![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
|
Vol. 10, Issue 5, 597-610, May 2000
REVIEW
|
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 ARTICLE
REFERENCES
|
|---|
Chromosome-specific low-copy repeats, or duplicons, occur in multiple regions of the human genome. Homologous recombination between different duplicon copies leads to chromosomal rearrangements, such as deletions, duplications, inversions, and inverted duplications, depending on the orientation of the recombining duplicons. When such rearrangements cause dosage imbalance of a developmentally important gene(s), genetic diseases now termed genomic disorders result, at a frequency of 0.7-1/1000 births. Duplicons can have simple or very complex structures, with variation in copy number from 2 to >10 repeats, and each varying in size from a few kilobases in length to hundreds of kilobases. Analysis of the different duplicons involved in human genomic disorders identifies features that may predispose to recombination, including large size and high sequence identity between the recombining copies, putative recombination promoting features, and the presence of multiple genes/pseudogenes that may include genes expressed in germ cells. Most of the chromosome rearrangements involve duplicons near pericentromeric regions, which may relate to the propensity of such regions to accumulate duplicons. Detailed analyses of the structure, polymorphic variation, and mechanisms of recombination in genomic disorders, as well as the evolutionary origin of various duplicons will further our understanding of the structure, function, and fluidity of the human genome.
| |
ARTICLE |
|---|
|
TOP
ABSTRACT
ARTICLE
REFERENCES
|
|---|
Although the primary cause of many genetic
diseases is single gene defects, a number of syndromes have been
defined as genomic disorders (Lupski 1998
), as the majority of cases
arise from structural rearrangements of certain regions of a
chromosome. In such cases, disease is due to dosage imbalance of one or
more genes, caused by chromosomal structural change and not by mutation
of the gene(s) itself. Rather than the disease etiology, this review
focuses on the structural mechanisms of chromosome rearrangements.
Structural changes of certain chromosomal segments were first
associated with several human genetic disorders in the early 1980s,
such as Prader-Willi syndrome at 15q11-q13 (Ledbetter et al. 1981) and
DiGeorge syndrome at 22q11 (de la Chapelle et al 1981). Different terms
(e.g., contiguous gene, microdeletion, contiguous deletion, and
segmental aneusomy) were given to describe these and other syndromes
later associated with a specific structural change of a chromosome,
including the more apt term, genomic disorder (Lupski 1998). Whereas
the term genomic disorder (Lupski 1998) includes all syndromes
involving a change at the genomic level, the following discussion is
restricted to genomic disorders arising by homologous recombination
between region-specific, low-copy repeats during meiosis. Recent
studies show that duplications of small chromosome segments may not be
a rare event. Analysis of long stretches of available sequence from
human chromosomes 22 and X suggest that 5%-10% of the genome may be
duplicated (Mazzarella and Schlessinger 1998; Dunham et al. 1999).
Sequence homology between duplicated DNA segments provides a chance for
misalignment during meiosis, leading to unequal exchange and chromosome
rearrangement, by either inter- or intrachromosomal or sister chromatid
homologous recombination (Fig. 1). This process with
divergence between duplicated segments is essential to the generation
of diversity and new genes over evolutionary time, although the more
typical, short-term effect is genetic disease.
|
Region-specific, low-copy repeat elements, or duplicons (Eichler 1998),
have been identified flanking many regions susceptible to human genomic
disorders (for recent reviews, see Purandare and Patel 1997; Lupski
1998; Mazzarella and Schlessinger 1998). These disorders arise from
chromosomal rearrangements mediated by homologous recombination between
different copies of a duplicon with deletions (Fig. 1a,b), duplications
(Fig. 1a), or inversions (Fig. 1c), the outcome depending on unknown
factors, although the orientation of duplicons is clearly one important
criterion. Characterization of various genomic disorders showed that
they can be classified into three groups. In the first group, the
length of the repeated segments and the distance between the repeat
units are both small. The duplicon in such cases is either a gene
cluster or repeated sequences flanking a gene. Homologous recombination between the duplicons generates an imbalance of a dosage-sensitive gene, leading to a disease phenotype. This group includes a number of
disorders such as red-green color blindness and 
- or
-thalassemia (Lupski 1998), and will not be discussed further
here. The other two groups of genomic disorders involve rearrangement
of a large segment of the chromosome, yet the flanking duplicons are
structurally simple in the second group, whereas the third group of
disorders involve complex duplicons. In both cases, if the
rearrangement is a deletion or duplication, the expression level of
multiple genes may be affected, although the phenotypes can arise from dosage imbalance of a single gene. The affected gene(s) may or may not
be part of the duplicon, and can thus be situated quite far from the repeats.
Simple Duplicons Mediating Genomic Disorders
S232 Elements and X-Linked Ichthyosis
Deletions of the steroid sulfatase (STS) gene in males cause a skin disorder, X-linked ichthyosis. The incidence of this disorder is 1 in 2-5000 males, with ~90% of the patients having a deletion of 1.9 Mb that includes the STS locus. The deletions are mediated by homologous recombination between duplicons called S232 elements, situated at either end of the deletion (Ballabio et al 1990;
Yen et al. 1990). There are 6-14 S232 duplicons, 4-12 in Xp22.3 and 2 on the Y chromosome, and each is composed of a 5-kb unique sequence and
two VNTR (variable-number tandem repeat) sequences (Li et al. 1992;
Lahn and Page 2000). One mapped S232 duplicon lies distal to the
STS locus and three proximal to STS, with
recombination proposed to occur between the former and only one of the
latter in tandem (Fig. 2a; Li et al 1992). No cases
involve recombination with the Y chromosome (Yen et al 1990). Recent
studies show that a copy of the VCX gene is present within the
S232 duplicon, and that the VCX gene family is expressed
exclusively in male germ cells (Lahn and Page 2000).
|
CMT1A-REP and Two Distinct Peripheral Neuropathies
Reciprocal chromosome rearrangements of 17p12 occur in Charcot-Marie-Tooth syndrome type 1A (CMT1A) and hereditary neuropathy with liability to pressure palsies (HNPP) (Chance and Fischbeck 1994;
Patel and Lupski 1994). A 1.5-Mb tandem duplication of 17p12 occurs in
70% of inherited and 90% of sporadic CMT1A cases (Hoogendijk et al.
1992; Wise et al. 1993; Nelis et al. 1996), in which 10%-18% of
CMT1A cases overall are de novo (Blair et al. 1996; Bort et al. 1997).
A deletion of the same chromosomal segment occurs in 82% of familial
and 86% of sporadic HNPP cases (Nelis et al. 1996). The protein
encoded by the PMP22, or peripheral myelin protein-22, gene in
the 1.5-Mb region appears to be sensitive to subtle dosage changes and
is responsible for both syndromes (for review, see Chance and Fischbeck
1994; Patel and Lupski 1994).
The 1.5-Mb region is flanked by a duplicon termed CMT1A-REP (Fig. 2b;
Pentao et al. 1992; Chance et al. 1993), with 98.7% sequence identity
across 24,011 bp (Reiter et al. 1997). The COX10 gene encoding
heme A:farnesyltransferase overlaps the distal duplicon, and a
duplicated internal segment of COX10 forms the proximal CMT1A-REP (Reiter et al. 1997). Homologous recombination between these
repeat elements lead to the duplications in CMT1A and deletions in HNPP
(Lupski 1998). In both CMT1A and HNPP rearrangements, the majority of
crossovers happen within a single recombination hot spot (Kiyosawa et
al. 1995; Kiyosawa and Chance 1996; Lopes et al. 1996, 1998; Reiter et
al. 1996, 1998). Because all crossovers occur in long stretches of
identity, the minimum efficient processing segment (MEPS) for efficient
meiotic homologous recombination was proposed to be between 337 and 456 bp long. This is similar to other observations of an ~200 bp MEPS
for homologous recombination in mammalian cells (Waldman and Liskay
1988). The finding of a mariner transposon-like element with
its 3' end 1774 bp from the hot spot led to a hypothesis that it
could mediate strand exchange events via double-strand breaks induced
by a transposase at or near the 3' end of the element (Reiter et
al. 1996, 1998). Multiple sequences similar to human minisatellite
consensus sequences [4 copies similar to 5'
-GGGCAGGA(A/G)G-3', and one copy similar to 5' -AGAGGTGGG
CAGGTGG-3'] and a 
-like sequence (5'
-GCTGGTGG-3') were also identified in a 270-bp region within the
hot spot (Lopes et al. 1998), and suggested to promote homologous
recombination events. However, no direct evidence supports the involvement in recombination of a mariner or other specific sequences.
Interestingly, most de novo CMT1A duplications contain both paternal
haplotypes and therefore arose from unequal crossing over involving
both homologs (Palau et al 1993; Lopes et al., 1997,1998). In contrast,
all four maternal cases involving CMT1A-REP recombination (two
duplications and two deletions) are intrachromosomal events (Lopes et
al. 1997), which suggests a mechanistic difference during
spermatogenesis and oogenesis for 17p12 rearrangements. Using the
mariner-like sequence as a probe, a 2.2-kb transcript was
claimed to be present in testis but not in ovary (Kiyosawa and Chance
1996), although the genomic source(s) of the putative transcripts are
unknown and the data is not convincing. Nevertheless, an open chromatin
structure due to transcription of COX10 or the mariner could be associated with generation of CMT1A
duplications during spermatogenesis.
int22h Elements and Severe Hemophilia A
Hemophilia A is an X-linked coagulation disorder affecting ~1 in 5000 males (Soucie et al. 1998), due to mutations of the factor VIII
gene (F8). About 45% of severe hemophilia A males carry a
large (~500 kb) inversion on Xq, which disrupts F8 (Lakich et al. 1993; Naylor et al. 1993). Most inversions are intrachromosomal (Lakich et al. 1993) and originate almost exclusively in male germ
cells (Rossiter et al. 1994). The inversion occurs by homologous recombination between 9.5-kb duplicons (int22h, or intron 22 homologous region), one in intron 22 of F8, and either of two
homologs oriented in the opposite direction and located 500 kb more
distal (Fig. 2c; Naylor et al. 1995). In one case, the intron 22 copy
recombined with a fourth, truncated copy of int22h, to
generate a smaller inversion (Naylor et al. 1996). An 8-kb segment of
int22h is 99.9% similar among the three copies, with a 1.8-kb
intronless gene (F8A) within each of the repeat elements
(Levinson et al. 1990), although it is not known whether F8A
is expressed during spermatogenesis.
Germ-line Deletions of 17q11.2 in Neurofibromatosis Type 1
Recent studies indicate that a minority (2%-13%) of neurofibromatosis type l (NF1) cases result from an NF1 gene deletion of 1.5 Mb that is mediated by homologous recombination between duplicons termed NF1REPs (Dorschner et al. 2000). These duplicons are
direct repeats that span 15-100 kb and contain several
genes/pseudogenes (a third copy maps distally in 17q24). Of the
NF1 deletions that were de novo, 75% were maternal in origin,
although somatic mosaicism for NF1 deletions have been
described suggesting that mitotic recombination between NF1REPs may
also occur (Dorschner et al. 2000).
Genomic Disorders Arising from Chromosome Rearrangements Involving Complex Duplicons
Recent molecular and cytological studies have identified large, complex duplicons in the breakpoint regions of several human genomic disorders. The rearranged DNA segments usually span several megabases and are flanked by large duplicons that often contain several genes and/or pseudogenes. Multiple copies are present both at the common breakpoint regions and elsewhere within the same subchromosomal region. On the basis of limited analyses, sequence identity between different duplicon copies are usually >90% and up to 99%. As the duplicons are large in size, have high homology, and multiple copies, the characterization of these duplicons has proven to be difficult. With the recent development of long-range mapping and sequencing techniques, the nature of some of these rearrangements has begun to emerge.
Chromosome 7q11.23 Deletions and Williams syndrome
Williams syndrome (WS) is a multisystem developmental disorder affecting connective tissue and the central nervous system, with an estimated incidence of 1 in 20,000 live births (Morris et al. 1993;
Sadler et al. 1993). Most (95%) patients have a de novo ~1.6-Mb
deletion of chromosome 7q11.23, including the elastin (ELN) gene (Ewart et al. 1993; Nickerson et al. 1995), with
common deletion breakpoints (Perez Jurado et al. 1996; Robinson et al. 1996). The deletions arise equally often from both paternal and maternal meiosis, although a potential role for genomic imprinting in
clinical features of WS (Perez Jurado et al. 1996) remains controversial (Francke 1999; Wang et al. 1999).
Duplicons of 200-300 kb that flank the common WS deletion (Fig.
3a; Francke 1999) each contain up to three
transcripts, including at least three duplicons that have sequences
related to the PMS2 mismatch repair gene (Osborne et al.
1997). Two of the duplicons contain copies of the transcription factor
GTF2I (Perez Jurado et al. 1998), in which the distal copy is
the ancestral GTF2I gene with a unique 5' end within the
common deletions, whereas the proximal copy represents a truncated,
expressed pseudogene (GTF2IP1). The neutrophil cytosolic
factor 1 (NCF1) gene maps telomeric to GTF2I in the
distal duplicon, whereas an NCF1 pseudogene is present at both
the proximal and the distal duplicons (Hockenhull et al. 1999). A
shared >3-Mb junction NotI fragment in WS deletions indicates clustering of breakpoints (Perez Jurado et al. 1998), suggesting that homologous recombination between the duplicons is
responsible for generating the WS deletion, with an interchromosomal event in most cases (Dutly and Schinzel 1996; Baumer et al. 1998).
|
Chromosome 17p11.2 Deletions (SMS) and Duplications
Smith-Magenis syndrome (SMS) is a mental retardation/multiple congenital anomalies syndrome that occurs in 1/25,000 births (Greenberg et al. 1991). Most patients (>90%) have an ~5-Mb common deletion
of 17p11.2 (Juyal et al. 1996; Chen et al. 1997; Lupski 1998). A
cluster of four genes/ pseudogenes (SRP, TRE,
KER, CLP) spanning >200 kb was found to be
duplicated and was mapped to the common SMS deletion breakpoint regions
(Chen et al. 1997), with a third copy mapping within the 5-Mb deleted
region (Fig. 3b). The gene CLP (for coactosin-like
protein) maps only to the proximal (P) and distal (D) breakpoint
regions (SMS-REPP and D), whereas the other three genes also map to
the middle duplicon (SMS-REPM). Additional copies of CLP
occur in chromosomes 15, 16, and Y, but are uncharacterized. A novel
NotI junction fragment of 1.2 Mb is identified, using a
CLP cDNA probe, in every SMS patient with the common deletion,
suggesting recombination between the outer SMS-REP duplicons (Chen et
al. 1997). No chromosome rearrangement involving the middle copy of
SMS-REP has been identified to date. Interestingly, a 210 bp sequence
in the 3' UTR of CLP has >95% identity to a human cDNA
sequence with autonomously replicating activity (Wu et al. 1993).
Initiation of DNA replication from the latter element, or transcription
through the CLP region, may be sufficient to unwind DNA and
render the region prone to homologous recombination (Chen et al. 1997).
Several 17p11.2 duplication patients have been identified, with mild
mental retardation and minor dysmorphism (Chen et al 1997; Potocki et
al 2000). Using the CLP cDNA probe, a novel ~1.1-Mb NotI junction fragment was detected in the duplication
patients. This result and other molecular studies suggested that the
17p11.2 duplications involve the SMS-REP and represent reciprocal
interchromosomal recombination events to those in SMS (Chen et al.
1997; Potocki et al. 2000). The 17p11.2 duplications were de novo and
preferentially paternal in origin (Potocki et al. 2000).
Low Copy Repeats and Chromosome 22q11 Rearrangements
Several congenital anomaly disorders are associated with de novo germ-line rearrangements of 22q11 that involve duplicon-mediated homologous recombination (Edelmann et al. 1999 a,b,c; Shaikh et al.
2000). DiGeorge and velocardiofacial syndromes (VCFS) are the most
common microdeletion syndrome, with an incidence of 1 in 4000 births
(Wilson et al. 1993). The cardiac and craniofacial features of VCFS may
result from haploinsufficiency for the UFD1L gene (Yamagishi
et al. 1999), although this remains controversial (Baldini 1999).
Cat-eye syndrome (CES) is a rare multiple anomalies disorder (Schinzel
et al. 1981) associated with a dicentric bisatellited supernumerary
chromosome 22 (pter>q11.2::q11.2->pter). Derivative 22 [der(22)] syndrome is a rare mental retardation/multiple anomaly disorder, due to partial trisomy of chromosome 22q11 by unbalanced segregation of a translocation (Funke et al. 1999; Edelmann et al. 1999c).
About 90% of VCFS deletion patients have a similar deletion of 3 Mb,
whereas 7% of patients have a 1.5-Mb deletion with the same proximal
breakpoint region as in the 3-Mb deletions (Carlson et al. 1997;
Edelmann et al. 1999b; Shaikh et al. 2000). A low-copy repeat (LCR22)
ranging from 40 to 350 kb and 97% to 98% identity occurs in the
vicinity of all three 22q11 breakpoint regions, as well as at six
adjacent locations over a 6.5-Mb region (Fig. 3c; Dunham et al. 1999;
Shaikh et al. 2000). Each LCR22 duplicon differs in content and
organization of complex sequence modules (Dunham et al. 1999; Shaikh et
al. 2000), including variable content and arrangements of 11 genes/
pseudogenes (GGT, BCRL, V7-Rel, HMPLPL, GGT-REL, E2F6L, ISG43,
KIAA0049, DGCR6, PRODH, NF1L). Several potential recombination sequences within the LCR22 duplicons include VNTR sequences and AT-rich repeats (Edelmann et al. 1999c; Shaikh et al. 2000). Most recombination events appear to occur within
the LCR22 repeats (Edelmann et al. 1999a b c; Funke et al. 1999; Shaikh
et al. 2000). There is no bias for the parental origin of 22q11
deletions leading to VCFS, with both inter- and intra-chromosomal
events observed (Baumer et al. 1998; Edelmann et al. 1999a).
Interestingly, misalignment of different LCR22s during meiosis may
generate both deletions and reciprocal duplications; an interstitial
duplication of the same 3-Mb DNA segment as in the VCFS deletion has
been identified in one family (Edelmann et al. 1999b). Two CES
breakpoint clusters also map to the LCR22 duplicons associated with the
proximal and distal breakpoint regions of the common VCFS deletions
(McTaggart et al. 1998). Der (22) chromosomes break within the nested
middle copy of LCR22 (Edelmann et al. 1999b, c; Funke et al. 1999).
Other LCR22 duplicons may also be used in generating rare chromosomal
rearrangements, including somatic rearrangements that lead to malignant
diseases (Edelmann et al. 1999b). The identification and sequence
analysis of exact breakpoints in patients with these diverse
rearrangements of 22q11.2 will shed significant light on the molecular
events that predispose to recombination within complex modular duplicons.
END Repeats and Chromosome 15q11-15q13 Deletions
Proximal chromosome 15q is also involved in an astonishing variety of cytogenetic aberrations with a high frequency. Prader-Willi and Angelman syndromes (PWS and AS) occur at a frequency of ~1 in 15, 000 births, and most cases are due to an ~4-Mb deletion of 15q11-q13 of paternal or maternal origin, respectively (Nicholls et al. 1998; Jiang et al. 1999). PWS and AS are models for genomic imprinting in human (Nicholls et al. 1998), although this is not considered further here. The vast majority (>95%) of the deletion patients have clustered deletion breakpoint regions, with a single distal breakpoint and two proximal breakpoint regions (Christian et al.
1995; Amos-Landgraf et al. 1999). There is no apparent mechanistic
differences for paternal and maternal deletions, as both inter- and
intra-chromosomal rearrangements are seen in PWS and AS patients
(Carrozzo et al. 1997; Robinson et al. 1998).
Recent molecular and cytogenetic analyses demonstrated at least 10 copies of a complex duplicon, termed the END repeat, that maps
to the three common deletion breakpoint regions, with two chromosome
16p11 copies (Fig. 3d; Buiting et al. 1992, 1998; Amos-Landgraf et al.
1999; Ji et al. 1999). The END repeats are composed to a large
extent of duplications of a huge gene, HERC2, although multiple rearrangements within the derived duplicons have occurred during evolution (Amos-Landgraf et al. 1999; Ji et al. 1999, 2000). The
ancestral HERC2 gene encodes a giant protein, recessive
mutations of which cause a neurological and developmental syndrome in
mouse (Ji et al. 1999). A simple model to explain the generation of PWS
and AS common deletions proposes homologous misalignment and meiotic
recombination between different END repeat copies in proximal and distal 15q11-q13 (Amos-Landgraf et al. 1999). Active transcription of HERC2 and related pseudogenes in male and female germ-line tissues suggest that they are transcribed in germ cells (Amos-Landgraf et al. 1999), which is supported by jdf2 mouse phenotypic
studies (Ji et al. 1999), and thus the resultant open chromatin
structure may further stimulate recombination. Identification of a
putative expressed breakpoint junction (using a HERC2 probe)
in one PWS patient of only five tested supports this model
(Amos-Landgraf et al. 1999).
Christian et al. (1999) independently identified duplicons that map to
the breakpoint regions of PWS/AS common deletions, by STS content
mapping of genomic clones at or near the proximal and distal PWS/AS
breakpoints. Several STSs and ESTs were mapped to duplicons at
breakpoints 2 and 3 (BP2 and BP3). Interphase FISH using a PAC clone
containing duplicon sequences identified ~3 signals in 15q11-q13
from normal chromosomes, and a reduced number of signals in deletion
chromosomes (as well as signals on 16p11 and 15q24; Christian et al.
1999). However, the difficulty in distinguishing closely mapping,
highly homologous, and partial (truncated) sequences resulted in an
underestimation of the number of copies of the duplicon.
One end of the END-repeat duplicons is within the 3'
part of HERC2 (Ji et al. 1999). The HERC2-related
content of different duplicons is 36.6 kb in the most
rearranged/deleted copy in 16p11.2 to an estimated ~150-200 kb in
some of the chromosome 15 duplicons (Ji et al. 2000). Nevertheless, the
end point of the duplicon 5' of HERC2 is unknown.
Additional genes are present in the duplicons, including the small
MYLE gene (Christian et al. 1999), although only four of the
seven potential transcripts identified may represent new
genes/pseudogenes in the END repeats (discussed in Ji et al. 2000). Furthermore, there is evidence that independent duplicons flank
the HERC2 (and MYLE)-containing duplicons in 15q11.2
(Buiting et al. 1999). Additional experiments, particularly complete
genomic sequence, are needed to further characterize the content,
orientation, and copy number of the END repeat duplicons that
may be involved in the mechanism for PWS/AS common deletions.
Other 15q11-15q13 Rearrangements
The most common of these involve marker chromosomes. FISH studies indicate that ~57% of all marker chromosomes in newborns originate from chromosome 15, with most an inv dup(15), that occur in ~ 0.4/1000 births (Blennow et al. 1995). Inv dup(15) chromosomes are
bisatellited chromosomes derived from chromosome 15, which contain two
centromeric regions, although usually only one centromere is active
(Schwartz and Depinet 1996). The inv dup(15) is a mirror image
chromosome that is duplicated about a central axis (e.g., pter
cenq13cenpter). It is a misnomer to call it an inv dup(15), and preferred terms would be either a derivative or dicentric chromosome; however, it is normally referred to as an inv dup(15) because of historical context. These chromosomes come in a variety of
size classes. Small marker chromosomes either have only one centromere
or one of the two proximal breakpoints associated with PWS and
AS; that is, for type 1, a breakpoint between the centromere (D15Z1) and markers located between BP1 and BP2, or for type
2, a breakpoint within the BP2 region (Cheng et al. 1994; Huang et al.
1997). Larger marker chromosomes can have variable breakpoints (Mignon
et al. 1996), although most either share the distal PWS/AS breakpoint
(the BP3 region) or have a breakpoint distal of this and within YAC
810f11 (Huang et al. 1997; Wandstrat et al. 1998). Because the vast
majority of the larger markers share this latter breakpoint, and given
the high frequency of these, the findings suggest the possible presence
of another duplicon or recombination-promoting sequence. Interestingly,
the larger inv dup(15) are all de novo and exclusively maternal in origin.
Three phenotypes associated with the presence of a supernumerary marker
chromosome derived from chromosome 15 are as follows: normal, the inv
dup(15) syndrome, or PWS/AS. The inv dup(15) syndrome refers to a
consistent phenotype of mental and developmental retardation, hypotonia, behavior disturbances, seizures, abnormal dermatoglyphics, and mild somatic abnormalities (Wisniewski et al. 1979; Leana-Cox et
al. 1994). It is associated with the presence of the PWS/AS critical
region as seen in the larger markers described above (Leana-Cox et al.
1994). The smaller markers (type 1 and type 2) are usually associated
with a normal phenotype and may be familial, although some have been
associated with a PWS or AS phenotype (Robinson et al. 1993). The
latter are due to uniparental disomy of the normal chromosomes 15 (Robinson et al. 1993).
Although less common, direct duplications of proximal 15q11-15q13 have
been observed (Pettigrew et al. 1987; Clayton-Smith et al. 1993a;
Browne et al. 1997; Cook et al. 1997, 1998; Mohandas et al. 1999), some
with the same breakpoints as PWS/AS deletions, whereas others have a
more distal breakpoint similar to the large inv dup(15) (Repetto et al.
1998; Robinson et al. 1998). Patients with maternal duplications that
include at least the PWS/AS critical region show developmental delay,
or autism (Browne et al. 1997; Cook et al. 1997, 1998; Repetto et al.
1998), whereas a few individuals with paternal duplications have a
nonspecific developmental delay (Mohandas et al. 1999). Finally, a few
triplication cases have been reported (Schinzel et al. 1994; Cassidy et
al. 1996; Long et al. 1998; Robinson et al. 1998), as well as
paracentric inversions of 15q11-15q13 (Clayton-Smith et al. 1993b). It
remains unclear whether these rearrangements are also mediated by
homologous recombination of different copies of the END repeat
duplicon or whether several different duplicons exist in 15q and are
involved in the various rearrangements.
Other Examples of Genomic Disorders and Chromosomal Duplicons
Beside the examples discussed above, several other diseases are also
caused by common DNA rearrangements. Proximal spinal muscular atrophy
(SMA), a recessive-lethal condition affecting 1/10,000 births, is
associated in ~18% of cases with rearrangements in an inverted
duplication of ~500 kb in 5q13 (Melki et al. 1994; Lefebvre et al.
1995; Scharf et al. 1998). A second example is the recessive Hunter
syndrome, due to mutations of the iduronate-2-sulfatase (IDS)
gene in Xq28. Twenty percent of Hunter syndrome cases involve an
inversion that disrupts the IDS gene, caused by homologous recombination between IDS and an IDS pseudogene 90 kb
downstream (Bondeson et al. 1995). Although less well characterized,
most cases of juvenile nephronophthisis (NPH) involve deletion of a 250-kb region on chromosome 2q13, flanked by repeat sequences (Konrad
et al. 1996). Lastly, similar chromosome rearrangement events that are
benign also occur in the genome. In studying a rare condition
(Emery-Dreifuss muscular dystrophy, Xq28), Small et al (1997)
identified a common 48-kb inversion that is present in the heterozygous
state in 33% of females, arising from homologous recombination between
flanking 11.3 kb duplicons of >99% sequence identity. As the human
genome sequence is completed, the structure and role in chromosome
rearrangements of these and other region-specific duplicons, including
those currently unrecognized, will likely emerge.
Transchromosomal Duplicons
In addition to the chromosome-specific duplicons described above, a
large number of genomic segments have been identified in the last few
years that map to multiple non-homologous chromosomes (termed
transchromosomal duplicons) (Borden et al. 1990; Wong et al. 1990;
Buiting et al. 1992; Bernardi et al. 1993; Tomlinson et al. 1994;
Eichler et al. 1996, 1997, 1999; Frippiat et al. 1997;
Kehrer-Sawatzki et al. 1997; Regnier et al. 1997; Zimonjic et al. 1997;
Eichler, 1998; Trask et al. 1998a,b; Brand-Arpon et al. 1999; Grewal et
al. 1999; Horvath et al. 2000). Two regions of the human genome that
seem unusually enriched for such transchromosomal duplication events
are the pericentromeric (defined as the first Giemsa light or dark band
flanking the primary constriction) and subtelomeric portions of
chromosomes (Eichler 1998; Trask et al. 1998a). Detailed analyses of
pericentromeric DNA in 10q11, 16p11, and 2p11 (Jackson et al. 1999a;
Loftus et al. 1999; Horvath et al. 2000; J. Horvath, S. Schwartz, E.E.
Eichler, unpubl.), for example, have identified zones of genomic
duplicons located distal to -satellite DNA. These regions are
composed almost entirely of genic/pseudogenic segments that have been
transposed from diverse regions of the hominoid genome. As one example,
analysis of a 225-kb pericentromeric BAC sequence from chromosome 16p11
identified four different duplicons, including portions of the
creatine-transporter, HERC2, variable 
, and variable
heavy chain immunoglobulin genes (Loftus et al. 1999). All of these
segments have been duplicated and transposed to 16p11 independently
from their ancestral loci in Xq28, 15q13, 2p11, and 14q32,
respectively. Comparative mapping and phylogenetic analysis of various
pericentromeric segments reveals that they began to transpose within
the hominoid genome relatively recently [estimates range from 3 to 20 million years ago (mya)]. Consequently, many of the paralogous
segments share a high degree of sequence similarity (~93%-98%).
Furthermore, differences in map position and copy number have been
observed among closely related primates (chimpanzee, human, and
gorilla). Population studies of subtelomeric duplications have
identified differences among different human ethnic groups (Trask et
al. 1998a), suggesting that dynamic spread of these segments may be an
ongoing process in the evolution of the human genome.
The molecular mechanism underlying the duplication and transposition of
genomic segments to the pericentromeric or subtelomeric regions of
chromosomes is not well understood. However, analysis of sequences
located at the presumptive sites of insertion within the
pericentromeric regions have uncovered several unusual classes of
GC-rich repeats located in close (<1 kb) proximity to the boundaries of paralogy (Borden et al. 1990; Eichler et al., 1996, 1997; Eichler 1999). One class of these repeats (CAGGG; Fig. 4)
localizes specifically to the pericentromeric regions of several
different human autosomes, with an evolutionary origin predating the
arrival of the duplicons themselves (Eichler et al. 1999). Moreover,
the sequences bear some similarity both in sequence and organization to
-recombination elements and immunoglobulin switch-recombination
signal sequences, which are capable of mediating recombination events
among non-allelic sequences. These repeats may play a role in
pericentromeric duplications, by serving as a preferred transposition
integration signal for the duplicons. It is interesting that similar
GC-rich elements have also been mapped within the vicinity of recent
subtelomeric duplications (Amann et al. 1996; Chute et al. 1997). In
addition to these initial transposition seeding events within
pericentromeric DNA, it is now clear that secondary events transfer
larger segments of the duplicons among other pericentromeric regions
(Fig. 4)a process that has been termed pericentromeric swapping
(Horvath et al. 2000; J. Horvath, S. Schwartz, and E.E. Eichler,
unpubl.). The net effect of these duplication events is the generation
of large (>100 kb) blocks of paralogy among multiple nonhomologous chromosomes, each of which is a mosaic of smaller duplicated segments (Fig. 4).
|
Is there a link between the dynamic duplicative nature of
pericentromeric DNA and genomic duplicons mediating genomic disease? A
direct connection is yet to be established, but there are some tantalizing associations. (1) CAGGG repeats have been mapped to several
cytogenetic bands associated with recurrent chromosomal structural
rearrangement and genomic disease (particularly within the proximal
regions of 15q11 and 22q11) (Eichler et al. 1999). (2) There is a
significant positional bias for duplicons and genomic diseases, as a
group, to be near the pericentromeric region (e.g., 7q11.23, 15q11.2,
17p11.2, 17q11.2, and 22q11.2). Three of the most common rearrangements
discussed above (SMS, PWS/AS, and VCFS), which involve complex
duplicons, involve rearrangement between a duplicon situated near a
proximal pericentromeric region and a duplicon located more distally.
(3) Analyses of pericentromeric DNA on chromosome 10 (Jackson et al.
1999a,b), chromosome 22 (Dunham et al. 1999; Shaikh et al. 2000) and
chromosome 16 (E.E. Eichler, unpubl.) suggest that intrachromosomal
(chromosome-specific) duplicons become more prevalent as one approaches
unique DNA. Furthermore, analysis of DNA from regions of
pericentromeric DNA from 16p11.2 and 15q11 have identified
HERC2 duplicons, derived from an ancestral locus in 15q13, as
one component in a series of duplications that constitute 16p11 (Fig.
4; Amos-Landgraf et al. 1999; Ji et al 1999, 2000; Loftus et al. 1999).
These data suggest that the processes responsible for transchromosomal
and intrachromosomal pericentromeric duplication and for
disease-causing rearrangements may be molecularly interrelated. One
scenario may be that pericentromeric DNA has created a favorable
evolutionary milieu for the accumulation and transfer of duplicated
segments. Recurrent chromosomal structural rearrangement associated
with disease then arises due to the fact that both donor and acceptor
recombining loci flank unique gene regions that are haplo-insufficient.
Of course, not all duplicon-mediated rearrangements can be described as
pericentromeric, nor is the process of recent genomic duplication
restricted to pericentromeric and subtelomeric DNA (DeSilva et al.
1999; Loftus et al. 1999). The process of pericentromeric duplication,
therefore, is likely one of several different mechanisms capable of
creating paralogous sequences that can mediate genomic instability.
Conclusions and Perspectives
From studies discussed above and by others (Purandare and Patel
1997; Mazzarella and Schlessinger 1998; Lupski 1998), it is clear that
the presence of duplicons predisposes to abnormal pairing and unequal
crossing-over, leading to chromosome rearrangements in specific regions
of the human genome. Combined, at least 0.7-1 per 1000 births will
have a genomic disorder due to a de novo rearrangement between
duplicons, attesting to the impact on medical care and economic burden.
Deletions, duplications, inversions, translocations, and supernumerary
marker chromosomes can all result from duplicon-mediated homologous
recombination, suggesting this disease mechanism is a general
phenomenon. The identification of novel junction fragments in most
patients with CMT1A/HNPP, WS, SMS, duplication 17p11.2, and some
22q11.2 rearrangements, supports the idea that precise recombination
events are associated with rearrangements involving even complex
duplicons. It is likely that double-strand breaks might be the
initiating event for these homologous recombination events (Lupski
1998), although it will be necessary to isolate the breakpoints from
multiple patients in each class of disorders to more accurately model
the mechanism of recombination.
Three general features of these duplicons can be proposed. (1)
Large-size and high-sequence identity are important characteristics. The duplicons span at least thousands of base pairs to several hundred
kilobases. The identity between the two recombining duplicons can be as
high as 99.9% (int22h). Recombination hot spots are usually
in regions of identical nucleotide sequence, of at least 200 to ~450
bp in length, although this has not yet been examined in complex
duplicon rearrangements. When genomic disorders are examined according
to the distance between the flanking duplicons, it appears that the
length of the duplicon increases as the distance between the duplicons
increases (Lupski 1998). Large duplicons with high sequence identity
may be necessary to initiate pairing and/or stabilize the recombination
complex, for unequal crossing-over and recombination to occur. (2)
Recombination promoting features may be present within duplicons. Most
characterized duplicons contain one or more genes/pseudogenes. Two of
the loci (VCX and HERC2) are expressed in germ cells,
although it is unknown whether these or genes in other duplicons are
transcribed in meiotic or post-meiotic germ cells, in which unequal
crossover must occur. In yeast, double-strand breaks have been
found primarily in promoter regions, suggesting a role for an open
chromatin structure in homologous recombination (Osman and
Subramani 1998). The presence of a sequence in SMS-REP that is
homologous to one with autonomously replicating activity also suggests
a role for open chromatin structure, and/or DNA replication. Sequences
related to those that can promote recombination have been identified in
recombination hot spots in CMT1A-REP and 22q11 rearrangements, and
VNTRs are found within each copy of the S232 and LCR22 repeats, as well
as within the END repeats (Ji et al. 2000). (3) In cases in
which more than two duplicons exist in the same chromosome region,
there are preferences as to which copies are involved in specific
rearrangements. Further studies are needed to understand why some
copies are chosen over others, but this may relate to sequence
identity, orientation of the duplicons, presence of recombination
promoting sequences, or chromatin structure.
Although much has been learned about the mechanisms for genomic
disorders, many questions remain. (1) Some rearrangements occur almost
exclusively in paternal meiosis (CMT1A duplication and F8 inversion),
whereas others happen equally often in male and female meiosis, despite
normal meiotic recombination being twice as frequent in females as in
males. It has been suggested that there are male factors for, or female
protection against, aberrant homologous recombination (Purandare and
Patel 1997). One such factor may be the continued proliferation of
spermatogonia after puberty, which may lead to a much greater
likelihood of a structural rearrangement eventually occurring in sperm
(Chandley 1991). (2) Most cases of genomic disorders are deletions,
whereas duplications and inversions seem to be less common. This may
arise from the mechanisms producing inter- or intra-chromosomal events between direct repeats (Fig. 1). A bias of ascertainment may also contribute, because deletions usually cause a more severe phenotype. (3) It is not known why most of these chromosome rearrangements involve
duplications near pericentromeric regions (Eichler 1998), but as
discussed above this may relate to the propensity for duplicons to
cluster in these regions and for duplicons to generally disperse to
adjacent chromosomal positions. Also not known is how duplicons several
megabases apart search for and recombine with each other, nor what
role higher order chromosome structures play in the process. (4) The
origin and dispersal of various duplicons appears to be limited to
primates. Although the CMT1A-REP seems to be a human- and
chimpanzee-specific repeat (Kiyosawa and Chance 1996; Keller et al.
1999), preliminary studies on the END repeats suggests an
origin of ~15-20 mya (Christian et al. 1999; Ji et al. 1999), and
the LCR22 duplicons arose at least 20-25 mya on the basis of their
presence in Old World monkeys (Shaikh et al. 2000). Thus, the age of
the duplicons appears to correlate with their complexity and the
frequency of chromosome rearrangements. A comparative genomic study of
the various duplicons first identified in the human genome will
determine their evolutionary origin and provide insight into the
mechanisms of formation and rearrangement of these duplicons.
Finally, it is possible that environmental factors may contribute to
the chromosome rearrangement events that lead to genomic disorders.
Preliminary studies suggest a link between paternal hydrocarbon
exposure and PWS (Strakowski and Butler 1987; Cassidy et al. 1989;
Åkefeldt et al. 1995), but other genomic disorders have not been
examined. Several studies show an increased rate of mitotic homologous
recombination between duplicated DNA segments. X-rays and other
carcinogens were shown to significantly increase the reversion
frequency of the pink-eyed unstable (pun) mutation,
a 70-kb internal tandem duplication of p (Schiestl et al.
1994, 1997). Reversion of mutations caused by tandem duplication have also been found in yeast, Chinese hamster and human cell lines,
and were drastically increased by carcinogen treatments (Schiestl et
al. 1988; Zhang and Jenssen 1992; Aubrecht et al., 1995). These events
are all mediated by intrachromosomal recombination. It will be
necessary and important to perform molecular and epidemiological studies to determine whether environmental factors (such as
hydrocarbons) increase germ-line chromosome rearrangements that
lead to the various genomic disorders in human. Given the remarkably
high frequency of genomic disorders, as a group, if environmental risk factors were recognized it would allow proper measures to be taken to
prevent preconceptional exposure, and thereby would significantly reduce the frequency as well as the economic and social costs of
genomic disorders.
| |
ACKNOWLEDGMENTS |
|---|
We thank James Amos-Landgraf for the original art of Figure 1. Work on duplicons in the author's laboratories is funded by a Neuromuscular Disease Research Program Grant from the Muscular Dystrophy Association (R.D.N.), by NIH grant GM58815-01 (E.E.E.), and by a Basil O'Connor Scholar award from the March of Dimes Birth Defects Foundation to E.E.E. (FY98-0712).
The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.
| |
FOOTNOTES |
|---|
1 Present address: Genentech, Inc., Department of Bioinformatics, South San Francisco, California 94080 USA.
2 Corresponding author.
E-MAIL rxn19{at}po.cwru.edu; FAX (216) 368-3432.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
ARTICLE
REFERENCES
|
|---|
