(Received for publication, January 31, 1997, and in revised form, March 12, 1997)
From the Washington University School of Medicine, Department of Cell Biology & Physiology, St. Louis, Missouri 63110
Rab7 has been shown to localize to late endosomes and to mediate transport from early to late endosome/lysosome in mammalian cells and in yeast. We developed a novel assay to quantify transport from early to late endosomes using the Xenopus oocyte. Oocytes were pulsed with avidin after which the oocytes were incubated to allow avidin transport to a late compartment. The oocytes were then allowed to internalize biotin-horseradish peroxidase (HRP). The oocytes were then injected with test proteins and incubated further to allow transport of biotin-HRP from early endosomes to late endosomal/lysosomal compartments. Transport was quantified by assessing the formation of HRP-biotin-avidin complexes. Injection of Rab7:wild-type (WT) and Rab7:Q67L, a GTPase defective mutant, stimulated transport. Rab5:WT had no effect. Rab7:WT-stimulated transport was inhibited by nocodazole, suggesting a role for intact microtubules. Wortmannin, a phosphatidylinositol 3-kinase inhibitor, blocked Rab7:WT-stimulated transport, but Rab7:Q67L-stimulated transport was unaffected by the drug. Rab7:Q67L is constitutively activated and may not require phosphatidylinositol 3-kinase activity for activation. Rab7-stimulated transport requires N-ethylmaleimide-sensitive factor (NSF) activity as transport was blocked by N-ethylmaleimide and ATPase defective NSF mutants. Our results indicate that sequentially acting endocytic Rab GTPases utilize similar factors although their modes of action may be different.
Endocytosis occurs in all eukaryotic cells and is essential for the uptake of nutrients, the turnover of membranes proteins (1) and, in some cells such as oocytes, the formation of yolk platelets (2). After internalization, endocytosed solute and receptor-bound proteins are delivered to a common early endosomal compartment where they are sorted and targeted to other intracellular destinations, including a putative recycling compartment, a trans Golgi network pathway and diversion to a lysosomal degradative/storage compartment (3). Currently, there is only a rudimentary understanding of the mechanisms that mediate endosomal sorting and transport. Nevertheless, much progress has been made toward a general understanding of vesicular transport. Vesicles containing cargo bud from donor compartments and deliver their contents to acceptor compartments by fusion with the latter. The processes of vesicle budding, docking, and fusion are regulated by a series of generic and compartment-specific proteins (4, 5). Sorting and transport at the level of endosomes may involve budding or tubule formation, or both.
Ras-related Rab GTPases regulate intracellular trafficking during endocytosis and secretion (6-10). Rab proteins are specifically localized to the cytoplasmic surface of specific intracellular compartments where they regulate the individual steps of vesicular transport by controlling vesicle docking and fusion (11-13). A number of Rab proteins have been localized to the early endocytic pathway including Rab4, Rab5, Rab11, Rab18, and Rab20, suggesting that the endocytic compartment is highly complex with multiple functions (6-9, 11, 14-16). Morphologically, the endosomal compartment is thought to be comprised of at least two functionally distinct sub-compartments (3, 17). Internalized receptor and ligand first enter the peripheral sorting endosome where membrane proteins destined for degradation or the trans Golgi network are sorted away from membrane proteins targeted for recycling back to plasma membrane. A substantial body of evidence indicates that Rab5 regulates transport from the plasma membrane to the early endosome (6-10). Rab4 appears to control the recycling from early endosomes to plasma membrane (14), and recent studies indicate that Rab11 regulates recycling through the perinuclear endosomes (16). Rab7 and Rab9 are associated with the late endosomal compartment (11, 18-22). Rab7 and Rab9 have been shown to regulate the traffic from early endosomes to late endosomes (20, 21) or lysosomes (23) and from the late endosome to trans-Golgi network (TGN) (18), respectively. Deletion of the YPT7 gene, the yeast homologue of Rab7, leads to fragmentation of the vacuole and to a delay in the processing of vacuolar proteins (24). Ypt7 has been shown to regulate transport between late endosome and vacuole (25). Moreover, Ypt7 is also involved in the homotypic fusion between vacuoles in yeast (26). Among the endocytic Rabs, Rab5 has been most extensively studied. Rab5 is active in the GTP form and does not require GTP hydrolysis for mediating docking and fusion (9, 27). Rab5 is regulated by multiple factors including phosphatidylinositol (PI)-3-kinase1 (28) and other signal transducing elements including RAS (29). Rab5 also requires NSF for function (30). However, the exact relationship among all these factors in the overall mechanism of intracellular trafficking during endocytosis is largely unknown.
Recently, we have shown that both Rab5 and Rab7 stimulate HRP uptake in oocytes and that Rab7 functions downstream of Rab5 (30). Moreover, Rab7 stimulates the inactivation/degradation of preloaded HRP in oocytes by enhancing transport to a late endocytic compartment. To further characterize Rab7-mediated transport, we developed a ligand mixing assay to measure transport from early to late endosome/lysosome compartments. The results demonstrate that Rab7 stimulates transport to the late compartment and that Rab7-stimulated transport is NSF-, microtubule-, and PI-3-kinase-dependent.
Recombinant NSF wild type and mutant proteins were generously provided by Sidney W. Whiteheart (University of Kentucky, Lexington, KY). Adult female Xenopus laevis were purchased from Nasco, Fort Atkinson, WI. Avidin and biotin-HRP were obtained from Pierce (Rockford, IL). All other reagents were from Sigma (St. Louis, MO).
Expression and Purification of Recombinant Rabs and MutantsThe fusion proteins used in this study, including Rab5:Q79L, Rab5:S34N, Rab7:WT, Rab7:Q67L, and Rab7:T22N, were described previously (8, 9). Recombinant proteins were expressed as glutathione S-transferase fusion proteins in Escherichia coli strain JM 101. GST fusion proteins were affinity purified with glutathione-Sepharose.
Horseradish Peroxidase Uptake in Microinjected Xenopus OocytesOvaries were dissected from adult female X. laevis anesthetized with 3-aminobenzoic acid ethyl ester (1 g/liter) in ice water, and oocytes were prepared as described (30). Oocytes were injected with test molecules in 50 nl of PBS and then allowed to recover in modified Barth's saline at 18 °C for 2 h. Healthy oocytes were then selected for HRP uptake. The injected oocytes were incubated with HRP (2 mg/ml) for 2 h at 18 °C. HRP uptake was stopped by washing the cells 3 times with modified Barth's saline containing 5% BSA. Individual oocytes were then lysed in 200 µl of PBS containing 0.1% Triton X-100 and centrifuged. The lysate was assayed for HRP in 96-well plates (Costar Corp.) using o-phenylenediamine as the chromogenic substrate (31). The products were quantified at A490 nm in a Bio-Rad microplate reader. Results were expressed as nanograms of HRP uptake per oocyte.
Assay for the Transport to Late CompartmentTo monitor the Rab7-mediated transport to the late endosome/prelysosome compartment, oocytes were incubated with avidin (2 mg/ml) for 5 h at 18 °C. The oocytes were then washed and incubated for 16 h at 18 °C to label the late endocytic compartment. Subsequently, cells were washed and incubated with biotin-HRP (1 mg/ml) at 18 °C for 30 min to mark the early endocytic compartment. Indicated amounts of test proteins were then injected into the oocytes in 50 nl of PBS, and the oocytes were further incubated at 18 °C. The oocytes were then washed, and three oocytes were lysed in 100 µl of PBS containing 0.3% Triton X-100, 0.2% methylbenzethorium chloride, and 400 µg/ml of BSA-biotin as scavenger. Finally, the lysates were centrifuged for 30 s in a microfuge, and the avidin-biotin complexes were immunoprecipitated using anti-avidin antibody-coated plates by incubation at 4 °C overnight. The plates were then washed at least 5 times to remove the unbound proteins. Transport from the early endosomes to the late compartments was quantified by measuring the amount of immunoprecipitated HRP. Background was determined by the amount of HRP immunoprecipitated in the presence of scavenger at the 0 time point, which was found to be very low. The total amount of possible ligand mixing was measured in a similar way but in the absence of scavenger.
By analogy with RAS, dominant negative and GTPase defective mutants of Rab5 and Rab7 have been prepared and characterized by several laboratories (6, 8, 12, 14, 27, 32). For microinjection experiments with Xenopus oocytes, we generated the GTPase defective mutant, Rab7:Q67L, and the dominant negative mutant, Rab7:T22N, in addition to the wild-type protein as described. All the proteins were prepared as glutathione S-transferase fusion proteins, and purified proteins were injected directly into oocytes.
Rab7 Stimulates Transport from Early to Late Endosomes in Xenopus OocytesRecently, we have shown that Rab7 regulates endocytosis
in oocytes in concert with Rab5 (30). Microinjection of Rab7:WT, Rab7:Q67L, and Rab5:WT stimulated HRP uptake into oocytes (30) (data
not shown). Moreover, co-injection of Rab5 and Rab7 produced an
additive effect on HRP uptake. Rab7-mediated HRP uptake was inhibited
by Rab5:S34N. Rab7:T22N inhibited Rab7-stimulated uptake but had little
effect on Rab5-stimulated HRP uptake. These results demonstrate that
Rab7 plays a crucial role in the early endocytic process in the oocyte
acting downstream of Rab5 (30). In the present study, we developed an
assay to detect and quantify transport from the early to late endocytic
compartments using ligand mixing as a biochemical readout. Oocytes were
preloaded with avidin (2 mg/ml for 5 h) after which they were
incubated for 16 h to label the late endocytic compartment,
including lysosomes. Subsequently, cells were pulsed with biotin-HRP
for a short period of time (30 min) to mark the early endocytic
compartment. The cells were then injected with different Rab
preparations, and following incubation for various periods of time, the
formation of avidin-biotin-HRP complexes was determined in cell
lysates. Oocyte lysis and preparation of avidin-biotin-HRP immune
complexes was routinely carried out in the presence of biotin-BSA to
scavenge unoccupied avidin molecules. The results presented in Fig.
1 show that Rab7:WT stimulates transport of biotin-HRP
to the late compartment prelabeled with avidin. Rab7-stimulated
transport was maximal at about 90 min. The maximal efficiency of
Rab7-mediated transport efficiency was calculated to be about 80%
(total signal was determined in the absence of the scavenger
biotin-BSA). In oocytes injected with control buffer, transport was
linear over 90 min, and the efficiency of transport was about 20% at
90 min. All subsequent transport assays were carried out for 60 min at
18 °C.
Rab7 is localized to late endosomes where it appears to function in early to late endosomal transport (20, 21). It has been shown in Hela cells that Rab7:WT does not colocalize with markers of the early endocytic compartment but colocalizes with MPR, a marker for the late endocytic compartment (20, 21, 23). Moreover, Rab7:Q67L was found to co-localize with LAMP1 and LAMP2 as well as the lysosomal enzyme cathepsin D (23). Using subcellular fractionation, Rab7- and Rab4-like GTPases were found to be localized on the lysosomal membrane in Dictyostelium discoideum (33). Thus, Rab7 is commonly found on terminal endocytic compartments.
The results presented in the Fig. 2A show
that both Rab7:WT and Rab7:Q67L stimulate early to late endosomal
transport. In contrast, Rab7:T22N, the dominant negative mutant, had no
effect. Furthermore, Rab5, which stimulates uptake and is associated
with the early endocytic compartment, did not affect transport (Fig. 2B). In fact, Rab5:Q79L and Rab5:S34N were found to be
slightly inhibitory (Fig. 2B). GTP hydrolysis by Rab5 is not
required for entrance into the early endosomal compartment, but it is
possible that GTP hydrolysis and guanine nucleotide cycling on Rab5 is required for exit from the early compartment. A current model suggests
that Rab recycling is coupled to guanine nucleotide exchange (34, 35)
and that GDP dissociation inhibitor delivers cytosolic Rabs to
appropriate donor compartments in a reaction that is coupled to the
exchange of GDP for GTP. The GTP bound protein assists in the
recruitment of appropriate fusion factors and GTP hydrolysis occurs
after vesicle docking and fusion with the acceptor membrane. The GDP
form of the protein is then recycled back to the donor compartment by
GDP dissociation inhibitor. According to this model, Rab7:Q67L, which
has reduced GTPase activity, should stimulate transport to the late
compartment (20, 21, 23). Indeed, both Rab7:WT and Rab7:Q67L stimulated
endocytic transport. As indicated above, Rab7:T22N, the dominant
negative mutant was not active and did not substantially inhibit
transport. Rab7:T22N is "functional" since we demonstrated earlier
that this dominant negative mutant blocks endocytosis induced by Rab7
injection into oocytes (30). It is possible that the
Rab7-dependent pathway is relatively inactive in oocytes.
Injection of Rab7 would then activate the pathway. Under these
conditions, one would not expect the dominant negative mutant of Rab7
to have much effect on normal transport. Alternatively, it is possible
that injected recombinant mammalian Rab7 competes poorly with the
endogenous equivalent frog Rab7 for accessory factors necessary for
function.
Rab7-mediated Transport Depends on Microtubules
It is known that endocytosed materials are transported from the peripheral to a perinuclear location as they progress from early to late endosomes and that this process depends on intact microtubules (36). When microtubules are depolymerized, the markers failed to reach the late endosomal compartment. Thus, if the Rab7 mediates transport from the early to the late compartments then the process should be sensitive to microtubule depolymerizing agents like nocodazole. When oocytes were preincubated with 50 µM nocodazole, Rab7-stimulated uptake was inhibited almost to the control levels (data not shown). In contrast, Rab5-mediated uptake was much less affected by the exposure of the cells to nocodazole (data not shown). These data confirm earlier observations that transport from the plasma membrane to early endosome is not dependent on microtubules.
To determine whether Rab7-mediated transport is dependent on
microtubules, we tested the effect of 50 µM nocodazole in
the transport assay. The results presented in Fig. 3
show that Rab7-mediated transport from the early to late compartment
was totally blocked by nocodazole, indicating that Rab7stimulated
early to late compartment transport in oocytes is
microtubule-dependent. The complete inhibition of transport
to essentially basal levels suggests that other microtubule-sensitive pathways may exist in the oocyte.
Rab7 Mediated Early to Late Transport Dependent on PI-3 Kinase
Recent work indicates that PI-3-kinases and their lipid products may play key roles in vesicular transport. The first evidence for the involvement of PI-3-kinase came from the studies of VPS34p, a yeast protein essential for protein targeting to the yeast vacuole and for vacuole morphogenesis (37). VPS34p shows sequence homology with the mammalian PI-3-kinase and exhibits PI-3-kinase activity (38). Several studies in mammalian cells also indicate that inhibitors of PI-3-kinase block fluid phase endocytosis as well as phagocytosis (28, 39, 40). Wortmannin, a fungal metabolite, is a potent PI-3-kinase inhibitor which binds covalently to the catalytic subunit of PI-3-kinase (p110) and irreversibly inhibits enzymatic activity at nanomolar concentrations (41). The inhibition is highly specific since at these low concentrations (50-100 nM), wortmannin has no effect on other kinases. We first examined the effect of wortmannin on the endocytosis of HRP by oocytes following microinjection of Rab5:WT, Rab7:WT, and Rab7:Q67L. Rab5:WT, Rab7:WT, and Rab7:Q67L all stimulated HRP uptake by 2-2.5-fold. Rab5:WT and Rab7:WT stimulation of HRP uptake was nearly completely blocked by wortmannin treatment. These results are consistent with previous findings in baby hamster kidney cells transiently expressing Rab5:WT and Rab5:Q79L (28, 29). Wortmannin inhibited Rab5-enhanced uptake, but the drug had no effect on enhanced endocytosis produced by Rab5:Q79L (28). Wild-type Rab proteins require an activation step (e.g. guanine nucleotide exchange) to become GTP bound and biologically active. Constitutively activated mutants presumably bypass the activation step. Recent work2 indicates that both wortmannin and recombinant PI-3-kinase affect the nucleotide status of Rab5, the former favoring the GDP-bound form and the latter favoring the GTP-bound form.
The data presented in the Fig. 3 show that transport from early endosomes to late endosomes requires PI-3-kinase activity as wortmannin was found to inhibit this transport. We have also found that wortmannin inhibits the transport from the early to late compartments in mammalian cells.3 Moreover, Rab7:Q67L, the constitutively activated mutant, does not require PI-3-kinase activity, suggesting that PI-3-kinase activates Rab7:WT either directly or through some intermediate to stimulate transport. In sum, the present data with oocytes along with earlier data suggest that PI-3-kinase activity is required for the activation of Rab5:WT and Rab7:WT.
Role of NSF in Rab7 Mediated Early to Late TransportThe NEM-sensitive fusion protein was originally identified as a factor required to restore the in vitro transport activity of Golgi membrane fractions treated with NEM (42). NSF is required for intra-Golgi transport, endoplasmic reticulum-Golgi transport, as well as fusion between the early endosomes suggesting that NSF is a general component of the fusion machinery (43-45). NSF is a heterotrimer whose polypeptide subunits are made up of three distinct domains, an amino-terminal domain and two homologous ATP (D1 and D2) binding domains (46). The ability of the D1 domain to hydrolyze ATP is required for NSF-mediated membrane fusion. The D2 domain is required for trimerization, but its ability to hydrolyze ATP is not absolutely required for NSF function. Two mutations in the first ATP binding site (D1), i.e. D1-K266A and D1-E329Q, are known to affect ATP binding and hydrolysis, respectively, resulting in inactive forms of NSF (46).
Given that NSF is an ubiquitous protein required for multiple vesicular
transport events, we tested the effect of NSF in the early to late
transport assay using anti-NSF antibodies and NSF mutants. To confirm
the participation of NSF in Rab7-mediated HRP endocytosis, we
co-injected oocytes with Rab7:WT and His6NSF wild-type
protein or Rab7:WT with the NSF mutants. Our results show that
Rab7:WT-stimulated endocytosis HRP is not affected by the presence of
NSF:WT protein but that His6 D1-E329Q and His6 D1-K266A significantly
inhibit HRP uptake mediated by Rab7:WT (data not shown). We then
explored the role of NSF in Rab7-stimulated transport from the early to
late endosomal compartments. As shown in Fig. 4, both
the DIEQ and DIKA mutants of NSF significantly inhibited the transport
of biotin HRP from early to late endosomal compartments. Rab7
stimulation of the HRP uptake as well as transport inhibition by both
NSF mutants indicates that ATP binding and hydrolysis are required for
both processes. (Moreover, NEM was found to inhibit early to
late transport in mammalian cells (data not shown).) These
results are also consistent with the finding that YPT7, which is
involved in the early to late transport events in yeast, also requires
Sec17p (yeast -snap) and Sec18p (yeast NSF) for function (47,
48).
In summary, our results demonstrate that early to late endosomal transport is stimulated by Rab7 and that this transport requires intact microtubules. Rab7:WT stimulation is blocked by wortmannin, suggesting that PI-3-kinase functions as a regulator of transport. Rab7:Q67L-stimulated transport is not blocked by wortmannin, suggesting that PI-3-kinase is directly or indirectly associated with nucleotide exchange as suggested for Rab5 (28). Furthermore, our results show that ATP binding and hydrolysis by NSF is also required in Rab7-stimulated transport. These findings also support the notion that NSF is a generic factor required for multiple transport events. It is well documented that Rab5 is required for transport between plasma membrane and early endosomes and that it regulates the dynamics of early endosome fusion (7, 8, 9, 27, 36). Overexpression of GTP bound mutant, Rab5:Q79L, in cultured cells increases the rate of endocytosis and leads to the appearance of very large endocytic vesicles that may be due to fusion between early endosomes. However, transport out of the large endosome is delayed,4 suggesting that hydrolysis of GTP by Rab5 is required for subsequent transport events. Rab5 and Rab7 appear to act in sequence along the endocytic pathway toward lysosomes. A key question that must be framed and resolved is the influence of a proximal Rab on the activation of an immediately distal Rab in the regulation of complex intracellular pathways.
We thank Rita Boshans for technical assistance and Marisa Colombo and Alejandro Barbieri for advice.