Dynamin and Rab5a-dependent Trafficking and Signaling of the Neurokinin 1 Receptor*

Understanding the molecular mechanisms of agonist-induced trafficking of G-protein-coupled receptors is important because of the essential role of trafficking in signal transduction. We examined the role of the GTPases dynamin 1 and Rab5a in substance P (SP)-induced trafficking and signaling of the neurokinin 1 receptor (NK1R), an important mediator of pain, depression, and inflammation, by studying transfected cells and enteric neurons that naturally express the NK1R. In unstimulated cells, the NK1R colocalized with dynamin at the plasma membrane, and Rab5a was detected in endosomes. SP induced translocation of the receptor into endosomes containing Rab5a immediately beneath the plasma membrane and then in a perinuclear location. Expression of the dominant negative mutants dynamin 1 K44E and Rab5aS34N inhibited endocytosis of SP by 45 and 32%, respectively. Dynamin K44E caused membrane retention of the NK1R, whereas Rab5aS34N also impeded the translocation of the receptor from superficially located to perinuclear endosomes. Both dynamin K44E and Rab5aS34N strongly inhibited resensitization of SP-induced Ca2+ mobilization by 60 and 85%, respectively, but had no effect on NK1R desensitization. Dynamin K44E but not Rab5aS34N markedly reduced SP-induced phosphorylation of extracellular signal regulated kinases 1 and 2. Thus, dynamin mediates the formation of endosomes containing the NK1R, and Rab5a mediates both endosomal formation and their translocation from a superficial to a perinuclear location. Dynamin and Rab5a-dependent trafficking is essential for NK1R resensitization but is not necessary for desensitization of signaling. Dynamin-dependent but not Rab5a-dependent trafficking is required for coupling of the NK1R to the mitogen-activated protein kinase cascade. These processes may regulate the nociceptive, depressive, and proinflammatory effects of SP.

ferrin receptor and is required for endosomal translocation to a perinuclear region (41)(42)(43). However, very little is known about the role of Rab5a in trafficking of GPCRs other than recent observations that Rab5a participates in the formation of endosomes containing the dopamine D2 and ␤ 2 -adrenergic (␤ 2 -AR) receptors (38,44) and contributes to down-regulation of the -opioid receptor (24). Nothing is known about the role of dynamin and Rab5a in trafficking of the NK1R, and the relative roles of dynamin and Rab5a in desensitization, resensitization and mitogenic signaling of GPCRs are unknown.
Our aims were to (a) determine the role of dynamin and Rab5a in SP-induced endocytosis and intracellular trafficking of the NK1R by expressing dominant negative mutants of these GTPases; (b) compare the importance of dynamin and Rab5adependent trafficking for desensitization and resensitization of signal transduction; (c) define the role of this trafficking for mitogenic signaling; and (d) establish whether dynamin and Rab5a could contribute to trafficking of the NK1R in neurons that naturally express this receptor.

EXPERIMENTAL PROCEDURES
Reagents and Antibodies-Enhanced green fluorescent protein (EGFP) expression vector pEGFP-C1, expression vector pIRES2-EGFP, and JM109 bacteria were from CLONTECH (Palo Alto, CA). Restriction enzymes, T4 ligase, and Lipofectin were from Life Technologies, Inc. or New England Biolabs, Inc. (Beverly, MA). Exp Tag DNA polymerase was from Stratagene (La Jolla, CA). A QiaEx extraction kit was from Qiagen (Hilden, Germany). G418 was from Gemini Bio-Products, Inc. (Calabashes, CA). Kanamycin and protease inhibitor mixture were from Calbiochem. Glutathione-agarose was from Amersham Pharmacia Biotech. The ECL detection kit was from Amersham Pharmacia Biotech. The Alexa594 protein labeling kit was from Molecular Probes, Inc.

FIG. 2. Characterization of cells expressing FLAGNK1R plus Rab5a or
Rab5aS34N. KNRK cells were stably transfected with FLAGNK1R plus Rab5a-GFP or Rab5aS34N-GFP. A, analysis by flow cytometry confirmed coexpression of NK1R (detected with FLAG antibody) and Rab5a or Rab5aS34N (detected with EGFP). B, Western blotting using Rab5a or EGFP antibodies revealed high expression of Rab5a-GFP and Rab5aS34N-GFP chimeras of the predicted mass. C, confocal microscopy showed localization of NK1R (detected with FLAG antibody) at the cell surface (arrowheads), Rab5a-GFP in endosomes (arrows), and Rab5aS34N-GFP throughout the cytosol (arrows). Scale bar, 10 m.
Generation of Rab5a-GFP and HA-Dynamin Constructs-Canine Rab5a and the GTPase-defective binding mutant Rab5aS34N were fused to EGFP at the N terminus by PCR using the forward primer 5Ј-GGCCGGAATTCCATGGCTAATCGAGGAGCAACA-3Ј (EcoRI site underlined followed by translation initiation site) and the reverse primer 5Ј-GGCCGGGGATCCTTAGTTACTACAACACTGACTCCT-3Ј (BamHI restriction site underlined followed by stop codon). Rab5a cDNA was used as template to amplify an 800-bp fragment that was separated on an agarose gel and purified using a QiaEx extraction kit. The PCR fragment and the vector pEGFP-C1 were digested with EcoRI and BamHI and ligated with the T4 ligase overnight at 16°C. Chemically competent JM109 bacteria were transformed using the heat shock method and selected on LB medium containing 30 g/ml of Kanamycin. Correct sequences were verified. Rab5a tagged with EGFP is fully functional (42). Generation and use of pcDNA3.1-HA-dynamin I and HA-dynamin I K44E (N-terminal HA.11 epitopes), a dominant negative mutant that lacks GTPase activity, have been described (34). HAdynamin-pIRES-EGFP constructs were generated by enzymatic digestion of pcDNA3.1-HA-dynamin with XhoI and XbaI and insertion of DNA into pIRES2-EGFP. The pIRES-EGFP construct enabled convenient identification of transfected cells using EGFP while avoiding attachment of EGFP to dynamin, which impairs its function (49).
Generation of Transfected Cells-Kirsten murine sarcoma virustransformed rat kidney epithelial cells (KNRK) were from the American Type Tissue Culture Collection (Manassas, VA). Generation of KNRK cells stably expressing rat NK1R with N-terminal FLAG epitope (KNRK-FLAGNK1R cells), which does not affect receptor function, has been described (50). To generate a cell line stably expressing FLAGNK1R plus Rab5a-GFP or Rab5aS34N-GFP, KNRK-FLAGNK1R cells were transfected with cDNA encoding Rab5a-GFP or Rab5aS34N-GFP (11). Clones were screened by fluorescence microscopy and flow cytometry to detect NK1R and Rab5a. Cells were also transiently transfected with Rab5a and dynamin, since stable expression of dominant negative dynamin caused cell death, and to enable simultaneous comparison of transfected and untransfected cells by microscopy. KNRK-FLAGNK1R cells were transiently transfected by overnight incubation with 5 g/ml cDNA encoding Rab5a-GFP, Rab5aS34N-GFP, HAdynamin, or HA-dynamin K44E by lipofection. The medium was replaced, and cells were studied 24 h later. To transiently express HA-dynamin-pIRES-EGFP or HA-dynamin K44E-pIRES-EGFP, KNRK-FLAGNK1R cells were transfected by electroporation (Bio-Rad) and plated for 18 h, and positive transfected cells were sorted and plated 24 h before experiments. Cell lines are designated KNRK-NK1RϩRab5a, KNRK-NK1RϩRab5aS34N, KNRK-NK1RϩDYN, or KNRK-NK1RϩDYNK44E (where DYN represents dynamin). Cells were prepared for experiments as described (11,27) and were usually incubated with 5 mM sodium butyrate for 18 h before use to boost expression of transfected genes. As a control, KNRK-FLAGNK1R cells were transfected with vectors without the Rab5a or dynamin inserts. Expression of empty vectors or EGFP alone did not affect signaling or trafficking of the NK1R (see Ref. 11; data not shown) or related receptors (23).
Flow Cytometry-Flow cytometry was used to monitor expression of NK1R, Rab5a, and dynamin and to enrich populations of transfected cells (11,23). To detect the NK1R, cells were resuspended in 1 ml of Iscove's medium containing 1 mg/ml bovine serum albumin, with 10 g/ml M2 antibody to the extracellular FLAG for 1 h at 4°C, washed, and incubated with 2 g/ml phycoerythrin-conjugated goat anti-mouse IgG for 1 h at 4°C. Expression of other constructs was determined using EGFP. Cells were analyzed using a Facscan flow cytometer (Becton Dickinson Co., Franklin Lakes, NJ). Fluorophores were excited at 488 nm, and emission was collected at 530/30 nm for EGFP and 575/25 nm for phycoerythrin.
Endocytosis of 125 I-SP-The rate of NK1R endocytosis was quantified with 125 I-SP (27). Cells were incubated in Hanks' balanced salt solution containing 50 pM 125 I-SP, 0.1% bovine serum albumin for 60 min at 4°C. They were washed and incubated at 37°C for 0 -10 min. Cells were washed with ice-cold PBS and incubated in 250 l of ice-cold 0.2 M acetic acid containing 500 mM NaCl (pH 2.5) on ice for 5 min to separate acid-labile (cell surface) from acid-resistant (internalized) label. Nonspecific binding was measured in the presence of 1 M SP and was subtracted to give specific binding. Observations were in triplicate in n Ͼ 3 experiments. KNRK cells expressing empty vector without NK1R insert do not bind or take up 125 I-SP (50).

Measurement of [Ca 2ϩ ] i -[Ca 2ϩ
] i was measured in populations of transfected cells using Fura-2/AM (11,27). Fluorescence was measured at 340-and 380-nm excitation and 510-nm emission, and the results were expressed as the ratio of the fluorescence at the two excitation wavelengths, which is proportional to the [Ca 2ϩ ] i . Cells were exposed once to SP to generate concentration-response curves. For desensitization experiments, cells were exposed to SP or vehicle (control) for 2 min, washed, and exposed again to SP 5 min after the first exposure. To examine resensitization, cells were incubated with SP or vehicle (control) for 10 min, washed, and challenged with SP 0 -180 min after washing. All observations were in n Ͼ 3 experiments. KNRK cells expressing empty vector without the NK1R insert do not mobilize [Ca 2ϩ ] i in response to SP (50).
stripped and reprobed with antibody to total ERK1/2 to ensure that equal levels of ERK1/2 were present at each time point. The intensities of the combined ERK1 and ERK2 bands for phosphorylated and total proteins were determined by histogram analysis (mean density ϫ number pixels) (15,16). Intensities of the pERK bands were normalized to total ERKs to correct for any differences in loading. Results are expressed as the -fold increase in pERK1/2 relative to unstimulated cells.
All observations were in n ϭ 4 experiments.
NK1R Trafficking in Enteric Neurons-Organotypic cultures were prepared from the rat ileum (Harlan Sprague-Dawley, male, 200 -250 g, Harlan Laboratories, San Diego, CA) as previously described for guinea pigs (48). Excised segments of distal were opened along the mesentery, pinned flat, and incubated in Krebs solution containing 100 nM SP for 1 h at 4°C. Specimens were either immediately fixed or incubated in SP-free medium for 20 min at 37°C (48). Specimens were fixed in 4% paraformaldehyde for 2 h and washed, and whole mounts of the longitudinal muscle with attached myenteric plexus were prepared. Tissues were processed for sequential immunofluorescence using a C-terminally directed rabbit antibody to NK1R (1:500, 48 h, 4°C) and a monoclonal dynamin antibody (1:50, 48 h, 4°C), followed by secondary antibodies labeled with contrasting fluorophores. The sequential double FIG. 4. Localization of Alexa594-SP and wild-type dynamin. KNRK-FLAGNK1R cells were transiently transfected with HA-dynamin. Cells were incubated with Alexa594-SP for 60 min at 4°C, washed, and incubated at 37°C for 0 or 10 min. Dynamin was localized by immunofluorescence using the HA.11 antibody. The same cells are shown in each row, and the images in the right panels are superimpositions of the images in the same row. After 0 min at 37°C, Al-exa594-SP was at the cell surface (arrowheads), and dynamin was in the cytoplasm and at the plasma membrane (arrowheads). After 10 min, Alexa594-SP was detected in perinuclear endosomes (arrows), and dynamin remained at the plasma membrane (arrowheads). Scale bar, 10 m. Results shown are representative of three experiments.

FIG. 5. Localization of Alexa594-SP and dynamin K44E. KNRK-FLAGNK1R
cells were transiently transfected with HAdynamin K44E. Cells were incubated with Alexa594-SP for 60 min at 4°C, washed, and incubated at 37°C for 0 or 10 min. Dynamin K44E was localized by immunofluorescence using the HA.11 antibody. The same cells are shown in each row, and the images in the right panels are superimpositions of the images in the same row. After 0 min at 37°C, Alexa594-SP was at the cell surface (arrowheads), and dynamin was in the cytoplasm and at the plasma membrane (arrowheads). After 10 min at 37°C, in cells expressing dynamin K44E, Alexa594-SP was found at the cell surface (arrowheads). In contrast, Alexa-SP was detected in perinuclear endosomes in untransfected (asterisks, arrows). Dynamin K44E remained at the cell surface (arrowheads). Scale bar, 10 m. Results shown are representative of three experiments. KNRK-FLAGNK1R cells were transiently transfected with Rab5aS34N-GFP. Cells were incubated with Alexa594-SP for 60 min at 4°C, washed, and incubated at 37°C for 0 -10 min. The same cells are shown in each row, and the images in the right panels are superimpositions of the images in the same row. After 0 min at 37°C, Alexa594-SP was at the cell surface (arrowheads). In cells expressing Rab5aS34N-GFP, Alexa594-SP remained at the cell surface at 2.5 min (arrowheads). At 5 and 10 min, Al-exa594-SP was found at the cell surface (arrowheads) and in superficial endosomes (yellow arrows) that did not proceed to a perinuclear location. In contrast, in untransfected cells, Alexa594-SP rapidly internalized into superficial and then perinuclear endosomes (asterisks, white arrows). Scale bar, 10 m. Results shown are representative of three experiments.
labeling procedure was chosen because preliminary experiments showed that mixing the NK1R and dynamin antibodies did not give clear staining of individual antigens. Myenteric neurons from the small intestine of newborn male guinea pigs (Duncan-Hartley; Simonsen, Gilroy, CA) were dispersed and cultured exactly as described (30,31). Neurons were studied at days 7-14 of culture. Neurons were incubated with 100 nM Alexa594-SP or Alexa594-[Sar 9 ,MetO 2 11 ]SP for 120 min at 4°C, washed at 4°C, and either fixed immediately or incubated in SP-free medium at 37°C for 5-30 min (30,31). Neurons were fixed with 4% paraformaldehyde for 20 min at 4°C. NK1R was detected using Alexa-SP, and Rab5a was localized by immunofluorescence with a polyclonal antibody (1:50, 2 h, 4°C) (30,31).
Microscopy-Specimens were observed using a Zeiss Axiovert 100 microscope (Carl Zeiss Inc., Thornwood, NY) with an MRC 1000 confocal microscope (Bio-Rad) or a Zeiss 410 confocal microscope (30,48). Images were collected at 0.5-m intervals using a ϫ 100 PlanApo 1.4 NA objective and a zoom of 1-2 and were processed using Adobe Photoshop (Adobe Systems, Mountain View, CA).
Statistical Analysis-Results are expressed as means Ϯ S.E. Differences between values were analyzed by one-way analysis of variance and the Student-Newman-Keul's test, with p Ͻ 0.05 considered to be significant.

RESULTS
Characterization of Cell Lines Expressing NK1R, Dynamin, and Rab5a-We selected KNRK cells, since they regulate NK1R signaling and trafficking similarly to neurons that naturally express the NK1R (11,18,27,30,31). We transiently transfected KNRK-FLAGNK1R cells with HA-dynamin-pIRES-GFP or HAdynamin K44E-pIRES-GFP and sorted EGFP-expressing cells (Fig. 1A). Western blotting identified an intensely immunoreactive band corresponding to dynamin in cells expressing dynamin and dynamin K44E, which was far more intense than that observed in nontransfected cells (Fig. 1B). NK1R was detected at the cell surface in close proximity to dynamin and dynamin K44E, which were also detected in the cytosol (Fig. 1C). We stably expressed the FLAGNK1R in a hygromycin-resistant vector and Rab5a-GFP or Rab5aS34N-GFP in a neomycin-resistant vector. We used KNRK-NK1RϩRab5a clone 10 and KNRK-NK1RϩRab5aS34N clone 3 in all experiments. Flow cytometry revealed a single population of cells highly expressing NK1R plus Rab5a or Rab5aS34N ( Fig. 2A). Western blotting with a Rab5a antibody detected endogenous Rab5a (ϳ30 kDa) in all cell lines and also detected Rab5a-GFP (ϳ55 kDa) corresponding to Rab5a (ϳ30 kDa) plus EGFP (ϳ27 kDa) in transfected cells (Fig. 2B, left  panel). The EGFP antibody detected only a band at ϳ55 kDa cells expressing Rab5a-GFP or Rab5aS34N-GFP (Fig. 2B, right panel). Immunoreactive NK1R was localized to the plasma membrane, Rab5a was detected in superficial and perinuclear vesicles, and Rab5aS34N was mostly present in the cytosol (Fig. 2C). To verify that cells expressed functional NK1R, we measured SP-induced Ca 2ϩ mobilization. In all cells, SP stimulated a prompt increase in [Ca 2ϩ ] i (not shown, but see Fig. 8) with EC 50 values similar to that measured in cells expressing NK1R alone (ϳ0.6 nM).
Dynamin Mediates Translocation of the Alexa-SP from the Cell Surface to Superficial Early Endosomes-To identify the steps of endosome formation and trafficking that are mediated by dynamin, we localized the NK1R using Alexa594-SP with dynamin or dynamin K44E by immunofluorescence. Cells were incubated with Alexa594-SP for 60 min at 4°C, washed, and incubated at 37°C for 0 -10 min. In cells expressing wild-type dynamin (KNRK-NK1RϩDYN) after 60 min at 4°C, dynamin was detected at the plasma membrane and in the cytosol, and Alexa594-SP and thus the NK1R were at the cell surface (Fig.  4, arrowheads). After 10 min at 37°C, Alexa594-SP was detected in superficial and then perinuclear endosomes and dynamin remained at the plasma membrane (Fig. 4). In cells expressing dominant negative dynamin (KNRK-NK1RϩDYNK44E) at 4°C, dynamin K44E was detected in the cytosol at the plasma membrane, and Alexa594-SP was at the cell surface (Fig. 5). However, after 10 min at 37°C, Al-exa594-SP was retained at the cell surface in cells expressing dynamin K44E (Fig. 5, arrowheads). This result is in marked contrast to that observed in cells expressing wild-type dynamin (Fig. 4) or to untransfected cells (Fig. 5, asterisk). Retention at the cell surface was even observed after 30 min (not shown). Thus, dynamin mediates formation of early endosomes containing the NK1R.

Rab5a Mediates Translocation of the Alexa-SP from the Cell Surface to Superficial and Perinuclear Early Endosomes-We
similarly localized the NK1R and Rab5a. In cells expressing wild-type Rab5a (KNRK-NK1R-Rab5a) after 60 min at 4°C, Rab5a-GFP was detected in multiple prominent vesicles and weakly at the plasma membrane, and Alexa594-SP was confined to the cell surface (Fig. 6). After 2.5 min at 37°C, Al-exa594-SP was localized to superficial vesicles, which also contained Rab5a-GFP (Fig. 6, arrows). After 5 and 10 min, Alexa594-SP was also detected in perinuclear vesicles containing Rab5a-GFP. We have previously reported that vesicles containing SP and the NK1R also contain the transferrin receptor and are thus early endosomes (11,18). Similarly, Al-exa594-SP underwent rapid endocytosis in cells not expressing Rab5a-GFP (Fig. 6, asterisk). In cells expressing dominant negative Rab5a (KNRK-NK1R-Rab5aS34N) at 4°C, Rab5aS34N-GFP was in the cytosol and Alexa594-SP was confined to the cell surface (Fig. 7). After 2.5 min at 37°C, Al-exa594-SP was retained at the cell surface in KNRK-NK1R-Rab5aS34N cells (Fig. 7, arrowheads). In marked contrast, in KNRK-NK1R-Rab5a cells (Fig. 6) and in untransfected cells (Fig. 7, asterisk) Alexa594-SP was detected in superficial endosomes at 2.5 min. After 5 and 10 min in KNRK-NK1R-Rab5aS34N cells, Alexa594-SP was present at the cell surface (Fig. 7, arrowheads) and in superficial endosomes that did not proceed to a perinuclear location (Fig. 7, yellow arrows). In contrast, in KNRK-NK1R-Rab5a cells (Fig. 6) and untransfected cells (Fig. 7, asterisk) Alexa594-SP was found in perinuclear endosomes at 5 and 10 min. Thus, Rab5a mediates formation of early endosomes and their translocation to a perinuclear location.
Dynamin and Rab5a-mediated Endocytosis of the NK1R Are Not Required for Desensitization of SP-induced Ca 2ϩ Mobilization-Exposure of KNRK-NK1R cells co-expressing dynamin or dynamin K44E (Fig. 8A) and Rab5a or Rab5aS34N (Fig. 8B) to 10 nM SP for 2 min caused a prompt and transient increase in [Ca 2ϩ ] i , confirming expression of functional receptors. When cells were washed and exposed again to 10 nM SP 5 min after the first challenge, the response was strongly desensitized in all cell lines (Fig. 8, A and B). We have previously reported a similar desensitization in cells expressing NK1R alone and shown that desensitization is not due to depletion of intracellular Ca 2ϩ stores (27,28). Desensitization occurred at a time when the NK1R was internalized in cells expressing wild-type dynamin and Rab5a but was at the cell surface in cells expressing dominant negative dynamin and Rab5a. Thus, dynaminand Rab5a-mediated endocytosis are not required for desensitization of the NK1R.
Dynamin-and Rab5a-mediated Endocytosis of the NK1R Are Required for Resensitization of SP-induced Ca 2ϩ Mobiliza- tion-To examine resensitization, cells were incubated with 10 nM SP or vehicle for 10 min, washed, and challenged with 10 nM SP at 0 -180 min after washing. In cells expressing FLAGNK1R alone, exposure to 10 nM SP for 10 min strongly desensitized SP-induced Ca 2ϩ mobilization, which gradually resensitized when the interval between SP challenges increased beyond 10 min, such that complete resensitization occurred after 180 min (Fig. 9A). A similar degree of desensitization was observed in cells expressing wild-type or dominant negative dynamin or Rab5a (Fig. 9B). Resensitization was almost complete after 180 min in cells expressing wild-type dynamin (91 Ϯ 23% resensitization) or Rab5a (79 Ϯ 34% resensitization) (Fig. 9B). In marked contrast, after 180 min in cells expressing dynamin K44E or Rab5aS34N, responses to a second challenge with SP were resensitized by only 36 Ϯ 12% and 12 Ϯ 5%, respectively (p Ͻ 0.05) (Fig. 9B). Compared with cells expressing wild-type dynamin and Rab5a, dynamin K44E inhibited resensitization by 60%, and dominant negative Rab5aS34N inhibited resensitization by 85%. Thus, dynaminand Rab5a-mediated trafficking of the NK1R are required for resensitization of responses to SP.
Dynamin but Not Rab5a-mediated Endocytosis Is Required for Mitogenic Signaling of the NK1R-To determine the role of dynamin and Rab5a in SP activation of the MAP kinase cascade, we compared SP-induced ERK1/2 activation in cell lines expressing wild-type and dominant negative dynamin or Rab5a. Cells were incubated with SP for 0 -30 min, and activation was determined by Western blotting using antibodies to pERK1/2 (activated) and total ERK1/2. Under basal conditions, there was a low level of phosphorylation of ERK1/2 in all cell lines. Densitometric analysis of the pERK1/2 and total ERK1/2 indicated that the basal level of activation was similar in all cells (pERK1/2 versus total ERK1/2 ratios: KNRK-NK1RϩDYN, 0.6 Ϯ 0.1; KNRK-NK1RϩDYNK44EK, 0.5 Ϯ 0.1; NRK-NK1RϩRab5a, 0.6 Ϯ 0.1; KNRK-NK1RϩRab5aS34N, 0.5 Ϯ 0.1). In cells expressing wild-type dynamin, SP stimulated a prompt increase in ERK1/2 phosphorylation that was maximal at 5 min (2.4-fold over basal) and remained elevated in the continued presence of SP after 30 min (Fig. 10A). We have previously reported similar results in KNRK cells expressing the NK1R alone and endothelial cells that naturally express the NK1R (16). Expression of EGFP alone did not alter this response (not shown). Expression of dynamin K44E strongly inhibited SP-induced phosphorylation of ERK1/2 (1.1fold over basal at 5 min, p Ͻ 0.05). In cells expressing wild-type Rab5a, SP stimulated a rapid phosphorylation of ERK1/2 that was maximal after 2.5 min (2-fold over basal) and remained elevated for 30 min (Fig. 10B). Expression of Rab5aS34N had no effect on SP-induced phosphorylation of ERK1/2 (2-fold over basal at 2.5 min). Thus, dynamin-but not Rab5a-mediated trafficking of the NK1R is required for mitogenic signaling.
Alexa594-SP Colocalizes with Rab5a in Endosomes in Enteric Neurons Expressing the NK1R-To confirm a potential role of dynamin and Rab5a in nontransfected cells, we studied myenteric neurons that naturally express the NK1R (30,31,46). We used rat neurons for studies of dynamin and guinea pig neurons for studies of Rab5a due to limitations in antibody specificity. Organotypic cultures of rat ileum were incubated with SP for 60 min at 4°C, washed, and incubated for 0 -20 min at 37°C. The NK1R and dynamin were localized by immunofluorescence. At 4°C, the NK1R was detected at the cell surface in the soma of a subpopulation of neurons (Fig. 11). Dynamin was detected in NK1R-positive neurons and in most other neurons, where it was predominantly cytosolic and also in close proximity to the plasma membrane. After 20 min at 37°C, the NK1R was detected in endosomes, whereas the distribution of dynamin was unchanged. Cultures of guinea pig neurons were incubated with Alexa594-SP or Alexa594-[Sar 9 ,MetO 2 11 ]SP (similar results were obtained with both peptides) for 120 min at 4°C, washed, and incubated for 0 -30 at 37°C. Rab5a was localized by immunofluorescence. After 0 -2 min at 37°C, Al-exa594-SP and, thus, the NK1R were detected at the cell surface, and Rab5a was in endosomes (Fig. 12). After 5-10 min at 37°C, Alexa594-SP was colocalized with Rab5a in endosomes in the soma and neurites. We have previously shown that these vesicles contain the transferrin receptor and are thus early endosomes (30). After 30 min, Alexa594-SP was present in perinuclear endosomes that did not contain Rab5a (not shown). Thus, myenteric neurons expressing the NK1R also express dynamin and Rab5a, and these proteins are appropriately localized within the cell to regulate trafficking of the NK1R. FIG. 10. SP-induced activation of ERK1/2. Cells expressing the NK1R with wild-type or dominant negative dynamin (A) or Rab5a (B) were incubated with 10 nM SP for the indicated times. Cells were analyzed by Western blotting using antibodies to pERK1/2 or total ERK1/2, for standardization. The upper panels show representative Western blots, and lower panels show the -fold increase over basal phosphorylation. *, p Ͻ 0.05 compared with wild-type cells, n ϭ 4 experiments.

DISCUSSION
We found that expression of dominant mutants of dynamin 1 and Rab5a retarded SP-induced endocytosis of the NK1R. Dominant negative Rab5a also impeded the intracellular trafficking of endosomes containing the NK1R from a superficial to a perinuclear region. These mutants strongly inhibited resensitization of SP-induced Ca 2ϩ mobilization but did not affect desensitization. Whereas dominant negative dynamin almost abolished SP-stimulated activation of ERK1/2, dominant negative Rab5a had no effect. Thus, our results indicate that dynamin and Rab5a are required for the formation of endosomes containing the NK1R and that Rab5a mediates their translocation to a perinuclear region. Dynamin and Rab5a-de-pendent trafficking of the NK1R is essential for resensitization of responses to SP, and dynamin-mediated endosome formation is required for coupling of the NK1R to the MAP kinase pathway. Moreover, dynamin and Rab5a are appropriately localized in neurons expressing the NK1R to suggest that they participate in the physiological regulation of NK1R trafficking and signaling. To our knowledge, this study is the first to demonstrate a role for dynamin and Rab5a in trafficking and signaling of the NK1R and the first report of the relative roles of dynamin and Rab5a in mitogenic signaling of a GPCR.
Molecular Mechanisms of SP-induced Endocytosis and Intracellular Trafficking of the NK1R-Dynamin was detected at or near the plasma membrane under basal conditions, and dy- FIG. 11. Localization of NK1R and dynamin in rat myenteric neurons. Neurons were incubated with SP for 60 min at 4°C, washed, and incubated at 37°C for 0 or 20 min. NK1R and dynamin were detected by immunofluorescence. The same neurons are shown in each row, and the images in the right panels are superimpositions of the images in the same row. At 0 min, the NK1R was at the cell surface, and dynamin was in the cytosol close to the plasma membrane. After 20 min, NK1R was detected in endosomes (arrows), and the distribution of dynamin was unchanged. Scale bar, 10 m. Neurons were incubated with Alexa594-SP for 120 min at 4°C, washed, and incubated at 37°C for 2 or 10 min. Rab5a was detected by immunofluorescence using a fluorescein isothiocyanateconjugated secondary antibody. The same neurons are shown in each row, and the images in the right panels are superimpositions of the images in the same row. At 2 min, the NK1R was at the cell surface (arrowheads), and Alexa-SP was in endosomes. After 10 min, Alexa594-SP was in endosomes containing Rab5a (arrows). Scale bar, 5 m. namin K44E inhibited endocytosis of 125 I-SP by ϳ50% at early time points (5 min) and caused retention of most Alexa-SP at the cell surface for prolonged periods (up to 30 min). The larger effect of dynamin K44E on endocytosis of Alexa-SP than on endocytosis of 125 I-SP, suggests that dynamin K44E permits 125 I-SP to be sequestered into a compartment at or near the plasma membrane, where it is resistant to removal by the acid wash that was used to separate cell surface from internalized peptide. However, the consensus of our results is that dynamin plays an important role in SP-induced endocytosis of the NK1R, and our findings support the hypothesis that dynamin mediates the final stages of endosome formation from clathrincoated pits (32,33). The observation that SP stimulates endocytosis of the NK1R by a clathrin-mediated mechanism in KNRK cells and enteric neurons supports a role for dynamin in endosome formation (18,30). In support of our results, dynamin mediates endocytosis of receptors that constitutively internalize, such as the transferrin receptor (34), as well as GPCRs that internalize in response to agonist binding, including the ␤ 2 -AR; ␦-opioid; muscarinic m1, m3, and m4; and dopamine D1 receptors (12,35,(37)(38)(39). However, dynamin is not required for endocytosis of angiotensin II type 1A, muscarinic m2, dopamine D2, and ␣ 2B -adrenergic receptors (12,37,39,40). The mechanisms of dynamin-independent endocytosis and the receptor domains that specify dynamin dependence remain to be determined.
Rab5a prominently colocalized with the internalized NK1R in superficial and perinuclear endosomes and is thus appropriately localized within cells to regulate NK1R trafficking. Rab5aS34N inhibited endocytosis of 125 I-SP by ϳ35% and retarded endocytosis of Alexa-SP, supporting a role for Rab5a SP-induced endocytosis of the NK1R. In support of our results, Rab5a mediates the formation of endosomes containing the transferrin receptor (41)(42)(43), dopamine D2 receptor (38), and the ␤ 2 -AR (44). Rab5aS34N also caused retention of the NK1R in superficial endosomes, indicating a requirement of Rab5a for translocation of endosomes from a superficial to a perinuclear region. Dominant negative Rab5a also impedes the kinetics of membrane trafficking in the endocytic pathway (41) and similarly causes retention of the ␤ 2 -AR in endosomes close to the plasma membrane (44).
Although expression of dynamin K44E and Rab5aS34N markedly retarded SP endocytosis of the NK1R, they did not abolish this trafficking. Explanations include the possibility that the mutants were not expressed at high enough levels to completely inhibit the endogenous proteins or that additional mechanisms also contribute to trafficking. ␤-Arrestins couple certain GPCRs to clathrin, including ␤ 2 -AR, m2 muscarinic receptor, protease-activated receptor 2, and the NK1R (9 -11, 23, 51). To interact with ␤-arrestins, receptors must undergo phosphorylation, and G-protein receptor kinases 2 and 3 phosphorylate the NK1R in reconstituted systems and in cell lines (52)(53)(54)(55). Dynamin acts downstream from ␤-arrestins, by participating in endosome formation at clathrin-coated pits (32,33). Rab5a is likely to participate in both the formation of these endosomes and their translocation to a perinuclear location (56,57).
Role of SP-induced Endocytosis and Intracellular Trafficking of the NK1R for Signal Transduction-Agonist-induced trafficking of GPCRs and associated proteins regulates cellular responses (1,2). SP induces translocation of ␤-arrestins to the plasma membrane, where they interact with the NK1R to mediate desensitization and endocytosis (11,31). Endocytosis could contribute to desensitization by depleting the cell surface of receptors. However, dynamin K44E and Rab5aS34N, which inhibited NK1R endocytosis, did not affect desensitization of SP-induced Ca 2ϩ mobilization, indicating that NK1R endocytosis is not the main mechanism of desensitization. In support of these findings, inhibition of NK1R endocytosis by dominant negative ␤-arrestin-(319 -418) or with drugs does not prevent desensitization (11,20,27). Thus, uncoupling of the NK1R from G-proteins is the main mechanism of desensitization. In a similar manner, uncoupling is the principal mechanism of desensitization of the ␤ 2 -AR (1). However, down-regulation of the -opioid receptor following prolonged activation requires endocytosis by dynamin, ␤-arrestin and Rab5a-dependent mechanisms (24).
Dynamin K44E and Rab5aS34N markedly inhibited resensitization of SP-induced Ca 2ϩ mobilization by ϳ60 -85%. These results indicate that endocytosis and trafficking of the NK1R is necessary for recovery of responses to SP. In support of these findings, pharmacological inhibition of endocytosis and recycling and the NK1R and ␤ 2 -AR disrupts their resensitization (19,20), and Rab5aS34N also inhibits resensitization of ␤ 2 -AR (44). Thus, resensitization of GPCRs requires endocytosis, dissociation from the receptor from ligand in acidified endosomes, receptor dephosphorylation and dissociation from ␤-arrestins, and recycling.
Although agonist-stimulated endocytosis of some GPCRs is necessary for activation of the MAP kinase cascade, little is known about the relative contributions of ␤-arrestin, dynamin and Rab5a to this process. We found that dynamin K44E, but not Rab5aS34N, abolished SP-induced activation of ERK1/2. We have previously reported that dominant negative ␤-arrestin-(319 -418) similarly inhibits this process (16). In support of our results, ␤-arrestinand dynamin-dependent endocytosis is also required for coupling the ␤ 2 -AR, ␦-opioid receptor, m1 muscarinic receptor, and proteinase-activated receptor-2 to the MAP kinase pathway (14,15,36,58,59). ␤-Arrestins serve as scaffolds that recruit and organize components of the MAP kinase cascade including Src, in the case of the NK1R and ␤ 2 -AR (14,16), and Raf-1 in the case of the proteinase-activated receptor-2, where this interaction also determines the location and specificity of activated ERK1/2 (15). We found that inhibition of the proximal steps of endocytosis, by expression of dominant negative mutants of ␤-arrestin and dynamin, inhibited SP-induced ERK1/2 activation, whereas disruption of Rab5a, which acts more distally, was without effect. Thus, for the same GPCR, inhibition of distinct steps of endocytosis and trafficking has different effects on coupling to the MAP kinase cascade. However, expression of dominant negative mutants of ␤-arrestin and dynamin does not affect coupling of the CXCR2, -opioid, or ␣ 2 -adrenergic receptors to ERK1/2 (60 -62). Dynamin may have independent roles in GPCR internalization and MAP kinase activation, since dominant negative dynamin inhibits phorbol ester-stimulated MAP kinase activation, which is independent of receptor endocytosis and transactivation (63).
Physiological Relevance of NK1R Trafficking-An understanding of the molecular mechanisms that regulate signaling by SP in the nervous system is of great interest, since the NK1R plays a major role in pain, depression, neurogenic inflammation, and gastrointestinal motility and secretion (25,26). We found that myenteric neurons expressing the NK1R also express dynamin and Rab5a and that dynamin and Rab5a are appropriately located to regulate endocytosis and trafficking of the NK1R in the soma. SP induces membrane translocation of G-protein receptor kinases 2 and 3 and ␤-arrestin 1/2 in these neurons, and disruption of the formation of clathrincoated pits inhibits receptor endocytosis (30,31). Thus, it is likely that clathrin, ␤-arrestins, dynamin, and Rab5a mediate SP-induced endocytosis of the NK1R in neurons, as they do in transfected KNRK cells. Defects in these regulatory mechanisms may result in unregulated signaling and disease. Deletion of neutral endopeptidase EC 3.4.24.11, which degrades SP at the cell surface and thereby terminates signaling, results in exaggerated inflammatory effects of SP (64,65). Alterations in the expression of other regulatory proteins, such as ␤-arrestins, dynamin and Rab5a, may also result in uncontrolled signaling and disease.