 |
INTRODUCTION |
Large dense core vesicles
(LDCV)1 or
secretory granules are specialized compartments for storing
biologically active neuropeptides and polypeptide hormones (1, 2).
However, the signals involved in sorting the membrane proteins that
function in secretory granules have not been well defined. Even the
relative importance of granule entry signals versus signals
for removal of non-granule membrane proteins is not clear.
Peptidylglycine 
-amidating monooxygenase (PAM; EC 1.14.17.3), a
bi-functional enzyme involved in the -amidation of peptides (3, 4),
is a type I integral membrane protein processed and stored in secretory
granules. Many type I membrane proteins use signals in their COOH
terminus to control trafficking (5-9), and the COOH terminus of PAM
contains several distinct trafficking signals (10, 11).
The role of phosphorylation as a routing determinant for membrane
proteins has been well established (8, 12-14). In particular, phosphorylation of acidic clusters in the cytosolic domains of proteins
such as furin (6, 15), carboxypeptidase D (16), and prohormone
convertase 6B (17) plays a crucial role in trafficking. For furin, the
binding of both adaptor protein 1 (AP-1) and phosphofurin acidic
cluster sorting protein 1 (PACS-1) is increased by casein kinase II
(CKII)-catalyzed phosphorylation of Ser residues located within an
acidic cluster (Fig. 1A) (18, 19). In the cytosolic domain
of PAM, the region located between Lys953 and
Lys971 is extremely acidic, forming a PEST motif, with
Thr959 and Ser961 identified as potential
phosphorylation sites (NetPhos 2.0 Prediction, Center for Biological
Sequence Analysis, www.cbs.dtu.dk/services/). However, truncation of
PAM-1 at Ser961, which interrupts the acidic cluster,
yields a protein that is trafficked normally (20). Therefore, the
phosphorylation of an acidic cluster has not been demonstrated to be
important in PAM trafficking.
When membrane PAM is expressed in the neuroendocrine cell line, AtT-20,
it is efficiently cleaved in the secretory pathway to yield soluble
peptidylglycine -hydroxylating monooxygenase (PHM) and membrane
peptidyl--hydroxyglycine -amidating lyase (PAL). Although intact
membrane PAM exhibits both enzymatic activities, soluble PHM has a
turnover number severalfold higher than intact PAM, revealing a
regulatory role for the COOH-terminal regions of PAM (21). The majority
of membrane PAM and PAL accumulates in a distal region of the
trans-Golgi network (TGN) at steady state, although some of
the protein is located in secretory granules (22). The small amount of
PAM expressed on the surface of AtT-20 cells at steady state undergoes
rapid internalization and is recycled to the TGN via perinuclear
endosomes. The internalized protein may undergo lysosomal degradation
or be packaged into new granules (11, 23, 24).
PAM is phosphorylated on multiple residues in its cytosolic
COOH-terminal domain (CD) (10). Test tube assays utilizing recombinant PAM-CD were used to determine that protein kinase A phosphorylates Ser921 and protein kinase C phosphorylates
Ser932 and Ser937 (10). Mutation of
Ser937 to Ala937 or Asp937
demonstrated that the inability to phosphorylate Ser937
limits the secretion of soluble PHM and increases the turnover rate of
newly synthesized PAM-1 (11). Furthermore, phosphorylation of PAM-1 at
Ser937 directs the protein away from lysosomes after
internalization from the cell surface, whereas dephosphorylation of
this site is needed for a later step in the endocytic pathway (10, 11). The novel protein kinase, PAM cytosolic
interactor protein-2 (P-CIP2), selectively
phosphorylates Ser949 in the CD of PAM (25). P-CIP2
interacts with the CD of PAM at several sites that are closer to the
transmembrane domain than Ser949 (25). Mutation of the
PAM-CD at these juxtamembrane sites eliminates the ability of PAM to
bind tightly to P-CIP2 and eliminates the ability of PAM to modify
regulated secretion (26). These mutations do not eliminate the ability
of PAM to be phosphorylated by P-CIP2 (25). Since Ser949
lies within a fairly acidic region of the PAM-CD (Fig. 1), this raised
the possibility that this region functions as an acidic cluster in PAM
trafficking. Therefore, our goals were to identify sites where CKII
phosphorylates the CD of PAM and to determine whether these sites could
be involved in mediating PAM trafficking. In this study, we identified
Thr946 and Ser949 (Fig. 1) as sites
phosphorylated by sea star CKII. We then explored the roles of
phosphorylation and de-phosphorylation of these sites in the
biosynthetic and endocytic trafficking of PAM-1 by mutating Thr946 and Ser949 (Fig. 1) to
Ala946 and Ala949 or to Asp946 and
Asp949. Mutant PAM proteins were expressed in AtT-20 cells
to evaluate their biosynthesis, catalytic activity, steady-state
localization, and internalization.
| |
MATERIALS AND METHODS |
In Vitro Phosphorylation of Purified Recombinant PAM
COOH-terminal Domain--
Purified, recombinant rat PAM-CD
(Met-Ala-Ser-Met-Thr-Gly-Gly-Gln-Gln-Met-Gly-Arg-Ile-Pro-rPAM-1(898-976))
(40 µg) was incubated with 0.4 µg of CKII from sea star,
Pisaster ochraceus (Upstate Biotechnology, Inc.), in 20 mM HEPES, 10 mM MgCl2 (pH 7.0)
buffer containing 10 mM ATP and 100 µCi of
[
-32P]ATP for 2 h (37 °C). The reaction was
stopped by adding 0.1% trifluoroacetic acid, and the peptide was bound
to a C18 SepPak cartridge, washed with 0.1%
trifluoroacetic acid, and eluted with 80% acetonitrile, 0.1%
trifluoroacetic acid. The dried eluate was redissolved in 25 mM Tris-HCl, 1 mM EDTA (pH 8.5) and digested with endoproteinase LysC (2.0 µg; Roche Molecular Biochemicals) for
18 h at 37 °C. The lyophilized sample was applied to a series of Bio-Sil TSK-DEAE (Bio-Rad) columns equilibrated and eluted in 32%
acetonitrile, 0.1% trifluoroacetic acid. The single major peak of
labeled material included in the column was pooled and dried before
application to a Bio-Sil TSK-DEAE column equilibrated in 50 mM sodium acetate (pH 4.5) containing 30% acetonitrile. This column was eluted with a linear gradient of 0.25 M
NaCl, 50 mM sodium acetate (pH 4.5), and 30% acetonitrile.
The phosphopeptides were pooled and dried before desalting on the
Bio-Sil TSK-DEAE columns. The labeled Bio-Sil TSK-DEAE eluates were
covalently linked to arylamine membranes (Sequelon-AA; Waters
Instruments) for Edman degradation sequencing in an ABI model 477A
instrument. The sequencer was configured to divide the successive
phenylthiohydantoin fractions, 40% for on-line high pressure liquid
chromatography detection, and 60% for determination of 32P
radioactivity by liquid scintillation counting.
Expression Vectors--
pBluescript plasmids encoding either
PAM-1 with Thr946 and Ser949 mutated to
Ala946 and Ala949 or to Asp946 and
Asp949 were made using SOEing polymerase chain reaction
(27). Amplification for both constructs involved an upstream sense
primer (rPAM-1(2954-3250)) and mutagenic antisense primers with codons
(underlined) either for Ala946 and Ala949
(5'-GCCCCCTCTGCGCTCACTCGG -3') or for
Asp946 and Asp949
(5'-CATCCCCTCGTCGCTCACTCGG-3') and mutagenic
sense primers for Ala946 and Ala949
(5'-GAGCGCAGAGGGGGCTGACCAAGAG-3') or for
Asp946 and Asp949
(5'-GAGCGACGAGGGGGATGACCAAGAG-3') and T7
antisense primers. pBS.KrPAM-1 TS/AA and pBS.KrPAM-1 TS/DD were both
made by inserting the final respective amplified fragment into
pBS.KrPAM-1 using the restriction enzymes HindIII and
XmaI. The DNA regions derived from polymerase chain reaction
were verified by sequencing. Constructs were placed into the pCI.neo
mammalian expression vector (Promega Corp.) using SalI and
NotI cloning sites.
Transfection and Generation of Stable Cell Lines--
AtT-20
cells were transfected using Lipofectin (Sigma) with either
pCI.neo.KrPAM-1/T946A/S949A (TS/AA) or pCI.neo.KrPAM-1/T946D/S949D (TS/DD) (10, 11). The stably transfected AtT-20 cell lines were
maintained in Dulbecco's modified Eagle's medium/F-12 medium (Life
Technologies, Inc.) with 10% fetal clone serum (HyClone, Logan, UT),
10% NuSerum (Collaborative Research, Bedford, MA), and antibiotics.
Medium containing 0.5 mg/ml G418 was used for selection and maintenance
of transfected cells. Three clones from each cell line were examined,
with the levels of PAM expression not varying more than 2-fold. Assays
for PHM and PAL in medium and cell extracts were performed as described
(11).
Antibodies--
The following rabbit polyclonal antisera were
diluted 1:1000 for use in immunostaining and Western blot analyses: PHM
(JH1764) directed against PAM-1(37-382); exon A (JH629) directed
against PAM-1(409-497); and PAL (JH 471) directed against
PAM-1(463-864) (11, 22). The following monoclonal antibody was used,
syntaxin 6 (BD Transduction Laboratories, 1:500).
Generation of Phospho-Ser949 Antibody--
A
peptide,
Asp942-Arg-Val-Ser-Thr-Glu-Gly-(P)Ser949-Asp-Gln-Glu-Lys953
(Fig. 1, italics), from the cytosolic domain of PAM was
synthesized with an NH2-terminal Cys residue using
monobenzyl-protected Fmoc (N-(9-fluorenyl)methoxycarbonyl)
phospho-Ser. The peptide was purified as a monomer and conjugated to
keyhole limpet hemocyanin (Pierce) with glutaraldehyde or with
maleimidobenzoyl-N-hydrosysuccinimide ester. Two rabbits
were immunized by Covance (Denver, PA) using an equal mixture of the
two conjugates. Serum (JH2541) was depleted of antibody against the
non-phosphorylated PAM-1(942-953) peptide by incubation with Affi-Gel
10 to which non-phosphorylated peptide had been linked. The titer and
specificity of the P-Ser949 antibody (JH2541) were
determined with a solid phase assay using the synthetic peptides
PAM-1(942-953) and PAM-1(942-953)P-Ser949. The depleted
phosphoserine specific antibody, JH2541, was used at a dilution of
1:250.
Immunofluorescence Microscopy--
Immunofluorescence microscopy
was performed as described previously using 4% paraformaldehyde or
100% methanol (24). For experiments with the microtubule destabilizing
drug, nocodazole (Sigma), cells were plated on
poly-L-lysine-coated glass slides and grown for 36 h
before incubation for 20 min in complete serum-free medium (CSFM)
(37 °C). The cells were then incubated in CSFM containing 10 µM nocodazole for 20 min (37 °C) before fixation in
100% ice-cold methanol. Following fixation, the cells were blocked
with 2 mg/ml bovine serum albumin in phosphate buffer, before
incubation with the exon A and syntaxin 6 antibodies. Visualization was
done using a Cy3-conjugated donkey anti-mouse IgG and a fluorescein
isothiocyanate-conjugated goat anti-rabbit IgG. Cells were observed
under epifluorescence optics on a Zeiss Axioskop (Carl Zeiss Inc.,
Thornwood, MT). Cells were photographed and analyzed with a Princeton
Instruments Micromax digital camera.
Antibody Internalization and the Microtubule-destabilizing Drug,
Nocodazole--
Antisera to PAL or exon A (JH471 or JH629; diluted
1:50) were used as described (11, 24, 28). Briefly, cells were
incubated with antiserum at different temperatures (4, 20, or 37 °C)
and chased in antibody-free CSFM without sodium bicarbonate (CSFM-Air) for various lengths of time to trap the internalized PAM-antibody complexes in either early or late stages of internalization (29). Nocodazole was dissolved in CSFM-Air and used at a concentration of 10 µM. After incubation with exon A antiserum (1:50), for
1 h at 20 °C to load early and late endosomes with PAM-antibody complexes, cells were rinsed in antibody-free CSFM, warmed to 37 °C,
and incubated with 10 µM nocodazole for 60 min before
fixation and double immunostaining with the syntaxin 6 monoclonal antibody.
Antibody Internalization Utilizing the Protein Kinase Inhibitor,
DRB--
The CKII inhibitor,
5,6-dichloro-1-
-D-ribofuranosylbenzimidazole (DRB;
Calbiochem) (30), was dissolved in CSFM-Air. AtT-20 PAM-1 cells were
incubated with exon A antiserum for 30 min (20 °C) to load early and
late endosomes with PAM-antibody complexes before the addition of 100 µM DRB to the medium (30 min; 20 °C). The medium was
then replaced with 37 °C antibody-free CSFM-Air containing 100 µM DRB for either 30 or 60 min before fixation and immunostaining.
Stimulated Secretion--
Stably transfected AtT-20 cells were
plated in triplicate on poly-L-lysine-coated culture
plates. Prior to the experiment, cells were incubated in three changes
of CSFM-Air over 1.5 h. Basal collections of medium were for 30 min each; these were followed by one 30-min collection in the presence
of 1 mM BaCl2 (31). Collected medium was
centrifuged after addition of protease inhibitors. Cells were extracted
in 20 mM sodium
N-tris[hydroxymethyl]methyl-2-aminoethanesulfonic acid, 10 mM mannitol, and 1% Triton X-100 (pH 7.4). To
measure PHM activity, aliquots of medium or cell extracts were assayed using 0.5 µM CuSO4, 0.5 mM
ascorbate, 0.5 µM Ac-Tyr-Val-Gly, trace amounts of
125I-labeled Ac-Tyr-Val-Gly, 0.1 mg/ml catalase, and 150 mM NaMES (pH 5.5) (32, 33).
PAM-1 Extraction and Sucrose Gradient Sedimentation--
One
confluent well of a 6-well dish from each of the three stably
transfected AtT-20 cell lines was grown for 2 days before preincubation
in 500 µl of CSFM-Air for 5 min (37 °C). The cells were then
incubated on ice in ice-cold buffer (20 mM PIPES, 2 mM Na2EDTA, 50 mM NaF, 10 mM Na4P2O7 and 1 mM Na3VO4, pH 7.5) with protease
inhibitors (30 µg/ml phenylmethylsulfonyl fluoride, 2 µg/ml
leupeptin, 2 µg/ml pepstatin, 16 µg/ml benzamidine, 5 µg/ml lima
bean trypsin inhibitor) and 1% Triton X-100 (500 µl) added (15 min)
(31). A suspension of the scraped cells was centrifuged at 250,000 × g for 20 min (4 °C). Aliquots (200 µl) from the
supernatant of each cell type were further fractionated on 2 ml of
5-20% linear sucrose gradients buffered to pH 7.5 or pH 5.5 with the
PIPES buffer minus detergent. A 50% sucrose "pad" (169 µl) was
placed at the bottom of each gradient, and the following molecular
weight markers (50-250 µg) were used as internal standards:
cytochrome c, ovalbumin, bovine serum albumin, catalase, and
apoferritin (Sigma). Gradients were centrifuged in a Ti-55 swinging
bucket rotor in a TL-100 centrifuge (Beckman Instruments) for 5 h
(4 °C) at 50,000 rpm (214,000 × g). Fractions (169 µl) from each gradient were collected from the top to the bottom and
particulate matter at the bottom of the gradient was resuspended in
Laemmli sample buffer. Gradient fractions were analyzed by Western blot using the exon A antibody, and molecular weight markers were visualized using Coomassie Brilliant Blue R-250.
Cell Extract Immunoprecipitation and Western Blot
Analysis--
A 100-mm dish of confluent AtT-20 PAM-1 cells was
extracted as described above (31), and a suspension of the scraped
cells was centrifuged at 304,000 × g for 20 min
(4 °C). Aliquots (200 µl) from the supernatant were incubated with
10 µl of the exon A antibody or the P-Ser949 antibody for
90 min (4 °C). Following centrifugation of the immunoprecipitates at
10,000 × g (20 min), they were incubated with 60 µl
of protein A coated in 0.25% bovine serum albumin (60 min). Washed
immunoprecipitates were boiled into 200 µl of sample buffer and
analyzed by Western blot with the exon A antibody. The use of twice as
much P-Ser949 antibody and exon A antibody
immunoprecipitated the same amount of PAM. Cell extracts,
immunoprecipitates, and sucrose gradient samples were fractionated on
polyacrylamide (10, 15, or 4-15%), 0.25%
N,N'-methylene-bisacrylamide/SDS gels (34). Proteins
transferred to Immobilon-P membranes (Millipore) were visualized with
one of several primary antisera and the Amersham Pharmacia Biotech ECL
Kit: PAL (JH471; 1:1000), PHM (JH1764; 1:1000), exon A (JH629; 1:1000),
or P-Ser949 (JH2541, 1:250) (21).
Metabolic Labeling and Immunoprecipitation--
By using
[35S]Met labeling mix (Amersham Pharmacia Biotech) or
[32P]PO
, cells were
metabolically labeled, and pulse-chase experiments were performed (10,
11, 24, 35). Ice-cold 20 mM sodium
N-tris[hydroxymethyl]methyl-2-aminoethanesulfonic acid, 10 mM mannitol, and 1% Triton X-100 (pH 7.4) buffer with protease inhibitors plus protein phosphatase inhibitors (5 mM EGTA, 5 mM EDTA, 1 mM sodium
orthovanadate, 10 mM sodium pyrophosphate and 50 mM NaF) was used for cell extraction. Immunoprecipitates were prepared as described (10, 24). Films were analyzed using Scion
Image (Scion Corp.).
In Vitro Kinase Assay--
Phosphorylation of recombinant PAM-CD
was carried out as described (25). The reaction volume was 30 µl and
consisted of the following reagents: kinase buffer (50 mM
HEPES, 1 mM MgCl2 (pH 7.5), 0.1% Thesit); 5 µg of recombinant PAM-CD; ± 100 µM DRB; purified
recombinant P-CIP2, CKII (Calbiochem), or protein kinase A
(Calbiochem); and 1 or 2 µCi of [-32P]ATP (3000 Ci/mmol, Amersham Pharmacia Biotech). Reactions were incubated at
37 °C for 30 or 60 min, fractionated by SDS-PAGE (15%
polyacrylamide), transferred to Immobilon-P membranes, and visualized
by autoradiography.
| |
RESULTS |
Recombinant PAM-CD Is Phosphorylated on Thr946 and
Ser949 by Casein Kinase II--
Since acidic clusters
often play a crucial role in the trafficking of membrane proteins and
may serve as sites for phosphorylation by CKII, we wanted to determine
whether the cytosolic domain of PAM is a substrate for CKII. Purified
recombinant PAM-CD incubated with [-32P]ATP and sea
star CKII underwent phosphorylation primarily on Ser residues (data not
shown). To identify the phosphorylation sites(s), endoproteinase
LysC-digested samples were subjected to gel filtration. The single
major peak observed was then applied to an anion exchange high pressure
liquid chromatography resin, yielding two major peaks, A and B (Fig.
1B). Edman degradation of each
peak yielded NH2-terminal sequences commencing with
Gly940-Phe941-; [32P] was
released from peak A primarily at cycle 10, which corresponded to
Ser949, whereas [32P] was released from peak
B at both cycles 7 and 10, corresponding to Thr946 and
Ser949 (Fig. 1C). To investigate the role of
phosphorylation at these two sites, vectors encoding PAM-1 in which
both residues were mutated to Ala (PAM-1 TS/AA) or to Asp (PAM-1 TS/DD)
were created and individually transfected into AtT-20 cells.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 1.
In vitro phosphorylation of PAM-1
with CKII. A, the sequences of the cytosolic,
COOH-terminal domains of rat (RPAM;
GenBankTM/EBI accession number P14925) and
Xenopus (XPAM; GenBankTM/EBI
accession number S17855) PAM, mouse cation-dependent
mannose 6-phosphate receptor (MCDM6PR;
GenBankTM/EBI accession number P24668), human furin
(HFUR; GenBankTM/EBI accession number KXHUF),
mouse prohormone convertase 6B (MPC6B;
GenBankTM/EBI accession number BAA04507), and human
carboxypeptidase D (HCPD; GenBankTM/EBI
accession number NP001295) were aligned based on their acidic clusters
(identified in bold) using ClustalW multiple sequence
alignment (BCM Search Launcher). Gap opening penalty was set at 10, and
the Gap extension penalty was 0.05. Tyr or Phe residues of importance
in trafficking are indicated by Y or
F, respectively. Known phosphorylated residues
are shaded. The peptide to which the
P-Ser949-specific antibody was raised is shown in
italics in RPAM. B, recombinant PAM-CD
phosphorylated by purified sea star CKII was digested with
endoproteinase LysC, and peptides were fractionated by gel filtration.
The single peak of radioactivity was pooled and fractionated on a
Bio-Sil TSK-DEAE column. C, pools A and
B were subjected to Edman degradation with 60% of each
phenylthiohydantoin fraction counted for 32P radioactivity.
Plot shows counts obtained at successive cycles. The major residue
released at each cycle is indicated, Thr946 and
Ser949.
|
|
PAM-1 TS/AA and PAM-1 TS/DD Mutants Display an Altered Steady-state
Localization--
The steady state localization of each mutant PAM-1
protein in AtT-20 cells was compared with that of PAM-1. As shown
previously, wild-type PAM-1 was localized in the perinuclear region of
the cell (Fig. 2A, top panel, wide
arrows) in a compact structure identified as the TGN on the basis
of immunoelectron microscopy (11, 22, 26). Slight staining at the tips
of the AtT-20 PAM-1 cells (narrow arrows) was also observed.
Immunostaining in the PAM-1 TS/AA cells (Fig. 2A, middle
panel) was only subtly different, with signal observed in the
perinuclear region (wide arrow) along with more diffuse
vesicular staining throughout the cell and more readily detectable
staining at the tips of processes (narrow arrows). However,
in the PAM-1 TS/DD cells (Fig. 2A, bottom panel),
immunostaining for PAM was clearly more widely distributed, with
vesicular structures observed at the tips of the processes (narrow arrows) along with intense staining of what
generally appeared to be more diffuse structures in the perinuclear
region (wide arrows). Thus, PAM-1 TS/DD has a unique
steady-state localization in AtT-20 cells.
View larger version (101K):
[in this window]
[in a new window]
|
Fig. 2.
Steady-state localization of PAM-1, PAM-1
TS/AA, and PAM-1 TS/DD. AtT-20 cells expressing PAM-1, PAM-1
TS/AA, or PAM-1 TS/DD were visualized by indirect immunofluorescence
using a PHM antibody (PHM Ab; A). AtT-20 PAM-1
cells were incubated with CSFM-Air medium with (+Noc) or
without ( Noc) 10 µM nocodazole for 20 min
before immunostaining with either a syntaxin 6 monoclonal antibody
(Syntaxin 6 mAb; B, top panels) or a TGN38
polyclonal antibody (TGN38 Ab; B, bottom panels).
C, the three AtT-20 cell types were incubated with 10 µM nocodazole for 20 min and immunostained with the exon
A antibody before visualization (exon A Ab). Immunostaining
in the perinuclear region is indicated by wide arrows and in
the cellular processes and tips by narrow arrows. All cells
were photographed under the same conditions. The scale bar
for all photographs is shown in the bottom panel of
C.
|
|
Since the TGN and late/recycling endosomes are both localized in the
perinuclear region and cannot be distinguished immunocytochemically (36), we used nocodazole, a microtubule-destabilizing drug that causes
dispersal of the late/recycling endosomes from the TGN and the eventual
dispersal of the TGN, to compare the three proteins (37-39). The
effects of nocodazole on the TGN depend on drug dosage, time of
treatment, the marker protein analyzed, and cell type (40-42).
Consistent with its demonstrated localization to the TGN, much of the
PAM-1 was still localized to a compact structure in the perinuclear
region following a short, low-dose incubation with nocodazole for 20 min (Fig. 2C, top panel, wide arrows). A similar structure
was also visualized by antisera to two TGN markers, syntaxin 6 and
TGN38 (Fig. 2B, wide arrows) (43, 44). Staining for syntaxin
6, which is localized to a discrete region of the TGN, was not affected
by this nocodazole treatment (Fig. 2B, top panels, wide
arrows). However, following treatment of the AtT-20 PAM-1 cells
with nocodazole, staining for TGN38 was localized to distinct vesicular
structures (Fig. 2B, bottom panel, narrow arrows).
Furthermore, nocodazole treatment did increase PAM staining in
vesicular structures distributed throughout the cells (Fig. 2C,
top panel, narrow arrow), and this effect increased with a longer
nocodazole treatment time (data not shown).
Nocodazole treatment affected PAM-1 TS/AA cells in much the same way it
affected the PAM-1 cells. Some PAM-1 TS/AA remained in the perinuclear
area (Fig. 2C, middle panel, wide arrows), coincident with
syntaxin 6 (data not shown), and PAM-1 TS/AA staining increased in
vesicular structures dispersed throughout the cell (Fig. 2C,
middle panel, narrow arrows). In contrast, nocodazole treatment brought about a dramatic alteration in the localization of
PAM-1 TS/DD. Although syntaxin 6 staining was unaffected (data not
shown), little perinuclear staining for PAM-1 TS/DD was observed following nocodazole treatment (Fig. 2C, bottom panel, wide
arrow). Broadly distributed PAM-1 TS/DD-containing vesicles were
present throughout the cell, with staining apparent in the processes
and along the margins of the cells (Fig. 2C, bottom panel, narrow arrows). Overall, these data suggest that PAM-1 TS/DD, unlike PAM-1, is largely localized to a nocodazole-sensitive,
microtubule-dependent endosomal compartment at steady state.
PAM-1 Is Phosphorylated on Ser949--
PAM-1
synthesized in AtT-20 cells is phosphorylated primarily on Ser residues
(10). To determine if Thr946 and/or Ser949
represent major phosphorylation sites in the CD of PAM, AtT-20 PAM-1
and PAM-1 TS/DD cells were biosynthetically labeled with medium
containing either [35S]Met or
[32P]PO. Full-length
PAM-1 and PAM-1 TS/DD were immunoprecipitated from
[35S]Met and
[32P]PO-labeled cell
extracts (Fig. 3A, 120 kDa). The mutation of these two potential phosphorylation sites
did not eliminate the phosphorylation of PAM, with similar ratios of
32P-labeled to 35S-labeled protein observed for
PAM-1 and for PAM-1 TS/DD. Therefore, the phosphorylation of
Thr946 and/or Ser949 accounts for only a small
fraction of PAM phosphorylation at steady state.
View larger version (64K):
[in this window]
[in a new window]
|
Fig. 3.
PAM-1 phosphorylation and characterization of
the P-Ser949 antibody. A, AtT-20 PAM-1 and
PAM-1 TS/DD cells were biosynthetically labeled with either 250 µCi
of [35S]Met or
[32P]PO for 2 h.
Cell extracts were immunoprecipitated with the polyclonal PHM antibody
(JH1764) and fractionated by SDS-PAGE. B, AtT-20 PAM-1,
PAM-1 TS/AA and PAM-1 TS/DD cells were fixed and visualized with the
P-Ser949 antiserum. Immunostaining in the perinuclear
region is indicated by wide arrows and in the tips of the
cellular processes by narrow arrows. All cells were
photographed under the same conditions; the scale bar is
shown in the 3rd panel. C, extracts of AtT-20 PAM-1 (1.3 µg), PAM-1 TS/AA (25 µg), and PAM-1 TS/DD (0.9 µg) cells were
subjected to Western blot analysis and visualized with the PAL and
P-Ser949 antisera. Due to the larger amount of PAM-1 TS/AA
cell extract loaded, a nonspecific band (*) was apparent with the
P-Ser949 antibody only in this sample. D, AtT-20
PAM-1 cells were extracted and the supernatant centrifuged and
immunoprecipitated with either the exon A antibody or the
P-Ser949 antiserum. An input sample and the
immunoprecipitates (IPT) were fractionated on a 4-15% SDS
gel and analyzed by Western blot using the exon A antibody. The
percentage of the total extract loaded on the gel is shown.
|
|
To determine if Ser949 in the PAM-CD is phosphorylated
in vivo, a polyclonal antibody specific for
P-Ser949 was generated (Fig. 1, italics). After
depleting the P-Ser949 serum of antibody against
non-phosphorylated PAM-1(942-953), a solid phase assay demonstrated
that the antiserum recognized peptide phosphorylated at
Ser949 but exhibited no cross-reactivity with the
non-phosphorylated peptide or with
PAM-1(942-953)P-Thr946/P-Ser949. The
specificity of the P-Ser949 antiserum was examined further
by immunostaining (Fig. 3B) and Western blot analysis (Fig.
3C) of cells expressing PAM-1, PAM-1 TS/AA, or PAM-1 TS/DD.
AtT-20 PAM-1 cells immunostained with the P-Ser949
antiserum exhibited staining in a compact, perinuclear region (Fig.
3B, 1st panel, wide arrows) that
corresponded to immunostaining with the syntaxin 6 antibody (data not
shown). Only a small amount of P-Ser949-specific staining
was observed at the tips of the PAM-1 cells (Fig. 3B,
1st panel, narrow arrows). Importantly, no signal was observed in PAM-1 TS/AA cells (Fig. 3B, middle
panel). Thus, the P-Ser949 antibody can be used to
demonstrate that a fraction of PAM-1 is phosphorylated at
Ser949 in AtT-20 cells.
To our surprise, the P-Ser949 antiserum visualized PAM-1
TS/DD, with staining observed in the perinuclear region and tips of the
cells (Fig. 3B, 3rd panel, wide and narrow
arrows), as observed with the PHM antibody (Fig. 2C).
It is not unprecedented for a phosphopeptide-specific antibody to
recognize an aspartate-substituted peptide. Previously, we demonstrated
that another PAM-1-specific P-Ser antibody (P-Ser937)
detected PAM-1 in which Ser937 was replaced by
Asp937 (11). Overall, these data suggest that the
P-Ser949 antibody recognizes a conformation imparted by the
negative charge present at position 949 rather than the presence of
P-Ser949.
For a final test of antibody specificity, extracts of PAM-1, PAM-1
TS/AA, and PAM-1 TS/DD cells were analyzed by Western blot. Intact
120-kDa PAM and 70-kDa membrane PAL were detected in all three cell
lines with the PAL antibody (Fig. 3C). The
P-Ser949 antiserum also detected intact PAM and membrane
PAL in the PAM-1 and PAM-1 TS/DD cells. The P-Ser949
antiserum did not detect PAM-1 TS/AA. Thus, the P-Ser949
antiserum has the required specificity to establish that PAM-1 is
phosphorylated on Ser949 in AtT-20 cells.
The P-Ser949 antiserum was used to evaluate the extent to
which Ser949 on the PAM-CD is phosphorylated. Intact
120-kDa PAM-1 and 70-kDa membrane PAL were detected with the exon A
antibody in the supernatant (input) produced from extracted PAM-1 cells
centrifuged at 304,000 × g and from immunoprecipitates
with the exon A and P-Ser949 antibodies (Fig.
3D). By using Scion Image for quantitation, we were able to
determine that less than 5% of the 120-kDa PAM-1 and 70-kDa PAL are
phosphorylated on Ser949.
PHM Produced from PAM-1 TS/AA and PAM-1 TS/DD Is Active and Varies
in Its Responsiveness to a Secretagogue--
Enzyme assays and Western
blots were used to compare protein expression in the different cell
lines. Extracts of AtT-20 PAM-1, PAM-1 TS/AA, and PAM-1 TS/DD cells
were assayed for PHM activity, and aliquots containing equal units of
PHM activity (150 pmol/h) were compared (Fig.
4A). PAL produced from each
cell line was also analyzed and determined to be active (data not
shown). While both enzymatic domains of PAM-1 are catalytically active,
endoproteolytic cleavages that release monofunctional PHM from membrane
PAM result in an increase in the turnover number of the monooxygenase
(45). The PHM antibody identified 120-kDa PAM-1 and 45 kDa PHM in all three cell lines, whereas the PAL antibody detected 120-kDa PAM-1 and
70-kDa PAL. Similar amounts of 120-kDa PAM-1, 70-kDa PAL, and 45-kDa
PHM were observed for the PAM-1 and PAM-1 TS/DD cells. In contrast, the
products generated from 120-kDa PAM-1 TS/AA were distinctly different;
less 70-kDa PAL was present, and processing of 120-kDa PAM-1 TS/AA into
monofunctional PHM was more extensive. The total amount of protein
detected by the PHM antiserum in the PAM-1 TS/AA extract was reduced,
suggesting that PHM produced from PAM-1 TS/AA has a higher specific
activity than PHM produced from PAM-1 or PAM-1 TS/DD. Elucidation of
the underlying mechanism will require an analysis of intact PAM-1 TS/AA
and the monofunctional PHM produced from it.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 4.
Mutant PAM-1 catalytic activity and
secretion. A, aliquots containing equimolar PHM
activity (150 pmol/h) from extracts of AtT-20 PAM-1, PAM-1 TS/AA, and
PAM-1 TS/DD cells were fractionated by SDS-PAGE and analyzed by Western
blot using the PHM and PAL antibodies. The specificity of each
antiserum is shown between the Western blots. B, AtT-20
PAM-1, PAM-1 TS/AA, and PAM-1 TS/DD cells were washed and incubated in
CSFM-Air for three 30-min periods. Medium was collected following two
additional 30-min incubations in CSFM-Air and assayed for basal
secretion of PHM. The cells then were incubated for 30 min in CSFM-Air
containing 1 mM BaCl2 (Stimulation).
The levels of PHM from the two basal secretion samples (Average
Basal) were averaged separately for each cell line. Experiments
were performed in triplicate.
|
|
The expression of PAM-1 is known to alter the ability of AtT-20 cells
to store soluble proteins in secretory granules, perhaps through the
interaction of PAM-1 with cytosolic proteins that regulate the actin
cytoskeleton (26, 46). Since entry into secretory granules is essential
for regulated secretion, we quantified the basal and
BaCl2-stimulated secretion of PHM activity from the three
cell lines (Fig. 4B). Secretion of PHM activity is
stimulated 1.8-fold in PAM-1 and PAM-1 TS/AA cells but 3.1-fold in
PAM-1 TS/DD cells. Overall, these data demonstrate that the PAM-1 TS/DD cells are able to store the mutant PAM protein in regulated secretory granules to a greater extent than either the PAM-1 or the PAM-1 TS/AA cells.
The PAM-1 Thr946/Ser949 Sites Affect PAM-1
Degradation and Storage--
The metabolism of PAM-1 TS/AA and PAM-1
TS/DD was compared with that of PAM-1 using pulse/chase metabolic
labeling and an antibody specific for PHM. As expected, a 117-kDa PAM
protein was immunoprecipitated from extracts of all three cell lines
following the pulse (Fig. 5A, 0 h). In all three cell lines, the newly synthesized 117-kDa protein
increased in size to 120 kDa during the 1-h chase, presumably due to
oligosaccharide maturation (10, 23, 28). A paucity of methionine
residues in PAL (only 6 of the 26 Met residues in PAM-1 are in 70-kDa
PAL) makes its detection during chase incubations difficult, so
quantification is based on the immunoprecipitation of PHM (11).
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 5.
Metabolism of mutant PAM-1.
A, wild-type and mutant PAM-1 cells were biosynthetically
labeled with 250 µCi of [35S]Met for 15 min and either
immediately extracted (0 h) or chased for 1, 2, or 4 h
in label-free medium before extraction. Cell extract and medium
(M) samples were immunoprecipitated with the PHM antibody.
B and C, data from 4 to 5 experiments were
quantified using Scion Image. Since the amount of labeled 120-kDa PAM
protein often increased during the 1st h of chase, the amount of newly
synthesized PAM-1 protein present after the 1-h chase was set to 100%.
B, the total recovery of PAM protein after the 4-h chase was
determined by adding the amount of 120-kDa protein in the cell extract
to the total amount of 45-kDa PHM present in both the cell extract and
medium. C, total recovery of 120-kDa PAM and 45-kDa PHM in
the cell extract (cell) and medium (mdm) is
shown.
|
|
Newly synthesized 120-kDa PAM-1 is cleaved to 45-kDa PHM only after it
enters immature secretory granules (24). As shown previously, 45-kDa
PHM was first detected in AtT-20 PAM-1 cells following the 2-h chase,
with more 45-kDa PHM present after the 4-h chase (Fig. 5A,
1st panel) (11). Similarly, 45-kDa PHM was first detected in
the PAM-1 TS/AA cells after the 2-h chase (Fig. 5A, middle
panel). In contrast, in the PAM-1 TS/DD cells, cleavage of 120-kDa
PAM-1 to 45-kDa PHM was detected earlier, following the 1-h chase,
suggesting enhanced access of the mutant protein to the cleavage
competent compartment (Fig. 5A, last panel). Whereas newly
synthesized 45-kDa PHM accumulated in the medium of PAM-1 and PAM-1
TS/DD cells during the 4-h chase, very little newly synthesized 45-kDa
PHM was secreted from the PAM-1 TS/AA cells.
Pulse/chase metabolic labeling followed by quantitative
immunoprecipitation of cell extracts and media allows one to quantify both secretion and turnover. We quantified the amount of radiolabeled PAM protein recovered after the 4-h chase; although approximately two-thirds of the newly synthesized protein was recovered from the
PAM-1 and PAM-1 TS/DD cells, only one-third was recovered from the
PAM-1 TS/AA cells (Fig. 5B, p 
0.05). The amount of newly synthesized 120-kDa PAM protein remaining in the cells after 4 h was similar in the PAM-1 and PAM-1 TS/DD cells (Fig. 5C,
open bars). In contrast, the PAM-1 TS/AA cells contained little
intact PAM-1 after the 4-h chase (p 0.04). The
45-kDa PHM generated from PAM-1 can be stored (Fig. 5C,
hatched bars) or secreted (Fig. 5C, filled bars);
substantially more newly synthesized 45-kDa PHM is stored in PAM-1
TS/DD cells (p 0.08) than in PAM-1 or PAM-1 TS/AA
cells. The basal secretion of newly synthesized PHM from PAM-1 and
PAM-1 TS/DD cells was similar, whereas PAM-1 TS/AA cells secreted less
45-kDa PHM (p 0.04). Overall, these data suggest
that phosphorylation of Thr946/Ser949 in the
PAM-CD both increases the cleavage of PAM-1 and its storage.
The Aggregation of PAM-1 TS/AA and PAM-1 TS/DD Is Differentially
Affected by pH--
The pH-dependent aggregation of
proteins like furin (47), carboxypeptidase E (48, 49), prohormone
convertase 2 (50), chromogranins A and B (51-54), insulin (55), and
prolactin (56) is thought to affect trafficking through the regulated
secretory pathway. In a recent study, we used an in vitro,
detergent-solubilized system to demonstrate that PAM-1 aggregates
more at pH 5.5 than at pH 7.5 (57). In this simplified
single compartment system, the juxtamembrane domains of PAM play a key
role in its aggregation. Therefore, we compared the
pH-dependent aggregation of solubilized PAM-1, PAM-1 TS/AA,
and PAM-1 TS/DD using linear sucrose gradients at pH 7.5 to mimic
conditions in the Golgi and at pH 5.5 to mimic conditions in the lumen
of mature secretory granules (58-60).
At pH 7.5 (Fig. 6A), PAM-1
fractionated as expected based on its monomeric molecular weight of
120,000; only 6% of the total intact PAM-1 was recovered from the
bottom of the gradient (particulate matter). At pH 5.5, 17% of the
intact PAM-1 sedimented to the bottom of the gradient. In comparison,
almost none (1.7%) of the intact PAM-1 TS/AA pelleted at pH 7.5, whereas 41% pelleted at pH 5.5 (Fig. 6B). Most strikingly,
11% of the intact PAM-1 TS/DD was recovered from the bottom of the
gradient at pH 7.5, whereas 13% of the intact PAM-1 TS/DD was pelleted
at pH 5.5 (Fig. 6C). The increased ability of PAM-1 TS/DD to
aggregate at neutral pH could facilitate its entry into immature
secretory granules, whereas the inability of PAM-1 TS/AA to aggregate
at pH 7.5 could result in its exclusion from secretory granules and its
trafficking to degradative pathways. If the behavior of PAM-1 TS/DD
mimics the effects of phosphorylation at this site,
pH-dependent aggregation and secretory granule entry may be
facilitated by phosphorylation.
View larger version (81K):
[in this window]
[in a new window]
|
Fig. 6.
Sucrose gradient sedimentation of PAM-1 and
mutant PAM-1 extracts. AtT-20 PAM-1 (A), PAM-1 TS/AA
(B), and PAM-1 TS/DD (C) cells were extracted in
buffer with protease inhibitors and 1% Triton X-100 (pH 7.5) and
centrifuged at 250,000 × g at 4 °C. Aliquots (200 µl) of this supernatant were further fractionated on 5-20% linear
sucrose gradients buffered either to pH 7.5 or pH 5.5 with 50% sucrose
"pads" placed at the bottom of the gradients. Fractions were
collected from the top of the gradients down, with the pellet
(P) fraction solubilized from the bottom of each gradient.
Western blot analysis of the fractions was done using the exon A
antibody; only the 120-kDa region of the blot is shown. Marker proteins
of known Svedberg value and native molecular weight (81) were analyzed
at the same time as follows: cytochrome c (12.6 kDa, 1.8 S),
ovalbumin (44 kDa, 3.6 S), bovine serum albumin (66 kDa, 5 S), catalase
(250 kDa, 11.4 S), and apoferritin (450 kDa, 17.6 S).
|
|
AtT-20 PAM-1 TS/AA and PAM-1 TS/DD Cells Have Distinct
Internalization Phenotypes--
The internalization of antibodies
directed against a lumenal domain of PAM was used to assess endocytic
trafficking in the PAM-1 TS/AA and PAM-1 TS/DD cells (11, 20). We
manipulated temperature as a simple way to halt the internalized
PAM-antibody complexes in specific cellular compartments. At 4 °C,
PAM-antibody complexes remain on the cell surface, and internalization
is blocked (29). Raising the temperature to 20 °C allows entry of
the PAM-antibody complexes into both early and late endosomes. As shown
in Fig. 7, A-C, all three
cell types had similar immunostaining profiles when the temperature was
raised to 20 °C, with large vesicular structures present in the
processes and tips of the cells (narrow arrows) and in the
perinuclear region (wide arrows).
View larger version (82K):
[in this window]
[in a new window]
|
Fig. 7.
Internalization of
PAM-antibody-complexes. A-C, following incubation with
PAL antibody at 4 °C for 60 min (pulse [P] 60 min),
cells were rinsed with antibody-free CSFM and maintained at 20 °C
for 60 min. The cells then were fixed and observed using a
Cy3-conjugated secondary antiserum. D-I, cells
were incubated with PAL antibody at 20 °C for 60 min (pulse
[P] 60 min) and then chased (C) for 30 (D-F, P60 min 20 °C/C30 min
37 °C) or 60 min (G-I, P60 min 20 °C/C60
min 37 °C) at 37 °C in antibody-free medium. Immunostaining
in the perinuclear region is indicated by wide arrows and in
the cellular processes and tips by narrow arrows. All cells
were photographed under the same conditions. The scale bar
for all photographs is shown in I.
|
|
Internalization steps beyond this point were then examined (Fig. 7,
D-I). Cells were incubated with PAL antibody at
20 °C for 60 min to load both early and late endocytic compartments with PAM-antibody complexes; as expected, no differences were apparent
(data not shown). Cells were then chased in antibody-free medium at
37 °C, permitting the complexes to move out of the endocytic compartments. Distinct internalization profiles were observed for all
three cell types. In AtT-20 PAM-1 cells, a 30-min chase at 37 °C
allowed efficient collection of the internalized PAM-antibody complexes
in the perinuclear TGN region (Fig. 7D, wide arrow) with
only a few large vesicular structures remaining spread throughout the
cell body and processes (Fig. 7D, narrow arrow). Fewer
vesicular structures were present following the 60-min chase (Fig.
7G).
The staining profile in the PAM-1 TS/AA cells was, however, very
different from that in the PAM-1 cells (Fig. 7, E and
H). Large immunostained vesicular structures remained spread
throughout the cell with a heavy concentration in the processes and
tips (narrow arrows), following both the 30- and 60-min
chases. Furthermore, the intensity of the signal decreased
substantially following the 60-min chase, suggesting that degradation
or recycling of the PAM-1 TS/AA-antibody complexes to the cell surface
had occurred. The PAM-1 TS/DD cells (Fig. 7F) exhibited an
intermediate staining pattern after the 30-min chase at 37 °C, with
large vesicular structures still spread throughout the cellular
processes and tips (narrow arrows) and some accumulation of
complexes in the perinuclear region (wide arrow). This trend
toward collection in the perinuclear region (wide arrows)
continued following the 60-min chase, yet a significant number of
vesicular structures remained spread throughout the cells (Fig.
7I, narrow arrows). Overall, these data suggest that both
phosphorylation and dephosphorylation of Thr946 and/or
Ser949 play key roles in routing to the perinuclear region
from late endosomes. Neither modification seems to play a key role
earlier in the endocytic pathway.
Internalized PAM-1 TS/DD Is Not Recycled to the TGN--
Since the
bulk of the internalized PAM-antibody complexes recycle to the
perinuclear region in both the PAM-1 and PAM-1 TS/DD cells, we again
used nocodazole to distinguish between PAM-antibody complexes that had
recycled to the TGN (short incubation time, low-dose
nocodazole-insensitive) versus PAM-antibody complexes that
remained in the late/recycling endosomes (nocodazole-sensitive) (22,
37, 38). Both the TGN and late/recycling endosomes localize to the
perinuclear region (36).
AtT-20 PAM-1 and PAM-1 TS/DD cells were incubated with exon A antibody,
and PAM-antibody complexes were internalized for 60 min at 20 °C to
load the early and late endocytic compartments while preventing exit
from the late endocytic compartments (Fig. 8, A-D). Nocodazole was
introduced only after the first incubation with antibody, since it can
inhibit the maturation of endosomes (61, 62). The cells were then
incubated with 10 µM nocodazole in antibody-free CSFM-Air
for 60 min (Fig. 8) at 37 °C. In the PAM-1 cells, a significant
fraction of the internalized PAM-antibody complexes co-localized with
the syntaxin 6 marker (Fig. 8, A and B, wide
arrows), indicating that this dose of nocodazole in AtT-20 cells
still allowed proteins to move from late endosomes to the TGN. Small
vesicular structures remained spread throughout the nocodazole-treated
PAM-1 cells (Fig. 8A, narrow arrows), suggesting that
nocodazole is retarding trafficking from late endosomes to the TGN. In
the PAM-1 TS/DD cells, only a small fraction of the internalized
PAM-antibody complexes co-localized with the syntaxin 6 marker after
the 1-h chase; instead, the PAM-antibody complexes remained in distinct
punctate structures spread throughout the cells including the processes
and tips. These data suggest that PAM-1 TS/DD remains in the more
nocodazole-sensitive late or recycling endosomes and cannot recycle to
the TGN as well as PAM-1. Assuming that PAM-1 TS/DD can mimic PAM-1
phosphorylated at Thr946 and/or Ser949, these
data support a role for the dephosphorylation of Thr946
and/or Ser949 in the exit of PAM from the
nocodazole-sensitive, late endosomes.
View larger version (67K):
[in this window]
[in a new window]
|
Fig. 8.
Nocodazole causes a redistribution of
internalized PAM-antibody complexes. AtT-20 PAM-1 and PAM-1 TS/DD
cells were incubated with exon A antibody at 20 °C for 60 min (pulse
[P] 60 min), rinsed with antibody-free CSFM, and chased
(C) in CSFM containing 10 µM nocodazole for 60 min (A and C, P60 min 20 °C/ C60
min 37 °C +Noc). Following antibody internalization, the cells
were fixed in 100% methanol, double immunostained with the syntaxin 6 antibody (B and D), and visualized using indirect
immunofluorescence. Wide arrows indicate immunostaining in
the perinuclear region, and narrow arrows show staining in
the cellular processes and tips. All cells were photographed under the
same conditions, and the scale bar for all photographs is
shown in D.
|
|
DRB Prevents Internalized PAM-1 from Recycling to the TGN--
To
understand better the trafficking steps affected by the phosphorylation
of Thr946 and/or Ser949, we looked for
pharmacological agents that could be used on intact cells. To this end,
the effectiveness of the CKII inhibitor, DRB, was first evaluated in
test tube assays using recombinant PAM-CD protein (Fig.
9A). When tested at 100 µM, DRB effectively reduced the CKII-mediated
phosphorylation of recombinant PAM-CD to ~38% of the control.
However, DRB proved to be an even more potent inhibitor of P-CIP2, a
kinase endogenous to AtT-20 cells that phosphorylates PAM at
Ser949, reducing the phosphorylation of recombinant PAM-CD
to 5.6% of the control (25). Finally, protein kinase A-mediated
phosphorylation of recombinant PAM-CD was reduced to ~59% of the
control by 100 µM DRB.
View larger version (102K):
[in this window]
[in a new window]
|
Fig. 9.
DRB inhibits P-CIP2 and CKII
phosphorylation and prevents PAM-1 recycling. A,
recombinant PAM-CD was phosphorylated with purified recombinant
P-CIP2, CKII (Calbiochem), or protein kinase A (PKA;
Calbiochem) either in the absence (0 µM) or presence (100 µM) of the kinase inhibitor, DRB. Reactions were
incubated at 37 °C, fractionated by SDS-PAGE on 15% polyacrylamide
gels, transferred to Immobilon-P membranes, and visualized by
autoradiography. B-G, AtT-20 PAM-1 cells were incubated
with PAL antibody for 60 min at 20 °C. Control cells are shown in
B-D. Treated cells (E-G) were incubated with
DRB (100 µM) for the final 30 min of incubation with PAL
antibody at 20 °C. Next, the cells in B and E
were immediately fixed and prepared for immunostaining as described,
whereas in the remaining panels, cells were warmed to 37 °C and
chased for either 30 (C and F) or 60 min
(D and G) in antibody-free medium. Cells in
F and G were chased in the presence of 100 µM DRB. Wide arrows indicate immunostaining in
the perinuclear region, and narrow arrows show staining in
the cellular processes and tips. All cells were photographed under the
same conditions; the scale bar is shown in
G.
|
|
If a DRB-sensitive protein kinase (P-CIP2, CKII, etc.) is involved in
the endocytic trafficking of PAM, its role should be most apparent as
proteins move out of late endosomes. To test this possibility, control
AtT-20 PAM-1 cells were incubated with exon A antibody, and
PAM-antibody complexes were internalized for 60 min at 20 °C to load
late endocytic compartments (Fig. 9, B-D). As expected,
vesicular structures containing PAM-antibody complexes were distributed
throughout the cells (Fig. 9B, wide and
narrow arrows). As observed using the PAL antibody (Fig. 7, D and G), after warming the control cells to
37 °C and chasing for 30 or 60 min, the PAM-antibody complexes
proceeded to collect in the TGN (Fig. 9, C and D, wide
arrows).
To assess the effect of DRB on the endocytic trafficking of
PAM-antibody complexes, 100 µM DRB was added to PAM-1
cells that had been incubated for 30 min at 20 °C with exon A
antibody (Fig. 9, E-G). Since DRB lacks specificity, PAM-1
TS/DD cells were treated in a similar fashion as a control. We reasoned
that any effects of DRB on other phosphorylation events would still be
apparent with the mutant PAM-1 TS/DD protein (data not shown). At the
end of a 30-min (20 °C) incubation with DRB, the DRB-treated PAM-1 cells differed only slightly from control cells (Fig. 9E);
PAM antibody-containing vesicles were more dispersed and more prevalent at the tips of processes (narrow arrows). When the
DRB-treated PAM-1 cells were chased for 30 or 60 min in the presence of
the drug at 37 °C, the cells showed little ability to collect
internalized PAM-antibody complexes in the TGN region (Fig. 9,
F and G). PAM antibody-containing vesicles
remained dispersed throughout the cells, with some concentration at the
tips of processes (narrow arrows). DRB had no effect on the
PAM-1 TS/DD cells (data not shown). Therefore, both our pharmacological
and mutagenesis studies indicate the importance of the phosphorylation
and dephosphorylation of Thr946 and/or Ser949
in trafficking from late endosomes into the TGN.
| |
DISCUSSION |
PAM Is Phosphorylated at Ser949--
For many membrane
proteins, phosphorylation of Ser/Thr residues located in acidic
clusters in cytosolic COOH-terminal domains plays a crucial role in
trafficking (6). The region between Lys953 and
Lys971 in the PAM-CD is extremely acidic and has typically
been aligned with known acidic cluster trafficking signals. However,
since truncation of PAM in the middle of this acidic region does not alter its trafficking, this acidic region is not a key trafficking determinant for PAM (20). We demonstrate in this paper that Thr946 and Ser949 in the CD of PAM can be
phosphorylated by CKII and by P-CIP2, transiently forming a short
acidic cluster
(Thr946-Glu-Gly-Ser949-Asp-Gln-Glu), and
thus play a role in the trafficking of integral membrane PAM.
We used an antibody specific for P-Ser949 to demonstrate
that PAM-1 expressed in AtT-20 cells is phosphorylated at this site. Based on SDS-PAGE, both intact PAM-1 and the 70-kDa PAL fragment derived from it can be phosphorylated at Ser949. Based on
the immunoprecipitation of PAM-1 with the P-Ser949-specific
antibody, less than 5% of Ser949 is phosphorylated at
steady state. Nevertheless, the small fraction of PAM-1 that is
phosphorylated on Ser949 is localized to a discrete
perinuclear area. Similarly, the single CKII phosphorylation site in
CD-Man-6-PR (EESEERDD) serves as a plasma membrane sorting signal
despite the fact that only 3% of the protein is phosphorylated at this
site at steady state (13).
Mutagenesis studies suggest a role for the phosphorylation of
Thr946 and Ser949 in the PAM-CD in the
activation of PHM and access to secretory granules and endocytosis. The
ability of the COOH-terminal domain of PAM to affect the specific
activity of PHM was first appreciated when we observed an increase in
PHM activity following partial proteolytic digestion of PAM (21). Based
on the ability of our P-Ser949 antibody to recognize a
conformation common to the PAM-1 TS/DD mutant protein and the increased
aggregation of PAM-1 TS/DD at neutral pH, phosphorylation at this site
may bring about a conformational change that affects the catalytic
activity of the lumenal PHM domain.
The enzyme or enzymes responsible for phosphorylating PAM at
Ser949 in vivo have not been identified. Both
CKII and P-CIP2 could be responsible for the phosphorylation of PAM at
Ser949 and are expressed in many of the tissues that
express PAM (63, 64). CKII, a highly conserved, ubiquitously expressed
heterotetramer composed of catalytic ( and ') and regulatory
() subunits, phosphorylates a wide range of substrates (65, 66). In
contrast, P-CIP2 is a highly selective kinase whose only known
substrate is Ser949 in the PAM-CD (25). Both P-CIP2 and
CKII are inhibited by DRB, which blocks a late stage in the endocytosis
of PAM-1 (Fig. 9). The interaction of PAM-1 with P-CIP2 is complex. The
P-CIP2-binding sites identified in the PAM-CD with the yeast two-hybrid
system (K919R, L926Q, and F929A/F930A) are distinct from the residue phosphorylated by P-CIP2 (25, 26, 64). In addition, PAM mutants unable
to interact with P-CIP2 in a yeast two-hybrid assay can still be
phosphorylated by P-CIP2. Finally, these P-CIP2 phosphorylated PAM
mutants as well as PAM mutants that cannot be phosphorylated exhibit
altered secretory phenotypes (26).
Phosphorylation of Thr946 and Ser949
Affects Multiple Steps in PAM Trafficking--
A simplified model of
PAM-1 routing through the secretory pathway in AtT-20 cells is shown to
aid in this discussion (Fig. 10). PAM-1
is synthesized in the endoplasmic reticulum, and the N-linked oligosaccharide attached to PAM-1 acquires
endoglycosidase H resistance as it traverses the Golgi (step
1) (67). Cleavage to yield soluble PHM and membrane PAL occurs
only after entry into immature secretory granules, and it can be
blocked by incubation at 20 °C (step 2) (35). Soluble PHM
is either stored in mature secretory granules until release is
stimulated (steps 3 and 4) (68) or it exits in
constitutive-like vesicles that bud from immature secretory granules,
accounting for basal secretion (not shown) (69, 70). Intact PAM-1 and
membrane PAL are also retrieved from immature secretory granules in
constitutive-like vesicles, and they recycle to the TGN through late
endosomes (steps 5 and 8). Any PAM-1 or membrane
PAL on the cell surface is internalized, traversing early and
late/recycling endocytic compartments before returning to the TGN
(steps 6-8) (11, 22).
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 10.
A model for routing in AtT-20 cells.
The minimum number of PAM-1 trafficking events involved in its
metabolism and the major subcellular compartments that contain PAM-1
are shown. Late/recycling endosomes are the most sensitive and disperse
from the perinuclear region first in a low-dose, short incubation with
nocodazole and increased nocodazole dosage and incubation time causes
dispersal of the TGN. PALm, 70-kDa membrane PAL;
P, PAM-1 phosphorylation.
|
|
Although the steady-state localization of PAM-1 TS/AA in AtT-20 cells
is somewhat more diffuse than that of PAM-1, the response of each
protein to nocodazole treatment is similar. In contrast, in the AtT-20
PAM-1 TS/DD cells, PAM staining was partially localized to the
perinuclear region with more PAM localized to vesicular structures at
the tips of processes. These vesicles are thought to be secretory
granules since regulated secretion is enhanced in these cells, and
adrenocorticotropin-containing vesicles are localized to this region in
AtT-20 PAM-1 TS/DD cells (26). Following nocodazole treatment, PAM-1
TS/DD staining in the perinuclear region was dispersed, suggesting that
PAM-1 TS/DD was largely localized to late/recycling endosomes (36).
Overall, the altered steady-state localization of the mutant PAM
proteins reflects changes in both biosynthetic and endocytic
trafficking with PAM-1 phosphorylated at
Thr946/Ser949 having an increased anterograde
and retrograde movement through the cell.
The initial synthesis and subsequent oligosaccharide maturation (Fig.
10, step 1) of PAM-1 are not affected by the phosphorylation state of Thr946 and Ser949. In contrast, the
ability of PAM to enter immature secretory granules (Fig. 10,
step 2) is increased by phosphorylation of
Thr946 and Ser949. We base this conclusion on
the fact that cleavage of PAM-1 TS/DD to soluble PHM, which begins in
immature secretory granules, occurs earlier and is completed sooner
than cleavage of PAM-1 or PAM-1 TS/AA (Fig. 5A, 3rd
panel) (24). In addition, PHM produced from PAM-1 TS/DD is stored
in mature secretory granules more efficiently than PHM produced from
PAM-1 or PAM-1 TS/AA (Fig. 5C). Finally, stimulation of
PAM-1 TS/DD cells with BaCl2 increases release of PHM
3-fold (Fig. 10, step 4); in contrast, release of PHM from PAM-1 and PAM-1 TS/AA cells is less responsive to BaCl2
(Fig. 4B). If the increased ability of PAM-1 TS/DD to
aggregate at the pH of the Golgi/TGN mimics that of PAM-1
phosphorylated at Ser949, phosphorylation at this site
could facilitate access to secretory granules (Fig. 10, step
2) or retention in secretory granules (Fig. 10, step
5). The fact that PAM-1 recognized by the
P-Ser949-specific antibody is localized to a compact
perinuclear compartment suggests that phosphorylation could be
occurring in the correct subcellular compartment to affect granule
entry. The increased degradation of PAM-1 TS/AA could reflect
diminished access to or retention in secretory granules, with the
inability of this protein to aggregate at neutral pH playing a key role.
The fact that PAM-1 truncated immediately following the transmembrane
domain is cleaved only half as well as wild-type PAM-1 and that it
rapidly accumulates on the cell surface (24) is consistent with the
presence of a granule entry/retention signal in the cytosolic domain.
Only a few resident secretory granule membrane proteins have signals
for sorting to regulated secretory granules (71). P-selectin is
efficiently recycled from the cell surface to secretory granules where
it is sequestered in endothelial and neuroendocrine cells (72, 73).
When the cytosolic domain of P-selectin is linked to the transmembrane
and extracellular domains of a plasma membrane protein, the chimera is
targeted to secretory granules (74). Several membrane proteins also
have signals that facilitate their retrieval from immature secretory granules. Furin is normally localized to the TGN, and the
phosphorylation of the two CKII sites in its acidic cluster results in
its rapid retrieval from LDCVs (6, 18). In the cytosolic COOH-terminal acidic motif of the vesicular monoamine transporter 2, replacement of
two serines with aspartates to mimic phosphorylation also increases its
removal during LDCV maturation (75). The unique trafficking role of the
acidic cluster in PAM may reflect its ability to interact in multiple
ways with cytosolic proteins like P-CIP2 and kalirin (25, 64).
Phosphorylation of PAM at Thr946 and Ser949
also alters its endocytic trafficking. Internalization of PAM-1 from
the cell surface, as for many proteins, depends upon a Tyr motif,
Gly-Tyr936-Ser-Arg-Lys (Fig. 1) (20). As for furin and
CD-Man-6-PR, phosphorylation of acidic cluster residues does not alter
the early endocytic trafficking of PAM (Fig. 10, step 6) (6,
76). Even when allowed to internalize PAM-antibody complexes for 90 min
at 20 °C to access late endocytic compartments (Fig. 10, step
7), no phenotypic differences were observed between cells
expressing PAM-1, PAM-1 TS/DD, and PAM-1 TS/AA.
Both phosphorylation and dephosphorylation of Thr946 and
Ser949 do, however, play major roles late in the endocytic
pathway (Fig. 10, steps 5, 8, and 9) (11). We
base this conclusion on four observations. First, PAM-antibody
complexes internalized by the PAM-1 TS/AA cells exhibit a diffuse
vesicular pattern with no concentration in the perinuclear area and a
more rapid loss of signal (Fig. 7). Internalized PAM-antibody complexes
in the PAM-1 TS/AA cells may be degraded in lysosomes (Fig. 10,
step 9) or recycled to the cell surface and released. This
observation suggests that phosphorylation at
Thr946/Ser949 is required at some step late in
the endocytic pathway. Second, and consistent with this conclusion,
treatment of PAM-1 cells with DRB, an inhibitor of protein kinases that
can phosphorylate Ser949, blocked the accumulation of
PAM-antibody complexes in the TGN; internalized antibody remained in
distinct punctate vesicles spread throughout the cell (Fig. 9). Third,
the localization of PAM phosphorylated at Ser949 to the TGN
(Fig. 3) suggests that phosphorylation at this site results in the
entry and/or retention of PAM in the TGN (Fig. 10, step 8).
Finally, the fact that PAM-antibody complexes are internalized by the
PAM-1 TS/DD cells but remain in dispersed, late endocytic structures
(Fig. 8) demonstrates the need to have a protein that is not modified
at this site. Differences in the trafficking of PAM-1 TS/AA and PAM-1
TS/DD become apparent only late in the endocytic pathway, after the
point at which internalized proteins accumulate at 20 °C. PAM-1 that
is not phosphorylated at Ser949 may be the preferred cargo
for transit from late endosomes to the TGN (Fig. 10, step
8).
Endocytic compartments have a lower pH than the TGN, suggesting that
regulation of the ability of PAM-1 to exhibit pH-dependent aggregation by phosphorylation of Ser949 could also play a
role in endocytic trafficking. In addition, PACS-1, a cytosolic sorting
protein that directs the TGN localization of furin by binding to CKII
phosphorylation sites in the COOH-terminal acidic cluster region (19),
might also facilitate the return of PAM-1 from late endosomes to the
TGN. Recently, we have determined that PACS-1 binds to both PAM-1 and
PAM-1 TS/DD GST fusion proteins, with enhanced binding to the PAM-1
TS/DD GST fusion protein (data not shown).
The Thr946/Ser949 acidic cluster signal in PAM
is readily distinguishable from similar trafficking signals in other
proteins. For example, the CKII-mediated phosphorylation of Ser
residues in the furin acidic cluster (SDSEEDE), which regulates the
binding of PACS-1 and AP-1, facilitates the removal of furin from
immature granules and its subsequent return to the TGN (6, 18, 19). Similarly, the CKII-mediated phosphorylation of an acidic cluster in
vesicular monoamine transporter 2 diminishes its localization to LDCVs
(75). A single CKII phosphorylation site in the acidic cluster of
cation-independent Man-6-PR regulates the binding of AP-1, which in
turn mediates the trafficking of the receptor (77, 78). Recently, it
was determined that trafficking of both cation-independent and
CD-Man-6-PR is mediated by Golgi-localized, -ear-containing, ADP-ribosylation factor binding proteins (
,
TOP
ABSTRACT
INTRODUCTION
