Limits...
A fully atomistic model of the Cx32 connexon.

Pantano S, Zonta F, Mammano F - PLoS ONE (2008)

Bottom Line: The lack of high resolution data for connexon structures has hampered so far the study of the structure-function relationships that link molecular effects of disease-causing mutations with their observed phenotypes.Our results provide new mechanistic insights into the effects of numerous spontaneous mutations and their implication in connexin-related pathologies.This model constitutes a step forward towards a structurally detailed description of the gap junction architecture and provides a structural platform to plan new biochemical and biophysical experiments aimed at elucidating the structure of connexin channels and hemichannels.

View Article: PubMed Central - PubMed

Affiliation: Institut Pasteur of Montevideo, Montevideo, Uruguay. spantano@pasteur.edu.uy

ABSTRACT
Connexins are plasma membrane proteins that associate in hexameric complexes to form channels named connexons. Two connexons in neighboring cells may dock to form a "gap junction" channel, i.e. an intercellular conduit that permits the direct exchange of solutes between the cytoplasm of adjacent cells and thus mediate cell-cell ion and metabolic signaling. The lack of high resolution data for connexon structures has hampered so far the study of the structure-function relationships that link molecular effects of disease-causing mutations with their observed phenotypes. Here we present a combination of modeling techniques and molecular dynamics (MD) to infer side chain positions starting from low resolution structures containing only C alpha atoms. We validated this procedure on the structure of the KcsA potassium channel, which is solved at atomic resolution. We then produced a fully atomistic model of a homotypic Cx32 connexon starting from a published model of the C alpha carbons arrangement for the connexin transmembrane helices, to which we added extracellular and cytoplasmic loops. To achieve structural relaxation within a realistic environment, we used MD simulations inserted in an explicit solvent-membrane context and we subsequently checked predictions of putative side chain positions and interactions in the Cx32 connexon against a vast body of experimental reports. Our results provide new mechanistic insights into the effects of numerous spontaneous mutations and their implication in connexin-related pathologies. This model constitutes a step forward towards a structurally detailed description of the gap junction architecture and provides a structural platform to plan new biochemical and biophysical experiments aimed at elucidating the structure of connexin channels and hemichannels.

Show MeSH

Related in: MedlinePlus

Overview of the connexon model in a membrane patch and dynamical behavior.a) Cartoon of the side view of the connexon. Each connexin is depicted with different colors. The central connexin is additionally colored by structure: pink for helices, white for coils and light pink for β–strands. The phospholipid membrane, embedding the hemi–channel, is represented in gray. Four phospholipids randomly chosen are colored by atom. b) Root mean square deviations (RMSD) calculated over the Cα atoms of the whole connexon (black) and over the sole TM domain of the connexon (red). Individual RMSD traces of each of the six connexin TM domains are practically indistinguishable and remain below 0.17 nm (lower traces). c) Structural superposition between initial (red) and final (blue) conformers. The Tyr151 residues that delineate the maximum constriction point of the channel pore are shown in space filling mode. d) Minimum diameter of the pore measured as the distance between the OH atoms of opposite Tyr151 in different connexins, averaged over the three possible pairs. The nearly horizontal red dotted line corresponds to a linear fit over the entire simulation, with slope of −0.004 and an y-ordinate of 1.45 nm. e) Stereo view of the rigid docking of the Cα models by Fleishman et al. (red tube) and ours (blue tube) onto the electron microcopy map of Cx43 for the TM part of one connexin seen from the membrane side. Cα atoms are also shown.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2481295&req=5

pone-0002614-g001: Overview of the connexon model in a membrane patch and dynamical behavior.a) Cartoon of the side view of the connexon. Each connexin is depicted with different colors. The central connexin is additionally colored by structure: pink for helices, white for coils and light pink for β–strands. The phospholipid membrane, embedding the hemi–channel, is represented in gray. Four phospholipids randomly chosen are colored by atom. b) Root mean square deviations (RMSD) calculated over the Cα atoms of the whole connexon (black) and over the sole TM domain of the connexon (red). Individual RMSD traces of each of the six connexin TM domains are practically indistinguishable and remain below 0.17 nm (lower traces). c) Structural superposition between initial (red) and final (blue) conformers. The Tyr151 residues that delineate the maximum constriction point of the channel pore are shown in space filling mode. d) Minimum diameter of the pore measured as the distance between the OH atoms of opposite Tyr151 in different connexins, averaged over the three possible pairs. The nearly horizontal red dotted line corresponds to a linear fit over the entire simulation, with slope of −0.004 and an y-ordinate of 1.45 nm. e) Stereo view of the rigid docking of the Cα models by Fleishman et al. (red tube) and ours (blue tube) onto the electron microcopy map of Cx43 for the TM part of one connexin seen from the membrane side. Cα atoms are also shown.

Mentions: By construction, the overall structural determinants of the connexon (Figure 1a) meet those of recombinant, C–terminal truncated, Cx43 channels obtained by cryo–electron microscopy [10], [11]. Molecular mechanics protocols lead to conformations corresponding to local energy minima in the proximity of the starting configuration. We opted instead for MD simulations since temperature effects intrinsic in this technique avoid trapping by shallow local minima, and thus permit the entire complex to explore better the energy landscape. In the course of the MD simulations performed on the Cx32 model, the TM region featured RMSD deviations from the initial conformation ranging from 0.1 up to 0.2 nm for individual connexins, indicative of a very stable behavior. RMSD values for the whole connexin stabilized around 0.3 nm after a simulation time of nearly 5 ns (Figure 1b). The differences between the RMSD calculated over the TM domain of the connexon and those of the separated connexins indicates that the structure of the helix bundles remains stable while global quaternary structure rearrangements take place in the course of the simulation. The observed dynamical behavior is in agreement with a general analysis of the convergence achieved by MD simulations of different membrane proteins within a time window of the order of 10 ns [20].


A fully atomistic model of the Cx32 connexon.

Pantano S, Zonta F, Mammano F - PLoS ONE (2008)

Overview of the connexon model in a membrane patch and dynamical behavior.a) Cartoon of the side view of the connexon. Each connexin is depicted with different colors. The central connexin is additionally colored by structure: pink for helices, white for coils and light pink for β–strands. The phospholipid membrane, embedding the hemi–channel, is represented in gray. Four phospholipids randomly chosen are colored by atom. b) Root mean square deviations (RMSD) calculated over the Cα atoms of the whole connexon (black) and over the sole TM domain of the connexon (red). Individual RMSD traces of each of the six connexin TM domains are practically indistinguishable and remain below 0.17 nm (lower traces). c) Structural superposition between initial (red) and final (blue) conformers. The Tyr151 residues that delineate the maximum constriction point of the channel pore are shown in space filling mode. d) Minimum diameter of the pore measured as the distance between the OH atoms of opposite Tyr151 in different connexins, averaged over the three possible pairs. The nearly horizontal red dotted line corresponds to a linear fit over the entire simulation, with slope of −0.004 and an y-ordinate of 1.45 nm. e) Stereo view of the rigid docking of the Cα models by Fleishman et al. (red tube) and ours (blue tube) onto the electron microcopy map of Cx43 for the TM part of one connexin seen from the membrane side. Cα atoms are also shown.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2481295&req=5

pone-0002614-g001: Overview of the connexon model in a membrane patch and dynamical behavior.a) Cartoon of the side view of the connexon. Each connexin is depicted with different colors. The central connexin is additionally colored by structure: pink for helices, white for coils and light pink for β–strands. The phospholipid membrane, embedding the hemi–channel, is represented in gray. Four phospholipids randomly chosen are colored by atom. b) Root mean square deviations (RMSD) calculated over the Cα atoms of the whole connexon (black) and over the sole TM domain of the connexon (red). Individual RMSD traces of each of the six connexin TM domains are practically indistinguishable and remain below 0.17 nm (lower traces). c) Structural superposition between initial (red) and final (blue) conformers. The Tyr151 residues that delineate the maximum constriction point of the channel pore are shown in space filling mode. d) Minimum diameter of the pore measured as the distance between the OH atoms of opposite Tyr151 in different connexins, averaged over the three possible pairs. The nearly horizontal red dotted line corresponds to a linear fit over the entire simulation, with slope of −0.004 and an y-ordinate of 1.45 nm. e) Stereo view of the rigid docking of the Cα models by Fleishman et al. (red tube) and ours (blue tube) onto the electron microcopy map of Cx43 for the TM part of one connexin seen from the membrane side. Cα atoms are also shown.
Mentions: By construction, the overall structural determinants of the connexon (Figure 1a) meet those of recombinant, C–terminal truncated, Cx43 channels obtained by cryo–electron microscopy [10], [11]. Molecular mechanics protocols lead to conformations corresponding to local energy minima in the proximity of the starting configuration. We opted instead for MD simulations since temperature effects intrinsic in this technique avoid trapping by shallow local minima, and thus permit the entire complex to explore better the energy landscape. In the course of the MD simulations performed on the Cx32 model, the TM region featured RMSD deviations from the initial conformation ranging from 0.1 up to 0.2 nm for individual connexins, indicative of a very stable behavior. RMSD values for the whole connexin stabilized around 0.3 nm after a simulation time of nearly 5 ns (Figure 1b). The differences between the RMSD calculated over the TM domain of the connexon and those of the separated connexins indicates that the structure of the helix bundles remains stable while global quaternary structure rearrangements take place in the course of the simulation. The observed dynamical behavior is in agreement with a general analysis of the convergence achieved by MD simulations of different membrane proteins within a time window of the order of 10 ns [20].

Bottom Line: The lack of high resolution data for connexon structures has hampered so far the study of the structure-function relationships that link molecular effects of disease-causing mutations with their observed phenotypes.Our results provide new mechanistic insights into the effects of numerous spontaneous mutations and their implication in connexin-related pathologies.This model constitutes a step forward towards a structurally detailed description of the gap junction architecture and provides a structural platform to plan new biochemical and biophysical experiments aimed at elucidating the structure of connexin channels and hemichannels.

View Article: PubMed Central - PubMed

Affiliation: Institut Pasteur of Montevideo, Montevideo, Uruguay. spantano@pasteur.edu.uy

ABSTRACT
Connexins are plasma membrane proteins that associate in hexameric complexes to form channels named connexons. Two connexons in neighboring cells may dock to form a "gap junction" channel, i.e. an intercellular conduit that permits the direct exchange of solutes between the cytoplasm of adjacent cells and thus mediate cell-cell ion and metabolic signaling. The lack of high resolution data for connexon structures has hampered so far the study of the structure-function relationships that link molecular effects of disease-causing mutations with their observed phenotypes. Here we present a combination of modeling techniques and molecular dynamics (MD) to infer side chain positions starting from low resolution structures containing only C alpha atoms. We validated this procedure on the structure of the KcsA potassium channel, which is solved at atomic resolution. We then produced a fully atomistic model of a homotypic Cx32 connexon starting from a published model of the C alpha carbons arrangement for the connexin transmembrane helices, to which we added extracellular and cytoplasmic loops. To achieve structural relaxation within a realistic environment, we used MD simulations inserted in an explicit solvent-membrane context and we subsequently checked predictions of putative side chain positions and interactions in the Cx32 connexon against a vast body of experimental reports. Our results provide new mechanistic insights into the effects of numerous spontaneous mutations and their implication in connexin-related pathologies. This model constitutes a step forward towards a structurally detailed description of the gap junction architecture and provides a structural platform to plan new biochemical and biophysical experiments aimed at elucidating the structure of connexin channels and hemichannels.

Show MeSH
Related in: MedlinePlus