Nanomotion of bacteria to determine metabolic profile

Nanotechnology is a valuable and promising tool for medicine, especially in diagnostics and the fight against difficult-to-cure diseases.^1–3 However, for its use in the prevention and treatment of diseases, there are significant limitations associated with the need for long-term preclinical testing of materials and drugs;⁴ for example, high levels of acute and chronic toxicity has been detected for many nanoparticles.⁵ Fortunately, no such restrictions apply to in vitro diagnostic tests, for which the most important criteria are high sensitivity, specificity, and reproducibility, which can be achieved with modern nanotechnology tools.

For example, in 2003, a micromechanical system made it possible to detect Salmonella enterica with only 25 bacteria attached to a microcantilever.⁶ In 2013, it was discovered that bacteria have characteristic nanomotion that can be detected by analyzing the oscillations of an AFM cantilever.⁷ This nanomotion corresponds to the microorganism’s subnanometer-scale movements, potentially as a result of metabolic activity and cell membrane dynamics (e.g., undulations of mesosomes during the synthesis of adenosine triphosphate). The possible factors of nanomotion are active transport of organic materials and organelles inside cells, energy transformation in biochemical reactions, oscillations of molecular motors (in bacteria with flagella), and nanomotility.⁸ One reason for cantilever nanomotion is hydrodynamic interaction with partially attached bacteria. In this case, the kinetic energy of the bacteria attached to the cantilever’s surface is transformed into vibrational motion of the cantilever beam. AFM-based determination of microorganism nanomotion allows not only to detect the presence of living bacteria but also to determine their antibiotic resistance within an hour or even less.⁹ Nowadays, this method of detecting cellular nanomotion is even more widespread. In particular, it is possible to detect the reaction of cancer cells to different chemotherapy,^10,11 detect metabolic patterns of neuroblastoma cells,¹² evaluate the effects of oxidants on genetically modified fibroblasts,¹³ and register certain physiological processes, such as phagocytosis.¹⁴ Moreover, AFM can be used as a nano-stethoscope to study oscillations coming from the internal live processes of larger creatures such as insects.¹⁵ 

The method of determining bacterial antibiotic resistance by analyzing cantilever oscillations has the advantages of being highly sensitive and relatively fast compared to traditional clinical methods, thereby allowing the rapid prescription of etiotropic antibiotic therapy. Lupoli et al. showed that the nature of bacterial nanomotion depends on their biochemical profile and enzyme activity,¹³ but the specificity and limitations of this method are yet to be determined. The purpose of the study reported herein was to develop a method for obtaining a nanomotion-based footprint of bacterial metabolic activity and to show that bacterial metabolism is stimulated if cultivated in a nutrient-rich medium.

To collect nanomotion signals from bacteria samples, the microorganisms were attached to functionalized cantilevers. Figure 1 shows SEM images of the cantilevers obtained using a JSM-IT300LV electron microscope (JEOL, Japan) with low-energy secondary electrons at an operating voltage of 5–20 kV and an electron beam probe current of less than 0.25 nA to minimize the effect of charge accumulation on the biological samples. The resolution of the obtained images was 12 nm. As can be seen, the density of microorganisms fixed on the cantilevers is quite high to allow us to detect additional motion coming from the bacteria.

The cantilever motion was recorded using an AFM (NTEGRA; NT-MDT Spectrum Instruments, Russia) as the detection system and soft cantilevers (MLCT-A; Bruker, USA; f = 22 kHz, k = 0.07 N/m). The deflection signal from the cantilever (hereinafter referred to as the DFL signal, being the voltage difference between the upper and lower sections of the AFM photodiode) was recorded at 100 Hz using the Nova SPM software (NT-MDT Spectrum Instruments). The cantilever can sense an external force produced by bacteria and convert it into nanomotion that is detectable by the AFM detection system. However, this method is also affected by external sounds and vibrations, which must be accounted for before analyzing the signal caused by bacteria.

The signals of the cantilever motion with and without bacteria attached were recorded by the AFM detection system. The laser spot was positioned on the free end of the cantilever, and the reflected laser beam was collected by a four-section photodiode. All cantilever oscillations translated into laser beam motion on the photodiode were recorded as the DFL signal (Fig. 2).

An AFM has three main elements: (i) a mechanical part in the form of a piezoelectric element that ensures precise movement of the sample along the X, Y, and Z axes and a cantilever whose deflections allow the surface of the sample to be studied while scanning; (ii) an optical part that allows the cantilever deflections to be recorded: when the interaction force between the cantilever tip and the sample changes, the cantilever bends and the reflection angle of the AFM laser focused on the cantilever tip changes; (iii) a feedback and signal correction system that provides a mechanism for positioning the tip, comparing the signal with the reference, and correcting it. The canonical use of AFM is surface scanning, obtaining additional information about tribological,¹⁶ profilometric,¹⁷ viscoelastic,¹⁸ and other properties of the scanned surfaces.¹⁹ Recently, thanks to its use in biomedical research, the range of AFM applications has increased drastically, such as inserting and fishing for molecules,²⁰ studying intercellular adhesive contacts,²¹ nanosurgical manipulations,²² and much more. A possible non-canonical use of AFM is the present work, where the cantilever oscillates freely without entering into feedback with the surface.

To determine the viability and metabolic activity of bacteria, cantilevers with bacteria were placed in culture dishes containing different media. Normal saline solution (NSS) was used as a medium that provides no nutrients to bacteria and so does not support their metabolic activity. To show that bacterial nanomotion is related to metabolic activity, nutrient-rich media (which satisfy all nutritional needs and ensure maximum metabolic activity) were also used in the form of meat-peptone broth (MPB) and lysogeny broth (LB) with differing salt concentration. In the event of bacterial death, the nanomotion stops and the DFL signal becomes almost that for an empty cantilever. If the bacteria are alive and actively metabolizing, this is reflected as an increased amplitude of cantilever oscillations.

The raw oscillation signals collected by the AFM were filtered using a high-pass fast-Fourier-transform filter (Origin 2021b; OriginLab Corp., USA) to remove the DC background component. Typical signals before and after filtering are shown in Fig. 3 for the control experiment (empty cantilever) [Fig. 3(a)] and the cantilever motion with E. coli 321 bacteria attached [Fig. 3(b)].

As can be seen, the overall amplitude of the signal sampled from the cantilever with bacteria attached is much higher than that for the empty cantilever, and the motion caused by external noise and vibration is much smaller than the signal generated by the bacteria. Therefore, to compare signals from different bacterial strains, we use the variance as a function of time $σt2=∑xt−μ2/N$ (squared standard deviation) of the filtered signal for a small time period, where μ is the average value of the signal and N is number of processed data points. The typical variance values are shown in Fig. 4.

Comparing the control data (empty cantilever) and the data collected from the cantilever with bacteria attached shows that the power density during bacterial nanomotion increases sharply by up to 100 times in the range of 3–5 Hz, while the peak around 8–10 Hz increases only slightly (Fig. 6). Therefore, it would appear that the characteristic frequencies of vibrations caused by bacterial nanomotion are in the range of 3–5 Hz. Because the power of the signal at 3–5 Hz is much higher than that elsewhere in the spectrum, it makes sense to focus on the overall amplitude characteristic of the signal.

To show that bacterial metabolic activity generates nanomotion that can be sensed by a cantilever, we compare the signals coming from the cantilever (with bacteria attached) when immersed in nutrient-poor and nutrient-rich media (Fig. 7). When immersed in NSS, the S. aureus 2879M bacteria stopped functioning (moving) after 5 min, and the signal dropped to the control level (empty cantilever). Therefore, it would appear that bacteria cannot survive in pure NSS. By contrast, placing bacteria in a nutrient-rich medium [MPB and LB (Lennox) for S. aureus 2879M] led to a stable signal from the cantilever during the full time of the experiment (1000 s). However, using a nutrient medium that was not optimal for the particular bacterial strain led to a decrease in the overall amplitude of the DFL signal. For example, for the S. aureus 2879M bacteria, increasing the salt concentration in the medium to 10 g/l (LB Miller) resulted in a suppressed bacterial metabolism and a decrease in the overall DFL signal amplitude, although its level remained statistically distinguishable from the control (empty cantilever). It is interesting to note that the bacteria have particular food preferences. Analyzing the DFL signals shows that the amplitude decreased significantly in the experiments with NSS for all strains, compared with the richer media such as LB.

The results of statistical analyses for the different types of media are presented in Fig. 8. As can be seen, the cantilever motion amplitude was quite different for the different media. In particular, S. aureus 2879M and Proteus mirabilis 649-2 were highly sensitive to high salt concentrations and exhibited a weak response to the LB (Miller) medium, whereas E. coli 321 was indifferent to the salt concentration but experienced sharply reduced metabolic activity and nanomotion in the MPB solution (the DFL signal was even lower than that in NSS). By contrast, Klebsiella pneumoniae 173-p2 preferred a high salt concentration, and its maximum metabolic activity was recorded in the LB (Miller) medium.

The existence of bacterial nanomotion has been shown previously by many methods, including direct ones such as plasmonic imaging of the Z motion of bacteria,²³ tracking the submicrometer XY motion of attached bacteria,²⁴ sensing of attached bacterial vibrations with a resonant crystal,²⁵ and subcellular fluctuation imaging.²⁶ Willaert et al. presented results rather similar to ours, where the metabolic activity of yeast was inhibited by phosphate-buffered saline and blocked by ethanol.²⁷ However, profiling by nutrient media not only shows that nanomotion occurs primarily because of the metabolic activity of bacteria but also allows us to identify the specific reaction of different strains to the composition of the nutrient medium, moreover emphasizing the specificity of the detected nanomotion depending on the bacterial strain.

The AFM-based approach described herein offers detection of nanomotion caused by living organisms such as bacteria. By analyzing the spectra of the recorded oscillations, it is possible to comprehend the physiological processes of bacteria non-invasively. The nanomotion sensed by the AFM cantilever was found to be specific for different strains of bacteria and was linked to their metabolic profile. Such high specificity in detecting the level of bacteria metabolism could be used for express antibiotic resistance profiling as well as for identifying the bacterial type. Moreover, the proposed method offers significantly reduced times for diagnosis and appointment of etiotropic therapy (from several days to several minutes), which will contribute to reducing mortality from infectious diseases.

This work was supported by the Russian Science Foundation (Grant No. 22-14-20001).

The authors have no conflicts to disclose.

The data that support the findings of this study are available within the article.

Svetlana N. Pleskova received a Ph.D. degree in human physiology, immunology, and allergology in 2001 and an M.D. degree in 2010, both from Nizhny Novgorod State University. She is now a Professor in the Department of Physiology and Anatomy at Lobachevsky State University of Nizhny Novgorod and the Department of Nanotechnology and Biotechnology at Alekseev Technical State University of Nizhny Novgorod. She is also a leader of the scanning probe microscopy laboratory. She was the laureate of the business award “Person of the Year” in 2022 in Nizhny Novgorod. Her research areas are biomedical studies using AFM and ion-conductance microscopy and studying antibiotic resistance using AFM.

Ekaterina V. Lazarenko is currently a Ph.D. student in the Department of Nanotechnology and Biotechnology at Alekseev State Technical University of Nizhny Novgorod. She also has a bachelor’s degree and a master’s degree in industrial biotechnology and bioengineering. A member of Professor Pleskova’s scientific group since late 2019, her scientific interests are in studying the nanomotion of bacteria to determine their metabolic profiles.

Nikolay A. Bezrukov is a junior researcher in the Research Laboratory of Scanning Probe Microscopy at Lobachevsky State University of Nizhny Novgorod and has been a member of Professor Pleskova’s scientific group since late 2018. He received a bachelor’s degree in biotechnology from Alekseev State Technical University of Nizhny Novgorod in 2020, then a master’s degree in the same specialty in 2022. Since 2022, he has been a Ph.D. student at Lobachevsky State University of Nizhny Novgorod, specializing in human and animal physiology. His interests include human physiology, immunology, microbiology, biotechnology, and engineering, and his main scientific topic is interactions in the human endothelial–blood cellular system, especially transendothelial migration of neutrophil granulocytes. He is currently one of the main operators of the scanning ion-conductance microscope.

Ruslan N. Kriukov received a Ph.D. degree in condensed matter physics in 2019 and is now an associate professor in the Department of Semiconductor Physics, Electronics and Nanoelectronics at Lobachevsky State University of Nizhny Novgorod. His current research is focused on AFM, chemical analysis of semiconductors, and X-ray photoelectron spectroscopy.

Aleksey V. Boryakov was born in 1986 and obtained an M.Sc. degree in 2010 and a Ph.D. degree in 2015. He is currently a senior researcher in the Research Educational Center for Physics of Solid-State Nanostructures at Lobachevsky State University of Nizhny Novgorod. His research interests are in characterization of nanostructures, biophysics systems, and nanoelectronics using microscopic and spectroscopic methods of investigation. The main methods of physicochemical interest are SEM, X-ray microanalysis, X-ray fluorescence analysis, atomic emission spectroscopy with inductively coupled plasma, X-ray photoelectron spectroscopy, and preparation and modification of the surfaces of nanostructures with ion beams.

Maxim E. Dokukin received a Ph.D. degree in condensed matter physics from Moscow State University in 2004 and is currently an associate professor in the Sarov Physics and Technology Institute at the National Research Nuclear University MEPhI. His research interests are in the mechanics of cells and developing methods for characterizing soft biological and polymeric materials at the nanoscale with the help of AFM.

Sergei I. Surodin was born in 1991 and obtained an M.Sc. degree in 2014 and a Ph.D. degree in 2019. He is currently Chief of Unit of Branch of FSUE “Russian Federal Nuclear Center – All Russian Research Institute of Experimental Physics” Research Institute of Measuring Systems named after Yu. Ye. Sedakov. His research interests are in characterizing nanostructures and biophysics and nanoelectronics systems using microscopic and spectroscopic methods of investigation, including SEM, X-ray microanalysis, X-ray fluorescence analysis, X-ray photoelectron spectroscopy, and preparation and modification of the surfaces of nanostructures with ion beams.