Insights into Kinesin-1 Activation from the Crystal Structure of KLC2 Bound to JIP3

Abstract

Kinesin-1 transports numerous cellular cargoes along microtubules. The kinesin-1 light chain (KLC) mediates cargo binding and regulates kinesin-1 motility. To investigate the molecular basis for kinesin-1 recruitment and activation by cargoes, we solved the crystal structure of the KLC2 tetratricopeptide repeat (TPR) domain bound to the cargo JIP3. This, combined with biophysical and molecular evolutionary analyses, reveals a kinesin-1 cargo binding site, located on KLC TPR1, which is conserved in homologs from sponges to humans. In the complex, JIP3 crosslinks two KLC2 TPR domains via their TPR1s. We show that TPR1 forms a dimer interface that mimics JIP3 binding in all crystal structures of the unbound KLC TPR domain. We propose that cargo-induced dimerization of the KLC TPR domains via TPR1 is a general mechanism for activating kinesin-1. We relate this to activation by tryptophan-acidic cargoes, explaining how different cargoes activate kinesin-1 through related molecular mechanisms.

Keywords: ARF6; JIP3; KLC; activation; cargo; kinesin; kinesin-1; molecular; motor; regulation.

Figure 1

The Crystal Structure of the KLC2^TPR:JIP3^LZ Complex (A and B) Schematics of the KLC2 (A) and JIP3 (B) molecules. (C) The KLC2^TPR:JIP3^LZ complex. The KLC2^TPRs are shown in teal (chain A) and cyan (chain B). The KLC2^TPR α helices are shown as cylinders and labeled 1A through 6B. JIP3^LZ subunits are shown in magenta (chain C) and pink (chain D) cartoon. Blue/red spheres are N- and C-terminal C_α atoms, respectively. See also Figure S1.

Figure 2

Biophysical Studies of KLC2^TPR Binding to the JIP3^LZ (A–D) ITC thermograms and isotherms for KLC2^TPR binding to GST3C-JIP3^LZ. The error bars in the isotherms show the errors associated with integration of the injection peaks in the corresponding thermograms. Molar ratios correspond to the GST3C-JIP3^LZ dimer concentration. The curves show the sequential binding model that was globally fitted across all the isotherms in the dataset. The lower panel shows the residuals between the isotherm data points and the fitted model. (E) CD spectra of the Trp-JIP3^LZ and KLC2^TPR-myc. (F and G) Secondary structure composition of the Trp-JIP3^LZ (F) and KLC2^TPR-myc (G). The black bars show the secondary structure composition determined from deconvolution of the CD spectra (mean and SD from three experiments). The gray bars show the Trp-JIP3^LZ and KLC2^TPR-myc secondary structure compositions calculated from the KLC2^TPR:JIP3^LZ crystal structure. (H) SEC-MALLS chromatogram for the JIP3^LZ domain showing the light scattering (LS) and differential refractive index (dRI) traces and molecular weight (black curve). (I) Schematic showing the sequential binding model for KLC2^TPR binding to the JIP3^LZ. See also Tables 2 and S1.

Figure 3

The JIP3^LZ Binds to KLC2 TPR1 (A) Amino acid sequence of the murine JIP3^LZ. Heptad repeats are numbered and the amino acid in the “d” position is written in red. Residues that bind to KLC2^TPR chain A are highlighted in magenta (chain C) and pink (chain D). JIP4 residues not conserved with JIP3 are written beneath. (B) The KLC2 binding site on the JIP3^LZ domain (chain C, magenta; chain D, pink). The side chains of residues that interact with the KLC2^TPR domain are shown as sticks (carbon, main chain color scheme; nitrogen, blue; oxygen, red). (C) The JIP3 binding site on the KLC2 TPR1. The side chains of residues that interact with the JIP3^LZ are shown as sticks. (D) Stereo image of the KLC2^TPR:JIP3^LZ interface involving KLC2^TPR chain A. Putative hydrogen bonds are shown as dashed lines. (E) Amino acid sequence alignment for murine KLC1-4 over TPR1. Conserved and semi-conserved residues are highlighted in red and yellow, respectively. KLC2 residues that interact with JIP3 and their counterparts in KLC1/3/4 are outlined in blue. Numbering corresponds to KLC2. (F) The KLC2^TPR from the KLC2^TPR:JIP3^LZ crystal structure showing the JIP3 (magenta) and tryptophan-acidic (orange) binding sites. The tryptophan-acidic peptide from SKIP (sticks; carbon, yellow; nitrogen, blue; oxygen, red) was modeled based on PDB: 3ZFW (Pernigo et al., 2013). TPR1 residues that were mutated in the binding studies in (G) are labeled. (G) Representative ITC thermograms and isotherms for titration of the Trp-JIP3^LZ domain (left panel) or FITC-CSTN-WD2 peptide (right panel) into wild-type and mutant KLC2^TPR-myc. Molar ratios for the Trp-JIP3^LZ domain correspond to the Trp-JIP3^LZ dimer concentration. The error bars in the isotherms show the errors associated with integration of the injection peaks in the corresponding thermograms. See also Table S2.

Figure 4

The JIP3^LZ Does Not Disrupt LFP Motif Binding to the KLC2^TPR (A) View of the KLC2^TPR from the KLC2^TPR:JIP3^LZ complex. The bound LFP motif (red cartoon) was modeled using PDB: 5FJY (Yip et al., 2016). (B) Fluorescence anisotropy from the FITC-LFP peptide with a fixed concentration of KLC2^TPR-myc (21 μM) and varying concentrations of Trp-JIP3^LZ or CSTN-WD1 peptide as indicated. For the Trp-JIP^LZ data, concentration values correspond to the total concentration of monomeric Trp-JIP^LZ subunits. The calsyntenin-1 data are fitted with a competitive inhibition model (see the STAR Methods). The JIP3 data points are joined by straight lines. Data from a single experiment is shown in each case. Data points and error bars correspond to the mean and SD of triplicate measurements. Where error bars are not visible they are smaller than the marker.

Figure 5

Conservation of the KLC2^TPR:JIP3^LZ Interface across The Animal Kingdom (A) Canonical animal species phylogeny complete with the evolutionary timeline. The timing of the emergence of Metazoa and Bilateria in the fossil record is indicated by orange bars. The number of species in our dataset for each phylum is given in parentheses. (B) Bar graph showing the percentage of species in our datasets containing the indicated KLC or JIP3 motifs. Solid and hatched bars show data for Bilaterian and non-Bilaterian species, respectively. Bars are color-coded by motif as shown in the key. See also Tables S3 and S4.

Figure 6

KLC^TPR Dimerization via TPR1 Mimics JIP3 Binding (A–C) Structural comparison of the KLC2^TPR:JIP3^LZ interface with the crystallographic TPR1:TPR1 dimer interface in crystals of the KLC2^TPR (PDB: 5OJF). (A–C) show three views of the PDB: 5OJF dimer interface, with chain A in teal and the dimer mate in yellow. The JIP3^LZ (pink and magenta) was positioned onto PDB: 5OJF chain A by superposing KLC2 chain A of the KLC2^TPR:JIP3^LZ complex onto subunit A of PDB: 5OJF via TPR1. (D–E) TPR1 from KLC2 (D) and KLC1 (E) shown in gray loop representation (PDB: 5OJF and 3NF1, respectively). Interfacial residues are shown in stick representation with carbon atoms colored according to the interface(s) they participate in. See also Figure S2.

Figure 7

Molecular Mechanisms by which ARF6 Regulates Kinesin-1 Recruitment by JIP3/4 (A) Molecular surface of the JIP4^LZ (PDB: 2W83) showing the KLC and ARF6 binding sites (Isabet et al., 2009). (B) Model of the JIP4^LZ bound to the KLC2^TPR and active, membrane-associated ARF6. GTP, sticks with carbon/oxygen/nitrogen/phosphorus, gray/red/blue/orange; N-terminal myristoyl group, black zigzag; N-terminal amphipathic helix, yellow cylinder.

Figure 8

A Unified Framework for Kinesin-1 Activation by Different Cargoes (A) Model for the autoinhibited state of the kinesin-1 heterotetramer, showing how binding of the KLC LFP motif to the TPR domain would steer the JIP3 binding site on KLC toward the KHC motor domains and tails. (B) Proposed pathway for kinesin-1 activation by JIP3 (LZ domain, oval; tail-binding region, square). (C) Proposed pathway for kinesin-1 activation by a general dimeric cargo that induces TPR dimerization via TPR1. (D) Proposed pathway for kinesin-1 activation by a tryptophan-acidic cargo (tryptophan-acidic motif, hexagon). This has been conceptually broken down into two steps to show relief of KLC and KHC tail regulation of the motor domains, but both steps could occur simultaneously (Yip et al., 2016).