Tetramer formation in Arabidopsis MADS domain proteins: analysis of a protein-protein interaction network

Abstract

Background: MADS domain proteins are transcription factors that coordinate several important developmental processes in plants. These proteins interact with other MADS domain proteins to form dimers, and it has been proposed that they are able to associate as tetrameric complexes that regulate transcription of target genes. Whether the formation of functional tetramers is a widespread property of plant MADS domain proteins, or it is specific to few of these transcriptional regulators remains unclear.

Results: We analyzed the structure of the network of physical interactions among MADS domain proteins in Arabidopsis thaliana. We determined the abundance of subgraphs that represent the connection pattern expected for a MADS domain protein heterotetramer. These subgraphs were significantly more abundant in the MADS domain protein interaction network than in randomized analogous networks. Importantly, these subgraphs are not significantly frequent in a protein interaction network of TCP plant transcription factors, when compared to expectation by chance. In addition, we found that MADS domain proteins in tetramer-like subgraphs are more likely to be expressed jointly than proteins in other subgraphs. This effect is mainly due to proteins in the monophyletic MIKC clade, as there is no association between tetramer-like subgraphs and co-expression for proteins outside this clade.

Conclusions: Our results support that the tendency to form functional tetramers is widespread in the MADS domain protein-protein interaction network. Our observations also suggest that this trend is prevalent, or perhaps exclusive, for proteins in the MIKC clade. Because it is possible to retrodict several experimental results from our analyses, our work can be an important aid to make new predictions and facilitates experimental research on plant MADS domain proteins.

Figure 1

Network of protein-protein interactions among Arabidopsis MADS domain proteins. Each node represents an Arabidopsis MADS domain protein encoded by a different gene. Edges bind pairs of nodes that physically interact.

Figure 2

Subgraphs composed of four nodes. a. Subgraph that represents a connection pattern associated to MADS domain protein heterotetramers. Each node binds two other nodes in this subgraph. b. Subgraphs that include all those interactions in a. Curved edges with both ends on the same node represent interactions between identical copies of a protein. c. Subgraphs that do not include all the interactions in a.

Figure 3

Network of protein-protein interactions among Arabidopsis TCP proteins. Each node represents a TCP protein encoded by a different gene. Edges bind pairs of nodes that physically interact.

Figure 4

Joint expression of genes encoding for proteins in tetramer-like subgraphs. The genes that encode for the four proteins in most tetramer-like subgraphs are expressed jointly in one or more Arabidopsis tissues.

Figure 5

The fraction of co-expression in the MADS_exp network is higher for tetramer-like subgraphs than for random ensembles of subgraphs. We define the fraction of co-expression as the proportion of subgraphs composed of proteins whose genes are expressed jointly in at least one tissue. The black dot indicates the fraction of co-expression for tetramer-like subgraphs (0.84). The histogram represents the frequency distribution of the fraction of co-expression for a sample of 10⁴ ensembles, each composed of 1351 4-node subgraphs picked randomly from the MADS_exp network.

Figure 6

Distributions of protein sets co-expressed in n tissues. The distribution for protein sets in tetramer-like subgraphs has a broader tail than that of the distribution associated to subgraphs that are not tetramer-like.