The retroviral structural protein Gag is the primary driving force of assembly of retroviruses. Gag contains three major domains: the N-terminal Matrix domain (MA) for association with the inner leaflet of the plasma membrane (PM), the Capsid domain (CA) for Gag multimerization, and the C-terminal Nucleocapsid domain (NC) for packaging two copies of genomic RNA. The Gag-RNA, Gag-Gag, and Gag-membrane interactions together result in viral assembly. The work presented in this thesis focuses on human immunodeficiency virus type-1 (HIV-1) and Rous sarcoma virus (RSV) Gag interactions with membranes, aiming to understand how Gag binds to the inner leaflet of the PM and how the biophysical properties of the PM contribute to Gag assembly. For membrane association, Gag may exploit electrostatic interactions, hydrophobic interactions, recognition of specific lipid headgroups, protein multimerization, and sensitivity to membrane order. HIV-1 Gag is naturally myristoylated while RSV Gag is not. HIV-1 MA is reported to be sensitive to membrane order, but the mechanism remains elusive. In a comparative study to probe effects of membrane charge and membrane order on Gag-membrane interactions, I developed model membranes with equal phosphatidylserine (PS) concentrations in both disordered (Ld) and ordered (Lo) phases. I found that RSV MA membrane association is primarily based on electrostatic interactions, being only sensitive to membrane charge but not to membrane order. In contrast, myristoylated HIV-1 MA, and proteins containing hydrophobic regions, such as MARCKS, all show preference for disordered membranes. Phosphatidylinositol 4,5-bisphosphate (PIP2) comprises approximately 2 mol% of total inner leaflet lipid. HIV-1 is hypothesized to assemble at PIP2-rich microdomains. To probe whether PIP2-rich domains exist under physiological conditions, I exploited inner leaflet model membranes and used biophysical methods such as self-quenching. I discovered that PIP2 forms clusters at remarkably low concentrations in a multivalent cation-dependent manner. Under physiological cation conditions, free PIP2 and cation-bridged PIP2 clusters can co-exist. Interestingly, I found that HIV-1 not only targets PIP2 clusters, but also further enriches PIP2 at assembly sites as Gag multimerizes in model membranes. I propose that PIP2 cation-bridged cluster formation and protein-induced PIP2 clusters could explain the distinct pools of PIP2 in biological membranes.