The basic task of molecular biology is to determinethe relationship between the structures and functionsof biological molecules. A convenient model is pro-vided by proteins with similar functions in differentprocesses (e.g., in mitosis and meiosis) within a singleorganism. In particular, any cell division is precededby an association (cohesion) of sister chromatids,which involves cohesins, specific protein complexes.Some cohesins operate in both mitosis and meiosis.Some mitotic cohesins are replaced by similar butnonidentical proteins in meiosis. For example, SMC1is replaced by SMC1β; Rad21, by Rec8; and Rec11,by STAG3 [1]. It is suggested that Rec8 provides abasis for the formation of axial elements in both mei-otic chromosomes and synaptonemal complexes [2].Cohesins are rather conserved, although a significantsimilarity is observed only in the Rec8/Rad21 func-tional domain [3]. For example, REC-8 of nematodeCaenorhabditis elegans is similar to Rec8p of asco-mycete Saccharomyces cerevisiae by 41.6% (17.3%identity) [3]. According to current evolutionarynotions, meiosis emerged later than mitosis, and it isreasonable to suggest that Rec8 is a modified Rad21.Specific features of cohesin operation in meiosis incomparison with mitosis (binding sites and partnerproteins) can cause differences in protein structureboth within and beyond the Rec8/Rad21 functionaldomain.To verify this suggestion, we studied in silico thefollowing Rec8 proteins: Rec8 from yeast S. cerevi-siae, a Rec8-like protein from mouse Mus musculus,REC8L1 from human (chosen from the homologsfound for the mouse protein by the BLAST program),W02A2.6p from nematode C. elegans (homolog ofRad21/Rec8; called REC-8 [4]), and variant 1 ofSyn1/Dif1 from thale cress Arabidopsis thaliana [5].In addition, we sought homologs of mouse Rec8 in theproteome of zebrafish Danio rerio and found eightproteins. One of them, annotated as a meiosis-specificcohesin, was chosen for the study. For the sake of con-venience, all these proteins are designated as Rec8 inthe figures. In addition, we chose some Rad21 pro-teins: the known mouse and yeast proteins, a Rad21-like human protein, A. thaliana Rad21-1 variant 1(chosen from several proteins annotated as Rad21homologs in the database), the zebrafish protein anno-tated as Rad21, and the hypothetical nematode proteincontaining the Rad21/Rec8 domain.We tested the similarity between the amino acidsequences of Rad21 and Rec8 with the TreeTop pro-gram (Phylogenetic Tree Prediction of the GeneBeepackage, Institute of Physico-Chemical Biology,Moscow State University, http://www2.genebee.msu.su/services/ phtree_reduced.html). The closestsimilarity was observed in the N-terminal fragment of80 amino acid residues and in a slightly shorter C-ter-minal fragment. It is notable that Rad21 from differentspecies proved to be more conserved than Rec8,because the program detected common motifs forsome regions of Rad21. No common motifs werefound for Rec8. The CDART program(http://www.ncbi.nlm.nih.gov/Structure/lexington/lex-ington.cgi?cmd=rps) confirmed the results obtainedfor Rad21 and detected two Rad21/Rec8 domains: alonger one at the N end and a shorter one at the C end.The domain structures of the Rad21 proteins of A.thaliana and D. rerio were quite similar to those ofother proteins of this group. All Rec8 proteins alsohad two Rad21/Rec8 domains except for the proteinsfrom C. elegans and D. rerio, possessing only the N-terminal domain. The fish Rec8 is annotated as a mei-osis-specific cohesin in the NCBI database, and nem-atode REC-8 has been chosen as a meiotic cohesinfrom four similar proteins on the grounds of experi-