Gene Conversion During Primate Evolution Yoko Satta, The Graduate University for Advanced Studies, Hayama, Kanagawa, Japan Gene conversion is a nonreciprocal recombination pro- cess; the term originally referred to distorted segregation of alleles in gametocytes. In evolutionary studies, how- ever, gene conversion often means ‘the transfer of deoxyribonucleic acid (DNA) sequence information from one locus to another’. In evolutionary processes, gene conversion is frequently observed between tandemly duplicated sequences or between homologous sequences on the same or on different chromosome(s). Gene con- version between functional loci has three significant roles: (1) Gene conversion generally works to maintain sequence and functional similarity in the ‘coevolution’ of interacting molecules. (2) Gene conversion often takes place between a functional gene and a pseudogene, and such events mainly cause diseases, especially in humans. Rarely, such conversions may confer a novel function to a converted gene. (3) Gene conversion erases advantageous sequence divergence between genes. In these cases, negative selection against for conversion maintains the advantageous divergence, and the converted genes are eliminated from a population. Introduction Definition of gene conversion Gene conversion was originally defined in fungi, such as Neurospora crassa. When gametocytes of an individual that is heterozygous for a pair of alleles undergo meiosis, the expected ratio of segregating alleles is 2:2. However, the ratio is sometimes 3:1, indicating that one allele converted to the other. The mechanism for this conversion is believed to be related to deoxyribonucleic acid (DNA) repair in the heteroduplexes that form during resolution of Holliday structures. Figure 1 illustrates the relationship among het- eroduplexes, recombination and gene conversion. A dou- ble-strand break that initiates recombination can have two different outcomes. One is reciprocal recombination, in which the entire segment extending from the breakpoint to the end of the DNA molecule or to the next breakpoint is exchanged between the two aligned duplexes (‘resolution A’ in Figure 1). Alternatively, gene conversion takes place, and a short segment of the donor DNA molecule is intro- duced into the recipient molecule without the former receiving anything in return from the latter (‘resolution B’ in Figure 1). Depending on how the Holliday structure is resolved, the outcome, either reciprocal recombination or gene conversion, is determined. In evolutionary studies, however, it is not clear whether the molecular mechanism involved is the same as that originally observed in molecular biological studies of alleles at a single locus. See also: Gene Conversion Example of gene conversion in primate evolution The evolution of the globin gene cluster is a well known example of gene conversion in primate evolution (Koop et al. , 1989). The cluster of globin genes is generally composed of five genes; one (HBBP) of these became inactive before the diver- gence between prosimians and euprimates, whereas the remaining four (HBE, HBG, HBD, and HBB) remain func- tional in all primates. As a duplication of HBG took place in the stem lineage of Catarrhini (Old World monkeys and homi- noids), there are now six genes in total including the pseudo- gene. Nucleotide sequence comparisons between a pair of loci reveals that nucleotide substitutions are not evenly distributed along the genes (Figure 2). When the HBD and HBB genes of humans are compared, each gene is divided into four cor- responding regions, and two regions (I and III) are quite similar between HBD and HBB (Figure 2). The nucleotide difference between HBD and HBB genes in regions I and III is approximately 10%, whereas the difference between regions II and IV is 40–60%. The latter probably corres- ponds to the original nucleotide difference since the emer- gence of HBD and HBB, but the former value is similar to the difference between genes from humans and New World Advanced article Article Contents . Introduction . Roles of Gene Conversion in Evolution Online posting date: 13 th June 2013 eLS subject area: Evolution & Diversity of Life How to cite: Satta, Yoko (June 2013) Gene Conversion During Primate Evolution. In: eLS. John Wiley & Sons, Ltd: Chichester. DOI: 10.1002/9780470015902.a0020832.pub2 eLS & 2013, John Wiley & Sons, Ltd. www.els.net 1
Transcript
Gene Conversion DuringPrimate EvolutionYoko Satta, The Graduate University for Advanced Studies, Hayama, Kanagawa, Japan
Gene conversion is a nonreciprocal recombination pro-
cess; the term originally referred to distorted segregation
of alleles in gametocytes. In evolutionary studies, how-
ever, gene conversion often means ‘the transfer of
deoxyribonucleic acid (DNA) sequence information from
one locus to another’. In evolutionary processes, gene
conversion is frequently observed between tandemly
duplicated sequences or between homologous sequences
on the same or on different chromosome(s). Gene con-
version between functional loci has three significant
roles: (1) Gene conversion generally works to maintain
sequence and functional similarity in the ‘coevolution’ of
interacting molecules. (2) Gene conversion often takes
place between a functional gene and a pseudogene, and
such events mainly cause diseases, especially in humans.
Rarely, such conversions may confer a novel function to a
sequence divergence between genes. In these cases,
negative selection against for conversion maintains the
advantageous divergence, and the converted genes are
eliminated from a population.
Introduction
Definition of gene conversion
Gene conversion was originally defined in fungi, such asNeurospora crassa. When gametocytes of an individualthat is heterozygous for a pair of alleles undergo meiosis,the expected ratio of segregating alleles is 2:2.However, theratio is sometimes 3:1, indicating that one allele convertedto the other. The mechanism for this conversion is believedto be related to deoxyribonucleic acid (DNA) repair in the
heteroduplexes that form during resolution of Hollidaystructures. Figure 1 illustrates the relationship among het-eroduplexes, recombination and gene conversion. A dou-ble-strand break that initiates recombination can have twodifferent outcomes. One is reciprocal recombination, inwhich the entire segment extending from the breakpoint tothe end of the DNA molecule or to the next breakpoint isexchanged between the two aligned duplexes (‘resolutionA’ in Figure 1). Alternatively, gene conversion takes place,and a short segment of the donor DNA molecule is intro-duced into the recipient molecule without the formerreceiving anything in return from the latter (‘resolution B’in Figure 1). Depending on how the Holliday structure isresolved, the outcome, either reciprocal recombinationor gene conversion, is determined. In evolutionary studies,however, it is not clear whether the molecular mechanisminvolved is the same as that originally observed inmolecular biological studies of alleles at a single locus.See also: Gene Conversion
Example of gene conversion in primateevolution
The evolution of the globin gene cluster is a well knownexample of gene conversion in primate evolution (Koop et al.,1989). The cluster of globin genes is generally composed of fivegenes; one (HBBP) of these became inactive before the diver-gence between prosimians and euprimates, whereas theremaining four (HBE, HBG, HBD, and HBB) remain func-tional inallprimates.AsaduplicationofHBG tookplace in thestem lineage of Catarrhini (Old World monkeys and homi-noids), there are now six genes in total including the pseudo-gene. Nucleotide sequence comparisons between a pair of locireveals that nucleotide substitutions are not evenly distributedalong the genes (Figure2). When theHBD andHBB genes ofhumans are compared, each gene is divided into four cor-responding regions, and two regions (I and III) are quitesimilar between HBD and HBB (Figure 2). The nucleotidedifference betweenHBDandHBB genes in regions I and IIIis approximately 10%, whereas the difference betweenregions II and IV is 40–60%. The latter probably corres-ponds to the original nucleotide difference since the emer-gence ofHBD andHBB, but the former value is similar tothe difference between genes from humans andNewWorld
Advanced article
Article Contents
. Introduction
. Roles of Gene Conversion in Evolution
Online posting date: 13th June 2013
eLS subject area: Evolution & Diversity of Life
How to cite:Satta, Yoko (June 2013) Gene Conversion During Primate Evolution.In: eLS. John Wiley & Sons, Ltd: Chichester.
monkeys. To explain this finding, the idea of a ‘gene con-version’ event before the divergence of humans and NewWorld monkeys is invoked. However, the molecularmechanism and biological significance of this conversionevent have not been fully investigated.
Roles of Gene Conversion in Evolution
To maintain homogeneity among sequences
Gene conversion and concerted evolution
Duplicated genes are generally divergent because of pointmutations. However, in some gene families, such as his-tones or ribosomal ribonucleic acids (rRNAs), the reten-tion of functional homogeneity among members in thefamily is necessary. The process of this homogenisation iscalled ‘concerted evolution’, and themechanismmediatingthis process is believed to be an ‘unequal crossing-over’ or‘gene conversion’. Different from the conversion in theglobin gene cluster, the unit of conversion (conversiontract) in these events is an entire gene. However, recentstudies demonstrate that this functional homogenisationdoes not always require complete nucleic acid sequencehomogeneity, and that the functional constraint againstthe amino acid replacement is sufficient to retain the samefunction among members within these families (Nei and
Rooney, 2005). Members in such gene families con-sequently show high similarity at the amino acid sequencelevel, but a high extent of nucleotide divergence at syn-onymous sites. See also: Concerted Evolution
Sialic acid-recognising receptors
One biological role of gene conversion is thought to be aninvolvement in ‘coevolution’ between interacting mol-ecules. Evolution of two sialic acid-recognising receptors,SIGLEC14 and SIGLEC5, is such an example; the desig-nation SIGLEC is from sialic acid immunoglobulin (Ig)-like lectin. SIGLEC14 and SIGLEC5 are immune cellsurface receptors, and both recognise glycan chains thatcontain sialic acids. In great apes, 13 functional SIGLECShave been identified, andmost of them are located in a genecluster; in humans they are located on 19q13.3–13.4.Immune cell surface receptors with similar ligand binding,but counteracting signal properties, are called ‘pairedreceptors’, namely inhibitory and activating receptors, andthey are believed to play an important role in the fine-tuning of immune responses. SIGLEC14 and SIGLEC5are one such pair of receptors. SIGLEC14 comprises threeIg-like domains, a single transmembrane (TM)domain anda short cytoplasmic tail; SIGLEC5 comprises four Ig-likedomains, a single TM domain and a long cytoplasmic tail.Of these domains, the first three Ig-like domains of thesetwomolecules in humans are almost identical, with a single
(1) Double-strand break
Resolution B
(3) Heteroduplex
Resolution A
Resolution A
Recombination
Resolution B
Gene conversion
(2) Holiday junction
Figure 1 The relationship between recombination and gene conversion. Double DNA strands are depicted as double arrows. The orientation of each arrow
indicates a 5’–3’ orientation. When a double-strand break occurs, a Holliday junction is produced to repair this break. For further repair, the junction moves
and produces a heteroduplex. When the heteroduplex is resolved, and depending on the point of the DNA cut (a red arrow at ‘resolution A’ or ‘resolution
B’), the outcome is either recombination or gene conversion.
amino acid difference in 199 amino acids. The nucleotidesequence similarity of exons encoding these domains isextended into the adjacent introns. Phylogenetic studies ofthese sequences revealed that similar gene conversionevents took place independently in each species of greatapes (chimpanzees, gorillas and orangutans) and in theOldWorldmonkeys, suggesting the importance ofmaintainingsequence similarity between the Ig-domains of the tworeceptors and consequently retaining a similar extent ofligand binding by both activating and inhibitory receptors(Angata et al., 2006). See also: CD Antigens
Chemokine receptors
Another example of gene conversion that maintains func-tional similarities is that between two chemokine receptors,CCR2 and CCR5 (Vazquez-Salat et al., 2007). Differentfrom frequent conversion between siglecs in each primatespecies, the conversion betweenCCR2 andCCR5 seems tohave taken place in the early stage of primate evolution.This conversion event in the primate lineage becomesapparent when othermammalian orthologues are includedin a phylogenetic analysis.CCR2 and CCR5 are both chemokine receptors and 7
TM G-protein-coupled receptors (GPCR) and areexpressed mainly in leucocytes. Within primates, the 7 TMdomains and the respective intracellular loops of CCR2and of CCR5 have sequence similarity, but the similarity isnot evident between a pair of the corresponding ortholo-gues within other mammalian lineages. Contrary to this,sequences from the N- to C-terminus and extracellularloops reveal the monophyletic (orthologous) relationshipof each gene family. CCR2 and CCR5 sometimes form aheterodimer, producing synergistic effects that can activate
a cellular response in the presence of a much reducedconcentration of their ligands. Because this hetero-dimerisation requires that molecules share a high degree ofsimilarity, it is plausible that the gene conversion betweenCCR2 and CCR5 maintains a similar TM domain. How-ever, because the N- and C-terminus and extracellularloops are important in ligand binding and signal trans-duction, these functions must be kept independently of thegene conversion event. See also: Chemokines and Chemo-kine Receptors
Conversion between a functional gene and arespective pseudogene
Disease-causing gene conversions
Bischof et al. (2006) searched for potential gene con-versions between functional genes and the respectivepseudogenes and found that among the 14 476 pseudo-genes (both processed and nonprocessed) in humans, 1945are localised near their parental gene, suggesting theirpossible involvement in gene conversion. Eleven knowncases of disease-causing gene conversions were identified.However, a separate search of the OMIM databaserevealed that multidrug resistance-associated protein 6(MRP6) did not show any evidence of gene conversionresulting in disease (pseudoxanthoma elasticum). Theremaining 10 genes have been reanalysed with a focus onevolutionary aspects, such as genetic distances betweenfunctional genes and pseudogenes, the physical distancesbetween them, estimated time of pseudogenisation and thepresence of orthologues in chimpanzees and rhesusmacaques (Table 1).
Figure 2 Sliding window analysis of the number of nucleotide differences (p-distances) between the HBD and HBB. The sliding window size is 100 bp, and
the sliding interval is 10 bp. Depending on the nucleotide differences, the entire gene is divided into four regions. In the schematic below the graph under
the X-axis, boxes and lines represent the positions of exons and introns, respectively.
Autosomal dominant polycystic kidney disease is relatedto at least three loci, two of which are polycystic kidneydisease (PKD) 1 on chromosome 16 and PKD2 onchromosome 4. Mutations in PKD1 are the most commoncause of the disease. One such mutation results in multipleamino acid substitutions in exon 23. Watnick et al. (1998)showed that thesemutated nucleotideswere identical to thecorresponding nucleotides in a PKD1 duplicate. Inhumans, there are six PKD1 duplicates (PKD1P1–PKD1P6) in the 13–16Mb upstream of PKD1, and each isa truncated pseudogene. A schematic diagram of the geneorganisation in this region is shown in Figure 3. On the basisof the orientation of genes, the similar physical distances,and the similar structures among these pseudogenes, itseems likely that each is derived from the same singlepseudogene, and not functional PKD1.Chimpanzees show a different pattern of the gene
organisation (Figure 3). In humans, all six pseudogenes lack14 exons at the 3’ end; in contrast, chimpanzees possessthree duplicates, and they appear to lack more than 10exons at both the 5’ and 3’ ends. The orthologous rela-tionship between human and chimpanzee genes is notdefinite; nevertheless, the chimpanzee pseudogenes aretentatively namedPKD1P1–P3. A phylogenetic tree that isbased on neutral sites (substitutions at synonymous sitesand those in introns) is shown (Figure 3). The alignablesegments betweenhumanand chimpanzee pseudogenes arequite limited; therefore, the number of compared neutralsites is only 1486 bp (1202 bp for introns and 284 bp forsynonymous sites). Both introns and synonymous sitesshow that six PKD1 pseudogenes are classified into threegroups: P1/P3/P5, P2/P4 and P6. Compared with therelatedness between genes within each group, the diver-gence between groups is large. The nucleotide diversitybetween groupsP1/P3/P5 andP2/P4 is approximately 2%and that between these two groups and P6 is 3.6%. Fur-thermore, human functional PKD1 is again relatively dis-tantly related to each of the six pseudogenes: the divergenceranges from 3.6% to 4.7%. The relationship betweenPKD1 and its pseudogenes in chimpanzees differs from thatin humans: The three chimpanzee pseudogenes are classi-fied into two groups,P1/P3 andP2, and the sequence ofP2is more closely related to human and chimpanzee PKD1than to other human and chimpanzee pseudogenes. Thesephylogenetic studies show that gene conversion betweenthe neighbouring pseudogenes does not always occur fre-quently in humans.
Gene conversion between differentchromosomes
Gene conversion more likely occurs between homologousgenes on the same chromosome than those on differentchromosomes. The case of the von Willebrand factor(VWF) gene is an example of the latter case. The humanVWF gene is located on chromosome 12, and theT
pseudogene (VWFP) is located on chromosome 22.VWF isa causal gene for type III vonWillebrand disease, which is ahaemophilia caused by low antihaemophilic globulin(factor VIII). The VWF gene encodes a glycoprotein thatfunctions as an antihaemophilic factor and a platelet-vesselwallmediator in blood coagulation. There are several typesof this disease, and type III is the most severe form. In typeIII, two closely situatedmutations in exon 28,G1263A andC1266T, are related to the disease. Because these twomutated nucleotides are exactly the same as themutation inVWFP, they are possibly introduced by gene conversionbetween VWF and VWFP. If this conversion hypothesis istrue, this represents quite an exceptional case of conversionbetween different chromosomes. However, because bothmutations are observed at CpGdinucleotide sites, at whichthe mutation rate is 410 times higher than that at non-CpG sites, the possibility of independent mutations inVWF cannot be ruled out. From the nucleotide divergencebetween VWF and VWFP (3.4%), the presence of VWFPin chimpanzees is expected.However, in the present version
of the chimpanzee database (build 2.1), any sequence that ishomologous to VWFP in the chimpanzee genome cannotbe detected. Because the location of humanVWFP is closeto the centromere of the chromosome and this regionmightbe unstable, the chimpanzeeVWFPmayhave been deleted.
Other disease-causing gene conversions onautosomes
In agammaglobulinaemia, patients lack B cell develop-ment, and the causal gene is immunoglobulin lambda-likepolypeptide 1 (IGLL1). Each patient carries four pointmutations in a 64-bp region in exon 3, and two of themutations cause amino acid replacements. Four nucleotidemutations in a limited region are unlikely to be due to de-novo point mutations, although two of them (at positions359 and 425) are at CpG dinucleotide sites. Because themutated nucleotides are exactly the same as those in one ofthe two nearby pseudogenes, these mutations are likelyintroduced into IGLL1 by gene conversion from the
100100
HosaPKD1P1HosaPKD1P5
HosaPKD1
HosaPKD1P5
HosaPKD1P3
HosaPKD1P1HosaPKD1P2
HosaPKD1P4
HosaPKD1P4
HosaPKD1
HosaPKD1P3HosaPKD1P2HosaPKD1P4
HosaPKD1P698
67
94
PatrPKD1P1PatrPKD1P3
PatrPKD1P2
PatrPKD1P1
PatrPKD1P3
PatrPKD1P2
PatrPKD1
PatrPKD1
78
6799
CLD02
6481
99
86
47
84
8449
0.005
PKD1
Human
Chimpanzee
P6
PKD1P3
PKD1P2
PKD1P1 PKD1P3
P1 P2
P4 P5
PKD1
PKD intron 1202 bp
PKD synonymous 284 bp(exon 25−30)
Figure 3 Gene organisation of PKD1 and the PKD1 pseudogene clusters in humans and chimpanzees. The length of an arrow roughly represents the size of
the respective gene. The orientation of an arrow indicates the orientation of a gene from the 5’–3’ direction. Phylogenies are constructed using the
Neighbour-Joining Method based on p-distances. A scale bar for each tree is indicated at the bottom of each tree.
Gene Conversion During Primate Evolution
eLS & 2013, John Wiley & Sons, Ltd. www.els.net 5
pseudogene. In the human genome, there are two pseudo-genes near the functional IGLL1: one includes only exon 3and is 58 kb upstream of IGLL1: the other includes exons 2and 3 and is 1.8 kb downstream of the gene. However, boththe pseudogenes are too diverged from IGLL1, and alsodiffer from IGLL1 in the four mutations; therefore, thesepseudogenes cannot be the donor in the conversion.However, the pseudogene in the 1.8Mb downstream ofIGLL1 possesses the four mutations, suggesting that this isthe donor in the disease-causing gene conversion. Theextent of nucleotide divergence between the recipient(IGLL1) and the donor (the pseudogene) is 5%, andchimpanzees possess the pseudogenes. However, thechimpanzee has duplicates of the downstream pseudogene.As expected from the divergence between the functionalgene and thepseudogene, the orthologue of the pseudogenecannot be detected in the macaque genome.From the viewpoint of the genetic distance between the
recipient and the donor in disease-causing gene conversion,a high extent of similarity is expected between them.However, the genetic distance between the recipient and thedonor in the cases of b-crystalline 2 (CRYBB2) and folatereceptor 1 (FOLR1) is not small. A pseudogene ofCRYBB2, designated CRYBP1, is located in a 0.2Mbregion proximal to CRYBB2, lacks exons 1 and 2, and is90% homologous with CRYBB2. The homology is not sohigh as in the cases of steroid 21-hydroxylase (CYP21A2:98%), iduronate 2-sulfatase (IDS: 99%) or neutrophilcytosolic factor 1 (NCF1: 99.5%). The genetic distancebetween FOLR1 andFOLR1P is also large (approximately14%). Dominant cataracts result from a clinically andgenetically heterogeneous group of disorders that causeblindness. These disorders are considered to be related tomore than 13 loci. One of the causal genes isCRYBB2, andpatients have twomutations, C475T andC483T, in exon 6;these changes are identical to those inCRYBP1. When thephylogenetical relationship between the orthologues inhumans and chimpanzees are examined, each pair oforthologues forma cluster, suggesting that gene conversionhas not been frequent between CRYBB2 and CRYBP1 inboth lineages after the divergence of humans and chim-panzees. A disorder caused by the gene conversion betweenFOLR1 andFOLR1P is a neural tube defect. Patients showa replacement of a segment (position 7497–7662) inFOLR1 with the corresponding FOLR1P sequence;FOLR1P is located in an approximately 30 kb regionproximal to FOLR1. This replacement introduced 3 aminoacid changes into exon 7 as well as a replacement into the 3’untranslated region (UTR) that leads to messenger RNA(mRNA) instability. As expected from the relatively largenucleotide divergence of 14%, the emergence of FOLR1Pgoes back to before the divergence of Old World monkeysand hominoids, and FOLR1Pmight be present in the NewWorldmonkey genomes. However, when the phylogeny oforthologues in humans, chimpanzees and macaques areexamined, frequent conversions between FOLR1 andFOLR1P are again not observed in their evolutionarycourses.
Gene conversion between CYP21A2 and CYP21AP is awell known example of disease-causing conversion.CYP21A2 encodes the adrenal steroid 21-hydroxylase,which is a member of the cytochrome P-450 supergenefamily and plays a crucial role in the synthesis of steroidhormones. Congenital adrenal hyperplasia (CAH) is amonogenic metabolic disorder, and the most commoncause of CAH is a defect in 21-hydroxylase. In humans,CYP21A2 and CYP21AP are located on chromosome 6 inan approximately 1Mb class III region within the majorhistocompatibility complex (MHC) loci region, and theyare approximately 30 kb apart from each other. Donohoueet al. (1989) proposed that gene conversion was a cause ofthe 21-hydroxylase deficiency based on a Southern blotanalysis ofCYP21A2 andCYP21AP.When the phylogenyof CYP21A2 and CYP21AP in humans and chimpanzeesare examined, the physical position of these genes inchimpanzees shows clear synteny with the position inhumans, suggesting their orthologous relationship. How-ever,when a tree based on the nucleotide differences amongthese sequences is constructed, the result reveals that twoconspecific paralogues are more closely related to oneanother than to homologues in other species. The presenceof two cyp genes in the mouse genome has been reported,suggesting that the emergence ofCYP21APmight go backto the emergence of mammals. However, due to frequentand independent gene conversions in human and mouselineages, the origin cannot be traced from a sequencecomparison.Gaucher disease is one of the most common glycolipid
storage diseases. It is caused by an inherited deficiency ofglucosidase beta acid (GBA). The deficiency is reported to bedue to four different mutations. These four changes areG476A, A1226G, C1361G and T1448C. All four changesmatchexactly the sequence found in theadjacentpseudogene,GBAP, which is located in the 16kb region proximal toGBA.However, despite the observation of gene conversion as acause of the disease, the conversion cannot be observed in anevolutionary time scale. When the phylogenetic relationshipof GBA and GBAP in humans, chimpanzees and macaquesare examined, orthologous sequences form a single cluster.Neutrophil cytosolic factor 1 (NCF1) is a component
of the nicotinamide adenine dinucleotide phosphate(NADPH) oxidase complex in phagocytes. The phagocyteNADPH oxidase plays an important role in the hostdefence system by releasing a large amount of superoxide.The predominant genetic defect causing NCF1-deficientchronic granulomatous disease is a guanine–thymine (GT)deletion at the beginning of exon 2. The high incidence ofthis mutation is reported in an unrelated, ethnically diversepopulation. One of the reasons for this high incidence isgene conversion betweenNCF1 andNCF1P, which are onchromosome 7 and approximately 400 kb apart from oneanother. In humans, another pseudogene (NCF1B1) res-ides approximately 1.5Mb proximal to NCF1, but it doesnot possess the deletion. Thus, in humans, there are threecopies of NCF1 in total (two pseudogenes and one func-tional gene), whereas in chimpanzees, only a functional
Gene Conversion During Primate Evolution
eLS & 2013, John Wiley & Sons, Ltd. www.els.net6
gene (NCF1) is detected. The twopseudogenes appear to behuman specific, and therefore, this disease is human spe-cific. The small nucleotide divergence (0.5%) betweenNCF1 and NCF1P in humans also supports this.Shwachman–Diamond syndrome is an autosomal
recessive disorder and is characterised by exocrine pan-creatic insufficiency and bone marrow dysfunction.Mutations that cause this disorder in the SBDS locus arereported in several ethnic groups. For example, mostEuropean patients have mutations within an approxi-mately 240-bp region of exon 2, whereas Japanese patientshave mutations at various sites between intron 1 and exon3. In both cases, most mutations are likely introduced bygene conversion with a pseudogene, SBDSP, which islocated in a 5.8Mb region distal to functional SBDS. Inchimpanzees, however, two additional segmental dupli-cations that include exons 1 and 2 are observed nearSBDS;these are not present in the humangenome.As suggestedbyphylogenetic studies, the sequence similarity betweenSBDS andSBDSP in both humans and chimpanzees is notextremely high (3.2%); consequently, each orthologouspair of genes forms a cluster in a phylogenetic tree.
Conversion between homologues on X andY chromosomes
Sex chromosomes in mammals were originally a pair ofautosomes that diverged into sex chromosomes through agradual or stepwise suppression of recombination betweengametologues on chromosomes X and Y. Depending onwhen recombination between X and Y ceased, the diver-gence of gametologous genes between X and Y differs(Lahn and Page, 1999). One of the most recent divergedsegments is called ‘strata 4’, and this segment shows a 10%nucleotide difference on average.However, in this segment,VCX and VCY seem to undergo frequent exchangesbetween X and Y chromosomes. The nucleotide differenceat neutral sites (synonymous and intron sites) betweenVCX andVCY is 2% in chimpanzees and 3.7% in humans,which is less than the apparent 10% for other X–Yhomologues in this segment. This pattern is shown in aphylogenetic tree ofVCY andVCX from chimpanzees andhumans (Bhowmick et al., 2007) and suggests that geneconversion across sex chromosomes occurred after thedivergence between humans and chimpanzees. KALX andY are also involved in this conversion (Iwase et al., 2010).BecauseKALY is a pseudogene in primates, the conversionfrom KALY to KALX leads to a deficiency of function,and sometimes becomes a cause of Kalmann syndromein humans. See also: Mammalian Sex ChromosomeEvolution
Conversion between genes on the Xchromosome
Hunter syndrome is a sex-linked mucopolysaccharidosis.Structural alterations or gross deletions of the IDS gene aredetected in many clinically severe Hunter syndrome
patients. A second IDS locus (IDSP) is reported to belocated within 90 kb proximal to the IDS gene, and somepatients show recombination between intron 7 of IDS andsequences close to exon 3 of IDSP. IDSP has a largedeletion, and is composed of sequences corresponding toexons 2 and 3 as well as introns 2, 3 and 7 of IDS. In aphylogenetic analysis, IDS and IDSP form a cluster inhumans and one in chimpanzees; moreover, the nucleotidedivergence between conspecific IDS and IDSP is almostnull and 1.3% at the synonymous sites between humansand chimpanzees. Taken together, these data indicate arecent gene conversion within each species.
Gene conversion to add novel function to aconverted gene
Itmightnotbe frequent, butanexampleof a conversion toaddnovel function to a converted gene has been reported (Haya-kawaetal., 2005;Wangetal., 2012).Theexample is conversionbetween SIGLEC11 and SIGLEC16 in hominoids.The ancestralSIGLEC11 is observed in rhesusmonkeys,
dogs and cows, suggesting that SIGLEC11 existed in thestem lineage of eutherianmammals.SIGLEC11 duplicatedand produced SIGLEC16 in a hominoid common ancestorapproximately 20 Mya (Hayakawa et al., 2005). Sub-sequently, 3Mya in the human lineage,SIGLEC16 becamea pseudogene, SIGLEC16P, via a 4-bp deletion in exon 2(Hayakawa et al., 2005). After this pseudogenisation, anapproximately 2 kb region in the 5’ part that contains thefirst five exons of SIGLEC16P converted the homologousregion of SIGLEC11 (Hayakawa et al., 2005). Interest-ingly, although the causal deletion for the pseudogenisa-tion in SIGLEC16P is located in the putative conversiontract, the mutation was not present in SIGLEC11 (Wanget al., 2012). The 170-bp region encompassing themutationdoes not seem to have been converted. A detailed analysisof the conversion tract reveals that the conversion occurredindependently in the 5’ (706 bp) and 3’ (1226 bp) parts of the170-bp region (Wang et al., 2012). Interestingly, the 3’region of the gene (intron 5 and its downstream) exhibits4–5% nucleotide diversity between SIGLEC11 andSIGLEC16P, suggesting that no gene conversion hasoccurred since its emergence (Wang et al., 2012). Whenchimpanzee siglec11 and siglec16 are compared, the geneticdistance in the approximately 2 kb 5’ region between thesegenes is 2.2%, suggesting that an additional and inde-pendent gene conversion in the 5’ region occurred in theancestor of humans and chimpanzees.Furthermore, the human population survey demon-
strated the presence of a functional SIGLEC16 at a fre-quency of 0.17. The presence of homozygotes andheterozygotes of SIGLEC16 and SIGLEC16P means thatSIGLEC16P is not fixed in the extant human population(Cao et al., 2008). In fact, the phylogenetic analysis ofhuman SIGLEC11, SIGLEC16 and SIGLEC16P revealsthat in addition to conversion from SIGLEC16P toSIGLEC11, SIGLEC16 received a SIGLEC11 segmentfrom the 5’ part that encompassed exons 1–3 and included
the 170-bp fragment (Figure 4). However, the frequency ofsuch conversions in other hominoids has not been wellexamined; therefore, a detailed analysis of the sequence
surrounding the conversion tract in humans and chim-panzees may be necessary to examine the molecular causeof this frequent conversion.
SIGLEC11
SIGLEC16
SIGLEC16P
SIGLEC16
SIGLEC16P
SIGLEC16
SIGLEC11
SIGLEC11
SIGLEC16P
1.0~1.2 Mya
< 1.0 Mya
Present
Figure 4 Proposed scenario of gene conversion between the human SIGLEC11 and SIGLEC16P/16 genes. SIGLEC11 and SIGLEC16P/16 are positioned in a
head-to-head orientation with an approximately 9 kb interval. The arrowhead on SIGLEC16P represents a 4-bp deletion. The black, grey and white rectangles
indicate the exons in the SIGLEC11, SIGLEC16P and SIGLEC16 genes, respectively. The position of ‘exon 1’ is located at the right-most for SIGLEC11 or the
left-most for both SIGLEC16P and SIGLEC16 in the figure. Reprinted by permission of Oxford University Press from Wang et al. (2012) Figure 9. & Oxford
University Press.
0.0
0 20 000
3′/3A2A2B
A3A6
psA2′/2 1′/1
40 000 60 000
0.04
0.08
Per
site
nuc
leot
ide
dive
rgen
ce
0.12
Figure 5 Nucleotide sequence divergences in a palindrome on Xq28. The divergence (ordinate) was examined in a window of 500 bp that did not overlap.
The position (abscissa) is relative to the middle of the loop of the palindrome (indicated by a blue arrow). Coloured rectangles at the bottom of the figure
indicate the duplicated sequences including the MAGE genes (light pink arrows). The area within the red, dotted line indicates the highly diverged region in
MAGE-A3 and -A6. Reproduced from Katsura and Satta (2011). & PloS.
Gene Conversion During Primate Evolution
eLS & 2013, John Wiley & Sons, Ltd. www.els.net8
The conversion from SIGLEC16P to SIGLEC11 inhumans might lead to a novel SIGLEC11 expression pat-tern (Hayakawa et al., 2005; Wang et al., 2012). The con-version tract inSIGLEC16P contains aUTR in exon 1 thatpossesses several translational regulatory elements.SIGLEC11 is expressed in brain microglia specifically inhumans, but not in chimpanzees or orangutans. Althougha role forSIGLEC11 in the humanbrain is not documentedyet, this human specificity could be the result of gene con-version from SIGLEC16P.
Negative selection on gene conversion
MAGE-A genes on the X chromosome
Sometimes, negative selection against gene conversion canwork to maintain gene function, as shown in the casebetween SIGLEC11 and SIGLEC16P (Hayakawa et al.,2005; Wang et al., 2012).Another example of such negative selection is observed
in the melanoma antigen gene (MAGE) family on theX chromosome. The MAGE gene family comprisesapproximately 10 subfamilies. Most ancestors of eachsubfamily arose in a stem lineage of placental mammals,and then extensive gene duplication within each subfamilyoccurred in anorder-specificway (Katsura andSatta, 2011)within the placental mammals. In particular, theMAGE-Asubfamilies in the genome formed a palindrome in theprimate lineage. The MAGE-A subfamily comprises 16genes in humans that are located in a 5Mb region on Xq28and are clustered into three blocks. These three blocks areseparated from each other by an approximately 1Mbnoncoding or intervening sequence and are tentativelycalled A, B and C, from the proximal to the distal regions.Block B contains 10 genes, of which seven form a palin-drome structure. Of the seven, three pairs of genes – psAand psAL, A3 and A6, A2 and A2B – are on arms of thepalindrome, and a single gene, A12, is on a loop. Asexpected, genes on the arms experienced frequent geneexchanges between pairs (Katsura and Satta, 2011).However,when the sequencesof the armsare compared, a
peak in the nucleotide divergence was found in the paircomprising A3 and A6 (Figure 5; Katsura and Satta, 2011).Excluding the region containing the A3 and A6 pair, thenucleotide divergence is almost null, but this pair shows anapproximately 2% divergence. Negative selection againsthomogenisation, or against erasing the accumulateddivergence, must take place. MAGE-A, especially A3 andA6, expressed exclusively in cancer cells and testis, is asource of peptides bound to human leucocyte antigens(HLA, human MHC); possibly, this binding might be atrigger for cancer immunity. Because HLA shows thelargest extent of polymorphism in the human genome dueto balancing selection andHLA is located on chromosome6, the maintenance of genetic diversity in A3 and A6 isimportant (Katsura and Satta, 2011) to maintain peptidebinding for HLA. Actually, many peptides derived from
MAGE-A3 andA6 can bind to several HLA from differentalleles.
Physical distance and sequence similaritybetween donors and recipients
It is expected that homologous gene pairs with closephysical proximity and with a high extent of nucleotidesequence similarity have greater chances of gene con-version. However, based on evidence from earlier cases,this does not always hold true. For example, the conversionbetween different chromosomes is observed between VCYand VCX or between VWF and VWFP. The genetic dis-tances between donors and recipients also vary greatly,from 0.5% to 14%, suggesting that high similarity, such asis the case for allelic differences (less than 0.1% in humans),is not a necessary condition. Moreover, the gene con-version in a human population, especially disease-causingconversion, is not always detected in an evolutionary timescale. Conversion observable in both a population and aspecies is only evident at CYP21A and IDS; in these cases,donors and recipients from a species are more closelyrelated to homologues in other species. For other cases,orthologous relationships between each of the donor andrecipient genes from humans and chimpanzees areobserved. As, in general, a small segment is involved inmany conversions, conversions related to diseases cannotbe detected by phylogenetic analysis. An alternativeexplanation is that functional genes converted by pseudo-genes cannot survive in a population due to disease. If thelatter holds true, a reduction of nucleotide diversity com-pared with other neutral regions might be observed insegments containing functional genes due to negativeselection (background selection).
References
Angata T, Hayakawa T, Yamanaka M, Varki A and Nakamura
M (2006) Discovery of Siglec-14, a novel sialic acid receptor
undergoing concerted evolution with Siglec-5 in primates.
FASEB Journal 20: 1964–1973.
Bhowmick BK, Satta Y and Takahata N (2007) The origin and
evolution of human ampliconic gene families and ampliconic
structure. Genome Research 17: 441–450.
Birot AM, Bouton O, Froissart R, Maire I and Bozon D (1996)
IDS gene–pseudogene exchange responsible for an intragenic
deletion in a Hunter patient. Human Mutation 8: 44–50.
Bischof JM, Chiang AP, Scheetz TE et al. (2006) Genome-wide
identification of pseudogenes capable of disease-causing gene
conversion. Human Mutation 27: 545–552.
Bogdanova N, Markoff A, Gerke V et al. (2001) Homologues to
the fist gene for autosomal dominant polycystic kidney disease
are pseudogenes. Genomics 74: 333–341.
Cao H, Lakner U, de Bono B et al. (2008) Siglec 16 encodes a
DAP12-associated receptor expressed in macrophages that
evolved from its inhibitory counterpart Siglec 11 and has
functional and non-functional alleles in humans. European
Journal of Immunology 38: 2303–2315.
Gene Conversion During Primate Evolution
eLS & 2013, John Wiley & Sons, Ltd. www.els.net 9
De Marco P, Moroni A, Merello E et al. (2001) Folate pathway
gene alterations in patients with neural tube defects. American
Journal of Medical Genetics 95: 216–223.
Donohoue PA, Jospe N, Migeon CJ and Dop CV (1989) Two
distinct areas of unequal crossing over within the steroid 21-
hydroxylase genes produce absence of CYP21B. Genomics 5:
397–406.
GorlachA,Lee PL,Roesler J et al. (1997)Ap47-phox pseudogene
carries the most common mutation causing p47-phox-deficient
