Small GTPase Rab17 Regulates the Surface Expression of Kainate Receptors but Not α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic Acid (AMPA) Receptors in Hippocampal Neurons via Dendritic Trafficking of Syntaxin-4 Protein*

Background: Rab17 regulates dendritic trafficking in hippocampal neurons, but the cargo molecules are unknown. Results: Using GluK2 and GluA1 as markers, we show that Rab17 or Syntaxin-4 knockdown reduces dendritic surface expression of KARs but not AMPARs. Conclusion: Rab17 mediates membrane insertion of GluK2 but not GluA1 via dendritic trafficking of Syntaxin-4. Significance: Syntaxin-4 specifically regulates GluK2 trafficking.

Glutamate receptors are fundamental for control synaptic transmission, synaptic plasticity, and neuronal excitability. However, many of the molecular mechanisms underlying their trafficking remain elusive. We previously demonstrated that the small GTPase Rab17 regulates dendritic trafficking in hippocampal neurons. Here, we investigated the role(s) of Rab17 in AMPA receptor (AMPAR) and kainate receptor (KAR) trafficking. Although Rab17 knockdown did not affect surface expression of the AMPAR subunit GluA1 under basal or chemically induced long term potentiation conditions, it significantly reduced surface expression of the KAR subunit GluK2. Rab17 co-localizes with Syntaxin-4 in the soma, dendritic shaft, the tips of developing hippocampal neurons, and in spines. Rab17 knockdown caused Syntaxin-4 redistribution away from dendrites and into axons in developing hippocampal neurons. Syntaxin-4 knockdown reduced GluK2 but had no effect on GluA1 surface expression. Moreover, overexpression of constitutively active Rab17 promoted dendritic surface expression of GluK2 by enhancing Syntaxin-4 translocation to dendrites. These data suggest that Rab17 mediates the dendritic trafficking of Syntaxin-4 to selectively regulate dendritic surface insertion of GluK2-containing KARs in rat hippocampal neurons.
Glutamate receptors are critical for excitatory synaptic transmission and plasticity. They are broadly classified into ionotropic and metabotropic types, and ionotropic glutamate receptors are further categorized into NMDA, AMPA, and kainate receptors (1). AMPARs 3 are tetrameric complexes of combinations of four separate subunits (GluA1-4). They are highly mobile proteins that undergo constitutive-and activity-dependent translocation to, and removal from, synapses (2). Crucially, changes in the number and properties of functional synaptic AMPARs results in the long term potentiation (LTP) or long term depression of synaptic efficacy (3).
KARs are also tetrameric assemblies of combinations of five possible subunits (GluK1-5). Presynaptic KARs modulate neurotransmitter release; postsynaptic KARs mediate excitatory neurotransmission, and extrasynaptic KARs are involved in controlling neuronal excitability (4,5). Importantly, KARs participate in the regulation of neuronal network activity and are involved in processes ranging from neuronal development to neurodegeneration and neuronal cell death (4,5).
Multiple proteins have been identified as participating in AMPAR trafficking. These include AP-4 and KIF5, which are involved in transporting AMPARs from the soma to dendrites (6,7); SNARE complex family proteins Syntaxin-3, Syntaxin-4, and SNAP-47 are implicated in AMPAR surface insertion (8,9); and small GTPase Rab and Arf family proteins (i.e. Rab11 and Arf1) play roles in AMPAR internalization and/or recycling (10,11). In contrast, relatively few interacting proteins have been implicated in KAR trafficking, although Rab11 is involved in KAR recycling and SNAP-25 in internalization (12,13). Furthermore, the similarities and differences between the mechanisms and pathways that regulate AMPAR and KAR trafficking and surface expression have not been well defined.
Rab proteins are small GTPases that are conserved in all eukaryotes. They mediate trafficking steps, including vesicle budding, translocation, docking to specific membranes, and fusion (14,15). Rab17 was originally described as an epithelial cell-specific protein that regulates polarized trafficking (16,17). However, we showed previously that Rab17 is also expressed in mouse brain where it is involved in the regulation of dendritic morphogenesis and postsynaptic development of hippocampal neurons (18). Interestingly, Rab17 is the only reported somatodendritic-compartmentalized Rab, but its functional roles in dendritic trafficking remain unclear.
We investigated the possible relationship between Rab17 and AMPAR and KAR trafficking in hippocampal neurons. Rab17 knockdown did not affect trafficking of the AMPAR subunit GluA1, but it markedly reduced surface expression of the KAR subunit GluK2. Surprisingly, GluK2 and Rab17 do not co-localize. Rather, Rab17 is highly co-localized with Syntaxin-4 in somatodendritic compartments of developing neurons and in spines of mature neurons. Rab17 knockdown decreased Syntaxin-4 in dendrites and increased mis-targeted Syntaxin-4 in axons of developing neurons. Syntaxin-4 knockdown reduced GluK2 surface expression but did not affect GluA1. Furthermore, Syntaxin-4 knockdown prevented the increase in surface GluK2 induced by expression of a constitutively active form of Rab17. Taken together, these results suggest Rab17 mediates polarized trafficking of Syntaxin-4 and that dendritic Syntaxin-4 is important for surface insertion of GluK2-containing KARs but not GluA1-containing AMPARs in the somatodendritic compartment of rat hippocampal neurons.
Hippocampal Neuron Culture, Cell Line Culture, and Transfections-Rat embryonic hippocampal neuronal cultures were prepared from E18 Wistar rats. Neurons were then plated at a density of 75,000 -100,000 onto 22-mm glass coverslips in a 6-well dish coated with 1 mg/ml poly-L-lysine (Sigma). The culture medium was composed of Neurobasal medium supplemented with 10% horse serum, 2% B27, 1% GlutaMAX, and 50 units/ml penicillin/streptomycin (all reagents from Invitrogen). On the 2nd day, the media were changed for Neurobasal medium supplemented with 2% B27 and 0.6% GlutaMAX, and 2.5 M cytosine-␤-D-arabinofuranoside (Sigma) was add to the medium after the 4th day of plating. Plasmid DNAs were transfected into the neurons by using Lipofectamine2000 (Invitrogen) according to the manufacturer's instructions.
HEK293T cells and Neuro2A cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Lonza) supplemented with 10% fetal bovine serum (FBS), 50 units/ml penicillin/streptomycin. The cells were plated onto a 6-well plate. Plasmid DNAs were transfected into JET-PEI (Poly-Plus transfection) according to the manufacturer's instructions.
Immunostaining-Neurons were fixed for 10 min with 4% paraformaldehyde (PFA), which was diluted by phosphatebuffered saline (PBS) from 16% PFA (Electron Microscopy Sciences) at room temperature. After permeabilizing the cells with 0.1% Triton X-100 or 0.01% digitonin (for endogenous Syntaxin-4 staining) in PBS for 10 min, they were blocked with the blocking buffer (10% horse serum in PBS) for 1 h. The neurons were then immunostained for primary antibody for 1 h at room temperature or overnight at 4°C after which they were incubated for 1 h with fluorescence-conjugated secondary IgG at room temperature. The neurons were examined for fluorescence with an LSM510 META confocal laser-scanning microscope (Zeiss), and the images were processed with Adobe Photoshop CS software. Fluorescent intensity or the number of fluorescent dots were quantified with ImageJ software (version 1.42q; National Institutes of Health), and Pearson's correlation  was determined manually by using the co-localization indices plug-in (24) to the ImageJ software program. Representative images of neurons are shown in each figure.
Immunoblotting-Cell lysates were subjected to SDS-PAGE and proteins transferred to PVDF membrane (Millipore) by electroblotting. The blots were blocked with 0.3% skim milk and 0.2% Tween 20 in PBS, and after incubating them with a primary antibody, they were washed with PBS containing 0.2% Tween 20 and incubated with a peroxidase-conjugated secondary antibody. Immunoreactive bands were detected by enhanced chemiluminescence or fluorescence with exposure to x-ray film or ODYSSEY FC dual-imaging system (LI-COR Biotechnology) for quantifying the bands by using LI-COR Image Studio software.
Chem-LTP-Chem-LTP was induced as essentially described previously with slight modification (25). Briefly, at 18 DIV, neuronal culture was changed into extracellular solution (ECS: 150 mM NaCl, 2 mM CaCl 2 , 5 mM KCl, 10 mM HEPES (pH 7.4), 30 mM glucose) containing 0.5 M tetrodotoxin. After 5 min in ECS, neurons were treated with 200 M glycine (chem-LTP condition) or without glycine (basal condition) for 3 min in ECS and then incubated ECS containing tetrodotoxin without glycine for 5 min.
GluA1 and GluK2 Surface Expression-GluA1 and GluK2 were surface immunostained using Myc-GluA1, Myc-GluK2, or SEP-GluK2 as described previously with slight modifications (23,25). Neurons were fixed in 4% PFA in PBS for 10 min under the nonpermeabilized conditions. For visualizing surface Myc-GluA1 or Myc-GluK2, the neurons were incubated with Myc antibody in the blocking buffer (1:500 dilution) for 20 min and then Cy3-conjugated secondary antibody (1:1000 dilution) for 20 min. Surface SEP-GluK2 was visualized by incubating neurons with Alexa Fluor 555-conjugated anti-GFP antibody (Invitrogen) in the blocking buffer (1:2000 dilution) for 5 min and then Cy3-conjugated secondary antibody (1:1000 dilution) for 1 h. The neurons were additionally fixed with 4% PFA in PBS for 10 min and permeabilized, and they were subjected to immunostaining with anti-Myc antibody (1:200 dilution), anti-GFP antibody (1:200 dilution), and Cy5-or Cy2-conjugated secondary antibody again to visualize total levels. Surface expression levels were quantified by dividing the surface fluorescence intensity by the total fluorescence intensity.
Quantification of Somatic, Axonal, and Dendritic Distributions-Protein compartmentalization was quantified by setting image thresholds to exclude pixels outside the neuron. After subtracting the background intensity values from each image before quantification, the integrated fluorescence intensity of the protein in the entire neuron or just the region of soma, axon, or dendrite region was obtained. The proportion (%) of the protein in the region was calculated by dividing the fluorescence intensity of the region by the fluorescence intensity of the entire neuron.
Statistical Analyses-Student's unpaired t test was used to evaluate every result for statistical significance in comparison with the results obtained in control. The single asterisk and double asterisk in the bar charts indicate a t test p value of Ͻ0.05 and Ͻ0.0025, respectively, and comparisons that yielded a p value Ͼ0.05 are indicated by NS (not significant).

Rab17 Is Required for Surface Insertion of GluK2 but Not
GluA1-We first investigated the effects of knocking down Rab17 on surface expression of the AMPAR subunit GluA1 under basal and glycine-stimulated chem-LTP conditions. Rab17-shRNA prevented expression of exogenously expressed mCherry-Rab17 and knocked down endogenous Rab17 by almost 80% in rat hippocampal neurons (levels compared with control-shRNA were 3.0 Ϯ 2.3% for mCherry-Rab17 and 21.6 Ϯ 31.1% for endogenous Rab17) (Fig. 1, A-D).
We next examined whether Rab17 is involved KAR surface expression. Because it has been reported that the KAR subunit GluK2 is present at both the pre-and postsynapse (4), we assessed the ratio of surface distribution of Myc-GluK2 in the soma, dendrites, and axons of 14 DIV hippocampal neurons. Myc-GluK2 was predominantly surface-expressed in the soma and dendrite with much less present in axons (ratio of surfaceexpressed Myc-GluK2 50.7 Ϯ 1.5% in dendrite, 3.2 Ϯ 0.9% in axon, and 46.6 Ϯ 3.6% in soma; Fig. 2, A and B). Thus, almost all of the surface signal of Myc-GluK2 can be attributed to the somatodendritic compartment under these conditions. Knockdown of Rab17 dramatically reduced GluK2 surface expression in the rat hippocampal neurons at 14 DIV (to 14.7 Ϯ 4.1% of the control neurons; Fig. 2, C and D). This loss was completely rescued by re-expression of shRNA-resistant mutant Rab17 SR (to 106.7 Ϯ 28.8% of the control neurons) (Fig. 2, C and D).
Interestingly, Rab17 knockdown did not change the total fluorescence intensity of GluK2 in the dendrites (to 103.0 Ϯ 24.7% of the control neurons; Fig. 2, E and F). Furthermore, Rab17 did not co-localize with GluK2 in dendrites (Fig. 2G). Thus, Rab17 is involved in GluK2 surface expression but is not required for the transport of GluK2 from the soma to the dendrite.
Interestingly, Rab17 knockdown resulted in the somatic accumulation of endogenous Syntaxin-4 with a consequent reduction in dendrites in 11 DIV hippocampal neurons (Syntaxin-4 in the dendrites, 26.0 Ϯ 3.1% of the control neurons versus 14.1 Ϯ 2.8% of the Rab17-knockdown neurons) (Fig. 6, I and J). However, unlike Myc-Syntaxin-4, endogenous Syntaxin-4 was not mis-targeted to axons in the Rab17 knockdown neurons (Fig. 6I). These results demonstrate that Rab17 is required for the translocation of Syntaxin-4 to the dendrites in hippocampal neurons.

DISCUSSION
Our results identify Rab17 as an important regulator of GluK2-containing KAR but not GluA1-containing AMPAR surface expression in cultured rat hippocampal neurons. We show that Rab17 is involved in the polarized localization of Syntaxin-4, which in turn regulates GluK2 trafficking. As represented schematically in Fig. 10, we interpret these data to

Rab17 Regulates GluK2 Surface Insertion via Syntaxin-4
suggest that Syntaxin-4 is a cargo molecule of Rab17-containing vesicles and that it is an important determinant of KAR insertion in dendritic membrane.
Several Rab proteins have been reported to directly transport AMPAR-or KAR-containing vesicles (10,12). However, because Rab17 is the only reported dendrite-specific Rab and is involved in dendritic morphogenesis and postsynaptic development (18), we tested whether it is also involved in receptor forward trafficking. We showed that the AMPAR subunits GluA1 and GluA2 do not co-localize with Rab17 (18). Similarly, here we show that the KAR subunit GluK2 also does not co-localize with Rab17 (Fig. 2G), indicating that AMPAR-or KARcontaining vesicles do not contain Rab17. Nonetheless, we show that Rab17 and Syntaxin-4 contribute to GluK2 (but not GluA1) surface insertion at the dendrite in developing hippocampal neurons. Given the dynamic regulation of KAR surface expression (13,23,27), we expect that future studies will investigate the activity dependence of Rab17 and Syntaxin-4 involvement in processes.
Because GluK2 is predominantly surface-expressed in dendrites in our experiments (Fig. 2, A and B), we have not determined whether Rab17 and Syntaxin-4 also regulate presynaptic KAR. Extensive future studies will be necessary to determine the transport, insertion, and internalization mechanism of AMPARs and KARs at the dendrite and axon.
Syntaxin-4 is present in dendritic spines in hippocampal neurons (8), but there are significant discrepancies between reports regarding the physiological function of Syntaxin-4 and AMPAR regulation. Kennedy et al. (8) used fluorophore-tagged transferrin receptor (TfR-SEP) as a marker for GluA1 trafficking and reported that Syntaxin-4 knockdown or overexpression of a dominant-negative Syntaxin-4 prevents activity-dependent surface insertion of TfR-SEP in spines and blocks LTP. Based on these observations, they concluded that Syntaxin-4 is involved in the synaptic surface insertion of AMPAR. In direct contrast however, Jurado et al. (9) state that knockdown of Syntaxin-3, but not of Syntaxin-4, inhibits surface insertion of endogenous GluA1 and blocks LTP (9). In agreement with Jurado et al. (9), our data indicate that Syntaxin-4 is not involved in AMPAR trafficking. Nonetheless, because KARs are implicated in postsynaptic LTP (28), Syntaxin-4 may still contribute to this process.
We interpret our results to suggest that Syntaxin-4 is a cargo molecule of Rab17-containing vesicles, which are specifically sorted for polarized trafficking to the dendrite in developing hippocampal neurons. Although we detected mis-trafficked axonal Myc-Syntaxin-4 in Rab17 knockdown neurons (Fig. 5, D and F), we did not observe similar mislocalization of endogenous Syntaxin-4 (Fig. 6I). One explanation for this is that the comparatively low levels of mislocalized endogenous Syntaxin-4 in axons is rapidly degraded, which occur less quickly for the much higher levels of overexpressed Myc-Syntaxin-4. However, future studies will be necessary to determine whether this is correct and define the mechanisms underlying Syntaxin-4 polarization.
Syntaxin-4 is known to be involved in recycling via the endosomal pathway (29,30), and Rab17 co-localizes with Rab11 in dendritic puncta (18). Furthermore, Syntaxin-4 directly inter-acts with GTP-bound form of Rab4 (31). Thus, an attractive hypothesis is that Rab17 directly interacts with Syntaxin-4 for sorting into the dendrite-specific recycling endosome at the soma.
Neurons lacking Rab17 have fewer and shorter dendrites and are deficient in dendritic filopodia and spines (18), and kainate stimulation alters dendrite length, branching, filopodia number, and elongation (32)(33)(34). However, Syntaxin-4 is not involved in dendrite morphogenesis (Fig. 9). Therefore, we conclude that Syntaxin-4-mediated GluK2 trafficking is not involved in Rab17-dependent dendrite morphogenesis. We speculate that not Syntaxin-4 but instead other cargo molecules of Rab17 or membrane additions from Rab17-containing vesicles to the plasma membrane are important for dendritic morphology in developing hippocampal neurons.
In conclusion, Syntaxin-4 is a cargo molecule of Rab17 in hippocampal dendrites, and both Rab17 and Syntaxin-4 regulate surface insertion of GluK2.