| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
a Cell and Developmental Biology Program, Rutgers University, Piscataway, New Jersey 08854
b Department of Genetics, Southwest Foundation for Biomedical Research, San Antonio, Texas 78245-0549
c Division of Urology, Robert Wood Johnson Medical School, New Brunswick, New Jersey 08901
ABSTRACT
The DNA in eukaryotic cells is organized into loop domains that are 25 to 100 kilobases long and attached at their bases to the nuclear matrix. This organization plays major roles in DNA replication and transcription. We examined changes in DNA loop structure of the 5S rDNA gene cluster in the Syrian golden hamster as a function of cellular differentiation by direct visualization with fluorescent in situ hybridization. The 5S rDNA cluster is large enough to encompass more than one loop domain but small enough that individual loop domains can still be resolved. We found that the sizes of the 5S rDNA loops are much smaller, and that the numbers of loops per locus are larger, in all pluripotent cell types than they are in adult somatic tissue. Within the pluripotent spermatogenic cell lineage, the loop domain organization was cell specific. The loop size decreased during the early stages of spermatogenesis but did not change during spermiogenesis, suggesting that DNA loop structure is independent of the chromatin condensation that occurs when protamines replace histones. In early embryonic cells, the loop structure remained small, but in differentiated somatic cells, it became much larger. We suggest that these changes in the 5S rDNA loop domain structure may be related to the maintenance or loss of developmental potential.
developmental biology, implantation/early development, sperm, spermatid
INTRODUCTION
In recent years, it has become increasingly clear that chromatin structure in somatic nuclei has a profound effect on function. One well-described example of gene regulation by chromatin structure is at the level of nucleosome organization. The exact positioning of the nucleosomes along the promoters of genes can directly affect the ability of transcription factors to access their binding regions within those promoters [1, 2]. Moreover, the position of nucleosomes within specific genes can be inherited from one cell type to another or be specifically altered during development [3, 4]. Thus, changes in chromatin structure probably contribute to tissue-specific gene expression during development. Another, less well-understood level of chromatin structure is the organization of DNA segments, ranging in size from 25 to 100 kilobases (kb), into loop domains. These loops are formed by the attachment of the ends of each DNA segment to a structure made of proteins and RNA called the nuclear matrix [5, 6]. The DNA loop domain organization plays important roles in DNA replication [7] and in gene regulation [5, 8].
We have been interested in the functional consequences of the structural organization of mammalian sperm DNA. During spermiogenesis, histones are replaced by protamines [9, 10]. The nucleosome-positioning information is lost, because protamines organize DNA into highly condensed toroids containing roughly 50 kb of DNA, structures that are very different from nucleosomes [11, 12]. We have demonstrated, however, that fully mature spermatozoa retain the organization of DNA into loop domains by the nuclear matrix, and that this organization is very specific [13]. Therefore, the only aspect of chromatin structure that is present during all stages of spermatogenesis and embryogenesis is the organization of DNA into loop domains.
In this work, we have examined the loop domain organization of the hamster 5S rDNA gene cluster during gametogenesis and embryogenesis. The hamster 5S rDNA gene cluster is a single locus that is large enough, approximately 1 megabase, to be organized into several loop domains but also small enough that quantitation of individual loops is possible [14]. We tested the hypothesis that pluripotent cells retain a generally smaller loop size than that found in terminally differentiated cells, and that this loop organization undergoes dynamic changes during development. We did this by directly visualizing the 5S rDNA loop domains in four different cell types during hamster spermatogenesis: premeiotic spermatogonia, meiotic spermatocytes, postmeiotic spermatids, and mature spermatozoa. We also examined the DNA loop structure of the same gene cluster in hamster embryonic stem (ES) cells, and in adult liver and brain nuclei, as a very broad indication of the changes that might occur during embryogenesis and cellular differentiation.
MATERIALS AND METHODS
Preparation of Nuclei from Hamster Spermatogonia, Pachytene Spermatocytes, and Round Spermatids
Enriched populations of specific spermatogenic cell types were prepared using the Sta Put unit gravity sedimentation technique as originally developed for isolation of cells from mouse testes [15–17]. Briefly, testes were dissected from male adult or 5-day-old golden Syrian hamsters and decapsulated to remove the tunica albuginea. The tissue was then dissociated by sequential digestions in 0.5 mg/ml of collagenase (Worthington Biochemical, Chapel Hill, NC) and 0.5 mg/ml of trypsin (Sigma, St. Louis, MO), each for 15 min at 33°C in EKRB buffer (120.1 mM NaCl, 4.8 mM KCl, 25.2 mM NaHCO3, 1.2 mM KH2PO4, 1.2 mM MgSO4-7H2O, 1.3 mM CaCl2, 11.1 mM glucose, 1 mM glutamine, 10 ml/L of essential amino acids, 10 ml/L of nonessential amino acids, 100 µg/ml of streptomycin, and 100 U/ml of penicillin [K+ salt]). Following the trypsin digestion, DNase I was added up to a concentration of 1 µg/ml, and the cells were further dissociated by physical pipetting and filtration through nylon mesh (pore size, 80 µm for adult testicular tissue and 53 µm for prepubertal testicular tissue). Dissociated cells were recovered by centrifugation and resuspended in 0.5% w/v BSA in EKRB buffer. Preparations from two to three adult males yielded 1–1.5 x 109 dissociated seminiferous tubule cells. These were layered onto an 1100-ml gradient of 2%–4% BSA in EKRB buffer as described elsewhere [15] and allowed to sediment at unit gravity for 4 h at 4°C. Next, 100 fractions of 11 ml each were collected and viewed under phase optics in a light microscope. Fractions enriched for pachytene spermatocytes (5–10 x 107 cells) or round spermatids (2–5 x 108 cells) were pooled to give populations that were >90% pure for each cell type.
A single preparation from forty-six 5-day-old males was prepared in a similar manner and yielded 3.8 x 107 seminiferous tubule cells. These were layered onto a 550 ml of 2%–4% BSA gradient and allowed to settle for 4 h, after which 100 fractions of 5 ml each were collected, viewed, and pooled to yield an enriched population containing 
60% premeiotic spermatogonia. Nuclei were prepared from each enriched cell population by resuspending cells in 0.5 M sucrose plus 0.25% Triton X-100 in TEKSS buffer (25 mM KCl, 1 mM EDTA, 1 mM EGTA, 0.15 mM spermine, 1.5 spermidine, and 50 mM Tris-HCl [pH 7.5]), followed by physical pipetting to lyse cells. Nuclei were recovered by centrifugation through 0.88 M sucrose in TEKSS buffer. The pelleted nuclei were resuspended in buffer (10 mM Tris-HCl [pH 7.5], 100 mM NaCl, 0.3 M sucrose, 3 mM MgCl2, 0.5% Triton X-100, and 50% glycerol) and then stored frozen at -80°C. These cells were prepared for fluorescent in situ hybridization (FISH) by placing 20 µl of isolated nuclei on prechilled microscope slides, incubating for 20 min at 4°C, and extracting with 2 M NaCl as described below.
ES Cells
Hamster ES cells were obtained by kind donation from Dr. Nobuyo Maeda at North Carolina State University, Chapel Hill, NC [18]. Mouse embryo feeder (MEF) cells were obtained by kind donation from Dr. Michael Shen from the Center for Advanced Biotechnology and Medicine, Robert Wood Johnson School of Medicine. The MEF cells were treated with mitomycin C (10 mg/ml) to prevent cell division in culture and were stored in liquid nitrogen until use. Cells were rapidly thawed by gentle shaking back and forth in a 37°C water bath, suspension in 5 ml of MEF cell medium (500 ml of Dulbecco modified Eagle medium [DMEM], 55 ml of fetal bovine serum, and 5 ml of pen/strep), and centrifugation at 1000 rpm for 2 min. Cells were resuspended in 2 ml of MEF medium and transferred to a 100-mm culture dish that had been previously coated with 0.2% gelatin (Sigma). Dishes were brought to a final volume of 12 ml of MEF medium and incubated overnight at 37°C in 5% CO2. The following day, hamster ES cells were removed from liquid nitrogen and rapidly thawed in a 37°C water bath. Cells were centrifuged at 1000 rpm for 2 min and resuspended in ES cell culture medium (500 ml of DMEM, 100 ml of fetal calf serum, 0.01 M nonessential amino acids, 0.2 M L-glutamine, 0.2 M pen/strep, and 500 U/ml of leukemia inhibitory factor (LIF) [ESGRO; Gibco]) and transferred to the MEF cell culture dishes. Cells were fed daily with ES cell culture medium containing LIF until ES cell colonies reached sufficient size. The ES cells were then harvested by detaching with trypsin/EDTA (Gibco, Rockville, MD) for 2 min at 37°C in 5% CO2, centrifuged for 2 min at 1000 rpm, and then resuspended in ES cell medium. The ES cells were next frozen in liquid nitrogen (freezing medium was ES cell culture medium + 10% dimethyl sulfoxide) until use for FISH.
To prepare ES cells for FISH, cells that had been stored in liquid nitrogen were rapidly thawed in a 37°C water bath, centrifuged at 2000 rpm for 10 min, and resuspended in 50 mM Tris-HCl (pH 7.4). Droplets of 20 µl were placed on prechilled, polylysine (Sigma)-coated slides and allowed to adhere for 30 min at 4°C. Nuclear halo formation and FISH was carried out as described below.
Preparation of Brain and Liver Nuclei
Nuclei from hamster brain and liver were prepared according to the method described by Blobel and Potter [19] with modifications. Briefly, the liver or brain was minced in 40 ml of ice-cold medium (250 mM sucrose, 50 mM Tris-HCl [pH 7.4], 5 mM MgCl2, and 0.1 mM PMSF) and centrifuged at 2000 rpm in a Sorval HS-4 (DuPont, Wilmington, DE) for 10 min. The pellets were resuspended in 180 ml of ice-cold 2.0 M sucrose and 50 mM Tris (pH 7.4), then layered onto six 10-ml 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. The nuclear pellets were suspended in 100 mM NaCl, 0.3 M sucrose, 3 mM MgCl2, 10 mM Tris (pH 7.4), and 0.5% Nonidet P-40 (Sigma) [20]. To prepare these nuclei for FISH, 20-µl aliquots of this suspension were placed on prechilled microscope slides and incubated for 20 min at 4°C.
Preparation of Nuclear Halos for FISH
For all cell types used, except spermatozoa, nuclear halos were prepared according to the method described by Gerdes et al. [20] with modifications. Slides with adhered nuclei as described above were rinsed by dipping in 10 mM Tris (pH 7.4) and incubating in 50 mM Tris/0.5% Triton-X 100 (pH 7.4) for 10 min at room temperature (RT) to remove membranes. Nuclear halo formation was induced by incubating the slides in 2 M NaCl, 10 mM Tris (pH 7.4), 10 mM EDTA, and 0.125 mM spermidine for 4 min on ice. The slides were then rinsed by dipping for 1 min each in 10x, 5x, and 2x PBS and then for 2 min in 1x PBS. Halos were dehydrated in an ethanol series (10%, 30%, 70%, and 95% for 2 min each). Slides were allowed to air-dry completely, then were fixed by baking for 2 h at 70°C. The RNA was removed from the cellular remains by incubation in 1x RNase (40 ml of 2x SSC and 40 µl of 1000x RNase; Oncor, Gaithersburg, MD; 1x SSC: 0.15 M sodium chloride and 0.015 M sodium citrate) for 1 h at 37°C (spermatogonia nuclei were not treated with RNase). Slides were then rinsed four times in 2x SSC for 2 min each. Halos were dehydrated in ethanol (70%, 80%, and 95% for 2 min each) on ice, then dried with a hair dryer set on low. Finally, slides were incubated in 1x PBS for 10 min, after which FISH was performed on nuclear halos as described elsewhere [13].
Isolation of Mature Sperm Nuclei and Preparation of Nuclear Halos for FISH
Mature spermatozoa were isolated and prepared for FISH as described elsewhere [13]. Briefly, the cauda epididymidis was dissected from Syrian golden hamster testes. Vertical incisions were made in the distal portion, and the caudal fluid was teased out with a pair of forceps. The spermatozoa-containing fluid was homogenized in a Dounce homogenizer (Wheaton Scientific, Millville, NJ) with 50 mM Tris-HCl (pH 7.4) and 0.5% SDS to separate the heads from the tails. The suspension was layered onto a two-layer sucrose step gradient in a Beckman SW-28 tube. The bottom layer contained 2.0 M sucrose, 0.075 g/ml of CsCl, and 25 mM Tris-HCl (pH 7.4), and the second layer contained the same solution without the CsCl. The three-layer step gradient was centrifuged at 25 000 rpm for 1.5 h, and the supernatant was discarded. Nuclear halos were induced to form through resuspending the pelleted nuclei by incubation in 2 M NaCl, 10 mM dithiothreitol, and 25 mM Tris-HCl (pH 7.4) for 30 min on ice, then for 10 min at 37°C. Next, 20-µl drops of the resulting nuclear halo suspension were placed on prechilled microscope slides and allowed to adhere for 30 min at 4°C. The slides were then rinsed in 10 mM Tris-HCl (pH 7.4) and dried overnight at RT, after which they were fixed by incubation in ice-cold 70% ethanol for 20 min, followed by two washes of 3:1 methanol:acetic acid for 20 min each on ice. Slides were allowed to dry completely at RT, then dehydrated in cold ethanol (70%, 80%, and 95% v/v for 2 min each). Slides were then dried with a hair dryer set on low. The DNA was denatured by incubation in 70% formamide and 2x SSC for 4 min at 74°C. Slides were again dehydrated in ethanol (70%, 80%, 90%, and 100% v/v), then dried with a hair dryer set on low. Slides were then used immediately for FISH.
Fluorescent In Situ Hybridization
Fluorescent in situ hybridization was performed as described elsewhere [13]. A double-stranded DNA probe was used for all FISH experiments. The probe was generated from a 2.2-kb hamster 5S rRNA repeat [14] and was labeled with biotin-conjugated adenosine by nick translation using the Oncor Non-Isotropic Probe Labeling Kit. The probe was suspended in Hybrizol 7 (Oncor) at a concentration of 1 ng/µl with 1 mg/µl of Cot-1 DNA. The probe was then denatured by boiling for 5 min and again incubated for 2 h at 37°C. Probe was stored at -20°C until use. Ten microliters of the biotinylated probe was added to slides containing adsorbed, fixed, and denatured nuclear halos (prepared as described above), and glass coverslips were affixed using rubber cement around the edges. The slides were incubated overnight at 37°C in a humidified chamber. The following morning, the slides were washed in 55% formamide at 40°C for 20 min, then rinsed twice in 2x SSC buffer for 4 min each at 37°C. Fluorescein isothiocyanate (FITC)-avidin binding and propidium iodide labeling was performed with the In Situ Chromosome Systems Kit (Oncor) according to the manufacturer's instructions. Fluorescent signals were detected using an upright Zeiss Akioskop fluorescent microscope (Thornwood, NY) using Zeiss filters (FITC: catalog no. 487709; excitation spectrum, 450–490 nm; emission spectrum, 520 nm; propidium iodide: catalog no. 487715: excitation spectrum, 534–558 nm; emission spectrum, 590 nm). Images for all cell types, except ES cells, were captured on 35-mm, 400-speed color film. The ES cell images were captured on a three-channel, cooled, digital camera (Dage, Thornwood, NY).
Image Analysis
Photographs of all cell types, except ES cells, were converted into digital images using an Agfa DuoScan Scanner (Agfa, Ridgefield Park, NJ). The ES cell images were collected digitally, so conversion was not necessary. Each photograph was scanned along with a photograph of a micron scale taken at the same magnification. Loop sizes were measured with Scion Image software (Scion Corp., Frederick, MD) using the segmented line tool. Each loop measured was calibrated by a micron bar taken from the micron scale image.
Statistical Analysis
Means with 95% confidence intervals were calculated, and means were compared using the Student t-test (SAS 6.12; SAS Institute, Inc., Cary, NC). Histograms were generated in Microsoft Excel (Redmond, WA).
RESULTS
Direct Visualization and Quantitation of the DNA Loop Domain Organization of the 5S rDNA Gene Cluster
The structure of the loop domain organization of the 5S rDNA gene cluster was examined directly by FISH. Nuclei were prepared from each cell type and were extracted with 2 M NaCl to remove the histones or protamines. In these extractions, the DNA forms a halo around the nucleus of the cell. This halo is made up of loops of DNA attached at their bases to a proteinaceous nuclear structure, the nuclear matrix [7]. Examples of such nuclear halos are shown in Figure 1, A and B, for sperm and liver nuclei. When such nuclear halos are hybridized with a biotinylated 5S rDNA probe, the individual loops of the 5S rDNA gene can be detected with fluorescently tagged avidin. Previous work from our laboratory suggested that the 5S rRNA gene cluster is organized very differently in sperm and liver nuclei in the hamster [13]. In the sperm nucleus, the 5S rDNA is organized into three loops (or three groups of loops), whereas in the liver, the same gene is organized into a single, large loop domain (Fig. 1, C and D). These data suggest that the 5S rDNA loop domain structure is cell-type specific.
|
The complexity of the 5S rDNA loop domain organization makes it an ideal candidate for studying changes in complete loop domain structure during spermatogenesis. An example of the changes in DNA loop domain structure detectable using a single gene probe, that for the CAD gene [21], is shown in Figure 1, E and F. This gene is only 47 kb in length, and the probe that was used in this study was a cosmid that contained the entire genomic sequence. The length of the CAD gene FISH signal shown in Figure 1E corresponds to a loop of approximately 46 kb of DNA, assuming the DNA was stretched out completely, with no coiling or secondary structure. Therefore, the CAD gene is organized into no more than a single DNA loop domain. In the liver, the CAD gene is organized into a small focus, still tightly bound to the nuclear matrix, even though the histones have been extracted (Fig. 1F). This type of association with the nuclear matrix was previously described by Gerdes et al. [20], who demonstrated that actively transcribing genes are more likely to be associated with the nuclear matrix as a single focus, as is the CAD gene in the liver (Fig. 1F), whereas inactive genes would be extended into the halo, as is the CAD gene in spermatozoa (Fig. 1E).
In the current work, we focused on the more complex organization of the 5S rRNA gene cluster, because it included several adjacent loop domains that appeared to be organized very differently in two cell types (Fig. 1, C and D). We tested the hypothesis that the 5S rDNA loop domain structure was cell-type specific by quantitating the loop domain structure of several different cell types. Figure 2 demonstrates the method used to quantitate DNA loop domain structures. For each image, the number of visible loops was determined, and then each loop was measured. We counted each linear signal as a single DNA loop domain. For large loops, contiguous line segments were used to measure the digitized images of the loop signals. The size of each loop was affected by the level of supercoiling that the individual loops retained after salt extraction and by the completeness of the salt extraction. The number of visible loops was affected by several factors, including how the loops fall when the nucleus is extracted and whether two loops happen to lie so close together that they would be counted as one. This is illustrated in Figure 2, C and D, for pachytene spermatocytes. In these cell types, the 5S rDNA was organized into several loop domains, but in some cases, all of the loops were clustered together on the slide. For this reason, we considered that loop size (i.e., the length of the loop from base to tip [µm]) was a more accurate variable than the number of loops in each focus. However, as discussed below, even with these uncontrollable variations, cell-specific patterns did emerge.
|
Changes in the 5S rDNA Loop Structure During Spermatogenesis
We first examined the structure of the 5S rDNA loop domain in four different cell types of spermatogenesis: spermatogonia, pachytene spermatocytes, round spermatids, and fully mature spermatozoa. The 5S rDNA loop domain structures for these cell types are shown in Figure 3. Several images for each cell type are shown to represent the reproducibility of the structure and size for each cell type. The FISH images of the 5S rDNA loop domains in these cells types were similar, in that most had more than a single loop domain. From these representative examples, the 5S rDNA loop domain structure in spermatogonia clearly is larger and more varied than any of the other three spermatogenic cell types. Most of the 5S rDNA loci of these 2N cells were organized into two loop domains. The pachytene spermatocytes had a much more consistently sized loop domain, but, because they were often grouped together, the number of visible 5S rDNA loops varied greatly (Figs. 2, C and D, and 3, D–F). These were the most complicated 5S rDNA loci to analyze, both because each individual 5S rDNA signal contained four adjacent copies of the gene cluster and because the replicated homologous chromosomes were paired into bivalents. The 5S rDNA organization of the round spermatids and spermatozoa were very similar (Fig. 3), both being smaller than either the spermatogonia or pachytene spermatocytes.
|
The comparison of the loop sizes for each cell type is shown in Figure 4. Spermatogonia contained the largest-size loop domains of the four spermatogenic cell types examined, with an average loop size of 17.9 ± 3.4 µm (all loop size data are presented as the mean ± 95% confidence interval; P < 0.05). This group also had the highest variation of loop size within the population examined. A histogram of the distribution of loop sizes within each cell type, as shown in Figure 5, suggests the possibility of two distinct subpopulations in the spermatogonial preparation. When these data were divided into two populations of less than and greater than 17.5 µm, the two means of 10.3 ± 1.2 µm (SG1) and 29.5 ± 15.5 µm (SG2) had nonoverlapping 95% confidence intervals. These data are complicated by the fact that this preparation was only approximately 60% pure, with the rest being largely contaminating Sertoli cells. The smaller of the two populations had a normal distribution around the mean, and this population contained the majority (60%) of the samples measured. The rest of the data were distributed widely over a very large range, and this population had the larger loop size. Therefore, we suggest that the smaller of these two populations, SG1, actually represents spermatogonia, and that the true loop size for the spermatogonia is 10.3 ± 1.2 µm. Note that the size and distribution of this smaller population resemble those of pachytene spermatocytes and ES cells (Fig. 5). Pachytene spermatocytes contained loop sizes of 8.0 ± 0.8 µm. The loop size of the round spermatids and of mature spermatozoa were 3.7 ± 0.5 µm and 3.7 ± 0.3 µm, respectively. These last three cell types all seemed to have normal distributions of loop size (Fig. 5). These data suggest that for the 5S rDNA loop size, a reduction in the loop size occurs during the postmeiotic stages of spermatogenesis, with the loop size then remaining essentially constant and indistinguishable in spermatids and spermatozoa.
|
|
As described above, the analysis of the number of loops per locus was complicated by the tendency of the loops to lie directly adjacent to each other. This is best illustrated for pachytene spermatocytes, in which the 5S rDNA was clearly organized into several loops, but in many cases, these loops were not distinguishable (Figs. 2C and 3F). Figure 6 shows the distribution of loops per focus for all of the spermatogenic cell types. Most of the 5S rDNA loci in spermatogonia were organized into two loops (mean, 1.9 loops), whereas in round spermatids and spermatozoa, they were organized into three loops (mean, 2.7 and 2.9 loops, respectively). The data for pachytene spermatocytes was complicated by the fact that each 5S rDNA loop signal contained four immediately adjacent gene clusters. However, the 5S rDNA clearly is organized into several loops, and even with our conservative estimate for each cell analyzed, pachytene spermatocytes contained the largest number of 5S rDNA signals per locus (mean, 3.6 signals).
|
5S rDNA Loop Structure in Embryonic and Adult Cells
We next analyzed the 5S rDNA loop domain structure in hamster ES cells and in adult liver and brain nuclei. Examples of the loop structures for these cell types are shown in Figure 7, with one example of spermatozoa for comparison (all images in Fig. 7 are shown at the same magnification). The 5S rDNA was organized into two loop domains in the majority of the ES cells examined, and this was most apparent when the DNA loops were somewhat extended (Fig. 7, B and D). The 5S rDNA of all adult, somatic nuclei examined was organized into a single, large loop domain (Fig. 7, E–G).
|
The sizes of the loop domains for these cell types were compared with those of the spermatogenic cells in Figure 4. The 5S rDNA loop domains of ES cells were 10.1 ± 1.0 µm, which was indistinguishable from those of pachytene spermatocytes. The distribution of loop sizes within the ES cells appeared to be normal (Fig. 5; the distribution of 5S rDNA loop sizes in ES cells is shown with those for the spermatogenic cell types, because they were smaller and distributed on the same scale). The number of 5S rDNA loops per locus for ES cells was close to that for spermatogenic cells, with a mean of 2.2 loops (Fig. 6). Liver and brain 5S rDNA loop sizes were much larger, with sizes of 37.1 ± 5.9 µm and 41.3 ± 11.2 µm, respectively (Fig. 4). The distributions of loop sizes for somatic liver and brain nuclei are shown in Figure 8. Note that these loop sizes are much larger than those for early embryonic and germline cells, and that the histograms in Figures 5 and 8 have different scales. For liver and brain nuclei, the 5S rDNA was organized into only one loop per locus.
|
Comparison of Loop Size and Loop Number
The data in Figure 4 suggested that the 5S rDNA is organized into at least three different loop domain sizes, as determined by the overlap of the 95% confidence intervals for the means of each group. Round spermatids and spermatozoa contained the smallest 5S rDNA loops. The ES cells, pachytene spermatocytes, and spermatogonia had the next largest, and the liver and brain contained the largest 5S rDNA loops.
We next determined the relationship between loop size and the number of loops within each locus. It was assumed that the length of the 5S rDNA gene cluster was the same in each cell type examined. Therefore, if the same length of DNA was organized into smaller loop domains, an inverse relationship should be found between the number of loops and the average size of each loop domain. This prediction was verified by calculating a simple linear regression analysis between loop size and the number of loops per locus (Fig. 9). When all cell types were included, the correlation constant was -0.89, showing a strong, inverse relationship between these two values. The least confident value for number of loops per locus was that of the pachytene spermatocytes, for the reasons discussed above. When this value was omitted from the linear regression analysis, the correlation constant was -0.98.
|
This regression analysis also suggested that the 5S rDNA loop organization is much more complicated than the usual model for DNA loop domain structure, as has been predicted by us and others [13, 20]. For example, the smallest-size 5S rDNA loop domains, as found in the round spermatids and spermatozoa, were more than 10-fold smaller than those of the largest 5S rDNA loop domains, as found in liver and brain, but they were only divided into 3 rather than 10 loop domains per locus. Possible reasons for this are provided in the Discussion.
DISCUSSION
We found that the 5S rDNA gene cluster is organized into several different types of loop structures in different cell types, which allowed us to follow changes in this type of chromatin structure throughout spermatogenesis and, broadly, during embryogenesis. Examining this chromosomal segment led us to several general conclusions regarding DNA loop domain structure in eukaryotes.
5S rDNA Loop Domain Structure Is Cell Specific and Complex
We previously demonstrated that the 5S rDNA gene cluster in the hamster is organized into three loop domains in sperm nuclei, and that the same gene cluster is organized into a single, large loop in liver and brain nuclei [13]. The striking difference between the DNA loop organization in spermatozoa as compared to that in liver and brain nuclei suggested that the 5S rDNA gene cluster was organized in a cell-specific manner.
In this work, we identified at least three different types of loop structure for the 5S rRNA gene cluster. In spermatozoa and spermatids, the 5S rDNA was organized into three loop domains, with an average size of 3.7 µm. The 5S rDNA loop domains of spermatogonia, pachytene spermatocytes, and ES cells were of similar size, approximately 8–10 µm, with the ES cells being clearly organized into two loops. Brain and liver contained the largest 5S rDNA loop size, approximately 40 µm and always organized into a single loop domain. This is diagramed in Figure 10. As noted in Figure 9, a clear, linear, and inverse relationship was found between loop size and loop number per locus, which was expected. However, this relationship is complex, in that the 5S rDNA loop size in sperm is approximately 10-fold smaller than that in liver but is organized into only threefold as many loops. If the DNA loop domain structure was, as current models suggest, a 25- to 100-kb segment of DNA with a 200-base pair attachment site, one would have predicted that the sperm 5S rDNA would be organized into 10 smaller loops. This suggests a more complex organization for DNA loop domains than has been commonly thought [13, 20].
|
DNA Loop Domain Structure Is Independent of Histone Association and Changes in Nuclear Morphology
Round spermatids are the first haploid cell type that result from meiosis during spermatogenesis. These cell types then go through a tremendous morphological change, in which virtually every aspect of the cell is altered, including the chromatin structure. Round spermatid nuclei contain histones, but during spermiogenesis, these histones are replaced by protamines, which condense the DNA into tightly packed toroids [22]. This morphogenesis is evident in Figure 3. Spermatid nuclei are rounded, as are most eukaryotic nuclei, but hamster sperm nuclei are flat, hook-shaped structures. However, we found that the 5S rDNA loop domain structures of these two cell types were essentially the same. Thus, the complete reorganization of chromatin structure that occurs during spermiogenesis by the replacement of histones with protamines does not affect the DNA loop domain structure. These data suggest that DNA loop domain formation is independent of histone binding and of gross nuclear morphogenesis. This raises the possibility that the nuclear matrix, to which the loop domains are attached, also undergoes compaction during spermiogenesis. It also suggests that the reason for the reorganization of the 5S rDNA loop domain structure that occurs between the pachytene spermatocyte and the round spermatid stages is not to facilitate the condensation of sperm chromatin that occurs during spermiogenesis, because the reorganization to smaller loops occurs before the round spermatids begin to condense.
5S rDNA Loop Domain Structure Undergoes Alterations During Spermatogenesis and Embryogenesis
Our results suggest that the DNA loop domain structure for the 5S rRNA gene cluster undergoes one or more alterations during spermatogenesis. The data suggest that the 5S rDNA loop size is similar in spermatogonia (10.3 ± 1.2 µm) and pachytene spermatocytes (8.0 ± 0.8 µm). A transition in 5S rDNA loop domain organization occurs between the pachytene spermatocyte and the round spermatid stages, in which the 5S rDNA is reorganized into a smaller loop size. As mentioned above, however, the 5S rDNA loop domain structure does not change during the subsequent morphogenic phase of spermatogenesis (i.e., spermiogenesis). This suggests that the 5S rDNA loop domain structure is tightly regulated during the initial stages of spermatogenesis.
In addition to the changes in 5S rDNA loop structure noted during spermatogenesis, another change occurs following fertilization, when the three short loops found in spermatozoa become two, slightly larger loops in early embryonic cells. The latter structure is exemplified in ES cells, which are representative of the inner cell mass in the blastula. We chose the ES cell line to represent embryonic tissue, because our techniques for examining loop domain structure required us to be able to isolate a homogeneous preparation of nuclei. The 5S rDNA loop domain structure of ES cells was strikingly similar to that of spermatogonia. A final and, perhaps, more significant change is found in differentiated cells, in which the small, multiple loop structure of the 5S rDNA seen in ES cells is reorganized into a much larger, single loop, as seen in adult brain and liver cells, which is consistent with previous observations [13, 20].
Relationship of DNA Loop Domain Structure to Embryogenesis
We found it significant that all of the germline and early embryonic cells we examined had significantly smaller loops than the adult somatic cell types. Gerdes et al. [20] showed that the 5S rDNA in adult human fibroblasts is also organized into a single, large loop domain. This suggests that organization of this region into a single, large loop may be characteristic of differentiated somatic cell types, whereas maintenance of this region in small loop domains may be a common feature of pluripotent cells. This is consistent with the presence of relatively small loops in the ES cells and their subsequent retention in germline but not in differentiated somatic cells. The changes that occur in the 5S rDNA loop domain structure during gametogenesis are diagramed in Figure 10. Of particular interest is the similarity in 5S rDNA loop structure between the ES cells and spermatogonia. During early embryogenesis, the germline derives from the pluripotent epiblast tissue [23–25], which, in turn, derives from the inner cell mass [26]. Thus, the similarity in 5S rDNA loop size seen in ES cells and spermatogonia represents the developmental maintenance of a constant loop size. However, somatic cell lineages also derive from the epiblast, and the small loop structure clearly is lost in these cells. We suggest that the increased loop size seen in terminally differentiated somatic cells correlates with a loss of developmental potential, and that retention of the small loop structure in germline cells reflects the maintenance of developmental pluripotency in these cells.
This concept is similar to a model originally proposed by McLaren [27] to explain continuity of pluripotency in the germline from generation to generation. She suggested that the embryonic epiblast acts as the pluripotent intermediate, from which both germ and somatic cell lineages are derived. Thus, the germline/epiblast/germline can be viewed as a continuous pluripotent lineage from generation to generation, with terminally differentiated somatic lineages also deriving from the epiblast at each generation. The general idea that nuclear matrix structure is an important contributor to developmental capability is supported by several other recent observations. We previously described results from intracytoplasmic sperm injection experiments, in which we demonstrated that mouse spermatozoa with unstable nuclear matrices are unable to participate in embryogenesis when injected into oocytes, whereas oocytes injected with similarly prepared sperm nuclei with stable nuclear matrices do develop into live births [28, 29]. Ogura et al. [30] previously showed that mouse eggs injected with round spermatid nuclei will undergo normal development into live mice, and this is consistent with our finding that round spermatid nuclei have the same DNA loop structure as spermatozoa (Fig. 10). Thus, a growing set of observations suggests that organization of DNA into the specific loop domains found in spermatozoa may contribute to proper embryogenesis in the subsequent generation.
The data presented in this work suggest that the three-dimensional structure of the 5S rDNA is specifically regulated during spermatogenesis and embryogenesis. The generally smaller loop organization seen throughout spermatogenesis may reflect a combination of cellular circumstances, including active DNA replication during the early stages of spermatogenesis, predisposition of this state during the latter stages for early embryogenesis, and maintenance of developmental pluripotency.
FOOTNOTES
First decision: 30 October 2000.
1 Supported by NIH grant HD28501 and the Howard K.L. Castle Foundation to W.S.W. and by HD 23126 to J.R.M. 
2 Correspondence and current address: W. Steven Ward, Institute for Biogenesis Research, Department of Anatomy and Reproductive Biology, 1960 East-West Road, University of Hawaii Medical School, Honolulu, HI 96822. FAX: 808 956 9481; wward{at}hawaii.edu
Accepted: November 28, 2000.
Received: September 28, 2000.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  A. M. Codrington, B. F. Hales, and B. Robaire Chronic Cyclophosphamide Exposure Alters the Profile of Rat Sperm Nuclear Matrix Proteins Biol Reprod, August 1, 2007; 77(2): 303 - 311. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Yamauchi, J. A. Shaman, and W. S. Ward Topoisomerase II-Mediated Breaks in Spermatozoa Cause the Specific Degradation of Paternal DNA in Fertilized Oocytes Biol Reprod, April 1, 2007; 76(4): 666 - 672. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Muriel, V. Goyanes, E. Segrelles, J. Gosalvez, J. G. Alvarez, and J. L. Fernandez Increased Aneuploidy Rate in Sperm With Fragmented DNA as Determined by the Sperm Chromatin Dispersion (SCD) Test and FISH Analysis J Androl, January 1, 2007; 28(1): 38 - 49. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. Lin, Q. Chen, L. Lin, and J. Zhou The Promoter Targeting Sequence mediates epigenetically heritable transcription memory Genes & Dev., November 1, 2004; 18(21): 2639 - 2651. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Muriel, E. Segrelles, V. Goyanes, J. Gosalvez, and J. L. Fernandez Structure of human sperm DNA and background damage, analysed by in situ enzymatic treatment and digital image analysis Mol. Hum. Reprod., March 1, 2004; 10(3): 203 - 209. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
