Nanoscopic acoustic vibrational dynamics of a single virus captured by ultrafast spectroscopy

Significance

The natural vibrations of microorganisms such as viruses, bacteria, and fungi, encode information on their mechanical properties, morphology, and three-dimensional structure. An all-optical method is demonstrated to detect and dynamically track the natural vibrational frequencies of a microorganism across the megahertz to terahertz spectral range. We uncover long-lived coherent oscillations in a single virus at room temperature that persist for many nanoseconds. These coherent signals give rise to an acoustic spectrum that is highly sensitive to the virus morphology and interactions between its glycoproteins and the environment. The methodology promises to shed light on viral dynamics without labeling and could serve as a means for viral fingerprinting.

Abstract

The natural vibrational frequencies of biological particles such as viruses and bacteria encode critical information about their mechanical and biological states as they interact with their local environment and undergo structural evolution. However, detecting and tracking these vibrations within a biological context at the single particle level has remained elusive. In this study, we track the vibrational motions of single, unlabeled virus particles under ambient conditions using ultrafast spectroscopy. The ultrasonic spectrum of an 80 to 100 nm lentiviral pseudovirus reveals vibrational modes in the 19 to 21 GHz range sensitive to virus morphology and 2 to 10 GHz modes with nanosecond dephasing times reflecting viral envelope protein interactions. By tracking virus trajectories over minutes, we observe acoustic mode coupling mediated by the local environment. Single particle tracking allows the capture of viral disassembly through correlated mode softening and dephasing. The sensitivity, high resolution, and speed of this approach promise deeper insights into biological dynamics and early-stage diagnostics at the single microorganism level.

The low-frequency vibrations of biological systems such as proteins, viruses, and bacteria arise from collective motion of all their constituent atoms. The vibrational spectra of these biological systems, therefore, reflect their three-dimensional structure and conformational flexibility, as well as critical interactions with their environment. However, detecting these low-frequency vibrations which occur in the hundreds of MHz to THz range within a biological environment has remained out of reach (1). Current methods capable of detecting narrow spectral regions of these low-frequency vibrations rely on external devices such as ultrahigh frequency (UHF) mechanical resonators (2, 3) or plasmonic nanostructures (4) that are specifically tuned to the analyte of interest. However, these platforms are incompatible within a biological environment. Further, these methods restrict the motion of the particles, and, therefore, do not allow tracking of dynamics as the particle environment changes. While studies using low-frequency Raman spectroscopy have been carried out on highly concentrated virus suspensions (5, 6), these ensemble measurements resulted in broad, poorly resolvable spectra. On the other hand, theoretical calculations using an atomistic approach (7, 8) suggested that virus identification should be possible using low-frequency Raman scattering if sufficient resolution is achieved.

In this study, we demonstrate an all-optical, ultrahigh resolution, method for detection and tracking of quantized acoustic vibrations in a small, biological particle—a single virus smaller than 100 nm. We measure acoustic spectra in the 2 to 50 GHz range with sub-GHz resolution that are especially sensitive to morphology and interactions of the viral envelope proteins with the environment. The concept of this approach, which we have termed BioSonic spectroscopy, is shown in Fig. 1. The virus sample is placed on a cover slip without further modification and interrogated with a pair of ultrashort laser pulses inside a microscope: a nonresonant pump pulse (<100 fs, 1,040 nm) to excite collective vibrations in a single virus particle, and a second, time-delayed probe pulse (<100 fs, 785 nm) to detect changes in light scattering induced by the coherent vibrations. The weak signal is isolated from the large background of backscattered light by using balanced detection and asynchronous optical sampling (ASOPS) (9, 10), a method by which the interpulse delays are rapidly scanned up to the laser pulse period (~10 ns) in sub-milliseconds to reduce laser and environmental noise. Further details of the optical setup and the signal isolation procedures are provided in supplementary information (SI Appendix, sections I and II).

Fig. 1.

Principle of BioSonic spectroscopy. (A) Natural frequencies of biological systems at different length scales. Simplified microscope setup showing illumination and detection geometry. The ASOPS sequence uses a fixed frequency offset to generate a rapid, linear scanning of the time delays up to the laser pulse period. (B, i) Motion of a nanoparticle (NP) on a substrate. (B, ii) Time-domain response showing simultaneous excitation of breathing (br), angular, and axial (ax) modes in the NP. (B, iii) Corresponding spectrum by Fourier transformation of the time-domain signal. (B, iv) Different types of vibrational modes that the NP may experience. (C) Representative time response and spectra of a single ~100 nm gold particle (AuNP), Top, and a single lentivirus particle, Bottom. The Inset shows the VSV-G glycoprotein and membrane envelope (blue) and capsid containing the green fluorescence protein (GFP) gene and protein (maroon).

The time-domain signal includes various contributions depending on the resonance condition. The coherence-only (i.e., oscillatory) component of the signal reflects the acoustic phonon spectrum, composed of low-frequency Raman-active vibrational modes in the 0.1 GHz to 2 THz spectral region. Measurement of the pump–probe response provides information on both the acoustic frequency, $v$, as well as the phonon dephasing time, $\mathrm{\Gamma }$. This dephasing is dictated by an interplay between the intrinsic anharmonicity of the lattice and other extrinsic factors such as material impurities or defects, and interactions with the local environment. Depending on the particle size, shape, and composition, three different types of vibrational modes are typically observed (Fig. 1B): 1) an axial or contact mode (11) at low frequency (<10 GHz), which represents interactions of the particle with the local environment, 2) nonspherically symmetric angular modes (10 to 20 GHz), which correspond to higher-order vibrational motion represented by spherical harmonics, and 3) a breathing or radial mode (>20 GHz), which represents radially symmetric vibrational motion (12). Angular modes may also be induced by breaking of the particle spherical symmetry near the substrate or from coupling of modes due to the local environment as will be discussed later. For comparison, we show the spectra (Fig. 1C) of a ~100 nm spherical gold nanoparticle (AuNP) and an 80 to 100 nm lentiviral pseudovirus particle with a GFP gene inserted into its RNA genome (LentiGFP). For these particles, both spectra exhibit fast oscillations that persist for ~500 ps and slower oscillations that persist for at least 5 ns. These oscillations correspond primarily to the breathing and axial modes, respectively, with the virus particle spectra exhibiting a more complex structure at frequencies below 10 GHz.

To better understand the complex acoustic spectrum of the virus, we first investigated a similarly sized spherical AuNP under identical experimental conditions. We chose AuNPs due to their well-characterized, structurally defined acoustic modes, particularly in the GHz range, which enabled us to effectively benchmark and analyze the virus’s acoustic response. While AuNPs differ from viruses in the excitation mechanism, both systems exhibit environmental sensitivity in their acoustic modes, allowing us to draw meaningful comparisons on spectral shifts and symmetry effects. The properties of the prominent radial breathing mode have been extensively studied using time-resolved ultrafast spectroscopy in a wide range of metallic nanoparticles including gold, silver, and bimetallics, and in varying sizes and shapes (12–14). When metallic NPs are illuminated with an ultrashort pulse, electrons within the Fermi energy are excited to higher lying states in the conduction band forming hot carriers which rapidly thermalize in tens of femtoseconds. This is followed by thermalization with the lattice through electron–phonon interactions, which generates a photo-induced stress in the NP, launching coherent mechanical vibrations that are detected by a time-delayed probe pulse. Above a few nanometers, the acoustic properties of these systems are well-described by an elastic continuum model (15), which predicts that the radial breathing frequency scales inversely with the characteristic dimension of the nanostructure and with the square root of the ratio of the shearing modulus, $G$, to the particle density, $\rho$: $v_{\mathit{br}}\propto {\mathrm{D}}^{-1}{\left(G/\rho \right)}^{1/2}$, where D is the particle diameter. For a free particle with radius, $R$, using stress-free boundary conditions [free sphere model (FSM); see SI Appendix, section XIII], the energy eigenvalues may be obtained for Lamb’s equation of motion for a three-dimensional elastic body (16, 17). These eigenvalues depend on the orbital angular momentum quantum number $l$, and harmonic $n$ (Fig. 1B). Since the excitation process occurs through Raman scattering, selection rules dictate that for a free spherical particle only the spherical modes are allowed: $l=0$ corresponds to a purely radial mode with spherical symmetry, while $l=2$ is a quadrupolar mode. When the NP is in contact with a surface, in addition to the radial breathing mode, a low-frequency axial mode is observed which depends on the adhesion force between the NP and the local environment. This interaction gives rise to a periodic motion of the particle position relative to the environment (e.g., substrate) which manifests as a few GHz mode for ~100 nm AuNPs. Studies of the breathing and axial modes for different diameter spherical AuNPs have shown that they are related to one another according to classical Hertzian contact mechanics, where by $v_{\mathit{ax}}\propto v_{\mathit{br}}^{7/6}$, where $v_{\mathit{br}}$ is the radial breathing mode frequency and $v_{\mathit{ax}}$ is the axial mode frequency (11, 18). Therefore, the scaling between the two modes is independent of the particle size.

Fig. 2 A, i shows the acoustic response from a single AuNP that is electrostatically bound to the substrate by a 0.5 to 1-nm-long (3-Aminopropyl)triethoxysilane (APTES) molecule absorbed on the -OH terminated fused silica substrate (19). The molecular tether ensured weak coupling between the AuNP and the substrate and established a more stable and “matrix-free,” environment under laser illumination. The sample preparation procedure (SI Appendix, section IV) ensures minimal residual contaminants and organic impurities. The measured radial breathing mode at $v=28.92 \mathrm{GHz}$ agrees with the FSM model for a 105 nm diameter spherical AuNP. At low frequency, the spectrum shows two axial modes near 3.65 GHz and 4.70 GHz: the splitting may be explained by considering the tether acting like a mass on a spring in series between the NP and the substrate. We also observed an angular mode near 18 GHz region which is in close agreement to $l=2,n=2$ angular mode expected by FSM.

Fig. 2.

Tracking of acoustic spectra for tethered (A) from untethered (B) gold nanoparticles (AuNPs). (i) Top: Average time-domain response (black curve) and SD (filled red curve) over the trajectory. The Inset shows the AuNP interaction with the substrate. Bottom: Acoustic spectra recorded in 1-min increments as the laser irradiation modifies the matrix [blue region in (B)] environment. Note, in the tethered case, the environment is considered matrix-free. (ii) Frequency peak shift relative to first time trace for peaks color-coded in (i). (iii) Relative peak broadening as a function of lab time. Note, the spectra are normalized to the highest peak in the spectral window. $v$: frequency, $\mathrm{\Gamma }$: dephasing rate. Gray regions show modes that are due to NP/matrix interactions.

Next, we considered an “untethered” AuNP where the adhesion forces between the NP and substrate are much stronger due to direct contact (Fig. 2B). Unlike the tethered case, the drop cast procedure resulted in the presence of residual contaminants on the surface which we loosely refer to as the matrix, giving rise to variance in the acoustic spectra of individual AuNPs due to heterogeneity in the local environment. For the particles shown in Fig. 2, the average frequency of the breathing mode is similar (<0.35 GHz across the time series), implying that the characteristic diameters are within 1 nm of one another. For the untethered AuNP, the amplitudes of the axial and angular modes were much stronger relative to the breathing mode than those for the tethered particle. We also observed additional acoustic modes for the untethered particle due to breaking of the spherical symmetry near the substrate. The large number of modes observed (within 6 to 25 GHz, Fig. 2B) cannot be accounted for by considering only the FSM, which suggests strong coupling between the axial, angular, and breathing modes induced by the local environment. In a spherical particle, there is an $2l+1$ degeneracy from the z-component of the angular momentum. However, deviations from spherical symmetry give rise to a splitting into $l+1$ modes (17, 20).

To gain further insight into the origin of the substrate-induced matrix coupling, we performed dynamic single-particle spectroscopy (d-SPS) experiments. The lasers, in addition to executing the pump–probe experiments, served to alter the matrix of the NP and the resulting spectral evolution revealed correlations and couplings among the observed acoustic modes. For the tethered AuNP, tracking the particle over 10 min revealed only a minor shift (<200 MHz) of the axial, angular, and breathing modes in the 2 to 30 GHz range, as well as no discernable correlation among the mode frequencies (Fig. 2 A, ii and iii) due to the more stable environment. In the untethered AuNP, however, evolution of the spectral features over several minutes revealed correlations among the mode frequencies as shown in Fig. 2 B, ii/iii. For the five modes analyzed, selected due to their relative isolation from nearby peaks, the acoustic frequencies all blue-shifted (shift to higher energy) with experimental time. This blueshift occurred to varying degrees for many of the single particle measurements performed (see SI Appendix, section VI, for other single particle trajectories). The laser light slowly ablated or otherwise removed the matrix surrounding the NP, which both lowered the effective mass and caused stronger association between the NP and the substrate. Such correlations are ideally measured by d-SPS because variations in size, shape, and other material properties such as defects, multiple crystal facets and dislocations, and ligand coverage obscure spectral changes originating from changes in the NP–environment interactions (21). As shown in Fig. 2 B, ii, the blueshift of the axial mode (labeled in cyan) is most pronounced, while the radial breathing mode shift (labeled in blue) is smallest both in absolute ($\mathrm{\Delta }v$) and relative terms ($\mathrm{\Delta }v/v$). We also observed that the breathing, axial, and one of the angular modes (labeled in red) narrowed over the 10-min trajectory (Fig. 2 B, iii). Additionally, we observed that the amplitudes of the axial mode, relative to the breathing and angular modes, increase over the time series (note, the plots are shown on a normalized scale for clarity). The axial mode amplitude depends on the distance between the NP and substrate as well as the displacement of the breathing mode. d-SPS provides a means to distinguish modes that arise from NP–matrix interaction from those due to NP–substrate interactions. As the matrix is removed and the particle-substrate distance decreases, the breathing mode induces a larger axial amplitude vibration. For the main axial mode and the angular mode in red (12.5 to 12.9 GHz), the dephasing rate decreased (i.e., exhibited a longer lifetime) with increased NP/substrate interaction, which is commensurate with an increase in the breathing mode lifetime (blue marker). Therefore, damping of these surface-sensitive modes are strongly affected by the matrix environment. In contrast, for the two modes shown in green (8.3 to 8.8 GHz) and brown (18 to 18.5 GHz), we observed that the frequency shifts were nearly identical with time. The dephasing of these two modes also followed the same trend, whereby the dephasing rates remained largely unchanged. This suggest that these modes are associated with internal modes (e.g., the $l=2,n=2$ mode shown in Fig. 1B), where the surrounding matrix has minimal effect on damping. In some of the broader modes (gray shaded regions), the amplitudes rapidly decrease with experiment time, indicating that these modes arise primarily from NP/matrix interactions (22).

With a deeper understanding of how the different acoustic modes in the AuNP influence one another and their dependence on the particle/environment interaction, we turned our attention to the virus particles. In contrast to the metal NPs, the virus vibrations are excited under nonresonant conditions (electronic transitions of biomolecules composing a virus range from about 4.4 to 4.8 eV; compared to 1.2 eV pump energy). Therefore, the excitation proceeds by stimulated Raman scattering (23), whereby Raman-active normal modes are excited on the electronic ground state. The signal strength is far weaker than a comparatively sized metal NP (see SI Appendix, section III for comparison), and an isolated virus particle does not experience an appreciable temperature change from direct photon absorption. As will be discussed later, in some cases, indirect heating of the virus particles via their surrounding matrix over long exposure times may cause viral damage or rupture (Fig. 3A).

Fig. 3.

Single-particle trajectories provide insight into the virus–substrate interactions. (A) Virus #1 trajectory over 12 min showing gradual phonon softening and sudden virus particle rupturing. (B) Virus #2 trajectory showing both axial and breathing modes as well as weaker angular modes. (C) Cartoon of the state of the virus particle at select time points. (D) Top: zoomed in spectrum of Virus #2 at T = 3 min in the 0 to 10 GHz region showing the axial modes formed from interaction of the glycoproteins with the surrounding matrix. Bottom: zoomed in spectrum of nanourchin particle showing axial mode splitting. (E, i) Estimated stiffness of the Virus #1 with lab time (horizontal red dashed line is average over first 9 min; slanted red dashed line is a fit over the last 3 min). (ii) Correlations between the breathing mode frequency and dephasing rates (closed triangles and circles are for Virus#2; open markers are for Virus #1. (iii) Frequency shift of the breathing and axial modes referenced to the T = 1 min. time trace. (iv) Change in the relative dephasing rate for the breathing mode peak (red) and axial mode peak (orange) for Virus #2 as a function of time.

Many single particle trajectories were examined (see SI Appendix, section VIII for other trajectories), but here, we focus on two that clearly illustrate the environmental effect on the virus acoustic spectra. Fig. 3A shows the dynamic tracking of a virus particle (Virus #1) that is well isolated and positioned atop a bare fused silica substrate over a span of 12 min [see SI Appendix, section VII for correlated atomic force microscopy (AFM) image]. The notable absence of axial modes suggests that the virus particle was weakly or noninteracting with its environment. The spectra reveal a single breathing mode near 21.8 GHz, with a broad peak at 1 to 2 GHz lower energy. As with the NPs, basic features of the acoustic spectra may be described by FSM. For the lowest-order radial breathing mode ($l=0,n=1$) observed at ~21.8 GHz, the estimated particle diameter is 83 nm using assumptions of the longitudinal sound velocity in the virus (24), which agrees with the estimated diameter of 80 to 100 nm for the LentiGFP virus. The next lowest allowed mode at $l=2,n=2$ for a spherical, model virus occurs at 20.5 GHz, which is in close agreement with the observed mode at $T\ge 1 \min$ near 20.8 GHz (Fig. 3A). We also note that slight deviations of the virus shape from nominally spherical may also induce asymmetric line shapes and additional features near the main breathing mode (15, 25). The spectra remain largely unchanged until about 6 min when the breathing mode begins to weaken and broaden. The mode exhibits a dramatic red shift and acute broadening at 10 min before the signal disappears below the noise floor. This red shift and broadening indicate a softening of the phonon modes which occurs as the viral capsid begins to weaken, swell, and suddenly ruptures (26). Assuming the radial breathing mode may be modeled by an underdamped harmonic oscillator (SI Appendix, section XII), the virus initially has a stiffness of $k=\mathrm{6,490}\pm 62 \mathrm{N}/\mathrm{m}$, which then suddenly drops at a rate of $-608 \mathrm{N}/\mathrm{m}/\min$ before rupturing (last three time points in Fig. 3 E, i). While most virus particles measured stayed intact during the measurement and despite employing nonresonant excitation conditions, the virus particles occasionally ruptured after several minutes of laser exposure, possibly through multiphoton ionization or localized and excess thermal accumulation. Further insight may be gained by examining the evolution of the breathing mode (Fig. 3 E, ii), where we observe that the frequency and dephasing rate are anticorrelated with each other (open blue circles and open red triangles). This indicated that weakening of the capsid caused rapid dephasing as the symmetry was reduced and the surface area increased.

We then measured a second virus particle (Fig. 3B), with the breathing mode and shoulder very similar to Virus #1 initially, but with a much stronger axial mode that changes over time. For this sample, labeled as Virus #2, correlative AFM imaging shows that the particle is buried in a matrix environment (see AFM images in SI Appendix, section VII). The localized environment of each virus particle measured may be different due to material remaining after purification—buffer, impurities, and crystallization of paraformaldehyde which is used for virus deactivation. After 1 min, the breathing mode in Virus #2 redshifted by 1.5 GHz, broadening significantly. This large redshift of the breathing mode is due to changes in the virus particle/matrix interaction upon repeated laser irradiation. While challenging to characterize, the redshift suggested an increase in the effective mass of the virus, which could be due to some matrix material adhering to it temporally. At 2 min, a prominent axial peak became evident (orange circle near 2 GHz region) which started to blueshift as the interaction between the virus particle and matrix decreased. Meanwhile, both the breathing and axial modes experienced a blueshift along with line shape narrowing, corresponding to increasing dephasing time. Angular modes appeared in the 5 to 10 GHz region but were weak and difficult to quantify, so they were not evaluated further. The signal starts to weaken at about 8 min and never fully recovers to their original state indicating that the virus particle has undergone some damage during the tracking experiment. Other trajectories showed that the virus remained intact for long periods or exhibited partially reversible dynamics in the limit of weak virus/environment interactions (SI Appendix, sections VIII and IX).

The axial modes show a far more complex pattern of peaks than in the case of spherical AuNPs (Fig. 3 D, Top). The high resolution of the acoustic spectrum in the sub-10 GHz spectral region arises from the long lifetime of the axial modes. The features may be qualitatively rationalized by considering the structure of the envelope proteins. The virus particle has the vesicular stomatitis virus G protein (VSV-G, 58.4 kDa without glycosylation) composing the envelope. These glycoproteins may be considered as a series of coupled oscillators whose interactions with the local environment are reflected in mode splitting. In the case of the tethered AuNP, we observed that the axial modes arising from splitting due to the tether acting as an additional mass on a spring. While examining the detailed interactions of the viral envelope proteins with the matrix is beyond the scope of the current study, we considered a biomimetic virus-like NP (27)—a single Au “nanourchin” particle (28) which has spiky protrusions coming out of its surface (see SI Appendix, section XI for additional details). As shown in Fig. 3 D, Bottom, the asymmetric shape of the nanourchin leads to a broad breathing mode, while exhibiting axial mode splitting analogous to those observed in Virus #2.

To gain more insights into the coupling of these modes, we looked at correlations among the spectral features. As with the trajectory of virus #1, the spectral shift of the breathing mode in virus #2 was anticorrelated to the dephasing rate. We then identified which axial modes most strongly coupled to the breathing modes. For the peak marked in orange (2 to 3 GHz), both the axial and breathing mode increased their lifetime with lab time at nearly the same rate, while both experienced correlated spectral shifts at different relative magnitudes (Fig. 3 E, iii/iv). This indicates that the breathing and axial modes are affected by the matrix environment in a correlated manner. The axial modes here are likely less influenced directly by the fused silica substrate as in the case of the AuNP because of the weaker van der Waals interactions between the virus envelope proteins and the untreated glass. To test this idea, we performed experiments with the same conditions of single virus particles on a mica substrate (grade V-1) where the interaction with envelope proteins is expected to be stronger. We observed an ~3 GHz redshift (SI Appendix, section X) of the breathing mode peak compared to fused silica, which is due to the increase in the effective mass of the particle from the stronger van der Waals interactions.

We then performed a statistical analysis on 12 individual single virus particle trajectories measured in this study (see correlative AFM in SI Appendix, section VII). The AFM imaging showed a large variance in the environment of each virus particle, which is reflected in a commensurate variance in the measured acoustic spectra. This is attributed to strong environmental effects due to coupling of the axial, angular, and breathing modes as described in detail above. To analyze the large volume of data acquired, we performed a Bayesian analysis method (29) on each trajectory that extracted the frequency, dephasing/decay rate, phase, and amplitude of more than 200 individual acoustic spectra (see SI Appendix, section XIV for details). Fig. 4 shows a 2D histogram and its projection of all the frequencies and decay rates extracted from the statistical analysis. Colored regions in the 2D plot indicate the number of counts within the specified 2D bin region. The most common mode observed is the radial breathing mode near 22 GHz with a dephasing rate of ~1 GHz (1 ns lifetime). We also see the redshifted mode appear in most of the traces although its position relative to the zero-order breathing mode varies by ~1 GHz. We observe relatively few instances of modes in the 5 to 15 GHz range, where we would expect other spheroidal modes (besides $l=0,n=1$) presumably because these modes are weak and more difficult to identify in the analysis. In the axial mode region, we observe a wide range of frequencies in the 2 to 5 GHz region with a wide range of dephasing times. We note that the extreme sensitivity of the axial mode to the virus/environment interactions causes these modes to shift appreciably, thereby giving rise to a far wider distribution of frequencies and dephasing rates compared to the relatively stable breathing mode. Ensemble measurements average out over these strong local effects, leading to broad and featureless spectra. This underscores the need to make single particle measurements with high time resolution.

Fig. 4.

Statistical analysis highlights the matrix effect. Summary depiction of all frequencies and decay rates extracted from the statistical analysis. Top: Two-dimensional histogram resulting from analysis of all frequencies recovered by Bayesian analysis. The dephasing rate, $\mathrm{\Gamma }$, and frequency, v, are binned by 0.5 GHz × 0.25 GHz. The color bar shows the number of counts per bin. Bottom: The projection of the 2D histogram onto the frequency dimension shows the main breathing mode near 22 GHz and lower-frequency axial and angular modes. Red vertical lines show predicted acoustic modes based on Lamb’s model for a free spherical particle.

In conclusion, we have measured the acoustic phonon spectra of individual virus particles by ultrafast spectroscopy. In contrast to optical spectroscopic measurements of local bond vibrations (30), the acoustic spectra are a measure of the collective oscillations of all the atoms that compose the virus particle. Our method captures structural features that are sensitive to virus morphology, surface proteins, and environmental conditions, providing insights into viral mechanics and interactions that are not possible with other single-particle methods. The nanosecond lifetimes of these collective vibrations impart to them a remarkable sensitivity to viral stiffness, shape, and glycoprotein properties, which are critical for understanding viral dynamics. Furthermore, tracking the evolution of the acoustic spectra provides unique insight into the effect of the surrounding environment on the vibrational motion (31). The method described has sufficient time resolution to examine a single virus particle through its life cycle (32) which occurs on the second to hour time scale. Future studies will focus on assigning specific features in the virus particle acoustic spectrum through examination of structure and matrix effects by correlative electron microscopy and other single particle characterization methods (33). Investigations on the effect of damping of acoustic vibrations in a liquid environment (34–36) will be important for assessing the viability of this approach for in vivo applications. In these cases, BioSonics may serve as a sensitive, high-resolution local sensor of microscopic environments. Future work will focus on understanding the specific role of the system–matrix interactions and the local temperature on the acoustic vibrations. While the studies here were focused on viruses, the broad spectral range of BioSonic spectroscopy could capture other microorganisms including bacteria and fungi, smaller nanoscale molecular machines such as molecular motors, and large proteins. The ability to detect single, unlabeled virus particles without physical contact enables a long list of applications from ultrasensitive viral detection to fundamental studies of viral dynamics including self-assembly and infection, paving the way for a comprehensive understanding of many biological processes by correlating both static structures and dynamics.

Materials and Methods

Experimental Setup and Signal Processing.

The ASOPS (ASOPS, Menlo Systems) laser system generates two pulse trains with a fixed repetition rate offset $\mathrm{\Delta }f$ (typically 1 to 10 kHz). The pump (1,040 nm, 100 fs, 1 W) pulse is focused to the sample by a reflective objective (74×/0.65, Beck Optronic Solutions) from the bottom of the homebuilt microscope, and the probe is focused to the sample by a transmissive objective (100×/0.9, MPlanFL N, Olympus) from the top. The probe pulse is split prior to the sample and 10% of it serves as a reference for balanced detection. The probe light scattered from the sample is collected in an epi configuration and directed toward one channel of the balanced detector, and the reference probe beam is collected by the other channel. The signal is then amplified and sent to a high-speed digitizer (GaGe RazorPlus) whose clock is set by the repetition rate electronics (RRE) and triggered by the offset either generated from the RRE electronically or through an optical trigger generation system.

Raw data are averaged using the on-board averaging function of the digitizer and then organized into a set number of record segments (J) upon receiving a rising-edge trigger event with a rate of a few kHz. Further data averaging is done offline in MATLAB. Signal in the lab time is then converted to the molecular time using the down-conversion factor $\mathrm{\Delta }f/f_r$, where $f_r$ = 100 MHz is the laser repetition rate. Using the Bayes inference approach (29), Raman signals arising from acoustic vibrations are extracted after removing the electric filter effect in the sub-GHz frequency by fitting the signal to a sum of exponentially decaying sinusoidal model functions, $g_i\left(t\right)=cos\left({\omega }_{i}t+{\varphi }_i\right)exp\left(-{\mathrm{\Gamma }}_{i}t\right)$ in an iterative manner, where ${\omega }_i$, ${\varphi }_i$, and ${\mathrm{\Gamma }}_i$ denote angular frequency, phase, and linewidth of signal $i$, respectively.

Sample Preparation.

Lentiviral pseudovirions were produced in 293FT cells by transfecting with packaging constructs pCMV-VSV-G and pCMV-Delta 8.2 (gifts from Jerome Schaack). Cell culture supernatant containing virions was collected 72 h posttransfection and filtered using a 0.45-µm filter to remove cell debris. The virions in the supernatant were concentrated by ultracentrifugation at 20,000 rpm for 2 h. The pelleted virions were inactivated and fixed using 4% paraformaldehyde.

For the tethered AuNPs, the precleaned glass slides were incubated in 5% (V/V) (3-aminopropyl) triethoxysilane (APTES) in ethanol for 3 h and then dried under N₂ stream. The slides were then subjected to thermal annealing in a vacuum oven at 110 °C for 2 h to obtain APTES-silanized slides with amine groups, which were subsequently immersed into the AuNPs solution (Sigma-Aldrich, #742031, 100 nm, stabilized suspension in citrate buffer, <0.2 polydispersity index) for 6 h adsorption. After the Au NP adsorption, the glass slides were washed and dried under N₂ stream. For untethered AuNPs, the sample was simply drop cast onto the glass slide.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.