Fig 1.
Structural modeling chains A and B of human PLD2.
Sequence alignments used to model chains A (residues 64–315) and B (residues 327–881) of human PLD2 (PDB ids: 6fsf and 3hsi). The secondary structure prediction (SSP) for hPLD2 is listed at the top each alignment by grey (coil), red (strand) and blue (helix) bars; the secondary structure from the structure of each template (dssp) is below the sequence. The loop in chain B for which there is no template is indicated with a gray bar. Residues in the sequence are highlighted as follows: basic (blue), acidic (red), polar (green), apolar (white) aromatic (orange), glycine and proline (pink). The final models for chains A and B are shown in C and D in ribbon style.
Fig 2.
Electrostatic potential analysis.
Electrostatic potential mapped on the surface representation of chains A and B of hPLD2. The cluster of positive residues proposed by Sciorra et al [48] to be part of a PIP2 binding site in chain B is highlighted with a red arrow (side view) and circle (top view). The segment is rendered in red in the ribbon representation of chain B model and the Cα of any lysine or arginine are rendered as blue spheres (third column).
Fig 3.
PIP2 structures used in MD simulations (A and B) and docking calculations (C and D).
A PIP2 headgroup singly-protonated at 4-phosphate on inositol ring and with stearic (18:0) and arachidonic (20:4) acid chains; B preceding PIP2 singly-protonated at 5-phosphate; C headgroup of A; D headgroup of B. Coloring: hydrogen, white; oxygen, red; phosphate, orange; saturated carbon on steric acid chain, yellow; unsaturated carbon on arachidonic acid chain, blue; all other saturated carbons, green.
Table 1.
Lipid component numbers in MD systems.
Fig 4.
Full-length modeling of hPLD2.
The models of chains A and B are oriented relative to a membrane (tan rectangular slab) using the electrostatic potential and putative PIP2 binding residues (regions highlighted in red) as guides. Residues 315 (the C-terminal of chain A) and 327 (the N-terminal of chain B) are connected by a 12 residue linker. The residues not modeled in the template-based step are indicated as dashed lines and were subsequently modeled using ab initio folding in I-TASSER server. The initial chain A and B models and the amino acid sequence of hPLD2 were used as inputs during the ab initio step and the resulting full-length model is shown in the lower panel.
Fig 5.
Docking of tail-truncated PIP2 on the hPLD2 model.
(top) The modular architecture of hPLD2 gene: domains PX (pbox consensus sequence) and PH (pleckstrin homology) are located in N-term (residues from 1 to 325), while C-term (326 to 933) contains four evolutionary conserved sequences (I−IV), where II and IV are HDK motifs. (bottom) Five most stable binding poses of tail-truncated PIP2 (singly-protonated at P4) shown as orange/yellow ball representations. The orange represents the most stable pose for each segment (N-term or C-term). Coloring of named protein regions; PX, light green; PH, green; HDK1, cyan; HDK2, blue.
Fig 6.
Time profiles of the numbers of contacts between PIP2 and PLD2 residues within 3.5 Å.
Data were obtained from the three replicas of the 10% system (red, green, and blue for R1, R2, and R3 in S5 Fig in S1 File, respectively). R1 has been extended to 3μs, while R2 and R3 were simulated for 500 ns. It was calculated for the two segments of hPLD2 separately: ‘N-term’ in the first row and ‘C-term’ in the second row. The vertical dotted line indicates the border between N-term and C-term. The horizontal lines are average numbers for 2–3 μs: 83.3 ± 7.9 (standard deviation) for N-term and 31.5 ± 5.9 for C-term.
Fig 7.
Snapshots of 10% PIP2 bilayer system (R1).
(left) side view at 0, 1, 2, and 3μs and (right) corresponding top view. Coloring is as follows: PIP2, red; POPA, blue; other lipids, cyan; hPLD, gray.
Fig 8.
The numbers of contacts between heavy atoms of PIP2 and PLD2 residues within 3.5 Å.
Data were obtained from the last 1μs trajectories of the replica R1 of the 10% system, dividing into 5 blocks of 200ns to estimate the average and standard error (SE). The vertical dotted line indicates the border between N-term and C-term.
Table 2.
PIP2 binding residues.
Fig 9.
Network representation of hPLD2-lipid bilayer interaction networks.
The networks were created based on the average connectivity during 2–3 μs of R1 of the 10% PIP2 system (not every link is present at every time point). Green and cyan rectangles denote hPLD2 residues in N-term and C-term, respectively; red circles are PIP2, and light blue circles are POPA. The size of the rectangle/circle corresponds to the degree (i.e., the number of links incident to the node). (a) The largest residue cluster (mostly in the PH domain) mediated by PIP2 and POPA, (b) A cluster of 5 residues mostly in the PX domain linked by PIP2, (c) A cluster of 4 residues linked by POPA in C-term, and (d) monomers in C-term, mediated by either PIP2 or POPA.
Fig 10.
Molecular image of the largest cluster on the PH domain at 3μs.
A side view of protein and membrane; B side view of the molecular surface plot of the binding pocket; C view looking from the bilayer midplane up at the protein. Coloring is as follows: (for lipids) PIP2, red small spheres; POPA, blue small spheres; other lipids, gray lines; (for amino acids, space-filling) R, green; T, red; S, magenta; M, blue; A, yellow; V, violet; Y, orange; H, pink; L, cyan; G, silver. Other protein residues in gray ribbon style. The molecular surface was calculated based on van der Waals radii with a probe radius of 1.4 Å using the “Surf” module in VMD [72]. Water and ions are not displayed.