Limits...
Rationalisation of the differences between APOBEC3G structures from crystallography and NMR studies by molecular dynamics simulations.

Autore F, Bergeron JR, Malim MH, Fraternali F, Huthoff H - PLoS ONE (2010)

Bottom Line: In the course of these simulations, we observed a general trend towards increased definition of the beta2 strand for those structures that have a distorted starting conformation of beta2.We also demonstrate that the identification of a pre-defined DNA binding site is prevented by the inherent flexibility of loops that determine access to the deaminase catalytic core.We discuss the implications of our analyses for the as yet unresolved structure of the full-length A3G protein and its biological functions with regard to hypermutation of DNA.

View Article: PubMed Central - PubMed

Affiliation: Randall Division of Cell and Molecular Biophysics, King's College London, London, United Kingdom.

ABSTRACT
The human APOBEC3G (A3G) protein is a cellular polynucleotide cytidine deaminase that acts as a host restriction factor of retroviruses, including HIV-1 and various transposable elements. Recently, three NMR and two crystal structures of the catalytic deaminase domain of A3G have been reported, but these are in disagreement over the conformation of a terminal beta-strand, beta2, as well as the identification of a putative DNA binding site. We here report molecular dynamics simulations with all of the solved A3G catalytic domain structures, taking into account solubility enhancing mutations that were introduced during derivation of three out of the five structures. In the course of these simulations, we observed a general trend towards increased definition of the beta2 strand for those structures that have a distorted starting conformation of beta2. Solvent density maps around the protein as calculated from MD simulations indicated that this distortion is dependent on preferential hydration of residues within the beta2 strand. We also demonstrate that the identification of a pre-defined DNA binding site is prevented by the inherent flexibility of loops that determine access to the deaminase catalytic core. We discuss the implications of our analyses for the as yet unresolved structure of the full-length A3G protein and its biological functions with regard to hypermutation of DNA.

Show MeSH

Related in: MedlinePlus

Structure of the β1-β2 sheet during MD simulations.Number of H-bonds between main-chain atoms of β1 and β2 during the simulated time. For clarity, data for simulations with NMR and crystal structures are shown in separate plots. Colour coding: (A) NMR1* in light green; NMR1-2K3A in pink; NMR2 light bue; NMR2-2K3A* in orange, NMR3* in cyan; NMR3-2K3A in black; (B) XRAY1 in purple; XRAY1-2K3A* in dark green; XRAY2* in yellow and XRAY2-2K3A dark blue. (C) Comparison of the β1-β2 sheet from the starting structures with the most representative structures derived clustering analysis of the MD simulations for NMR1-2K3A, (D) NMR2 and (E) NMR3-2K3A (F) XRAY1 and (G) XRAY2-2K3A.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2902501&req=5

pone-0011515-g006: Structure of the β1-β2 sheet during MD simulations.Number of H-bonds between main-chain atoms of β1 and β2 during the simulated time. For clarity, data for simulations with NMR and crystal structures are shown in separate plots. Colour coding: (A) NMR1* in light green; NMR1-2K3A in pink; NMR2 light bue; NMR2-2K3A* in orange, NMR3* in cyan; NMR3-2K3A in black; (B) XRAY1 in purple; XRAY1-2K3A* in dark green; XRAY2* in yellow and XRAY2-2K3A dark blue. (C) Comparison of the β1-β2 sheet from the starting structures with the most representative structures derived clustering analysis of the MD simulations for NMR1-2K3A, (D) NMR2 and (E) NMR3-2K3A (F) XRAY1 and (G) XRAY2-2K3A.

Mentions: Since the formation of a β-sheet is dependent on the formation of H-bonds between the two β-strands, we also analysed the number of H-bonds between the main-chain atoms of the β1-β2 sheet during the simulations (Figure 6). This analysis confirmed that XRAY1 and XRAY1-2K3A* structures maintain a β1-β2 sheet with a predominantly regular geometry and a minimum of eight H-bonds that remained stable throughout the simulation (Figure 6B). Hydrogen bonding also remained stable during the simulations of XRAY2* and XRAY2-2K3A, and during these simulations the bulged conformation of β2 was maintained. Increases in the number of β1-β2 H-bonds from the starting structure were observed with the NMR1*, NMR2, NMR2-2K3A*, NMR3* and NMR3-2K3A structures (Figure 6A). Of these, NMR1* showed the weakest stabilisation of the β1-β2 sheet by 1 to 2 H-bonds. For the NMR1-2K3A structure there was an initial loss of H-bonds during the simulation, but in the final stages this recovered to the same number of H-bonds as present in the starting structure. Interestingly, during this simulation a different part of β2 gained in definition at the expense of the defined part of the β-sheet present in the starting structure (Figure 6C). Together, these results demonstrate that the A3G C-CDA NMR structures with a poorly defined β2-strand showed a general tendency towards the formation of a more stable β1-β2 sheet and this was observed both in the absence and presence of the five solubility-enhancing mutations (Figure 6). We note, however, that the most dramatic stabilisation of β2 was observed with NMR2-2K3A* and NMR3-2K3A that both contain the solubility enhancing mutations.


Rationalisation of the differences between APOBEC3G structures from crystallography and NMR studies by molecular dynamics simulations.

Autore F, Bergeron JR, Malim MH, Fraternali F, Huthoff H - PLoS ONE (2010)

Structure of the β1-β2 sheet during MD simulations.Number of H-bonds between main-chain atoms of β1 and β2 during the simulated time. For clarity, data for simulations with NMR and crystal structures are shown in separate plots. Colour coding: (A) NMR1* in light green; NMR1-2K3A in pink; NMR2 light bue; NMR2-2K3A* in orange, NMR3* in cyan; NMR3-2K3A in black; (B) XRAY1 in purple; XRAY1-2K3A* in dark green; XRAY2* in yellow and XRAY2-2K3A dark blue. (C) Comparison of the β1-β2 sheet from the starting structures with the most representative structures derived clustering analysis of the MD simulations for NMR1-2K3A, (D) NMR2 and (E) NMR3-2K3A (F) XRAY1 and (G) XRAY2-2K3A.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2902501&req=5

pone-0011515-g006: Structure of the β1-β2 sheet during MD simulations.Number of H-bonds between main-chain atoms of β1 and β2 during the simulated time. For clarity, data for simulations with NMR and crystal structures are shown in separate plots. Colour coding: (A) NMR1* in light green; NMR1-2K3A in pink; NMR2 light bue; NMR2-2K3A* in orange, NMR3* in cyan; NMR3-2K3A in black; (B) XRAY1 in purple; XRAY1-2K3A* in dark green; XRAY2* in yellow and XRAY2-2K3A dark blue. (C) Comparison of the β1-β2 sheet from the starting structures with the most representative structures derived clustering analysis of the MD simulations for NMR1-2K3A, (D) NMR2 and (E) NMR3-2K3A (F) XRAY1 and (G) XRAY2-2K3A.
Mentions: Since the formation of a β-sheet is dependent on the formation of H-bonds between the two β-strands, we also analysed the number of H-bonds between the main-chain atoms of the β1-β2 sheet during the simulations (Figure 6). This analysis confirmed that XRAY1 and XRAY1-2K3A* structures maintain a β1-β2 sheet with a predominantly regular geometry and a minimum of eight H-bonds that remained stable throughout the simulation (Figure 6B). Hydrogen bonding also remained stable during the simulations of XRAY2* and XRAY2-2K3A, and during these simulations the bulged conformation of β2 was maintained. Increases in the number of β1-β2 H-bonds from the starting structure were observed with the NMR1*, NMR2, NMR2-2K3A*, NMR3* and NMR3-2K3A structures (Figure 6A). Of these, NMR1* showed the weakest stabilisation of the β1-β2 sheet by 1 to 2 H-bonds. For the NMR1-2K3A structure there was an initial loss of H-bonds during the simulation, but in the final stages this recovered to the same number of H-bonds as present in the starting structure. Interestingly, during this simulation a different part of β2 gained in definition at the expense of the defined part of the β-sheet present in the starting structure (Figure 6C). Together, these results demonstrate that the A3G C-CDA NMR structures with a poorly defined β2-strand showed a general tendency towards the formation of a more stable β1-β2 sheet and this was observed both in the absence and presence of the five solubility-enhancing mutations (Figure 6). We note, however, that the most dramatic stabilisation of β2 was observed with NMR2-2K3A* and NMR3-2K3A that both contain the solubility enhancing mutations.

Bottom Line: In the course of these simulations, we observed a general trend towards increased definition of the beta2 strand for those structures that have a distorted starting conformation of beta2.We also demonstrate that the identification of a pre-defined DNA binding site is prevented by the inherent flexibility of loops that determine access to the deaminase catalytic core.We discuss the implications of our analyses for the as yet unresolved structure of the full-length A3G protein and its biological functions with regard to hypermutation of DNA.

View Article: PubMed Central - PubMed

Affiliation: Randall Division of Cell and Molecular Biophysics, King's College London, London, United Kingdom.

ABSTRACT
The human APOBEC3G (A3G) protein is a cellular polynucleotide cytidine deaminase that acts as a host restriction factor of retroviruses, including HIV-1 and various transposable elements. Recently, three NMR and two crystal structures of the catalytic deaminase domain of A3G have been reported, but these are in disagreement over the conformation of a terminal beta-strand, beta2, as well as the identification of a putative DNA binding site. We here report molecular dynamics simulations with all of the solved A3G catalytic domain structures, taking into account solubility enhancing mutations that were introduced during derivation of three out of the five structures. In the course of these simulations, we observed a general trend towards increased definition of the beta2 strand for those structures that have a distorted starting conformation of beta2. Solvent density maps around the protein as calculated from MD simulations indicated that this distortion is dependent on preferential hydration of residues within the beta2 strand. We also demonstrate that the identification of a pre-defined DNA binding site is prevented by the inherent flexibility of loops that determine access to the deaminase catalytic core. We discuss the implications of our analyses for the as yet unresolved structure of the full-length A3G protein and its biological functions with regard to hypermutation of DNA.

Show MeSH
Related in: MedlinePlus