(Received for publication, July 11, 1996, and in revised form, September 9, 1996)
From the Departments of Genetics and
¶ Biochemistry, Faculty of Natural Sciences, Comenius
University, 842 15 Bratislava, Slovakia, and the
Institut
Curie, Section de Recherche, Batiment 110, Centre Universitaire
Paris XI, 91 405 Orsay, France
Terminal segments (telomeres) of linear
mitochondrial DNA (mtDNA) molecules of the yeast Candida
parapsilosis consist of large sequence units repeated in tandem.
The extreme ends of mtDNA terminate with a 5 single-stranded overhang
of about 110 nucleotides. We identified and purified a
i
ochondrial
elomere-
inding
rotein (mtTBP)
that specifically recognizes a synthetic oligonucleotide derived from
the extreme end of this linear mtDNA. MtTBP is highly resistant to
protease and heat treatments, and it protects the telomeric probe from
degradation by various DNA-modifying enzymes. Resistance of the complex
to bacterial alkaline phosphatase suggests that mtTBP binds the very
end of the molecule. We purified mtTBP to near homogeneity using DNA
affinity chromatography based on the telomeric oligonucleotide
covalently bound to Sepharose. Sodium dodecyl sulfate-polyacrylamide
gel electrophoretic analysis of the purified fractions revealed the
presence of a protein with an apparent molecular mass of ~15 kDa. UV
cross-linking and gel filtration chromatography experiments suggested
that native mtTBP is probably a homo-oligomer. MtTBP of C. parapsilosis is the first identified protein that specifically
binds to telomeres of linear mitochondrial DNA.
Although circularity is accepted as the most common form of
mitochondrial DNA (mtDNA), there are numerous examples of organisms with linear mtDNA, raising questions about their evolutionary origin
and mode of replication (1). Mitochondrial genomes of many yeast
species are represented by linear DNA molecules with defined terminal
structures (2-5). Two different types of yeast linear mtDNA have been
described. Type 1 (Pichia pijperi, Pichia jadinii, Williopsis mrakii) possesses inverted terminal
repeats with a covalently closed single-stranded hairpin at each ends (6). In contrast, terminal structures of type 2 linear mtDNA (Candida parapsilosis, Candida salmanticensis,
Pichia philodendra) have inverted terminal repeats
consisting of units repeated in tandem. The variable number of repeat
units generates a population of mtDNA molecules of heterogeneous size.
The largest fraction of mtDNA molecules of C. parapsilosis
is represented by the shortest molecules, which contain only one
incomplete tandem repeat unit. Regardless of the number of repetitions,
both ends of the mtDNA molecule terminate at a unique position in the
last repeat unit sequence thus leaving an open structure with a 5
single-stranded extension of about 110 nucleotides (7). This is in
contrast to typical nuclear telomeres of most eukaryotes consisting of simple arrays of tandem G-rich repeats that run 5
to 3
at the chromosome ends and protrude as 3
overhangs (8). The structure of the
mitochondrial telomeres of C. parapsilosis is more
reminiscent of those of Tetrahymena mtDNA, although the
telomeric tandem repeat unit of the latter is considerably shorter (9,
10). The role of terminal structures in the replication and maintenance
of C. parapsilosis mtDNA is not known. Their unique
organization raised many intriguing issues, e.g. (i) how the
terminal 5
single-stranded extensions are stabilized in mitochondria;
(ii) why DNA polymerase stops at the 3
end and does not fill the
protruding 5
overhang; (iii) how the 5
overhang is generated; and
(iv) how the shortest molecules restore the missing sequence, to name a
few.
The search for proteins specifically recognizing mitochondrial telomeres of C. parapsilosis may reveal essential components of the telomere replication machinery (e.g. analog of the nuclear telomerase) and proteins that play a role in the stabilization, maintenance, and proper segregation of the linear mtDNA. For nuclear genomes, specific telomere-binding proteins (TBPs)1 have been identified in protozoa (11-16), slime molds (17, 18), yeasts (19-26), algae (27), plants (28), and vertebrates (29-32). These proteins presumably protect the telomere DNA from nucleolytic degradation, mediate telomere-telomere and telomere-nuclear matrix interactions, promote the formation of a typical telomeric structure, and regulate the accessibility of telomeres to the replication machinery (33). They may also participate in the events associated with the cell cycle (34) and cellular senescence (35).
In this paper we describe the identification and purification of the
first mitochondrial telomere-binding protein (mtTBP). MtTBP from yeast
C. parapsilosis is a heat- and protease-resistant protein
that interacts specifically with the sequence of the 5 single-stranded
extension from the extreme ends of a linear mtDNA molecule and protects
it against various DNA-modifying enzymes. The possible role of mtTBP in
telomere stabilization is also discussed.
Oligonucleotides were synthesized by Genset
(Paris, France) (their sequences and positions in C. parapsilosis mtDNA are shown in Fig. 1). DNase I
and proteinase K were purchased from Boehringer (Mannheim, Germany).
Cyanogen bromide-activated Sepharose 4B and RNase A were obtained from
Sigma. Trypsin was from Serva (Heidelberg, Germany), exonuclease III
and bacterial alkaline phosphatase were provided by Life Technologies,
Inc. (Paisley, UK). T4 polynucleotide kinase and mung bean nuclease
were from New England Biolabs (Beverly, MA), and S1 nuclease was
obtained from Pharmacia (Uppsala, Sweden). [-32P]ATP
was from Amersham (Braunschweig, Germany) or New England Nuclear (Les
Ulis, France). Rabbit antisera against mitochondrial malate
dehydrogenase and porin of Saccharomyces cerevisiae were generously provided by Dr. Gottfried Schatz (Biocenter, University of
Basel, Switzerland). Rabbit antisera against aconitase of S. cerevisiae were obtained from Dr. P. Haviernik (Department of Biochemistry, Comenius University, Bratislava). Restriction enzymes were from New England Biolabs.
Preparation of Mitochondrial Membrane Protein Extracts
Strain SR23 (CBS 7157) of C. parapsilosis is a stock from the Department of Biochemistry, Comenius University, Bratislava. Yeasts were grown in complete liquid medium (1% Bacto-yeast extract (Difco, Detroit, MI), 1% Bacto-peptone (Difco), 2% glucose) at 28 °C with shaking until early stationary phase. For standard preparation, 4 liters of culture were harvested by centrifugation, and cells were washed with cold water and resuspended in 300 ml of 25 mM EDTA (pH 7) containing 4.5% (v/v) 2-mercaptoethanol. After a 20-min incubation at room temperature cells were washed with SCE buffer (1.2 M sorbitol, 0.1 M sodium citrate, 60 mM EDTA (pH 7)), resuspended in 300 ml of SCE containing 50 mg of zymolyase 20T (Kirin Brewing Ltd., Tokyo, Japan) and incubated at 37 °C until about 90% of cells converted to spheroplasts (60-90 min). The reaction was stopped by the addition of an equal volume of ice-cold 1.2 M sorbitol, and spheroplasts were washed twice with 1.2 M sorbitol.
The pellet was resuspended in 200 ml of 0.7 M sorbitol, 60 mM Tris-Cl (pH 7.4), 1 mM EDTA and mixed with a
magnetic stirrer for 10 min at 4 °C. Spheroplasts were mechanically
broken in Kenwood mixer (30 s at maximal speed), the pH was adjusted to
7.0 with several drops of 1 M Tris base, and cell debris
was pelleted by centrifugation for 10 min at 2,000 × g. Mitochondria were pelleted from the supernatant by a
10-min centrifugation at 10,000 × g and washed three
times with 0.5 M sorbitol, 50 mM Tris-Cl (pH 7.4), 1 mM EDTA. Fractions from spheroplast crude lysate,
postmitochondrial supernatant, and mitochondria were stored in small
aliquots at 70 °C.
For the extraction of mtTBP, mitochondria were resuspended in 10 volumes of buffer A (10 mM Tris-Cl (pH 7.4), 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride,
10 µg/ml leupeptin, 5 µg/ml pepstatin A) and mixed for 20 min at
4 °C with a magnetic stirrer. The suspension was centrifuged for 15 min at 10,000 × g, and the pellet was resuspended in
10 volumes of buffer A using a Potter homogenizer. After a 15-min
incubation on ice, the suspension was homogenized using a Potter
homogenizer and incubated on ice for an additional 30 min. Triton X-100
was added to a final concentration of 0.1% followed by a 60-min
incubation on ice. Triton-insoluble material was pelleted by
centrifugation for 45 min at 55,000 × g. The resulting
pellet was resuspended in 5 volumes of buffer A containing 2 M NaCl using a Potter homogenizer and incubated on ice for
an additional 60 min. The suspension was centrifuged for 45 min at
55,000 × g. The supernatant containing the majority of
mtTBP was dialyzed overnight against 200 volumes of 10 mM
Tris-Cl (pH 7.4), 1 mM EDTA, aliquoted, and frozen at
70 °C. Protein concentration was determined as described by
Bradford (36). The preparation is called crude mitochondrial membrane
extract.
Oligonucleotides were 5 end labeled using
[
-32P]ATP by T4 polynucleotide kinase and separated
from nonincorporated nucleotides by chromatography on a BioSpin-6
column (Bio-Rad).
5 µg of crude mitochondrial membrane protein extract or 1 ng of the purified mtTBP was incubated with 1 ng of labeled oligonucleotide in 10 µl of DNA-binding buffer (10 mM Tris-Cl (pH 7.4), 50 mM NaCl) containing 1 mg/ml poly(dI·dC) for 10 min at room temperature. Poly(dI·dC) was omitted when treated with DNA-modifying enzymes (see below). DNA-protein complexes were resolved by electrophoresis on 4% polyacrylamide gel in 0.5 × TBE (4.5 mM Tris borate, 1 mM EDTA) at 10 V/cm for 90 min. Gels were dried under vacuum at 80 °C and exposed to Kodak X-Omat film. In some experiments, samples were treated with different protein- or nucleic acid-modifying enzymes before or after the DNA-protein binding reaction as indicated in corresponding figure legends. If necessary, different ions were added prior to the addition of enzymes.
UV Cross-linking in SolutionSamples indicated in
corresponding figure legends were after the DNA binding reaction
(10-µl total volume, see above) illuminated by UV light (312 nm,
Vilbert-Lourmat transilluminator) from a distance of 3 cm for 45 min.
DNA-protein complexes were resolved by 12% SDS-PAGE by the method of
Laemmli (37). Gels were heat dried under vacuum and exposed to Kodak
X-Omat film at 70 °C between intensifying screens.
For UV cross-linking in
situ, a modification of the protocol of Williams et al.
(38) was employed. Briefly, after the mobility shift assay, gel was
irradiated for 30 min from a distance of 5 cm by UV light (312 nm),
positions of the free and shifted probe were visualized by
autoradiography, and bands corresponding to both free and complexed
probe were excised. Gel slices were boiled for 5 min in 3 ml of a
modified SDS-PAGE sample buffer (1% SDS, 3 mM
dithiothreitol, 125 mM Tris-Cl (pH 6.8)) and then placed between gel plates in the stacking gel area. Covalently bound DNA-protein complexes were separated from a free probe by 12% SDS-PAGE
by the method of Laemmli (37), and the gel was dried at 80 °C under
the vacuum and exposed to Kodak X-Omat film at 70 °C between
intensifying screens.
TEL51-Sepharose was
prepared according to a protocol described by Kadonaga and Tjian (39).
The concentration of TEL51 was estimated to be 15 µg/ml of Sepharose.
3 ml of beads was packed into a column and equilibrated by 20 ml of
buffer B (10 mM Tris-Cl (pH 7.4), 50 mM NaCl,
0.1 mM EDTA). Dialyzed mitochondrial membrane protein
extract (total protein of about 10 mg) was combined with poly(dI·dC)
(final concentration 65 µg/ml) and incubated at room temperature for
15 min. The DNA-protein mixture was loaded on the column and incubated
with beads for 5 min. The flow-through was reloaded on the column, and
this step was repeated three times. Beads were then washed with 50 ml
of buffer B. Bound proteins were eluted stepwise with buffer B
containing 0.3, 0.5, 0.7, 1.0, 1.5, 2.0, and 3.0 M NaCl. In
some experiments, beads were washed with 50 ml of buffer B containing
0.5 M NaCl to remove contaminants weakly bound to the
ligand. Selected samples were dialyzed against 1,000 volumes of 10 mM Tris-Cl (pH 7.4), 0.1 mM EDTA, 20% glycerol and stored at 70 °C. For SDS-PAGE analysis, selected samples were
dialyzed against double distilled water, concentrated by lyophilization, and dissolved in a SDS-PAGE sample buffer. Proteins were then separated by 15% SDS-PAGE according to the method of Laemmli
(37) and stained with silver using a silver stain kit (Pharmacia).
Proteins (50 µg) from the corresponding fractions were separated by 13.5% SDS-PAGE according to the method of Laemmli (37) and electrotransferred to a nitrocellulose filter (BA85, Schleicher & Schuell) in buffer consisting of 25 mM Tris, 192 mM glycine (pH 8.3) for 60 min at 250 mA. The blots were first blocked for 2 h in TBS buffer (10 mM Tris-Cl (pH 7.4), 150 mM NaCl) containing 2% skim milk (Difco) and then probed with anti-malate dehydrogenase (diluted 1:1,000 in a blocking solution), anti-porin (1:500), and anti-aconitase (1:500) antisera, respectively, followed by anti-rabbit immunoglobulin G-alkaline phosphatase conjugate (Sigma). The blots were developed with 0.1 M NaHCO3, 1 mM MgCl2 (pH 9.8) containing 0.15 mg/ml 5-bromo-4-chloro-3-indolyl phosphate toluidinium (Sigma), 0.3 mg/ml nitro blue tetrazolium (Sigma).
Enzyme AssaysPublished procedures were used to assay glucose-6-phosphate dehydrogenase (40) and cytochrome c oxidase (41).
DNA Isolation and Restriction AnalysisAppropriate fractions were treated with proteinase K in 50 mM Tris-Cl (pH 7.5), 5 mM EDTA, 150 mM NaCl, 1% N-laurylsarcosine for 2 h at 50 °C. Nucleic acids were extracted by phenol:chloroform:isoamyl alcohol (25:24:1) and precipitated with ethanol. Samples were treated with RNase A following the phenol:chloroform:isoamyl alcohol extraction and ethanol precipitation. DNA samples (250 ng) were digested with restriction enzymes HindIII and PvuII according to the manufacturer's instructions and separated by agarose gel electrophoresis (42).
Detection of the Protein Component of DNA-Protein Complex by SDS-PAGEA gel retardation assay was performed as described above except that 100 ng of the probe and a corresponding amount of protein were subjected to DNA-protein binding reaction. After separation of the DNA·mtTBP complex, the gel was exposed, and the shifted band and bands from the corresponding areas in control lanes (probe without mtTBP and mtTBP without a probe) were excised and denatured by boiling for 5 min in 3 ml of a modified SDS-PAGE sample buffer (1% SDS, 3 mM dithiothreitol, 125 mM Tris-Cl (pH 6.8)). The excised band was then placed between gel plates in the stacking gel area, proteins were separated by 15% SDS-PAGE by the method of Laemmli (37) and stained with silver using a silver stain kit.
Renaturation of mtTBP Isolated from SDS-PAGE GelElution of mtTBP from the gel, SDS removal, guanidine hydrochloride treatment, and renaturation were performed essentially as described by Hager and Burgess (43) except that renaturation of the protein was achieved by overnight dialysis of the denatured sample against 50,000-fold excess of 10 mM Tris-Cl (pH 7.4), 1 mM EDTA at 4 °C.
Determination of Molecular Mass of Native mtTBPThe apparent molecular mass of native mtTBP was determined using gel filtration FPLC on a Superose 12 column (Pharmacia) equilibrated with 10 mM Tris-Cl (pH 7.4), 0.1 mM EDTA, 150 mM NaCl. The column was calibrated with catalase (240 kDa), aldolase (158 kDa), bovine serum albumin (68 kDa), ovalbumin (45 kDa), chymotrypsin (25 kDa) and cytochrome c (12.5 kDa).
Reproducibility of DataAll experiments were repeated at least twice with similar results.
Preliminary experiments using the Southwestern assay
indicated that most of the DNA binding activity is associated with the inner mitochondrial membrane (data not shown). The majority of membrane-associated DNA-binding proteins were extracted by treatment with 2 M NaCl. To assess the presence of a putative mtTBP
in the extract, an oligonucleotide identical to the sequence of 51 nucleotides (TEL51) from the extreme 5 end of the mitochondrial
telomere (Fig. 1) was used in a gel retardation assay (Fig.
2). Mixing of TEL51 with mitochondrial membrane extract
resulted in a DNA-protein complex that migrated slower than a free
probe (Fig. 2, lanes 1 and 2). A 5-100-fold
molar excess of unlabeled TEL51 competed quantitatively with the
labeled probe for complex formation (Fig. 2, lanes 3-5).
Similarly, an oligonucleotide representing the last 31 telomeric bases
(TEL31) proved to be a competitor (Fig. 2, lanes 6-8). In
contrast, oligonucleotides derived from other parts of mtDNA of
C. parapsilosis did not compete with labeled TEL51 as
effectively as homologous competitors (Fig. 2, lanes 9-17).
Moreover, oligonucleotides complementary to TEL51 (oligonucleotide 4 in
Fig. 1), yeast tRNA, and total mitochondrial RNA from C. parapsilosis were found to be ineffective competitors (data not shown). In addition to a major DNA-protein complex representing a
specific interaction between the telomeric probe and a putative mtTBP,
a slower migrating complex was also observed in some experiments (Fig.
2, lane 2). However, in this case all nonspecific
competitors competed successfully with a probe (Fig. 2, lanes
3-17). Moreover, the formation of this complex was not observed
when purified mtTBP was used for the gel retardation assay (see below),
suggesting that the slower migrating complex probably resulted from
nonspecific DNA-protein interaction.
MtTBP Activity Is Enriched in the Mitochondrial Fraction
Several assays were performed to assess the purity of
mitochondria used for the extraction of mtTBP (Fig. 3).
First, mtTBP activity was shown to be completely absent in the
postmitochondrial supernatant (Fig. 3A). Second, the
mitochondrial fraction was highly enriched for the mitochondrial
immunological markers porin, malate dehydrogenase, and aconitase (Fig.
3B). Third, mitochondrial preparation was almost devoid of
the cytosolic enzymatic marker glucose-6-phosphate dehydrogenase, which
was in contrast to the distribution of cytochrome c oxidase
activity (Fig. 3C). Finally, mtDNA isolated from the
mitochondrial fraction was not substantially contaminated by nuclear
DNA (Fig. 3D). All of these results strongly suggest that
mtTBP is localized in mitochondria of C. parapsilosis.
MtTBP and mtTBP·mtDNA Telomere Complex Are Resistant to Heat and Several Protein- and Nucleic Acid-modifying Enzymes
C.
parapsilosis is a petite negative yeast, and the stability of its
mtDNA is probably essential for cell viability. Resistance of the
mitochondrial telomeric structure to various damaging agents including
proteases and DNA-modifying enzymes might reflect the need to maintain
the integrity of the mitochondrial genome. The susceptibility of mtTBP
to proteinase K and trypsin and the effect of heat on mtTBP activity
were assayed by pretreatment of the extract followed by a gel
retardation analysis using labeled TEL51 (Fig.
4A). Interestingly, mtTBP was highly
resistant to proteinase K (Fig. 4A, lane 4) and
partially resistant to both trypsin and to 15-min incubation at
62 °C (Fig. 4A, lanes 5 and 6).
Moreover, neither an overnight incubation at 30 °C nor treatment
with 0.1% SDS destroyed the mtTBP activity (data not shown).
Next, the susceptibility of the mtTBP·TEL51 complex to various DNA-modifying enzymes was evaluated (Fig. 4B). None of the tested enzymes completely destroyed the complex, whereas a free probe that served as an internal control was degraded completely in each case. To demonstrate that the enzymes are active in the mitochondrial extracts, reactions were also performed in the presence of the molar excess of homologous competitor (Fig. 4B, 3rd, 6th, 9th, 12th, 15th, and 18th lanes). Since the probe was terminally labeled by T4 polynucleotide kinase, resistance of mtTBP·TEL51 to bacterial alkaline phosphatase (Fig. 4B, 4th-6th lanes) suggests that mtTBP protects the very end of the mitochondrial telomere of C. parapsilosis. These results imply that mtTBP may play a similar protective role in vivo thus preventing mtDNA from degradation.
UV Cross-linking of mtTBP to the Extreme End of mtDNATo
estimate the apparent molecular mass of a mtTBP·TEL51 complex, UV
cross-linking experiments both in solution and in situ were
performed (Fig. 5, A and B). Both
of these assays gave rise to two prominent bands with apparent
molecular masses of ~32 and 40 kDa, respectively (Fig. 5, panel
A, lane 4, and panel B, lanes 2 and 4). The larger complex was more abundant in the sample
after UV cross-linking in solution, whereas irradiation of
mtTBP·TEL51 complex in mobility shift gel resulted in a higher
proportion of the faster migrating complex (Fig. 5B). The
presence of a 100-fold excess of either TEL51 or TEL31 completely
competed out a labeled probe from the complex (Fig. 5A,
lanes 5 and 6). On the other hand,
oligonucleotides derived from different regions of mtDNA of C. parapsilosis were not as effective as specific competitors (Fig.
5A, lanes 7-9), which is consistent with the
results obtained from gel retardation assays.
Purification of mtTBP on DNA Affinity Chromatography
To
purify mtTBP from a crude mitochondrial membrane extract, DNA affinity
chromatography based on TEL51 covalently bound to Sepharose was
employed. MtTBP activity was eluted from the column with relatively
high salt concentrations as shown by both gel retardation (Fig.
6A) and UV cross-linking (Fig.
6B). Formation of a novel complex observed in a sample
eluted by 0.5 M NaCl (Fig. 5A, lane
4) was probably due to the presence of a less specific TEL51-binding protein and the absence of mtTBP in this fraction. Silver
staining of proteins after SDS-PAGE revealed the presence of a major
protein band with an apparent molecular mass of about 15 kDa (Fig.
6C). Strong interaction between mtTBP and immobilized telomeric oligonucleotide caused most contaminants to elute with 0.3-0.5 M NaCl. Purified preparations of mtTBP from
TEL51-Sepharose contained only minor impurities with a molecular mass
above 40 kDa.
15-kDa Protein Represents mtTBP
Because of an apparent
contradiction between the results of the UV cross-linking assay (which
gave rise to two prominent bands with apparent molecular masses of
~32 and 40 kDa, respectively) and the abundance of 15-kDa protein
(p15) in the purified fractions, two experiments were performed to
prove that p15 represents mtTBP. First, the mtTBP·TEL51 complex was
excised from the mobility shift gel, denatured, and separated by
SDS-PAGE. As controls gel slices were used from the same regions of the
lanes with a probe and mtTBP activity containing fraction alone,
respectively. A protein band comigrating with purified p15 was detected
only in a sample containing both probe and mtTBP (Fig.
7A).
Second, affinity-purified p15 was isolated from the gel after SDS-PAGE, precipitated with acetone to remove SDS, resuspended in 6 M guanidine hydrochloride, renatured by dialysis, and tested for mtTBP activity (Fig. 7B, lane 3). As a control, the same protocol was performed on a gel slice in which no protein was present (Fig. 7B, lane 4). Although it lost some of the mtTBP activity, renatured p15 was able to bind the TEL51 probe in a gel mobility shift assay. Taken together these data demonstrate that p15 represents mtTBP and is responsible for the generation of the complexes in both gel retardation and UV cross-linking assays.
FPLC-Superose 12 ChromatographyIt was shown previously that
under standard SDS-PAGE conditions, covalently linked protein-DNA
complexes usually migrate with the same electrophoretic mobility as the
protein alone (38). However, purified mtTBP migrated in SDS gel as a
15-kDa protein band (Fig. 6C), whereas the UV cross-linked
mtTBP·TEL51 complex exhibited an apparent molecular mass above 30 kDa
(Fig. 5). The possibility that this discrepancy might be due to an
oligomeric state of the native mtTBP was tested using FPLC gel
filtration chromatography on Superose 12 (Fig. 8).
TEL51-retarding activity was eluted between bovine serum albumin (68 kDa) and ovalbumin (45 kDa), suggesting that native mtTBP is either a
trimer or tetramer. The absence of any major protein band other than
p15 in purified mtTBP preparations suggests that mtTBP is a
homo-oligomer.
Several lines of evidence indicate that nuclear telomeres form specific nucleoprotein structures (telosomes), consisting of telomeric DNA and various proteins involved in both replication and stabilization of the chromosome ends. Moreover, these structures are responsible for distinct interactions in the nucleus and influence the expression of subtelomeric genes (8, 44-47).
Specific nuclear TBPs from a variety of organisms were recently
identified (33, 48). The first isolated and best characterized is TBP
from Oxytricha nova which is a heterodimer of two different polypeptides and
(12, 13). It binds specifically to the 3
overhang both in vitro and in vivo, protects the
telomeric DNA from chemical modification and nuclease digestion, and
promotes the formation of a G4 structure typical for telomeric DNA
(49). Another well characterized telomeric protein is RAP1 of S. cerevisiae which is a multifunctional polypeptide that binds to
the duplex region of the telomeric sequence both in vitro
and in vivo and plays not only a protective but also
regulatory role at telomeres (20-22, 26, 50-53). The crystal
structure of its DNA binding domain in complex with telomeric DNA was
reported recently (54). Although several TBPs were isolated from
vertebrates (29-32), only the protein described recently by Chong
et al. (29) has been shown to be at telomeres in
vivo.
Linear mtDNAs were described recently in more organisms than were
expected previously. Their genomic organization and terminal structures
differ greatly. Different types of mitochondrial telomeres may imply
different strategies of their replication which do not resemble the
replication of typical nuclear telomeres. For example, linear mtDNAs
found in several species from yeast genera Pichia and
Williopsis possess a terminal hairpin that may allow the
formation of circular replication intermediates (2, 6). In contrast, termini of linear mtDNA of C. parapsilosis consist of tandem
repeats that end with a 110-nucleotide-long 5 single-stranded overhang that does not permit the formation of circles (7). It is not clear how
the 5
overhang is formed, how it is stabilized, and why DNA polymerase
does not use this sequence as a template.
Although data about nuclear TBPs accumulate rapidly, no similar
proteins were identified from mitochondria with linear genomes. We
considered the possibility that specific mtTBPs are responsible for the
stability and/or replication of mtDNA termini of C. parapsilosis. By both gel retardation and UV cross-linking
experiments using a 51-base-long oligonucleotide derived from the 5
overhang of the extreme end of linear mtDNA we detected a specific
mtTBP in the salt extracts from mitochondrial membranes of this yeast. The interaction of mtTBP with the membrane was relatively strong since
lower salt concentrations (<2 M NaCl) or nonionic
detergents (1% Triton X-100) did not quantitatively extract mtTBP from
the membrane environment. It is possible that the high salt is needed to remove mtTBP from endogenous DNA. However, treatment of the membranes with nucleases did not result in the extraction of mtTBP (data not shown). The mtTBP of C. parapsilosis may at least
partly mediate the association of mtDNA with the inner mitochondrial membrane, which might be necessary for its replication and proper segregation into newly formed mitochondria. Association of mtDNA with
the mitochondrial membrane was demonstrated for Physarum polycephalum and S. cerevisiae, and several protein
components responsible for this interaction were detected (55-57).
One of the interesting characteristics of mtTBP was its resistance to
protease and heat treatments. This resistance was higher in a crude
fraction than in purified form. The crude fraction did not contain
general protease inhibitors since more than 90% of other proteins were
degraded by the two proteases tested (data not shown). It is possible
that mtTBP in a crude fraction is in association with a partner that
protects it from the action of proteases or which in a diluted state
mtTBP is more susceptible to proteolytic cleavage. Moreover, mtTBP
activity was not affected by pretreatment of the sample with RNase A
(data not shown), excluding the possibility that mtTBP is a
ribonucleoprotein complex similar to nuclear telomerase (58). The
resistance of mtTBP to heat (<7 min at 62 °C) was helpful in
removal of contaminating Mg2+-dependent
nucleases. This allowed us to test the sensitivity of the mtTBP·TEL51
complex to various DNA-modifying enzymes, some of which require
magnesium ions (DNase I, exonuclease III). None of the used enzymes
modified DNA in the complex. Surprisingly, free TEL51 was sensitive to
exonuclease III (Fig. 4B, 16th lane) as this
enzyme is specific for double-stranded DNA molecules with protruding 5
ends. This suggests that TEL51 may adopt a double-stranded structure
under the reaction conditions employed in this experiment. The
relevance of such structures for in vivo function(s) of
mitochondrial telomere remains to be determined. Since the probe
was labeled terminally with polynucleotide kinase and the label was not
lost by the action of bacterial alkaline phosphatase, mtTBP probably protects the very end of the telomere. Tight interaction between mtTBP
and mitochondrial telomere may also play a role in preventing DNA
polymerase from resuming synthesis of the 3
end. In addition, mtTBP
may be involved in regulating the length of the mitochondrial telomere.
Preliminary experiments demonstrated that the formation of a complex
between mtTBP and a telomeric oligonucleotide is resistant to
relatively high salt concentrations (data not shown). This allowed us
to purify mtTBP to near homogeneity in a single step using DNA affinity
chromatography. Identification of a single protein with a molecular
mass of about 15 kDa in the active fractions was in apparent
contradiction with the data from UV cross-linking experiments, where
the mtTBP·TEL51 complex migrated as two bands with molecular masses
of about 32 and 40 kDa, respectively. Three different chemical
cross-linkers
(3,3-dithiobis(sulfosuccinimidylpropionate), bis(sulfosuccinimidyl)suberate, potassium
permanganate) gave similar results when compared with cross-linking by
UV light (data not shown). Since these two bands were formed also by UV
cross-linking of a mobility shift gel in situ and since the
15-kDa protein was extracted from the retarded band after gel
retardation assay we tested the possibility that native mtTBP is a
homo-oligomer. Data from FPLC gel filtration chromatography supported
this hypothesis. The nature of two differently migrating bands after UV
cross-linking of mtTBP to a probe remains to be determined.
MtTBP identified in this study is the first example of a protein specifically recognizing the terminal structures of linear mtDNA. It remains to be shown that the mtTBP functions effectively in vivo as a termini-binding protein or fulfills some other tasks. To answer this question, the gene encoding mtTBP will be cloned,2 and mutations will be created for functional analysis. Such studies are complicated by the absence of a genetic transformation system and by the asexual mode of reproduction of C. parapsilosis which does not allow the use of powerful molecular and genetic tools available for S. cerevisiae. In spite of the lack of in vivo data the specificity of mtTBP binding to a telomeric sequence in vitro suggests that it plays a specific role in mitochondria. Beside the protective role that is implied by the resistance of the mtTBP-telomere complex to various DNA-modifying enzymes mtTBP may play a part during the replication of mtDNA termini. Because of their unusual structure, mitochondrial telomeres of C. parapsilosis may reveal a novel strategy to solve the end replication problem associated with linear DNA genomes. In addition, mtTBP might be involved in the recombination and segregation of mtDNA molecules in C. parapsilosis, thus serving as a guard of mtDNA integrity of this petite negative yeast. It will be also interesting to find the mtTBP-encoding gene and its homologs in other organisms. At this stage we cannot rule out the possibility that mtTBP is encoded by mtDNA since only one-third of the C. parapsilosis mtDNA sequence is known, and linear cytoplasmic genomes (viruses, plasmids) often code for their terminal proteins (59). All of these questions are addressed in experiments that are currently in progress in our laboratory.
We thank Prof. L. Ková
(Department of Biochemistry, Comenius University, Bratislava) for
continuous support and helpful discussions and comments, Prof. J. Kolarov for reading the manuscript, L'. Adamíková and
Dr. K. Bederková (Department of Biochemistry, Comenius
University) for technical assistance, Dr. E. Kutejová (Institute of Molecular Biology, Bratislava) for help with FLPC, and R. Resnick (Cornell University, Ithaca, NY) for valuable technical suggestions and for reading the manuscript.