Two phases of chromatin decondensation during dedifferentiation of plant cells: distinction between competence for cell fate switch and a commitment for S phase.

Cellular dedifferentiation is the major process underlying totipotency, regeneration, and formation of new stem cell lineages in multicellular organisms. In animals it is often associated with carcinogenesis. Here, we used tobacco protoplasts (plant cells devoid of cell wall) to study changes in chromatin structure in the course of dedifferentiation of mesophyll cells. Using flow cytometry and micrococcal nuclease analyses, we identified two phases of chromatin decondensation prior to entry of cells into S phase. The first phase takes place in the course of protoplast isolation, following treatment with cell wall degrading enzymes, whereas the second occurs only after protoplasts are induced with phytohormones to re-enter the cell cycle. In the absence of hormonal application, protoplasts undergo cycles of chromatin condensation/decondensation and die. The ubiquitin proteolytic system was found indispensable for protoplast progression into S phase, being required for the second but not the first phase of chromatin decondensation. The emerging model suggests that cellular dedifferentiation proceeds by two functionally distinct phases of chromatin decondensation: the first is a transitory phase that confers competence for cell fate switch, which is followed, under appropriate conditions, by a second proteasome-dependent phase representing a commitment for the mitotic cycle. These findings might have implications for a wide range of dedifferentiation-driven cellular processes in higher eukaryotes.

A high proportion of mature plant cells retain characteristic features of totipotent stem cells, i.e. they have the capability to dedifferentiate, re-enter the cell cycle, and proliferate, eventually giving rise to the various organs that make up a new plant (1). In multicellular organisms, cellular dedifferentiation is the major process enabling the regeneration of complex tissues and organs as well as the establishment of new stem cell lineages; in animals it is often associated with carcinogenesis. The molecular mechanism(s) underlying re-entry of differentiated animal cells into the cell cycle and reactivation of DNA synthesis have been intensively studied. Yet, little is known about the early events that accompany cellular dedifferentiation, i.e. the establishment of competence for cell fate switch and the determination of cell fate. The tobacco protoplast system provides an outstanding experimental tool for the study of the biochemical and molecular basis for cellular dedifferentiation. The fully differentiated, non-dividing mesophyll cells of tobacco leaves can be separated from their original tissue by cell wall-degrading enzymes. This treatment results in the formation of a large population of protoplasts (plant cells devoid of cell wall) that, following treatment with phytohormones (auxin and cytokinin) can re-enter the cell cycle and proliferate (2)(3)(4)(5). This system demonstrates a unique attribute of plant cells, totipotency, i.e. the capability for cloning in plants (1).
Dedifferentiation of mature tobacco cells was shown to be accompanied by a sharp increase in ubiquitin gene expression (6), the biochemical tag that marks proteins for degradation by the ubiquitin proteasome system (7). Elevation in ubiquitin gene expression probably represents a critical point in cellular reorganization, namely, selective destruction of proteins involved in maintaining the previous function of a cell and a concomitant activation of proteins that are essential for cell proliferation (8 -11). Genetic studies in Arabidopsis demonstrated that some of the factors involved in auxin response are components of the cell cycle-regulated ubiquitin-protein-ligase complex (E3), known as SCF (Skp1-Cdc53-F-box protein) (Ref. 12 and references therein). The SCF complex is required for the ubiquitination of specific phosphorylated substrates during G 1 to S transition (9 -11).
In the course of cellular dedifferentiation, mature cells undergo remarkable changes in their pattern of gene expression, switching from a program that drives the specific function of a given somatic cell to a new one that directs cell multiplication. How do differentiated cells remodel their gene expression program in a rapid and orderly manner? Conceivably, chromatin reorganization could serve that purpose. Indeed, it has been suggested that one of the first stages in cellular dedifferentiation is the unfolding of chromatin supercoiling (13). We currently study mechanism(s) underlying dedifferentiation of mature plant cells, focusing on changes in chromatin structure. Using FACS 1 and MNase analyses, we found that differentiated tobacco mesophyll cells undergo two distinct phases of chromatin decondensation prior to entry into S phase. We also report the involvement of the ubiquitin proteasome system in chromatin decondensation and entry of cells into S phase.

EXPERIMENTAL PROCEDURES
Protoplast and Nuclei Preparation-Protoplasts were isolated from sterile leaves of Nicotiana tabacum ("Samsun NN") essentially as * This research was supported by a grant from the Israel Science Foundation (ISF). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 (14). Freshly prepared protoplasts were washed and incubated at 22°C in VKM medium containing 0.5 g/ml 6Ј-benzylaminopurine (BAP) and 2 g/ml ␣Ϫnaphthalene acetic acid (NAA). In certain assays, protoplasts or leaves were incubated in the presence of 100 M of the ubiquitin proteasome inhibitor MG132 (N-carbobenzyloxy (N-CBZ)-Leu-Leu-Leu-al, Sigma), or the protease inhibitor leupeptin (acetyl-Leu-Leu-Arg-al, Sigma), or 5 mM hydroxyurea (Sigma). All FACS measurements were performed on isolated nuclei to avoid interference of plastid and mitochondrial DNAs. Nuclei were isolated in a Hamilton buffer essentially as described (15,16), washed twice with FACS buffer (10 mM MES, 0.2 M sucrose, 0.01% Triton X-100, 2.5 mM EDTA, 2.5 mM dithiothreitol (16)) to remove soluble contaminants, and passed through two layers each of 150-and 100-m filters to remove cell debris. Nuclei were precipitated (1000 ϫ g, 7 min, 4°C), resuspended in FACS buffer supplemented with 50 g/ml DNase-free RNase A (Roche Molecular Biochemicals) and 50 g/ml PI (Sigma), viewed under the microscope, and subjected to FACS analysis using FACSort (Becton Dickinson). The position of PI fluorescence intensity for G 0 /G 1 nuclei has been changed from one experiment to another as a result of alteration in the amplifier gains for FL-2, which was necessary to accommodate fluorescence intensity of both G 0 /G 1 and G 2 nuclei.
Bromodeoxyuridine (BrdUrd) Labeling-Protoplasts reactivated for S phase (72 h after their preparation) were pulse-labeled with 10 M BrdUrd (Sigma) for 30 min, after which nuclei were isolated, stained with PI, and analyzed by FACS. Nuclei were then sorted, and DNA was prepared (from about 50,000 nuclei) as described below, resolved on 0.8% agarose gel, transferred onto nitrocellulose membrane, and probed with anti-BrdUrd (Becton Dickinson).
RNA Preparation and Northern Blot Analysis-Total RNA was prepared from protoplasts by using the Tri-Reagent kit (Sigma) according to the manufacturer's protocol. Equal amounts of RNA were resolved on 1.4% agarose formaldehyde gel, transferred to a BioTrace pure nitrocellulose membrane (Gelman Sciences), and hybridized with 32 P-labeled Arabidopsis Skp1 antisense RNA. Arabidopsis Skp1 was obtained by polymerase chain reaction using PWO DNA polymerase (Roche Molecular Biochemicals) and the following primers: Skp1-sense, 5Ј-G-TCGACGAATTCGAGCTCTCACTCGAGTCATTCAAAAGCCCATTGA-TTCTCC-3Ј flanked with the EcoRI and BamHI restriction sites, and Skp1-antisense primer, 5Ј-GTCGACGAATTCGAGCTCTCACTCGAGT-CATTCAAAAGCCCATTGATTCTCC-3Ј flanked with HincII and EcoRI sites. As a template we used the BAC clone T4L20 (GenBank TM /EBI accession number AL023094), kindly provided by the Arabidopsis Biological Resource Center (ABRC). The amplified DNA fragment was digested with BamHI and EcoRI, subcloned into the same sites of pBluescript SK, and sequenced to confirm identity. To prepare antisense RNA probes, the plasmid DNA was linearized with BamHI and subjected to in vitro transcription using the Riboprobe System (Promega) and T7 RNA polymerase according to the manufacturer's protocol.

First Phase Chromatin Decondensation Occurs in Freshly
Prepared Protoplasts-Tobacco leaves were treated with cell wall-degrading enzymes to produce a large population of protoplasts. We compared the FACS histograms of PI-stained nuclei isolated from freshly prepared protoplasts (t-0 protoplasts) versus nuclei prepared from leaves. Both t-0 protoplasts and leaf nuclei displayed a single peak corresponding to G 0 /G 1 content of DNA (Fig. 1, A and B). However, nuclei prepared from t-0 protoplasts (Fig. 1B, G1 nuclei) reproducibly showed an upward shift in fluorescence intensity compared with leaf nuclei (Fig. 1A, G0 nuclei). To verify that these changes in PI fluorescence reflected intrinsic differences in the PI binding capabilities of chromatin, both populations of nuclei were mixed and incubated on ice for 15-30 min prior to FACS analysis. Fig. 1C shows that even after nuclei mixing, each population retained its position in the FACS histogram supporting an intrinsic difference in PI binding capabilities of chromatin. Because PI intercalates between the strands of DNA without sequence specificity, its fluorescence intensity is directly proportional to the level of accessible DNA, hence reflecting the condensation state of chromatin. We next applied a biochemical test, namely the MNase assay, to test whether the observed increment in fluorescence intensity of protoplasts nuclei reflected an increase in DNA accessibility to PI conferred by DNA decondensation (17). To this end, permeabilized nuclei from leaves or from t-0 protoplasts were treated with MNase for various time periods, and protected DNA fragments were resolved by agarose gel electrophoresis. As shown in Fig. 1D, digestion by MNase resulted in a typical nucleosomal ladder visualized by ethidium bromide staining, where each band represents DNA protected by one or several nucleosomes. Chromatin of t-0 protoplasts showed a higer sensitivity to MNase than chromatin from leaf nuclei (Fig. 1D, compare lane 6 with lane 11). This verified that chromatin in nuclei of t-0 protoplasts is at a relatively decondensed configuration compared with chromatin in leaf nuclei. By analyzing the effect of leaf and protoplast nuclear protein extracts on the activity of MNase, we excluded the possibility that the reduced sensitivity of leaf chromatin to MNase resulted from the presence of nuclease inhibitors, otherwise absent from protoplasts nuclei (data not shown). Taken together, we suggest that the removal of the cell wall and the separation of differentiated mesophyll cells from their original tissue induce a stress response leading to conformational changes in chromatin.
Progression of Protoplasts into S Phase Is Accompanied by a Second Phase of Chromatin Decondensation-Following preparation, protoplasts re-enter the cell cycle and reactivate DNA synthesis only upon application of the phytohormones auxin and cytokinin. In the absence of phytohormones protoplasts undergo cycles of chromatin condensation/decondensation and die (Fig. 2). FACS analysis showed that within 24 h chromatin became condensed (Fig. 2C), shifting from the G 1 to G 0 position (Fig. 2, G1 and G0), which is typical for leaf nuclei (Fig. 2A). The chromatin is then gradually decondensed, and sub-G 1 population, which is typical for cells undergoing apoptotic cell death (18), starts to accumulate (Fig. 2, D and E), eventually leading to protoplast death (not shown). In the presence of phytohormones, protoplasts re-enter the cell cycle (Fig. 3); division patterns were first evident at 72 h after hormonal application (Fig. 3D, right panel). FACS analysis showed that most freshly prepared protoplasts (t-0 protoplasts, Fig. 3A) as well as those incubated with hormones for 24 h (Fig. 3B) had DNA content corresponding to the G 1 phase. By 48 h (Fig. 3C), about 10% of protoplasts approached S phase, and by 72 h (Fig.  3D) 30 -40% were in S and G 2 phases of the cell cycle. An extra G 1 peak (referred to as supra-G 1 ; Fig. 3, *G1), revealing an increase in PI fluorescence intensity, was reproducibly evident at 72 h after protoplast preparation (Fig. 3D). In contrast, cycling tobacco cells showed monophasic G 1 (see Fig. 4C), suggesting that the occurrence of a supra-G 1 peak is unique to differentiated cells re-entering the cell cycle. When protoplasts, however, were incubated for a longer duration (96 h), almost no original G 1 nuclei were apparent; the majority of nuclei shifted to the supra-G 1 position (Fig. 3E). This pattern suggests that the supra-G 1 peak represents a stage at which nuclei are competent to re-enter S phase, and, therefore, may be equivalent to G 1 nuclei of cycling cells. To examine this possibility, we established a cell suspension culture of N. tabacum (cycling cells) and compared the FACS histograms of nuclei prepared from this culture with that of N. tabacum protoplasts. Fig. 4 shows that G 1 nuclei of cycling cells had a higher fluorescence intensity than nuclei of t-0 protoplasts (Fig. 4, compare A with C), coinciding with the position of the supra-G 1 nuclei (Fig. 4, compare B with C). Each population of nuclei retained its position in the histogram after mixing (data not shown). These results indicate that the supra-G 1 peak represents chromatin relaxation rather than increased DNA content. This conclusion is strengthened by the finding that treatment of protoplasts with HU (a blocker of DNA replication) resulted in accumulation of both G 1 and supra-G 1 nuclei (Fig. 5A, G1 and *G1), and by the absence of BrdUrd labeling from supra-G 1 nuclei (Fig. 5,  B and C). These results verify that the increased fluorescence intensity of PI-stained nuclei (supra-G 1 ) prior to entry of cells into S phase resulted from chromatin decondensation, apparently an essential event for replication factors to approach and be assembled onto origins to initiate DNA replication.
The Ubiquitin Proteasome System Is Required for the Second Phase of Chromatin Decondensation and Reactivation of S Phase-The expression of various ubiquitin genes is enhanced in the course of dedifferentiation of mature tobacco cells (6). To investigate the involvement of the ubiquitin pathway in dedifferentiation, we tested the progression of protoplasts into S phase in the presence of MG132, a specific inhibitor of the ubiquitin proteasome system. MG132 had no effect on the occurrence of the first phase chromatin decondensation (data not shown). Freshly prepared protoplasts induced to re-enter the cell cycle were incubated with MG132 at various time points (0, 24, and 48 h), and their progression into S phase was monitored by FACS 72 h after preparation. Addition of MG132 at times 0 and 24 h completely abolished progression of protoplasts into S phase (Fig. 6, A and B). When MG132 was added after 48 h (the time at which about 10% of the protoplasts approached S phase (see Fig. 3C)), the FACS histogram showed two peaks corresponding to G 1 and G 2 content of DNA; no or very small amount of S phase protoplasts could be detected (Fig. 6C). Notably, treatment of protoplasts with MG132 diminished the supra-G 1 peak (Fig. 6, compare C with E). In contrast, the protease inhibitor leupeptin, used as a control in these experiments, enhanced protoplast progression into S phase (Fig. 6D). These findings demonstrate the indispensability of the ubiquitin proteolytic system for the second phase chromatin decondensation and for protoplasts entry into S phase; progression from S to G 2 phase is proteasome-independent. Consistent with the requirement for the proteasome system, Skp1 mRNA, a component of the cell cycle-regulated E3 ubiquitin ligase (SCF), is absent in leaves, but its level is markedly increased in protoplasts progressing into S phase (Fig. 6F). DISCUSSION The results presented here demonstrate that chromatin decondensation is an integral part in cellular dedifferentiation. Tobacco mesophyll cells re-entering the cell cycle displayed two successive phases of increased PI fluorescence intensity of nu-FIG. 1. First phase chromatin decondensation revealed in freshly prepared protoplasts. Nuclei were prepared from tobacco leaves (A) or from freshly prepared protoplasts (B), stained with PI, and subjected to FACS analysis. Note the higher PI fluorescence intensity in protoplast nuclei (G1) compared with leaf nuclei (G0). C, FACS histogram of a mixture of both types of nuclei. D, chromatin from freshly prepared protoplasts shows increased sensitivity to MNase. MNase (1000 units/ml) was applied for the indicated time periods (min) to nuclei prepared from tobacco leaves (lanes 2-6) or from freshly isolated protoplasts (lanes 7-11). DNA was extracted and separated by 1.6% agarose gel, followed by ethidium bromide staining. None (lanes 2 and 7) is DNA prepared from untreated nuclei. Positions of DNA size markers (M, lane 1) are indicated in base pairs (bp) on the left. clei prior to reactivation of S phase. These phases reflect increased accessibility of DNA to PI conferred by chromatin relaxation rather than DNA synthesis. The first phase (transition from G 0 to G 1 ), which occurs in the course of protoplast isolation, is ubiquitin-independent and does not require exogenous application of auxin and cytokinin. The second phase (transition from G 1 to supra-G 1 ) is auxin/cytokinin-dependent and occurs just before entry of cells into S phase. In agreement with these findings, an extra G 1 peak, not attributed to DNA synthesis but rather to conformational changes in chromatin, has been recorded by FACS in dedifferentiating red blood cells of frog (13). Also, two phases of chromatin decondensation, prior to reactivation of DNA synthesis, were found in chicken erythrocyte nuclei incubated in Xenopus egg extract (17). The mechanism underlying chromatin remodeling of somatic nuclei transplanted into Xenopus egg extract has recently been shown to require the activity of the chromatin-remodeling nucleosomal adenosine triphosphatase (ATPase) ISWI (19), a member of the SWI2/SNF2 superfamily. The ISWI is a subunit of several distinct nucleosome remodeling complexes that increase the accessibility of DNA in chromatin (20). Taken together, it appears that chromatin decondensation is a common mechanism underlying cellular dedifferentiation both in plants and animals. We propose that the first phase of chromatin decondensation represents a transitory phase that is necessary for activation of genes whose products are required for the establishment of competence for cell fate switch (Fig. 7). The ensuing cell fate determination may be governed by growth factors and environmental signals. In plant cells, auxin and cytokinin in- Protoplasts were induced to reenter the cell cycle by auxin and cytokinin. At the indicated time points after hormonal induction (t-0 to t-96), nuclei were isolated, and progression of protoplasts into S phase was monitored by FACS analysis (A-E). The appearance of protoplasts is shown (right panels, size bars ϭ 20 m); cell division patterns are seen at 72 and 96 h. Note the additional G 1 peak designated supra-G 1 (*G1) at D, a characteristic feature of protoplasts re-entering the cell cycle. To compare changes in fluorescence intensity in both G 1 and G 2 nuclei, we changed the amplifier gains for FL-2 to put the G 1 peak at about channel 360. duce dedifferentiation and re-entry into the cell cycle, auxin by itself may induce redifferentiation (21), whereas in the absence of phytohormones, cells die.
The occurrence of the supra-G 1 peak appears to be a unique feature of differentiated cells re-entering the cell cycle not found in cycling cells. The accumulation of protoplasts at the supra-G 1 , just before entry into S phase, indicates that the transition rate from G 1 to supra-G 1 is faster than from supra-G 1 to S phase. This progression pattern may mark a checkpoint control known as the restriction point in animals or START in yeast, a point of commitment to the mitotic cell cycle (22).
The stepwise manner by which chromatin decondensation occurs points to various levels of heterochromatin compaction, i.e. heterochromatin may not be a homogeneous but rather a heterogeneous structure displaying different degrees of chromatin condensation. This assumption is supported by recent findings demonstrating variations in chromatin (de)condensation within heterochromatin segments in meiotic chromosomes of Arabidopsis (23). We propose that the first phase of chromatin decondensation occurs in a less condensed heterochromatin, often called facultative heterochromatin, which is likely to be located at the boundary region between eu-and heterochromatin. This border chromatin might have an important biological significance in cellular plasticity, i.e. the interplay between differentiation, proliferation, and cell death. The role played by the ubiquitin system in regulating cell cycle progression has been well documented in animals (10 -11) and recently described also in plants (24). By using a specific inhibitor to the proteasome system we demonstrated its role in driving protoplasts into S phase. The fact that treatment of protoplasts with the proteasome inhibitor MG132 eliminated the supra-G 1 and, consequently, entry into S phase supports the notion that DNA decondensation is essential for reactivation of DNA synthesis (25). Furthermore, our results suggest that the activity of the ubiquitin proteolytic system plays a role, either directly or indirectly, in DNA relaxation, thereby enabling the assembly of replication proteins onto origins to initiate DNA replication.
It is long standing that auxin and cytokinin are required to induce cell division in plant tissue culture (26) and are indispensable for the progression of protoplasts into S phase (3)(4)(5).
Auxin and cytokinin presumably evoke signaling pathway(s) leading to a commitment for the mitotic cell cycle and reactivation of DNA replication (4,27,28). Notably, several mutations that confer resistance to auxin were found to encode components of the cell cycle-regulated ubiquitin ligase complex SCF (12) required for G 1 -S transition (10,11). This implies that auxin exerts its effect, at least in part, by inducing an orderly regulated protein destruction via the ubiquitin proteasome system. Given that the ubiquitin proteolytic system is required for redifferentiation of Zinnia mesophyll cells into tracheary elements (29), we suggest that the ubiquitin proteolytic system is not only involved in cell cycle progression but also in regulating cell fate determination.