Sunday Driver/JIP3 binds kinesin heavy chain directly and enhances its motility

Abstract

Neuronal development, function and repair critically depend on axonal transport of vesicles and protein complexes, which is mediated in part by the molecular motor kinesin-1. Adaptor proteins recruit kinesin-1 to vesicles via direct association with kinesin heavy chain (KHC), the force-generating component, or via the accessory light chain (KLC). Binding of adaptors to the motor is believed to engage the motor for microtubule-based transport. We report that the adaptor protein Sunday Driver (syd, also known as JIP3 or JSAP1) interacts directly with KHC, in addition to and independently of its known interaction with KLC. Using an in vitro motility assay, we show that syd activates KHC for transport and enhances its motility, increasing both KHC velocity and run length. syd binding to KHC is functional in neurons, as syd mutants that bind KHC but not KLC are transported to axons and dendrites similarly to wild-type syd. This transport does not rely on syd oligomerization with itself or other JIP family members. These results establish syd as a positive regulator of kinesin activity and motility.

Figure 1

syd interacts with KHC independently of KLC. (A) Schematic illustration of full-length or truncated syd constructs used. Domains of known function located within the amino-terminal portion of syd (N-syd) are indicated. cc: coiled-coil domains; JBD: JNK binding domain; LZ: leucine-zipper domain (KLC binding domain). (B) Mouse brain lysate was used in pulldown experiments using recombinant GST, GST N-syd, or GST C-syd. Western blot analysis was performed with the indicated antibodies. VAMP3 was used as a negative control. Both KHC and KLC are pulled down with GST N-syd, but not with GST C-syd. (C) As in (B), but the indicated N-syd deletion mutants were used. N-syd lacking the KLC binding site (N-sydΔLZ) interacts with KHC but not KLC. The asterisk points to non-specific reaction with molecular marker loaded in this lane. (D) As in (B), but the indicated syd fragments were used. The syd fragment aa3–239, which does not contain the LZ domain is sufficient to bind KHC. Densitometry analysis of ponceau staining for (C) and (D) shows the relative amount of recombinant GST proteins used in each condition. Input for (B–D) is 10% of total starting material. (E) Active microtubule-bound kinesin-1 was released by ATP. The soluble kinesin-1 fraction was collected and separated by sucrose density gradient. The fractions were collected and analysed by western blot on two separate gels. The black separating line marks where separate gels have been spliced together. The asterisk points to non-specific crossreacting bands. A population of KHC is detected in lighter fractions at the top of the gradient, which does not contain KLC.

Figure 2

Mapping the syd domain necessary for KHC interaction. (A) Scheme of the constructs used in (B), (C), and E). (B) GST-pulldown analysis was conducted as described in Figure 1B. Residues 50–80 are required for syd interaction with KHC. VAMP3 is used as a negative control. (C) The indicated N-syd deletion mutants were used in a GST-pulldown assay and analysed by western blot with the indicated antibodies. Deletion of both LZ (KLC binding domain) and KBD (KHC binding domain) was required to prevent the syd–kinesin-1 interaction. (D) Recombinant GST–KHC tail (aa 807–956) was used in a GST-pulldown assay. The GST–KHC tail, but not GST, interacts with syd. Densitometry analysis of ponceau staining for (B–D) shows the relative amount of recombinant GST proteins used in each condition. Input is 10% of total starting material. (E) The indicated purified His-tagged syd fragments were incubated with purified GST or the GST–KHC tail. Direct binding of syd to the KHC tail domain was assessed using an anti-polyhistidine antibody. (F, G) Fixed amounts of recombinant GST N-syd (50 nM) were incubated with His–KLC1-TPR (F) or with His–KHC tail (G) at the indicated concentrations. The amount of bound His–KLC1-TPR or His–KHC tail was analysed by western blot with an anti-polyhistidine antibody and quantified using ImageJ. To estimate the dissociation constant, we plotted the band intensity for each concentration and used the first order binding equation: fraction of bound kinesin=[kinesin]/(Kd+[kinesin]), to fit the data generated.

Figure 3

syd binding to KHC is functional. TIRF microscopy was used to visualize the KHC-dependent transport of GFP–syd along microtubules in vitro. (A) Scheme of the experimental procedures. COS-7 cells were transfected with the indicated GFP–syd constructs or Flag–KHC. Extracts were prepared and mixed just before TIRF assay. Non-transfected extract was used as a control for KHC-dependent movement. (B) Selected images from movie recorded with GFP–syd wt in the presence of KHC. The arrowheads mark the GFP–syd wt molecule movement on a microtubule over time (in seconds), bar: 2 μm. (C) Cell lysate expressing the indicated GFP–syd construct was mixed with lysate expressing Flag–KHC or non-transfected lysate as a control. Kymographs were generated from movies recorded for the indicated GFP–syd construct in the presence or absence of Flag–KHC. Vertical bar=50 s, horizontal bar=2 μm. (D) The proportion of motile versus non-motile events observed for each condition is shown. The difference in the probability for being motile or non-motile in each condition was analysed with χ² analysis. Data represent results of 3–4 independent experiments. One asterisk, P<0.05; compared with GFP–syd wt with KHC. (E) Calculated transport velocity for each condition. Velocity values represent mean±s.e.m. of 15–101 motile events from 3 to 4 independent experiments. Compared with GFP–syd wt with KHC (Student's t-test), three asterisks, P<0.001. (F) Calculated run length for each condition. Run length values represent mean±s.e.m. of 15–66 motile events from 3 to 4 independent experiments. Two asterisks P<0.01; compared with GFP–syd wt with KHC.

Figure 4

syd regulates kinesin-1 activity and motility. (A) Scheme of the experimental procedures. COS-7 cells were transfected with the indicated Flag–syd constructs or fluorescently tagged KHC at the C-terminus (KHC–mCit). (B) Kymographs were generated from movies recorded for KHC–mCit in the presence or absence of Flag–syd wt or in the presence of Flag–sydΔKBD or Flag–syd3–239. Vertical bar=20 s, horizontal bar=5 μm. (C) The proportion of motile versus non-motile events observed for each condition is shown. The difference in the probability for being motile or non-motile in each condition was analysed with χ² analysis. (D, E) Velocity (D) and run length (E) were calculated from the kymographs generated. Velocity and run length values represent mean±s.e.m. of n=52 for KHC–mCit alone, n=358 for KHC–mCit with Flag–syd wt, n=76 for KHC–mCit with Flag–sydΔKBD, n=191 for KHC–mCit with Flag–syd3–239. Data were collected from 3 to 4 independent experiments. Asterisks (^*): compared with non-transfected condition (^**P<0.01, ^***P<0.001). Pound (#), compared with each other (^###P<0.001). (F, G) Raw distribution data are shown for velocity (F) and run length (G).

Figure 5

syd interaction with KHC is functional in neurons and mediates transport to both dendrites and axons. (A) E18 hippocampal neurons were stained for endogenous syd and the axonal marker tau. syd accumulates at the axonal tip. (B, C) Hippocampal neurons were transfected with the indicated GFP–syd constructs and stained with the axonal marker tau in (B) or the dendritic marker MAP2 in (C). Deletion of both LZ and KBD is required to prevent syd transport to axonal tips and proximal dendrites. (D) Deletion of both kinesin binding sites prevents exit from the cell body and leads to accumulation of syd in proximity of the Golgi apparatus stained with giantin antibody. (E) Quantification of syd accumulation in axon or dendritic tips. The ratio of GFP intensities between axon or dendrite tips and the somatic region was measured for each construct. Values are expressed as mean±s.e.m. (n=12–18), ^***P<0.001. Statistical difference was analysed for axon or dendrites categories between all conditions with the Student's t-test. (A–C) Bar=50 μm, except that bar=10 μm for high magnification (high mag.); arrows: axon tips; DAPI: nuclear marker.

Figure 6

KHC-dependent syd transport does not depend on oligomerization with endogenous JIP family members. (A) Cell lysates were prepared from N2A cells transfected with Flag–syd together with GFP, GFP–syd3–239, GFP–syd wt, or GFP–sydΔΔ. Immunoprecipitation was performed with GFP antibodies or rabbit IgG as a negative control; western blots were probed with Flag and GFP antibodies. GFP–syd3–239 and GFP–sydΔΔ retain the ability to homo-oligomerize. (B) E18 cortical neurons from syd +/+ or syd−/− pups were stained with syd and tau, confirming the lack of expression in syd−/− neurons. (C) E18 brain lysates from wt (syd+/+) or syd knockout (syd−/−) were analysed by western blot. β-Tubulin is used as a loading control. (D) The ratio of fluorescence intensities between axon tips and the somatic region was measured for endogenous syd (stained with anti-syd antibody in syd+/+ neurons) and for the indicated GFP-constructs (GFP intensity in syd−/− neurons). (E, F) syd−/− cortical neurons were transfected with mCherry–syd3–239 (E) or GFP–sydΔLZ (F) and stained with the axon marker tau. Both constructs were transported to the axon tip. (G) Cell lysates were prepared from N2A cells transfected with Flag-tagged JIP2 together with GFP, GFP–syd3–239, GFP–syd wt, or GFP–sydΔΔ. Immunoprecipitation was performed with GFP antibodies or rabbit IgG as a negative control; western blots were probed with Flag and GFP antibodies. GFP–syd3–239 does not hetero-oligomerize with JIP2. (H) Cell lysates were prepared from N2A cells transfected with myc-tagged JIP1 together with Flag–syd wt, Flag–syd3–239, or Flag–sydΔΔ. Immunoprecipitation was performed with Flag antibodies or rabbit IgG as a negative control; western blots were probed with Flag and myc antibodies. All syd constructs oligomerize with JIP1. Arrows: axon tips; arrowheads: cell body. (B–E) Bar=50 μm, except that bar=10 μm for high magnification (high mag.).