The Hong group develops and applies magic-angle-spinning solid-state NMR spectroscopy to elucidate the structure and dynamics of biological macromolecules. We seek to understand how molecular conformation, motion, and intermolecular interactions enable proteins and carbohydrates to carry out their functions such as ion conduction across lipid membranes, membrane fusion between viruses and cells, membrane scission during virus budding, and maintenance and expansion of plant cell walls. We also investigate the structure and dynamics of amyloid proteins in neurodegenerative diseases. To answer these mechanistic biophysical questions, we innovate multinuclear (1​​H, 1​3​C, 1​5N, 1​9​F, 2​​H, & 3​1​P) and multidimensional solid-state NMR techniques that measure molecular structure, inter-atomic distances, and amplitudes and rates of molecular motions with high sensitivity.


Ion channels and transporters are integral membrane proteins that transport ions and polar compounds across the hydrophobic interior of lipid bilayers. In this large family of membrane proteins, virus ion channels, also called viroporins, are important for the lifecycles and pathogenicity of viruses, while bacterial multidrug-resistance transporters pump out toxic compounds to cause antibiotic resistance. Elucidating the molecular conformation and dynamics of these membrane proteins is critical for advancing our fundamental knowledge about the mechanism of membrane transport, and for aiding the design of antiviral and antibiotic compounds to combat infectious diseases.

Figure 1. Schematic models of the influenza M2 four-helix bundle, the SARS-CoV-2 E channel, and the dimeric E. Coli transporter EmrE, which are investigated using solid-state NMR.

We are currently studying three ion channels and transporters involved in infectious diseases: the influenza M2 proton channel, the Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2) envelope (E) protein cation channel, and the bacterial transporter EmrE (Fig. 1). All three proteins are oligomeric α-helical bundles that require proton transport for function. The acid-activated and amantadine-targeted M2 proton channel is essential for the influenza virus lifecycle, and is the prototype of viroporins 1. The SARS-CoV-2 E protein forms a cation channel 2 whose calcium-conducting activity stimulates the host inflammasome. EmrE is a proton-coupled transporter that exports polyaromatic cations to cause multidrug resistance 3-4.

Solid-state NMR spectroscopy is well suited to structure elucidation and dynamics investigation of these viral and bacterial channels and transporters in lipid bilayers. We conduct multidimensional 1H, 13C and 15N correlation experiments to extract conformation-dependent chemical shifts. We apply 19F-based solid-state NMR experiments to measure inter-atomic distances to ~2 nm 5 to determine the oligomeric structure of these membrane proteins. We employ spin-polarization transfer and intermolecular correlation experiments to investigate which amino acid residues line the channel pore and which residues face lipids. In addition to structure, we examine the dynamics of these channels and transporters to gain insight into how protein sidechain and backbone motion mediates transport. We also measure the dynamics of water molecules inside the channel pore to understand how water facilitates ion conduction 6. With suitably isotopically labeled substrates, we probe substrate dynamics and correlate them with protein dynamics during transport. These studies of protein, substrate and water dynamics exploit the sensitivity of NMR lineshapes and nuclear spin relaxation times to motions from picoseconds to milliseconds, and yield detailed insight into the mechanism of membrane transport.


Membrane curvature is essential to many biological processes such as endocytosis, vesicle trafficking, and cell division. Proteins can sense, stabilize, and induce membrane curvature. An important class of membrane proteins that induce membrane curvature is viral fusion proteins, which merge the virus envelope and the target membrane to enable virus entry into cells. They accomplish this task by undergoing complex conformational rearrangements, as seen in X-ray crystal structures of water-soluble ectodomains of these proteins (Fig. 2A). These protein conformational changes presumably lower the free energy barriers for membrane dehydration and membrane structural changes from the lamellar state to hemifused intermediates to the final fused state. However, this conceptual framework excludes two key hydrophobic domains: the N-terminal fusion peptide (FP) domain and the C-terminal transmembrane domain (TMD), which play important roles in destabilizing the lamellar structures of the two lipid membranes.


Using solid-state NMR spectroscopy and complementary techniques such as small-angle X-ray scattering (SAXS), we investigate the conformations and oligomeric structures of the hydrophobic FP and TMD of viral fusion proteins in biologically relevant lipid membranes. Our studies of the parainfluenza virus fusion protein indicate that both the FP and the TMD have membrane-dependent structures (Fig. 2B), and the β-sheet conformation is responsible for generating saddle-splay curvature to the membrane. The latter is manifested in 31P NMR spectra (Fig. 2C) and SAXS data. We also measure water-lipid and water-protein interactions to obtain information about membrane dehydration during fusion. By coupling these protein structure measurements with membrane morphology experiments, we obtain comprehensive information about the protein and membrane structural changes along the fusion pathway (Fig. 2D). We are also investigating the structure of the HIV fusion protein, gp41.

Figure 2. Solid-state NMR studies of virus-cell fusion. (A) Virus-cell fusion model. The fusion protein undergoes a series of conformational changes to transition from a compact prefusion structure to a membrane-bound post-fusion structure. The N-terminal fusion peptide is encapsulated in the globular head in the prefusion state but inserts into the target cell membrane in the extended intermediate state, while the C-terminal TMD is anchored in the viral envelope. Hairpin formation by the protein ectodomain pulls the cell membrane and the virus envelope together, causing putative membrane intermediates such as the hemifusion diaphragm. Action of the FP and TMD eventually causes full merger of the two membranes. (B) Conformations of the parainfluenza virus 5 TMD in different lipid membranes determined by 2D 13C-13C correlation NMR spectra 7-8. The TMD is α-helical in POPC membranes, adopts a mixed helix/sheet structure in DOPC/DOPG membranes, and becomes predominantly β-sheet in negative-curvature POPE membranes. These membrane-dependent conformations suggest that the local lipid composition of the membrane has significant influence on the site of virus-cell fusion. (C) 31P NMR spectra show that the PIV5 F TMD converts the DOPE membrane from the hexagonal phase to a cubic phase, as manifested by a strong isotropic peak 7. (D) Model of the membrane-dependent conformations of the FP and TMD of the PIV5 fusion protein, likely in a hemifusion intermediate. Local enrichment of phosphatidylethanolamine lipids causes β-strand conformations in both the FP and TMD, which in turn cause negative Gaussian curvature and dehydration to the membrane.


Many peptides and proteins can form well-ordered and insoluble β-sheet fibrils. These so-called cross-β amyloid fibrils, whose long axis is perpendicular to the β-strand backbone, can serve biological functions but can also be the result of neurodegenerative disorders. In human brains, protein misfolding into cross-β amyloid fibrils underlies many diseases such as Alzheimer’s disease (AD) and Parkinson’s disease. Understanding the structures and mechanism of fibril formation of these proteins is therefore important for therapeutic intervention of neurodegenerative disorders.


In this project we aim to understand the structure and dynamics of the protein tau. Tau is an intrinsically disordered protein whose function is to associate with and stabilize neuronal microtubules. However, tau aggregates into cross-β amyloid fibrils called paired-helical filaments in many neurodegenerative diseases such as AD, chronic traumatic encephalopathy and corticobasal degeneration. Tau has six isoforms in adult human brains, which mainly differ in whether three or four microtubule-binding repeats are present in the protein (Fig. 3A). Tau aggregation in diseased brains is correlated with hyperphosphorylation and other posttranslational modifications. Cryoelectron microscopy data have shown the structures of the β-sheet core of tau in the brains of several tauopathies. But they do not give information about the dynamics of the protein or the structure of the domains outside the rigid β-sheet core. It is also not known how tau misfolds from its functional, microtubule-bound state, to the dysfunctional, aggregated β-sheet state.


We are employing high-resolution 2D and 3D solid-state NMR spectroscopy to elucidate the structures and dynamics of tau in the amyloid fibril state as well as the microtubule-bound state (Fig. 3A). We have investigated in vitro fibrillized four-repeat (4R) tau and three-repeat (3R) tau 9-10, and obtained low-resolution structural models for the β-sheet cores (Fig. 3B). These studies revealed that the domains outside the β-sheet core exhibit a pronounced dynamical gradient, ranging from being semi-rigid to isotropically mobile. These studies employ the full repertoire of multidimensional correlation NMR experiments at magic-angle-spinning frequencies of 10 to 60 kHz.


In addition to the amyloid fibril structure of tau, we are interested in the microtubule-bound structure of tau. We use solid-state NMR to determine which domain in tau binds these microtubules with the highest affinity, how this binding is affected by posttranslational modification, and how the weakened tau-microtubule interactions potentially lead to β-sheet formation. Using NMR, we are also investigating the mechanism of tau fibril formation in AD brains. How are different tau isoforms incorporated into the paired-helical filament? What is the structure of the toxic seed that is propagated in the diseased brain? These studies integrate NMR spectroscopy, electron microscopy, protein biochemistry, as well as computational modeling and molecular dynamics simulations.


In addition to tau, we study the structures of other amyloid-forming peptides and proteins. One recent example is the peptide hormone glucagon (Fig. 3C) 11, which is used to treat diabetic hypoglycemia. At the high concentration and acidic pH required for pharmaceutical formulation, glucagon rapidly aggregates into well ordered cross-β fibrils. Our study showed that the low-pH glucagon fibril is an antiparallel β-sheet that contains two coexisting molecular conformations. This unique structure gives numerous insights into the molecular interactions that stabilize this fibril fold, and suggest the conformational changes of this peptide hormone from an α-helical structure at low concentrations to the β-sheet structure at high concentrations.

Figure 3. Investigating amyloid protein structures and dynamics by NMR. (A) The tau protein contains multiple domains with different charge properties, which have significant influence on the structure and dynamics of the protein. Tau interacts with microtubules, membranes, as well as self-associates to form ordered β-sheet fibrils.  (B) In vitro heparin-fibrillized tau appears as straight filaments. Detailed 2D and 3D ssNMR experiments showed that the four-repeat tau has a β-sheet core that spans the R2 and R3 domain to form a hairpin. In comparison, three-repeat tau has a large b-sheet core that includes the C-terminus (CT). This three-dimensional fold suggests a protective effect of the CT against toxic fibril formation. (C) High-resolution structure of the glucagon amyloid fibril formed under the pharmaceutical condition of acidic pH and high concentration. The antiparallel β-sheet fibril contains two coexisting molecular conformations, as shown by the peak doubling in the ssNMR spectra.


Figure 4. Solid-state NMR studies of the structure and dynamics of plant cell walls. (A) Cartoon representation of the primary cell walls of dicots, in which cellulose microfibrils act as the scaffold of the cell wall, surrounded by pectins and hemicelluloses. (B) Representative 2D 13C-13C correlation spectra of primary cell walls of Arabidopsis. High magnetic fields enhance spectral resolution and allow resonance assignment, thus permitting the detection of intermolecular contacts between different polysaccharides 12-13. (C) Expansin binding to cellulose in plant cell walls 14. By measuring expansin-to-polysaccharide 1H spin diffusion under DNP, we showed for the first time that expansin’s binding target in the cell wall is cellulose rather than matrix polysaccharides. Molecular dynamics simulation shows how expansin may dock onto the cellulose microfibril. (D) Schematic of the single-network model of the dicot primary wall based on our solid-state NMR data 15-16.

Plant cell walls provide mechanical strength to plant cells while at the same time allowing plants to grow rapidly. Plant cell walls primarily contain three types of polysaccharides: cellulose, hemicellulose, and pectins (Fig. 4A). Although the chemical composition of plant cell walls is relatively well known, the three-dimensional architecture and the dynamics of cell wall polysaccharides have long been elusive due to the lack of high-resolution structural techniques to characterize this insoluble and disordered material.


We are pioneering the application of multidimensional solid-state NMR to elucidate the structures and dynamics of the polysaccharides of intact primary cell walls. By enriching whole plants with 13C, we are able to employ 2D and 3D correlation MAS NMR techniques to detect and resolve the signals of the complex mixture of polysaccharides in intact cell walls (Fig. 4B), and determine their spatial contacts and mobilities. By using sensitivity-enhancing dynamic nuclear polarization (DNP) and paramagnetic relaxation enhancement NMR techniques, we elucidate how polysaccharides interact with proteins to loosen the cell walls during growth (Fig. 4C). Using model plants of both dicot (e.g. Arabidopsis thaliana) and grass (e.g. Brachypodium distachyon and Zea mays) families, we have shown that cellulose, hemicellulose and pectins form a single three-dimensional network (Fig. 4D) instead of two separate networks, thus revising the long-held view of the plant cell wall structure. With our collaborators, we also investigate the structural polymorphism of cellulose microfibrils, hydration of wall polysaccharides, interactions of cellulose with matrix polysaccharides, and the effects of genetic mutations on cell-wall structure.


Driven by our interest in answering fundamental biological questions, we continue to develop and expand the capabilities of solid-state NMR spectroscopy. We have a long-standing interest in increasing the distance reach of NMR. Using nuclear spins with high gyromagnetic ratios such as 19F and 1H, we are now extending the measurable distance upper limit of NMR to ~2 nm (Fig. 5a). Combining 2D correlation experiments with 1H-19F and 13C-19F distance measurements, we can now extract tens to hundreds of nanometer-range distances in 2-3 pairs of 2D spectra 5, 17-18 (Fig. 5b, c). These 19F MAS NMR techniques open the avenues for elucidating the structures of fluorinated compounds bound to their target proteins (Fig. 5d, e), structures of homo-oligomeric membrane proteins and other biological complexes. 1H-detected fast MAS 3D correlation experiments enable resonance assignment in a short amount of time, thus allowing this distance extraction.

Figure 5. Solid-state NMR methods for rapid measurement of nanometer-range distances. (a) Distance rulers in NMR depend on the nuclear spin gyromagnetic ratio g. Commonly used spin pairs in biomolecules with 30-Hz dipolar couplings are shown. 1H and 19F spins have the highest g, and hence longest distance reach. (b) Dipolar coupling strength as a function of internuclear distance for common spin pairs. The longest distances measurable for the intermediate difficulty regime of 30-150 Hz are ~7 Å for 1H-15N and 19F-15N spin pairs, ~10 Å for 1H-13C and 19F-13C spin pairs, and ~15 Å for 1H-1H and 1H-19F spin pairs. (c) Pulse sequence for the 2D hNH resolved 1H-19F REDOR technique. (d) 1H-19F REDOR dephasing measured for the bacterial membrane protein EmrE. These REDOR dephasing curve are obtained from 2D REDOR control (S0) and difference (DS) spectra and are fit to give precise distances. (e) High-resolution structure of EmrE bound to a fluorinated substrate, TPP+. EmrE transports such cationic aromatic antibiotic compounds to cause multidrug resistance.


(1) Mandala, V. S.;  Loftis, A. R.;  Shcherbakov, A. A.;  Pentelute, B. L.; Hong, M, Nat. Struc. Mol. Biol 2020, 27, 160–167.
(2) Mandala, V. S.;  McKay, M. J.;  Shcherbakov, A. A.;  Dregni, A. J.;  Kolocouris, A.; Hong, M., Nat. Struc. Mol. Biol 2020, 27, 1202-1208.
(3) Shcherbakov, A. A.;  Hisao, G.;  Mandala, V. S.;  Thomas, N. E.;  Soltani, M.;  Salter, E. A.;  Davis Jr., J. H.;  Henzler-Wildman, K. A.; Hong, M., Nat. Commun. 2021, 12, 172.
(4) Shcherbakov, A. A.; Spreacker, P. J.; Dregni, A. J.; Henzler-Wildman, K. A.; Hong, M., Nat. Commun. 2022, 13, 991
(5) Shcherbakov, A. A.;  Mandala, V. S.; Hong, M., J. Phys. Chem. B. 2019, 123, 4387-4391.
(6) Gelenter, M. D.;  Mandala, V. S.;  Niesen, M. J. M.;  Sharon, D. A.;  Dregni, A. J.;  Willard, A. P.; Hong, M., Commun. Biol. 2021, 4, 338.
(7) Yao, H.; Lee, M. W.; Waring, A. J.; Wong, G. C.; Hong, M. Proc. Natl. Acad. Sci. USA 2015, 112, 10926-10931.
(8) Yao, H.; Lee, M.; Liao, S. Y.; Hong, M. Biochemistry 2016, 55, 6787-6800.
(9) Dregni, A. J.; Mandala, V. S.; Wu, H.; Elkins, M. R.; Wang, H. K.; Hung, I.; DeGrado, W. F.; Hong, M. Proc. Natl. Acad. Sci. U.S.A 2019, 116, 16357-16366.
(10) Dregni, A. J.; Wang, H. K.; Wu, H.; Duan, P.; Jin, J.; DeGrado, W. F.; Hong, M. Proc. J. Am. Chem. Soc. 2021, 143, 7839-7851.
(11) Gelenter, M. D.; Smith, K. J.; Liao, S. Y.; Mandala, V. S.; Dregni, A. J.; Lamm, M. S.; Tian, Y.; Xu, W.; Pochan, D. J.; Tucker, T. J.; Su, Y.; Hong, M. Nat. Struc. Mol. Biol 2019, 26, 592-598.
(12) Wang, T.; Zabotina, O. A.; Miller, R. C.; Hong, M. Biochemistry 2012, 51, 9846-9856.
(13) Wang, T.; Hong, M. J. Exp. Botany 2016, 67, 503-514.
(14) Wang, T.; Park, Y. B.; Caporini, M. A.; Rosay, M.; Zhong, L.; Cosgrove, D. J.; Hong, M. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 16444-16449.
(15) Dick-Pérez, M.; Zhang, Y.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M. Biochemistry 2011, 50, 989-1000.
(16) Wang, T.; Phyo, P.; Hong, M. Solid State Nucl. Magn. Reson. 2016, 78, 56-63.
(17) Roos, M.; Wang, T.; Shcherbakov, A. A.; Hong, M. J. Phys. Chem. B 2018, 122, 2900-2911.
(18) Shcherbakov, A. A.; Hong, M. J. Biomol. NMR 2018, 71, 31-43.

View Hong Labs complete list of publications