MBE Advance Access originally published online on August 16, 2007
Molecular Biology and Evolution 2007 24(10):2254-2265; doi:10.1093/molbev/msm168
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
A Comparative and Phylogenetic Analysis of the 
-Actinin Rod Domain
Department of Biochemistry, Umeå University, Umeå, Sweden
E-mail: lars.backman{at}chem.umu.se.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
-Actinin is a ubiquitous actin-binding protein, composed of 3 domains; an actin-binding domain and a calcium-binding domain at the termini, connected by a rod domain composed by 1, 2, or 4 spectrin repeats (SRs). To understand how the rod domain has evolved during evolution, we have analyzed and compared the amino acid residue heterogeneity and phylogeny of the SRs of -actinins of vertebrates, invertebrates, fungi, and several protozoa.
The repeats of vertebrate -actinins show a high degree of similarity, whereas repeats of invertebrates, fungi, and, in particular, of protozoa are more divergent.
In the phylogeny, SR1 of all species were clustered together, independent of the number of repeats in the protein. It was also obvious that the second and last repeat in fungi (SR2) grouped with the fourth and last repeat of vertebrates and invertebrates (SR4).
Therefore, the phylogeny implied that the rod domain of the cenancestral -actinin only contained one SR. It was also obvious that SR2 of fungi are related to SR4 of vertebrates and invertebrates, implying that in the second intragenic duplication 2 repeats (i.e., what become SR2 and SR3) were inserted between the initial 2 repeats that become SR1 and SR4.
Key Words: -actinin • phylogeny • evolution • spectrin superfamily • spectrin repeat
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
-Actinin, spectrin, dystrophin, and utrophin are the major members of the spectrin superfamily (Broderick and Winder 2005
). Members of this family are usually involved in organizing the infrastructure within the cell, often by attaching actin filaments to various membrane structures or cross-linking them.
This group of proteins has a common structure: an N-terminal actin-binding domain, composed of 2 calponin homology domains, a C-terminal with at least 2 EF-hand motifs and an interjacent rod domain containing a varying number of spectrin repeats (SRs) (Blanchard et al. 1989; Bennett and Baines 2001; Broderick and Winder 2002; Sutherland-Smith et al. 2003; Broderick and Winder 2005). SRs are also present in several other proteins, such as muscle A-kinase-anchoring protein (Dodge-Kafka and Kapiloff 2006), dystonia/Bpag1 (Pool et al. 2006), microtubule actin cross-linking factor (Leung et al. 1999), and Syne-1 (Beck 2005).
The SR motif forms a left-handed triple helix bundle, where the 3 helices, A, B, and C, wrap around each other (Speicher and Marchesi 1984; Pascual et al. 1996; Broderick and Winder 2002). Although the sequence identity between repeats sometime is low, their structure appears to be well conserved (Yan et al. 1993; Grum et al. 1999; Tang et al. 2001; Ylänne et al. 2001; Park et al. 2003; Kusunoki, Minasov, et al. 2004).
In the repeat structure, certain positions are conserved. There are 2 tryptophans (one in helix A at position 17 and the other in helix C, around position 92–95) that are present in nearly all repeats. These 2 tryptophans have been suggested to be important for the stability of the triple helix bundle (Kusunoki, MacDonald, and Mondragon 2004) although more recently their importance has been questioned (An et al. 2006). Another highly conserved residue is a leucine, 2 residues from the end of the repeat. Figure 1 shows the solution structure of a typical SR, with these 3 conserved residues highlighted. In addition, the surface of each helix in the repeat unit has conserved charged and nonpolar residues that are important not only for stabilizing the repeat through intramolecular interactions but also for interactions with other proteins and cytoplasmic structures, which imply specific functions for each of the SRs (Otey and Carpen 2004).
|
The linker that connects one repeat with the next is continuous between certain but not all repeats (Djinovic-Carugo et al. 1999
; Grum et al. 1999; MacDonald and Cummings 2004). Because the length of the linker varies, it can be difficult to determine the beginning and ending of each repeat. Therefore, the repeat length has been suggested to range from 106 to 122 residues (Parry et al. 1992).
-actinin, the smallest member of the spectrin superfamily, is a ubiquitous actin-binding protein. It is found in almost all organisms with the exception of plants and prokaryotes. Due to its ability to form antiparallel dimers, -actinin is able to cross-link actin filaments into extended networks or actin bundles. In addition, -actinin is also able to attach actin filaments to membrane-bound structures such as focal adhesion contacts and to the Z-disk in the sarcomeres.
It has been suggested that -actinin is the ancestor of the other members of the spectrin superfamily. Phylogenic analyses have indicated that the family members arose after several gene duplications and rearrangements of an ancestral -actinin isoform (Pascual et al. 1997). The evolution of this group of proteins is suggested to involve 2 evolutionary phases: an initial dynamic phase characterized by intragenic duplications, which increased the number of SRs from the 4 present in -actinin to 20 or 24 present in spectrin and dystrophin, respectively (Broderick and Winder 2005). This was followed by a stable phase, with constant number of repeats, where the isoforms evolved independently (Thomas et al. 1997).
Up to the invertebrate–vertebrate bifurcation, organisms seemed to have a single calcium-dependent -actinin. After the split, invertebrates have kept this single isoform in contrast to vertebrates that have acquired 4 isoforms. Two of these isoforms are found in muscle tissue, and both are calcium independent, and the other 2 isoforms are found mainly in nonmuscle tissue and both bind calcium, which in turn regulates the affinity for actin (Blanchard et al. 1989).
The completion of the genomes of several different organisms has led to the discovery of new and atypical -actinin isoforms (Virel and Backman 2004). The protozoa Trichomonas vaginalis (Addis et al. 1998; Bricheux et al. 1998) has an -actinin with one single putative SR. Also the fungus Encephalitozoon cuniculi appears to have an -actinin with a single SR in contrast to other fungi, such as Schizosaccharomyces pombe (Wu et al. 2001), Neurospora crassa, and Aspergillus fumigatus that have -actinins with 2 SRs in the rod domain.
The genome of the parasite Entamoeba histolytica contains genes for 2 different -actinins. Entamoeba -actinin1 is a genuine -actinin with ability to cross-link actin filaments and bind calcium ions (Virel and Backman 2006). However, at present it is uncertain whether the rod domain forms a typical SR or a coiled-coil structure. The other Entamoeba isoform has a longer rod domain that probably forms 2 repeats (Virel et al. 2007).
The discovery of new and atypical -actinin isoforms from ancestral organism has given important insights to the origin of this protein family. During the evolution of -actinin, it is apparent that it is the rod domain that has changed mostly, its length and composition varies considerably from early to more complex organisms. The actin-binding domain and in most cases also the C-terminal EF-hands have been conserved throughout evolution (Virel and Backman 2004).
In contrast to the origin and evolution of spectrin and dystrophin (Muse et al. 1997; Baines 2003), little is known about the processes that have led to the modern -actinin. We have therefore extended our previous analysis (Virel and Backman 2004) and compared the amino acid residue heterogeneity and phylogeny of SRs of -actinin of protozoa, vertebrates, and invertebrates. Repeats of vertebrate -actinins show a high degree of similarity whereas repeats of invertebrates and in particular protozoa are much more divergent. This has allowed us to follow and analyze how the SRs evolved and gave rise to the repeats present in -actinin. Our phylogenic analysis indicate that the -actinin progenitor contained a single SR and that during evolution a first intragenic duplication gave rise to an -actinin with 2 SRs that are closely related to repeats 1 (SR1) and 4 (SR4) of modern -actinins. A second duplication event added 2 more repeats (SR2 and SR3) giving rise to the 4 repeats of modern -actinins.
|
Material and Methods |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
Sequences and Repeats
-Actinin sequences were retrieved from National Center for Biotechnology Information, The Institute for Genomic Research, UniProt Knowledgebase at Swiss Institute of Bioinformatics, Ensembl, DOI Joint Genome Institute (http://genome.jgi-psf.org/), and Broad Institute (http://www.broad.mit.edu/).
SRs were assigned using SUPERFAMILY (Wilson et al. 2007). SMART (Letunic et al. 2006), InterproScan (Quevillon et al. 2005), and Pfam (Bateman et al. 2004) were also used to assign and localize repeats. However, only repeats recognized by SUPERFAMILY were included in the final alignment and phylogenic analysis. Therefore possible repeats, such as those in the rod domain of -actinin1 of E. histolytica and E. cuniculi that were not recognized as typical SRs were not included in the analysis.
The Pfam definition of a typical SR was used to define the beginning and ending of each repeat. Therefore, there is a tryptophan residue at position 17 and a leucine 2 residues from the end in all repeats. The numbering used in all sequences is based on the final alignment and therefore also includes any gaps present in the sequence.
The MODELLER server at the Computational Biology Service Unit at Cornell University (http://cbsuapps.tc.cornell.edu/modeller.aspx) was used for molecular modelling (Sali and Blundell 1993). WHAT CHECK (Hooft et al. 1996) and PROCHECK (Laskowski et al. 1993) available at NIH MBI Laboratory for Structural Genomics and Proteomics (http://nihserver.mbi.ucla.edu/SAVS/) were used to validate obtained structural models.
Sequence Alignment and Analysis
The amino acid sequences of all SRs were aligned using the EMBL-EBI ClustalW www service, using default settings. The alignment of the corresponding nucleotide sequences was achieved using the AA to NA alignment www service (http://www.ii.uib.no/
matthewb/tools/).
Pairwise identities were calculated with the sequence identity matrix tool in BioEdit (Hall 1999). To reduce the data set in the identity matrix, consensus sequences representing vertebrate, invertebrate, and fungi were used when compared with the protozoa sequences.
DnaSP (Rozas et al. 2003) was used to calculate the ratio of the number of nonsynonymous substitutions per nonsynonymous site (Ka) to the number of synonymous substitutions per synonymous site (Ks).
Phylogenetic Analysis
Phylogenetic analyses were done using Bayesian inference, Neighbor-Joining, maximum parsimony, unweighted pair group method with arithmetic mean (UPGMA), and maximum likelihood. For the analysis, MEGA 3.1 (Kumar et al. 2004), MrBayes (Ronquist and Huelsenbeck 2003), and Phyml (Guindon and Gascuel 2003) software packages were used.
MrBayes was run for 10,000,000 or more generation, in order to reach an acceptable standard deviation of split frequencies (around or less than 0.017), using a mixed amino acid model at a fixed rate. The first 25% samples of the cold chain were discarded. Neighbor-Joining consensus trees were established based on 5,000 bootstrap replications, using the Poisson correction model with gaps and missing data treated by complete deletion.
Phyml was run using the Hasegawa-Kishino-Yano (HKY) model, with the gamma model set to 4 categories, for at least 100 replicates.
|
Results |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
-Actinin SequencesAll major databases were interrogated to obtain available
-actinin sequences. Only full length sequences or sequences covering the rod domain were included in the phylogenic analysis. There are several incomplete sequences of -actinin and -actinin–related proteins available but these covers in most cases only the actin-binding domain and were therefore not included.
All -actinin sequences used in this analysis are listed in table 1. The final alignments of all included sequences are shown in supplementary tables (additional file 1 and additional file 2, Supplementary Material online).
|
The SUPERFAMILY as well as InterProScan, SMART, or Pfam servers were used to detect and phase the SRs in the
-actinin sequences. Four independent SRs were detected in all -actinin isoforms of vertebrates and invertebrates as well as in -actinins of Dictyostelium discoideum and Monosiga brevicollis. Although the rod domain of the pathogens Phytophthora infestans, Phytophthora ramorum, and Phytophthora sojae are long enough to contain 4 repeats, SUPERFAMILY only detected a single SR in the -actinins of these organisms.
For the -actinins of fungi and isoform 2 of E. histolytica, the SUPERFAMILY and InterProScan servers, but not the SMART server, identified 2 SRs.
Because neither of the web services detected any typical SRs in -actinins of E. histolytica (isoform 1) and E. cuniculi, these sequences were not included in the phylogenic analysis.
The rod domain of -actinin1 of E. histolytica contains a stretch of 8 residues (K/LQ/AREEQER) that are repeated 9 times. Among these 8 repeats, it is also possible to find 2 perfect repeats of 30 residues. On the DNA level, 92 nt (starting at base pair 729) are repeated from base pair 850. Interestingly, similar regions of multiple 8-residue repeats are present in several other proteins, and these regions are predicted to form coil-coiled structures.
We also tried to model the rod domain of -actinin1 of E. histolytica on the known structure of an SR (SR3 of human -actinin 4, pdb file: 1wlx) using MODELLER. Although sequence identity was low (only 13 of 113 residues were identical), the molecular modelling yielded a triple-helical structure similar to the template structure. In this model, nearly all residues were within allowed regions of the Ramachandran plot. The G-factors calculated by PROCHECK was positive and thus in the acceptable range. Although WHAT CHECK reported that several interatomic distances were too short, thereby reducing the reliability of the structural models, it appears possible that the rod domain of the -actinin1 of E. histolytica forms a triple-helical–like structure.
Pairwise Comparative Analysis
In order to reduce the complexity of the pairwise comparative analysis, initially SRs of vertebrates were only compared with other vertebrate repeats. From the analysis, it was obvious that the differences between repeats within a single isoform were much larger than between the same repeat in -actinins of different organisms. For instance, repeat 1 (SR1) of human and mouse isoform 2 are 100% identical in contrast to SR1 and SR2 of human -actinin2 that are only 13% identical. Further, it was also apparent that SR1 is the most conserved repeat; among all vertebrate SR1 sequences analyzed, the pairwise identity was at least 71% (mean value 87%). The other repeats were less conserved although some pairs were identical; the pairwise identities were at least 67% (82%), 52% (72%), and 54% (79%) for SR2, SR3, and SR4, respectively (mean values within parenthesis).
Similar to the vertebrate sequences, SR1 in invertebrates is also the most conserved repeat, with a pairwise identity of 54% or more among all analyzed sequences (mean value 71%). When omitting the Schistosoma mansoni sequence, the pairwise identity increased by approximately 4%. Again, the other repeats are less identical; in this case, the pairwise identities were at least 40% (67%) for SR2, 35% (56%) for SR3, and 32% (61%) for SR4 (mean values in parenthesis).
In figure 2, consensus sequences of vertebrate, invertebrate, and fungi repeats are compared with each other as well as with protozoa repeats. Although the level of identity is lower when comparing repeats of different phyla, it is obvious that SR1 is again the best-conserved repeat. For instance, the single repeats of T. vaginalis and the Phytophora species are on average 17–18% (±2%) identical to SR1 of all other species but less than 13% (±3%) identical to SR2, SR3, and SR4.
|
Despite of the lower identity, the invertebrate SRs contain the typical conserved residues as defined in Pfam. For instance, the tryptophan at position 17 is conserved in all invertebrate sequences. However, this residue is not conserved in SR1 and SR2 of the fungi, and in SR4 of D. discoideum and S. mansoni, there is a tyrosine in this position. In M. brevicollis SR3 and SR4, the tryptophan is exchanged for phenylalanine. The other characteristic tryptophan (close to the C-terminus), as well as the lysine 2 residues from the end, is also conserved in nearly all repeats.
Although SR1 and SR2 of fungi lack the typical tryptophan at position 17 and SR2 lacks that at position 110 (in our alignment), these repeats are predicted to form typical SRs by the SUPERFAMILY, InterProScan, Pfam, and SMART servers (fig. 1). The tryptophan residues have been suggested to be important for formation and stability of the triple-helical structure (Kusunoki, MacDonald, and Mondragon 2004) although recently their importance has been questioned and instead it has been proposed that hydrophobic interactions between the 3 helices account for the stability of the helical bundle (An et al. 2006). Therefore, the correctness of this structure prediction will have to await further structural determination.
Even though the D. discoideum and M. brevicollis sequences show less identity with other species, the trend is the same; the differences between SRs in the same protein are larger than between the same repeat in different -actinins. It is also evident that repeat 1 of the fungal isoforms are most similar to repeat 1 of the invertebrate (and vertebrate) isoforms, whereas repeat 2 of the fungal -actinins are most similar to repeat 4 of all the other isoforms.
In the crystal structure of human -actinin2, a charge gradient along the dimer interface was apparent, from slightly basic in SR1 to acidic in SR4 (Ylänne et al. 2001). Inspection of the vertebrate and invertebrate sequences indicated that about half of the basic residues of SR1 are located to helix B. Because helix B is involved in the dimer interface with the acidic SR4, this probably reduces any repulsive electrostatic interactions in the antiparallel dimer. The fungal SR1 is also more basic than the fungal SR2, but in this case, the charge difference is much larger. Because the whole rod domain is believed to be involved in the dimer formation, this large charge difference may assure that also the fungal dimers are sufficiently stabilized albeit the shorter rod domain.
Analysis of Nucleotide Sequences
The multiple amino acid sequence alignment was used to convert the corresponding nucleotide sequences into a codon alignment. We then used DnaSP to identify synonymous and nonsynonymous sites and to obtain the Ka/Ks ratio of each sequence pair. To facilitate the analysis of the very large data set (nearly 49,000 lines of data), the data were divided in 4 groups and color coded depending on the ratio: Ka/Ks < 0.3, 0.3 
Ka/Ks < 0.7, 0.7 Ka/Ks < 2.0, and Ka/Ks 
2.0, as shown in supplementary figure (additional file 3, Supplementary Material online). For about one-third of the sequence pairs, the ratio could not be calculated due to too large sequence differences. The calculable Ka/Ks ratios were distributed such that 15% of the values were less than 0.3, 28% were between 0.3 and 0.7, 54% were between 0.7 and 2.0, and 3% were larger than 2.0.
Interestingly, Ka/Ks ratios less than 0.3 were in principle only obtained when an SR of one species was compared with the same repeat of another species. This was particularly true among vertebrate and invertebrate SR1s and SR2s as well as fungi SR1s where nearly all Ka/Ks ratios were less than 0.3. This trend was also seen among SR3 and SR4 of vertebrates and invertebrates and SR2 of fungi, although a significant proportion of values were in the span 0.3–0.7.
The Ka/Ks ratio of a certain repeat compared with any another repeat were in most cases between 0.7 and 2.0, even though both higher and lower ratios were seen. Most noticeable were the large fraction of higher values (>2.0) when vertebrate SR1s were compared with vertebrate SR4s.
Phylogenetic Analysis
The aligned SRs were analyzed using the MrBayes software package. The unrooted consensus tree shown in figure 3 was obtained after running Bayesian inference for 10,000,000 generation, using a fixed rate and mixed amino acid model. The placements of most branches were supported by high posterior probabilities, in most cases higher than 0.90, thereby giving the obtained phylogram reasonable credibility.
|
We also tried to use a variable rate model but the calculation did not reach an acceptable standard deviation of split frequencies within a reasonable number of generations (and computer time). Analysis by Neighbor-Joining gave phylogenic trees with very similar appearance. Also trees obtained by maximum parsimony, UPGMA, and maximum likelihood had a similar overall appearance although some of the branch points were slightly different. However, in all models, each SR was clustered together, despite any minor differences among the different phylograms.
We also run analysis on the aligned nucleotide sequences using Neighbor-Joining and maximum likelihood. The obtained phylograms had an appearance very similar to the trees obtained using the amino acid sequence alignment. Some differences were noticed compared with the protein-based trees; for instance, D. discoideum SR2 and SR4 both distributed to the SR4 leaf and the branch order within each leave was somewhat different (additional file 4, Supplementary Material online).
Independent of whether the amino acid or nucleotide sequence alignment was used, all trees showed that each repeat grouped together; 4 different branches were clearly distinguished corresponding to the 4 SRs (SR1, SR2, SR3, and SR4). In each branch of the unrooted tree in figure 3, vertebrates and invertebrates formed separate but closely related groups. It was also evident from the phylograms as well as from the pairwise identity analysis that the fungal SR1 is related to the first repeat of all metazoa, whereas the second fungal repeat clustered with the fourth metazoa repeat. Although SR2 and SR3 formed 2 distinct groups, the phylogeny implied that these repeats have a common ancestry.
The fungal branches of our tree compared favorably with a recent phylogeny based on nearly 200 fungal species (James et al. 2006) considering the branch order of the ascomycota and basidiomycota phyla. However, the zygomycota Rhizopus oryzae differed in that it was placed together with the basidiomycota in our tree and not on a separate branch (fig. 4).
|
The single repeat present in
-actinin of T. vaginalis appeared to be most closely related to SR1. When the rod domains of E. histolytica -actinin1 and the E. cuniculi isoform were included in the phylogenic analysis, these plausible repeats were placed close to SR1 independent of the model used.
It was also evident that the repeats of the 4 vertebrate -actinin isoforms showed an isoform and not a species-related relation (fig. 5). For instance, SR1 of human -actinin1 was closer to SR1 of chicken -actinin1 than to SR1 of any of the other 3 human isoforms.
|
We also noted that the branch lengths of SR1 and SR2 of the vertebrates were considerably shorter than those of SR3 and SR4. This would imply that fewer sequence changes have occurred in SR1 and SR2, which in turn may indicate that these repeats have been under greater evolutionary pressure than the other 2 repeats.
By rooting the tree using the single repeat of T. vaginalis as an outgroup, we obtained a tree indicating a possible order of branching. The tree shown in figure 6 is an abbreviated form of the complete tree (additional file 5, Supplementary Material online).
|
Surprisingly, the branch of SR4 of
-actinin2 of Pan troglodytes was much longer than all other vertebrate branches, implying a higher rate of sequence changes in this repeat. However, inspection of the alignment (additional file 1, Supplementary Material online) indicated that the first 25 residues of this sequence probably are wrong. The difference compared with all other repeats is much too large to be accounted for by a higher evolutionary rate in this particular part of the sequence.
When we interrogated the databases to retrieve sequences, we noted that some of the D. rerio -actinin sequences appeared to be annotated incorrectly; for instance, accession number Q6TNW2 annotated as -actinin2 is rather a type 3 isoform. Further, contrary to the other species included in this study, 8 distinct isoforms of -actinin have been annotated in the zebra fish genome. This is probably a consequence of a whole-genome duplication believed to have occurred during the evolution of ray-finned fish (Van de Peer 2004).
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
-actinins can be divided into 3 different groups based on the number of SRs present in the rod domain; all vertebrate and invertebrate -actinins belong to the group that has 4 repeats, most, if not all, fungal -actinins belong to the second group that has 2 repeats, and -actinins of some unicellular organisms such as T. vaginalis with only a single repeat constitutes a third group. When the amino acid sequences of the repeats are aligned, it is obvious that all the vertebrate sequences are very similar with a high degree of identity. The invertebrate sequences as well as the fungi sequences also show a high, albeit lower, degree of identity. However, the identities between these 3 groups of organisms are much lower. The single-repeat -actinins display an even lower sequence identity, both within this group and in relation to the other groups. This may indicate that there has been less evolutionary pressure on these single-repeat -actinins compared with the 4-repeat isoforms where SR1 and SR4 should complement one another.
It seems rational to divide the evolution of -actinins into 2 phases; an initial and highly dynamic phase followed by a consolidation phase. During the dynamic phase, 2 rounds of intragenic duplication occurred, thereby creating an -actinin with a rod domain containing 4 SRs from an ancestor with a short rod domain, comprising only a single spectrin repeat or possibly a coiled coil. Thereafter did not the number of repeats change but instead began the different -actinins to evolve independently. It is during this phase and after the vertebrate–invertebrate bifurcation that the 4 different -actinin isoforms present in vertebrates evolved.
Spectrin and dystrophin, the other members of the spectrin superfamily, appear to follow a similar evolutionary plan (Thomas et al. 1997). In this case, it has been suggested that the dynamic phase involved several nested and nonuniform duplications from an ancestral -actinin with 4 SRs (Pascual et al. 1997). Presence of spectrin is a distinct feature of metazoan, neither spectrin nor dystrophin have been found in any protozoa (Bennett and Baines 2001).
As the pairwise comparative analysis showed, there is a high sequence divergence across the different repeats within the same isoform (fig. 2). In fact, a certain repeat is more similar to the same repeat in other isoforms than with the other repeats in the same isoform. Furthermore, of the 4 repeats, it is clear that SR1 is the most conserved repeat in all -actinins. SR1 in vertebrates is extremely well conserved; in the 50 sequences included in the analysis, 68 of 112 residues are identical and this number increase to 75 if the fish sequences are excluded, compared with 40% or less identical residues in the other 3 repeats. This clearly indicates that SR1 has been under strong evolutionary pressure, probably due to its role in dimer formation. It has also been suggested that the highly similar SR1 in ß-spectrin has a role in the binding of actin filaments (Li and Bennett 1996; Djinovic-Carugo et al. 2002). Therefore this repeat may have ceased to evolve earlier than the other repeats due to functional pressure. Characterization of -actinins of E. histolytica has shown that the 2-repeat isoform (Virel et al. 2007) as well as the single-repeat isoform can form dimers (Virel and Backman 2006). Therefore, it seems quite likely that SR1 acquired a specific function early in evolution. This is supported by the low values of the Ka/Ks ratio for all SR1s and in particular the vertebrate SR1s.
Further, support is given by the phylogeny. In the phylogenetic trees, all SR1 were grouped together, independent on whether in an -actinin with 1, 2, or 4 repeats (fig. 2). When the rod domains of E. histolytica -actinin1 and that of E. cuniculi were included in the phylogeny also, these were placed close to SR1 in the phylogeny.
The rod domain of E. cuniculi -actinin is not only less conserved than that of E. histolytica -actinin1 but also much longer that required for a single repeat. Therefore, it can be questioned whether this protein should be regarded as an -actinin until it has been characterized further.
In invertebrates, the sequence conservation follows the same trend as in vertebrates, although on a slightly lower level; again SR1 is the most conserved repeat, whereas SR3 is the least conserved. This implies that the selection pressure may have been different and, perhaps, less during the evolution of these -actinin isoforms. It is also possible that the rate of evolution between the different invertebrate species has been larger than among the vertebrates leading to a greater divergence.
SR2 and SR4 of -actinins with 2 and 4 repeats, respectively, also grouped together. Therefore, SR1 and SR2 of fungi and E. histolytica 2 are closely related to SR1 and SR4 of the 4-repeat -actinins. Similarly, SR2 and SR3 of 4-repeat -actinins were placed on separate but very close branches in the phylogenetic trees. Together with the pairwise distance analysis, this suggests a plausible evolutionary pattern of the rod domain of -actinins (fig. 7).
|
As suggested in previous work, we believe that the most ancestral repeat was SR1. This ancestral repeat did probably not yet contain the conserved residues or structures of a typical SR. For instance, in E. histolytica
-actinin1, it is predicted that the rod domain forms a coiled coil though not necessarily a triple-helical structure (Virel and Backman 2006). After a first intragenic duplication, giving rise to SR1 and SR2 in all 2-repeat -actinins, these repeats acquired the typical SR properties. Due to a second intragenic duplication, a gene fragment of SR1 and SR2 was inserted between these 2 repeats. Thereby the initial SR2 became SR4 and the inserted fragment gave rise to SR2 and SR3. Spectrin has been suggested to follow a similar evolutionary path. When spectrin evolved from -actinin, 4 repeats were added between SR2 and SR3 by unequal crossing-over (Pascual et al. 1997; Thomas et al. 1997). This scenario implies that fungi branched off before this second intragene duplication. The lack of certain typical residues in the fungi repeats (such as the tryptophan at position 17) not only supports this suggestion but also implies that SR1 and SR2 evolved into typical SRs after this bifurcation.
Although less likely, we cannot exclude the possibility that fungi have followed the same evolutionary plan as invertebrates and vertebrates but instead of keeping all 4 repeats lost 2 of them.
Interestingly, all SRs of vertebrate and invertebrate have a tryptophan at position 17 with the exception of SR4 of Fugu rubripes and Tetraodon nigroviridis -actinin4 that have tyrosine and histidine, respectively, in this position. SR1, SR2, and SR3 of Dictyostelium all have a tryptophan at this position, whereas there is a tyrosine in SR4. Similarly, the 3 first repeats of a hypothetical -actinin of S. mansoni appear to have a tryptophan at position 17 and a tyrosine at this position in the forth repeat. The significance of this tryptophan has to await the completion of more -actinin sequences from organism at the bottom of the evolutionary tree.
|
Conclusions |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
Protozoa such as E. histolytica, T. vaginalis, and E. cuniculi are so far the most basal organisms where an
-actinin or -actinin–like protein has been identified. Therefore, it is likely that protozoa -actinin constitutes the cenancestor of this family of proteins. It also seems possible that the first -actinins appeared due to a necessity of a complex cytoskeletal network to improve motility as well as resistance toward different environmental conditions.
Our analysis supports the suggestion that the evolution of -actinin can be divided into 2 phases. The first phase occurred in invertebrates and was characterized by 2 rounds of intragenic duplication that increased the number of repeats from 1 to 4 as well as a high evolutionary rate. In this phase, the sequence conservation across a specific repeat was low. The second phase appears to have been more stable, with a lower evolutionary rate. It was during this phase the 4 different isoforms of -actinin present in vertebrates appeared.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary tables and figures as additional files 1–5 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
We thank Drs Pär Ingvarsson and Magnus Wolf-Watz for their discussions. This work was supported by grants from Carl Tryggers Stiftelse.
|
Footnotes |
|---|
Michele Vendruscolo, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterial and MethodsResultsDiscussionConclusionsSupplementary MaterialAcknowledgementsReferences |
|---|
Addis MF, Rappelli P, Delogu G, Carta F, Cappuccinelli P, Fiori PL. Cloning and molecular characterization of a cDNA clone coding for Trichomonas vaginalis alpha-actinin and intracellular localization of the protein. Infect Immun (1998) 66:4924–4931.
An X, Guo X, Zhang X. (12 co-authors). Conformational stabilities of the structural repeats of erythroid spectrin and their functional implications. J Biol Chem (2006) 281:10527–10532.
Baines AJ. Comprehensive analysis of all triple helical repeats in beta-spectrins reveals patterns of selective evolutionary conservation. Cell Mol Biol Lett (2003) 8:195–214.[Web of Science][Medline]
Bateman A, Coin L, Durbin R. (13 co-authors). The Pfam protein families database. Nucleic Acids Res (2004) 32:D138–D141.
Beck KA. Spectrins and the Golgi. Biochim Biophys Acta (2005) 1744:374–382.[Medline]
Bennett V, Baines AJ. Spectrin and ankyrin-based pathways: metazoan inventions for integrating cells into tissues. Physiol Rev (2001) 81:1353–1392.
Blanchard A, Ohanian V, Critchley D. The structure and function of -actinin. J Muscle Res Cell Motil (1989) 10:280–289.[CrossRef][Web of Science][Medline]
Bricheux G, Coffe G, Pradel N, Brugerolle G. Evidence for an uncommon -actinin protein in Trichomonas vaginalis. Mol Biochem Parasitol (1998) 95:241–249.[CrossRef][Web of Science][Medline]
Broderick MJ, Winder SJ. Towards a complete atomic structure of spectrin family proteins. J Struct Biol (2002) 137:184–193.[CrossRef][Web of Science][Medline]
Broderick MJ, Winder SJ. Spectrin, -actinin, and dystrophin. Adv Protein Chem (2005) 70:203–246.[CrossRef][Web of Science][Medline]
Djinovic-Carugo K, Gautel M, Ylanne J, Young P. The spectrin repeat: a structural platform for cytoskeletal protein assemblies. FEBS Lett (2002) 513:119–123.[CrossRef][Web of Science][Medline]
Djinovic-Carugo K, Young P, Gautel M, Saraste M. Structure of the -actinin rod: molecular basis for cross-linking of actin filaments. Cell (1999) 98:537–546.[CrossRef][Web of Science][Medline]
Dodge-Kafka KL, Kapiloff MS. The mAKAP signaling complex: integration of cAMP, calcium, and MAP kinase signaling pathways. Eur J Cell Biol (2006) 85:593–602.[CrossRef][Web of Science][Medline]
Grum VL, Li DN, MacDonald RI, Mondragon A. Structures of two repeats of spectrin suggest models of flexibility. Cell (1999) 98:523–535.[CrossRef][Web of Science][Medline]
Guindon S, Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol (2003) 52:696–704.[CrossRef][Web of Science][Medline]
Hall TA. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser (1999) 41:95–98.
Hooft RW, Vriend G, Sander C, Abola EE. Errors in protein structures. Nature (1996) 381:272.[Medline]
James TY, Kauff F, Schoch CL, et al. Reconstructing the early evolution of Fungi using a six-gene phylogeny. Nature (2006) 443:818–822.[CrossRef][Medline]
Koradi R, Billeter M, Wuthrich K. MOLMOL: a program for display and analysis of macromolecular structures. J Mol Graph (1996) 14:51–55.[CrossRef][Web of Science][Medline]
Kowalski K, Merkel AL, Booker GW. 1H, 13C and 15N resonance assignments of the third spectrin repeat of alpha-actinin-4. J Biomol NMR (2004) 29:533–534.[CrossRef][Web of Science][Medline]
Kumar S, Tamura K, Nei M. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Bioinformation (2004) 5:150–163.
Kusunoki H, MacDonald RI, Mondragon A. Structural insights into the stability and flexibility of unusual erythroid spectrin repeats. Structure (2004) 12:645–656.[Medline]
Kusunoki H, Minasov G, Macdonald RI, Mondragon A. Independent movement, dimerization and stability of tandem repeats of chicken brain alpha-spectrin. J Mol Biol (2004) 344:495–511.[CrossRef][Web of Science][Medline]
Laskowski RA, MacArthur MW, Moss DS, Thornton JM. PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Cryst (1993) 26:283–291.[CrossRef][Web of Science]
Letunic I, Copley RR, Pils B, Pinkert S, Schultz J, Bork P. SMART 5: domains in the context of genomes and networks. Nucleic Acids Res (2006) 34:D257–D260.
Leung CL, Sun D, Zheng M, Knowles DR, Liem RK. Microtubule actin cross-linking factor (MACF): a hybrid of dystonin and dystrophin that can interact with the actin and microtubule cytoskeletons. J Cell Biol (1999) 147:1275–1286.
Li XL, Bennett V. Identification of the spectrin subunit and domains required for formation of spectrin/adducin/actin complexes. J Biol Chem (1996) 271:15695–15702.
MacDonald RI, Cummings JA. Stabilities of folding of clustered, two-repeat fragments of spectrin reveal a potential hinge in the human erythroid spectrin tetramer. Proc Natl Acad Sci USA (2004) 101:1502–1507.
Muse SV, Clark AG, Thomas GH. Comparisons of the nucleotide substitution process among repetitive segments of the a- and b-spectrin genes. J Mol Evol (1997) 44:492–500.[CrossRef][Web of Science][Medline]
Otey CA, Carpen O. -actinin revisited: a fresh look at an old player. Cell Motil Cytoskelet (2004) 58:104–111.[CrossRef][Web of Science][Medline]
Park S, Caffrey MS, Johnson ME, Fung LW. Solution structural studies on human erythrocyte alpha-spectrin tetramerization site. J Biol Chem (2003) 278:21837–21844.
Parry DAD, Dixon TW, Cohen C. Analysis of the three-a-helix motif in the spectrin superfamily of proteins. Biophys J (1992) 61:858–867.[Web of Science][Medline]
Pascual J, Castresana J, Saraste M. Evolution of the spectrin repeat. BioEssays (1997) 19:811–817.[CrossRef][Web of Science][Medline]
Pascual J, Pfuhl M, Rivas G, Pastore A, Saraste M. The spectrin repeat folds into a three-helix bundle in solution. FEBS Lett (1996) 383:201–207.[CrossRef][Web of Science][Medline]
Pool M, Rippstein P, McBride H, Kothary R. Trafficking of macromolecules and organelles in cultured Dystonia musculorum sensory neurons is normal. J Comp Neurol (2006) 494:549–558.[CrossRef][Web of Science][Medline]
Quevillon E, Silventoinen V, Pillai S, Harte N, Mulder N, Apweiler R, Lopez R. InterProScan: protein domains identifier. Nucleic Acids Res (2005) 33:W116–W120.
Ronquist F, Huelsenbeck JP. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (2003) 19:1572–1574.
Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics (2003) 19:2496–2497.
Sali A, Blundell TL. Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol (1993) 234:779–815.[CrossRef][Web of Science][Medline]
Speicher DW, Marchesi VT. Erythrocyte spectrin is comprised of many homologous triple helical segments. Nature (1984) 311:177–180.[CrossRef][Medline]
Sutherland-Smith AJ, Moores CA, Norwood FL, Hatch V, Craig R, Kendrick-Jones J, Lehman W. An atomic model for actin binding by the CH domains and spectrin-repeat modules of utrophin and dystrophin. J Mol Biol (2003) 329:15–33.[CrossRef][Web of Science][Medline]
Tang J, Taylor DW, Taylor KA. The three-dimensional structure of a-actinin obtained by cryoelectron microscopy suggests a model for Ca2+-dependent actin binding. J Mol Biol (2001) 310:845–858.[CrossRef][Web of Science][Medline]
Thomas GH, Newbern EC, Korte CC, Bales MA, Muse SV, Clark AG, Kiehart DP. Intragenic duplication and divergence in the spectrin superfamily of proteins. Mol Biol Evol (1997) 14:1285–1295.[Abstract]
Van de Peer Y. Tetraodon genome confirms Takifugu findings: most fish are ancient polyploids. Genome Biol (2004) 5:250.[CrossRef][Medline]
Virel A, Addario B, Backman L, Forthcoming. Characterization of Entamoeba histolytica -actinin2. Mol Biochem Parasitol (2007).
Virel A, Backman L. Molecular evolution and structure of -actinin. Mol Biol Evol (2004) 21:1024–1031.
Virel A, Backman L. Characterization of Entamoeba histolytica -actinin. Mol Biochem Parasitol (2006) 145:11–17.[CrossRef][Web of Science][Medline]
Wilson D, Madera M, Vogel C, Chothia C, Gough J. The SUPERFAMILY database in 2007: families and functions. Nucleic Acids Res (2007) 35:D308–D313.
Wu JQ, Bahler J, Pringle JR. Roles of a fimbrin and an -actinin-like protein in fission yeast cell polarization and cytokinesis. Mol Biol Cell (2001) 12:1061–1077.
Yan Y, Winograd E, Viel A, Cronin T, Harrison SC, Branton D. Crystal structure of the repetitive segments of spectrin. Science (1993) 262:2027–2030.
Ylänne J, Scheffzek K, Young P, Saraste M. Crystal structure of the -actinin rod reveals an extensive torsional twist. Structure (2001) 9:597–604.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Wilson, R. Pethica, Y. Zhou, C. Talbot, C. Vogel, M. Madera, C. Chothia, and J. Gough SUPERFAMILY--sophisticated comparative genomics, data mining, visualization and phylogeny Nucleic Acids Res., January 1, 2009; 37(suppl_1): D380 - D386. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||