Limits...
Conformational dynamics and ligand binding in the multi-domain protein PDC109.

Kim HJ, Choi MY, Kim HJ, Llinás M - PLoS ONE (2010)

Bottom Line: The effective PDC109-PhC association constant of 28 M(-1), estimated from their potential of mean force is consistent with the experimental result.Principal component analysis of the long timescale MD simulations was compared to the significantly less expensive normal mode analysis of minimized structures.The present study illustrates the use of detailed MD simulations to clarify the energetics of specific ligand-domain interactions revealed by a static crystallographic model, as well as their influence on relative domain motions in a multi-domain protein.

View Article: PubMed Central - PubMed

Affiliation: Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States of America.

ABSTRACT
PDC109 is a modular multi-domain protein with two fibronectin type II (Fn2) repeats joined by a linker. It plays a major role in bull sperm binding to the oviductal epithelium through its interactions with phosphorylcholines (PhCs), a head group of sperm cell membrane lipids. The crystal structure of the PDC109-PhC complex shows that each PhC binds to the corresponding Fn2 domain, while the two domains are on the same face of the protein. Long timescale explicit solvent molecular dynamics (MD) simulations of PDC109, in the presence and absence of PhC, suggest that PhC binding strongly correlates with the relative orientation of choline-phospholipid binding sites of the two Fn2 domains; unless the two domains tightly bind PhCs, they tend to change their relative orientation by deforming the flexible linker. The effective PDC109-PhC association constant of 28 M(-1), estimated from their potential of mean force is consistent with the experimental result. Principal component analysis of the long timescale MD simulations was compared to the significantly less expensive normal mode analysis of minimized structures. The comparison indicates that difference between relative domain motions of PDC109 with bound and unbound PhC is captured by the first principal component in the principal component analysis as well as the three lowest normal modes in the normal mode analysis. The present study illustrates the use of detailed MD simulations to clarify the energetics of specific ligand-domain interactions revealed by a static crystallographic model, as well as their influence on relative domain motions in a multi-domain protein.

Show MeSH

Related in: MedlinePlus

Projections of the principal components on the normal modes for PDC109.(A) p = 1, (B) p = 2, and (C) p = 3. Insets show a magnified view for low frequency modes.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2823774&req=5

pone-0009180-g013: Projections of the principal components on the normal modes for PDC109.(A) p = 1, (B) p = 2, and (C) p = 3. Insets show a magnified view for low frequency modes.

Mentions: Both columns of Fig. 12 show NMs and principal components (PCs) orientated via best-fits of NM and PC reference (red) against the PDC109/a structures. Fig. 12 A depicts the three lowest normal modes of PDC109, which are hinge-bend (n = 1), twist (n = 2), and tilt (n = 3) motions. The modes are shown as overlaid snapshots, where the blue snapshot is time-shifted relative to the red snapshot as described in the Fig. 12 A caption. Fig. 12 B displays the three largest eigenvalue PCs of ligand-free PDC109 as structural displacements (blue) relative to the average reference structure (red). The index increases as the eigenvalue (or the amplitude of atomic fluctuations) decreases. PCs were calculated by diagonalizing the covariance matrix generated from long timescale (400 ns) explicit solvent MD simulations. These PCs therefore account for an anharmonic diffusion motions due to complicated solvent effects. The PC associated with the largest eigenvalue corresponds to the largest atomic fluctuations. Due to solvent damping, PCs have smaller vibrational amplitudes as compared to NMs calculated in vacuum. Compared to NM, loop 2 of each PC rotates about 180 relative to loop 1. This difference stems from the average reference structure of ligand-free PDC109, obtained from the 400 ns trajectory where the two domains mostly remain in perpendicular relative orientation (Fig. 9). Each of the three largest eigenvalue PCs involves rotation of the PDC109/a domain relative to the PDC109/b domain similar to the lowest frequency NMs. To evaluate the relatedness between PCs and NMs, PCs with the three largest eigenvalues were projected onto all NMs of PDC109 as shown in Fig. 13. These projected values are displayed for n 1000 with insets magnifying values for n 24. We find that PCs with large eigenvalues are highly correlated to low frequency NMs. The n = 1 and n = 2 NMs account for 19.1% and 16.5% of the first PC, respectively. The n = 2 and n = 3 NMs account for 37.7% and 11.9% of the second PC, respectively. Similarly, the n = 1 and n = 2 NMs account for 26.5% and 13.2% of the third PC, respectively (see insets of Fig. 13 AC). The contribution from the other NMs on the PCs is negligible, reflecting the fact that the amplitude of NM motion decreases rapidly as the NM index increases.


Conformational dynamics and ligand binding in the multi-domain protein PDC109.

Kim HJ, Choi MY, Kim HJ, Llinás M - PLoS ONE (2010)

Projections of the principal components on the normal modes for PDC109.(A) p = 1, (B) p = 2, and (C) p = 3. Insets show a magnified view for low frequency modes.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2823774&req=5

pone-0009180-g013: Projections of the principal components on the normal modes for PDC109.(A) p = 1, (B) p = 2, and (C) p = 3. Insets show a magnified view for low frequency modes.
Mentions: Both columns of Fig. 12 show NMs and principal components (PCs) orientated via best-fits of NM and PC reference (red) against the PDC109/a structures. Fig. 12 A depicts the three lowest normal modes of PDC109, which are hinge-bend (n = 1), twist (n = 2), and tilt (n = 3) motions. The modes are shown as overlaid snapshots, where the blue snapshot is time-shifted relative to the red snapshot as described in the Fig. 12 A caption. Fig. 12 B displays the three largest eigenvalue PCs of ligand-free PDC109 as structural displacements (blue) relative to the average reference structure (red). The index increases as the eigenvalue (or the amplitude of atomic fluctuations) decreases. PCs were calculated by diagonalizing the covariance matrix generated from long timescale (400 ns) explicit solvent MD simulations. These PCs therefore account for an anharmonic diffusion motions due to complicated solvent effects. The PC associated with the largest eigenvalue corresponds to the largest atomic fluctuations. Due to solvent damping, PCs have smaller vibrational amplitudes as compared to NMs calculated in vacuum. Compared to NM, loop 2 of each PC rotates about 180 relative to loop 1. This difference stems from the average reference structure of ligand-free PDC109, obtained from the 400 ns trajectory where the two domains mostly remain in perpendicular relative orientation (Fig. 9). Each of the three largest eigenvalue PCs involves rotation of the PDC109/a domain relative to the PDC109/b domain similar to the lowest frequency NMs. To evaluate the relatedness between PCs and NMs, PCs with the three largest eigenvalues were projected onto all NMs of PDC109 as shown in Fig. 13. These projected values are displayed for n 1000 with insets magnifying values for n 24. We find that PCs with large eigenvalues are highly correlated to low frequency NMs. The n = 1 and n = 2 NMs account for 19.1% and 16.5% of the first PC, respectively. The n = 2 and n = 3 NMs account for 37.7% and 11.9% of the second PC, respectively. Similarly, the n = 1 and n = 2 NMs account for 26.5% and 13.2% of the third PC, respectively (see insets of Fig. 13 AC). The contribution from the other NMs on the PCs is negligible, reflecting the fact that the amplitude of NM motion decreases rapidly as the NM index increases.

Bottom Line: The effective PDC109-PhC association constant of 28 M(-1), estimated from their potential of mean force is consistent with the experimental result.Principal component analysis of the long timescale MD simulations was compared to the significantly less expensive normal mode analysis of minimized structures.The present study illustrates the use of detailed MD simulations to clarify the energetics of specific ligand-domain interactions revealed by a static crystallographic model, as well as their influence on relative domain motions in a multi-domain protein.

View Article: PubMed Central - PubMed

Affiliation: Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States of America.

ABSTRACT
PDC109 is a modular multi-domain protein with two fibronectin type II (Fn2) repeats joined by a linker. It plays a major role in bull sperm binding to the oviductal epithelium through its interactions with phosphorylcholines (PhCs), a head group of sperm cell membrane lipids. The crystal structure of the PDC109-PhC complex shows that each PhC binds to the corresponding Fn2 domain, while the two domains are on the same face of the protein. Long timescale explicit solvent molecular dynamics (MD) simulations of PDC109, in the presence and absence of PhC, suggest that PhC binding strongly correlates with the relative orientation of choline-phospholipid binding sites of the two Fn2 domains; unless the two domains tightly bind PhCs, they tend to change their relative orientation by deforming the flexible linker. The effective PDC109-PhC association constant of 28 M(-1), estimated from their potential of mean force is consistent with the experimental result. Principal component analysis of the long timescale MD simulations was compared to the significantly less expensive normal mode analysis of minimized structures. The comparison indicates that difference between relative domain motions of PDC109 with bound and unbound PhC is captured by the first principal component in the principal component analysis as well as the three lowest normal modes in the normal mode analysis. The present study illustrates the use of detailed MD simulations to clarify the energetics of specific ligand-domain interactions revealed by a static crystallographic model, as well as their influence on relative domain motions in a multi-domain protein.

Show MeSH
Related in: MedlinePlus