Divergent Intracellular Sorting of FcγRIIA and FcγRIIB2*

The human low affinity FcγRII family includes both the activating receptor FcγRIIA and the inhibitory receptor FcγRIIB2. These receptors have opposing signaling functions but are both capable of internalizing IgG-containing immune complexes through clathrin-mediated endocytosis. We demonstrate that upon engagement by multivalent aggregated human IgG, FcγRIIA expressed in ts20 Chinese hamster fibroblasts is delivered along with its ligand to lysosomal compartments for degradation, while FcγRIIB2 dissociates from the ligand and is routed separately into the recycling pathway. FcγRIIA sorting to lysosomes requires receptor multimerization, but does not require either Src family kinase activity or ubiquitylation of receptor lysine residues. The sorting of FcγRIIB2 away from a degradative fate is not due to its lower affinity for IgG and occurs even upon persistent receptor aggregation. Upon co-engagement of FcγRIIA and FcγRIIB2, the receptors are sorted independently to distinct final fates after dissociation of co-clustering ligand. These results reveal fundamental differences in the trafficking behavior of different Fcγ receptors.

The human low affinity Fc␥RII family includes both the activating receptor Fc␥RIIA and the inhibitory receptor Fc␥RIIB2. These receptors have opposing signaling functions but are both capable of internalizing IgG-containing immune complexes through clathrin-mediated endocytosis. We demonstrate that upon engagement by multivalent aggregated human IgG, Fc␥RIIA expressed in ts20 Chinese hamster fibroblasts is delivered along with its ligand to lysosomal compartments for degradation, while Fc␥RIIB2 dissociates from the ligand and is routed separately into the recycling pathway. Fc␥RIIA sorting to lysosomes requires receptor multimerization, but does not require either Src family kinase activity or ubiquitylation of receptor lysine residues. The sorting of Fc␥RIIB2 away from a degradative fate is not due to its lower affinity for IgG and occurs even upon persistent receptor aggregation. Upon co-engagement of Fc␥RIIA and Fc␥RIIB2, the receptors are sorted independently to distinct final fates after dissociation of co-clustering ligand. These results reveal fundamental differences in the trafficking behavior of different Fc␥ receptors.
Fc␥ receptors (Fc␥R) 2 are key players in immune responses. Widely expressed on cells of the hematopoietic system, these receptors mediate a multitude of biological responses that are triggered upon receptor engagement by multivalent IgG-containing immune complexes (1). These responses include production of inflammatory cytokines, antibody-dependent cellular cytotoxicity, and induction of dendritic cell maturation. Fc␥R also mediate the internalization of immune complexes. Soluble immune complexes are internalized by clathrin-mediated endocytosis whereas large (Ͼ0.5 m) antibody-coated particles are internalized via phagocytosis. These uptake processes are important in host defense against infection; moreover, defects in immune complex clearance are associated with the development of autoimmunity (2).
Fc␥Rs can be categorized into two functional groups. The activating Fc␥R have an immunoreceptor tyrosine-based activation motif (ITAM) within the cytoplasmic domain of the receptor itself or an associated FcR␥ signaling subunit. Phos-phorylation of tyrosine residues in the ITAM by Src family kinases after receptor aggregation initiates signaling cascades that trigger downstream effector responses (3). In contrast, inhibitory Fc␥R contain an immunoreceptor tyrosine-based inhibitory motif (ITIM), which recruits phosphatases that antagonize ITAM-mediated signaling. Therefore, the responsiveness of effector cells to immune complexes is determined by the balance between activating and inhibitory receptors (1).
A feature of the human immune system that distinguishes it from that of mouse is the presence of both activating and inhibitory members of the low affinity Fc␥RII subfamily. Fc␥RIIA is an activating Fc␥R unique to humans and other primates, and is the most widely expressed Fc␥R on human leukocytes (4). In addition, Fc␥RIIA is unusual among activating Fc␥R in having both ligand binding and ITAM signaling domains contained in a single polypeptide chain. Fc␥RIIA exists as two codominantly expressed polymorphic variants, with either histidine or arginine at residue 131; the His-131 form has a higher affinity for several IgG subclasses (5). Fc␥RIIA can mediate phagocytosis of large IgG-opsonized particles, which depends on ITAM-mediated signaling and rearrangements of the actin cytoskeleton (6). Fc␥RIIA is also able to mediate clathrin-dependent endocytosis of soluble IgG-containing immune complexes (7,8).
Human leukocytes also express the inhibitory receptor Fc␥RIIB. Whereas the extracellular domains of Fc␥RIIA and Fc␥RIIB have 92% amino acid identity, their intracellular domains are divergent, with Fc␥RIIB containing an ITIM. Two different isoforms of Fc␥RIIB can be generated through alternative splicing. Fc␥RIIB1 is expressed mainly in B cells, where it inhibits signaling from the B cell receptor when the two receptors are coengaged by antigen-antibody complexes (9). Fc␥RIIB2 is expressed mainly in myeloid cells. Human Fc␥RIIB2 lacks the ability to support phagocytosis of IgG-opsonised large particles; on the contrary, it negatively regulates signaling for phagocytosis and other ITAM-dependent responses when coengaged with activating Fc␥R (10). Fc␥RIIB2 is, however, able to mediate endocytosis of soluble immune complexes because of the presence of a di-leucine motif in its cytoplasmic domain (11)(12)(13).
As both of these receptors can mediate immune complex uptake, we sought to investigate how they traffic within the cell after such internalization. This is important for understanding both the overall regulation of inflammatory signaling as well as the processing of immune complexes internalized via Fc␥R. We demonstrate that, in addition to their opposing effects on cell activation, Fc␥RIIA and Fc␥RIIB2 exhibit divergent trafficking behavior after internalization, with Fc␥RIIA, but not Fc␥RIIB2, being targeted for degradation.
Cell Culture-ts20 cells were grown at 34°C and 5% CO 2 in ␣-minimal essential medium ϩ10% fetal bovine serum. For preparation of human monocyte-derived macrophages, peripheral blood mononuclear cells were obtained from blood of healthy donors by Ficoll-Paque Plus (GE Healthcare) density centrifugation, followed by isolation of monocytes by adherence to tissue culture plastic. Cells were then cultured for 6 days in RPMIϩ10% fetal bovine serum with GM-CSF (1000 units/ml).
DNA Constructs and Transfection-cDNAs for wild-type or mutated forms of Fc␥RIIA (His131 variant) and Fc␥RIIB2 were cloned into pcDNA3.1/Myc-His or pcDNA3.1/His (Invitrogen) for expression with C-terminal Myc-His or His tags. Chimeric Fc␥RIIA/Fc␥RIIB2 Myc-His-tagged receptors were generated by PCR amplification of receptor sequences from GFP-tagged chimeric receptors (14) and cloning into pcDNA3.1/Myc-His. Transfections were performed with FuGENE 6 (Roche Applied Science) or Lipofectamine 2000 (Invitrogen) according to the manufacturers' instructions. Stable cell lines expressing Fc␥RIIA, Fc␥RIIB2 and mutant receptors were selected with G418 (0.5 mg/ml). For treatment of cells with inhibitors, cells were pretreated for 30 min by addition of the inhibitor directly to the culture medium at 30 M (PP1) or 300 nM (bafilomycin).
Development of Rabbit Polyclonal Antibodies Specific to the Intracellular Domains of Fc␥RIIA or Fc␥RIIB2-The intracellular domains of human Fc␥RIIA (residues 251-316) or Fc␥RIIB2 (residues 255-290) were individually cloned into the pGEX-3X vector to express the domains as fusion proteins to glutathione S-transferase (GST). The GST-tagged proteins were expressed in BL21 bacteria, purified on GST-agarose and used for immunization of New Zealand White rabbits (Division of Comparative Medicine, University of Toronto). Anti-GST antibodies were removed from antisera using GST-agarose.
Endocytosis Assay-Human IgG (10 mg/ml) was aggregated at 62°C for 20 min followed by centrifugation at 16,000 ϫ g for 10 min to precipitate insoluble IgG aggregates; supernatants containing soluble aggregates were used at 1:100 dilution to induce endocytosis. In experiments where endocytosis was triggered with anti-receptor antibodies, IV.3 (anti-Fc␥RIIA) or AT10 (pan anti-Fc␥RII) were added to transfected ts20 cells at 0.5 g/ml for 20 min at 4°C. After washing, cells were incubated for 20 min at 4°C with 1 g/ml Cy5-donkey anti-mouse antibody to cross-link the receptors, followed by warming to 34°C to trigger endocytosis. In some experiments where subsequent immunofluoresence was performed using mouse antibodies, isotype-specific secondary antibodies were employed. For transferrin loading, rhodamine-or Alexa488-conjugated transferrin was added at 50 g/ml for the last 10 min of incubation. For dextran loading, Alexa647-conjugated dextran was added at 50 g/ml for 1 h followed by a chase for 1 h at 37°C.
Microscopy and Immunofluorescence-Cells were washed and fixed with 4% paraformaldehyde for 30 min, then permeabilized with 0.1% Triton X-100 at room temperature for 20 min. To detect Fc␥R, cells were blocked with 5% BSA in phosphate-buffered saline (PBS) and incubated with anti-Myc antibody 9E10 or anti-Fc␥RIIB2 antibody at 1:1000 dilution in blocking buffer for one hour. For LAMP1 immunofluorescence, cells were permeabilized with methanol at Ϫ20°C for 20 min followed by blocking with 5% BSA and treatment with UH1 antibody at 1:1 dilution. Samples were then treated with fluorophore-conjugated secondary antibodies at 0.8 g/ml in PBS for 30 min before mounting cells with DAKO mounting medium for microscopy analysis. AgIgG was detected with fluorophoreconjugated anti-human secondary antibody.
Cells were analyzed using a Zeiss Axiovert 200 M microscope with 40ϫ or 100ϫ objectives or a Zeiss LSM 510 confocal scanning microscope with a 63ϫ objective. Alexa488, Cy3, and Cy5 signals were detected using standard filter sets.
Flow Cytometry-ts20 cells expressing Myc-His-tagged Fc␥RIIA or FcRIIB2 were detached from culture dishes and dispersed in PBS. Following fixation with 2% paraformaldehyde and permeabilization with 0.1% Triton X-100, cells were blocked with 5% BSA and stained with anti-Myc antibody 9E10 at 1:1000 dilution in blocking buffer for 30 min at room temperature. After washing, cells were stained with Alexa488-conjugated anti-mouse antibody (0.8 g/ml for 15 min) and analyzed by flow cytometry using a FACSCalibur (Becton Dickinson). The background fluorescence observed with 9E10stained untransfected ts20 cells was subtracted from the mean fluorescence intensity (MFI) values. For detection of aggregated IgG, cells were stained with 1.5 g/ml Cy5-anti-human secondary antibody, and background fluorescence observed with untransfected ts20 cells was subtracted.
Immunoprecipitation and Western Blotting-ts20 or THP-1 cells were lysed in lysis buffer (1% Triton X-100, 0.1% SDS, 0.5% deoxycholate, 1 mM NaF, 0.1% protease inhibitor mixture in PBS). Lysates were incubated on ice for 20 min, insoluble material was removed by centrifugation at 16,000 ϫ g for 10 min at 4°C, and lysates were frozen for future analysis. For immunoprecipitations, 1 g of IV.3 antibody and 25 l of protein G beads were mixed with cell lysates followed by overnight nutation at 4°C. After washing beads two times with lysis buffer and two times with PBS-T (PBS with 0.05% Tween 20), beads were resuspended in Laemmli's sample buffer. Samples were analyzed by SDS-PAGE, transferred to nitrocellulose membrane (Bio-Rad), blocked with 5% BSA in PBS-T for 1 h at room temperature and probed with primary antibody at 1:1000 dilution in blocking buffer overnight. Blots were washed in PBS-T three times, incubated with anti-mouse or anti-rabbit horseradish peroxidase for 30 min, washed again, and developed with Supersignal West Pico chemiluminescent substrate (Pierce). A Genius2 Bioimager (Syngene) was used to visualize the blots.

Aggregated IgG Internalized by Either Fc␥RIIA or Fc␥RIIB2 Is
Degraded in Lysosomes-To better understand how these two Fc␥ receptors traffic, Fc␥RIIA or Fc␥RIIB2 carrying Myc-His tags were stably transfected into ts20 cells, a Chinese hamster fibroblast cell line. This heterologous expression system allows us to study the trafficking behavior of a single class of Fc␥R either in its wild type or mutated forms (8,14,15). Soluble complexes of heat aggregated IgG (agIgG), a mimic of soluble multivalent immune complexes, were used to trigger endocytosis of either Fc␥RIIA or Fc␥RIIB2. Both receptors mediated uptake of agIgG ( Fig. 1, B and E), and by 80 min the agIgG was delivered to late endosomal/lysosomal compartments, as indicated by its co-localization with LAMP1 ( Fig. 1, A-F). No significant uptake of agIgG was observed in untransfected cells (data not shown). Consistent with the observed delivery to lysosomes, agIgG internalized via either receptor underwent degradation (Fig. 1G).
Fc␥RIIA, but Not Fc␥RIIB2, Undergoes Degradation in Lysosomes after Internalization-Whereas agIgG ligand internalized by either Fc␥RIIA or Fc␥RIIB2 underwent similar degradation, a second important question is the fate of the receptors themselves. To address this, immunoblotting was performed to determine the amount of receptors remaining in cells after stimulation with agIgG. For Fc␥RIIA, receptor levels declined after induction of endocytosis, such that by 5 h after addition of agIgG very little Fc␥RIIA remained ( Fig. 2A). In contrast, no drop in Fc␥RIIB2 levels was observed over the same time frame ( Fig. 2A). The total amounts of Fc␥RIIA and Fc␥RIIB2 were also quantified by flow cytometry using intracellular staining with anti-myc antibody, which similarly showed loss of Fc␥RIIA but not Fc␥RIIB2 (Fig. 2, B and C and supplemental Fig. S1A). No drop in Fc␥RIIB2 level was observed after agIgG stimulation even in the presence of cycloheximide, indicating that the persistence of Fc␥RIIB2 following endocytosis is not due to new protein synthesis (data not shown). The loss of Fc␥RIIA after endocytosis was also analyzed in THP-1 cells, a human monocytic cell line that expresses endogenous Fc␥RIIA. As in the ts20 transfectants, Fc␥RIIA largely disappeared within 3 h of treatment with agIgG (Fig. 2D). This loss was inhibited by treatment with bafilomycin, indicating that degradation occurs in the endolysosomal system.
To characterize the subcellular localization of Fc␥RIIA and Fc␥RIIB2 after endocytosis, the tagged receptors were detected by immunofluorescence. Under unstimulated conditions, both Fc␥RIIA and Fc␥RIIB2 were observed to reside on the cell surface and in diffuse perinuclear vesicles (Fig. 3, A and J). This intracellular compartment is a recycling endosome pool, as judged by its co-localization with loaded transferrin (Fig. 3, A-C, J-L). After 80 min of agIgG treatment, Fc␥RIIA was largely co-localized with internalized agIgG ligand, which, as noted above (Fig. 1), has moved by this time into punctate perinuclear lysosomes (Fig. 3, G-I). These puncta, while located in the central region of the cell like the transferrinpositive recycling endosomes, are adjacent to these endosomes rather than co-localized with them (Fig. 3, D-F). Thus, consistent with the observed receptor degradation, Fc␥RIIA is targeted to lysosomes along with its ligand. In contrast, at the same time point after agIgG addition, Fc␥RIIB2 showed a localization similar to that seen in unstimulated cells, namely, it was found at the cell surface and in transferrin-positive endosomes (Fig. 3, M-O), and did not follow internalized agIgG to lysosomes (  Images are representative of three experiments. G, ts20 cells expressing Fc␥RIIA or Fc␥RIIB2 were incubated with agIgG at 4°C, washed, and chased at 34°C for the indicated times. Cells were then fixed and permeabilized. Total agIgG remaining was detected with Cy5-anti-human antibody and quantified by flow cytometry. agIgG level is expressed as a percentage of the initial mean fluorescence intensity at time 0 after background subtraction of fluorescence intensity of untransfected ts20 cells incubated with agIgG and stained with Cy5-anti-human antibody. Triangles: Fc␥RIIA; squares: Fc␥RIIB2. Error bars indicate S.D. n ϭ 3. the affinities of the His-131 form of Fc␥RIIA for IgG1, IgG2, and IgG3 are severalfold higher than those of Fc␥RIIB2 (5,16). The divergent localization of Fc␥RIIB2 and agIgG at later times after endocytosis (Fig. 3, P-R) implies that the agIgG dissociates from the receptor after internalization. Thus, one possible determinant of the differential sorting of Fc␥RIIB2 and Fc␥RIIA could be a difference in release of ligand due to their differing extracellular domains. To address this possibility, chi-meric receptors were generated in which the extracellular domain of Fc␥RIIA was fused to the transmembrane and cytoplasmic domains of Fc␥RIIB2 ("IIA-IIB2") or vice versa ("IIB2-IIA"). Western blotting was performed at different time points after induction of agIgG endocytosis in cells expressing these receptors. As shown in Fig. 4, the chimeric receptor with Fc␥RIIA extracellular domain and Fc␥RIIB2 intracellular domain (IIA-IIB2) showed no obvious drop in receptor level after agIgG internalization, while conversely the chimeric receptor with Fc␥RIIB2 extracellular domain and Fc␥RIIA intracellular domain (IIB2-IIA) was degraded. Thus, the differential sorting of these receptors following endocytosis is presumably due to their divergent cytoplasmic domains.
Sorting of Fc␥RIIA and Fc␥RIIB2 Upon Persistent Cross-linking with Antibody Complexes-Whereas a difference in affinity for IgG apparently does not explain the difference between Fc␥RIIA and Fc␥RIIB2 sorting, the question nonetheless remains whether release from clustering ligand is necessary for sorting of the low affinity Fc␥RIIB2 away from a degradative fate, since persistence of ligand binding is thought to be an important factor determining lysosomal sorting of receptors (17). To address this question, we attempted to induce a more persistent clustering of Fc␥RIIB2 by engaging it with anti-Fc␥RII antibody AT10 followed by cross-linking receptors with secondary antibody, rather than through the low affinity binding of multivalent agIgG. When Fc␥RIIB2-expressing cells were incubated with AT10 alone (with no secondary antibody), the AT10 was delivered to transferrin-positive endosomes, suggesting that either there is constitutive cycling of Fc␥RIIB2 between cell surface and recycling endosomes in the unclustered state or that dimerization of receptors by bivalent antibody is sufficient to trigger such cycling (Fig. 5, A-D). Upon clustering receptors with AT10 and secondary antibody, in contrast to the release of ligand that occurs with agIgG, AT10antibody complexes remained co-localized with the receptors, presumably due to their binding through a higher affinity interaction. Moreover, both Fc␥RIIB2 and the AT10 complexes were directed to recycling endosomes (Fig. 5, E-H). Consistent with this localization, neither Fc␥RIIB2 nor AT10 antibody complexes were degraded, in contrast to the case for Fc␥RIIA similarly engaged with AT10 and secondary antibody (supplemental Fig. S2). Thus, even under conditions of persistent oligomerization, Fc␥RIIB2 avoids a degradative fate.
Of note, when Fc␥RIIA-expressing cells were incubated with Fab fragments of the anti-Fc␥RIIA antibody IV.3 and chased at 34°C, the Fab was internalized to transferrin-positive endosomes, suggesting that constitutive cycling of Fc␥RIIA can occur in the absence of clustering (Fig. 5, J-L). In contrast, when whole IV.3 antibodies were used, Fc␥RIIA was sorted away from transferrin-positive recycling endosomes (Fig. 5, M-P) and showed localization in LAMP1-positive lysosomes (Fig. 5, Q-S). Similar results were obtained using whole AT10 (data not shown). Thus, dimerization of Fc␥RIIA by bivalent antibodies (or possibly trimerization through additional engagement of the Fc region) appears to be sufficient to drive this sorting event. With whole IV.3 followed by cross-linking with secondary antibody, robust delivery of Fc␥RIIA to lysosomes was seen (Fig. 5, T-V).

JOURNAL OF BIOLOGICAL CHEMISTRY 34253
To recapitulate these findings in primary human cells, monocyte-derived macrophages were used. Fc␥RIIA was engaged with IV.3 with or without secondary antibody cross-linking. Cells were preloaded with fluorescent dextran to label the lysosomal compartment. At early times after engagement, IV.3 or IV.3 complexes were internalized into peripheral early endosomes that were positive for EEA1 (data not shown) and negative for dextran (Fig. 6, A-C, G-I). By 60 min post-internalization, both IV.3 alone and IV.3-secondary antibody complexes moved into central lysosomes (Fig. 6, D-F, J-L).
Fc␥RIIA Sorting for Degradation Does Not Require Src Family Kinase Activity or Receptor Lysines-Our results with chimeric receptors indicate that it is most likely the divergent intracellular domains of Fc␥RIIA and Fc␥RIIB2 that dictate their divergent sorting. The intracellular domain of Fc␥RIIA contains an ITAM that is phosphorylated by Src family kinases to induce downstream signaling cascades (3). We have previously reported that Fc␥RIIA engagement in transfected ts20 cells leads to receptor phosphorylation, and that both phosphorylation of Fc␥RIIA and phagocytosis driven by this receptor are inhibited by the Src family kinase inhibitor PP1 (8). Here, we tested the effect of PP1 on Fc␥RIIA intracellular sorting. While degradation was blocked by bafilomycin treatment, PP1 did not prevent agIgG-induced Fc␥RIIA degradation (Fig. 7, A and B), indicating that Src family kinase-mediated receptor tyrosine phosphorylation is not essential to direct Fc␥RIIA to lysosomes.
Ubiquitylation of surface receptors can function both to trigger endocytosis and as a sorting signal driving their delivery to lysosomes (18 -20). Fc␥RIIA is ubiquitylated upon engagement by immune complexes and an active ubiquitylation machinery is required for receptor endocytosis (8,15). This raises the possibility that receptor ubiquitylation may also account for its lysosomal sorting. The intracellular domain of Fc␥RIIA contains five lysine residues that can serve as potential ubiquitylation sites. We assessed the degradation of a mutated version of Fc␥RIIA in which all five lysine residues were mutated to arginine (IIA-5KR). ts20 cells expressing either wild-type Fc␥RIIA or IIA-5KR were treated with IV.3 and cross-linking secondary antibody to trigger endocytosis. Antibody complexes internalized by either wild-type Fc␥RIIA or IIA-5KR were both delivered to LAMP1-positive lysosomes (Fig. 8, A-F). Moreover, the lack of lysines does not prevent Fc␥RIIA degradation (Fig. 8G), indicating that direct ubiquitylation of lysines in Fc␥RIIA is not essential for its degradation, and suggesting that other sorting signals in the cytoplasmic domain of Fc␥RIIA are involved in this step of its trafficking.
Fc␥RIIA and Fc␥RIIB2 Trafficking upon Co-engagement with agIgG-Our heterologous transfection model has the advantage of allowing the analysis of the intracellular trafficking capabilities of Fc␥RIIA and Fc␥RIIB2 in isolation. However, in cells expressing both Fc␥RIIA and Fc␥RIIB2, both receptors would be expected to be simultaneously co-engaged by multivalent immune complexes. If Fc␥RIIA and Fc␥RIIB2 carry different sorting signals, what is the result of such receptor co-engagement? On the one hand, Fc␥RIIA might have a dominant effect, pulling Fc␥RIIB2 toward lysosomal degradation. Conversely, co-engaged Fc␥RIIB2 might reroute Fc␥RIIA away from such a fate. A third possibility is that dissociation of immune complexes from the receptors allows them to sort to their respective fates independently of each other. To address this question, ts20 cells stably expressing Myc-His-tagged Fc␥RIIA were transiently transfected with untagged Fc␥RIIB2. AgIgG was added to these cells to co-engage Fc␥RIIA and Fc␥RIIB2 and the subcellular localization of Fc␥RIIA, Fc␥RIIB2, and agIgG were determined by immunofluorescence (Fig. 9, A-H). After 10 min of agIgG stimulation, Fc␥RIIA and Fc␥RIIB2 both co-localized with agIgG in dispersed peripheral endosomes (Fig. 9, A-D). However, by 90 min after agIgG stimulation, Fc␥RIIA and agIgG were colocalized in central puncta but Fc␥RIIB2 showed  a distinct localization in a diffuse central endosomal pool, as was seen with individually expressed receptors (Fig. 9, E-H). As an alternative approach, the effect of co-engagement on receptor degradation was assessed by transiently overexpressing Fc␥RIIB2 in cells that stably express Fc␥RIIA, or vice versa, and measuring degradation of the stably expressed receptor by flow cytometry (Fig. 9, I and J). Expression of cotransfected GFP was used as a measure of transient receptor expression (Fig. 9, I and J and supplemental Fig. S1, B-D). Fc␥RIIA was still degraded in cells expressing a range of levels of Fc␥RIIB2 (Fig. 9I), and conversely there was no obvious degradation of Fc␥RIIB2 in cells expressing a range of levels of Fc␥RIIA (Fig. 9J). Thus, our results suggest that dissociation of Fc␥RII from immune complexes after internalization allows independent sorting of Fc␥RIIA and Fc␥RIIB2 receptors to distinct final fates.

DISCUSSION
In this study, we have directly compared the intracellular trafficking of Fc␥RIIA and Fc␥RIIB2. Our results demonstrate that in addition to their opposing signaling functions, they have very different trafficking behaviors. Degradation of the activating receptor after its internalization may serve as a mechanism for ensuring appropriate termination of inflammatory signaling after initial encounter with immune complexes. In contrast, persistence of Fc␥RIIB2 may help to maintain the ability of myeloid cells to continue clearing immune complexes.
Incubation of Fc␥RIIA-expressing cells with Fab fragments of anti-Fc␥RIIA antibody IV.3 led to labeling of a recycling endosomal pool. This suggests that Fc␥RIIA, like human Fc␥RI (21), has the capacity to cycle between the cell surface and endosomal compartments in the absence of receptor clustering. Notably, we found that engagement of Fc␥RIIA with bivalent whole antibodies was sufficient to induce sorting of the receptors to lysosomes. It would be interesting to investigate to what extent such a minimal degree of clustering is also able to stimulate inflammatory signaling from Fc␥RIIA. It may be that small immune complexes (e.g. dimeric IgG) could have an overall anti-inflammatory effect by triggering down-regulation of activating Fc␥R with relatively little concomitant proinflammatory signaling. This may contribute to the therapeutic effect seen with small immune complexes or immune complex-like  agents in the context of autoimmune diseases in which activating Fc␥R on effector cells play deleterious roles (22,23).
Our experiments with Fc␥RIIB2 highlight the fact that the receptor can dissociate from multivalent agIgG complexes after internalization. On first thought, one might expect that the interaction of receptors with multivalent complexes would be highly stable. Indeed, immune complexes can bind strongly to the cell surface due to the simultaneous interaction of multiple Fc domains with multiple Fc receptors (i.e. avidity). However, once the receptor-ligand complexes have been delivered to endosomal compartments, dissociation of individual receptors from the internalized immune complexes will be determined solely by their affinity for Fc, because each receptor binds only one Fc portion. In this context the multivalent nature of the complex may primarily have the effect of increasing the local concentration of IgG. The low affinity of the Fc␥RII thus will serve to facilitate release of receptors from internalized immune complexes, allowing their spatial segregation from the complexes through endosomal fission events. Our observations suggest that following co-engagement of Fc␥RIIA and Fc␥RIIB2 these two receptors are sorted independently, imply-ing that dissociation of receptors from agIgG can occur before committed sorting events. The low affinity of the Fc␥RII may therefore confer two distinct evolutionary advantages. As is often noted, the low affinity of Fc␥R allows cells to respond specifically to multivalent immune complexes, rather than being permanently occupied by monomeric IgG present in serum. A second advantage, however, may be that low affinity allows Fc␥R to access distinct intracellular compartments even after co-engagement by immune complexes.
As for the effect of affinity not on the fate of the receptors but rather on that of internalized immune complexes, our findings with anti-Fc␥RII antibody imply that release of immune complexes internalized via Fc␥RIIB from the receptor is required for the immune complexes to undergo lysosomal degradation. The low affinity of human Fc␥RIIB for IgG would facilitate this release. It is noteworthy that among the murine low affinity Fc␥R, the inhibitory receptor Fc␥RII binds to mouse IgG1 with higher affinity than the activating receptors Fc␥RIII and Fc␥RIV, and binds to IgG2a and IgG2b with a comparable affinity to Fc␥RIII (24). In contrast, among the human low affinity Fc␥R, Fc␥RIIB has a substantially lower affinity than the activating receptors Fc␥RIIA and Fc␥RIII for all human IgG isotypes except IgG4 (for which affinities are similarly low for all receptors) (5). Fc␥R-mediated uptake of antigen in immune complexes by antigen presenting cells can lead to greatly increased efficiency of subsequent antigen presentation to T cells (25). However, the consequences of antigen uptake via Fc␥RIIB in particular are unclear; in some studies murine  Fc␥RIIB has been shown to be capable of facilitating antigen presentation (26 -29), while in others it has been seen to have an inhibitory effect (30,31). One means by which Fc␥RIIB can impair antigen presentation is by suppressing ITAM-induced dendritic cell maturation (32). In addition, however, Fc␥RIIB in mouse dendritic cells can recycle bound immune complexes to the cell surface and prevent their delivery to degradative compartments that is required for antigen processing (30,31). Recycling of a portion of internalized immune complexes to the cell surface was also observed with rat Fc␥RIIB in liver sinusoidal endothelial cells (33). Such routing of immune complexes away from a degradative fate will depend on the persistence of their binding to receptors. If immune complexes are released from Fc␥RIIB after internalization in human antigen-presenting cells, with delivery of the released immune complexes to lysosomes, the uptake via Fc␥RIIB would be expected to facilitate antigen processing. The pH sensitivity of receptor binding to immune complexes may also be an important factor determining their release in endosomal compartments, and it would be interesting to determine the extent to which this varies among different Fc receptors.
Studies of Fc␥RIIA in neutrophils have shown an extremely rapid degradation of the receptor occurring in less than a minute following its ligation; this degradation appears to involve the action of the proteasome (34). Our results with monocytic THP-1 cells indicate that, as in the transfected cell model, degradation of Fc␥RIIA occurs in lysosomes over the longer time frames typical for such degradation. Moreover, trafficking of anti-Fc␥RIIA antibodies to lysosomes was observed in monocyte-derived macrophages. These results suggest that while loss of Fc␥RIIA after its ligation is a common theme in both monocytic cells and neutrophils, the very rapid degradation of Fc␥RIIA seen in neutrophils may be a mechanistically unique feature of these cells. Another recent study concluded that Fc␥RI, but not Fc␥RIIA, can be delivered to lysosomes in monocytes within 10 min of receptor engagement (35). Our results indicate that Fc␥RIIA can also traffic to lysosomes, with the slower kinetics more typical of lysosomal sorting of surface receptors.
While the ITAM of the common FcR␥ subunit that associates with Fc␥RI is important for triggering downstream signaling upon stimulation of cells via this receptor, the cytoplasmic tail of Fc␥RI itself also has functional effects, controlling ligand binding and endocytosis through interactions with periplakin (36). Our finding that Src family kinase activity is not essential for Fc␥RIIA degradation suggests that the cytoplasmic tail of

Trafficking of Fc␥RIIA and Fc␥RIIB2
OCTOBER 29, 2010 • VOLUME 285 • NUMBER 44 this receptor also interacts with as yet unidentified cytosolic proteins distinct from the ITAM signaling pathway. The lack of requirement for lysine residues also argues against a requirement for direct ubiquitylation of the receptor for its degradation, though we cannot exclude a role for non-conventional ubiquitylation of nonlysine residues (37), or ubiquitylation of a receptor-associated protein. It is also possible that sorting is influenced by the transmembrane domains of Fc␥RIIA and Fc␥RIIB, which differ from each other in three amino acids and which were swapped along with the cytoplasmic domains in our chimeric receptors. The transmembrane domains of both receptors have been shown to affect their localization to membrane microdomains (38,39).
Preferential degradation of the activating Fc␥RIIA coupled with persistence of its inhibitory counterpart implies that after immune complex stimulation, not only does the level of the activating receptor decline, but the ratio of activating to inhibitory receptor also decreases. This reduction should serve to accentuate the termination of signaling from activating receptors, while maintaining cellular ability to clear immune complexes. Thus, these fundamental differences in the sorting of Fc␥RIIA and Fc␥RIIB2 add an additional level of modulation of signaling over the longer term beyond the initial down-regulation of activating signals by ITIM-mediated dephosphorylation.