Identification of potential tyrosine-containing endocytic motifs in the carboxyl-tail and seventh transmembrane domain of the neurokinin 1 receptor.

Although agonist-induced endocytosis of G-protein-coupled receptors is critical for receptor desensitization and resensitization, receptor motifs that interact with the endocytic apparatus have not been adequately characterized. We examined the effects of mutating the rat neurokinin-1 receptor on endocytosis using 125I-substance P, fluorescent substance P, and receptor antibodies. Substance P induced rapid internalization of wild-type receptors that were targeted to perinuclear endosomes. Truncation of the C-tail at residues 324, 342, and 354 reduced internalization up to 60% and caused retention of receptors at the cell surface and in superficial endosomes. Mutation of Tyr-341 and Tyr-349 in potential tyrosine-containing endocytic motifs of the C-tail also impaired internalization. A Y305A mutant within the putative NPX2-3Y endocytic motif of the seventh transmembrane domain showed impaired signaling and was minimally expressed at the plasma membrane but was found in cytoplasmic vesicles. In contrast, a Y305F mutant signaled normally and was normally expressed at the plasma membrane but showed impaired internalization. Thus, endocytosis of the neurokinin 1 receptor relies on several tyrosine-containing sequences in the C-tail and seventh transmembrane domain.

Receptor endocytosis is a fundamental process with many diverse functions. Endocytosis of some receptors, such as the transferrin receptor, is required for the uptake of nutrients by many cell types. These receptors internalize constitutively, without the need to bind ligand. In contrast, agonist binding is required for internalization of receptors for many hormones, neurotransmitters, and growth factors (1). Endocytosis of these receptors contributes to receptor desensitization, resensitization, and down-regulation, critical processes that regulate cellular responses.
The main pathway for constitutive and agonist-induced endocytosis is by clathrin-coated pits, although alternative pathways involve non-coated vesicles or caveolae (2). Clathrin-mediated endocytosis probably requires interaction of specific receptor motifs with components of the endocytic machinery (3). Tyrosine-containing motifs of six residues forming a tight ␤-turn are common, and interchangeable endocytic signals are found in the intracellular domains of constitutively internalized and tyrosine kinase receptors (1,4). In constitutively internalized receptors, they may be continuously exposed to the endocytic machinery, whereas agonist-induced internalization of the signaling receptors may require receptor modification to expose a cryptic endocytic motif. Thus, a point mutation that inactivates the tyrosine kinase activity of the epidermal growth factor receptor abolishes agonist-induced internalization, probably by preventing exposure of an endocytic motif (5). In contrast, deletion of the entire tyrosine kinase domain abolishes the endocytic switch and results in constitutive endocytosis, which is probably due to continuous exposure of endocytic motifs.
Although regions in the C-tail and TMD 1 VII have been implicated as important for endocytosis of some G-proteincoupled receptors (6 -12), a common endocytic motif has not been identified, and the nature of the molecular switch for agonist-induced internalization is unknown. Agonist-induced internalization may require activation of signaling mechanisms or receptor phosphorylation. Arguing against the first possibility are observations with receptor mutants that either signal normally and show impaired internalization or exhibit impaired coupling to G-proteins and normal internalization (13,14). This suggests that different domains are involved in internalization and G-protein coupling. Agonist binding induces phosphorylation of Ser and Thr residues of many Gprotein-coupled receptors, and Ser/Thr-rich regions are critical for internalization (6, 8 -10, 15). However, phosphorylation of the ␤ 2 -AR is not required for internalization (16 -18).
Recently, there has been considerable interest in the mechanism and function of endocytosis of the NK1R, a G-proteincoupled receptor for the neuropeptide SP (19 -22). SP is a neurotransmitter in the central and peripheral nervous systems, stimulates smooth muscle contraction and exocrine gland secretion, and induces neurogenic inflammation, actions that are mediated in part by the NK1R (23). SP stimulates clathrinmediated endocytosis and recycling of the NK1R in multiple cell types (19 -22). It is important to identify endocytic motifs of the NK1R since receptor endocytosis and recycling determine the ability of a cell to respond to SP (24). Although much is known about the extracellular and transmembrane domains of the NK1R that interact with receptor agonists and antagonists (25,26), there is little information about NK1R domains that are necessary for endocytosis. Thus, the aim of this investigation was to identify motifs for SP-induced endocytosis of the NK1R. Based on our knowledge of endocytic motifs of single TMD receptors and on the limited information on G-proteincoupled receptors, we mutated the rat NK1R and investigated the effects of the mutations on agonist-induced endocytosis. We identified Tyr-containing motifs in the C-tail and TMD VII of the NK1R that are important for endocytosis and which resemble motifs of the single TMD receptors.

EXPERIMENTAL PROCEDURES
Reagents-SP was labeled with Bis-functional cyanine 3.18 as described (27). A rabbit polyclonal antiserum (number 11884-5) was raised to a 15-residue peptide (K 393 TMTESSSFYSNMLA 407 ), corresponding to the intracellular C terminus of the rat NK1R (28). A mouse monoclonal antibody (M2) to the Flag peptide (DYKDDDDK) was from International Biotechnologies, Inc. (New Haven, CT). A mouse monoclonal antibody (X22) to the heavy chain of clathrin was a gift from Dr. Francis Brodsky (University of California, San Francisco).
Mutagenesis of Rat NK1R-Mutants were generated by PCR using the rat NK1R cDNA with an N-terminal Flag-peptide in pcDNAI/neo as a template (28). The Flag epitope allowed detection of NK1R with the Flag antibody as well as the antibody to the C terminus of the receptor. The Flag does not alter the signaling or trafficking of the NK1R (19,28). Two overlapping fragments of the coding region were amplified using SP1Flag (5Ј-GATCGAAGCTTGCCACCATGGACCAAGG-3Ј, HindIIIsite underlined) as 5Ј primer and NheI673B (5Ј-GGATCTCGCTAGC-CCACAGTG-3Ј) as 3Ј primer in one reaction, and NheI673A (5Ј-CAC-TGTGGGCTAGCGAGATCC-3Ј) and SP2 (5Ј-CATAAGCGGCCGCCTA-GGCCAGCATGTTAG-3Ј, NotI-site underlined) in the other reaction. Primers NheI673A and B introduced an NheI-site (underlined) at base 673 of the coding region by silent mutagenesis (mutated residues in bold). The reaction included 0.1 g of template DNA, 0.1 M of each primer, 50 M each dNTP, 10 ϫ cloned Pfu buffer and 2.5 units Pfu polymerase in a volume of 100 l. The PCR conditions were 1 min at 94°C, 2 min at 55°C, and 3 min at 72°C for 30 cycles. Both reactions were combined and the products purified by ultrafiltration. The purified fragments were used as templates in a PCR reaction with primers SP1Flag and SP2. The product of this reaction was cut with HindIII/ NotI and cloned into pcDNA3. This construct was used as NK1R-wt. All other mutations were introduced by the aforementioned strategy using primers to the Sp6 and T7 regions in pcDNA3 as flanking primers and other primers to introduce the mutations. Truncated receptors were generated by amplifying rat NK1R with SP1Flag and the primers ␦311, ␦324, ␦342, and ␦354 containing a stop codon and a NotI restriction site. The fidelity of PCR amplification was verified by sequencing all constructs in both directions. All mutated receptors were cloned into Hin-dIII/NotI sites of pcDNA3.
Expression Vectors and Cell Lines-We stably expressed receptors in KNRK cells, which do not express endogenous NK1R, using Lipofectin (28). Clones were selected in medium containing 800 g/ml G418 and screened by 125 I-SP and cy3-SP binding or NK1R immunofluorescence. Clones were maintained in medium containing 400 g/ml G418. Cells were plated 24 -48 h before experiments on poly-D-lysine-coated glass coverslips (for microscopy or measurement of Ca 2ϩ mobilization) or on 35-mm plastic wells (for 125 I-SP binding experiments, ϳ10 5 cells plated per well).
Binding and Endocytosis of 125 I-SP-Cells were incubated in Hank's balanced salt solution, 0.1% bovine serum albumin, and 50 pM 125 I-SP for 60 min at 4°C, washed, and incubated at 37°C for 0 -10 min (19). An acid wash procedure was used to separate cell-surface (acid-sensitive) from internalized (acid-resistant) 125 I-SP, exactly as described (19). The K D and B max were determined by competition binding and Scatchard analysis (28). Cells were incubated with 50 pM 125 I-SP and 30 pM to 100 nM SP for 60 min at 4°C. Nonspecific binding was measured by preincubating the cells with 1 or 2.5 M SP and was subtracted to give specific binding.
Examination of Cy3-SP and NK1R Endocytosis by Microscopy-Cells were incubated in Dulbecco's modified Eagle's medium, 0.1% bovine serum albumin, and 100 nM cy3-SP for 60 min at 4°C, washed, and incubated at 37°C for 0 -30 min (19,21). Cells were fixed with 4% paraformaldehyde in 100 mM phosphate-buffered saline for 20 min at 4°C and mounted. Cy3-SP specifically binds to the NK1R in these cells (27). We also used an NK1R and Flag M2 antibodies to localize NK1R. Cells were incubated with 100 nM SP at 4°C, washed, incubated at 37°C for 0 -30 min and fixed. NK1R and clathrin were localized by indirect immunofluorescence (19,21). Cells were examined with an MRC 1000 Confocal Microscope using a Zeiss plan-Apochromat x100 oil-immersion objective. Images were collected at an aperture of 2-5 mm and a zoom of 2. Typically, 10 -20 optical sections were taken at 0.5-m intervals.
Measurement of Ca 2ϩ Mobilization-Cells were washed with Hank's balanced salt solution, 0.1% bovine serum albumin and incubated with 2.5 M Fura-2/AM for 20 min at 37°C (28). Fluorescence was measured in a fluorimeter at 340-and 380-nm excitation and 510-nm emission. A single injection of SP (30 pM to 0.1 M) was made. Results are expressed as the ratio of the fluorescence at 340 and 380 nm, which is proportional to [Ca 2ϩ ] i .
Statistical Analysis-Results are expressed as mean Ϯ standard error. Differences between multiple groups are examined by analysis of variance and Student-Newman Keuls test and between two groups by Student's t test. A p Ͻ 0.05 is considered statistically significant.

RESULTS
To identify regions of the NK1R that are important for endocytosis, we generated a series of mutants (Fig. 1).
C-terminal Truncations-First, we progressively truncated the C-tail (designating receptors as "␦" followed by number of the last residue). The intracellular C-tail of the rat NK1R extends 99 residues from the plasma membrane (309 -407). Tyr, Ser, and Thr residues, usually 30 -50 residues from the plasma membrane, have been implicated in endocytosis of several G-protein-coupled receptors (8, 29 -32). The C-tail of the NK1R contains three Tyr residues (Tyr-331, Tyr-341, and Tyr-349) that are conserved between species and which may be important for endocytosis. Tyr residues 341 and 349 are surrounded by six Ser and Thr residues. We designed four truncated receptors to remove varying portions of the C-tail ( Fig. 1): ␦354, lacking part of the tail with no conserved Tyr residues; ␦342, lacking Tyr-349 and a portion of the surrounding Ser and Thr residues; ␦324, lacking Tyr-349, Tyr-341, Tyr-331 and the remaining Ser and Thr residues; and ␦311, in which the entire C-tail including the putative palmitoylation site was deleted. NK1R-␦311 corresponds to the naturally truncated, potentially alternatively spliced NK1R variant (33,34). Immunofluorescence with the Flag antibody revealed that NK1R-wt, NK1R-␦354, NK1R-␦342, and NK1R-␦324 were normally located at the plasma membrane (not shown). However, NK1R-␦311 was mostly found intracellularly in vesicles, suggesting mistargeting, and was not further evaluated.
For all mutants, we initially measured the time course of receptor endocytosis using 125 I-SP. Cells were incubated with 125 I-SP for 60 min at 4°C, washed, and incubated at 37°C for 0, 5, or 10 min. They were washed with acid to separate cell surface from internalized label. We used these results to identify mutants with significant alterations in endocytosis. These were analyzed in more detail by measuring internalization of 125 I-SP at 0, 1, 2, 4, or 8 min. The data points from 0, 1, and 2 min were fitted to a linear function; the initial rate of internalization was calculated from the slope and is expressed as % internalization/min. When NK1R-wt cells were incubated with 125 I-SP at 4°C and immediately washed with acid, 13.8 Ϯ 1.7% of the specifically bound counts were in the acid-resistant, internalized fraction. After 5 min at 37°C, 57.2 Ϯ 2.0% of specific counts were internalized. The initial internalization rate was 9.9 Ϯ 0.9%/min (r 2 for linear regression ϭ 0.995) ( Fig. 2A). At 4°C, 8 -20% of specific counts were internalized by NK1R-␦354, -␦342, or -␦324 cells. After 5 min the proportion of specific counts that were internalized was 50.6 Ϯ 0.5% for NK1R-␦354, 21.1 Ϯ 0.8% for NK1R-␦342, and 32.8 Ϯ 3.5% for NK1R-␦324, significantly less than for NK1R-wt (p Ͻ 0.05 compared with NK1R-wt, Table I). These differences were also significant at 10 min. The initial rates of internalization were measured for NK1R-␦342 and NK1R-␦324 cells, since they showed the largest defects in endocytosis. The initial rat of internalization was 2.4 Ϯ 0.5%/min (r 2 ϭ 0.999) for NK1R-␦342 cells and 3.0 Ϯ 0.6%/min (r 2 ϭ 0.971) for NK1R-␦324 cells, significantly slower than for NK1R-wt cells (p Ͻ 0.05 compared with NK1R-wt, Fig. 2A).
We used cy3-SP, immunofluorescence, and confocal microscopy to determine the effects of the mutations on the subcellular distribution of NK1R. We have shown that cy3-SP and NK1R are co-localized in the same endosomes up to 30 min after internalization in KNRK cells expressing NK1R-wt (19 -21). Thus, cy3-SP can be used to localize the NK1R at these times. When NK1R-wt cells were incubated with cy3-SP at 4°C and immediately fixed, cy3-SP was confined to the plasma membrane (Fig. 3). After 2 min at 37°C, there was minimal detectable surface labeling, and cy3-SP was found in superficial endosomes. After 5 min, cy3-SP was mostly present in perinuclear endosomes (Fig. 3), and the distribution was similar at 10 and 30 min. In NK1R-␦354, -␦342, and -␦324 cells at 4°C, cy3-SP was confined to the plasma membrane (Fig. 3). In NK1R-␦354 cells, a substantial amount of cy3-SP remained at the plasma membrane after warming, and endosomes containing cy3-SP remained in a superficial location and did not proceed to a perinuclear region (Fig. 3). The largest defect in internalization was observed in NK1R-␦342 cells, in which cy3-SP was retained at the plasma membrane and was rarely detected in endosomes even after 30 min (Fig. 3). In NK1R-␦324 cells, cy3-SP was also detected at the plasma membrane at all time points and there was minimal internalization (Fig.  3).
We assessed the function of the NK1R-wt and the mutated receptors by measuring SP-induced changes in [Ca 2ϩ ] i and determining the EC 50 and maximal response. In NK1R-wt cells, SP induced a prompt increase in [Ca 2ϩ ] i with an EC 50 of 0.55 nM (Table I). SP also increased [Ca 2ϩ ] i in NK1R-␦354, -␦342, and -␦324 cells, but SP was more potent, and the doseresponse curves were shifted to the left between 3-and almost 7-fold ( Fig. 4A, Table I). Maximal responses were similar for NK1R-wt and the truncated receptors. Those receptors with the largest defects in internalization were further characterized by determining the K D and B max of the receptors. The K D for 125 I-SP binding was 7.4 Ϯ 1.6 nM and the B max was 58.4 Ϯ 4.9 fmol/10 5 cells for NK1R-wt (Table I). The K D and B max were similar in NK1R-␦342 and -␦324 cells. Thus, although we do not know the reason for increased sensitivity of truncated receptors to SP, their impaired ability to internalize cannot be explained by defective signaling, and defects in ligand binding cannot explain the impaired endocytosis of these truncated receptors.
Tyrosine Mutations in the C-tail-Tyr-containing endocytic motifs, in which the Tyr is of critical importance, have been

FIG. 2.
Internalization of 125 I-SP by wild-type and mutated NK1R. Cells were incubated with 125 I-SP for 60 min at 4°C for equilibrium binding. They were either acid-washed immediately or were incubated at 37°C for 1, 2, 4, or 8 min and then washed with acid to separate acid-labile (cell surface) from acid-resistant (internalized) label. Results are expressed as the percentage of the specifically bound counts in the internalized fractions and are the mean Ϯ standard error of triplicate observations from n ϭ 3 experiments. The inset bar graphs show the rate of internalization calculated from the initial 2 min. *, Ͻ0.05 compared with NK1R-wt cells.
At 4°C, cy3-SP was confined to the plasma membrane in all mutants involving Tyr residues in the C-tail (NK1R-Y341A and NK1R-Y349A, Fig. 5; NK1R-Y331A and NK1R-Y341A/Y349A not shown). After 5 min at 37°C, cy3-SP was retained at the plasma membrane and found in superficial and perinuclear endosomes in cells expressing NK1R-Y331A (not shown) and NK1R-Y341A (Fig. 5). In NK1R-Y349A cells, cy3-SP was mostly retained at the plasma membrane and in superficial endosomes after 5 min, although there were a few perinuclear vesicles containing cy3-SP (Fig. 5). This is in contrast to NK1R-wt cells, in which cy3-SP was mostly found in many perinuclear endosomes after 5 min (Fig. 3).
EC 50 values for SP-induced Ca 2ϩ mobilization ranged from 0.037 to 0.156 nM for NK1R-Y331A, -Y341A, and -Y349A cells compared with 0.55 nM for the NK1R-wt (Fig. 4B, Table I). Maximal Ca 2ϩ responses were similar. The K D and B max were similar to NK1R-wt except for the Y341 mutant, where the K D was ϳ4-fold lower. Thus, defects in endocytosis are unlikely to be the result of markedly abnormal binding and signaling.
Tyrosine Mutations in TMD VII-A Tyr-containing motif (NPX 2-3 Y) in TMD VII is highly conserved within G-proteincoupled receptors and may be a common endocytic motif (7). We examined this possibility for the NK1R by mutating Tyr-305 in TMD VII to Phe or Ala. Phe can replace Tyr in many, but not all, proteins without a significant effect on internalization, whereas Ala substitutions usually abolish the signal (1). Mutation of Tyr-305 to Phe significantly reduced the extent of endocytosis of NK1R, as measured by internalization of 125 I-SP. Thus, after 5 min at 37°C, only 40.3 Ϯ 4.7% of specific counts were internalized by NK1R-Y305F cells, which represents a reduction of internalization by 30% compared with NK1R-wt cells (p Ͻ 0.05 compared with NK1R-wt, Table I). Mutation of Tyr-305 to Ala resulted in increased internalization at 4°C, although there was low specific binding compared with the other mutants. After 60 min at 4°C, 29.4 Ϯ 7.1% of specific counts were internalized in NK1R-Y305A cells, compared with 13.8 Ϯ 1.7% in NK1R-wt cells (Table I). However, after 5 min at 37°C, the proportion of internalized 125 I-SP was similar for NK1R-Y305A and NK1R-wt cells. The initial internalization rates were 7.9 Ϯ 1.1%/min (r 2 ϭ 1.000) for NK1R-Y305F cells and 7.1 Ϯ 1.0%/min (r 2 ϭ 0.921) for NK1R-Y305A cells (Fig. 2C). These rates were slower than for NK1R-wt cells but not statistically significant different.
The low surface binding and apparent internalization of 125 I-SP and cy3-SP by NK1R-Y305A cells at 4°C prompted us to directly localize the NK1R by immunofluorescence with an antibody to the C-tail of the receptor. When NK1R-wt and NK1R-Y305F cells were incubated with SP at 4°C and immediately fixed, receptor immunoreactivity was mostly confined To examine internalization, cells were incubated with 125 I-SP for 60 min at 4°C for equilibrium binding, washed, and incubated at 37°C for the indicated times. Cells were washed with acid to separate cell surface from internalized counts. Results are expressed as the percentage of the total specific binding in the internalized fraction and are the mean Ϯ S.E. of triplicate observations, with (n) as the number of experiments.
To examine mobilization of intracellular Ca 2ϩ , cells were loaded with Fura-2/AM, and changes in fluorescence after exposure to a single concentration of SP were measured. Results are from triplicate observations. The EC 50 is expressed in nM and the maximal response is expressed relative to that of NKlR-wt. The K D and B max were only determined for cells with marked defects in the internalization that were selected for further experimentation. The K D is the mean Ϯ standard error of duplicate observations from n ϭ 3 experiments, and the B max is expressed as relative to that of NKlR-wt. ND, not done. to the plasma membrane with very few immunoreactive endosomes (Fig. 7). In sharp contrast, in NK1R-Y305A cells at 4°C, receptor immunoreactivity was mostly found in perinuclear vesicles with minimal surface labeling (Fig. 7). After 10 min at 37°C, receptor immunoreactivity was detected in larger perinuclear vesicles in NK1R-wt cells, whereas immunoreactivity was also retained at the plasma membrane in NK1R-Y305F cells (Fig. 7). The subcellular location of receptor immunoreactivity was similar in NK1R-Y305A cells at 4°C and at 37°C (Fig. 7). In NK1R-Y305F cells, the EC 50 and efficacy of SP-induced Ca 2ϩ mobilization were similar to NK1R-wt cells (Fig. 4C, Table I). However, SP was over 4-fold less potent and 7-fold less efficacious in NK1R-Y305A than in NK1R-wt cells (Fig. 8, Table I). The K D values for NK1R-Y305 and NK1R-Y305A were similar to NK1R-wt, but the B max for NK1R-Y305A cells was markedly lower. Thus, the Y305F mutation does not affect SP-induced Ca 2ϩ mobilization, whereas the Y305A mutation interferes with Ca 2ϩ mobilization.
Serine/Threonine-Mutations and a Mutation of the EMKST Motif in the C-tail-Ser/Thr-rich regions are critical for internalization of several G-protein-coupled receptors (6, 8 -10, 15). Based on our results with the truncated receptors and on studies suggesting that Ser and Thr residues in the vicinity of Tyr residues may be a more common determinant of internalization signals than their possible location in a consensus site for phosphorylation (6, 12), we mutated Ser and Thr residues in the C-tail of the NK1R. We also mutated Lys-337 because it is situated in a EMKST motif, which closely resembles the internalization signal DAKSS of the yeast ␣-pheromone receptor (35). Mutation of the lysine in this motif prevents endocytosis.
Cy3-SP was confined to the plasma membrane at 4°C for all Cells were incubated with cy3-SP at 4°C, and either fixed immediately or washed and incubated at 37°C for 5 min and then fixed. At 4°C, cy3-SP was confined to the plasma membrane in both cell lines. After 5 min at 37°C, cy3-SP was retained at the plasma membrane (see NK1R-Y349A) or superficial endosomes (see NK1R-Y341A), with minimal progression to perinuclear endosomes compared with NK1R-wt (Fig. 3). Scale bar ϭ 10 m. mutants (Fig. 9). Surprisingly, after 5 min at 37°C cells expressing NK1R-⌬ST338-352 demonstrated a halo-like retention of cy3-SP in small peripheral vesicles, seemingly in con-trast to the data obtained by using 125 I-SP. To a lesser extent, this phenomenon was also detectable in cells expressing NK1R-S338G/T339A. Compared with NK1R-wt, the amount of cy3-SP in perinuclear endosomes was not markedly reduced in these mutants. Endocytosis of cy3-SP by cells expressing NK1R-T339A/S347A (Fig. 9) and NK1R-K337A (not shown) was similar to that observed in NK1R-wt cells. Cells expressing NK1R-S338G/T339A, NK1R-⌬ST338-352, and NK1R-K337A responded to SP with EC 50 values for Ca 2ϩ mobilization that were similar to NK1R-wt cells (Fig. 4D, Table I). In cells expressing NK1R-T339A/S347A, SP was about 3-fold less potent for Ca 2ϩ mobilization. The K D for NK1R-⌬ST338-352 was similar to NK1R-wt. Despite the reduced B max of this mutant, the efficacy for Ca 2ϩ mobilization was similar to NK1R-wt.

Effects of Mutations on the Mechanism of Endocytosis-To
FIG. 6. Confocal photomicrographs showing distribution of cy3-SP in NK1R-Y305A and NK1R-Y305F cells. Cells were incubated with cy3-SP at 4°C and either fixed immediately or washed and incubated at 37°C for 5 min and then fixed. In NK1R-Y305A cells at 4°C, there was weak binding of cy3-SP to the plasma membrane (arrow) and detection of some cy3-SP in endosomes (arrowhead), whereas after 5 min at 37°C there was stronger labeling in endosomes. In NK1R-Y305F cells at 4°C, there was strong binding of cy3-SP to the plasma membrane, whereas after 5 min at 37°C cy3-SP was retained at the cell surface and found in superficial vesicles, with minimal progression to perinuclear endosomes compared to NK1R-wt (Fig. 3). Scale bar ϭ 10 m.

FIG. 7. Confocal photomicrographs showing distribution
NK1R immunoreactivity in NK1R-wt, NK1R-Y305A, and NK1R-Y305F cells. Cells were incubated with unlabeled SP at 4°C and either fixed immediately or washed and incubated at 37°C for 10 min and then fixed. The NK1R was localized by immunofluorescence using an antiserum to the C terminus of the receptor. In NK1R-wt cells, NK1R immunoreactivity was confined to the plasma membrane at 4°C and detected in perinuclear endosomes with minimal surface staining after 10 min at 37°C. In NK1R-Y305A cells, NK1R immunoreactivity was predominantly detected in endosomes in a perinuclear location both at 4°C and after 10 min at 37°C. In NK1R-Y305F cells, NK1R immunoreactivity was confined to the plasma membrane at 4°C and detected at the plasma membrane and in endosomes after 10 min at 37°C. Scale bar ϭ 10 m. determine if the mutated receptors with abnormal rates of endocytosis were internalized similarly to NK1R-wt at sites of clathrin (21), we colocalized cy3-SP and clathrin. Cells were incubated with cy3-SP for 60 min at 4°C, washed, incubated at 37°C for 2 or 5 min, and fixed, and clathrin was localized by immunofluorescence. In cells expressing NK1R-wt, most superficial vesicles containing cy3-SP were initially stained with the clathrin antibody, as indicated by superimposition of confocal images, where yellow denotes colocalization (Fig. 10). In contrast, cy3-SP was mostly retained at the cell surface of cells expressing truncated receptors such as NK1R-␦342, and there was no clathrin colocalization (Fig. 10). Although cy3-SP was internalized more slowly in cells expressing NK1R-Y349A, NK1R-Y341A, or NK1R-Y305F, superficial vesicles containing cy3-SP were still stained with the clathrin antibody (Fig. 10). In cells expressing NK1R-Y305A, there were few superficial vesicles detected containing cy3-SP, although these were occasionally stained with the clathrin antibody (Fig. 10). In cells expressing NK1R-⌬ST338-352, cy3-SP was retained in numerous small, very superficial vesicles that were also stained by the clathrin antibody (Fig. 10). Thus, despite the finding that many mutants affected the internalization rate, endocytosis still proceeded at sites of clathrin.

DISCUSSION
The combination of receptor mutagenesis, analyses using radiolabeled and fluorescently labeled SP, and use of antibodies to the NK1R and clathrin permitted detailed investigation of endocytic motifs of the NK1R. The results show that there are several sequences within the C-tail and TMD VII of the rat NK1R that are important for SP-induced endocytosis. We compared SP-induced endocytosis of mutant and wt NK1R stably expressed in KNRK cells. We have previously examined trafficking of the NK1R in KNRK cells in detail and found that it behaves like the NK1R in neurons and endothelial cells (19 -21, 36).
The C-Tail of the NK1R Contains Endocytic Motifs-Truncation of the NK1R at residue 355 had a relatively small effect on internalization, indicating that either residues 355 to 407 are not critical for endocytosis or this region contains positive and negative endocytic signals whose influences are offset when both are removed. Further deletion at residue 342 markedly slowed the rate of endocytosis and trafficking to a perinuclear region, indicating that more proximal regions of the C-tail are critical for internalization. The dramatically reduced rate of internalization of NK1R-␦342 compared with NK1R-␦354 or NK1R-wt indicates that residues 343-354 include an important endocytic domain. Thus, the NK1R resembles the receptors for gastrin-releasing peptide, thyrotropin-stimulating hormone, angiotensin II, and parathyroid hormone, in which truncation of the C-tail also decreases the rate of internalization (8,12,37,38). In contrast, removal of the C-tail of the luteinizing hormone/chorionic gonadotropin receptor and the avian ␤ 1 -or ␤ 2 -AR increases the rate of internalization (29,39,40). Further deletion of residues 325 to 342 (NK1R-␦324) slightly enhanced the degree of endocytosis compared with NK1R-␦342. This may be explained by the presence of a negative endocytic signal between residues 325 and 342. Such a signal (EVQ) is found at the membrane-cytoplasmic interface of TMD VII of the parathyroid hormone receptor (12). Alternatively, incremental shortening of the C-terminal tail may increase the lateral mobility of the receptor in the plasma membrane and, therefore, its probability of becoming trapped in a clathrin-coated pit (41). With the exception of NK1R-␦311, all of the truncated receptors were appropriately expressed at the plasma membrane. This lack of surface expression may explain the poor functional responses of the naturally occurring trun-cated NK1R to SP (33,42). Similar mistargeting occurs for other receptors truncated N-terminal to the putative palmitoylation site or lacking the entire C-tail (12,29).
Conserved Tyrosine Residues Contribute to Endocytic Motifs of the C-tail-Tyr-containing endocytic motifs, in which the Tyr is of critical importance, have been identified for many of the single TMD proteins (1,32). They usually contain six residues forming an exposed ␤-turn. Positions 1, 3, and 6 are frequently occupied by aromatic or large hydrophobic residues, with positions 3 and 6 more important to the function of the motif than position 1. Either position 3 or 6 must contain an aromatic residue, and residues in positions 2, 4, and 5 tend to be polar and are often found in turns.
All of the conserved Tyr residues in the C-tail of the NK1R contribute to endocytosis, since each mutant showed defective internalization, albeit to different degrees. Mutation of Tyr-349 had the largest effect on internalization. This is located between residues 343 and 354, and removal of these residues in NK1R-␦342 resulted in the greatest defect in internalization between the three truncated mutants. Tyr-341 and Tyr-331 are probably part of signals of lesser importance. However, only the residues surrounding Tyr-331 fit the typical consensus sequence for Tyr endocytic motifs, with aromatic or large hydrophobic residues in positions 3 and 6 (GD 331 YEGL). Tyr-349 could be placed in a hexapeptide motif with Tyr-349 in position 1 and Val and Leu in positions 3 and 6 ( 349 YKVSRL). Tyr-341 is not surrounded by any other aromatic or large hydrophobic residues that fit a 1-3-6 pattern. Tyr-containing motifs in the C-tail of other G-protein-coupled receptors, such as the parathyroid hormone receptor and angiotensin II 1a receptor, are also important for internalization (9,12). However, mutation of surrounding residues and the exchange of motifs between Gprotein-coupled and single TMD receptors will be required to determine the extent to which these motifs resemble the Tyrcontaining endocytic motifs of single TMD receptors. In contrast, Tyr residues in the C-tails of the ␤ 2 -AR and m 2 muscarinic receptor are not critical for internalization (43,44).
The NK1R-⌬ST338-352 was internalized marginally more quickly than NK1R-wt in response to SP. Thus, in contrast to some other G-protein-coupled receptors, in which Ser/Thr-rich regions are critical for internalization (6, 8 -10, 15), the ST338-352 region plays only a subtle role in internalization of the NK1R.
The TMD VII NPX 2-3 Y Sequence Is Not a General Endocytic Motif-A Tyr-containing motif (NPX 2-3 Y) in TMD VII is highly conserved within the G-protein-coupled receptors. It closely resembles endocytic motifs of the low density lipoprotein and insulin receptors, both of which fit the six-residue consensus sequence for Tyr-containing endocytic motifs (1,32). It has been suggested that this may be a common endocytic motif for G-protein-coupled receptors (7).
We investigated this possibility by mutating this region of the NK1R. The NK1R-Y305F and NK1R-Y305A mutants behaved very differently. The conservative Y305F mutation did not affect the subcellular distribution of the receptor in the basal state after binding SP at 4°C but caused diminished internalization after incubation at 37°C. The more drastic Y305A mutation markedly altered the subcellular distribution of the receptor so that it was in intracellular vesicles even in the basal state. This could be due to mistargeting of newly synthesized receptors to the plasma membrane. However, several observations indicate that some receptors reach the plasma membrane. First, there was a low level of detectable binding of 125 I-SP and cy3-SP to the cell surface at 4°C. Second, there was internalization of 125 I-SP after warming at a similar rate as for the NK1R-wt. Finally, the prominent intra- cellular pools of cy3-SP observed after incubation at 37°C must have undergone endocytosis, since SP is hydrophilic and can only bind surface receptors. An alternative explanation for the detection of NK1R-Y305A in intracellular vesicles in the basal state is that this receptor internalizes constitutively. In support of this possibility is the elevated internalization of 125 I-SP and cy3-SP at 4°C. Thus, the Y305A mutation may induce a conformational change in the receptor that permits its interaction with the endocytic apparatus at 4°C or even without agonist binding. If this is the case, the NPX 2-3 Y sequence in TMD VII could resemble the switch function of the tyrosine kinase domain for agonist-induced internalization of the epidermal growth factor receptor (5). The tyrosine kinase point mutation and the Y305F mutation may interfere with agonistinduced changes in receptor conformation, thereby preventing exposure of cryptic endocytic motifs. In contrast, deletion of the tyrosine kinase domain and the Y305A mutation may cause the receptors to assume a conformation in which endocytic motifs are constitutively exposed. Further experimentation is required to test this hypothesis.
Mutation of the TMD VII NPX 2-3 Y domain also affected signaling. Although the Y305F mutation did not affect SPinduced Ca 2ϩ mobilization, both potency and efficacy of SP were reduced in NK1R-Y305A cells as compared with NK1R-wt cells. The impaired signaling of NK1R-Y305A may be explained by its low level of expression at the cell surface. Alternatively, the NPX 2-3 Y motif may have induced a conformational change in the receptor that alters its ability to interact with mechanisms of signal transduction.
The role of the NPX 2-3 Y domain in endocytosis and signaling has been examined for other receptors. Mutation of the corresponding Tyr in TMD VII of the ␤ 2 -AR to Ala (Y326A) abolishes internalization, reduces agonist-induced phosphorylation by Gprotein receptor kinase 2, and depresses activation of adenylyl cyclase (7,45,46). Overexpression of this kinase rescues receptor phosphorylation and internalization. Thus, the ␤ 2 -AR-Y326A mutation alters the ability of the agonist-occupied receptor to achieve a conformation required for phosphorylation, signaling, and internalization. The more conservative Y326F mutation of the ␤ 2 -AR reduces internalization by only 25% and also depresses phosphorylation and activation of adenylyl cyclase (45). A tyrosine to alanine mutation of the NPX 2-3 Y sequence in TMD VII of the angiotensin II 1A receptor also reduces internalization by ϳ25% (9,11). In contrast, an equivalent mutation does not affect endocytosis of the receptor for gastrin releasing peptide (47).
These observations and the results of the present study suggest that the NPX 2-3 Y sequence of TMD VII is not an endocytic motif but may be important for maintaining the appropriate conformation of the receptor for endocytosis or signaling to occur. Therefore, Tyr-305 of the NK1R may be critical for transducing agonist binding at the outer face of the receptor into a conformational change in the C-tail that is necessary to trigger endocytosis or signaling. Recent structural models of G-protein-coupled receptors suggest that the NPX 2-3 Y motif is in an appropriate location to receive a signal from agonist-induced conformational changes in the ligandbinding region (48).
Mechanisms of SP-Induced Endocytosis of the NK1R-The NK1R internalizes by a clathrin-mediated mechanism (21), although some G-protein-coupled receptors internalize by caveolin-dependent pathways (49,50) or by mechanisms that are independent of clathrin and caveolin (51). Despite the finding that many of the mutations affected the rate of internalization, or even resulted in retention of the receptor in very small, superficial vesicles, many of these internalized vesicles were colocalized with clathrin at early time points. Thus, the effect of the mutations is to alter the rate of clathrin-mediated endocytosis, rather than to divert the receptor into a different endocytic pathway. Further analyses will be required to identify proteins that interact with potential endocytic domains of the NK1R. However, Tyr-containing motifs in the C-tails of some membrane proteins interact directly with the clathrinassociated protein complex AP-2 (52,53). The 2-chain of AP-2 interacts with the SDYQRL motif of the C-tail of the integral membrane protein TGN38, as well as with similar motifs from lamp-1, CD68, H2-Mb, and the transferrin-receptor (54). The importance of the interaction between Tyr-containing endocytic motifs and the AP-2 complex has recently been questioned, since deletion of the high affinity AP-2 binding site in the C-tail of the epidermal growth factor receptor abolishes AP-2 binding to the receptor without affecting internalization (55).
Conclusions-Multiple domains in the intracellular C-tail and TMD VII of the NK1R are important for SP-induced endocytosis. All of the conserved Tyr residues in the C-tail participate in endocytosis although Tyr in positions 341 and 349 are most important. Although mutation of the Tyr residue of the NPX 2-3 Y sequence in TMD VII alters endocytosis, it is probable that this region is important for maintaining the correct conformation of the receptor and that it is not an endocytic motif. None of the point mutations reproduced the marked inhibition of endocytosis observed with the truncated receptors or abolished internalization, emphasizing that NK1R internalization relies on multiple endocytic motifs.