Limits...
The Patterns of Coevolution in Clade B HIV Envelope's N-Glycosylation Sites.

Garimalla S, Kieber-Emmons T, Pashov AD - PLoS ONE (2015)

Bottom Line: Indications of pressure to preserve the evolving glycan shield are seen as well as strong dependencies between the majority of the potential N-glycosylation sites and the rest of the structure.The map we propose fills the gap in previous attempts to tease out sequon evolution by providing a more general molecular context.Thus, it will help design strategies guiding HIV gp120 evolution in a rational way.

View Article: PubMed Central - PubMed

Affiliation: University of Michigan Health System, Ann Arbor, MI, United States of America.

ABSTRACT
The co-evolution of the potential N-glycosylation sites of HIV Clade B gp120 was mapped onto the coevolution network of the protein structure using mean field direct coupling analysis (mfDCA). This was possible for 327 positions with suitable entropy and gap content. Indications of pressure to preserve the evolving glycan shield are seen as well as strong dependencies between the majority of the potential N-glycosylation sites and the rest of the structure. These findings indicate that although mainly an adaptation against antibody neutralization, the evolving glycan shield is structurally related to the core polypeptide, which, thus, is also under pressure to reflect the changes in the N-glycosylation. The map we propose fills the gap in previous attempts to tease out sequon evolution by providing a more general molecular context. Thus, it will help design strategies guiding HIV gp120 evolution in a rational way.

No MeSH data available.


Difference in the pair distances of the couplings between sequon and non-sequon positions based on the structure 2B4C.The plot represents medians and interquartile ranges. The intersequon pairings (between positions 1 and 3 in each sequon) were excluded from the analysis. All pairs of positions were grouped first according to the significance of their mfDCA coupling: A not significantly correlated, B with significant mfDCA couplings. Then all positions were grouped in the following groups: non-sequon related—nonSRP, frequent sequon related positions—F, rare sequon related positions—R and variable sequon related positions—V The pairings were further grouped in those between nonSRP, between nonSRP and SRP, the homotypic pairings between F or R and finally—all other intersequon couplings. In this way, the typical distances for the interacting pairs were compared to the remaining pairs of the same type existing in the structure. The distances were compared using Kruskal-Wallis based multiple comparisons test (* - p<0.05). The mfDCA significant couplings were at shorter distances than the non-significant ones except for the homotypic pairs between frequent or rare sequons (F/F and R/R). The interacting homotypic pairs F/F and R/R were also at a longer distance than all other interacting SRP pairs Overall, these results indicate that homotypic interactions between frequent and rare SRP are at long distances and this is not because they are positioned distinctly in the structure since the mfDCA non-significant pairs among them are not different from the rest.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4482261&req=5

pone.0128664.g009: Difference in the pair distances of the couplings between sequon and non-sequon positions based on the structure 2B4C.The plot represents medians and interquartile ranges. The intersequon pairings (between positions 1 and 3 in each sequon) were excluded from the analysis. All pairs of positions were grouped first according to the significance of their mfDCA coupling: A not significantly correlated, B with significant mfDCA couplings. Then all positions were grouped in the following groups: non-sequon related—nonSRP, frequent sequon related positions—F, rare sequon related positions—R and variable sequon related positions—V The pairings were further grouped in those between nonSRP, between nonSRP and SRP, the homotypic pairings between F or R and finally—all other intersequon couplings. In this way, the typical distances for the interacting pairs were compared to the remaining pairs of the same type existing in the structure. The distances were compared using Kruskal-Wallis based multiple comparisons test (* - p<0.05). The mfDCA significant couplings were at shorter distances than the non-significant ones except for the homotypic pairs between frequent or rare sequons (F/F and R/R). The interacting homotypic pairs F/F and R/R were also at a longer distance than all other interacting SRP pairs Overall, these results indicate that homotypic interactions between frequent and rare SRP are at long distances and this is not because they are positioned distinctly in the structure since the mfDCA non-significant pairs among them are not different from the rest.

Mentions: As expected, the mfDCA values decreased with the increase of the interatomic distances (Fig 7). SRP couplings decreased faster than the rest. The large radius of the glycans increases the distance, at which direct contacts would affect fitness Thus, it seemed that SRP coevolution was predominantly affected by direct contacts between amino acid residues or by the attached glycans. At the same time, several long range interactions were found too (Fig 8, S3 File). These included E293–N462, E293–N355, S334–D230, R444–D230, R444–Q363 and N197–N276. Most of them involved the outer domain SRP E293, S334 and R444 and linked them to membrane proximal regions (D230, N355 and N462), and the core near V4/V5 (Q363 and N355). Although these long range interactions were few, they represented homotypic pairing between low entropy positions (both frequent and rare) but not highly variable sequon positions (Fig 9 and Table 1). The long range homotypic frequent to frequent and rare to rare sequon pairings together with much closer frequent to rare interdependencies confirmed that rare sequons most probably are versions of the nearby frequent ones and probably have a similar role. Homotypic coupling is predominant also for variable loops high entropy SRP, which are mostly connected to their neighbors.


The Patterns of Coevolution in Clade B HIV Envelope's N-Glycosylation Sites.

Garimalla S, Kieber-Emmons T, Pashov AD - PLoS ONE (2015)

Difference in the pair distances of the couplings between sequon and non-sequon positions based on the structure 2B4C.The plot represents medians and interquartile ranges. The intersequon pairings (between positions 1 and 3 in each sequon) were excluded from the analysis. All pairs of positions were grouped first according to the significance of their mfDCA coupling: A not significantly correlated, B with significant mfDCA couplings. Then all positions were grouped in the following groups: non-sequon related—nonSRP, frequent sequon related positions—F, rare sequon related positions—R and variable sequon related positions—V The pairings were further grouped in those between nonSRP, between nonSRP and SRP, the homotypic pairings between F or R and finally—all other intersequon couplings. In this way, the typical distances for the interacting pairs were compared to the remaining pairs of the same type existing in the structure. The distances were compared using Kruskal-Wallis based multiple comparisons test (* - p<0.05). The mfDCA significant couplings were at shorter distances than the non-significant ones except for the homotypic pairs between frequent or rare sequons (F/F and R/R). The interacting homotypic pairs F/F and R/R were also at a longer distance than all other interacting SRP pairs Overall, these results indicate that homotypic interactions between frequent and rare SRP are at long distances and this is not because they are positioned distinctly in the structure since the mfDCA non-significant pairs among them are not different from the rest.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4482261&req=5

pone.0128664.g009: Difference in the pair distances of the couplings between sequon and non-sequon positions based on the structure 2B4C.The plot represents medians and interquartile ranges. The intersequon pairings (between positions 1 and 3 in each sequon) were excluded from the analysis. All pairs of positions were grouped first according to the significance of their mfDCA coupling: A not significantly correlated, B with significant mfDCA couplings. Then all positions were grouped in the following groups: non-sequon related—nonSRP, frequent sequon related positions—F, rare sequon related positions—R and variable sequon related positions—V The pairings were further grouped in those between nonSRP, between nonSRP and SRP, the homotypic pairings between F or R and finally—all other intersequon couplings. In this way, the typical distances for the interacting pairs were compared to the remaining pairs of the same type existing in the structure. The distances were compared using Kruskal-Wallis based multiple comparisons test (* - p<0.05). The mfDCA significant couplings were at shorter distances than the non-significant ones except for the homotypic pairs between frequent or rare sequons (F/F and R/R). The interacting homotypic pairs F/F and R/R were also at a longer distance than all other interacting SRP pairs Overall, these results indicate that homotypic interactions between frequent and rare SRP are at long distances and this is not because they are positioned distinctly in the structure since the mfDCA non-significant pairs among them are not different from the rest.
Mentions: As expected, the mfDCA values decreased with the increase of the interatomic distances (Fig 7). SRP couplings decreased faster than the rest. The large radius of the glycans increases the distance, at which direct contacts would affect fitness Thus, it seemed that SRP coevolution was predominantly affected by direct contacts between amino acid residues or by the attached glycans. At the same time, several long range interactions were found too (Fig 8, S3 File). These included E293–N462, E293–N355, S334–D230, R444–D230, R444–Q363 and N197–N276. Most of them involved the outer domain SRP E293, S334 and R444 and linked them to membrane proximal regions (D230, N355 and N462), and the core near V4/V5 (Q363 and N355). Although these long range interactions were few, they represented homotypic pairing between low entropy positions (both frequent and rare) but not highly variable sequon positions (Fig 9 and Table 1). The long range homotypic frequent to frequent and rare to rare sequon pairings together with much closer frequent to rare interdependencies confirmed that rare sequons most probably are versions of the nearby frequent ones and probably have a similar role. Homotypic coupling is predominant also for variable loops high entropy SRP, which are mostly connected to their neighbors.

Bottom Line: Indications of pressure to preserve the evolving glycan shield are seen as well as strong dependencies between the majority of the potential N-glycosylation sites and the rest of the structure.The map we propose fills the gap in previous attempts to tease out sequon evolution by providing a more general molecular context.Thus, it will help design strategies guiding HIV gp120 evolution in a rational way.

View Article: PubMed Central - PubMed

Affiliation: University of Michigan Health System, Ann Arbor, MI, United States of America.

ABSTRACT
The co-evolution of the potential N-glycosylation sites of HIV Clade B gp120 was mapped onto the coevolution network of the protein structure using mean field direct coupling analysis (mfDCA). This was possible for 327 positions with suitable entropy and gap content. Indications of pressure to preserve the evolving glycan shield are seen as well as strong dependencies between the majority of the potential N-glycosylation sites and the rest of the structure. These findings indicate that although mainly an adaptation against antibody neutralization, the evolving glycan shield is structurally related to the core polypeptide, which, thus, is also under pressure to reflect the changes in the N-glycosylation. The map we propose fills the gap in previous attempts to tease out sequon evolution by providing a more general molecular context. Thus, it will help design strategies guiding HIV gp120 evolution in a rational way.

No MeSH data available.