Integrating FRET and Molecular Dynamics Simulation for Single‐Molecule Aptameric Detection of Staphylococcus aureus IsdA Surface Protein

ABSTRACT

Staphylococcus aureus is ranked among the top five most common foodborne pathogens affecting public health and the economy worldwide. To improve detection and reduce diagnostic burdens, several detection methods from traditional culture‐based techniques to biosensing platforms have evolved. Among several markers, surface proteins are considered to be the most important markers due to the specific roles they play in the survival and colonization of the bacterium on hosts. Here, we have developed a detection platform for a key surface protein, iron‐regulated surface determinant protein A (IsdA), using a combination of computationally developed aptamer and single‐molecule fluorescence resonance energy transfer (smFRET). Computationally generated RNA aptamer incorporated into the FRET‐based sensor show high specificity detection of IsdA with a detection limit down to 0.6 pM and dynamic range extending to ∼10 nM. Molecular dynamics (MD) simulations show distinct conformational flexibility of the unbound aptamer and a reduced flexibility for the aptamer‐IsdA complex, corresponding to the experimentally observed higher FRET efficiencies. The FRET‐based single‐molecule aptasensor that we developed has great potential for rapid monitoring S. aureus. Further, the developed approach has the potential to be broadly applicable across diverse fields of biotechnology including environmental monitoring, forensic analysis, and clinical diagnostics.

Graphical Abstract and Lay Summary

A single‐molecule FRET platform for high sensitivity detection of IsdA protein using a computationally derived RNA aptamer.   

1. Introduction

Food safety has become a major concern around the world due to many prevalent foodborne diseases (FBD) affecting human health. The World Health Organization estimates around 600 million illnesses (1 in 10 people) due to contaminated foods [1] and around 420,000 people die each year. The contamination can occur during food preparation, processing, distribution, or transportation [2] or due to poor personal hygiene [3]. Among foodborne pathogens, Staphylococcus aureus is recognized as a significant one causing nearly 250,000 cases each year in the United States alone [4, 5].

S. aureus is a gram‐positive cocci shaped bacteria. Human invasion of S. aureus causes mild skin infections to life‐threatening diseases such as pneumonia, endocarditis, osteomyelitis, arthritis, and sepsis [6, 7]. Moreover, S. aureus produces toxins including staphylococcal enterotoxins (SEs), which cause food poisoning (SFP) [3, 8], which causes nausea, vomiting, diarrhea, abdominal cramps, and in some cases death [9]. While antibiotics (e.g., methicillin) are the main treatment option, S. aureus has rapidly developed antibiotic resistance [10] thereby making it difficult to avoid the fatal infections [11]. Therefore, there is a pressing need of a platform to enable reliable early detection of S. aureus [12] to prevent the contaminated foods entering into the market.

Current promising methods for S. aureus detection are based on the detection of whole cells, nucleic acids, surface proteins, and/or enterotoxins [13, 14]. The present culture‐based methods for the whole cell detection such as Baired Parker (B‐P) plate count and most probable number are considered as the gold standard [15] and are inexpensive and provides quantitative information such as bacteria counts on foods. However, these methods typically require 3–7 days of incubation for results [16]. More recent immunological methods use enterotoxins for the identification and quantification of S. aureus. These involve immuno‐labeling, enzyme‐linked immunosorbent assay (ELISA) [17], immunofluorescence biosensing [18], and immunomagnetic separation (IMS) methods [19], which are based on antigen‐antibody interactions and exhibit high‐specificity detection however, they often require labeling of antibodies. In addition, they are susceptible to heat damage of antibodies, specially for processed foods [20]. Techniques such as colorimetric and fluorescence‐based methods, which involve conjugating aptamers with nanomaterials [21], carbon dots [22], and quantum dots [23]. Though these applications are promising, the complexity of nanomaterial synthesis, difficulty in controlling size that provides desirable properties, and in some cases, unwanted aggregation can hinder detection [24]. Similarly, although commonly used and is rapid and inexpensive, the polymerase chain reaction (PCR)‐based techniques cannot distinguish between live and dead cells, which results in difficulty in the detection of certain markers such as SEAs [25].

Recent search for novel and sensitive detection methods has shown great promise for aptamer‐based technologies as they are proven to offer several advantages over traditional methods [26]. For example, unlike antibodies that are relatively expensive to produce and have limited shelf life, aptamers can be stored for a long period of time under ambient environment without compromising affinity and specificity [27, 28, 29]. Therefore, since the development of the first DNA aptamer for Staphylococcal enterotoxin B [30], there was a significant extension of the approach [31, 32]. More recently, aptamers have been developed for the detection of S. aureus whole cell [33, 34, 35], surface protein A [36], and enterotoxin A and B [37, 38]. Although these are important accomplishments, the aptamer‐based detection of S. aureus surface proteins is still underdeveloped.

Surface proteins play critical roles for the survival and colonization of S. aureus [39]. There are around 20 cell‐wall anchored surface proteins identified in S. aureus, which function as adhesions to the host tissue [40]. Among them, iron‐regulated surface determinant protein A (IsdA) is one of the key surface proteins in S. aureus that facilitates heme uptake by binding to human hemoproteins and removing heme molecules [41]. Located on the cell surface, IsdA represents as one of the best markers when thinking about protein‐based detection of S. aureus. We recently reported a proof‐of‐concept aptamer‐based single‐molecule sensor for IsdA, however, the binding was somewhat hindered due to design limitation [42]. Unlike the partially blocked aptamer used before, in this study, we demonstrated a simple yet more sensitive sensor with ∼18‐fold higher sensitivity (Figure 1). The sensor enables single‐step detection of IsdA with great specificity over interfering surface proteins. This simple yet sensitive sensor for S. aureus can be adapted for single‐step detection of many other pathogens via aptamer‐protein interactions. To complement the smFRET experimental approach and provide molecular insights into the binding behavior of the aptamer, computational simulations will be employed to reveal the structural and dynamic properties of the IsdA aptamer and its interaction with the IsdA protein. These studies will aim to identify key features influencing binding specificity and affinity, providing an understanding of the aptasensing mechanism to guide the rational optimization of the sensor design. We will perform MD simulations of the aptamer under free and protein‐bound states to offer binding predictions of the conformational flexibility and stability, which are critical for enhancing sensor performance. The MD simulations will provide a theoretical framework to interpret binding dynamics and systematically improve the sensitivity and adaptability of the sensor for S . aureus detection.

FIGURE 1.

Working principle of the IsdA sensor using single‐molecule platform where hundreds of sensors are immobilized on a microscope slide. Only four sensor molecules are shown for example. The sensor (DNA construct) is composed of four oligonucleotides, two of which are labeled with a Cy3 or Cy5 fluorophore. The construct is expected to show a low FRET state in the absence of IsdA but switch to a high FRET state upon IsdA binding.

2. Materials and Methods

2.1. Chemical Reagents

Biotinylated bovine serum albumin (bBSA), 6‐hydroxy‐2,5,7,8‐tetramethylchroman‐2‐carboxylic acid (Trolox), magnesium chloride, protocatechuate 3,4‐dioxygenase (PCD) and ethylenediaminetetraacetic acid disodium salt (EDTA) were purchased from Fisher Scientific. PCD was prepared by dissolving it in a PCD stock buffer (pH 8.0) consisting of 100 mM Tris‐HCl, 50 mM KCl, 1 mM EDTA, and 50% glycerol. Streptavidin, sodium chloride, and protocatechuic acid (PCA) were purchased from VWR. Recombinant S. aureus IsdA, recombinant S. aureus clumping factor A (ClfA), and Klebsiella pneumoniae outer membrane protein A (OmpA) were purchased from My BioSource. Bovine serum albumin (BSA) and fibronectin binding protein (FnbP) were purchased from Thermo Scientific. Protein A from S. aureus was purchased from MP Biomedicals.

2.2. Aptamer Selection

Aptamer was first selected computationally and custom synthesized from IDT DNA. While experimental SELEX is a common platform for aptamer selection, computational SELEX uses free energy calculations to screen aptamers. The use of computational simulations for aptamer discovery is uniquely advantageous for its high potential, as it enables reagent‐less, speedy search of suitable aptamer candidates using a much larger RNA/DNA library [43, 44]. Our previous studies– both computationally and experimentally—identified that the T3‐117558 RNA aptamer binds efficiently with the IsdA protein (reported sensitivity 11 pM) [42] and computationally [45], and hence we explored a new sensor using this aptamer (Table S1). The aptamers were selected computationally by generating an enriched library of RNA sequences according to the pattern library method [46]. These selections are majorly based on the probability of forming complex secondary structures. Approximately 120,000 RNA sequences were screened based on their predicted minimum free energies using RNAfold from the ViennaRNA package [47]. Sequences with the lowest free energy structures were compiled into a refined pool for further analysis. We used RPISeq to predict the binding propensities between each RNA sequence and the IsdA protein. The screening process was repeated iteratively using the support vector machine and random forest models associated with RPIseq until the RNA sequences with predicted binding probabilities exceeding 85% were identified. The sequences showing strong binding were then ranked based on their binding free energies and then the best binder was used for the detailed MD simulation and FRET experiments [48].

2.3. Single‐Molecule Fluorescence Microscopy

All FRET experiments were performed using custom‐made flow cells using quartz slides, parafilm, coverslips, and tubing as described in our previous publications [50, 51]. The slides were functionalized by injecting ∼300 µL of 1 mg/mL biotinylated‐BSA (5 min incubation), followed by ∼300 µL of 0.2 mg/mL streptavidin (3 min incubation), and finally ∼300 µL of 1 × TAE buffer to remove any unbound molecules.

The fluorescence imaging experiments were conducted following the previously optimized method [50, 52]. The detailed preparation of the aptasensor for single‐molecule experiments is described in supplementary information. Briefly, the aptasensor (∼10 pM final concentration) was dissolved in an imaging buffer containing 1 × TAE, 150 mM NaCl and an oxygen scavenging system (OSS) (4 mM Trolox, 10 mM PCA, 100 nM PCD) [53, 54] and was injected into the flow cell. The OSS is needed to slow down the photobleaching of fluorophores. After ∼1 min incubation, the flow cell was flushed with ∼400 µL of imaging buffer to get rid of any unbound aptasensor molecules. In this process, the sensor molecules are immobilized on the slide surface via biotin/streptavidin interaction. After immobilization of the sensor molecules, the solution containing the protein target at a given concentration and OSS was injected and FRET movies were recorded after 20 min of incubation. Control experiments were performed in the absence of the target protein—IsdA. All FRET experiments were performed using a prism‐based total internal reflection fluorescence microscope IX73 from Olympus with an UPLSAPO 60× water immersion objective. The fluorescence emissions were acquired using an iXon Ultra897 EMCCD camera [50]. During FRET movies, the Cy3 fluorophores were continuously excited using a 532 nm green laser while recording the fluorescence emission from both the Cy3 and Cy5 fluorophores simultaneously on the green and red channels (512 × 256 pixels) using an EMCCD camera. The exposure time used was 100 ms. The presence of an active FRET pair was confirmed toward the end of each movie by directly exciting the Cy5 fluorophore using the red laser. All of the FRET experiments were performed at room temperature (23°C).

2.4. Single‐Molecule Data Acquisition and Analysis

Data acquisition and processing were done using the single‐molecule FRET data acquisition and analysis procedures described previously [50]. Briefly, Single.exe software generates a.pma file for each recorded FRET movie. Next, the.pma files were fed into the IDL program to obtain fluorescence signals from each individual molecule by pairing the corresponding donor and acceptor in both channels. Once this mapping is done, Matlab was used to manually screen the generated intensity−time and FRET−time traces. The single molecules that show evidence for the presence of both the Cy3 and Cy5 fluorophores with single‐step photobleaching were manually selected for further analysis. The FRET efficiency was calculated using background‐corrected fluorescence intensities of the acceptor (I_A) and donor (I_D) fluorophores using the equation, FRET Efficiency (E_FRET) = I_A/(I_D + I_A). The dynamic and static molecules were visually distinct from one another (Figure 2).

FIGURE 2.

A) Microscope field of view of the surface immobilized sensor molecules in the absence (−Target) and presence of the target (+Target). Images belong to the movie when the green laser was ON but the red laser was OFF. The scale bar is 5 µm. (B) Typical intensity−time and corresponding FRET efficiency traces. The left panel shows FRET traces in the absence of IsdA. Three types of traces (low‐FRET static, mid‐FRET static, and low‐FRET dynamic) were observed. The right panel shows FRET traces in the presence of IsdA. Two types of traces (high‐FRET static, high‐FRET dynamic) were observed. FRET represents FRET efficiency.

3. Results and Discussion

3.1. Computational Predictions of Aptamer Behavior

We first report how the mean structure of the aptamer changes upon sensing IsdA during the 100‐ns simulation. As shown in Figure 3A, the overall conformation of the aptamer core is largely preserved upon protein binding. However, an interesting conformational switch of the aptamer was observed. The unpaired middle part of the free aptamer was in a relatively “down” position, while upon protein binding it raises up and interacts with GLN42, LYS31, GLN27, THR77, SER4, PRO6, ASN72, and PRO29. Beyond the comparisons at the level of mean structures, dynamic fluctuation provides additional important information. The overall RMSD evaluates the global stability of molecular systems by comparing the deviation of atomic positions over time from a reference structure. This metric is useful in assessing the stability of protein‐ligand complexes and understanding the dynamic behavior of biomolecules during simulations [55]. RMSD of the aptamer and complex were plotted as a function of simulation time for visualizing the overall stability of molecular systems by comparing the deviation of atomic positions over time from a reference structure. This metric is useful in assessing the stability of protein‐ligand complexes and understanding the dynamic behavior of biomolecules during simulations [55]. RMSD of the aptamer and complex were plotted as a function of simulation time for visualizing the overall stability of the RNA aptamer system. Each frame was fitted to the first frame as the reference structure, and this was used for the RMSD calculation. The ptraj module of Amber 20 was used to perform the alignment to the reference structure. The RMSD plot for the aptamer alone shows the conformational changes during a 100 ns MD simulation (Figure 3B). The aptamer exhibits an initial rapid increase in RMSD of ∼8 Å within the first 5 ns, corresponding to the relaxation and equilibration of the system. After this initial period, the RMSD stabilizes, with the system fluctuating around a consistent average of ∼7.5 Å. This plateau indicates that the structure ensemble of the aptamer converges into a stable free‐energy basin, and no drastic conformational changes occur during the remainder of the simulation. Minor fluctuations observed between 48–50 ns could arise from the conformational flexibility within the loop region, where the RNA folds back on itself to form a double‐stranded region stabilized by intramolecular interactions [56, 57]. Despite these minor fluctuations, the overall structural framework of the aptamer remains intact, highlighting its stability during the simulation. In comparison, the RMSD of the aptamer‐IsdA complex (Figure 3B) demonstrates enhanced stability relative to the free aptamer. The RMSD for the complex shows a significantly lower average (∼3 Å) throughout the simulation, suggesting that the binding of IsdA stabilizes the conformation of the aptamer. The lower deviation in RMSD for the complex reflects a reduction in conformational flexibility due to interactions between the aptamer and the protein, which constrain the dynamic behavior of the aptamer.

FIGURE 3.

MD simulation and data analysis of the free aptamer and aptamer complexed with the IsdA protein. (A) The superposition of the ensemble averaged free aptamer (blue) structure and the corresponding aptamer‐IsdA complex structure (red) in a 3D cartoon representation. (B) The deviation of each conformation snapshot from the starting conformation, indicated by the overall RMSD, as a function of time. One can observe how each system dynamically evolves into a more stable state during the simulation. The lower RMSD for the complex highlights the stabilizing effect of the protein binding to the aptamer, whereas the higher RMSD for the free aptamer reflects its greater flexibility in solution. (C) The B‐factor comparison of the free aptamer (blue) and aptamer in the presence of IsdA (red), showing changes of flexibility at a residue resolution. The peak positions in the plot such as 18A, 21U, and 29C correspond to relatively flexible regions of the free aptamer, while the trough positions such as 11U, 14G, and 34U indicate more rigid and well‐structured segments. Upon IsdA recognition, regions with increased mobility (17U, 21U, and 29C) are likely due to dynamic interactions between the aptamer and the protein, while the troughs (3U, 20A, 22A, and 33C) indicate stable binding regions. (D) MM/GB‐SA calculation provides the breakdown of binding energy contribution at the residue resolution.

Figure 4 illustrates the binding‐induced expansion of the aptamer. In the apo distance matrix (Figure 4A), contacts along the diagonal are uniformly intense, reflecting a compact core. Upon IsdA binding (Figure 4B), the same diagonal pattern persists, but the pixels at the terminal rows/columns (nt 4–8 vs. nt 33–38) shift toward lighter (bluer) colors, indicating increased separations. The difference map (Figure 4C) confirms that these near‐terminal positions experience the largest positive changes; blue hues up to the top of the color‐bar scale, while most other contacts remain near zero (white). Contact‐probability heatmaps (Figure 4D,E) likewise show that end‐to‐end interactions are significantly reduced in the bound ensemble. Finally, PCA projection (Figure 4F) separates apo (red cluster) from bound (purple cluster) frames, and the PC1 loading map (Figure 4G) pinpoints nt 4–8 and nt 33–38 as the dominant contributors to the motion. Altogether, these panels demonstrate that protein binding selectively pries open the aptamer near terminal positions while leaving the central core intact. These are great qualitative insights related to binding interactions. It is important to note that the measured FRET in the single‐molecule experiment is the reflection of the distance between the 5' and 3' ends of the aptamer (not the near‐terminal and core distance, discussed in Figure S1). Based on the FRET results, the aptamer ends come slightly closer to one another upon IsdA binding, which is possible due to the less involved terminal nucleotides in binding. To demonstrate convergence within our 100 ns trajectories, we have performed a block‐averaging analysis (Figure 4) to illustrate the results: Each 100 ns trajectory was divided into blocks, specially the final 20 ns (80–90 ns and 90–100 ns). We computed the mean structure for each block and superimposed them (Figure S1). The near‐perfect overlap of Block A (80–90 ns) and Block B (90–100 ns) indicates that, during the final 20 ns, the aptamer‐protein complex remains confined to a single conformational basin with only minor fluctuations.

FIGURE 4.

MD‐derived conformational changes of the aptamer upon IsdA binding. (A) Mean distance matrix for the 40‐nt aptamer core in the apo simulation. (B) Mean distance matrix in the IsdA‐bound simulation. (C) Difference map (bound—apo) with blue regions marking inter‐nucleotide separations. (D, E) Contact probability maps for the apo (top) and bound (bottom) ensembles. (F) Projection of all frames onto principal components 1 and 2. (G) PC1 loading map.

3.2. Single Molecule Detection of IsdA

The aptasensor that we described here has shown significantly improved detection of pathogen protein IsdA. The sensor construct is simple in design with three DNA strands and one RNA aptamer strand. The detection is based on direct binding to the aptamer rather than commonly used strand displacement strategies [42]. In the absence of the target the aptasensor shows three different intensity‐time traces (below 0.5 FRET), which are interpreted by FRET efficiency calculations (Materials and Methods) and molecular dynamic simulations (Figure 3 and Figure S1). The inter‐dye distance (R) was calculated using the following equation, where E is the FRET efficiency and R ₀ is the Förster radius—the distance for 50% energy transfer from the donor to the acceptor fluorophore.

$$E=1/〈1+{\left(\frac{R}{Ro}\right)}^6〉$$

Figure S1A shows the computationally modeled IsdA aptamer. Figure S1B represents the estimated distance between the two fluorophores of the fully closed aptamer (assuming the formation of a duplex region due to aptamer self‐folding) is ∼2 nm relating to the calculated 0.9 FRET efficiency. Any FRET efficiency lower than 0.9 corresponds to a somewhat open aptamer increasing the donor‐acceptor distance (Figure S1C–E). The static low‐FRET trace (0.25, Figure 2 left panel) corresponds to the calculated distance of 6.4 nm (contour length of 3‐bp distance), static mid‐FRET trace (0.5) corresponds to the calculated distance of 5.4 nm (contour length of 2‐bp distance) while the dynamic FRET trace (between 0.25 and 0.5) corresponds to the interchange of the conformation from the 3‐bp to 2‐bp open state. Additionally, in the absence of the target, we noted that the population of the low‐FRET traces was predominant–indicating weaker interaction between the end nucleotides. This observation was consistent with the relatively higher RMSD observed for the end nucleotides using computational simulation. Figure S1F represents the computationally predicted IsdA‐aptamer complex (see Materials and Methods). In the presence of IsdA, the sensor showed two characteristic high‐FRET traces (FRET efficiency above 0.5). Figure S1G shows a schematic of the complex corresponding to static high‐FRET (0.7, Figure 2 right panel). This observation is consistent with the computational simulations that the 5' and 3' end bases of the aptamer interact with IsdA [45]. The high FRET dynamic state shows relatively dynamic interaction between the IsdA and bases at the 5' and 3' of the aptamer (Figure S1H). Population analysis of single‐molecule FRET traces provided some valuable insights in interpreting these observations. Particularly, the low‐FRET static molecules (∼60%) in the absence of IsdA translated to high‐FRET target‐bound population (∼60%) in the presence of IsdA. Interestingly, there was also a high‐FRET dynamic population (∼20%) in the presence of IsdA, but the extent of dynamics was much lower than without protein, reflecting some dynamic interactions between end nucleotides and protein (Figure S1H). This observation aligns well with computational analysis, which showed reduced RMSD of nucleotides when the protein is bound.

3.3. Analytical Sensitivity

The analytical sensitivity of the sensor was determined by conducting a series of experiments at various concentrations of the target. The calibration curve shows the average %high‐FRET (y‐axis) and standard deviation (as error bars) from those three groups of data for a given concentration of IsdA (Figure 5). The percentage of high FRET molecules was calculated by dividing the number of high‐FRET molecules over the total number of single molecules. The standard deviations were calculated by randomly assigning single molecules from the same experimental condition into three separate groups. A clear correlation is observed between the high FRET molecules and target concentration. Around 60% high‐FRET molecules were observed at the saturating concentration of the target with a dynamic range extending to 10 nM. The concentrations of IsdA in real‐life samples is not expected to be that high and hence this dynamic range should be suitable for real‐life testing. A linear response was observed up to 1 pM target. Interestingly, this IsdA sensor shows sensitive detection due to unrestricted binding. However, a small high‐FRET population (∼10%) was observed for the control experiment defining the experimental background. This observation indicates some kind of self‐folding of the aptamer in some of the sensor molecules yielding a high FRET state.

FIGURE 5.

Determination of analytical sensitivity. A calibration curve was obtained by plotting the percentage of high FRET (target‐bound) molecules as a function of IsdA concentration. The inset shows the linear range of the calibration curve. The error bars represent the standard deviation.

The LOD of the sensor was calculated using the following equation:

$$\mathrm{LOD}=\frac{3\times \mathrm{S}{\mathrm{D}}_{\mathrm{blank}}}{\mathrm{Slope}}$$

The linear fitting function in OriginLab was applied for fitting the linear range data (inset, Figure 5) to determine the slope for LOD. The calculated LOD was 0.64 pM, which is ∼18‐fold lower compared to the strand‐displacement‐based sensor developed before [42]. Further, the dissociation constant (K _d) for the aptamer‐IsdA complex was determined from the binding data using the built‐in Hill 1 function in OriginLab (Figure 5) assuming 1:1 binding [58, 59] and found to be 3.9 ± 1.6 pM, which is also a significant improvement from previously reported K _d using a different sensor construct [42]. The Hill equation used to fit the binding data is shown below, where HF is %high FRET population, HFmax is the maximum %high FRET, HF₀ is the %high FRET in the absence of IsdA, and K _d is the dissociation constant.

$$\mathrm{HF}={\mathrm{HF}}_0+\frac{\left({\mathrm{HF}}_{\max }-{\mathrm{HF}}_0\right)\left[\mathrm{IsdA}\right]}{{\mathrm{K}}_{\mathrm{d}}+\left[\mathrm{IsdA}\right]}$$

3.4. Specificity

Specificity is an important attribute of a sensor and is typically conducted to evaluate the ability of the sensor to detect a target even in the presence of other interfering target‐like molecules. Therefore, to determine the specificity of our IsdA sensor, we selected different surface proteins of S. aureus; fibronectin‐binding protein (FnbP), clumping factor A (ClfA) and Protein A. BSA and outer membrane protein A (OmpA) of K. pneumoniae were used as non‐specific targets (Figure 6). The experiments were carried out using saturating concentrations (10 nM) of these proteins, first individually and then in the mixture containing all of these proteins.

FIGURE 6.

Determination of specificity. Specificity analysis was conducted using a nonspecific protein BSA, surface proteins of S. aureus (fibronectin‐binding protein: FnBP, clumping factor A: clfA, protein A), K. pneumoniae outer membrane protein A (OmpA), and IsdA all independently using 10 nM of these proteins and in a mixture containing all these proteins at 10 nM each. The dotted line represents the background observed without the target protein.

We conducted a series of experiments with each protein at the saturating concentration of 10 nM (previously determined with IsdA) separately as well as in a combination of all proteins. As observed, it was clear that (Figure 6) the new sensor was specific and selective towards IsdA protein. Error bars indicate standard deviation determined from randomly assigned three groups of molecules collected for each protein (target or non‐target) and control. Briefly, the %high‐FRET population in non‐specific proteins resulted in quite similar outcomes compared to the no‐protein experiment (only ∼1.3‐fold change). In the presence of IsdA alone or in the mixture of IsdA and other non‐target proteins, there is a significant (∼4‐fold) increase in the %high‐FRET, demonstrating the high specificity and selectivity of the sensor for IsdA.

4. Conclusions

IsdA is an important marker for the detection of the pathogen S. aureus, and hence, its detection is very important in the food industry, healthcare, and in biotechnology field. Despite the development of several methods for IsdA detection, a vast majority of them require a complicated experimental design in order to achieve a sensitive and specific detection. Due to the emerging need for reliable pathogen detection and monitoring capabilities, here we presented an aptasensor for high‐accuracy single‐molecule detection of IsdA and showed that it allows sensitive detection of IsdA down to 0.64 pM (∼0.064 femtomoles, considering 100 µL sample volume) with a large dynamic range extending to ∼10 nM. The MD simulations demonstrated that the aptamer exhibits inherent flexibility, particularly in its unpaired middle loop region, which plays a critical role for binding. Upon binding to the relatively rigid IsdA, this flexibility is significantly reduced, resulting in a stable aptamer:IsdA complex. Notably, the binding interaction promotes a conformational switch in the aptamer core, while retaining sufficient conformational dynamics to facilitate effective detection and sustained target engagement. These findings underscore the aptamer's malleable yet functionally robust nature, which is essential for its role as a high‐sensitivity biosensing element. The combination of molecular dynamics simulation and single‐molecule studies of aptamer and aptamer‐protein complexes enabled robust analysis of FRET data. Specifically, single‐molecule analysis aligned well with the computational predictions that the binding of the aptamer results in reduced flexibility of nucleotides within the aptamer. Additionally, we also verified the selectivity of the IsdA aptasensor in the presence of potential interfering proteins. Therefore, this simple aptasensor with single‐step detection of IsdA provides a sensitive and adaptable sensing platform for various pathogen proteins. While DNA and RNA‐based sensors like the ones used in this study offer great stability under physiological conditions, the sensor performance may be affected when using real‐life samples due to possible nonspecific interactions from other proteins. Significant changes in the pH and ionic strength can also have a negative impact on the structural integrity and 3D confirmation of both the aptamer and the protein, affecting the binding interaction, and thus some optimization would be necessary for successful detection. With optimization, the sensing strategy that we developed should enable the detection of IsdA in biological samples. More broadly, this sensing strategy can be adopted for the detection of other proteins by incorporating the protein‐specific aptamers, and hence, the scope of the strategy is broad. Overall, this work will benefit the food and healthcare industries and contribute to the advancement of biotechnology.

Author Contributions

Chamika Harshani Algama: investigation, writing – original draft, methodology, validation, visualization, writing – review and editing, formal analysis, data curation. Tracy A. Bruce‐Tagoe: formal analysis, data curation, visualization, writing – review and editing, writing – original draft, investigation, methodology. Joy Adetunji: investigation, writing – review and editing, formal analysis, supervision. Tongye Shen: formal analysis, data curation, writing – review and editing, investigation. Michael K. Danquah: methodology, funding acquisition, writing – original draft, writing – review and editing, supervision, project administration. Soma Dhakal: conceptualization, funding acquisition, writing – original draft, writing – review and editing, project administration, supervision, resources.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting information file 1: biot70101‐sup‐0001‐SuppMat.pdf.

Harshani Algama C., Bruce‐Tagoe T. A., Adetunji J., Shen T., Danquah M. K., and Dhakal S., “Integrating FRET and Molecular Dynamics Simulation for Single‐Molecule Aptameric Detection of Staphylococcus aureus IsdA Surface Protein.” Biotechnology Journal 20, no. 9 (2025): 20, e70101. 10.1002/biot.70101

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.