Dynein and dynactin leverage their bivalent character to form a high-affinity interaction

Abstract

Cytoplasmic dynein and dynactin participate in retrograde transport of organelles, checkpoint signaling and cell division. The principal subunits that mediate this interaction are the dynein intermediate chain (IC) and the dynactin p150(Glued); however, the interface and mechanism that regulates this interaction remains poorly defined. Herein, we use multiple methods to show the N-terminus of mammalian dynein IC, residues 10-44, is sufficient for binding p150(Glued). Consistent with this mapping, monoclonal antibodies that antagonize the dynein-dynactin interaction also bind to this region of the IC. Furthermore, double and triple alanine point mutations spanning residues 6 to 19 in the yeast IC homolog, Pac11, produce significant defects in spindle positioning. Using the same methods we show residues 381 to 530 of p150(Glued) form a minimal fragment that binds to the dynein IC. Sedimentation equilibrium experiments indicate that these individual fragments are predominantly monomeric, but admixtures of the IC and p150(Glued) fragments produce a 2:2 complex. This tetrameric complex is sensitive to salt, temperature and pH, suggesting that the binding is dominated by electrostatic interactions. Finally, circular dichroism (CD) experiments indicate that the N-terminus of the IC is disordered and becomes ordered upon binding p150(Glued). Taken together, the data indicate that the dynein-dynactin interaction proceeds through a disorder-to-order transition, leveraging its bivalent-bivalent character to form a high affinity, but readily reversible interaction.

Figure 1. Dynein and dynactin schematic.

A. Dynein and dynactin macromolecular complex organization. Both complexes are drawn approximately to scale from EM reconstruction images. B. Schematic of domain structure and binding sites of both the dynein intermediate chain isoform 2C and dynactin p150^Glued. The dynactin binding domain, spanning residues 10–44 (indicated by a black line), is within the predicted N-terminal coiled-coil domain of the IC, residues 1–45. The intermediate chain binding interface has been mapped to fragments spanning residues 217–548, denoted CC1, or 600–811 (indicated by a black line).

Figure 2. Dynein intermediate chain minimal binding domain.

A. Native PAGE indicates that residues 10–44 are sufficient for binding to p150^Glued CC1. A gel shift indicates IC2C fragments spanning residues 1–124, 10–124 and 1–44 are capable of binding CC1 (indicated by arrows). However, fragments spanning 20–124 and 1–32 are not able to bind to CC1. It is important to note that due to the large negative charge of the IC some constructs to not enter the gel. B. Epitope mapping of IC antibodies: The epitopes of α-IC mAb 70.1 and 74.1 are located within the p150^Glued binding domain. Specifically, α-IC 70.1 recognizes the region between residues 1–18 and α-IC 74.1 recognizes 10–30.

Figure 3. Alanine scanning mutagenesis of Pac11.

A. Sequence alignment of R. norvegicus IC and S. cerevisiae Pac11 (IC). Alanine point mutations were introduced into Pac11 (indicated in red). A (*) indicates identical amino acids, (∶) indicates highly conserved, similar amino acids, (.) indicates amino acids that are somewhat similar and blank indicates dissimilar amino acids. B. Native PAGE indicates that S. cerevisiae wt-Pac11^1–86 is capable of binding p150^Glued CC1B alone (indicated by the loss of CC1B) and in the presence of dynein light chain 8 (LC8) (indicated by the gel shift of LC8). Due to the charge and hydrodynamic properties of wt-Pac11^1–86, it does not enter the native PAGE. An asterisk indicates gel shift upon wt-Pac11^1–86-LC8-CC1B binding. Pac11^1–86 triple point mutant Pac11-R12A,Q13A,L14A (12AAA) is unable to bind p150^Glued CC1B alone or in the presence of LC8. Arrow indicates formation of Pac11-R12A,Q13A,L14A-LC8 (12AAA) binding. No gel shift occurs upon addition of CC1B. C. Size exclusion chromatography indicates wt-Pac11^1–86 forms a complex with LC8 and CC1B (purple), with an elution volume of 8.07 mL, while Pac11-R12A,Q13A,L14A (12AAA) is unable to bind to CC1B (red). The Pac11-R12A,Q13A,L14A(12AAA)-LC8 complex elutes at 8.97 mL and CC1B elutes at 11.45 mL. D. Spindle positioning in wild-type and mutant cells expressing GFP-labeled microtubules. The percentage of cells exhibiting spindle position defects (see Materials and Methods) was determined for wild-type (yJC5919), p150 ^Glued/nip100Δ (yJC6047), DIC/pac11Δ (yJC6354), DIC/pac11-4A (L4A,K5A,Q6A, yJC6916), pac11-6A (Q6A,L7A,E8A, yJC6917), pac11-9A (E9A,K10A,R11A, yJC6918), pac11-12A (R12A,Q13A,L14A, yJC6846), pac11-17A (L17A,R18A, yJC6847), and pac11-19A (E19A,R20A,R21A, yJC6919) strains. Error bars denote SEM. P-values are shown in Fig. S3B.

Figure 4. p150^Glued minimal binding domain.

A. Native PAGE indicates that residues 381–530 and 415–530 of p150^Glued are sufficient to bind to the intermediate chain as shown by the gel shift (indicated by arrows) upon incubation of p150^Glued with IC^1–44. Each binding experiment, p150^Glued fragment + IC, was performed on an individual gel. Lanes are representative of individual binding experiments. (B–D) Size exclusion chromatography of p150^Glued fragments in complex with IC2B^1–158-TcTex1-LC8. B. CC1, IC2B^1–158-TcTex1-LC8 alone or in complex. C. CC1A, IC2B^1–158-TcTex1-LC8 alone or in complex. D. CC1B, IC2B^1–158-TcTex1-LC8 alone or in complex. The change in elution volume of the IC complex is 2.22, 0.08 and 1.26 ml in the presence of CC1, CC1A and CC1B, respectively.

Figure 5. Biophysical characteristics of intermediate chain fragments.

A. Cartoon of two potential mechanisms to create a dimer-of-dimers through a coiled coil. Magenta highlights the IC binding interface on CC1. Blue indicates the CC1 binding site on the IC. The top route assumes that the IC is homodimer before binding to CC1. The bottom route assumes that the IC is monomeric before binding the CC1. B. Circular dichroism spectra of 20 μM IC^1–44 in 1− PBS 1 mM TCEP at 4°C with increasing concentrations of trifluoroethanol results in a minimal increase of α–helical characteristics. C. Thermal denaturation of 20 μM IC^1–44 and IC^1–124. Thermal denaturation was monitored at 222 nm. D/E. Sedimentation equilibrium: sedimentation equilibrium analytical ultracentrifugation (SE-AUC) at 10000, 20000, 30000, 40000 rpm and 20°C show that both IC^1–44 (C) and IC^1–124 (D) are monomeric.

Figure 6. Biophysical characteristics of p150^Glued fragments.

A. Circular dichroism spectra of 5 μM p150^Glued CC1, CC1A and CC1B fragments in 1 x PBS 1 mM TCEP indicate they are predominantly helical at 4°C. B. Thermal denaturation of 10 μM p150^Glued CC1, CC1A and CC1B. Thermal denaturation was monitored at 222 nm. C–E. Sedimentation equilibrium: SE-AUC at 20000, 25000, 30000, 35000 rpm and 4°C show that all three fragments, CC1 (25000, 30000 and 35000 rpm only) (C), CC1A (D) and CC1B (E) are dimeric.

Figure 7. Stoichiometry of the IC-p150^Glued interaction.

A–C. Sedimentation equilibrium: A. SE-AUC at 12000, 16000 and 20000 and 4°C of CC1 and IC^1–124. B–C. SE-AUC at 10000, 20000, and 30000 rpm and 4°C of CC1A (B) and CC1B (C) with IC^1–124. Both CC1 and CC1B associate much more strongly with IC^1–124 than CC1A.

Figure 8. Salt and pH dependence of IC-p150^Glued interaction.

A. Sedimentation equilibrium: SE-AUC of CC1, IC^1–124 and the CC1-IC^1–124 complex in the presence of 0, 50, 100, 250, 500 mM and 1.0 M sodium chloride. No change in the oligomeric state of either CC1 or IC^1–124 occurs with increasing salt (inset). The CC1-IC^1–124 interaction is strongest at 100 mM sodium chloride and decreases upon increasing salt concentration (fitting analysis is shown in Fig. S8). B. SE-AUC of CC1, IC^1–124 and the CC1-IC^1–124 complex was run at pH 6.0, 7.0, 8.0 and 9.0. No change in oligomeric state is seen in either CC1 or IC^1–124. (asterisk denotes that CC1 precipitates at pH 6.0). A strong pH dependence is seen for formation of the CC1-IC^1–124 complex, where the interaction is the strongest at pH 8.0 and weaker at pH 7.0 and 9.0.

Figure 9. Proposed mechanism of dynein-dynactin binding.

Based on the refinement of the dynein-dynactin interaction and the physical characterization of the individual fragments, we propose that the N-terminus of the dynein IC is disordered in the in the absence of dynactin. Further, based on recent data from our lab and others we propose that the dynein LCs affect the N-termini by reducing and/or orienting the N-termini of the IC to optimally bind to dynactin. Taken together, we suggest the following model: A. The N-terminus of the intermediate chain exists in a disordered state. B. Upon IC binding to the light chains the radius of gyration of the disordered region is reduced. C. The N-termini of the IC bind to p150^Glued CC1B located in the shoulder.