| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, May 23, 1994; and in revised form, December 13,
1994) From the
Biochemical, structural, and functional properties of Rab5
wild-type (WT) protein were compared with those of Q79L and N133I
mutants. The detergent
3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate
increased guanine nucleotide binding to Rab5 WT
Eukaryotic cells maintain a highly compartmented organization,
and are capable of ordered and specific transport among different
intracellular compartments. A large number of Ras-related GTPases,
termed Rabs, have been implicated in distinct steps of
intercompartmental transport (for review, see (1) ). Regulatory
GTPases shuttle between two activity states, which are determined by
the phosphorylation status of bound guanine nucleotides(2) . It
has been proposed that a cycle of regulated nucleotide exchange and GTP
hydrolysis is superimposed on the cycling of Rab proteins between donor
and acceptor compartments to ensure accurate and directional vesicle
transport. Among the known Rab GTPases, Rab5 is of great interest
because it appears to be rate-influencing for receptor-mediated and
fluid-phase endocytosis. Lateral fusion between endocytic vesicles is
stimulated by Rab5 both in vitro and in vivo, and
antibodies against Rab5 inhibit fusion in
vitro(3, 4, 5) . Endosome fusion in
vitro is inhibited by cytosol containing overexpressed mutant Rab5
N133I protein that has impaired guanine nucleotide binding (3) . In cells overexpressing Rab5 N133I, the rate of
receptor-mediated and fluid-phase endocytosis is significantly
decreased compared with control(4, 5) . The N133I
mutant is believed to interfere with endocytosis by interacting
nonproductively and competitively with some important component of the
endocytic apparatus. If Rabs must cycle between GDP- and GTP-bound
forms in order to function, then a mutation reducing the GTPase
activity (such as the cognate of H-Ras Q61L) is predicted to also be a
dominant inhibitor. This is the case with Rab2, but not with
Rab1(6) , and the Sec4 mutation has an intermediate
phenotype(7) . Wild-type Rab25 actually contains Lys at the
cognate position(8) . The molecular basis for the functional
differences among these related proteins is not known. Although
several in vivo and in vitro functional studies of
Rab5 have been published, this protein has not yet been subjected to a
detailed biochemical analysis. Correct interpretation of in vivo and in vitro experiments using WT (
Figure 1:
Purification of recombinant Rab5 WT. Lane1, molecular weight protein standards (M
Figure 2:
Binding of [
The
effect of 0.1% CHAPS on [
Figure 3:
Effects of CHAPS on guanine nucleotide
binding to Rab5. A, association of
[
The findings with Rab5
contrast with the lack of effect of CHAPS on nucleotide binding to
unprocessed Rab6 (31) but are not unique to Rab5 in that
similar effects were observed by us with bacterial recombinant Rab4
(not shown). Of note, both nucleotide binding and GTPase activities of
Rab6 differ dramatically between the processed and unprocessed
forms(31) . Together with our results, this suggests that
conformational effects on the nucleotide binding and hydrolyzing site
of Rab proteins may be induced either by covalently attached prenyl
groups or by noncovalently associated detergents, and such effects may
vary from one Rab protein to another. Supporting the involvement of the
carboxyl terminus of Rab5 in guanine nucleotide binding was our finding
that Rab5 truncated after amino acid 184 (cognate of
H-Ras
Figure 8:
Proteolysis of Rab5 proteins by trypsin.
Purified recombinant Rab5 WT (A), Q79L (B), or N133I (D), were preincubated in the absence or the presence of 10
mM nucleotides and 5 mM MgCl
The
pseudo-first order association rate constant of
[
Figure 4:
Association of
[
Figure 5:
Dissociation of guanine nucleotides from
Rab5 WT and Q79L. Proteins were preincubated with 500 nM labeled nucleotide for 3 h and then chased with 1 mM cold
nucleotide at 30 °C. One hundred-µl aliquots were diluted with
cold washing buffer and filtered through nitrocellulose at the
indicated times. A, [
Figure 6:
Rab5 WT and Q79L GTPase activities. A, steady-state [
The net GTPase reaction
may be schematized as in Fig. 7, in which the transit time is
given by ,
Figure 7:
Rab5 GTPase cycle. A simplified scheme of
the GTPase cycle is illustrated, which does not resolve the chemical
step of GTP hydrolysis from P
where k` is the single net rate constant for each step (40) . If the rate of nucleotide binding to apo-Rab5 (k The rate for Rab5
WT is intermediate among the rates of other Rabs measured by charcoal
sedimentation (summarized in (35) ). Since the hydrolytic step
for Rab5 WT (transit time, 16 s) is considerably faster than the GDP
dissociation step (transit time, 256 s), the overall rate constant
approaches that of GDP dissociation. In contrast, due to the
accelerated GDP dissociation rate and the retarded GTPase rate of Rab5
Q79L, both the hydrolytic step (transit time, 196 s) and the GDP
dissociation step (transit time, 71 s) contribute substantially to the
overall rate constant.
In
the absence of added nucleotides, proteolysis of Rab5 WT yielded two
major fragments of 12 and 8 kDa (Fig. 8A). It should be
noted that Rab5 is not expected to be stoichiometrically free of bound
guanine nucleotide in this circumstance because Ras and other Rabs are
purified from E. coli with GDP
bound(17, 34, 35) , and the concentration of
GDP in solution was estimated at 30 nM after dilution of the
protein as described under ``Materials and Methods.''
Nonetheless, the conditions of proteolysis were optimized to
demonstrate the nucleotide-dependence of proteolysis. An ``empty
state'' distinct from the GDP-bound state was effectively
demonstrated since there was virtually none of the 20-kDa fragment,
which was almost stoichiometrically generated in the presence of added
GDP (Fig. 8A). It was not clear what the relationship
is between the tryptic fragments observed by us and the one reported by
Steele-Mortimer et al.(47) while this manuscript was
in preparation. In that report, the major fragment appears by SDS-PAGE
to be approximately 20 kDa, but amino-terminal sequencing indicates
that only four amino acids had been digested. The results for Rab5
Q79L were almost identical to those for WT (Fig. 8B),
suggesting that the mutation did not result in major conformational
changes. The only difference was that for Rab5 Q79L, GDP was less
effective in preventing degradation to 12- and 8-kDa peptides,
consistent with the accelerated dissociation of GDP from Rab5 Q79L
compared with WT (Fig. 5A). Rab5 N133I did not
undergo any observable shift in mobility by SDS-PAGE in the presence of
10 mM GDP (Fig. 8D) or GTP (not shown), but 10
mM GTP Our
results also suggest that the mechanism of dominant inhibition of
endocytic function by Rab5 N133I (3, 4, 5) is
via nonproductive interaction with downstream (i.e. Rab5
Figure 9:
In vitro prenylation of Rab5
proteins. Prenylation of Rab5 WT, Q79L, and N133I proteins was
accomplished using rabbit reticulocyte lysate and
[
While both Rab5 WT and Q79L become
post-translationally modified in vitro, the extent of protein
processing is limited (<2%). Other investigators have reported
similar results in studies of the in vitro prenylation of
recombinant Rab proteins isolated from E.
coli(29, 30) . This is in contrast to
modification of Rab proteins overexpressed in mammalian cells or
expressed in a reticulocyte lysate, wherein such proteins are
apparently processed quite
efficiently(4, 5, 20, 39) . What is
clear from our results is that post-translational modification of Rab5
is essential for its function in endosome fusion (see below), but what
remains to be determined is in what ways the in vitro prenylation of the E. coli-expressed protein differs from
that synthesized by mammalian systems. For example, when translated by
the very same reticulocyte lysate used to support the modification of
the recombinant proteins, newly synthesized Rab5 rapidly incorporates
geranylgeranyl and is converted to a fully processed form within
3-4 h of incubation(20) . In contrast, similar amounts of
bacterially expressed Rab5 proteins failed to become fully processed (Fig. 9), and it is possible that other events of co- or
post-translational processing may occur in the reticulocyte lysate that
are not supported by E. coli. Such is, in fact, the case for
the post-translational modification of H-Ras, which is known to be
palmitoylated on a cysteine residue upstream from the carboxyl-terminal
site of isoprenylation(49) . In support of a similar situation
for Rab5, there is at least one report of fatty acid acylation of Rab
family members(50) , and it is possible that other protein
processing steps can occur, including proteolysis and
carboxymethylation at the carboxyl terminus. Another limiting factor
for in vitro prenylation of recombinant Rab5 may result from
the rather unique features of Rab geranylgeranyl
transferase(51, 52) . A component of this enzyme, REP
(Rab escort protein), continues to interact with Rab proteins after
geranylgeranylation is complete(53) . The complex between REP
and Rab proteins is thought to dissociate only upon interaction with
other cellular factors, such as GDI guanine nucleotide dissociation
inhibitor(54) . It is possible that for the E.
coli-expressed protein, interactions with factors such as REP or
GDI are somehow limited in the in vitro system, accounting for
incomplete processing.
Our cell-free system has
been previously shown to support fusion between early endocytic
vesicles(21) , an activity that is inhibited by the presence of
GTP
As noted above, Rab5 N133I was a poor
substrate for geranylgeranylation and progressively precipitated after
purification. It was therefore not possible to determine whether its
failure to significantly affect in vitro endosome fusion (not
shown) was due to inadequate prenylation, structural instability, or an
intrinsic property of the protein. Rab5 Q79L, in contrast, was
processed to a similar extent to Rab5 WT and was stable in the presence
of guanine nucleotides and magnesium. The modified Q79L mutant not only
failed to inhibit fusion activity but actually stimulated fusion on
every occasion tested, although the results for stimulation were not
statistically different from control at 95% confidence intervals (p = 0.065). The stimulatory effect of Rab5 Q79L was never
greater than or equal to that observed for WT in matched experiments,
consistent with the results of Stenmark et al.(39) . One inference that can be drawn from our results, assuming that the
Q79L protein is not functionally inactive despite its activity in the
above biochemical assays, is that Rab5 Q79L does not inhibit in
vitro endosome fusion. This is surprising given that the
equivalent GTPase-defective mutant of Ras is a dominant promoter of
transformation(23) , and Rab cognates could be expected to be
dominant inhibitors of vesicle transfer. However, in vivo studies of cells overexpressing Rab5 Q79L document that this
mutant instead acts as does Rab5 WT to stimulate
endocytosis(5, 39) . The cognate GTPase defective
mutant of Rab1B also fails to interfere with in vivo ER to
Golgi transport, although Rab2 Q65L is a potent inhibitor of this
transport step when overexpressed(6) . One explanation for
these apparently conflicting results might be the near-normal k
Volume 270,
Number 10,
Issue of March 10, 1995 pp. 5048-5056
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS AND DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
10-fold. The
single-step catalytic rate of Rab5 WT exceeded that of Q79L 12.2-fold,
but the steady-state GTPase rate was only 2.8-fold greater because GDP
dissociation was rate-limiting and GDP dissociation was 3.6-fold slower
than for Q79L. In contrast, dissociation rates of GTP were
indistinguishable. Binding to Rab5 N133I was not detectable. GTP
protected Rab5 WT and Q79L from any apparent proteolysis by trypsin. A
20-kDa fragment was the major product of digestion in the presence of
GDP, and 12- and 8-kDa fragments were the major products in the absence
of added guanine nucleotides. Rab5 N133I underwent no apparent
proteolysis with 10 mM GTP or GDP, suggesting a
``triphosphate'' conformation may be induced in Rab5 N133I by
either GTP or GDP. Partially geranylgeranylated Rab5 WT stimulated
endosome fusion in vitro, whereas unmodified Rab5 WT did not.
Processed Rab5 Q79L failed to inhibit endosome fusion, and Rab5 N133I
could not be geranylgeranylated. These findings identify biochemical
and structural features of Rab5 proteins, providing data for the
interpretation of functional assays.
)and mutant
Rab5 proteins depends on thorough knowledge of the biochemistry of
these reagents. As a step toward understanding the role of Rab5 in
endosome fusion and endocytosis, we have expressed in Escherichia
coli Rab5 WT, its N133I mutant, and a putative GTPase defective
mutant (Q79L). The biochemical properties, particularly the kinetics of
nucleotide binding and GTPase activities, of purified recombinant
proteins were characterized; functional properties of the purified
proteins were studied using an in vitro endosome fusion assay
to verify predicted phenotypes; and nucleotide-dependent structural
properties of the proteins were analyzed by limited proteolysis.
Construction of Rab5 Mutants
Recombinant
plasmids with human Rab5 and Rab4 cDNAs were obtained from A.
Tavitian(9) . The cDNA inserts were amplified by add-on
polymerase chain reaction and cloned into M13mp18 for nucleotide
sequencing(10) . DNA sequence analysis of Rab5 clones from
several independent polymerase chain reactions identified three
discrepancies from the published sequence(9) : Arg replacing
Gly at position 81, Val replacing Ala-86, and Arg replacing Gly-197.
The Gly to Arg conversions are identical to residues found in these
positions (amino acids 81 and 197) in canine Rab5(11) . The
remaining alteration, Val to Ala, is conservative and most likely does
not affect the functional integrity of the protein; in fact, the
product of this cDNA can bind and hydrolyze GTP, can be
geranylgeranylated, and functions in an endosome fusion assay (see
``Results and Discussion''). Site-directed point mutagenesis
of Rab5 was performed by the method of Kunkel et
al.(12) . Two point mutants, Q79L and N133I, were
generated using the oligonucleotides 5`-GGTATCGTTCTAGACCAGCTGTA-3` and
5`-AGGTCCGGCCTTGATTCCCGATAAAG-3`, respectively. Mutants were verified
by nucleotide sequence analysis in both directions. WT and mutant Rab5
cDNAs and the Rab4 WT cDNA were cloned into the T7-polymerase
expression plasmid pT7.7 (13) utilizing the NdeI
cloning site to ensure the production of recombinant proteins with
native amino termini.Purification of Rab Proteins
E. coli BL21(DE3) cells were transformed with recombinant plasmids and
grown in LB medium containing 50 µg/ml ampicillin. When the A
was 1.0,
isopropyl-1-thio-
-D-galactopyranoside was added to a
final concentration of 0.8 mM. Cells were harvested 3-4
h later by centrifugation at 1500 
g for 30 min. The
cell pellet from a 2-liter culture was resuspended in Tris-buffered
saline (20 mM Tris-HCl, pH 8.0, 150 mM NaCl),
centrifuged, and then washed with the same buffer. The washed pellet
was suspended in 10 ml of hypertonic buffer (2.4 M glucose, 20
mM Tris-HCl, pH 8.0, 1 mM EDTA) and kept on ice for
30 min. One hundred ml of a lysozyme solution (0.5 mg/ml lysozyme, 0.02
mg/ml DNase, 40 mM Tris-HCl, pH 8.0, 5 mM MgCl
, 1 mM EDTA) was then added, and the
incubation was continued on ice for a further 15 min. One hundred ml of
lysis buffer (20 mM Tris-HCl, pH 8.0, 1 mM EDTA, 5
mM MgCl, 1 µM GDP, 2 mM -mercaptoethanol, 0.2 mM phenylmethylsulfonyl
fluoride) was then added. The lysate was centrifuged in a GSA rotor
(Sorvall) at 11,000 rpm for 30 min, and the supernatant was filtered
through a glass fiber filter. CHAPS was added to a final concentration
of 0.1% at this point, except where noted. This solution was applied to
a 500-ml DEAE-Sepharose FF column equilibrated with several volumes of
Buffer A (20 mM Tris-HCl, pH 8.0, 1 mM EDTA, 2 mM -mercaptoethanol, 5 mM MgCl, 1
µM GDP, 0.1% CHAPS). Protein that did not bind to the
resin was collected and concentrated on an Amicon YM-10 membrane and
then applied to a 1500-ml Sephacryl S-200 column equilibrated with
Buffer A plus 150 mM NaCl. Rab proteins in eluting fractions
were identified by SDS-PAGE, and concentrated by membrane filtration.
The N133I mutant was insoluble when expressed in E. coli,
similar to the cognate Ras mutant(14) . Therefore, the protein
was extracted from the bacterial pellet with 8 M urea, which
was removed during chromatography on the gel filtration column. Rab4
was purified by sequential chromatography as for Rab5 WT, except that
Rab4 bound to DEAE-Sepharose and was eluted with a linear gradient of
NaCl in Buffer A. Protein concentrations were determined with the
Coomassie Plus assay (Pierce), using bovine serum albumin as a
standard. Proteins were stored at 2-5 mg/ml in Buffer A at
-80 °C.Binding of Nucleotides in Solution
Nucleotide
binding was determined by a rapid filtration technique(15) .
Binding buffer contained 20 mM Tris-HCl, pH 8.0, 250 mM NaCl, 1 mM EDTA, 5 mM MgCl, 1 mM dithiothreitol, 0.1% CHAPS, 500 nM guanine nucleotide,
and either [
S]GTP
S,
[
-
P]GTP, or [
H]GDP
(all 2,000 cpm/pmol). The protein concentration was 50 nM in a
volume of 100 µl. After incubation at 30 °C for the indicated
times, samples were diluted with 4 ml of iced buffer, filtered through
BA85 nitrocellulose (Schleicher and Schuell), and then washed twice
with 4 ml of iced buffer. Filters were then dried, immersed in
scintillation fluid, and counted by scintillation counting. Data points
in all figures represent the mean of triplicate determinations from a
representative experiment that was repeated 2-4 times. Except
where noted, maximum binding was determined by extrapolation from
experimental data using the program Enzfitter (Biosoft, Ferguson, MO).
In the presence of 0.1% CHAPS and 5 mM MgCl,
maximum binding averaged 1.2 ± 0.2 mol of nucleotide/mol of
Rab5.Measurement of Dissociation Rates
These were
measured as described previously(16) . Protein was preincubated
for 3 h in the conditions described above for nucleotide binding, and
then the bound labeled nucleotide was competed with 1 mM unlabeled nucleotide. At indicated times, samples were diluted
with cold buffer, filtered, and counted, as above. In some experiments,
Rab5 was rapidly loaded with radiolabeled guanine nucleotides by a
modification of the procedure of Tucker et al.(17) as
follows. Protein was incubated in the presence of 1.25 mM EGTA
for 10 min at 30 °C to release prebound nucleotide and then
radiolabeled nucleotide was added, followed by 5 mM MgCl.GTPase Activity
Steady state GTPase activity was
measured by the release of [P]P
using the ``charcoal assay'' as described(18) .
Protein was incubated with 500 nM [-P]GTP (2,000 cpm/pmol) in the
presence of 5 mM MgCl at 30 °C, and at the
indicated times aliquots were withdrawn and mixed with charcoal.
Following centrifugation, the [P]P in the supernatants was assayed by scintillation counting.
Pre-steady state GTPase rates were estimated by the charcoal method at
early times following rapid loading of Rab5 as above. The single-step
GTPase rate was determined by a filtration assay following rapid
nucleotide loading. This assay was performed essentially as described
for the measurement of dissociation rates, with the exceptions that
[-P]GTP was used, and the GTPase rate was
calculated by subtracting the [S]GTPS
dissociation rate from the apparent GTPase rate(19) . All rates
were calculated using the program Enzfitter.In Vitro Processing of Recombinant Rab5
Proteins
Post-translational modification of Rab proteins with
geranylgeranyl was accomplished by an in vitro reaction
supported by rabbit reticulocyte lysate (Promega, Madison, WI), as
described(20) . Briefly, Rab proteins (final concentration,
0.1-0.8 mg/ml) were added to lysate together with 5 µM [H]geranylgeranyl pyrophosphate (33,300
dpm/pmol, American Radiolabeled Chemicals) in 10 mM Tris, pH
8.0, 0.1% CHAPS, 1 µM GDP, 5 mM MgCl,
1 mM EDTA, 2 mM -mercaptoethanol; the final
mixture contained 40-60% reticulocyte lysate (v/v). The reaction
mixtures were incubated at 37 °C for 3-4 h, after which time
aliquots (3.4 µl) were removed for analysis by PAGE and
fluorography. The amount of [H]geranylgeranyl
incorporated into the proteins was determined by excising the bands
from the dried gel, dissolving samples in 30% HO at 65 °C, and scintillation counting. After processing but
prior to the addition to in vitro endosome fusion assays, the
remaining reaction mixture was chromatographed over a Sephadex G25 spin
column to remove excess radioactivity and to exchange the Rab protein
into the appropriate buffer (20 mM Hepes, pH 7.4, 0.1 M KCl, 85 mM sucrose, 20 µM EGTA).In Vitro Endosome Fusion Assay
Cell-free vesicle
fusion assays were performed exactly as described
previously(21, 22) . Postnuclear supernatant fractions
were prepared from K562 cells that had endocytosed either biotinylated
transferrin (BTf) or avidin--galactosidase (AvGal) at 20
°C for 45 min to load early endosomes. The postnuclear supernatant
fractions were then dialyzed against 20 mM Hepes, pH 7.4, 0.1 M KCl, 85 mM sucrose, 20 µM EGTA, and
frozen at -80 °C until use. Postnuclear supernatant fractions
were rapidly thawed immediately prior to the assay, and 5-µl
aliquots of each were added to a reaction mixture that contained 1
mM MgATP, 50 µg/ml creatine kinase, 8 mM phosphocreatine, 10 µg/ml biotin-insulin, 1 mM dithiothreitol, 400 µM GTP, with appropriate
additions of Rab proteins and/or rabbit reticulocyte lysate as detailed
in the legend to Table 1. Fusion activity was promoted by
incubation of the reaction mixtures at 37 °C, while control
reactions (4 °C) were held on ice. Specific vesicle fusion results
in co-localization of AvGal and BTf within fused vesicles, and the
resulting avidin:biotin complex between the two probes was measured by
a modified enzyme-linked immunosorbent assay using a fluorogenic
-galactosidase substrate as detailed previously (21) . The
signal in fluorescence units is directly proportional to the extent of
endocytic vesicle fusion in the in vitro reaction. Samples
(15-20 µl) of desalted prenylation reaction mixtures were
tested in each assay.
Expression of Rab5 and Two of Its
Mutants
Induction of E. coli containing recombinant
plasmids resulted in the accumulation of a major protein of 25 kDa in
lysates as assessed by SDS-PAGE and Coomassie Blue staining (Fig. 1, lane2). Prior to purification, the
predicted biochemical phenotypes of WT and mutant proteins were
qualitatively confirmed by [P]GTP overlay
assays. Lysates were Western blotted onto nitrocellulose filters and
incubated with [-P]GTP (Fig. 2A) or [-P]GTP (Fig. 2B). Autoradiography revealed that Western
blotted Rab5 WT bound and hydrolyzed GTP, as indicated by reduced
density of the 25-kDa band with -labeled GTP as compared with
-labeled GTP. In contrast, Rab5 Q79L showed a much smaller
difference in band intensity between -labeled and -labeled
GTP, suggesting a reduced GTPase activity as expected from similar
effects of the cognate mutation in Ras (23) and in other Rabs (7, 19, 24) . Rab5 N133I had a severe guanine
nucleotide binding defect, as indicated by the complete absence of an
autoradiographic signal with nucleotide labeled in either position,
consistent with previous results (4, 5) . Cognate
mutations in Ras (14, 25) and in other Rabs (6, 24, 26, 27) result in
undetectable guanine nucleotide binding as assayed by rapid filtration
or GTP overlay. This severe defect in nucleotide binding is not
surprising in view of the central role of the highly conserved
asparagine in linking three structural elements directly involved in
nucleotide binding, as inferred from the crystal structures of Ras and
EF-Tu(28) .
10
); lane2, whole cell lysate of E. coli BL21(DE3)-pT7.7-Rab5 culture (60 µg); lane3, supernatant after 20,000 g spin and
filtration through a glass fiber filter (40 µg); lane4, DEAE-Sepharose FF flow-through pool (8 µg); lane5, Sephacryl S-200 HR peak pool (4 µg).
Proteins were resolved by electrophoresis through a 12% polyacrylamide
SDS gel.
P]GTP
to Western blotted proteins. Crude proteins (2 µg/lane)
were resolved by SDS-PAGE, electrotransferred to nitrocellulose, and
then incubated with (A) [-P]GTP or (B) [-P]GTP (both 10
cpm/pmol) for 1 h at room temperature. The nitrocellulose was
then washed for 1 h, dried, and exposed to x-ray film as
described(57) .
Detergent Effects on Guanine Nucleotide Binding to
Rab5
During our initial purification of Rab5, much of the
partially fractionated protein precipitated. Efforts were therefore
directed at stabilizing Rab5 in solution without losing
[S]GTPS binding activity. Several nonionic
and zwitterionic detergents caused a dramatic increase in
[S]GTPS binding at a concentration of
0.05%, as did 0.01% bovine serum albumin, but there was no detectable
binding in the presence of the ionic detergent SDS (not shown). When
the concentration of the detergents was increased to 0.5%, only CHAPS
still supported high [S]GTPS binding.
Hence, subsequent protein purification and biochemical experiments were
performed in buffers containing 0.1% CHAPS unless indicated.S]GTPS binding to
Rab5 was time-dependent as shown in Fig. 3A. To test
the hypothesis that the increase in [S]GTPS
binding induced by CHAPS might be due to acceleration of prebound
nucleotide dissociation, 100 nM [H]GDP
was loaded onto Rab5 by transient magnesium chelation. However, the
dissociation of [H]GDP from Rab5 in the presence
of 0.1% CHAPS was slower than in its absence (Fig. 3B, diamonds). When excess unlabeled GDP was not added after
loading [H]GDP, binding remained stable in the
presence of CHAPS but declined rapidly in its absence (Fig. 3B, circles). These results suggest that
without added detergent, Rab5 quickly assumes a conformation that does
not bind guanine nucleotides with high affinity. This conclusion is
consistent with the decline in the small amount of
[S]GTPS initially bound to Rab5 in the
absence of CHAPS during the course of prolonged incubation (Fig. 3A, opencircles). It should be
noted that the recombinant protein is not post-translationally modified
when expressed in E. coli, and therefore lacks geranylgeranyl
groups that are attached to Rab proteins in
eukaryotes(29, 30) .
S]GTPS. Recombinant Rab5 WT purified in
the absence of CHAPS was incubated with 500 nM [S]GTPS either in the absence or
presence of 0.1% CHAPS for 240 min at 30 °C or was incubated
initially in the absence of CHAPS for 90 min and then in the presence
of 0.1% CHAPS for the next 150 min. At the indicated times, aliquots
were removed and rapidly filtered through nitrocellulose. B,
dissociation of [H]GDP. Rab5 WT was loaded with
[H]GDP by transient magnesium chelation followed
immediately by the addition of binding buffer with or without 0.1
mM GDP and with or without 0.1% CHAPS (final concentrations).
Aliquots were removed at the indicated times and assayed by vacuum
filtration through nitrocellulose.
) was insoluble in E. coli (not shown),
similar to Rab5 N133I and to several Ras mutants that are both
defective in nucleotide binding and insoluble in E. coli(14) . In contrast, truncated H-Ras is
soluble in E. coli and has guanine nucleotide binding and
hydrolyzing properties indistinguishable from those of the full-length
protein (32) . In addition, reciprocal interactions between
guanine nucleotide binding and guanine nucleotide dissociation
inhibitor binding to carboxyl-terminally prenylated Rabs have been
described for other Rab proteins(1) , suggesting an interaction
between the carboxyl terminus and the guanine nucleotide-binding
domains. Of interest, guanine nucleotide binding by ADP-ribosylation
factor, another small GTPase that regulates vesicular traffic, is
highly dependent on interactions with lipids and detergents when its
acylated amino terminus is intact (33) .Purification of Bacterial Recombinant Rab5 and Two of Its
Mutants
After stabilization of Rab5 WT and Q79L in solution with
0.1% CHAPS, 1 µM GDP, and 5 mM MgCl,
these proteins were readily purified from the supernatants of E.
coli lysates. Rab5 WT (calculated pI, 7.05) did not bind to
DEAE-Sepharose at pH 8.0 and was recovered in the flow-through, while
the bulk of contaminating E. coli protein remained on the
column (Fig. 1, lane4). Most of the residual
contaminating protein was resolved by molecular sieve chromatography,
yielding Rab5 WT of approximately 95% purity as estimated by Coomassie
staining after SDS-PAGE (Fig. 1, lane5). A
minor 20-kDa contaminant appeared to be a degradation product of Rab5
due to carboxyl-terminal proteolysis since it bound
[P]GTP on a GTP overlay, it reacted with
affinity-purified antiserum to Rab5 holoprotein, and the amino terminus
was intact by Edman degradation (not shown). Rab5 Q79L was purified by
an identical scheme and yielded protein of similar purity (see Fig. 8B). Rab5 N133I was insoluble in E. coli lysates and was extracted with 8 M urea. Urea was removed
during the molecular sieve chromatography, but approximately 90% of the
protein subsequently precipitated. The remaining Rab5 N133I (see Fig. 8D) was fairly stable in solution in Buffer A
containing 1 mM GDP.
for 1 h at
30 °C. Proteins (2.5 µg) were then incubated for 1 h with or
without 0.25 µg of trypsin in the presence of the indicated
nucleotides in a total volume of 50 µl at 30 °C, as described
previously(42) , and boiled for 5 min in sample buffer, and the
resulting peptides were resolved by SDS-PAGE in a Tris-Tricine buffer
system(58) . PanelC depicts an identical
experiment to panelA, except that the concentration
of trypsin was reduced to 0.05 µg.
Kinetics of Binding of Guanine Nucleotides to Rab5 WT and
Q79L
In preliminary studies, the binding of
[S]GTPS to Rab5 N133I could not be detected
by rapid filtration assay. This is consistent with the Western blot
binding assay (Fig. 2) and with results from other Ras family
proteins and was therefore probably not simply due to extraction in
urea. Data acquired subsequently from proteolysis (see below) indicate
that purified Rab5 N133I does effectively interact with guanine
nucleotides but with an apparent affinity that is several orders of
magnitude lower than those of Rab5 WT and Q79L. Therefore, binding and
GTPase studies were performed only with Rab5 WT and Q79L.S]GTPS with Rab5 Q79L (0.013
s
) was 2.3-fold faster than that with WT (0.0057
s) (Fig. 4). Since previously purified Ras
family proteins had GDP stoichiometrically bound due to their very high
affinities(17, 34, 35) , we hypothesized that
the difference in association rates was due to differences in
dissociation rates of GDP. As predicted, the rate of dissociation of
GDP from Rab5 Q79L exceeded that from WT 3.6-fold (Fig. 5A). This is similar to the 3.2-fold increase in
the k
of the cognate mutant of Rab3A (22) and the 6.7-fold increase in the Ram cognate(19) ,
but it contrasts with the 2-fold decrease in the k of the cognate mutant of
G
(36) . Differences in the effects of the Gln
Leu mutation on nucleotide exchange rates among different
GTPases may account for some of the phenotypic differences
observed(6, 7, 36, 37) . In contrast
to GDP, the dissociation rates of GTP and GTPS did not
significantly differ between Rab5 WT and Q79L (Fig. 5, B and C); such comparisons have not been reported for other
Rabs. The first-order dissociation rate constants for each nucleotide
are shown in Table 1and are intermediate among those of other
Rabs (summarized in (35) ). The ratio k/k
is 0.93 for
Rab5 WT, similar to that of Ras and other Rabs with the exception of
Sec4, which is unique in having a ratio of
120(31, 35, 38) .
S]GTPS with Rab5 WT and Q79L. Rab proteins
(50 nM) were incubated in the presence of 500 nM [S]GTPS at 30 °C for the indicated
periods of time. One hundred-µl aliquots were diluted with cold
washing buffer and filtered through
nitrocellulose.
H]GDP; B,
[-P]GTP; C,
[S]GTPS.
GTPase Activities of Rab5 WT and Q79L
The GTPase
activities of Western blotted Rab5 WT and Q79L appeared dramatically
different (Fig. 2), but GTPase rates measured under steady-state
conditions by charcoal sedimentation revealed only a modest 2.8-fold
difference (Fig. 6A and Table 2). We hypothesized
that in view of the relatively slow dissociation rate of GDP (Fig. 5A), a more rapid catalytic rate of Rab5 WT
compared with Q79L might be masked by a slower GDP dissociation rate.
To test this, pre-steady-state GTPase rates were measured by rapidly
loading Rab5 proteins with GTP by magnesium chelation and then
measuring P release at early times by charcoal
sedimentation (Fig. 6B). This yielded a rate for Rab5
WT (0.026 s), which exceeded that of Q79L (0.0038
s) by 6.8-fold, supporting the hypothesis. The
single step first-order rate constant for GTP hydrolysis was then
determined by rapid filtration (Fig. 6C) and was found
to be 12.2-fold higher for Rab5 WT than for Q79L (Table 2).
(During preparation of this manuscript, a single step rate constant of
0.05 min at 37 °C for Rab5 WT was reported,
which agrees well with our data, but the rate for Rab5 Q79L was found
to be about 100-fold slower(39) ).
-P]GTPase
activities were measured over 4 h at 30 °C by the charcoal method. B, following rapid loading of Rab5 proteins with
[-P]GTP, pre-steady state GTPase rates were
measured at early times by the charcoal method. C, single-step
GTPase rates were measured by the filtration method. See
``Materials and Methods'' for experimental details and
calculations.
dissociation because these
are not readily analyzed independently and which does not show any
conformational states. Each step is numbered such that the forward and
reverse rate constants for Step i are k and k
, but if the reverse reaction is
negligible, it is not included. Transit times for individual steps were
calculated as the inverse of the individual step net rate constants,
and the complete cycle transit time was calculated neglecting k` due to its presumed negligible contribution to
the overall rate (see text for details).
) is exceedingly fast compared with the rates of
GTP dissociation (k), GTP hydrolysis (k), and GDP dissociation (k)
(for H-Ras, k
= 1.4
10
M s(41) , and for Rab9, k = 1.2 10
M s(35) );
the hydrolysis of GTP is essentially irreversible under physiologic
conditions; and the association of GDP with apo-Rab5 may be ignored in
the presence of excess GTP, then reduces to
1/k
= 1/k +
1/k. This rearranges to k = k
k/k + k. Calculated this way from the
measured first-order rate constants k and k, the turnover number (k)
of both Rab5 WT and Q79L is 0.0037 s (Table 2). The calculated and measured k of Rab5 WT are nearly identical, but those of Q79L differ
approximately 3-fold (Table 2). Since measurement of the
first-order rate constants k and k by filtration does not depend on independent determination of the
number of active Rab5 molecules, these constants should be highly
accurate, and determination of k by calculation
from single step rate constants may be preferable to direct measurement
of steady state kinetics by the charcoal method.Nucleotide-dependent Proteolysis
The structures of
Rab5 proteins in solution were assessed by limited tryptic digestion.
The products of degradation were strongly dependent on the presence of
nucleotides and magnesium, similar to what has been observed for other
GTPases(36, 37, 42, 43) . In the
absence of added nucleotides, magnesium chelation with 1 mM EDTA resulted in the progression of proteolysis such that no
peptide fragments were visible by SDS-PAGE with Coomassie staining (not
shown). The addition of 10 mM GDP, GTP, or GTPS in the
presence of 5 mM MgCl resulted in the appearance
of a proteolytic doublet of approximately 20 kDa (Fig. 8A). However, protection by these different
guanine nucleotides was not comparable, since GDP consistently
permitted the appearance of faint 12- and 8-kDa fragments,
characteristic of the empty state (see below). GTP allowed almost no
appearance of the 12- and 8-kDa peptides under the conditions used, and
GTPS showed an intermediate pattern. Furthermore, when the
trypsin/Rab ratio was increased from 1:10 to 1:50 (w:w), a band
indistinguishable from the 25-kDa holoprotein was observed in the
presence of GTP but not in the presence of GDP or GTPS (Fig. 8C). At even higher ratios, the protection of the
25-kDa protein was almost complete (not shown). These results suggest
that limited proteolysis may be used to distinguish the Rab5GTP
conformation from the Rab5GDP conformation and that GTPS
does not efficiently induce a conformation comparable with
Rab5GTP. The latter is distinct from the situation with trimeric
G-proteins in which GTPS promotes greater functional activation
and confers greater protection from trypsin than
GTP(36, 37) . This difference between trimeric
G-proteins and a Ras-related protein to a ligand that differs only at
the -phosphate position is reminiscent of the sensitivity of
G-proteins but not small GTPases to aluminum fluoride(44) . The
conformational differences between Rab5GTP and
Rab5GTPS detected by limited proteolysis raise the question
as to whether the inhibitory function of GTPS in endosome fusion
reactions (22, 45, 46) is due to the failure
of GTPS to induce a GTP conformation in Rab5 rather than to the
inhibition of the GTPase activity of Rab5 or some other GTPase.S conferred no protection (Fig. 8D). In contrast to Rab5 WT and Q79L, which were
highly protected by 100 µM guanine nucleotides, Rab5 N133I
was unprotected at this concentration (not shown). These results
confirm that Rab5 N133I has an extremely low affinity for guanine
nucleotides and suggest that its conformation with either GTP or GDP
bound is similar to the Rab5 WTGTP conformation. This is
consistent with the transforming phenotype (implying an active GTP-like
conformation) of H-Ras N116I (14, 25) but raises the
question of whether the biochemical activities of Asn Ile
mutants are dependent on the phosphorylation state of bound guanine
nucleotides. Similarly, it has been suggested that activation of the
downstream signaling pathway by H-Ras G12V may be less guanine
nucleotide-dependent than activation by H-Ras WT(48) .GTP-interactive) vesicle transfer regulatory components,
since the product of proteolytic protection of Rab5 N133I by native
guanine nucleotides was similar to that of Rab5 WTGTP. The fact
that GTPS offered no protection from proteolytic digestion (Fig. 8D) may simply reflect the relative inability of
GTPS to induce a GTP-like conformation of Rab5 N133I, as was the
case for Rab5 WT (see above), or it may additionally reflect structural
instability of the N133I mutant with heightened sensitivity to subtle
differences among ligands.In Vitro Prenylation of Recombinant Rab5
Proteins
Prenylation of purified Rab5 proteins was supported by
rabbit reticulocyte lysate upon incubation for 3 h in the presence of
[H]geranylgeranyl pyrophosphate, and the extent
of processing was assessed by SDS-PAGE and fluorography (Fig. 9). The incorporation of
[H]geranylgeranyl into each protein was as
follows: Rab5 WT, 7,366 cpm, or 1.2% of the protein modified; Rab5
Q79L, 7,816 cpm, or 1.3% of the protein modified; Rab5 N133I, 2,793
cpm, or less than 0.4% of the protein modified. Previous studies have
established that co-translational prenylation of Rab5 is dependent on
guanine nucleotide binding and that Rab5 N133I is a very poor substrate
for this reaction (20) .
H]geranylgeranyl pyrophosphate as described
under ``Materials and Methods.'' Proteins were resolved on a
urea (4-8 M)/acrylamide (10-15%) gel, and then
processed for fluorography.
Stimulation of in Vitro Endosome Fusion by Recombinant
Rab5 Protein and Its Mutants
In order to characterize the
recombinant WT and mutant Rab5 proteins functionally, their activity in
regulating endosome fusion in vitro was assessed in a
well-established assay(21) . Previous studies have documented
the ability of Rab5 to stimulate endocytosis when overexpressed in
vivo(4, 5, 39) , and in vitro studies utilizing cytosol from transfected, Rab5-overexpressing
cells have demonstrated the capacity of this preparation to stimulate
endosome fusion(3, 39) . Our investigation of the
functional properties of Rab5 is unique in that the recombinant protein
was highly purified, it was not modified by fusion to another
polypeptide such as glutathione S-transferase, and the exact
role of post-translational processing of the cognate protein could be
assessed. Using this approach, it was also possible, for the first
time, to directly compare the properties of Rab5 WT with the Q79L
mutant within the same experimental system.S in vitro(22) . As demonstrated by the
results of Table 3, processed (geranylgeranylated) Rab5 WT
stimulated in vitro vesicle fusion when added to postnuclear
supernatant fractions containing endosomes that were not depleted of
any protein factors. This result suggests that Rab5 is rate-limiting
for endosome fusion, an idea consistent with observations reported for
transfected, Rab5-overexpressing
cells(4, 5, 39) . Unprocessed Rab5 either had
no activity in our assay or suppressed vesicle fusion slightly (Table 3). This result is consistent with in vivo deletion studies performed with overexpressed Rab5, which
indicated that removal of the terminal 4 amino acids from WT resulted
in nonfunctional protein(5) . Our observations with the
unprocessed, full-length Rab5 molecule support the idea that the latter
effect can be entirely attributed to the lack of post-translational
geranylgeranylation of the 2 cysteine residues contained within this
domain of the molecule.
for Rab5 Q79L despite a 12-fold reduction in
the single-step GTPase rate constant (Fig. 7). However, by
analogy with other well-studied GTPases(2) , each step of the
Rab5 GTPase cycle is likely to be regulated by interactive proteins in vivo, and a more likely explanation for the apparently
discrepant results among various Gln Leu mutants might lie in
differences in their interactions with endogenous target proteins and
GAPs (GTPase activating proteins). The transforming potential of H-Ras
Q61L is primarily due to an alteration in the protein's ability
to interact with GAPs (p120-GAP and neurofibromin) such that the mutant
is predominantly found in the GTP-bound state(55) . The
corresponding point mutants of several Rab proteins result in defective
intrinsic GTPase activities, yet the Rabs are still capable of
interacting with GAPs such that the hydrolysis of GTP is not impaired in vivo relative to WT. Thus, even though Rab3A Q81L displays
a defective GTPase activity, it remains sensitive to the action of
Rab3A GAP to stimulate hydrolysis, the ratio of GDP/GTP bound in
vivo is the same as for WT, and its action in cells is similar to
that of WT (24, 56) . Likewise, Sec4 Q79L displays
defective GTPase activity, yet hydrolysis can be stimulated by a
GAP(7) . These examples suggest that Rab5 Q79L might
effectively interact with a target protein and act as WT in the fusion
process. This is also consistent with our proteolysis data, which
suggest that the guanine nucleotide-dependent conformations of Rab5
Q79L are similar to those of Rab5 WT. It remains to be determined in
future studies exactly how Rab5's function in endocytosis relies
on interactions with GAPs and other proteins and precisely what role
the cycling between GTP- and GDP-bound forms of Rab5 plays in endosome
fusion.
)S, guanosine
5`-O-(3-thiotriphosphate); GAP, GTPase activating protein.
We thank Dr. Ted Wensel and Dr. Lutz Birnbaumer for
helpful discussions.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Zhu, G. Zhu, J. Liu, Z. Liang, X. C. Zhang, and G. Li |
