|Home | About | Journals | Submit | Contact Us | Français|
Protrudin is a FYVE (Fab 1, YOTB, Vac 1, and EEA1) domain-containing protein involved in transport of neuronal cargoes and implicated in the onset of hereditary spastic paraplegia. Our image-based screening of the lipid binding domain library revealed novel plasma membrane localization of the FYVE domain of protrudin unlike canonical FYVE domains that are localized to early endosomes. The membrane binding study by surface plasmon resonance analysis showed that this FYVE domain preferentially binds phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2), phosphatidylinositol 3,4-bisphosphate (PtdIns(3,4)P2), and phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P3) unlike canonical FYVE domains that specifically bind phosphatidylinositol 3-phosphate (PtdIns(3)P). Furthermore, we found that these phosphoinositides (PtdInsP) differentially regulate shuttling of protrudin between endosomes and plasma membrane via its FYVE domain. Protrudin mutants with reduced PtdInsP-binding affinity failed to promote neurite outgrowth in primary cultured hippocampal neurons. These results suggest that novel PtdInsP selectivity of the protrudin-FYVE domain is critical for its cellular localization and its role in neurite outgrowth.
Hundreds of signaling proteins recognize specific phospholipids via lipid binding domains (LBD)3 with diverse structures and lipid specificity (1–3). Phospholipid-LBD interaction is critical in regulating the appropriate subcellular localization of proteins involved in diverse cellular processes, including cell signaling, vesicle trafficking, organization of the cytoskeleton, autophagy, and neurite outgrowth (1–3). Defects of proteins regulating these lipid-mediated processes cause various human diseases exact causative mechanisms of which have not been fully understood (4).
Patients with hereditary spastic paraplegia (HSP), an inherited neurological disease, suffer from progressive dysfunction of the nerves (5). The large group of HSP-related proteins appears to be involved in membrane trafficking (5). Previous reports showed that HSP-related proteins, spastin and kinesin superfamily protein 5 (KIF5), interact with protrudin (6). Protrudin is a FYVE domain-containing protein that also harbors a Rab11-binding domain (RBD11), two hydrophobic domains (HP-1 and HP-2), an FFAT motif, and a coiled-coil domain (7, 8). Protrudin is reported to play a role in membrane recycling (8) and in induction of neurite outgrowth through the delivery of neuronal cargoes (6, 7). It has also been suggested that protrudin is implicated in an autosomal dominant form of HSP (AD-HSP) (6, 8, 9). However, it is not known how the cellular function of protrudin is regulated and how protrudin is involved in neurite formation and the onset of HSP.
The FYVE domain is a zinc-containing module of 60–80 amino acids that specifically bind PtdIns(3)P (10–12). As expected from the endosomal localization of PtdIns(3)P and its role in vesicle trafficking, a large number of FYVE domain-containing proteins, including EEA1, Hrs, and FENS-1, are involved in endocytic vesicle trafficking (13–15). Some FYVE domain-containing proteins also function in cytoskeletal regulation (FGD1) (16) and growth factor signaling (SARA (17) and endofin (18)). High-resolution structures of different FYVE domains showed that they achieve high PtdIns(3)P specificity via the lipid binding pocket composed of three conserved motifs, the N-terminal WXXD, the central R + HHC + XCG, and the C-terminal RVC motifs, where + represents a basic residue and X any residue (19–22). To date, the specificity of FYVE domains for other PtdInsPs has not been reported. In this study, we found that the FYVE domain of protrudin has unique subcellular localization and novel selectivity for PtdIns(4,5)P2, PtdIns(3,4)P2, and PtdIns(3,4,5)P3 by means of imaging-based cell screening (23–25) and SPR analysis. Our study suggests that binding of PtdIns(4,5)P2, PtdIns(3,4)P2, and PtdIns(3,4,5)P3 to the FYVE domain may differentially regulate the shuttling of protrudin between the endosome and PM, which is critical for the role of protrudin in mediating neurite outgrowth.
LBDs were tagged with enhanced yellow fluorescence protein (EYFP) by a combination of pENTR-207 with pDSA03 (pEYFP-XB) using LR recombinase (Invitrogen). LBDs generally contain additional flanking sequences on both sides of the domain ends. Detailed information on constructed LBDs is listed in supplemental Table S1. Protrudin was a kind gift of Keiichi I. Nakayama (8) and transferred into pCitrine-C1 vector. To make FYVE domain-deleted protrudin, the FYVE domain was removed from positions 348 to 409 in full-length protrudin by using HindIII and KpnI. EYFP-Cb5 (26) and EYFP-Btk-PH domains were kindly provided from Takanari Inoue. PLCδ-PH domain and Akt-PH domain constructs were created in pEYFP-C1 vector (Clontech). 2× FYVE domain of FENS-1 was constructed by using compatible ends in the EYFP-C1 vector (Clontech). ARF6(Q67L)-EYFP and ARF6(T27N)-EYFP were generated in the pEYFP-N3 vector using BglII and KpnI sites, and EYFP-Rab5B(Q79L) was inserted in pEYFP-C1 vector (Clontech) using EcoRI and BamHI sites. Lyn was constructed using EcoRI and BamHI sites into the pEYFP-N1 vector (Clontech) and then EYFP of Lyn-EYFP was replaced with mCherry. The FYVE domain (318–409 amino acids) of protrudin was constructed in pRSET-B vector (Invitrogen) for protein purification and surface plasmon resonance (SPR) analysis.
NIH3T3 was obtained from American Type Culture Collection (ATCC, Manassas, VA) and maintained in Dulbecco's modified Eagle's medium (DMEM, PAA Laboratories GmbH) supplemented with 10% fetal bovine serum (FBS, Invitrogen) at 37 °C in an atmosphere of 10% CO2 and 95% humidity. Rat hippocampus was obtained from embryonic rat brain at embryonic day 18 (KOATECH). Primary rat hippocampal neurons were plated in Neurobasal medium containing B27 supplement (Invitrogen) and GlutaMAX (Invitrogen). Transfection was performed by using Lipofectamine 2000 (Invitrogen), Neon® Transfection System 10 μl (Invitrogen) or the CalPhosTM Mammalian Transfection kit (Clontech) (27) according to the manufacturer's instructions.
NIH3T3 cells expressing LBDs were observed using a ×60 lens (oil immersion objective, NA 1.40; Nikon) and primary cultured hippocampal neuron expressing protrudin was imaged using ×20 lens (air objective, NA 0.75; Nikon) of an A1R confocal microscope (Nikon) at room temperature. Images were acquired at a resolution of 512 × 512. Time-lapse images were captured with 30-s intervals.
PLCδ-PH, Btk-PH, and Akt-PH domains were used as PtdIns(4,5)P2, PtdIns(3,4,5)P3, and PtdIns (3,4)P2/PtdIns(3,4,5)P3 biosensors, respectively (28–30). 2× FYVE domain of FENS-1 served as a biosensor for PtdIns(3)P (31). ARF6(Q67L)-EYFP and ARF6(T27N)-EYFP were applied to induce and block PIP5K activation followed by formation of PtdIns(4,5)P2-containing endocytic vesicles (32). EYFP-Rab5B(Q79L) was used to accumulate PtdIns(3)P in the endocytic vesicles (33). 50 μm LY29 (Sigma) or 5 nm PDGF (PreproTech) was used to inhibit and induce phosphoinositide 3-kinase (PI3K) activation at the PM, respectively (23, 34). LY30 (Sigma) served as a structural inactive analog for LY29 (23).
Image analyses were performed using the NIS-element AR 64 bit version 3.1 and MetaMorph offline version 220.127.116.11 from MDS Analytical Technologies. Neurite length of primary culture hippocampal neurons was measured using NeuronJ 1.4.2 (National Institutes of Health). The relative abundance of the fluorescence protein-labeled protein at the PM in each cell was assessed by calculating the ratio of fluorescence intensity at the PM to that at the cytosol (=FPM/Fcytosol) as described (23). The fold-change of fluorescence intensity of a protein at the PM caused by each treatment was quantified as (F*PM/F*cytosol)/(FPM/Fcytosol) in which FPM/Fcytosol and F*PM/F*cytosol indicate the fluorescence ratio before and after each treatment, respectively. Fluorescence fold-change in cytosol (F*cytosol/Fcytosol) indicates the degree of translocation of each domain from the PM to cytosol upon PDGF treatment: i.e. a lower value means more PM translocation.
The protrudin FYVE domain (residue number 318–409) was expressed as an N-terminal His6 fusion protein in Escherichia coli BL21(DE3) pLysS (Novagen). Cells were inoculated at 37 °C in 500 ml of Luria broth containing 50 μg/ml of kanamycin until the A600 reached 0.6–0.8. The overexpression of recombinant FYVE domains was induced by adding isopropyl β-d-thiogalactopyranoside to a final concentration of 1 mm. Then the cells were grown for 16 h at room temperature (16 °C) and harvested by centrifugation at 4 °C. The recombinant proteins were purified with affinity chromatography using nickel-nitrilotriacetic acid-agarose beads (Qiagen). Briefly, the cell pellets were resuspended in 20 ml of lysis buffer (50 mm Tris, 300 mm NaCl, 10 mm immidazole, 10% (v/v) glycerol, 50 μm phenylmethylsulfonyl fluoride, 5 mm 2-mercaptoethanol, pH 7.9), and the solutions were sonicated on an ice bath for 5 min (15 s of sonication followed by 15 s of cooling on ice). The cell lysates were centrifuged at 61,000 × g for 30 min at 4 °C, the supernatants were collected and mixed with 1 ml of 50% (v/v) nickel-nitrilotriacetic acid-agarose beads and incubated for 1 h at 4 °C with mild shaking at 60 rpm. The protein-bound beads were then loaded to a column and washed with an excess volume of washing buffer 1 (50 mm Tris, 300 mm NaCl, 20 mm immidazole, pH 7.9) followed by washing buffer 2 (20 mm Tris, 160 mm NaCl, 20 mm immidazole, pH 7.9). The FYVE domain proteins were then eluted with an elution buffer (20 mm Tris, 160 mm NaCl, 300 mm immidazole, pH 7.9). Finally, the buffer solution for the FYVE domains was exchanged to the SPR elution buffer (20 mm Tris-HCl, pH 7.4, 160 mm NaCl) using a gel filtration column, PD-10 (GE Healthcare), equilibrated with the same buffer. Protein purities were checked on a 20% polyacrylamide gel, and the protein concentrations were determined by a Bradford assay (Bio-Rad).
Large unilamellar vesicles were prepared using a Liposofast (Avestin) microextruder with a 100-nm polycarbonate filter. All SPR measurements were performed at 23 °C in 20 mm Tris-HCl, pH 7.4, containing 0.16 m NaCl using a lipid-coated L1 chip in the BIACORE T100 system. 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC)/1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoserine (POPS)/PtdInsP (77:20:3) vesicles and POPC vesicles were coated onto the active surface and the control surface, respectively. Vesicles were coated onto the corresponding sensor chip surfaces to yield the identical resonance units, ensuring the equal concentration of the coated lipids. Equilibrium SPR measurements were done at the flow rate of 5 μl/min to allow sufficient time for the response in the association phase to reach near-equilibrium values (Req) (35). A minimum of 5 different protein concentrations were injected to collect a set of Req values that were plotted against the protein concentrations (Po). An apparent dissociation constant (Kd) was then determined by nonlinear least squares analysis of the binding isotherm using the following equation, Req = Rmax/(1 + Kd/Po), where Rmax indicates the maximal Req value (36). Because the concentration of lipids coated on the sensor chip cannot be accurately determined, an apparent dissociation constant Kd is defined as the protein concentration yielding half-maximal binding with a given lipid concentration. The measurement was repeated at least three times to determine average and S.D. values.
The molecular modeling was performed using the program CCP4 molecular graphics (37) by superposing the protrudin FYVE structure (PDB code 1X4U) onto the structure of 1,3-bisphosphate bound human EEA1-FYVE domain (PDB code 1JOC). The root mean square deviation score was calculated using the same program. For modeling the protrudin-FYVE domain structure in a PtdIns(4,5)P2-bound form, inositol 1,4,5-triphosphate was incorporated to the lipid binding pocket of the apo-form structure in a similar manner with 1,3-bisphosphate bound to the human EEA1-FYVE domain.
We generated a library of 94 LBDs tagged with EYFP (Fig. 1A) and then imaged NIH3T3 cells expressing each EYFP-LBD. 94 LBDs were divided into five groups based on the subcellular localization pattern (Fig. 1B). Our results revealed that all endosome-localized LBDs were FYVE and PX domains, consistent with previous studies (31, 38–41) (Fig. 1E). They include the FYVE domains of HGS, FYCO1, WDFY1 (also known as FENS-1), and EEA1, and the PX domain of KIF6B. Among six LBDs showing PM localization (Fig. 1, C and D), five of them belonged to the ENTH and Tubby domains (Fig. 1, C and E), which are known to interact with PtdIns(4,5)P2 (42–44). Intriguingly, the other PM-localized LBD was the FYVE domain of protrudin (Fig. 1, C and E), which was unexpected in light of the other FYVE domains that display endosomal localizations (45). Therefore, we further examined lipid binding of the FYVE domain of protrudin. It should be noted that a significant degree of nuclear localization of the protrudin-FYVE domain was seen in our study. Similar nuclear localization has been reported for many other isolated LBDs, including PH domains of Gab1, Gab2, and Grp1 (30, 46–48), which has been attributed to smaller sizes of isolated LBDs, potential exposure of a cryptic nuclear localization sequence, or nonspecific binding to nuclear lipids or nucleic acids.
To validate our cell imaging data, we quantitatively measured lipid binding of the protrudin-FYVE domain by equilibrium SPR analysis. Specifically, we measured the PtdInsP selectivity of the protrudin-FYVE domain by determining its Kd values for POPC/POPS/PtdInsP (77:20:3) vesicles (see Fig. 2 for example). 30% PS was included to simulate the PS content of the inner layer of the PM of mammalian cells. As summarized in Table 1, the protrudin-FYVE domain has high affinity for vesicles containing PtdIns(3,4)P2, PtdIns(4,5)P2, and PtdIns(3,4,5)P3, respectively. However, it shows very low affinity for vesicles containing PtdIns(3)P, PtdIns(4)P, or PtdIns(5)P. The Kd values for binding of the protrudin-FYVE domain to vesicles containing PtdIns(3,4)P2, PtdIns(4,5)P2, and PtdIns(3,4,5)P3 range from 100 to 200 nm, which are comparable with that for PtdIns(3,4,5)P3-specific PH domains determined under similar conditions (30). These values also compare well with the Kd value for binding of other prototypical FYVE domains to vesicles containing PtdIns(3)P (49). Collectively, our SPR analysis shows that the protrudin-FYVE domain has high affinity for PtdIns(3,4)P2, PtdIns(4,5)P2, and PtdIns(3,4,5)P3.
Comparison of the amino acid sequence of the protrudin-FYVE domain with that of the prototypical EEA1-FYVE domain reveals unique features of the protrudin-FYVE domain (Fig. 3A). To identify which residues of the FYVE domain of protrudin are involved in lipid selectivity, three-dimensional structures of this domain and the inositol 1,3-bisphosphate-bound human EEA1-FYVE domain were superposed with a root mean square deviation of 0.68 Å over 34 Cα atoms. Based on the structural superposition, we modeled the lipid binding pocket of the protrudin-FYVE domain, which is larger than that of the EEA1-FYVE domain, accommodating an inositol 1,4,5-triphosphate molecule (a PtdIns(4,5)P2 headgroup) (Fig. 3B). The model suggests that PtdIns(4,5)P2 associates with the predominantly positive surface of the FYVE domain and that Lys-367, Arg-369, Arg-381, and Lys-386 might be the residues that directly interact with the phosphate groups of PtdIns(4,5)P2.
To determine whether these four cationic residues of the protrudin-FYVE domain control its PtdInsP selectivity, we mutated them to alanine, individually or in combination, expressed them in NIH3T3 cells, and monitored their subcellular localization. The levels of protrudin-FYVE domain at the PM were quantified by calculating the ratio of fluorescence intensity at the PM to that in the cytoplasm. In cells expressing the FYVE domain with single alanine mutations such as K367A, R381A, and K386A, the relative fluorescent intensity at the PM decreased 77.8, 57.2, and 45.4%, respectively, when compared with that of the wild type (WT) (Fig. 3, C and D). PM localization of a double-site mutant K367A/R369A was completely abolished, as was K367A/R369A/R381A/K386A (referred to as 4A hereafter). The R381A/K386A mutant remained at the PM to a similar extent as K367A (Fig. 3D). When measured by SPR analysis, K367A/R369A and 4A mutants showed dramatically lower affinity than WT for vesicles containing PtdIns(4,5)P2, PtdIns(3,4)P2, or PtdIns(3,4,5)P3 (Table 1). These results indicate that two cationic residues, Lys-367 and Arg-369, are critically involved in PtdInsP selectivity of the protrudin-FYVE domain for PtdIns(4,5)P2, PtdIns(3,4)P2, and PtdIns(3,4,5)P3.
To test the effect of the novel lipid-binding specificity of the FYVE domain on subcellular localization of protrudin, the above mutations were introduced to the full-length protrudin and the mutants were expressed in NIH3T3 cells. Although protrudin WT was localized weakly to the PM, vesicles, and the endoplasmic reticulum (ER), three mutants, i.e. K367A/R369A, 4A, and FYVE domain-deleted protrudin (ΔFYVE), with low to no affinity for PtdIns(4,5)P2, PtdIns(3,4)P2, and PtdIns(3,4,5)P3, were seen only in the ER (Fig. 3E).
We then modulated the level of PtdIns(3,4)P2 and PtdIns(3,4,5)P3 at the PM by PDGF and LY29 treatments, which activates and inhibits PI3K, respectively. Levels of the protrudin-FYVE domain at the PM were increased when accumulation of PtdIns(3,4,5)P3 and PtdIns(3,4)P2 was induced by PDGF treatment (Fig. 4A) (34). Interestingly, the protrudin-FYVE domain exhibited sustained translocation from the cytosol to PM (Fig. 4B). It was previously reported that the Akt-PH domain interacting with both PtdIns(3,4)P2 and PtdIns(3,4,5)P3 showed similar sustained PM localization, whereas the PtdIns(3,4,5)P3-specific Btk-PH domain exhibited transient translocation to the PM (Fig. 4B) (30). Thus, the PM recruitment pattern of the protrudin-FYVE domain seems to be consistent with its PtdInsP specificity. As was the case with the isolated FYVE domain, full-length protrudin was translocated to the PM upon PDGF-triggered PtdIns(3,4)P2 and PtdIns(3,4,5)P3 increases at the PM (Fig. 4, C and E, and supplemental Movie 1). Treatment of NIH3T3 cells with LY29 reduced the PtdIns(3,4)P2 and PtdIns(3,4,5)P3 levels at the PM, as evidenced by displacement of the Akt-PH domain from the PM (Fig. 5A) but had no effect on the PM PtdIns(4,5)P2 level as seen by retention of the PLCδ-PH domain (Fig. 5B). Fig. 5C shows that LY29 treatment caused protrudin to localize at endosomes (see also supplemental Movie 1). Also, when cells were preincubated with LY29, protrudin did not show PM localization even after PDGF treatment (Fig. 5D). Treatment with LY303511 (LY30), an inactive structural analog of LY29, had no effect in NIH3T3 cells expressing the protrudin-FYVE domain (Fig. 5B). The LY29 treatment, which inhibits production of PtdIns(3)P as well as PtdIns(3,4)P2 and PtdIns(3,4,5)P3, did not interfere with formation of endocytic vesicles carrying protrudin (Fig. 5, C and D), suggesting that these vesicles contain primarily PtdIns(4,5)P2 instead of PtdIns(3)P. Furthermore, protrudin (ΔFYVE) remained in the ER and did not show any response to PDGF treatment (Fig. 4, D and E). Collectively, these results suggest that PtdIns(4,5)P2 and PtdIns(3,4)P2/PtdIns(3,4,5)P3 play differential roles in controlling protrudin trafficking through interaction with the FYVE domain: i.e. PtdIns(4,5)P2 is important for vesicular transport of protrudin, and PtdIns(3,4)P2 and PtdIns(3,4,5)P3 play a more direct role in the PM recruitment of protrudin.
ARF6 activates PtdIns(4)P 5-kinase (PIP5K) inducing local synthesis of PtdIns(4,5)P2 and regulates membrane trafficking between PM and endosomes (32, 50, 51). Rab5B induces PtdIns(3)P formation and vesicle trafficking through recruitment of class III PI3K (33). Thus, one would expect that overexpression of a constitutively active form of ARF6 (i.e. Q67L) and Rab5B (i.e. Q79L) will increase the local concentration of PtdIns(4,5)P2 and PtdIns(3)P, respectively, which can be monitored by a PtdIns(4,5)P2 probe (i.e. PLCδ-PH domain) or a PtdIns(3)P probe (i.e. tandem FENS1-FYVE domain (2× FYVE)) (Fig. 5, E and F). We found that, in cells expressing either ARF6(Q67L) or Rab5B(Q79L), the protrudin-FYVE domain was localized in PtdIns(4,5)P2-enriched vesicles (Fig. 5E), but not in PtdIns(3)P-accumulated vesicles (Fig. 5F), which is consistent with its PtdInsP specificity. As was the case of the isolated FYVE domain, vesicle-bound protrudin colocalized with PtdIns(4,5)P2-enriched endocytic vesicles resulting from PIP(5)K activation (Fig. 5G) and these protrudin-containing vesicles disappeared in the presence of a dominant-negative mutant of ARF6 (i.e. T27N) (Fig. 5H). These results suggest that vesicular localization of protrudin is dependent on the PtdIns(4,5)P2 binding activity of the FYVE domain and that PtdIns(4,5)P2-mediated endocytosis regulates endosomal trafficking of protrudin.
Previous studies suggest that spastin and KIF5, protrudin-interacting proteins, induce neurite formation through microtubule-related membrane trafficking (6, 52). To demonstrate the physiological significance of novel PtdInsP selectivity of the FYVE domain of protrudin, we measured the neurite length of primary cultured hippocampal neurons expressing either protrudin (WT) or various protrudin mutants 36 h post-transfection. As expected, protrudin (WT) promoted 2.56- and 2.26-fold longer neurite outgrowth than the control when the total and maximum length of neurite was measured, respectively (Fig. 6, A and B). However, expression of protrudin harboring K367A/R369A or 4A mutations, or the ΔFYVE mutant had no significant effect on promotion of neurite outgrowth (Fig. 6, A and B). Overall, these data suggest that interaction of the FYVE domain of protrudin with PtdIns(4,5)P2, PtdIns(3,4)P2, and PtdIns(3,4,5)P3 is essential for the function of protrudin in promoting neurite outgrowth.
Many signaling proteins are regulated by complex interactions between their modular domains and numerous factors. LBDs play a key role in precise spatiotemporal regulation of proteins through their interaction with various lipids. To study roles of key signaling lipids, PtdIns(4,5)P2, and PtdIns(3,4,5)P3 in the regulation of cellular proteins in a simple, comprehensive, and explicit manner, we generated a library of distinct LBDs obtained from diverse cellular proteins and determined their subcellular localization. Our study showed that most FYVE and PX domains are endosome-localized, whereas ENTH and Tubby domains are PM-associated, as expected from their reported lipid specificity (42–44). The study also revealed that the FYVE domain of protrudin is localized to PM and the nucleus instead of endosomes, unlike typical FYVE domains (45).
The FYVE domains are known to show endosomal localization and interact with PtdIns(3)P (10–12). Unlike typical FYVE domains, the FYVE domain of protrudin did not show endosomal localization but was instead found at the PM. It was recently reported that the protrudin-FYVE domain binds to PtdIns(3)P when measured by a lipid blot assay (52). Our quantitative SPR analysis demonstrate, however, that the protrudin-FYVE domain has high affinity for PtdIns(4,5)P2 and two main products of the class I PI3K (34), PtdIns(3,4)P2 and PtdIns(3,4,5)P3, but not for PtdIns(3)P. Comparison of the amino acid sequences of the protrudin-FYVE domain and the prototypical EEA1-FYVE domain reveals that the protrudin-FYVE domain has unique structural variations in lipid binding motifs. Furthermore, molecular modeling suggests that a larger lipid binding pocket of the protrudin-FYVE domain may allow interaction with bulkier multi-phosphorylated PtdInsPs. Our modeling as well as mutational analysis indicates that Lys-367 and Arg-369 are most critically involved in specific recognition of 3-, 4-, and/or 5-phosphate groups of PtdIns(3,4)P2, PtdIns(4,5)P2, and PtdIns(3,4,5)P3.
The novel PtdInsP selectivity of the protrudin-FYVE domain is consistent with the observed subcellular localization pattern of the domain and the intact protrudin. PDGF triggered PM localization of protrudin, implying that PI3K activation is involved in dynamic PM recruitment of protrudin. Also, reduction of the PtdIns(3,4)P2 and PtdIns(3,4,5)P3 levels at PM by LY29, which inhibits production of 3-phosphorylated PtdInsPs, caused protrudin to localize at endosomes. Thus, PtdIns(4,5)P2 binding activity may be more important for vesicular trafficking of protrudin than its PM recruitment. ARF6 regulates membrane trafficking between PM and endosomes by activating PIP5K that locally synthesizes PtdIns(4,5)P2 from PtdIns(4)P (51). We observed that the PtdIns(4,5)P2-rich endocytic vesicles formed by a constitutively active ARF6 were colocalized with protrudin-containing vesicles, and coexpression of a dominant-negative mutant of ARF6 blocked localization of protrudin to the vesicles. This suggests that PIP5K-mediated PtdIns(4,5)P2 synthesis and endocytosis carry protrudin to endosomes for recycling. The fact that the protrudin mutant lacking the FYVE domain is localized at the ER indicates that ER localization of protrudin is driven not by lipid binding but by protein binding. In our model, ER-resident or vesicle-bound protrudin is recruited to the PM in response to local synthesis of PtdIns(3,4)P2 and PtdIns(3,4,5)P3 through PI3K activation and recycled to endosomes via PtdIns(4,5)P2-enriched vesicles when PtdIns(3,4)P2 and PtdIns(3,4,5)P3 levels decrease. By this mechanism, PtdIns(4,5)P2 and PtdIns(3,4)P2/PtdIns(3,4,5)P3 differentially modulate the endosome to PM trafficking of protrudin (Fig. 6C).
Involvement of protrudin in an AD-HSP has been suggested (6, 8, 9). The large group of HSP-related proteins is related to membrane trafficking. Previous studies on protrudin showed that protrudin plays a role in promoting neurite outgrowth through dynamic movement from the ER to the PM (7, 8). It is also involved in delivery of neuronal cargoes by interacting with KIF5, a molecular motor protein, suggesting that microtubules can be involved in protrudin-regulated vesicle trafficking. Our present study provides new insight into how its FYVE domain with novel PtdInsP selectivity contributes to the cellular function and regulation of protrudin. That is, differential regulation of endosome to PM trafficking of protrudin by binding of PtdIns(4,5)P2 and PtdIns(3,4)P2/PtdIns(3,4,5)P3 to its FYVE domain is crucial for its function in facilitating transport of neuronal cargo proteins. Undoubtedly, further studies are necessary to fully understand the mechanism underlying this complex process. Nevertheless, our discovery of the unique PtdInsP specificity of the FYVE domain of protrudin should help better understand how protrudin is linked to the onset of neurodegenerative disease such as AD-HSP via lipid signaling.
We thank Dr. Keiichi I. Nakayama and Dr. Takanari Inoue for kindly providing construct and localization marker for ER, respectively. We are also thankful to Min Jee Jang and Dr. Yoonkey Nam for helping isolate hippocampus from rat brain and to all of my lab members for helpful discussion.
*This work was supported by Korea Healthcare Technology R&D Project, Ministry of Health & Welfare, Republic of Korea Grant A101204.
This article contains supplemental Table S1 and Movie S1.
3The abbreviations used are: