(Received for publication, March 3, 1997, and in revised form, April 14, 1997)
From the Laboratories of Molecular Biochemistry of
Hypertension and § Biochemical Neuroendocrinology, Clinical
Research Institute of Montreal (IRCM), Montreal, Quebec, Canada
H2W 1R7
PC1, a neuroendocrine member of the prohormone convertase family of serine proteinases, is implicated in the processing of proproteins in the secretory pathway. PC1 is synthesized as a zymogen and cleaves not only its own profragment in the endoplasmic reticulum, but a subset of protein substrates in the Golgi apparatus and in the Golgi-distal compartments of the regulated secretory pathway. Likewise, mouse PC1 (mPC1) has previously been shown to cleave human prorenin in GH4 cells (that contain secretory granules) while being unable to cleave prorenin in cells, such as Chinese hamster ovary (CHO) or BSC-40, which are devoid of secretory granules. In the current study, we show that removal of a C-terminal tail of mPC1 allows the efficient cleavage of prorenin in the constitutive secretory pathway of CHO cells. The C-terminal tail thus appears to act as an inhibitor of PC1 activity against certain substrates in the endoplasmic reticulum and Golgi apparatus, and its removal, which occurs naturally in secretory granules, may explain the observed granule-specific processing of certain proproteins. These results also demonstrate that PC1 is present in a partially active state prior to the secretory granules where it is processed to a maximally active state.
Proprotein convertases are serine proteinases that cleave their substrates at specific basic amino acids (1, 2). They share a common domain structure: an N-terminal signal peptide followed by a profragment region, a conserved subtilisin-like catalytic domain, a conserved domain of unknown function named P-domain, and a divergent C-terminal tail. They are differently distributed in tissues and show maximal activity in different cellular compartments. For instance, furin, PACE4, PC5-B, and PC7 are active in the constitutive secretory pathway, whereas PC1 (also named SPC3), PC2, and PC5-A are preferentially expressed in endocrine and neuroendocrine tissues where they are thought to process prohormones in the secretory granules of the regulated secretory pathway (1-4). PC1 has been implicated in the processing of many proproteins including proopiomelanocortin (POMC)1 and proinsulin. Whereas proinsulin is only cleaved in secretory granules (5), the processing of POMC begins in the trans-Golgi network (TGN) (6-8), clearly demonstrating that PC1 can be active prior to entering secretory granules. Following its synthesis as a 753-amino acid zymogen (9, 10), PC1 undergoes an autocatalytic intramolecular processing of its N-terminal profragment in the endoplasmic reticulum, resulting in an 87-kDa enzyme (11, 12). This 87-kDa form is targeted to the regulated secretory pathway where it is further shortened by removal of 135 amino acids of its C-terminal tail leading to a 66-kDa form (11-14). The C-terminal cleavage occurs at the dibasic Arg-Arg617-618 site (15), possibly by an autocatalytic event (15, 16). The C-terminal tail has been proposed to play a role in sorting of PC1 to the regulated secretory pathway (17), but its exact role on PC1 activity is still poorly understood.
Prorenin is an inactive precursor that is sorted to secretory granules where it is converted to active renin by a unique proteolytic cleavage at a specific pair of basic amino acids (Lys-Arg42,43), resulting in the removal of a 43-amino acid profragment (18-20). Cotransfection of expression vectors for mouse PC1 (mPC1) and human prorenin in cells that contain secretory granules (GH4 cells) leads to the secretion of active renin in the medium (21). In contrast, co-expression of prorenin and mPC1 in cells that are devoid of secretory granules (CHO or BSC-40 cells) does not lead to active renin secretion, suggesting that the environment of the secretory granules is essential for prorenin processing by mPC1 (21). Previous results have shown that the kinetics of the N-terminal prosegment cleavage of mPC1 are similar in the two cell types (21), raising the possibility that the granule dependance of prorenin processing is due to additional granule-induced modifications on prorenin and/or mPC1.
Because of its unique PC1 cleavage site and its granule-dependance for activation, we have used human prorenin as a model to understand how PC1 discriminates between substrates to be cleaved in the early (TGN) and late (secretory granules) compartments of the secretory pathway. Our results suggest that the C-terminal tail of mPC1 has an inhibitory effect on PC1 activity in the constitutive secretory pathway and that the removal of this C-terminal tail in secretory granules is necessary for efficient prorenin processing.
Construction of the expression vector for native human prorenin (Proren, also referred to as pRhR1100) under the control of the Rous sarcoma virus promoter/enhancer has been described previously (22). The mutation of lysine to alanine in the native Lys-Arg42,43 cleavage site at the junction of the profragment and the renin moiety (Proren-K/A+42 (19); numbering relative to amino acid 1 of the prosegment) was carried out by overlap extension polymerase chain reaction (23).
The cDNA for mPC1 (9) was inserted in the expression vector
pRSV-globin (22). Site-directed mutagenesis of the C-terminal tail of
mPC1 (Fig. 1) was carried out using the following oligonucleotides: mPC1-C (converts Arg617 to a stop codon),
5
-CGGGATCC CTA GTC ATT CTG GAC TG-3
(antisense
primer; generation of a stop codon (underlined) and a BamHI
site (bold)); and mPC1-KA618 (Arg-Arg617,618 to
Lys-Ala), 5
-G AAT GAC AAG GCA GGA GTG G-3
(sense primer) and 5
-C CAC TCC TGC CTT GTC ATT
C-3
(antisense primer).
An expression vector was also generated for the C-terminal tail of mPC1
(see Fig. 1, mPC1-Cterm). A BamHI site was inserted in mPC1
at a site found to correspond to the unique BamHI site of
human PC1 (by alignment of both sequences) using the polymerase chain
reaction with the following oligonucleotide: 5-CGGGATCC TGT
GGA GAA GCG G-3
(sense primer; BamHI site (bold) and mPC1 sequence starting at Val627).
The expression vector for mPC1 containing the BamHI site was
cleaved with HindIII and BamHI to delete the
N-terminal portion. The coding sequence for the signal peptide followed
by 6 amino acids of the prosegment of prorenin was inserted 5 of the
C-terminal sequence of mPC1 by a cohesive 5
HindIII 3
BamHI ligation to assure effective insertion of the
recombinant protein in the endoplasmic reticulum. All recombinant
plasmid constructions were verified by sequencing of double-stranded
DNA.
The rat GH4 somatomammotrophic cell line was grown at 37 °C in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% fetal calf serum and 0.1% Serxtend (Irvine Scientific) in a humidified incubator at 5% CO2, 95% air. CHO cells were grown in DMEM supplemented with 10% fetal calf serum and 20 µg/ml proline.
Measurement of Renin and ProreninGH4 or CHO cells were plated at a density of 8 × 105 cells in a 6-well dish and were transfected 24 h later with a calcium-phosphate precipitate (300 µl) containing 2 µg/ml of the prorenin expression vector and 8 µg/ml of either the appropriate mPC1 expression vector or a control vector (for transfection experiments with 2 plasmids), or 1 µg/ml of the prorenin expression vector, 4 µg/ml of the appropriate mPC1 expression vector, and 5 µg/ml of the PC1-Cterm expression vector (for transfection experiments with 3 plasmids). In all cases, the control vector consisted of an analogous expression vector, which coded for an inert portion of the mouse immunoglobulin heavy chain constant region (24). Twenty-four h after transfection, the cells were glycerol shocked. After a 16-h secretion period, the supernatants were collected for determination of prorenin and renin levels by the angiotensin I generation assay described previously (19). Briefly, supernatants were incubated with an excess of the renin substrate angiotensinogen either directly (active renin content) or following an incubation with trypsin (total renin content = prorenin + renin). The percentage of active renin was calculated as (active renin content/total renin content) × 100.
Secretion Kinetics of Active ReninGH4 cells were plated at a density of 1 × 106 cells in a 6-well dish and were transfected 24 h later with Lipofectin Reagent (Life Technologies, Inc.) using 10 µg of the prorenin expression vector and 10 µg of the appropriate mPC1 expression vector. Sixty h after transfection, the cells were depleted of methionine for 1 h, labeled with 300 µCi of [35S]methionine/well (pulse), and incubated in complete medium for 15, 30, or 60 min (chase). Culture supernatants were immunoprecipitated with protein G-agarose-coupled anti-human renin antibody as described previously (24). Immunoprecipitated proteins were fractionated by SDS-PAGE (10% acrylamide gel) and gels were subjected to fluorography.
Biosynthetic Labeling of PC1 ProteinsGH4 cells were plated at a density of 8 × 105 cells in a 6-well dish and were transfected 24 h later with a calcium-phosphate precipitate (300 µl) containing 10 µg/ml of the appropriate mPC1 expression vector. Twenty-four h after transfection, the cells were glycerol shocked. The next day the cells were depleted of methionine for 1 h in methionine-free DMEM containing 10% dialyzed fetal calf serum and labeled for 2 h with 300 µCi of [35S]methionine/well. Culture supernatants were then immunoprecipitated with protein A-Sepharose-coupled anti-N-terminal or C-terminal mPC1 antibody as described previously (11). Immunoprecipitated proteins were fractionated by SDS-PAGE (7.5% acrylamide for mPC1 or related proteins and 12% acrylamide for mPC1-Cterm) and gels were subjected to fluorography.
As shown previously (21), cotransfection of GH4 cells with an
expression vector for human prorenin and a control plasmid led to the
secretion of inactive prorenin into the medium (see Fig. 2,
proren + control), indicating that GH4 cells do
not express an endogenous prorenin processing enzyme. In contrast,
cotransfection of GH4 cells with expression vectors for both prorenin
and mPC1 led to the secretion of about 20% of active renin (see Fig.
2, proren + mPC1) (21). To investigate the
dependance of prorenin processing on the secretory granule environment
and the role of the C-terminally truncated form of mPC1 in active renin
generation, we constructed an expression vector for mPC1 containing a
deletion of its C-terminal tail starting at the
Arg-Arg617,618 cleavage site (Fig. 1,
mPC1-C). This construct should lead exclusively to the
synthesis of the 66-kDa form. Cotransfection of GH4 cells with both
proren and mPC1-
C led to a dramatic increase in the secretion of
active renin (62.6 ± 2.9% of active renin), suggesting that the
absence of the C-terminal tail enhances PC1 activity in prorenin
processing. Although mutation of the Arg-Arg617,618
cleavage site into the noncleavable Lys-Ala site in the full-length mPC1 (Fig. 1, mPC1-KA618) would have been
expected to prevent PC1 activity, transfection of this mutant with
prorenin did not significantly decrease prorenin processing activity
compared with native mPC1 possibly due to cleavage of the C-terminal
tail at alternative dibasic sites (Fig. 4,
mPC1-KA618). When mPC1-
C was co-expressed
with human prorenin mutated at the native profragment cleavage site, no
active renin was secreted (Fig. 2,
proren-K/A+42 + mPC1-
C),
indicating that mPC1-
C retained its specificity for the previously
reported unique cleavage site at Lys-Arg42,43.
To further investigate the effects of mPC1 and mPC1-C on prorenin
processing, pulse-chase experiments were performed on GH4 cells
cotransfected with the expression vectors for prorenin and either mPC1
or mPC1-
C. Following a 2-h pulse-labeling period with radioactive
methionine, the cells were chased by incubation in complete medium for
15, 30, or 60 min. Cell supernatants were immunoprecipitated with an
anti-renin antibody and submitted to SDS-PAGE (24). As shown in Fig.
3, while co-expression of mPC1 and prorenin leads to the
production of active renin whose relative abundance gradually
accumulates over the chase period, the relative amount of active renin
generated by mPC1-
C has already reached a maximum by the first time
point (2 h + 15 min). Thus, removal of the C-terminal tail of mPC1
accelerates the generation and secretion of active renin from
transfected GH4 cells.
To exclude the possibility that the differences observed in the
efficiency of prorenin processing resulted from differential expression
of the various mPC1 proteins (mPC1, mPC1-C, and
mPC1-KA618), GH4 cells were transfected with the expression
vectors for the mPC1 proteins and were pulse-labeled for 2 h with
radioactive methionine. Media were immunoprecipitated with an anti-mPC1
antibody directed against the N-terminal part of the protein (11), and proteins were fractionated by SDS-PAGE. As expected, pro-mPC1 was
cleaved to the 87-kDa form (mPC1) and to the 66-kDa C-terminally truncated form that were both secreted in the medium (Fig.
4, mPC1). In contrast, transfection of GH4
cells with mPC1-
C resulted in the secretion of the 66-kDa form only.
Mutation of the dibasic Arg-Arg617,618 site prevented the
formation of the 66-kDa form, indicating that the double point mutation
was effective in inhibiting the cleavage of the C-terminal tail.
However, two minor bands were also detected, probably resulting from
alternative cleavages at the Lys-Arg629,630 and
Arg-Arg654,655 dibasic sites (Fig. 4,
mPC1-KA618). Thus, C-terminal processing at
these alternative sites results in active mPC1 since
mPC1-KA618 was able to cleave prorenin in GH4 cells (Fig.
2, proren + mPC1-KA618).
To investigate the role of the C-terminal tail on PC1 activity in the
constitutive secretory pathway, transfection experiments were performed
in CHO cells that are devoid of secretory granules and a regulated
secretory pathway. The 87-kDa form of mPC1 has been shown to be the
predominant form present in constitutive cells, whereas the 66-kDa form
is not detected (25). As described previously (21), PC1 was virtually
unable to process prorenin in these cells (Fig. 5,
proren + mPC1), confirming that prorenin processing by PC1 does not occur in cells lacking secretory granules. Interestingly, secretion of about 25% of active renin was observed when CHO cells were cotransfected with expression vectors for both
prorenin and mPC1-C (Fig. 5), indicating that the C-terminally truncated PC1 is active in the constitutive secretory pathway. Similar
to the results obtained in GH4 cells, mPC1-
C specifically cleaved
the prorenin processing site (Fig. 5,
proren-K/A+42 + mPC1-
C). These
results demonstrate that PC1 deleted of its C-terminal tail can cleave
prorenin in the constitutive secretory pathway.
An expression vector for the C-terminal tail of mPC1 was next
constructed to test its effect on the activity of mPC1 (Fig. 1,
mPC1-Cterm). The design of this construct assured targeting of the protein to the secretory pathway. Transfection of GH4 cells with
the mPC1-Cterm expression vector led to the secretion of a protein
having the expected molecular mass of 20 kDa (Fig. 4, mPC1-Cterm). Fig. 6A shows that
transfection of GH4 cells with mPC1-Cterm caused a 2.5-fold decrease in
the level of active renin generated by mPC1-C, resulting in a level
of active renin similar to the one observed with native PC1 (from
56.3 ± 6 to 22.3 ± 2.9% of active renin). Likewise, the
C-terminal tail of mPC1 was able to inhibit processing of human
prorenin in the constitutive secretory pathway of CHO cells (Fig.
6B). The same inhibitory effect was obtained using an
expression vector coding for the C-terminal tail of human PC1 (hPC1)
(not shown). These results provide the evidence that the C-terminal
tail of PC1 has an inhibitory effect on PC1 activity in the
constitutive secretory pathway.
Previous cotransfection studies conducted with different cell
lines have shown that prorenin processing by the prohormone convertase
mPC1 could occur only in cells containing secretory granules (21).
Following the cleavage of its profragment in the endoplasmic reticulum,
mPC1 undergoes an additional cleavage in the secretory granules that
results in the removal of its C-terminal tail (11-13). In this study,
we have investigated the role of the C-terminal tail of mPC1 on
prorenin processing. We show that truncation of the C-terminal tail of
mPC1 (mPC1-C) greatly augments the cleavage of prorenin in GH4 cells
and allows PC1 to cleave prorenin in CHO cells, which are devoid of
secretory granules. Moreover, pulse-chase experiments in GH4 cells
indicate that processed renin is secreted more rapidly when mPC1-
C
is coexpressed compared with native mPC1, consistent with processing in
the rapidly secreting constitutive secretory compartment. These results
suggest that the environment of the secretory granules is not essential
for the activity of mPC1-
C. However, in CHO cells, mPC1-
C
generates less active renin than in GH4 cells (24.6 ± 1.2 versus 62.6 ± 2.9%) although the mPC1-
C protein is
as efficiently expressed (not shown). This observation suggests that
the environment of the TGN or of the immature granules is more
favorable for the activity of mPC1-
C in GH4 cells. It is also
possible that a portion of mPC1-
C is sorted to the secretory
granules of GH4 cells; however, pulse-chase experiments show that the
bulk of mPC1-
C is secreted constitutively and thus never reaches the
mature secretory granules (not shown).
The current study demonstrates that the C-terminal tail of mPC1 can inhibit PC1 activity in the constitutive secretory pathway by an intermolecular mechanism. Our results do not allow us to determine whether in the natural state the C-terminal tail inhibits PC1 activity via an intra- or intermolecular mechanism or, in fact, if this inhibition is directly on PC1 or on a necessary cofactor for its action. However, these results and the proposed implication of the C-terminal tail of mPC1 in the sorting of the 87-kDa form to the regulated secretory pathway (17) raise the possibility that the C-terminal tail contributes significantly to targeting the maximal activity of PC1 to the secretory granules.
Using AtT-20 cells transfected with human prorenin, we have previously shown that the processing site for a granule-specific endoprotease between the profragment and the renin molecule seems necessary for appropriate prorenin sorting to the regulated secretory pathway (24). This result suggests that prorenin can be driven in the regulated secretory pathway by binding to the active site of the processing enzyme prior to sorting. As prorenin activation occurs only in the secretory granules, it is possible that prorenin binds mPC1 in the TGN but is not immediately processed due to the inhibition of mPC1 by its C-terminal tail. The complex could then be targeted to the regulated secretory pathway where the environment of the secretory granules might favor the cleavage and dissociation of the C-terminal tail rendering mPC1 capable of processing prorenin.
The mechanism by which the C-terminal tail inhibits PC1 activity could
involve the entire tail or a more specific region. The region between
amino acids 625-660 shows an important divergence between species; for
instance, only 48.8% identity is found between mPC1 and hPC1 compared
with 92.7% identity in the total amino acid sequence (Fig.
7). Nevertheless, the C-terminal tail of human PC1 was
effective in inhibiting the activity of mPC1-C (not shown). A
possible inhibitory mechanism could involve binding of the C-terminal tail to the active site of mPC1 by one of the several dibasic residues
present in the C-terminal region. However, site-directed mutagenesis of
the conserved dibasic Arg-Arg site at position 654-655 did not prevent
the inhibitory effect of mPC1-Cterm on mPC1-
C (not shown), making it
unlikely that such a mechanism accounts for the inhibition of PC1 by
its C-terminal tail. Another possible mechanism could involve induction
of a conformational change around the active site by binding of the
C-terminal tail elsewhere on the mPC1 protein. Future studies will be
necessary to test this and other possible mechanisms.
The functional importance of the 66-kDa C-terminally truncated form of
mPC1 has been investigated in various systems. In vitro studies have shown that the 66-kDa form of mPC1 was more active against
fluorogenic peptide substrates but less stable than the nontruncated
mature form (15). In addition, the C-terminally truncated form of mPC1
requires calcium concentrations above 20 mM and pH values
between 5.0 and 5.5 for maximal enzymatic activity, conditions
compatible with an optimal activity in mature secretory granules. Zhou
et al. (15) have reported that the 66-kDa form of PC1 was
more active than the 87-kDa form in generating certain products of
proneurotensin in secretory granule-containing PC12 cells. Moreover,
expression of a C-terminally truncated form of mPC1 in AtT-20 cells
that endogenously express POMC and mPC1 greatly accelerated POMC
conversion to its smaller products (17). However, it is important to
note that POMC processing is not strictly restricted to the secretory
granules since processing at the junction between the 23-kDa fragment
and -LPH begins in the TGN (8). This suggests that, in the context
of the TGN, the 87-kDa form of PC1 may have a high enough affinity for
certain substrates such as POMC to overcome the inhibition by its
C-terminal tail. In contrast, processing of other substrates such as
prorenin and proinsulin by mPC1 apparently requires the removal of the
C-terminal tail within the granules, thereby limiting the activity of
the processing protease to this compartment.
Two other prohormone convertases, PC2 and PC5-A, are also preferentially active in the regulated secretory pathway (1-3). In contrast to PC1, pro-PC2 slowly exits the endoplasmic reticulum and is processed to PC2 only in the TGN where its activity is regulated by the binding protein 7B2. It has been suggested that binding of PC2 to 7B2 may prevent premature activation of PC2 in the regulated secretory pathway (26-28). Similar to PC1, PC5-A has been shown to carry out granule-specific processing of human prorenin in transfected cells (29). In addition to the rapid removal of an N-terminal prosegment in the early compartments of the secretory pathway, PC5-A undergoes truncation of its C-terminal tail in secretory granules (3). We may thus speculate that the C-terminal tail of PC5-A also inhibits PC5 in the constitutive secretory pathway.
In conclusion, our results suggest that the C-terminal tail of mPC1 could act as a partial inhibitor of mPC1 activity in the constitutive secretory pathway and that maximal activity of mPC1 requires the removal of its C-terminal tail. PC1 activity is thus dependent on two different sequencial processing steps: first, the N-terminal profragment is processed in the endoplasmic reticulum, and second, the C-terminal tail is cleaved in secretory granules. Such a mechanism may explain the observed secretory granule-specific processing of certain prohormones or proproteins such as prorenin.