Using a pericentromeric interspersed repeat to recapitulate the phylogeny and expansion of human centromeric segmental duplications

Article (PDF Available)inMolecular Biology and Evolution 20(9):1463-79 · October 2003with55 Reads
DOI: 10.1093/molbev/msg158 · Source: PubMed
Despite considerable advances in sequencing of the human genome over the past few years, the organization and evolution of human pericentromeric regions have been difficult to resolve. This is due, in part, to the presence of large, complex blocks of duplicated genomic sequence at the boundary between centromeric satellite and unique euchromatic DNA. Here, we report the identification and characterization of an approximately 49-kb repeat sequence that exists in more than 40 copies within the human genome. This repeat is specific to highly duplicated pericentromeric regions with multiple copies distributed in an interspersed fashion among a subset of human chromosomes. Using this interspersed repeat (termed PIR4) as a marker of pericentromeric DNA, we recovered and sequence-tagged 3 Mb of pericentromeric DNA from a variety of human chromosomes as well as nonhuman primate genomes. A global evolutionary reconstruction of the dispersal of PIR4 sequence and analysis of flanking sequence supports a model in which pericentromeric duplications initiated before the separation of the great ape species (>12 MYA). Further, analyses of this duplication and associated flanking duplications narrow the major burst of pericentromeric duplication activity to a time just before the divergence of the African great ape and human species (5 to 7 MYA). These recent duplication exchange events substantially restructured the pericentromeric regions of hominoid chromosomes and created an architecture where large blocks of sequence are shared among nonhomologous chromosomes. This report provides the first global view of the series of historical events that have reshaped human pericentromeric regions over recent evolutionary time.

Full-text (PDF)

Available from: John Mcpherson, Apr 16, 2014
Using a Pericentromeric Interspersed Repeat to Recapitulate the Phylogeny
and Expansion of Human Centromeric Segmental Duplications
J. E. Horvath,* C. L. Gulden,* J. A. Bailey,* C. Yohn,* J. D. Mcpherson, A. Presco tt,à
B. A. Roe,à P. J de Jong,§ M. Ventura,k D. Misceo,k N. Archidiacono,k S. Zhao,{
S. Schwartz,* M. Rocchi,k and E. E. Eichler*
*Department of Genetics and Center for Human Genetics, Case Western Reserve University School of Medicine and University
Hospitals of Cleveland; Washington University School of Medicine Genome Sequencing Center, St. Louis; àDepartment of
Microbiology and Immunology, University of Oklahoma Health Sciences Center; §Children’s Hospital Oakland Research
Institute, Oakland, California; kSezione di Genetica, DAPEG, University of Bari, Bari, Italy; and {The Institute for Genomic
Research, Rockville, Maryland
Despite considerable advances in sequencing of the human genome over the past few years, the organization and
evolution of human pericentromeric regions have been difficult to resolve. This is due, in part, to the presence of large,
complex blocks of duplicated genomic sequence at the boundary between centromeric satellite and unique euchromatic
DNA. Here, we report the identification and characterization of an approximately 49-kb repeat sequence that exists in
more than 40 copies within the human genome. This repeat is specific to highly duplicated pericentromeric regions with
multiple copies distributed in an interspersed fashion among a subset of human chromosomes. Using this interspersed
repeat (termed PIR4) as a marker of pericentromeric DNA, we recovered and sequence-tagged 3 Mb of pericentromeric
DNA from a variety of human chromosomes as well as nonhuman primate genomes. A global evolutionary
reconstruction of the dispersal of PIR4 sequence and analysis of flanking sequence supports a model in which
pericentromeric duplications initiated before the separation of the great ape species (.12 MYA). Further, analyses of this
duplication and associated flanking duplications narrow the major burst of pericentromeric duplication activity to a time
just before the divergence of the African great ape and human species (5 to 7 MYA). These recent duplication exchange
events substantially restructured the pericentromeric regions of hominoid chromosomes and created an architecture
where large blocks of sequence are shared among nonhomologous chromosomes. This report provides the first global
view of the series of historical events that have reshaped human pericentromeric regions over recent evolutionary time.
With the sequence of many organisms complete or
nearly so, comparative work between species promises to
expand our knowledge of genome organization and
evolution. Pericentromeric regions are particularly in-
teresting because these regions demarcate the transition
between the heterochromatic alpha satellite DNA at the
centromere and the euchromatic gene-containing chromo-
some arm sequences. Further, such regions are sites of
rapid evolutionary turnover, reduced gene expression, and
suppressed genetic recombination (Mahtani and Willard
1998; Eichler 2001a; Yan et al. 2002). An understanding
of the genetic and functional properties requires a detailed
understanding of the sequence structure.
Pericentromeric Organization
In general, resolution of the organization and
evolution of these regions has been hampered by unusual
constellations of repetitive sequences when compared with
other regions of the genome. Sequence analysis of Dro-
sophila melanogaster pericentromeric regions indicated
that they are mainly composed of simple satellite
sequences, transposons, retroposons, and rRNA genes
(Sun, Wahlstrom, and Karpen 1997; Adams et al. 2000).
Similarly, Arabidopsis thaliana pericentromeric regions
are largely composed of retroelements, transposons,
microsatellites, and various classes of middle repetitive
DNA (Copenhaver et al. 1999; The Arabidopsis Genome
Initiative 2000). In addition to simple satellites and
retroposons, directed analyses of human pericentromeric
regions on chromosomes 2, 10, and 16 reveal a pre-
ponderance of partial gene duplications (Jackson et al.
1999; Loftus et al. 1999; Guy et al. 2000; Horvath et al.
2000; Horvath, Schwartz, and Eichler 2000). Although the
occurrence of mobile genetic elements and duplicated
sequence within pericentromeric regions is a common
property shared by these distant species, the structure of
the human genome appears to be unique in the proportion
and extent of these blocks of duplications, which may be
as large as a few Mb in size (Dunham et al. 1999; Jackson
et al. 1999; Hattori et al. 2000; Bailey et al. 2001; IHGSC
2001; Bailey et al. 2002a, 2002b; Crosier et al. 2002).
Human Pericentromeric Duplications
The structure of these large mosaic blocks of
duplication is complex. For nearly half of human
chromosomes, an estimated zone of duplication extends
from the satellite-repeat sequence to the unique euchro-
matic region (Bailey et al. 2001; IHGSC 2001). These
regions are composed of a mosaic of duplicated genomic
segments that originate from diverse areas of the genome.
A large number of partial-gene and whole-gene duplica-
tions have been recently characterized in detail (Jackson et
al. 1999; Horvath et al. 2001). These segmental duplica-
tions share conserved exon-intron structure and have been
termed duplicons (Eichler 2001b). In most cases, the
duplicons originate from an ancestral expressed locus,
range in copy number from 2 to 15, and show an
Key words: pericentromeric DNA, segmental duplications, genome
architecture, nonhuman primates, genome evolution.
Mol. Biol. Evol. 20(9):1463–1479. 2003
DOI: 10.1093/molbev/msg158
Molecular Biology and Evolution, Vol. 20, No. 9,
Ó Society for Molecular Biology and Evolution 2003; all rights reserved.
by guest on June 1, 2013 from
interchromosomal distribution restricted largely to peri-
centromeric regions. Comparative analyses of a few
regions indicate that these transposed duplicated segments
are found only in humans and closely related nonhuman
primates (Arnold et al. 1995; Eichler et al. 1996, 1997;
Regnier et al. 1997; Zimonjic et al. 1997; Orti et al. 1998;
Horvath et al. 2000). With the exception of these few
anecdotal studies focused on individual duplicated seg-
ments, a global synopsis of this property of genome
evolution and chromosome structure has been lacking. The
molecular basis for the duplicative transposition bias
toward pericentromeric regions is unknown.
Pericentromeric-Specific Repeat Sequences
In addition to duplicated gene segments, a variety of
primate-specific degenerate repeat sequences have been
identified between the duplicons (Eichler et al. 1997;
Eichler, Archidiacono, and Rocchi 1999; Guy et al. 2000;
Horvath, Schwartz, and Eichler 2000). The fact that they
demarcate the transition between unrelated pericentro-
meric genic duplication events and that, in at least one
case, they existed before the evolutionary transfer of the
duplicated segments has been taken as circumstantial
evidence that these repeats may play a role in the
duplication process (Eichler, Archidiacono, and Rocchi
1999). Unlike the genic duplications described above,
these pericentromeric interspersed repeat sequences (PIRs)
do not exhibit obvious exon-intron structure. They,
therefore, do not appear to be derived from ancestral gene
sequences that have been transposed from nonpericentro-
meric regions of the genome. Several types of pericentro-
meric repeat sequences have been described, including
CAGGG, GGGCAAAAGCCG, and chAB4 repeats (As-
sum et al. 1991; Eichler et al. 1996; Wohr, Fink, and
Assum 1996; Eichler et al. 1997; Eichler, Archidiacono,
and Rocchi 1999; Horvath, Schwartz, and Eichler 2000).
Unlike satellite sequences, these sequences are not
composed of repetitive tandem arrays. In some cases, the
underlying sequence structure of the interspersed repeats is
reminiscent of degenerate subtelomeric repeat tracts (Flint
et al. 1997; Riethman et al. 2001). Indeed, telomeric
associated–repeats have occasionally been reported in
close proximity to these sequence elements (Eichler,
Archidiacono, and Rocchi 1999). In addition, the pericen-
tromeric interspersed repeats often exist at multiple
locations within the same chromosome, separated by tens
to hundreds of kb of intervening duplicated sequence.
Here, we characterize a novel pericentromeric in-
terspersed repeat, termed PIR 4, that is specific to the
genomes of humans and apes. This element represents one
of the most abundant recent segmental duplications within
the human genome. Among humans, this repeat occurs on
more than half of all chromosomes, it is found in
association with other segmental duplications, and it is
restricted almost exclusively to pericentromeric regions.
The purpose of this study was to take advantage of the
multichromosomal and pericentromeric distribution of this
interspersed repeat, using it as a marker to (1) recover
additional sequence from these intractable regions of the
genome, (2) map existing sequences generated as part of
the Human Genome Project that were ambiguously placed,
and (3) reconstruct the series of evolutionary events that
occurred in the distribution of this repeat among primate
chromosomes. Our analysis provides a global snapshot of
the dynamic evolutionary history of these regions and the
series of nonhomologous sequence exchanges that created
the architecture of contemporary human chromosomes.
Materials and Methods
Computational Analyses
To identify all sequenced copies of PIR4 in the
genome, BlastN sequence similarity searches were initially
performed against both nr (nonredundant) and htgs (high
throughput genomic sequence) divisions of GenBank
using representative PIR4 masked (RepeatMasker version
07/13/2002, A. F. A. Smit and P. Green [http://ftp.genome.]) sequence from
AC002038.1 (coordinates 140,121 to 161,973). All
sequenced accessions were then searched against each
other to identify the longest copy that was AC073318
(positions 71,401 to 120,576). Repeat-masked sequence
from AC073318 was used as query against both nr and
htgs divisions of GenBank and identified 170 finished and
working draft GenBank accessions containing at least 1 kb
and more than 90% sequence identity to the query
Of the 170 accessions containing PIR4, there were
only 37 that were distinct finished copies and could be
used for further analyses. These 37 GenBank accessions
were analyzed using our previously described algorithm
(Bailey et al. 2001), which is designed to capture large
genomic alignments despite the presence of retroposon-
induced large insertions and deletions. Here, the PIR4
reference (AC073318) was compared with each of the 37
finished clones after the high copy repeats identified by
RepeatMasker (version 07/13/2002) were spliced out and
then a pairwise comparison using gap Blast was generated
(Altschul et al. 1990). For each alignment, repeats were
subsequently reinserted, the end-points were heuristically
trimmed, and optimal global alignments were generated
using the program ALIGN (Myers and Miller 1988).
Based on these alignments, we extracted a PIR4 sequence
for each sequence accession based on the orientation and
extent of the putative ancestral locus, AC073318, where
overlapping sequences compared with the reference
sequence were removed in favor of the higher scoring
alignment within the clone’s sequence. These extracted
segments served as the basis for constructing an optimal
global alignment for all PIR4 pairs of sequence. We
limited our analysis to alignments that were at least 10 kb
(a total of 25 GenBank accessions). We estimated the
number of substitutions/site/year (substitution rate) by
correcting the divergence for multiple substitutions using
Kimura’s two-parameter model (Kimura 1980).
To study the characteristics of other duplicons
flanking the PIR4 sequences, we performed a second all-
by-all BlastN comparison of the 25 accessions that
included the entire GenBank accession. We defined
flanking alignments as alignments within 7 kb (the full
length of L1 element insertion) of PIR4 sequence.
1464 Horvath et al.
by guest on June 1, 2013 from
Alignment statistics were only calculated for the non-PIR4
alignment portions. For each clone comparison, we
selected the largest global alignment (minus PIR4) that
was at least 10 kbp. To compare the divergence of PIR4
with the largest flanking alignment, we calculated
the difference (K
). In this case, a positive
) value indicates that PIR4 is more
divergent, whereas a negative (K
) value
reflects a more divergent flanking sequence. A value at or
near zero indicates that both PIR4 and flanking duplicons
were equally divergent and likely duplicated at or near the
same timepoint in evolution.
Identification of duplicons within accessions (fig. 1a,
b, and c) was performed using a repeat masked accession
against the EST division of GenBank. All ESTs exhibiting
exon-intron structure to the given accession were searched
against the Unigene database (http://www.ncbi.nlm.nih.
gov/entrez/query.fcgi?db¼unigene). A representative EST
(for each cluster with more than one EST hit to the given
accession) and all ESTs without Unigene hits were
subsequently queried against the nr and htgs divisions of
GenBank. The accession with a 100% match to the query
EST was considered the ancestral locus of the duplication,
which then was used in comparisons with the original
PIR4 accession to determine the extent of overlap and
percent identity of this segment as shown in figure 1.
The RPCI-11 human BAC library (segments 1, 2, 4,
and 5), RPCI-43 chimpanzee BAC library, RPCI-41 baboon
BAC library (segments 1 and 2), CHORI-253 orangutan
(segment 1) and cosmid libraries LLNL-01AH, LLNL-
02AE, LLNL-07Y, LLNL-09P, LA13NC01, LA14NC01,
LA15NC01, LA16NC01, LLNL-18AD, LLNL-21Q, and
LLNL-22N (table 1) were hybridized with a PCR-generated
probe using forward primer 32 and reverse primer 49 (see
table 2) amplified from 2p11 BAC DNA (AC002038).
Known PUC false positives were removed from all BAC
positive lists before PCR analysis. Since no false positive
lists exist for the cosmid libraries, all positives were used in
PCR assays and only those amplifying with 32N49 were
used in further analyses. The RPCI-41 (segment 2) baboon
BAC library was also hybridized using a long-range PCR-
generated probe using forward primer (62) and reverse
primer (64) (see table 2). This hybridization yielded no
positives. Hybridization probes were purified from pooled
PCR product using Qiagen’s QIAquick
PCR purification
kit (250) according to the manufacturer’s recommendations.
Twenty-five to 50 nanograms of purified product was
random-hexamer labeled with [a-
P] dCTP using Amer-
sham’s Megaprime kit according to the manufacturer’s
recommendations. All membranes were blocked with 1mg
sonicated salmon sperm DNA (Stratagene, La Jolla, Calif.).
High-density arrayed BAC and cosmid membranes were
hybridized at 658C for at least 16 h in 25 ml hybridization
solution (0.25M NaPO
, 0.25M NaCl, 5% SDS, 10% PEG,
and 1mM EDTA) heated to 658C. Membranes were washed
three times (for 30 min each) at 658C in wash solution
(0.05M NaPO
, 0.5% SDS, and 1mM EDTA) at room
temperature for the first wash and heated to 658 C for the
second and third washes. Genomic southern blots were
performed using 5lg PstI digested genomic DNA from two
chimpanzees, two bonobos, one baboon, one orangutan,
two gorillas and two humans. Genomic DNA was trans-
ferred to Zeta-Probe
membranes (BIO-RAD). ‘Genomic
blots’ were hybridized at least 16 h in QuikHyb
(Stratagene, La Jolla, Calif.) at 658C and then washed four
times (1 min at room temperature, 1 min at room
temperature, 15 min at room temperature, and 15 min at
658C) in 2 3 SSC/0.1% SDS and then four times (10 min
each at 658C) in 0.1 3 SSC/0.1%SDS.
PCR and Sequencing
The BAC and cosmid clones used for PCR analysis
were grown from single-colony isolates in 5 ml overnight
cultures. The DNA was isolated using Qiagen (Valencia,
Calif.) QIAwell 8 DNA isolation kit and resuspended
in water and 1/25 (BAC) or 1/250 (cosmid) of the total
volume (;15 ng) was used in subsequent PCR assays.
Gibbon genomic DNA was isolated from cell lines using
DNA isolation kit (Gentra systems, Min-
neapolis, Minn.) according to the manufacturer’s recom-
mendations, and 100 ng was used as a template in PCR
assays. Long-range gibbon PCR products (;1 kb) were
amplified from primate genomic DNA (Hylobates Lar and
Hylobates klossii), subcloned into PGEM
-T easy cloning
vector using the Promega Rapid ligation kit according to
the manufacturer’s recommendations, transformed into
XL1-Blue supercompetent cells (Stratagene, La Jolla,
Calif.), and screened by PCR to identify transformants
containing full-length inserts. Positive transformants were
then amplified with three sets of forward and reverse
primers (32N49, 120N123, and 124N129 [see table 3]).
All PCR conditions entailed a 2 min initial denaturation at
958C, followed by 35 cycles of: 958C for 30 s, 558C for 30
s, and 728C for 45 s followed by a final extension at 728C
for 7 min and then a 48C hold. Long-range PCR using
primers 62N64 were amplified as above with a 60 s
extension time at 728C for each of 35 cycles. PCR
products were directly sequenced using the forward and
reverse primers following a modified dye-terminator
sequencing protocol (Horvath et al. 2000). To remove
single-stranded DNA and deoxynucleotide triphosphates
from the PCR after the cycling steps were completed, 8 ul
of PCR product was treated with 1.50 U exonuclease I and
0.30 U shrimp alkaline phosphatase (Amersham Corpora-
tion) at 378C for 5 min and then heat inactivated at 728C
for 15 min followed by a 48C hold. Cycle sequencing
conditions were performed in 8 ll: 5 ll exonuclease I/
shrimp alkaline phosphatase treated PCR product, 1 ll
primer (20 lM), and 2 ll dichlororhodamine dye-
terminator reaction mix (ABI). All fluorescent traces were
analyzed using an Applied Biosystems PRISM
377 DNA
Sequencing System (Perkin-Elmer Applied Biosystems,
Norwalk, Conn.), and the quality of the sequence data was
assessed with PHRED/PHRAP/CONSED software (http:// Sequence data used in this manuscript
is available from Genbank accession numbers CC466014
through CC466043.
Expansion of Human Centromeric Segmental Duplications 1465
by guest on June 1, 2013 from
1466 Horvath et al.
by guest on June 1, 2013 from
Phylogenetic Analysis
Fasta formatted sequence files for all BAC and
cosmid sequences were created after comparison of both
forward and reverse sequences of each PCR product using
CONSED. Fasta formatted sequence files from accessions
used to generate the 1-kb and 3-kb trees were extracted
from the most updated GenBank accession, and coor-
dinates are listed in tables 1 and 2, respectively, of
Supplementary Material online. Multiple pairwise align-
ments were generated using ClustalW (version 1.82)
(Higgins, Thompson, and Gibson 1996). For the 1-kb
tree, 945 bp of PIR4 sequence was generated from 67
human, two chimpanzee, four orangutan, and four gibbon
loci (fig. 2). For the 3-kb tree, 3,000 bp of sequence was
extracted from 32 human accessions (fig. 1 in Supple-
mentary Material online). Phylogenetic analyses were
performed using MEGA (Molecular Evolutionary Genetic
Analysis) version 2.1 (
(Kumar et al. 2001). Neighbor-joining analysis was used
with complete deletion parameters and bootstrap (1,000
iterations) to provide confidence of each branchpoint in the
phylogenetic trees. We chose to use the neighbor-joining
method (although minimum evolution was also used and
yielded a tree with similar results) because we were
interested in calculating divergence times between se-
quence taxa and neighbor joining methods were amenable
to this task. Also, maximum-likelihood and parsimony
methods are too cpu-intensive with 77 taxa. Determination
of orthologous chimpanzee and orangutan BAC sequences
was conducted by BAC end sequence placement with
respect to NCBI, build31, (November 2002). Chimpanzee
BAC 145P5 end sequences and orangutan BACs 220O24
and 1J18 end sequences placed at positions orthologous to
human AC073318 on build31. We were unable using
build31 to determine the orthologous placement of
chimpanzee BAC 109O6. Because the rates of nucleotide
substitution vary for pseudogenic sequences, the rate of
nucleotide substitution was calibrated based on ortholo-
gous PIR4 sequence comparisons between human and
primate sequences using a divergence of 18 MYA for
FIG. 1.—The genomic organization of sequences flanking PIR4. Three examples of the duplication architecture surrounding PIR4 loci are shown.
For each, the horizontal black line depicts the genomic sequence drawn to scale (tick marks occur every 20 kb along the sequence). Black rectangles
represent PIR4 loci, dark gray rectangles represent satellite sequences, and light gray rectangles indicate duplicons shared by two or more genomic loci
(as identified by RepeatMasker and Blast searches; see Materials and Methods). Slanted black lines between accessions indicate regions containing
PIR4, and gray lines indicate regions duplicated between two or more genomic loci. The program PARASIGHT (Bailey, unpublished) was used to
generate this output. (a) A comparison of ancestral AC073318 sequence with AC127380 indicates that these two accessions have only the 35 kb of
PIR4 in common. Whereas PIR4 on AC073318 is flanked by ‘‘unique’ sequence, PIR4 on AC127380 is flanked by ‘unique’ sequence on one side
and a stretch of nearly 160 kb of satellite sequences on the other side. (b) Comparing two chromosome 2 loci with ancestral AC073318 indicates that
both chromosome 2 loci share more sequence in common than PIR4 alone. Although PIR4 is truncated at the end of both chromosome 2 clones, on one
side PIR4 is flanked by a duplicated genomic segment common to both clones. Both of these clones also contain duplicons not shared by the other and
AC026273 contains approximately 40 kb of alpha satellite sequences. Ancestral loci are indicated above or below each light gray rectangle as well as
percent identity and length of alignment. Interestingly, AC026273 contains a 100% match over five exons to EST BQ651044 (Crosier et al. 2002). (c)
Two loci from different chromosomes containing PIR4 indicate that PIR4 is often flanked by duplicons shared by different chromosomes. These two
loci on chromosomes 13 and 21 share almost 100 kb of sequence that is composed of numerous duplicons, including PIR4.
Table 1
Genomic Distribution of PIR4 in Representative Primates
Library Library Name Fold Coverage Positives PCR Amplified Estimated Copy Number PSVs
Human BAC RPCI-11* 25.5 768 205 30.1 43þ
1 cosmid LL01 4 95** 43 10.8 9
2 cosmid LL02 3.9 40** 19 4.9 3
7 cosmid LL07 4.8 32** 16 3.3 7
9 cosmid LL09 5.6 43** 37 6.6 7
13 cosmid LA13 7 16** 15 2.1 4
14 cosmid LA14 5 33** 29 5.8 6
15 cosmid LA15 6 5** 5 0.8 1
16 cosmid LA16 5.9 51** 25 4.2 8
18 cosmid LL18 5.7 5** 0 0 0
21 cosmid LL21 7.4 8** 4 0.5 1
22 cosmid LL22 17.1 39** 24 1.4 4
Chimp BAC RPCI-43 3.5 82 57 23.4 20þ
Orangutan BAC CHORI-253 6.4 33 25 3.9 4
Baboon BAC RPCI-41 10.4 0 0 0 0
NOTE.—A summary of the number of positives identified by radioactive colony hybridization of human, chimpanzee, orangutan, and baboon BAC and human cosmid
libraries using a 300-bp PCR product as a probe (amplicon 32N49 [see table 2]). All positively hybridizing clones were screened by PCR (with the exception of RPCI-11,
where only a subset of the positive BACs were chosen for further analyses). Copy number estimates for the BAC libraries were calculated by dividing the number of
positives by the fold coverage of the library. Copy number estimates for the cosmid libraries were estimated by dividing the number of cosmids PCR amplifying by the fold
coverage of the library. Direct sequencing was performed on all PCR products (Materials and Methods). Based on the genomic coverage of each library and the number of
positive BACs or positive cosmids that successfully amplified by PCR, an estimate of the copy number (expected) of PIR4 in each library was calculated. The number of
PSVs (paralogous sequence variants) observed was determined by comparison of all directly sequenced PCR products to determine the number of distinct variants
represented in each library. Occasionally a BAC sequence would have heterozygous peaks, suggesting there were two copies of PIR4 within a single BAC. * indicates
analyzed segments 1, 2, 4, and 5. ** indicates number of true positives unknown because no false positive list available. þ indicates some BACs have at least two copies of
PIR4 that could not be distinguished by direct sequencing. PSV (paralogous sequence variant) is a distinct sequence signature when compared with other paralogous copies.
Expansion of Human Centromeric Segmental Duplications 1467
by guest on June 1, 2013 from
gibbon-human and 6 MYA for chimpanzee-human di-
vergence. Duplication timing events were calculated using
the equation T ¼ K/2r (Li 1997). We conducted relative
rate tests to determine whether molecular clock estimates
would be valid. Relative rate tests were performed in
MEGA using human AC073318 and chimpanzee BAC
145P5 or orangutan BAC 220O24 in comparison with
orangutan BAC 220O24 or gibbon as an outgroup,
respectively. The relative rate test (Tajima’s test) was not
rejected as chi-square values ranged between 0.5 and 0.27
with a probability of 0.819 to 0.602, respectively.
Fluorescent in Situ Hybridization
Human metaphase chromosomes (fig. 3) were pre-
pared as described previously (Horvath, Schwartz, and
Eichler 2000) and hybridized with BAC DNA isolated
using the Nucleobond
DNA isolation kit from Clontech
(Palo Alto, Calif.) according to manufacturer’s recom-
mendations. Human and primate metaphase chromosomes
(fig. 4) from H. sapiens, P. trogl odytes, G. gorilla, P.
pygmaeus, H. lar, and M. fascicularis were prepared from
lymphoblastoid lines as described previously (Horvath,
Schwartz, and Eichler 2000).
Identification and Characterization of PIR4 Within the
Human Genome
PIR4 (pericentromeric interspersed repeat number 4)
was initially identified during the sequence characterization
of a clone that mapped to the pericentromeric region of
human chromosome 2p11 (AC002038 [Horvath, Schwartz,
and Eichler 2000]). Using reiterative DNA searches against
GenBank (see Materials and Methods) a putative full-length
copy of this element (49kb) was subsequently identified on
chromosome 7 (AC073318). Sequence similarity searches
of GenBank (12/15/02) revealed highly identical (90% to
99% sequence identity) copies of this element (.1kbin
length) on more than 170 human accessions representing at
least 44 distinct loci. The majority of these genomic
sequences (70%) were not assigned to a chromosome within
the working draft assembly of the human genome (NCBI,
build 31, November 2002). Sequence similarity searches of
expressed sequence databases revealed a single significant
HSP (E ¼ 2e47) for a potential unprocessed EST that did
not have mRNA support (BQ082091.1). Other than this
Table 2
PCR Oligonucleotide Sequences
Primer Orientation Sequence
NOTE.—All oligonucleotides were designed based on the 2p11 BAC reference
sequence (AC002038). Sequences are presented in 59 to 39 orientation. Both for-
ward (F) and reverse (R) oligonucleotides for each assay are presented.
Table 3
FISH Localizations of PIR4-Positive BACs
BAC Library
Location Metaphase FISH Locations
101B6 CIT AC002038 2p11 1cen 2cen* 4q24 7pcen 9p/qcen 10cen 16p11 17q11 21cen 22q11 Ycen
2i21 RPCI-11 none unk 1cen 2cen 7cen* 9p/qcen 10cen 13q11 14pcen 15q11 16cen 17cen 18cen 19cen 20cen 21cen 22cen
2053H7 CIT AC025223 2 1cen 2cen* 7cen 9cen 10cen 13cen 14cen 15cen 16cen 17cen 18cen 19cen 21cen 22cen
168j1 RPCI-11 AC034151 2 1qh 2pcen* 4q24 7cen 9cen 10cen 13cen 14cen 15cen 16pcen 21cen 22cen* Ycen
165d20 RPCI-11 AC027612 2 1cen/qh* 2cen* 7cen/qter 9cen 13cen 14cen 15cen 16cen 21cen 22 doublet
28o7 RPCI-11 AC129782 1 1p32/
1qh* 2p11/2q14 7cen 9p11 13cen 14cen 15cen 16p11/q11 17cen 18cen 21cen 22cen
51a19 RPCI-11 AC129338 7 1cen 2cen 7cen* 16p11
1429e17 RPCI-11 none 1or7 1qh 2p/qcen 7cen/qter 9qh 10qtel 13cen* 14cen 15cen 16cen 17cen 21cen* 22cen*
1390m18 RPCI-11 none 1or9 1p/qh* 16pcen
1386h14 RPCI-11 none 1 1qh* 2cen* 4pter 7cen 9qh 10cen 13pcen 14pcen 15qcen 16cen 17cen 21cen 22cen Ycen
3m10 RPCI-11 none 13 13cen* 14pcen 15cen* 20cen 21pcen 22pcen
1360m22 RPCI-11 AC127381 15 1qh 2cen/q21 7cen 9qh 10cen 13cen 14cen/ qter 15qcen* 16cen* 17cen 18cen 21cen 22cen*
1391n9 RPCI-11 none 22 1qh 2cen/q21 7cen 8cen 10cen 13cen 14cen 16qh* 17cen 18cen 21cen 22cen*
1390a11 RPCI-11 AC127384 16 1p36/ pcen/ qh* 2pcen 5cen 7cen 13cen 14cen 16pcen 21pcen 22pcen
1221g12 RPCI-11 AC129778 unk 1qh 2cen/q21* 7cen 9qh/p* 10cen 13cen 14cen 15cen 16pcen 21cen 22cen Ycen
1360o11 RPCI-11 none unk 1qh 2pcen 7pcen 14cen 16pcen* 22pcen
1363e3 RPCI-11 none 2 1qh 2p/qcen* 4q26 7cen 9qh 10cen 13cen 14cen 15cen 16pcen 22qcen* Ycen
NOTE.—CIT D ¼ California Institute of Technology, Library D; RPCI-11¼ Roswell Park Cancer Institute; * indicates largest signal(s). To assess genomic distribution of PIR4, 17 RPCI-11 (human) BACs were selected as probes for
FISH against human metaphase chromosomes. Accession numbers and chromosomal placement (determined if a BAC PSV matches a cosmid PSV) are indicated when known. Chromosomes 1, 2, 7, 9–17, 21, and 22 were observed in more
than half of all BAC FISH experiments.
1468 Horvath et al.
by guest on June 1, 2013 from
single EST, there was no evidence that the region was
transcribed or that it possessed ancestral exon-intron
structure. Sequence content analysis of the 49-kb element
revealed repeat and GC-content only slightly lower (34.4%
GC and 42.5% repeat content) than the genomic average
(IHGSC 2001). Among the interspersed repeat classes, only
LTR and LINE content were slightly (albeit not significant-
ly) increased (LTR; 10.24% versus 8.29% genome average,
and LINE; 24.73% versus 20.42% genome average).
Overall, there is no obvious sequence property of this
element that would easily account for its proliferation within
the genome.
FIG. 2.—Phylogeny of PIR4. A neighbor-joining phylogram rooted on gibbon using approximately 950 bp of PIR4 sequence from human,
chimpanzee, orangutan, and gibbon sequences was constructed using MEGA (Materials and Methods). Bootstrap values greater than 80% from 1,000
replicates are indicated on each respective branch. The branch separating clades A and B has a bootstrap value of 80 for the 1-kb tree and 100 for the 3-
kb tree (fig. 1 in Supplementary Material online), indicating high confidence. Sequences with an asterisk indicate those that have been mapped to
a chromosome based on cosmid support. Human accession or clone sequences are followed by chromosomal location, when known. Nonhuman
primate sequences generated from BAC clones for chimpanzee and orangutan and genomic DNA for gibbon are shaded light gray, dark gray, and black,
respectively. Two clades are readily distinguished (clade A and B). Most nonhuman primate sequences (orangutan and gibbon), as well as the putative
ancestral locus on chromosome 7, map to clade B.
Expansion of Human Centromeric Segmental Duplications 1469
by guest on June 1, 2013 from
PIR4 Copy Number Estimates
A variety of methods (Southern analysis, FISH, library
depth-of-coverage) were used to estimate copy number in
the genome. Initially, a ‘‘unique’ 300-bp PCR (32N49)
amplicon was designed specific to the repeat and screened
against a 25.5-fold redundant human BAC library (RPCI-
11, segments 1, 2, 4, and 5). We obtained 768 strongly
hybridizing positives, suggesting there were at least 30
copies of this element in the human genome (table 1). An
independent analysis of depth of coverage using whole-
genome shotgun sequence data (Bailey et al. 2002a)
showed a 40-fold (1972/47.2) excess of sequence read
depth when compared with unique regions of the genome
( Of all segmental dupli-
cations characterized within the human genome, only DNA
sequence corresponding to rDNA duplications from
acrocentric chromosomes surpassed PIR4 in both depth
of coverage and degree of sequence identity. PCR analysis
of a monochromosomal hybrid DNA panel confirmed
copies of PIR4 on chromosomes 1, 2, 7, 9, 10, 13, 14, 15,
16, 17, 21, 22, and Y (fig. 5a). Subsequent sequence
analysis of the amplicons, in many cases, revealed the
presence of ‘heterozygous’ sequence signatures (fig. 5b).
Since each DNA sample was derived from a monochromo-
somal somatic cell hybrid source, it is likely that multiple
copies of the element are present on many chromosomes.
To improve the copy number estimate and to recover
genomic clones specific for each chromosome, cosmid
libraries from flow-sorted human chromosomes 1, 2, 7, 9,
13, 14, 15, 16, 18, 21, and 22 were hybridized with
amplicon 32N49 (table 1). Based on the depth of coverage
for each library, the results suggested that most human
chromosomes contained multiple copies of the PIR4 repeat
sequence (mean ¼ four copies per chromosome for
chromosomes with PIR4) with chromosomes 1 and 9 being
particularly enriched (estimated seven to nine copies).
Mapping PIR4 Sequences to Chromosomes
Two approaches were undertaken to determine the
location of PIR4 sequences in the human genome. First, 17
RPCI-11 BACs were individually probed against meta-
phase spreads of human chromosomes (table 3 and fig. 3).
Chromosomes 1, 2, 7, 9–17, 21, and 22 were observed in
FIG. 3.—Examples of metaphase FISH from PIR4-containing BACs. FISH was conducted for all BACs listed in table 2. The results for a subset
are shown here. All BACs hybridize to multiple pericentromeric loci. Signal intensity alone is not a good indicator of chromosomal origin, because
some chromosomes consistently demonstrate large signals (chromosome 1 in panel b), whereas others show discrete signals flanking the centromere,
suggesting more than one copy of PIR4 exists on certain chromosomes (chromosome 2 in panels a and c and chromosome 9 in panels b and c).
1470 Horvath et al.
by guest on June 1, 2013 from
more than half of all BAC FISH experiments, consistent
with monochromosomal hybrid PCR data. Occasionally,
two signals were observed on the same chromosome
(chromosome 2, fig. 3a and c and chromosome 9, fig. 3b
and c). In both of these cases, the noncentromeric signal
(9q12 and 2q21) corresponded to the site of an ancient
vestigial centromere recently euchromatized as a result of
evolutionary chromosomal rearrangements within the
human lineage (Baldini et al. 1993). To exclude the
possibility that duplicated sequences flanking PIR4 were
primarily responsible for these cross-hybridization results,
these experiments were repeated using a chromosome 22
cosmid clone (N20B5, AC093314) which had been
sequenced in its entirety and was found to contain PIR4
as its sole duplicon, as well as common repeats that could
be easily blocked by C
1-DNA that hybridizes to almost
all chromosomes identified in table 2 (fig. 4a).
Although FISH data confirmed a pericentromeric map
location for the vast majority of PIR4-containing BAC
clones, it was impossible using this approach to un-
ambiguously assign a specific BAC clone to its chromo-
some of origin. The variable copy number of the repeat
within specific chromosomes, furthermore, made assign-
ment based on signal intensity unreliable (table 3 and fig.
3). As a secondary means of resolving the chromosomal
location of PIR4-containing clones, we implemented
a sequence-based strategy (termed paralogous sequence
tagging) (Horvath, Schwartz, and Eichler 2000) that
depends on the identification and characterization of
paralogous sequence variants (PSVs) specific to chromo-
somes with PIR4. Since the BAC libraries were con-
structed from two chromosomal haplotypes (maternal and
paternal), sequence variants between two BACs may be
due to either allelism (one maternal variant and one
FIG. 4.—Primate metaphase FISH of PIR4-containing clones. (a) A human chromosome 22 cosmid (N20B5, AC093314) containing only the PIR4
duplicon was hybridized to human (Hsa), common chimpanzee (Ptr), and orangutan (Ppy) metaphase spreads after C
1 blocking. In humans, this probe
hybridized to pericentromeric regions of chromosomes 1, 2, 7, 9, 13, 14, 15, 16, 17, 18, 21, and 22, whereas in chimpanzee, it hybridized to pericentromeric
regions of chromosomes I, IIp, VII, X, and XVI. In contrast, this probe hybridized only to syntenic chromosome VII in orangutan, the putative ancestral
locus. All chromosomal designations are with respect to the human phylogenetic group designations (ISCN 1985). (b) Orangutan BAC CHORI-253-1J18
containing PIR4 sequences hybridizes only to chromosome 7 in human and orangutan, whereas (c) orangutan BAC CHORI-253-346B14 hybridizes solely
to chromosome VII in orangutan but to multiple pericentromeric regions (chromosomes 1, 2, 7, 14, 16, 17, 21, and 22) in humans.
Expansion of Human Centromeric Segmental Duplications 1471
by guest on June 1, 2013 from
paternal variant at the same locus) or paralogy (two
variants at different loci as the result of a duplication
event). In contrast to BAC libraries, cosmid libraries
constructed early in the Human Genome Sequencing
Project at Lawrence Livermore and Los Alamos National
Laboratories were constructed from a single flow-sorted
chromosome (from a somatic cell line containing a single
human chromosome) and represent in theory a single
haplotype, thereby excluding allelism as a possible source
for the variation (Trask et al. 1991). Thus, sequence
variants between two cosmid clones from the same library
identify paralogous, not allelic, copies. Sequence identity
matches between BAC and cosmid sequence signatures
further allow the assignment of large-insert BAC clones to
specific pericentromeric regions. We reasoned that clones
identified from each of the cosmid libraries could,
therefore, be informative as a mapping resource for these
intractable, highly duplicated areas of the genome.
To implement this approach, we selected 205 BACs
(RPCI-11) and 176 cosmids for sequence analyses. Each
clone was PCR amplified using oligonucleotides specific
to the PIR4 repeat and all PCR products were directly
sequenced to obtain a catalog of sequence signatures that
distinguish various contigs of clones. A total of 67 distinct
cosmid and BAC-derived sequence signatures were
identified. A BAC sequence signature was considered
distinct if the number of sequence differences was greater
(two differences/252 bp) than that expected for allelic
variation. Only high-quality sequence differences present
on both forward and reverse sequencing of the PCR
amplicon were considered in this analysis. Twenty-one of
these cosmid signatures matched a BAC signature,
allowing for chromosomal assignment of the BAC and
providing an anchor for future sequence assembly.
However, 20 BAC signatures were left unassigned to
any chromosome, and 27 cosmid signatures had no
evidence of BAC sequence support, suggesting extensive
levels of allelic variation for these pericentromeric loci.
Using the collection of experimentally derived sequence
signatures, sequence similarity searches were performed
against both the nonredundant (nr) and high throughput
genomic sequences (htgs) divisions of GenBank. Forty-
nine of the 67 variants matched an accession in the
database (zero or one variant in 252 bp), 19 of which could
be unambiguously assigned to a chromosome. In contrast,
20 BAC/cosmid signatures were not represented within
GenBank, suggesting considerable underrepresentation of
this segment within the current genome assembly, and
additional sequence tag information (total ¼ 945 bp) was
obtained for future sequence comparisons. Further,
analysis of the most recent assembly of the human genome
(NCBI, build 31, November 2002) using these PIR4
sequences revealed that only 20 of the estimated 49
database copies of PIR4 were currently represented within
this assembly, confirming considerable underrepresenta-
tion of these sequences within human genome assemblies.
In total, our analysis allowed the unambiguous chromo-
somal assignment of 19 distinct PIR4 loci. Sixteen RPCI-
11 BAC clones (AC127362, AC127380, AC127381,
AC127384, AC127387, AC127389, AC127391,
AC127701, AC128674, AC128676, AC128677,
AC129338, AC129778, AC129779, AC129782, and
AC092854) and three chromosome 22 cosmids
(AC093314, AC103582, and AC093091) were placed in
the sequence queue. Notwithstanding, half of the clones
characterized in this study could not be assigned to
a specific chromosome. This may be due to extreme levels
of allelic variation, structural heteromorphism, or clone
gaps within existing libraries.
Analysis of PIR4 Flanking Sequences
Since FISH analyses indicated that PIR4 occurred
exclusively in pericentromeric regions, we tested more
directly its association with satellite DNA (classical
centromeric DNA markers). A subset (306) of PIR4-
containing RPCI-11 BACs was selected for end-sequence
analysis. These sequences were then searched against
GenBank, revealing that at least one end sequence placed
within centromeric satellite DNA for 83 of these BACs
(27.1%), a significantly higher proportion than expected
based on random sampling of human BAC end sequences
(,1% satellite repeats). This association with satellite
DNA was further supported by analysis of existing Human
Genome Project data. Twenty of the 37 (54%) distinct
FIG. 5.—Genomic distribution of PIR4. (a) PCR analysis using
primer pair 32N49 against a monochromosomal somatic cell hybrid panel
of DNAs. Chromosomes 1, 2, 7, 9, 10, 13, 14, 15, 16, 17, 21, 22, and Y
amplified a product of identical size (;300 bp, Gibco-BRL 100-bp
ladder). (b) The chromosome 2 monochromosomal hybrid DNA
chromatogram shows four putative variant sites suggestive of multiple
copies of PIR4 on chromosome 2. Subsequent sequencing of cosmids
from the chromosome 2 LLNL02 library (92M18 and 67N8) resolves
these variant sites.
1472 Horvath et al.
by guest on June 1, 2013 from
BAC clones, for which finished sequence was available,
contained at least 1 kb (and most often more than 10 kb) of
centromerically associated satellite sequences, including
designations of centromeric DNA). These data are
consistent with PIR4 sequences lying within the euchro-
matin/heterochromatin transition zone in close proximity
to human centromeres.
Similarly, the segmental duplication content within
the vicinity of PIR4 loci was assessed by comparing the
sequences of the 37 large-insert PIR4 BAC clones that
had been completely sequenced and a comparison of the
flanking genomic sequences to the segmental duplication
database of the human genome (Bailey et al. 2002a).
With the exception of alpha satellite containing clones,
only AC073318 contained PIR4 as its sole duplication
element (fig. 1a). The organization of most clones
showed complex patterns of segmental duplications
(both interchromosomally and intrachromosomally) with
the PIR4 sequence most often associated with a larger
block of duplicated sequence (fig. 1b and c). This
organization of duplications embedded within duplica-
tions is consistent with the previously proposed two-step
model for the origin of pericentromeric duplications
(Eichler et al. 1997) (Horvath, Schwartz, and Eichler
2000). Based on this analysis, it therefore is not
surprising that these clones had ambiguous chromosomal
assignments in the public build31. Further, the lack of
unique sequence in the vicinity of PIR4 and the high
degree of sequence identity among the duplicates
indicates that most of the available PIR4-containing
sequences within GenBank could not be mapped using
traditional methods.
PIR4 as a Marker of Pericentromeric Duplications
Based on the multichromosomal distribution and the
pericentromeric specificity of PIR4, we reasoned that this
interspersed repeat might serve as an informative phylo-
genetic marker to reconstruct the series of evolutionary
events that have restructured these regions of the human
genome. Moreover, since most of the PIR4 elements were
associated with larger blocks of segmental duplication, the
PIR4 elements might also provide insight into these larger
secondary duplication events. This assumes that PIR4
sequences have not been preferential targets of gene
conversion and therefore represent ‘neutral’’ markers of
pericentromeric evolution. To test this assumption, the
pairwise genetic distance between each finished copy of
PIR4 within GenBank was calculated (fig. 6). Here, 10 kb
or more of aligned sequence was compared with the 25
copies of PIR4 for a total of 234 comparisons. Next, we
examined the largest flanking sequence excluding PIR4
and calculated the genetic distance between these
duplicated flanks. We then compared the genetic distance
of the PIR4 element to the genetic distance of the flanking
duplicated material as the difference of these two estimates
(see fig. 2 in Supplementary Material online). A difference
of zero (identity) between K values would suggest that
both PIR4 and flanking sequences had diverged equally
and arose at approximately the same time in evolution. A
negative K value would suggest that the PIR4 copies were
more similar than the flanking DNA and had therefore
undergone conversion events. Because we assessed only
flanking (within 7 kb of PIR4) and not nearby duplications
(.7 kb away), our sample size was small (18) and we
likely excluded some duplications that could have been
separated from PIR4 due to secondary rearrangement
events. However, since nearly half (7/18) of the PIR4
elements showed genetic distances consistent with those of
the flanking duplications (a difference of 0.005 changes/
bp, or less than 1% difference), many of the PIR4 elements
act as a marker of pericentromeric DNA.
Comparative Primate Analysis of PIR4
To provide evolutionary points of reference in our
analysis of PIR4, we employed complementary molecular
and cytogenetic approaches among representative non-
human primate species. Colony hybridizations were
performed using the 32N49 amplicon as a probe against
the chimpanzee (RPCI-43), orangutan (CHORI 253), and
baboon (RPCI-41) BAC libraries (table 1). Numerous
BAC clones were identified within the orangutan and
chimpanzee libraries, suggesting multiple (albeit a reduced
number of) PIR4 copies. Interestingly, hybridization
experiments against the baboon BAC library (10.8x
coverage) failed to yield a single positive. Subsequent
hybridizations using a larger amplicon as well as Southern
hybridization experiments against genomic DNA provided
no evidence of PIR4 within the baboon. To determine the
accuracy of the copy number estimates for the chimpanzee
and orangutan hybridizations, DNA was isolated from all
positive BACs and PCR products corresponding to the
32N49 amplicon were sequenced. All sequences were
compared within each species to identify the contiguous
sets of clones linked by a common set of sequence
variants. Through these studies, the 57 amplifying
chimpanzee BAC sequences could be grouped into 20
distinct sequence classes (one or more differences within
252 bp sampled), whereas the 25 orangutan BAC clones
fell into only four sequence classes.
Since this molecular evidence points to multiple
copies of PIR4 in chimpanzee and orangutan, two sets of
comparative FISH experiments were undertaken to de-
termine the copy number and distribution of PIR4
sequences on these primate chromosomes. In the first
study, a human chromosome 22 cosmid probe, N20B5,
which contained a single copy of the PIR4 sequence was
probed against chromosome spreads of chimpanzee and
orangutan metaphases (fig. 4a). Multiple pericentromeric
signals were observed on chimpanzee chromosomes (I,
IIp, VII, X, and XVI with respect to the human
phylogenetic group designations). In contrast to human
and chimpanzee metaphases, a single robust signal was
observed in orangutan metaphases, corresponding to
phylogenetic group VII. Since the absence of signal on
orangutans might presumably be due to sequence di-
vergence, a reciprocal set of experiments was conducted
using orangutan BACs as probes on both human and
orangutan chromosomes. A representative orangutan BAC
from each of the four sequence classes was assessed. Two
Expansion of Human Centromeric Segmental Duplications 1473
by guest on June 1, 2013 from
of the orangutan BACs (CHORI-253 220o24 and
CHORI-253 1j18) yielded identical results, hybridizing
to a single locus on chromosome VII in both human and
orangutan (BAC 1J18 [fig. 4b]). In contrast, orangutan
BAC 346B14 hybridized to a single locus in orangutan
(chromosome VII [fig. 4c]) but multiple chromosomes in
human (1, 2, 7, 14, 16, 17, 21, and 22), whereas 321D4
hybridized to two discrete but nearby loci on chromosome
VII in orangutan and multiple loci in humans (2, 7, 14,
16, and data not shown). BAC-end sequencing and
subsequent similarity searches of the orangutan PIR4-
containing BAC clones revealed that they mapped to two
different positions within the human chromosome 7
reference sequence.
Phylogenetic Analyses of PIR4 Sequences
As the final step in our analysis, a phylogenetic tree
was generated using MEGA2 to compare 67 human, two
chimpanzee, four orangutan, and four gibbon loci (fig. 2)
(Kumar et al. 2001). At least two major clades could be
distinguished. One clade (termed A) consists almost
entirely of human sequences from many different
chromosomes. This clade is further stratified into relatively
chromosome-specific subgroups of PIR4 (see chromo-
somes 1, 2, 7, and 9) as well as an acrocentric chromosome
subclade (13, 14, and 21). In contrast, clade B consists of
human, chimpanzee, orangutan, and gibbon sequences as
well as the putative ancestral human sequence on
chromosome 7 (AC073318). With the exception of
chromosome 7, very little evidence of chromosome-
specific amplification is observed within this clade. It
should be noted that chromosomes 2, 7, 13, 16, and 22
have representative sequences in both clade A and clade B.
To increase confidence of the two separate clades on the
1-kb tree, we generated a 3-kb tree from a subset of the
accessions (fig. 1 in Supplementary Material online). This
increased bootstrap support from 80% to 100% for the
existence of two clades. In total, these data suggest a rapid
dispersal of PIR4 sequences over a narrow window of
primate evolution followed by more recent chromosome-
specific duplication events. To examine this in more detail,
another phylogenetic tree was constructed from a shorter
multiple sequence alignment (252 bp), incorporating an
additional 18 distinct chimpanzee BAC sequences. These
chimpanzee sequences distributed throughout clade A and
clade B, showing, in general, closer phylogenetic relation-
ship to other human loci rather than other chimpanzee
sequences (data not shown). Thus, it is likely that PIR4
sequences populated the hominoid genome before the
divergence of the two lineages.
In the absence of a robust genome assembly near
centromeric regions, we conducted a global analysis of
half of all human pericentromeric regions using a single
element (PIR4), providing insight into the biology of these
complex regions of our genome. Within the human
genome, we estimate approximately 40 copies of this (20
to 40 kb) element, which share, on average, 95.2%
sequence identity (range 90.2% to 99.5% sequence
identity) (table 1). The available data suggest that PIR4
represents one of the most prolific and highly homologous
segmental duplications within the human genome. Cyto-
genetic and molecular evidence confirm that the repeat
localizes almost exclusively to the pericentromeric regions
of more than half of all human chromosomes (1p, 2p, 7p,
9p, 9q, 10q, 13q, 14q, 15q, 16p, 17q, 18q, 21q, 22q, and
Ypcen, [tables 1 and 3 and fig. 3]). With the exception of
the ancestral copy on AC073318 from 7p12 and a few
copies flanking alpha satellite DNA, all PIR4 elements
map within 100 kb of other duplicated segments. PIR4
elements themselves almost always are a component of
a larger duplication block with a more limited pericentro-
meric distribution pattern. Interestingly, the evolutionary
age of the PIR4 sequences was often consistent with the
evolutionary age of the duplicated flanking sequencing.
This suggested that phylogenetic analysis of PIR4 not only
would be valuable in reconstructing the evolutionary
history of the repeat but also would provide insight into the
series of large-scale duplications that have reshaped
hominoid pericentromeric regions.
PIR4 Ancestral Sequence
Several lines of evidence point to chromosome 7 as
the ancestral origin of PIR4. First, it is the only
chromosome commonly hybridizing to human, chimpan-
zee, and orangutan metaphase spreads (fig. 4ac). Second,
it is the only locus for which a clear ortholog can be
identified within each great ape species examined (fig. 2).
This is supported both by phylogeny as well as BAC end-
sequence analysis. Third, it is one of the only PIR4-
containing loci (GenBank AC073318) devoid of other
segmental duplications. Characterization of numerous
other segmental duplications (Eichler et al. 1996, 1997;
Regnier et al. 1997; Zimonjic et al. 1997; Horvath et al.
2001; Crosier et al. 2002) suggest that the progenitor loci
most often occur outside of the pericentromeric duplica-
tion zone surrounded by unique sequence. Subsequent
duplicative transposition events become associated with
FIG. 6.—Genetic distances between PIR4 sequences. The number of
substitutions per site (K) among PIR4 elements as a function of the
number of pairwise alignments. Based on available finished sequence, 25
PIR4 sequences were extracted. A total of 254 pairwise alignments were
performed, and the genetic distance for each pairwise was calculated.
Each alignment was at least 10 kb in length. Distance estimates were
computed using the Kimura two-parameter model.
1474 Horvath et al.
by guest on June 1, 2013 from
other pericentromeric duplications. Finally, size estimates
of PIR4 from AC073318 indicate that it represents the
largest and most complete copy (49 kb). Other copies of
PIR4 have become truncated with respect to this locus,
perhaps as a result of deletion of secondary progenitors
before subsequent rounds of duplication (fig. 7). For
example, AC093787 from chromosome 2 contains 44.2 kb
of PIR4, AC025223 from chromosome 2 has 17.3 kb, and
AC073210 and AC104057 from chromosome 7 contain
16.2 kb and 26.3 kb, respectively. While the extent of
PIR4 rearrangement with respect to the putative ancestral
locus (AC073318) is not always a good indicator of degree
of nucleotide sequence identity, it is noteworthy that
similar deletion patterns share monophyletic origins
consistent with the placement within the phylogenetic tree
(see AC128674, AC127384, AC0024500, and AC006359
for an example [fig. 7]). These data not only validate the
phylogeny but also provide insight into different trajecto-
ries of evolutionary duplication, where irreversible de-
letion/rearrangement events tagged a progenitor copy and
its descendants. Finally, the phylogenetic data are
consistent with a strict division of PIR4 sequences into
two clades, an ancestral (clade B), which contains the
putative human donor locus as well as representatives from
each great ape species, and derivative clade (clade A),
which contains only chimpanzee and human copies of this
PIR4 Duplication Timing
Since gibbon and orangutan lineages contain only
four copies of PIR4, whereas chimpanzee and human have
20 and 40 copies, respectively, the data suggest that
a major transpositional burst of PIR4 sequences likely
occurred before the divergence of the African great ape
and human lineage (5 to 8 MYA). To determine a more
precise estimate of when the burst of PIR4 duplications
occurred, we first calculated the specific neutral sub-
stitution rate for this duplication since previous molecular
clock estimates among primates have varied greatly (1 3
mutations/site/year for human-chimpanzee versus 2
3 10
mutations/site/year for human-lemur comparisons
[Liu et al. 2003]). Using gibbon sequences as an outgroup
(which diverged from the lineage leading to humans ap-
proximately 18 MYA) (Goodman 1999) and the average
K value (0.068) between all gibbon and all human
sequences, the rate of neutral substitution for this repeat is
FIG. 7.—Genomic structure of PIR4 sequences. The genomic structure of 47 human PIR4 loci with respect to the putative ancestral locus
(AC073318) is depicted using the program PARASIGHT (Bailey, unpublished). Briefly, 50 kb of the ancestral copy of PIR4 (AC073318) is
represented by the top horizontal line. Regions with significant sequence homology are shown below using gray rectangles to indicate sequence identity
to AC073318 (black rectangles represent segments 100% identical, dark gray rectangles represent .94% identity, and light gray rectangles represent
,94% identity). Some accessions (AC093787) share considerable sequence in common (.40 kb for AC073318), whereas others (AL138715) share
much less (13 kb with AC073318). The vertical gray lines through the middle indicates the approximate position of the approximately 1-kb tree (rooted
on gibbon), which is provided on the left side of the schematic for comparison.
Expansion of Human Centromeric Segmental Duplications 1475
by guest on June 1, 2013 from
1.89 3 10
(þ/0.17 3 10
)(r¼ k/2T). Similarly,
calculating the rate based on chimpanzee 145P5 and its
inferred ortholog AC073318 gave a result of 1.75 3 10
(þ/0.7 3 10
) (based on a separation of 6 Myr between
human and chimpanzee [Goodman 1999]), and the rate
determined using the orthologous orangutan BACs (1J18
and 220O24) is 1.5 3 10
(þ/0.25 3 10
). The higher
substitution rates of 1.89 3 10
and 1.75 3 10
seen for
this repeat agree with the 2.1 3 10
estimated for the Old
World monkey comparison of the CAGGG (Eichler,
Archidiacono, and Rocchi 1999) and could indicate that
pericentromeric repeats and other sequences devoid of
genes may have different rates of substitution than
previously determined for noncoding sequences (Li and
Tanimura 1987).
Using the average rate (between 1.5 3 10
, 1.75 3
, and 1.89 3 10
)of1.713 10
year, the most divergent human sequences in figure 2
(AC127701B and AC073318) have a K value of 0.087,
suggesting approximately 25 Myr of change between
them. Interestingly, the phylogenetic tree in figure 2 has
some sequences clustered together (e.g., 7cos43b6 and
7cos33f1), further suggesting recent intrachromosomal
duplication or conversion events. This suggests that
whereas some PIR4 copies have existed for over 20
Myr of evolution, others have arisen recently and the
process of PIR4 duplication may be ongoing. However,
the mean genetic distance between all human sequences
is 0.047 (Kimura’s estimate), suggesting that many
duplications occurred 14 MYA ( just before the di-
vergence of humans from our great ape ancestors) (fig.
6). Surprisingly, sequences from chromosomes 2, 7, 13,
16, and 22 are found in both clades of the tree,
suggesting very different evolutionary histories exist on
the same chromosome.
Consequences of PIR4 Duplications
Of the pericentromeric regions identified in this study,
most harbor multiple copies of PIR4 (table 1). Based on
available data within GenBank, it appears that intra-
chromosomal copies of PIR4 are separated by at least 100
to 150 kb, as BAC clones rarely contain two distinct
elements. Based on their high identity and close proximity
within pericentromeric regions, PIR4 elements have the
potential to undergo gene conversion. This is supported in
part by our analysis of cosmid and BAC PSV signatures
that we were sometimes unable to match to one another,
suggesting that PIR4 elements have rapidly diverged
between individuals or have been effectively deleted
within the population (see table 3 in Supplementary
Material online). In some cases, such as chromosomes 9p/
9q12 and 2p/2q21, individual copies of PIR4 may be
separated by multiple Mb as evidenced by distinct
metaphase FISH signals. This organization is presumably
due to recent evolutionary centromeric rearrangements that
have occurred within these two specific chromosomes
(Baldini et al. 1993). The organization of these intra-
chromosomal copies of PIR4 is reminiscent of low-copy
repeat (LCRs) sequences that have been implicated in
chromosomal instability associated with more than two
dozen genomic disorders. It is possible that intrachromo-
somal PIR4 sequences separated by hundreds of kb could
similarly facilitate nonhomologous recombination events,
leading to secondary deletions, duplications, and inver-
sions (Bailey et al. 2002a; Stankiewicz and Lupski 2002).
Such dynamic mutational events, if they exist, might
account for the considerable heteromorphism observed for
these regions of the genome (Buiting et al. 1992; Barber et
al. 1998, 1999). Although the clinical and evolutionary
significance of such germline/somatic instability is un-
known, it is noteworthy that many of the same
pericentromeric regions containing PIR4 duplications (1,
2, 8, 9, 14, 15, 16, 17, 18, 21, and 22) are regions
associated with common breakpoints in solid tumor cell
lines, suggesting that the presence of PIR4 may be
associated with somatic instability (Padilla-Nash et al.
2001). Finally, the unusual architecture of PIR4 repeats on
chromosomes 2 and 9 could help explain the high
frequency of large-scale inversions. Chromosome 9
inversion events are the most common karyotype variation
seen in humans, and chromosome 2 inversion events are
the second most commonly diagnosed events (Kaiser
1984). Although PIR4 has not yet been directly implicated
in these common rearrangements, its existence in many
regions of instability necessitates a more thorough in-
vestigation of the genomic architecture.
Based on BAC end-sequencing data as well as large-
scale sequencing of PIR4-containing clones, we estimate
that approximately 25% of PIR4 copies abut large tracts
(.10 kb) of satellite repeat sequences (alpha, HSATII,
etc). Such repetitive sequences have been postulated to
play a pivotal role in the recent nonhomologous exchanges
that have dynamically shaped human pericentromeric
regions (Mashkova et al. 1998; Guy et al. 2000; Horvath,
Schwartz, and Eichler 2000), as they often demarcate the
boundaries of large-scale interchromosomal duplications.
The proximity of PIR4 sequences to blocks of satellite may
have contributed to their proliferation within the human
genome. Of the chromosomes known to contain pericen-
tromeric duplications (Bailey et al. 2001; Cheung et al.
2001), detailed pericentromeric analyses have only been
conducted for chromosomes 2, 10, and 16 and the
completely sequenced chromosomes 14, 20, 21, and 22
(Dunham et al. 1999; Jackson et al. 1999; Hattori et al.
2000; Horvath et al. 2000; Horvath, Schwartz, and Eichler
2000; Deloukas et al. 2001; Heilig et al. 2003). Our
analysis predicts that many duplicon-rich pericentromeric
regions, such as chromosomes 1, 5, 7, 9, 11, 12, 13, 15, 17,
18, and Y, still remain uncharacterized with respect to the
full extent of their duplicated architecture. Interestingly,
even among chromosomes that have been deemed
completed (21, 22, and 14), our analysis has identified
additional clones that have not yet been sequenced.
Presumably, these clones map centromerically to the most
proximal sequence within the sequence assembly.
Using PIR4 to Fill Genome Gaps
Utilizing PIR4 as a marker of pericentromeric DNA,
we have used paralogous-sequence tagging to begin to
successfully map approximately 40% of these relatively
1476 Horvath et al.
by guest on June 1, 2013 from
intractable regions of the genome. In addition, our analysis
recovered additional candidate clones for targeted se-
quencing. As part of a collaboration with the Washington
School of Medicine Genome Sequencing Center, we have
submitted an additional 15 RPCI-11 BAC clones whose
sequence signature did not match an accession within
the NCBI database (at least three variants over 950 bp of
sequence analyzed) (AC127362, AC127380, AC127381,
AC127384, AC127387, AC127389, AC127391,
AC127701, AC128674, AC128676, AC128677,
AC129338, AC129778, AC129779, and AC129782). In
collaboration with Oklahoma’s Advanced Center for
Genome Technology, we have sequenced one RPCI-11
BAC clone (AC092854) as well cosmid clones from
chromosome 22 (AC093314, AC103582, and AC093091).
These clones have effectively added over 2 Mb of human
pericentromeric sequence to GenBank, although their
integration into the final human genome assembly is still
ongoing. In cases where chromosome-assigned pericen-
tromeric clones have been dropped during the assembly
process, we are working with the sequence community to
ensure that such clones are reincorporated into the minimal
tiling path of the final human genome sequence. Although
it is unlikely that complete closure of these regions will be
achieved by the finish target date (2003), these sequences
should provide valuable anchor points from which to seed
future mapping, sequencing and assembly.
Is the additional effort within these regions war-
ranted? Although biological and evolutionary arguments
may be easily mustered, the primary motivation of the
Human Genome Project has been to identify all genes
within the context of its genomic sequence (Collins et al.
1998). Many pericentromeric regions have been recalci-
trant to closure due to their unusual duplication archi-
tecture. Pericentromeric genes embedded within these
highly duplicated regions have been difficult to identify
because of a lack of available sequence, difficulties in as-
sembly of underlying genomic DNA, and/or ambiguities
of paralogous gene annotation. Furthermore, pericentro-
meric regions have been operationally classified as
heterochromatic DNA, since they are located in the
vicinity of centromeres. As such, they are considered
gene-poor genomic environments. Although heterochro-
matin is typically devoid of transcription presumably due
to its compact nature (Dillon and Festenstein 2002; Donze
and Kamakaka 2002), several recent studies have
challenged the notion that DNA sequence in the vicinity
of heterochromatic DNA is transcriptionally silent. For
example, a mammalian artificial chromosome study
indicated that a gene placed in close proximity to and
between centromeric and telomeric satellites can still be
readily expressed (Bayne et al. 1994). Within Drosophila,
essential genes such as the MAP-kinase were recovered
embedded within satellite sequences (Adams et al. 2000).
Similarly, recent articles by Crosier et al. (2002) and
Bailey et al. (2002b) provide strong evidence of human
transcripts from pericentromeric regions on chromosomes
2 and 22. Our own analysis of 89 GenBank accessions
containing (.5 kb) PIR4 reveals that 31 of these genomic
sequences contain at least one transcript (exon-intron
structure over at least two exons .99% identity to an EST).
Seven out of these 31 accessions also contain tracts of
satellite sequence (.3 kb). These transcripts map to 19
different Unigene clusters that have been assigned to
chromosomes 2, 7, 9, 10, 16, and 22. Although these data
do not prove the existence of pericentromerically located
genes associated with PIR4 in humans, they do suggest
transcriptional potency of these genomic regions. This
underscores the importance of complete human genome
sequence and assembly up to the higher order alpha
satellite arrays to provide a comprehensive transcription
and, ultimately, a gene map of the human genome.
We thank Laurie Christ, Jason Carter, and Andrew
Grow for technical assistance and Dr. Norman Doggett for
kindly providing access to chromosome 16 cosmid filters
and clones. This work was supported by grants NIH
GM58815, NIH HG002385, DOE ER62862 to E.E.E. and
NIH HG02152 to B.A.R., the financial support of
Telethon, CEGBA (Centro di Eccellenza Geni in campo
Biosanitario e Agroalimentare), MIUR (Ministero Italiano
della Universita’ e della Ricerca), European Commission
(INPRIMAT, QLRI-CT-2002-01325) to N.A. and M.R.
J.E.H. was supported in part by NIH GM08613, Genetics
Training grant and J.A.B. was supported by NIH Career
Development Program in Genomic Epidemiology of
Cancer (CA094816) and Medical Scientist Training Grant.
Literature Cited
Adams, M. D., S. E. Celniker, R. A. Holt et al. (194 co-authors).
2000. The genome sequence of Drosophila melanogaster.
Science 287:2185–2195.
Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J.
Lipman. 1990. Basic local alignment search tool. J. Mol. Biol.
Arnold, N., J. Wienberg, K. Emert, and H. Zachau. 1995.
Comparative mapping of DNA probes derived from the Vk
immunoglobulin gene regions on human and great ape
chromosomes by fluorescence in situ hybridization. Genomics
Assum, G., T. Fink, C. Klett, B. Lengl, M. Schanbacher, S. Uhl,
and G. Wohr. 1991. A new multisequence family in human.
Genomics 11:34–41.
Bailey, J. A., Z. Gu, R. A. Clark, K. Reinert, R. V. Samonte,
S. Schwartz, M. D. Adams, E. W. Myers, P. W. Li, and
E. E. Eichler. 2002a. Recent segmental duplications in the
human genome. Science 297:1003–1007.
Bailey, J. A., A. M. Yavor, H. F. Massa, B. J. Trask, and
E. E. Eichler. 2001. Segmental duplications: organization and
impact within the current human genome project assembly.
Genome Res. 11:1005–1017.
Bailey, J. A., A. M. Yavor, L. Viggiano, D. Misceo,
J. E. Horvath, N. Archidiacono, S. Schwartz, M. Rocchi,
and E. E. Eichler. 2002b. Human-specific duplication and
mosaic transcripts: the recent paralogous structure of
chromosome 22. Am. J. Hum. Genet. 70:83–100.
Baldini, A., T. Ried, V. Shridhar, K. Ogura, L. D’Aiuto,
M. Rocchi, and D. C. Ward. 1993. An alphoid DNA sequence
conserved in all human and great ape chromosomes: evidence
for ancient centromeric sequences at human chromosomal
regions 2q21 and 9q13. Hum. Genet. 90:577–583.
Expansion of Human Centromeric Segmental Duplications 1477
by guest on June 1, 2013 from
Barber, J. C., I. E. Cross, F. Douglas, J. C. Nicholson, K. J.
Moore, and C. E. Browne. 1998. Neurofibromatosis pseudo-
gene amplification underlies euchromatic cytogenetic dupli-
cations and triplications of proximal 15q. Hum. Genet.
Barber, J. C., C. J. Reed, S. P. Dahoun, and C. A. Joyce. 1999.
Amplification of a pseudogene cassette underlies euchromatic
variation of 16p at the cytogenetic level. Hum. Genet.
Bayne, R. A., D. Broccoli, M. H. Taggart, E. J. Thomson, C. J.
Farr, and H. J. Cooke. 1994. Sandwiching of a gene within 12
kb of a functional telomere and alpha satellite does not result
in silencing. Hum. Mol. Genet. 3:539–546.
Buiting, K., V. Greger, B. H. Brownstein, R. M. Mohr, I.
Voiculescu, A. Winterpacht, B. Zabel, and B. Horsthemke.
1992. A putative gene family in 15q11-13 and 16p11.2:
possible implications for Prader-Willi and Angelman syn-
dromes. Proc. Natl. Acad. Sci. USA 89:5457–5461.
Cheung, V. G., N. Nowak, W. Jang et al. (65 co-authors). 2001.
Integration of cytogenetic landmarks into the draft sequence
of the human genome. The BAC Resource Consortium.
Nature 409:953–958.
Collins, F. S., A. Patrinos, E. Jordan, A. Chakravarti, R. Gesteland,
and L. Walters. 1998. New goals for the U.S. Human genome
project: 1998–2003. Science 282:682–689.
Copenhaver, G. P., K. Nickel, T. Kuromori et al. (14 co-authors).
1999. Genetic definition and sequence analysis of Arabidopsis
centromeres. Science 286:2468–2474.
Crosier, M., L. Viggiano, J. Guy et al. (11 co-authors). 2002.
Human paralogs of KIAA0187 were created through in-
dependent pericentromeric-directed and chromosome-specific
duplication mechanisms. Genome Res. 12:67–80.
Deloukas, P., L. H. Matthew, J. Ashurst et al. (127 co-authors).
2001. The DNA sequence and comparative analysis of human
chromosome 20. Nature 414:865–871.
Dillon, N., and R. Festenstein. 2002. Unravelling heterochro-
matin: competition between positive and negative factors
regulates accessibility. Trends Genet. 18:252–258.
Donze, D., and R.T. Kamakaka. 2002. Braking the silence: how
heterochromatic gene repression is stopped in its tracks.
Bioessays 24:344–349.
Dunham, I., N. Shimizu, B. A. Roe et al. (25 co-authors). 1999. The
DNA sequence of human chromosome22. Nature 402:489–495.
Eichler, E., N. Archidiacono, and M. Rocchi. 1999. CAGGG
repeats and the pericentromeric duplication of the hominoid
genome. Genome Res. 9:1048–1058.
Eichler, E. E. 2001a. Recent duplication, domain accretion and
the dynamic mutation of the human genome. Trends Genet.
———. 2001b. Segmental duplications: What’s missing, mis-
assigned, and misassembled—and should we care? Genome
Res. 11:653–656.
Eichler, E. E., M. L. Budarf, M. Rocchi, L. L. Deaven, N. A.
Doggett, A. Baldini, D. L. Nelson, and H. W. Mohrenweiser.
1997. Interchromosomal duplications of the adrenoleukodys-
trophy locus: a phenomenon of pericentromeric plasticity.
Hum. Mol. Genet. 6:991–1002.
Eichler, E. E., F. Lu, Y. Shen, R. Antonacci, V. Jurecic,
N. A. Doggett, R. K. Moyzis, A. Baldini, R. A. Gibbs, and
D. L. Nelson. 1996. Duplication of a gene-rich cluster
between 16p11.1 and Xq28: a novel pericentromeric-directed
mechanism for paralogous genome evolution. Hum. Mol.
Genet. 5:899–912.
Flint, J., K. Thomas, G. Micklem, H. Raynham, K. Clark, N.
Doggett, A. King, and D. Higgs. 1997. The relationship
between chromosome structure and function at a human
telomeric region. Nature Genet. 15:252–257.
Goodman, M. 1999. The genomic record of humankind’s
evolutionary roots. Am. J. Hum. Genet. 64:1–39.
Guy, J., C. Spalluto, A. McMurray et al. (15 co-authors). 2000.
Genomic sequence and transcriptional profile of the boundary
between pericentromeric satellites and genes on human
chromosome arm 10q. Hum. Mol. Genet. 9:2029–
Hattori, M., A. Fujiyama, T. D. Taylor et al. (26 co-authors).
2000. The DNA sequence of human chromosome 21. The
chromosome 21 mapping and sequencing consortium. Nature
Heilig, R., R. Eckenberg, J. L. Petit et al. (98 co-authors). 2003.
The DNA sequence and analysis of human chromosome 14.
Nature 421:601–607.
Higgins, D. G., J. D. Thompson, and T. J. Gibson. 1996. Using
Clustal for multiple sequence alignments. Methods Enzymol.
Horvath, J., S. Schwartz, and E. Eichler. 2000. The mosaic
structure of a 2p11 pericentromeric segment: a strategy for
characterizing complex regions of the human genome.
Genome Res. 10:839–852.
Horvath, J., L. Viggiano, B. Loftus, M. Adams, M. Rocchi, and
E. Eichler. 2000. Molecular structure and evolution of an
alpha/non-alpha satellite junction at 16p11. Hum. Mol. Genet.
Horvath, J. E., J. A. Bailey, D. P. Locke, and E. E. Eichler. 2001.
Lessons from the human genome: transitions between eu-
chromatin and heterochromatin. Hum. Mol. Genet. 10:2215–
IHGSC (International Human Genome Sequencing Constortium).
2001. Initial sequencing and analysis of the human genome.
Nature 409:860–921.
ISCN (International System of Chromosome Nomenclature).
1985. Report of the standing committee on human cytogenetic
nomenclature. Birth Defects 21:1–117.
Jackson, M. S., M. Rocchi, G. Thompson et al. (13 co-authors).
1999. Sequences flanking the centromere of human chromo-
some 10 are a complex patchwork of arm-specific sequences,
stable duplications and unstable sequences with homologies to
telomeric and other centromeric locations. Hum. Mol. Genet.
Kaiser, P. 1984. Pericentric inversions. Problems and significance
for clinical genetics. Hum. Genet. 68:1–47.
Kimura, M. 1980. A simple method for estimating evolutionary
rates of base substitutions through comparative studies of
nucleotide sequences. J. Mol. Evol. 16:111–120.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001.
MEGA2: molecular evolutionary genetics analysis software.
Bioinformatics 17:1244–1245.
Li, W. 1997. Molecular evolution. Sinauer Associates, Sunder-
land, Mass.
Li, W. H., and M. Tanimura. 1987. The molecular clock runs
more slowly in man than in apes and monkeys. Nature
Liu, G., S. Zhao, J. A. Bailey, S. C. Sahinalp, C. Alkan, E.
Tuzun, E. D. Green, and E. E. Eichler. 2003. Analysis of
primate genomic variation reveals a repeat-driven expansion
of the human genome. Genome Res. 13:358–368.
Loftus, B., U. Kim, V. Sneddon et al. (19 co-authors). 1999.
Genome duplications and other features in 12 Mbp of DNA
sequence from human chromosome 16p and 16q. Genomics
Mahtani, M. M., and H. F. Willard. 1998. Physical and genetic
mapping of the human X chromosome centromere: repression
of recombination. Genome Res. 8:100–110.
Mashkova, T., N. Oparina, I. Alexandrov, O. Zinovieva, A.
Marusina, Y. Yurov, M. H. Lacroix, and L. Kisselev. 1998.
1478 Horvath et al.
by guest on June 1, 2013 from
Unequal cross-over is involved in human alpha satellite DNA
rearrangements on a border of the satellite domain. FEBS Lett.
Myers, E. W., and W. Miller. 1988. Optimal alignments in linear
space. Comput. Appl. Biosci. 4:11–17.
Orti, R., M. C. Potier, C. Maunoury, M. Prieur, N. Creau, and J.
M. Delabar. 1998. Conservation of pericentromeric duplica-
tions of a 200-kb part of the human 21q22.1 region in
primates. Cytogenet. Cell Genet. 83:262–265.
Padilla-Nash, H. M., K. Heselmeyer-Haddad, D. Wangsa, H.
Zhang, B. M. Ghadimi, M. Macville, M. Augustus, E.
Schrock, E. Hilgenfeld, and T. Ried. 2001. Jumping trans-
locations are common in solid tumor cell lines and result
in recurrent fusions of whole chromosome arms. Genes
Chromosomes Cancer 30:349–363.
Regnier, V., M. Meddeb, G. Lecointre, F. Richard, A. Duverger,
V. C. Nguyen, B. Dutrillaux, A. Bernheim, and G. Danglot.
1997. Emergence and scattering of multiple neurofibromatosis
(NF1)-related sequences during hominoid evolution suggest
a process of pericentromeric interchromosomal transposition.
Hum. Mol. Genet. 6:9–16.
Riethman, H. C., Z. Xiang, S. Paul, E. Morse, X. L. Hu, J. Flint,
H. C. Chi, D. L. Grady, and R. K. Moyzis. 2001. Integration
of telomere sequences with the draft human genome sequence.
Nature 409:948–951.
Stankiewicz, P., and J. R. Lupski. 2002. Genome architecture,
rearrangements and genomic disorders. Trends Genet. 18:
Sun, X., J. Wahlstrom, and G. Karpen. 1997. Molecular structure
of a functional Drosophila centromere. Cell 97:1007–
The Arabidopsis Genome Initiative. 2000. Analysis of the
genome sequence of the flowering plant Arabidopsis thaliana.
Nature 408:796–815.
Trask, B. J., G. van den Engh, M. Christensen, H. F. Massa, J.
W. Gray, and M. Van Dilla. 1991. Characterization of
somatic cell hybrids by bivariate flow karyotyping and
fluorescence in situ hybridization. Somat. Cell Mol. Genet.
Wohr, G., T. Fink, and G. Assum. 1996. A palindromic structure
in the pericentromeric region of various human chromosomes.
Genome Res. 6:267–279.
Yan, C. M., K. W. Dobie, H. D. Le, A. Y. Konev, and G. H.
Karpen. 2002. Efficient recovery of centric heterochromatin
P-element insertions in Drosophila melanogaster. Genetics
Zimonjic, D., M. Kelley, J. Rubin, S. Aaronson, and N. Popescu.
1997. Fluorescence in situ hybridization analysis of keratino-
cyte growth factor gene amplification and dispersion in
evolution of great apes and humans. Proc. Natl. Acad. Sci.
USA 94:11461–11465.
Keith Crandall, Associate Editor
Accepted April 28, 2003
Expansion of Human Centromeric Segmental Duplications 1479
by guest on June 1, 2013 from
    • "We assessed the possible association between regions with significantly different epigenetic panorama in humans when compared with other primates (Shulha et al. 2012 ) and genomic segments around loci that were structurally modified during the recent evolution of the human genome, i.e., HSA2 fusion point and ancestral centromere, as well as HSA1 and HSA18 inversion breakpoints (Yunis et al. 1980; Dennehey et al. 2004; Szamalek et al. 2006 ), with HSA1 also encompassing pericentromeric heterochromatin that is absent in its chimpanzee homolog (Yunis et al. 1980). Additionally, we assessed highly plastic segments such as human-specific segmental duplications (Sudmant et al. 2013) together with subtelomeric and pericentromeric regions (Yunis et al. 1980; Bailey et al. 2001; Bailey et al. 2002; Horvath et al. 2003; The Chimpanzee Sequencing and Analysis Consortium 2005; Linardopoulou et al. 2005; Locke et al. 2005). The human-specific prefrontal cortex H3K4me3-enriched sites significantly accumulate at subtelomeric (fivefold) and pericentromeric sites (threefold) in both the number of peaks (P-value = 4 3 10 À73 and 3 3 10 À15 , respectively; permutation P-value = 0.001) and the amount of base pairs covered (P-value = 2 3 10 À56 and 2 3 10 À13 , respectively; permutation P-value = 0.001) (Fig. 2A,B;Table 1). "
    [Show abstract] [Hide abstract] ABSTRACT: Human and chimpanzee genomes are 98.8% identical within comparable sequence. They however differ structurally in nine pericentric inversions, one fusion that originated human chromosome 2 and content and localization of heterochromatin and lineage-specific segmental duplications. The possible functional consequences of these cytogenetic and structural differences are not fully understood and their possible involvement in speciation remains unclear. We show that subtelomeric regions - that have a species-specific organization, are more divergent in sequence, and are enriched in genes and recombination hotspots - are significantly enriched for species-specific histone modifications that decorate transcription start sites in different tissues in both human and chimpanzee. Human lineage-specific chromosome 2 fusion point and ancestral centromere locus as well as chromosome 1 and 18 pericentric inversion breakpoints showed enrichments of human-specific H3K4me3 peaks in prefrontal cortex. Our results reveal an association between plastic regions and potential novel regulatory elements.
    Full-text · Article · Jun 2014
    • "PCR products were purified with the Roche Diagnostics Corporation High Pure PCR Product Purification Kit, and 25–50 ng of purified DNA was individually labeled with [a-32 P]dCTP using the High Prime DNA labeling Kit (Roche Diagnostics Corporation). BAC library membranes from the black and ring-tailed lemurs (CHORI-273 and LBNL-2, respectively [BACPAC Resources]) were hybridized as described previously (Horvath et al. 2003). Hybridized membranes were imaged for at least 16 h using a PhosphorImager (Amersham Biosciences), and positives were called by hand. "
    [Show abstract] [Hide abstract] ABSTRACT: Here we provide a detailed comparative analysis across the candidate X-Inactivation Center (XIC) region and the XIST locus in the genomes of six primates and three mammalian outgroup species. Since lemurs and other strepsirrhine primates represent the sister lineage to all other primates, this analysis focuses on lemurs to reconstruct the ancestral primate sequences and to gain insight into the evolution of this region and the genes within it. This comparative evolutionary genomics approach reveals significant expansion in genomic size across the XIC region in higher primates, with minimal size alterations across the XIST locus itself. Reconstructed primate ancestral XIC sequences show that the most dramatic changes during the past 80 million years occurred between the ancestral primate and the lineage leading to Old World monkeys. In contrast, the XIST locus compared between human and the primate ancestor does not indicate any dramatic changes to exons or XIST-specific repeats; rather, evolution of this locus reflects small incremental changes in overall sequence identity and short repeat insertions. While this comparative analysis reinforces that the region around XIST has been subject to significant genomic change, even among primates, our data suggest that evolution of the XIST sequences themselves represents only small lineage-specific changes across the past 80 million years.
    Full-text · Article · Jun 2011
    Julie E HorvathJulie E HorvathChristina B SheedyChristina B SheedyStephanie L Merrett+4 more authors ...Huntington F WillardHuntington F Willard
    • "Moreover, our evolutionary reconstruction identified species-specific duplicated sequences on the Y chromosome, especially among the African great apes. Detailed analyses of pericentromeric regions of several human chromosomes (Eichler et al. 1996; Jackson et al. 1999; Horvath et al. 2000 Horvath et al. , 2003 Horvath et al. , 2005 Bailey et al. 2001 Bailey et al. , 2002 Crosier et al. 2002; She et al. 2004; Kirsch et al. 2005; Locke et al. 2005) have illuminated a general principle of human genome evolution: complex mosaics of segmental duplications originating from diverse euchromatic chromosomal regions are created by recurrent duplicative transpositions of diverse euchromatic segments into pericentromeric regions (Jackson et al. 1999; Guy et al. 2000 Guy et al. , 2003 Horvath et al. 2000 Horvath et al. , 2005 Locke et al. 2005). Differential dispersal of larger duplicon cassettes among human and great ape pericentromeric regions subsequently leads to lineage-specific quantitative and qualitative differences. "
    [Show abstract] [Hide abstract] ABSTRACT: Human chromosomal regions enriched in segmental duplications are subject to extensive genomic reorganization. Such regions are particularly informative for illuminating the evolutionary history of a given chromosome. We have analyzed 866 kb of Y-chromosomal non-palindromic segmental duplications delineating four euchromatin/heterochromatin transition regions (Yp11.2/Yp11.1, Yq11.1/Yq11.21, Yq11.23/Yq12, and Yq12/PAR2). Several computational methods were applied to decipher the segmental duplication architecture and identify the ancestral origin of the 41 different duplicons. Combining computational and comparative FISH analysis, we reconstruct the evolutionary history of these regions. Our analysis indicates a continuous process of transposition of duplicated sequences onto the evolving higher primate Y chromosome, providing unique insights into the development of species-specific Y-chromosomal and autosomal duplicons. Phylogenetic sequence comparisons show that duplicons of the human Yp11.2/Yp11.1 region were already present in the macaque-human ancestor as multiple paralogs located predominantly in subtelomeric regions. In contrast, duplicons from the Yq11.1/Yq11.21, Yq11.23/Yq12, and Yq12/PAR2 regions show no evidence of duplication in rhesus macaque, but map to the pericentromeric regions in chimpanzee and human. This suggests an evolutionary shift in the direction of duplicative transposition events from subtelomeric in Old World monkeys to pericentromeric in the human/ape lineage. Extensive chromosomal relocation of autosomal-duplicated sequences from euchromatin/heterochromatin transition regions to interstitial regions as demonstrated on the pygmy chimpanzee Y chromosome support a model in which substantial reorganization and amplification of duplicated sequences may contribute to speciation.
    Full-text · Article · Aug 2008
Show more