The evolution of filamin-a protein domain repeat perspective

Abstract

Particularly in higher eukaryotes, some protein domains are found in tandem repeats, performing broad functions often related to cellular organization. For instance, the eukaryotic protein filamin interacts with many proteins and is crucial for the cytoskeleton. The functional properties of long repeat domains are governed by the specific properties of each individual domain as well as by the repeat copy number. To provide better understanding of the evolutionary and functional history of repeating domains, we investigated the mode of evolution of the filamin domain in some detail. Among the domains that are common in long repeat proteins, sushi and spectrin domains evolve primarily through cassette tandem duplications while scavenger and immunoglobulin repeats appear to evolve through clustered tandem duplications. Additionally, immunoglobulin and filamin repeats exhibit a unique pattern where every other domain shows high sequence similarity. This pattern may be the result of tandem duplications, serve to avert aggregation between adjacent domains or it is the result of functional constraints. In filamin, our studies confirm the presence of interspersed integrin binding domains in vertebrates, while invertebrates exhibit more varied patterns, including more clustered integrin binding domains. The most notable case is leech filamin, which contains a 20 repeat expansion and exhibits unique dimerization topology. Clearly, invertebrate filamins are varied and contain examples of similar adjacent integrin-binding domains. Given that invertebrate integrin shows more similarity to the weaker filamin binder, integrin β3, it is possible that the distance between integrin-binding domains is not as crucial for invertebrate filamins as for vertebrates.

Fig. 1

Long repeat domains in vertebrate species. Hierarchically clustered heatmap of the commonality of the long repeat domains (LRD) (x-axis) in the various species listed on the y-axis. Each cell in the heatmap reflects the fraction of the LRD among all LRDs in the species. The brighter the red, the more common the repeat is in the species while green indicates that the domain is less common. The frequencies are clustered by row and then by column using euclidian distance and average linkage.

Fig. 2

Examples of repeat domain expansions. Each plot shows the internal sequence identity domainwise within the protein. Black shows complete sequence identity, while white indicates no sequence identity. (A) Tandem casette duplication of spectrin domains in Xenopus tropicalis, (B) two separate clustered tandem duplications of the immunoglobulin domain in Ailuropoda melanoleuca. (C) The every other similarity pattern seen in the Filamin of Strongylocentrotus purpuratus. (D) A Ciona savignyi protein containing KH domain repeats.

Fig. 3

Schematic illustration of human filamin in the cell. The light blue bar represents the membrane, the blue beads illustrate the filamin domains and the red beads illustrate the actin-binding domain. The green bars represent β-integrins and the orange rectangles their transmembrane regions. The numbers show the positions of the filamin domains relative to the N-terminus. Note that the image is a schematic and may not reflect the mechanics of the filamin–integrin interaction.

Fig. 4

Species tree with domain assignments for putative filamins. The tree is based on 31 ubiquitous proteins, as described by Cicarelli (Ciccarelli et al., 2006). The yellow boxes show CH domains and the pink boxes show filamin domains. The color intensity indicates the degree of sequence similarity to the 21st filamin domain of human filamin A, the strongest integrin binding domain. Blue boxes indicate domains other than CH and filamin domains.

Fig. 5

Internal similarity matrices for filamin proteins. Each cell represents a filamin domain. The darker the cell, the higher sequence similarity between the two domains compared in that cell. Opacity values can only be compared within one protein, since the values are normalized against the highest score for that particular protein. BFL – Branchistoma floridae, DME – Drosophila melanogaster, HME – Hirudo medicinalis, HSA – Homo sapiens, NVE – Nematostella vectensis and TAD – Trichoplax adhaerans.

Fig. 6

Pattern (LOGO) for the non-binding (A), binding (B) and dimerizing (C) domains of Filamin. The pattern was created using weblogo (Crooks et al., 2004).

Fig. 7

Structural modeling of filamin and integrin. Filamin domains are colored grey in non-binding regions and green in the binding region (known as the CD β-strands). Cytoplasmic tails of β-integrins are shown in purple. Positions in the binding filamin pattern (LOGO) with high information content are shown as sticks, with residue names and pattern positions in green. Positions in the integrin pattern with high information content are shown as sticks, with residue names and pattern positions in black.

Fig. 8

Pattern (LOGO) for the integrin cytoplasmic tail. The pattern was created using weblogo (Crooks et al., 2004).

Fig. 9

Domain architechtures of Filamin (A), β-integrin (B), Migfilin (C) and GP1bα (D).

Fig. 10

Phylogenetic tree for the cytoplasmic integrin tail. The tree was built based on multiple sequence alignment created using the program Muscle (Edgar, 2004) using the neighbor joining method.