ABSTRACT
DNA-protein complexes (nucleoids) are believed to be the segregating unit of
mitochondrial DNA (mtDNA) in
Saccharomyces cerevisiae
. A mitochondrial HMG box protein, Abf2p, is needed for maintenance of mtDNA in cells grown on rich dextrose medium, but is dispensible in glycerol grown cells. As visualized by 4
'
,6
'
-diamino- 2-phenylindole staining, mtDNA nucleoids in mutant cells lacking
Abf2p (
[Delta]
abf2
) are diffuse compared with those in wild-type cells. We have isolated mtDNA nucleoids and characterized two mtDNA-protein complexes, termed NCLDp-2 and NCLDs-2, containing distinct but overlapping sets of
polypeptides. This protocol yields similar nucleoid complexes from the
[Delta]
abf2
mutant, although several proteins appear lacking from NCLDs-2. Segments of mtDNA detected with probes to
COXII
,
VAR1
and ori5 sequences are equally sensitive to DNase I digestion in NCLDs-2 and NCLDp-2 from wild-type cells and from the
[Delta]
abf2
mutant. However,
COXII
and
VAR1
sequences are 4- to 5-fold more sensitive to DNase I digestion of mtDNA in toluene-permeabilized mitochondria from the
[Delta]
abf2
mutant than from wild-type cells, but no difference in DNase I sensitivity was detected with the
ori5 probe. These results provide a first indication that Abf2p influences
differential organization of mtDNA sequences.
In situ
observations of
mitochondrial DNA (mtDNA) in a variety of organisms reveal that mtDNA is highly
organized within the organelle. Electron microscopic analysis of animal mtDNA
often shows clumped or thickened fibers (
1
) in two to six spatially distinct centers, called nucleoids (
2
), by analogy with the organized DNA structures in bacteria (
3
,
4
). Similar observations have been made from serial sections of stationary phase
cells of the yeast
Saccharomyces cerevisiae
: bundles of filaments are distributed among one to 20 discrete areas within
mitochondria (
5
). Examination of intact yeast cells by fluorescence microscopy with the DNA
binding dye 4',6'-diamidino-2-phenylindole (DAPI) indicates that mtDNA exists
as small, discrete spots (called nucleoids or chondriolites) in mitochondria,
usually at the periphery of the cell (
5
,
6
). These spots, each containing an estimated two to eight mtDNA molecules (
5
), are characteristic of the mtDNA staining pattern in vegetative cells. During
meiosis and sporulation the number of spots decreases and mtDNA is reorganized
into a branched network surrounding the nucleus (
5
).
Wild-type haploid yeast strains have been estimated to contain 25-50 mtDNA molecules that are stably maintained during mitotic
growth. In sexual crosses, mtDNA from each parent is faithfully transmitted to
the progeny and segregates rapidly to yield homoplasmic cells (
7
). It is generally believed that the nucleoid is the unit of mtDNA segregation (
8
). Recently a mutant of the
MGT1
gene, which is defective in resolving yeast mtDNA recombination junctions (
10
), was found to have alterations in the number of mtDNA nucleoids and in the
pattern of mtDNA transmission (
9
-
11
).
MtDNA nucleoids, which have been isolated and characterized from several
organisms (
12
,
13
), are isolated from sucrose gradients as rapidly sedimenting complexes. Based
on electron and fluorescence microscopy, these complexes appear to retain their
morphological structure
in vitro
(
12
,
14
). Fluorescence intensity analysis of isolated nucleoids from yeast suggests
that they contain the same number of mtDNA molecules as
in vivo
estimates (
14
). Proteins were shown to be required to maintain the structural integrity of
isolated yeast nucleoids and only a small subset of total mitochondrial protein
is recovered in isolated nucleoids (
14
).
One yeast mtDNA binding protein that functions in mtDNA stability is a 20 kDa
basic protein called Abf2p. This protein, initially called HM, was first purified from isolated mitochondria by DNA-cellulose chromatography (
15
). Abf2p was later identified in yeast extracts based on its interaction with
origins of nuclear DNA replication (
16
). Abf2p is a relatively abundant mitochondrial protein, which, based on DNA
binding assays, could bind to every ~30 bp of mtDNA (
17
). Abf2p is a member of the family of chromosomal non-histone high mobility group (HMG) proteins (
18
) and contains two HMG box domains (
16
). A similar HMG box protein (mtTFA) is also present in animal cell mitochondria
(
19
,
20
). Null mutations of the
ABF2
gene lead to [rho]
+
mtDNA instability in yeast cells grown on dextrose medium, where respiration is
dispensable (
16
,
21
). However, mutant cells lacking Abf2p grow on glycerol medium (where
respiration is obligate), indicating that the protein is not essential for
mtDNA replication, expression or transmission. At least one HMG box domain of
Abf2p is required for maintenance of [rho]
+
mtDNA (
22
).
As a class these proteins can bend and wrap DNA and, given the other known
properties of HMG box proteins (
23
), are likely to function in DNA packaging. DNA packaging is probably an
important function of Abf2p, since the
Escherichia coli
DNA packaging protein HU, which is not an HMG box protein, when expressed and
targeted to mitochondria can partially restore [rho]
+
mtDNA stability in yeast cells lacking Abf2p (
21
). In addition, HU and HMG box proteins are functionally interchangeable for a
variety of DNA transactions (
24
).
As an initial approach to characterizing the role of Abf2p in mtDNA
organization, we have examined mtDNA in a wild-type (
ABF2
) strain and an
abf2
null mutant ([Delta]
abf2
) by microscopic and biochemical approaches. We find that Abf2p affects the
in vivo
morphology of mtDNA nucleoids as seen by DAPI staining. The absence of the
protein in [Delta]
abf2
cells was found to have little effect on the sedimentation properties or
nuclease sensitivity of isolated nucleoids. In contrast, using toluene-permeabilized mitochondria from [Delta]
abf2
cells some mtDNA sequences, but not all, are ~4-fold more sensitive to DNase I compared with the wild-type. These experiments indicate a role for Abf2p in the
differential organization of mtDNA sequences.
The wild-type yeast strain used is 14WW (MAT
a
ade2 trp1 leu2 ura3-52 cit1
::
LEU2
). The
ABf2
gene of strain 14WW was disrupted by transformation with a 2.0 kb
Eco
RI fragment isolated from plasmid pAM1A::TRP1 (
16
). These strains contain [rho]
+
mtDNA from strain D273-10B (
25
). [rho]o derivatives were generated by passage through YPD medium containing
10 [mu]g/ml ethidium bromide. [rho]
+
and [rho]o strains were grown on YPG and YPD respectively, at 30oC as previously described (
26
), except that 2% Bacto-peptone was used.
Cells were grown overnight in YPD or YPG medium and stained with DAPI as
previously described (
26
,
27
).
Mitochondria were isolated from sphereoplasts from 9-18 l YPG cultures (OD
600
0.8-1.1). All procedures were carried out at 4oC except where noted. Cells were harvested, resuspended in ~200 ml 0.1 M Tris-HCl, 0.02 M EDTA, pH 9.3 and shaken with 5 [mu]l/ml [beta]-mercaptoethanol (14.4 M stock) for 15
min at 30oC. The cells were collected, washed once with SCE, pH 5.8 (0.6 M sorbitol, 0.3 M mannitol, 20 mM K
2
HPO
4
, 20 mM citric acid, 1 mM EDTA) and resuspended in 1 ml/g cells SCE, pH 5.8. Yeast lytic enzyme (+70 000 U/g; ICN Biomedicals) was added to 0.5 mg/ml and the cells treated at 30oC for 30 min, mixing gently every 10 min. Spheroplasts were washed twice
with SCE, pH 7.0, containing 1 mM spermidine, 7 mM [beta]-mercaptoethanol and 1 mM phenylmethyl sulfonyl fluoride (PMSF) (~20 ml/wash) and broken by vigorous shaking at 4oC with glass beads in the same buffer. Cellular debris and
unlysed cells were removed by three 5 min centrifugations twice at 1600
g
and once at 2500
g
. Crude mitochondria were collected by centrifugation at 12 000
g
for 20 min. These mitochondria were further purified by flotation (
28
), with the following changes: mitochondria were resuspended in Tricine buffer containing 80% sucrose (w/v) before being floated; after
centrifugation mitochondria at the 45/55% sucrose interface were collected,
diluted with 2 vol. ice-cold Tricine buffer and pelleted as above. These purified mitochondria
were stored frozen at -80oC.
We modified a published isolation procedure for mtDNA nucleoids from yeast. Purified mitochondria were thawed on ice, resuspended in NE2 buffer (0.25 M sucrose, 20 mM Tris-HCl, pH 7.6, 2 mM EDTA, 7 mM [beta]-mercaptoethanol) (
14
) and diluted with an equal volume of 0.5* NE2 buffer to a final concentration of 5-7 mg mitochondrial protein/ml. Spermidine (1.0 M) was added to a
final concentration of 3 mM and mitochondria were lysed by adding 20% NP-40 to a final concentration of 0.5%. After 5 min on ice with gentle mixing
the lysate was fractionated at 12 000
g
for 20 min into supernatant (S) and pellet (P) fractions. The P fraction was
resuspended as above.
Aliquots from both S and P fractions were layered on top of step gradients,
initially comprised of 3.5 ml 20%/2.5 ml 40%/1.8 ml 60%/0.9 ml 70% sucrose, but
later comprised of 4 ml 20%/2 ml 60%/1 ml 80% sucrose in gradient buffer (20 mM
Tris-HCl, pH 7.6, 1 mM EDTA, 1 mM spermidine, 7 mM [beta]-mercaptoethanol, 1 mM PMSF) in SW41 tubes and centrifuged at
111 000
g
(30 000 r.p.m.) for 75 min. Gradients were fractionated and analyzed for
distribution of
mtDNA and protein. MtDNA peaks derived from the S and P fractions from an
initial NP-40 extraction are hereafter referred to as NCLDs-1 and NCLDp-1 respectively. NCLDs-1 and NCLDp-1 nucleoid fractions banding at the 20/60%
interface (see Fig.
1
) were collected, diluted with 2 vol. ice-cold gradient buffer, treated with 0.5% NP-40 and centrifuged through a second (20/40/60/70% sucrose) step
gradient (49 000
g
, 3 h) to yield NCLDs-2 and NCLDp-2.
The position of mtDNA in gradients was determined by dot blots using CsCl-purified [rho]
+
mtDNA as probe. MtDNA in purified and lysed mitochondria was extracted with
PCIA (phenol/chloroform/isoamyl alcohol, 25:24:1), precipitated, resuspended and the concentration
determined using dot blots, with known concentrations of mtDNA, purified in a CsCl-bisbenzimide gradient (
29
) as a standard and quantitated by PhosphorImager analysis. Protein
concentration was determined by the Bradford assay (BioRad), using bovine serum
albumin as a standard. SDS-PAGE was performed according to Laemmli (
30
) using 12.5 or 15% gels or 7.5-17.5% gradient gels containing 29.2:0.8% acrylamide:bis-acrylamide. Some gels were stained with Coomassie blue. Others were
electrotransferred to nitrocellulose and blocked with 5% non-fat milk in TBS (25 mM Tris, pH 7.6, 0.2 M NaCl, 0.02% KCl) for at least 1
h. Polyclonal antisera against porin (provided by G.Schatz) and monoclonal
antibodies against CoxIIIp (Molecular Probes, Eugene, OR) were used at 1:10 000
and 1:1250 dilution respectively. Antisera against Abf2p were prepared in
rabbits using protein derived from the cloned gene in a pT7-7 bacterial expression system and used at a 1:10 000 dilution. Antisera
against MAS70, cytochrome c, malate dehydrogenase and the ATPase complex were provided by Drs L.Pon, F.Sherman, L.Henn and M.Douglas respectively. All blots were incubated at room temperature
in TBS and 5% non-fat milk with primary antisera (overnight) and secondary (goat anti-rabbit or goat anti-mouse) antisera (2-3 h). Blots were washed extensively with TBS after each
incubation, prior to treatment with ECL reagents (Amersham) and exposure to
film for 1-5 s.
The concentrations of DNA in nucleoid fractions were determined by fluorimetry
and by quantitating Southern blots of
mtDNA. Nucleoid fractions and deproteinized mtDNA (PCIA-extracted nucleoid DNA) were treated in 50 mM Tris, pH 7.5, 10 mM MnCl
2
, 20% sucrose at 30oC with 0.012 U/ml DNase I (Promega).
Reactions were terminated by addition on ice of EDTA to 20 mM and treated with
1% SDS and 200 [mu]g/ml proteinase K for at least 2 h at 37oC. Reactions were adjusted to 2.5 M ammonium acetate and 50 [mu]g/ml yeast tRNA prior to PCIA extraction and ethanol precipitation.
Samples were collected, dried, restricted with
Hae
III or
Hin
fI and subjected to electrophoresis through 0.8% agarose gels in TBE (89 mM
Tris, pH 8.3, 89 mM boric acid, 2 mM EDTA). Gels were diffusion blotted to
nylon membranes and hybridized in Rapid-Hyb buffer (Amersham) at 65oC to probes of total [rho]
+
mtDNA or gel-purified restriction fragments [0.26 kb
Pvu
II fragment from pSB20 (
31
) for
COXII
, 1.0 kb
Hpa
II fragment for
VAR1
(
32
), 0.76 kb
Nde
I fragment from the hypersuppressive petite HS40 genome (
33
) for ori5] random primed (Boehringer Mannheim) with [
32
P]dATP. Blots were washed according to the manufacturer's recommendations and
quantified.
Flotation gradient-purified mitochondria (corresponding to 5-8 mg protein) were resuspended in 1 vol. NE2 buffer and 1 vol. 0.5* NE2 buffer to a final volume of ~1.5 ml and spermidine added to 3 mM. Toluene was added
to 1% and mixed gently for 15 min at room temperature. Permeabilized
mitochondria were collected in a microfuge tube, resuspended in NE2 buffer,
treated with 0.2 U/ml DNase I and processed as described above. MtDNA in
mitochondria not treated with toluene was completely resistant to DNase I and
the mtDNA in toluene-treated mitochondria not treated with DNase I remained undegraded after 20
min at 30oC (data not shown).
As an initial test of whether the mtDNA binding protein Abf2p functions in mtDNA
organization, we compared the morphology of mtDNA nucleoids in wild-type and [Delta]
abf2
cells using fluorescence microscopy. This technique has been employed
frequently to determine the presence or absence of mtDNA in mutant yeast
strains (
6
), but has only recently been used to examine the effects of specific mutations
on staining of [rho]
+
mtDNA (
11
). Wild-type (wt,
ABF2
) and mutant ([Delta]
abf2
) [rho]
+
cells were grown on glycerol medium and then stained with DAPI. As shown in
Figure
1
,
ABF2
cells show typical mtDNA nucleoids: brightly staining, punctate cytoplasmic
structures (indicated by arrows) often located around the periphery of the
cell. As a control cytoplasmic staining is absent in a [rho]o petite mutant (which lacks mtDNA) grown on 2% dextrose medium. In
contrast to
ABF2
[rho]
+
cells, nucleoids in [rho]
+
[Delta]
abf2
cells appear diffuse and clearly different from the bright, punctate staining of
wild-type mtDNA nucleoids. When grown on dextrose medium [Delta]
abf2
cells rapidly lose their mtDNA and lack any fluorescent cytoplasmic staining (
16
,
25
). The altered DAPI-staining pattern of [rho]
+
mtDNA nucleoids in [Delta]
abf2
cells suggests that the absence of Abf2p affects the organization of [rho]
+
mtDNA nucleoids.
We next analyzed mtDNA organization
in vitro
by comparing the properties of mtDNA nucleoids isolated from wild-type and [Delta]
abf2
cells. Initial studies of yeast mtDNA nucleoids employed a published
purification procedure (
14
), however, we found that only a minority fraction of mtDNA was recovered from
the lysed organelles and that the nucleoid fractions were contaminated with
mitochondrial membrane proteins (see below). Therefore, we modified that
procedure to improve both the properties and yield of mtDNA nucleoids.
As shown in Table
1
, the NP-40 soluble fraction (S) contains ~30% of the starting mtDNA, while 70% sediments in the insoluble (P)
fraction. The S fraction (as defined here) was used exclusively by Miyakawa
et al.
(
14
) in their characterization of yeast mtDNA nucleoids, thus it appears that a
majority of the starting mtDNA was not examined by them. Figure
2
shows that mtDNA from both the S and P fractions sediments through the sucrose
step gradient, banding in both cases at the 40/60% sucrose interface (corresponding to fractions 10 and 11). These mtDNA peaks derived from
the S and P fractions from an initial NP-40 extraction are hereafter referred to as NCLDs-1 and NCLDp-1 respectively. We also consistently observe that more mtDNA
is recovered in NCLDp-1 and mtDNA and protein sediment together as a much sharper peak than in
NCLDs-1 fractions (Fig.
2
).
To determine if the
in vivo
morphological alterations of mtDNA nucleoid organization in [Delta]
abf2
cells (Fig.
1
) are reflected in the properties of the isolated nucleoids, NCLDs-2 and NCLDp-2 fractions from wild-type and mutant cells were compared. Sedimentation properties
of nucleoids from [Delta]
abf2
cells (data not shown) were indistinguishable from the wild-type (Fig.
2
), indicating that Abf2p is not critical in maintaining the nucleoid as a
rapidly sedimenting mtDNA-protein complex. However, a comparison of the protein profiles of NCLDs-2 and NCLDp-2 from wild-type and [Delta]
abf2
cells reveals some differences in the polypeptides associated with mtDNA
nucleoids (Fig.
4
). As expected, Abf2p is absent from both nucleoid fractions from [Delta]
abf2
cells (Fig.
4
, lanes 2 and 5). In addition, 60 and 46 kDa polypeptides of comparable
abundance to Abf2p are absent or greatly reduced in NCLDs-2 from [Delta]
abf2
cells relative to the wild-type (Fig.
5
, lane 2). Some minor polypeptides are also reduced in abundance in NCLDs-2 fractions from [Delta]
abf2
cells. While these differences were consistently observed in independent NCLDs-2 nucleoid preparations, no reproducible differences between wild-type and mutant were observed in protein profiles of NCLDp-2 fractions.
The sensitivity of protein-DNA complexes to nuclease digestion has proven useful for examining DNA
packaging and organization (
34
). Therefore, to further characterize the organization of mtDNA in nucleoids
isolated from wild-type and [Delta]
abf2
cells we analyzed their sensitivity to DNase I digestion. First we determined
if mtDNA in isolated NCLDs-2 and NCLDp-2 fractions is more nuclease resistant than mtDNA not complexed with
proteins. This analysis should measure the maximum difference that could be
observed between nucleoids isolated from wild-type and [Delta]
abf2
cells. Deproteinized mtDNA, NCLDs-2 and NCLDp-2 were digested with DNase I, treated with SDS and proteinase K,
PCIA extracted and precipitated. Samples, either undigested or digested with
restriction enzymes, were analyzed by Southern hybridization using radiolabeled
mtDNA restriction fragments as probes. A representative Southern blot of
sequences hybridizing to the
COXII
gene (encoding subunit 2 of cytochrome c oxidase) following DNase I and
Hae
III digestion is presented in Figure
5
A and quantified results are summarized in Figure
5
B.
In the representative experiment shown in Figure
5
, the
t
½
of DNase I digestion of
COXII
sequences in deproteinized mtDNA is 2.5 min, compared with
t
½
values of 6 and 9 min for those same sequences in NCLDs-2 and NCLDp-2 respectively. Analysis of three independent preparations of
deproteinized mtDNA (
t
½
2.7 +- 0.3 min), NCLDs-2 (
t
½
5.5 +- 1.3 min) and NCLDp-2 (
t
½
7.0 +- 2.3 min) from the wild-type reveal that
COXII
sequences in both nucleoid fractions are 2- to 3-fold more resistant to DNase I digestion than in deproteinized
mtDNA. Thus this approach measures the protection of mtDNA by one or more of
the proteins present in the isolated nucleoid fractions and represents a level
of organization not found in deproteinized mtDNA.
These data show no statistically significant difference in the DNase I
sensitivity of
COXII
sequences in the nucleoid NCLDs-2 and NCLDp-2 fractions (Fig.
5
A). The DNase I sensitivity of two other mtDNA sequences was also examined:
VAR1
is an A+T-rich gene (unlike
COXII
) that encodes a ribosomal protein (
35
); ori5, which is also very A+T-rich and is a putative origin of mtDNA replication (
36
). These loci were chosen to examine any bias that DNase I might have for A+T-rich sequences and to determine if DNase I sensitivity differs between
gene and non-coding sequences. As with
COXII
, no statistically significant difference was detected in the DNase I
sensitivity (
t
½
)
of the
VAR1
(7.0 +- 0.5 min and 11 +- 1.0 min) and ori5 (6.5 +- 1.3 min and 8.5 +- 4.0 min) sequences in NCLDs-2 and NCLDp-2 fractions respectively (data not
shown). Therefore, no differential nuclease protection was observed among these
sequences between NCLDs-2 and NCLDp-2 fractions.
Next we compared the DNase I sensitivity of nucleoids isolated from [Delta]
abf2
cells with those isolated from wild-type cells, but no significant differences were found. As with wild-type nucleoids, similar DNase I sensitivities (
t
½
) were observed for NCLDs-2 (4.0 +- 0.5 min, 3.8 +- 0.3 min and 4.3 +- 0.8 min) and NCLDp-2 (5.2 +- 2.9 min, 5.0 +- 1.0 min and 5.3 +- 0.6 min) from [Delta]
abf2
cells for all three sequences (
COXII
,
VAR1
and ori5 respectively) analyzed. Also, no statistically significant differences
were detected in the DNase I sensitivity among these three sequences. These
data indicate that despite the presence of high concentrations of Abf2p in wild-type mtDNA nucleoids, this protein is not a major determinant of nuclease
resistance in isolated nucleoids. Therefore, major differences in mtDNA
organization that might exist between wild-type and [Delta]
abf2
cells may not be retained in isolated nucleoids. Alternatively, since
deproteinized mtDNA is at most only 2- to 3-fold more DNase I sensitive than mtDNA nucleoids from wild-type cells (Fig.
5
), the maximum difference in DNase I sensitivity of isolated nucleoids from wild-type and [Delta]
abf2
cells must certainly be less. Therefore, small differences in DNase I
sensitivity of nucleoids isolated from wild-type and [Delta]
abf2
cells may not be detectable by this
in vitro
assay.
In an effort to determine potential differences in mtDNA organization between
wild-type and [Delta]
abf2
cells we next analyzed DNase I sensitivity of mtDNA
in organello
. The use of isolated mitochondria for the analysis of transcription initiation
and replication (
37
,
38
) has proven successful. Gradient-purified mitochondria were permeabilized with toluene, treated with DNase
I and processed as described above. Preliminary experiments used total [rho]
+
mtDNA as a probe to determine the effect of DNase I treatment on mtDNA in
permeabilized mitochondria from wild-type and [Delta]
abf2
cells (Fig.
6
A). Undigested DNA from isolated mitochondria (Fig.
6
A) migrates as a >45 kb species on a 0.8% agarose gel, similar to DNA from
isolated nucleoids (not shown). mtDNA in toluene-permeabilized mitochondria from wild-type cells is clearly sensitive to DNase I, although a 10-fold higher concentration of enzyme is required to achieve a
comparable extent of digestion to that of isolated nucleoids. MtDNA in
mitochondria from [Delta]
abf2
cells appears more sensitive (~4-fold, data not shown) to DNase I than mtDNA from wild-type cells (Fig.
6
A).
We have examined the organization of [rho]
+
mtDNA in a wild-type and an
abf2
null mutant ([Delta]
abf2
) strain of yeast. Our experiments show that in the [Delta]
abf2
strain mtDNA nucleoids, visualized
in situ
by DAPI staining, are morphologically less distinct than nucleoids in wild-type cells: they appear more diffuse and lack the characteristic bright,
punctate staining of wild-type nucleoids (Fig.
1
). These data indicate that Abf2p has a role in mtDNA organization, in addition
to its previously ascribed function in mtDNA stability (
16
).
To compare mtDNA nucleoids from wild-type and [Delta]
abf2
cells
in vitro
we modified the published nucleoid isolation procedure (
14
) in several important ways. First, mitochondrial fractions were prepared by flotation gradient centrifugation: mitochondria generated solely by
differential centrifugation contain ~10 times more nuclear DNA contamination than do flotation-purified mitochondria. Second, we found that the soluble NP-40 mitochondrial extract, used by others to prepare mtDNA
nucleoids (
14
), contains only a fraction of the total mtDNA: we routinely recover ~70% of the starting mtDNA in the NP-40
insoluble
fraction and that material is readily purified as a DNA-protein complex, comparable with the mtDNA nucleoids present in the NP-40 soluble extract. Third, a second NP-40 extraction quantitatively eliminates contamination by inner and outer
mitochondrial membrane proteins, without any significant loss of Abf2p or mtDNA
from the final nucleoid preparations.
At present the biochemical significance of the partitioning of mtDNA nucleoids
between the S and P fractions is not known. NCLDp-2 and NCLDs-2 fractions may reflect mtDNA nucleoids present in the organelle as
membrane-bound and -free protein-DNA complexes respectively. Membrane association of bacterial
nucleoids has been recognized for some time (
3
,
4
). Additionally, studies of HeLa cell mtDNA have suggested a specific membrane
attachment at the D-loop origin of replication and the presence of a unique protein structure bound to that region of the mitochondrial
genome (
39
). While proteins present in the NCLDp-2 and NCLDs-2 fractions are similar, one notable difference is the presence of an
abundant ~110 kDa species in NCLDp-2 that is essentially absent from NCLDs nucleoids (Figs
3
and
4
). It is tempting to speculate that this protein could be a component
originating from membranes and be involved in nucleoid membrane attachment.
Studies are currently under way to characterize this 110 kDa species, as well
as other proteins in these fractions. To date few proteins associated with
mtDNA nucleoids have been identified. In the present work Abf2p has been shown
to be present in both NCLDs-2 and NCLDp-2 nucleoid preparations. Miyakawa
et al.
(
40
) identified a 48 kDa polypeptide that is loosely associated with yeast
mitochondrial nucleoids and a series of small (15-21 kDa) proteins that were chemically cross-linked to kinetoplast DNA has been described (
41
), however, in neither case has the function of these proteins been determined.
Given the difference in morphology between DAPI stained mtDNA nucleoids in wild-type and
[Delta]
abf2
cells and the known function of HMG box proteins in DNA organization, we
attempted to detect biochemical differences between mtDNA nucleoids of wild-type and
[Delta]
abf2
cells. The partitioning of mtDNA between the S and P fractions and the sedimentation profile difference of
NCLDs-2 and NCLDp-2 fractions that we observed for mtDNA nucleoids isolated from wild-type were similar to those observed for mtDNA nucleoids
isolated from [Delta]
abf2
cells. Therefore, Abf2p cannot be a major protein needed for maintaining the
presence of fast sedimenting mtDNA-protein complexes in yeast. However, we found several polypeptides in
NCLDs-2 fractions of the wild-type strain that were significantly depleted or absent from this
fraction from [Delta]
abf2
cells (Fig.
5
). This observation is consistent with studies in both prokaryotes (
42
-
44
) and eukaryotes (
45
,
46
), indicating that some DNA packaging proteins influence the DNA binding of
other proteins.
We have measured DNase I sensitivity of mtDNA nucleoids both
in vitro
(Fig.
5
) and
in organello
(Fig.
6
). Clearly, the presence of proteins associated with mtDNA in both the NCLDs-2 and NCLDp-2 nucleoid fractions contributes to their 2- to 3-fold increased DNase I resistance compared with
deproteinized mtDNA (Fig.
5
). However, Abf2p alone cannot be responsible for this increased resistance,
since purified nucleoids from [Delta]
abf2
cells are still 2- to 3-fold more resistant than deproteinized mtDNA. No statistically
significant difference was observed in the nuclease sensitivity between NCLDs-2 and NCLDp-2 or between nucleoids from wild-type and [Delta]
abf2
cells.
While these nuclease digestion experiments of mtDNA in nucleoid fractions did
not provide a correlation with the morphological differences between nucleoids
observed in wild-type and [Delta]
abf2
cells, the
in organello
DNase I digestion experiments revealed possible organizational differences. In
this case both
COXII
and
VAR1
sequences were 4- to 5-fold more sensitive to DNase I digestion in mitochondria from [Delta]
abf2
cells than from the wild-type. It is quite conceivable, therefore, that Abf2p-dependent features of mtDNA nucleoid organization are lost upon
extraction and isolation of these structures. However, we believe that differences in
nuclease sensitivity
in organello
between the wild-type and mutant is more representative of the actual [Delta]
abf2
phenotype than the
in vitro
digestion pattern. While purified nucleoids will be valuable in characterizing
specific proteins associated with mtDNA, the
in organello
analysis opens up additional possibilities for the experimental evaluation of
the organization of mtDNA. Recent studies (
37
) indicate differences in the
in vitro
and
in organello
binding properties of the human Abf2p homolog, mtTFA, to mtDNA.
The finding that the
in organello
DNase I sensitivity of ori5 sequences is the same in wild-type and in [Delta]
abf2
cells suggests fundamental differences in the organization of those sequences
relative to the
COXII
and
VAR1
genes.
In vitro
Abf2p binding to ori sequences is phased, binding every 25-30 bp and separated by short stretches of poly(dA) or poly(dT) (
17
). Although both ori5 (
36
,
47
) and
VAR1
(
35
) are ~90% A+T sequences, the increased density of poly(dA) or poly(dT) sequences
surrounding ori5, relative to the
VAR1
gene, could result in less efficient binding of Abf2p to ori5 than to
VAR1
. However, our data indicate that Abf2p is not a factor in conferring DNase I
resistance to ori5. Therefore, the binding of other proteins to ori5 must
confer the observed DNase I resistance. For example, origins of DNA replication in bacteria (
48
), nuclei (
49
) and mitochondria (
39
,
50
) are probably compartmentalized to membrane- or scaffold-associated regions and hence subject to constraints that may differ
from bulk DNA. We anticipate that future experiments comparing the properties
of mtDNA nucleoids
in vitro
and
in organello
will be fruitful in defining the functional organization of the mitochondrial
genome.
This work was supported by grant GM33510 from the NIH and grant I-0642 from The Robert A.Welch Foundation. We thank Betty Key for expert
technical assistance and Beverly Rothermel for preparation of the Abf2p
antibody. We also thank our colleagues for generous gifts of antisera and
D.Clayton for the Abf2p expression vector.
REFERENCES
Return