The Genom of Homo sapiens.pdf

120 BAILEY AND EICHLERand other investigators have suggested that the perceivedenrichment of duplications within the pericentromeric regionsmay simply be due to the lack of deleterious consequencesthat would select against their fixation (Jacksonet al. 1999; Guy et al. 2000). Other gene-poor regions ofthe genome that are non-pericentromeric and lack duplicationsdo not support this claim (Hattori et al. 2000). Itis likely that reduced selection and increased generationof duplications act in concert to create the large blocks ofduplications within pericentromeric regions.Structure and Dynamics of Subtelomeric RegionsTelomeres are essential for chromosome stability andconsist of a TTAGGG repeat that is created and maintainedby the enzyme telomerase. Between these “stabilizingrepeats” and the unique euchromatic sequences ofthe chromosome arms lie the subtelomeric regions thatconsist of divergent telomeric sequence interspersed withsegmental duplications (Riethman et al. 2001; Meffordand Trask 2002). Like pericentromeric regions, subtelomericregions appear to preferentially duplicate materialto other subtelomeric regions (Monfouilloux et al.1998; Grewal et al. 1999; van Geel et al. 2002). Many ofthese regions, when examined by FISH, differ both betweenspecies and also within the human population(Monfouilloux et al. 1998; Trask et al. 1998a; Park et al.2000), suggesting that gene differences caused by thepresence or absence of these regions could explain phenotypicdifferences both within and between species. Unlikepericentromeric regions, subtelomeric regions aregene-rich, and the duplications often contain expressedgenes (Riethman et al. 2001). Examples of transcribedgenes that have been duplicated include RAB-like genes,TUB4Q members, and myosin light-chain kinase (Brand-Arpon et al. 1999; Wong et al. 1999; van Geel et al.2000). The most striking gene duplications found in thesubtelomeric regions involve the olfactory receptors. Olfactoryreceptors are found throughout the genome (Glusmanet al. 2001), but many appear to have spread recentlythrough subtelomeric segmental duplication (Rouquier etal. 1998; Trask et al. 1998b; Glusman et al. 2001; Meffordet al. 2001). The evolutionary instability of subtelomericregions and their propensity to accumulate segmentalduplications may have been critical in theexpansion of this gene family.Segmental Duplications and Human DiseaseIt has long been recognized that repetitive DNA canmediate misalignment resulting in aberrant homologousrecombination that creates rearrangements—mainlydeletions and duplications. Clinical diseases caused bythis mechanism have been labeled genomic disordersand have been reviewed extensively (Mazzarella andSchlessinger 1997, 1998; Lupski 1998; Ji et al. 2000;Stankiewicz and Lupski 2002). The thalassemias werethe first genomic disorders to be characterized as an eventoccurring as a result of unequal-homologous recombination.For both α and β thalassemia, misalignment and recombinationarise due to the highly similar tandemly repeatedsequence within and between the genes (Kunkel etal. 1969; Phillips et al. 1980). In the case of β-globin, recombinationtakes place within the 5´ situated γ-globinand the 3´ β-globin, resulting in a γ–β deletion fusioncalled Hb Lepore. The corresponding meiotic homolog isa chromosome with 3 genes: γ-globin, β–γ fusion, and β-globin.In addition to tandemly duplicated genes, another classof genomic disorders is due to interspersed duplications.Since the identification of the CMT-1A REPs (Pentao etal. 1992), a growing list of genomic disorders have beencharacterized that involve rearrangements between distantlyspaced, highly similar (>98%) sequences. Such interspersedsegmental duplications are often referred to aslow-copy repeats (LCRs) or duplicons and are also associatedwith nonallelic homologous recombination(NAHR) (Stankiewicz and Lupski 2002) between copies.When NAHR occurs between these distantly spacedcopies, it results in rearrangements not only of the duplicons,but also of the intervening sequence, often resultingin gains and losses of blocks of intervening unique sequence(segmental aneusomy). If these unique regionscontain genes that are triplosensitive, haploinsufficient,or imprinted, disease may occur. Characterized interspersedgenomic disorders now include the 17p11 duplicationof Charcot-Marie-Tooth disease 1A (Reiter et al.1996); the 17p11 deletion of hereditary neuropathy withliability to pressure palsies (HNPP) (Chen et al. 1997);the 22q11 deletion in velocardiofacial syndrome (VCFS),the most common microdeletion syndrome (Morrow etal. 1997; Shaikh et al. 2000); and the 15q11-q13 deletionin Prader-Willi/Angelman syndrome (PW/AS) (Amos-Landgraf et al. 1999). An extensive list of genomic disordersmediated by nonallelic homologous recombinationhas been described previously (Stankiewicz and Lupski2002).From the examination of the sequence properties ofsegmental duplications associated with genetic disease, itbecomes apparent that the size, relative orientation, degreeof sequence identity, and length of intervening sequenceare key factors in determining the frequency andtype of rearrangement. In general, as segmental duplicationsbecome larger and share higher sequence identity,there is an increased propensity for rearrangement (Reiteret al. 1996, 1998; Edelmann et al. 1999; Lopez-Correa etal. 2001). We examined the human genome for all occurrenceswhere duplications had >95% sequence identityand were larger than 10 kb in length, and where the interveningsequence separating two duplications was between50 kb and 5 Mb in length (Bailey et al. 2002b).This analysis identified 163 potential regions of the humangenome that would be predicted to be sites of structuralrearrangement—23 of these regions were alreadyassociated with some form of genomic disease. A detailedassessment, thus, offers considerable clinical valuein the sense that it provides a road map for the identificationand characterization of sites of recurrent structuralrearrangement within the human population associatedwith both genetic disease and large-scale polymorphism.