| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Divisions of Cytokine Biology and Cellular and Gene Therapies, Center for Biologics Evaluation and Research, Food and Drug Administration, Rockville, MD 20852
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
for expression. To determine
whether post-transcriptional aspects of IL-12 production might be
regulated, we examined intracellular protein processing of each
subunit. We report here that p40 and p35 subunits are processed by
disparate pathways. Whereas processing of p40 conforms to the
cotranslational model of signal peptide removal concomitant with
translocation into the endoplasmic reticulum (ER), processing of p35
does not. Translocation of the p35 preprotein into the ER was not
accompanied by cleavage of the signal peptide; rather, removal of the
p35 signal peptide occurred via two sequential cleavages. The first
cleavage took place within the ER, and the cleavage site localized to
the middle of the hydrophobic region of the signal peptide. Although
the preprotein was glycosylated upon entry into the ER, its
glycosylation status did not affect primary cleavage. Subsequently, the
remaining portion of the p35 signal peptide was removed by a second
cleavage, possibly involving a metalloprotease, concomitant with
additional glycosylation and secretion. Secretion could be inhibited by
mutation of the second cleavage site or by inhibition of glycosylation
with tunicamycin. In contrast, p40 secretion was not affected by
inhibition of glycosylation. Our findings demonstrate that IL-12
subunits are processed by disparate pathways and suggest new modalities
for regulation of IL-12 production. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To better understand the regulation of IL-12 production during immune responses, we investigated the intracellular protein processing of each of the IL-12 subunits. Signal peptides target proteins for secretion and are thought to direct transport across the endoplasmic reticulum (ER)4 membrane by a cotranslational mechanism (reviewed in Ref. 16). Signal peptides have a common structure comprised of a short, positively charged amino-terminal region (n-region), a central hydrophobic region of approximately 10 aa (h-region), and a more polar carboxyl-terminal region of 4–6 aa (c-region) that is typically the site of cleavage by signal peptidase (11). All the information necessary to direct processing of the signal peptide by signal peptidase is thought to be contained within the signal peptide (12). In eukaryotes, but not prokaryotes, partial glycosylation may occur concomitant with translocation. The functional activity of signal peptides may be examined in vitro using a combination of protein translation in the presence or the absence of microsomes that perform many, if not all, of the translocation and cleavage functions of the ER in vitro. Alternatively, cDNA constructs may be transfected transiently into cells in vitro and monitored for the presence of the protein encoded by the cDNA within the cell or secreted into the culture media. Both of these approaches were employed to evaluate the post-translational processing and secretion of IL-12 p35 and p40 proteins. Key aspects of the process were also tested by examining the stable processing intermediates of the native gene products in monocytes/macrophages following gene induction of IL-12. Our results indicate that while p40 follows this model, the p35 preprotein does not function as a typical secretory peptide as would be predicted by its amino acid sequence (12). Instead, processing of the p35 secretory signal peptide involves two sequential cleavages, which occur after migration of the intact preprotein into the ER. The first cleavage is located within the hydrophobic portion of the predicted signal peptide. The second cleavage occurs at a later stage, possibly coupled to secretion and complex glycosylation, and releases the amino-terminus of the secreted p35 protein. Discrimination of this alternate pathway is not confined solely to the p35 signal sequence, but appears to also be influenced by other regions of the p35 preprotein. Our results demonstrate that the IL-12 subunits are processed by disparate protein processing pathways and suggest new therapeutic intervention points for modulation of IL-12 production.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Human peripheral blood monocytes were purified from single-donor leukapheresis preparations by centrifugal counterflow elutriation. Monocytes were <95% pure by Giemsa and nonspecific esterase staining. Cells were cultured in six-well tissue culture plates (Costar, Cambridge, MA) in RPMI 1640 supplemented with L-glutamine, gentamicin sulfate, HEPES, and 10% FBS (Life Technologies, Gaithersburg, MD). All culture reagents were tested and were free of detectable endotoxin.
COS-7 cells were obtained from American Type Culture Collection (Manassas, VA) and grown in DMEM (Life Technologies) containing 10% FCS, L-glutamine (2 mM), nonessential amino acids (0.1 mM), and gentamicin (0.1 mg/ml) in an atmosphere of 6% CO2.
Complementary DNA constructs
The cDNAs of both human p35 and p40 in Bluescript (Stratagene,
La Jolla, CA) were provided by U. Gubler (Hoffmann-La Roche, Nutley,
NJ; GenBank accession no. M65271 and M65272). Two forms of the p35 open
reading frame (ORF) were constructed, representing each of the two
potential forms of the signal peptide fused to the ORF encoding the
secreted portion of p35. The first, termed
p35-57, contains two potential initiator
methionines at positions -57 (M-57) and -22
(M-22) relative to the first amino acid (R) of
the secreted form of p35. This form corresponds to the ORF described in
lymphoblastoid cells (13). The second, termed
p35-22, corresponds to ORF resulting from
transcription arising from the TATA-like element described for
IFN-/LPS-stimulated monocytes/macrophages (14). This
latter form was constructed by removing nucleotides 1–156 from the p35
cDNA. Both constructs were excised from Bluescript by
NotI/KpnI treatment and ligated into pcDNA3.1(+)
(Invitrogen, San Diego, CA), which had been digested with
NotI/KpnI. In vitro transcription was performed
from the T7 promoter. The p40 cDNA was excised from Bluescript by
XhoI/NotI treatment and ligated into pcDNA3.1(+)
following treatment with XhoI/NotI.
Site-directed mutagenesis of the p35 secretory peptide leader sequence
Mutations within the signal peptide (Table II
) were constructed
using PCR-mediated site-directed mutagenesis (15) and
p35-57 as template. Briefly, the left PCR
product was generated with the forward priming
p35-57-specific oligonucleotide gene-specific
primer (GSP) 61 (5'-AGCGGTACCTTATAAAAATGTGGCCCCCT-3'; sequence
complementary to bases 60–79 and containing a novel KpnI
site) and the reverse GSP containing the base change for the particular
mutant (see Table I). The right PCR
product was generated using the corresponding forward gene-specific
primer (Table I) and vector-specific primer (VSP) 59
(5'-ACGGGCCCTCTAGACTCGAGCGGCCGC-3'; complementary to bases 978-1004
of pcDNA3.1(+)). To generate the complete sequence, corresponding left
and right PCR products were gel purified, added in equimolar amounts to
a PCR reaction, annealed to each other, and then amplified using the
external oligonucleotides, GSP61 and VSP59, and Advantage Klentaq
polymerase (Clontech, Palo Alto, CA). The resultant PCR products were
gel purified, digested with KpnI/NotI, and
ligated into pcDNA3.1(+), which had been linearized with
KpnI/NotI. This was then transfected in DH5
cells (Life Technologies), selected, and purified using Plasmid Maxi
kits (Qiagen, Santa Clarita, CA). A double mutant
(L-10
R/R+1G) was
generated using the primers listed to generate mutant
L-10R, but the template was
R+1G. All DNA manipulations were confirmed by
DNA sequence analysis using Sequenase version 2.0 (Amersham, Arlington
Heights, IL).
|
|
Plasmids were linearized with NotI (p35) or XhoI (p40). The RNA was transcribed in vitro with T7 polymerase according to instructions supplied by the manufacturer (Stratagene, La Jolla, CA). Rabbit reticulocyte lysate and canine pancreatic microsomes were obtained from Promega (Madison, WI) and used to translate RNA in the presence of 20 µCi of [35S]methionine as recommended by the manufacturer. Proteinase K treatment of translation reactions was performed at a final concentration of 0.1 mg/ml (with added 10 mM CaCl2) for 45 min on ice with or without 0.1% Triton X-100. The reaction was terminated by addition the protease inhibitor 4-(2-aminoethyl)-benzenesulfonyl-fluoride to 10 mM.
Transient transfection in vitro
COS-7 cells were transfected with Lipofectin reagent (Life
Technologies) according to the manufacturer’s instructions, and
48 h later cells were labeled with 100 µCi of
[35S]methionine for the indicated time periods.
Labeled cultures were washed twice in PBS and extracted with lysis
buffer (150 mM NaCl, 50 mM Tris (pH 7.5), 50 mM NaF, 0.5 mM
Na4P2O7,
and 1% Triton X-100) supplemented with the protease inhibitors
4-(2-aminoethyl)-benzenesulfonyl-fluoride (1 mM; ICN Biomedicals,
Aurora, OH), aprotinin (2 µg/ml; Sigma, St. Louis, MO), leupeptin (1
µg/ml; Sigma), and pepstatin A (1 µg/ml; Sigma). In some
experiments, transfected COS-7 cells were treated with the metabolic
inhibitors tunicamycin (1 µg/ml; Roche, Indianapolis, IN) or benzyl-
2-acetamido-2-deoxy--D-galactopyranoside (2 mM, Sigma)
for 24 h before and during the
[35S]methionine labeling period.
Human elutriated monocytes were plated at a density of 3.3 x
106/ml and cultured overnight in the presence of
IFN- (100 ng/ml; Genentech, South San Francisco, CA). The following
morning LPS (1 µg/ml; catalogue no. L2755, Sigma) was added to the
cultures for 4 h. Cells were then washed in PBS, placed in
methionine-free culture medium containing LPS (1 µg/ml) and 100 µCi
of [35S]methionine, and incubated for either
4 h or overnight, at which time the cells and culture fluids were
harvested.
Immunoprecipitation
Immunoprecipitations were performed in final volumes of 500 µl. When necessary, in vitro translation mixtures were diluted to the final volume with lysis buffer. Cellular supernatants were supplemented with the above-mentioned protease inhibitors and concentrated using Centricon concentrators (Amicon, Beverly, MA). Samples were treated first with protein G-Sepharose beads before incubation with either mouse anti-human p35 Ab mixture or mouse anti-human p40 Ab (anti-p35, catalogue no. 20501D, anti-p40, catalogue no. 20512D, isotype control, catalogue no. 03171; PharMingen, San Diego, CA). Immunocomplexes were collected with protein G-Sepharose beads (Amersham-Pharmacia, Uppsala, Sweden), heated to 95°C for 5 min, and analyzed by 12% SDS-PAGE and autoradiography.
Glycosidase treatment
Immunocomplexes were resuspended in 50 µl of 0.2 M phosphate buffer (pH 6.0), containing 0.2% SDS and 2% 2-ME, and heated at 95°C for 5 min, then cooled on ice. After this, Nonidet P-40 (2%) and AEBSF (10 mM) *were added, and the samples were incubated overnight at 37°C individually or in combination with the following enzymes: 20 mU endoglycosidase (Endo) H (Roche), 250 mU Endo F/N-glycosidase F-free (Roche), 3 mU of O-glycosidase (Roche), and 100 mU neuraminidase (Sigma). Samples were then subjected to SDS-PAGE analysis following treatment for 5 min at 95°C in reducing sample buffer.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We first examined processing of the p40 signal peptide and found
that p40 processing conformed to the accepted model of signal peptide
function (Fig. 1). In vitro translation
of p40 RNA resulted in a preprotein of 42.5 kDa. Addition of microsomes
to the translation reaction resulted in a size reduction to
approximately 41.5 kDa, representing both cleavage of the signal
peptide as well as partial glycosylation, which was confirmed by
removal of carbohydrates with Endo H to yield a 39.5-kDa protein (Fig. 1A). Thus, approximately 3 kDa was removed from the nascent
p40 protein during translocation into the microsomes, consistent with
removal of the entire 22-aa signal peptide (13) by signal
peptidase. Transfection in COS-7 cells showed that the intracellular
and secreted forms were similar in size to the 41.5-kDa form observed
in vitro, suggesting that p40 was not subjected to complex
glycosylation (Fig. 1B). This was confirmed by subjecting
the intracellular and secreted forms of p40 to deglycosylation with
Endo H and Endo F. The initial stage of protein glycosylation within
the ER results in the addition of high mannose sugars that are
susceptible to removal by either Endo H or Endo F. However, once within
the Golgi, additional carbohydrate modifications may occur that result
in loss of susceptibility to Endo H, but not to Endo F. Both
intracellular and secreted forms of p40 were sensitive to
deglycosylation with Endo H, demonstrating that both contained only
noncomplex carbohydrates. Secreted p40 did not undergo complex
glycosylation, because the results with Endo F, which cleaves both
complex and noncomplex carbohydrates, were identical with those with
Endo H. When p40 transfected cells were treated with tunicamycin, an
inhibitor of lipid-linked oligosaccharide formation, a single band of
39.5 kDa was immunoprecipitated from intracellular and secreted,
indicating that glycosylation was not necessary for secretion (Fig. 1C). Taken together, these studies demonstrate that the p40
signal peptide is processed in vitro and in living cells consistent
with removal of the signal peptide by signal peptidase.
|
The p35 gene can be transcribed from one of two possible
promoters: either a CpG-like element functional in EBV-transformed
lymphoblastoid cells (13) or a downstream TATA-like
element following LPS stimulation of IFN--primed monocytes
(14). Depending upon which promoter is used, the resultant
ORF may contain either one or two initiator methionines (diagrammed in
Fig. 2a). Transcripts of p35
in IFN-/LPS-stimulated monocytes (p35-22)
contain only the second initiator methionine
(M-22) and are predicted to encode a typical
signal peptide of 22 aa and a predicted preprotein size of 25 kDa.
Lymphoblastoid cell-derived p35 mRNA (p35-57)
contains an additional in-frame initiator methionine
(M-57) encoding an additional 34 aa between
M-57 and M-22 with a
predicted preprotein size of 28 kDa. Both forms contain the same
predicted signal peptide sequence between M-22
and R+1. The functional significance of these two
potential forms of p35 is unknown at present (13).
|
B). RNA transcribed from p35-22 and
p35-57 cDNAs encoded proteins of either 30 kDa
(p35-57) or 27 kDa
(p35-22), which displayed apparent molecular
sizes on SDS-PAGE approximately 2 kDa larger than the predicted
theoretical sizes and suggested that the first initiator methionine
encountered was used by in vitro translation. Each form also functioned
similarly, in that comparable amounts of protein were produced from
equivalent amounts of mRNA. In the presence of microsomes, both
transcripts yielded additional products approximately 5 kDa larger.
These larger species were glycosylated forms of the proteins, since
they were susceptible to treatment with the glycosidase Endo H.
However, upon deglycosylation the proteins were identical in size to
their counterparts translated in the absence of microsomes, indicating
that the signal sequences were not removed from either
p35-57 or p35-22 during
translocation into the microsomes (Fig. 2B).
To determine whether failure to remove the signal peptides of p35
preproteins in vitro reflected integration into the membrane of the
microsome or failure to traverse into the lumen of the microsome, we
employed a proteinase K protection assay (Fig. 2C). If p35
protein were completely translocated across the microsomal membrane, it
should be protected from digestion by proteinase K by the lipid bilayer
of the microsome. If p35 were a transmembrane protein, only those
domains located within the lumen of the microsome vesicle should be
protected. Following treatment with proteinase K, the glycosylated
forms of form of p35-22 (32 kDa) and
p35-57 (35 kDa) were protected from proteolysis,
indicating that each form of the p35 secretory leader was able to
direct translocation into the lumen of the microsome. In contrast, the
unglycosylated forms of p35-22 (27 kDa) and
p35-57 (30 kDa) were not protected, indicating
that these forms were not translocated. This is not completely
unexpected, as translocation of proteins into microsomes by the in
vitro system is not quantitative. However, taken together, these
results demonstrated that both forms of the p35 signal peptide were
capable of directing translocation across the ER, but were not
concomitantly removed during this process.
To determine whether similar events occurred in living cells, COS-7
cells were transiently transfected with cDNA constructs encoding each
potential form of the preprotein, and the resultant proteins were
characterized (Fig. 3). Regardless of
whether cells were transfected with p35-57 or
p35-22, immunoprecipitates from cell lysates
contained a predominant species of 31 kDa rather than the 32-kDa
(p35-22) or 35-kDa
(p35-57) forms observed following in vitro
translation in the presence of canine microsomes. The different size
suggested that additional processing occurred within the endoplasmic
reticulum regardless of the form of p35 signal peptide used (Fig. 3A). A second band, approximately 5 kDa larger than the
predominant band, appeared in immunoprecipitates from
p35-57 transfected cell lysates. This 36-kDa
protein had a half-life of 30 min and disappeared over a prolonged time
period, whereas the 31-kDa species had a half-life of approximately
3 h regardless of whether it was encoded by
p35-57 or p35-22 (data
not shown). We interpret this finding as representing initiation of
protein synthesis from M-57 or
p35-57, which was then subsequently processed
within the ER to yield the more stable 31-kDa species. However, the
31-kDa intracellular form does not represent the secreted form, because
immunoprecipitated p35 protein from transfected COS-7 cell culture
supernatants contained a single diffuse band of 36 kDa (Fig. 3A) regardless of whether the cells were transfected with
p35-57 or p35-22,
demonstrating that additional processing events were required.
|
- and
LPS-stimulated monocytes were metabolically labeled with
[35S]methionine and immunoprecipitated from
cell lysates or culture medium supernatants with anti-p35 Ab (Fig. 3B). The stable intracellular form of p35 in
monocytes/macrophages was identical in size to the 31-kDa species
observed in COS-7 cells transfected with each cDNA. Likewise the
secreted form of p35 was 36 kDa. In addition, a 42.5-kDa protein
corresponding in size to p40 was also observed in anti-p35
immunoprecipitates from cell lysates, suggesting that
heterodimerization had occurred between the partially processed p35
preprotein and the fully processed p40. A small amount of the 31-kDa
species was also observed in the supernatant at 24 h, possibly due
to cell death/leakage during the prolonged metabolic labeling period.
Similar results were seen in COS-7 cells cotransfected with p35 and p40
(data not shown). Taken together, these data demonstrate that the
results obtained by transient transfection of cDNA in COS-7 cells were
reflective of IL-12 processing in monocytes/macrophages. The p35 complex glycosylation is coupled to terminal secretion
Our studies suggested that the most stable form of the p35 preprotein translocated into the ER was processed to a 31-kDa form, which was subsequently secreted as a 36-kDa protein regardless of the initiation methionine used (M-22 or M-57) or the cell type assayed. To clarify what role glycosylation played in this observation we determined the effects of the deglycosylation enzymes Endo H, Endo F, O-glycosidase, and neuraminidase on intracellular and secreted forms of p35. The initial stage of protein glycosylation within the ER results in the addition of high mannose sugars that are susceptible to removal by either Endo H or Endo F. However, once within the Golgi, additional carbohydrate modifications occur that result in loss of Endo H susceptibility, but not that of Endo F. O-linked glycosylation, which is thought to also occur within the Golgi compartment, is susceptible to removal by O-glycosidase. Addition of N-acetylglucosamine, which occurs after exiting the ER, but not before, is removed by neuraminidase.
When cell lysates from COS-7 cells transfected with each p35 form were
digested with either Endo H or Endo F, the stable intracellular 31-kDa
form of p35 was reduced in size to 25 kDa. No difference could be
discerned between the products resulting from treatment with either
Endo H or Endo F (Fig. 4A,
upper panel), consistent with the presence of only high
mannose sugars. In contrast, no change in the size of the stable 31-kDa
intracellular form ofp35 was observed when the immunoprecipitate was
digested with O-glycosidase and/or neuraminidase (data not
shown). Moreover, only the 25-kDa species was observed when cocktails
of these enzymes were used, and no difference was found from the use of
either Endo H or Endo F alone. These results confirm that the stable
intracellular 31-kDa form of p35 was subjected only to noncomplex
glycosylation and, based upon the sensitivity to Endo H, revealed that
this form resided within the ER and not within the Golgi. In contrast,
the secreted 36-kDa form p35 was resistant to Endo H (Fig. 4A, lower panel), suggesting that additional,
complex glycosylation occurred during later stages of the secretion
process. This was confirmed by treatment of secreted p35 with Endo F,
which resulted in a size reduction of approximately 10 kDa (from 36 to
26.5 kDa; Fig. 4A, lower panel). The 1.5-kDa
disparity between the intracellular and secreted deglycosylated forms
of p35 (25 vs 26.5 kDa, respectively) suggested that additional
post-translational protein processing may occur. Neuraminidase
treatment alone reduced the size of secreted p35 by approximately 3 kDa
(Fig. 4A). However, O-glycosidase treatment had
no effect either by itself (data not shown) or in a mixture with Endo F
and neuraminidase, as the observed reduction in size was the same as
that after treatment with Endo F alone (Fig. 4A, lower
panel). These studies demonstrated that the 31-kDa intracellular
form of p35 contained noncomplex carbohydrates, while the secreted form
contained complex, N-linked carbohydrates. However, since
the 31-kDa glycosylated species was sensitive to both Endo H and Endo
F, we concluded that this stable intracellular form of the preprotein
must reside within the ER, since complex glycosylated proteins become
insensitive to the effects of Endo H, but not Endo F, as they migrate
out of the ER and are modified by further glycosylation within the
Golgi (17).
|
B).
Benzyl
2-acetamido-2-deoxy--D-galactopyranoside, an
inhibitor of O-glycosylation, did not affect the size of the
31-kDa intracellular form or affect secretion, confirming the results
obtained with O-glycosidase. However, treatment with
tunicamycin, a metabolic inhibitor of N-glycosylation,
completely blocked secretion and resulted in the appearance of 29- and
24-kDa intracellular forms for p35-57 (Fig. 4B). In a parallel experiment only a single species of 24
kDa was observed for p35-22 (data not shown).
These findings suggested that the 24-kDa species resulting from
tunicamycin treatment most likely represented the unglycosylated
counterpart of the 31-kDa stable intracellular form of p35, while the
29-kDa species most likely represented the unglycosylated counterpart
to the transient 36-kDa species. The 1-kDa difference observed between
the 25-kDa Endo F-treated intracellular product and the 24-kDa product
after tunicamycin treatment may be due to the different modes of action
of these agents: removal of oligosaccharide additions by Endo F vs
prevention of the first step in lipid-linked oligosaccharide precursor
formation by tunicamycin. Together, these results demonstrated that 1)
p35 was a complex glycosylated protein; 2) complex glycosylation was an
integral part of the p35 secretory process; 3) primary cleavage of the
preprotein occurred regardless of the state of glycosylation; but 4)
subsequent protein processing and/or secretion required
glycosylation. Generation of the 31-kDa p35 species is due to cleavage within the hydrophobic (h-) region of the signal peptide
To date we have shown that the stable intracellular intermediate
in p35 biosynthesis is a 31-kDa glycoprotein that is observed with
p35-57 or p35-22 or in
monocytes/macrophages following induction by IFN- and LPS. To
exclude formally the possibility that the 31-kDa protein observed for
p35-57 in COS-7 cells might be due to
translation initiation arising internally at
M-22, mutational analysis was performed (Table II and Fig. 5). Substituting L for
M-22 (M-22L) did not
abrogate production of the 31-kDa band, indicating that the 31-kDa
protein was not generated from M-22, and
supported our interpretation that partial cleavage of the preprotein
occurred after translocation into the ER (Fig. 5A).
Occasionally in eukaryotes, leucine may initiate protein translation;
however, there is a substantial decrease in the amount of protein
translated relative to that when initiation occurs at a methionine
residue (18). Since we could not detect any difference in
translation efficiency between p35-57 and
M-22L, we concluded that translation
initiation at M-22 was not responsible for the
31-kDa preprotein.
|
B).
Treatment of p35-57 protein with Endo H revealed
a protein approximately 2 kDa smaller than the protein generated by in
vitro translation from M-22 in the absence of
microsomes (25 vs 27 kDa). This suggested that cleavage occurred distal
to M-22, within the signal peptide between
M-22 and R+1 (see Fig. 2A). Complementing this result was the observation that
approximately 5 kDa of leader sequence was removed from
p35-57 (Fig. 2C), whereas the
predicted size difference between M-57 and
M-22 was 3.4 kDa. This prediction was shown to
be correct, since there was approximately a 3-kDa difference between in
vitro products using either M-57 or
M-22 (Fig. 2B). Based upon these
findings, we postulated that the initial cleavage site must occur
within a region between 10–17 aa downstream of
M-22, corresponding to the middle portion of the
h-region of the signal peptide (diagrammed in Fig. 2A). To
evaluate the potential function of this region, aa -13 to -6
(V-13ATLVLLD-6, 
h,
Table II) within the h-region of the signal peptide were deleted. The
resultant p35 protein detected in COS-7 cells using anti-p35 Ab was
a single band of 28 kDa (Fig. 6A), corresponding to the
expected size of the unmodified preprotein translation product (29 kDa
minus approximately 1 kDa of the h region). This form of the p35
preprotein was consistent with the failure to translocate into the ER
and undergo subsequent glycosylation to produce the 31-kDa
form.
|
h region, three L were conserved among 10 animal species
of p35 examined. Thus, we decided to mutate each L in turn to determine
its possible function in this process. Mutating
L-10 to either R
(L-10R) or Q
(L-10Q) resulted in a 29-kDa protein,
suggesting that the h-region of the signal peptide was involved with
entry into the ER. Mutating L-8
(L-8R) or L-7
(L-7R) yielded a mixture of the 29-, 31-, and
36-kDa proteins (Fig. 6A). These results suggested that
mutation of L-10 abrogated entry into the ER,
while mutation of either L-8 or
L-7 only partially inhibited this process. Thus,
the mutations h, L-10R,
L-10Q, and
L-10R/R+1G appeared
to disrupt the ability of the p35 peptide to translocate from the
cytosol into the ER.
To examine further the role of L-10 in vitro
translation in the presence or the absence of microsomes was performed
(Fig. 6B). In the absence of microsomes, 30-kDa proteins
were produced for each form of p35-57
(L-10R,
L-10R/R+1G, and
p35-57). However, when microsomes were present,
only the unmodified p35-57 proteins had
increased in m.w., suggesting that only this protein was translocated
and subsequently glycosylated. As the 29-kDa proteins produced in COS-7
cells by h, L10R, and
L-10R/R+1G were
identical in size to the proteins observed in tunicamycin-treated COS-7
cells transfected with wild-type p35-5 (shown in
Fig. 4B), we compared results obtained from
tunicamycin-treated cells transfected with either
p35-57 or L-10R (Fig. 6C). Tunicamycin treatment of p35-57
transfected cells resulted in 29- and 24-kDa proteins, whereas COS-7
cells transfected with the L-10R mutation of
p35-57 produced a single 29-kDa species,
regardless of treatment with tunicamycin. It is unlikely that the
introduced mutations somehow affected preprotein glycosylation, which,
in turn, inhibited cleavage, since our studies with tunicamycin
demonstrated that primary cleavage within the signal peptide occurred
regardless of the glycosylation state. Taken together, these results
demonstrated that processing of the p35 secretory peptide within the ER
resulted in cleavage within the mid-h-region of the signal peptide and
not at the site predicted for cleavage of a signal peptide by signal
peptidase.
A second cleavage of the signal peptide is necessary for secretion
The amino terminus of secreted p35 has been mapped to an R
(12, 19), designated +1 (see Table II and Fig. 2A). Since cleavage within the mid-h-region of the secretory
peptide would not release a protein with an amino terminus comparable
to that of the mature protein, we hypothesized that a second cleavage
must have occurred after exiting the ER. Additional sequence analysis
revealed that the c-region of the signal peptide resembled a monobasic
protease recognition site, with cleavage predicted for the N-terminus
of R+1: 1) cleavage occurs at R; 2) there is H at
the -5 position; and 3) L and A are present immediately preceding the
single R cleavage site (20, 21). Therefore, we examined
this possibility using site-directed mutagenesis (Fig. 7A) of essential elements
within this sequence. As expected, replacing R+1
with G (R+1G) or L
(R+1L) did not affect cleavage within mid-h to
produce the stable 31-kDa p35 preprotein contained within the cell
lysate. However, subsequent p35 protein secretion was abolished by
R+1G and was greatly reduced by
R+1L (Fig. 7B). Taken together,
these results confirmed that removal of the p35 signal peptide required
two sequential cleavages. Translocation of the preprotein into the ER
was accompanied by the addition of 7 kDa of carbohydrate. The first
cleavage arose after entry into the ER, within the mid-h-region of the
signal peptide. The second cleavage, which generated the amino terminus
of the secreted protein, occurred either immediately before or upon
exiting the ER into the Golgi, where an additional 5 kDa of
carbohydrate was added. The latter steps appeared to be quite rapid, as
none of these forms could be detected intracellularly.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
30 min). Upon
exiting the ER into the Golgi compartment, a second cleavage generates
the amino terminus of the mature protein. This step appears to be
critical, because mutation of the cleavage site abrogates secretion,
but does not affect processing within the ER. This process must also be
quite rapid, as none of this form of the protein can be detected
intracellularly by immunoprecipitation. Unlike p40 and many other
secreted proteins (22, 23, 24), p35 appears to require
glycosylation for secretion, since inhibition of N-linked
glycosylation by tunicamycin completely inhibits secretion. Our current understanding of signal peptide function suggests that the signal sequence of a nascent protein emerges from the ribosome and binds to a signal sequence binding site on the methionine-rich domain of SRP54, one of the six different polypeptides that together with one 7S RNA molecule make up the signal recognition particle (SRP). This complex of ribosome-nascent protein and SRP then interacts with SRP receptor on the ER membrane, facilitating translocation of the peptide for further processing into the secretory pathway (16, 25). Similar to the findings of other investigators (26), we have shown that disruption of the hydrophobic core of the signal peptide inhibits translocation, suggesting that a similar process is required for p35. In particular, a single amino acid substitution at the -10 position completely inhibited this process, whereas similar substitutions at positions -8 and -7 were less inhibitory. Since the first cleavage site and the signal recognition site appear to be so closely related, we are unable to precisely identify the position of the first cleavage by the methods employed.
How the p35 signal peptide manages to bypass signal peptidase is unknown at this time. We do not think that the p35 signal peptide mediates this process by itself, because when the p35 signal peptide was used to direct secretion of p40 in vitro, it behaved as a typical secretory peptide, much as the p40 signal peptide. However, when we used the p40 signal peptide to direct secretion of p35 in vitro, the p40 signal peptide was not removed upon entry into microsomes (data not shown), indicating that the ability to evade signal peptidase was associated or at least influenced by some portion of the main body of the protein and was not mediated solely by the signal peptide.
What is the mechanism for generation of the amino terminus of secreted p35? The first amino acid residue of the secreted p35 peptide was shown previously by amino acid sequencing to be an R (13, 19). Consequently, the second site of cleavage must occur amino terminal to this residue. Upon closer examination of the amino acid sequence around R+1, we noted that the sequence conformed to a monobasic cleavage recognition site, as proposed by Benoit et al. (20) and more recently by Devi (21); namely, there is H at the -5 position and L and A immediately precede the single R cleavage site. Using site-directed mutagenesis we showed that reducing the length and pKa of the basic amino acid side chain (L) diminished, while removing the functional group altogether (G) abrogated, secretion, demonstrating the importance of this functional group in cleavage and/or site recognition. Proteases cleaving amino terminal to monobasic sites are often metalloproteases (27, 28, 29). We are currently investigating a role for metalloproteases in cleavage at R.
Proteases recognizing monobasic cleavage sites appear to have specific subcellular distributions (including the Golgi compartments), and different enzymes have been reported for every preprotein examined (21). This raises the possibility that specific inhibition of injurious IL-12 production, as in the case of toxic shock (30), might be achievable by blocking the function of a p35 monobasic-specific endopeptidase without globally affecting immune responses. Moreover, we do not know at this time the tissue distribution of the p35 monobasic-specific endopeptidase or whether all cell types can process and secrete IL-12. This raises the concern that in vivo delivery of IL-12 by gene transduction methods to some tissue types may result in atypical production of one or both of the subunits, with possible unintended effects.
In summary we demonstrated that p40 is processed in a manner consistent with the signal peptide/signal peptidase model, but that p35 is processed atypically. Initial cleavage of p35 occurred within the hydrophobic domain of the signal peptide and was accompanied by but not dependent upon partial glycosylation within the ER. Cleavage to release the amino terminus of the secreted form of the protein appeared to be a rapid, transient process, with no detectable cellular accumulation, and appeared concomitant with additional complex glycosylation. Glycosylation appeared to be requisite for p35, but not for p40, secretion.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: AAI, Inc., 2320 Scientific Park Drive, Wilmington, NC 28405.
3 Address correspondence and reprint requests to Dr. Parris R. Burd at the current addresss Maxygen, Inc., 515 Galveston Drive, Redwood City, CA 94063. E-mail address:
4 Abbreviations used in this paper: ER, endoplasmic reticulum; ORF, open reading frame; GSP, gene-specific primer; AEBSF, 4-(2-aminoethyl)-benzenesulfonyl-fluoride; Endo, endoglycosidase; h-, hydrophobic; SRP, signal recognition particle.
Received for publication August 27, 1999. Accepted for publication October 29, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
of lipopolysaccharide-inducible p35 and p40 genes. Blood 86:646., and IL-10: role of B cells in the maintenance of T cell responses. J. Exp. Med. 189:1.-dependent inducible expression of the human interleukin-12 p35 gene in monocytes initiates from a TATA-containing promoter distinct from the CpG-rich promoter active in Epstein-Barr virus transformed lymphoblastoid cells. Blood 91:4645.: differences in glycosylation and proteolytic processing lead to heterogeneity in batch culture. Biochem. J. 272:333.[Medline]
production and lethality in lipopolysaccharide-induced shock in mice.
Eur. J. Immunol. 25:672.
This article has been cited by other articles:
|
|
 |
|
  E. H. Duitman, Z. Orinska, E. Bulanova, R. Paus, and S. Bulfone-Paus How a Cytokine Is Chaperoned through the Secretory Pathway by Complexing with Its Own Receptor: Lessons from Interleukin-15 (IL-15)/IL-15 Receptor {alpha} Mol. Cell. Biol., August 1, 2008; 28(15): 4851 - 4861. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Gerosa, B. Baldani-Guerra, L. A. Lyakh, G. Batoni, S. Esin, R. T. Winkler-Pickett, M. R. Consolaro, M. De Marchi, D. Giachino, A. Robbiano, et al. Differential regulation of interleukin 12 and interleukin 23 production in human dendritic cells J. Exp. Med., June 9, 2008; 205(6): 1447 - 1461. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Saito, M. Matsuura, and Y. Hirai Regulation of Lipopolysaccharide-Induced Interleukin-12 Production by Activation of Repressor Element GA-12 through Hyperactivation of the ERK Pathway. Clin. Vaccine Immunol., August 1, 2006; 13(8): 876 - 883. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Liu, X. Guan, T. Tamura, K. Ozato, and X. Ma Synergistic Activation of Interleukin-12 p35 Gene Transcription by Interferon Regulatory Factor-1 and Interferon Consensus Sequence-binding Protein J. Biol. Chem., December 31, 2004; 279(53): 55609 - 55617. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Nagai, O. Devergne, T. F. Mueller, D. L. Perkins, J. M. van Seventer, and G. A. van Seventer Timing of IFN-{beta} Exposure during Human Dendritic Cell Maturation and Naive Th Cell Stimulation Has Contrasting Effects on Th1 Subset Generation: A Role for IFN-{beta}-Mediated Regulation of IL-12 Family Cytokines and IL-18 in Naive Th Cell Differentiation J. Immunol., November 15, 2003; 171(10): 5233 - 5243. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Liu, S. Cao, L. M. Herman, and X. Ma Differential Regulation of Interleukin (IL)-12 p35 and p40 Gene Expression and Interferon (IFN)-{gamma}-primed IL-12 Production by IFN Regulatory Factor 1 J. Exp. Med., October 20, 2003; 198(8): 1265 - 1276. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Kollet, C. Witek, J. D. Gentry, X. Liu, S. D. Schwartzbach, and T. M. Petro Deletional Analysis of the Murine IL-12 p35 Promoter Comparing IFN-{gamma} and Lipopolysaccharide Stimulation J. Immunol., November 15, 2001; 167(10): 5653 - 5663. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Lehmann, S. Bellmann, C. Werner, R. Schroder, N. Schutze, and G. Alber IL-12p40-Dependent Agonistic Effects on the Development of Protective Innate and Adaptive Immunity Against Salmonella Enteritidis J. Immunol., November 1, 2001; 167(9): 5304 - 5315. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Shiratori, M. Matsumoto, S. Tsuji, M. Nomura, K. Toyoshima, and T. Seya Molecular cloning and functional characterization of guinea pig IL-12 Int. Immunol., September 1, 2001; 13(9): 1129 - 1139. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. K. Wesa and A. Galy IL-1{beta} induces dendritic cells to produce IL-12 Int. Immunol., August 1, 2001; 13(8): 1053 - 1061. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Grazia Cappiello, F. S. Sutterwala, G. Trinchieri, D. M. Mosser, and X. Ma Suppression of IL-12 Transcription in Macrophages Following Fc{{gamma}} Receptor Ligation J. Immunol., April 1, 2001; 166(7): 4498 - 4506. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
