• Here mitochondrial morphology and dynamics were investigated in Medicago truncatula cell-suspension cultures during growth and senescence.
• Cell biology techniques were used to measure cell growth and death in culture. Mitochondrial morphology was investigated in vivo using a membrane potential sensor probe coupled with confocal microscopy.
• Expression of a senescence-associated gene (MtSAG) was evaluated in different cell-growth phases. Mitochondria appeared as numerous, punctuate organelles in cells at the beginning of the subculture cycle, while interconnected networks were observed in actively growing cells. In senescent cells, giant mitochondria were associated with dying cells. The release of cytochrome c from mitochondria was detected in different growth phases of cultured cells.
• Studies on plant cell cultures allowed us to identify physiological and molecular markers of senescence and cell death, and to associate distinct mitochondrial morphology with cells under different physiological conditions.
The study of plant cell cultures may help to clarify numerous general aspects of organ senescence and contribute to better understanding of its relationship with programmed cell death (PCD). Senescence and PCD are associated with developmental processes in plants such as leaf senescence and petal wilting, although the relationship between the two terms is not yet well defined. Whereas van Doorn & Woltering (2004) argued that senescence is part of the programme leading to cell death, Thomas et al. (2003) maintain that senescence and PCD are, at best, only distantly related. Investigations using cultured plant cells have been performed, as cell growth, effects of depletion or addition of nutrients, cell autonomy and reversibility of processes can be analysed more easily in cultured cells compared with complex tissues.
The growth cycle of cultured plant cells is influenced by several physiological parameters, including nutrient availability, cell density, light, temperature and hormonal concentration. If cells are not subcultured at the end of the growth period, they begin senescing and PCD ensues. In cultured cells of Nicotiana plumbaginifolia, chromatin condensation, one of the hallmarks of PCD, occurs in cells during spontaneous senescence; however, if the senescing cell population is subcultured, chromatin condensation is reversed and cell death is prevented (O’Brien et al., 1998). Senescence has been associated with PCD in two other cell-cultured species: carrot and Arabidopsis thaliana, where the presence of oligonucleosomal DNA degradation has been shown to occur (Lo Schiavo et al., 2000). Recently we reported that Arabidopsis cultured cells express senescence-associated gene 12 (SAG12), a gene encoding a cysteine protease, at the end of subculturing cycle before entering PCD (Carimi et al., 2004). The expression of this gene, initially identified by Lohman et al. (1994), appears to be closely linked to leaf senescence and is not increased by many of the stress treatments that induce other senescence-associated genes (Weaver et al., 1998; Noh & Amasino, 1999). In cell cultures, senescence can occur spontaneously (cells allowed to grow without subculturing), but it can also be induced by chemicals: in Arabidopsis cell-suspension cultures, we demonstrated that high levels of 6-benzylaminopurine (BA) induce early expression of SAG12, which precedes cell death and DNA fragmentation (Carimi et al., 2003, 2004). The above data suggest that many molecular events occurring in plant cell cultures are similar to those occurring in plant organs, and support the conviction that cultured cells can be used as a model system for studies of cellular senescence.
Various cell death-signalling pathways are critically dependent on mitochondria, which are key players in the regulation mechanism in both animals (Kroemer & Reed, 2000) and plants (Zottini et al., 2002; Yao et al., 2004), not only through the release of pro-apoptotic factors (such as cytochrome c, which constitutes a key event in cell death), but also by altering their morphology (Frank et al., 2001). In animal cells, mitochondria can form dynamic, interconnected networks, and the relative rates of fusion and fission have been implicated in regulation of their number, size and shape (Mozdy & Shaw, 2003). During apoptosis, mitochondrial networks are fragmented and activation of the fission machinery is one of the primary triggers of this process (Bossy-Wetzel et al., 2003). A marked decline in mitochondrial functionality and an increase in abnormal mitochondrial cristae structures have been associated with cell ageing (Lenaz, 1998).
Recently it has been reported that the fusion–fission process also occurs in plants, as described in onion epidermal cells and BY-2 tobacco cells (Arimura et al., 2004b). In Arabidopsis, mitochondrial morphology appears to be controlled by dynamin-like proteins (Arimura et al., 2004a; Logan et al., 2004). In Arabidopsis leaves, it has been reported recently that morphological changes in mitochondria are one of the features of cell death that is induced by reactive oxygen species (Yoshinaga et al., 2005).
Medicago truncatula is a model legume species with several features that make it attractive for basic and applied research studies, in addition to the fact that sequencing of its genome is nearly complete (Cannon et al., 2005; Young et al., 2005). In this study we characterized M. truncatula cell-suspension cultures by defining several parameters of cell growth and ageing, eventually leading to PCD. Specifically, we focused our studies on changes in mitochondrial morphology during different growth phases and during senescence. In particular, we observed a constant association between spontaneous/induced cell death and the presence of giant mitochondria in cells showing condensed and stretched nuclei. In addition, cytoskeletal organization was analysed in healthy and senescent cells.
Materials and Methods
Cell cultures and treatments
The cell line JR was generated from roots of plantlets of Medicago truncatula L. cv. Jemalong (genotype 2HA) and routinely subcultured in modified Murashige & Skoog (1962) liquid medium (MSR4: 2.70 mm KH2PO4, 40 µm nicotinic acid, 33 µm thiamine hydrochloride, 60 µm pyridoxal hydrochloride) supplemented with 0.5 g l−1 malt extract, 30 g l−1 sucrose, 18 µm BA and 4.5 µm 2,4-dichlorophenoxyacetic acid (2,4-D). For subculture cycles, 1.2 ml packed cell volume was placed in 100-ml Erlenmeyer flasks containing 20 ml liquid medium. Cells were subcultured in fresh medium at 10-d intervals and maintained in a climate chamber on a horizontal rotary shaker (80 rpm) at 25 ± 1°C under a 16 h day length.
The pH of the media was adjusted to 5.7 ± 0.1 with NaOH before autoclaving at 121°C for 15 min. Growth regulators [2,4-D, BA, 6-benzylaminopurine riboside (BA-R), 1,3-diphenylurea (DPU)], when needed, were filter-sterilized and added directly to the medium. To determine the effect of high concentrations of cytokinin, 4-d-old cells were incubated with 27 µm DPU, 27 µm BA or 27 µm BA-R. Cell death was determined by spectrophotometric measurements of the uptake of Evan's blue stain, as described by Shigaki & Bhattacharyya (1999).
To determine dry weight, integer cells were separated from the culture medium and cell debris through a vacuum filtration unit (Sartorius, Florence, Italy). The collected cells were dried overnight at 60°C. For cell suspension-culture experiments, a randomized complete block design was used with three replicates (individual Erlenmeyer flasks). Each experiment was repeated three times.
Medicago truncatula cells were prepared for microscope analysis according to a previously described procedure (Traas et al., 1992) with minor modifications: cells were fixed by adding 0.5 ml of a solution containing 4% (v/v) paraformaldehyde in Pipes, EGTA, MgSO4 (PEM) buffer (100 mm Pipes pH 6.9, 10 mm EGTA, 10 mm MgSO4) to 0.5 ml culture. After 15 min, cells were washed three times in PEM buffer and finally resuspended in PEM containing 0.2% (w/v) Triton X-100 (Sigma-Aldrich, Milan, Italy) and 1 µg ml−1 of the DNA specific dye 4′,6-diamidino-2-phenylidone (DAPI) (Alexis Chemical, Vinci, Italy). The cells were overlaid on poly-l-lysine-coated (Sigma-Aldrich) microscope slides and nuclei were visualized using a Leica DMR epifluorescence microscope (Leica Microsystems Wetzlar GmbH, Wetzlar, Germany) with an excitation filter of 330–380 nm and a barrier filter of 400 nm. For nuclear morphology experiments, a randomized complete block design was used with three replicates (individual Erlenmeyer flasks). Each experiment was repeated at least three times. For each time point and treatment, 100 representative nuclei were counted. Images acquired by fluorescence microscopy were processed using Corel photo-paint (Corel Corporation, Dallas, TX, USA).
Analysis of mitochondria
A Nikon PCM2000 laser scanning confocal microscope (Nikon, Sesto Fiorentino, Italy) was used for analysis of mitochondrial morphology. The tetramethylrhodamine methyl ester dye (TMRM) (Molecular Probes, Leiden, the Netherlands), a mitochondrial membrane potential sensor, was used for visualizing mitochondria in cell culture as described by Zottini et al. (2002). Cell suspensions (300 µl) were collected at different times during their growth cycle and following different cytokinin treatments, and incubated in 700 µl MSR4 medium containing 1 µm TMRM for 15 min on a rotary shaker. Cells were centrifuged for 3 min at 10 000 g, the supernatant was discarded and the pellet washed twice with 700 µl MSR4. Cells were then resuspended in 500 µl MSR4. For microscope analysis, 100 µl cell suspension was placed on a microscope slide and visualized under a confocal microscope (excitation 548 nm, emission 573 nm). Control experiments were performed in the presence of the uncoupler carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (Zottini et al., 2002). Images were processed using Corel photo-paint. Mitochondrial dimensions were estimated with the ImageNT-MicroImage software (Casti Imaging, Venice, Italy). For mitochondrial morphology experiments, a randomized complete block design was used with three replicates (individual Erlenmeyer flasks). Each experiment was repeated three times.
Detection of cytochrome c release
For protein extraction, 3 g (FW) cells were harvested, frozen, powdered in liquid N2 and homogenized in two volumes of protein extraction buffer [0.3 m sucrose, 0.1 m Tris pH 7.5, 1 mm ethylenediaminetetraacetic acid (EDTA), 5 mm dithiothreitol (DTT) containing 2 mm phenylmethylsulfonyl fluoride (PMSF), 1 mm benzamidine, 5 mmɛ-aminocaproic acid, 5 µg ml−1 leupeptin, 2 µg ml−1 aprotinin and 0.7 µg ml−1 pepstatin]. The homogenate was centrifuged at 1500 g for 15 min at 4°C to eliminate debris. The supernatant was centrifuged at 10 000 g for 15 min at 4°C, and mitochondria were collected in the pellet.
Protein concentration was determined in each fraction studied by the Bradford method, using the Bio-Rad protein assay (Bio-Rad, Segrate, Italy). Mitochondrial (30 µg) and cytoplasmic (40 µg) proteins were separated by 15% (w/v) SDS–PAGE, transferred to a nitrocellulose membrane (Sartorius) and analysed with antibodies raised against human cytochrome c (Santa Cruz, CA, USA) that allow identification of a 12.4-kDa band (Bossy-Wetzel et al., 1998; Zottini et al., 2002). Equal loading and transfer of proteins were checked by staining the membranes with Ponceau S. Densitometric analyses of the blots were performed with a digital imaging analysis system (Chemi Doc; Bio-Rad).
Medicago truncatula suspension cells (1 ml) were fixed for 1 h in freshly prepared 3% (v/v) paraformaldehyde (Sigma-Aldrich) solution in microtubule-stabilizing buffer (MSB: 0.05 m Hepes pH 6.9, 1 mm EGTA, 0.5 mm MgCl2). After fixation, cells were washed three times for 15 min each in MSB, laid on poly-l-lysine-coated microscope slides and treated for 10 min with 1% bovine serum albumin (BSA) in PBS (137 mm NaCl, 2.7 mm KCl, 4.3 mm Na2HPO4 × 7H2O, 1.4 mm KH2PO4 pH 7.2). Cells were collected at 7 and 14 d of a subculture cycle and at 3 d after 27 µm 6-BA treatment. Each sample was incubated overnight at 4°C with the primary mouse monoclonal anti-β-tubulin antibody [1 : 200 dilution into PBS containing 1% (w/v) BSA; Sigma-Aldrich] and rinsed three times (15 min each) with PBS containing 1% (w/v) BSA. Cells were then incubated for 2 h at room temperature in the secondary fluorescein isothiocyanate-conjugated antibody (1 : 50 dilution into PEM-BSA; Sigma-Aldrich). After application of the Slow Fade Antifade reagent (Molecular Probes), cells were overlaid on a microscope slide and visualized with a Nikon PCM2000 laser scanning confocal microscope (excitation 488 nm, emission 510–540 nm). The images acquired with confocal microscopy were processed using Corel photo-paint. Each experiment was repeated three times.
Cells were harvested, frozen in liquid N2 and stored at −80°C. RNA was isolated using Trizol (Invitrogen, San Giuliano Milanese, Italy) following the manufacturer's instructions, and treated with DNase I (Ambion Inc., Austin, TX, USA). Total RNA from each sample (2.5 µg) was reverse-transcribed using PowerScript reverse transcriptase (Clontech Laboratories, Palo Alto, CA, USA) following the manufacturer's instructions. RT–PCR was carried out by following the manufacturer's instructions (Taq DNA Polymerase, Eppendorf, Hamburg, Germany). The 18S rRNA primers and competimers of the Quantum RNA Universal 18S Internal Standards Kit (Ambion) were used as an internal standard. The competimers were specially modified primers that anneal to the 18S rRNA templates but cannot be extended, resulting in the production of an attenuated 315-bp internal fragment (He et al., 2002). The primers used for RT–PCR analysis of MtSAG were 5′-GAAGGCTGCAATGGTGGTCTCA-3′ (forward) and 5′-CGGCCTCAATAGCAACGCTCAC-3′ (reverse). The following cycling conditions were used: 95°C for 30 s followed by 26 cycles at 94°C for 20 s, 65°C for 30 s, and 72°C for 60 s using a Hybaid PCR express thermal cycler (VWR, Milan, Italy). PCR products were visualized on 2% (w/v) agarose gels containing ethidium bromide (Sambrook et al., 1989), and densitometric analysis of the gels was performed using quantity one software (Bio-Rad). The relative abundance of the transcript within the samples was calculated as the ratio of the intensities of the MtSAG amplicon relative to the 18S rRNA amplicon.
Products of RT–PCR were recovered from agarose gels using the Qiaex II gel extraction kit (Qiagen, Milan, Italy), cloned into the plasmid pCR2.1-TOPO (Invitrogen), and sequenced.
Cell growth and spontaneous senescence in M. truncatula suspension cell culture
Medicago truncatula cells were subcultured with a cycle of approx. 10 d. If the medium was not changed at this time, a rapid decrease (decline phase) in dry cell weight was observed (Fig. 1a). This decrease correlated well with a rapid increase in cell death, as suggested by analysis of the survival curve (Fig. 1b).
Nuclear morphology, observed at different times after culture initiation, was investigated using DAPI staining coupled with fluorescence microscopy (Houot et al., 2001). Healthy cell nuclei showed diffuse staining (Fig. 1d). Two morphotypes of nuclei could be distinguished in old cells (17 and 21 d): nuclei with highly condensed chromatin in micronuclei (Fig. 1e); and stretched nuclei (Fig. 1f). In the decline phase of the culture, the number of condensed/stretched nuclei increased, reaching 20% at 14 d and up to 80% at 21 d (Fig. 1c).
A search in the TIGR Gene Indices M. truncatula database identified a transcript (TC89773), subsequently referred to as MtSAG, which codes for a protein displaying an amino acid residue identity of 50% with Arabidopsis SAG12 (Fig. 1h). In M. truncatula cell-suspension cultures, a low basal level of MtSAG expression was detected by RT–PCR analysis during the first 7 d of the subculture cycle. RT–PCR analysis of the expression of MtSAG showed that the transcript accumulated at day 8, defining the beginning of the reversible senescent phase (Fig. 1g). The relative abundance of the MtSAG transcript in 11-d-old cells was 80% higher than in 7-d-old cells, and this level of expression was maintained during the senescent phase. The increase of the MtSAG transcript preceded that of cell death and nuclear chromatin condensation. In the standard culture conditions mentioned above, the medium is renewed after 10 d. However, the medium change should be performed before reaching 40–50% cell death and 60% condensed/stretched nuclei, a physiological condition in which the reversible senescence phase ends and the entire cell culture enters PCD.
Mitochondrial morphology and dynamics during cell-culture growth and spontaneous senescence
Mitochondrial morphology and dynamics during the growth cycle were determined in cultured cells using TMRM staining coupled with confocal microscopy. At different times after culture initiation, cells were harvested from the flasks and stained with TMRM to visualize mitochondria. After 4 d of subculture, M. truncatula cells initiated the exponential growth phase, and mitochondria appeared as numerous, punctuate spheres distributed uniformly throughout the cytoplasm (Fig. 2a). In dividing cells, punctiform mitochondria were observed at the newly formed cell plate (Fig. 2b). At 7 d, when cells are in the exponential phase, mitochondria were present as interconnected networks (Fig. 2c,i; vermiform mitochondria). At day 10, at the end of the exponential phase when the cells stop dividing, the number of mitochondrial network structures decreased considerably (Fig. 2d). At this time, if cells are kept in spent medium they proceed in the senescence pathway and enter a cell-death program. At 14 d, mitochondria appeared as small, punctiform organelles (Fig. 2e,l) and the total number of mitochondria maintaining the membrane potential decreased progressively (Fig. 2e–g). At day 21, giant spherical mitochondria were observed (Fig. 2g,m), sometimes associated with a clumped nuclear distribution (Fig. 2h).
Throughout all phases of cell culture and senescence, we observed alterations in both the morphology and dynamics of mitochondria. In Fig. 3, a representative time-lapse series of confocal images of 4-d-old cells is shown, demonstrating that the mitochondrial distribution pattern mutates very rapidly (Video Clip S1 in Supplementary Material). These alterations are probably accountable for the active fusion–fission process in which the organelles are involved. By contrast, in steady-state cells, mainly punctiform mitochondria are observed that show very rare movements, completely absent in senescent cells (data not shown).
Release of cytochrome c in M. truncatula cell culture
An early event triggered in several forms of PCD is the release of cytochrome c into the cytoplasm after permeabilization of the mitochondrial outer membrane (Sun et al., 1999; Carimi et al., 2003). In order to define the different physiological phases of M. truncatula cell culture in more detail, we evaluated the release of cytochrome c by Western blot analysis. As shown in Fig. 4, the release of cytochrome c into the cytosolic fraction was seen clearly at the end of the log phase (at day 8), in the initial reversible phase of senescence. Its level remains stable in the cytoplasm for few days, and increased consistently in the very last phase of cell culture. However, it could also be detected as early as 5–7 d.
Cytokinin-induced PCD in M. truncatula cell culture
Recently we reported that high dosage of the cytokinin BA was able to induce PCD by accelerating the senescence process in both Arabidopsis and carrot cell lines (Carimi et al., 2004). Treatment of 4-d-old M. truncatula cell cultures with 27 µm BA similarly induced a threefold increment of cell death after 3 d. The ribosylated form of BA or DPU, an inhibitor of cytokinin oxidase (both at 27 µm), induced a similar increase in cell death (Fig. 5a). We observed a concomitant increase in the percentage of condensed plus stretched nuclei (Fig. 5b).
We next analysed mitochondrial morphology by confocal microscopy to determine if cytokinin-induced cell death was also associated with structural alterations at the mitochondrial level. After treatment for 3 d (27 µm DPU, BA, BA-R), the typical tubular mitochondrial structure observed in healthy cells disintegrated into round organelles (Fig. 5c). Giant, spherical mitochondria were also observed in treated cells, showing the morphology typically observed in culture cells when the spontaneous senescence phase proceeds into the PCD phase. In addition, cytokinins also stopped mitochondrial movement after 48 h of treatment (data not shown).
When cells were treated with 27 µm BA-R for 48 h, MtSAG transcript levels increased by 65% (Fig. 5d) concomitant with an increased release of cytochrome c in the cytoplasm (Fig. 5e).
Mitochondria and cytoskeleton organization
The mitochondrial association with the cytoskeleton prompted us to examine its organization in different phases of the subculture cycle, corresponding to differences in mitochondrial morphology.
Cells at 7 d with mitochondria observed as interconnected networks, contain a fine microtubule network as determined by immunofluorescence experiments using an antibody against β-tubulin (Fig. 6a). At 14 d, when the cells are senescent and mitochondria are present as small, punctiform organelles, cells showed disorganization of the microtubule network (Fig. 6b). A larger extent of microtubule degradation was observed in 3-d cytokinin (27 µm BA-R)-treated cells (Fig. 6c).
Here we show that different growth phases of cells in culture are associated with alterations in mitochondrial dynamics and morphology. In M. truncatula (in our opinion the best plant experimental system for visualizing mitochondrial morphology in vivo), we have observed differences in mitochondrial morphology ranging from interconnected networks to punctiform mitochondria.
In M. truncatula cell cultures, three phases can be distinguished: initial, in which cells condition their medium; log, in which cell division takes place; and final, in which cells stop dividing, elongate and senesce (Fig. 1a). Following the expression of MtSAG (an orthologue of Arabidopsis SAG12), we observed that it was already detectable at the beginning of the cell-subculture cycle and increased, as in Arabidopsis, at the end of the log phase (Fig. 1g). The difference in expression pattern of the two SAG species could be caused by the presence of a somewhat higher cell-death background level in M. truncatula (20%, cf. 8–10% in Arabidopsis), but still compatible with the ongoing subculture cycle. A Western blot analysis was then performed to evaluate the release of cytochrome c from mitochondria into the cytoplasm. This analysis revealed, in the initial phase of the culture, a modest presence of this marker protein, which increased at the end of the log phase and reached a stable level in later phases (Fig. 4). The cytochrome c pattern (Fig. 4) was similar to that observed for MtSAG (Fig. 1g). This result suggests that initial low levels of cytochrome c detected in the cytoplasm could depend on the cell-death background level present in M. truncatula cultures. Its successive increase coincides with the initial reversible phase of senescence. This could suggest a cell-autonomous cell-death process, at least during the initial phases of cell culture.
During the senescence phase, the level of cell death in the cell population increases slowly but constantly. If the medium is renewed, another subculture cycle reinitiates. In Arabidopsis cell cultures, the reversibility period is over when DNA laddering becomes detectable and the level of cell death reaches approx. 40% (Carimi et al., 2005). In the final phase of M. truncatula cell culture, the increase in cell death (Fig. 1b) is concomitant with the increased number of nuclei showing chromatin condensation (Fig. 1c); unless the medium is renewed before reaching 40–50% of cell death and 60% of condensed/stretched nuclei, the entire population of cells undergoes a massive, irreversible PCD.
Healthy, growing cells are characterized by a typical reticular arrangement of mitochondria; when cells go into senescence the network disintegrates and punctiform mitochondria are observed (Fig. 2). The number of punctiform mitochondria appears reduced with respect to that detected at the beginning of the log phase, probably because of a decrease in their membrane potential. Giant mitochondria are observed when high levels of cell death are reached in the cell culture and, at this stage, cytochrome c is detected principally in the cytoplasm, the entire cell population has entered into PCD and cannot be rescued. Giant mitochondria have been reported to appear during hypoxic stress in tobacco cells (Van Gestel & Verbelen, 2002).
During the process of senescence and cell death induced by cytokinins, the reticular arrangement of mitochondria disintegrates rapidly, MtSAG transcript levels increase, and giant mitochondria are detected together with an increasing release of cytochrome c in the cytoplasm (Fig. 5). This scenario is similar to that observed when spontaneous senescence proceeds into PCD. Hence mitochondrial changes in morphology and release of cytochrome c support a role of these organelles in the death process.
Mitochondria have been shown to associate with both microtubules and actin (Logan, 2003; Sheahan et al., 2004), and for this reason we analysed microtubule organization in cultured cells under different physiological conditions. In M. truncatula, during cell growth a normal microtubule array is observed that appears disorganized when senescence takes place, becoming almost disrupted in cytokinin-treated cells induced to enter PCD (Fig. 6). These results suggest a relationship between morphological alterations of mitochondria and cytoskeletal organization.
In conclusion, the analysis of mitochondrial changes in morphology, performed at different cell-growth phases, prompted us to identify giant mitochondria as a marker for senescent cells entering PCD. In addition, two significant results from the characterization of senescence in M. truncatula cell cultures allowed us to confirm MtSAG as a senescence marker, and to hypothesize that the release of cytochrome c in the cytoplasm could be an autonomous signal for cell death.
We are grateful to Professor Mario Terzi for helpful scientific discussion and critical reading of the manuscript. This work was supported by the national FIRB program: Post Genomica di Leguminose Foraggere.