Spontaneous Assembly of Gold Nanostars onto Magnetotactic Bacteria: Morphological Characterization and SERS Response of Magneto-Plasmonic Biohybrids

Abstract

Magnetotactic bacteria (MTB) are a distinct group of microorganisms that naturally biomineralize nanoparticles in the form of magnetite (Fe₃O₄) or greigite (Fe₃S₄). These nanocrystals, known as magnetosomes, give MTB natural magnetic properties and the ability to orient precisely in external magnetic fields. In parallel, gold nanoparticles are extensively used in biomedicine due to their high biocompatibility, chemical stability, and the possibility of functionalization with various biomolecules. The exposure of MTB to plasmonic nanoparticles is only now beginning to be seen as a promising avenue of research, one that could pave the way for the self-assembly of magneto-plasmonic structures. In our study, the combination of MTB (MS-1 strain) with symmetrical gold nanostars (AuNS) or silica-coated gold nanostars (AuNS-SiO₂) was initially conceived as “possible interaction”, but it was unclear whether the two would “recognize” each other. As expected, scanning electron microscopy (SEM) analysis showed that there is indeed a visible association. The nanostars were observed at multiple locations on the bacterial surface, including regions corresponding to the cell body and polar areas, suggesting a more complex anchoring preference than we had anticipated. The AuNS strongly enhance the Raman spectra of MTB via surface-enhanced Raman scattering (SERS) and provide better image contrast for MTB in SEM imaging. This shows that the use of gold nanostars is not only effective but also opens up new horizons for exploring the optical and structural properties of MTB. Such observations provide the foundation for future research on the spontaneous assembly of MTB–plasmonic nanoparticles as potential components in biosensing and bioimaging platforms. The SERS method could provide a more sustainable and efficient framework for combining MTB with gold nanoparticles in advanced biomedical applications.

Introduction

The growing interest in creating hybrid materials with advanced functional properties has led researchers to study more closely the interactions between metal nanostructures and microorganisms, so to speak, natural “nano-reactors”. − Among all these microorganisms, magnetotactic bacteria (MTB) form a separate category, being a unique type of bacteria. They biomineralize magnetite (Fe₃O₄) or greigite (Fe₃S₄) crystals, forming the magnetosomes organized inside the cell in ordered chains. These natural crystals give MTB strong magnetic properties and the ability to orient themselves in external magnetic fields. ,

The structural and functional particularities of MTB have attracted a lot of interest in recent decades, opening perspectives for their use in fields such as biomedical imaging, magnetic sensing, targeted delivery of therapeutic agents, and nanoscale materials engineering. − For some time, gold nanoparticles (AuNS) have become important tools in biomedical applications due to their high biocompatibility, chemical stability, and their ability to be functionalized with various biomolecules. − They can offer unique opportunities in the research related to biological processes at the molecular scale. In this regard, gold nanostars with defined symmetry are of particular interest since the final shape can be finely tuned during the synthesis process. According to Niu et al., the presence of sharp tips and high-index crystalline facets leads to a large increase in the local electromagnetic field , and the fact that the tips are symmetrical makes the plasmonic response much more stable and reproducible, which is extremely important for biomedical applications and sensitive detection techniques. Nowadays, such intense plasmonic behavior is widely exploited in the development of advanced biosensors, high-resolution diagnostic techniques, and innovative therapeutic applications. ,−

The interaction between plasmonic nanoparticles and MTB is a strategy that is just beginning to be explored (Figure ). A recent study has demonstrated the functionalization of MTBs with plasmonic nanostructures, specifically gold nanoprisms, for the development of hybrid biological systems. These investigations have primarily focused on bacterial functionalization strategies for photothermal applications. In contrast with it, the present study explores the spontaneous association between native, unfunctionalized MS-1 cells and plasmonic gold nanostars, including both PVP-coated and silica-coated variants. The choise of gold nanostars reside in their strong plasmonic activity and well-established ability to generate intense localized electromagnetic fields at multiple sharp tips. , Gold nanostars provide higher/defined SERS enhancement, and also have the potential for ultradetection of nonfunctionalized analytes. The goal here was not to engineer the bacterial surface, but rather to determine whether naturally occurring MTB-AuNS assemblies can be formed and subsequently characterized via SEM and Raman/SERS spectroscopy. As such, this work should be viewed as a proof-of-concept study aimed at establishing the feasibility of forming MTB-AuNS biohybrids and evaluating their spectroscopic implications. When MTBs are exposed to these nanostructures in a controlled manner, hybrid assemblies can be observed. It is worth mentioning that we also notice a considerable improvement in the visibility of magnetosomes in SEM images and the rise of the sensitivity for Raman fingerprint analysis of MTB due to the fulfillment of the SERS experimental conditions.

1.

System under investigationAuNS stabilized with PVP (1) and silica (2) respectively, which are expected to interact with magnetotactic bacteria, MS-1 strain.

Although MTBs are more difficult to isolate and culture than conventional bacterial strains, they offer a unique biological architecture due to the presence of intracellular magnetosomal chains organized along the cell’s longitudinal axis. Unfortunately, common microbial models or simplified biomimetic systems cannot easily replicate such structural features. Therefore, MTBs are attractive for exploring nanoparticle-cell interactions.

What sets our study apart comes mainly from the way MTBs, MS-1 strain, are viewed in relation to symmetrical gold nanostars. We do not treat them merely as a passive support to which particles adhere, but suggest that they could serve as a biological platform model for studying magneto-plasmonic assemblies, something that has not been analyzed much in similar biological systems until now.

The purpose of MTB-AuNS combinatory approach in the present study was not to surpass conventional AuNS-based SERS substrates, but rather to investigate whether a stable association could be achieved between a natural biological magnetic system and plasmonic nanostructures. Proof of such an interaction may be a prerequisite for future studies exploring magnetic guidance, spatial organization, or detection applications of MTB-plasmonic hybrid assemblies.

Results

SEM/TEM Characterization of AuNS/AuNS-SiO₂

Morphological characterization of both gold nanostars and gold nanostars coated with silica by SEM allowed for precise evaluation of their geometry, diameter, as well as their morphological and size dispersity (Figure ). The obtained SEM images highlight the symmetrical star-shaped structure, defined by multiple sharp extensions from a compact icosahedral core; a characteristic associated with a high surface-to-volume ratio and enhanced optical activity. The high contrast provided by SEM facilitates the observation of tips and edges of the extensions.

2.

SEM and TEM images of AuNS: (A) SEM image of PVP-coated AuNS; (B) TEM image of AuNS-SiO₂.

The homogeneous dimensional distribution and the structural integrity suggest a controlled synthesis. The absence of aggregation also indicates colloidal stability suitable for further applications in biomedical and plasmonic detection. The mean diameter is 171 ± 24 nm. The silica-shell thickness of AuNS-SiO₂ was measured to be 14 ± 2 nm.

Extinction Measurements of AuNS/AuNS-SiO₂

The extinction spectra, carried out on a standard spectrophotometer, of the ethanol-based suspension of the PVP-stabilized AuNS (red curve) and silica-coated AuNS (blue curve) are presented in Figure . The two distinct peaks in both spectra correspond to the LSPR of the star’s “body” and branches. The redshift and widening of the spectrum are clearly observed in the case of AuNS-SiO₂. The spectra agree well with preliminary discrete dipole approximation (DDA) calculations for symmetric nanostars and experimental spectra for almost monodisperse, symmetric icosahedral gold nanostars in the literature.

3.

Extinction spectra of two types of AuNS dispersed in ethanol: PVP-stabilized AuNS (red line) and AuNS-SiO_2.(blue line).

SEM Characterization of MTB (MS-1 strain)

Morphological characterization of the magnetotactic strain MS-1 by scanning electron microscopy provides a detailed insight into the cell dimensions, membrane conformation, and organization of magnetosome chains, visible on the cell surface. , The deposition of bacteria on a silicon substrate is an optimal method for visualizing the cells during analysis, since the flat and chemically inert, conductive surface of silicon minimizes artifacts and allows obtaining high-resolution images.

MS-1 cells were analyzed in a stand-alone form, without prior treatment or coating with a conductive layer (gold or carbon). Adopting this strategy allowed for a close examination of the way these bacteria present themselves on the substrate used, and is an imminent and preliminary condition for the actual experimental process in which the interaction between them and gold nanostars is desired.

SEM analysis allowed the identification of several bacterial morphologies. As can be seen in Figure (top left), the bacteria are found isolated, and the bacterial morphologies present a spiral shape, with a small degree of variability, some being slightly elongated. The dimensions of the analyzed bacterial shapes vary between 2.3 and 3.3 μm.

4.

SEM images of the MS-1 strain of MTB on a silicon substrate. Overview (with several bacteria, indicated by red arrowstop left) and three bacterial morphologies with lengths of (1) 2.9 μm, (2) 3.3 μm, and (3) 2.3 μm. Magnetosome chains with 18–21 crystals/cell and sizes between 31 and 76 nm; MTBmagnetotactic bacteria, MS-1 strain, MCmagnetosome chain.

The magnetosome chains, visible on the cell surface (Figure , morphologies 1, 2, and 3), were analyzed, and the findings revealed about 18–21 magnetite particles per bacterial cell with dimensions ranging between 31 and 76 nm. The specific contrast provided by SEM also allows the observation of possible changes at the cell surface, in particular, the presence of flagella, which are relevant for magnetotactic behavior in the environment.

SEM Analysis of MS-1 and AuNS Assemblies

Characterization of the interaction between the magnetotactic bacterial strain MS-1 and PVP-coated gold nanostars by SEM provides a clear view on how the particles attach and organize themselves on the surface of bacterial cells (Figure ). Samples were specifically prepared on a silicon substrate to avoid image distortion caused by charging effects.

5.

SEM images of the interaction between the magnetotactic bacterial strain MS-1 and AuNS on silicon substrate; (1–4) four different MTB morphologies with varying densities of nanoparticles distributed across the bacterial surface; MTBmagnetotactic bacteria, MS-1 strain; MCmagnetosome chain; AuNSgold nanostars.

SEM images show the observable association of the nanostars to the cell membrane, also noticeable along the longitudinal axis, where the internal chains of magnetosomes are located (Figure , morphologies 1 and 2). This spatial association could suggest the existence of electrostatic interactions and the possible contribution of functional groups exposed at the membrane level, that could favor the anchoring of the gold particles. At this stage, however, this should be regarded as a working hypothesis requiring dedicated further investigations.

It is essential to mention, in this context, that no obvious morphological alterations were observed following AuNS exposure under the investigated experimental conditions. SEM observations also showed local differences in particle density, especially in the polar regions (Figure , morphologies 3 and 4), which may reflect the involvement of structures associated with motility (flagella), or possible surface charge variations in these areas.

In order to provide a preliminary assessment of AuNS spatial distribution, a semiquantitative analysis was performed on representative SEM images (n = 20). The average number of AuNS observed in regions corresponding to the magnetosome-chain axis (on average 4 particles per region) was comparable to that observed in the polar regions of the cells (on average 4 particles per region). These observations suggest that AuNS can be found at multiple locations on the bacterial surface rather than being exclusively associated with a single cellular region.

SEM Analysis of MS-1 and AuNS-SiO₂ Assemblies

In comparison, the interaction between the MS-1 strain and the silica-coated gold nanostars presents relevant features revealing the preservation of structural stability and controlling the anchoring of the cell surface (Figure ). The silica layer may act as a biocompatible barrier, reducing the impact of direct interactions between gold in the nanoparticles and the bacterial membrane. Thus, it may limit any potential toxic effects (such as membrane disruption or oxidative stress) and, at the same time, it appears to contribute to preserving the morphological integrity of MS-1. SEM images reveal that AuNS-SiO₂ are arranged on the bacterial surface, without inducing apparent deformations of the cell contour, confirming a moderate and controlled interaction.

6.

SEM images of the interaction between the magnetotactic bacterial strain MS-1 and AuNS_SiO₂ on silicon substrate; (1–4) four different MTB morphologies with varying densities of AuNS-SiO₂ distributed across the bacterial surface; MTBmagnetotactic bacteria, MS-1 strain; MC magnetosome chain; AuNS-SiO₂gold nanostars coated with silica.

Similar to AuNS, the preliminary semiquantitative analysis of AuNS-SiO₂ on SEM images (n = 20) showed comparable particle densities in regions corresponding to the bacterial longitudinal axis (on average 3 per region) and the polar regions (on average 4 particles per region).

Raman and Surface-Enhanced Raman Spectroscopic Investigations

Raman scattering has been used in microbiology for several years − for the detection and discrimination of tiny living organisms. This method can provide abundant information on the chemical composition of microorganisms and help gain insights into its growth, metabolism, and the process of organelle formation. Surface-enhanced Raman is known to provide very good quality signal-to-noise spectra and show greater differentiation as compared to bulk Raman spectra. ,

The obtained Raman and SERS spectra of MS-1 bacteria, are in absence and presence of gold nanostars and are shown in Figure . Specifically, the excitation wavelength of 785 nm was chosen for following reasons: to preserve the organics from decomposition or burning by highly energetic irradiation, to eliminate the parasitic fluorescence from the glass substrate, to enhance the signal by exciting the proper LSPR and to take advantage of the so-called tip-effect in the vicinity of the nanostars tips (thus, avoiding undesirable artifacts related to interband transitions in gold nanoparticles).

7.

Raman and SERS spectra of MTB (A), MTB_AuNS (B) and MTB_AuNS-SiO₂ (C). All spectra were acquired on a glass substrate under an excitation wavelength of 785 nm, with an estimated spot size of ∼1.5 μm.

Careful analysis of the spectra allows us to distinguish two sets of characteristic peaks: one set may be associated with iron oxide phases commonly reported in MTB, including magnetite and its transformation products. Another set associated with the organic constituents of MTB. The overlaying of the peaks complicates straightforward interpretation of the results, nevertheless we summarize our findings in Table , along with peaks frequently reported in the literature. Although magnetite is the predominant mineral phase reported for magnetosomes produced by MS-1, Raman bands in this spectral region may overlap with those of other iron oxide phases. Therefore, the present assignments (corresponding to our study in Table ) should be regarded as tentative. Definitive phase identification would require complementary crystallographic and magnetic analyses.

1. Main Raman Peaks Reported for MTB, and Organic Components from Bacteria and Gold Nanostars.

Raman peak (cm–1) literature	Raman peak (cm–1) this work	Item corresponding	Refs
250, 302 (Eg), 350, 543 (T2g), 653 (A1g), 668	302, 542	magnetite (Fe3O4)	−
346 (T1), 505 (E), 653 (LO Eu), 660, 710, 738	505, 653, 737	maghemite (γ-Fe2O3)	−
346, 379, 491, 736, 740	489	ferrihydrite (5Fe2O3·9H2O)	,
901	901	C–O–C stretch
1060	1060	C–N and C–C stretch
1084	1084	C–O stretch
1132	1131	C–N and C–C stretch, unsaturated fatty acids in lipids
1151	1151	C–C, C–O ring breath asymmetric
1206	1205	tyrosine, phenylalanine, proteins, amide III
1266		lipids
1297	1296	CH2 deformation
1332	1356	broad band overlapping with the glass background; possible contribution from adenine/guanine/tyrosine/tryptophan cannot be excluded
1454	1412	CH2 deformation
1579	1578	guanine, adenine
1614	1613	tyrosine
1645	1645	amide I

The band observed at ∼1356 cm^–1 in the Raman spectrum of MTB should be interpreted with caution, as control measurements revealed a broad contribution from the glass substrate in the same spectral region. Consequently, this feature cannot be unambiguously assigned to bacterial biomolecules, although contributions from nucleobases and aromatic amino acids reported in the literature cannot be excluded.

As expected, the bacteria themselves yield only a weak Raman signal (Figure A). In the presence of AuNS-SiO₂ the signal is only slightly enhanced (Figure C), since a relatively thick (about 14 nm) silica shell screens the surface enhancement effect. This observation is consistent with prior reports indicating that the SERS intensity diminishes to nearly zero when the separation between the substrate and the Raman probe exceeds 10 nm. , In contrast, the AuNS without a silica shell, which are simply stabilized by a thin PVP layer provided a pronounced SERS response (Figure B). Our results are consistent with previously reported findings. ,

SERS Fingerprinting of MTB in the Presence of AuNS and AuNS-SiO₂

A relevant perspective is that the intensity of the Raman signal is strongly influenced by the local arrangement of nanoparticles relative to the bacterial surface (Figure ). Variations in nanoparticle density and distribution seem to contribute to heterogeneous hotspot formation, with different levels of signal enhancement across the sample. AuNS presence consistently leads to a stronger Raman response compared to MTB alone, confirming their role as effective SERS-active nanostructures within the hybrid assemblies (Figure A). In the presence of silica-coated AuNS, the signal intensity of the MTB varies (Figure B), although it is much weaker than in the presence of AuNS with an isolating shell (Figure A). These variations could also be explained by local variations in the surface concentration of the bacteria and the nanoparticles, and possibly also in the thickness of the silica shells. This approach has the potential to contribute to the development of sensitive and “on-spot” spontaneous assembly of nanoparticles onto MTB.

8.

Representative SERS spectra from different regions of the MTB_AuNS (A) and MTB_AuNS-SiO₂ (B) assemblies, which illustrates the local variability of the signal resulting from differences in nanoparticle distribution. All spectra were acquired on a glass substrate under an excitation wavelength of 785 nm, with an estimated spot size of ∼1.5 μm.

To provide a semiquantitative assessment of the SERS performance, the average intensity of the characteristic band at 1578 cm^–1, one of the most intense and well-defined Raman bands observed in the spectra, was compared between the conventional Raman spectrum of MTB and the corresponding SERS spectra using five independent measurements acquired under identical experimental conditions. The MTB_AuNS hybrid exhibited an average signal enhancement of approximately 4.4-fold, whereas the MTB_AuNS-SiO₂ hybrids showed an average enhancement of approximately 3.2-fold. These values are relative signal enhancement ratios rather than absolute SERS enhancement factors and provide a comparative estimate of the signal amplification achieved by the two hybrid systems.

To provide a quantitative estimate of spectral reproducibility, the intensity of the band at 542 cm^–1 was evaluated across five spectra acquired from different regions of the same sample. The relative standard deviation (RSD) was ∼22%, confirming the expected local signal variability arising from the heterogeneous distribution of AuNS on the bacterial surface (Table ).

2. Statistical Analysis of the Selected Raman Marker Band Intensity for MTB, MTB_AuNS, and MTB_AuNS-SiO₂ Samples.

Sample	Marker band (cm–1)	Mean intensity	SD	RSD (%)
MTB	542	3.56	1.66	46.6
MTB_AuNS	542	8.47	1.89	22.4
MTB_AuNS-SiO 2	542	6.10	3.35	54.9

Discussion

SEM analysis of the MTB, MS-1 strain, revealed well-defined structures with a typical, spiral-shaped morphology. The magnetosome chains were clearly visible, confirming the cellular integrity and internal organization of the crystals. The bacterial forms analyzed had dimensions between 2.30 and 3.30 μm, and the magnetosomes were composed on average of 18–21 crystals/cell with sizes between 31 and 76 nm. These values are characteristic of MTB, MS-1 species, and are consistent with data reported in the literature. −

In terms of conjugation with gold nanostars, SEM revealed the interaction of nanoparticles with bacteria, but without affecting the cell morphology. The distribution of gold nanostars before the addition of bacteria, as shown in Figure , was relatively uniform, and the important aspect in this regard was the absence of agglomeration.

The lack of nanoparticle internalization should not necessarily be considered a limitation. The localization of AuNS on the cell surface is of particular interest, as it may allow for close interaction with membrane-associated biomolecules and promotes the formation of localized plasmonic hot spots. Such a configuration was expected to facilitate the amplification of Raman signals originating from bacterial components, which constitute the primary targets of the present investigations.

A similar study, conducted by Cerezo-Collado et al., used a different strain of magnetotactic bacteria (Magnetospirillum gryphiswaldense, MSR-1) to functionalize it with gold nanoprisms by two approachesboth direct attachment and an indirect approach involving exopolysaccharidesfor photothermal applications. In our case, we employed a similar direct interaction method, but with the aim of analyzing spontaneous association between native, unfunctionalized MS-1 cells while using AuNS and AuNS-SiO₂. The focus in this instance was on how nanoparticles interact with bacteria, as well as SEM and Raman/SERS characterization of the assemblies. The interaction between gold nanostars and the MS-1 strain, as analyzed by SEM, revealed a spatial association of AuNS with regions corresponding to the bacterial body and magnetosome-containing axis. Despite it, the current experiments do not allow us to determine whether this distribution originates from biological interactions, physicochemical surface effects, or sample-preparation-related phenomena. PVP-coated gold nanostars, as well as those coated with silica, as observed in the SEM images (Figures and ), are also observed in the vicinity of the polar ends of the bacteria, an aspect that could be associated with the involvement of structures associated with motility (flagella), or possible surface charge variations in these areas. The density of nanostars that attach to the surface and to the polar ends of the bacteria shows a degree of variability. A preliminary semiquantitative evaluation of representative SEM images (n = 20) showed comparable numbers of AuNS associated with regions corresponding to the magnetosome-chain axis (on average 4 particles per region) and the polar regions of the cells (on average 4 particles per region). A similar tendency was observed for AuNS-SiO₂, with average values of 3 and 4 particles per region for chain-associated and polar regions, respectively. Even though AuNS and AuNS-SiO₂ were consistently observed on the bacterial surface, the present data do not support a strong preferential localization to a specific cellular region. Instead, the results suggest that the interaction may occur at multiple accessible sites on the cell surface.

An important particularity that we observed is that, in the case of AuNS coated with silica, the distribution seems to be slightly more homogeneous, an aspect that could be suggestive for a potentially more increased stability of the interaction with MTB. The latter is due to the silica layer, which may have reduced particle agglomeration and minimized the possible toxic effects of gold nanoparticles on MS-1.

The attachment and interaction between bacteria and nanostars or nanostars and bacteria may suggest favorable interactions between the bacterial surface and these particles. They may be mediated by functional groups on the bacterial membrane. The mechanism responsible for the observed AuNS distribution cannot be established from the present data set and will require dedicated studies involving additional biological and physicochemical controls.

Regarding SERS, the interaction between MS-1 and AuNS resulted in changes in the Raman signals characteristic of MS-1 bacterial biomolecules. This aspect demonstrated the efficiency of gold nanostars as SERS promoters for the detection of cellular molecules. The spectra indicated the enhancement of the bands assigned to proteins and membrane lipids, highlighting the direct interaction between the nanoparticles and cellular components. In the case of AuNS-SiO₂, the SERS signal was observable and stable, although much weaker compared to AuNS-PVP-induced SERS. This could be attributed to the large distance created by the silica layer between the gold nanoparticle and the bacteria and is in line with earlier findings. ,

However, this effect may be compensated by better stability and probably biocompatibility, which confers another utility of silica-coated nanostructures, making them possibly more suitable for long-term biological applications.

These observations highlight the advantage of the silica layer in maintaining the optical properties of AuNS when associated with biological systems. This latter aspect is of particular importance in the design and direction of applications for which these bacteria, respectively nanostars, are intended to be used. Thus, the obtained data suggest that MS-1 could be used for functionalization with gold nanostars for SERS signal amplification, while the silica coating may provide an optimal balance between stability, biocompatibility, and optical efficiency. SEM, and SERS provided complementary information on the morphology, physicochemical interactions, and the efficiency of attachment of nanostructures on this biological matrix, emphasizing the importance of choosing the appropriate nanoparticle depending on the targeted application.

Conclusions

The present study investigated whether the MS-1 strain of MTB can efficiently interact with plasmonic nanostructures while preserving both its morphological integrity and the organization of magnetosomes in chains. As expected, the obtained data suggest that MS-1 underwent spontaneous interactions with AuNS and AuNS-SiO₂ used as hybrid metallic nanoparticles, AuNS serve for SERS signal amplification, while AuNS-SiO₂ provide an optimal balance between stability, biocompatibility, and optical efficiency. On the other hand, SEM and SERS investigations provide complementary information regarding the morphology and physicochemical interactions of the nanostructures with the magnetotactic bacteria. At the same time, the results also point out the importance of selecting the appropriate composition of nanoparticles for an intended application.

These results indicate that hybrid platforms based on MS-1 and AuNS could be promising alternative for future research in the field of magneto-plasmonic biohybrid assemblies, which may include sensitive molecular detection in complex environments, where stability and biocompatibility will play a key role. Thus, the unique combination of the natural magnetic properties of MTB with the plasmonic properties of gold nanostructures undoubtedly provides an ambitious foundation for the development of ultrasensitive, addressable magnetic sensing platforms.

Methods Section

Synthesis of Two Types of AuNS

The experiments were performed using gold nanostars (AuNS) prepared according to the synthesis established by Niu et al. They were stabilized during synthesis with a thin layer of polyvinylpyrrolidone (PVP) with Mw = 55000 g/mol (Sigma-Aldrich, lot number MKCQ4695) or subsequently additionally coated with a silica shell (AuNS-SiO₂). The AuNS size (mean diameter: 171 ± 24 nm) was selected based on their plasmonic properties and their ability to generate intense localized electromagnetic fields at the sharp tips. These characteristics are highly advantageous for SERS applications. The concentration of all gold nanostar samples was 4 mg/L, as determined by flame atomic absorption spectroscopy (AAS) using an ICE 3000 Series AAS spectrometer from Fisher Scientific.

Description of MTB Culture

In this work, the MTB strain, Paramagnetospirillum magnetotacticum MS-1 (DSMZ: DSM 3856), grown in standard culture medium (as shown and provided by the supplier) containing 100 μM of Fe­(III) quinate solution, seven vitamins solution, and mineral solution was used. ,

Two Models Support for the Deposition of AuNS/AuNS-SiO₂/MTB and Their Mixture

The MTB, MS-1 bacterial suspension was systematically dropped (3 μL) on clean silicon substrates and glass slides after mixing with colloidal solutions of gold nanostars. The actual experimental samples consisted of the bacterial suspension and the gold nanostars (named MTB_AuNS), and the bacterial suspension and the gold nanostars coated with silica (named MTB_AuNS-SiO₂).

Experimental Controls

The first control configuration involves separate SEM characterization of MTB, and AuNS (under identical experimental conditions used for MTB_AuNSa drop of 3 μL) (Figure A,B). For direct and specific comparison, two MS-1 bacteria in the presence of AuNS at the same selected concentration are presented (Figure C,D).

9.

SEM images obtained on silicon substrate for control (A, B) and specific (C, D) experiments: AuNS at the selected concentration in the absence of MS-1 bacteria (A), MS-1 bacteria in the absence of AuNS (B), two MS-1 bacteria in the presence of AuNS (at the same concentration used for (A) image (C, D).

For the second control configuration: Raman spectra of the bare glass and PVP solutionto identify possible contributions from the substrate, or nanoparticle stabilizer, were recorded (Figure ). These measurements allowed for a more reliable interpretation of the spectral features observed the MTB_AuNS and MTB_AuNS-SiO₂ assemblies.

10.

Raman spectra of four individual control samples: bare glass, PVP, AuNS and AuNS-SiO₂.

For the third control configuration, AuNS, at the same selected concentration, were incubated with nonmagnetotactic bacterial cells (E. coli), and droplets (3 μL) were collected from the mixture and dried on the silicon substrate and Formvar coated copper TEM grid (300 mesh, carbon type-b; Ted Pella, Redding, CA, USA), respectively for analysis by SEM (Figure ).

11.

STEM and SEM controls in the presence of nonmagnetic bacterium E. coli: STEM (A, B) images on a Formvar coated copper TEM-grid for better visualization of the interaction with AuNS­(A) and AuNS-SiO₂ (B); SEM images (C, D, E) obtained on a silicon substrate, with obvious difficulties in distinguishing the shape of individual E. coli cells (E) in the presence of AuNS.

SEM investigations performed on a membrane-based TEM grid enabled visualization of the interaction between AuNS and E. coli bacterial cells, without the need for metallic coating and high annealing temperature, as previously reported. No obvious attachment or preferential accumulation of AuNS or AuNS-SiO₂ on the bacterial surface was observed (Figure A, B), in contrast to the behavior recorded for the MS-1 strain.

It should be noted that, due to the pronounced charging effects observed during SEM imaging on the silicon substrate (Figure C, D, E), there were obvious difficulties in distinguishing the shape of individual E. coli cells (E) in the presence of AuNS.

Microscopic and Spectroscopic Characterization of AuNS/AuNS-SiO₂

The characterization of freshly prepared AuNS and AuNS-SiO₂ was carried on carbon-coated copper grids (carbon coating type A, 6–10 nm thickness, Cu, 200 mesh, Plano GmbH) using a Hitachi SU5000 Schottky field emission scanning electron microscope in secondary electron (SE) imaging mode with an electron acceleration voltage of 10 kV and a spot intensity of 3. For the size analysis, more than 200 AuNS at different spots were analyzed using the software FIJI. Characterization of AuNS-SiO₂ dried on carbon-coated copper grids was performed with a transmission electron microscope (TEM) EM 109 from Zeiss at an acceleration voltage of 80 kV.

UV/vis/NIR extinction measurements of AuNS in dispersion were performed using a Cary 5000 UV–vis-NIR spectrometer from Agilent Technologies. The AuNS or AuNS-SiO₂ were dispersed in ethanol. The measurements were taken in quartz cuvettes (type 26.715/Q/10/Z20; Starna) within the range of 400 to 1200 nm.

Microscopic Characterization of MTB and MTB Mixed with AuNS/AuNS-SiO₂

Microscopic characterization was performed using a scanning electron microscope, a Hitachi FEG SU8030 (Hitachi High-Technologies Corporation, Tokyo, Japan) at 5−25 kV, and working distance of 10 mm in the SE imaging measurement mode.

Spectroscopic characterization was performed using a LabRAM HR Evolution RAMAN spectrometer (HORIBA Scientific, HORIBA Jobin Yvon S.A.S France), equipped with a diode laser module (ONDAX, USA). Raman spectra were acquired using a 785 nm excitation source, a laser power of 1% at the sample, an acquisition time of 10 s, and 4 accumulations. Measurements were performed using a 100× objective, with a spot size of ∼1.5 μm.

Sample preparation for characterization involved mixing the MTB suspension with AuNS and AuNS-SiO₂ dispersions (concentration: 4 mg/mL) using a 2:1 MTB-to-nanoparticle volume ratio (6 μL MTB: 3 μL NS). The resulting mixtures were vortexed for 15 min, then droplets (3 μL) were deposited on a silicon substrate (for SEM characterization) and on a glass substrate (for Raman/SERS characterization), respectively. The prepared samples were allowed to dry for 30 min under ambient laboratory conditions (room temperature). The Raman/SERS spectra were acquired from 3–5 different regions, with 4 consecutive spectral acquisitions collected per region.

Glossary

Abbreviations

MTB

magnetotactic bacteria

SEM

scanning electron microscopy

TEM

transmission electron microscope

SERS

surface-enhanced Raman scattering

AuNS

gold nanostars

PVP

polyvinylpyrrolidone

MS-1

Paramagnetospirillum magnetotacticum

SE

secondary electron

N.L.P. conducted the investigation, performed SEM/SERS characterization, and wrote the original manuscript; D.N. and P.S. contributed to the synthesis of gold nanostars and conducted extinction and SERS/TEM measurements; A.R. contributed to methodology and the collection of SERS data/references; C.G. contributed to validation, visualization, and interpretation of optical spectra; C.O.P. contributed to validation, visualization, and data curation; R.E.I. contributed to conceptualization, supervision, project management, methodology, investigation, and writingreview and editingof the manuscript.

The authors declare no competing financial interest.