Fig. 1: High-resolution reconstructions of MA in immature, mature and MA–SP1 particles.

a, A schematic representation of the Gag cleavage state for each sample. b, Representative image of immature (PR⁻) HIV-1 virion from cryo-EM data (Extended Data Table 1). Arrowheads highlight specific Gag ___domain layers (red, MA; white, immature CA; green, NC). c, Side-view 2D class of the Gag layer. Arrowheads as in b. d, Representative MA–SP1 HIV-1 virion (Extended Data Table 1). e, Side-view 2D class of the Gag layer. As expected, MA and immature CA layers are seen, but NC has been cleaved. Arrowheads as in b. f, Representative mature (WT) HIV-1 virion (Extended Data Table 1). g, Side-view 2D class of the MA layer. Only the MA layer is visible because CA has been cleaved and undergone maturation. h–j, 3D cryo-EM reconstructions of the MA lattices, viewed from the membrane. As previously observed¹, the MA lattice was found to be immature in immature (PR⁻) particles (h), and mature in MA–SP1 (i) and mature (WT) (j) particles. Additional ligand density is observed (orange) in both mature MA lattices, but was absent in the immature lattice. k, Top: top view of the model of the mature MA lattice, including the ligand density. α1, α-helix 1. One MA monomer is blue; surrounding monomers are white. Bottom: rotated view with helices labelled. l, Zoom in of the trimer–trimer interface in the atomic model of the mature MA lattice. The interface is mediated by electrostatic interactions between HBR residues and an electronegative patch along the intra-trimeric interface. m, Zoomed-in view of i showing ligand binding across the trimer–trimer interface with fitted atomic model (grey). n, Rigid fitting of the deposited model of MA bound to PtdIns(4,5)P₂ (Protein Data Bank accession code 2H3Q)¹² shows that the ligand density is not consistent with PtdIns(4,5)P₂ bound as previously reported (the PtdIns(4,5)P₂ atomic model is not accommodated within the orange EM density). Scale bars, 50 nm.