A possible role of mitochondria in the apoptotic-like
programmed nuclear death of Tetrahymena thermophila
Takashi Kobayashi and Hiroshi Endoh
Division of Life Science, Graduate School of Natural Science and Technology, Kanazawa University, Japan
Mitochondria are known to play a major role in apop-
tosis or programmed cell death (reviewed in [1,2]).
Multiple cell death-associated factors have been identi-
fied in mitochondria. These factors may be divided into
three categories based on their functions: cyto-
chrome c, Smac ⁄ DIABLO, and Omi ⁄ HtrA2, all of
which are involved in caspase activation [3–7], while
apoptosis-inducing factor (AIF) and endonuclease G
(EndoG) are direct effectors of nuclear condensation
and DNA degradation [8,9]. The pro- and antiapoptotic
members of the Bcl-2 family proteins regulate loss of
mitochondrial inner membrane potential, which results
in the release of these apoptogenic factors [1,10]. The
involvement of mitochondria in apoptosis is common
among metazoans and plants [11]. Homologues of the
aforementioned mitochondrial apoptosis factors have
been identified even in protistans, such as the cellular
slime moulds and kinetoplastids [12,13]. Taking these
discoveries into consideration, the crucial role played
by mitochondria in apoptosis appears to have an early
evolutionary origin.
The ciliated protozoan Tetrahymena thermophila
undergoes a unique process during conjugation, i.e.
programmed nuclear degradation. Unicellular Tetra-
hymena has two morphologically and functionally dif-
ferent nuclei within the same cytoplasm. One is the
germinal micronucleus and the other is the somatic

and the degenerating macronucleus were colocalized in autophagosome
using two dyes for the detection of mitochondria. In addition, an endo-
nuclease with similarities to mammalian endonuclease G was detected in
the isolated mitochondria. When the macronuclei were incubated with iso-
lated mitochondria in a cell-free system, DNA fragments of 150–400 bp
were generated, but no DNA ladder appeared. Taking account of the pre-
sent observations and the timing of autophagosome formation, we conclude
that mitochondria might be involved in Tetrahymena PND, probably with
the process of oligonucleosomal laddering.
Abbreviations
AIF, apoptosis-inducing factor; DAPI, 4,6-diamino-2-phenylindole; DePsipher, 5,5¢,6,6¢-tetrachloro-1,1¢,3,3¢-
tetraethylbenzimidazolylcarbocyanine iodide; EndoG, endonuclease G.
5378 FEBS Journal 272 (2005) 5378–5387 ª 2005 FEBS
process known as ‘programmed nuclear death’ (PND),
because it is controlled by specific gene expression [16].
Programmed nuclear death resembles apoptosis in cer-
tain aspects: nuclear condensation, chromatin conden-
sation, and DNA laddering are observed during the
destruction of the parental macronucleus [16–18], and
several studies have demonstrated the involvement of
caspase-like enzymes [19,20]. Caspase family proteins
are essential for eukaryotic apoptosis, so it seems likely
that PND and apoptosis are regulated by similar
molecular mechanisms.
Previously, we identified caspase-8- and caspase-9-
like activities, which appear to be involved in the final
resorption of the parental macronucleus during PND
in T. thermophila, and suggested the involvement of
mitochondria in this process [19]. In mammalian apop-
tosis, caspase-8 and caspase-9 are known to be associ-

E’
F
F’
Fig. 1. DePsipher staining of cells during conjugation. The cells are
stained with DAPI (left) and the mitochondrial membrane potential-
dependent dye DePsipher (right). (A) A preconjugating cell. Micro-
nucleus (mic) and macronucleus (Mac) sets are observed by DAPI
staining. Most of the mitochondria show red fluorescence, while
green fluorescence is occasionally visible in cells that are stained
with DePsipher. (B) Nuclear selection-stage cell (6 h after mating
induction). One of four meiotic products is positioned at the paroral
zone. (C) Post-zygotic division I (PZD I)-stage cell (7 h). (D) PZD
II-stage cell (7.5 h). The program for degeneration of the old paren-
tal macronucleus begins at this stage. Degenerating meiotic prod-
ucts are observed in the posterior region of the cells (arrowheads
in C and D). Some of these nuclei are stained green by DePsipher
(white arrowheads), while others are not (yellow arrowheads). (E)
Mac IIp-stage cell (12 h). The degenerating old macronucleus
(dOM) is stained green by DePsipher. The micronuclei and macro-
nuclear anlagen (MA) do not display this staining pattern. (F) Mac
IIe-stage cell (16 h). The dOM also stains green during its degrada-
tion. The scale bar indicates 10 lm.
T. Kobayashi and H. Endoh Mitochondria in nuclear death of Tetrahymena
FEBS Journal 272 (2005) 5378–5387 ª 2005 FEBS 5379
Results
Co-localization of mitochondria and the
degenerating macronucleus in the
autophagosome
Previously we proposed an involvement of mitochon-
dria in PND from the results of preliminary experi-

not have diffused into the cytosol through autophago-
some membrane, resulting in specific localization to
the autophagosome containing degenerating macro-
nucleus. Small spots of green fluorescence, where
some mitochondria are thought to be incorporated
into small autophagosomes for turnover, were sporad-
ically observed, and some of them correspond to the
degenerating meiotic products (Fig. 1C and D; white
arrowheads).
A macronucleus that is committed to degeneration
is initially surrounded by the autophagosome, and is
eventually resorbed [17]. Thus, an autophagosome
that contains a degenerating macronucleus is called
‘the large autophagosome’ here. The large autophago-
some fuses with lysosomes, and becomes acidic in the
final step of PND [22,23]. DePsipher staining of the
macronucleus appeared initially during the stage of
autophagosome formation, and persisted until resorp-
tion of the parental macronucleus (Fig. 1D–F). Based
on these observations, we examined the possibility
that the monomeric forms of DePsipher localize to
the large autophagosome merely in response to low
pH. In order to exclude this possibility, conjugating
cells were stained with acridine orange (AO), which is
an indicator dye for acidic organelles [22]. Numerous
acidic organelles ) stained in orange ) were observed
Fig. 2. Distribution of acidic organelles dur-
ing degeneration of parental macronucleus.
The living cells during conjugation were stai-
ned with AO, which has different staining

previously [22].
Green fluorescence of DePsipher did not directly
show the localization of mitochondria in the autophag-
osome, as the red fluorescence corresponding to intact
mitochondria was not observed in the area. Therefore,
to confirm further the localization, the MitoTracker
Green ) a dye that accumulates in the lipid environ-
ment of mitochondria ) was used. With this dye, mito-
chondria can be easily localized, irrespective of
membrane potential. In the nonconjugating cells, the
mitochondria were arranged mainly along ciliary lows
(Fig. 3A). Similar staining patterns were observed for
conjugating cells (Fig. 3B–E). MitoTracker stained the
degenerating parental macronucleus, but not the other
nuclei (Fig. 3C–E). Moreover, the density of staining
was high around the degenerating macronucleus, pre-
sumably corresponding to the space between the
autophagosomal membrane and nuclear envelope
(Fig. 3C–E). In a previous study, mitochondria were
not observed in or outside the large autophagosome
using the electron microscope [17]. Considering this
report and our observations of the monomeric form of
DePsipher in the autophagosome together, the mito-
chondria taken in the autophagosome might be broken,
once they were incorporated into the autophagosomes.
These observations led us to an idea that the appar-
ently dead mitochondria (or broken membrane frag-
ments) that have lost membrane potential, together
with the degenerating parental macronucleus, are taken
up preferentially by the autophagosome. This, in turn,

FEBS Journal 272 (2005) 5378–5387 ª 2005 FEBS 5381
coincubated with the isolated mitochondria at neutral
pH, and an experimental condition was surveyed
(Fig. 4A). All of the following experiments were car-
ried out in the following conditions: 200 lL reaction
containing 20 lg protein, incubated for 120 min at
30 °C. The putative DNase had an optimum pH of
6.0–6.5 for the digestion of circular DNA (Fig. 4B,
lane 3 and 4). The divalent cation requirement for the
mitochondrial DNase activity was investigated
(Fig. 4C). As shown by inhibition with EDTA
(Fig. 4C, lanes 6–8), the mitochondrial nuclease activ-
ity required divalent cations. However, higher concen-
trations (5 and 10 mm)ofMg
2+
inhibited the DNA
cleavage activity (Fig. 4C, lane 4 and 5) and weak inhi-
bition was observed even in 1 mm of Mg
2+
(compare
lane 2 and 3 in Fig. 4C), indicating a different nature
from most other DNases. On the other hand, nicking
activity was unaffected by Mg
2+
, as shown by the
increased amounts of open circular DNA (Fig. 4C,
lanes 4 and 5). The addition of Mn
2+
and Ca
2+

of the mitochondrial nuclease activity. Reaction mixtures that contained 50 m
M Mops (pH 6.5) and 10 mM KCl, together with 1, 5, and 10 mM
MgCl
2
(lanes 3, 4, and 5, respectively), 1, 5, and 10 mM EDTA (lanes 6, 7, and 8, respectively), and 0.1, 1, and 5 mM ZnCl
2
(lanes 9, 10, and 11,
respectively) were assayed at 30 °C for 120 min. A standard reaction (S) was performed with 50 m
M Mops (pH 6.5) and 10 mM KCl (lane 2).
The undigested sample (U) was similar to the standard reaction, but contained no test sample (lane 1).
A
B
Fig. 5. (A) Fractionation PCR. A partial fragment of the mitochond-
rial large subunit ribosomal RNA (23S rRNA) was amplified by PCR,
using fraction samples that contained equal amounts of protein.
Lane, 1 pre-mitochondrial fraction; lane 2, mitochondrial fraction;
lane 3, post-mitochondrial fraction 1; lane 4, post-mitochondrial frac-
tion 2; lane 5, cytosolic fraction. PCR products were observed in
fractions 1–3 (lanes 1–3). (B) The nuclease activities of the fractions
under two different pH conditions. The reaction mixtures (200 lL)
contained 50 m
M sodium acetate (pH 5.0) or Mops (pH 6.5), 10 mM
KCl, 20 l g plasmid DNA as substrate, and 20 lg protein from each
fraction. The isolation of each fraction is described in Experimental
procedures. Lanes 1 and 6, pre-mitochondrial fraction; lanes 2 and
7, mitochondrial fraction; lanes 3 and 8, post-mitochondrial fraction
1; lanes 4 and 9, post-mitochondrial fraction 2; lanes 5 and 10, cyto-
solic fraction.
Mitochondria in nuclear death of Tetrahymena T. Kobayashi and H. Endoh
5382 FEBS Journal 272 (2005) 5378–5387 ª 2005 FEBS

(3 and 4) also showed nuclease activities, probably due
to low-level contamination with mitochondria and ⁄ or
the lysosomal enzyme itself (Fig. 5B, lanes 6, 8, 9).
Taking these results into consideration, it can be
judged that the DNase activity was derived mainly
from mitochondria rather than lysosomes.
To determine whether chromatin-associated DNAs,
as opposed to naked DNAs, are degraded by this
DNase the mitochondria were incubated with isolated
macronuclei as the substrate (Fig. 6). Under the pre-
sent experimental conditions of low osmotic pressure
and ⁄ or freeze–thawing of the mitochondrial fraction,
mitochondria are usually burst, resulting in the release
of the putative DNase as well as divalent cations. Pro-
longed incubation enhanced DNA cleavage, thereby
generating fragments of approximately 150–400 bp
(Fig. 6 lanes 3–5). Although the chromatin-sized lad-
ders were not identified, their sizes corresponded
roughly to the monomeric and dimeric forms of the
DNA ladder, as demonstrated previously for Tetra-
hymena [16,19].
Discussion
In the ciliated protozoan Tetrahymena, apoptosis-like
cell death is known to occur following treatment with
staurosporine [24], C
2
ceramide [25], or Fas-ligand
[26]. On the other hand, PND is a process in which
only the parental macronucleus is removed from the
cytoplasm of the next generation. This degradative

2 Mitochondrial 0.7196 ± 0.0435 1.00
3 Post-mitochondrial 1 3.1093 ± 0.1531 4.32
4 Post-mitochondrial 2 2.0750 ± 0.2412 2.88
5 Cytoplasmic 0.4356 ± 0.2008 0.61
Fig. 6. Nuclear DNA degradation by mitochondrial nucleases. The
isolated nuclei were incubated with mitochondria. The reaction was
carried out for 0 min (lane 1), 30 min (lane 2), 60 min (lane 3),
90 min (lane 4), and 120 min (lane 5). M represents the 100-bp
DNA ladder.
T. Kobayashi and H. Endoh Mitochondria in nuclear death of Tetrahymena
FEBS Journal 272 (2005) 5378–5387 ª 2005 FEBS 5383
rial degeneration may play a crucial role in autophago-
some formation, as the scattered small autophago-
somes shown by green fluorescence are probably
formed prior to the formation of the large autophago-
some (Fig. 1C and D). In either case, the autophago-
some can acquire some key molecule from the
sequestered mitochondria. This notion is supported by
the presence of a nuclease activity in the mitochondria
of Tetrahymena.
DNase activities of isolated mitochondria
In general, mitochondria have signalling pathways that
involve either AIF or EndoG, in which these molecules
execute apoptosis in a caspase-independent manner [2].
To identify mitochondrial factors in Tetrahymena,we
focused on EndoG-like enzyme activities, as EndoG is
a nuclease and AIF is not. In this study, we detected
strong nuclease activities in isolated mitochondria
(Fig. 4). This activity required divalent cations and
was strongly inhibited by the addition of Zn

uptake of the mitochondria in the autophagosome, as
discussed below.
Mitochondria as a possible executor of PND
The process of DNA degradation during PND can be
divided into three different steps, based on the sizes of
the DNA fragment generated [16–19]: (a) initial gen-
eration of high-molecular-weight (30-kb) DNA frag-
ments, followed by (b) oligonucleosome-sized ladder
formation, and (c) eventual complete degradation of
the DNA. The initial higher-order DNA fragmentation
precedes nuclear condensation [18]. Moreover, this
DNA fragmentation is a prerequisite for nuclear con-
densation. An as yet unidentified enzyme has been sug-
gested to act as a Ca
2+
-independent, Zn
2+
-insensitive
nuclease [18]. In mammalian apoptosis, AIF is known
to act as a caspase-independent death effector that
localizes to the mitochondrial intermembrane space
and translocates to the nucleus after its release from
mitochondria. Apoptosis-inducing factor causes chro-
matin condensation and degrades DNA into fragments
of sizes > 50 kb. To date, there has been no evidence
of an association between mitochondria and Tetrahym-
ena cell death, and mitochondrial homologues of mam-
malian apoptosis factors, such as AIF, have not been
identified in the Tetrahymena genome, despite the
ongoing Tetrahymena genome sequencing project.

into the enclosed nucleus, where the second step of
DNA degradation occurs, resulting in DNA laddering.
Evidence for this stage is provided by the observation
showing the localization of mitochondria at the
circumference of the nucleus (Fig 3.C–E). This hypo-
Mitochondria in nuclear death of Tetrahymena T. Kobayashi and H. Endoh
5384 FEBS Journal 272 (2005) 5378–5387 ª 2005 FEBS
thesis is consistent with our previous finding that the
initial degradation of DNA into the chromatin-sized
ladder is suspended once for a few hours, after which
period final DNA loss occurs rapidly [19]. In the final
stage, during which the macronucleus is resorbed, acid
phosphatase activity becomes localized deeper inside
the nucleus, as supported by acridine orange staining,
which reveals that the most highly condensed macro-
nuclei are acidic [22]. In addition, the caspase-8- and
caspase-9-like activities increase dramatically just
before this stage [19].
These three steps of DNA degradation are similar to
those seen in the apoptotic nucleus [29,30]. The large-
fragment-size DNA fragmentation and DNA laddering
are characteristics of the apoptotic nucleus, and the
final DNA degradation step in the autophagosome may
correspond to the phagocytosis of apoptotic bodies by
macrophages. The machinery for apoptosis may have
originated in the era of unicellular protistans, whereas
the apoptotic function of mitochondria is thought to
have evolved relatively recently. For instance, the
nematode Caenorhabditis elegans seems to have no
pathway for caspase activation by cytochrome c.In

diately under a fluorescence microscope with fluorescein
isothiocyanate (FITC) and green filters. For photography,
the cells were fixed with formalin (final concentration
0.5%) and stained with DAPI (4,6-diamino-2-phenylindole)
to visualize the nucleus. Acridine orange staining was per-
formed as described in Mpoke and Wolfe (1997) [22]. Mito-
Tracker Green (Molecular Probes Inc., Eugene, OR) stain-
ing has been described previously [33].
Subcellular fractionation
The late log phase cells were harvested by centrifugation at
1000 g for 5 min and washed with cold 10 mm Tris ⁄ HCl
pH 7.5. The washed cells were resuspended in a cold solu-
tion of 0.35 m sucrose, 10 mm Tris ⁄ HCl pH 7.5, 2 mm
EDTA (MIB; mitochondria isolation buffer), and homo-
genized using a Polytron homogenizer. To remove nuclei
and unbroken cells, the homogenate was centrifuged twice
at 1000 g for 5 min, and the precipitate was used as frac-
tion 1. To sediment the mitochondria, the supernatant
(fraction 1; premitochondrial fraction) was centrifuged at
8700 g for 10 min. To increase the purity, the crude mito-
chondria were resuspended in MIB that contained 10%
Percoll (Amersham Pharmacia Biotech AB, Uppsala, Swe-
den) and centrifuged at 5300 g for 5 min. The purified
mitochondria were washed once to remove Percoll and re-
suspended in MIB (fraction 2; mitochondrial fraction). The
supernatant of the crude mitochondrial fraction was centri-
fuged at 10 700 g for 10 min, and then the obtained super-
natant was further centrifuged at 18 100 g for 10 min. Both
precipitates were resuspended in MIB (fraction 3 designated
as postmitochondrial fraction 1, and fraction 4 as designa-

nol phosphate [35,36]. Each fraction sample (10 lL) was
mixed with 190 lL5mm p-nitrophenol phosphate dissolved
in 50 mm sodium acetate buffer (pH 5.0), and the mixture
was incubated at 30 °C for 60 min. To stop the reaction,
1 mL 0.4 m NaOH was added. The amount of liberated
p-nitrophenol was determined spectrophotometrically
at 410 nm.
Agarose gel assay for mitochondrial nuclease
activity
The standard nuclease reaction (200 lL) contained 20 lgof
the protein in the subcellular fraction, 2 lg substrate DNA
[pT7Blue (R) vector; Novagen Inc., San Diego, CA],
50 mm Hepes ⁄ NaOH pH 7.0, 10 mm KCl. The reaction
was incubated at 30 °C for 120 min. To stop the reaction,
300 lL of stop solution (100 mm Tris ⁄ HCl pH 7.5, 50 mm
EDTA, 2% SDS, 0.2 mgÆmL
)1
proteinase K) was added to
the reaction, and the mixture was incubated at 50 °C for
60 min. The stopped reaction was deproteinized with phe-
nol ⁄ chloroform (1 : 1), and the DNA was precipitated with
an equal volume of isopropanol. The precipitated DNA
was washed with 70% ethanol and diluted with 50 lLof
TE buffer (pH 8.0). The DNA samples (10 lL) were loaded
onto a 1% agarose gel, electrophoresed, and visualized by
staining with ethidium bromide.
In vitro nuclear apoptosis
Tetrahymena nuclei were isolated by the modified method
of Mita et al. [37]. Late log phase cells were harvested, and
washed with cold solution 1 (0.25 m sucrose, 10 mm

were incubated with mitochondrial fractions (20 lg pro-
teins) in 200 lL of reaction buffer (50 mm Mops pH 6.5,
10 mm KCl) at 30 °C. To stop the reaction, 300 lLof
stop solution (100 mm Tris ⁄ HCl pH 7.5, 50 mm EDTA,
2% SDS, 0.2 mgÆmL
)1
proteinase K, 100 lg ÆmL
)1
RNase
A) was added to the reaction, and the mixture was incu-
bated at 50 °C for 60 min. The stopped reaction was de-
proteinized with phenol ⁄ chloroform (1 : 1), and the DNA
was precipitated with an equal volume of isopropanol.
The precipitated DNA was washed with 70% ethanol
and diluted in TE buffer (pH 8.0). The DNA samples
were loaded onto a 2% agarose gel, electrophoresed, and
visualized by staining with ethidium bromide.
References
1 Kroemer G (1999) Mitochondrial control of apoptosis:
an overview. Biochem Soc Symp 66, 1–15.
2 Gurp M, Festjens N, van Loo G, Saelens X & Vandena-
beele P (2003) Mitochondrial intermembrane proteins
in cell death. Biochem Biophys Res Commun 304, 487–
497.
3 Liu X, Kim CN, Yang J, Jemmerson R & Wang X
(1996) Induction of apoptotic program in cell free
extracts: requirement for dATP and cytochrome c.
Cell 86, 147–157.
4 Li P, Nijhawan D, Budihardjo I, Srinivasula SM,
Ahamd M, Alnemri ES & Wang X (1997) Cytochrome

11 Ameisen JC (2002) On the origin, evolution, and nature
of programmed cell death: a timeline of four billion
years. Cell Death Differ 9, 367–393.
12 Arnoult D, Tatischeff I, Estaquier J, Girard M, Sureau
F, Tissier JP, Grodet A, Dellinger M, Traincard F,
Kahn A, Ameisen JC & Petit PX (2001) On the evolu-
tionary conservation of the cell death pathway: mito-
chondrial release of an apoptosis-inducing factor during
Dictyostelium discoideum cell death. Mol Biol Cell 12,
3016–3030.
13 Arnoult D, Akarid DA, Grodet A, Petit PX, Estaquier
J & Ameisen JC (2002) On the evolution of pro-
grammed cell death: apoptosis of the unicellular eukar-
yote Leishmania major involves cysteine proteinase
activation and mitochondrion permeabilization. Cell
Death Differ 9, 65–81.
14 Martindale DW, Allis CD & Bruns PJ (1982) Conjuga-
tion in Tetrahymena thermophila. A temporal analysis of
cytological stages. Exp Cell Res 140, 227–236.
15 Orias E (1986) Ciliate conjugation. In The Molecular
Biology of Ciliated Protozoa (Gall, JG, ed.), pp. 45–84.
Academic Press, San Diego.
16 Davis MC, Ward JG, Herrick G & Allis CD (1992)
Programmed nuclear death: apoptotic-like degradation
of specific nuclei in conjugating Tetrahymena. Dev Biol
154, 419–432.
17 Weiske-Benner A & Eckert WA (1987) Differentiation
of nuclear structure during the sexual cycle in Tetra-
hymena thermophila; II. Degeneration and autolysis of
macro- and micronuclei. Differentiation 34, 1–12.

2
ceramide on the inositol phospholipid
metabolism (uptake of
32
P,
3
H-serine and
3
H-palmitic
acid) and apoptosis-related morphological changes in
Tetrahymena. Comp Biochem Physiol C 122, 215–224.
26 Jaso-Friedmann L, Leary JH III & Evans DL (2000)
Role of nonspecific cytotoxic cells in the induction of
programmed cell death of pathogenic protozoans: Parti-
cipation of the Fas ligand-Fas receptor system. Exp
Parasitol 96, 75–88.
27 Widlak P, Li LY, Wang X & Garrard WT (2001)
Action of recombinant human apoptotic endonuclease
G on naked DNA and chromatin substrates: coopera-
tion with exonuclease and DNase I. J Biol Chem 276,
48404–48409.
28 Balk J, Chew SK, Leaver CJ & McCabe PF (2003) The
intermembrane space of plant mitochondria contains a
DNase activity that may be involved in programmed
cell death. Plant J 34, 573–583.
29 Robertson JD, Orrenius S & Zhivotovsky B (2000)
Nuclear events in apoptosis. J Struct Biol 129, 346–358.
30 Nagata S, Nagase H, Kawane K, Mukae N &
Fukuyama H (2003) Degradation of chromosomal
DNA during apoptosis. Cell Death Differ 10, 108–116.