Conservation And Evolution Of SRF And MEF2with Their Cognate DNA Binding Sites

Serum response factor (SRF) and myocyte enhancer factor2(MEF2) represent two typesof members of the MCM1, AGAMOUS, DEFICIENS, and SRF (MADS)-box transcriptionfactor family present in animals and fungi. Each type has distinct biological functions, whichare reflected by the distinct specificities of the proteins bound to their cognate DNA-bindingsites and activated by their respective cofactors.SRF-and MEF2-type MADS domains bind to A/T-rich DNA sequences in common, anddifferently, SRF binds as a homodimer to CC(A/T)6GG, called SRF site, whereas MEF2bindsas homo-and heterodimer to CTA(A/T)4TAG, called MEF2site. SRF binds a10-bps ciselement of CC(A/T)6GG, which is famous known as CArG box. More than three millionCArG boxes are distributed in the mouse genome, while SRF can discriminate the functionalCArG boxes from millions of functionless ones and bind to the functional to promote geneexpression. MEF2gene subfamily has four members referred to as MEF2A,2B,2C, and2Din vertebrate genomes, and the four members display distinct but overlapping expressionpatterns in body development and maintenance. Accordingly, we mainly investigated thefollowing three topics. First, functional CArG flanking sequences were investigated toelucidate SRF bound to CArG box. Second, conservation and evolution of SRF-andMEF2-type MADS domains with their cognate binding sites were investigated to explore therelationship of cis and/or trans changes with phenotype changes. Finally, the origin,conservation and evolution of the four MEF2genes in vertebrates were investigated todemonstrate MEF2A-D subfunctionalization and neofunctionalization.Part I: By Chi-square test and computer simulation, we investigated the characteristicsof the CArG-SRF binding context.The results showed:1) In addition to the10highly conserved positions of CArG box,some other positions are significantly conserved (e.g.15,8and8).2) Comparing theadjacent positions of consensus CArG and CArG-like boxes reveal more conserved positionsin the latter (e.g.12,7and6).3) A frequency of CpG dinucleotides is much higher within CArG-SRF binding context than that within introns or genome.Conclusion: These results suggest that there are some special pre-existing features in theflanking sequences of functional CArG boxes, probably contributing to SRF selectivelyrecognizing and binding to the functional CArG from millions of functionless CArG boxes inmammalian genomes.Part II: By different methods including phylogenetic analyses, information content (IC)and Z-test, we reported the conservation and evolution of SRF-and MEF2-type MADSdomains with their cognate DNA-binding sites.The results showed:1) There are great similarities with highly conserved positionsbetween the two types of proteins, which are critical for binding to the DNA sequence and forthe maintenance of the3D structure.2) In contrast to MEF2-type MADS domains, distinctconserved residues are present at some positions in SRF-type MADS domains, determiningspecificity and the configuration of the MADS domain bound to DNA sequences.3) Theinferred ancestor sequence of SRF-and MEF2-type MADS domains is more similar toMEF2-type than to SRF-type.4) In the case of DNA-binding sites, MEF2site has a T-richcore in one DNA sequence and an A-rich core in the reverse sequence as compared with SRFsite, no matter whether where either A or T is present in the two complementary sequences.5)Comparing SRF sites in mammals reveals that the evolution rate of CArG boxes is faster inmice than in humans.6) A CArG-like sequence, which is probably functionless, couldpotentially mutate to a functional CArG box that can be bound by SRF and vice versa.Conclusion: These results significantly improve our knowledge on the conservation andevolution of the MADS domains and their binding sites to date and provide new insights toinvestigate the MADS family, which is not only on evolution of MADS factors but also onevolution of their binding sites and even on coevolution of MADS factors with their bindingsites.Part III: By phylogenetic analyses, we investigated the origin, conservation, andevolution of the four MEF2genes in vertebrates.The results showed:1) Among the four MEF2branches, MEF2B is clearly distant fromthe other three branches, mainly because it lacks the HJURP_C (Holliday junction recognitionprotein C-terminal) region.2) Three duplication events have occurred to produce the fourMEF2paralogous genes and the first duplication event occurred before the origin ofvertebrates producing MEF2B and the ancestor of the other three members.3) MEF2Bevolves faster than the other three MEF2proteins despite purifying selection on all of the fourMEF2branches.4) Variable selection exists among MEF2proteins, and positions53and64along MEF2B branch are likely under positive selection.6) Type II MADS genes (i.e. MEF2 type) evolve as slowly as type I MADS genes (i.e. SRF type) in animals, which is inconsistentwith the fact that type II MADS genes evolve much slower than type I MADS genes in plants.Conclusion: Our findings shed light on the relationship of MEF2A,2B,2C and2D withfunctional conservation and evolution in vertebrates, providing a rationale for futureexperimental design to investigate distinct but overlapping regulatory roles of the four MEF2genes in various tissues.Taken together, by various methods of statistics, computer simulation, phylogeneticanalyses, natural selection and comparative genomics, we investigated functional CArGflanking sequences, SRF-and MEF2-type MADS domains with their cognate DNA bindingsites, and the four MEF2A-D genes in vertebrates. The studies provide fundamentalknowledge to reveal the coevolution of MADS factors with their binding sites, and what’smore intriguing and significant is that the studies give us a new insight to investigate othertranscriptional factors as well as their corresponding binding sites.