Bacterial flotation devices enhance ultrasound imaging

Cyanobacteria thrive near the surface of their aquatic environment, where a rich mixture of carbon dioxide, water, and sunlight produces the optimal conditions for photosynthesis. These microorganisms regulate their buoyancy by producing rigid, cylindrical gas vesicles (GVs), granting them mobility in the water column (1). Like any pressure vessel, there is a threshold pressure difference between the internal and surrounding media, beyond which the GVs will break. Naturally, the proteinaceous structures found in cyanobacteria and several archaea have evolved to endure the hydrostatic pressure of their environment (1). The exceptional stability of GVs is partly attributed to beta-sheet structures formed by the self-assembly of GV subunits. This results in a highly ordered protein shell that exhibits Young’s moduli comparable to glassy polymers (a few GPas). When the threshold or critical pressure is surpassed and the vesicles rupture, they disappear under a light microscope. This dramatic change in optical properties due to rupture was the first example of GV use as a signaling agent, operating as a pressure gauge in this case. Further characterization of GV geometry and composition as well as the invention of high-resolution optical and acoustic imaging tools like electron microscopy, NMR, and ultrasound imaging has expanded the possible use cases for GVs as sensors and reporters. Although assembly of the gas-permeable membrane is a complex process requiring a cluster of 14 genes, the vesicle membrane itself is constructed from just two proteins. The alpha-helical hydrophilic protein GvpC forms rod-like reinforcements over the smaller, beta-sheet-forming hydrophobic protein GvpA, generating a membrane that promotes gas transport, inhibits water permeation, and is mechanically robust despite being roughly a couple of nanometers thick. Accurate measurement of GV dimensions is critical for predicting rupture pressure and signal calibration. The size distributions of GVs from some of the most commonly studied species have been approximated using multiple techniques, but most recently, Dutka et al. demonstrated that cryogenic electron microscopy (cryo-EM) preserves morphology better for dimensional characterization than other forms of electron microscopy (2). The exceptional structural details of GVs revealed by cryo-EM paves the way for further understanding of the role of geometry on the physics of these systems, for instance by molecular dynamics simulations and other structural analysis methods.

While GVs can be functionally expressed in E. coli by encoding the 14 necessary genes in the plasmid, GVs are also easily purified directly from their natural host by cell lysis and centrifugal-assisted flotation (3). In this issue of Biophysical Journal, Salahshoor et al. take this purification process one step further by fractionating GVs into different size distributions, which rupture under different critical pressures, unlocking the possibility for multiplexed signaling (4).

The most immediate application of GVs arises from their natural function. Expression of GVs beyond a certain point will cause the cell that is producing them to float, which is a straightforward method of cell sorting. Since GV production is genetically encoded, effectiveness of gene-suppression strategies, for example, can be easily quantified by counting the cells that float and those that sink. However, applications extend well beyond their role as cellular buoys. When ultrasound waves are transmitted through buffer solutions, there is little to no acoustic scattering of the reflected signal. On the other hand, since gas bubbles are compressible, they respond to incident pressure waves by oscillating about an equilibrium radius, causing substantial backscatter and a robust signal. In the context of ultrasound imaging, there are two favorable behaviors of gas compartments. First, the magnitude of backscattering increases nonlinearly with transmitted frequency. Second, gas compartments, whether they are vesicles or microbubbles, collapse when the peak amplitude of incident pressure waves exceeds the strength of the shell, eliminating nearly all scattering effects. Both of these mechanisms have inspired strategies to enhance contrast in ultrasound imaging (3, 5). Protein- and lipid-stabilized microbubbles are the preferred contrast agent for ultrasound imaging because they have high echogenicity and can be stabilized and administered directly. Unfortunately, they degrade quickly, which is a severely limiting factor. For deep tissue imaging, where a longer half-life of the contrast agent is required, nanoparticles are often used instead. At high intensities of ultrasound irradiation, cavitation can damage surrounding tissue, but the nanoparticle surface provides a nucleation site that causes cavitation at lower intensities. Oscillations of the cavitating nanobubbles can then be detected by scattering the signal that created them (6). By inserting the gene cluster necessary for GV production into mammalian cells and using ultrasound imaging to detect contrast from those GVs, Farhadi et al. propounds GVs as a stable alternative contrast agent (7). Scattering signals are also compounded through GV aggregation. Solutions of intact cyanobacteria where GVs are concentrated within the membrane have a higher ratio of signal to background scattering than the same solution after cell lysis disperses the GVs (3). This and other experiments that induced GV aggregation suggest that for equal GV concentration, there is greater scattering contrast in the solutions with aggregated GVs over solutions with evenly dispersed GVs.

Membrane mechanical properties, protein composition, and GV dimensions are the predominant features influencing the buckling pressure and nonlinear scattering signal. The first two features are closely related. Although the sequences of GvpA and GvpC are largely conserved across species, stability of the amyloid-like cross-beta sheet structure of the assembled GVs is highly sensitive to point mutations and changes in sequence length. Sequence differences are therefore responsible for a broad range of average critical pressures from 0.1 to 1.0 MPa in cyanobacteria and a halobacterium (1), which reflects the tunability of GVs as signaling agents.

Molecular and multiscale computational methods are indispensable for predicting structure from sequence (8), for further resolving predictions with methods like cryo-EM, and to help understand the mechanical behavior of microbially fabricated protein materials (9). Currently, there is need for synergy between modeling and experiments to understand buckling mechanisms, sequence and buffer dependence of properties, and mechanical response at molecular and vesicle scales. In ”Geometric effects in gas vesicle buckling under ultrasound,” Salahshoor et al. employ finite element analysis (FEM) to simulate the buckling behavior of GVs with physiologically relevant membrane mechanical properties, dimensions, and conditions. Consistent with previous experimental observations and the theory of cylindrical shells, the authors confirmed an inverse cubic relationship between threshold buckling pressure of GVs and their diameter and observed an insignificant change in buckling pressure with length. To validate the FEM model, the authors fractionated GVs into different size distributions and showed that buckling pressure can be accurately calculated using their model.

Since the FEM model depends on prior knowledge of mechanical properties of the GV membrane, it will be most effective when combined with a method for characterizing GVs with different protein sequences across species. The authors also note the possibility of multiplexed detection, which is supported by their success in sorting GVs by size. In the future, multiplexed detection in mammalian cells could be tested by encoding gene clusters from two different microorganisms whose GVs rupture at distinct ultrasound frequencies and isolating the signal between them. The methods put forth in this paper for fractionating GVs and predicting their buckling pressures with FEM are promising steps toward advancing engineering applications of GVs as ultrasound contrast agents and reporter genes. In the field of nanomedicine, diagnostics must be utilized safely and effectively in environments that are extremely difficult to observe. Simultaneous employment of computational and experimental methods enables multiscale comprehension of phenomena and improves confidence in emerging technologies.

Editor: Guy Genin.