| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, April 4,
1994; and in revised form, September 13, 1994) From the
According to current concepts, new peroxisomes are formed by
division of pre-existing peroxisomes or by budding from a peroxisomal
reticulum. Recent cytochemical and biochemical data indicate that
protein content in peroxisomes are heterogenous and that import of
newly synthesized proteins may be restricted to certain protein
import-competent peroxisomal subcompartments (Yamamoto, K., and Fahimi,
H. D.(1987) J. Cell Biol. 105, 713-722; Heinemann, P.,
and Just, W. W.(1992) FEBS Lett. 300, 179-182;
Lüers, G., Hashimoto, T., Fahimi, H. D., and
Völkl, A.(1993) J. Cell Biol. 121,
1271-1280). We have observed that substantial amounts of
peroxisomal proteins are found together with ``microsomes''
(100,000
Peroxisomes are nearly ubiquitous organelles, generally
spherical in shape with a finely granular matrix surrounded by a single
membrane. The abundance, size, and appearance in electron microscopy of
peroxisomes vary considerably. Peroxisomes contain enzymes which
produce and degrade hydrogen peroxide, and it is clear that fatty acid
degradation is a common function of peroxisomes in eukaryotic cells
(for reviews, see (1) and (2) ). Earlier studies on
the incorporation and turnover of peroxisomal proteins suggested that
peroxisomes form a homogenous population without any signs of
maturation and with an apparently random degradation of the
organelle(3, 4) . Subsequent studies showed that
peroxisomal proteins are synthesized on unbound ribosomes and appear in
the cytosol before import into pre-existing
organelles(4, 5, 6, 7) . These data,
together with cytochemical studies(8) , showed that the
biogenesis of peroxisomes is independent of ER. ( Recent
studies on inherited peroxisomal disorders and peroxisome assembly
mutants have added much new information on the structure, function, and
biogenesis of peroxisomes. The first ultrastructural observations on
Zellweger syndrome indicated that the defect may be due to a total lack
of peroxisomes(20) . However, subsequent studies utilizing
immunocytochemistry demonstrated that peroxisomal membranes are present
in cells of Zellweger patients(21, 22) , and
subcellular fractionations of fibroblasts from Zellweger patients
demonstrated that peroxisomal proteins are particulate to varying
extents(23, 24) . Some fibroblast cell lines from
Zellweger patients were able to import thiolase into peroxisomes (25) but failed to import proteins containing the
carboxyl-terminal-SKL targeting signal(26) . Thus, studies on
Zellweger patients have provided evidence of at least two, possibly
three, distinct pathways for import of peroxisomal matrix proteins.
Immunocytochemical studies on the heterogeneity of peroxisomes, and the
characterization of peroxisome assembly mutants from various yeasts
have shed light on both functional and developmental aspects of
peroxisomes(27, 28, 29, 30, 31) .
Taken together these observations now call for a revision of present
peroxisome biogenesis models. A review presenting current concepts on
the ultrastructural basis of the biogenesis of peroxisomes was recently
published(32) . In our earlier unpublished experiments on
subcellular fractionations of liver homogenates from
di(2-ethylhexyl)phthalate (DEHP)-treated rats, we found substantial
amounts of peroxisomal proteins recovered in the microsomal fraction.
In the present study we have further fractionated microsomal fractions
by modified gradient centrifugations in Nycodenz. The enzyme activity
distributions of the gradients were analyzed by Western blotting and
fractions enriched in peroxisomal proteins were analyzed by
immunoelectron microscopy. We conclude that with the present protocol
we can separate several classes of ``peroxisomes'' from
livers of DEHP-treated rats (that can be distinguished by different
protein contents and sedimentation properties which are interpreted to
represent peroxisomal subcompartments, probably formed during
homogenization of the tissue), rather than populations of peroxisomes.
These subcompartments are discussed in relation to peroxisome structure
and biogenesis.
In some experiments linear Nycodenz gradients
with density range of 1.15-1.27 g/ml were prepared and
centrifuged for 1, 20, and 29-32 h at 60 000
Antibodies against peroxisomal acyl-CoA oxidase
(anti-Aox) and thiolase (anti-thiolase) were kindly provided by Dr. W.
W. Just. Affinity-purified antibodies against peroxisomal MFE
(anti-MFE) were kindly provided by Dr. J. K. Hiltunen. Catalase was
purified from livers of DEHP-treated rats. Briefly, after
solubilization and centrifugation (at 200 000
Figure 1:
Subcellular fractionation of a LM
fraction in Nycodenz. A LM fraction was prepared from livers of
DEHP-treated rats and subsequently layered on top of a linear Nycodenz
density gradient ranging from 25 to 50% Nycodenz. The gradient was
centrifuged at 60,000
Figure 2:
Immunoelectron microscopy on low density
peroxisomes isolated by centrifugation of a LM fraction in Nycodenz.
The low density fraction containing highest Aox activity was prepared
for immunoelectron microscopy as described under ``Methods.''
Double labeling experiments were carried out with anti-catalase (5-nm
gold particles) and anti-MFE (10-nm gold particles, a and b), anti-Aox (10-nm gold particles, c and d), and anti-PMP70 (10-nm gold particles, e-g). Bar = 0.2 µm.
Figure 3:
Subcellular fractionation of a microsomal
fraction in Nycodenz. Microsomal fractions were prepared from livers of
rats treated with DEHP, layered on top of linear Nycodenz gradients,
and centrifuged at 60,000
Western blot analysis of the same gradient
fractions (Fig. 3, lower panel) verified the enzymatic
data for catalase and 3-hydroxyacyl-CoA dehydrogenase (MFE). In
addition, Western blotting showed that peroxisomal thiolase distributed
similarly to catalase and that substantial amounts of PMP70 were found
in fractions 15-21, indicating the presence of peroxisomal
membranes in these fractions. The very close association of most of the
Aox and MFE found in the microsomal fraction with microsomes after
gradient centrifugation demonstrates that these proteins are not
associated with peroxisomes of normal size and/or density. From these
experiments it was not possible to rule out the possibility that Aox
and MFE were sticking to microsomes. However, in parallel experiments
where microsomal fractions were centrifuged in Percoll gradients, a
clear separation of Aox activity from microsomes was obtained (data not
shown). The results shown are typical of more than 15 fractionation
experiments on enzyme activity distributions and at least 5 experiments
by Western blot analysis.
Figure 4:
Immunoelectron microscopy on high density
peroxisomes isolated by centrifugation of microsomal fractions in
Nycodenz. Gradient fractions corresponding to the high density peak of
peroxisomes obtained after centrifugation of microsomal fractions in
Nycodenz were prepared and incubated with antibodies and colloidal gold
as described under ``Methods.'' a, incubated with
anti-catalase; b, incubated with anti-peroxisomal MFE; c, incubated with anti-PMP70. The peroxisomes found at high
densities were generally 0.1-0.5 µm in diameter. Bar = 0.2 µm.
Figure 5:
Immunoelectron microscopy on low density
gradient fractions after centrifugation of microsomal fractions in
Nycodenz. Microsomal fractions were centrifuged in Nycodenz gradients,
and fractions corresponding to the low density peak of Aox (also
containing catalase) were prepared and incubated with antibodies and
colloidal gold as described under ``Methods.'' Labeled for: a and b, catalase (10 nm gold particles) and MFE (5
nm gold particles), showing a MFE-negative vesicle labeled for catalase (a) and a double-labeled vesicle positive for MFE (arrows,
b); c and d, MFE (10-nm gold particles) and
thiolase (5-nm gold particles). MFE-positive vesicles lacking thiolase (c) and vesicles labeled for both proteins (thiolase at arrows, d); e and f, catalase (10-nm gold
particles) and PMP70 (5-nm gold particles), catalase-positive particles
also labeled for PMP70 (arrows, e) and membrane material
labeled for PMP70 apparently lacking catalase (f); g and h, catalase (10-nm gold particles) and thiolase (5-nm
gold particles), catalase-positive particles lacking thiolase (g) and vesicles strongly labeled for thiolase (arrows,
h) containing catalase. The sizes of the labeled vesicles are
generally below 200 nm. Bar = 0.1
µm.
The immunolabeling
experiments on the low density fractions clearly demonstrated that the
peroxisomal proteins found at low density during gradient
centrifugation are at least partially confined to membrane surrounded
vesicles, rather than becoming released proteins in soluble form or
sticking to other membranes such as microsomes. The double labeling
experiments also indicated the existence of profound heterogeneities in
the contents of peroxisomal enzymes in these vesicles.
Figure 6:
Comparison of the sedimentation of
catalase and Aox after centrifugation of microsomal fractions for 1 and
20 h in Nycodenz gradients. Liver microsomal fractions were prepared
from DEHP-treated rats and centrifuged in linear Nycodenz gradients as
described under ``Methods.'' The microsomal fraction was
divided into three aliquots which were layered on top of Nycodenz
gradients and centrifuged for 1, 20, and 29 h (not shown),
respectively. The gradients were fractionated from the bottom (left to right) and analyzed for catalase and
Aox activities, protein, and density. a, centrifugation for 1
h; b, centrifugation for 20 h.
Aox was also distributed into two
peaks after 1-h centrifugation, one peak at about 1.22 g/ml and the
other peak at about 1.15 g/ml, clearly different from catalase. After
20 and 29 h of centrifugation, Aox still tended to band in two peaks,
one at about 1.22 g/ml and the other still at about 1.15 g/ml. However,
it is evident that a larger part of the Aox activity also sedimented to
a density of 1.25 g/ml or more, clearly higher than catalase. These
experiments imply that there exists differences also in the equilibrium
densities among the vesicles containing catalase and Aox. Isolation and characterization of peroxisomes have been
hampered by the facts that peroxisomes have densities in gradient media
similar to, or only slightly different from, the densities of other
organelles and because peroxisomal protein normally constitutes only 2%
(in rat liver) or less of the cellular protein. It is also well
established that peroxisomes are leaky and that the leakiness is
selective. Catalase and thiolase are very easily released, whereas
other proteins, such as Aox and the MFE, are much less prone to be
released(42) . Other peroxisomal proteins may be expected to
cover the range. This leakiness appears to be time-dependent and
dependent on the treatment of peroxisomes. It has also been implied
that peroxisomes are very sensitive to hydrostatic pressure, since
enzyme distributions in gradients often show activity of peroxisomal
proteins throughout the gradients. However, in the light of the results
presented here, the peroxisomal proteins found in the low density
fractions are at least partially present in vesicles.
The most likely interpretation of our
results is that the different peroxisomal subcompartments described
here are formed during homogenization of a peroxisome reticulum. Fig. 7illustrates how the different peroxisomal populations,
found during subcellular fractionation of rat liver homogenates, may
form by vesiculation. Besides normal peroxisomes (1), the
peroxisomes found at a density of about 1.19 g/ml (corresponding to the
high density peaks after centrifugation of microsomes) may be formed by
the pinching off of buds that are not yet complete (2). The
very small vesicles with heterogenous protein contents may be formed by
vesiculation of the ``stalks'' connecting the buds to the
``body'' of the reticulum (3, 4, 5). The
heterogeneity of these vesicles could be explained by a nonrandom
distribution of the receptors involved in binding and import of
peroxisomal proteins. Such a heterogeneity could also explain why we
are able to isolate different populations of peroxisomes enriched in
proteins containing different PTS.
Figure 7:
Hypothetical model for the formation of
various populations of peroxisomes during subcellular fractionation of
rat liver. If peroxisomes are formed by budding from a peroxisomal
reticulum, fragmentation of such a reticulum by homogenization could
explain the observed heterogeneity of vesicles containing peroxisomal
proteins. Normal peroxisomes (1) are formed by budding and
correspond to the bulk of peroxisomes found at densities around 1.24
g/ml. Fragmentation of ``buds'' (at the stalks) may thus
correspond to ``immature'' peroxisomes with a complete set of
peroxisomal proteins sedimenting to about 1.19 g/ml (2).
Fragmentation of the stalks can be expected to result in the formation
of ``microperoxisomes'' (3, Aox- and MFE-enriched
vesicles; 4, catalase-enriched vesicles; and 5 thiolase-enriched vesicles). The heterogeneity could be explained
by uneven distribution of receptors for import of peroxisomal proteins.
It is possible that the ``body'' of the reticulum forms
another population (?) of vesicles that is not yet identified
due to lack of an appropriate marker.
Recent information concerning
peroxisome structure and protein content in Zellweger syndrome has
demonstrated that peroxisomal particles may have a selective set of
proteins. The presence of peroxisomal ghosts containing peroxisomal
membrane proteins(21, 22) , and the finding of a low
density particle (w-particle) in fibroblasts of Zellweger patients
which contained catalase(24) , also demonstrated a
heterogeneity in peroxisomal protein content. Suzuki et al.(23) showed by indirect immunofluorescence staining and
subcellular fractionation that some catalase and Aox was associated
with organelles in some fibroblasts from Zellweger patients.
Interestingly, substantial amounts of 3-ketoacyl-CoA thiolase and PMP70
were detected in organelles in all cells. These data suggest that the
transport and processing of peroxisomal proteins carrying different
PTS's are different, which is compatible with the present
knowledge on sequence requirements for import of these proteins into
peroxisomes. Several peroxisomal proteins contain a conserved
COOH-terminal -Ser-Lys-Leu sequence, or acceptable variations of this
sequence, that directs proteins to microbodies in mammalian cells,
yeasts, plants, insects, and
trypanosomes(54, 55, 56, 57, 58, 59, 60, 61) .
Import of proteins containing this conserved tripeptide is probably
mediated by a common receptor. However, import of peroxisomal thiolase
appears to be dependent on a cleavable NH Fahimi et al.(32) suggested that synthesis and incorporation
of membranous material, containing PMP70, is an early event, resulting
in the growth of the membrane followed by import of newly synthesized
proteins. Our results are compatible with such a model favoring
synthesis of membrane material as an early event which is followed by
incorporation of highly expressed
Strong support for this model is also
obtained from studies in yeasts where a functional heterogeneity among
microbodies has been suggested to exist(27, 28) .
These studies demonstrated that addition of methanol to Candida
boidinii cells grown on oleic acid resulted in incorporation of
newly synthesized proteins into smaller, presumably
``immature'' peroxisomes. This heterogeneity was only
transient, and continuous cultures grown on a mixture of oleic acid and
methanol at steady-state conditions showed that both the enzymes of the
The formation of
a peroxisomal reticulum distinguishes peroxisomes from the other
intracellular membrane systems of mitochondria, chloroplasts, and
lysosomes. Mitochondria are generally believed to represent an ancient
endosymbiont, probably after ``invasion'' of a bacteria. ER
and lysosomal proteins are synthesized by co-translational insertion of
the proteins into ER followed by glycosylation. ER resident proteins
are apparently retrieved in the ER by a receptor-mediated process
involving a COOH-terminal tetrapeptide (KDEL or HXEL) (see (68) for review). Lysosomal proteins are transported to the
Golgi apparatus where the lysosomes are formed by budding of vesicles
containing lysosomal proteins which are then transported to the acidic
compartment. Peroxisomal proteins may be transported from the site of
import (at the peroxisome-forming sheet) by ``bulk flow''
(similar to protein transport in the rough ER) to the buds. Experiments
aimed at functionally characterizing the different peroxisomal
subcompartments, which are under way in our laboratory, may clarify the
origin and functions of these vesicles. It still remains to
biochemically identify the peroxisome-forming sheet during subcellular
fractionation of rat liver homogenates.
Volume 270,
Number 12,
Issue of March 24, 1995 pp. 6949-6958
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
IMPLICATIONS FOR PEROXISOME STRUCTURE AND BIOGENESIS (*)
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
g pellet) after subcellular fractionation of
rat liver homogenates. In this study we have investigated the origin of
these peroxisomal proteins by modified gradient centrifugation
procedures in Nycodenz and by analysis of enzyme activity
distributions, Western blotting, and immunoelectron microscopy. It is
concluded that much of this material is confined to novel populations
of ``peroxisomes.'' Immunocytochemistry on gradient fractions
showed that some vesicles were enriched in acyl-CoA oxidase and
peroxisomal multifunctional enzyme (``catalase-negative'')
whereas others were enriched in catalase and thiolase (``acyl-CoA
oxidase-negative''). Double immunolabeling experiments verified
the strong heterogeneity in the protein contents of these vesicles and
also identified peroxisomes varying in size from about 0.5 µm
(``normal peroxisomes'') to extremely small vesicles of less
than 100 nm in diameter. The possibility that these vesicles may be
related to different subcompartments of a larger peroxisomal structure
involved in protein import and biogenesis will be discussed.
)The
current view implies that peroxisomes are formed by division of
preexisting organelles or by budding from a peroxisomal
reticulum(9, 10) . The finding that proliferation of
peroxisomes results in a heterogenous induction of peroxisomal proteins (11) may offer a useful tool in the exploration of the
mechanisms of peroxisome biogenesis. Several reports have now
demonstrated profound heterogeneity under proliferative conditions. The
first biochemical indications were obtained by analytical differential
centrifugation demonstrating that clofibrate treatment induced a
polydispersity of peroxisomes in rat liver(12) . Similar
polydispersities in rat liver occur after thyroxine
treatment(13) , ischemia reperfusion(14) , and cold
exposure(15) . A catalase-negative subpopulation of
peroxisomes, which was induced by clofibrate treatment, was identified
in mouse liver by subcellular fractionation (16) . In addition,
cytochemical studies have indicated heterogenous labeling of
peroxisomal proteins(17, 18, 19) .
Materials
DEHP was obtained from Aldrich (Steinheim, Germany). Fatty
acyl-CoA esters, o-nitrophenyl acetate, cytochrome c (type II from horse heart), and horseradish peroxidase were all
obtained from Sigma. Nycodenz was purchased from Nycomed AS (Oslo,
Norway). ECL chemiluminescence detection kit and nitrocellulose
membranes were from Amersham (Buckinghamshire, United Kingdom), and
Percoll was obtained from Pharmacia (Uppsala, Sweden).Methods
Animals
Male Sprague-Dawley rats were obtained
from Eklunds (Stockholm, Sweden). Rats were fed ordinary laboratory
chow (R3, Ewos, Södertälje,
Sweden) or chow supplemented with 2% (w/w) of DEHP for at least 10 days
in order to induce peroxisomes. After fasting overnight, the animals
were sacrificed by CO
anesthesia followed by decapitation.Subcellular Fractionations
The livers were
homogenized with four up and down strokes in a Potter-Elvehjem type
homogenizer in ice-cold 0.25 M sucrose, containing 10 mM Tris-HCl (pH 7.4), 1 mM EDTA (pH 7.4), and 0.1% ethanol.
The homogenates were diluted to 20% (w/v) and centrifuged twice for 10
min at 750 g to obtain the nuclear fraction
(N-fraction). A heavy mitochondrial fraction (HM fraction) was prepared
by centrifugation of the combined supernatants at 3,670 g for 10 min, followed by re-centrifugation of the resuspended
pellet. The peroxisome-enriched light-mitochondrial fraction (LM
fraction) was prepared by centrifugation at 24,500 g for 20 min. The pellet was resuspended and centrifuged again. The
microsomal fraction (P-fraction) was prepared by centrifugation of the
combined supernatants at 48,400 g for 2 h.Gradient Centrifugations
Linear Nycodenz gradients
were prepared with density ranges of 1.09-1.19 g/ml for
fractionations of P-fractions and 1.15-1.25 g/ml for
fractionation of LM fractions. Nycodenz solutions were prepared by
solving Nycodenz in 250 mM sucrose containing 10 mM Tris, pH 7.4, 1 mM EDTA, and 0.1% ethanol. The total
volume of the gradients was 23 ml resting on 8-ml cushions consisting
of 60% Nycodenz. The total volume of the applied sample and overlay
(consisting of homogenization buffer) was 8 ml, making up to a total
volume of 39 ml. Gradient centrifugations of microsomes were performed
at 60,000 g for 60 min, and LM fractions were
centrifuged for 35 min at the same g-force in a Beckman VTi 50
rotor with slow acceleration and deceleration. The density of the
gradient fractions was calculated from the measured refractive index of
gradient fractions. g to compare the enzyme distributions after rate sedimentation and
equilibrium density centrifugations.Enzyme Assays
Acyl-CoA oxidase(33) ,
catalase(34) , esterase(35) , protein(36) ,
cytochrome c oxidase(37) , 3-hydroxyacyl-CoA
dehydrogenase(33) , and acid phosphatase (38) were
determined as described.Gel Electrophoresis and Western Blotting
Sodium
dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was
carried out with 10% acrylamide gels and stained with Coomassie
Brilliant Blue or silver. For Western blotting, proteins were separated
by SDS-PAGE and transferred to nitrocellulose membranes in a semi-dry
blotter for 1 h at 150 mA. The membranes were blocked overnight in TST
buffer (50 mM Tris base, 150 mM NaCl, and 0.5% Tween
20) containing 1% bovine serum albumin, followed by incubation of
primary antibodies diluted in TST buffer containing 0.1% bovine serum
albumin. After washing of the membranes in TST, the membranes were
incubated with peroxidase-conjugated goat anti-rabbit IgG and
visualized by a chemiluminescence system.Preparation of Antibodies
Antibodies against PMP70
were prepared by immunization of rabbits with a synthetic peptide
corresponding to a predicted cytosolic domain (amino acids
403-417) of the sequence reported by Kamijo et
al.(1990). The peptide was synthesized with an extra
NH-terminal cysteine which was used for the coupling of the
peptide to activated keyhole limpet hemocyanin. The preparation and
characterization of this antibody is described elsewhere. () g for 45
min in a Beckman TL-100 table top ultracentrifuge) of isolated
peroxisomes, the matrix fraction was applied to chromatography on a
DEAE column. The bound activity was eluted with a linear NaCl gradient,
and the fractions containing activity were pooled and precipitated by
35% (NH
)SO. The pellet was
dissolved in 50 mM Tris buffer containing 200 mM NaCl
and subjected to size exclusion chromatography in Sephacryl S-300
(Pharmacia) at a flow rate of 0.8 ml/min. The fractions containing
activity were precipitated with 65%
(NH)SO and applied to a MEMSEP 1010
(DEAE cartridge, Millipore). Catalase activity was eluted with a linear
NaCl gradient (0-0.5 M) in 50 mM Tris-HCl (pH
7.4). One rabbit (of the Loop strain) was immunized intramuscularly
with 330 µg of purified catalase emulsified with Freund's
complete adjuvant, followed by three booster injections of 165 µg
of catalase protein emulsified with Freund's incomplete adjuvant.
The rabbit was bled from the ear vein at 2-week intervals after the
third booster injection, and sera were prepared. This antiserum was
monospecific for catalase at dilutions up to 1:500 000 on Western blot. (
)Immunocytochemistry
Aliquots of
gradient fractions were immediately embedded in an equal volume of 8%
luke warm gelatin (Merck, Darmstadt, Germany) and allowed to congeal.
The samples were fixed in 1% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4), overnight at 4 °C. Small pieces
were infiltrated with 2.0 M sucrose containing 15%
polyvinylpyrrolidone and frozen in liquid nitrogen. Sectioning was
performed according to (39) at -110 °C. Sections were
placed on carbon-reinforced formvar-coated 50-mesh grids and placed
directly on drops of 2% bovine serum albumin (Fraction V, Sigma) in 0.1 M phosphate buffer (pH 7.4), containing 20 mM glycine
for 30 min. Subsequently, the sections were incubated with the primary
antibody for 2 h in a humidified chamber at room temperature. All
antibodies were diluted in 0.1 M phosphate buffer (pH 7.4),
containing 0.1% bovine serum albumin (PBB). The sections were
thoroughly washed in PBB, and bound antibodies were detected with
protein A coated with 10-nm gold (Janssens, Olen, Belgium, diluted
1:100). In the double labeling procedure, the protein A was added to a
final concentration of 20 µg/ml for at least 5 min, after which the
sections were rinsed in PBB and incubated with the second antibodies.
Sections were then washed, contrasted by uranyl acetate-oxalate for 5
min, and embedded in 2% methyl cellulose containing 0.3% uranyl
acetate. The grids were then examined in a Philips 420 electron
microscope at 80 kV. Specificity of labeling was checked with normal
peroxisomes prepared by centrifugation of LM fractions (from
DEHP-treated rats) in Nycodenz gradient fractions and with pre-immune
rabbit antiserum.
Subcellular Fractionation
After fractionation of
rat liver homogenates into nuclear, HM, LM, microsomal, and supernatant
fractions, it was noted that substantial amounts of the peroxisomal
enzymes catalase and Aox were present both in the LM fractions and
microsomal fractions. After DEHP treatment, almost 10% of the recovered
Aox activity was found in the microsomal fractions, and about 17% of
the activity was found in the LM fractions. Since DEHP treatment causes
a 10-20-fold induction of peroxisomal 
-oxidation
enzymes(40) , the amount of peroxisomal Aox and catalase
sedimenting with microsomes must be considered to be substantial. In
order to study eventual heterogeneity of peroxisomes, LM fractions and
microsomal fractions were further fractionated by gradient
centrifugation, and enzyme distributions were compared by means of
enzyme activity determinations, Western blot analysis and
immunoelectron microscopy.Fractionation of LM Fractions by Nycodenz Gradient
Centrifugation
LM fractions prepared from livers of DEHP-treated
rats were centrifuged in Nycodenz gradients according to standard
procedures (see ``Methods''). Electron microscopy on the peak
peroxisomal fraction showed a mixture of peroxisomes, some containing
crystalloid cores of urate oxidase (not shown). The distributions of
catalase and Aox in the gradients were as expected with most of the
activities found at high density (fractions 3 and 4) with small
additional peaks near the top of the gradient (Fig. 1, upper
panel). However, it was apparent that the small peaks of catalase
and Aox in the low density region were displaced. Catalase activity
peaked in fractions 19 and 20 and Aox activity peaked in fractions 17
and 18. These data were confirmed by Western blot analysis of the same
fractions (Fig. 1, lower panel) which also showed that
the distribution of peroxisomal thiolase closely resembled the
distribution of catalase. Western blot analysis with anti-PMP70 showed
signal in both the high and low density fractions, indicating that
peroxisomal membranes are present also in the low density fractions of
the gradient.
g for 35 min, and the gradient
was fractionated from the bottom (left to right). The gradient was analyzed for catalase and Aox
activities, and aliquots of the fractions were analyzed by Western
blotting after SDS-PAGE. The blots were probed with primary antibodies
against catalase, Aox, PMP70, and peroxisomal
thiolase.
Immunoelectron Microscopy of Peroxisomal Peak Fractions
after Nycodenz Gradient Centrifugation
Immunocytochemistry on
the high density gradient fractions showed intense labeling with all
antibodies tested (anti-MFE, anti-thiolase, anti-catalase, and
anti-PMP70) (data not shown). Additionally, the low density fraction
containing highest Aox activity was analyzed by immunoelectron
microscopy (Fig. 2). The immunogold labeling was highly specific
for vesicular structures. The antibody labeling was intense in the
labeled vesicles, whereas other vesicles were devoid of labeling, as
could be expected from the presence of microsomes in this region of the
gradient. Double labeling experiments showed profound heterogeneity in
the ratios of the different peroxisomal proteins. Sections incubated
with anti-MFE and anti-catalase showed varying ratios of the number of
gold particles between different structures. Some were strongly
positive for catalase with relatively weak labeling for MFE (Fig. 2a). Other vesicles were strongly labeled for MFE
and weakly labeled for catalase (Fig. 2b). Similar data
were obtained in double labeling experiments with anti-Aox and
anti-catalase (Fig. 2. c and d) in which Aox
labeling resembled labeling for MFE. Double labeling experiments with
anti-PMP70, compared with anti-catalase and anti-Aox, verified that the
immunolabelings were confined to vesicles which were labeled for the
peroxisome-specific membrane protein PMP70 (Fig. 2, e-g). In some cases we observed structures which were
very strongly labeled for PMP70, although the morphology of these
structures more resembled aggregates (Fig. 2g). It is
possible that these structures are related to the
``double-membraned loops'' that were recently described by
Baumgart et al.(18) .
Centrifugation of Microsomal Fractions in Nycodenz
Gradients
Microsomal fractions were fractionated by
centrifugation for 60 min into Nycodenz gradients (density range
1.09-1.19 g/ml). The distributions of peroxisomal and marker
enzyme activities are shown in Fig. 3(upper panel).
Most of the gradient protein was found near the top, almost completely
co-sedimenting with esterase activity. Catalase separated into two main
peaks, the first was a high density peak in fractions 1-4
corresponding to a density of about 1.20 g/ml, with a shoulder of
catalase at densities of 1.15-1.20 g/ml. The second peak banded
at about 1.06-1.10 g/ml. In these gradients it was obvious that
catalase did not band at distinct densities but rather distributed in
broad peaks, indicating a profound heterogeneity. Aox activity showed a
bimodal distribution with some of the activity found in fractions
1-4 and most of the activity found at a density of about 1.12
g/ml (fractions 18 and 19). In contrast to catalase, most of the
activities of Aox (about 75%) and 3-hydroxyacyl-CoA dehydrogenase were
associated with the low density peaks near the top. The mean
distribution of catalase between the high and low density peaks was 35%
(range: 25-52%) and 27% (range: 21-30%), respectively; and
the mean distribution of Aox was 18% (range: 10-28%) and 53%
(range: 41-75%), respectively (means of four different
fractionations). Marker enzymes for mitochondria and lysosomes
(cytochrome oxidase and acid phosphatase, respectively) distributed
mainly in fractions between the two peaks of Aox and 3-hydroxyacyl-CoA
dehydrogenase, respectively. Esterase, the marker enzyme for
microsomes, banded in fractions 18 and 19 similar to Aox and MFE. In
control experiments (not shown) microsomal fractions pretreated with
different concentrations of Triton X-100 showed that released catalase
sedimented to the same fractions as the main low density peak of
catalase. Centrifugation of 100,000 g supernatants
showed that ``free'' catalase banded in these fractions. Thus
the low density peak of catalase is mainly due to protein which is
apparently not associated with vesicles. Fractions corresponding to the
high and low density peaks of Aox and MFE activities were analyzed by
transmission and immunoelectron microscopy. The morphology of
peroxisomal structures found at high density was similar to normal
peroxisomes, although the average sizes appeared somewhat smaller with
most peroxisomes having diameters less than 0.5 µm. Urate oxidase
cores appeared less frequently in these peroxisomes (data not shown).
The structure of the vesicles observed in the low density peak enriched
in Aox and MFE resembled smooth ER and the sizes of these vesicles were
usually below 100 nm.
g for 60 min. Fractions were
collected from the bottom of the tubes (from left to right) and analyzed for enzyme activities (upper
panels). It should be noted that the amount of mitochondria found
in the microsomal fractions is very low and does not contribute
significantly to the distributions of protein and 3-hydroxyacyl-CoA
dehydrogenase activity in the gradients. Gradient fractions (10 µl
of each) were electrophoresed by SDS-PAGE in 10% acrylamide gels. The
separated proteins were transferred to nitrocellulose membranes and
probed with the indicated antibodies. The blots were visualized by a
chemiluminescence method (lower
panels).
Immunoelectron Microscopy on High Density Nycodenz
Fractions
Immunocytochemistry on high density fractions from
Nycodenz gradients (corresponding to fractions 1-3 in Fig. 3a) showed that the labeling with the different
antibodies were highly specific (Fig. 4). Antibodies to catalase
specifically labeled membrane surrounded organelles evenly (Fig. 4a) as did antibodies to MFE (Fig. 4b). Labeling for PMP70 was confined to the
membranous structures of the particles (Fig. 4c). The
immunocytochemistry on these gradient fractions showed highly specific
labeling of vesicles with average diameters of about 0.2-0.3
µm. It thus appears to be an enrichment of smaller peroxisomes in
the microsomal fraction as compared with the LM fraction.
Immunoelectron Microscopy of Low Density Fractions from
Nycodenz Gradients
The low density fractions from Nycodenz
gradients contained mainly microsomes, and it was not possible to
identify peroxisomes by means of morphology. However, immunolabeling
experiments showed that few particles, with a ghost-like appearance,
and some particles, with a more granular matrix, contained peroxisomal
proteins. The structures of these particles resembled peroxisomes in
that they were surrounded by a single membrane, relatively round in
shape with some of them also having irregular protrusions (Fig. 5b) as reported by Lüers et al.(41) . In double immunolabeling experiments,
some particles were positively labeled for catalase but were apparently
negative for MFE (Fig. 5a), whereas other particles
were labeled by both antibodies (Fig. 5b).
Heterogeneous staining was also observed in double labeling experiments
for MFE and thiolase, demonstrating MFE-positive particles (apparently
negative for thiolase) and particles containing both enzymes (Fig. 5, c and d, respectively). We also
observed heterogeneities in the labeling for catalase and PMP70 (Fig. 5, e-f), and catalase and thiolase (Fig. 5, g and h). Fig. 5g shows a catalase-positive particle, lacking thiolase, and Fig. 5h shows a particle labeled for both proteins. The
ratio of labeling for thiolase and catalase was very different in
various particles (data not shown). Although the frequency of labeled
vesicles were low in these fractions, the antibody labeling was highly
specific for vesicular structures. The low frequency is due to the
small amount of peroxisomal vesicles present in these fractions which
mainly contain microsomes and to the method of choice for preparing the
fractions for immunoelectron microscopy. We considered it important to
avoid fixation of dilute fractions, followed by pelleting by high speed
centrifugation. Instead, the milder treatment of embedding peroxisomes
in gelatin and fixing in formaldehyde probably reduced the release of
proteins from the vesicles, although this procedure resulted in a final
2-fold dilution of the fraction. It should also be emphasized that the
immunolabeling experiments were carried out on consecutive cryosections
from the same block within each experiment.
Effect of Prolonged Density Gradient Centrifugation of
Microsomal Fractions
In order to investigate the effect of
centrifugation of the various peroxisomal subcompartments to their
equilibrium in Nycodenz, microsomal fractions were divided into three
aliquots and layered on top of three identical gradients that were
centrifuged for 1, 20, and 29 h, respectively. The gradients were
fractionated and catalase and Aox activities were measured. Fig. 6shows the enzyme distributions after 1 and 20 h of
centrifugation from one of these experiments. The enzyme distributions
after 20- and 29-h centrifugation were identical, indicating that these
enzymes had reached equilibrium at 20 h. After 1-h centrifugation,
catalase was distributed into two peaks of high (about 1.22 g/ml) and
low (about 1.08 g/ml) densities. After 20 h, catalase activity was
found in only one, although rather broad, peak at a density of about
1.16-1.22 g/ml. Essentially no catalase activity was found at
densities higher than 1.25 g/ml.
Identification and Separation of Different Peroxisomal
Subcompartments
In this study we report on the separation and
identification of peroxisomal subcompartments
(``populations'') of different sizes, equilibrium densities,
and heterogenous contents of peroxisomal proteins after centrifugation
of LM and microsomal fractions in Nycodenz gradients. Besides the
normal peroxisomes sedimenting to a density of >1.20 g/ml (after
fractionation of LM fractions), catalase, Aox, and MFE each separated
into two peaks after fractionation of LM and microsomal fractions in
Nycodenz gradients. One peak was found at a density around 1.19 g/ml,
possibly representing ``immature'' peroxisomes. The second
peak sedimented to a lower density. Equilibrium density centrifugations
for 20 and 29 h in Nycodenz-sedimented catalase into one broad peak at
1.15-1.22 g/ml, whereas Aox still showed a somewhat bimodal
distribution. By electron microscopy it was evident that the structures
of the low density organelles were quite different from normal
peroxisomes which typically contained crystalline cores and granular
matrix, indicating a substantial content of protein. The organelles
found at low density, that were labeled with antibodies to peroxisomal
proteins, often consisted of a membrane surrounded vesicle in which
most of the matrix area was apparently empty and sometimes had only a
rim of protein associated with the inside of the membrane. Although it
is well established that peroxisomes are leaky, which may result in the
formation of ``ghost-like'' peroxisomes, we conclude that the
peroxisomal proteins present in the low density fractions are at least
in part confined to peroxisomal vesicles. These are not likely to be
the result of selective protein release based on the following
arguments: (i) Immunogold labeling showed that the peroxisomal proteins
were mainly confined to vesicles surrounded by membranes that contained
PMP70. (ii) Our previous findings showed that thiolase and catalase are
the most leaky enzymes of those studied so far(42) , which was
supported by subsequent studies of Thompson and Krisans(43) .
It is unlikely that extraction of peroxisomal proteins should result in
the formation of catalase- and thiolase-positive particles which are
nearly devoid of MFE and Aox, proteins known to be retained in
peroxisomes. However, independent of the mechanisms for release of
proteins, any heterogeneity in the protein contents between various
vesicles indicates hetrogeneity of peroxisomal structures. (iii) The
peroxisomal ``ghosts'' isolated after extraction with
pyrophosphate (at pH 9) still contained the crystalline cores of urate
oxidase(42) . The low density particles found in the present
study lack such crystalline cores, but cores were observed in
peroxisomes banding at high density in these gradients. The very small
size of the vesicles lacking cores also argues against the presence of
cores in these vesicles. (iv) Although we have looked only at a limited
number of peroxisomal proteins so far, there appears to be a rather
strong correlation between the selective protein content of these
vesicles and proposed targeting mechanisms for protein import.Formation of Peroxisomal Subcompartments by Fragmentation
of a Peroxisome Reticulum
So how do these membranous structures
originate? The classical model for peroxisome biogenesis is based on
the assumption that peroxisomes are formed by budding from ER. This
assumption was derived from ultrastructural studies describing apparent
membrane connections between peroxisomes and ER (44, 45, 46) and similarities in membrane
composition(47) . Later biochemical data (on import of
peroxisomal proteins) as well as cytochemistry on catalase and glucose
6-phosphatase, support a concept where peroxisomal proteins are
synthesized on free ribosomes and posttranslationally imported into
peroxisomes(5, 6, 7, 8) . Recent
ultrastructural work showed that peroxisomes are, at least temporarily,
interconnected to form a
reticulum(10, 48, 49, 50, 51, 52) .
Staining of peroxisomes for various peroxisomal enzymes have
demonstrated heterogeneity both in the intensity between different
peroxisomes and in the staining intensity in
interconnections(18, 51, 52, 53) .
The weak labeling of interconnections suggest that a barrier may exist
that prevents diffusion of matrix proteins in these interconnections or
that matrix proteins are transported through the interconnections and
accumulating in the buds.
-terminal
presequence(59, 62, 63) . The targeting
sequence for catalase is not clear at present. It appears that the
internal -Ser-His-Leu-sequence may not target catalase to peroxisomes,
as it was demonstrated that the addition of one or two amino acids
COOH-terminal of the -SKL PTS abolishes import into
peroxisomes(64) . This implies that catalase may follow
another, still unknown, mechanism for targeting and import.-oxidation enzymes (Aox, MFE,
and thiolase). This is reflected by the high content of Aox and MFE in
the low density compartment after DEHP treatment. However, neither this
model nor our present results can fully explain the detailed mechanism
for the heterogenous distribution of peroxisomal proteins in different
peroxisomes.Proposed Model for Peroxisome Biogenesis
Based on
previous data in the literature and our present data, we favor a model
where peroxisomes are formed by budding from a peroxisome reticulum (Fig. 7). Import of proteins may be restricted to specific
locations in the reticulum with Aox and the multifunctional protein,
both carrying a COOH-terminal -SKL PTS, and being imported at a common
site, possibly in a ``body'' of the reticulum. A zonation of
protein import could be due to nonrandom distributions of receptors for
different PTS or differences in lipid composition affecting protein
targeting. Catalase and thiolase, probably carrying different
PTS's, may be imported at different sites by specific receptors.
However, the nature of this hypothetical peroxisomal reticulum remains
to be characterized. In a recent study, Ohno and Fujii (10) showed that in cultured hepatocytes, the peroxisomal
reticulum consisted of a smooth ER-like structure which they termed
``peroxisome-forming sheet.'' By DAB staining for catalase
they identified three types of peroxisomal segments: one type strongly
DAB-positive, a second type containing weak DAB reaction, and the third
type of sheet-like extensions containing no DAB reaction. They
concluded that the smooth membranous structures lacking DAB reaction
may represent the peroxisome forming sheet. Assuming that import of
peroxisomal proteins is restricted to specific locations in such a
reticulum, this model can explain the observations of small particles
(formed by homogenization of the narrow interconnections) containing
mainly Aox, MFE, and PMP70. It may also explain the formation of
catalase/thiolase-rich vesicles formed by homogenization, assuming that
the import of these proteins takes place, e.g. closer to the
``peroxisome bud.'' Urate oxidase import is also mediated by
the -SKL PTS(65) , but crystalloid cores may only form in the
growing buds after substantial accumulation of the protein. Withdrawal
of peroxisome proliferators quickly reverses the effects of most
parameters. It should be noted that heterogenous staining for at least
catalase appears after withdrawal of clofibrate treatment in mice (66) . Such a heterogeneity could be explained by deterioration
of the reticulum containing immature buds. If there is only a limited
import of protein into released peroxisomes, the heterogeneity may
persist for some time. One could speculate that the limited success so
far in achieving a high efficiency of in vitro import of
peroxisomal proteins into normal peroxisomes may not be due to
leakiness of these organelles, but may in fact be due to the presence
of few adequate PTS receptors.-oxidation pathway and methanol metabolism were found in one and
the same compartment(28) . This transient heterogeneity is
likely to reflect an incompetence of mature peroxisomes to incorporate
newly synthesized proteins. This idea is further supported by the
recent pulse-chase experiments on in vitro import which showed
that Aox is imported into a compartment of intermediate density (pulse)
and subsequently chased into a high density compartment (normal
peroxisomes)(67) . During our present investigation,
Lüers et al.(41) reported on a
modified scheme of subcellular fractionation of regenerating livers
which resulted in the separation of two apparently different
populations of peroxisomes. They found that
[
S]methionine injected into rats was first
incorporated into a ``light peroxisomal'' fraction
(5-30 min after injection) and that the label appeared in the
``heavy peroxisomal'' fraction at 90 min.
)))
We thank Dr. Björn Afzelius for
helpful discussions and Dr. Henrik Garoff and Dr. J. Kalervo Hiltunen
for critical reading of this manuscript.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  V. D. Antonenkov, R. T. Sormunen, and J. K. Hiltunen The behavior of peroxisomes in vitro: mammalian peroxisomes are osmotically sensitive particles Am J Physiol Cell Physiol, December 1, 2004; 287(6): C1623 - C1635. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Guo, Y. Y. Kit, J.-M. Nicaud, M.-T. Le Dall, S. K. Sears, H. Vali, H. Chan, R. A. Rachubinski, and V. I. Titorenko Peroxisome division in the yeast Yarrowia lipolytica is regulated by a signal from inside the peroxisome J. Cell Biol., September 29, 2003; 162(7): 1255 - 1266. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Schrader Tubulo-Reticular Clusters of Peroxisomes in Living COS-7 Cells: Dynamic Behavior and Association with Lipid Droplets J. Histochem. Cytochem., November 1, 2001; 49(11): 1421 - 1430. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. I. Titorenko, H. Chan, and R. A. Rachubinski Fusion of Small Peroxisomal Vesicles In Vitro Reconstructs an Early Step in the In Vivo Multistep Peroxisome Assembly Pathway of Yarrowia lipolytica J. Cell Biol., January 10, 2000; 148(1): 29 - 44. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M Schrader, S. King, T. Stroh, and T. Schroer Real time imaging reveals a peroxisomal reticulum in living cells J. Cell Sci., January 10, 2000; 113(20): 3663 - 3671. [Abstract] [PDF]
|
|
|
|
|
|
H. D. Fahimi and E. Baumgart Current Cytochemical Techniques for the Investigation of Peroxisomes: A Review J. Histochem. Cytochem., October 1, 1999; 47(10): 1219 - 1232. [Abstract] [Full Text]
|
|
|
|
|
|
A. Völkl, H. Mohr, and H. D. Fahimi Peroxisome Subpopulations of the Rat Liver: Isolation by Immune Free Flow Electrophoresis J. Histochem. Cytochem., September 1, 1999; 47(9): 1111 - 1118. [Abstract] [Full Text]
|
|
|
