A novel internalization motif regulates human IFN-γR1 endocytosis

A new type of internalization domain highly conserved in IFNγ receptors across species, regulates human IFNγR1 endocytosis.

Abstract

This study tested the hypothesis that the IFN-γR1 287-YVSLI-91 intracellular motif regulates its endocytosis. IFN-γ exerts its biological activities by interacting with a specific cell-surface RC composed of two IFN-γR1 and two IFN-γR2 chains. Following IFN-γ binding and along with the initiation of signal transduction, the ligand and IFN-γR1 are internalized. Two major types of consensus-sorting signals are described in receptors, which are rapidly internalized from the plasma membrane to intracellular compartments: tyrosine-based and dileucine-based internalization motifs. Transfection of HEK 293 cells and IFN-γR1-deficient fibroblasts with WT and site-directed, mutagenesis-generated mutant IFN-γR1 expression vectors helped us to identify region IFN-γR1 287-YVSLI-291 as the critical domain required for IFN-γ-induced IFN-γR1 internalization and Y287 and LI290–291 as part of a common structure essential for receptor endocytosis and function. This new endocytosis motif, YxxLI, shares characteristics of tyrosine-based and dileucine-based internalization motifs and is highly conserved in IFN-γRs across species. The IFN-γR1 270-LI-271 dileucine motif, previously thought to be involved in this receptor endocytosis, showed to be unnecessary for receptor endocytosis.

Introduction

IFN-γ exerts its biological activities by interacting with a specific cell-surface RC composed of two IFN-γR1 (or IFN-γ-binding chain) and two IFN-γR2 (IFN-γR signal-transducing chain) molecules, which are associated with JAK1 and JAK2, respectively. Upon ligand binding, JAK1 phosphorylates Y440 in the IFN-γR1 intracellular domain, which serves as the STAT1 docking site.

STAT1 is also phosphorylated at Y701 and S727, probably by JAK2. Next, two P-STAT1 molecules homodimerize to form the transcription factor GAF. In the currently accepted IFN-γ signal transduction model, GAF rapidly dissociates from the IFN-γ RC in the cytosol and translocates to the nucleus without recognized transport proteins to induce IFN-γ target genes (reviewed in ref. [1]). It is also well documented that upon IFN-γ binding and along with the initiation of signal transduction, the ligand and IFN-γR1 are internalized (receptor-mediated endocytosis; reviewed in ref. [1]). In this model, receptor-mediated endocytosis is a mechanism for ligand-receptor endocytosis and signal attenuation, without further contribution to the pathway. Contrary to this model, other studies show that following IFN-γ stimulation, complexed IFN-γ/IFN-γR1/STAT1 colocalizes at the nuclear membrane and subsequently, accumulates in the nucleus, suggesting that the IFN-γ/IFN-γR1 complex may function as an intracellular chaperone facilitating STAT1 nuclear translocation for gene transcription [2–4]. These latter observations suggest that the endocytosis process is not limited to receptor recycling and degradation but is integral to signaling.

Sorting of transmembrane proteins into endosomal-lysosomal compartments is thought to be mediated by signals present within the cytosolic domains of the proteins in conjunction with the molecular machinery that recognizes those signals and delivers the proteins to their intended destinations. The two major consensus internalization motifs recognized in receptors, which are rapidly internalized and delivered to endosomes, are tyrosine-based (NPxY or YxxØ; where N stands for asparagine, P for proline, x for any amino acid, Y for tyrosine, and Ø for an amino acid with a bulky hydrophobic side-chain), and dileucine-based (LL or LI; where L stands for leucine and I for isoleucine) [5].

Farrar and coworkers [6] identified the region encompassing aa 256–303, within the cytoplasmic domain of IFN-γR1, as necessary for internalization. Dipeptide LI270–271 was later assumed to be the region responsible for this process, as it fit the recognized motif, and deletion of that large region led to impaired surface removal of mutant molecules [6–8]. In support of this assumption, deletions of the intracellular domain of IFN-γR1, caused by common mutations, lead to overaccumulation of truncated receptors on the cell surface [9]. For IFN-γR2, a member of the same class II cytokine receptor family as IFN-γR1 and its partner in IFN-γ signaling, the dileucine motif LI255–256 was identified as the receptor internalization consensus region [10, 11]. Interestingly, the sequence surrounding IFN-γR2 LI255–256 showed characteristics of tyrosine-based and dileucine-based endocytosis motifs: 252YRGLI256, where YRGL corresponds to the tyrosine-based endocytosis motif (YxxØ) and blends into the dileucine-based endocytosis motif LI255–256.

Amino acid analysis of the IFN-γR1 cytoplasmic domain shows a second LI dipeptide (LI290–291) downstream from the originally described IFN-γR1 endocytosis domain (LI270–271). As with IFN-γR2, with which IFN-γR1 shares multiple structural features, the region surrounding IFN-γR1 LI290–291 has characteristics of tyrosine-based and dileucine-based endocytosis motifs: 287-YVSLI-291, where YVSL corresponds to the tyrosine-based endocytosis motif (YxxØ) and blends into the dileucine-based endocytosis motif LI290–291.

As this distal LI motif in IFN-γR1 is similar to the intracellular domain of IFN-γR2, which has been identified as the receptor endocytosis domain, we sought to determine the role of IFN-γR1 287-YVSLI-291 in receptor function.

MATERIALS AND METHODS

Constructs

The cDNA sequences for the IFN-γR1 MP and SP were cloned separately into pDSRed2-C1 (bicistronic vector coding for RFP and a Kana/Neo-resistance gene; BD Clontech, Palo Alto, CA, USA), resulting in the construct 5′ … SP/RFP/IFN-γR1 MP … 3′, expressing RFP-IFN-γR1 as a fusion protein. ATCC (Manassas, VA, USA) IFN-γR1 (ATCC 59,872) was used as the template for IFN-γR1 MP and SP cloning (Supplemental Data). The inserts and boundaries were sequenced and confirmed to be appropriate, as was the orientation. Expression of the RFP-IFN-γR1 fusion protein was detected by immunoblot using anti-RFP antibody (Millipore, Chemicon, Billerica, MA, USA) and mouse anti-human IFN-γR1 mAb (CD119; US Biological, Swampscott, MA, USA), confocal microscopy, and also, flow cytometry with anti-RFP antibody (Millipore, Chemicon; Fig. 1).

Figure 1. rRFP-IFN-γR1 fusion protein detection in transiently transfected HEK 293 cells.

(A) Immunoblot (representative of two independent experiments). HEK 293 cells were transfected with WT or 286X mutant vector. Identical bands were detected with anti-RFP antibody and anti-IFN-γR1 mAb (CD119). (B) Confocal microscopy (left) and flow cytometry (right; representative of two independent experiments). rRFP-IFN-γR1 fusion protein could be localized by confocal microscopy to the cell surface and the cytosol, but not to the nucleus, of HEK 293 cells transfected with the WT vector. Cell surface expression was confirmed by flow cytometry using a FITC-conjugated anti-RFP mAb without cell permeabilization. FL2, Fluorescence 2.

Mutagenesis.

Mutations 270Δ2, 286X, 293X, 290Δ2, LI290–291AA, Y287A, Δ290, L290A, Δ291, I291A, and 288Δ2 were introduced separately into the WT construct by site-directed mutagenesis using Pfx (Invitrogen Life Technologies, Carlsbad, CA, USA) and DpnI (New England BioLabs, Ipswich, MA, USA; Supplemental Data). Mutant constructs are shown in Figs. 2 and 3.

Figure 2. Expression, accumulation, localization, and internalization of WT IFN-γR1 and mutants 286X, 270Δ2, and 293X.

(A) Schematic representation of IFN-γR1 WT and mutants 286X, 270Δ2, and 293X. (B and C) Immunoblot and densitometry, respectively. RFP-IFN-γR1 fusion protein accumulation was evaluated in HEK 293 cells. The fusion proteins (WT or mutant) were detected with an anti-RFP antibody. The intensity of each RFP-IFN-γR1 fusion protein band was normalized with β-actin and compared with the WT:β-actin ratio, arbitrarily defined as “1”. The median mutant:β-actin/WT:β-actin ratio values are indicated. *Significant differences. Each vector was tested independently at least four times (dots). Mutant 286X overaccumulated 1.9 times in comparison with the WT vector (P=0.0352). (D) Confocal microscopy (representative of three independent experiments). RFP-IFN-γR1 fusion protein localization was evaluated. WT, 270Δ2 and 293X vectors showed a comparable membrane accumulation level, whereas mutant 286X showed increased plasma membrane accumulation. All pictures were taken at identical settings. (E) Confocal microscopy (representative of three independent experiments). RFP-IFN-γR1 fusion protein internalization was evaluated. HEK 293 cells transiently transfected with different mutants were incubated with a rabbit anti-RFP antibody, reincubated with a FITC-conjugated goat F(ab′)2 anti-rabbit IgG (H+L) green fluorescent antibody and stimulated with IFN-γ to promote receptor internalization. The WT, 270Δ2, and 293X receptors showed preserved internalization: the green fluorescent antibody was detected on the cell surface and intracellularly. The 286X mutant vector showed impaired internalization: the green fluorescent antibody was detected almost exclusively on the cell surface. (F) Immunoblot. RFP-IFN-γR1 fusion protein accumulation was evaluated in IFN-γR1-DF. The fusion proteins (WT or mutant) were detected with an anti-RFP antibody. WT, 270Δ2, and 293X vectors showed a comparable accumulation level, whereas mutant 286X showed increased accumulation.

Figure 3. Expression, accumulation, localization, and internalization of WT and mutant IFN-γR1 vectors involving region 287-YVSLITS-293.

(A) Schematic representation of WT IFN-γR1 and mutants 290Δ2, LI290–291AA, Y287A, Δ290, L290A, Δ291, I291A, and 288Δ2. (B and C) Immunoblot and densitometry, respectively. RFP-IFN-γR1 fusion protein accumulation was evaluated on transiently transfected HEK 293 cells. The fusion proteins (WT or mutant) were detected with an anti-RFP antibody. The intensity of each RFP-IFN-γR1 fusion protein band was corrected for β-actin expression and compared with the WT:β-actin ratio, arbitrarily defined as “1”. The median mutant:β-actin/WT:β-actin] ratio values are indicated. *Significant differences. Each vector was tested independently (dots) at least four times. WT and mutants Δ290, Δ291, I291A, L290A, and 288Δ2 vectors showed comparable accumulation levels. Mutants Y287A, 290Δ2, and LI290–291AA overaccumulated 2.2, 2.2, and 1.5 times, respectively, in comparison with the WT vector (P=0.0156, P=0.0313, and P=0.0078, respectively). (D) Confocal microscopy (representative of three independent experiments). RFP-IFN-γR1 fusion protein localization was evaluated. Mutants 290Δ2 and Y287A show increased plasma membrane accumulation when compared with the WT. All pictures were taken at identical settings. (E) Confocal microscopy (representative of three independent experiments). RFP-IFN-γR1 fusion protein internalization was evaluated. HEK 293 cells, transiently transfected with different mutants, were incubated with a rabbit anti-RFP antibody, reincubated with a FITC-conjugated goat F(ab′)2 anti-rabbit IgG (H+L) green fluorescent antibody, and stimulated with IFN-γ to promote receptor internalization. The WT receptor showed preserved internalization: the green fluorescent antibody was detected on the cell surface and intracellularly. The Y287A and 290Δ2 (continued on next page) mutant vectors showed impaired internalization: the green fluorescent antibody was detected almost exclusively on the cell surface after IFN-γ stimulation. (F) RT-PCR. HEK 293 cells were transiently transfected with WT or mutant Y287A and 290Δ2 vectors. Relative quantification of gene expression was performed by a comparative threshold method between RFP (target gene) and kanamycin (endogenous control). No significant differences in gene transcription were detected between the WT and the overaccumulating mutant vectors. The experiment was repeated twice, and each condition was reproduced by triplicate.

Cell lines

HEK 293 (ATCC, CRL-1573) and human IFN-γR1-deficient fibroblasts (IFN-γR1-DF, compound heterozygous 107ins4/200+1G->A) [12] were grown in DMEM (Invitrogen Life Technologies, Gibco) with 10% FCS (Invitrogen Life Technologies), 10 mM HEPES (Invitrogen Life Technologies), 2 mM L-glutamine (Invitrogen Life Technologies), and antibiotics (Pen/Strep; Invitrogen Life Technologies). Cells were split every 3–4 days.

Transfections

HEK 293 were transfected with Lipofectamine 2000 (Invitrogen Life Technologies) following the Lipofectamine 2000-adherent cell-transient transfection protocol. IFN-γR1-deficient fibroblasts were transfected with the AMAXA Cell Line Nucleofector Kit R (Lonza Cologne AG, Koln, Germany), following the manufacturer's recommendations (Supplemental Data).

IFN-γR1 accumulation

RFP-IFN-γR1 fusion protein accumulation was evaluated by immunoblotting in HEK 293 and IFN-γR1-DF cells, transiently transfected with WT and mutant vectors. WT and mutant RFP-IFN-γR1 gene transcription was evaluated by RT-PCR on transiently transfected HEK 293 cells (Supplemental Data).

IFN-γR1 distribution and internalization

RFP-IFN-γR1 fusion protein localization and internalization were evaluated by confocal microscopy on HEK 293 cells transiently transfected with WT and mutant vectors. For IFN-γ-induced IFN-γR1 internalization studies, media were removed, 48 h after transfections, and cells were incubated with rabbit anti-RFP antibody (Millipore, Chemicon; 1/200 dilution in fresh media) overnight. Media were removed, and cells were washed, 2 times with 37°C-preheated PBS media, and then reincubated for 1 h at room temperature with a FITC-conjugated goat F(ab′)2 anti-rabbit IgG(H+L) green fluorescent antibody (Caltag, Invitrogen Life Technologies; 1/80 dilution). All steps were performed without cell permeabilization. To induce IFN-γR1 internalization, transfected cells were incubated with IFN-γ (1000 U/ml for 15 min) prior to confocal microscopy evaluation. The green fluorescent antibody should be detected on the cell surface as well as intracellularly on cells with preserved internalization capacity but only on the cell surface on cells expressing receptors with impaired internalization (Supplemental Data).

Cellular response to IFN-γ

IFN-γ responsiveness was evaluated by P-STAT1 (IFN-γR1-DF, immunoblotting), and MHC-I up-regulation (HEK 293 cells, flow cytometry; Supplemental Data).

RESULTS AND DISCUSSION

WT IFN-γR1 expression vector reproduces the natural expression pattern of IFN-γR1

HEK 293 cells were transfected with the WT IFN-γR1 vector to evaluate RFP-IFN-γR1 fusion protein expression and localization. The fusion protein was detected with anti-RFP and anti-IFN-γR1 antibody by immunoblot. Confocal microscopy and flow cytometry studies showed the fusion protein on the cell surface and in the cytosol with nuclear exclusion (Fig. 1). This expression pattern is the same as that described previously for natural IFN-γR1 [1].

IFN-γR1 LI270–271 is not critical for receptor internalization; rather, the receptor endocytosis motif resides at 287-YVSLI-291

To evaluate the role of regions LI270–271 and 287-YVSLI-291 on IFN-γR1 biology, HEK 293 cells were transiently transfected with WT and mutant vectors 270Δ2, 286X, and 293X IFN-γR1 to determine receptor accumulation, localization, and endocytosis (Fig. 2A–F). There were no significant differences among the WT, 270Δ2, and 293X vectors. In contrast, mutant receptor 286X showed significant overaccumulation and was not internalized after IFN-γ stimulation. Therefore, neither LI270–271 nor the domain distal to aa 292 is necessary for IFN-γR1 endocytosis, accumulation, or localization, but the region downstream of aa 286 and upstream of aa 292 (IFN-γR1 287-YVSLIT-292) is.

To determine the effect of the endogenous HEK 293 naturally expressed IFN-γR1 on the transfected vectors, IFN-γR1-DF were transiently transfected. IFN-γR1-DF transfected cells showed the same pattern of receptor accumulation seen with HEK 293 cells: WT, 270X, and 293X expression vectors were indistinguishable, whereas mutant 286X overaccumulated. Therefore, the endogenous IFN-γR1 expressed on HEK 293 cells is unlikely to be interfering with the accumulation dynamics of the labeled expression vectors.

The IFN-γR1 287-YxxLI-291 motif is responsible for receptor internalization and accumulation

HEK 293 cells were transiently transfected with WT and mutant IFN-γR1 vectors Y287A, 290Δ2, LI290–291AA, Δ290, L290A, Δ291, I291A, and 288Δ2 to evaluate their accumulation, localization, and endocytosis (Fig. 3A–F). When compared with WT, mutant receptors Y287A, 290Δ2, and LI290–291AA all showed similar overaccumulation. Interestingly, neither single amino acid mutations affecting the dipeptide LI290–291 (Δ290, L290A, Δ291, and I291A) nor deletion of aa 288–289 (288Δ2) showed significant differences from WT. The mutant vectors Y287A, 290Δ2, and LI290–291AA all showed plasma membrane-localized overaccumulation and impaired endocytosis by confocal imaging studies. Increased transcription of the overaccumulating mutant vectors Y287A or 290Δ2 was excluded by RT-PCR. Therefore, the domain IFN-γR1 287-YVSLI-291 is necessary for receptor internalization, and Y287 as well as dipeptide LI290–291 are critical for the IFN-γR1 endocytosis process.

IFN-γ-induced P-STAT1 is not dependant on IFN-γR1 internalization

IFN-γR1-DF were transiently transfected with WT or mutants 270Δ2, 286X, Y287A, 288Δ2, 290Δ2, LI290–291AA, and 293X to evaluate IFN-γ responsiveness by P-STAT1 (Fig. 4). As expected, neither untransfected nor STAT1 binding-site-devoid mutants 286X and 293X were able to support IFN-γ signaling. In contrast, cells transfected with WT IFN-γR1 or mutant vectors 270Δ2 and 288Δ2 appropriately phosphorylated STAT1. Interestingly, internalization-impaired mutants Y287A, 290Δ2, and LI290–291AA also supported P-STAT1 similarly to the WT vector. These results indicate that IFN-γ-induced P-STAT1, an early step in IFN-γ signaling, is not dependant on IFN-γR1 endocytosis; besides, they also reinforce the established concept that IFN-γR2, through its tightly regulated plasma membrane expression and not IFN-γR1, is the bottleneck for IFN-γ signaling [1]. This mechanism prevents IFN-γR1 overaccumulating vectors Y287A, 290Δ2, or LI290–291AA from exerting dominant gain-of-function effects.

Figure 4. Cellular response to IFN-γ: P-STAT1 accumulation in IFN-γR1-DF immunoblot (representative of three independent experiments).

IFN-γR1-DF was transfected with WT and mutant IFN-γR1 vectors, stimulated with IFN-γ, and P-STAT1 accumulation was evaluated as a marker of reconstitution of IFN-γ responsiveness. One out of three representative experiments is shown. WT and mutant constructs 270Δ2, Y287A, 288Δ2, 290Δ2, and LI290–291AA supported IFN-γ signaling. No significant differences in terms of P-STAT1 accumulation were detected between these vectors (densitometry, not shown).

Overaccumulating signaling mutants Y287A and 290Δ2 do not exert dominant negative effects on MHC-I expression

MHC-I up-regulation upon IFN-γ stimulation was evaluated by flow cytometry on untransfected HEK 293 cells and those transiently transfected with WT, Y287A, and 290Δ2 (Fig. 5). Significant MHC-I up-regulation was found on untransfected and transfected cells, without significant differences between the other conditions tested. These results suggest that neither the WT nor the mutant IFN-γR1 vectors exerted dominant gain or loss-of-function effects over the signaling of the native WT IFN-γ RC expressed by HEK 293 cells. Altogether, these findings show that in opposition to other overaccumulating IFN-γR1 mutants described previously [13], mutants 290Δ2 and Y287A produce no significant effects on native WT IFN-γ RC signaling.

Figure 5. MHC-I up-regulation in HEK 293 cells in response to IFN-γ.

Flow cytometry. Human HEK 293 cells were left untransfected or transfected with different IFN-γR1vectors. MHC-I up-regulation was evaluated after 48 h of IFN-γ stimulation (1000 U/ml). Only positive-transfected cells were gated; the mean fluorescent intensity (MFI) was measured and compared between the different conditions. (Upper) Histograms from one of three independent experiments are shown (solid black line, IFN-γ-stimulated cells; solid gray line, unstimulated cells; dotted black line, isotype control). (Lower) Statistical analysis is shown (*P<0.05). Untransfected and transfected HEK 293 cells significantly up-regulated MHC-I expression upon IFN-γ stimulation; no other differences between the experimental conditions were detected.

As reported previously, endocytosis is the fundamental cellular process by which eukaryotic cells internalize material, including that from the extracellular medium and fragments that have budded or been pinched off from the plasma membrane. Upon IFN-γ binding, along with the initiation of signal transduction, the IFN-γ/IFN-γR1 complex is internalized by a process of receptor-mediated endocytosis [9, 14–17]. Typically, receptor-mediated endocytosis is a mechanism for ligand/receptor recycling and/or signal attenuation, without further contribution to signaling. Sorting of transmembrane proteins into the endosomal-lysosomal system is mediated by signals present within the cytosolic domains of those proteins coupled to molecular machinery that recognizes those signals and delivers the proteins to their intended destinations. There are different types of sorting signals, including tyrosine-based and dileucine-based motifs [5, 18].

Almost two decades ago, Farrar et al. [6] reported that IFN-γR1 aa 256–303 were required for receptor internalization. Since then, the IFN-γR1 intracellular LI dipeptide at 270–271 has been thought to be involved in receptor endocytosis [6–8]. Interestingly, at 20 aa downstream of LI270–271, another dileucine motif is found (IFN-γR1 LI290–291). Similar to what has been described for the IFN-γR2 endocytosis motif [10], the IFN-γR1 LI290–291-flanking sequence shares characteristics of tyrosine-based- and dileucine-based-sorting motifs. When Jouanguy et al. [13] first identified human mutations leading to IFN-γR1 overaccumulation and dominant inhibition of signaling (e.g., mutation 818del4, introducing 5 missense aa and an early stop codon at position 262), it was thought that overaccumulation was secondary to the receptor being devoid of the LI motif at 270–271. However, our experiments revealed that the mutant construct 286X, which retained LI270–271 but lacked LI290–291 motif, showed plasma membrane overaccumulation and impaired receptor endocytosis. In stark contrast, mutant 293X, which retained the LI290–291 motif, showed normal plasma membrane accumulation levels and endocytosis. Therefore, amino acids, encompassed between positions 287 and 292, must mediate IFN-γR1 plasma membrane accumulation and endocytosis.

Systematic mutational analyses led us to define this tyrosine- and dileucine-based internalization motif (YxxLI) as the determinant domain for IFN-γR1 endocytosis. Mutant receptors, in which aa Y287 or LI290–291 were modified, showed the same magnitude of overaccumulation as 286X, in which the whole 287-YVSLI-291 motif was eliminated. Neither Y287 (preserved in mutants 290Δ2 and LI290–291AA) nor LI290–291 (preserved in mutant Y287A) were individually sufficient to support normal receptor endocytosis and plasma membrane accumulation. These data indicate that IFN-γR1 Y287 and LI290–291 are part of a common structure, and both are required for receptor endocytosis. In addition, when particular amino acids where deleted or Ala-substituted (Δ290-L290A, Δ291-I291A, 290Δ2-LI290–291AA), no differences between the paired mutants were detected, adding to our understanding of the biology of this novel type of endocytosis motif. This newly recognized internalization motif therefore includes characteristics of tyrosine-based and dileucine-based endocytosis domains, which are equally necessary for function. Moreover, phylogenetic analysis shows that Y287 and LI290–291 are immersed in a highly conserved region among IFN-γR1 receptors across species. The homology region starts ∼15 aa upstream of Y287 and extends up to the carboxy-terminus end of the receptor, indicating evolutionary conservation of the sorting (and possibly signaling) mechanisms in which they participate (Fig. 6). However, neither the conserved region upstream of Y287 nor the one downstream of LI290–291 appears to play a significant role in the receptor endocytosis based on our experiments in human IFN-γR1; besides, Y287 and LI290–291 are the only amino acids preserved in this particular region between both human IFN-γRs.

Figure 6. IFN-γR1 YxxLI endocytosis motif, phylogenetic analysis.

The IFN-γR1 sequences from different species were compared using ClustalX 8.1. IFN-γR1 YxxLI is highly conserved among species.

Amino acid sequence is not the only variable determining the function of an endocytosis motif; for tyrosine-based endocytosis signals, position within the cytosolic domain also determines receptor fate. The purely endocytic tyrosine-based motifs, YxxØ, are most often situated 10–40 residues from the transmembrane domain but not at the carboxy-termini of proteins. In contrast, lysosomal-targeting, tyrosine-based endocytosis signals are conspicuous for their presence, six to nine residues from the transmembrane domain, as well as at the carboxy-termini of proteins [5]. Like tyrosine-based endocytosis signals, dileucine-based endocytosis signals are also influenced by their position within the intracellular domain [5]. Based on its intracellular position, the IFN-γR1 287-YVSLI-291 motif may have a predominantly endocytic rather than lysosomal targeting role.

Previous reports have shown that clathrin and caveolae-associated mechanisms are involved in the IFN-γR1 internalization process [19–22]. In addition, AP2 simultaneously recognizes and binds intracellular tyrosine-based and dileucine-based motifs, as well as clathrin [5]. As a result of the structural analogy between this newly described hybrid endocytosis motif and the classic tyrosine-based and dileucine-based motifs, 287-YVSLI-291 may also bind AP2 and be implicated in clathrin-dependent endocytosis.

Several mutations in the intracellular domain of IFN-γR1 have been described to affect its internalization and transduction capacity [6, 8, 23–25]. These mutants generate stop codons proximal to LI270–271 (and therefore, proximal to Y287 and LI290–291) and result in overaccumulating receptors devoid of intracellular domains exerting dominant-negative effects over the remaining native WT allele. Mutants Y287A and 290Δ2, which only affect the 287-YVSLI-291 IFN-γR1 endocytosis motif but do not truncate the receptor, also overaccumulate on the plasma membrane but do not affect native WT IFN-γ RC signaling (Figs. 5). Therefore, we have created IFN-γR1 overaccumulating mutants, which do not affect signaling, providing valuable reagents to dissect the relationship between IFN-γR1 internalization and signaling [2–4, 19–22].

In summary, IFN-γR1 287-YVSLI-291 is the critical region for IFN-γR1 IFN-γ-induced receptor internalization. This novel endocytosis motif, YxxLI, shares characteristics of tyrosine-based and dileucine-based internalization sequences and is highly conserved in IFN-γRs across species.

AUTHORSHIP

J.Y., M.A.S., N.A., and A.B. performed the experiments and discussed the results. J.Y. and S.D.R. designed the experiments, analyzed data, and wrote the manuscript. S.M.H. and S.D.R. conceived of the study and reviewed the manuscript.

DISCLOSURES

The authors have no conflict of interests to declare.