Structural Basis For Kinesin-1:Cargo Recognition

Abstract

Kinesin-mediated cargo transport is required for many cellular functions and plays a key role in pathological processes. Structural information on how kinesins recognize their cargoes is required for a molecular understanding of this fundamental and ubiquitous process. Here we present the crystal structure of the tetratricopeptide repeat of kinesin light chain 2 in complex with a cargo peptide harboring a ‘tryptophan-acidic’ motif derived from SKIP, a critical host determinant in Salmonella pathogenesis and a regulator of lysosomal positioning. Structural data together with biophysical, biochemical and cellular assays allow us to propose a framework for intracellular transport based on the binding by kinesin-1 of W-acidic cargo motifs through a combination of electrostatic interactions and sequence-specific elements, providing direct molecular evidence of the mechanisms for kinesin-1:cargo recognition.

The plus-end directed motor, kinesin-1 plays a critical role in the intracellular transport of diverse protein, ribonuclear protein complexes and membrane compartments on microtubules (1). Its functions are also usurped by bacteria and viruses to aid in their replication (2, 3). Kinesin-1 can perform this diverse range of functions by virtue of its ability to interact with many different cargo proteins (4). Diversity of cargo recognition is accomplished largely through the kinesin light chains (KLC) which harbour a tetratricopeptide repeat (TPR) domain, a versatile protein interaction platform (5, 6). The KLC^TPR domain can recognise short peptide stretches within relatively disordered regions of its targets. These peptides are characterized by a tryptophan residue flanked by acidic residues (e.g. EWD) and are found in a growing list of KLC binding proteins (7-16). Although W-acidic motifs often occur in pairs, single motifs are also functional and can support microtubule-based transport even when explanted from their host protein (7, 12, 17). We set out to solve the structure of a KLC^TPR domain bound to a cargo W-acidic motif. We focused our attention on the SKIP cargo for its importance in Salmonella pathogenesis. SKIP contains a pair of W-acidic motifs centred at amino acid positions 207-208 (WD) and 236-237 (WE) that fall within the N-terminal kinesin-1 binding region (residues 1-310)(3, 7, 13) (Fig. 1A). To assess the relative importance of the SKIP W-acidic motifs for KLC binding we co-transfected HeLa cells with wild-type and WD/WE mutant constructs expressing GFP-SKIP(1-310) and HA-KLC2 (Fig. 1B). Disruption of the WD motif significantly reduced GFP-SKIP interaction with HA-KLC2 whereas abrogation of the WE motif had no obvious effects. A double mutant with both motifs disrupted displayed HA-KLC2 binding similar to the single WD mutant. Thus the WE motif has a very low affinity for KLC2. Indeed, a ten amino acid long peptide centered on the WD motif (SKIP^WD, Figure 1A) bound to KLC2^TPR with a K_D of 24 μM whereas the affinity of the equivalent SKIP^WE peptide was above 110 μM (Fig. 1C). The presence of both motifs improves the affinity for KLC2^TPR because a 40 amino acid long peptide SKIP^WDWE which encompasses the W-acid pair sequence bound with an apparent affinity higher than that of the single SKIP^WD motif (K_D = 4.9 μM). Binding affinity measurements at varying NaCl concentrations showed that electrostatic interactions play an important role in the recognition process (Fig. S1). For structural studies we focused on the SKIP^WD motif and to facilitate the crystallization process, we engineered a chimeric construct in which the SKIP^WD peptide was fused N-terminal to KLC2^TPR via a flexible (TGS)₄ linker (18). Crystals of SKIP^WD-KLC2^TPR grew using vapour diffusion techniques and diffracted at 2.9 Å using synchrotron radiation. The final model of the cargo complex is characterized by R and R_free value of 20.3 and 24.5%, respectively (Table S1). KLC2^TPR consists of six TPR repeats (TPR1-6) each contributed by a classical helix-turn-helix structural motif arranged in a right-handed super-helical conformation with a non-TPR helix is positioned between TPR5 and TPR6 (14). The structure of SKIP^WD-KLC2^TPR revealed the cargo peptide bound in an extended conformation at the N-terminal portion of KLC2^TPR concave surface with a direction parallel to the external helix of the repeat (Fig. 2A,B and Fig. S2). Although our construct includes the external helix of TPR1, this region, as well as other flexible stretches, could not be unambiguously interpreted in the electron density. Thus, KLC2^TPR starts from TPR2 in our model. A comparison between apo KLC2^TPR (3CEQ) and cargo-bound KLC2^TPR revealed structural differences considered to arise from a rigid jaw movement of the N-terminal TPR2-3 region which closes upon cargo recognition engendering the binding surface and pockets for the SKIP^WD peptide. (Fig. 2B, S3). An analysis using the PISA algorithm (19) indicates that all residues of the SKIP^WD peptide are involved in formation of the complex, which is stabilized by residues from TPR2-3 and the internal helix of TPR4 (Fig. 2C). This concave groove surface displays a positive electrostatic charge ideally poised to complement the negatively charged W-acidic cargo motifs (Figure 2C and Fig. S4). Overall, the interface area of the SKIP^WD-KLC2^TPR complex is about 770 Ų, stabilized by a mixture of H-bonds, salt bridges and hydrophobic interactions (Figure 3A). The W residue central to the motif is generally flanked by amino acids bearing a carboxylate side-chain. SKIP^W207 (position 0, p0) is buried within a leucine-rich pocket positioned roughly in the middle along the TPR length and contributed by side-chains from TPR2 (L248, R251, L263) and TPR3 (N287, L290, L291). In particular, the side-chain of N287 is in a conformation such that it serves the dual purpose of stabilizing SKIP^W207 indole group by lining one side of the pocket whilst engaging at the same time in hydrogen bonds with the main chain amide and carbonyl oxygen of the residue at p+1. The latter position is almost invariably occupied by a glutamic or aspartic acid residue (7). The carboxylate side-chain of SKIP^D208 points in the opposite direction to that of SKIP^W207 engaging in a network of salt bridges and H-bonds with the positively charged R312 and K325 side-chains of α3E and α4I, respectively. An acidic side-chain at position p−1 (SKIP^E206) of the motif is also very conserved in other W-acidic motifs. Like SKIP^D208, the carboxylate side-chain of SKIP^E206 faces the TPR3-4 side of the KLC2^TPR recognition groove where it is stabilized by an ionic interaction with K325. Position p-2 of SKIP^WD features a leucine residue (SKIP^L205). The hydrophobic side chain of SKIP^L205 is deeply buried in a hydrophobic pocket formed by residues from TPR3 and 4. The side-chain of N329 plays a similar dual role as N287. Whilst lining the SKIP^L205 pocket it also H-bonds with the main chain at position p−1. Together, N287 and N329 act like a clamp on opposite sides of SKIP^WD providing four of the ten hydrogen bonds that stabilize the complex. Residues at positions p-(3,4) are less important for complex stability consistent with their general lack of sequence conservation amongst W-acidic motif cargo (7). The C-terminal stretch of SKIP^WD encompassing p+(2,3,4,5) is observed in a more compact, turn-like conformation with SKIP^D209,S210 at p+(2,3) making very minor contacts with the groove. As for the N-terminal peripheral region of SKIP^WD the exact nature of the amino acids at p+(4,5) does not seem critical for complex stability. Alternative side-chains can be positioned at this topological position possibly involving a rearrangement of the main-chain. To investigate the effect of amino acid replacements within the KLC2^TPR region on the interaction with SKIP we used immunoprecipitation as described above. All amino acid substitutions tested resulted in abrogation or near abrogation of complex formation between KLC2 and SKIP (Fig 3B,C). The importance of electrostatic interactions in cargo recognition is underscored by charge-reversal mutations (Fig 3B,C). When co-expressed in cells, SKIP and its small GTPase binding partner Arl8 associate strongly with lysosomes and promote their trafficking to the cell periphery. GFP-KLC2 associates with the same lysosomes (Fig. 3D, S5) (13). This is essentially abolished by KLC2 mutations (R251D, N287L and R312E) (Fig. 3D, S5). The same mutations strongly inhibited the binding of KLC2^TPR to the WD motif peptide (Fig. 3C). We conclude that optimal SKIP:KLC2 stability critically depends on a very conserved cargo recognition-interaction groove. We have extended our analysis to the well characterized Calsyntenin (CSTN) cargo which also exhibits two W-acidic motifs (9, 11, 12). Differently from SKIP, both CSTN motifs bind to KLC2^TPR with similar affinity (Fig S6). However, the mutations in the KLC2 binding groove which disrupt SKIP^WD binding also abrogate recognition of CSTN (Fig S6). A sequence alignment shows total conservation of the KLC^TPR residues interacting with the cargo across the kinesin light chain family (Fig. S7). Thus, the concave groove within the TPR domain where the SKIP^WD peptide binds is likely to be the primary site of interaction for W-acidic cargo motifs in general. However, we cannot exclude that secondary sites also exist. In the context of intracellular transport where kinesin-1 functions as a tetramer containing two KLCs held together by a coiled-coil region and both SKIP motifs contribute to transport (7, 13), it is tempting to speculate that both chains can contribute to the binding of the W-acidic cargo pair. For SKIP, the higher affinity WD motif would direct the first binding event to one of the KLC^TPR with avidity effect promoting the association of the WE motif to the other KLC (20, 21). It will be important to determine this relationship to the kinesin-1 tetramer and examine the importance of cargo induced TPR conformational change in motor activation.

Figure 1.

Binding of SifA-kinesin binding protein (SKIP) to kinesin light chain 2 (KLC2). (A) Scheme of SKIP and KLC2. The two W-acidic motifs and KLC2 TPR repeats are highlighted. (B) Co-immunoprecipitation shows that alanine replacement in the first W-acidic motif (WD) abrogates SKIP:KLC2 binding whereas the same substitution in the second motif (WE) has virtually no effect. (C) Fluorescence polarization measurements confirm that the first SKIP W-acidic motif (SKIP^WD) has a higher affinity for the TPR domain of KLC2 than the second one (SKIP^WE). A peptide encompassing both motifs (SKIP^WDWE) binds with higher affinity than the best single motif. K_D values reported here were calculated at 150 mM NaCl. Binding affinity values at varying ionic strength values are given in Fig. S1.

Figure 2.

Structure of the SKIP^WD-KLC2^TPR cargo complex. (A,B) Cartoon representations of the SKIP^WD motif (displayed as stick model in green) bound to KLC2^TPR domain (orange cartoon) in two orthogonal orientations. Simulated annealing (F_o-F_c) omit map for the W-acidic cargo motif is contoured at the 3σ level. Individual TPR repeats composed by helix-turn-helix elements (I and E for internal and external, respectively) are highlighted in A. A non-TPR helix (αN) is between TPR5 and TPR6. Panel B also shows the apo KLC2^TPR structure (grey transparent cartoon) with an orange arrow indicating the movement of the TPR2-3 region with respect to the common TPR4-6 reference frame. (C) Electrostatic potential surface representation of KLC2^TPR with its SKIP^WD-bound cargo. Positive and negative potential is shown in blue and red, respectively. Cargo recognition is achieved by a combination of charge complementarity and sequence specificity.

Figure 3.

The SKIP^WD-KLC2^TPR interface and the effect of KLC2^TPR mutations in cargo binding and cellular recruitment. (A) Details of the SKIP^WD:KLC2^TPR interface. KLC^TRP side-chains stabilizing the SKIP^WD cargo peptide (green) are shown as grey sticks emanating from the orange cartoon. Non-carbon elements are: nitrogen, blue; oxygen, red; sulfur, yellow. Hydrogen bonds are represented by dotted cyan lines. (B) Co-immunoprecipitation assay showing the effect of KLC2^TPR mutations at the SKIP^WD:KLC2^TPR interface on the interaction. (C) Fluorescence polarization measurements showing that R251D, N287L and R312E mutations in KLC2 dramatically reduce the affinity of the TPR domain for the SKIP^WD peptide. K_D values reported here were calculated at 150 mM NaCl. (D) Replacement of key KLC2^TPR residues results in loss of GFP-KLC2 association with Arl8/SKIP positive lysosomal membranes. Scale bar shows 10μm. In merge panels, GFP-KLC2, Arl8 and SKIP are shown in green, red and blue respectively.