In this study, we investigated the contribution of JIP1 binding proteins to the interaction between JIP1 and kinesin-1. Although the binding of JNK to the JBD of JIP1 showed no significant effect on the kinesin-1 binding, the binding of JIP3 to the PTB domain of JIP1 enhanced it significantly. DLK, another JIP1-PTB domain binding protein, did not show such an effect. However, the over-expression of kinase-inactive DLK perturbed the formation of the JIP1–JIP3–kinesin-1 ternary complex, suggesting that DLK might compete with JIP3 for the JIP1-PTB domain. Taken together with the reported function of JIP1 as a cargo adaptor , these results suggest that the potency of the interaction between JIP1 and kinesin-1 depends on which protein binds to the JIP1-PTB domain.
JIP1 and JIP3 were originally identified as JNK binding proteins that function as a scaffold for the JNK activating protein kinase cascade in mammalian cells [18, 19]. While JIP1 is now recognized as a cargo adaptor that connects cargo and kinesin-1 , the relationship between these distinct JIP1 functions remained unclear. It has been reported that the activation of JNK leads to the unloading of kinesin-1 from MTs . However, the contribution of JIP1 or JIP3 to this process is ambiguous. A previous study in Drosophila suggested the possibility that JIP1–kinesin-1 binding is disrupted by JNK activation via an unidentified JNK target . Because the Drosophila JIP1 ortholog, APLIP1, lacks the JBD, the effect of JNK may not be dependent on JIP1–JNK binding. We show here that the binding of JIP1 to kinesin-1 is not affected by the binding of JNK to JIP1, regardless of JNK activity, in mammalian cells. JIP1–JIP3 binding was also independent of JIP1–JNK binding (see Figure 2B and Additional file: 1 Figure S1A). We still cannot rule out the possibility that JIP3–JNK binding affects JIP1–JIP3 binding or JIP1–kinesin-1 binding. However, such scenarios may be unlikely because it has been reported that deletion of the JBD in JIP3 does not affect JIP3 transport in neural cells , and because the JIP1–JIP3 interaction in Neuro 2a cells was independent of JNK activity (see Additional file: 1 Figure S1B). Taken together, JNK seems to be a cargo of the JIP1–JIP3–kinesin-1 complex rather than a regulator of JIP1-JIP3-kinesin-1 complex formation.
In contrast to the dispensability of the JIP1 JBD, the PTB domain is essential for kinesin-1 binding by JIP1 in differentiated Neuro2a cells. We identified JIP3 as a major PTB-dependent JIP1 binding protein that was largely essential for the neurite tip localization and kinesin-1-binding of JIP1. JIP1 or JIP3 bound to kinesin-1 via their kinesin-1-binding motifs, without forming a JIP1–JIP3 complex. However, this “basal” kinesin-1 binding activity was significantly enhanced by JIP1–JIP3 complex formation, which was totally dependent on the JIP1-PTB domain. A previous study showed that distinct regions of the TPR domains in KLC are responsible for JIP1 and JIP3 binding . Therefore, JIP1 and JIP3 may bind to kinesin-1 in a co-operative manner to form a stable ternary complex. This notion is further supported by the observation that the neurite tip localization of JIP1 is decreased by JIP3 knockdown (see Figure 3) . It has been reported that JIP3 binds not only to kinesin light chain (KLC) but also to kinesin heavy chain (KHC) . While we cannot rule out a possibility that JIP3–KHC binding contributes to the ternary complex formation, our results with a mutant JIP3 bearing amino acid substitution in the KLC contact site indicate that JIP3–KLC binding is essential for the ternary complex formation. As for JIP1–JIP3 binding, it has been reported that the central 189 amino acids of JIP3, which include the JBD and kinesin-1 binding regions, and the C-terminal 158 amino acids of JIP1, which include the PTB domain and kinesin-1 binding region, are sufficient for their binding . Our results using JIP1-PTB domain point mutants further delineate the JIP3-binding domain in JIP1. By using domain specific mutants of JIP1 and JIP3, we showed that JIP1–JIP3–kinesin-1 ternary complex formation depends on JIP1–JIP3-binding, JIP1–kinesin-1-binding, and JIP3–kinesin-1-binding. Although the abrogation of one of these interactions prevented ternary complex formation, JIP1–JIP3 binding was independent of their binding to kinesin-1. This indicates that JIP1–JIP3 complex formation leads to the formation of the ternary complex. It has been reported that JIP1–kinesin-1 binding can trigger the activation of kinesin-1 motor activity in the presence of additional factors such as FEZ1 . Therefore, JIP1–JIP3 complex formation might also be critical for the JIP1-dependent activation of kinesin-1.
Because the kinesin-1 binding activity of JIP3 is crucial for the formation of the ternary complex, other proteins that compete with JIP3 for binding to the JIP1-PTB domain and lack the capacity to bind kinesin-1 may negatively regulate the potency of JIP1 for kinesin-1 binding. Using DLK, we demonstrated a potential regulatory role for JIP1-PTB domain binding proteins in the formation of the JIP1–JIP3–kinesin-1 ternary complex. In addition, kinesin-1 or JIP3-binding proteins may also affect the formation of the JIP1–JIP3–kinesin-1 complex. The inhibition of JIP1–kinesin-1 binding by Ca2+−dependent binding of S100A6 to KLC, and the inhibition of JIP3–kinesin-1-binding by GTP-dependent binding of ARF6 to JIP3, have been reported [24, 25]. Taken together, these observations suggest that kinesin-1-dependent transport is regulated by multiple signaling pathways via JIP1, JIP3 and KLC.
In this study, we have shown that JIP3 is a major binding protein of JIP1 in Neuro2a cells, and that JIP3 has the highest JIP1-binding capacity among various known JIP1-PTB domain binding proteins. Because JIP1 and JIP3 are highly expressed in neurons [26, 27], these observations imply that JIP1 and JIP3 play a common role in kinesin-1 dependent intracellular transport. In fact, genetic studies in Drosophila and Caenorhabditis elegans have indicated that both JIP1 and JIP3 support vesicle transport in neural cells [4, 28, 29]. However, knockdown of JIP1 or JIP3 result in different phenotypes in mammalian neurons: JIP1 knockdown results in the partial inhibition of axon elongation, while JIP3 knockdown results in the stimulation of neurite elongation and branching [30, 31]. In our experiments, although JIP3 was a major JIP1 binding protein in Neuro2a cells, only a fraction of the endogenous JIP3 was co-precipitated with endogenous JIP1 (see Figure 2E). This implies that a considerable fraction of JIP3 exists in protein complexes free of JIP1, and may have an additional function unrelated to JIP1. Therefore, the JIP1–JIP3 complex might not be a universal kinesin-1 regulatory module, but rather a specific regulator of intracellular transport in neuronal cells.