* Department of Pediatrics, Case Western Reserve University School of Medicine, and
Division of Nephrology, Veterans Administration Hospital, Cleveland, Ohio
Correspondence: Melvin Berger, M.D., Ph.D., Division of Pediatric Immunology, Rainbow Babies’ and Children’s Hospital, Room 594, 11100 Euclid Ave., Cleveland, OH 44106. E-mail: mxb12{at}po.cwru.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
S. Mg2+ at 0.4
µM is necessary for half-maximal binding of ARFs to CR1 storage
vesicles. Artificial phospholipid vesicles and primary and secondary
granules from human neutrophils do not bind ARFs themselves and do not
compete for recruitment of ARFs to CR1 vesicles, suggesting that
specific membrane environments and/or proteins on these vesicles
stabilize the ARF-GTP-Mg2+ complex. Free Ca2+
at 300 nM does not inhibit ARF binding to CR1 storage vesicles, but 10
mM Ca2+ does reduce such binding. These findings suggest
that ARF-GTP specifically and reversibly interacts with CR1 storage
vesicles in human polymorphonuclear leukocytes and may play a role in
regulating their transport.
Key Words: low-molecular-weight GTP-binding proteins • transport vesicles • phagocytosis
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Low-molecular-weight monomeric GTP-binding proteins including adenosine
diphosphate-ribosylation factors (ARFs) and Rabs are members of the Ras
superfamily [6
7
8
9
10
11
12
13
14
]. Thirty Rabs [6
] and
six ARFs [15
] have been identified so far in mammalian
cells. The common feature of these GTP-binding proteins is their
ability to shuttle between a soluble guanosine diphosphate (GDP)-bound
form and a membrane-associated GTP-bound state as they regulate
vesicular trafficking. Some Rabs are expressed universally in cells
that share common transport pathways, and others are found only in
highly differentiated cells with unique transport pathways
[16
, 17
]. We observed that Rabs 3a, 4, and
5a were constitutively present in the immunoisolated CR1 vesicles,
whereas the association of ARF with these vesicles was variable and
could be increased by incubation with cytosol, suggesting regulatable
binding. Of six known mammalian ARF proteins with molecular masses of
21 kDa, ARF1 is the best characterized. It participates in a wide
range of functions, including regulation of vesicle traffic along both
exocytic and endocytic pathways, maintenance of organelle structure,
and assembly of coat proteins, and it may also be a cofactor for
cholera toxins [7
8
9
]. In this study we present evidence
for the recruitment of an ARF from the cytosolic fraction of
neutrophils to isolated CR1 vesicles in the presence of GTPS, and
the requirements for Mg2+ versus Ca2+ in this
binding.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
band," and
a mixture of primary and secondary granules [3
]. Cytosol
and the band (5–6 mL) were freed from Percoll by centrifugation
without a cushion of Percoll according to a procedure described
previously [4
]. Cytosol was stored at -70°C until
further use. CR1 storage vesicles were subsequently immunoisolated from
the Percoll-free band by using a magnetic bead system (Dynabeads,
M280; Dynal, Inc., Lake Success, NY) with rabbit antibodies against the
cytoplasmic tail of CR1 and sheep anti-rabbit antibodies on the beads,
exactly as described previously [4
]. Vesicles purified
in this way are free from contamination, as judged by the absence of
marker enzymes of primary and secondary granules [4
]. In
addition, Golgi proteins such as p230 [18
], golgin-84
[19
], and GS28 [20
] were absent from the
immunoisolated CR1 vesicles (data not shown).
Assay for myeloperoxidase and preparation of primary- and
secondary-granule mixture
Myeloperoxidase activity was determined with 0.006% hydrogen
peroxide, 2.5 mM 4-aminoantipyrine, and 16 mg/mL of phenol in 100 mM
Tris-HCl buffer (pH 8.5) in 96-well microtiter plates. The change in
absorbance was monitored continuously at 490 nm at room temperature for
2–3 min.
A mixture of primary and secondary granules was prepared on 2-layer
Percoll gradients from nitrogen cavitates of PMN as described above.
One-milliliter fractions were collected after removal of cytosol and
the band from the top of the gradient. Fractions with
myeloperoxidase activity were pooled and freed from Percoll by
centrifugation at 180,000 g for 1 h on a cushion of 0.5
mL of undiluted Percoll according to the published procedure
[1
, 4
]. This mixture of primary and
secondary granules, which was positive for myeloperoxidase, was kept at
4°C until further use.
Association of low-molecular-weight GTP-binding
proteins with immunoisolated CR1 storage vesicles
As a first step in studying the interaction of small GTP-binding
proteins with CR1 vesicles, we checked for the presence of Rabs and
ARFs in isolated CR1 vesicles, cytosol, and the mixture of primary and
secondary granules by Western blotting as described below.
To determine if additional Rabs and ARF could be recruited from cytosol to immunoisolated CR1 vesicles, we performed incubations in the presence and absence of cytosol at 37°C for 30 min in relaxation buffer [100 mM KCl, 3 mM NaCl, 1 mM ATP, 3.5 mM MgCl2, 10 mM PIPES (pH 7.3)] containing protease inhibitors [4 ] and appropriate guanine nucleotides. Reactions were terminated with 4 volumes of ice-cold phosphate-buffered saline (PBS) containing protease inhibitors. In initial experiments, the beads were harvested with a magnet, washed five times with PBS buffer containing protease inhibitors, and prepared for sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). In later experiments, 3.5 mM Mg2+ was included in the wash buffer. Additional experiments with varying amounts of cytosol, different nucleotides, and different concentrations of cations (see below) were also performed using this protocol. Proteins attached to immunoisolated CR1 vesicles on the magnetic beads were eluted by suspension in 3 volumes of nonreducing SDS sample buffer and boiling for 2 min. Proteins were then separated by SDS-PAGE [21 ] and detected by Western blotting.
Similar experiments were performed to determine if primary and secondary granules could also recruit ARF proteins from cytosol on incubation at 37°C for 30 min in relaxation buffer as above. Granules from the reaction mixture were separated on modified 2-layer Percoll gradients containing 3 mL of 1.065-g/mL Percoll with a cushion of 0.5 mL of 1.12-g/mL Percoll. Fractions (0.5 mL) were assayed for myeloperoxidase activity as described above. Myeloperoxidase-positive fractions were pooled and freed from Percoll, and the recovered granules were concentrated to 50 µL on microconcentrators (Microcon 10; Amicon, Beverly, MA) before being subjected to SDS-PAGE for analysis of ARF binding by Western blotting.
Magnesium-dependent binding of ARF to immunoisolated CR1
vesicles
To determine of the effect of Mg2+ on the binding of
ARF to CR1 storage vesicles, Mg2+ was omitted from the
relaxation buffer before nitrogen cavitation. No attempt was made to
remove intracellular Mg2+ or nucleotides from cytosol.
Mg2+-EDTA buffer was made with varying concentrations of
Mg2+ in relaxation buffer in the presence of 5 mM EDTA.
Mg2+-EGTA buffer was made likewise. The concentrations of
free Mg2+ at 37°C for both buffers (ionic strength =
0.109) were calculated as described by Portezehl et al.
[21
].
Effects of granules and phospholipid vesicles on ARF recruitment to
immunoisolated CR1 vesicles
In some experiments we included granules or artificial
phospholipid vesicles in the reaction mixture with cytosol and CR1
vesicles on beads to determine if the presence of these other types of
lipid membranes would influence the recruitment of ARF to CR1 storage
vesicles. A mixture of primary and secondary granules containing
myeloperoxidase activity was isolated from PMN cavitate on 2-layer
Percoll gradients as described above and used for these experiments to
determine if they would compete with the CR1 vesicles for binding of
ARF.
To provide further assurance that the observed binding of ARF was not
just a hydrophobic effect, we also used phospholipid vesicles
containing phosphatidylethanolamine, phosphatidylinositol
4,5-diphosphate, and dipalmitoyl phosphotidylcholine in a molar ratio
of 10:1.5:1. This composition was selected to allow determination of
activation of phospholipase D (PLD) upon binding of ARF, as has been
reported by others [10
, 14
,
23
]. PLD activity in these vesicles would result in
liberation of choline, which would have been detected by
chemiluminiscence of luminol in the presence of horseradish peroxidase
by H2O2 from the oxidation of liberated choline
by choline oxidase. The chemicals were obtained from Sigma Chemical Co.
(St. Louis, MO), and phospholipid vesicles were prepared
according to the work of Brown et al. [14
]. Phospholipid
vesicles (56 and 168 µL) were incubated for 30 min at 37°C with CR1
vesicles and cytosol in a final volume of 1 mL of incubation buffer
containing 50 mM HEPES buffer (pH 7.5) with 3 mM EGTA, 80 mM KCl, 3 mM
MgCl2, and 2 mM CaCl2, with or without 400 µM
GTPS. After incubation, CR1 vesicles were isolated with a magnet,
washed as before, and prepared for SDS-PAGE. Incubations without
granules or phospholipids were run in parallel as controls for the
amount of ARF bound to CR1 vesicles alone.
Immunoblotting
Protein samples were analyzed on SDS-PAGE gels (12 or 15%
polyacrylamide) under nonreducing conditions in a Mini Protean II Dual
Slab Cell (Bio-Rad, Hercules, CA) using prestained low-molecular-weight
markers from Bio-Rad. Separated proteins were transferred onto
nitrocellulose membranes (Pharmacia, Uppsala, Sweden) using 192 mM
glycine in 25 mM Tris (pH 8.3) containing 20% methanol
[23
]. The membrane was blocked with 5% nonfat dry milk
and 1% bovine serum albumin in PBS with 0.05% Tween 20 (Fisher
Scientific, Pittsburgh, PA) for 1 h at room temperature and was
subsequently washed four times with PBS buffer containing 0.05% Tween
20. The membrane was then probed with primary antibody for 1 h at
room temperature. Polyclonal rabbit antibody to human ARF1, as well as
recombinant human ARF1 (used as a positive control) were kind gifts of
Dr. Sylvain Bourgoin of Centre de Recherche en Rhumatologie et
Immunologie, Quebec, Canada. Isoform-specific polyclonal and monoclonal
antibodies to human Rabs were purchased from Santa Cruz Biotechnology,
Inc. and BD Transduction Laboratories (San Diego, CA), respectively.
Monoclonal antibodies to Golgi proteins were also obtained from
Transduction Laboratories. Control proteins were always run in parallel
to identify Rabs and ARF. Horseradish peroxidase-conjugated donkey
anti-rabbit immunoglobulin G (Amersham, Little Chalfont,
Buckinghamshire, UK) and anti-mouse immunoglobulin G (BD Transduction
Laboratories) were used as secondary antibodies to probe the membrane
for 1 h, and bound antibody was detected by enhanced
chemiluminescence using a kit from Amersham. In some cases we used
alkaline phosphatase-conjugated secondary antibody.
Estimation of protein
Bradford reagent from Bio-Rad was used for the estimation of
proteins in a 96-well microtiter plate using a Thermomax reader
(Molecular Devices, Sunnyvale, CA) according to the manufacturer’s
protocol.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
S resulted
in no significant increase in the amount of Rab proteins bound to
vesicles (Fig. 1
, lane 5 vs. lanes 6 and 7). Rabs 1, 2, 7, 9, and 11
were absent from CR1 storage vesicles. Occasionally, small amounts of
Rab6 were also found in association with these vesicles isolated on
magnetic beads, with or without the addition of cytosol (data not
shown). Rabs 3a, 4, and 5a were also present in the mixture of primary
and secondary granules (Fig. 1
, lane 4). In addition, Rabs 1, 2, 6, 7,
9, and 11 were also detectable in the mixture of primary and secondary
granules (data not shown), and all except Rab1a were present in cytosol
(data not shown). We thus conclude that Rabs 3a, 4, and 5a are
constitutively associated with this subpopulation of vesicles and that
Rabs 2, 7, 9, and 11, which are associated with primary and secondary
granules, are absent from the CR1 storage vesicles.
 View larger version (30K): [in a new window] |
Figure 1. Detection of Rabs 3a, 4, and 5a on immunoisolated CR1 storage vesicles
by Western blotting. Incubations were carried out at 37°C for 30 min
in 10 mM PIPES, pH 7.3, containing 3.5 mM Mg2+, 1 mM ATP,
and protease inhibitors. Lane 1, control; lane 2, PMN lysate; lane 3,
cytosol; lane 4, mixture of primary and secondary granules; lane 5, CR1
vesicles alone; lane 6, CR1 vesicles recovered after incubation with
cytosol; lane 7, similar to lane 6, but with GTPS in the
incubation mixture.
|
S (Fig. 2A
, lane 4). In
contrast to the CR1 storage vesicles, ARF was not present in the
granule fraction (Fig. 2B
, lane 4, and Fig. 2A
, lane 5) and could not
be recruited to granules by incubation with cytosol in the presence of
GTPS (Fig. 2B
, lane 5) under the same conditions that promoted its
association with CR1 vesicles. These results therefore suggest that the
variable binding we observed in early experiments was due to variable
association of cytosolic ARF with the CR1 vesicles and not to
contamination of the CR1 vesicles with granule membranes.
 View larger version (35K): [in a new window] |
Figure 2. ARF is recruited to CR1 vesicles from the cytosol, but not from
granules. Reaction conditions were the same as those for Fig. 1
. ARF
was detected by enhanced chemiluminescence on Western blots as
described in Materials and Methods. (A) Interaction of CR1 vesicles
with and without cytosol. Proteins were isolated by
SDS–12% PAGE and transferred to a nitrocellulose membrane for
detection of ARF. Lane 1, ARF control; lane 2, CR1 vesicles alone; lane
3, CR1 vesicles recovered after incubation with cytosol; lane 4,
similar to lane 3, with 400 µM GTPS in the incubation mixture;
lane 5, CR1 vesicles with granules and 400 µM GTPS, without
cytosol. (B) Interaction of granules with cytosol. Proteins isolated on
an SDS–15% PAGE gel were transferred to a nitrocellulose membrane,
which was probed for ARF by Western blotting. Lane 1, ARF
control; lanes 2 and 3, 3 and 9 µg of cytosol, respectively; lane 4,
1.8 µg of the mixture of primary and secondary granules; lane 5,
recovered granules after incubation with cytosol in the presence of
GTPS.
|
S. Brefeldin A (10 µg/mL) did not inhibit ARF
recruitment to CR1 storage vesicles (data not shown), and we have
previously found that brefeldin treatment of intact PMN did not alter
the net up-regulation of CR1 on the plasma membrane in response to
chemoattractants (unpublished data). We speculate that this may be due
to the presence of brefeldin A-resistant guanine nucleotide exchange
proteins such as cytohesin 1 and ARNO [24
]. Figure 3
shows that increasing amounts of ARF are recruited to CR1 vesicles
with increasing amounts of cytosol. In this experiment, the bound ARF
was stabilized by inclusion of 3.5 mmol/L of MgCl2 in the
buffer used for washing the CR1 vesicle-bearing beads after their
incubation with or without cytosol.
 View larger version (15K): [in a new window] |
Figure 3. Effect of increasing concentrations of cytosolic proteins on ARF
binding to CR1 vesicles. ARF was recruited from the cytosol to
immunoisolated CR1 vesicles under experimental conditions similar to
those described for Figure 2
, but the beads were washed with a buffer
containing 3.5 mM MgCl2 after the incubation
with or without cytosol. Proteins were separated on an
SDS–12% PAGE gel, and ARF was detected by enhanced chemiluminescence.
Lane 1, ARF control; lane 2, CR1 vesicles incubated without cytosol,
and without GTPS; lanes 3–8, CR1 vesicles incubated with 10, 50,
75, 100, 125, and 150 µg of cytosol protein, respectively, each with
400 µM GTPS.
|
S enhanced
ARF binding (lane 5). The amount of ARF bound to CR1 storage vesicles
does not change from 1 to 400 µM GTPS (unpublished observation).
The effect of 200 µM GTPS in recruiting ARF to CR1 storage
vesicles was increased further in the presence of 10 mM EDTA (lane 6).
Although the increase in expression of CR1 on the plasma membrane
following stimulation of PMN with agonists such as
N-formyl-methionyl-leucyl-phenylalanine is Ca2+
dependent, we found that addition of excess Ca2+ (10 mM) in
the incubation buffer resulted in decreased binding of ARF to the CR1
vesicles in these studies (Fig. 4
, lane 7). However, there was no
inhibition at other concentrations tested, and most experiments were
carried out with 300 nM free Ca2+ (see Fig. 7
below), a concentration frequently employed in assays of PLD
activity [14
]. We speculate that the requirement for
Ca2+ in up-regulation of CR1 expression on the plasma
membrane during PMN activation is due to involvement in a step other
than low-molecular-weight GTPase binding to the vesicles.
 View larger version (12K): [in a new window] |
Figure 4. Effects of nucleotides and divalent cations on recruitment of ARF to
CR1 vesicles. Reaction conditions were the same as those for Figure 1
except that 10 mM CaCl2 was added in the incubation for
which results are shown in lane 7. Samples were separated by SDS–12%
PAGE and transferred to a nitrocellulose membrane, and ARF was detected
with an alkaline phosphatase-conjugated secondary antibody. Lane 1, ARF
control; lane 2, CR1 vesicles alone; lane 3, CR1 vesicles with cytosol;
lane 4, CR1 vesicles with cytosol and 200 µmol/L of GDP; lane 5, CR1
vesicles with cytosol and 200 µM GTPS; lane 6, CR1 vesicles with
cytosol, 200 µM GTPS, and 10 mM EDTA; lane 7, CR1 vesicles with
cytosol, 200 µM GTPS, and 10 mM CaCl2.
|
 View larger version (18K): [in a new window] |
Figure 7. Effects of varying concentrations of phospholipid vesicles on ARF
binding to immunoisolated CR1 vesicles. ARF bound to recovered CR1
vesicles was detected by enhanced chemiluminescence after separation of
proteins by SDS–15% PAGE and transfer to a nitrocellulose membrane.
Lane 1, ARF control; lane 2, CR1 vesicles with 400 µM GTPS but
without cytosol; lane 3, CR1 vesicles with cytosol and 400 µM
GTPS; lane 4, as in lane 3 with 56 µL of phospholipid vesicles;
lane 5, as in lane 3 with 168 µL of phospholipid vesicles. The final
volume of the incubation mixture was 1 mL in all cases.
|
S-vesicle complex. The
apparent enhancement of ARF binding by EDTA (Fig. 4
, lane 6) may be due
to its ability to promote exchange of GTPS for GDP bound to
cytosolic ARF for the subsequent formation of a CR1
vesicle-GTP/GTPS-Mg2+ complex. These results suggest
that GTP-dependent binding of ARF to CR1 storage vesicles is a discrete
process which is likely independent of the mechanisms of the increases
in the intracellular free Ca2+ concentration that are
necessary for up-regulation of CR1 expression on the plasma membrane
during PMN activation [5
]. None of the experiments with
EDTA or EGTA caused removal of CR1 from the vesicles, suggesting that
at the concentrations of these chelators used, none of the components
of the reaction mixtures developed detergent-like effects.
 View larger version (20K): [in a new window] |
Figure 5. Magnesium-dependent recruitment of ARF to CR1 vesicles. (A) Western
blot for ARF binding to CR1 vesicles with varying concentrations of
free Mg2+ in the presence of 5 mM EDTA. Reaction conditions
were the same as those described for Fig. 1
, including 400 µmol/L of
GTPS, except that Mg2+ was left out of the relaxation
buffer and the buffer for each reaction was made with specific
concentrations of Mg2+ in order to achieve the final free
Mg2+ concentration indicated in the figure. (B) The Western
blot shown in panel A was scanned by densitometer. A plot of ARF
density in each lane against the calculated free Mg2+
concentration at half-maximal ARF binding gave a value of 0.4 µM for
Mg2+ at 37°C.
|
S (Fig. 2B
, lane 5). To provide
further assurance that the binding of ARF to CR1 vesicles did not occur
through nonspecific hydrophobic interactions alone, which would also
allow it to interact promiscuously with other types of lipid membranes,
we determined whether traditional granules would compete with CR1
vesicles for binding of ARF from the cytosol. Addition of a mixture of
traditional granules, separated on 2-layer Percoll gradients, together
with Mg2+ (3.5 mM) and GTPS (400 µM), did not hinder
ARF recruitment to CR1 vesicles from the cytosol (Fig. 6
, lane 4). Changing the order of addition of cytosol and granules
had no observable effect (Fig. 6
, lanes 4 vs. 5 and 6 vs. 7), but
inclusion of MgCl2 in the wash buffer used for tubes 6 and
7 at the end of the incubation stabilized the bound ARF so that it was
more easily seen on the immunoblot. These findings further indicate
that ARF does not bind to traditional granules under conditions that
promote its binding to CR1 vesicles, and thus suggest that the binding
to CR1 vesicles is specific.
 View larger version (15K): [in a new window] |
Figure 6. Western blot analysis of the effect of granules on ARF binding to CR1
vesicles. The conditions were same as those for Fig. 1
, including 3.5
mmol/L of MgCl2 in the incubation buffer only,
except in lanes 6 and 7. Lane 1, ARF control; lane 2, CR1 vesicles
alone; lane 3, CR1 vesicles with cytosol and GTPS; lane 4, CR1
vesicles with cytosol, granules, and GTPS; lane 5, CR1 vesicles with
granules, cytosol, and GTPS in that order of addition; lanes 6 and
7, same conditions as in lanes 4 and 5 but with 3.5 mM
MgCl2 in the wash buffer as well.
|
S and Mg2+, which in this
experiment would be detected by inhibition of the binding of ARF to CR1
vesicles (Fig. 7
, lane 3 vs. lanes 4 and 5). We did not observe any
liberation of choline, which would have indicated that ARF binding
induced activation of PLD (data not shown). This led us to conclude
that the binding of ARF to CR1 vesicles is not influenced by the
presence of excess phospholipids, and it confirms the hypothesis that
the recruitment of ARF to CR1 vesicles is not due to nonspecific
hydrophobic interactions of ARF with any type of membrane lipids, but
is predominantly specific and dependent on Mg2+ and GTP. |
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We found that, unlike Rabs, one or more ARF family members can be variably recruited from the cytosol by immunoisolated CR1 storage vesicles, implying a reversible association. In many types of cells ARFs 1, 3, 4, and 5 are found in the cytosol [15 , 23 ], and neutrophils have previously been reported to contain ARFs in their cytosol [23 ]. These proteins are 84% homologous with each other, and the polyclonal antibody against human ARF1 that we used, like the monoclonal antibody ID9 [15 ], can react with multiple isoforms. Since ARF1 and ARF3 are reported to be present in greater abundance than ARF4 and ARF5 [15 ], and since our Western blots showed a single immunoreactive protein that correlated with human recombinant ARF1 protein, we believe that the major ARF protein(s) we are observing is most likely ARF1.
We investigated the conditions that favored association versus dissociation of ARF in greater detail, because we speculated that the reversibility of this interaction might be of functional significance in regulating the trafficking of these unique vesicles during cell activation in vivo. ARF is known to be required for numerous biological activities, including regulation of vesicular trafficking between intracellular organelles as well as exocytosis in other cell types [8 , 9 ].
The mechanism of exocytosis varies with cell type, and the different
mechanisms leading to fusion of export vesicles with the plasma
membrane may be sensitive to GTP, calcium, and/or protein
phosphorylation [26
, 27
]. Since
Ca2+ is required for increased CR1 expression during PMN
activation, we thought it might play a role in regulating the
association of GTP-binding proteins with our isolated CR1 vesicles. In
general, GTP-mediated processes require cytosol and are not inhibited
by reagents that block Ca2+-mediated membrane fusion. We
found that ARF protein recruitment to CR1 vesicles was enhanced by EDTA
and that Ca2+ did not promote its recruitment (Fig. 4)
. In
contrast, with increasing concentrations of free Mg2+,
there was a marked increase in the efficiency of ARF recruitment (Fig. 5)
. ARF, like all other small GTP-binding proteins, is generally found
free only in its GDP-Mg2+ form, whereas the
GTP-Mg2+-ARF complex is the form that becomes bound to
membranes. ARF goes back to its soluble form upon hydrolysis of GTP.
Thus, the use of the nonhydrolyzable analog, GTPS, stabilizes ARF in
its bound form and facilitates observation of the bound complex (Fig. 2A , lane 4). A recent crystallographic study shows that the nucleotide
binding site and the Mg2+ binding site are near each other
and that the role of Mg2+ is to promote the nucleotide
binding [27
]. In order for ARF to be active and to bind
to membranes, the bound GDP has to be replaced by GTP/GTPS.
Dissociation of GDP could be facilitated artificially either by EDTA,
by different concentrations of Mg2+, or by guanine
nucleotide exchange proteins in the cytosol [24
,
28
29
30
]. This is in agreement with our findings that on
regulation of the concentration of free Mg2+, more ARF was
recruited to CR1 vesicles (Fig. 5)
, and that the stability of the
complex increased (Fig. 3
and 7)
on inclusion of magnesium in the wash
buffer [29
, 31
]. The half-maximal effective
concentration (0.4 µM) we observed for the effect of Mg2+
in ARF binding to CR1 vesicles is within the same order of magnitude as
that reported in previous studies, showing that there is effective
guanine nucleotide exchange and stable complex formation by isolated
components at concentrations of free Mg2+ below 1 µM
[26
, 32
]. These results strongly suggest
that GTP-dependent recruitment of ARF to CR1 storage vesicles is a
distinct process independent of the fluxes in intracellular free
Ca2+ concentrations that are necessary for up-regulation of
CR1 expression on the plasma membrane during PMN activation
[5
].
As we found here with CR1 vesicles, ARF proteins are reported to be
readily recruited to purified cellular fractions such as ER, Golgi, or
endosome membranes in the presence of GTPS, with the sole exception
of ARF6 [15
]. This is also the case with coatomer
proteins, which are involved in vesicle budding [33
]. In
some studies, ARF has been shown to bind to phospholipid vesicles in
the absence of other proteins [34
]. However, our results
suggest that the recruitment of ARF to CR1 storage vesicles is
specific, since ARF was not contained in traditional granules and did
not bind to granules or to artificial phospholipid vesicles under
conditions that readily promote its association with CR1 vesicles (Fig. 6
and 7)
. The composition of phospholipid vesicles we selected was
designed to enable us to detect PLD if it was activated upon ARF
binding [23
], but this was not detected. If nonspecific
interactions and/or lipid-lipid binding of the myristate moiety
provided the predominant binding energy, we should have observed
decreased ARF binding to CR1 storage vesicles in the presence of
phospholipid vesicles or mixtures of traditional granules as a result
of competitive binding to these other sources of lipids, but that was
not the case. Since only minimal amounts of ARF were recruited in the
absence of GTPS, its binding to the CR1 vesicles is not likely due
to simple electrostatic or hydrophobic interactions. Moreover, the
Mg2+ dependence of the binding of ARF in the presence of
GTPS provides additional evidence that the interaction between CR1
storage vesicles and ARF is predominantly specific, although we cannot
exclude the possibility that other proteins on the cytoplasmic face of
CR1 vesicles, which could be absent from granule membranes, might
contribute to ARF binding.
Secretory vesicles are more readily mobilized than any other compartments in neutrophils, and we have shown that once CR1 reaches the plasma membrane, it is rapidly internalized in small endocytic vesicles upon PMN activation [25 ]. Our results are consistent with the hypothesis that the interaction of CR1 storage vesicles with the plasma membrane is affected by guanine nucleotide exchange on ARFs at the site of interaction. It is possible that the physiological role of ARF binding to CR1 storage vesicles is similar to its role in the assembly of coatomer complexes on Golgi-derived vesicles for retrograde transport [35 ], and that ARF does not have a role in exocytosis. A role for ARF in endocytosis of CR1 would not be surprising, because the generation of carrier vesicles from donor membranes such as plasma membranes and/or other intracellular membranes characteristically involves GTP-dependent recruitment of coat complexes [35 , 36 ]. ARF has been associated with coat formation in other systems, for example, in assembling a multiprotein complex on the surfaces of vesicles that initiate paxillin-rich focal adhesions at the ends of prominent actin stress fibers [37 ]. Synaptic vesicle formation in neuroendocrine cell lines has been shown to occur by both clathrin-dependent and clathrin-independent pathways, and both pathways require ARF-GTP [38 ]. Previous results from our own and other laboratories suggest that CR1 can also be internalized in clathrin-dependent and clathrin-independent pathways [25 , 39 ]. Spang and Schekman [35 ] have shown that, like anterograde transport in vitro, retrograde transport from Golgi to the ER in yeast requires ARF-GTP and guanine nucleotide exchange-activating proteins (Gea1p and Gea2p). In view of our findings, it is quite conceivable that the major role of ARF in this system is to serve as a regulator of CR1 storage vesicle trafficking because of its GTP-dependent binding, and that this may be mainly involved in endocytosis rather than exocytosis. Gaynor et al. [40 ] have shown in yeast mutants that a fundamental role for ARF is to maintain membrane morphology both in ER-Golgi transport and in endocytosis. The present study establishes that ARF specifically binds to CR1 storage vesicles in a Mg2+- and GTP-dependent manner. The question of whether this is involved in anterograde or retrograde transport (i.e., exocytosis vs. endocytosis) will be the target of continuing investigation.
Received February 21, 2001; revised May 31, 2001; accepted June 4, 2001.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
S-stimulated phospholipase D activity in membrane fractions of HL-60 cells J. Biol. Chem. 270,22795-22800
-soluble NSF attachment proteins (SNAP) mediate dissociation of GS28-syntaxin 5 Golgi SNAP receptors (SNARE) complex J. Biol. Chem. 272,25441-25444This article has been cited by other articles:
 |
|
 |
|
  M. J. Herrero-Turrion, J. Calafat, H. Janssen, M. Fukuda, and F. Mollinedo Rab27a Regulates Exocytosis of Tertiary and Specific Granules in Human Neutrophils J. Immunol., September 15, 2008; 181(6): 3793 - 3803. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
