g*2 = 2.027, and *g*3 = 2.002, while the C-radical species have *g*avg ~ 2.000 [68].

Therefore, the main aim of this study is to thoroughly investigate the properties of the AuBSA system with the aid of selected techniques (EPR/LEPR, HR-TEM/STEM). The following key findings are presented: (i) redox processes take place upon entrapment of Au(III) in BSA, in which the protein backbone, in absence of other reducing equivalents, provide the electrons needed for Au(III) reduction; (ii) the Au uptake process by BSA in the solution occurs on two types of sites, i.e., cysteine (Cys) and oxygen/carboxylate donor residues (e.g., tyrosine, Tyr); (iii) the magnetic interaction among cationic forms of Au arises from an admixture of Au(0), Au(I), and Au(II); and (iv) UV irradiation of the AuBSA system induces irreversible damages, and radicals are generated in BSA upon UV light irradiation. It is disclosed that no energy transfer between entrapped Au clusters and the BSA protein occurs under photoexcitation, thus the photoluminescence phenomenon is governed by the intrinsic properties of the Au system (its size and the presence of various spin active redox states) and both variables equally matter. On the other hand, in view of the extended application of these materials in biomedical scenarios, correlation between their experimental parameters (e.g., concentration in solution) and cell's toxicity (with and upon photoexcitation) should be analyzed with grea<sup>t</sup> caution, because observation of an enhancement of toxic effects on cells by light irradiation might be hampered by light-induced radical's formation on the protein itself and not solely to the photoexcited states of the gold nanoassembly.

## **2. Materials and Methods**
