|
|
||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||



* Department of Experimental Immunology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands;
Department of Biochemistry, Radboud University Nijmegen, Nijmegen, The Netherlands; and
Departments of Comparative Genetics and Refinement, and Animal Science, Biomedical Primate Research Centre, Rijswijk, The Netherlands
1Correspondence: Department of Experimental Immunology, K0–144, Academic Medical Center, University of Amsterdam, P.O. Box 22700, 1100 DE Amsterdam, The Netherlands. E-mail: j.hamann{at}amc.uva.nl
SPECIFIC AIMS
Evolution of the heptahelical epidermal growth factor (EGF)-TM7 receptors CD97 and EMR1–4 is a recent and ongoing process. EMR2 is a chimeric molecule with a seven-span transmembrane (TM7) region most related to EMR3 and an EGF-like domain region nearly identical to CD97. While both EMR2 and CD97 interact with chondroitin sulfate (CS), another ligand, CD55, is bound only by CD97. Here we investigated the evolutionary history of EMR2 in relation to its ligand binding properties.
PRINCIPAL FINDINGS
1. EMR2 evolved by regional and frequent DNA transfer with CD97 and EMR3
The striking homology of EMR2 with CD97 and EMR3 implied the possibility that the chimeric structure of EMR2 evolved only recently by incidental exchange of DNA sequences between the encoding, adjacent genes. However, identification of the complete cDNA sequence of rhesus macaque EMR2 did not reveal essential differences with human EMR2. 88% amino acid identity and a comparable regional similarity with CD97 and EMR3 showed that the chimeric nature of EMR2 originates prior to hominoid evolution. We thus searched for EMR2 in phylogenetically more distant genomes by performing basic local alignment search tool (BLAST) surveys of several known mammalian genome drafts. In the dog genome, we identified EMR2 together with CD97 and EMR3 on chromosome 20. Unexpectedly, regional homology of EMR2 with CD97 and EMR3 in the dog closely matches the situation in humans. The five EGF domains of EMR2 and CD97 in the dog are > 95% identical. For the TM7 region, intra- and interspecies homology are in the same range (
85% identity).
Concerted evolution due to the exchange of DNA sequences within gene families leads to an "inverted" homology by which paralogs are more similar than orthologs. To test the possibility that EMR2, CD97, and EMR3 in mammals did not evolve independently, we performed a phylogenetic analysis of the nucleotide sequences encoding the EGF domains, the stalk region, and the TM7 region. Maximum likelihood trees for the compared sequences are shown in Fig. 1
. These trees strongly support concerted evolution of EMR2 and CD97 for the EGF domain region and, more weakly, of EMR2 and EMR3 for the TM7 region. The stalk region of the three genes, in contrast, evolved monophyletically. Further statistical analysis (Kishino-Hasegawa and Shimodaira-Hasegawa test) supported these conclusions. The absence of homologous flanking sequences and the opposite transcriptional orientation of EMR2 with CD97 imply gene conversion as a mechanism of the concerted evolution.
|
2. The chondroitin sulfate binding fourth EGF domain of EMR2 and CD97 is highly conserved in mammals
The EGF domains of EMR2 and CD97 mediate interactions with other cell surface molecules, suggesting that ligand binding formed the driving force for the coevolution of EMR2 and CD97. If true, sequence conservation should be especially high in EGF domains that harbor ligand binding sites. A systematic comparison of EGF domains of CD97 and EMR2 between species from different mammalian orders revealed EGF domain 4, which mediates CS binding in primates, to be significantly more conserved than any other EGF domain (Fig. 2
A). We thus predicted that canine EMR2 and CD97 will also interact with CS. Because the composition of the EGF domain region of both receptors is determined through alternative splicing, we first examined the composition of isoforms expressed in the dog. As shown in Fig. 2B
, the dominantly expressed isoform of both receptors contains EGF domains 4 and so might potentially interact with CS. To test the functionality of the CS binding site in the canine receptors, a multivalent fluorescent probe loaded with recombinant Fc protein of the extracellular part of the dominant, middle isoform of canine EMR2 was generated and tested for binding to glycosaminoglycan-deficient (PgsB-618) and exclusively CS-expressing (PgsD-677) CHO cell mutants. As depicted in Fig. 2C
, canine EMR2 binds CS in a manner comparable to its human homologue.
|
3. Different molecular mechanisms prevent the interaction of EMR2 with CD55 in hominoids
Despite a nearly identical EGF domain region, human EMR2, in contrast to CD97, does not bind CD55. To investigate the reason for this difference, we compared the EGF domain region of CD97 and EMR2 in humans and chimpanzees. Surprisingly, we noted that the three amino acid substitutions within the first two EGF domains that prevent CD55 binding by human EMR2 are not present in chimpanzees. Another difference with human EMR2 was observed in the isoform ratio of chimpanzee EMR2. Other than in humans, chimpanzee transcripts dominantly encode the largest five EGF domain-containing isoform.
To test whether chimpanzee EMR2 possesses specificity for CD55, we generated a multivalent fluorescent probe loaded with the extracellular part of the largest isoform of chimpanzee EMR2. Efficient binding to the glycosaminoglycan-deficient CHO cell mutant PgsB-618, transfected with CD55, revealed that chimpanzee EMR2 indeed has the ability to bind CD55. The question remained as to whether chimpanzee EMR2 will also interact with CD55-expressing primary cells (human leukocytes that express both CD55 and CS). Binding of chimpanzee EMR2–5EGF was efficiently blocked by preincubation of the cells with chondroitinase ABC while pretreatment with CD55 monoclonal antibody (mAb) had little effect.
CONCLUSIONS AND SIGNIFICANCE
We demonstrate molecular mechanisms that have been involved in shaping the chimeric structure of the EGF-TM7 receptor EMR2. Comparing sequences from humans and dogs led to the unexpected finding that the evolution of EMR2 has been continuously linked with that of CD97 and EMR3 since mammal radiation. What makes the concerted evolution of EMR2 unusual is the exchange of DNA sequences in a regional mode with two related genes. It is likely that the frequency of DNA transfer has been higher for EMR2-CD97 than EMR2-EMR3, which can be deduced from a higher homology in the EGF domain region compared with the TM7 region.
Concerted evolution homogenizes paralogs within a genome and thus allows advantageous changes to rapidly spread within multigene families. Our study suggests that concerted evolution can also increase the structural diversity within a gene family. In the case of EMR2, a chimeric receptor evolved that might combine the ligand specificity of CD97 with the signaling capacity of EMR3. We investigated ligand binding and found a high conservation of the CS binding fourth EGF domain in mammals. In contrast to the interaction with CS, CD55 binding evolved differently in the two receptors. We show that different molecular mechanisms (mutations vs. alternative splicing) prevent CD55 binding by EMR2 in hominoids.
In summary, the evolutionary history of the EGF-TM7 family (Fig. 3
) is an amazing combination of molecular mechanisms including gene duplication, exon shuffling, gene conversion, alternative splicing, functional mutations, and even deletion of genes. Due to the late appearance of the EGF-TM7 family only in mammals, we can follow these events from an unusually close perspective.
|
FOOTNOTES
To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.06-6500fje
This article has been cited by other articles:
![]() |
J. Q. Davies, G.-W. Chang, S. Yona, S. Gordon, M. Stacey, and H.-H. Lin The Role of Receptor Oligomerization in Modulating the Expression and Function of Leukocyte Adhesion-G Protein-coupled Receptors J. Biol. Chem., September 14, 2007; 282(37): 27343 - 27353. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |