Maintenance of genetic information in stem cells

The available data on DNA cosegregation in some stem cells are reviewed. Cairns was the first to assume cosegregation of template DNA strands for adult stem cells; i.e., all maternal DNA strands are preserved in one daughter cell, which remains a stem cell, while the newly synthesized DNA strands, which may contain errors, appear in the daughter cell that is committed to differentiation and passes to the transitory compartment of the cell population. The role of asymmetric mitosis in DNA cosegregation and maintenance of genetic information in stem cells is discussed.     

Asymmetric division and cosegregation of template DNA strands in adult muscle satellite cells

Satellite cells assure postnatal skeletal muscle growth and repair. Despite extensive studies, their stem cell character remains largely undefined. Using pulse-chase labelling with BrdU to mark the putative stem cell niche, we identify a subpopulation of label-retaining satellite cells during growth and after injury. Strikingly, some of these cells display selective template-DNA strand segregation during mitosis in the muscle fibre in vivo, as well as in culture independent of their niche, indicating that genomic DNA strands are nonequivalent. Furthermore, we demonstrate that the asymmetric cell-fate determinant Numb segregates selectively to one daughter cell during mitosis and before differentiation, suggesting that Numb is associated with self-renewal. Finally, we show that template DNA cosegregates with Numb in label-retaining cells that express the self-renewal marker Pax7. The cosegregation of 'immortal' template DNA strands and their link with the asymmetry apparatus has important implications for stem cell biology and cancer.     

Stem cell identity and template DNA strand segregation

The quest for stem cell properties to distinguish their identity from that of committed daughters has led to a re-investigation of the notion that DNA strands are not equivalent, and 'immortal' DNA strands are retained in stem cells whereas newly replicated DNA strands segregate to the differentiating daughter cell during mitosis. Whether this process occurs only in stem cells, and also in all tissues, remains unclear. That individual chromosomes can be also partitioned non-randomly raises the question if this phenomenon is related to the immortal DNA hypothesis, and it underscores the need for high-resolution techniques to observe these events empirically. Although initially postulated as a mechanism to avoid DNA replication errors, alternative views including epigenetic regulation and sister chromatid silencing may provide insights into this process.     

Immortal DNA or epigenetic signature ?

During mitosis each daughter cell inherits a full copy of the maternal genomic material. DNA replication, however, is an imprecise process, thus errors can arise resulting in potentially deleterious mutations over extended rounds of cell division and these may lead to cancinogenesis. Over thirty years ago, J. Cairns proposed that a cell could avoid the accumulation of mutations arising from DNA replication if all template DNA strands are inherited in one daughter cell during cell division, thus giving rise to the notion of < immortal > DNA strands. In this model the stem cells would retain the template DNA (older) strands. Proving or disproving this notion experimentally has been challenging. Further, it has recently become apparent that epigenetic regulation of gene expression plays a critical role in governing cell states, self-renewal and differentiation. In light of these data, can the phenomenon on template DNA strand segregation also reflect this epigenetic signature? In this review we explore these notions, discuss the evidence in support of this theory, the implications, and some of the mechanisms which could explain this phenomenon.     

Random Segregation of Chromatids at Mitosis in Saccharomyces Cerevisiae

Previous experiments suggest that mitotic chromosome segregation in some fungi is a nonrandom process in which chromatids of the same replicative age are destined for cosegregation. We have investigated the pattern of chromatid segregation in Saccharomyces cerevisiae by labeling the DNA of a strain auxotrophic for thymidine with 5-bromodeoxyuridine. The fate of DNA strands was followed qualitatively by immunofluorescence microscopy and quantitatively by microphotometry using an anti-5-bromodeoxyuridine monoclonal antibody. Chromatids of the same replicative age were distributed randomly to daughter cells at mitosis. Quantitative measurements showed that the amount of fluorescence in the daughter nuclei derived from parents with hemilabeled chromosomes diminished in intensity by one half. The concentration of 5-bromodeoxyuridine used in the experiments had little effect on the frequency of either homologous or sister chromatid exchanges. We infer that the 5-bromodeoxyuridine was distributed randomly due to mitotic segregation of chromatids and not via sister chromatid exchanges.     

The segregation of DNA in epithelial stem cells

It has recently been suggested that stem cells may invariably keep, from one division to the next, the daughter DNA molecules that contain the older of the two parental strands—that is, they may retain a complete set of “immortal strands,” through successive cell divisions (Cairns, 1975). We can test this hypothesis by labeling either the old immortal strands at the time the stem cells are created or the newly synthesized strands during subsequent divisions of the stem cells. In the former case, the stem cells should become permanently labeled; in the latter case, they should eliminate their label on their second division.Experiments of this sort have been conducted with the tongue papilla under steady state conditions and with the regenerating small intestinal crypts. The results clearly show that by far most of the multiplying cells in tongue and intestinal epithelium segregate their DNA “randomly” at mitosis. Nevertheless, the results, though far from conclusive, suggest that there are a small number of cells (1–5 in the stem cell region of each crypt and one at the base of each column of cells in the tongue) that selectively segregate their old and new DNA strands in the expected way. Thus in the immortal strand labeling experiments, there are a few labeled cells that retain their label for up to 4 weeks; conversely, in the new strand labeling experiments, a few cells appear to rid themselves of label after intervals equivalent to approximately two cell cycles.     

DNA methyltransferase-3–dependent nonrandom template segregation in differentiating embryonic stem cells

Asymmetry of cell fate is one fundamental property of stem cells, in which one daughter cell self-renews, whereas the other differentiates. Evidence of nonrandom template segregation (NRTS) of chromosomes during asymmetric cell divisions in phylogenetically divergent organisms, such as plants, fungi, and mammals, has already been shown. However, before this current work, asymmetric inheritance of chromatids has never been demonstrated in differentiating embryonic stem cells (ESCs), and its molecular mechanism has remained unknown. Our results unambiguously demonstrate NRTS in asymmetrically dividing, differentiating human and mouse ESCs. Moreover, we show that NRTS is dependent on DNA methylation and on Dnmt3 (DNA methyltransferase-3), indicating a molecular mechanism that regulates this phenomenon. Furthermore, our data support the hypothesis that retention of chromatids with the “old” template DNA preserves the epigenetic memory of cell fate, whereas localization of “new” DNA strands and de novo DNA methyltransferase to the lineage-destined daughter cell facilitates epigenetic adaptation to a new cell fate.     

Low-level DNA exchanges in normal human and xeroderma pigmentosum cells after UV irradiation

We have used a T4 endonuclease V assay method for UV-induced pryrimidine dimers in cellular DNA in vivo to obtain evidence for recombinational DNA exchanges after UV irradiation of normal human and Xeroderma pigmentosum (XP) cells. Our data indicate that the endonuclease-sensitive sites in excision-defective XP cells are removed very slowly from the irradiated parental strands and appear concomitantly in daughter strands newly synthesized during post-UV incubation. In the defective XP cells, the extent of appearance of sensitive sites in daughter strands synthesized during a period of 24 h after 10 J/m² appears to be small, probably less than 15% of the initial number of sensitive sites detected in cellular parental strands. Demonstration of such exchanges between normal-density parental and 5-bromodeoxyuridine-labeled daughter strands by alkaline CsCl isopycnic centrifugation was unsuccessful. Further, the extent is much lower in normal human cell because of their efficiet excision repair of the dimers before and after exchanges than in the defective XP cells.     

Formation of single-stranded breaks in newly-synthesized Escherichia coli DNA following treatment with bleomycin

Changes in molecular weight of newly synthesized DNA was studied after bleomycin treatment of Escherichia coli cells. The treatment by this drug causes only the increase of dispersion in sedimentation profiles of daughter DNA strands in wild type cells. There are two alternative explanation of this fact. First, single-strand breakage does not occur in newly synthesized DNA, i.e. bleomycin-induced athyminic sites do not block cellular DNA polymerases. Second, it is possible to explain it by quick rejoining of given breaks by cell repair systems. The sedimentation profile of daughter DNA strands of recA mutant rules out the first possibility. Observed shift to low molecular weight fractions region strongly indicates the formation of single-strand breaks in newly synthesized DNA. Extensive daughter DNA degradation in xthA mutant supports the idea of the existence of very effective excision repair in the case of apyrimidinic sites. Thus, non-eliminated bleomycin-induced damage causes the formation of single-strand breaks in newly synthesized DNA strands. These breaks may be repaired in the course of recA-dependent post-replication repair.     

Chromosome segregation in an asymmetrically dividing bacterium,Caulobacter crescentus

The mode of chromosome segregation in an asymmetrically dividing bacterium, Caulobacter crescentus, was studied by examining the fate of labeled DNA strands. Swarmer cells (one type of Caulobacter daughter cell), in which single strands of DNA had been labeled with [³H]thymidine during the previous round of chromosome replication, were grown synchronously in a non-radioactive medium for two generations. The distribution of radioactivity among the cells was visualized by autoradiography under a phase-contrast microscope. The labeled DNA strands in each cell were found to consist of two conserved units. From this, we propose a model in which the swarmer cell has two identical chromosomes, which are segregated into the progeny swarmer cell and the progeny stalked cell after chromosome replication.     

Cosegregation of cell wall and DNA in Bacillus subtilis.

Cosegregation of cell wall and DNA of a lysis-negative mutant of Bacillus subtilis was examined by continuously labeling (i) cell wall, (ii) DNA, and (iii) both cell wall and DNA. After four to five generations of chase in liquid media it was found by light microscope autoradiography that the numbers of wall segregation units per cell are 29 and 9 in rich and minimal medium, respectively. Under the same conditions the numbers of segregation units of DNA were almost 50% lower: 15 and 5, respectively. Simultaneous labeling of cell wall and DNA (iii) provided figures almost identical to those obtained for cell wall alone, (i), implying cosegregation of the two components. Statistical analysis ruled out their random distribution into daughter cells. Measurements of the positions of grain clusters at the end of the chase period along chains of cells, each derived from a single cell at the beginning of chase, show that cell wall units are localized according to a symmetrical pattern, whereas those of DNA are distributed in an asymmetrical but highly regular way. It appears that of two cell wall units of the same age one only has a strand of DNA attached to it. We present a simple diagrammatic model of cell wall organization and DNA-cell wall association which is compatible with our observations. Finally, we discuss previous experiments pertinent to cosegregation of cell wall and DNA obtained with cells grown on solid media as well as with germinating spores; an explanation for the independent segregation of cell wall and DNA observed in the latter case is advanced.     

A subpopulation of adult skeletal muscle stem cells retains all template DNA strands after cell division

1. Download : Download high-res image (335KB)
2. Download : Download full-size image

Highlights? Pax7^Hi quiescent stem cells constitute a metabolically low “dormant” subpopulation ? Both Pax7^Hi and Pax7^Lo quiescent stem cells are serially transplantable ? Only proliferating Pax7^Hi cells segregate old DNA strands to the renewing stem cell ? CO-FISH demonstrates that template DNA strand cosegregation engages all chromosomes     

Metakaryotic stem cell nuclei use pangenomic dsRNA/DNA intermediates in genome replication and segregation

Bell shaped nuclei of metakaryotic cells double their DNA content during and after symmetric and asymmetric amitotic fissions rather than in the separate, pre-mitotic S-phase of eukaryotic cells. A parsimonious hypothesis was tested that the two anti-parallel strands of each chromatid DNA helix were first segregated as ssDNA-containing complexes into sister nuclei then copied to recreate a dsDNA genome. Metakaryotic nuclei that were treated during amitosis with RNase A and stained with acridine orange or fluorescent antibody to ssDNA revealed large amounts of ssDNA. Without RNase treatment metakaryotic nuclei in amitosis stained strongly with an antibody complex specific to dsRNA/DNA. Images of amitotic figures co-stained with dsRNA/DNA antibody and DAPI indicated that the entire interphase dsDNA genome (B-form helices) was transformed into two dsRNA/DNA genomes (A-form helices) that were segregated in the daughter cell nuclei then retransformed into dsDNA. As this process segregates DNA strands of opposite polarity in sister cells it hypothetically offers a sequential switching mechanism within the diverging stem cell lineages of development.     

Metakaryotic stem cell nuclei use pangenomic dsRNA/DNA intermediates in genome replication and segregation

Bell shaped nuclei of metakaryotic cells double their DNA content during and after symmetric and asymmetric amitotic fissions rather than in the separate, pre-mitotic S-phase of eukaryotic cells. A parsimonious hypothesis was tested that the two anti-parallel strands of each chromatid DNA helix were first segregated as ssDNA-containing complexes into sister nuclei then copied to recreate a dsDNA genome. Metakaryotic nuclei that were treated during amitosis with RNase A and stained with acridine orange or fluorescent antibody to ssDNA revealed large amounts of ssDNA. Without RNase treatment metakaryotic nuclei in amitosis stained strongly with an antibody complex specific to dsRNA/DNA. Images of amitotic figures co-stained with dsRNA/DNA antibody and DAPI indicated that the entire interphase dsDNA genome (B-form helices) was transformed into two dsRNA/DNA genomes (A-form helices) that were segregated in the daughter cell nuclei then retransformed into dsDNA. As this process segregates DNA strands of opposite polarity in sister cells it hypothetically offers a sequential switching mechanism within the diverging stem cell lineages of development.     

T4-endonuclease V-sensitive sites in DNA from ultraviolet-irradiated human cells.

Human diploid cells (WI38) were pre-labeled with 32Pi, exposed to ultraviolet irradiation and then pulse labeled with [3H]thymidine. The extracted DNA from these cells was subsequently treated with the T4-endonuclease V, an enzyme which specifically nicks DNA strands at positions adjacent to pyrimidine dimers. Sedimentation in alkaline sucrose gradients revealed that the DNA synthesized after irradiation, as well as that made before, contained endonuclease-sensitive sites. Our results suggest that pyrimidine dimers are transferred from parental to daughter DNA strands during post-irradiation incubation. Sedimentation in neutral sucrose gradients showed that the molecular weight of native DNA was not affected by the endonuclease treatment, suggesting that the gaps appearing in daughter strands after irradiation are not opposite dimers or that the enzyme cannot recognize dimers in the gap regions.     

Support for the immortal strand hypothesis: neural stem cells partition DNA asymmetrically in vitro

The immortal strand hypothesis proposes that asymmetrically dividing stem cells (SCs) selectively segregate chromosomes that bear the oldest DNA templates. We investigated cosegregation in neural stem cells (NSCs). After exposure to the thymidine analogue 5-bromo-2-deoxyuridine (BrdU), which labels newly synthesized DNA, a subset of neural precursor cells were shown to retain BrdU signal. It was confirmed that some BrdU-retaining cells divided actively, and that these cells exhibited some characteristics of SCs. This asymmetric partitioning of DNA then was demonstrated during mitosis, and these results were further supported by real time imaging of SC clones, in which older and newly synthesized DNA templates were distributed asymmetrically after DNA synthesis. We demonstrate that NSCs are unique among precursor cells in the uneven partitioning of genetic material during cell divisions.     

The coupling of epigenome replication with DNA replication

In multicellular organisms, each cell contains the same DNA sequence, but with different epigenetic information that determines the cell specificity. Semi-conservative DNA replication faithfully copies the parental nucleotide sequence into two DNA daughter strands during each cell cycle. At the same time, epigenetic marks such as DNA methylation and histone modifications are either precisely transmitted to the daughter cells or dynamically changed during S-phase. Recent studies indicate that in each cell cycle, many DNA replication related proteins are involved in not only genomic but also epigenomic replication. Histone modification proteins, chromatin remodeling proteins, histone variants, and RNAs participate in the epigenomic replication during S-phase. As a consequence, epigenome replication is closely linked with DNA replication during S-phase.     

recF-dependent and recF recB-independent DNA gap-filling repair processes transfer dimer-containing parental strands to daughter strands in Escherichia coli K-12 uvrB.

The processes for repairing DNA daughter-strand gaps were studied in UV-irradiated uvrB, uvrB recB, uvrB recF, and uvrB recB recF cells of Escherichia coli K-12. The dimer-containing parental DNA was found to be joined to daughter strands during postreplication repair in all four strains examined. Therefore, both the major (recF-dependent) and the minor (recF recB-independent) gap-filling processes repair DNA daughter-strand gaps by transferring parental strands into daughter strands.     

Persistence of pyrimidine dimers during post-replication repair in ultraviolet light-irradiated Escherichia coli K12

We have used a new assay for pyrimidine dimers to obtain evidence regarding the mechanism of post-replication repair of ultraviolet light-induced damage in excision-deficient (uvr) mutants of Escherichia coli. Our data indicate that dimers are gradually removed from the irradiated DNA under conditions permitting post-replication repair. Concomitantly, dimers appear in daughter strands synthesized after irradiation. The daughter strands initially contain gaps. During post-replication repair the gaps are filled and the originally discontinuous DNA is joined into long molecules resembling those observed in unirradiated control cells. Density transfer experiments reported by other investigators have provided evidence that the gap-filling involves exchanges between irradiated parental DNA and unirradiated daughter strands. The results of our experiments are in accord with this possibility and suggest that some dimers are included in the exchanged regions. Our data imply that intact, dimer-free DNA molecules are not necessarily generated by gap-filling and may not appear in uvr cells until several hours after u.v. irradiation. Instead, dimers may be gradually diluted among successive generations of DNA molecules synthesized after irradiation.     

Preirradiation of host cells does not alter blockage of simian virus 40 replication forks by pyrimidine dimers

Do damage-inducible responses in mammalian cells alter the interaction of lesions with replication forks? We have previously demonstrated that preirradiation of the host cell mitigates UV inhibition of SV40 DNA replication; this mitigation can be detected within the first 30 min after the test irradiation. Here we test the hypotheses that this mitigation involves either (1) rapid dimer removal, (2) rapid synthesis of daughter strands past lesions (trans-dimer synthesis), or (3) continued progression of the replication fork beyond a dimer. Cells preirradiated with UV were infected with undamaged SV40, and the effects of UV upon viral DNA synthesis were measured within the first hour after a subsequent test irradiation. In preirradiated cells, as well as in non-preirradiated cells, pyrimidine dimers block elongation of daughter strands; daughter strands grow only to a size equal to the interdimer distance along the parental strands. There is, within this first hour after UV, no evidence for trans-dimer synthesis, nor for more rapid dimer removal either in the bulk of the parental DNA or in molecules in the replication pool. Progression of the replication forks was analyzed by electron microscopy of replicating SV40 molecules. Dimers block replication-fork progression in preirradiated cells to the same extent as in non-preirradiated cells. These experiments argue strongly against the hypotheses that preirradiation of host cells results in either the rapid removal of dimers, trans-dimer synthesis, or continued replication-fork progression beyond dimers.