PAK Kinases Target Sortilin and Modulate Its Sorting

The multifunctional type 1 receptor sortilin is involved in endocytosis and intracellular transport of ligands. The short intracellular domain of sortilin binds several cytoplasmic adaptor proteins (e.g., the AP-1 complex and GGA1 to -3), most of which target two well-defined motifs: a C-terminal acidic cluster dileucine motif and a YXXΦ motif in the proximal third of the domain. Both motifs contribute to endocytosis as well as Golgi-endosome trafficking of sortilin.

Trafficking of sortilin is controlled by interactions between its C-terminal tail and cytoplasmic adaptor proteins. Its cytoplasmic domain is short (53 amino acids) but contains binding sites for several adaptors, including adaptor protein complex 1 (AP-1) and AP-2; Golgi apparatus-localized, gamma ear-containing, ARF-binding proteins (GGAs) (GGA1 to -3 [GGA1-3]); elements of the retromer complex; Ras-related protein (Rab7b); and phosphofurin acidic cluster-sorting protein 1 (PACS-1) (20)(21)(22)(25)(26)(27). In regard to sortilin trafficking, the role of PACS-1 (and, to some extent, that of the GGAs) is still unclear, but several studies have established that AP-2 is involved in the endocytosis of sortilin, whereas the remaining adaptors participate in its Golgiendosome cycling (21,26,27). A single functional site (F 787 LV 789 ), possibly part of an as-yet-unidentified bipartite retromer-binding site, for retromer binding has been reported (26,28), but two of the most important sorting sites in sortilin are particularly well defined. The first, which contributes to Golgi-endosome trafficking (and more modestly to endocytosis) and is made up of a C-terminal acidic cluster combined with a dileucine (D 823 DSDEDLL 830 ), is targeted by PACS-1, AP-1 and -2, as well as the GGA proteins (20,21). The sequence of this motif also harbors a casein kinase phosphorylation site (Ser 825 ), and phosphorylation is known to modulate adaptor binding (29,30). The second site (Y 792 SVL 795 ) is a classical tyrosine-based motif (AP-1 and -2 binding) positioned in the upper one-third of the cytoplasmic domain. This motif constitutes the main site for endocytosis (AP-2) but is also of major importance for Golgi-endosome transport (AP-1) (21). Notably, the tyrosine-based motif also contains a serine residue, i.e., a potential site for phosphorylation and modulation of adaptor protein binding. In the present study, we have identified this motif as part of a site for binding of p21-activated kinases 1 to 3 (PAK1-3) and examined the functional implications of this interaction.
The PAK family consists of six members that are all expressed in brain but otherwise differ in terms of tissue expression patterns. Based on sequence and molecular function, the family is divided into two groups: group A, consisting of PAK1-3, and group B, including PAK4-6. The group A PAKs show extensive sequence similarity, share several conserved motifs, and have similar molecular functions (31,32). The monomers of each group A member contain an N-terminal GTPase-binding domain with a Cdc42/Racinteractive binding (CRIB) motif partly overlapped by an autoinhibitory domain (AID) and a separate kinase domain in the C-terminal half of the molecule (Fig. 1A). Inactive kinases are homodimers in which the monomers are positioned "head to toe" and stabilized by the AIDs, which bind and inhibit the kinase domain of the opposing monomer. Following binding of a Rho GTPase (Cdc42/Rac), this interaction is disrupted, the dimer dissociates, and the free monomers are subsequently converted to active kinases by autophosphorylation.
The activated PAKs exhibit a diversity of functions reflecting their role as Cdc42 and Rac effectors. In general terms, they influence the growth, shape, and motility of neuronal as well as nonneuronal cells (for a review, see reference 33). More specifically, their activity affects the actin cytoskeleton, microtubule dynamics, membrane trafficking, neuronal connectivity, axon guidance, dendritic spine formation, and synaptic plasticity, i.e., functions that may affect or implicate Vps10p-D receptors, including sortilin.
Here, we show for the first time that PAK1-3 have a direct impact on sortilin and that each of the three group A kinases binds and phosphorylates a serine residue positioned in a sorting motif within the sortilin cytoplasmic domain. We further demonstrate that phosphorylation of sortilin reduces its adaptor protein affinity and alters its intracellular trafficking. Our findings provide new insight into the mechanisms that regulate sortilinmediated sorting and add a new aspect to the functional repertoire of PAK1-3.

RESULTS
The p21-activated kinases PAK1-3 were identified as potential sortilin adaptor proteins in a yeast two-hybrid (Y2H) screen using a human brain library of fragmented cDNA and the human sortilin cytoplasmic domain (sortilin-cd) as bait. Several hits were obtained with clones expressing any one of the three kinases. The cDNA expressed by the positive clones differed but represented overlapping fragments of the respective PAK kinase domains (Fig. 1A). The overlapping segments all comprised a 121-aminoacid-residue segment with almost complete (91 to 99%) sequence identity (residues I 298 to Q 418 of human PAK1; GenBank accession number NP_002567.3), indicating that this region is responsible for the sortilin-cd interaction. The interaction was subsequently confirmed by a positive response in Y2H tests using the sortilin-cd and the original library plasmids purchased from Hybrigenics encompassing the C-terminal kinase domains of PAK1-3 containing this 121-amino-acid-residue segment (Fig. 1B). Negative-control plasmids were included to verify that the PAK plasmids do not exhibit autoreactivity.
The kinase domains of PAK1-3 target the sortilin-cd. To establish the PAK-sortilin interaction at the protein level, a PAK1 fragment (Y 270 -H 545 ) containing the entire kinase domain segment (Fig. 1A) was expressed in Escherichia coli as an N-terminally glutathione S-transferase (GST)-tagged protein. The resulting fusion protein was purified, immobilized on glutathione-Sepharose, and tested for its ability to mediate the pulldown of full-length sortilin and chimeric receptors carrying the sortilin-cd from cell and tissue lysates (Fig. 1C). Pulldown experiments using Sepharose resin coated with GST only (negative control) or GST-GGA2 (positive control) were performed in parallel (34). As apparent from Fig. 1C, Western blot (WB) analysis of the respective precipitates demonstrated that the GST-PAK1 kinase domain conveyed efficient pulldown (comparable to that of GST-GGA2) of full-length sortilin from CHO cell transfectants as well as of endogenous sortilin from human brain lysates. Moreover, the fact that chimeric receptors, combining the sortilin-cd with alternative luminal receptor domains (IL-2R [interleukin-2 receptor]-sortilin-cd and MPR300 [cationindependent mannose-6-phosphate receptor]-sortilin-cd) (21), were precipitated with similar efficiencies demonstrates that the cytoplasmic, and not the luminal, domain is responsible for the interaction with PAK1.
The affinity of PAKs for sortilin is similar to that of established adaptor proteins. To explore the kinetics of the interaction, we next examined sortilin-kinase binding by isothermal titration calorimetry (ITC). The complete cytoplasmic domain of sortilin was expressed as a GST-tagged protein and purified prior to the removal of the GST tag by enzymatic cleavage with thrombin. Titration of 25 M GST-PAK with 100 M sortilin-cd was performed in a buffer containing 100 mM NaCl and 50 mM Tris-HCl at pH 7.4. The analysis showed a dissociation constant (K d ) of ϳ2 M (Fig. 2), reflecting an affinity similar to that of the interaction between sortilin and GGA2 (30). Corresponding ITC analyses of the interactions between PAK1 and five partly overlapping peptides (peptides A to E, each comprising 12 to 15 amino acid residues) covering the entire sequence of the sortilin-cd further indicated that the responsible binding site(s) was located in the N-terminal part of the sortilin-cd (Fig. 2). In agreement with this conclusion, a Y2H test of the sortilin-cd and three C-terminally truncated cytoplasmic domain constructs demonstrated that the binding site(s) for PAK1 resides within the first 27 residues of the cytoplasmic domain and the removal of the last (C-terminal) 26 amino acid residues have little or no effect on the interaction (Fig. 3A).
The PAK-binding segment of sortilin includes a functional sorting motif. Based on these data, we next set up a Y2H screen of wild-type (wt) and mutant sortilin-cd constructs in order to identify the residues involved in PAK binding. Positive PAK cDNA clones (the translated products are shown in Fig. 1A) were tested against a series of sortilin mutants in which two or three consecutive residues were replaced with alanine. Taken together, the mutations cover the sequence from R 786 to T 812 (residues 8 to 34 of the sortilin-cd), i.e., close to the collected sequences of the binding-active sortilin-cd peptides A, B, and C ( Fig. 2) (below, the residues are simply designated residues 1 to 53 according to their position in the cytoplasmic domain). As displayed in Fig. 3B, the outcome of the alanine-scanning experiment identified the cytoplasmic domain sequence R 8 FLV 11 as a key determinant in the binding of all three PAKs and further suggested an additional and significant contribution by Y 14 SVL 17 and H 20 AE 22 . In other words, the residues involved in the binding of the PAK A kinase domains encompass a (B) Alanine scanning for PAK interaction sites in sortilin by Y2H analysis using a series of sortilin mutants in which two to three consecutive residues were replaced with alanine residues. Taken together, the mutations cover residues 8 to 34 of the sortilin-cd, as indicated. Amino acid residues marked in red affect binding in Y2H assays. (C) Y2H analysis investigating the effect on PAK1-3 binding when Ser 15 in the sortilin-cd is replaced with either Ala (S15A) or Asp (S15D) (mimicking phosphorylation). His, histidine; 3-AT, 3-amino-1,2,4-triazole. key adaptor-binding site (YSVL) and, notably, a sequence that complies with the basic requirements for phosphorylation of Ser 15 by PAK.
Interestingly, an additional Y2H screen ( Fig. 3C) revealed that whereas the replacement of Ser 15 with Ala had no effect on PAK1-3 binding, an Asp-for-Ser substitution (imitating phosphorylation) caused significantly weaker signals (in particular with regard to PAK2). Thus, the sortilin tail appears to lose its affinity for all three PAKs upon the phosphorylation of Ser 15 . The Y2H analysis further reveals that peptides A, B, and C ( Fig. 2) all comprise parts of the, but not the complete, binding sequence, which explains their lower affinities for PAK than the full-length cytoplasmic domain and furthermore why the nonoverlapping peptides A and C independently confer binding to PAK.
PAKs phosphorylate the sortilin-cd. It is well known that Ser 47 in the C-terminal GGA-binding sequence of sortilin is phosphorylated in vivo by casein kinase and that this serves to modulate adaptor binding and sorting of the receptor. To confirm previous findings and to pursue the possibility of alternative phosphorylation sites in the sortilin-cd, we therefore performed 32 P labeling of CHO and SY5Y cells stably transfected with wt sortilin or, as a negative control, a sortilin mutant lacking the cytoplasmic domain (sortilin-delta-cd). The cells were incubated with carrier-free 32 Plabeled o-phosphate in phosphate-free Dulbecco's modified Eagle's medium (DMEM) for 4 h prior to 1 h of incubation in the presence of the phosphatase inhibitor calyculin. The cells were then lysed in a 1% Triton X-100 buffer supplemented with phosphatase and proteinase inhibitors, and sortilin was immunoprecipitated from the lysates using antibodies directed against the receptor ectodomain. Analysis by SDS-PAGE and autoradiography established that sortilin was indeed phosphorylated and that the specific site of phosphorylation was localized to the cytoplasmic domain (Fig. 4A). Similar To clarify whether Ser 47 was responsible for all the observed labeling or if other sites contributed, we next compared 32 P labeling of sortilin in SY5Y cells transfected with either wt sortilin or a mutant receptor in which Ser 47 was replaced with an Asp residue. Cell labeling and analysis of immunoprecipitated proteins were performed as described above, and the two transfectants were examined in parallel (Fig. 4B). The results showed that the phosphorylation of sortilin persisted despite the absence of Ser 47 , and based on a comparison between the degrees of labeling in the two constructs (radioactivity relative to the amount of protein precipitated), alternative sites appear to be responsible for about 50% of the observed phosphorylation of wt sortilin.
PAKs specifically phosphorylate Ser 15 of the sortilin-cd and alter its trafficking. Mass spectrometry (MS) analysis of phosphopeptides recovered from a tryptic digest of the in vitro-phosphorylated sortilin-cd demonstrated the presence of a single phosphorylation site encompassed in the Y 14 -K 40 tryptic fragment, which includes the potential PAK target (RY)S 15 . Our data suggest that the cytoplasmic domain of sortilin indeed encompasses a PAK-specific site of phosphorylation, but the tryptic peptide contains three additional Ser or Thr amino acid residues, and we were not able to assign the specific site of phosphorylation by mass spectrometry. However, in vivo phosphorylation of Ser 15 was recently reported (35), and to determine if the sortilin-cd, and Ser 15 in particular, may in fact be subject to PAK kinase activity, we tested PAK's capacity for in vitro phosphorylation of the wt sortilin-cd and selected mutant constructs of the cytoplasmic domain (Fig. 5). The receptor constructs were generated with an containing only a single potential Ser phosphorylation site, and two single point mutants of Ser15 in the sortilin-cd, sortilin-cd S15A and sortilin-cd S15D. The in vitro phosphorylation assay was performed using 32 P-labeled ATP, full-length GST-PAK2 T402E, and the purified fusion protein or GST as a negative control. The phosphorylation reaction was stopped after 30 min and analyzed by SDS-PAGE and subsequently autoradiography. (B) Evaluation of all isolated GST-tagged fusion proteins used for the in vitro phosphorylation assay analyzed by SDS-PAGE and Coomassie brilliant blue staining. Amino acid sequences of the sortilin cytoplasmic domain are shown, with the potential Ser and Thr phosphorylation sites underlined. Amino acid residues mutated in sortilin-cd S15D/A are marked in red.
N-terminal GST tag, expressed in E. coli, and purified from culture medium by glutathione affinity chromatography. Each of the purified sortilin-cd variants was then incubated with constitutively active PAK2 (T402E) in the presence of [␥-32 P]ATP. After 30 min, the reaction was stopped, and the samples were examined by SDS-PAGE and autoradiography. As apparent from Fig. 5, wt sortilin and a deletion mutant comprising only Ser 15 (sortilin-cd 1-32) was readily phosphorylated, whereas mutants carrying a single substitution at position 15 (Asp or Ala for Ser) showed no trace of radioactivity.
It can be concluded that PAK1-3 may indeed instigate the phosphorylation of sortilin and that they target a single serine residue (Ser 15 ) located in the kinase domain-binding site of the sortilin-cd.
Phosphorylation of Ser 15 in the sortilin-cd alters its trafficking. Given that Ser 15 is part of the short signal sequence which governs endocytosis and, to a large extent, also Golgi-endosome sorting of sortilin, we decided to examine if the phosphorylation of Ser 15 impacts sortilin trafficking. To this end, we first compared the internalization of wt sortilin to that of a mutant receptor in which Ser 15 had been replaced with an Asp residue to mimic permanent phosphorylation. HEK cells expressing either the wt or the mutant receptor were incubated for 1 h with antisortilin antibodies at 4°C and then washed in unsupplemented warm medium. After a maximum of 40 min at 37°C, the cells were fixed and stained with a fluorescent secondary antibody, and the degree of receptor endocytosis was determined by fluorescence microscopy. The results (Fig. 6A) revealed the almost complete internalization of both receptors, signifying little or no difference between their capacities for endocytosis.
However, the detection (immunofluorescence) of the two receptors at steady state ( Fig. 6B) suggested that mutant sortilin (sortilin S15D) was more widely distributed in the cells than wt sortilin, which was mainly concentrated in paranuclear compartments. Also, sortilin S15D-associated compartments included large vesicular structures that were not seen upon staining for wt sortilin.
The overall cellular localization of the receptors (reflecting trafficking) was therefore next determined by subcellular fractionation. Cells expressing either wild-type or mutant (S15D) sortilin were examined in parallel, and as apparent by subsequent WB of the fractionated samples, the distributions of the two receptors clearly differed (Fig.  7). Similar results were obtained in two separate but identical experiments.
Despite the difference in localization/trafficking, the sortilin S15D mutation did not appear to affect the turnover of sortilin. Thus, in pulse-chase experiments with biolabeled receptors, wild-type sortilin and the mutant receptor displayed similar half-lives (not shown).
Phosphorylation (Ser 15 ) downregulates the sortilin-AP-1 interaction. Since the S15D mutation did not significantly affect the endocytosis or turnover of sortilin, we deduce that the observed difference in subcellular distribution must result from a change in intracellular trafficking induced by the point mutation mimicking phosphorylation. As mentioned above, the Y 14 XXL 17 motif partakes in both endocytosis and the Golgi-endosome transport of sortilin. Tyr 14 plays a dominant role in endocytosis but does not affect the AP-1-dependent Golgi-endosome transport of chimeric receptors carrying the luminal domain of MPR300 (cation-independent mannose-6-phosphate receptor) in combination with the sortilin cytoplasmic domain. In contrast, substitution of Ala for both Tyr 14 and Leu 17 significantly reduced the MPR300/sortilin chimeramediated sorting of newly synthesized lysosomal enzymes, suggesting that Val 16 Leu 17 is the key residue(s) in the latter context (21). We therefore performed a Y2H analysis to test if the phosphorylation of Ser 15 could alter the function of the dileucine-like (Val 16 -Leu 17 ) motif, i.e., its interaction with the 1A and 1B subunits of AP-1. The outcome is depicted in Fig. 8, which clearly demonstrates that compared to wt sortilin, the S15D mutant exhibits a distinct reduction in its ability to interact with each of the two 1 subunits. However, the capacity for 1 subunit interactions of another mutant, S15A, appeared to be equal to that of the wt sortilin construct, signifying that the reduced interaction between the AP-1 subunits and the S15D mutant owed itself to the introduction of a negative charge, i.e., phosphorylation or, in this case, Asp.
We conclude that the PAK1-3 kinases target and phosphorylate Ser 15 of the sortilin-cd and thereby alter the intracellular trafficking of the receptor predominantly by reducing its affinity for AP-1, which mediates Golgi-endosome sorting of wild-type sortilin.

DISCUSSION
Sortilin is a multifunctional sorting receptor that binds a variety of extracellular ligands, including both soluble and transmembrane proteins. It partakes in the endocytosis of ligands, but the major pool of the receptor is intracellular and engaged in trafficking between different compartments, notably between the Golgi apparatus and endosomes (21,22). Accordingly, sortilin interacts with several cytoplasmic adaptor proteins via specific motifs in its short cytoplasmic domain.
The present study was undertaken in an attempt to identify new binding partners FIG 6 Cellular internalization and localization of the sortilin S15D mutant mimicking constitutive PAK phosphorylation. (A) Cellular internalization. HEK cells stably expressing the wild-type or an S15D mutated sortilin receptor were incubated for 1 h on ice with monoclonal mouse antisortilin clone F11 antibodies. Cells were washed; incubated at 37°C for 0, 5, or 10 min; and finally fixed in 4% paraformaldehyde and permeabilized. Following fixation, cells were stained with Alexa Fluor 488-conjugated donkey anti-mouse antibody (green), and the nuclei were visualized by using Hoechst stain (blue). (B) Immunocytochemistry of sortilin in transfected HEK cells showing the localization of wt sortilin and sortilin S15D. Cells were stained with monoclonal mouse antisortilin clone F11 antibody (green), and nuclei were visualized with Hoechst stain (blue).
for the sortilin-cd, and here, we report its interaction with the p21-activated kinases PAK1-3 and the functional implications of the interaction. A novel motif in the sortilin-cd. The established adaptor-binding motifs of sortilin are localized in the extreme C terminus of the cytoplasmic domain and in the upper one-third of its juxtamembrane (Fig. 9). The C-terminal site comprises an acidic cluster combined with a dileucine motif (AC-dileucine) and constitutes a target for AP-1 complexes, GGA1-3, and PACS-1 (20,21). The role of PACS-1 is still unresolved, but AP-1, and most likely GGAs, is involved in Golgi-endosome transport (21,29). Moreover, the dileucine contributes to endocytosis and interacts with AP-2 complexes. The juxtamembrane segment harbors at least two separate sites, i.e., a short segment (F 9 LV 11 ), which may serve as a binding site for the retromer complex, and a second segment (Y 14 SVL 17 ) fitting the canonical AP-1-and -2-binding motif YXX⌽ (where X is any residue and ⌽ is a hydrophobic residue) (21,26). Our findings now reveal that both motifs are part of yet another binding sequence forming a motif (R 8 FLV-[XX]-YSVL-[XX]-HAE 22 ) that mediates interaction with the kinase domain in each of the three PAKs. Analysis of the interaction showed that sortilin and PAK coprecipitate from cells and brain tissue, and ITC measurements of binding established a K d of about 2 M, signifying an affinity FIG 7 Effect of phosphorylation on the subcellular localization of sortilin. Subcellular fractionation of HEK cells transfected with either wt sortilin or S15D mutated sortilin (HEK/sortilin and HEK/sortilin S15D cells, respectively) was performed by centrifugation in a discontinuous iodixanol density gradient. Twenty-four fractions were collected by ultracentrifugation and subsequently analyzed by Western blotting using mouse antisortilin (catalog number 612100; BD Biosciences). Bars indicate the distribution of fractions with the largest amount of sortilin.

FIG 8
Phosphorylation of sortilin inhibits its interaction with AP-1. Yeast two-hybrid (Y2H) analysis investigating the effect of phosphorylation of Ser 15 in the cytoplasmic domain of sortilin on the interaction with the 1A and 1B subunits of AP-1. Y2H analysis was performed by using the entire cytoplasmic domain of sortilin (sortilin-cd), sortilin S15D, or sortilin S15A against the 1A and 1B subunits of AP-1. Plasmids without inserts were included as negative controls. Growing colonies indicate interactions between expressed proteins. His, histidine; 3-AT, 3-amino-1,2,4-triazole.
comparable to that of GGA adaptors for sortilin and confirming its functional relevance (30).
The sortilin-cd is a PAK substrate. Interestingly, the YSVL sorting motif, which is part of the PAK-binding sequence, also harbors a serine that appeared as a potential site for phosphorylation by PAK. It is well known that the sortilin-cd is subjected to phosphorylation in vivo and that casein kinase 2 mediates the phosphorylation of Ser 47 (29,30). However, our MS analysis of in vivo 32 P-labeled sortilin demonstrates that only about 50% of the phosphorylation is accounted for by Ser 47 , and in vitro experiments with the purified sortilin-cd presented Ser 15 as a functional and specific site for PAK phosphorylation. These findings are in accordance with those of a recent study by Li et al., who reported the phosphorylation of both Ser 15 and Ser 47 in sortilin obtained from sortilin-expressing HepG2 cells as well as from mice overexpressing sortilin (35). In concert, these data establish that PAKs can phosphorylate Ser 15 40 ) and the resulting increase in sortilin degradation. In contrast, we find no significant changes between the half-life of S15D sortilin mutants and that of wild-type receptors. Instead, the present results obtained by cellular expression of S15D mutants and by Y2H analysis using a mutated sortilin-cd strongly indicate that Ser 15 phosphorylation alters the intracellular localization of sortilin due to a lowered affinity for AP-1 and a change in sorting. As mentioned above, Ser 15 is positioned in the YXX⌽ motif (Y 14 SVL 17 ), which serves in endocytosis as well as in Golgi-endosome trafficking. We previously demonstrated that both Tyr 14 (in particular) and Leu 17 are important for receptor endocytosis and account for about 80% of the total endocytic activity of sortilin (21). In contrast, a Y14A substitution had no apparent effect on Golgi-endosome transport, while double mutants (Y14A and L17A) presented an ϳ50% reduction, indicating that the YSVL motif might in fact be a combination of two functional motifs, i.e., Y 14 XXL 17 for endocytosis and dileucine-like V 16 L 17 for Golgi-endosome sorting (21). The present results seem to support this notion, as the S15D mutation (i.e., phosphorylation) affected intracellular sorting and AP-1 binding without significantly affecting endocytosis. Moreover, a previous study reporting a comparable half-life of sortilin in AP-1 knockout and wild-type cells supports our finding that the phosphorylation of Ser 15 results in a decreased affinity for AP-1 complexes and "missorting" but not in a shorter sortilin half-life (20). Recent findings suggest that retromer binding in yeast is conveyed by a bipartite site in receptor cytoplasmic domains (28). This could evidently also be the case for mammalian receptors such as sortilin. However, as the phosphorylation of Ser 15 did not increase sortilin turnover (degradation), as might be expected upon an insufficient retrieval of the receptor, it appears that phosphorylation does not affect the retromer-binding site(s).
It follows that each of the two major adaptor-binding sites in sortilin is targeted by kinases. Ser 47 in the C-terminal AC-dileucine motif is phosphorylated by casein kinase 2 and modulates GGA binding, and as shown here, PAK1-3 phosphorylate Ser 15 , thereby modulating AP-1 binding and the distribution of cellular sortilin. As PAKs are found in all tissues expressing sortilin and are activated via many pathways, it seems likely that they actively modulate sortilin trafficking under several conditions (36)(37)(38).
In summary, we have shown that sortilin interacts with the kinase domains of PAK1-3. The endocytic and Golgi-endosome sorting motif Y 14 SVL 17 is part of the PAK-binding sequence, and PAKs phosphorylate Ser 15 . Upon phosphorylation, sortilin loses its affinity for AP-1 complexes as well as for PAKs and alters its cellular localization and trafficking. PAKs have numerous targets, but as far as we know, sortilin is the first example of a PAK receptor substrate. Thus, the present findings present new functional aspects of both PAKs and sortilin.
GST pulldown of sortilin from transfected cells and brain homogenates. For pulldown experiments on human brain homogenates, 10-g frozen brain tissue samples were homogenized on ice in 10 ml lysis buffer (20 mM HEPES-KOH, 125 mM potassium acetate, 2.5 mM magnesium acetate, 320 mM sucrose, 0.1 mM EDTA [pH 7.6]) containing protease inhibitors (cOmplete mini; Roche). The homogenate was centrifuged (10 min at 1,000 ϫ g at 4°C), Triton X-100 was added to a final concentration of 1.25% (vol/vol), and the mixture was vortexed and rotated for 60 min at 4°C. Soluble protein fractions were generated by centrifugation for 15 min at 20,000 ϫ g at 4°C and subsequent ultracentrifugation for 60 min at 100,000 ϫ g at 4°C. For GST pulldowns from transfected cell lines, CHO cells stably expressing either full-length sortilin or an IL-2R-sortilin-cd chimera and HEK293 cells transfected with MPRsortilin-cd were lysed on ice in lysis buffer (150 mM NaCl, 2 mM MgCl 2 , 0.1 mM EGTA, 2 mM CaCl 2 , and 10 mM HEPES [pH 7.4] with cOmplete mini protease inhibitors and 1% [vol/vol] Triton X-100) and centrifuged for 10 min at 18,400 ϫ g at 4°C. A total of 500 l of the supernatant from the human brain homogenate was mixed with either 100 g of GST, GST-GGA2, or GST-PAK1, adjusted to a volume of 10 ml with lysis buffer (without Triton X-100), and incubated with rotation overnight at 4°C. Similarly, 100 l of the supernatant from cell lysates was mixed with 10 g GST, GST-GGA2, or GST-PAK1, adjusted to a volume of 1 ml in lysis buffer (without Triton X-100), and rotated overnight at 4°C. The following day, 100 l glutathione-Sepharose 4B beads was added, and samples were rotated for 4 h at 4°C. Beads were pelleted, washed four times for 5 min in lysis buffer (modified to 0.4 M NaCl, without Triton X-100), and finally eluted with 100 l SDS-PAGE sample buffer with dithioerythritol (DTE) and analyzed by WB as described above, using rabbit antisortilin (catalog number 5264), anti-MPR (2C2), or anti-IL-2R (anti-CD25; Roche).
Yeast two-hybrid screening. The human sortilin cytoplasmic domain was cloned into pB27, a Y2H vector optimized by Hybrigenics, and subsequently transformed into the L40GAL4 yeast strain (42). An adult human brain random-primed cDNA library, transformed into the Y187 yeast strain, was used for mating. Following selection on medium lacking leucine, tryptophan, and histidine (ϪLeu/ϪTrp/ϪHis), positive clones were picked, and the corresponding prey fragments were amplified by PCR and sequenced at their 5= and 3= junctions. To certify data from the primary screening, positive preys from the yeast-transformed cDNA library and bait plasmids were purchased from Hybrigenics for additional verification using control plasmids without inserts. Transformed cells were suspended in sterile doubledistilled water (ddH 2 O) and spot plated (5 l/spot) onto agar plates prepared from 46 g/liter yeast minimal SD agar base (Clontech) with 100 g/ml penicillin-streptomycin and supplemented with either 640 mg/liter ϪLeu/ϪTrp dropout supplement (Clontech) for control plates or 640 mg/liter ϪLeu/ϪTrp/ ϪHis dropout supplement (Clontech) for test plates, and 3-amino-1,2,4-triazole (3-AT; Sigma-Aldrich) was added to the plates to strengthen the conditions of the interaction.
Isothermal titration calorimetry. Binding of sortilin-cd and peptides A to E, each comprising 12 to 15 amino acid residues (CASLO) to GST-tagged PAK1 (Y 270 to H 545 ), was measured by isothermal titration calorimetry (ITC). The complete wt sortilin-cd was expressed as a GST-tagged peptide and purified prior to the removal of the GST tag by thrombin cleavage. The titration experiments were performed on a MicroCal VP-ITC isothermal titration calorimeter (MicroCal Inc.). GST-PAK1 for ITC was prepared as described above and dialyzed against 100 mM NaCl-50 mM Tris-HCl buffer (pH 7.5). Peptides were dissolved in the same buffer, and to ensure accurate measurements, concentrations of protein and peptides were determined by analysis of total amino acids. All solutions were filtered, degassed, and equilibrated to the corresponding temperature before each experiment. In a typical ITC experiment, 1.45 ml GST-PAK1 (25 M) was titrated at 34°C with peptides A to E (100 M) or the sortilin-cd (100 M) in 28 steps of 10 l. The time between injections was set to 150 s, and the syringe mixing speed was set at 300 rpm. Heat evolving from dilution was measured by injecting the ligand into Tris-HCl buffer. This heat of dilution was subtracted from the heat of the reaction to obtain the effective heat of binding. Finally, the equilibrium dissociation constant (K d ) for the binding processes was determined using ORIGIN software (OriginLab Corporation).
Phosphorylation assay. For 32 P labeling of hippocampal neurons, untransfected SY5Y and cells transfected with full-length sortilin, sortilin-delta-cd, or sortilin S47D were cultured in monolayers in poly-L-lysine-coated 6-well plates. At about 90% confluence, cells were washed 3 times with 5 ml phosphate-free DMEM (catalog number 11971025; Gibco) with Na-pyruvate (Invitrogen) added. Cells were then added to fresh phosphate-free DMEM supplemented with carrier-free 32 P-labeled orthophos-phate (0.4 mCi/ml), and cells were incubated at 37°C with 5% CO 2 for 4 h prior to an additional hour of incubation in the presence of 70 nM calyculin (Cell Signaling Technology). Cells were then washed with cold phosphate-buffered saline (PBS) buffer (HyClone Dulbecco PBS [DPBS]; Thermo Scientific) containing phosphatase inhibitors (PhosSTOP; Roche), and finally, the cells were lysed on ice with immunoprecipitation (IP) buffer (150 mM NaCl, 2 mM MgCl 2 , 0.1 mM EGTA, 2 mM CaCl 2 , and 10 mM HEPES [pH 7.4] with a cOmplete mini protease inhibitor cocktail [Roche] and PhosSTOP [Roche]) containing 1% Triton X-100. Lysates were collected from the wells, centrifuged at 18,000 ϫ g for 10 min at 4°C, and diluted four times in IP buffer. Sortilin was then immunoprecipitated for 3 h at 4°C from the supernatants using GammaBind G Sepharose beads (GE Healthcare) preincubated with polyclonal rabbit antisortilin antibodies (catalog number 5438; custom-made by Dako). Beads were pelleted, washed five times in IP buffer with 0.1% Triton X-100, and finally eluted with 40 l SDS-PAGE sample buffer with DTE. Precipitated proteins were separated by SDS-PAGE and electroblotted onto a PVDF membrane, and phosphorylation was analyzed by autoradiography using the FLA-3000 fluorescence image analyzer (Fujifilm). Precipitated proteins were identified by WB of the same membrane using mouse antisortilin (anti-NTR3, catalog number 612100; BD Biosciences). The degree of sortilin labeling (radioactivity relative to the amount of sortilin precipitated) was evaluated from WBs and autoradiographs using Multi Gauge v.3.2 software.
In vitro kinase assays. Phosphorylation assays were carried out using 6 M constitutively active GST-PAK2 (T402E) kinase in a 50-l volume containing 50 mM Tris-HCl (pH 7.5), 0.1 mM EGTA, 0.1% (vol/vol) 2-mercaptoethanol, 10 mM magnesium acetate, and 0.2 mM [␥-32 P]ATP (200 cpm/pmol). The following GST fusion proteins (7.5 M) were added as the substrate: GST, GST-sortilin-cd, GST-sortilin-cd 1-32, GST-sortilin-cd S793D, and GST-sortilin-cd S793A. The phosphorylation assays were carried out at 30°C for 30 min with mild shaking. Reactions were terminated by the addition of 20 l of 4ϫ NuPAGE LDS sample buffer (Novex) and 5 l DTE, followed by denaturation at 95°C for 5 min. Samples were separated by SDS-PAGE on NuPAGE 4 to 12% Bis-Tris protein gels, and gels were washed for 10 min in water and fixed according to the manufacturer's protocol. Phosphorylated products were visualized by autoradiography using the FLA-3000 fluorescence image analyzer. Subcellular fractionation. Subcellular fractionation of transfected HEK cells (HEK/Sortilin Fl and HEK/Sortilin S15D cells) by a discontinuous iodixanol density gradient was performed essentially as described previously by Chang et al. (44). In brief, transfected cells were harvested from two confluent T175 cell culture flasks by centrifugation for 10 min at 3,000 ϫ g at 4°C. Cells were homogenized in 3 ml ice-cold solution A (0.25 M sucrose, 1 mM EDTA, and 10 mM HEPES [pH 7.4] plus a cOmplete mini cocktail) and disrupted on ice by 6 passages through a 27-gauge needle, followed by 6 passes through a metal cell cracker with a 9-m gap. Nuclei and unbroken cells were removed by centrifugation at 1,500 ϫ g for 10 min. The postnucleated supernatant was centrifuged for 1 h at 65,000 ϫ g at 4°C. The resultant membrane pellets were resuspended in 0.8 ml of solution A. Discontinuous iodixanol density gradients were prepared starting from an OptiPrep gradient stock solution (catalog number D1556; Sigma) containing a final concentration of 50% OptiPrep in 250 mM sucrose, 1 mM EDTA, and 10 mM HEPES (pH 7.4). The resuspended vesicle fractions were loaded on top of the gradients and centrifuged in an SW41 rotor at 40,000 rpm for 2.5 h at 4°C. Twenty-four fractions of 0.5 ml were collected. Fractions (15 l) were subsequently analyzed by WB using NuPAGE 4 to 12% Bis-Tris gels, nitrocellulose iBlot transfer stacks (Invitrogen), and mouse antisortilin (catalog number 612100).
Immunocytochemistry. Cells grown on poly-L-lysine-coated coverslips were fixed for 15 min in 4% paraformaldehyde, washed 3 times with PBS (pH 7.4), and permeabilized using immunocytochemistry (ICC) buffer (0.25% [wt/vol] saponin, 10% [vol/vol] FBS, PBS [pH 7.4]) for 30 min at room temperature. Cells were then incubated in ICC buffer supplemented with primary monoclonal mouse antisortilin (clone F11; produced in-house) overnight at 4°C. Cells were washed 3 times and incubated with Alexa Fluor 488-conjugated donkey anti-mouse antibody (catalog number A21202; Invitrogen) in ICC buffer for 1 h at room temperature. Cells were washed 3 times in PBS, and nuclei were visualized with Hoechst stain (Sigma) and mounted using fluorescence mounting medium (catalog number S3023; Dako). Images were acquired with a Zeiss LSM710 confocal microscope.
Internalization. HEK cells expressing either the wild-type or S15D mutated sortilin receptor grown on poly-L-lysine-coated coverslips were incubated for 1 h on ice in DMEM with monoclonal mouse antisortilin clone F11 antibodies. Cells were washed and incubated with warm DMEM. After 0, 5, 10, 20, and 40 min at 37°C, cells were fixed in 4% paraformaldehyde and stained with Alexa Fluor 488conjugated donkey anti-mouse antibody (as described above for immunocytochemistry), and the degree of receptor endocytosis was evaluated by using a Zeiss LSM710 confocal microscope.
Mass spectrometry. The GST-tagged cytoplasmic domain of sortilin (sortilin-cd) (described above) was subjected to in vitro phosphorylation by using the constitutively active GST-PAK2 kinase (as described above for the in vitro kinase assay). Proteins were subsequently separated by SDS-PAGE and visualized by silver staining. Bands of interest were excised and prepared for in-gel digestion (45), and sequencing-grade porcine trypsin was added (Promega). To evaluate the presence of phosphopeptides in the digests, the material was processed using TiO 2 microcolumns (46). Peptides were eluted directly onto a stainless steel matrix-assisted laser desorption ionization (MALDI) target using 2,5dihydroxybenzoic acid as a matrix support and analyzed by using an Autoflex Smartbeam III instrument (Bruker) operated in linear and positive mode. Prior to analyses, the instrument was calibrated by external calibration using a peptide mix containing 7 calibrants (Bruker). The obtained data were evaluated by using GPMAW software.