Unveiling the real-time behaviour of Staphylococcus aureus through nanomotion sensing

Abstract

Iron is an essential micronutrient for bacterial growth and plays a crucial role in various physiological processes, including motility. This study investigates the influence of iron availability on the behaviour of Staphylococcus aureus, a significant human pathogen, using a nanomotion sensing technique. By correlating nanoscale vibrations with metabolically related behaviour, this sensor offers real-time insight into the activity of biological specimens from single proteins to living bacteria and mammalian cells. S. aureus cultures were grown under iron-repleted and iron-depleted conditions, and their nano-scale motility was monitored in real time. We compared the activity of wild type S. aureus with an isogenic deletion mutant, in which siderophore production was abrogated. Our results demonstrate distinct motility patterns of the two strains in response to varying iron concentrations, including different outlines of their activity. We identified the fingerprint of the iron intake pathway for the two strains: in the presence of iron, while the wild-type strain followed an exponential growth pattern, the mutant suspended its growth. Remarkably, in presence of iron, both strains showed some coordinated oscillations at well-defined frequencies, which could be related to specific behaviour of S. aureus. These results were corroborated by the phenotypic assays performed through optical microscopy and conventional growth methods. Such findings highlight the potential of nanomotion sensing, a powerful tool to study cellular vibrations and to investigate the complex relationship between iron homeostasis and S. aureus activity, suggesting this metabolic route as a possible fruitful target for the development of novel antibacterials.

Introduction

Staphylococcus aureus, a versatile and resilient Gram-positive bacterium, poses a significant threat to human health, causing a spectrum of infections ranging from minor skin infections to life-threatening conditions like sepsis and pneumonia^1–3. This pathogen has great clinical relevance because of its ability to develop resistance to last resort antibiotics^4,5. This, combined with its frequent occurrence in healthcare settings as a source of hospital-acquired (nosocomial) infections, makes S. aureus a key member of the ESKAPE group, which includes pathogens considered the most dangerous to global health^6,7.

Iron, an essential micronutrient for all living organisms, plays a pivotal role in the growth, metabolism, and virulence of S. aureus⁸. However, iron availability within the host environment is tightly controlled, limiting bacterial growth and dissemination⁹. To overcome this challenge, S. aureus has evolved sophisticated molecular mechanisms to acquire and utilize iron¹⁰. Among the multiple iron-acquisition mechanisms employed by the bacterium, siderophores are critical players in the process, particularly for the acquisition of inorganic iron forms when heme concentration is limited¹¹. The ability of S. aureus to utilize a range of Fe(III) scavengers, each with distinct functional groups involved in iron coordination, has been well-documented¹². The key siderophores produced by S. aureus are Staphyloferrin A (SA) and Staphyloferrin B (SB), which play pivotal roles during growth under iron-restricted conditions^13,14. After being synthesized and secreted into the extracellular environment, siderophores chelate ferric iron [Fe(III)] from the surrounding area, competing with the host proteins. Once the Fe(III)-siderophore complex is formed, it is actively transported into the bacterial cytoplasm through an energy-dependent process¹⁵.

Mutants unable to make both SB and SA exhibit severe growth defects and impaired viability in serum or in synthetic media containing human transferrin (Tf) as sole iron source¹⁶. Moreover, SB-deficient strains are less virulent in in vivo models of infections, highlighting the essential role of siderophores in S. aureus colonization, virulence, and survival^13,17, while strains unable to secrete SA display reduced fitness within subcutaneous abscesses and epithelial cells¹⁸. Investigating the real-time dynamics of iron collection and its impact on S. aureus behaviour through the generation of siderophore-deficient mutants provides a new outlook into the intricate mechanisms of iron metabolism in this species and could open up new directions towards the development of effective diagnostic and therapeutic strategies.

Bacterial motility, the ability of bacteria to move independently, is a fundamental aspect of microbial life, playing a crucial role in survival, environmental exploration, nutrient acquisition, and ultimately, the establishment of infection¹⁹. Historically, S. aureus was considered a non-motile bacterium due to the absence of flagella and pili, the well-characterized appendages responsible for swimming and twitching motility in many other species. However, recent investigations have unveiled several unconventional forms of surface-associated movement in S. aureus, challenging this long-standing dogma²⁰. One notable form of surface motility in S. aureus is spreading, a passive process observed on soft agar media²¹. S. aureus exhibits other forms of surface motility, such as comet formation, which resembles gliding motility away from the central colony²⁰. Finally, S. aureus frequently adopts a sessile lifestyle by forming biofilms, which are structured communities of bacteria encased within a self-produced extracellular polymeric substance matrix²². Nutrient availability plays a critical role in regulating biofilm growth and dispersal, with nutrient-rich conditions typically promoting biofilm formation and maturation²³. In addition to these motile characteristics, these bacteria exhibit smaller, even nanoscale vibrational motions, that are certainly correlated to the energetic status of the biosystem and reflect the internal motion of the cells. Thus, studying the motion and nanoscale vibrations of S. aureus can provide new insight into the evaluation of bacterial activity and unearth their response to variations in nutrient availability and environmental conditions.

Traditional microbiological techniques grant essential knowledge but often lack the temporal resolution to capture the rapid and dynamic changes in bacterial physiology associated with iron acquisition and utilization. Advanced microscopy techniques such as atomic force microscopy (AFM) and single-molecule localization microscopy are increasingly being used to study the dynamics of individual S. aureus cells²⁴. These methods have provided valuable insights into processes like peptidoglycan synthesis during cell division and the effects of antimicrobial agents on cell structure and mechanics. The nanoscale mechanical properties of S. aureus cells have been investigated using AFM, revealing information about cell wall stiffness²⁵ and turgor pressure during different growth phases²⁶ or dynamic remodelling processes^27,28. Vibrational spectroscopy, particularly FTIR, has proven to be a valuable tool for identifying and differentiating S. aureus strains based on the unique spectral signatures of their vibrations²⁹. Recent advancements in nanotechnology have provided novel procedures and tools to investigate microbial behaviour³⁰. Among these, the nanomotion sensing leverages the sensitivity of novel nanomechanical devices to investigate microbial and cellular behaviour at unprecedented spatial and temporal scales^31–33. Indeed, the nanomotion sensor can translate small vertical forces (in the order of 10^–18 J, equivalent to the energy produced by the hydrolyzation of three ATP molecules) into measurable signals³¹. This means that the sensor can transduce even minimal mechanical oscillations of the specimens, which are typical of any living system, and directly reflect the activity of the specimens attached to the sensor. These nanoscale movements and fluctuations can be generated by specific motors (such as flagella or pili)^34,35 but also by many other smaller internal movements of the cell, such as reorganizations of the cell wall and cytoskeleton, opening of ionic channels or even mass displacements in the cytoplasm³⁶. In fact, several works show how the nanomotion sensor allowed detecting vibrations of mitochondria³⁷ or even conformational changes of proteins³⁸. By monitoring the changes in nanomotion of microorganisms upon exposure to antibiotics or antifungals, several groups have demonstrated how this technique can quickly assess the microorganism’s viability and determine their susceptibility or resistance to drugs^39–41. This method is effective for a wide range of bacteria, including motile, non-motile, fast, and slow-growing species, offering a significant advantage over traditional growth-based antibiotic susceptibility test methods that can require a time that can span from several hours to days to yield results^42–44. Furthermore, the exquisite sensitivity of such nanomechanical devices, i.e., cantilevers, nanowires, and microfluidic channels, can also elicit real-time information on bacterial motility, vibrations, growth, and interactions with the surrounding environment, providing a glimpse of the immediate behaviour of the specimens^34,45,46.

This study represents the first application of nanomotion technology to investigate the real-time behaviour of S. aureus cells and the impact of iron availability on their activity and replication. We have generated a double mutant (i.e., S. aureus ∆sbnA∆sfaD), deficient in the production of both SB and SA, and we have compared its dynamic nanoscale motions with those of the wild-type strain. This approach allowed evaluating the real-time effects of iron on S. aureus activity and reproduction rates as well as investigating the emergence of complex behaviour such as periodic and collective movements.

The insights gained from this study will advance our understanding of S. aureus physiology and pathogenesis. Furthermore, the protocols developed in this research could be incorporated into the microbiological testing arsenal. This research could open the way to a tool to test new kinds of inhibitors and therapeutic strategies targeting nutrient-driven microbial behaviours, such as those related to iron acquisition pathways.

Results

Phenotypic characterization of the ΔsbnAΔsfaD strain

Since the expression of SA and SB is tightly regulated by iron availability, to characterize the phenotype of the double mutant, it is essential to culture S. aureus under iron-depleted conditions. To this aim, the chemically-defined iron-poor medium cTMS was used, and the ability of ΔsbnAΔsfaD to grow in cTMS was compared to that of the wild-type (WT) strain, using the same experimental setting recently described⁴⁷. S. aureus WT and its isogenic ΔsbnAΔsfaD mutant were pre-cultured in cTMS supplemented with 300 µM DIP, to further restrict iron availability, for 16–20 h at 37 °C with 180 rpm shaking. Bacterial cells were then cultivated in cTMS supplemented or not with 1 mM FeCl₃, as the sole iron source, for up to 36 h. Results showed that the WT growth was restricted in cTMS, since the addition of an excess of FeCl₃ promoted bacterial growth at 12 h post-incubation, confirming that iron is a limiting nutrient in this medium (Fig. 1)⁴⁷. Interestingly, the growth of ΔsbnAΔsfaD was completely abrogated in cTMS, suggesting that the presence of both siderophores is necessary for supporting bacterial growth (Fig. 1). Notably, when cultured in cTMS supplemented with 1 mM FeCl₃, the growth of ΔsbnAΔsfaD is promoted at 10 h post-inoculum, although to a lower extent compared to the WT (OD₆₀₀ ≃ 0.5 vs 1.2, respectively). A similar behaviour is also observed at 12 h, however the growth of the ΔsbnAΔsfaD resulted to be fully restored to levels comparable to WT at 24 h post-inoculum (Fig. 1). This suggests the activation of putative alternative iron-entrance route(s) into bacterial cells to compensate for the loss of siderophore production, ensuring sufficient iron to support bacterial growth.

Fig. 1.

Growth experiments comparing S. aureus WT and ΔsbnAΔsfaD mutant. Panel (a): Growth of the S. aureus WT and ΔsbnAΔsfaD mutant (blue and red, respectively) in cTMS. Panel (b): growth of the S. aureus WT and ΔsbnAΔsfaD mutant (blue and red, respectively) in cTMS supplemented with 1 mM FeCl₃. Data represent average values of two independent biological replicates, each one performed in triplicate, and error bars indicate the corresponding standard deviations from the mean values. The statistical analysis was performed by using two-way ANOVA, followed by Bonferroni’s multiple-comparison test. Differences between the WT and ΔsbnAΔsfaD strains were considered statistically significant with a p-value ≤ 0.02, indicated with *; not significant (ns) indicates a p > 0.02.

Non-synchronized nanomotion experiments

The nanomotion experiments involved immobilizing S. aureus on the sensor and monitoring the induced fluctuations of the sensor over time. Preliminary tests (Supplementary material 1) show that, in the absence of iron, the bacteria on the sensor can be monitored for several hours, with constant variance and spectral response (Figure S1). This indicates that nanomotion analyses can be carried out for several hours in a controlled fashion and that they are complementary to the conventional phenotypic assays that last several days.

At first, we performed nanomotion experiments on WT S. aureus and on the double mutant ΔsbnAΔsfaD deficient in siderophore production, both of which had undergone a non-synchronized growth to late-exponential phase in cTMS. We immobilized the bacteria on the nanomotion sensor and monitored their activity for 1 h in cTMS followed by several hours in a medium enriched with 1 mM FeCl₃. All time-resolved fluctuations were analysed by calculating their variance and evaluating the spectral response by determining the Power Spectral Density (PSD) in the 15–100 Hz range. Indeed, this is a very interesting spectral region for PSD where the slopes arising from the fluctuations of biological systems are directly correlated to the activity of the system itself. In general terms, the analysis of the spectra was performed by fitting them with a 1/f^α curve, where f is the frequency and α is a parameter correlated to the fluctuations produced by the biological system.

As depicted in Fig. 2, the WT does not respond immediately to the presence of iron ions in the medium but, after two hours, the amplitude of the oscillations starts to increase (Fig. 2, Panel a). The PSD on the other hand shows that bacterial activity increases constantly over time, even before the variance values increase in a significant matter (Fig. 2, Panel b). The spectra show linear trends which change their 1/f^α behaviour over time, with α values starting from 1.48 ± 0.06 before adding FeCl₃ and increasing gradually to 1.89 ± 0.04 in the iron-rich medium (Fig. 2, Panel e).

Fig. 2.

Control experiment: a comparison of the behaviour of WT and ΔsbnAΔsfaD mutant before and after the exposure to FeCl₃. Panel (a,c): Typical average variance of the data elicited from the WT and ΔsbnAΔsfaD nanomotion data divided into 30-min chunks. The green bars represent the fluctuations collected in cTMS, while the red bars represent the evolution of the nanomotion signal after exposure to FeCl₃. Panels (b,d): PSD graphs in the 1–100 Hz area depicting the evolution of the oscillations from WT and ΔsbnAΔsfaD mutant over time. Each curve represents a 40-min experiment where the black and red curves were measured in cTMS while light blue, light green, purple, yellow and blue spectra were collected after the exposure to iron. The spectra are represented on a log–log scale. Panel (e,f): Graphical representation of the α values for WT and ΔsbnAΔsfaD cases. The dashed line represents the average value measured in cTMS. The stars represent the significance of the differences between the iron chunks and the cTMS chunk, determined using a one-way Anova test: *p < 0.05, **p < 0.01, ***p < 0.005.

The ΔsbnAΔsfaD exhibits a remarkably different behaviour. The variance grows for a short period of time and then remains constant for the rest of the period investigated in the experiments (Fig. 2, Panel c). The corresponding PSD (Fig. 2, Panel d) shows a fast increase in the overall intensity of the spectra, followed by a series of almost identical spectra. Interestingly, in both cases the PSD spectra have similar 1/f^α behaviour, but the α value was higher compared to the WT data, with average values of 1.77 ± 0.09 before the addition of FeCl₃ and a constant value of 1.81 ± 0.08 after (Fig. 2, Panel f).

The optical images (Fig. 3), collected simultaneously to the nanomotion data, provide a visual representation of the variance data shown in Fig. 2, Panels a and c. The WT bacteria (Fig. 3, Panels a to c), after 90 min from the FeCl₃ introduction, do not show a large increase in number, while after three hours, the bacteria have completely covered the sensor. The ΔsbnAΔsfaD bacteria, after the exposure to the iron-enriched medium, increase rapidly, but after three hours the number of bacteria on the sensor is still comparable to the number present at the 90-min mark (Fig. 3, Panels d to f).

Fig. 3.

Optical images of the sensor bearing the S. aureus strains. Panels (a–c): time-lapse evolution of the WT bacteria over the sensor before the exposure of the iron ions, at 1.5 h from the injection and after 3 h from the injection. The insets show with higher contrast the growth of the bacteria, which can be seen in increasing numbers even in the external parts of the sensor. Panels (d–f): time-lapse evolution of the ΔsbnAΔsfaD mutant bacteria over the sensor before the exposure of the iron ions, at 1.5 h from the injection and after 3 h from the injection.

Synchronized nanomotion analyses

We submitted bacterial colonies from the WT and the ΔsbnAΔsfaD strains to a synchronization procedure in which, before performing the nanomotion experiments, they had grown to late-exponential phase in the presence of the iron chelator 2,2’-dipyridyl (DIP) to ensure they were depleted of iron and then had been exposed for two hours to cTMS for revitalization. Upon immobilization on the sensor, the nanomotion assays involved a first one-hour stabilization period in cTMS followed by 1 mM FeCl₃ enriched buffer (see Materials and Methods).

Figure 4 shows the evolution of the variance of nanomotion oscillations over a single experiment in the case of the WT bacteria (Fig. 4, Panel a) and in the case of the ΔsbnAΔsfaD strain (Fig. 4, Panel b). Both strains appear to have a comparable low base motion, which is constant over the preliminary cTMS period. Interestingly, the oscillations recorded in the first 30 min after the exposure to FeCl₃ were much higher in the case of the mutated strains (an increase of roughly 5x) compared to the WT, in which the increase of oscillations remained within the measurement error. Furthermore, the increase of the signal amplitude for the ΔsbnAΔsfaD strains was almost immediate, revealing that the bacteria respond very rapidly to the iron abundance, probably due to sensing of this chemical stimulus. After this first period, the mutated strain’s oscillations decrease and return to values comparable to those measured in the control part of the experiment. Indeed, the oscillation amplitude after the first few minutes remains constant or, in some cases, was reduced over longer time periods. The nanomotion signal of the WT bacteria, on the other hand, continued to rise, leading to a more than tenfold increase compared to the values measured in cTMS.

Fig. 4.

Synchronized experiment: a comparison of the behaviour of WT and ΔsbnAΔsfaD mutant before and after the exposure to FeCl₃. Panel (a,b): Typical average variance of the nanomotion data produced by the WT and ΔsbnAΔsfaD mutant, respectively. The bars represent 30-min chunks each, with green bars indicating the fluctuations collected in cTMS, and with the red bars representing the evolution of the nanomotion signal after exposure to FeCl₃. The dashed line represents the average value measured in cTMS. The stars represent the significance of the differences between the iron chunks and the average of the cTMS chunks, determined using a one-way ANOVA test: *p < 0.05, **p < 0.01, ***p < 0.005.

A great difference between the behaviour of the two bacterial strains before and after exposure to iron can be underlined by performing a PSD analysis of the time-dependent nanomotion data. A notable difference is that in the WT PSD, we can identify a wide bump in the 2 Hz area, with a width of approximately 0.5 Hz. This feature is much less evident, and almost invisible in the case of the ΔsbnAΔsfaD bacteria (Fig. 5, Panels a and b). The introduction of the iron ions causes, in both strains, a dramatic change of the spectral behaviour. Almost immediately, the PSDs of the oscillations of both bacteria strains show, in the 2 to 5 Hz spectral range, a much more complex spectral structure, with several sharp peaks (Fig. 5, Panels c for WT and d for ΔsbnAΔsfaD). Remarkably, the two spectra in this range are very similar. Furthermore, new smaller peaks appear in the 9–10 Hz area, which were only hinted at in the WT control spectra.

Fig. 5.

PSD spectra for synchronized experiments. 1–100 Hz section of the PSD graphs of the fluctuations of WT and ΔsbnAΔsfaD mutant bacteria which were treated with the synchronization protocol. The spectra in the 2-h cTMS period (Panel a for WT and b for ΔsbnAΔsfaD mutant) and after the exposure to iron (Panel c for WT and d for ΔsbnAΔsfaD mutant). Panel (e,f): Graphical representation of the α values for WT and ΔsbnAΔsfaD cases. All spectra are represented on a log–log scale limited to the 1–100 Hz range. The dashed line represents the average value measured in cTMS. The stars represent the significance of the differences between the iron chunks and the average of the cTMS chunks, determined with a one-way ANOVA test: *p < 0.05.

Furthermore, even the underlying spectra in the remaining 15–100 Hz range produced by the bacteria change in response to iron exposure. In the cTMS medium, the spectra of the WT and ΔsbnAΔsfaD bacteria span the 15–100 Hz range with a comparable 1/f^α behaviour with α values of 1.56 ± 0.07 and 1.69 ± 0.05, respectively (Fig. 5, Panels e and f for WT and ΔsbnAΔsfaD respectively). After the exposure to the iron enriched medium, the spectra in the 15–100 Hz range decreased their slope, with a 1/f^α behaviour with an α of 1.47 ± 0.09 in the case of the WT and 1.54 ± 0.08 for the ΔsbnAΔsfaD (Fig. 5, Panels e and f).

Discussion

The ability of S. aureus to acquire iron from the host environment is essential for its pathogenesis, particularly in iron-limited conditions such as those encountered during infection⁴⁸. In this context, siderophores play a pivotal role in facilitating iron uptake by chelating Fe(III) and delivering it into the bacterial cell, a process that is crucial under such conditions¹².

Understanding the real-time dynamics of iron acquisition and its impact on S. aureus behaviour is crucial for advancing diagnostics and therapies. Using nanomotion techniques we demonstrated that the variance and the PSD data of WT and ΔsbnAΔsfaD bacteria are very similar in intensity and remain constant during the control period. This suggests that both strains are viable but do not replicate under these conditions and timescales, in line with previous studies showing very little variance changes over time in Escherichia coli, S. aureus, Bacillus pertussis, Bacillus subtilis and other bacterial strains^33,42,45,49.

Upon the addition of iron to the medium, the variance data of the bacteria that did not undergo the synchronization protocol, coupled with the optical images, show that the WT exhibit a much larger increase in number compared to the ΔsbnAΔsfaD strain, which does not appear to grow in number in the timescale of our experiments (Figs. 2, 3). This is consistent with the growth data (Fig. 1) that showed a much faster response of the WT strain to iron supplementation compared to the ΔsbnAΔsfaD double mutant (OD₆₀₀ ≃ 0.5 vs 1.2, at 10 h post-inoculum). Indeed, the difference of the variance histograms between the WT and the ΔsbnAΔsfaD strain upon exposure to an abundant concentration of FeCl₃ is striking (Fig. 2, Panels a and c). The WT bacteria exhibit a typical exponential growth rate, characteristic of many other bacterial species, studied with similar nanomotion sensor setups^34,46,50. In contrast, the mutated bacteria have a fast and strong reaction to the iron intake, with a short increase in variance which stabilizes very rapidly. These dynamics may suggest that, in the absence of siderophores, the ΔsbnAΔsfaD strain has a transient ability to respond to iron influx, but it is unable to immediately activate the metabolic cycles leading to sustained growth, or cellular activity due to impaired iron metabolism.

For the synchronized microorganisms, the difference is even greater. After exposure to iron, the WT bacteria increase their activity very rapidly (Fig. 4, Panel a), probably in part due to a fast increase in number on the sensor. On the other hand, the ΔsbnAΔsfaD have a fast and strong reaction to the iron availability, but this is short-lived, and the fluctuations return to low levels very rapidly. These results strongly suggest that the WT can readily uptake iron, even when iron is scarce, such as in the first hour-long exposure to cTMS, while the ΔsbnAΔsfaD strain, impaired in siderophore biosynthesis, requires additional time to adapt its iron metabolism and internalize iron via less efficient pathways⁵¹. This delayed response in the ΔsbnAΔsfaD strain highlights the critical role of siderophores in iron acquisition and bacterial growth under iron-limiting conditions¹² and suggests that inhibition of siderophore biosynthesis can be a valuable strategy to reduce bacterial fitness and increase the efficacy of combined antibiotic treatments.

The information provided by the PSD is particularly interesting. A power-law 1/f^α behaviour with α between 1 and 2 (with α = 1 representing pink noise and α = 2 representing Brownian noise) is typical of biological systems involving complex active phenomena, for instance evidencing long-range correlation, self-organization or multiplicative noise^52,53. The α values measured in our study, ranging from 1.47 to 1.8, likely reflect the interplay between different phases of the bacterial activity, such as the transition from more passive, diffusion-like motion to active mechanisms driven by interactions with the environment. These values suggest the presence of intricate dynamic behaviours with temporal correlations, potentially involving mechanical interactions, signalling processes, or cell–cell communication, such as those underlying biofilm formation. Furthermore, the dynamics of the PSD curves over time provided additional insights into bacterial behaviour. In the absence of iron, S. aureus WT produces, over time, spectra with similar values of α, suggesting that no metabolic shifts occur under these conditions (Figure S1, Panel d).

In the control experiments (where the bacteria did not undergo the iron-depletion synchronization), the spectra reveal that an increase in variance is well reflected by a similar increase of PSD in the 1–100 Hz area. Interestingly, for the WT bacteria, spectra recorded after the exposure to iron, which are associated with the time periods in which optical and variance data indicate an increase in the number of specimens on the sensor, also exhibit a change in the 1/f^α behaviour, with an increase of the values of α. The shift from 1.47 to 1.81 suggests a change in bacterial activity, pointing to fluctuations that favour lower frequencies, which may reflect a transition toward less spatially coordinated, more heterogeneous movements or indicate change in the bacterial metabolic activity (Fig. 2, Panel b). On the other hand, the spectra of the ΔsbnAΔsfaD strain before and after the exposure to FeCl₃ show a similar behaviour in that spectral region, with α values remaining around 1.8. This indicates that, in the time of our experiments, the change in the bacterial activity has not occurred, also pointing to a potential correlation between spectral features and bacterial phenotype. This underscores the importance of dynamic spectral analysis, which can serve as a sensitive indicator of physiological states (Fig. 2, Panel d).

In the case of the S. aureus bacteria which were treated with the synchronization protocol, the values of the exponential behaviour of the PSD spectra show a very interesting trend which contrasts with that of the non-synchronized microorganisms. Indeed, in the 15–100 Hz region, the values of α decrease after exposure to the abundant iron, falling from 1.56 in the case of the WT and 1.69 for the mutant to 1.47 and 1.54, respectively (Fig. 5, Panels e and f). This suggests that the response of both bacterial strains to the presence of iron, in this synchronized scenario, tends to be more ordered, with an increase in correlation in the cellular activity.

Even more interesting are the PSD spectra collected from the synchronized bacteria in the 1–15 Hz range. In cTMS, the spectra relative to the WT bacteria show a wide structure at 2 Hz in the PSD, which is only barely visible in the case of ΔsbnAΔsfaD (Fig. 5, Panels a and b). This could be assigned to a fingerprint of the basic activity of the bacteria in the control environment. Indeed, a similar feature can also be seen, before the exposure to iron, in the non-synchronized WT case (Fig. 2, Panel b, black curve).

In the presence of abundant iron, both synchronized strains exhibit similar features centred around 2 Hz and 10 Hz (Fig. 5, Panel c and d), which could point to a similar response of the bacteria in this condition, with the only difference being the timescale of the iron intake mechanism, which leads to the activation of the increased bacterial activity and replication. While similar features in the spectral response of the movements of bacteria have already been evidenced for several different bacterial species and in particular for S. aureus⁵⁴, here we suggest an association to a specific phenotype. Indeed, we hypothesize that the 2-h waking period in cTMS with no DIP in addition to the 1 h period in cTMS, common to all nanomotion experiments, allow the WT to collect the small amount of iron ions in the medium and stimulate the iron intake machinery, while the mutations in the ΔsbnAΔsfaD make these bacteria unable to do the same in the timeframe of this first part of the experiments (i.e., 3 h). On the other hand, in presence of abundant iron ions, both strains can absorb iron by diffusion through the membrane, and, even if using a different and slower pathway, this is enough to stimulate the metabolism of ΔsbnAΔsfaD bacteria and probably their replication pathways. The result is reflected by the absence of the large bump at 2 Hz in the ΔsbnAΔsfaD in cTMS before exposure to iron and in the appearance of large peaks at 2 Hz and at 10 Hz, very similar to those exhibited by the WT PSDs, when iron is abundant.

In conclusion, we have evidenced different behaviour of the WT and mutant strains of S. aureus, which were chosen to study how the alteration in the intake mechanisms of iron can affect their activity. Overall, our results suggest that this metabolic pathway can be extremely important in devising new strategies to fight S. aureus infections. These results offer a first glimpse into bacterial activities that can be captured by the nanomotion sensor, which may represent a reliable addition for microbiologists and clinical researchers to perform the validation of new antibiotics or inhibitors (including bacteriophages) on time scales that are much faster than conventional growth techniques. The insights gained from the capability to look at the bacterial response in real time will advance our understanding of S. aureus physiology and pathogenesis. While further investigations to test new antibacterial treatments are underway, the protocols developed in this research can benefit our ability to fight these microorganisms since they allow following in real-time nutrient-driven microbial behaviour, such as iron acquisition pathways, which could be exploited to develop novel therapeutic strategies and antibacterial treatments.

Materials and methods

Bacterial strains and growth conditions

S. aureus Newman⁵⁵ and a ΔsbnAΔsfaD isogenic in-frame deletion mutant were used in this study. The ΔsbnAΔsfaD mutant was obtained following the mutagenesis protocol described in Schuster et al.⁵⁶. For inactivation of the sfaD gene, a 995-bp fragment and a 945-bp fragment overlapping ATG and the TAA of sfaD, respectively, were amplified by PCR using the primer couples UPFWsfaD: 5’-CCGCTCGAGTGAAATACCACACCTAACAAA-3’/UPRVsfaD: 5’-CCCAAGCTTGTTCAAGTTCATAATTCACCTC-3’ and DWFWsfaD: 5’-CCCAAGCTTCGAGAAAATAATTAAGGTTATAGAA-3’/DWRVsfaD: 5’-AAAACTGCAGTTCAATTTCGATTGCTGATTTT-3’, respectively, and cloned in pIMAY*, yielding pIMAY*ΔsfaD. Then, the latter was introduced by electroporation into S. aureus ΔsbnA (unpublished results). The functionality of the plasmid was then examined using a temperature-sensitivity test, confirming the ratio of colonies on plates incubated at 28 °C and 37 °C. Chromosomal integration via a single crossover event was achieved by incubating S. aureus ΔsbnA/pIMAY*ΔsfaD transformants at 37 °C with antibiotic selection and verified by a colony PCR using the T3/T7 primers. The second crossing-over event was encouraged by culturing/sub-culturing the selected clones at 28 °C for at least 50 generations, followed by verification of plasmid loss using para-chlorophenylalanine (PCPA), a toxic phenylalanine analogue. Once chloramphenicol sensitivity was confirmed, genomic DNA was extracted from potential candidates and the deletion was confirmed by PCR using OUTFWsfaD: 5’-TTCCCAATGGAATCATACCTC-3’/OUTRVsfaD:5’-CTTTTTTACGGGTCTGTTTTCT-3’ primers, and PCR fragments were checked by sequencing.

Bacterial strains were routinely cultured in Tryptic Soy Agar (TSA) at 37 °C for 18 h and stored with 15% glycerol (v/v) at − 80 °C. In this study, a chemically-defined minimal medium, namely the Tris Minimal Succinate (TMS) was prepared according to a previously described protocol⁵⁷, with minor modifications. Briefly, all glassware used to prepare the medium was pre-treated with 0.1 M HCl for 24 h and washed at least three times with Double Distilled H₂O (ddH₂O), to avoid iron contamination. TMS was then treated with 5% (w/v) Chelex resin (Bio-Rad) O/N at 4 °C to chelate all trace metals, including iron (this buffer was named cTMS). After Chelex removal by filtration (Whatman no. 1 paper filter), 1 ml of MgCl₂ (95.3 mg/ml), and 1 ml of CaCl₂ (11.1 mg/ml) were incorporated, and the pH of the medium was adjusted to 7.4, then sterilized by filtration using Corning filter system (Corning) and stored at 4 °C.

When required, 2,2’-dipyridyl (DIP; Sigma-Aldrich, an iron chelator with a very high affinity that is commonly used to remove iron in biological experiments⁵⁸) was prepared as a 30 mM stock solution in ddH₂O and stored at − 20 °C. FeCl₃ (Sigma-Aldrich) was prepared as a 0.1 M stock solution in 10 mM HCl and stored at − 20 °C.

S. aureus WT and its isogenic ΔsbnAΔsfaD mutant bacteria were pre-cultured in cTMS supplemented with 300 μM DIP, to further restrict iron availability, for 16–20 h at 37 °C with good aeration (shaking 180 rpm). Next day, bacterial cells were sub-cultured into 7 ml cTMS and cTMS supplemented with 1 mM FeCl₃ and incubated at 37 °C with vigorous shaking. Bacterial growth (OD₆₀₀) was measured at selected time points.

The nanomotion setup

The naonomotion experiments were performed on an Atomic Force Microscope, a FlexAFM (Nanosurf, Zurich, CH), which is mounted on an Olympus IX inverted optical microscope (Olympus, Tokio, JP). The images were collected using a 20 × objective through a high-speed Progres MFCool digital camera (Jenoptik, Germany). As nanomotion sensors we used AFM cantilevers, ONP-10 sensors from Bruker (Bruker Corporation, Massachusetts, USA), choosing the cantilever marked as B, with nominal elastic constant of 0.12 N/m. Each sensor was calibrated in air through the thermal noise method prior to being used in the experiments. The entire setup was placed in an environmental control chamber (a biological incubator) that ensured maintaining a 37 °C temperature throughout each experiment.

Sensor preparation

The preparation protocol involved exposing the sensor to 10 μl of (3-Aminopropyl)triethoxysilane (APTES) at 10% v/v concentration for 15 minutes⁵⁹. After the exposure, the sensor was thoroughly rinsed with ddH₂O and immediately exposed to a 10 μl droplet of PBS containing the bacteria and incubated for 10 min. The sensor bearing the bacteria was then transferred to the analysis chamber containing 3 ml of cTMS to start the nanomotion experiments.

The nanomotion experiment

The nanomotion experiments involved the immobilization of bacteria on the surface of the APTES-activated sensor and the monitoring of the sensor’s fluctuations induced by the nano-scale vibrations and fluctuations of the S. aureus.

To ensure the stability of the bacterial viability and the consistency of the resulting sensor’s fluctuations, we collected single bacterial colonies of the WT S. aureus from TSA plates. These were inoculated in 1 ml PBS and transferred directly to the sensor-immobilization procedure. The sensor bearing the bacteria was kept in cTMS for several hours (approximately 7 h; longer than the typical 3–4-h timeframe of the nanomotion experiments), while monitoring its oscillations and collecting optical images every 20 s to investigate presence and number of the bacteria on the sensor. These preliminary nanomotion experiments were performed in cTMS and lasted for several hours (more than 7 h) (Supplementary information 1).

For the subsequent nanomotion experiments (non-synchronized nanomotion), we collected WT and isogenic ΔsbnAΔsfaD mutant bacterial cells directly from TSA plates. The single bacterial colonies were inoculated in 1 ml cTMS and incubated at 37 °C for ca. 20 h. The following day, bacterial cultures were washed twice in PBS and diluted to an OD₆₀₀ = 0.4, prior to immobilization on the sensor.

At first, the sensor bearing the bacteria was kept in the analysis chamber with cTMS to perform a measurement of the base movement of the cells, while optical images were collected every 20 s. After two measurements in this condition, lasting 30 min each, we added FeCl₃, to achieve a final concentration of 1 mM. In this condition, we performed at least four, 30-min measurements, both with the nanomotion sensor and optical imaging. This allowed determining the variation of bacterial activity when exposed to an iron-enriched environment over a time window of more than 2 h.

Synchronized nanomotion analyses

The main nanomotion experiments involved a preliminary iron-starvation step and a subsequent reloading period. This ensured that all bacteria started the nanomotion measurements in a similar condition, with a uniform exposure time to cTMS and to FeCl₃. We named this condition synchronized nanomotion, since this protocol ensures all bacteria start from a common metabolic status. To achieve this, the single bacterial colonies obtained from the TSA plates were inoculated in 1 ml cTMS, supplemented with 300 µM DIP and incubated at 37 °C for ca. 20 h. The following day, bacterial cultures were pelleted upon centrifugation (10 min at 3,000 × g), resuspended in 1 ml of cTMS and incubated for 2 h to ensure bacterial wakening in absence of DIP. Finally, bacterial cells were washed twice in PBS and diluted to an OD₆₀₀ = 0.4, prior to immobilization on the sensor.

Similarly to the non-synchronized nanomotion, the sensor bearing the bacteria was kept in the analysis chamber with cTMS to perform a measurement of the base movement of the cells, while optical images were collected every 20 s. After two measurements in this condition, lasting 40 min each, we added FeCl₃, to achieve a final concentration of 1 mM. In this condition, we performed at least four 40-min measurements, both with the nanomotion sensor and optical imaging. This allowed determining the variation of bacterial activity when exposed to an iron-enriched environment over a time window of more than 2.5 h.

Data collection and data analysis

The time-dependent vertical fluctuations of the sensor, resulting from the motility of the biological specimens, were recorded using a custom LabView software to control a NI USB-4431 card (National Instruments, USA).

The software collects raw data at a 15 kHz rate, and, at the end of each experiment, the time dependent sensor fluctuations were analysed using a second custom software that calculates variance and Power Spectral Density (PSD) of the vertical fluctuations⁵⁹. We focused our analysis on the 1–100 Hz range, since these are the most interesting frequencies to correlate bacterial motility and activity. A particular feature of these spectra is the low-frequency behaviour, which can be fitted by a 1/f^α curve, where f is the frequency of the fluctuation. Indeed, 1/f^α fluctuations are widely found in nature and such curves are the fingerprint of a large number of stochastic behaviours which are typical of biological systems. During 80 years since the first observation by Johnson⁶⁰, long-memory processes with long-term correlations and 1/f^α behaviour of power spectra at low frequencies f have been observed in physics, technology, biology, astrophysics, geophysics, economics, psychology, language and even music⁶¹. Indeed, this function is a commonly used analysis tool of noise-like signals, where α (typically between 0 and 2) is a signal-dependent parameter which provides information on the noise-producing system. Our PSD graphs are represented in log–log scales, where such 1/f^α curves are translated to linear plots where α represents the slope of the fitted curves. In our analysis, we limited the fit to the 15–100 Hz area.

All nanomotion results, graphs, PSD, histograms and optical images are representative of a minimum of 10 independent experiments. All nanomotion results, namely the average variance values, as well as the determination of slope changes in the PSD spectra, were compared using a one-way Anova test. These analyses and fits were carried out using Origin Pro (OriginLab Corporation, Northampton, MA, USA).

All phenotypic experiments were performed in two independent biological replicates; each one performed in triplicate. The statistical analysis was performed with the GraphPad Prism software, version 8.0.1, by using two-way ANOVA, followed by Bonferroni’s multiple-comparison test.

Author contributions

SH optimized and performed the microbiology experiments, analysed the data and wrote the manuscript; MG performed the nanomotion experiments and contributed to writing the manuscript; SD contributed to performing the nanomotion experiments; SB provided funding and coordination; RM contributed to performing the microbiology experiments; EF developed and performed the microbiology experiments, analysed the data, provided funding and wrote the manuscript; GL developed and performed the nanomotion experiments, analysed the data and wrote the manuscript. All authors read, commented, and contributed to the discussion and manuscript preparation.

Data availability

The raw data supporting the conclusions of this article is stored at the ISM-CNR and the Urbino University servers and will be made available by the authors, upon reasonable request without undue reservation.

Declarations

Competing interests

The authors declare no competing interests.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-16164-5.