A novel split kinesin assay identifies motor proteins that interact with distinct vesicle populations

Abstract

Identifying the kinesin motors that interact with different vesicle populations is a longstanding and challenging problem with implications for many aspects of cell biology. Here we introduce a new live-cell assay to assess kinesin-vesicle interactions and use it to identify kinesins that bind to vesicles undergoing dendrite-selective transport in cultured hippocampal neurons. We prepared a library of "split kinesins," comprising an axon-selective kinesin motor domain and a series of kinesin tail domains that can attach to their native vesicles; when the split kinesins were assembled by chemical dimerization, bound vesicles were misdirected into the axon. This method provided highly specific results, showing that three Kinesin-3 family members-KIF1A, KIF13A, and KIF13B-interacted with dendritic vesicle populations. This experimental paradigm allows a systematic approach to evaluate motor-vesicle interactions in living cells.

Figure 1.

Linking a kinesin tail–vesicle complex to an active axon-selective motor domain causes a distinctive increase in axonal vesicle transport. (A) Hippocampal neurons were transfected with a split kinesin consisting of an FRB-GFP-KIF1A tail, which interacts with endogenous vesicles, and the KIFC⁵⁵⁹-tdTM-FKBP motor domain. The constructs were expressed for 18 h before live imaging to evaluate the transport of KIF1A-labeled vesicles. (B) Before addition of linker drug, the FRB-GFP-KIF1A–labeled vesicles were present in both the axon and the dendrites. In this and all subsequent figures, the contrast was inverted so that brightly labeled vesicles appear dark. Bars: (top panels) 20 µm; (high magnification panels) 5 µm. (C and D) Kymographs illustrate the transport of KIF1A-labeled vesicles before (0 min) and 25 min after addition of 1 µM AP 21967. Before addition of the linker, labeled vesicles were transported bi-directionally in the axon and the dendrites. After addition of the linker drug, there was a pronounced increase in long-range anterograde events in the axon. Graphs with red lines illustrate all anterograde events visible on the corresponding kymographs. In the kymographs, time is shown on the x axis and position along the neurite on the y axis. Diagonal lines with positive slope represent movements away from the cell body. Time and distance calibration are the same for all kymographs. See also Video 1.

Figure 2.

Transport characteristics of vesicles labeled by expressing full-length GFP-tagged Kinesin-3 motors. Fluorescently labeled full-length kinesins were expressed in cultured hippocampal neurons to label vesicles. (A–D) Representative images of axons and dendrites of hippocampal neurons expressing GFP-tagged KIF1A, KIF1Bβ, KIF13A, and KIF13B, respectively. Boxed regions showing high magnification views of a portion of the axon and a dendrite are shown below. Kymographs illustrating vesicle movements in the boxed regions show that vesicles labeled with KIF1A, KIF1Bβ, and KIF13A were transported bi-directionally in both axons and dendrites. In contrast, vesicles labeled with KIF13B were largely polarized to the dendrites. Bars: (top panels) 20 µm; (high magnification panels) 5 µm. Time and distance calibration are the same for all kymographs.

Figure 3.

Schematic diagram illustrating the split kinesin assay. (A) Three constructs are expressed together: a vesicle marker labeled with GFP, a constitutively active axon-selective kinesin motor domain tagged with tdTomato and fused to FKBP (KIF5C⁵⁵⁹-tdTM-FKBP), and a myc-tagged kinesin tail domain fused to FRB. The kinesin tail domain binds its endogenous cargo vesicle, but co-assembles with the kinesin motor domain only after addition of the linker drug. (B) In the absence of linker drug, endogenous motor proteins transport GFP-labeled vesicles in dendrites, but they do not enter the axon. The constitutively active motor domain (black arrows) translocates toward the axonal tip, but does not bind cargo. After the linker drug is added, the split kinesin is assembled, but without interaction between the kinesin tail and the labeled vesicle population, this does not result in any changes in the transport behavior of the vesicles. (C) In case of a positive interaction between the tail and the vesicle, the tail is able to bind the vesicle population immediately after expression. After the addition of linker drug the split kinesin is assembled on the labeled vesicle. (D) Upon the addition of the linker drug, GFP-labeled vesicles that bind the expressed kinesin tail become attached to the constitutively active axonal KIF5C⁵⁵⁹-FKBP motor domain, which transports them into the axon.

Figure 4.

Localization and level of expression of the split kinesin components. (A and B) A cell co-expressing a split kinesin tail (FRB-myc-KIF13B tail) and the KIF5C⁵⁵⁹-tdTM-FKBP motor domain. Immunostaining for the kinesin tail was present throughout the cell body and dendrites and extended far into the axon (arrows). The motor domain was concentrated at the tips of axonal branches (arrowheads) but not at dendritic tips (asterisks). (C) Comparison of the expression levels of the different split kinesin tails (based on anti-myc immunostaining). Average intensity of staining in the cell body was measured for at least 15 cells from two separate experiments in each condition. Bar, 20 µm.

Figure 5.

KIF13A and KIF13B tails bind transferrin receptor vesicles. (A) A schematic showing the three constructs expressed in this assay before and after assembly of the split kinesin. (B) Kymographs showing the transport of TfR vesicles in dendrites before assembly of the split kinesin in two different dendrites. (C) Kymographs showing the transport of TfR vesicles in the axon before and at varying times after adding the linker drug (AP 21967, 1 µM) in a cell expressing the FRB-KIF13B tail. Before adding the linker drug there was far less vesicle transport in the axon than the dendrites (compare with B). After drug-induced assembly of the split kinesin there was a pronounced increase in long-range anterograde vesicle transport in the axon. Time and distance calibration are the same for kymographs in B and C. (D) Images showing the cell body and proximal axon of the neuron imaged immediately before (0 min) and after 16 min of treatment with linker drug. Note the increase in intensity of TfR-GFP in the axon after 16 min. Bar, 20 µm. (E–L) Kymographs illustrating axonal transport of TfR-GFP vesicles in hippocampal neurons expressing different split kinesin tail constructs. The kymographs show the transport of TfR vesicles in the axon before (0 min) and 14–28 min after addition of the linker drug. There was no change in the overall transport of TfR vesicles when KIF1A, KIF1Bα, KIF1Bβ, KIF5C, KIF17, or KIF21B tails were expressed. In contrast, there was a large increase in the long-range anterograde transport events of TfR vesicles when KIF13A (H) or KIF13B (I) tails were used. See also Video 2. Time and distance calibrations are the same for kymographs in E–L.

Figure 6.

Quantification of changes in TfR vesicle traffic in neurons expressing different split kinesin tails. The figure plots the difference in the amount of anterograde transport before and 15–30 min after adding linker drug (number of anterograde events after drug minus number of anterograde events before drug; mean ± SEM). A statistically significant increase in axonal transport of TfR vesicles was observed in cells expressing FRB-KIF13A and FBR-KIF13B tails (Wilcoxon signed rank test; P < 0.01). n = 7–16 cells per condition.

Figure 7.

KIF1A and KIF13B tails bind low-density lipoprotein receptor vesicles. (A) The three constructs expressed in this assay before and after assembly of the split kinesin. (B–I) Kymographs illustrating axonal transport of LDLR-GFP vesicles in hippocampal neurons expressing different split kinesin tail constructs. The kymographs show the transport of LDLR vesicles in the axon before (0 min) and 13–21 min after addition of the linker drug. There was no change in the overall transport of TfR vesicles when KIF1Bα, KIF1Bβ, KIF13A, KIF5C, KIF17, or KIF21B tails were expressed. In contrast, there was a large increase in the long-range anterograde transport events of LDLR vesicles when KIF1A (B) or KIF13B (F) tails were used. (J) The difference in the amount of anterograde transport before and 15–30 min after adding linker drug (number of anterograde events after drug minus number of anterograde events before drug; mean ± SEM). A statistically significant increase in axonal transport of LDLR vesicles was observed in cells expressing FRB-KIF1A and FRB-KIF13B tails (Wilcoxon signed rank test; P < 0.01). n = 7–11 cells per condition. Time and distance calibration are the same for all kymographs.

Figure 8.

Accumulation of misdirected TfR in the axon. In cells expressing TfR-GFP, a KIF13B split kinesin, and soluble eBFP2 (to enable visualization of the entire axonal and dendritic arbor), incubation with the linker drug for 4 h significantly increased axonal TfR. To detect the TfR that reached the cell surface, live staining was performed using a monoclonal antibody against the extracellular GFP tag. (A) Representative control and treated cells showing the eBFP2 fill, TfR-GFP, and TfR that could be labeled from the cell surface in living cells. Note the prominent TfR fluorescence present in the treated cell. Bar, 20 µm. Boxed regions of the axon are shown at high magnification below. High magnification bar, 5 µm. (B) Line scans of the boxed regions of the axons in A show a pronounced increase in total TfR-GFP (left) and in TfR that reached the cell surface (right). A.U., arbitrary units. (C) Dot plots showing the average fluorescence intensity of TfR-GFP (left) and of TfR that could be labeled from the cell surface (right) in the axons of control cells and cells exposed to the linker drug (1 µM AP21967 for 4 h). Both total TfR and TfR that was accessible to extracellular antibody were significantly increased (t test, P < 0.001 and P < 0.005, respectively). Each point represents one cell; horizontal bars show means and SDs.