a) Biomolecular Engineering of Gas Vesicles

Gas vesicles (GVs) are a class of hollow gas-filled protein nanostructures evolved in photosynthetic microbes. They are hundreds of nanometers in size and possess a gaseous interior enclosed by a ~ 2 nm-thick protein shell that is permeable to air but excludes water (Fig. a-b). The native role of GVs is to regulate the buoyancy of their microbial hosts in bodies of water so that when the microbes need to increase their photosynthesis rate, they will express GVs to float to the surface of the water for an increasing amount of sunlight. GVs are encoded by a cluster of 8-14 genes, and transferring these genes from their native hosts to other organisms, such as E. coli Nissle 1917 (the common probiotic Mutaflor®), results in the expression of GVs in these cells (Fig. c).


What makes GV an interesting target for bioengineers are their unique biochemical and physical properties. The change of acoustic impedance at the air-water interface allows their detection by ultrasound imaging; the differential magnetic susceptibility between air and water enables magnetic resonance imaging (MRI); and lastly, the fast diffusion of xenon molecules across GV shells permits HyperCEST 129Xe MRI. Although GVs have been studied for decades by biologists, these new engineering efforts in the past few years have rendered them a new utility as genetically encoded agents for biomedical imaging, diagnosis and therapy. For example, a project that I led during my postdoc developed acoustically modulated magnetic resonance imaging (AM-MRI), a new deep-tissue imaging modality that explored GVs’ dual responsiveness to ultrasound and magnetic field to enable background-free MR imaging (Fig. d). Another project explored the acoustic properties of GVs to control gene expression in deep tissue, a project that recently received NIH Pathway to Independence Award (K99). In this nascent field, we envision that these new methods would find wide utilities in biomaterials, biomedical imaging and therapeutic development.

Read more: Nature Materials (2018) 17:456–463,Current Opinion in Chemical Biology (2018) 45:57-63, Annu. Rev. Chem. Biomol. Eng. (2018) 9:229-252, Nature (2018) 553:86-90.

b) Structural Biology of Membrane Proteins

Among structural biology methods, NMR is unique in its ability to study proteins in physiological pH, hydration and temperature. This has been crucial for cases such as intrinsically disordered proteins and amyloid fibrils, of which the cellular context is important for the correct folding, dynamics and function. Moreover, the recent studies of structure and function of proteins inside living cells by NMR offer an avenue towards an atomic-resolution description of cellular functions.

The overarching goal of my PhD research was to develop a method that could determine atomic-resolution structures of membrane proteins in their native lipid bilayer environment. To achieve the goal, I explored novel methods in solid-state nuclear magnetic resonance (ssNMR). I pushed the technological frontier of the method by solving the structure of a challenging protein at the time, the mercury transporter MerF. In addition to being a proof-of-concept for the methodology, the structure of MerF revealed large conformational rearrangement in response to terminal peptide truncation, which served as an example of the importance of studying full-length, unmodified membrane proteins in lipid bilayer.

Read more: J. Am. Chem. Soc. (2013) 135:9299-9302, Quarterly Reviews of Biophysics (2014) 47:249-283.

c) Method Development in Solid-state NMR Spectroscopy

NMR method can be subdivided into solution NMR (solNMR) and solid-state NMR (ssNMR). SolNMR requires the fast isotropic tumbling of proteins in a solution to achieve high spectral resolution, and therefore, large-size proteins and membrane proteins embedded in lipid bilayers are traditionally difficult to be studied by solNMR. ssNMR, on the other hand, abolishes the requirement of isotropic tumbling and instead utilizes radiofrequency pulse sequence and mechanical manipulation of samples (such as magic angle spinning) to achieve high resolution.

During PhD, I pioneered dipolar coupling correlated isotropic chemical shift (DCCICS) analysis for resonance assignment and developed MSHOT-Pi4/Pi pulse sequence that substantially improved 1H resolution. Driven by the finding that resolution enhancement of MSHOT-Pi4/Pi pulse sequence occurred selectively in membrane protein samples but not in microcrystalline samples, I analyzed the quantum mechanics theory underlying the pulse sequence and unveiled that MSHOT-Pi4/Pi sequence was resistant to the interference effects from the motion of membrane proteins. This “motion-adapted” feature was further generalized and, as another example, I showed how membrane protein motion could enable spin diffusion through the forbidden proton-relay mechanism.

Read more: J. Biomol. NMR (2014) 58:69-81, J. Magn. Reson. (2011) 209:195-206, J. Chem. Phys. (2014) 140:124201, J. Chem. Phys. (2013) 139:084203.