Download Cell-Specific Organization of the 5S Ribosomal RNA Gene Cluster

BIOLOGY OF REPRODUCTION 53, 1222-1228 (1995)
Cell-Specific Organization of the 5S Ribosomal RNA Gene Cluster
DNA Loop Domains in Spermatozoa and Somatic Cells'
Bonnie Nadel, Jocelyn de Lara, Scott W. Finkernagel, and W. Steven Ward 2
Division of Urology, Robert WoodJohnson Medical School, and The Cell and DevelopmentalBiology Program
Rutgers University, New Brunswick, New Jersey
ABSTRACT
DNA in eucaryotic cells is organized into loop domains, ranging in size from 25 to 100 kb, that are attached at their bases to the
structural component of the nucleus termed the nuclear matrix. These DNA loop domains have been shown to be important in the
regulation of both DNA replication and RNA transcription. In this study we have compared the structural organization of the DNA loop
domains of the 5S rRNA gene cluster in sperm, liver, and brain nuclei in the Syrian golden hamster. The individual loop domains were
visualized by fluorescent in situ hybridization to protamine (sperm)- and histone (somatic)-depleted nuclei, termed nuclear matrix halo
preparations. We found that in sperm nuclei, the 5S rRNA gene cluster was organized into three small loop domains that were approximately 48 kb each. In both types of somatic cell nuclei examined, the 5S rRNA gene cluster was organized into a single, much larger loop
domain that was up to 480 kb in length. The data suggest that at least some of the compaction that sperm DNA undergoes during
spermiogenesis is mediated by the nuclear matrix independent of protamine binding. Additionally, this sperm-specific DNA organization
may be involved inthe specific patterns of DNA replication and transcription of the paternal genome in the embryo.
INTRODUCTION
The organization of higher eucaryotic DNA into loops by
nuclear structures was first seen by electron microscopy in
amphibian lampbrush chromosomes [1] and mammalian
mitotic nuclei [2]. Evidence for their existence in interphase
nuclei was demonstrated by extracting nuclei with salt to
remove the histones and staining for DNA [3, 4]. In these
nuclei the DNA loop domains appeared as a halo of fluorescence surrounding the nuclei and had an average length
of 60 kb [3]. In interphase nuclei, these DNA loop domains
are attached at their bases to the nuclear matrix, the structural component of the nucleus [3, 5, 6]. Several laboratories
have demonstrated that DNA loop domains are functional
units, acting as replicons during DNA synthesis [3, 71, and
play a role in transcription [8-12]. The organization of DNA
into loop domains is different in different cell types
[8, 11, 12], and this may play a role in cell-specific transcription [13]. This has recently been vividly demonstrated by
fluorescent in situ hybridization (FISH) by Gerdes et al. [10].
Here, actively transcribed genes were shown to be associated with the nuclear matrix while inactive genes were localized to DNA that was within the extended loop domain.
We have previously shown that sperm DNA is also organized into loop domains by a sperm nuclear matrix [14].
These loop domains were demonstrated to be attached to
the sperm nuclear matrix by unique sequences [15, 16], indicating that at this level, sperm DNA is organized as speAcceptedJuly 7, 1995.
Received May 3, 1995.
'This work was supported by a grant from the National Institutes of Health (NICHD,
HD 28501) and by the Edwin A. Beer Program/New York Academy of Medicine.
2
Correspondence: W. Steven Ward, Ph.D., Division of Urology, Robert Wood Johnson
Medical School, MEB-588, 1 RWJ PI., New Brunswick, NJ 08903-0019. FAX: (908) 235-7013:
e-mail: [email protected]
1222
cifically as somatic cell DNA. Furthermore, the average size
of sperm loop domains was found to be 60% smaller than
that of somatic cell nuclei from the same species. The reason
for this difference may be related to the fact that embryonic
cells replicate their DNA much faster than do adult cells
[17, 18]. DNA loop domain sizes correspond to the sizes of
replicons [5, 7], so the smaller loop domains in sperm nuclei
may correspond to the smaller replicons in the embryo. The
smaller size of the sperm loop domains may also be important in affecting the tremendous degree of compaction
that sperm nuclei undergo during spermiogenesis.
In this study we have examined this difference in DNA
loop domain organization between sperm and somatic cell
nuclei much more specifically by visualizing the structural
organization of a particular DNA segment, the 5S rRNA gene
cluster. In the Syrian golden hamster, this gene cluster consists of a 2.2-kb segment of DNA that is repeated approximately 1350 times at one locus in the haploid genome [19].
It is an ideal probe for the architecture of DNA loop domains
for two reasons: it is small enough that individual loop domains can be visualized by FISH, but large enough to encompass more than one loop domain so that differences in
organization can be easily identified. We found a surprising
degree of specificity for the organization of DNA loop domains in both sperm and somatic cell nuclei.
MATERIALS AND METHODS
PreparationofMitotic Chromosomes
Primary hamster tail skin fibroblasts were obtained by
cutting the tail and stripping the skin with dissecting scissors. Both the tail and the skin were incubated with 1 mg/
ml collagenase in a-minimum essential medium without se-
5S rDNA LOOP DOMAINS IN SPERM AND SOMATIC CELLS
rum for 1 h at 37°C and then incubated in medium with 10%
fetal bovine serum in a tissue culture flask until the cells
were growing. The cells were replated onto slide chambers
and upon reaching approximately 50% confluence were
treated with 0.0225 gg/ml colcemid for 4 h. The medium
was then washed out, and 4 ml/slide of 75 mM KCI was
added; incubation was performed at 37 ° for 35 min, 2 ml of
3:1 methanol:acetic acid was added to each slide chamber,
and slides were incubated for 2 min at room temperature.
Slides were then fixed in four washes of 3:1 methanol:acetic
acid, the first two at 1 h and the last two at 45 min, all at
room temperature. Chromosomes were spread by gently
forcing a stream of humidified air onto the slide.
Preparationof Sperm Nuclear Structures
All sperm nuclear structures were prepared and fixed as
previously described [201. Briefly, hamster spermatozoa
were isolated from Syrian golden hamster caudae epididymides and washed in either 0.25% Nonidet P-40 (Sigma
Chemical Co., St. Louis, MO) (for decondensing nuclei) or
0.5% SDS (for condensed nuclei and nuclear matrices), and
nuclei were isolated by sucrose step gradients. Decondensing nuclei and nuclear matrices were prepared by extraction in 2 M NaCl, 25 mM Tris (pH 7.4), and 10 mM dithiothreitol on ice. Condensed nuclei were extracted with 300 mM
CaCl2 and 10 mM dithiothreitol on ice. All nuclear preparations except nuclear matrix halos were placed onto glass
slides, dried overnight, and fixed in 3:1 methanol:acetic
acid. The nuclear matrix halos were incubated on the slides
for 20 min, washed, and used immediately for FISH.
Preparationof Somatic Nuclear Halos
Nuclei from hamster brain and liver were prepared by
the method of Blobel and Potter [21] with modifications [14].
Briefly, the liver or brain was minced in 40 ml of ice-cold
medium (250 mM sucrose, 50 mM Tris [pH 7.4], 5 mM MgCl 2,
0.1 mM PMSF) and centrifuged at 2000 rpm in a Sorval
(DuPont, Wilmington, DE) HS-4 for 10 min. The pellets
were suspended in ice-cold 2.0 M sucrose, 50 mM Tris (pH
7.4) and layered onto 10 cushions of the same buffer. These
sucrose step gradients were centrifuged at 25 000 rpm in a
Beckman SW-28 rotor (Palo Alto, CA) for 1 h at 4°C.
Nuclear halos were prepared from these nuclei as described by Gerdes et al. [10], with modifications. The nuclear
pellets were suspended in 100 mM NaCl, 0.3 M sucrose, 3
mM MgCl 2, 10 mM Tris (pH 7.4), and 0.5% Nonidet P-40; 20
Ail of this suspension was placed on each of several slides.
The slides were incubated for 20 min at 40C; they were then
dipped twice in single-strength PBS (150 mM NaCl, 15 mM
Na2 PO 4, pH 7.2) at room temperature. Next, the slides were
dipped in 2 M NaC1, 10 mM Tris (pH 7.4), 10 mM EDTA, and
0.125 mM spermidine and incubated for 4 min at room temperature. The slides were rinsed by dipping for 1 min each
1223
in 10-strength, 5-strength, and double-strength PBS and
then for 2 min in single-strength PBS. They were dehydrated
for 1 min each in 10%, 30%, 70%, and 95% ethanol, allowed
to dry completely, and then baked at 70°C for 2 h. Finally,
the slides were incubated in single-strength PBS for 10 min,
then denatured in 70% formamide, and used for FISH.
FISH
The 2.2-kb hamster 5S rRNA probe, obtained from W.R.
Folk (University of Michigan, Ann Arbor, MI) [191, was labeled with biotin-conjugated adenosine by means of nick
translation using the Oncor Non-Isotopic Probe Labelling
Kit (cat. no. 24089-KIT; Gaithersburg, MD). The labeled
probe was solubilized in Hybrizol 7 (Oncor) at a concentration of 10 ng/Il. Ct-1 DNA (Sigma) was added to 100
ng/gl, and the probe was denatured by boiling and then
incubated at 370C for 2 h for prehybridization. The sperm
nuclear preparations were hybridized to the probe as described previously [20]. Briefly, slides were denatured in
70% formamide at 75°C for 4 min, dehydrated through successive steps of ethanol, and then dried. The probe was
added, and incubation was carried out overnight in a 37°C
humidified chamber. The slides were washed in 40% formamide at 40°C for 20 min and then in sodium citrate buffer.
The slides were next treated with avidin-fluorescein, then
with anti-avidin-fluorescein, and then again with avidin-fluorescein.
Measurement of DNA Loop Domain Length
Loop domain sizes were estimated by methods previously described for nuclear matrix halo preparations
[3, 10, 14] that were modified slightly for FISH signals. The
length of the fluorescent hybridization signal in the FISH
micrograph was measured in centimeters with a standard
ruler. This number was then converted to micrometers using
a constant calculated from a micrograph of a ruled microscope slide containing 10-$m increments that was photographed at the same magnification. Since the DNA within
the loop under these conditions is fully extended and not
bound by protein, the amount of DNA could then be calculated using the constant 0.34 nm/bp. For both sperm and
somatic cell nuclei, the signals in several nuclei were measured and the average size was calculated.
RESULTS
5S rDNA ChromosomalLocalization
The repeat unit for the 5S rRNA gene cluster (also referred to as 5S rDNA) in the Syrian golden hamster (Mesocricetus auratus) that was used in this study had not been
previously localized on the hamster karyotype [19]. In humans, the 5S rDNA is located predominantly on a single
1224
NADEL ET AL.
chromosomal locus [22], while in the Chinese hamster evidence for two loci has been presented [23]. Before we examined the organization of the 5S rDNA loop domains, it
was necessary to demonstrate by direct visualization that in
the Syrian golden hamster this gene cluster was located at
a single locus on one chromosome, as predicted. We prepared mitotic chromosomes from primary hamster tail fibroblasts and hybridized these to the 5S rDNA repeat. The
results demonstrated that the 5S rRNA probe reacted with
only one pair of homologous chromosomes (Fig. la). The
probe hybridized to the q arm of a large acrocentric chromosome (Fig. lb).
Structure of the 5S rDNA in Intact and Decondensed
Nuclei
The 5S rRNA gene cluster was localized as a single focus
in fully condensed sperm nuclei with dimensions of less
than 1 plm 2 (Fig. c). It should be noted that the 5S rDNA
does not occupy the same position in each nucleus. A more
detailed examination of the position of this gene in the fully
condensed nucleus has recently been completed and will
be published separately. When nuclei were decondensed,
the 5S rDNA extended into a relatively long, linear signal
(Fig. d, shown at the same magnification). In this example,
the DNA has decondensed and extends across Figure d
from the nuclear annulus (arrowhead), which is the point
at which the tail was attached to the sperm head [24]. The
gene cluster is not extended completely, as can be seen by
areas along the length of the 5S rDNA signal in Figure ld
that are thicker than others. This was confirmed by a few
cases in which the signal was a continuous line that
stretched well beyond the boundaries of the microphotographic field (Fig. le).
A comparison between Figure c and Figure 1, d and e,
readily illustrates the high degree of DNA packaging present
in mammalian sperm nuclei. All of the 5S rDNA shown in
the decondensed sperm nuclei (Fig. 1, d and e) was coiled
into a single locus when fully condensed (Fig. Ic), and twice
this amount was coiled into the mitotic chromosome (Fig.
la). These differences in the FISH images of the 5S rDNA
in the condensed and decondensed sperm nuclei were potentially misleading. The decondensed nuclei appeared to
have much more 5S rDNA, when in fact they contained the
same amount as the condensed nuclei. It was unclear
whether the highly localized images of 5S rDNA in condensed nuclei and in mitotic chromosomes actually had a
reduced amount of hybridization, since we did not directly
quantitate the fluorescence in this work. It was possible that
these nuclei had a reduced amount of hybridization due to
a decreased accessibility of coiled DNA in nuclei and chromosomes because of the presence of histones and protamines. These DNA-binding proteins likely inhibit both
probe hybridization and avidin binding to biotinylated
probes. The techniques used in this study allowed us to
examine the structure of the DNA sequences (that is,
whether they were extended or coiled), but our ability to
quantitate the DNA was limited to fully decondensed DNA
(see below).
Structure of the 5S rRNA Gene Cluster Loop Domains
Sperm DNA, like somatic cell DNA, is organized into
loop domains that are attached at their bases to the structural component of the nucleus, termed the nuclear matrix
[14, 20]. All the loop domains in the nucleus can be visualized as a group by extracting the protamines from the nucleus and staining with a DNA-specific dye such as propidium iodide. With use of these techniques on hamster
sperm nuclei, the DNA loop domains appeared as a halo of
fluorescence surrounding the nuclear matrix; they were
made up of approximately 70 000 loop domains with an
average size of 46 kb (Fig. 2a). When such halo preparations
were hybridized to the 5S rRNA probe, only the DNA loop
domains that made up the 5S rRNA gene cluster were visualized. Figure 2b presents the same nucleus shown in Figure 2a with the green filter used to visualize the fluorescein
isothiocyanate (FITC)-labeled 5S rRNA probe. A cluster of
loop domains can be seen within the halo emanating from
a single point within the nuclear matrix. Several more examples are shown in Figure 2, c-e. The average length of
the 5S rDNA loop domains was calculated by measuring the
length of several such structures and was found to be approximately 48 kb. This was well within the range of 25100 kb that has been reported for average eucaryotic loop
PLATE . Visualization of 5S rDNA.
FIG. 1. Localization of the 5S rDNA to a single locus. a,b) Primary hamster
fibroblast mitotic chromosomes, (c) condensed, and (d) decondensed hamster
sperm nuclei were hybridized to the 5S rRNA gene. All slides were counterstained
with propidium iodide to visualize the total DNA (red), and the biotin-labeled 5S
rRNA probe was visualized with FITC (green). The 5S rDNA signal appears yellow
in a and b. All micrographs are shown at the same magnification (bar = 10 m)
except for b,which is magnified 1.5 more.
FIG. 2. Visualization of the 5S rDNA loop domains. a) A single hamster sperm
nuclear matrix that has been stained with propidium iodide to stain all the DNA.
The DNA appears as a halo of fluorescence that surrounds the sperm nucleus. b)
The same nucleus as shown in a, here photographed with the green filter to visualize the hybridized 5S rRNA probe (green). c-e) Additional nuclear matrices hybridized to the 5S rRNA probe labeled with higher specific activity to reveal more
of the loop domain structure. Note that in all examples the 5S rDNA loop domains
have splayed into three discrete loop domains. (All micrographs are shown at the
same magnification as in Fig. 1; bar = 10 pm.)
FIG. 3. DNA loop domain structure of the 5S rDNA in liver and brain nuclear
matrices. a)A liver nuclear matrix stained with propidium iodide to visualize all the
DNA. The DNA loop domains are visible as a halo of fluorescence surrounding the
nucleus and have an average size of 60 kb. b) The same nucleus as shown in a, this
time in the green filter to visualize the hybridized 5S rRNA probe. Note that the two
signals extend well beyond the halo periphery. The allele on the right has broken
at one point. c) FISH of another liver nuclear matrix showing both alleles of the 5S
rDNA large loops. d) A liver nucleus showing one fully extended loop domain (left).
This loop domain contains 480 kDa of DNA. e) A fourth example of 5S rDNA loop
domains in liver. In this case the loop is clearly visible on the left allele. f) A brain
nuclear matrix showing the two large loop domains for the two 5S rDNA loci.
1226
NADEL ET AL.
domains. In most of these examples, the 5S rDNA signal in
the DNA halo was split into three or more signals. By examining 73 nuclear matrix preparations such as those
shown in Figure 2, we found that 94% were split into more
than one signal and that 58% clearly had three lines emanating from a single source. Examples of sperm halos with
three distinct loop domains are shown in Figure 2, c-e.
Structure of the 5S rDNA Loop Domains in Somatic Cells
Examination of the 5S rDNA loop domains in somatic
cells demonstrated a markedly different organization (Fig.
3). In diploid liver nuclei, the two 5S rRNA gene cluster
alleles were each organized into a single, long loop, extending well beyond the periphery of the fluorescent halo
that represented the average size of the loop domains. Figure 3a illustrates a single liver nuclear matrix stained with
propidium iodide to visualize this halo. The average size of
the loop domains in this halo is roughly 60 kb. Figure 3b is
a micrograph of the same nucleus taken with the green filter
to visualize the 5S rDNA. The two large loop domains extend well beyond the periphery of the fluorescent halo (the
5S rDNA loop on the right is severed in one place). Figure
3, c and d, contain two more examples of liver nuclear matrices. Because the size of the 5S rDNA is so large, most of
the individual loop domains did not fully uncoil from their
original superhelical state after histone extraction and were
therefore shorter than completely uncoiled DNA. The most
extended loop domain we were able to visualize is shown
in Figure 3d. In this nucleus, the 5S rDNA on the left was
positioned on the slide in such a way that the loop was
clearly visible. This DNA loop was 480 kb. A similar loop
organization for the 5S rDNA was seen in brain nuclei (Fig.
3f). Gerdes et al. [10] visualized the 5S rDNA in human fibroblasts with similar results, showing a single large loop
domain.
DISCUSSION
The data presented in this study indicate that the DNA
that composes the 5S rRNA gene cluster is organized into
two very different but specific structures by the sperm and
somatic cell nuclear matrices. The same DNA is organized
into three small loop domains in spermatozoa, but is present as a single, large loop domain in somatic cells. Thus, in
sperm nuclei the 5S rDNA is organized into a much more
compact structure than in somatic cells. This compaction
was not a consequence of protamine condensation [25, 26],
since the protamines were extracted from the sperm nuclear
matrix preparations before the hybridization was performed, as were the histones from the liver and brain nuclei.
Rather, the organization of the 5S rDNA into smaller loop
domains is mediated by the sperm nuclear matrix. The differences in the loop domain structure of the 5S rDNA between spermatozoa and somatic cells may be attributed to
the fact that the nuclear matrix protein components of different cell types vary significantly [27, 28]. These data suggest that at least some of the condensation that the genome
undergoes during spermiogenesis is mediated by the sperm
nuclear matrix, independent of protamine binding.
The exact nature of the different loop domain structures
is difficult to determine at this point. Our current working
model is illustrated in Figure 4, though this is almost certain
to be modified with further experimentation. The major
problem that must be addressed is that the total amount of
DNA in the three small 5S rDNA loop domains in the sperm
nucleus does not equal the total amount of DNA in the single, large 5S rDNA loop domain of the liver (compare Figs.
2c and 3d). The sperm nucleus does contain the entire repeated element, since completely decondensed sperm nuclei do contain continuous stretches of DNA that hybridize
the 5S rRNA probe (Fig. 1, d and e). These decondensed
signals from sperm nuclei were at least as long as the largest
single loop of 5S rDNA seen in the liver (Fig. 3d). Therefore,
there was more 5S rDNA present than was visible in the
three sperm loop domains seen in Figure 2, and this DNA
must still be present as coiled DNA within the sperm nuclear
matrix. Furthermore, neither sperm nor somatic nuclei contain the total amount of DNA predicted by Southern blot
data. The loop domain in the liver contains 480 kb of DNA,
while the three loop domains in the sperm nucleus contain
a total of 144 kb. But Hart and Folk [19] estimated that the
hamster genome contains approximately 2700 copies of the
2.2-kb 5S rRNA gene by Southern blot analysis; this is equivalent to 2.9 mb of DNA at each 5S rRNA gene locus. Thus,
not even the largest loop domain, that of the liver (Fig. 2d),
contains all of the 5S rRNA genes.
One possible explanation is that the 5S rDNA that was
not accounted for in the sperm loop domains remained
tightly coiled or folded within the nuclear matrix and that
this was seen as a small locus of fluorescence at the base
of the loops (Fig. 4). That a large amount of 5S rDNA could
be folded into such a small region (i.e., that located at the
base of the sperm 5S rDNA loops) was evident upon comparison of the 5S rDNA FISH signals in fully condensed
sperm nuclei to those of decondensed sperm nuclei. Similarly, in the somatic cell nuclei some of the 5S rDNA may
remain coiled within the nuclear matrix (Fig. 4). In the liver
and brain nuclei, this DNA may be coiled at one or both
ends of the visible loop domain. In sperm nuclei, this nuclear matrix-associated 5S rDNA may be present at the ends
of all three loop domains (Fig. 4). Evidence for this type of
organization comes from recent work by Gerdes et al. [10],
who demonstrated that actively transcribed DNA is organized into tightly compacted structures in histone-depleted
nuclei. In the liver, the 5S rRNA gene is active, so much of
the gene cluster-that being actively transcribed-may be
organized into a tightly compact structure. In the sperm nucleus, this compact DNA organization may serve a different
5S rDNA LOOP DOMAINS IN SPERM AND SOMATIC CELLS
1227
FIG. 4. Model for the 5S rDNA loop domain structure inliver and sperm nuclear matrices. The 5S rDNA is organized into a single
large loop domain inthe liver nucleus (left; only one allele isshown for clarity). The loop is larger than the average size of all the
DNA loop domains present, as delineated by the dashed circle surrounding the nucleus. Asignificant portion of the gene cluster
remains associated with the nuclear matrix ina manner not yet understood. The same DNA is organized into three, much smaller,
loop domains inthe sperm nucleus (right). Much more of the 5S rDNA is bound directly to the sperm nuclear matrix to create
more loop domains.
function, since mammalian sperm nuclei contain no measurable transcription [29]. Possible functions are discussed
below. Regardless of how this unaccounted-for DNA is
packaged within the nucleus, the presence of these three
loop domains in most of the sperm nuclei examined suggests that the 5S rDNA is organized into a specific structure
by the sperm nuclear matrix.
The linear 5S rDNA signal in the decondensed sperm
nuclei contained several interruptions (Fig. d). This may
represent competitive annealing between the biotin-labeled
probe and the two complementary strands of the chromosomal DNA. We observed a similar pattern using the telomere-specific repeat (TTAGGG)n, which is known to be
continuously tandem [20]. Additionally, these breaks in the
signal provide evidence for some interruptions in the DNA
of sequences that differ significantly from the 2.2-kb 5S
rRNA repeat [30]. These may include matrix attachment
regions that help to organize the gene cluster into loop domains, as discussed below.
Having identified the specific structural organization of
the 5S rDNA repeat, one is tempted to speculate on its func-
tion. DNA loop domains have been implicated in DNA replication [3-5, 7] and in RNA transcription [6, 8-12], neither
one of which is active in sperm nuclei. There are at least
three possibilities, none of which are mutually exclusive,
for the function of the 5S rDNA loop domains in sperm
nuclei. First, these loop domains may be involved in the
packaging of sperm DNA. The example shown in this work
suggests that this is a probable function since the 5S rDNA
in sperm nuclear matrices was organized into a much more
compact structure than was the same DNA in somatic nuclei. This compaction was independent of protamines, indicating that the sperm nuclear matrix may play a role in
sperm DNA condensation. Second, the organization of the
5S rDNA into three loop domains may be the result of a
functional structure in germ cell nuclei during spermatogenesis that is retained in an inactive form in the fully mature sperm. For example, the organization of the 5S rDNA
into three loop domains may be important for transcription
of the 5S rRNA in the meiotic spermatocytes or for both
transcription and DNA replication in the spermatogonia. Finally, these DNA loop domains may be important in the
NADEL ET AL.
1228
function of the paternal genome in the newly fertilized zygote, for either DNA replication, RNA transcription, or both.
One possible mechanism for the role of the sperm 5S
rDNA organization in embryonic DNA replication is that the
attachment site of the three loop domains represents a replication focus for the 5S rDNA. Somatic [31] and sperm nuclei [32] that have been induced to replicate DNA do so in
300-1000 discrete foci throughout the nucleus. Origins of
replication in mammalian cells have been shown to be attached to the nuclear matrix at the bases of DNA loop domains [33, 34]. It is possible that the organization of the 5S
rDNA loop domains into one cluster represents one such
replication focus that may be used during the replication of
the paternal genome shortly after fertilization. Furthermore,
it has been established that the size of the loop domain
corresponds to the size of the replicon [7]. Embryonic DNA
is replicated at a much faster rate than that of adult cells, so
the replicons, and therefore the loop domains, are smaller
[17, 18].
Even though the functions have yet to be elucidated, it
is clear that sperm DNA loop domains are organized in a
specific manner that differs markedly from that in adult, somatic cell nuclei. Further confirmation of these results with
other genes is necessary to determine how general these
structural differences are, and our laboratory is currently
investigating this. We are also extending these observations
of the 5S rDNA to determine the origin of this specific DNA
structure during spermatogenesis as well as its fate during
embryogenesis in order to address some of the questions
raised by these experiments.
REFERENCES
1. Callan HG, Gall JG, Berg CA. The lampbrush chromosomes of Xenopus laevis preparation, identification, and distribution of 5S DNA sequences. Chromosoma 1987; 95:236250.
2. Paulson JR, Laemmli UK. The structure of histone-depleted metaphase chromosomes.
Cell 1977 12:817-828.
3. Vogelstein B, Pardoll DM, Coffey DS. Supercoiled loops and eucaryotic DNA replication.
Cell 1980;22:79-85.
4. McCready SJ,GodwinJ, Mason DW, Brazell IA,Cook PR. DNAis replicated at the nuclear
cage. J Cell Sci 1980; 46:365-386.
5. Nelson WG, Pienta KJ, Barrack ER, Coffey DS. The role of the nuclear matrix in the
organization and function of DNA. Annu Rev Biophys Chem 1986; 15:457-475.
6. Mirkovitch J, Gasser SM, Laemmli UK. Relation of chromosome structure and gene expression. Philos Trans R Soc Lond 1987; 317:563-574.
7. Buongiorno-Nardelli M, Micheli G, CarrnMT, Marrilley M.A relationship between repli
con size and supercoiled loop domains in the eucaryotic genome. Nature 1982; 298:100102.
8. Robinson SI, Nelkin BD, Vogelstein B. The ovalbumin gene is associated with the nuclear
matrix of chicken oviduct cells. Cell 1982; 28:99-106.
9 Gasser SM, Laemmli UK. Cohabitation of scaffold binding regions with upstream/enhancer elements of three developmentally regulated genes of D. melanogaster Cell
1986: 46:521-530.
10. Gerdes M, Carter KC, Moen P. Lawrence J. Dynamic changes in the higher-level chromatin organization of specific sequences revealed by in situ hybridization. J Cell Biol
1994: 126:189-304.
11. Cockerill PN, Garrard WT. Chromosomal loop anchorage of the kappa immunoglobulin
gene occurs next to the enhancer in a region containing topoisomerase II sites. Cell
1986; 44:273-282.
12. Ciejek EM, Tsai MJ, O'Malley BW. Actively transcribed genes are associated with the
nuclear matrix. Nature 1983; 306:607-609.
13. Pienta KJ, Getzenberg RH, Coffey DS. Cell structure and DNA organization. Crit Rev Euk
Gene Express 1991; 1:355-385.
14. Ward WS, Partin AW, Coffey DS. DNA loop domains in mammalian spermatozoa. Chromosoma 1989; 98:153-159
15. Ward WS, Coffey DS. Specific organization of genes in relation to the sperm nuclear
matrix. Biochem Biophys Res Commun 1990; 173:20 25.
16. Kalandadze AG, Bushara SA,Vassetzky YS Jr, Razin SV.Characterization of DNA pattern
in the site of permanent attachment to the nuclear matrix in the vicinity of replication
origin. Biochem Biophys Res Commun 1990: 168:9-15.
17. Hyodo M, Flickinger RA. Replicon growth rates during DNA replication in developing
frog embryos. Biochim Biophys Acta 1973:299:29 33.
18. Flickinger RA, Givens R. Pine S, Sepanik P. Factors controlling the size of DNA loops in
frog embryos and Friend erythroleukemia cells. Cell Differ 1986; 19:59-71.
19. Hart RP,Folk WR. Structure and organization of a mammalian 5S gene cluster. J Biol
Chem 1982; 257:11706-1711.
20. de LaraJ, Wydner KL, Hyland KM, Ward WS. Fluorescent in situ hybridization of the
telomere repeat sequence in hamster sperm nuclear structures. J Cell Biochem 1993:
53:213-221.
21. Blobel G, Potter VR. Nuclei from rat liver: isolation method that combines purity with
high yield. Science 1966; 154:1662-1665.
22. Steffensen DM, Duffey P. Prensky W. Localisation of 5S ribosomal RNA genes on human
chromosome 1. Nature 1974; 252:741-743.
23. Stambrook PJ. Organization of the genes coding for 5S RNA in the Chinese hamster.
Nature 1976; 259:639-641.
24. Ward WS, Coffey DS. Identification of a sperm nuclear annulus: a spenn DNA anchor.
Biol Reprod 1989;41:361-370.
25. Balhorn R. A model for the structure of chromatin in mammalian sperm. J Cell Biol 1982:
93:298-305.
26. Ward WS, Coffey DS. DNA packaging and organization in mammalian spermatozoa:
comparison with somatic cells. Biol Reprod 1990; 44:569-574.
27 Getzenberg RH, Coffey DS. Tissue specificity and cell death are associated with specific
alterations in nuclear matrix proteins. Mol Endocrinol 1990; 4:1336-1342.
28. Fey EG, Penman S. Nuclear matrix proteins reflect cell type of origin in cultured human
cells. Proc Natl Acad Sci 1988;85:121-185.
29. Stewart TA, Bellv AR, Leder P. Transcription and promoter usage of the myc gene in
normal somatic and spermatogenic cells. Science 1984: 226707-710.
30. Moens PB, Pearlman RE. Satellite DNA 1in chromatin loops of rat pachytene chromosomes and in spermatids. Chromosoma 1989; 98:287-294.
31. Nakayasu H, Berezney R. Mapping replication sites in the eucarotic cell nucleus. J Cell
Biol 1989: 108:1-11.
32. Mills AD, Blow JJ, White JG, Amos WB, Wilcock D, Laskey RA. Replicationoccurs at
discrete foci spaced throughout nuclei replicating in vitro. J Cell Sci 1989: 94:471-477
33 Vaughn JP, Dijkwel PA. Mullender LHF,Hamlin JL. Replication forks are associated with
the nuclear matrix. Nucleic Acids Res 1990; 18:1965-1969.
34. Brylawski BP, Tsongalis GJ, Cordeiro-Stone M, May \XF'TComeaue LD).Kaufman DG.
Association of putative origins of replication with the nuclear matrix in normal human
fibroblasts. Cancer Res 1994;53:3865-3868.