Download Neocentromeres and alpha satellite: a proposed structural code for

© 2000 Oxford University Press
Human Molecular Genetics, 2000, Vol. 9, No. 2
149–154
COMMENTARY
Neocentromeres and alpha satellite: a proposed
structural code for functional human centromere DNA
Jørn Koch+
Danish Cancer Society, Department of Cytogenetics, Tage hansens Gade 2, DK-8000 Denmark
Received 1 September 1999; Revised and Accepted 5 November 1999
A comparative analysis of the conserved region of primate alpha satellite DNA and of a human neocentromere is
reported. The two types of centromere DNA share structural features, possibly defining mitotically functional
human centromere DNA. These features consist of double dyad symmetries of a particular size, as well as short
conserved base motifs adjacent to the dyad symmetries. The possible importance of these features is supported
by their presence also in the centromeres of yeast. The overall resemblance between primate alpha satellite DNA
and yeast centromere DNA is further strengthened by the presence of a second shared structure. This structure
consists of a single dyad symmetry incorporating a particular short base sequence. In the mitotically stable human
neocentromere, this second symmetry is not found in connection with the double symmetry. However, the yeast
homologue of the second symmetry is responsible for stability in meiosis I, rather than for mitotic stability. If the
human homolog serves a similar function, it would be expendable, except for reproduction.
INTRODUCTION
The centromere is a complex structural element needed for
proper segregation of eukaryotic chromosomes in two different
types of cell division: meiosis and mitosis. In each of these two
types of division it must perform quite diverse tasks. These
include keeping sister chromatids together until the appropriate
point in the division process, and then either releasing sister
chromatids (mitosis and meiosis II) or not (meiosis I). The
centromere must also provide the anchor points for the spindle
apparatus, which pulls the separating chromosomes to opposite
poles of the dividing cell.
In contrast to a growing body of knowledge on the proteinaceous part of the centromere, little is known about the requirement for a DNA element to be a functional part of the
centromere complex. Obviously, the exact base sequence is not
the determining factor, as centromere DNAs from different
species have no significant sequence similarity. Even within a
single species, it seems that more than one type of DNA may
be able to fulfil the function, as illustrated by a number of
human neocentromeres apparently lacking alpha satellite DNA
(1,2). However, considering that centromeres from one species
are functional inside cells from another species, as evidenced
by the relative stability of somatic cell hybrids, a universal
centromere code should exist. The absence of a centromerespecific base sequence thus merely implies that this class of
DNA should be defined and described differently from coding
DNA sequences. In line with this, physicists have found that
whereas coding DNA sequences are organized like the bits and
bytes in a computer, non-coding DNA is organized like
+Tel:
+45 86183566; Fax: +45 86183695; Email: [email protected]
language (3,4). Rather than trying to identify centromere DNA
through sequence matching, it may thus be appropriate to look
for such characteristics as rhythm (e.g. alternating purine and
pyrimidine elements), or sequence superstructure (e.g.
symmetries). The identification of such hypothetical centromere identifiers would obviously enhance our ability to elucidate the mechanics of centromere function. In fact, given the
present scarcity of data on this topic, the mere suggestion of
such structures might prove of value, as it will enable the
formation of models for molecular interactions at the centromere, and subsequent experimental testing of these models.
As far as knowledge about the structures of human centromere DNA is concerned, the most informative observation
may be the finding that alpha satellite DNA contains a region
of sequence conservation, whereas the remaining part of the
monomer sequence is more variable, possibly serving a spacer
function (5). In this respect, alpha satellite DNA may have
some similarity with budding yeast centromere DNA, which
also comprises conserved elements separated by a spacer (6).
As reported here, this similarity extends to the sequence structure of the conserved elements. Furthermore, a human neocentromere (1) shares certain sequence structures with these two
natural centromere DNAs. Interestingly, the sequence structure and composition of these conserved elements are such that
they could be imagined to form particular DNA structures, e.g.
hairpins or H-form DNA, which could possibly serve as
centromere identifiers. In fact, the centromere-binding proteins
HMGI (alpha protein) and topoisomerase II have already been
reported to bind such structures preferentially (7–9).
150
Human Molecular Genetics, 2000, Vol. 9, No. 2
Figure 1. Organization of centromere elements. (Top) Conserved region of primate alpha satellite DNA (sequence adapted from ref. 5). In each position, the base most
commonly found in that position is indicated. The left part of this sequence comprises a 2-fold dyad symmetry (symmetry 1), the center of which can be cleaved by the
restriction enzyme DdeI, recognizing the sequence CTNAG. The center of the symmetry is marked ‘O’. The inner symmetry is marked ‘IS’. Each wing of the inner
symmetry contains five bases, all of which are either pyrimidines (left wing) or purines (right wing). The beginning and ending of each wing are marked. The
outer symmetry is marked ‘OS’. Each wing of the outer symmetry contains seven bases, all of which are either purines (left wing) or pyrimidines (right wing). In the left
half of symmetry 1, a one base hinge is found between the inner and the outer symmetry, marked ‘H’. To the right of symmetry 1 is the GTGT motif (underlined). To the
right of the GTGT motif is symmetry 2. The center of this symmetry is marked ‘O’. Each wing of symmetry 2 contains seven bases. The TCACA motif (underlined) is
found in the right wing of the symmetry, whereas binding site I for the alpha protein crosses the border between the left wing of the symmetry and the center of it (here:
TGAATTCC). (Middle) Yeast centromere DNA (sequence adapted from ref. 13). In each position, the base most commonly found in that position is indicated. The
nomenclature is the same as used for alpha satellite DNA. Symmetry 2 (CDEI) is now on the left side of the sequence, and the TCACA motif constitutes the left wing and
the center of that symmetry. Symmetry 1 (CDEIII) is on the right side of the sequence, but again comprises a 2-fold dyad symmetry, as also noted by others (6). The inner
symmetry again contains five bases, as does also the outer symmetry in this case, and again each element is made up of only purines or only pyrimidines. In this sequence,
the inner and outer symmetries are in both wings separated by a short hinge. The GT element now occurs to the left of symmetry 1. In comparison with the sequence above,
it looks as if each of symmetry 1, symmetry 2 and the GT element have been rotated 180 on themselves, and then rotated 180 on each other. However, if the yeast
sequence was repeated in tandem, symmetry 2 would follow symmetry 1, as it does in alpha satellite DNA, and the CDEII spacer would assume the position of the variable
region of alpha satellite DNA. This poses the obvious question of whether a precursor of alpha satellite DNA similarly had symmetry 2 at some distance from symmetry
1, their present juxtaposition simply reflecting the tandem duplication of a unit beginning with symmetry 2 and ending with symmetry 1. (Bottom) The chromosome 10
neocentromere (adapted from ref. 1). Two units of the symmetry element are shown in tandem to illustrate how the GTGT motif is sandwiched between dyad symmetries
in the tandem array. Nomenclature as above. As for the yeast centromere, both the inner and the outer symmetries contain five bases, and the inner and the outer symmetries
are in both wings separated by a hinge. In this case, the symmetry elements are not homopurines or homopyrimidines.
RATIONALE
The sequence data used for this analysis come primarily from
three sources. The primate alpha satellite sequence is a crossspecies consensus sequence derived from Koch et al. (5), who
compared sequence data from four different primate species.
Consensus sequences for human alpha satellite DNA alone are
also available (10–12). These display similar structures to the
consensus sequence used here. The human neocentromere
sequence is derived from Barry et al. (1), describing the
sequence analysis of a human chromosome 10 neocentromere,
void of detectable alpha satellite DNA. The budding yeast
centromere sequence is derived from the consensus sequence
for Saccharomyces cerevisiae (13).
In each of the three cases, the consensus sequence has been
analyzed not primarily in terms of exact base sequence, but
rather in terms of overall sequence structure.
PRIMATE ALPHA SATELLITE DNA CONTAINS
COMPLEX CONSERVED ELEMENTS WITH A
SIMILARITY TO KNOWN ELEMENTS OF THE YEAST
CENTROMERE
As previously reported (5), a region of 54 bp is more conserved
between primate species than the remaining part of the 171 bp
alpha satellite monomer. This conservation can be visualized
with the restriction enzyme DdeI, which recognizes a site
within the conserved region (5). Unlike other restriction
enzymes, this enzyme is a universal cutter of genomic alpha
satellite DNA, predominantly reducing it to monomer size,
though leaving a minor fraction of the DNA as short
oligomers.
An even more striking feature of the conserved region is its
surprising complexity. As illustrated in Figure 1, and explained
in the figure legend, the conserved region is composed of two
Human Molecular Genetics, 2000, Vol. 9, No. 2
regions of dyad symmetry. The left region of dyad symmetry
(symmetry 1) is to the right flanked by two copies of the dinucleotide GT, followed by the right region of dyad symmetry
(symmetry 2), flanked by the trinucleotide GTT. Additional
levels of complexity are found, including that symmetry 1
contains an inner and an outer symmetry, and that the elements
in these symmetries are all either homopurines or homopyrimidines.
If such structures are of relevance for centromere function,
they should also appear in centromere DNA from non-primate
chromosomes. Unfortunately, although a number of AT-rich
non-primate satellite DNAs have been sequenced, yeast
species, such as our very remote relative S.cerevisiae, are the
only eukaryotes in which it has been firmly established which
elements carry the centromere function.
The S.cerevisiae centromere contains three elements: CDEI,
CDEII and CDEIII (summarized in refs 6,13) (Fig. 1). CDEII
is a 78–87 bp AT-rich spacer separating CDEI and CDEIII.
CDEI is very similar to symmetry 2 of alpha satellite DNA.
Both regions are dyad symmetries, and both regions contain
the pentanucleotide TCACA inside the dyad symmetry.
CDEIII, in turn, has significant similarity to symmetry 1: both
show a dyad symmetry composed of an inner and an outer
symmetry, both are composed of homopurine and homopyrimidine elements, and both are flanked by an element of alternating G and T residues, although the CDEIII symmetry is
preceded, rather than followed, by it. Indeed, if the yeast
centromere elements were placed in tandem, as the monomers
of alpha satellite DNA are, CDEI would follow CDEIII, just
like symmetry 2 follows symmetry 1. Similarly, the 120 bp
variable region separating the conserved regions in alpha satellite DNA might be the equivalent of the CDEII region in yeast
(a region of relatively fixed size and defined AT richness). The
profound similarity between these two distantly related centromere DNAs argues that the structures found in primate alpha
satellite DNA are likely to be functionally important, not least
as the corresponding yeast elements have been demonstrated to
carry the centromere function in that species.
Unfortunately, the high complexity of the shared structures
also makes their interpretation difficult. The dyad symmetries
could potentially fold into various hairpins. Alternatively, the
homopurine/homopyrimidine segments could be important, as
they could potentially form triplex DNA. Given the alternating
nature of these elements, and their presence on both strands of
all monomers of alpha satellite DNA, they could in principle
bind all monomers tightly to each other through alternate Hform formation (14), thus condensing the centromere through
simple DNA–DNA interactions. The two types of structure
could even be imagined to co-exist, the hairpins interacting via
the formation of triplex DNA. Such H-form formation between
hairpins have been described for oligopurine/oligopyrimidinecontaining dyad symmetries in vitro (15).
The interpretation of the structural similarities between
primates and S.cerevisiae obviously depends on whether such
structures are a common finding in the genome, or whether
they are a rare occurrence. Although intuitively it seems
exceedingly unlikely that such complex structures should exist
by chance in two types of eukaryote this distantly related, their
actual genomic frequency is not easily ascertained. Both for
man and yeast, much of the genome has been sequenced,
enabling computer searches for specific base sequences.
151
However, the proposed centromere structures are only vaguely
defined in terms of base sequence and are therefore not readily
searchable, except manually. If base sequences were entirely
random, a dyad symmetry with five bases in each wing (such
as the inner element of symmetry 1) would occur every 4–5 bp.
The frequency of the outer symmetry would be similar,
assuming a size of five bases in each wing, and 4–7 assuming
seven bases. The frequency of the two symmetries occurring
one inside the other would be the product of these probabilities
(4–10 or 4–12). The frequency of finding the double symmetry
next to a GT motif is of course even lower, and the expected
frequency is also lowered by the requirement for this DNA to
be AT rich. If the symmetry elements also need to be homopurines/homopyrimidines, and at a defined distance from a
symmetry 2 element, even the three gigabases of DNA in the
human genome would not be enough to make a chance occurrence anywhere in the genome likely.
It might, of course, be that such structures are a common
product of the processes forming satellite DNA (16). However,
none of the other major human satellite DNAs (satellites I, II
and III) contains these structures in the consensus sequence
(17). Satellite III is rather asymmetrical, having most of the G
residues in one strand and most of the C residues in the other.
The basic structure of this satellite thus makes the occurrence
of dyad symmetries rather unlikely. Satellite II is more degenerate and less asymmetrical, but also void of relevant dyad
symmetries. Satellite I contains long stretches of pure A and T
residues. These form single symmetries of the size in question
here, but not double symmetries, and the symmetries are not
flanked by a GT element. Considering the abundance of these
classical satellites, it cannot be excluded that an individual unit
could have more elaborate symmetries. Similar conclusions
apply to the major satellite in the mouse (18), another highly
repeated DNA not supporting centromere function. The mouse
major satellite sequence contains abundant homopurine/
homopyrimidine elements. However, as the homopurine
elements are primarily found in one strand, and the homopyrimidine elements in the other, elaborate symmetry elements
would not be a likely finding.
THE CHROMOSOME 10 NEOCENTROMERE DNA
SHARES SOME OF THE STRUCTURAL FEATURES
OF ALPHA SATELLITE DNA
Searching an 80 kb neocentromere from human chromosome
10 for AT-rich DNA repeats, Barry et al. (1) found no alpha
satellite DNA, but instead found 34 other tandem repeats,
including 21 tandem copies of a 28 bp repeat (AT28).
Comparing the AT28 repeat to alpha satellite DNA, they found
no similarities between the two DNA sequences. Interestingly,
if the AT28 sequence is phased as illustrated in Figure 1, and is
analyzed for sequence structure rather than base sequence, a
striking similarity to alpha satellite DNA becomes evident.
Like symmetry 1, one monomer of the AT28 repeat consists of
an inner and an outer dyad symmetry flanked on the right by
two copies of the dinucleotide GT. Two of these repeat
elements in between them form a virtually perfect copy of the
conserved region of alpha satellite DNA. The shared structure
thus comprises the GTGT sequence adjacent to, or sandwiched
between, regions of dyad symmetry.
152
Human Molecular Genetics, 2000, Vol. 9, No. 2
Table 1. Centromere elements (CDEII) in two species of budding yeast (13,19)
Species
Centromere sequence
Saccharomyces cerevisiae
TGTTTTTGTTTTCCGAAAATTAAAAA
Candida glaberata
TGGGTAAGAGTTCCGAAATAGAATCTTA
Inner and outer symmetries are underlined, and the seven bases at the center of the
symmetry and the TG motif are in bold. These motifs seem most stringently
preserved and were found to be necessary for centromere function in a deletion
experiment (19). In the human alpha satellite, these seven positions are
TCTCAGA (12), a single inversion of the first and second bases in each wing thus
apparently separates these two motifs. Note the variable size of the hinge region
between the inner and outer symmetries, and that in both species the GT element
is 11 bases from the center of the symmetry. In the alpha satellite this distance is
12 bases.
An interesting feature of the AT28 repeat is that the dyad
symmetry is not made from homopurine/homopyrimidine
units, thus possibly indicating that this additional feature of
symmetry 1 is not a necessity for human mitotic centromere
function. Instead, the requirement seems to be for a 2-fold
symmetry with an inner element of five bases and an outer
element of at least five bases. This conclusion obviously has
the caveat that it cannot be excluded that the actual centromere-forming DNA in this neocentromere is found outside the
80 kb sequenced.
It could be proposed that the structure of AT28 might be a
common product of the processes forming variable number of
tandem repeats (VNTRs). Furthermore, if such structures are
necessary for centromere formation, they should exist at a
number of sites throughout the genome, as neocentromeres can
be formed at such sites (1,2), and VNTRs could thus constitute
a recruitable source of spare centromeres. However, such
structures have not been found in other VNTRs, which is in
line with the fact that of the 34 VNTRs within the 80 kb fragment, only AT28 has the structure in question. Some of these
other VNTRs have single symmetries or GT elements, but
none has the right size double symmetry next to a GT element.
RELATIONSHIP OF THE DISCUSSED STRUCTURES
TO OTHER CENTROMERE DNAS AND SUGGESTED
CENTROMERE DNAS
Comparing sequence data from different yeast species, and
different chromosomes within a particular species, further
patterns appear. First of all, a comparison of data from all 16
chromosomes in S.cerevisiae (13) reveals minor variations in
the base sequence of the CDEIII element, although some positions seem invariable (e.g. the central part of the symmetry). A
similar pattern of absolute and relative conservation is found
when comparing S.cerevisiae with Candida glaberata (19)
(Table 1). This could be taken to indicate that centromeric
function in yeast is absolutely dependent on the strict conservation of these particular base positions, whereas some variation is allowable in the remainder of the structure. Such an
interpretation is in agreement with the findings of Kitada et al.
(19). Studying the effects of mutation and deletion of various
parts of the C.glaberata centromere, Kitada et al. (19) found
that mutating any of the two bases in the TG dinucleotide next
to the dyad symmetry in CDEIII completely abolished any
centromere function. A similar dramatic effect was seen if the
sequence at the center of the symmetry was changed. Mutating
the sequence of any other part of the centromere elements had
much less drastic effects, if any. Curiously, in spite of the
absence of long-range sequence similarities, the stringently
conserved central motif in budding yeast would be similar to
the corresponding element in human alpha satellite DNA if the
first two bases in each wing were inverted (Fig. 1, Table 1).
It may be worth noting that a major contributor to the seemingly lower conservation in the outer parts of the symmetry is
the plasticity of the hinge regions, which seem rather variable,
both in size and composition. This plasticity argues against the
view that the inner and outer symmetry elements should simultaneously bind a particular interacting partner. Instead they
could bind this partner alternately, bind two different partners
or serve different functions.
The fission yeast Schizosaccharomyces pombe has a centromeric organization which is quite different from the yeast
species discussed above. In S.pombe, the centromere is a
complex structure with a size of 40–100 kb, corresponding to
1–3% of the total chromosome length (described in ref. 20).
The centromere contains several distinct elements, each
several kilobases long. The so-called B, J, K and L elements
are organized in a strictly symmetrical fashion around a central
element named cc. In this organization, the S.pombe centromere has a peculiar resemblance to the symmetries discussed
above, except that the size is measured in kilobases rather than
in bases. However, it may be difficult to comprehend that the
overall structure should serve similar functions if it varies in
size by three orders of magnitude, especially if this function is
to bind certain centromere proteins specifically. Indeed, it
appears from deletion experiments that the cc element and one
K element in combination are sufficient for mitotic function
(20). The natural symmetry is thus not a functional requirement. Within the K and cc elements, neither any substantial
match with the Saccharomyces cerevisiae centromere nor any
symmetry structures have been reported. A possible explanation for this is that, whereas primates and budding yeast centromere DNA contain compact complex elements providing a full
complement of centromere functionalities, Schizosaccharomyces pombe may have these functionalities dispersed on
simpler elements, some residing in the K element and some in
the cc element. This interpretation suggests that it may be
worth noting that, although the AT28 repeat itself contains no
homopurine/homopyrimidine elements, at least half of the
VNTRs in the chromosome 10 neocentromere contain one or
more such element(s), placing the AT28 repeat in an environment rich in homopurine/homopyrimidine elements.
Among the higher eukaryotes, the mouse is the genus outside
primates where most information about centromeres has been
obtained. It is generally believed that the mouse minor satellite
is the centromere DNA, and thus the mouse equivalent of the
primate alpha satellite. There are, however, data arguing
against this view, including that primary constrictions form
unrelated to the minor satellite (21), and that the minor satellite
varies considerably in sequence among mouse strains (22).
Furthermore, the minor satellite is found in only a subset of
mouse species (23). Even in the species where it is present, no
consistent localization is apparent, the minor satellite residing
in the centromere of some species, but being distributed
throughout the chromosome in other species (23). Finally, the
minor satellite has a patchwork structure, part of the sequence
Human Molecular Genetics, 2000, Vol. 9, No. 2
153
Table 2. Satellite DNAs from the centromeric region in five species of higher eukaryotes
Species
Satellite DNA sequence
Mouse minor
AAACATGGAAAATGATGAAAACCACACT
Sea cucumber
GTAACTTTTAAAAGTTCCACCATTTCAAGTACCTTT
Parrot
GRACTTTRAGGTGGGAAARCRARGAGYGAGAG
Domestic dog
TCCTTAGCTCTCACAGGGAAGCATGTCTTGAAGG
Chironimus pallidivittatus
TTTTTTTTAAAGGCTTTGGAGCA(---)GTCTCTGTGGTAATATC
None of these satellite DNAs have been firmly established as centromere forming. Homopurine/
homopyrimidine motifs of at least four bases are underlined. The mouse minor satellite sequence is
from the region of homology with the major satellite [positions 205–231 (24)]. Outside the homology
region, the mouse minor satellite is less asymmetrical, but contains elements similar to those displayed
here and no extensive dyad symmetries. The sea cucumber sequence is from Holothuria polii, but a
similar repeat is found in H.tubulosa (25). The sequence segment chosen contains the only major dyad
symmetry found in this satellite (bold). It is only a single symmetry, although it could in theory be
considered a double symmetry with hinge regions of zero bases. It is also flanked by a GT element
(italics) and the overall AT content of the satellite is 66.8%, very close to the AT content of alpha
satellite DNA. The parrot sequence is a consensus sequence from 10 species, and has similarity to
sequences from crane, falcon and flamingo (26). This shared bird sequence, like the mouse satellite,
predominantly has the purines in one strand and the pyrimidines in the other. The dog sequence has
70–80% similarity to the corresponding satellite DNA from the grey wolf (27). The centromeric
satellite of C.pallidivittatus (28) contains a dyad symmetry (bold) looking like a mirror-reversed copy
of the central yeast element (TTCCGAA turned into AAGGCTT, the difference between a central G or
a central C residue only reflecting which strand is considered). This symmetry resides within a
homopurine/homopyrimidine region (underlined). At the right border of the symmetry a TG
dinucleotide is found. Elsewhere in each repeating unit another peculiar element is found. The left part
of this element has a stretch of six bases (TCTGTG, italics) with a complete match to the outer six
bases in the left wing of the primate symmetry 2 (Fig. 1). The right part of the C.pallidivittatus element
has a six base complete match (AATATC, italics) with binding site III of the alpha protein. This
protein has been found to bind the alpha satellite at three specific sites that would be brought into
mutual proximity by DNA folding in the nucleosome (29), and to recognize any stretch of six A:T base
pairs in duplex DNA (30). In the alpha satellite, the left wing of symmetry 2 is flanked on the right by
binding site I of the alpha protein, followed by the right wing of symmetry 1 (Fig. 1). In
C.pallidivittatus, no obvious symmetries are found in this region.
being similar to the major satellite and part of it not (24).
Looking at the actual sequence of the minor satellite, no relevant symmetry structures are apparent (Table 2). Indeed, the
minor satellite mimics the major satellite to a large extent in
being rather asymmetrical, having most of the G residues in
one strand, and most of the C residues in the other, making the
occurrence of dyad symmetries unlikely. The only obvious
similarity between the mouse minor satellite and the alpha
satellite is the presence of oligopurine/oligopyrimidine
elements, predominantly containing A residues in the purine
units and predominantly T residues in the pyrimidine units.
However, these units do not exhibit the orderly organization
found in symmetry 1 of the alpha satellite. Such elements are a
common finding in satellite DNAs from centromeric regions in
general, including mouse major satellite and satellite DNA
from echinoderms (25), parrots (26) and dogs (27) (Table 2).
Actually, the most striking similarity to the centromere
elements of primates and budding yeast is found not in the
mouse minor satellite, but in a centromeric satellite DNA from
the dipteran Chironimus pallidivittatus (28). This DNA is
restricted to the centromeric regions of all chromosomes, and
consists of tandem repetitions of an AT-rich 155 bp unit. Individual units contain several elements with similarities to established centromere DNAs (Table 2).
In summary, centromere DNAs outside primates and
budding yeast are generally both ill-defined and ill-described,
and thus difficult to interpret. Although this is frustrating, it is
not surprising considering the complexity of centromeres and
the seeming plasticity of centromere DNA. However, the identification of candidate elements from known centromeres
allows specific hypotheses regarding their functional interaction to be generated. Through the experimental testing of such
possible interactions, the partners of the DNA elements can
probably be identified and characterized. Known partners will,
in turn, allow centromere DNA from other species to be identified, thus propagating the initial findings beyond primates
and yeast.
REFERENCES
1. Barry, E.B., Howman, E.V., Cancilla, M.J., Saffery, R. and Choo, K.H.A.
(1999) Sequence analysis of an 80 kb human neocentromere. Hum. Mol.
Genet., 8, 217–227.
2. Wandall, A., Tranebjærg, L. and Tommerup, N. (1998) A neocentromere
on human chromosome 3 without detectable alpha satellite DNA forms
morphologically normal kinetochores. Chromosoma, 107, 359–365.
3. Mantegna, R.N., Buldyrev, S.V., Goldberger, A.L., Havlin, S., Peng, C.K., Simons, S. and Stanley, H.E. (1994) Linguistic features of noncoding
DNA sequences. Phys. Rev. Lett., 73, 3169–3172.
4. Flam, F. (1994) Hints of language in junk DNA. Science, 266, 1320.
5. Koch, J., Kolvraa, S., Gregersen, N. and Bolund, L. (1989) Oligonucleotide-priming methods for the chromosome-specific labeling of alpha satellite in situ. Chromosoma, 98, 259–265.
6. Clarke, L. (1990) Centromeres of budding and fission yeasts. Trends
Genet., 6, 150–154.
7. Pognan, F. and Paoletti, C. (1992) Does cruciform DNA provide a recognition signal for DNA-topoisomerase II? Biochimie, 74, 1019–1023.
8. Hill, D.A. and Reeves, R. (1997) Competition between HMG-I (Y),
HMG-1 and histone H1 on four way junction DNA. Nucleic Acids Res.,
25, 3523–3531.
154
Human Molecular Genetics, 2000, Vol. 9, No. 2
9. West, K.L. and Austin, C.A. (1999) Human topoisomerase II binds and
cleaves four-way junction DNA in vitro. Nucleic Acids Res., 27, 984–992.
10. Bachvarov, D.R., Markov, G.G and Ivanov, I.G. (1987) Sequence heterogeneity of the human alpha satellite DNA and thermal stability of mismatched alphoid DNA duplexes. Inst. J. Biochem., 19, 963–971.
11. Vissel, B. and Choo, K.H. (1987) Human alpha satellite DNA—consensus
sequence and conserved regions. Nucleic Acids Res., 15, 6751–6752.
12. Waye, J.S. and Willard, H.F. (1987) Nucleotide sequence heterogeneity of
alpha satellite DNA: a survey of alphoid sequences from different human
chromosomes. Nucleic Acids Res., 15, 7549–7567.
13. Fleig, U., Beinhauer, J.D. and Hegemann, J.H. (1995) Functional selection for the centromere DNA from yeast chromosome VIII. Nucleic Acids
Res., 23, 922–924.
14. Jayasena, S.D. and Johnston, B.J. (1992) Oligonucleotide-directed triple
helix formation at adjacent oligopurine and oligopyrimidine DNA tracts
by alternate strand recognition. Nucleic Acids Res., 20, 5279–5288.
15. Harvey, S.C., Luo, J. and Lavery, R. (1988) DNA stem–loop structures in
oligopurine–oligopyrimidine triplexes. Nucleic Acids Res., 16, 11795–
11809.
16. Galas, D.J. (1978) On the symmetries of multipalindromic DNA
sequences. J. Theor. Biol., 72, 57–73.
17. Prosser, J., Frommer, M., Paul, C. and Vincent, P.C. (1986) Sequence
relationships of three human satellite DNAs. J. Mol. Biol., 187, 145–155.
18. Vissel, B. and Choo, H.K. (1989) Mouse major satellite DNA is highly
conserved and organized into extremely long tandem arrays: implications
for recombination between nonhomologous chromosomes. Genomics, 5,
407–414.
19. Kitada, K., Yamaguchi, E., Hamada, K. and Arisawa, M. (1997) Structural analysis of a Candida glaberata centromere and its functional
homology to the Saccharomyces cerevisiae centromere. Curr. Genet., 31,
122–127.
20. Baum, M., Ngan, V.K. and Clarke, L. (1994) The centromeric K-type
repeat and the central core are together sufficient to establish a functional
Schizosaccharomyces pombe centromere. Mol. Biol. Cell, 5, 747–761.
21. Vig, B.K. and Richards, B.T. (1992) Formation of primary constriction
and heterochromatin in mouse does not require minor satellite DNA. Exp.
Cell Res., 201, 292–298.
22. Aker, M. and Huang, H.V. (1996) Extreme heterogeneity of minor satellite repeat arrays in inbred strains of mice. Mamm. Genome, 7, 62–64.
23. Wong, A.K., Biddle, F.G. and Rattner, J.B. (1990) The chromosomal distribution of the major and minor satellite is not conserved in the genus
Mus. Chromosoma, 99, 190–195.
24. Pietras, D.F., Bennett, K.L., Siracusa, L.D., Woodworth-Gutai, M.,
Chapman, V.M., Gross, K.W., Kane-Haas, C. and Hastie, N.D. (1983)
Construction of a small Mus musculus repetitive library: identification of
a new satellite sequence in Mus musculus. Nucleic Acids Res., 11, 6965–
6983.
25. Sainz, J. and Cornudella, L. (1990) Preservation of a complex satellite
DNA in two species of echinoderms. Nucleic Acids Res., 18, 885–890.
26. Madsen, C.S., de Kloet, D.H., Brooks, J.E. and de Kloet, S.R. (1992)
Highly repeated DNA sequences in birds: the structure and evolution of an
abundant, tandemly repeated 190-bp DNA fragment in parrots. Genomics,
14, 462–469.
27. Fanning, T.G. (1989) Molecular evolution of centromere-associated
nucleotide sequences in two species of canids. Gene, 85, 559–563.
28. Rovira, C., Beermann, W. and Edström, J.-E. (1993) A repetitive DNA
sequence associated with the centromeres of Chironimus pallidivittatus.
Nucleic Acids Res., 21, 1775–1781.
29. Strauss, F. and Varshavsky, A. (1984) A protein binds to a satellite DNA
repeat at three specific sites that would be brought into mutual proximity
by DNA folding in the nucleosome. Cell, 37, 889–901.
30. Solomon, M., Strauss, F. and Varshavsky, A. (1986) A mammalian high
mobility group protein recognizes any stretch of six A:T base pairs in
duplex DNA. Proc. Natl Acad. Sci. USA, 83, 1276–1280.