Structural Studies of Peripheral Membrane Proteins Involved in HIV-1 Infection and Propagation

Soluble peripheral membrane proteins form a class of proteins that transiently associate with the membrane. Even though the interactions are transient, the peripheral
membrane proteins often undergo conformational changes at the membrane interface in order to perform their function. As a result, studying the protein separate from a lipid environment often provides insufficient information to fully understand the biological processes in which these proteins are involved. Neutron reflectometry offers particular advantages to studying peripheral membrane proteins in a biomimetic membrane environment. In this thesis I discuss the application of surface sensitive techniques, including electrochemical impedance spectroscopy, surface plasmon resonance, and neutron reflection, to the study of peripheral membrane proteins involved in HIV-1 viral assembly and propagation. HIV-1 matrix is the myristoylated membrane-targeting domain of the Gag polyprotein, which is the structural factor required for viral assembly and capsid formation. Matrix uses multiple motifs for membrane association including hydrophobic, electrostatic,
and lipid specific interactions with PIP₂. As a result, matrix is conformationally flexible and can adopt multiple conformations at the membrane depending on which motifs are engaged. We identified experimental conditions that overcame the challenges presented by this flexibility and allowed us to determine the effect of myristoylation on the structural organization of matrix bound to charged membranes. The presence of the myristate resulted in shift in membrane contacts from helix II for non-myristoylated matrix on charged membranes to helix I and brought the basic patch used for electrostatic interactions into direct contact with the membrane. In addition the myristoyl-dependent re-orientation positioned key residues favorably for engagement of PIP₂. Relatively simple lipid compositions (two or three components) have been essential for dissecting the molecular mechanisms that drive membrane attraction and determining the membrane-bound structures of matrix and other peripheral membrane proteins. However, to bridge the gap to complementary studies in a cellular context it is desirable to also conduct measurements using model membranes with lipid compositions that more closely mimic the target cellular membranes. In the case of matrix,
this means a model membrane mimic for the inner leaflet of the plasma membrane. Complex, PE-containing tethered bilayer membranes were developed as model membrane
mimics of the inner leaflet of the plasma membrane using an adapted osmotic shock vesicle fusion method for bilayer formation. The effect of the PE-containing complex model membrane on the binding affinity of myristoylated matrix was then measured. Inclusion of PE increased the affinity of myristoylated matrix for charged, cholesterol containing membranes. A plasma membrane mimic containing relevant
phosphoinositides was also developed to be used in future matrix work. In addition to matrix, the membrane interactions of the HIV-1 accessory protein Nef were measured using tethered bilayer membranes. Full-length Nef consists of a well-folded core and a flexible, myristoylated N-terminal arm. The core is primarily responsible for interactions with binding partners essential to Nef function, such as tyrosine kinases, while the N-terminal arm drives membrane binding via the myristate and charged residues. The structure of myristoylated Nef was previously determined using neutron reflectometry on fully (100%) charged monolayers. To determine the conformation of myristoylated Nef on more physiologically relevant bilayers of moderate charge (30%), neutron reflectometry was again applied. For both model systems the Nef core was dynamic and displaced from the membrane while the N-terminal
myristate and charged residues anchored the protein on the membrane. The displacement of the core from the membrane interface is presumably amenable for interactions
with membrane-bound kinases and may also result in dimerization. To provide indirect evidence for dimerization, a dimerization defective mutant was also measured to
probe for conformational differences. The distance of the Nef core from the membrane differed slightly for wild-type Nef and the mutant, although it may simply be due to differences in surface concentration. The long-term goal of the Nef project is to determine the membrane-bound structure(s) of Nef in complex with host cell tyrosine kinases implicated in HIV-1 infection via their interactions with Nef. However, it is important to understand the membrane-interactions of the individual components of the complex before measuring them together. The membrane-association of two non-receptor tyrosine kinases that interact with Nef, Itk and Hck, were measured. Unfortunately, Itk was difficult to express and purify and was aggregation-prone during measurements. Measurements with Hck were more successful and yielded interesting results. We focused on the disordered N-terminal SH4-U region implicated in membrane-targeting and downstream function for which the mechanism(s) controlling these processes are not understood. For the isoform of Hck studied here, the SH4-U region was found to specifically target PA lipids, and the PA/protein interaction resulted in a persistently bound state in which the protein was partially inserted. Addition of a regulatory domain to the SH4-U construct appeared to change the lipid-binding behavior of the protein, which
suggests a role for the regulatory domains in modulating the membrane-association of the protein in addition to their known role in regulating the activity of the kinase
domain.