Limits...
The network organization of protein interactions in the spliceosome is reproduced by the simple rules of food-web models.

Pires MM, Cantor M, Guimarães PR, de Aguiar MA, Dos Reis SF, Coltri PP - Sci Rep (2015)

Bottom Line: This food-web model was as successful in reproducing the structure of protein interactions as it is in reproducing interactions among species.Moreover, because network models allow building ensembles of realistic networks while encompassing uncertainty they can be useful to examine the dynamics and vulnerability of intracelullar processes.Unraveling the mechanisms organizing the spliceosome interactions is important to characterize the role of individual proteins on splicing catalysis and regulation.

View Article: PubMed Central - PubMed

Affiliation: Departamento de Ecologia, Instituto de Biociências, 05508-090, Universidade de São Paulo, São Paulo, Brazil.

ABSTRACT
The network structure of biological systems provides information on the underlying processes shaping their organization and dynamics. Here we examined the structure of the network depicting protein interactions within the spliceosome, the macromolecular complex responsible for splicing in eukaryotic cells. We show the interactions of less connected spliceosome proteins are nested subsets of the connections of the highly connected proteins. At the same time, the network has a modular structure with groups of proteins sharing similar interaction patterns. We then investigated the role of affinity and specificity in shaping the spliceosome network by adapting a probabilistic model originally designed to reproduce food webs. This food-web model was as successful in reproducing the structure of protein interactions as it is in reproducing interactions among species. The good performance of the model suggests affinity and specificity, partially determined by protein size and the timing of association to the complex, may be determining network structure. Moreover, because network models allow building ensembles of realistic networks while encompassing uncertainty they can be useful to examine the dynamics and vulnerability of intracelullar processes. Unraveling the mechanisms organizing the spliceosome interactions is important to characterize the role of individual proteins on splicing catalysis and regulation.

No MeSH data available.


Related in: MedlinePlus

The yeast spliceosome networks.(a) Spliceosome protein-protein networks with two different cutoffs defining interactions (0.15 and 0.5). Nodes represent proteins and links represent the interactions among them. Node size is proportional to the protein connectivity (number of interactions a protein establishes with others). Different colors represent different network modules. Networks were built using Gephi (http://gephi.org). (b) Matrix representation of the empirical protein-protein interactions for each cutoff. Each row or column represents a protein, and the black squares represent an interaction between two proteins. (c) Matrix representation of protein-protein interactions yielded by the model. The color heat of squares corresponds to the probability of interactions according to the model. Note the correspondence between the observed interactions (b) and interactions predicted by the model (c).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4595644&req=5

f1: The yeast spliceosome networks.(a) Spliceosome protein-protein networks with two different cutoffs defining interactions (0.15 and 0.5). Nodes represent proteins and links represent the interactions among them. Node size is proportional to the protein connectivity (number of interactions a protein establishes with others). Different colors represent different network modules. Networks were built using Gephi (http://gephi.org). (b) Matrix representation of the empirical protein-protein interactions for each cutoff. Each row or column represents a protein, and the black squares represent an interaction between two proteins. (c) Matrix representation of protein-protein interactions yielded by the model. The color heat of squares corresponds to the probability of interactions according to the model. Note the correspondence between the observed interactions (b) and interactions predicted by the model (c).

Mentions: The S. cerevisiae spliceosome network analyzed contained 103 proteins. Spectral network analysis (see methods) showed that the description of structure was affected by the level of interaction reliability considered to build the network. With a permissive filtering of the network, in which we kept all the interactions with reliability greater than 0.15 (cutoff = 0.15; see Methods) the resulting network contained 2,538 interactions out of the 5,253 possible interactions (network connectance, C = 0.48; Fig. 1). The average connectivity of proteins was 49.28 ± 24.08 interactions per protein, i.e. a highly cohesive network. When we were more restrictive in filtering the network and considered only interactions with reliability greater than 0.5 (cutoff = 0.5), the number of interactions was reduced to 881 (C = 0.17), resulting in a sparser network with 17.10 ± 13.04 interactions per protein. In both configurations analyzed, larger proteins (inferred by their estimated molecular weight, MW) were more likely to have more interactions (0.15 cutoff: slope = 0.19, p < 0.001; 0.5 cutoff: slope = 0.09; p < 0.01; Supplementary Fig. S2). Moreover, the number of interactions of each protein in the networks built under the two different filtering schemes was correlated, suggesting the differences in connectivity among proteins were preserved between cutoffs (r = 0.62, p < 0.001; Supplementary Fig. S3).


The network organization of protein interactions in the spliceosome is reproduced by the simple rules of food-web models.

Pires MM, Cantor M, Guimarães PR, de Aguiar MA, Dos Reis SF, Coltri PP - Sci Rep (2015)

The yeast spliceosome networks.(a) Spliceosome protein-protein networks with two different cutoffs defining interactions (0.15 and 0.5). Nodes represent proteins and links represent the interactions among them. Node size is proportional to the protein connectivity (number of interactions a protein establishes with others). Different colors represent different network modules. Networks were built using Gephi (http://gephi.org). (b) Matrix representation of the empirical protein-protein interactions for each cutoff. Each row or column represents a protein, and the black squares represent an interaction between two proteins. (c) Matrix representation of protein-protein interactions yielded by the model. The color heat of squares corresponds to the probability of interactions according to the model. Note the correspondence between the observed interactions (b) and interactions predicted by the model (c).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4595644&req=5

f1: The yeast spliceosome networks.(a) Spliceosome protein-protein networks with two different cutoffs defining interactions (0.15 and 0.5). Nodes represent proteins and links represent the interactions among them. Node size is proportional to the protein connectivity (number of interactions a protein establishes with others). Different colors represent different network modules. Networks were built using Gephi (http://gephi.org). (b) Matrix representation of the empirical protein-protein interactions for each cutoff. Each row or column represents a protein, and the black squares represent an interaction between two proteins. (c) Matrix representation of protein-protein interactions yielded by the model. The color heat of squares corresponds to the probability of interactions according to the model. Note the correspondence between the observed interactions (b) and interactions predicted by the model (c).
Mentions: The S. cerevisiae spliceosome network analyzed contained 103 proteins. Spectral network analysis (see methods) showed that the description of structure was affected by the level of interaction reliability considered to build the network. With a permissive filtering of the network, in which we kept all the interactions with reliability greater than 0.15 (cutoff = 0.15; see Methods) the resulting network contained 2,538 interactions out of the 5,253 possible interactions (network connectance, C = 0.48; Fig. 1). The average connectivity of proteins was 49.28 ± 24.08 interactions per protein, i.e. a highly cohesive network. When we were more restrictive in filtering the network and considered only interactions with reliability greater than 0.5 (cutoff = 0.5), the number of interactions was reduced to 881 (C = 0.17), resulting in a sparser network with 17.10 ± 13.04 interactions per protein. In both configurations analyzed, larger proteins (inferred by their estimated molecular weight, MW) were more likely to have more interactions (0.15 cutoff: slope = 0.19, p < 0.001; 0.5 cutoff: slope = 0.09; p < 0.01; Supplementary Fig. S2). Moreover, the number of interactions of each protein in the networks built under the two different filtering schemes was correlated, suggesting the differences in connectivity among proteins were preserved between cutoffs (r = 0.62, p < 0.001; Supplementary Fig. S3).

Bottom Line: This food-web model was as successful in reproducing the structure of protein interactions as it is in reproducing interactions among species.Moreover, because network models allow building ensembles of realistic networks while encompassing uncertainty they can be useful to examine the dynamics and vulnerability of intracelullar processes.Unraveling the mechanisms organizing the spliceosome interactions is important to characterize the role of individual proteins on splicing catalysis and regulation.

View Article: PubMed Central - PubMed

Affiliation: Departamento de Ecologia, Instituto de Biociências, 05508-090, Universidade de São Paulo, São Paulo, Brazil.

ABSTRACT
The network structure of biological systems provides information on the underlying processes shaping their organization and dynamics. Here we examined the structure of the network depicting protein interactions within the spliceosome, the macromolecular complex responsible for splicing in eukaryotic cells. We show the interactions of less connected spliceosome proteins are nested subsets of the connections of the highly connected proteins. At the same time, the network has a modular structure with groups of proteins sharing similar interaction patterns. We then investigated the role of affinity and specificity in shaping the spliceosome network by adapting a probabilistic model originally designed to reproduce food webs. This food-web model was as successful in reproducing the structure of protein interactions as it is in reproducing interactions among species. The good performance of the model suggests affinity and specificity, partially determined by protein size and the timing of association to the complex, may be determining network structure. Moreover, because network models allow building ensembles of realistic networks while encompassing uncertainty they can be useful to examine the dynamics and vulnerability of intracelullar processes. Unraveling the mechanisms organizing the spliceosome interactions is important to characterize the role of individual proteins on splicing catalysis and regulation.

No MeSH data available.


Related in: MedlinePlus