JLP Associates with Kinesin Light Chain 1 through a Novel Leucine Zipper-like Domain*

Scaffolding proteins exist in eukaryotes to properly assemble signaling proteins into specific multimeric functional complexes. JLP is a novel leucine zipper protein belonging to a family of scaffolding proteins that assemble JNK signaling modules. JLP is a proline-rich protein that contains two leucine zipper domains and a highly conserved C-terminal domain. We have identified kinesin light chain 1 (KLC1) as a binding partner for the second leucine zipper domain of JLP using yeast two-hybrid screening. The interaction domain of KLC1 was mapped to its tetratripeptide repeat, which contains a novel leucine zipper-like domain that is crucial for the interaction with JLP. Mutations of Leu-280, Leu-287, Val-294, and Leu-301 within this domain of KLC1 disrupted its ability to associate with JLP. Immunofluorescence studies showed that JLP and KLC1 co-localized in the cytoplasm and that the localization of JLP was dependent on its second leucine zipper. Ectopic expression of a dominant negative form of KLC1 resulted in the mislocalization of endogenous JLP. Moreover, the association between JLP and KLC1 occurred in vivo and was important in the formation of ternary complex with JNK1. These results identify a novel protein-protein interaction between KLC1 and JLP that involves leucine zipper-like domains and support the role of motor proteins in the spatial regulation of signaling modules.

To respond properly to extracellular signals, eukaryotic cells have developed cascades of highly conserved protein kinases (MAPK 1 (s) and their activating kinases), which form the central elements of signal transduction pathways that activate transcription factors in the nucleus and other effectors throughout the cell in response to environmental signals. The MAP kinase cascades have recently been shown to be regulated by scaffolding proteins, which organize signaling modules by forming large protein complexes composed of MAPKs and their upstream kinases. One such family of scaffolding proteins, known as JIPs, assemble the c-Jun NH 2 -terminal kinase (JNK) signaling module. This family contains three members, termed JIP-1, JIP-2, and JIP-3/JSAP-1/Syd. These proteins have been shown to directly associate with kinesins (1)(2)(3)(4), which are a family of motor proteins that move cargo proteins along microtubules in an ATP-dependent manner (5,6). Conventional kinesin, also known as kinesin-I or kinesin, is a heterotetramer of two kinesin heavy chains (KHCs) and two kinesin light chains (KLC). KHC is believed to have movement potential, whereas the KLCs are the link between the specific cargo proteins (7). KHC is composed of a catalytic motor domain, a neck linker and neck domains, and a stalk domain that is responsible for dimerization of the KHCs and association with KLC (8,9). Finally, the tail domain plays a role in the regulation of motor activity (10).
KLC contains two notable domains, an N-terminal ␣-helical heptad repeat region (coiled-coil domain) that associates with the KHC stalk and a C-terminal domain that consists of a series of six tetratricopeptide repeat (TPR) motifs (9). TPR motifs are degenerate repeats, often arranged in tandem to form protein-protein interaction domains (11,12). It is known that the TPR motifs of KLC interact with cargo proteins. In addition, it has also been demonstrated that the coiled tail of KHC can also provide a binding site for cargo proteins (2,7,13).
JIP-1 and JIP-2, which are unrelated to JIP-3 by sequence homology, interact with the TPR motifs of KLC through their extreme C termini. Deletions or a point mutation (P704A or Y709A) within the C-terminal residues of JIP1 (PTEDIYLE) was sufficient to disrupt this association (2). It has also been demonstrated that the interaction between JIP-1 and KLC1 was important for the proper cellular localization of JIP-1. In the neuronal cell line N1E 115, deletion or mutation of the KLC binding domain of JIP-1 led to inappropriate localization of JIP-1. Although the full-length JIP-1 localizes normally at the tips of neurites, deletion of its KLC binding domain caused it to be diffusely localized throughout the cell. Conversely, the expression of dominant negative kinesin constructs in differentiated neuronal cells caused the localization of endogenous JIP-1 protein to shift from the normal localization at the tips of the neuronal processes to throughout the cell.
Because the C terminus of JIP-3/JSAP-1/Syd bears no sequence similarity to those of JIP-1 and JIP-2, it is not surprising that the TPR of KLC1 does not interact with the C termini of these proteins. Rather, it is the coiled-coil/leucine zipper domain in the N terminus of Syd that directly associates with KLC1. However, the precise domain that contributes to the association between JIP-3/JSAP-1/Syd with KLC1 has yet to be identified. Dominant negative forms of KLC1 have also been shown to cause mislocalization of JIP-3 in neuronal cells. Therefore, it is clear that proper localization of JIPs is dependent on their ability to associate with KLC1.
We have recently described a new member of this scaffolding family of proteins called JLP (JNK-associated leucine zipper protein), which contains two leucine zipper domains (LZI and LZII), and Domain C that shares a high degree of homology with the Caenorhabditis elegans protein UNC-16. JLP is 69% homologous to the JNK scaffolding proteins JIP-3, JSAP1, and the Drosophila homologue Syd. Furthermore, an alternatively spliced variant of JLP has recently been described (4). Like other JIPs, JLP can assemble a three-tiered JNK signaling complex through the association of the upstream kinases MKK4 and MEKK3. However, it is unique from the other JIPs in that it has been demonstrated to interact with both JNK and p38MAPK. In addition, JLP can tether signaling proteins to transcription factors such as Myc and Max (14).
To gain more insight into the biological function of JLP, we sought to determine potential binding partners for the second leucine zipper of JLP (LZII). Using only the LZII as a bait, we screened a mouse brain cDNA library in a yeast two-hybrid system. These studies showed that KLC1 strongly associated with the LZII of JLP. Mapping studies revealed that JLP bound to the TPR motifs of KLC1, which was in accordance with previous studies pertaining to JIP-1, JIP-2, and JIP-3/JSAP-1/Syd (1,2,15). However, our studies also identified a novel protein-protein interaction domain within TPR that consists of a heptad repeat resembling a leucine zipper-like domain. A KLC1 protein containing point mutations within this domain hindered the ability of KLC1 to associate with JLP. Our results provide the first evidence of a leucine zipper-like domain in KLC1 that plays a role in the association between KLC and JLP. Interestingly, exogenous expression of the mutant JLP containing mutations in LZII or dominant negative form of KLC1 also resulted in the mislocalization of JLP, suggesting a role of motor proteins in the spatial regulation of signaling modules.

MATERIALS AND METHODS
Yeast Two-hybrid Screening and cDNA Cloning-The cDNA encoding the LZII of JLP was generated using the following primers and subcloned into an EcoRI/BamHI-digested pNLX3 yeast expression vector: sense 5Ј-GAGAGAGAATTCGTGGAGAATCTCATACTG and antisense 5Ј-GAGAGAGGATCCTAAGTTCACAGCCGCCAGCTC. The construct was confirmed by sequence analysis, and the construct subsequently used as a bait for screening the mouse brain Matchmaker Yeast Two-Hybrid library (Clontech) according to manufacturer's instructions.
Cell Culture and Transfection-African green monkey kidney cells (COS-7) and NIH3T3 cells were grown in Dulbecco's Modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum and calf serum, respectively, in an incubator at 37°C under humidified conditions and 5% CO 2 . COS-7 cells were transfected using the DEAE-dextran method as described previously (17). NIH3T3 cells were transfected using calcium phosphate precipitation as described previously (18).
Association of Endogenous KLC1 and JLP and Ternary Complex Formation-NIH3T3 cells were lysed in the lysis buffer (PBS containing 0.25% Nonidet P-40, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium fluoride, 5 mg/ml protease inhibitors (apostatin, leupeptine, pepstatin)) and centrifuged at 10,000 ϫ g for 15 min. The supernatants were mixed with control goat IgG or goat anti-KLC1 antibodies (Santa Cruz) at 4°C for 1 h. Protein G was then added to the samples, and the incubation was continued at 4°C for 1 h. The immunoprecipitates were washed three times with lysis buffer, resolved by SDS-PAGE, and subjected to immunoblot analysis using a rabbit anti-JLP antibody described previously (14). For detection of ternary complex formation, the same procedures were followed with cell lysates prepared from COS-7 cells transfected with plasmids expressing HAtagged KLC1, S-tagged JLP, and JNK1. The JNK-S levels in the lysates and immunoprecipitates were quantified using the software MacBas.
P19 Cell Neuronal Differentiation-The P19 clone was obtained from the American Type Culture Collection (Rockville, MD). Cells were cultured in Dulbecco's Modified Eagle medium supplemented with 10% fetal bovine serum (Sigma) in 6% CO 2 humidified chamber. To promote cellular differentiation, cells were cultured as aggregates in bacterial Petri dishes using media supplemented with 1 M retinoic acid (RA) for 4 days. After 4 days, the cells were harvested, washed with PBS, and then transferred to tissue culture dishes in medium lacking RA. Arabinosidase C was added to inhibit cell proliferation and to enrich the differentiated population.
Analysis of Protein-Protein Interactions in Vivo-Transfected COS-7 cells were harvested 48 -72 h posttransfection and lysed in the association assay buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5% Nonidet P-40, 0.5 mM EDTA, 0.5 mM dithiothreitol, 2 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium fluoride, 5 g/ml protease inhibitors (apostatin, leupeptin, pepstatin)). Clarified lysates were rotated for 1 h at 4°C with S-protein-agarose (Novogen) or glutathioneagarose (Sigma), centrifuged, and washed three times with association buffer. The precipitates were the resolved using SDS-PAGE and transferred to nitrocellulose. The membranes were probed with anti-HA, anti-GST, and anti-S antibodies (Santa Cruz) as described in the figure legends and visualized by the PerkinElmer Life Sciences enhanced chemiluminescence detection system.
Immunofluorescence Studies-Cells were grown on glass coverslips and transfected with appropriate expression plasmids using the calcium phosphate method (18). Posttransfection, the cells were fixed by incubation with 4% paraformaldehyde in PBS for 30 min, washed with PBS, and permeabilized with 0.1% Triton X-100 in PBS. After an incubation with 2% bovine serum albumin in PBS for 15 min, the coverslips were incubated with primary antibodies in PBS with 2% bovine serum albumin for 1 h. The primary antibodies used were a rabbit polyclonal antibody to JLP and a mouse monoclonal antibody to the HA epitope as described in the figure legends. Immunocomplexes were detected using Texas Red-conjugated anti-mouse or fluoresceinconjugated anti-rabbit immunoglobulin secondary antibodies (Molecular Probes) in 2% bovine serum albumin in PBS. Coverslips were mounted using the ProLong Antifade Kit (Molecular Probes) and examined by confocal microscopy.
Retroviral Infection-The cDNAs encoding the HA-tagged wild-type, TPR1, and TPR2 mutants of KLC were subcloned into the pFBneo vector (Stratagene). The retroviral vectors were transfected into an amphotrophic packaging cell line derivative of BOSC 23 (19) to prepare retroviral supernatants. The retroviral supernatants were harvested, centrifuged, and used to infect HFL-1 cells for three consecutive days in the presence of Polybrene (8 g/ml).

Identification of KLC1 as a JLP Interacting
Protein-To gain an insight into the biological functions of JLP, we sought to identify potential binding partners of the LZII of JLP using a yeast two-hybrid screening (Fig. 1A). This approach resulted in the identification of a partial cDNA clone of KLC1 (missing 164 bp of the 5Ј-sequence) as a binding partner of JLP. To confirm the specificity of the interaction, the L40 yeast strain expressing this KLC1 clone was mated with the AMR70 yeast stain expressing LZII (bait), the LZII:3-4, LZI and LZI mutants, or Lamin C, which was used as a negative control. Yeast coexpressing KLC1 and each of the LZ constructs showed growth on -Leu/-Trp plates, whereas only yeast expressing KLC1 and LZII were able to grow on -Leu/-Trp/-His plates (Fig. 1B). This suggests that KLC1 associates specifically with LZII but not with LZI. Mutation of the LZII domain by replacing the third and fourth leucine residues with alanines disrupts the KLC1/ LZII interaction.
On average, six heptad repeats of leucines comprise a leucine zipper domain. However, a minimum of only four heptad repeats of leucines are required to make up a functional leucine zipper (20). LZII of JLP is unique in that it is relatively large; it contains nine heptad repeats of leucines with the exception of valine at the seventh position. This implies that more than one protein can potentially bind LZII of JLP at the same time. To gain an understanding of the nature of interaction between JLP and KLC1, we investigated the specificity within LZII that contributed to the association with KLC1. We systematically mutated the appropriate leucine and valine residues to alanines within LZII to disrupt the appropriate heptad repeats (Fig. 1A). We then tested whether these mutants were capable of interacting with KLC1 by performing yeast-mating experiments using one strain of yeast expressing KLC1 and another expressing LZII, LZII:1-2, LZII:3-4, LZII:5-6, LZII:7-8, or LZII:8 -9. After mating, yeast strains expressing the wild-type JLPLZII and mutants JLPLZII:7-8 and LZII:8 -9, but not the mutants JLPLZII:1-2, JLPLZII:3-4, and JLPLZII:5-6, were found grow on -Leu/-Trp/-His plates (Fig. 1C) indicating that mutations of any of the first six leucine residues abrogate the association between LZII of JLP and KLC1. These results demonstrate that the association between JLP and KLC1 is dependent on the first six heptad repeats within LZII of JLP.
In Vivo Interaction of JLP and KLC1-To validate the results performed in yeast, we sought to determine the ability of full-length JLP and KLC1 to interact when co-expressed in eukaryotic cells. GST pull-down assays using lysates derived from COS-7 cells expressing JLP and KLC1 demonstrated that GST-KLC1 could precipitate JLPWT-HA, whereas the JLPLZII-HA mutant was unable to do so ( Fig. 2A). In a reciprocal experiment, wild-type S-tagged JLP, but not the mutant JLPLZII:3-4, could precipitate KLC1 in pull-down experiments using S-protein-agarose (Fig. 2B). These results demonstrated that the association between JLP and KLC1 was dependent on LZII of JLP. Additional experiments were performed using the series of S-tagged JLPLZII mutants in conjunction with HA-KLC1 in COS-7 cells. The results of these experiments showed that although wild-type JLP, mutants JLPLZII:7-8, and JLPLZII:8 -9 could precipitate KLC1, mutants JLPLZII:1-2, JLPLZII:3-4 and JLPLZII:5-6 failed to do so (Fig. 2C). These in vivo association studies in COS-7 cells corroborate the results obtained from the yeast mating assays and confirm that formation of the JLP-KLC1 complex is dependent on the first six heptad repeats within LZII of JLP (Fig. 2C). To determine the role of LZII (which associates with KLC1) in the subcellular localization of JLP, we expressed HA-tagged wild-type, LZI, and LZII mutants of JLP in NIH3T3 cells. The transfected cells were fixed, and the subcellular localization of the exogenous JLP proteins was analyzed by immunofluorescence using a specific antibody against the HA-tag. As described previously, wild-type JLP was mainly localized throughout the cytoplasm (Fig.  2D). No significant change in subcellular localization was observed when LZI was mutated, suggesting that LZI is not involved in the subcellular localization of JLP. However, when LZII was mutated, the JLP mutant was localized to distinct patched areas within the cytoplasm (Fig. 2D), in contrast to that observed with wild-type JLP. This indicates that LZII is important for the proper subcellular localization of JLP. Because LZII associates with KLC1, it is possible that Identification of a Novel Leucine Zipper-like Domain in KLC1-KLC1 contains an N-terminal ␣-helical domain (also known as the heptad repeat domain or coiled-coil domain) that associates with the KHC stalk and a C-terminal domain that consists of six TPR motifs (Fig. 3A). The TPR motifs are involved in protein-protein interactions and represent a link between cargo and microtubule. These TPR motifs of KLC1 have been shown to associate with JIP-1, JIP-2, JIP-3/JSAP-1/Syd, and JIP-4, but the precise domain within KLC1 that contributes to the association has not yet been identified. To determine the precise domain of KLC1 that interacts with JLP, we generated several deletion mutants of KLC1 that encoded only the heptad repeat/coiled-coil domain, the TPR domain (TPR1) or the partial TPR domain (TPR2) (Fig. 3A). S-protein precipitation studies in COS-7 cells demonstrated that S-tagged JLPWT was only able to associate with full-length KLC1 and the TPR1 mutant (Fig. 3B). However, S-tagged JLPWT was not able to co-precipitate the coiled-coil domain or TPR2 mutants. This suggests that JLPWT binds to a region in KLC1 spanning from amino acids 181 to 331. Because the LZII of JLP was found to mediate the association between JLP and KLC1, we next searched for the presence of a leucine zipper-like motif within amino acid residues 181-331 of KLC1. This search revealed the presence of a peptide sequence at position 279 -300 that contained four heptad repeats (Fig. 4A). Mutation of Leu-280, Leu-287, Val-294, and Leu-301 of KLC1 to alanine (HA-KLC1LZ) abolished its association with JLP (Fig. 4B). The results of this study revealed the presence of a novel "leucine zipper-like" domain in KLC1 that was essential for binding to LZII of JLP. These results also suggest that JLP and KLC1 may from heteromers via LZII of JLP and the leucine zipperlike domain of KLC1.
Association of Endogenous KLC1 and JLP-To determine whether endogenous KLC1 and JLP could associate in vivo, we performed immunoprecipitations using cell lysates derived from NIH3T3 cells treated in the presence or absence of arsenite using control or anti-KLC1 antibodies, and subjected them to immunoblot analysis using a JLP-specific antiserum. Treatment with arsenite has been shown to stimulate the activities of JNK and p38MAPK (21,22). In addition, our studies showed that treatment of cells with arsenite results in the phosphorylation of JLP and its perinuclear localization (data not shown). The results presented in Fig. 5 showed that the anti-KLC1 antibody could specifically precipitate JLP, demonstrating that endogenous KLC1 and JLP associate in vivo. However, treatment with arsenite did not change the levels of JLP that are associated with KLC1 (Fig. 5).

Role of LZII Domain of JLP in Subcellular Localization-
The spatial organization of several cytoplasmic proteins has been shown to be dependent on their direct or indirect association with components of the cytoskeleton. KLC1 is one component that links cargo to motor proteins, which in turn are anchored on microtubules. Because our observations show that JLP associates with KLC1, we hypothesized that KLC1 was responsible for proper subcellular localization of JLP and that a dominant negative mutant of KLC1 (TPR1) could alter its subcellular localization. To address this question, we determined the subcellular distribution of JLP in human lung fibroblasts, HFL-1 (Fig. 6). The cells were infected with viral supernatants derived from packaging cells transfected with   FIG. 2. JLP and KLC1 associate in vivo. A, GST pull-down assays were performed using lysates derived from COS-7 cells expressing GST or GST-KLC1 together with HA-tagged JLPWT or JLPLZII:3-4. The precipitates were subjected to immunoblot analysis using an anti-HA antibody. B, lysates derived from COS-7 cells expressing S-tagged JLPWT, JLPLZII:3-4, or JLPLZI mutant together with HA-tagged KLC1 were precipitated using S-protein-agarose. The precipitates were immunoblotted using an anti-HA antibody. C, S-tagged JLPWT and JLPLZII mutants were expressed with HA-KLC1 in COS-7 cells. The cell lysates were precipitated with S-protein-agarose. The precipitates were immunoblotted using an anti-HA antibody. All protein expression in the cell lysates was detected with anti-S, anti-HA, or anti-GST antibodies. D, NIH3T3 cells were transfected with an empty expression vector (V) or a vector expressing HA-tagged wild-type (WT), LZI, or the LZII mutant of JLP. The cells were fixed and subjected to immunofluorescence studies using a specific antibody directed against the HA-tag. The exogenous JLP proteins appear red in color. retroviral vectors expressing HA-tagged wild-type KLC1 and the TPR1 and TPR2 mutants shown in Fig. 3. The infected cells were then analyzed by immunofluorescence using a specific JLP antibody and an HA antibody to detect the endogenous JLP and KLC1 proteins, respectively. The results of this study showed that JLP was mainly distributed homogenously throughout the cytoplasm (Fig. 6), which is consistent with our previously published results obtained in Swiss3T3 cells (14). Moreover, wild-type KLC1 was also similarly distributed throughout the cytoplasm, suggesting that both proteins are localized in the same region of the cell (Fig. 6). In contrast, when the TPR1 mutant of KLC1 was expressed, it was detected in both the cytoplasm and the nucleus (Fig. 6). This inappropriate localization of TPR1 resulted in the translocation of JLP to the nucleus. It is possible that TPR1 transported JLP into the nucleus because of its ability to bind JLP (Fig. 3). Similarly, when the TPR2 mutant was expressed, it was detected in both the cytoplasm and the nucleus (Fig. 6). However, in this case, the mislocalization of TPR2 did not result in JLP being localized to the nucleus. This is likely because of the fact that TPR2 does not bind to JLP (Fig. 3). Taken together, these results demonstrate that the proper subcellular localization of JLP depends on its interaction with KLC1.
Ternary Complex Formation by JLP, JNK, and KLC1-We have previously demonstrated that JLP can regulate JNK signaling and that this phenomenon can be attributed to the ability of JLP to associate with JNK and to function as a scaffolding protein for the JNK signaling module (14). To demonstrate that JLP, KLC1, and JNK form a complex, we expressed various combinations of HA-tagged KLC1, S-tagged JLP, and JNK1 in COS-7 cells (Fig. 7). Cell lysates derived from these cells were subjected to immunoprecipitation using a KLC1-specific antibody. Our results show that when JLP or JNK1 was expressed either individually or together in the absence of KLC1, neither protein was able to be immunoprecipitated. However, when JLP was co-expressed with KLC1, JLP could be co-immunoprecipitated using the KLC1 antibody; this result is in agreement with those shown in Fig. 2. Furthermore, when KLC1 and JNK1 were co-expressed, a small constructs. Amino acid numbers correspond to appropriate domains within the KLC1 deletion and mutation contracts. B, lysates derived from COS-7 cells expressing HA-tagged KLC1, coiled-coil (CC), TPR1, or TPR2 together with S-tagged JLPWT were precipitated using Sprotein-agarose. Immunoblot analysis was performed on the lysates and precipitates using anti-HA and anti-S antibodies.

FIG. 4. Identification of a novel leucine zipper-like domain in KLC1.
A, schematic representation of KLC depicting mutations to alanine at positions 280, 287, 294, and 301. B, wild-type HA-KLC1 or HA-KLC1LZ was expressed in COS-7 cells together with wild-type or mutants of S-tagged JLP as indicated. The cells were lysed and subjected to precipitation using S-protein-agarose. The precipitates were immunoblotted for associated proteins using anti-HA antibodies. Protein expression was detected with either anti-S or anti-HA antiserum.

FIG. 5. In vivo association of endogenous KLC1 and JLP.
NIH3T3 cells were treated with or without arsenite (0.5 mM) for 30 min. Lysates derived from these cells were immunoprecipitated (IP) with normal control goat IgG (cont) or a goat anti-KLC1 antibody (KLC1). The immunoprecipitates were washed three times with the lysis buffer, resolved by SDS-PAGE, and immunoblotted with the rabbit antibody directed against JLP as described previously (14). One-twentieth of the lysates used for immunoprecipitation were loaded in each lane.
FIG. 6. Effects of KLC1 mutants on the subcellular localization of JLP. HFL-1 cells were infected with supernatants derived from an empty retroviral vector (Vector) or vectors expressing HA-tagged wildtype (KLC1), TPR1, or TPR2 mutants of KLC1 as described in the legend to Fig. 3. After infection, the cells were grown on coverslips, fixed, and stained with a rabbit antibody directed against JLP (14) and a monoclonal anti-HA antibody. Secondary staining was performed using anti-rabbit fluorescein isothiocyanate (green) and anti-mouse Texas Red (red) secondary antibodies. The right panel shows the merged images obtained by the staining of JLP and the HA-tagged KLC1 forms. Overlapping localization of both proteins appears yellow.
amount of JNK1 could also be co-immunoprecipitated with KLC1, which is clearly visible in the long exposure. This association was likely because of the tethering property of JLP toward KLC1 and JNK1 and is supported by the fact that greater levels of JNK1 are precipitated when exogenous JLP is co-expressed with KLC1 and JNK1 (Fig. 7). These results support the notion that JLP acts as a bridging molecule that mediates the formation of a ternary complex with KLC1 and JNK1 and rejects the notion that two independent binding events occur between JLP and KLC1, and JLP and JNK1.
Induction of JLP and KLC1 Transcripts during Neuronal Differentiation-The observation that JLP and KLC1 are highly expressed in adult brain compared with other tissues (4,23) suggested that their expression might be induced during neuronal differentiation. To test this hypothesis, we examined the expression of JLP and KLC1 in the P19 cell model of neuronal differentiation in response to RA (Fig. 8A). A Northern blot analysis showed that both JLP and KLC1 were detected at low levels in proliferating P19 cells. However, when these cells were induced to differentiate, RNA levels of both JLP and KLC1 were increased as a function of time (Fig. 8B). Time course analysis indicated that JLP and KLC1 expression was not an early event following RA treatment. However, JLP and KLC1 expression was robust at day four following initiation of differentiation (Fig.  8B). The expression profiles of JLP and KLC1 are also similar at the protein level (Fig. 8C), and it is therefore likely that these proteins associate in neuronal cells as well. DISCUSSION There are essentially two classes of scaffolding proteins that assemble the JNK signaling module. The first class is comprised of the SH3-motif containing proteins known as JIP-1 and JIP-2 and the second class includes JIP-3/JSAP-1/Syd, JLP, and JIP-4, which contain leucine zipper motifs (4, 24 -27). Besides their role as scaffolding proteins for the JNK signaling pathway, both classes of proteins also associate with kinesin motor proteins through KLC1. However, because of the variance in amino acid sequence between the two classes of proteins, it is not surprising that they associate with KLC1 TPRs through different domains. JIP-1 and JIP-2 bind to KLC1 TPRs through their extreme C termini. This is consistent with previous studies on TPR domain-containing proteins demonstrating that TPR domains recognize short sequences at the C terminus of their partner protein (28). However, the TPRs of KLC1 associate with the internal sequences of the JNK scaffolding proteins, JIP-3/JSAP-1/Syd and JIP-4 (1, 4).
JLP is a leucine zipper-containing scaffolding protein that also associates with JNK and KLC1. Sequence comparison shows that it exhibits 69% homology to JIP-3/JSAP-1/Syd, with an exceptionally high homology in the two coiled-coil domains (leucine zipper domains) and Domain C. Therefore, it was not surprising that LZII of JLP was shown to associate with KLC1 when it was used as a bait in the yeast two-hybrid screening. This result is consistent with the finding that the GST-KLC1 was only able to pull down the recombinant N terminus of Syd Cell lysates derived from COS-7 cells expressing various combinations of HA-tagged KLC1, S-tagged JLP, and JNK1 were immunoprecipitated with a KLC1-specific antibody. The immunoprecipitates and total cell lysates were subjected to Western blot using antibodies specific for the HA-and S-tags. Short (5 min) and long (60 min) exposures of the immunoprecipitated S-tagged JNK1 levels are shown (short exp and long exp, respectively). The levels of S-tagged JNK1 in the lysates and immunoprecipitates were quantified and indicated below their corresponding panels. Based on the quantitation, the relative ratio of S-tagged JNK1 between precipitates and lysates is given (ratio: ppt/lysate). and not the C terminus (1). Furthermore, the alternative splicing variant of JLP, JIP-4, interacts with the TPR of KLC1 via a region containing the leucine zipper (4). Our results are therefore consistent with the idea that the domains of JIP proteins that interact with KLC1 are distinct and that the binding of each JIP protein to kinesin is mediated by the TPR region of kinesin light chain.
A leucine zipper consists of a stretch of amino acids with a leucine residue in every seventh position in a coiled-coil ␣-helical structure (20). These leucine residues are important for the leucine zipper to form a dimer with the leucine zipper of another polypeptide. Our results provide definitive evidence that the heptad repeats of leucine residues within LZII of JLP are required for the association with KLC1 both in yeast and in mammalian cells (Figs. 1 and 2). However, to date, there have been no prior reports that identified leucine zipper-like domains in the TPRs of KLC1. We have scanned the peptide sequence of the KLC1 TPRs and identified such a domain. One stretch of amino acids contained four heptad repeats of leucine residues, with the exception of a valine in the third heptad. A KLC1 construct containing the mutations L280A, L287A, V294A, and L301A was unable to associate with JLP in mammalian cells (Fig. 4). These experiments not only show an association between JLP and KLC1 but also provide evidence for a novel leucine zipper-like domain in KLC1 that is important for the association with JLP.
The connection between motor proteins and signaling pathways has many implications. It has been reported that KLC1 is the link between kinesin motor proteins and JIPs and that the significance of this complex is to mediate proper cellular localization (26). Furthermore, the formation of a ternary complex consisting of JLP, KLC1, and JNK1 (Fig. 7) suggests that JLP serves as a link between the kinesin motor proteins and their cargo JNK signaling complex proteins (Fig. 9). Interestingly, accumulation of JLP in the perinuclear region results when cells are exposed to stress-inducing agents such as arsenite or UV radiation (14). This may represent a mechanism by which JLP can transduce signals from the periphery of the cell to the nucleus. Alternatively, scaffolding proteins could allow signaling pathways to regulate motor activity. For example, the association of kinesin with the JIPs could enable the JNK pathway to induce phosphorylation of kinesin itself or an associated protein, thereby activating the motor at the point of departure and/or inactivating it at the point of destination (Fig. 9).
We and others have demonstrated that a dominant negative mutant of kinesin (DN kinesin) can alter the proper localization of JIPs and JLP. However, it is still unresolved as to whether DN kinesin can alter the localization of JNK. In addition, it would be interesting to determine whether DN kinesin can alter the activation of JNK. It has been speculated that there is a connection between kinesin, JNK signaling, and neurogenesis. For example mice lacking the ubiquitously expressed conventional kinesin gene kif5b and mice lacking the jnk1 and jnk2 genes exhibit embryonic lethality with severe defects in early brain development (29).
Although biochemical studies and transfection assays have demonstrated that putative JNK scaffold proteins can regulate JNK activation in cultured cells, the function of these scaffold proteins in vivo has not been extensively established. One such in vivo study involves genetic alterations in the syd gene, the Drosophila homologue of JIP-3, and JSAP. Mutations in syd interfered with axonal transport processes and resulted in massive accumulation of anterograde and retrograde membranous axonal cargo within the axons of the larval segmental nerves. Thus, proper Syd function is required for efficient axonal transport in Drosophila, and the Syd mutant phenotypes are nearly indistinguishable from mutants lacking the anterograde axonal transport motor kinesin-1 (1, 30 -32). These studies suggest an important role for scaffolding proteins such as JIP-3, JSAP-1, and JLP in motor transport.