Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
1
Import of a Cytosolic Protein into Lysosomes by Chaperone-Mediated
Autophagy Depends on its Folding State\\ §
Natalia Salvador‡\\ \\ , Carmen Aguado‡ || ||, Martin Horst¶ and Erwin Knecht‡DD
From the ‡Instituto de Investigaciones Citológicas, Fundación Valenciana de
Investigaciones Biomédicas, Amadeo de Saboya, 4, Valencia 46010, Spain and
the ¶ Faculty Phil. II, University of Basel, Missionstr. 64, CH-4055 Basel,
Switzerland
\This project was supported by the Ministerio de Educación y Ciencia (grants
PB97-1445 and PM98-0041) and Fundació La Caixa (grant 97/131-00).
§N. Salvador and C. Aguado contributed equally to this work.
\\A predoctoral fellow of the Consellería de Cultura Educación y Ciencia de la
Generalitat Valenciana.
||A postdoctoral fellow of the Fundación Bancaixa.
DSend correspondence to:
Erwin Knecht
Instituto de Investigaciones Citológicas, Fundación Valenciana de Investigaciones
Biomédicas, Amadeo de Saboya, 4, 46010-Valencia, Spain.
Tel.: 346-3391250; Fax: 346-3601453; E-mail: [email protected].
Copyright 2000 by The American Society for Biochemistry and Molecular Biology, Inc.
JBC Papers in Press. Published on June 20, 2000 as Manuscript M001394200 by guest on February 16, 2020
http://ww
w.jbc.org/
Dow
nloaded from
2
Running title: FOLDING STATE OF A CYTOSOLIC PROTEIN IMPORTED INTO
LYSOSOMES.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
3
SUMMARY.
We have analysed the folding state of cytosolic proteins imported in vitro
into lysosomes, using an approach originally developed by Eilers and Schatz,
(Nature 322 (1986), 228-232) to investigate protein import into mitochondria. The
susceptibility towards proteases of mouse dihydrofolate reductase (DHFR),
synthesised in a coupled transcription-translation system with rabbit reticulocytes,
decreased in the presence of its substrate analogue, methotrexate. This analogue
complexes with high affinity with the in vitro synthesised DHFR and locks it into a
protease-resistant folded conformation. DHFR was taken up by freshly isolated rat
liver lysosomes and methotrexate reduced this uptake by about 80%. A chimeric
DHFR protein, which carries the N-terminal presequence of subunit 9 of ATP
synthase preprotein from Neurospora crassa fused to its N-terminus, was taken
up by lysosomes more efficiently. Again, methotrexate abolished the lysosomal
uptake of the fusion protein, which was partially restored by washing of
methotrexate from DHFR or by adding together methotrexate and dihydrofolate,
the natural substrate of DHFR. Immunoblot analysis with-anti-DHFR of liver
lysosomes and of other fractions, isolated from rats starved for 88 h and treated
with lysosomal inhibitors, suggests that DHFR is degraded by chaperone-
mediated autophagy. Competition with ribonuclease A and stimulation by
ATP/Mg2+ and the heat shock cognate protein of 73 kDa show that the lysosomal
uptake of the fusion protein also occurs by this pathway. It is concluded that the
lysosomal uptake of cytosolic proteins by chaperone-mediated autophagy mainly
occurs by passage of the unfolded proteins through the lysosomal membrane.
Therefore, this mechanism is different from protein transport into peroxisomes, but
similar to the import of proteins into the endoplasmic reticulum and mitochondria.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
4
INTRODUCTION.
Proteins are continuously being degraded by both lysosomal (1) and non-
lysosomal (2) proteases. Lysosomes, which are found in almost all eukaryotic
cells, participate in intracellular protein degradation by various mechanisms (1):
endocytosis, crinophagy, direct conversion of endoplasmic reticulum (ER)1
cisternae into lysosomes, macroautophagy, microautophagy and a selective
pathway for the uptake and degradation of cytosolic proteins mediated by the heat
shock cognate protein of 73 kDa (hsc73), also known as chaperone-mediated
autophagy.
Endocytosis is the degradative route followed by extracellular proteins,
which are either recognised by specific receptors or simply trapped by nonspecific
uptake (3). This is also the degradative pathway followed by plasma membrane
proteins which, unlike the LDL or the transferrin receptors, do not recycle back to
the plasma membrane for reuse. By crinophagy (4), proteins originally destined for
secretion are delivered to lysosomes, when the demands for these proteins
decline, by a process involving fusion of secretory granules with endosomes
and/or lysosomes instead of with the plasma membrane. Also, proteins in transit
through the ER, as well as membrane and luminal resident proteins of the ER,
can be degraded, under certain conditions, by a direct conversion of cisternae of
the transitional part of the ER into lysosomes (5). In macroautophagy, or classical
autophagy, large areas of cytoplasm, typically including whole organelles, are
sequestered by a segregating structure and degraded by lysosomes (6, 7).
Microautophagy is a degradative route whereby portions of cytoplasm, including
certain organelles such as peroxisomes and/or cellular components down to the
level of macromolecules, are directly internalised into the lysosomal matrix by
various modifications of the lysosomal membrane which produce intralysosomal
vesicles (8). Finally, in serum-deprived confluent fibroblasts, a selective pathway
was described for the degradation of ribonuclease A (RNase A) which required
the pentapeptide sequence KFERQ, hsc73 and ATP/Mg2+ (9).
Since its original discovery, the chaperone-mediated autophagic pathway
for the uptake and degradation of cytosolic proteins has been found to be also
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
5
operative in rat liver, especially under long-term starvation (10, 11), and in kidney
and heart, but not in brain, testes and skeletal muscle (10). This selective uptake
of cytosolic proteins requires KFERQ-like sequences (9), ATP/Mg2+ and a
cytosolic (12) and an intralysosomal (13, 14) hsc73. In addition, the lysosomal
membrane glycoprotein lamp-2a was suggested as a receptor for this pathway
(15). The pathway has been also, at least partially, reconstituted in vitro with
lysosomes from rat liver (16, 17) and from human fibroblasts (12, 18), as well as
with yeast vacuoles (19). In this latter case, the relationship of the observed
uptake with the in yeast well established uptake of cytosolic proteins into vesicular
intermediates, which is followed by fusion with the vacuolar membrane (20, 21),
remains to be clearly demonstrated. From all this work, it appears that the
selective transport of proteins into mammalian lysosomes occurs by a process
which resembles, in some respects, the import of proteins synthesised on
cytosolic ribosomes into other cell compartments, such as mitochondria,
chloroplasts and the various types of microbodies (peroxisomes, glycosomes,
hydrogenosomes and glyoxysomes)(22-25).
Uptake of cytosolic proteins into lysosomes for degradation by the
chaperone-mediated autophagic pathway requires the movement of the protein
across the lysosomal membrane. Proteins can be transported into organelles
either in an unfolded (as for example occurs in mitochondria, chloroplasts or
endoplasmic reticulum, see (22-24) for review) or in a folded conformation (as is
the case, for example, for proteins transported into peroxisomes or into the
thylakoid membranes of the chloroplast, see (25, 26) for review). However, the
conformation of the protein as it passes from the cytosol into the lysosomal lumen
is still unknown. Since the original work of Eilers and Schatz (27), the cytosolic
enzyme dihydrofolate reductase (DHFR) has been used to investigate the
conformation of proteins as they pass across membranes, taking advantage of the
fact that its conformation can be stabilised by complexing it with methotrexate.
Here, we have analysed the effect of methotrexate on the uptake by isolated rat
liver lysosomes of the cytosolic enzyme DHFR synthesised in vitro in a coupled
transcription-translation system with rabbit reticulocytes. From the obtained
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
6
results we conclude that DHFR passes through the lysosomal membrane mostly
in an unfolded conformation.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
7
EXPERIMENTAL PROCEDURES.
Materials.
Metrizamide (grade I), trypsin, elastase, trypsin inhibitor, elastatinal,
phenylmethylsulfonyl fluoride, chymostatin, chloroquine, methotrexate,
dihydrofolate, Triton X-100, 3-[N-morpholino] propanesulfonic acid (MOPS), OsO4,
ATP, creatine phosphate, creatine phosphokinase, ATP-agarose, p-nitrophenol-N-
D-acetyl glucosaminide, 4-methylumbelliferyl-2-acetamido-2-deoxy-β-D-
glucopyranoside, goat anti-mouse IgM- and goat anti-mouse IgG (H+L)-alkaline
phosphatase conjugates, 5-bromo-4-chloro-3-indolyl phosphate, nitro blue
tetrazolium, RNase A and ribonuclease S-protein (RNase S-protein) were from
Sigma-Aldrich Química S.A. (Madrid, Spain). Leupeptin was from Peptide Institute
(Osaka, Japan). Antibodies against hsc73 were obtained from clone 13D3 (mouse
immunoglobulin M) from Maine Biotechnology Services (Portland, ME). Anti-
DHFR-polyclonal antibodies were raised against 6-His-tagged mouse DHFR
cloned in pQE16 (Qiagen, Hilden, Germany) and overexpressed in Escherichia
coli. Sodium-deoxycholate, uranyl acetate, proteinase K and sucrose were from
Merck (Darmstadt, Germany). Bovine serum albumin (fraction V) was from
Boehringer (Mannheim, Germany). Acrylamide/Bis 29:1 was from Bio-Rad
(Richmond, CA). Sodium dodecyl sulfate was from Serva (Heidelberg, Germany).
Cellulose nitrate paper (0.45 µm) was from Schleicher and Schuell (Dassel,
Germany). Glutaraldehyde was from Tousimis (Rockville, MD). Epon (Poly/Bed
812 resin) was from Polysciences (Warrington, PA). TNT-coupled rabbit
reticulocyte lysate system was from Promega (Madison, WI). Tran35S-label (70%
L-methionine) was from ICN Pharmaceuticals, Inc. (Irvine, CA) and 3H-leucine (44
Ci/mmol) was from NEN (Boston, MA). Other reagents were of the best analytical
quality available.
In vitro transcription and translation of mouse DHFR, a fusion protein between the
presequence of F0-ATPase subunit 9 and mouse DHFR and human
glyceraldehyde-3-phosphate dehydrogenase.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
8
A recombinant pSP65 plasmid containing a BamHI-HindIII fragment of 600
base pairs, carrying the full-length mouse DHFR with 6 histidines on its C-
terminus, was used for the expression of DHFR. Also, recombinant pGEM3
plasmid (28) containing a SmaI fragment of 777 base pairs, which includes the
fusion protein between the presequence and three amino acid residues of the
mature part of subunit 9 of the F0-ATP synthase of Neurospora crassa and the
mouse full length coding region of DHFR was used. This fusion protein is called
hereafter Su9-DHFR for brevity. For in vitro translation of glyceraldehyde-3-
phosphate dehydrogenase (GAPDH), a recombinant pSP64 plasmid, designed
pIC328, containing a PstI-BamHI fragment of about 1400 base pairs, which
includes the full-length human GAPDH coding region plus additional 5´- and 3´-
non-translated sequences (16), was used.
In vitro transcription and translation of the various DNAs with the TNT-
coupled rabbit reticulocyte system was carried out following the manufacturer´s
instructions.
Proteolytic susceptibility measurements.
Incubations were carried out at 37 ºC in a final volume of 20 µl. Assays
contained 5 µl of the standard synthesis reaction, with and without 250 nM
methotrexate, 0.1 M triethanolamine buffer, pH 7.6, and elastase (1% in terms of
protein and referred to the protein in the rabbit reticulocyte lysate, 111.5 mg
protein/ml). At the times indicated, portions were taken, 50 µM elastatinal was
added, and the samples were subjected to sodium dodecyl sulfate polyacrylamide
gel electrophoresis (SDS-PAGE) and fluorography. Experiments were also carried
out with trypsin and trypsin inhibitor with similar results (data not shown).
Preparation of lysosomes.
24 h-fasted male Wistar rats (Interfauna Ibérica S.A., San Feliu de Codines,
Spain) were used throughout. All rats were fed ad libitum for at least 7 days before
the experiments began. In some experiments, rats were treated with leupeptin (2
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
9
mg/100 g weight, intraperitoneally, 1.5 h before sacrifice). Lysosomes were
obtained from a light mitochondrial fraction in a discontinuous metrizamide
gradient by a procedure based on that originally developed by Wattiaux et al. (29).
Rats (200-250 g) were anaesthetised with ether and decapitated. 10 g of rat liver
were used and homogenate (7 ml of chilled 0.25 M sucrose/g liver) was filtered
through cheese-cloth and centrifuged in a Sepatech Biofuge 28RS (Heraeus
Sepatech GmbH, Osterode, Germany) at 4,800 x g for 10 min and the supernatant
at 17,000 x g for 10 min. After resuspension and washing once, the sediment (a
light mitochondrial-lysosomal fraction), suspended in 57% metrizamide, was
loaded on the bottom of a discontinuous metrizamide gradient (adjusted to pH
7.0). The discontinuous metrizamide gradient consisted of the following layers:
19.8% (top layer, 3.9 ml), 26.3% (3.5 ml), 32.8% (2.2 ml) and 57.0% (2.3 ml).
Centrifugation in the metrizamide gradient was for 90 min in an SW 40 Ti
rotor (Beckman) at 141,000 x g. Lysosomes were collected from the most upper
layer and the 26.3-19.8% interface, diluted at least five times with 0.3 M sucrose
and sedimented at 37,000 x g for 10 min in a Sorvall centrifuge (rotor SS-34).
Mitochondria were obtained from the 57-32.8% interface (16). All procedures were
carried out at 0-4 ºC. After a quick wash, lysosomes were carefully resuspended
in 10 mM MOPS, pH 7.2 and 0.3 M sucrose to a concentration of about 6 mg
lysosomal protein/ml and immediately used for the uptake experiments. Yields
were 0.3-0.5 mg of lysosomal protein per g of liver. The appearance of the
lysosomes by electron microscopy was not changed during the various
incubations used here. The integrity of the lysosomal membranes was also
estimated by measuring the latency of the lysosomal enzymes β-hexosaminidase
and β-N-acetyl glucosaminidase (11). It was >94% at the end of the incubations.
In some experiments, two lysosomal fractions were prepared by
centrifuging separately the top layer and the 26.3-19.8% interface of the
metrizamide gradient as described (14). These fractions (referred to hereafter as
HSC+ and HSC- lysosomes) contained or did not contain, respectively, hsc73
within the lysosomal lumen (14 and data not shown).
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
10
Cell-free assay of transport of cytosolic proteins into lysosomes.
Freshly isolated lysosomes (60 µg protein), treated or not with chymostatin
(see below), were incubated, for 10 min at 30 ºC in a final volume of 40 µl, in 0.3
M sucrose/10 mM MOPS buffer (pH 7.2), containing an ATP-regenerating system
(final concentrations: 10 mM MgCl2, 10 mM ATP, 2 mM phosphocreatine and 50
µg/ml creatine phosphokinase) (medium S) with the various in vitro synthesised
proteins (10 µl of the standard synthesis reaction, except where indicated), treated
or not with 250 nM methotrexate. Although lysosomes (11, 13, 14) and rabbit
reticulocyte lysates (data not shown) do contain hsc73 in enough amounts for
lysosomal transport, in most experiments additional hsc73 (5 µg/ml, final
concentration) was added to the incubation mixture. Chymostatin, when used,
was added to the lysosomes at 0 ºC for 10 min at three times its final
concentration (30 µM) and then diluted threefold with incubation buffer containing
the in vitro synthesised proteins. Experiments were also carried out with 10 mM
(final concentration) chloroquine with similar results (data not shown). Samples
were centrifuged in a Sepatech Biofuge 28 RS (rotor HFA 22.1) at 26,000 x g for 5
min at 4 ºC, the pellets were quickly washed once and pellets and supernatants
were subjected to SDS-PAGE and fluorography. In some experiments, a treatment
of proteinase K (20 µg per tube) in 0.3 M sucrose/MOPS buffer (pH 7.2) of the
washed sediments was carried out, with or without 1% Triton X-100, for 10 min at
0 ºC. After addition of 2 mM phenylmethylsulfonyl fluoride, samples were
centrifuged as above and the pellets and the supernatants were subjected to
SDS-PAGE and fluorography.
In experiments with HSC+ and HSC- lysosomes, prior to the standard
incubation with the in vitro synthesised proteins and the following treatments,
lysosomes were preincubated for 5 min at 30 ºC with 0.3M sucrose/10 mM MOPS
buffer (pH 7.2) containing or not hsc73 and/or the ATP-regenerating system.
Competition experiments between RNase A and RNase S-protein and the in vitro
synthesised proteins, were carried out as described (17).
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
11
Proteolysis measurements with freshly isolated lysosomes.
Chinese hamster ovary (CHO) cells were grown as described (30). They
were metabolically labelled for 48 h with 25 µCi/dish 3H-leucine and a cytosolic
fraction was prepared as described previously (31). Freshly isolated lysosomes
(60 µg of protein) were incubated for different times in medium S with the 3H-
labelled cytosolic proteins from CHO cells (10 µg protein with a specific activity of
about 10,000 dpm/µg protein) and 10 µl of the reticulocyte lysate transcription-
translation mixture, with or without DHFR plus 250 nM methotrexate, in a total
volume of 100 µl. At 0, 5, 10, 20, 30 and 40 min incubation, aliquots were taken
and mixed with 500 µl of 10% (w/v) trichloroacetic acid to terminate the reaction. 3
mg bovine serum albumin were added to each aliquot as carrier protein and the
samples were incubated for 10 min on ice and then pelleted by centrifugation.
Under these conditions, non-degraded proteins are precipitated, whereas
proteolytic fragments remain soluble. To control for the integrity of the isolated
lysosomes during the incubation period (16), lysosomes were incubated in parallel
as above but without the 3H-labelled soluble proteins. At 5, 10, 20, 30 and 40 min,
lysosomes were removed by centrifugation and supernatants were incubated with3H-labelled cytosolic proteins for the same times (5, 10, 20, 30 and 40 min,
respectively). The reaction was stopped with trichloroacetic acid plus bovine
serum albumin as above. The radioactivity of the acid-soluble and acid-insoluble
material (dissolved in 0.2 M NaOH containing 0.4% sodium deoxycholate) was
determined by liquid scintillation counting. Degradation was expressed as
percentage of the initial acid-insoluble radioactivity remaining at the different
incubation times.
General
DNA sequencing was performed using a 377 Automated DNA sequencer of
Applied Biosystems (Foster City, CA). Protein concentration was measured by a
modification, with sodium-deoxycholate (32), of the Lowry et al. (33) method and
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
12
using bovine serum albumin as the standard. Hsc73 was purified from rat liver
cytosol by ATP-agarose affinity chromatography (34). Cytosol was the supernatant
of three successive centrifugations of rat liver homogenates at 2,500 x g, 10 min,
17,000 x g, 10 min and 155,000 x g, 60 min in a Beckman L5-65 centrifuge, Ti-
70.1 rotor. SDS-PAGE was done according to Laemmli (35) and gels were used
for fluorography. The radioactivity associated with the various proteins in the
autoradiograms was quantified by phosphorimager analysis using a FLA-2000
image analyser and Science Lab 98 Image Gauge ver. 3.11 software from
FujiFilm España S.A. (Barcelona, Spain). Statistical analyses were carried out
with Student´s t test. Immunoblotting procedures were carried out as described
previously (14, 16). Lysosomal fractions, treated or not with methotrexate, were
fixed and embedded in Epon for conventional electron microscopy by standard
procedures (36). All data shown are representative results of at least three
separate experiments.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
13
RESULTS.
To test the effect of methotrexate on the conformation of DHFR, aliquots of
in vitro synthesised [35S]-labelled-DHFR, treated or not with methotrexate (250
nM), were incubated with elastase for increasing time periods (Fig. 1A). In the
absence of methotrexate (lanes 2-5), and under the conditions of the experiments,
DHFR was degraded, but in its presence (lanes 6-9), DHFR was protected from
elastase by the bound methotrexate.
To follow the import of DHFR into lysosomes, we mixed [35S]-labelled
DHFR with freshly isolated rat liver lysosomes in an isotonic medium at pH 7.2
(see Experimental Procedures) in the absence or in the presence of chymostatin
(to inhibit intralysosomal proteolysis) (Fig. 1B). This in vitro assay, to monitor the
uptake of cytosolic proteins into rat liver lysosomes, has been described in detail
(e.g. 16, 17). As it is the case with other proteins previously investigated, we also
found that, after incubation of the protein with the lysosomes, DHFR was
associated with the washed lysosomal pellets (lanes 2 and 5), especially in the
presence of chymostatin (lane 2). To eliminate the protein which is simply bound
to the external surface of the lysosomal membrane, we used a treatment with an
externally added protease (e.g. proteinase K, see Experimental Procedures). In
the presence of chymostatin, but not in its absence, part of the protein which was
untreated with methotrexate was protected from proteinase K (Fig. 1B, compare
lanes 3 and 6). In the presence of Triton X-100, DHFR associated with lysosomes
was degraded by proteinase K (lane 4). These results indicate that, in the absence
of methotrexate, DHFR is taken up into lysosomes where it is degraded by
lysosomal cathepsins unless chymostatin is present. When the same
experiments were carried out in the presence of methotrexate (Fig. 1C), almost no
uptake of DHFR could be detected, independently of the presence of the inhibitor
of lysosomal proteases. Thus, under the conditions of the experiment, there were
almost no difference in the amount of DHFR associated to the washed lysosomal
pellets with or without chymostatin (lanes 2 and 5), and proteinase K released
from the lysosomal membrane the associated DHFR (lanes 3 and 6).
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
14
Methotrexate neither affected the activity of elastase (as observed by
digestion of albumin by elastase in the presence of 250 nM methotrexate, data not
shown), nor the lysosomes which were used in these experiments. Thus, the
electron microscopic appearance of lysosomes incubated in the presence of
methotrexate for 10 min was not different from that of lysosomes incubated in
parallel but without methotrexate (Fig. 2). Also, the latency of the lysosomal
enzymes β-hexosaminidase and β-N-acetyl glucosaminidase was not modified by
the methotrexate treatment (data not shown). Finally, Fig. 3 shows that the
association of DHFR and methotrexate neither affects the lysosomal integrity
(compare s—s with ∆---∆), nor the proteolytic activity of lysosomes towards an
extract of cytosolic proteins (compare n—n with o---o).
Moreover, when a DHFR-unrelated protein (GAPDH) was used in similar
assays to those shown in Fig. 1 it was found that methotrexate neither modified
the proteolytic susceptibility of GAPDH towards elastase (data not shown) nor the
transport of GAPDH into lysosomes (Fig. 4).
The use of an exogenously added protease (proteinase K) in the assay
employed in Fig. 1B and C has some inconveniences when used to quantitatively
compare the lysosomal uptake of proteins with very different susceptibilities to
proteases, such as DHFR and DHFR-methotrexate. Thus, proteinase K, at the
conditions needed to eliminate DHFR-methotrexate from the lysosomal
membranes, also affects, to some extent, the lysosomal membrane. This,
together with the higher proteolytic susceptibility of DHFR without methotrexate to
proteinase K, determines that the amount of DHFR transported into lysosomes, in
the absence of methotrexate, might be underestimated. Therefore, to quantify
more accurately the effect of methotrexate on the binding and uptake of DHFR, we
incubated DHFR expressed in the coupled transcription-translation assay with
freshly isolated lysosomes with or without an inhibitor of lysosomal proteases. We
have already shown in previous papers (e.g. 11, 14, 16, 17, 37) that the protein
associated to lysosomes incubated without inhibitors of lysosomal proteases
represents the protein bound to the external surface of lysosomes because, under
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
15
these conditions, the internalised protein is degraded (as confirmed also here by
the proteinase K treatment, see Fig. 1B). In the presence of an inhibitor of
lysosomal proteases, part of the protein is insensitive to proteinase K if it is
transported into the lysosomes (Fig. 1B, lane 3). This part represents the protein
which has entered into lysosomes and is not degraded by the lysosomal
proteases (because they are inhibited) and which is also protected from the
exogenously added protease by the lysosomal membrane, unless a detergent
(Triton X-100) is added. Therefore, for a specific protein, the protein associated to
lysosomes in the presence of inhibitors of lysosomal cathepsins (here
chymostatin) represents both the protein bound to the lysosomal membrane
(released by proteinase K) and the internalised protein (insensitive to proteinase
K). Thus, it is possible to quantify the protein which has been taken up by
lysosomes by subtracting the protein associated to lysosomes in the presence
(protein bound to the lysosomal membrane and taken up by lysosomes) and in
the absence (protein bound to the lysosomal membrane) of inhibitors of lysosomal
proteases (11, 14, 16, 17, 37). Table 1 (two upper lines) shows that while the
binding of DHFR to the lysosomal membranes, with or without methotrexate, is
similar, the uptake of DHFR by lysosomes occurs much more efficiently in the
absence of methotrexate.
DHFR is not quite efficiently incorporated into lysosomes (see Table 1, two
upper lines), in agreement with previous observations from Dice´s lab (38, 39).
Therefore, we tested a chimeric DHFR protein (Su9-DHFR) with the N-terminal
presequence of the precursor polypeptide from Neurospora crassa ATP synthase
subunit 9 fused to the N-terminus of DHFR (see Experimental Procedures). The
fusion protein was quite efficiently incorporated into lysosomes (see below),
probably because this presequence contains two possible additional KFERQ-like
sequences (see Discussion). When this construct was incubated with elastase for
increasing time periods (Fig. 5A), we found that, in the absence of methotrexate
(lanes 2-5), DHFR was completely degraded. However, in the presence of
methotrexate (lanes 6-9), the DHFR moiety was protected from the elastase
attack by the bound methotrexate. Here we found, as also observed by others
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
16
(27), that the protease cleaves off the unprotected pre-sub 9 extension, because it
is in an unfolded state, leaving the DHFR moiety intact, because it is protected by
the bound methotrexate. The effect of methotrexate on the binding and uptake by
freshly isolated rat liver lysosomes of Su9-DHFR was quantified in four different
experiments similar to that shown in Fig. 5B (Table 1, two lower lines). These
results show again that while the binding of DHFR to the lysosomal membranes,
with or without methotrexate, is similar, the uptake of DHFR by lysosomes occurs
five times more efficiently in the absence of methotrexate. In both cases, the
uptake efficiency of the fusion protein was higher than when DHFR alone was
used. These results confirm the importance of an unfolded conformation for the
transport of DHFR.
To make sure that Su9-DHFR-methotrexate bound to the lysosomal
membrane was on the lysosomal import pathway, we tried to reverse the effect of
methotrexate, washing the drug from Su9-DHFR bound to the lysosomal
membrane (Fig. 6). During this procedure, part of the membrane-bound Su9-DHFR
was apparently released and there was also some degradation of the imported
Su9-DHFR. However, it was clear that partial removal of methotrexate from Su9-
DHFR bound to the lysosomal membrane led to an increased import of the fusion
protein (Fig. 6A, lanes 4 and 5, compare with lanes 6 and 7, where methotrexate
was present, compare also in Fig. 6B, uptake bars in mtx (+/-) and mtx (+/+)). This
import represents about one third of the total import of the fusion protein subjected
to two successive incubations without methotrexate (Fig. 6A, lanes 2 and 3,
compare also in Fig. 6B, uptake bars in mtx (+/-) and mtx (-/-)).
Dihydrofolate, the natural substrate of DHFR, binds to DHFR with lower
affinity than methotrexate. Thus, we reasoned that DHFR-dihydrofolate could
unfold easier than DHFR-methotrexate. As shown in Fig. 7, dihydrofolate had a
much less protective effect than methotrexate on the proteolysis of Su9-DHFR by
elastase (lanes 4 and 5 in Fig. 7A and +dhf -mtx in Fig. 7B, compare with lanes 6
and 7 in Fig. 7A and -dhf +mtx in Fig. 7B). When methotrexate and dihydrofolate
were used together (lanes 8 and 9 in Fig. 7A and +dhf +mtx in Fig. 7B), Su9-
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
17
DHFR was protected to a lower extent than with methotrexate alone, suggesting
competition of dihydrofolate and methotrexate for the binding site in DHFR. The
same results where obtained with DHFR without the N-terminal presequence of
the precursor polypeptide of subunit 9 of ATP synthase (data not shown).
To find out whether there is a correlation between the folding state of DHFR
and its lysosomal import, the effect of competition of dihydrofolate with
methotrexate on lysosomal import of Su9-DHFR was investigated. Without
methotrexate, dihydrofolate alone reduces the uptake of Su9-DHFR by about 50%
(Fig. 8A compare lanes 4-7 and lanes 2 and 3, compare also in Fig. 8B, + dhf and
- dhf uptake bars). This is probably because dihydrofolate keeps DHFR folded,
although less efficiently than methotrexate (see Fig. 7 and Fig. 8C and D). When
dihydrofolate was used together with methotrexate, it was found that the
lysosomal uptake of DHFR increased slightly (about 2 times) when compared
with methotrexate alone (compare, in Fig. 8C, lanes 4-7 and lanes 2 and 3,
compare also in Fig. 8D, + dhf and - dhf uptake bars). These results indicate that
dihydrofolate and methotrexate compete for the binding site of DHFR and that
dihydrofolate binding keeps DHFR into a more relaxed conformation, as
suggested also by the proteolytic susceptibility observations (Fig. 7). All these
observations stress again the importance of an unfolded conformation for an
efficient lysosomal transport of a cytosolic protein.
These experiments demonstrate that the folding state of both DHFR and
Su9-DHFR affects their uptake by isolated lysosomes. Although this experimental
system has been previously shown to reproduce the chaperone-mediated
autophagic pathway (11-15, 17, 18, 31, 37, 40), the following experiments were
carried out to show more clearly that both proteins are targeted to lysosomes by
this pathway. It is known that in rat liver prolonged starvation (64-88 h) activates
the chaperone-mediated autophagic pathway while reducing the activity of other
lysosomal pathways (10, 11). Therefore, we reasoned that if DHFR follows in vivo
the chaperone-mediated autophagic pathway it should be possible to detect, in
rats subjected to prolonged starvation, DHFR within liver lysosomes whose
proteases are inhibited. Thus, we isolated several liver fractions from rats starved
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
18
for prolonged periods (88 h) which were injected or not with the lysosomal
cathepsins inhibitor leupeptin (Fig. 9A). As expected, DHFR was detected, by
Western blot with an antibody towards DHFR, in whole homogenates (not shown)
and in the cytosol of both leupeptin-treated and untreated 88 h-starved rats (Fig.
9B, lanes 1 and 4). DHFR was not detected in mitochondria and in lysosomes
from untreated rats (Fig. 9A, lanes 2 and 3). However, it was detected in
lysosomes (Fig. 9B, lane 6), but again not in mitochondria (Fig. 9A, lane 5) of
leupeptin-treated rats (i.e. it was found in lysosomes with inhibited proteases).
Moreover, in leupeptin-treated rats, the lysosomal associated protein was partially
resistant to a proteinase K treatment which, under the same conditions,
completely degrades DHFR in the cytosol (Fig. 9C, compare lines 5 and 2,
respectively). Therefore, it appears that in liver of rats starved for 88 h, a condition
under which the chaperone-mediated autophagic pathway is activated, at least
part of the DHFR molecules are transported from the cytosol into lysosomes.
Similar experiments, carried out with rats starved for 20 h, showed that DHFR was
also found within isolated lysosomes from leupeptin-treated rats, but in relatively
smaller amounts as compared to lysosomes isolated from the liver of rats starved
for 88 h (data not shown). It should be noticed that in livers from rats starved for 20
h, although the chaperone-mediated autophagic pathway is also operative (11),
the macroautophagic pathway, which non-selectively degrades all kind of
proteins, is strongly activated (1, 6-8, 11).
Su9-DHFR is a chimeric protein which does not exist in vivo and which, in
the presence of mitochondria (i.e. under in vivo conditions), would be mainly
targeted to these organelles. Although this construct was used here merely as a
model to confirm the observations with DHFR, we carried out experiments to
investigate if the lysosomal uptake of this protein also occurs by chaperone-
mediated autophagy.
In a long series of experiments (reviewed in 9, 40) Dice´s laboratory
showed, both in vivo and in vitro, that RNase A (which contains a KFERQ
sequence at amino acids 7-11) follows the chaperone-mediated autophagic
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
19
pathway of protein degradation. The uptake of RNase A and of other cytosolic
proteins by the chaperone-mediated autophagic pathway requires both ATP/Mg2+
and cytosolic and lysosomal hsc73. Although lysosomes (11, 13, 14) and rabbit
reticulocyte lysates (data not shown) do contain hsc73 in sufficient amounts for
lysosomal transport, we separated, on the basis of their density, two different
lysosomal populations (14 and data not shown), one containing hsc73 (HSC+)
and another without this protein (HSC-). As shown in Fig. 10, without additions,
the lysosomal uptake of Su9-DHFR is considerably reduced, but HSC+ lysosomes
are slightly more effective than HSC- lysosomes in taking up the protein.
Preincubation (see Experimental Procedures) of both lysosomal populations with
an ATP-regenerating system and hsc73 increased considerably their uptake of
Su9-DHFR (Fig. 10A, lanes 2 and 5, compare with lanes 1 and 4, compare also in
Fig. 10B, +ATP +hsc73 bars with Control bars). When hsc73 was omitted from
the incubation mixture with the ATP-regenerating system, there was a decrease in
the uptake of Su9-DHFR by HSC- lysosomes (Fig. 10A, lanes 3 and 6, compare
with lanes 2 and 5, compare also in Fig. 10B, +ATP bars with +ATP +hsc73
bars). This effect was also observed with HSC+ lysosomes although it was less
evident. The same results were obtained when DHFR was used (data not shown).
Moreover, and as is the case with RNase A, Su9-DHFR and DHFR are
incorporated by lysosomes isolated from rats starved for 88 h more efficiently than
when using lysosomes from 20-h starved rats (data not shown). All these
observations are similar to those obtained, in similar experiments, with RNase A
(11, 14).
Finally, and in contrast to RNase A, RNase S-protein (amino acids 21-124
of RNase A and, thus, without a KFERQ sequence) is not a substrate of the
chaperone-mediated autophagic pathway (9, 40). Thus, we carried out
competition experiments between RNase A and RNase S-protein with Su9-DHFR
for lysosomal uptake. Increasing amounts of RNase A (Fig. 11A), but not of
RNase S-protein (Fig. 11B), reduced the lysosomal uptake of Su9-DHFR. Similar
results were obtained with DHFR (data not shown). All these results taken
together show that there are common elements in the lysosomal uptake
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
20
mechanisms of DHFR and Su9-DHFR and RNase A, which has been extensively
shown, both in vivo and in vitro (see 9, 40 for reviews), to be degraded by
chaperone-mediated autophagy.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
21
DISCUSSION.
In eukaryotic cells, cytoplasmic synthesised proteins can enter into their
organelles of residence in either an unfolded conformation (as for example in
mitochondria) or in a folded conformation (as for example in peroxisomes)(22-26,
41-43). The conformation of proteins which enter into lysosomes by chaperone-
mediated autophagy is unknown. We decided to approach this problem using
dihydrofolate reductase (DHFR) and a procedure first developed by Eilers and
Schatz (27) to study the transport of mitochondrial proteins into this organelle.
The mechanism whereby DHFR is degraded in rat liver is unknown.
Immunoblot analysis of rat liver fractions revealed the presence of DHFR in the
cytosol, as expected, and also within lysosomes from leupeptin-treated rats.
Leupeptin enters lysosomes by endocytosis (44) and it has been shown to be very
effective in inhibiting, under in vivo conditions, rat liver lysosomal cathepsins (45,
46). This has allowed the identification of several in vivo substrates of rat liver
lysosomes under a variety of conditions. Macroautophagy appears to rapidly
respond to starvation (1, 6-8) and degrades 30-40% of all rat liver cytosolic
proteins during the first 16-24 h of starvation, but obviously it can not continue for
longer times degrading proteins at such a high rate (11). Thus, the chaperone-
mediated autophagic pathway is primarily activated at later times of starvation,
when macroautophagy declines, and selectively degrades dispensable proteins
containing KFERQ-like sequences (10, 11). The fact that there is DHFR
associated to lysosomes, whose proteases are inhibited by leupeptin, specially
when isolated from rats starved for 88 h (Fig 9), suggests that, in rat liver, DHFR is
transported from the cytosol into lysosomes by chaperone-mediated autophagy.
Methotrexate binds to DHFR and stabilises the tertiary structure of the
protein so that it can not unfold efficiently. We were unable to find any indication
that methotrexate affected lysosomes or the transport of GAPDH (a DHFR-
unrelated protein which does not bind the drug)(Figs. 2-4). We found, in
agreement with others (38, 39), that DHFR is inefficiently transported in vitro into
lysosomes. Therefore, we also used in these experiments a fusion protein of
DHFR with the N-terminal presequence of the precursor polypeptide of the subunit
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
22
9 of the ATP synthase of Neurospora crassa (Su9-DHFR), which is much more
efficiently incorporated (see Table 1). Although this protein does not exist in vivo, it
turned out to be a useful model for the investigation of chaperone-mediated
autophagy. Figs. 10 and 11 and other experiments (see Results) show that the rat
liver lysosomal uptake of this protein also occurs by chaperone-mediated
autophagy. The reason why this fusion protein is incorporated into lysosomes with
such high efficiency is at present unknown. Mouse DHFR contains 2 KFERQ-
related sequences. One is KDRIN (beginning at amino acid 69), if Q can be
replaced by N, and the other is RLIEQ (beginning at amino acid 98). The precursor
polypeptide of the subunit 9 of the ATP synthase of N. crassa contains two
additonal motifs (KRTIQ and LKRTQ, starting at amino acids 33 and 45 in the
sequence) which could be also considered KFERQ-like sequences if T is
phosphorylated. Therefore, it is possible that these two additional KFERQ-like
motifs have a synergistic effect on the lysosomal uptake of DHFR, either by itself
or simply because in the fusion protein the KFERQ-like motifs present in the
DHFR molecule are more accessible to the lysosomal transport machinery. Since
the presequence is in an unfolded state, another explanation is that partially
unfolded proteins are imported with higher efficiencies compared to tightly folded
proteins. In any case, it is clear that addition of the precursor polypeptide
dramatically increases the lysosomal uptake of DHFR (Table 1).
We found that the transport of DHFR and of the DHFR fusion protein into
rat liver lysosomes in vitro was strongly inhibited by addition of methotrexate (Figs.
1B and C, 5B and Table 1). We also observed that inhibition of the uptake of
DHFR was directly related to the methotrexate-induced folding of DHFR (as
assessed by analyses of proteolytic susceptibility)(Figs. 1A and 5A). Moreover,
washing of the drug or addition of dihydrofolate, which competes with
methotrexate for DHFR binding and has a lower affinity of binding to DHFR,
increased the uptake efficiency of methotrexate-treated DHFR (Figs. 6 and 8,
respectively). These observations may probably be explained if both treatments
allow some degree of hsc73-induced unfolding of DHFR. Therefore, DHFR and
probably also other cytosolic proteins are transported into the lysosomes through
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
23
their membranes in an unfolded conformation. According to this model of
transport, and assuming that import of the protein starts at its N-terminal portion
(17), when the fusion protein treated with methotrexate interacts with the
lysosomal membrane, a membrane-spanning translocation intermediate could
accumulate, as the unfolded region crosses the bilayer but the folded moiety gets
"stuck". Since, in the absence of chymostatin, the lysosomal proteases are active,
they should destroy the unfolded region which has entered the lysosome.
Therefore, it may appear surprising that when Su9-DHFR treated with
methotrexate is incubated with lysosomes without chymostatin (i.e. which have
active the internal cathepsins, Fig. 5B, lane 6) no lower molecular mass band is
observed (as it occurs in the incubations with elastase, Fig. 5A, lanes 6-9 and 7A,
lanes 2-9, and with lysosomal cathepsins, data not shown). However, it could be
possible that cutting of this part of the molecule releases the fusion protein from
its binding to the membrane and, in fact, we found, in the supernatants of these
incubations with Su9-DHFR, a lower molecular mass band corresponding to the
DHFR size. Another, or an additional, possibility is that the unfolded portion of the
protein enters initially into a region of the lysosome which is protected from
lysosomal cathepsins. In this regard, it is noteworthy that in most lysosomes of
the dense type, it is possible to detect a narrow electron-lucent rim or band of
clear lysosomal lumen beneath the limiting lysosomal membrane which
separates the inner surface of the lysosomal membrane from the more dense
lysosomal core (see Fig. 3, inset). This area, which has been called a lysosomal
"halo" (47, 48), is believed to correspond to the sugar rich area of lysosomal
proteins on the lysosomal membrane which protects the lysosomal membrane
from the attack by their own hydrolases. Since the width of these area is about 30
nm, it may be possible that the unfolded part of the fusion protein (which is 69
amino acids long) were retained in this part of the lysosomes and, in this way, it
were not accessible to lysosomal proteases. Thus, assuming a size of 0.5 nm per
amino acid, the total length of the N-terminal presequence in its extended
conformation will be about 35 nm, which corresponds approximately to the size of
the lysosomal "halo". If, by analogy with the role in protein transport of other
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
24
intraorganellar heat shock proteins (22, 49), intralysosomal hsc73 (11, 13, 14)
also acts at the trans side of the lysosomal membrane to provide the driving force
to pull the transported protein emerging in the lysosomal lumen, this protective
"halo" may also explain why intralysosomal hsc73, which is susceptible to broken
lysosomes, is not immediately degraded within lysosomes by their cathepsins.
These possibilities merit further investigation, since a translocation intermediate
may be quite useful to identify other componentes of the putative lysosomal
machinery for protein translocation, by procedures similar to those employed with
mitochondria (50, 51).
In summary, DHFR and a DHFR fusion protein can be taken up by
lysosomes, and methotrexate strongly inhibits the uptake. This suggests, as is the
case for the import of proteins into the endoplasmic reticulum and mitochondria,
that DHFR enters lysosomes mainly in an unfolded state. This mechanism is
different from nonclassical protein-transport pathways (42), such as in
peroxisomes, and may explain our earlier unpublished observations that colloidal
gold particles, even when bound to GAPDH, are not efficiently imported in vitro
into isolated rat liver lysosomes. Since binding to the lysosomal membrane is not
affected by methotrexate, it appears that binding to lysosomes is independent of
the folding state of the protein, but that import requires unfolding. In this regard, it
has been shown (52), at least in mitochondria, that certain precursor proteins,
which assume their native forms before mitochondrial import, are unfolded during
translocation. Since hsc73 has been found both in the lysosomal lumen but also
associated to the cytosolic face of the lysosomal membrane (11, 13, 14), it may
be possible that this latter hsc73 fraction produces the required unfolding of the
proteins which are transported into lysosomes for degradation. In addition, DHFR
+ methotrexate can be also incorporated into lysosomes, albeit with low
efficiencies (about 20% of the import without methotrexate, according to the data
in Table 1). Therefore, it appears that either the transport system is not completely
strict in requiring an unfolded protein (as may also occur with other cell
organelles) or that, as previously suggested (37), two different lysosomal uptake
mechanisms coexist in the cell-free assay of lysosomal degradation: a direct
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
25
transfer through the lysosomal membrane corresponding to chaperone-mediated
autophagy and a less effective uptake, under the in vitro conditions used here,
which does not require protein unfolding (most probably microautophagy).
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
26
REFERENCES.
1. Knecht, E., Martín de Llano, J. J., Andreu, E. J., and Moreno Miralles, I.
(1998) Adv. Cell Mol. Biol. 27, 201-234
2. Coux, O., Tanaka, K., and Goldberg, A. L. (1996) Annu. Rev. Biochem. 65,
801-847
3. Mellman, I. (1996) Annu. Rev. Cell Develop. 12, 575-626
4. Lenk, S. E., Fischer, D. L., and Dunn Jr., W. A. (1991) Eur. J. Cell Biol. 56,
201-209
5. Noda, T., and Farquhar, M. G. (1992) J. Cell Biol. 119, 85-97
6. Seglen, P. O., and Bohley, P. (1992) Experientia 48, 158-172
7. Mortimore, G. E., and Kadowaki, M. (1994) in Cellular Proteolytic Systems
(Ciechanover, A. J., and Schwartz, A. L., eds.), pp. 65-87, Wiley-Liss, Inc.,
New York
8. Dunn, W. A., Jr. (1994) Trends Cell Biol. 4, 139-143
9. Dice, J. F. (1990) Trends Biochem. Sci. 15, 305-309
10. Wing, S. S., Chiang, H.-L., Goldberg, A. L., and Dice, J. F. (1991) Biochem. J.
275, 165-169
11. Cuervo, A. M., Knecht, E., Terlecky, S. R., and Dice, J. F. (1995) Am. J. Cell.
Physiol. 269, C1200-C1208
12. Chiang, H.-L., Terlecky, S. R., Plant, C. P., and Dice, J. F. (1989) Science
246, 382-385
13. Agarraberes, F. A., Terlecky, S. R., and Dice, J. F. (1997) J. Cell Biol. 137,
825-834
14. Cuervo, A. M., Dice, J. F., and Knecht, E. (1997) J. Biol. Chem. 272, 5606-
5615
15. Cuervo, A. M., and Dice, J. F. (1996) Science 273, 501-503
16. Aniento, F., Roche, E., Cuervo, A. M., and Knecht, E. (1993) J. Biol. Chem.
268, 10463-10470
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
27
17. Cuervo, A. M., Terlecky, S. R., Dice, J. F., and Knecht, E. (1994) J. Biol.
Chem. 269, 26374-26380
18. Terlecky, S. R., Chiang, H.-L., Olson, T. S., and Dice, J. F. (1992) J. Biol.
Chem. 267, 9202-9209
19. Horst, M., Knecht, E., and Schu, P. V. (1999) Mol. Biol. Cell 10, 2879-2889
20. Baba, M., Osumi, M., Scott, S. V., Klionsky, D. J., and Ohsumi, Y. (1997) J.
Cell Biol. 139, 1687-1695
21. Chiang, M. C., and Chiang, H.-L. (1998) J. Cell Biol. 140, 1347-1356
22. Schatz, G., and Dobberstein, B. (1996) Science 271, 1519-1526
23. Voos, M., Martin, H., Krimmer, T., and Pfanner, N. (1999) Biochim. Biophys.
Acta 1422, 235-254
24. Chen, X., and Schnell, D. J. (1999) Trends Cell Biol. 9, 222-227
25. Subramani, S. (1998) Physiol. Rev. 78, 171-188
26. Dalbey, R. E., and Robinson, C. (1999) Trends Biochem. Sci. 24, 17-22
27. Eilers, M., and Schatz, G. (1986) Nature 322, 228-232
28. Pfanner, N., Müller, H. K., Harmey, M. A., and Neupert, E. (1987) EMBO J. 6,
3449-3454
29. Wattiaux, R., Wattiaux De Conninck, S., Ronveaux-Dupal, M. F., and Dubois,
F. (1978) J. Cell. Biol. 78, 349-368
30. Martín de Llano, J. J., Andreu, E. J., and Knecht, E. (1996) Anal. Biochem.
243, 210-217
31. Aniento, F., Roche, E., and Knecht, E. (1997) Electrophoresis 18, 2638-2644
32. Bennett, J. P. (1982) in Techniques in the Life Sciences, Biochemistry
(Hesketh, T. R., Kornberg, H. L., Metcalfe, J. C., Northcote, D. H., Pogson, C.
X., and Tipton, K. F., eds.), Vol. B4/1, Section B-408, pp. 15-22, Elsevier
Biomedical, Amsterdam
33. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol.
Chem. 193, 265-275
34. Welch, F. J., and Feramisco, J. R. (1985) Mol. Cell Biol. 5, 1229-1237
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
28
35. Laemmli, U. K. (1970) Nature 227, 680-685
36. Robards, A. W., and Wilson, A. J. (eds.)(1993) Procedures in Electron
Microscopy, Section 5, pp. 0.1-4.26, John Wiley & Sons, New York
37. Aniento, F. Papavassiliou, A. G., Knecht, E., and Roche, E. (1996) FEBS Lett.
390, 47-52
38. Chiang, H.-L., and Dice, J. F. (1988) J. Biol. Chem. 263, 6797-6805
39. Terlecky, S. R., and Dice, J. F. (1993). J. Biol. Chem. 268, 23490-23495
40. Cuervo, A. M., and Dice, J. F. (1998) J. Mol. Med. 76, 6-12
41. Blobel, G. (1995) Cold Spring Harb. Symp. Quant. Biol. 60, 1-10
42. Teter, S. A., and Klionsky, D. J. (1999) Trends Cell Biol. 9, 428-431
43. Klionsky, D. J., and Ohsumi, Y. (1999) Annu. Rev. Cell Dev. Biol. 15, 1-32
44. Wilcox, D., and Mason, R. W. (1992) Biochem. J. 285, 495-502
45. Kominami, E., Hashida, S., Khairallah, E. A., and Katunuma, N (1983) J. Biol.
Chem. 258, 6093-6100
46. Ueno, T., Ishidoh, K., Mineki, R., Tanida, I., Murayama, K., Kadowaki, M., and
Kominami, E. (1999) J. Biol. Chem. 274, 15222-15229
47. Holtzmann, E. (1976). Lysosomes: A Survey. Springer Verlag, Wien
48. Daems, W. T., Wisse, E., and Brederoo, P. (1972) in Lysosomes: A
Laboratory Handbook (Dingle, J. T., ed.), pp. 156-199, North Holland
Publishing Company, Amsterdam
49. Neupert, W. (1997) Annu. Rev. Biochem. 66, 863-917
50. Vestweber, D., Brunner, J., Baker, A., and Schatz, G. (1989) Nature 341, 205-
209
51. Horst, M., Hilfiker-Rothenfluh, S., Oppliger, W., and Schatz, G. (1995) EMBO
J. 14, 2293-2297
52. Huang, S., Ratliff, K. S., Schwartz, M. P., Spenner, J. M., and Matouschek, A.
(1999) Nature Struct. Biol. 6, 1132-1138
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
29
FOOTNOTES.
Abbreviations used.1The abbreviations used are: ER, endoplasmic reticulum; hsc73, heat shock
cognate protein of 73 kDa; RNase A, ribonuclease A; RNase S-protein, amino
acids 21-124 of RNase A; DHFR, dihydrofolate reductase; MOPS, 3-[N-
morpholino] propanesulfonic acid; Su9-DHFR, a fusion protein between the
presequence and three amino acid residues of the mature part of subunit 9 of the
F0-ATP synthase of Neurospora crassa and the mouse full length coding region of
DHFR; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
Acknowledgements.
We are very grateful to J. F. Dice and A. M. Cuervo (Tufts University, Boston) and
M. E. Armengod for carefully reading the manuscript and valuable advice. We also
thank A. Montaner and D. Cerveró for technical assistance.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
30
FIGURE LEGENDS.
Fig. 1. Methotrexate protects DHFR from proteolysis by elastase (A) and inhibits
its uptake by freshly isolated lysosomes (B,C). A: Effect of methotrexate on the
proteolysis by elastase (1%, w/w) of in vitro synthesised DHFR. Lane 1:
unincubated DHFR; lanes 2-5: DHFR + elastase incubated at 37 ºC for 0, 15, 30
and 60 min; lanes 6-9: DHFR + elastase + methotrexate (250 nM) incubated as in
lanes 2-5, respectively. B and C: Uptake of in vitro synthesised DHFR by freshly
isolated rat liver lysosomes. Aliquots of an in vitro synthesis of DHFR (10 µl of
total synthesis), non-treated (B, - Methotrexate) or treated (C, + Methotrexate) with
methotrexate (250 nM), were incubated (lanes 2-7) in the cell-free assay with
freshly isolated lysosomes (60 µg protein), with (lanes 2-4) or without (lanes 5-7)
30 µM chymostatin, for 10 min at 30 ºC in medium S (see Experimental
Procedures). At the end of the incubation, samples were centrifuged and the
sediments were resuspended in ice-cold 0.3 M sucrose/10 mM MOPS buffer (pH
7.2) and treated (lanes 3, 4, 6, 7) or not (lanes 2, 5) for 10 min at 0 ºC with
proteinase K, in the presence (lanes 4, 7) or not (lanes 3, 6) of Triton X-100, as
described in Experimental Procedures. Then, lysosomes were centrifuged and the
pellets were subjected to SDS-PAGE and fluorography. Lane 1 corresponds to 3%
of the synthesis, which was incubated with the lysosomes. In the presence of
inhibitors of lysosomal proteolysis (lanes 2-4), but not in its absence (lanes 5-7),
DHFR untreated with methotrexate (B), is in part inside lysosomes and, therefore,
protected against proteinase K (lane 3), unless Triton X-100 (lane 4) is added. The
protein associated with lysosomes in the presence of chymostatin (lane 2) roughly
corresponds to the sum of lane 3 (internalised DHFR) and lane 5 (protein
adsorbed to the external surface of lysosomes). Without chymostatin (lanes 5-7),
DHFR associated to lysosomes remains outside, bound to the lysosomal
membrane and, thus, sensitive to exogenously added proteinase K (lane 6).
According to these same criteria, DHFR treated with methotrexate (C) is not
transported and remains outside the lysosomes (compare lanes 3 and 6 in B and
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
31
C). The positions of molecular-mass markers and their size in kDa are indicated
on the left.
Fig. 2. A: Typical ultrastructure of freshly isolated lysosomes from rat liver
incubated for 10 min at 30 ºC in the presence of 250 nM methotrexate. Bar = 0.5
µm. The inset shows an area of this preparation at higher magnification. Note the
clear "halo" (see Discussion) beneath the lysosomal membrane (arrow) which is
visible in most lysosomes of the fraction. Bar = 0.25 µm. No differences were
noticed in the electron microscopic appearance of lysosomes when compared
with an unincubated lysosomal preparation (B). Bar = 0.5 µm.
Fig. 3. Methotrexate plus DHFR neither affects the lysosomal integrity nor the
proteolytic activity of freshly isolated lysosomes. 3H-labelled cytosolic proteins
from CHO cells (about 100,000 cpm), prepared as described in Experimental
procedures, and 10 µl of the reticulocyte lysate transcription-translation mixture,
with (closed symbols) or without (open symbols) DHFR plus 250 nM
methotrexate, were incubated in the cell-free assay under standard conditions
with (n—n, o---o) or without (l—l, m---m) freshly isolated lysosomes (60 µg of
protein). At the indicated times, aliquots were taken to measure proteolysis of the
CHO cells soluble proteins as described in Experimental Procedures. To evaluate
the leakage of proteolytic activity into the medium (lysosomal integrity), lysosomes
were incubated in parallel as above but without the 3H-labelled soluble proteins. At
5, 10, 20, 30 and 40 min, lysosomes were removed by centrifugation,
supernatants were incubated as above with 3H-labelled soluble proteins for the
indicated times and proteolysis was measured (s—s, ∆---∆). The figure shows a
typical experiment. Similar results were obtained with two different preparations.
Fig. 4. A and B: Methotrexate does not affect the transport and uptake of in vitro
synthesised GAPDH by freshly isolated rat liver lysosomes. The experiments were
carried out as in Fig. 1B (Fig. 4A) and 1C (Fig. 4B), except that GAPDH cDNA was
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
32
used instead of DHFR cDNA. Lane 7 corresponds to 6% of the synthesis which
was incubated with the lysosomes. The positions of molecular-mass markers and
their size in kDa are indicated on the left. The histogram (C) below shows the
phosphorimager analysis of the fluorographies: uptake + binding (lane 1), binding
(lane 4), uptake (lane 2).
Fig. 5. A: Effect of methotrexate on the proteolysis of in vitro synthesised Su9-
DHFR by elastase (1%, w/w). Lane 1: unincubated Su9-DHFR; lanes 2-5: Su9-
DHFR + elastase incubated at 37 ºC for 15, 30, 45 and 60 min; lanes 6-9: Su9-
DHFR + elastase + methotrexate incubated, respectively, for the same time
periods as in lanes 2-5. B: Su9-DHFR is efficiently taken up by lysosomes and
methotrexate abolishes this uptake. A cDNA encoding Su9-DHFR was in vitro
transcribed and translated in a rabbit reticulocyte system. Lysosomes (60 µg),
prepared as described in Experimental Procedures, and treated (+) or not (-) with
an inhibitor of lysosomal cathepsins (chymostatin) were incubated in medium S
with Su9-DHFR (10 µl of the reticulocyte lysate transcription-translation mixture)
for 10 min at 30 ºC in the presence (+) or absence (-) of 250 nM methotrexate.
After incubation, the lysosomes were pelleted by centrifugation, washed in 0.3 M
sucrose/10 mM MOPS buffer (pH 7.2), and analysed by SDS-PAGE and
fluorography. Lysosomal uptake can be calculated from the difference in the
values measured for the Su9-DHFR radioactivity associated to the lysosomal
pellets treated (uptake + binding, lanes 3 and 5) or not (binding, lanes 4 and 6)
with chymostatin. Lanes 1 and 2 contain, respectively, 6 and 3% of the
reticulocyte lysate transcription-translation mixture added to the lysosomes. The
positions of molecular-mass markers and their size in kDa are indicated on the
left.
Fig. 6. A: Partial removal of methotrexate increases the uptake efficiency of
methotrexate-treated DHFR by freshly isolated lysosomes. Lysosomes (60 µg)
were incubated, for 10 min at 30 ºC, in a first standard incubation (Incubation 1,
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
33
see Experimental Procedures) with Su9-DHFR (10 µl of the reticulocyte lysate
transcription-translation mixture) in the presence or not of lysosomal protease
inhibitors (30 µM chymostatin) and methotrexate (250 nM), as indicated.
Afterwards, lysosomes were reisolated at 4 ºC, resuspended in medium S in the
absence (lanes 2-5) or in the presence (lanes 6, 7) of the same concentrations of
methotrexate and chymostatin as indicated, and incubated for 10 further minutes
at 30 ºC (Incubation 2). Then, lysosomes were pelleted by centrifugation, washed
in 0.3 M sucrose/10 mM MOPS buffer (pH 7.2), and analysed by SDS-PAGE and
fluorography. Lane 1 contains 6% of the reticulocyte lysate transcription-
translation mixture added to the lysosomes. The positions of molecular-mass
markers and their size in kDa are indicated on the left. The histogram below (B)
shows the phosphorimager analysis of the fluorography. Lysosomal uptake was
calculated as in Fig. 5. The presence (+) or absence (-) of methotrexate (mtx) in
the first and second incubations (separated by a slash) is indicated under
brackets.
Fig. 7. A: Effect of methotrexate (250 nM) and dihydrofolate (250 nM) on the
proteolysis of in vitro synthesised Su9-DHFR by elastase (1%, w/w). Incubations
were carried out at 37 ºC for 15 and 60 min. Lane 1: unincubated Su9-DHFR;
lanes 2-9: Su9-DHFR incubated with elastase for 15 and 60 min, respectively:
without additions (lanes 2, 3) or in the presence of 250 nM dihydrofolate (lanes 4,
5), 250 nM methotrexate (lanes 6, 7) or both together (lanes 8, 9). The positions of
molecular-mass markers and their size in kDa are indicated on the left. The
histogram below (B) shows the phosphorimager analysis of the fluorography. The
presence (+) or absence (-) of methotrexate (mtx) and dihydrofolate (dhf) in the
incubations is indicated.
Fig. 8. Addition of dihydrofolate without (A) or with (C) methotrexate affects the
uptake of in vitro synthesised Su9-DHFR by freshly isolated lysosomes. 5 µl
aliquots of an in vitro synthesis of Su9-DHFR were incubated, in the presence (C)
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
34
or not (A) of methotrexate and without (lanes 2, 3) or with (lanes 3-7) the indicated
concentrations of dihydrofolate, with freshly isolated lysosomes (60 µg protein),
treated (lanes 2, 4, 6) or not (lanes 3, 5, 7) with chymostatin, under standard
conditions. Lane 1 corresponds to 12% of the synthesis which was incubated with
the lysosomes. The positions of molecular-mass markers and their size in kDa are
indicated on the left. The histograms on the right (B, - Methotrexate, and D, +
Methotrexate) show the phosphorimager analysis of the fluorographies on the left
(A and C, respectively). Lysosomal uptake was calculated as in Fig. 5. The
presence (+) or absence (-) of dihydrofolate (dhf) and its concentration when
added (under brackets) are shown.
Fig. 9. Immunolocalisation of DHFR in rat liver fractions. Rats starved for 88 h
were treated or not with leupeptin and liver fractions were prepared as described
under Experimental Procedures. Proteins (50 µg for cytosol and 100 µg for
mitochondria and lysosomes) were separated by SDS-PAGE and stained with
Coomassie Blue R-250 (A) or immunoblotted with anti-DHFR (B and C). A and B:
Cytosol (CYT, lanes 1 and 4), mitochondria (MIT, lanes 2 and 5) and lysosomes
(LYS, lanes 3 and 6) from rats treated (+leupeptin, lanes 4-6) or not (-leupeptin,
lanes 1-3) with leupeptin. C: Cytosolic (lanes 1-3) and lysosomal (lanes 4-6)
fractions isolated from rats treated with leupeptin were incubated for 15 min at 0
ºC without (lanes 1 and 4) or with (lanes 2, 3, 5 and 6) proteinase K (10%, w/w)
and in the presence (lanes 3 and 6) or in the absence (lanes 2 and 5) of 1% Triton
X-100. The positions of molecular-mass markers and their size in kDa are
indicated on the left.
Fig. 10. Effect of ATP/Mg2+ and hsc73 on the uptake of Su9-DHFR by two different
lysosomal populations. A: Su9-DHFR was incubated, as described under
Experimental Procedures, with freshly isolated HSC- (lanes 1-3) and HSC+ (lanes
4-6) lysosomes (i.e. not containing or containing lysosomal hsc73, see
Experimental Procedures), without (lanes 1, 4) or with an ATP-regenerating
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
35
system (ATP) plus hsc73 (lanes 2, 5) or with an ATP-regenerating system alone
(lanes 3, 6). After incubation, samples were treated with proteinase K as
described in Experimental Procedures, centrifuged and pellets and supernatants
were subjected to SDS-PAGE and fluorography. Only lysosomal pellets are
shown; Su9-DHFR in supernatants was totally digested by added proteinase K
(data not shown). The positions of molecular-mass markers and their size in kDa
are indicated on the left. The histogram below (B) shows the mean values of the
phosphorimager analysis of three different fluorographies as in A.
Fig. 11. Effect of RNase A (with KFERQ sequence) and RNase S-protein (without
KFERQ sequence) on the uptake of Su9-DHFR by lysosomes. Lysosomes (60 µg
of protein) treated with 30 µM chymostatin were incubated at 30ºC for 10 min
under standard conditions with an in vitro synthesis of Su9-DHFR without (lane 1)
or with increasing amounts of RNase A (A) and RNase S-protein (B) as labelled
on the figure (lanes 2-5). At the end of the incubation, samples were centrifuged,
the sediments were treated with proteinase K as described in Experimental
Procedures and subjected to SDS-PAGE and fluorography. The positions of
molecular-mass markers and their size in kDa are indicated on the left.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
36
Table 1
Quantification of the binding and uptake of DHFR and Su9-DHFR, treated or not
with methotrexate, by freshly isolated lysosomes.
Binding + Binding (%) uptake (%) Uptake (%)
DHFR 1.2 ± 0.5 2.6 ± 0.8 1.4 ± 0.3
DHFR + methotrexate 1.1 ± 0.3 1.4 ± 0.4 0.3 ± 0.1
Su9-DHFR 3.4 ± 0.7 14.4 ± 3.1 11.0 ± 2.7
Su9-DHFR+ methotrexate 3.9 ± 0.7 6.1 ± 1.0 2.2 ± 0.4
Quantification, using phosphorimager analysis, of the autoradiograms from four
different experiments as in Fig. 5 (DHFR, two upper lines in the table) and in Fig.
6B (Su9-DHFR, two lower lines in the table). The protein associated to the
lysosomes in the absence (binding, Fig. 5 and 6B, lanes 4 and 6, which
represents the protein adsorbed to the external surface of the lysosomes) or in the
presence (binding + uptake, Fig. 5 and 6B, lanes 3 and 5, which represents
surface bound plus internalised protein) of chymostatin are shown in columns 1
and 2 of the table, respectively. Uptake (column 3 of the table) was calculated
from the difference in the values measured for columns 2 and 1. Values are
expressed as a percentage of the total protein added to the assay. Differences in
uptake, with and without methotrexate, were significant at p < 0.005.
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from
Natalia Salvador, Carmen Aguado, Martin Horst and Erwin Knechtdepends on its Folding State
Import of a Cytosolic Protein into Lysosomes by Chaperone-Mediated Autophagy
published online June 20, 2000J. Biol. Chem.
10.1074/jbc.M001394200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here
by guest on February 16, 2020http://w
ww
.jbc.org/D
ownloaded from