| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Research Department and
Institute for Pathology, Cantonal Hospital St. Gallen, St. Gallen, Switzerland; and
Institute of Experimental Immunology, Department of Pathology, University Hospital Zurich, Zurich, Switzerland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
treatment, which leads to a gradual
replacement of constitutive proteasomes by immunoproteasomes in vitro.
However, a quantitative analysis of changes in the steady state subunit
composition of proteasomes during an immune response against viruses or
bacteria in vivo has not been reported. Here we show that the infection
of mice with lymphocytic choriomeningitis virus or
Listeria monocytogenes leads to an almost
complete replacement of constitutive proteasomes by immunoproteasomes
in the liver within 7 days. Proteasome replacements were markedly
reduced in IFN--/- mice, but were only slightly
affected in IFN-
R-/- and perforin-/-
mice. The proteasome regulator PA28/
was up-regulated, whereas
PA28 was reduced in the liver of lymphocytic choriomeningitis
virus-infected mice. Proteasome replacements in the liver strongly
altered proteasome activity and were unexpected to this extent, since
an in vivo half-life of 12 days had been previously assigned to
constitutive proteasomes in the liver. Our results suggest that during
the peak phase of viral and bacterial elimination the antiviral
cytotoxic T lymphocyte response is directed mainly to
immunoproteasome-dependent T cell epitopes, which would be a novel
parameter for the design of vaccines. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The proteasome system consists of a proteolytic core unit, the 20S
proteasome, and regulators such as PA700, PA28/, PA28, and
PI31, which control substrate acquisition and catalytic activity of the
proteasome. The 20S proteasome is constructed like a barrel of four
stacked rings. The two outer rings contain seven different subunits of
the type, whereas the inner two rings contain seven different
subunits of the type that are numbered according to their positions
in the ring. The subunits 1 (
,Y), 2 (MC14, LMP9, Z), and 5
(MB1, X) bear the catalytically active sites of the 20S proteasome
facing the inner cavity of the protease. When cells are treated with
the inflammatory cytokines IFN- and TNF-, three additional active
site subunits, named 1i (LMP2), 2i (MECL-1), and 5i (LMP7),
are transcriptionally induced and take the place of their constitutive
homologues during proteasome neosynthesis. This substitution of active
site subunits changes the proteolytic specificity of the proteasome
(3, 4, 5, 6) and leads to marked changes in the distribution of
fragments that are produced from polypeptides (7, 8). The
exchange of LMP2 for was unanimously reported to down-regulate
cleavages C-terminal of acidic residues (the caspase-like activity) and
favor the cleavage C-terminal of hydrophobic residues (the
chymotrypsin-like activity). This change in cleavage specificity
conforms to structural requirements of class I peptide ligands that
possess hydrophobic C termini in the mouse and either hydrophobic or
basic C termini in the human. Conflicting data have been reported on
the effects of the two other subunit exchanges of LMP7 for MB1
(8, 9, 10) and MECL-1 for MC14 (Z) (11), and
structural analysis does not predict major changes in the specificity
of the P1 pockets (12), leaving the functions of these
latter two exchanges elusive. With respect to the bulk production of
MHC class I ligands, it was found in LMP2- and LMP7-deficient cell
lines (13, 14) that these subunits are not generally
required to maintain class I cell surface expression. For the
generation of specific epitopes, however, the expression of either LMP2
or LMP7 can be pivotal. Several epitopes were produced inefficiently or
not at all in the absence of LMP2 (15, 16) or LMP7
(17, 18, 19, 20, 21, 22), but the opposite case has recently been reported
for several epitopes from tumor Ags that were destroyed when LMP2 and
LMP7 were expressed at high levels (23). Consistent with
these in vitro results, it was recently found in
LMP2-/- mice that this subunit has a pivotal
influence on the hierarchy of T cell epitopes from influenza virus
(24).
A further component of the proteasome system that has a pronounced
effect on Ag processing is the proteasome regulator PA28/. Both,
the and subunits of this ring-shaped regulator are inducible
with IFN- at both mRNA and protein levels (25).
PA28/ stimulates the various proteolytic activities of the
proteasome to different extents and markedly alters the fragmentation
of polypeptides through the 20S proteasome (8, 26, 27).
Overexpression of PA28/ in cell lines resulted in a higher
presentation of some (28, 29, 30), but not all, T cell
epitopes (22), suggesting a role for PA28/ in Ag
processing that was recently confirmed in
PA28-/- mice (31). A homologue
of PA28/ that is not transcriptionally induced by IFN- is
named Ki Ag or PA28. The single subunit of PA28 forms
heptahomomeric rings that associate with the proteasome and are
predominantly found in the nucleus. Interestingly, the PA28 protein
appears to be down-regulated upon the IFN- treatment of cell lines
in vitro and in transgenic mice that constitutively express IFN- in
the liver (32). PA28-/- mice
were slightly retarded in growth (33), but the function of
PA28 that causes this phenotype has remained elusive.
Taken together it appears that the inducible proteasome subunits LMP2,
LMP7, and MECL-1 as well as the proteasome regulator PA28/ are
not needed to guarantee the bulk flow of MHC class I ligands, but for
the generation of unique epitopes they often have a decisive impact.
Based on these insights, two hypotheses can be envisaged as to why the
immunoproteasomes and PA28/ coevolved with the MHC class I
pathway. It may be that these factors are induced to generate different
sets of proteasome complexes in a cell to enlarge the diversity of
peptide processing, which would allow the processing of a given
antigenic protein in several ways. For this scenario it should be
useful to express approximately equal amounts of constitutive
proteasomes and immunoproteasomes in an inflamed tissue. Alternatively,
it could be that CTLs are meant to be directed to a different set of
peptides in an inflammatory site compared with noninflamed tissues, and
this could only be achieved if constitutive proteasomes were largely
replaced by immunoproteasomes. We have investigated in this study the
steady state composition of 20S proteasomes of the liver on several
days during the course of an infection of the mouse with the
hepatotropic WE strain of LCMV and the bacterium Lm. Unexpectedly, the
infections led to virtually complete replacement of constitutive
proteasomes by immunoproteasomes and strong induction of PA28/ as
well as the disappearance of PA28 on days 7 and 8 of the infection,
when the cytotoxic immune response is maximal and pathogen numbers
decline. Our results suggest that the cytotoxic immune response in the
infected liver will be predominantly directed to a set of T cell
epitopes that are produced by immunoproteasomes, while the T cell
response is clearing the virus. This finding argues against the maximal
diversity argument and suggests that it must be beneficial for the
immune response to convert epitope production to a set of peptides that
is produced by immunoproteasomes and PA28/.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
C57BL/6 mice (H-2b) and BALB/c mice
(H-2d) were purchased from Institut für
Labortierkunde, Tierspital Zurich (Zurich, Switzerland), and kept in a
specific pathogen-free environment. Mice deficient for IFN-
(34), IFN-R (35), and perforin
(36) have been described previously. MC57 is a
C57BL/6-derived methylcholantrene-induced fibrosarcoma cell line
(37), and B8 is a BALB/c-derived SV40-induced fibroblast
line (38). H2.35 is a mouse hepatocyte cell line that we
obtained from American Type Culture Collection (Manassas, VA). All cell
lines were grown in IMDM or RPMI 1640 medium supplemented 10% FCS, 2
mM L-glutamine, and 100 U/ml penicillin/streptomycin.
Flow cytometry
For surface staining, aliquots of 5 x 105 cells in PBS and 2% FCS were incubated for 15 min in a round-bottom 96-well plate on ice with the mouse mAb KL25 specific for the LCMV glycoprotein (39). The cells were washed three times and subsequently stained by FITC-conjugated goat anti-mouse Ig secondary Ab (Silenus, Melbourne, Australia). After two washing steps the cells were analyzed on a FACScan flow cytometer (BD Biosciences, Mountain View, CA).
LCMV, Listeria, and infection of mice
LCMV-WE was propagated in the L929 fibroblast line, and viral
stocks were kept at -70°C (40). The virus was grown and
titrated on L929 cells exactly as previously described
(41). Mice were infected i.v. with 200 PFU (low dose) or
106 PFU (high dose) LCMV-WE. After the indicated
days the mice were sacrificed, and the livers were excised for
purification of 20S proteasomes and for Western analysis. Spleen and
liver samples were used for determination of LCMV titers, and
peripheral blood was used for IFN- ELISA. Serum levels of IFN-
were below detection levels on all days of infection except day 4, when
10–30 U/ml was found.
For Listeria infection, mice were infected i.v. with 9800 CFU Lm 10304S in PBS. Lm were cultured overnight at 37°C in brain-heart infusion broth. On the day of infection, 10-fold serial dilutions of the inoculum were plated on brain-heart infusion agar plates and incubated overnight at 37°C. The exact amount of bacteria injected the day before was calculated from these plates.
Purification of 20S proteasomes from mouse liver and fluorogenic assays
20S proteasomes from two or three mouse livers were purified and
quantitated as detailed previously (42). Hydrolytic assays
for proteasome activity employing fluorogenic assays were performed as
previously described (43). PA28/ complexes were not
detected in the proteasome preparations according to Western analysis
(data not shown).
Two-dimensional gel electrophoresis
For
Figs. 2–5
aliquots of 60 µg purified 20S proteasomes were
separated on two-dimensional nonequilibrium pH gradient gel
electrophoresis (NEPHGE)/SDS-PAGE or isoelectric focusing/SDS-PAGE
exactly as previously described (44). After the run, the
gels were stained with Coomassie brilliant blue R250. For quantitation,
the gels were scanned and analyzed using AIDA software (Fuji, Tokyo,
Japan).
|
|
|
|
H2.35, B8, or MC57 cells were left untreated or were infected
with LCMV-WE at a multiplicity of infection of 0.01. One day after
infection cells were incubated with 100 U/ml mouse rIFN- for 16
h or were left unstimulated. LCMV infection was confirmed flow
cytometrically by LCMV glycoprotein expression on the cell surface, and
IFN- stimulation was confirmed by up-regulation of H-2 class I
expression. Cells were starved in methionine/cysteine-free RPMI 1640
and 10% dialyzed FCS with or without rIFN- for 30 min, labeled for
6 h with 0.2 mCi/ml Trans 35S label (ICN,
Eschwege, Germany), and chased for 18 h with or without rIFN-.
The cells were harvested by trypsinization and lysed by sonication in
buffer A (25 mM Tris (pH 7.6), 2 mM MgCl2, 17%
glycerol, 1 mM DTT, and 2 mM ATP). The postnuclear lysates were counted
for 35S incorporation, and equal aliquots were
used for immunoprecipitation. The lysate was precleared for 1 h
with protein G-Sepharose CL-4B (Pharmacia, Uppsala, Sweden), followed
by immunoprecipitation with an anti-proteasome serum
(30) bound to protein G-Sepharose for 4 h at 4°C.
The precipitates were washed with buffer A, separated by
NEPHGE/SDS-PAGE, and visualized on x-ray films by autoradiography.
Western blot analysis and sucrose gradient density centrifugation
In Fig. 6 liver specimens were lysed in buffer B (50 mM Tris-HCl
(pH 7.5), 5 mM MgCl2, 1 mM EDTA, 1% SDS, 0.75
µM aprotinin, 10 µM leupeptin, 2.8 µM pepstatin, and 0.85 mM
PMSF) in a Dounce homogenizer (Kontes, Vineland, NJ), and 100 µg
total protein/lane was separated by SDS-PAGE. In Fig. 2
D
liver samples from C57BL6 mice on day 8 of LCMV infection were lysed in
buffer C (25 mM Tris-HCl (pH 7.5) and 2 mM MgCl2)
in a Dounce homogenizer with subsequent sonication. Postnuclear
supernatants were loaded onto a 10–40% sucrose gradient and
centrifuged at 40,000 rpm in a Beckman SW40Ti rotor (Palo Alto, CA) for
16 h. Gradient fractions of 0.6 ml were drawn, proteasome activity
was determined to confirm proper separation, and proteins were
separated on SDS-PAGE. The proteins were blotted onto nitrocellulose
membrane (Schleicher & Schuell, Dassel, Germany), blocked with PBS/5%
(w/v) low fat dry milk/0.1% Tween 20 for 1 h, and agitated
overnight at 4°C with the described rabbit antisera specific for
peptides PA28 and PA28 (45). Rabbit antisera for
mouse LMP7 and PA28 were raised against the keyhole limpet
hemocyanin-coupled peptides SDVSDLLYKYGEAAL and ILLTNSHDGLDGPTYK,
respectively. The MC3 Ab has been described previously
(46). The blots were washed and incubated for 1 h
with HRP-conjugated secondary Ab. After extensive washing with
PBS/0.1% Tween 20, proteins were visualized on x-ray films by enhanced
chemiluminescence.
|
Total RNA was extracted from mouse liver tissue using TRIzol
reagent (Life Technologies, Basel, Switzerland). cDNA was synthesized
from 3 µg RNA using 2 mM dT15, 2 mM dNTPs, 0.6 U/µl RNasin
(Promega, Wallisellen, Switzerland), 200 U Moloney murine leukemia
virus RT (Promega)/reaction (1 h at 42°C). After cDNA synthesis, the
nucleic acids were precipitated in 0.3 M NaCl/80% ethanol, and the
pellet was washed twice in 70% ethanol. The samples were resuspended
in H2O. Real-time PCR was performed with a Light
Cycler (Roche, Basel) using the DNA Master SYBR Green I reaction mix
(Roche) and 0.5 pmol/µl of each primer in 3.5 mM
MgCl2. Sense and antisense primers used for PCR
amplification were (annealing temperatures (a.t.) given in
parentheses): mouse (mo) LMP2, 5'-TCCACACCGGGACAACC-3' and
5'-CCAGCCAGCTACTATGAGATGC-3' (a.t. = 62°C); moDELTA,
5'-TCGAGTGACTGACAAGCTGACC-3' and 5'-GAACAGAGTACACCTGCCCTCC-3' (a.t.
= 62°C); moLMP7, 5'-CTCCGTGTCTGCAGCATCC-3' and
5'-TCCACTTTCACCCAACCGTC-3' (a.t. = 58°C); moMB1,
5'-CTTGACGGAACCACCACC-3' and 5'-CCATAGACAGCCCCATGC-3' (a.t. = 58°C);
moTNF-, 5'-GCACAGAAAGCATGATCC-3' and
5'-TGTCCCTTGAAGAGAACC-3' (a.t. = 62°C); and mouse hypoxanthine
phosphoribosyltransferase (HPRT): 5'-GCTGGTGAAAAGGACCTC-3' and
5'-CACAGGACTAGAACACCTGC-3' (a.t. = 60°C). The runs were
programmed as follows: denaturation for 30 s at 95°C;
amplification (40 cycles, with readings of the fluorescence at the end
of each cycle) for 1 s at 95°C, 10 s at a.t., and 20 s
at 72°C; and analysis of the products (reading of the fluorescence in
a continuous mode) for 0 s at 95°C, 62 to 95°C transition with
0.1°C increment/sec. For mouse IFN- we used a conventional PCR mix
containing the PCR primers moIFN (5'-GACAATGAACGCTACACACTGC-3' and
5'-GGACAATCTCTTCCCCACC-3'; a.t. = 58°C) and the fluorescence
resonance energy transfer primers
(5'-TGCCAAGTTTGAGGTCAACAACCCACA-fluorescein-3' and
5'-LCred640-GTCCAGCGCCAAGCATTCAATGAGC-3'); reading of fluorescence was
performed after annealing. Analysis was performed with Light Cycler
software 3. The value calculated by the quantification analysis was
always within the range covered by three concentrations of sample that
were taken as arbitrary units to construct the standard curve for
linear regression with r = 1.0 and p
< 0.0001. The cDNAs were normalized to HPRT mRNA content. The
specificity of the amplification was checked with melting curve
analysis of the products as well as analysis on an agarose
gel.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-mediated induction of immunoproteasomes
To investigate how the steady state composition of 20S proteasome
subunits changes during the course of a viral infection it was
important to choose a virus that infects an entire organ in a
well-documented manner and reproducible time course. The infection of
the mouse with the hepatotropic strain WE of LCMV is a suitable model
system to address this question, because this virus readily replicates
in hepatocytes, Kupffer cells, fibroblasts, endothelial cells, and
other cell types of the liver, and it rapidly spreads throughout the
organ (47, 48, 49). Moreover, LCMV-WE elicits a strong and
well-characterized CTL response that is maximal on day 8 of infection
and leads to elimination of the virus. However, as some viruses
interfere with the induction of immunoproteasomes (50, 51), we decided to initially test whether this was also the case
for LCMV-WE. The mouse hepatocyte line H2.35 as well as B8 or MC57
fibroblasts were infected with LCMV-WE and after 1 day were cultivated
in the presence or the absence of IFN- for 16 h before
metabolic labeling. Proteasomes were immunoprecipitated, and the
subunit composition was analyzed by two-dimensional NEPHGE/SDS-PAGE
(Fig. 1A). Inductions of LMP2
or LMP7 were not observed in H2.35 cells (Fig. 1) or in B8 or MC57
cells (data not shown) upon LCMV-WE infection, although the cells were
productively infected, as documented by flow cytometric analysis of
LCMV glycoprotein expression on the cell surface (data not shown).
Moreover, an interference of LCMV-WE with the IFN- mediated
induction of LMP2 and LMP7 was not observed, which was a prerequisite
for monitoring their in vivo induction in infected mice.
|
Mice of the C57BL/6 strain were infected i.v. with a high dose
(106 PFU) of LCMV-WE, and two livers each from
uninfected mice and mice on days 2, 4, 8, 12, and 25 after infection
were used to purify 20S proteasomes. The subunits of the 20S
proteasomes were separated by two-dimensional NEPHGE/SDS-PAGE, and the
proteins were visualized by Coomassie staining. This method allows a
reliable quantification of the steady state amounts of proteasome
subunits and the relative intensity changes during subunit
replacements. The results shown in Fig. 2A reveal that uninfected mice
express very little LMP2 and LMP7 in the liver compared with the
constitutive subunits and MB1. The protein spots for these four
subunits as well as the constitutive and invariant -type subunit C6
were quantified by densitometry, and the data are shown in Fig. 2B. On day 2 postinfection the inducible subunits had
already begun to rise, and by day 4 LMP2 and were equally
prominent, whereas the LMP7 spots were already twice as intensive as
that for MB1. On day 8 of infection, when virus titers start to decline
(Table I) and the cytotoxic immune
response and the expansion of LCMV-specific CD8+
T cells are usually maximal (47, 52), we observed an
almost complete replacement of by LMP2 and of MB1 by LMP7. The
8-fold reduction of MB1 and the 4-fold reduction of by day 8 of
infection were unexpected, as the constitutive proteasome was reported
to possess an in vivo half-life of 12–15 days in rat livers
(53). The exchange of MC14 by MECL-1 on day 8 was analyzed
on isoelectric focusing-SDS-PAGE two-dimensional gels, because it was
not readily resolved by NEPHGE/SDS-PAGE and was found to occur to an
extent of only 50% (data not shown) as has been previously documented
for IFN--treated cells (44). On day 12 postinfection,
when virus titers were 100-fold reduced, the ratio of inducible to
constitutive subunits started to revert again, but even on day 25 of
infection the constitutive proteasome subunits and MB1 had not yet
reached preinfection levels.
|
, the mRNAs for LMP2 and LMP7 were
elevated by factors of 3.95 and 2.55 on day 8 of infection,
respectively, and declined thereafter, which correlated with the
incorporation of the respective proteins into proteasomes. Much less
expected was the well-reproducible decline in the mRNAs for the
constitutive subunits and MB1, which we have not observed in
previous studies when cells were treated with IFN- in vitro
(54). The factors responsible for this mRNA
down-regulation of and MB1 in vivo remain to be identified. To
examine whether IFN- and TNF-, which have been shown to enhance
LMP2 and LMP7 in vitro, are actually elevated during the infection we
quantified their mRNA from the same cDNA samples. The elevations of
IFN- and TNF- mRNAs on day 4 of infection were only 1.61- and
3.2-fold, respectively, thus raising some doubt as to whether such a
weak up-regulation can be sufficient to transcriptionally induce LMP2
and LMP7 or whether other cytokines are involved at that stage. On day
8 of infection, however, the 5.5-fold elevation of IFN- and the
43.0-fold elevation of TNF- are consistent with these cytokines
being responsible for immunosubunit induction at the peak of the CTL
response. The results were also congruent with our histological
examination, which revealed a massive infiltration of lymphocytes on
day 8, but not yet on day 4, of LCMV infection (data not shown) in
accordance with previous studies (47). Western analysis
was performed to determine the expression levels of LMP2/ and
LMP7/MB1 in lysates of liver samples throughout the course of LCMV
infection and fully confirmed the results obtained with
two-dimensional gels (Fig. 2C). The synthesis of LMP7
precursors on day 8 of infection was apparently so strong that not all
30-kDa precursors could be incorporated into assembling proteasomes, as
they were found on the top of sucrose gradients in contrast to the
-type subunit MC3 (Fig. 2D).
|
A). This
phenomenon was not dependent on infections with high doses of LCMV-WE,
as the subunit replacements were similar in the liver of BALB/c mice
that had been infected for 7 days with a low dose of LCMV-WE (200 PFU),
which is normally used as a physiological equivalent in this
experimental system. The quantitative analysis of the two-dimensional
gels from the high dose and low dose infections (Fig. 3B)
confirmed that the virus dose did not alter the rate of proteasome
subunit replacements, and the same result was obtained when C57BL/6
mice were infected with 200 PFU LCMV-WE (data not shown).
In vivo replacement of and MB1 by LMP2 and LMP7 is
largely dependent on IFN-, but only to a minor extent on IFN-
and perforin
Since IFN- is the cytokine that transcriptionally induces LMP2,
LMP7, and MECL-1 most vigorously (55), we decided to
determine the extent of subunit replacement in the liver of
IFN--deficient BALB/c mice (34) that had been infected
with a high or a low dose of LCMV-WE 7 days previously (Fig. 4). In the absence of IFN- the
induction of LMP2 and LMP7 was reduced by about 50% despite an
enhanced virus titer, indicating that a complete replacement of active
site proteasome subunits, as in the wild-type BALB/c mouse, could not
be obtained. Hence, other cytokines must be able to induce the 7-fold
induction of LMP2 and LMP7 observed in
IFN--/- mice. We tested the effects of
IFN- and - in IFN-R-/- mice and found
that the replacement of constitutive proteasomes by immunoproteasomes
in the liver was only slightly reduced. Apparently, type I IFNs do not
contribute much to the proteasome replacement in vivo, which is
consistent with a lack of LMP2 and LMP7 up-regulation after LCMV
infection in vitro that led to an induction of IFN-/ (data not
shown). Another gene-targeted mouse that we examined was the
perforin-/- mouse, because we wanted to
determine whether perforin-mediated destruction of LCMV-infected cells
in the liver by CTLs or NK cells would be required to promote the
removal of constitutive proteasomes and the replacement by
immunoproteasomes during tissue regeneration. As shown in the
bottom panel of Fig. 4, perforindeficiency affected
the replacement of constitutive proteasomes in the liver to only a
minor extent, suggesting that perforin-mediated cell death contributed
only marginally to these replacements.
Infection of mice with Listeria leads to an extensive replacement of constitutive proteasomes by immunoproteasomes in the liver
We investigated whether the proteasome replacements in the liver
during LCMV infection was a virus-specific phenomenon or whether an
immune response against bacteria would have a similar consequence.
BALB/c mice were infected i.v. with 9800 CFU of the Lm strain 10304S,
and livers were removed on days 3 and 7 postinfection for isolation of
20S proteasomes and analysis on Coomassie-stained two-dimensional gels.
As depicted in Fig. 5, Lm infection led
to a comparable exchange of constitutive proteasomes by
immunoproteasomes as had previously been observed with LCMV
infection. The onset of proteasome replacements appeared to be very
early after Lm infection, as it was more pronounced on day 3 of Lm
infection than on day 4 after LCMV infection, a finding that was
consistently obtained in both BALB/c (Fig. 5) and C57BL6 (data not
shown) mice. Taken together, the extensive proteasome exchanges
observed during the course of LCMV infection do not seem to be
a solitary phenomenon confined to this virus, but may be extended to
bacteria and possibly other pathogens.
PA28 and are up-regulated, whereas PA28 is
down-regulated, in the liver of LCMV-infected mice
Since PA28/ is known to be transcriptionally up-regulated by
IFN- and since PA28 was reported to be down-regulated by IFN-
by a to date uncharacterized post-transcriptional mechanism, we decided
to follow the protein levels of PA28-, -, and - by Western
analysis using peptide antisera against these three subunits of
proteasome regulators. Liver tissues were obtained from C57BL/6 mice
that were either uninfected or infected with a high dose of LCMV-WE on
the indicated days before analysis (Fig. 6). The Western blots revealed that the
amounts of PA28 and - increased by a factor of 3.5 from day 0 to
day 8 of infection and declined thereafter. As we have observed for the
induction of LMP2 and LMP7, the induction was already visible on day 3
after LCMV infection and reached a peak when the CTL response is
expected to be maximal on day 8. Interestingly, the level of PA28
protein was strongly reduced during days 4, 8, and 15 of infection
compared with that in uninfected mice, which has not previously been
documented for a viral infection. In contrast to the induction of LMP2
and LMP7 that occurred at least to some degree in
IFN--/- mice, neither the induction of
PA28/ nor the disappearance of PA28 occurred on day 7 after
LCMV infection in the absence of IFN-, indicating that other
cytokines cannot compensate for the lack of IFN- in this
respect.
The peptidolytic activity of liver 20S proteasomes is strongly altered during the infection with LCMV
To address what consequences the proteasome subunit replacements
during LCMV infection have on proteasome activity, we determined the
activity of purified 20S proteasomes from livers of uninfected BALB/c
mice as well as livers of LCMV-infected BALB/c wild type and
IFN--/- mice 7 days postinfection in
peptidolytic assays. The respective 20S proteasomes, which were the
same as those analyzed on the two-dimensional gels in Figs. 3 and 4,
were incubated with the fluorogenic peptide substrates
(Z)-LLE--nitroanilide (NA), Suc-LLVY-7-amido-4-methylcoumarin
(MCA), (Z)-GGL-MCA, and Bz-VGR-MCA, and the release of the fluorogenic
leaving groups NA or MCA was measured at different substrate
concentrations after 30, 60, and 90 min. The 60-min values of the
linear and very reproducible reaction are shown in Fig. 7 and indicate that the cleavage of the
(Z)-LLE-NA substrate at the C terminus of glutamic acid, which is
frequently used to determine the peptidylglutamyl peptide-hydrolyzing
or caspase-like activity of the proteasome, was almost completely lost
on day 7 of LCMV infection in wild-type mice and was reduced by
approximately 50% in IFN--/- mice.
According to our previous transfection experiments (8, 10), this dramatic change in the caspase-like activity must be
due to the replacement of by LMP2, which is almost quantitative in
liver proteasomes on day 7 post-LCMV infection. The trypsin-like
activity, as determined by hydrolysis of the Bz-VGR-MCA substrate, was
slightly reduced in LCMV-infected mice, whereas the chymotrypsin-like
activity measured by hydrolysis of (Z)-GGL-MCA was not significantly
different among the three proteasome populations. The hydrolysis of the
Suc-LLVY-MCA substrate, in contrast, was markedly reduced in liver
proteasomes of LCMV-infected mice, as we had observed consistently in
cell lines after IFN- treatment and in cell lines transfected
with LMP2, LMP7, and MECL-1 (7, 8, 10). Together, the
results obtained corresponded very well with data obtained in
LMP2/LMP7/MECL-1 triple transfectants (56), indicating
that these marked changes in proteasome activity are probably
attributed to the almost complete replacement of constitutive
proteasomes by immunoproteasomes in the liver of LCMV-infected
mice.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
/
is markedly enhanced, while the PA28 regulator disappears. The
extent of 20S proteasome replacement in the liver is striking and
indicates that immunoproteasomes can take over the numerous
housekeeping functions of constitutively expressed proteasomes at least
for a few days. In the priming phase of the cytotoxic immune response 4
days after LCMV infection, the proteasome population in the liver
consists of about equal amounts of immunoproteasomes and constitutive
proteasomes, which may enhance the diversity of Ag processing. In the
main effector phase of the CTL response on day 7 or 8 after infection,
however, the balance is shifted toward immunoproteasomes, suggesting
that the recognition of infected cells and the clearance of the
pathogen rely on the presentation of epitopes that are produced by
immunoproteasomes under an increased influence of the PA28/
regulator.
The short time period of 7 days in vivo until the constitutive
proteasomes containing the active site subunits and MB1 disappear
is unexpected, since 20S proteasomes were shown to have an
extraordinarily long half-life of 12–15 days in the liver of
metabolically labeled rats (53). Since the replacement of
active site subunits in the proteasome population cannot occur in
preformed and matured 20S proteasomes (55), it relies on
the de novo assembly of immunoproteasomes and the relative loss of
constitutive proteasomes. Because of the long half-life of 20S
proteasomes, we expected that immunoproteasomes would at best add a
population to the pre-exisiting constitutive pool until the virus is
cleared. A possible explanation for the rapid elimination of
housekeeping proteasomes in the inflamed liver would be a massive
destruction of LCMV-infected hepatocytes by virus-specific CTLs.
Indeed, serum levels of liver enzymes such as alanine aminotransferase
or aspartate aminotransferase were shown to transiently rise by a
factor of 10–20 in LCMV-WE-infected mice, indicating the lysis of
hepatocytes (47). However, based on the enhancement of
alanine aminotransferase levels, it has been estimated that only 5% of
hepatocytes are lysed during an acute immune response to LCMV-WE
(49) in accordance with the small effect that perforin
deficiency had on proteasome replacements in the liver (Fig. 4). A
histological assessment of LCMV infection by in situ hybridization
indicated that only 5–10% of hepatocytes contained the viral mRNA on
day 5 after infection, whereas all Kupffer cells were infected
(49). Even if the destroyed tissue would be immediately
replaced by new hepatocytes, a cellular turnover of 10% within 7 days
would not suffice to account for a complete proteasome exchange if the
half-life of constitutive proteasomes in the liver is 12–15 days. It
will therefore be interesting to determine the in vivo half-life of
constitutive proteasomes in the liver of LCMV-infected and uninfected
mice to test whether an accelerated degradation of constitutive 20S
proteasomes may occur in inflamed organs. An alternative contribution
to the change in proteasome composition could be the infiltration of
the liver with leukocytes that possess roughly equal amounts of
immunoproteasomes and constitutive proteasomes in the uninduced state.
The total lymphocyte number in the liver increases by a factor of 100
from day 4 until day 6, adding up to a total of 20 million
lymphocytes/liver throughout days 6–10 of infection (47).
However, even in the peak phase of the immune response the cell number
of lymphocytes is about 10-fold lower than the number of the much
larger hepatocytes and hence contributes only marginally to the total
proteasome pool of the liver.
Our experiments performed in IFN--deficient mice showed that the
induction of PA28/ as well as the disappearance of PA28 relied
entirely on IFN- production in the LCMV-infected liver. While a
vigorous induction of PA28/ by IFN- could be predicted from
previous in vitro data (25), the disappearance of PA28
after IFN- treatment has not been observed in all cell lines (M.
Groettrup, unpublished observations) and is still a matter of debate.
Our in vivo data clearly support this phenomenon first observed by
Tanahashi et al. (32) and pose the question of its
functional implication. One could imagine that PA28 competes with
PA28/ for binding to the proteasome, which may adversely affect
Ag processing, and we are currently addressing this possibility in our
laboratory. Also, the mechanism that leads to the down-regulation of
PA28 protein during LCMV infection needs to be investigated, because
mRNA levels of PA28 are not affected by IFN- treatment of cells
in vitro (25). In contrast to the alteration of PA28
regulators, the replacement of , MB1, and MC14 by LMP2, LMP7, and
MECL-1 did occur in IFN--deficient mice, albeit at a 50% reduced
level. Thus, IFN- is required for a full replacement of constitutive
proteasomes by immunoproteasomes, but other factors must be able to
cause a half-maximal exchange. TNF- is a prime candidate to mediate
this effect, as it is known to induce LMP2 and LMP7 in vitro
(55), and since TNF- mRNA has been found in the liver 2
days after LCMV-WE infection (49). The cells that produce
IFN- and TNF- during the first 4 days of infection are most
likely NK cells, NKT cells, and macrophages, respectively. NK cells are
found in 4-fold greater number between days 1–5 of LCMV infection in
the liver (57). However, our real-time RT-PCR analysis
revealed that the enhancement of TNF- and IFN- mRNAs was not very
prominent in the liver of C57BL/6 mice on day 4 after LCMV-WE infection
(Table II), which could indicate that other cytokines may also
contribute to the early induction of immunoproteasomes. Beginning on
day 5 of infection, LCMV-specific CTLs and Th cells can be found in the
liver (data not shown) (47), which are likely to account
for the strongly elevated levels of IFN- and TNF- observed on day
8 of infection.
Another class of cytokines that is immediately induced after LCMV
infection in vivo is IFN-/ (58, 59). LCMV is a (-)
strand RNA virus that replicates via dsRNA intermediates known to be
potent inducers of IFN-/. Nevertheless, we could not observe an
induction of PA28 (data not shown) or LMP2 and LMP7 after productive
infection of a fibroblast line, although the induction of IFN- could
be demonstrated by RT-PCR after LCMV-WE infection in vitro (K. Schwarz,
unpublished observations). This indicates that a cell-autonomous
production of IFN-/ does not effectively induce immunoproteasomes
and that other cytokines, such as IFN- and TNF-, are required. In
accordance with this result, a deficiency of the IFN-R only
marginally affected proteasome exchange in mouse liver after LCMV
infection (Fig. 4).
What consequences does virtually complete replacement of constitutive
proteasomes by immunoproteasomes have on our understanding of the
immune response to LCMV in particular and of the functions of LMP2,
LMP7, and MECL-1 induction in general? Although the CTL response
against LCMV has not yet been investigated in mice that are deficient
for LMP2 or LMP7, it is already clear from studies in
IFN-R-/- mice that the IFN--mediated
induction of these proteasome components is not required for
elimination of the virus, although virus titers were elevated and the
CTL response reduced when the IFN-R was genetically ablated
(35) or when IFN- was neutralized by Abs in vivo
(60, 61). With respect to the generation of LCMV epitopes,
we have recently shown that the transfection of fibroblasts with LMP2,
LMP7, and MECL-1 results in a markedly enhanced presentation of the
H-2Ld-restricted nucleoprotein epitope
nucleoprotein 118 and that purified immunoproteasomes produced 6-fold
more precursors of this epitope compared with constitutive proteasomes
in vitro (22). Recently, we have investigated the
generation of the H-2Db-restricted glycoprotein
epitopes GP33 and GP276 and found that IFN- treatment of fibroblasts
led to a 4-fold enhanced presentation of the immunodominant GP33
epitope, whereas the subdominant GP276 was 3-fold down-regulated (K.
Schwarz, unpublished observations), and we are currently investigating
whether these changes can be attributed to alterations in proteasome
composition.
If immunoproteasomes are not required for the bulk production of MHC
class I ligands and if some epitopes are dependent on LMP2, LMP7, and
MECL-1, while others, including important tumor epitopes, are
destroyed, why, then, would an inducible expression of immuoproteasomes
occur in sites of viral infection? The surprising finding of an almost
complete replacement of constitutive proteasomes by immunoproteasomes
suggests that rather than increasing the diversity of proteasome
populations, immunoproteasomes serve to generate different T cell
epitopes in inflamed as opposed to uninflamed tissues, which normally
express extremely low levels of LMP2, LMP7, and MECL-1
(54). We hypothesize that the replacement of proteasomes
in inflamed tissues serves to focus the immune response on T cell
epitopes that are preferentially or exclusively made by
immunoproteasomes. This change in epitope production may also
contribute to avoid autoimmune assaults if different peptide epitopes
are processed from endogenous housekeeping genes in uninflamed sites as
opposed to sites of viral infection. T cells that recognize an
immunoproteasome or PA28/ - dependent peptide from a household
protein in an inflammatory site may not find the same peptide in
uninflamed tissues and hence would not cause tissue damage in
inappropriate sites.
|
Acknowledgments |
|---|
-deficient mice, and
Peter-M. Kloetzel, Ulrike Kuckelkorn, Alice Sijts, and Klavs
Hendil for contributing Abs. We are grateful to Nathalie Oetiker and
Hans Christian Probst for help with LCMV infection of mice. |
Footnotes |
|---|
2 S.K., M.v.d.B., and K.S. contributed equally to this work and are all considered first authors.
3 Address correspondence and reprint requests to Dr. Marcus Groettrup, Kantonsspital St. Gallen, LFA, Haus 09, CH-9007 St. Gallen, Switzerland. E-mail address: marcus.groettrup{at}kssg.ch
4 Abbreviations used in this paper: LCMV, lymphocytic choriomeningitis virus; a.t., annealing temperature; ER, endoplasmic reticulum; HPRT, hypoxanthine phosphoribosyltransferase; Lm, Listeria monocytogenes; NEPHGE, nonequilibrium pH gradient gel electrophoresis; NA, -nitroanilide; MCA, 7-amido-4-methylcoumarin; moLMP, mouse LMP.
Received for publication April 3, 2001. Accepted for publication October 17, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-Interferon and expression of MHC genes regulate peptide hydrolysis by proteasomes. Nature 365:264.[Medline]
and major histocompatibility complex-encoded subunits on peptidase activities of human multicatalytic proteases. Proc. Natl. Acad. Sci. USA 92:584. stimulation modulates the proteolytic activity and cleavage site preference of 20S mouse proteasomes. J. Exp. Med. 179:901.-inducible 11S regulator (PA28) and the LMP2/LMP7 subunits govern the peptide production by the 20S proteasome in vitro. J. Biol. Chem. 270:23808.. Eur. J. Immunol. 25:2605.[Medline]
2 (MECL-1) demonstrates conservation of cleavage specificity between yeast and mammalian proteasomes. FEBS Lett 454:11.[Medline]
exposes a cryptic cytotoxic T lymphocyte epitope in HIV-1 reverse transcriptase. J. Immunol. 162:7075./ enhances the presentation of an immunodominant lymphocytic choriomeningitis virus T cell epitope. J. Immunol. 165:768.-interferon-inducible protein activator of the 20S proteasome. FEBS Lett. 366:37.[Medline]
in antigen presentation. Nature 381:166.[Medline]
/ can enhance antigen presentation without affecting 20S proteasome subunit composition. Eur. J. Immunol. 30:3672.[Medline]
-interferon regulation. Genes Cells 2:195.[Abstract]
. J. Biol. Chem. 274:38211. genes. Science 259:1739.-induced subunit exchange in the 20S proteasome. Eur. J. Immunol. 26:863.[Medline]
, MB1, and C6 are indicated; the
two spots assigned to LMP7 were confirmed by Western blotting (not
shown). B, Densitometric evaluation of indicated
proteasome subunits from the two-dimensional gels shown in
A. C, Western analysis of the proteasome
subunits MB1, LMP7, 
