SEARCH

SEARCH BY CITATION

Keywords:

  • dynein;
  • endosomes;
  • kinesin-1;
  • kinesin-2;
  • lysosomes;
  • microtubules

Abstract

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Important progress has been made during the past decade in the identification of molecular motors required in the distribution of early and late endosomes and the proper trafficking along the endocytic pathway. There is little direct evidence, however, that these motors drive movement of the endosomes. To evaluate the contributions of kinesin-1, dynein and kinesin-2 to the movement of early and late endosomes along microtubules, we made use of a cytosol-free motility assay using magnetically isolated early and late endosomes as well as biochemical analyses and live-cell imaging. By making use of specific antibodies, we confirmed that kinesin-1 and dynein move early endosomes and we found that kinesin-2 moves both early and late endosomes in the cell-free assay. Unexpectedly, dynein did not move late endosomes in the cell-free assay. We provide evidence from disruption of dynein function and latrunculin A treatment, suggesting that dynein regulates late endosome movement indirectly, possibly through a mechanism involving the actin cytoskeleton. These data provide new insights into the complex regulation of endosomes’ motility and suggest that dynein is not the major motor required to move late endosomes toward the minus end of microtubules.

The distribution of endocytic compartments in space and time governs both their biogenesis and their functions. Early endosomes receive cargo by fusing with vesicles arriving from the plasma membrane. As these endosomes grow, they gradually concentrate cargo destined for late endosomes while also recycling other cargo to the plasma membrane. Large early endosomes then move to the perinuclear region where they are converted into late endosomes (1). Finally, these late endosomes fuse with lysosomes that are also found in this region, so delivering their cargo for processing by these degradative organelles.

The subcellular locations of early and late endosomes and lysosomes, and their movements, result from their interactions with the actin and microtubule networks and the forces generated by actin- and microtubule-dependent motors. Research from several teams has produced a broad consensus that cytoplasmic dynein moves early and late endosomes and lysosomes toward the minus ends of microtubules and kinesin-1 moves them toward the plus ends (2–9). In addition, Schroer’s group has reported that kinesin-2 is required for the normal steady-state location of late endosomes/lysosomes but not early endosomes or recycling endosomes, suggesting that kinesin-2 moves late but not early endosomes (10). These three motors are protein complexes composed of ‘heavy chains’ and several other subunits – including ‘light chains’ and ‘intermediate chains’ in the case of the dynein – that regulate their motor activity and cargo attachment. Their role in the dynamics and cellular distribution of organelles has been assessed by dissociating the motor protein complex with a specific antibody, overexpressing dominant-negative mutants or gene knock out (2,3,5,6,8,11). Yet, these treatments have pleiotropic effects including the impairment of the distribution of various organelles such as Golgi complex, endoplasmic reticulum (ER), trans Golgi network and mitochondria and the distribution of microtubule and actin networks (5–16). Furthermore, deleting one subunit of a motor complex can increase the cytoplasmic concentration of the remaining subunits, which may interfere with cellular functions (17,18). These different treatments could not therefore distinguish unambiguously between a motor that drives directly the movements of endosomes and one that might influence endosome motility indirectly (19).

ATP-dependent movement of endocytic compartments along microtubules in vitro could, in principle, provide an assay for the direct role of motors in the movement of these compartments. Two in vitro motility assays for endosome movement along microtubules have been reported previously (20–25). One of these assays, because it requires cytosol, does not allow to discriminate between motors associated to the endosomes or motors possibly recruited from the cytosol during the experiment (25). In the second assay, the movements of asialoorosomucoid-loaded endosomes isolated from mouse or rat liver have been reconstituted in vitro, but the correlation between the in vitro movements of these endocytic vesicles and their cellular distribution or movements in a tissue such as liver is difficult to assess (20–24).

In this report, we used magnetic nanoparticles synthesized according to Massart (26) to isolate early and late endocytic compartments from HeLa cells and we reconstituted the ATP-dependent movements of these compartments along microtubules in vitro in the absence of cytosol. We observed that the early and late endocytic compartments have distinct microtubule-based motile properties in vitro, suggesting that these two populations of endosomes are equipped with different sets of microtubule-associated motors. We also used biochemical analyses and live-cell imaging to evaluate the direct role of dynein, kinesin-1 and kinesin-2 in the movement of early and late endosomes and lysosomes. We confirmed that dynein and kinesin-1 move early endosomes, and we found that kinesin-2 also moves this population of endosomes. Unexpectedly, we also found that dynein did not move late endosomes/lysosomes in our cell-free assay. Furthermore, we provide evidence from the disruption of dynein function and latrunculin A treatment, suggesting that dynein regulates indirectly late endosome/lysosome distribution potentially through its action on the actin network.

Results

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Isolation of magnetic early and late endocytic compartments

A recent report has shown that negatively charged iron oxide nanoparticles bind to the surface of HeLa cells at 4°C and are subsequently internalized into the cells (27). Depending upon the duration of internalization, after binding at 4°C, the magnetic nanoparticles appear in compartments, presenting specific proteins of early or late endosomes. After 30 min of internalization, the magnetic particles colocalized with early endosome antigen-1 (EEA1) an effector of Rab5 that is associated with early endosomes (Figure 1A,B) and after 90 min with lysosome-associated membrane protein-1 (Lamp-1) that is associated with late endosomes and lysosomes (Figure 1C,D). The magnetic nanoparticles do not interfere with the morphology of the endosomes (27). Wilhelm et al. observed them in clathrin-coated endosomes after 10 min of chase at 37°C and in multivesicular endosomes after 60 min of chase. They do not affect the transport of cargo: internalized BSA was observed in late endocytic and lysosomal compartments after 90 min whether in the presence or absence of magnetic nanoparticles (unpublished data). Furthermore, the organelles loaded with magnetic nanoparticles acquire magnetic properties. Endocytic compartments loaded with magnetic nanoparticles are oriented in the magnetic field when HeLa cells were maintained in a magnetic field during fixation (Figure 1, see arrows).

image

Figure 1. Depending upon the time of their internalization, magnetic nanoparticles colocalized with early or late endosomes. HeLa cells were allowed to bind magnetic nanoparticles 1 h at 4°C. After washing, magnetic nanoparticles have been internalized at 37°C for 30 or 90 min. The cells were then incubated in a magnetic field for 20 min prior to being fixed under the magnetic field. The cells that have internalized the magnetic nanoparticles during 30 min (A and B) or 90 min (C and D) were observed by phase-contrast microscopy (A and C) or processed for IF with an anti-EEA1 antibody (B) or with an anti-Lamp-1 antibody (D) and imaged by confocal microscopy. Average projected Z stacks are shown. The black arrows in (A) and (C) represent the orientation of the magnetic field. Note that black dots in (A) and (C) that correspond to compartments loaded with magnetic nanoparticles are aligned in the magnetic field two-by-two in (A) or more in (B) (see small white arrows). The so formed chains of magnetic compartments codistribute with EEA1 or Lamp-1-labeled compartments in (B) and (D), respectively. Bar represents 3 μm.

Download figure to PowerPoint

We therefore attempted to separate endosomes loaded with magnetic nanoparticles for various times from the nonmagnetic organelles after fractionation of HeLa cells. The purity of the membrane fraction isolated on the magnet was assessed by biochemistry, immunofluorescence (IF) and electron microscopy (EM). Depending upon the duration of internalization, the membrane fractions were enriched either in two specific markers of early endosomes, Rab5 and EEA1, or in two specific markers of late endocytic and lysosomal compartments, Lamp-1 and Rab7 (Figure 2A,B). After internalization for 15–45 min, the magnetic membrane fractions were enriched in Rab5, maximum enrichment being detected after 30 min internalization (Figure 2A,B). Early endosome antigen-1 was also detected in these fractions, and maximum enrichment of this marker was seen after 30–45 min of internalization (Figure 2A,B). By contrast, Rab7 and Lamp-1 were detected at maximal levels in the membrane fractions isolated after 60–180 min internalization of the magnetic nanoparticles (Figure 2A,B).

image

Figure 2. Biochemical and structural characterization of magnetic endosomes. A) HeLa cells were allowed to bind magnetic nanoparticles for 1 h at 4°C and, after washing, to internalize them at 37°C for periods ranging from 15 to 180 min, as indicated. The cells were then washed and homogenized, and the MF was isolated as described in the Materials and Methods. Forty micrograms of proteins for each of PNS, NMF and MF was analyzed by immunoblotting with antibodies against early endosomal markers Rab5 and EEA1 and late endosome/lysosome markers Rab7 and Lamp-1. B) The amounts of Rab5, Rab7, EEA1 and Lamp-1 detected in 40 μg of proteins of the MF were quantified as described in Materials and Methods and normalized to the amount detected in 40 μg of proteins of PNS. The means of three independent experiments are shown, and error bars indicate the standard deviation. C) Electron micrographs of the MFs isolated after 30 and 90 min internalization of the magnetic nanoparticles. D) Forty micrograms of proteins from the PNS, NMF and MF isolated after 30 (30 min) or 90 min of internalization (90 min), as indicated, was analyzed by immunoblotting with mouse anti-EEA1, mouse anti-Lamp-1, rabbit anti-ER and rabbit anti-GMAP210 antibodies. The polyclonal rabbit anti-ER serum contains several immunoglobulin-G that recognizes several proteins from the ER (54). Only one of these proteins that has an apparent molecular weight of 66 kD is shown here. E–L) The MFs isolated after 30 and 90 min internalization of the magnetic nanoparticles have been observed by phase-contrast microscopy (E, I, G and K) or labeled with anti-EEA1 (F and J) or anti-Lamp-1 antibodies (H and L) after injection in a flow cell. Arrowheads mark isolated endosomes, and arrows mark aggregates. Bars represent 9.7 μm.

Download figure to PowerPoint

We further compared the magnetic fraction isolated after 30 min internalization with that isolated after 90 min. The morphologies of the membrane vesicles contained in the fractions isolated after 30 and 90 min were characteristic of early and late endocytic compartments, respectively, when analyzed by EM (Figure 2C). The fraction isolated after 30 min presented small magnetic vesicles of 50–200 nm, whereas the fraction isolated after 90 min showed larger vesicles of 200–500 nm. These larger vesicles contained internal vesicles or multilamellar structures characteristic of late endosomes and lysosomes, respectively (Figure 2C).

After 30 and 90 min of internalization, the magnetic fractions were enriched for EEA1 or Lamp-1, respectively (Figure 2D). However, small amounts of proteins from the ER or the Golgi complex similar to that detected in the corresponding nonmagnetic fractions (NMFs) or the postnuclear fraction can be observed in these fractions (Figure 2D).

To assess the homogeneity of the magnetic fractions isolated after 30 and 90 min internalization for markers of early and late endocytic compartments, we immunolabeled them with anti-EEA1 and anti-Lamp-1 antibodies. After 30 min, 80% (n= 857) of the individual vesicles observed by phase-contrast microscopy were labeled with the anti-EEA1 antibody (Figure 2F arrowheads), while some aggregates were labeled with anti-Lamp-1 antibody (Figure 2H arrows). Whereas after 90 min, 86% (n= 562) of the individual vesicles were labeled with the anti-Lamp-1 antibody (Figure 2L arrowheads), while some aggregates were labeled with anti-EEA1 antibody (Figure 2J arrows). Few aggregates present in both fractions were also labeled with anti-ER and anti-Rab6 antibodies (that decorates the Golgi complex), but we never observed individual vesicles labeled by these two antibodies (Figure S1). It is very likely that these aggregates account for protein contaminants detected by Western blotting in the magnetic fractions.

Together, these experiments suggest that a large majority of individual magnetic vesicles isolated after 30 min of internalization corresponds to early endosomes, while the large majority of individual magnetic vesicles isolated after 90 min of internalization corresponds to late endosomes and lysosomes. We will refer to the membrane fraction enriched in early endosomes as ‘magnetic early endosomes’ and the membrane fraction enriched in late endosomes and lysosomes as ‘magnetic late endosomes’.

Magnetic early and late endosomes move with distinct properties along microtubules in vitro

To analyze the contributions of various microtubule-associated motors to the movements of magnetic early and late endosomes, we reconstituted in vitro the microtubule-based movements of these endocytic compartments. Rhodamine-labeled microtubules were allowed to adhere to the glass surface of a flow chamber coated with polylysine, and magnetic endosomes were then injected into the flow chamber. Both populations of magnetic endosomes bound to microtubules. Arrows in Figure 3A point to individual magnetic late endosomes bound to fluorescent microtubules. Upon injection of ATP into the chamber, the movements of the individual magnetic early and late endosomes were very often discontinuous; during pauses, they oscillated while still bound to the microtubules (see Movies S1–S3). In contrast, aggregates did not move in these experimental conditions. Both magnetic early and late endosomes moved bidirectionally along the microtubules, as previously observed for asialoorosomucoid-loaded endocytic vesicles isolated from rat liver (22,24). Figure 3B shows one magnetic late endosome changing direction (see also Movies S1, S2).

image

Figure 3. Magnetic early and late endosomes have different motile properties in vitro. A) Phase-contrast microscopy of magnetic late endosomes overlaid with the fluorescent microtubule image of a whole microscope field after injection of the magnetic late endosomes into the flow cell. Arrows point to endosomes bound to microtubules. B) A sequence of high-magnification phase-contrast images overlaid with the fluorescent microtubule image illustrating the movement of one late magnetic endosome. The images were taken from the area indicated by a white rectangle in (A). The asterisk on each frame marks the starting point of the endosome, and the arrows point to its successive positions. Note that this magnetic endosome moves back and forth. C) The time during which magnetic early and late endosomes displayed oscillatory movements was normalized (%) to the total duration of their movements observed during an observation period of 4 min for both populations of magnetic endosomes. Thirty-one and 32 moving compartments were observed in five and four independent experiments for early and late magnetic endosomes, respectively. Plots show the mean (%) of oscillatory endosomes ± standard deviation (SD). D) The numbers of magnetic early and late endosomes and kinesin-1-coated beads moving bidirectionally were normalized to the total number of magnetic endosomes or beads moving directionally and plotted as mean (%) ± SD. Four hundred and ninety-nine, 428 and 236 magnetic early and late endosomes and kinesin-coated beads were analyzed in 8, 12 and 2 independent experiments, respectively. *p < 0.001, **p = 0.002. E) Distribution of instantaneous velocities (IV) of magnetic early and late endosomes moving on microtubules in vitro. The total number of IVs computed was 834 and 866 for magnetic early and late endosomes, corresponding to 31 and 32 moving compartments observed in five and four independent experiments, respectively. The number of IVs with a given value was calculated as a percentage of the total number of IVs computed. F) Distribution of the duration of the movements performed by magnetic early and late endosomes, calculated as a percentage of the total movements analyzed (n= 143 and 156 for magnetic early and late endosomes, respectively). G) Distribution of the lengths covered by magnetic early and late endosomes as a percentage of the total movements analyzed (n= 143 and 156 for magnetic early and late endosomes, respectively).

Download figure to PowerPoint

Endosomes might switch direction either by using alternately plus-end- and minus-end-directed motors to move along an individual microtubule or by using a single motor but switching from one microtubule to another microtubule of opposite polarity in a bipolar bundle. We observed that early and late magnetic endosomes moved bidirectionally on single-polarity-marked microtubules (unpublished data). However, the statistical analysis of these movements was precluded by the limited number of movements observed per field. To identify the polarity of these microtubules, we had to dilute them 10 times more than the rhodamine-labeled ones. Consequently, we observed only 0.5 movements per field with these fluorescent microtubules compared with 15–20 movements per field with the rhodamine-labeled ones. We thus evaluated the contribution of bipolar bundles to the observed bidirectional movements on rhodamine-labeled microtubules. We compared the movement of polystyrene beads coated with a single type of motor kinesin-1 that moves bidirectionally only on bipolar microtubule bundles (28) with those of the magnetic endosomes. Of the kinesin-1-coated beads, 14% exhibited bidirectional movements in our experiments compared with 56% of magnetic early endosomes and 37% of magnetic late endosomes (Figure 3D). Thus, the majority of bidirectional movements observed in our experiments were probably because of the magnetic endosomes being moved in opposite directions by two or more different motors rather than by endosomes switching between microtubules of opposite polarities.

The magnetic early endosomes were more likely to move bidirectionally (56%) than were the magnetic late endosomes (37%), while both populations spent a similar proportion of their time in oscillation (Figure 3C,D), and they present similar distributions for the length and the duration of their movements (Figure 3F,G). Magnetic early endosomes also moved more slowly than did the late endosomes. The velocity of moving magnetic endosomes is not constant. We have compared the distribution of the instantaneous velocities (velocity measured between frames) of early and late magnetic endosomes. Thirty-one percent of the instantaneous velocities computed for the magnetic early endosomes were below 0.2 μm/s, compared with 15% of those for late magnetic endosomes (Figure 3E). The mean velocity of magnetic early endosomes was 0.27 ± 0.07 μm/s, whereas the mean instantaneous velocity of the magnetic late endosomes was 0.49 ± 0.1 μm/s. The value for early magnetic endosomes is on the same range of movements of transferrin-loaded compartments (29). The mean velocity of the magnetic late endosomes was very similar to the velocity of lysosomes and CD63GFP-positive compartments measured in vivo (0.45 ± 0.01 μm/s and 0.5 μm/s, respectively) (6,30).

Nielsen et al. (25) have reported that addition of cytosol to an in vitro microtubule-based motility assay increased the number of movements per minute and per field compared with movement in the absence of cytosol. The addition of cytosol from HeLa cells (2 mg/mL) containing dynein and kinesin-1 that had the ability to move microtubules in a gliding assay, to our in vitro motility assay had no effect on the frequency, bidirectionality or velocity of movements of the magnetic early or late endosomes (unpublished data). This suggests that microtubule-associated motors did not dissociate from the endocytic compartments during their isolation by the magnetic nanoparticle method and that no additional proteins are required for their movement on microtubules in our in vitro assay.

Together, these observations indicate that magnetic early and late endosomes have distinct motile properties: magnetic early endosomes move more slowly and change direction more frequently than do magnetic late endosomes. These distinct properties may reflect differences in the motor proteins they hold.

Kinesin-1 and dynein move early but not late endosomes in vitro

Kinesin from the kinesin-1 family and dynein are generally thought to mediate bidirectional movements of the endocytic compartments. We therefore asked whether these two types of motors contribute to the movement of the magnetic early and late endosomes that we see in vitro. We first investigated whether these motor proteins were associated with magnetic endosomes by immunoblotting. Both the magnetic early and late endosomes contained kinesin-1 and cytoplasmic dynein [Figure 4A, lanes magnetic fraction (MF)]. None of these proteins was enriched in the magnetic fractions, however, when compared with PNS (PNS lanes) or NMF (NMF lanes), consistent with the fact that these motors are associated with various organelles. The proportion of kinesin-1, DYNC1H1 (dynein heavy chain) and DYNC1I1 (dynein intermediate chain) detected in the magnetic early endosomal fraction represented 4, 9 and 8% of the amount detected in the PNS, respectively. The proportion of the kinesin-1 heavy chain in the magnetic late endosomal fraction represented 3% of the amount detected in the PNS. The amount of proteins from the dynein complex was slightly lower in the magnetic late fraction than in the early fraction (Figure 4B). The proportion of DYNC1H1 and DYNC1I1 (detected with the two different antibodies, 74.1 and 70.1, respectively) in the magnetic late endosomes represented 34, 45 and 32%, respectively, of the amount detected in the magnetic early endosomes (Figure 4B) and 2% of the total amount detected in the PNS.

image

Figure 4. Detection of kinesin-1, kinesin-2 and proteins of the dynein complex in magnetic early and late endosomes. A) Forty micrograms of protein from PNS, NMF and MF enriched in magnetic early and late endosomes, as indicated, was analyzed by immunoblotting with rabbit anti-kinesin-1, mouse anti-kinesin-2 (K2.4), mouse anti-DYNC1H1 and mouse anti-DYNC1I1 (70.1 and 74.1) antibodies. B) The amounts of kinesin-1, kinesin-2, DYNC1H1 and DYNC1I1 (detected with either antibody 70.1 or antibody 74.1) in 40 μg of protein from the MFs were estimated by quantification of the immunoblots using the procedure described in the Materials and Methods. The values were normalized to the amounts detected in 40 μg of proteins of the PNS as determined by the same method. Results of three independent experiments are shown as mean ± standard deviation. The p values are shown.

Download figure to PowerPoint

To determine the specific involvement of dynein and kinesin-1 in the motility of individual magnetic endosomes in the in vitro assay, we next monitored their movements in the presence of antibodies that specifically block the activity of kinesin-1 or DYNC1 (dynein complex). The monoclonal antibody SUK4 inhibits kinesin-1 activity (4,12,31); addition of SUK4 to the motility assay did not affect significantly the number of magnetic early and late endosomes displaying directed movements when compared with no treatment or with movement in the presence of antibodies against Lamp-1 or EEA1 (Figure 5A). By contrast, SUK4 induced a dose-dependent decrease (by up to 37%) in the proportion of magnetic early endosomes moving in a bidirectional fashion, whereas this proportion remained unchanged for late magnetic endosomes (Figure 5B).

image

Figure 5. Antibodies against kinesin-1 and DYNC1I1 impair the movement of magnetic early endosomes but not of magnetic late endosomes in vitro. A and B) Microtubule-based movements of magnetic early and late endosomes in vitro were examined in the absence of antibody (−AB) or after addition of 12.5 μg/mL control antibody (anti-Lamp-1 antibody for magnetic early endosomes and anti-EEA1 antibody for magnetic late endosomes, as indicated) and 6.25 μg/mL [SUK4 (a)] and 12.5 μg/mL [SUK4 (b)] antibody against kinesin-1. C and D) Microtubule-based movements of early and late magnetic endosomes were examined in vitro without antibody or after addition of 25 μg/mL control antibody [as in parts (A) and (B)] and 12.5 μg/mL [74.1 (a)] and 25 μg/mL [74.1 (b)] antibody 74.1 against DYNC1I1. The number of magnetic early and late endosomes moving directionally in the presence of antibodies per field observed during 4 min is expressed as a percentage of the number of untreated magnetic endosomes moving under the same experimental conditions (n= 495 and n= 195 for magnetic early and late endosomes, respectively, analyzed with SUK4 antibody; n= 292 and n= 242 for magnetic early and late endosomes, respectively, analyzed with 74.1 antibody) (A and C). The proportion of magnetic endosomes moving bidirectionally relative to the number of those moving directionally was plotted for magnetic early and late endosomes treated with antibodies as mean (%) ± standard deviation (SD) (B and D). Errors bars indicate the SD of the mean for number of movements observed per fields. *p < 0.04, **p < 0.004.

Download figure to PowerPoint

Monoclonal antibody 74.1 against DYNC1I1 has been reported to block cytoplasmic dynein activity by dissociating the dynein–dynactin complex (32). Addition of antibody 74.1 to the motility assay did not significantly affect the movement of magnetic late endosomes, but it decreased slightly the number of magnetic early endosomes displaying directed movements compared with the untreated magnetic endosomes (−AB) or with endosomes treated with the antibody against Lamp-1 (Figure 5C). By contrast, antibody 74.1 induced a marked dose-dependent decrease in the proportion of magnetic early endosomes moving in a bidirectional fashion, whereas this proportion remained unchanged for late magnetic endosomes (Figure 5D).

Consistent with previous observations, these data suggest that kinesin-1 and dynein are involved in moving early endosomes and contribute to their bidirectional movements (3,5), but, unexpectedly, these motors appear not to be the major motors that move late endosomes. Both antibody treatments inhibiting dynein and kinesin-1 function decreased the proportion of bidirectional movements of early magnetic endosomes, but they decreased slightly the total number of movements. This suggests that dynein and kinesin-1 are mostly associated with early magnetic endosomes that display also other motors that contribute to their movements.

Disruption of dynein function may affect the cellular distribution of late endocytic compartments indirectly

Our observations that dynein appears not to move late endosomes contradict the interpretations of previous studies showing that knock down of DYNC1H1 expression or dissociation of the dynein–dynactin complex delocalized the late endocytic compartments to the cell periphery, suggesting, therefore, that dynein is involved in moving late endosomes toward the perinuclear minus end of microtubules (2,3,9). To reconcile our study with the interpretation of the previous experiments, we investigated the impact of silencing DYNC1H1 expression upon the distribution of the Golgi complex and the lysosomes as well as upon the movements of the lysosomes. We used a mixture of four different small interfering RNA (siRNA) duplexes that target different regions of the messenger RNA encoding human DYNC1H1. Immunoblot analyses showed 85% depletion of DYNC1H1 4 days after siRNA transfection, whereas no depletion of DYNC1H1 was detected after transfection with control siRNA (Figure 6A). The level of DYNC1I1 was also significantly reduced while the level of the light chains LC81 and Tctex1 remained similar to that of untreated cells or of cells transfected with control siRNA (Figure 6A). The level of LC82 was also not affected (unpublished data). The observed reduction of DYNC1H1 and DYNC1I1 suggests that dynein complex is disrupted in these conditions, releasing the light chains in the cytosol.

image

Figure 6. DYNC1H1 siRNA impairs the distribution of the Golgi complex and Lamp-1 but has a minor effect on centrifugal and centripetal movements of lysosomes. A) Immunoblots of DYNC1H1, DYNC1I1, LC81 and Tctex1 in HeLa cells in absence of treatment (‘non treated’) or 4 days after transfection with control siRNA (‘cont. siRNA’) and DYNC1H1 siRNA (‘DYNC1H1 siRNA’). Immunoblotting of EEA1 serves as a control for loading of DYNC1H1 and tubulin as a control for loading of intermediate and light chains. B) The amount of DYNC1H1 detected in the cells transfected with control siRNA or DYNC1H1 siRNA was quantified as described in the Materials and Methods and normalized to the amount detected in cells transfected with control siRNA. The mean of three independent experiments is shown, error bars indicate standard deviation and the p value is shown on the graph. C–F) Cells 4 days after transfection with control siRNA (C and E) or with DYNC1H1 siRNA (D and F) were labeled for IF microscopy with an anti-Golgi antibody (C and D) or with an anti-Lamp-1 antibody (E and F). The outlines of the cells in the corresponding phase-contrast images are indicated. Bars represent 5 μm. G) Comparison of the number of cells with a Golgi complex in the perinuclear region, as shown in (C), with the number of cells with a dispersed Golgi complex, as shown in (D), after transfection with the control siRNA (n= 526) or with the DYNC1H1 siRNA (n= 635). The p value is shown on the graph. H and I) BODIPY-TR pepstatin A was internalized for 90 min in HeLa cells previously transfected with control siRNA or DYNC1H1 siRNA. The centripetal and centrifugal movements of the pepstatin-A-labeled compartments were then analyzed by time-lapse microscopy as described in the Materials and Methods. The average distances covered by pepstatin-A-labeled compartments to the center of the cell (centripetal) or to the cell periphery (centrifugal) during 4 min and the average of their instantaneous velocities in both directions are represented in (H) and (I), respectively. The p values are shown in the figure (H).

Download figure to PowerPoint

Consistent with a previous report, in 79% of the cells transfected with DYNC1H1 siRNA, the Golgi complex was dispersed throughout the cytoplasm (compare Figure 6C and D and see Figure 6G) (33). Also, 81% of the transfected cells had a large proportion of Lamp-1 staining (indicating late endosomes and lysosomes) concentrated at the cell periphery, whereas in 96% of the cells transfected with the control siRNA, the majority of Lamp-1 staining was concentrated in the perinuclear region with the reminder scattered throughout the cell (compare Figure 6E and F and see Figure 7H).

image

Figure 7. Latrunculin A treatment reversed the delocalization of Lamp-1-labeled compartments caused by DYNC1H1 siRNA transfection. A–D) Cells 4 days after transfection with control siRNA (A and C) or with DYNC1H1 siRNA (B and D) were labeled for IF microscopy with an anti-tubulin antibody (A and B) or with a fluorescent phalloidin (C and D). Bars represent 5 μm in (A) and (B) and 10 μm in (C) and (D). E) Comparison of the number of cells with a cortical actin cytoskeleton, as shown in (C), with the number of cells showing an increase in the number of actin cables, as shown in (D), after transfection with the control siRNA (n= 250) or with the DYNC1H1 siRNA (n= 259). The p values are shown. F–G) Cells transfected with DYNC1H1 siRNA were treated for 30 min with 0.5-μm latrunculin A prior to processing and then analyzed for the distribution of actin filaments (F) or Lamp-1 (G). Bars represent 10 μm, and arrows point to the perinuclear accumulation of Lamp-1. H) Comparison of the number of cells with a perinuclear pool of Lamp-1-labeled compartments, as shown in Figure 5E, with the number of cells with a high concentration of Lamp-1-labeled compartments at the cell periphery, as shown in Figure 5F after transfection with the control siRNA (n= 337) or with the DYNC1H1 siRNA (n= 411), or transfection with the DYNC1H1 siRNA followed by treatment with 0.5 μm latrunculin A for 30 min (n= 220). *p < 0.001.

Download figure to PowerPoint

We next analyzed the movement of the late endosomes/lysosomes upon depletion of the dynein heavy chain. We have previously shown that internalized pepstatin A conjugated covalently to BODIPY-TR cadaverine hydrochloride (BODIPY-TR pepstatin A) fills compartments that codistribute with Lamp-1 (30). Internalized BODIPY-TR pepstatin A codistributed also with Lamp-1-labeled compartments in cells transfected either with control siRNA or with DYNC1H1 siRNA and largely accumulated at the periphery of the cells transfected with DYNC1H1 siRNA (unpublished data). We compared the movements of these compartments after transfection of DYNC1H1 siRNA with those observed after transfection of control siRNA. Although the average velocities and the average distances covered by BODIPY-TR pepstatin-A-marked compartments of both centripetal and centrifugal movements were slightly reduced in cells transfected with DYNC1H1 siRNA, they were still undergoing directional movements in both centrifugal and centripetal directions between the cell center and the periphery (Figure 6H,I and Movie S4). These movements were totally inhibited by nocodazole treatment, supporting the fact that they were microtubule dependent (Movie S6).

We then reasoned that the delocalization of late endosomes/lysosomes and the slight decrease of the velocity and distance covered by these compartments in both directions after depletion of DYNC1H1 may be because of an indirect effect. For example, the release of the light chains in the cytosol under DYNC1H1 depletion may contribute to delocalizing late endosomes and lysosomes by a mechanism independent of the function of the dynein complex. Recent experimental evidence suggests a link between the actin dynamics and the dynein light chains TcTex1 and LC8 independent of their role in the dynein complex (17,18). In addition, actin filaments are known to co-operate with microtubules in the movement of late endosomes and lysosomes and to contribute to their normal subcellular distribution (30,34). We postulated that the redistribution of late endocytic compartments upon the depletion of DYNC1H1 might, therefore, be because of reorganization of the actin network.

To test this hypothesis, we investigated the impact of DYNC1H1 depletion on the actin and microtubule networks. The distribution of the microtubules appeared similar in cells transfected with DYNC1H1 siRNA and in those transfected with the control siRNA (Figure 7A,B). Cells transfected with the control siRNA displayed an important actin cortical network while actin cables were barely visible when we focused at the base of the nucleus (Figure 7C). In contrast, 84% of the cells transfected with DYNC1H1 siRNA displayed an increased number of actin cables compared with control cells when we imaged both cell populations at the same focal plane while the actin cortical network was barely visible (compare Figure 7C with D and see also Figure 7E).

We used latrunculin A, a drug that binds actin monomers and as a consequence moves the filamentous actin/globular actin (G-actin) equilibrium in cells toward G-actin to determine whether the actin network contributes to the redistribution of Lamp-1-marked compartments under DYNC1H1 depletion (35). The actin cables in DYNC1H1-depleted cells were depolymerized to a large extent by treatment with latrunculin A (Figure 7F). The treatment of DYNC1H1-depleted cells with latrunculin A did not inhibit movement of pepstatin-A-loaded endosomes, and it led to the redistribution of Lamp-1-labeled compartments from the periphery to the perinuclear region similar to that observed in control siRNA-treated cells (compare Figure 7G with Figure 6E and Movie S5 with Movie S4). Of the cells transfected with DYNC1H1 siRNA and treated with latrunculin A, 83% showed a similar distribution of Lamp-1-positive endosomes to that found in control siRNA-treated cells (Figure 7H). Although to a less extent (40% of the microinjected cells with the antibody 74.1 and treated with latrunculin), the delocalization of the Lamp-1-labeled compartments induced by the microinjection of the 74.1 antibody was also reversed by actin depolymerization (unpublished data). Thus, depolymerization of actin filaments by latrunculin A treatment reversed the effect of dynein heavy chain depletion or dissociation of the dynein dynactin complexes.

The return of Lamp-1-positive compartments to the perinuclear region following latrunculin A treatment of DYNC1H1-depleted cells suggests that disruption of the dynein complex interferes with the perinuclear distribution of late endosomes and lysosomes not directly by preventing their transport along microtubules but indirectly by reorganizing the actin network. The slight decrease of the average velocity and the average distance covered by the lysosomes under depletion of DYNC1H1 both in centrifugal and in centripetal movements may reflect the increase of the cytoplasmic viscosity because of the increase of actin filaments.

Kinesin-2 moves early and late endosomes

Previous reports have shown that kinesin-2 is required for the normal cellular distribution of late endosomes and that it moves asialoorosomucoid-loaded late endosomes in an in vitro assay (10,22). We therefore investigated the contribution of kinesin-2 to the movement of HeLa cell endocytic compartments in vivo and in vitro.

We first addressed whether one of the kinesin-2 heavy chains, KIF3A, that is associated with endosomes (10) contributes to the movement of late endosomes/lysosomes in vivo. We studied the movements of pepstatin-A-filled compartments in cells expressing the motor-less green fluorescent protein (GFP) fusion protein, GFP–kif3A-C2, that is dominant negative for the function of kinesin-2 (36). We compared the movement of endosomes filled with BODIPY-TR pepstatin A in HeLa cells expressing GFP–kif3A-C2 with those occurring in cells expressing GFP and in untransfected cells. The tracks of the BODIPY-TR pepstatin-A-labeled endosomes recorded in GFP–kif3A-C2-expressing cells tended to be shorter than those in GFP-expressing cells (compare Figure 8A,B). The mean square displacements (MSDs) of the labeled endosomes were computed from their tracks. MSD fits with a power law <r2(t)> =ktα in which the exponent α reflects the type of movement observed: directional, Brownian or confined (37–40). According to the value of the exponent, we determined the proportion of late endocytic compartments undergoing Brownian motion (exponent = 1), directional movement (exponent > 1) or confined movement (exponent < 1) in the different experimental conditions (40). Expression of GFP–kif3A-C2 reduced the proportion of labeled compartments undergoing directional movement (34% compared with 46% in GFP-expressing cells and 54% in untransfected cells) and increased the proportion of compartments undergoing confined movement (56% compared with 41% in GFP-expressing cells and 34% in untransfected cells) (Figure 8C). Thus, as suggested by the experimental evidence reported by Brown et al. (10), kinesin-2 is involved in the in vivo movement of the late endosomes/lysosomes.

image

Figure 8. Expression of the motor-less kinesin-2 protein impairs in vivo movement of late endosomes and lysosomes. A and B) BODIPY-TR pepstatin A was internalized for 1 h 30 min by GFP–kif3-C2-transfected HeLa cells or GFP-transfected cells or untransfected cells and the movements of the pepstatin-A-labeled compartments were subsequently analyzed by time-lapse microscopy. Eighty representative lysosomes tracks are shown for GFP-expressing cells (A) or GFP–kif3-C2-expressing cells (B). C) MSD analysis of tracked organelles allowed us to classify the movement as ‘confined’ (exponent < 0.95), ‘Brownian’ (0.95 < exponent < 1.05) or ‘directional’ (exponent > 1.05) (40). The percentage of organelles in each class in cells expressing GFP–kif3-C2 or GFP, as well as in untransfected cells, is represented as the mean ± standard deviation from 24 cells filmed in four independent experiments. The p values are shown on the graph.

Download figure to PowerPoint

To determine whether kinesin-2 contributes directly or indirectly to the movements of late endosomes/lysosomes, we analyzed the role of kinesin-2 in the movement of magnetic endosomes in vitro. We verified the presence of this motor in the early and late magnetic endosomes by immunoblotting. Kinesin-2 was more abundant than kinesin-1 in both early and late magnetic endosomal fractions. The total amount of kinesin-2 estimated in the magnetic early and late endosomes represented 14% of that detected in the PNS (Figure 4A,B).

We next monitored the movements of magnetic early and late endosomes in the presence of an anti-kinesin-2 antibody that was previously shown to inhibit specifically the activity of kinesin-2 (22). In agreement with our analyses in vivo, we observed that addition of anti-kinesin-2 antibody to the in vitro motility assay decreased the directional movement of magnetic late endosomes by up to 40% (Figure 9A). Interestingly, the proportion of bidirectional movements displayed by the remaining moving endosomes increased by 50% (Figure 9B). Furthermore, this antibody inhibited the directed movement of early magnetic endosomes by 70% (Figure 9A). The proportion of bidirectional movements performed by the remaining moving early endosomes in these experimental conditions was similar to that in untreated endosomes or in endosomes treated with antibodies against Lamp-1 (Figure 9B).

image

Figure 9. An antibody against kinesin-2 inhibits the movement of magnetic early and late endosomes in vitro.The microtubule-based movements of magnetic early and late endosomes were examined in vitro without antibody (−AB) and addition of 25 μg/mL control antibody (anti-Lamp-1 antibody for early magnetic endosomes and anti-EEA1 antibody for late magnetic endosomes) or with 12.5 μg/mL [Kif3 (a)] or 25 μg/mL [Kif3 (b)] anti-kinesin-2 antibody (from BD Transduction Laboratories). The number of magnetic early and late endosomes moving directionally in the presence or absence of antibodies per field observed during 4 min is expressed as a percentage of the number of untreated magnetic endosomes moving under the same experimental conditions (n= 212 and n= 242 for magnetic early and late endosomes, respectively, from two independent experiments) (A). The proportion of magnetic endosomes moving bidirectionally relative to the number of those moving directionally was plotted for magnetic early and late endosomes treated or not treated with antibodies, as indicated (B). The p values are shown on the graphs.

Download figure to PowerPoint

The absence of decrease of the proportion of bidirectional movements in presence of anti-kinesin-2 antibody may suggest that a substantial proportion of magnetic early and late endosomes bearing kinesin-2 does not have any other active microtubule-based motor. Alternatively, kinesin-2-bearing early magnetic endosomes may also present the dynein–dynactin complex. According to Deacon et al. (41), the function of these two motors may be coordinated by the p150-glued protein of the dynactin complex, and therefore, inhibition of the function of kinesin-2 may impair the function of the dynein . These results together indicate that kinesin-2 is involved directly not only in the movements of late endosomes but also in the movement of early endosomes.

Discussion

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Magnetic procedures have been used successfully in recent years to isolate phagosomes and, more recently, also certain populations of endosomes (42–44). We applied such a procedure to isolate early and late endocytic compartments together with their entire machinery for microtubule-based movement from cells in culture. Some of the magnetic early or late endosomes may aggregate with contaminating ER or Golgi membranes during cell fractionation. Nevertheless, the magnetic fraction isolated after 30 min of internalization contains 80% of individual magnetic vesicles presenting a marker of early endosomes and the magnetic fraction isolated after 90 min present 86% of individual magnetic vesicles with a marker of late endosomes/lysosomes. We devised a new in vitro motility assay for these magnetic endosomes. Importantly, the contaminating aggregates did not move in these experimental conditions, and thus, they cannot account for any movement that we have recorded using our in vitro motility assay.

Two other in vitro motility assays have been reported previously, but it is difficult to compare the observations obtained with our assay with those obtained with the previous ones because of important differences in the experimental conditions: one assay required cytosol and, in the other, the properties of the isolated endosomes were very probably different from our magnetic endosomes because of the method applied to isolate them (21–23,25). Our method has the advantage that we use cultured mammalian cells and so can carry out both the in vitro motility assay and the morphological studies, whether on fixed cells or in vivo, using a single cell type.

Dynein, kinesin-1 and kinesin-2 move early endosomes

We found that the bidirectional movements of magnetic endosomes from the fraction enriched in Rab5 and EEA1 are inhibited by a specific antibody (74.1) that dissociates the dynein–dynactin interaction and the anti-kinesin-1 antibody SUK4. These observations are in agreement with previous reports (10,45–47).

We also show here that a kinesin-2 is enriched in this population of endosomes and that an anti-kinesin-2 antibody decreased the number of directional movements. There is an apparent discrepancy between the data we obtained in vitro showing a role for kinesin-2 in early endosome movement and the morphological analyses on fixed cells showing that expression of a motor-less kinesin-2 mutant does not affect the normal steady-state location of EEA1-containing early endosomes [(10) and unpublished data]. It may suggest that complex mechanisms control the subcellular distribution of these organelles at steady state.

Dynein contributes indirectly to the movements of late endosomes and lysosomes

Perturbation of dynein’s activity strongly affects the normal cell distribution of late endosomes and lysosomes. However, the data we report here argue that dynein contributes indirectly to the movement of the late endocytic compartments. First, although we have detected dynein on individual magnetic early and late endosomes (unpublished data), it was less abundant in the magnetic late fraction than in early magnetic fraction. It is unlikely that proteins from the dynein complexes were released during cell fractionation because addition of cytosol containing active motors did not change the number of bidirectional movements we observed in vitro. Furthermore, the low amount of dynein in the magnetic late endosomes is corroborated by a previous analysis by EM (48). Second, in vitro movement of magnetic late endosomes was insensitive to an antibody against the dynein intermediate chain that disrupts the dynein–dynactin interaction and thereby the processivity of the dynein. By contrast, this reagent inhibited the in vitro movement of magnetic early endosomes in the same experimental conditions. Third, depletion of the dynein heavy chain DYNC1H1 that disrupts the dynein complex caused delocalization of Lamp-1-labeled compartments to the cell periphery, but these compartments still underwent centripetal and centrifugal microtubule-dependent movements. Fourth, depletion of the dynein heavy-chain DYNC1H1 also had a major effect on the distribution of the actin network, and both of the effects on actin filaments and Lamp-1-labeled compartments were reversed by the actin-depolymerizing drug latrunculin A.

This argues that the delocalization of Lamp-1-labeled compartments can be accounted for by the effect of the disassembly of the dynein complex. Recent studies suggest a role for dynein light chains in regulating the actin network independent of their role in the dynein complex. The light-chain DYNLT1 (Tctex1) was shown to modulate actin dynamics during neurite outgrowth, and the light-chain DYNLL1 (LC8) was found to regulate dynamin-mediated actin filament assembly during spermatogenesis in Drosophila(17,18). The release of the dynein light chains into the cytoplasm upon depletion of dynein heavy chain may reorganize the actin network, thereby affecting the subcellular localization of late endosomes and lysosomes. Although we cannot totally exclude that dynein moves late endosomes and lysosomes, our observations strongly support that dynein contributes indirectly to the movement of the late endocytic compartments possibly through some effect on the organization of the actin network rather than directly through its motor activity.

Kinesin-2 but not kinesin-1 moves late endosomes and lysosomes

A recent report showed that overexpression of a motor-less kinesin-2 delocalized late endosomes and lysosomes (10). Indeed, our experimental evidence indicates that kinesin-2 moves these endocytic compartments in vivo as well as in vitro. Inhibition of kinesin-1 activity has been reported to perturb the subcellular distribution of late endosomes and lysosomes (3,8), yet anti-kinesin-1 antibody did not inhibit the movement of magnetic late endosomes in vitro, suggesting that, like dynein, kinesin-1 does not directly move late endosomes and lysosomes but rather contributes indirectly to their subcellular localization. Both kinesin-1 and dynein are less abundant in the late magnetic fraction. Interestingly, a recent study using fluorescent imaging with one nanometer accuracy (FIONA) suggests that multiple kinesins or multiple dyneins work together to move peroxisomes (49). Therefore, one possibility is that the amounts of kinesin-1 and dynein associated with late magnetic endosomes are not sufficient to move these compartments.

In conclusion, we have shown that dynein, kinesin-1 and kinesin-2 contribute to the movement of early endosomes enriched in EEA1 and Rab5, whereas kinesin-2 contributes to the movement of late endosomes and lysosomes enriched in Rab7 and Lamp-1. Because late endosomes/lysosomes still move to the perinuclear region in cells transfected with DYNC1H1 siRNA and present bidirectional movements in vitro, it is very likely that a kinesin moving to the minus end of the microtubules (class 14) is involved in their movements. The perinuclear distribution of the late endocytic compartments at steady state may result from the balance between the function of kinesin-2 and this kinesin from the class 14. In addition, our data suggest that the actin network plays also an important role in the steady-state distribution of the lysosomes. Interestingly, short actin filaments have been observed previously in the perinuclear region (50). Furthermore, pepstatin-A-loaded compartments bind and can be retained on actin filaments (30). One possibility is that short actin filaments in the perinuclear region contribute to retain the late endosomes/lysosomes transiently in this region.

The new methodologies that we have developed should enable us in the future to identify the kinesin moving to microtubule minus end and to further elucidate the molecular mechanisms involved in the recruitment and the regulation of the microtubule-associated motors that contribute to the movement of early and late endosomes and lysosomes.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Antibodies and reagents

We used the following primary antibodies in this study: mouse anti-Lamp-1 and mouse anti-EEA1 (BD Transduction Laboratories); mouse anti-tubulin (clone DM1A) (Sigma-Aldrich); mouse anti-dynein intermediate chain (clone 70.1; Sigma-Aldrich); mouse anti-dynein intermediate chain (clone 74.1; Chemicon); mouse anti-dynein heavy chain (directed against amino acids 4320–4644 of rat DYNC1H1) and rabbit anti-Rab6 (Santa Cruz Biotechnology); mouse anti-p150 (BD Pharmingen); mouse anti-kinesin-2 (K2.4; a gift from Dr V. Gelfand, North Western University, Chicago and BD Transduction Laboratories); mouse anti-DYNC1L1 (LC81) (10D6; Alexis Biochemical); mouse polyclonal anti-DYNC1L2 (LC82) (Abnova Corporation); mouse anti-Golgi complex (a gift from Dr M. Bornens, Institut Curie, Paris, France) (51); rabbit anti-Rab5 (a gift from Dr M. Zerial, Max Planck Institute for Cell Biology and Genetics, Dresden, Germany); rabbit anti-Rab7 (a gift from Dr P. Chavrier, Institut Curie, Paris, France); rabbit anti-DYNC1LT (Tctex1) (a gift from Dr Sung Cornell University, New York, USA) (52); rabbit anti-kinesin-1 (a gift from Dr F. Navone, CNR, Milano, Italy) (53); rabbit anti-ER (54); and rabbit anti-GMAP210 (a gift from Dr M. Bornens) (55). The SUK4 monoclonal anti-kinesin-1 developed by Dr J. Scholey was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the National Institute of Child Health and Human Development (NICHD) and maintained by The University of Iowa (Department of Biological Sciences, Iowa City, IA, USA).

We also used a recombinant plasmid encoding the motor-less domain of a kinesin-2 (GFP–kif3-C2) (36). Kinesin-1-coated polystyrene beads were a generous gift from Dr Patricia Bassereau and Dr Cécile Leduc (Institut Curie, Paris, France) (28).

Cell culture and transfection

HeLa cells were cultured in DMEM medium (Seromed) supplemented with 10% fetal calf serum (Seromed), penicillin (5 U/mL; Seromed) and streptomycin (5 mg/mL; Seromed) at 37°C with 10% CO2. For transfection, cells were electroporated at 240 V and 0.95 mF using the Cell Projects kit (Cell Projects) and grown for 48 h on glass coverslips before analysis.

RNA interference (RNAi) experiments

Four different siRNA duplexes were designed against various regions of DYNC1H1 (human dynein cytoplasmic heavy chain 1; accession number NM_001376) making use of Dharmacon’s custom SMARTpool siRNA service (Lafayette) (Ref. M-006828-00): 5′-GCAAAGAAUUCACAAGUAA-3′, 5′-GAGAGGAGGUUAUGUUUAA-3′, 5′-GCAAGAAUGUCGCUAAAUU-3′ and 5′-GGAUGAAUCUAAUGUGUUA-3′. This pool of siRNAs or a ‘nontargeting’ control siRNA (siRNA-1; Ref. 001210-01-05) from Dharmacon were transfected into HeLa cells by making use of oligofectamin (Invitrogen) at a concentration of 12 μm. After 4 days to allow maximum depletion of DYNC1H1, the cells were processed for time-lapse microscopy, lysed for immunoblotting or fixed and processed for IF staining.

Internalization of magnetic nanoparticles

The magnetic nanoparticles used in this study were synthesized according to Massart (26). They possess a magnetic cobalt ferrite oxide core (CoFe2O4), and they are coated with negatively charged citrate ligands (56). The resulting negative surface charge ensures both the colloidal stability of the nanoparticles in aqueous suspension and their high electrostatic affinity for the cell plasma membrane. The magnetic nanoparticles had a mean diameter of 8 nm, and each nanoparticle was a ferrimagnetic single domain with a mean magnetic moment of 1.2 × 10–19 A m2.

To internalize the magnetic nanoparticles, 2.5 × 108 HeLa cells growing on Petri dishes or 20 000 cells growing on coverslips were incubated for 1 h at 4°C in RPMI supplemented with 2 mm sodium citrate and magnetic nanoparticles (iron concentration, 2 mm). Then they were washed in RPMI and chased at 37°C for various times, depending upon the experiment in the same medium. At the end of the period of internalization, the cells were washed three times in cold PBS and processed for IF labeling or cell fractionation.

Isolation of early and late magnetic endosomes

Cells that have internalized magnetic nanoparticles were washed in HB buffer (250 mm sucrose, 5 mm imidazole, pH 7.4, 5 mm MgSO4, 1 mm ethylenediaminetetraacetic acid, 0.5 mm EGTA, 0.15 mg/mL casein), scraped from the plastic support, centrifuged and resuspended in HB buffer supplemented with 1-mm DTT and protease inhibitor cocktail (Sigma-Aldrich). They were then homogenized by passing them several times through a 23-G needle. Unbroken cells and nuclei were removed from the cell homogenate by centrifugation at 1300 ×g for 7 min, and the PNS was loaded onto a syringe fixed to a magnet. After overnight incubation at 4°C, the subcellular compartments containing magnetic nanoparticles accumulated on the side of the syringe fixed to the magnet similar to the magnet of SuperMACS (Miltenyi Biotech). The NMF was collected in the presence of the magnet, and the concentrated magnetic fraction was resuspended in HB buffer. Both fractions were flash-frozen in liquid nitrogen and stored at −80°C.

HeLa cytosol

HeLa cell cytosol was prepared from a pellet of 5 × 109 HeLa cells (4C Biotech) as previously described by Plastino et al. (57).

Immunoblotting

Proteins separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis were transferred to nitrocellulose and processed for immunoblotting using a chemiluminescence blotting substrate (Super Signal West Pico Chemiluminescent substrate; Pierce Biotechnology). Images of the immunoblots were captured with a FUJI LAS-1000 charge-coupled device camera and quantified using scionimage software (Scion Corporation). To ensure that our enhanced chemiluminescence quantification was linear and reliable to the protein quantities, we determined the range of the immunoglobulin-G that was detected linearly in our experimental conditions and adjusted then the amount of analyzed proteins and of antibody to be in this range.

Electron microscopy

Purified magnetic compartments were fixed overnight with 2.5% glutaraldehyde in 0.1 m cacodylate buffer pH 7.2, postfixed with 2% osmium tetroxide, dehydrated in ethanol and embedded in Epon after being concentrated in a magnetic field. Ultrathin sections were counterstained with uranyl acetate and lead citrate and observed and photographed under a Philips CM 120 Electron Microscope (FEI Company).

IF staining of cell culture

Cells grown on coverslips that have internalized magnetic nanoparticles were incubated 20 min in a magnetic field prior to be fixed under the magnetic field and processed for IF labeling as described by Cordonnier et al.(30). Cells were viewed with an epifluorescence microscope or a Leica SP2 confocal microscope (Leica).

IF staining of isolated magnetic endosomes

Isolated magnetic early or late endosomes were injected in a flow chamber coated with polylysine as for the in vitro motility assay. After 6 min incubation, anti-EEA1 or anti-Lamp-1 antibodies diluted in HB buffer were injected in the flow chamber and incubated for 6 min at room temperature. The flow chamber was then washed three times with large volumes of HB buffer before injection of the fluorescent secondary antibody diluted in HB buffer. After 6 min incubation at room temperature, the flow chamber was washed four times with large volumes of HB buffer and imaged with a Leica SP2 confocal microscope.

In vitro motility assays

Fluorescent microtubule polymerization

Tubulin and rhodamine-coupled tubulin were purchased from Cytoskeleton (Denver). Oligomers of rhodamine-coupled tubulin were prepared by incubating 32 μg tubulin and 8 μg rhodamine–tubulin in 5 μL BRB buffer (80 mm piperazine-1,4-bis(2-ethanesulfonic acid) (PIPES), 1 mm MgCl2, 1 mm EGTA, pH 6.9) supplemented with 1 mm GTP and 10% glycerol at 37°C. Microtubules were elongated by diluting the oligomers of tubulin in 50 μL BRB containing 10% glycerol, 1 mm GTP, 40 μg tubulin and 10 μg rhodamine–tubulin and incubating them for 15 min at 37°C. Then taxol was added at a concentration of 20 μm and incubated 15 min at 37°C to stabilize the newly formed microtubules. These newly formed microtubules were finally incubated at room temperature for 48 h to reach a length of 10–20 μm before being centrifuged on a glycerol cushion (20 min at 100 000 ×g and 25°C, TLA 100 rotor) and resuspended in BRB complemented with 10% glycerol and 1 mm GTP.

Motility assays

Motility assays were performed in flow chambers coated with polylysine (Sigma-Aldrich). Rhodamine-labeled microtubules (500 μg/mL) or polarity-marked microtubules (40 μg/mL) were incubated in the chamber for 5 min before being washed with HB buffer. Purified magnetic compartments (3–4 ng total protein) or kinesin-1-coated beads were then injected into the flow chamber and incubated for 8 min before injection of the motility buffer, consisting of HB buffer supplemented with a 2-mm ATP, an ATP-regenerating system (80 μg/mL creatine kinase, 16 mm creatine phosphate) and an anti-fade system (10 mm b-mercaptoethanol, 2.5 mg/mL glucose, 20 μg/mL catalase, 0.1 mg/mL glucose oxidase). When necessary, antibodies or cytosol were added with the motility buffer. Microtubule-based movement of endosomes was monitored at 37°C by the use of a Leica inverted microscope equipped with a Plan Apochromat 100× oil immersion lens (1.4 numerical aperture) in the phase-contrast mode to visualize magnetic endosomes and with a fluorescence filter (N21 excitation BP515-560 dichroïc 565, emission LP590) to observe microtubules. Phase images of the endocytic compartments were acquired at one frame per second for 4 min, preceded by a single fluorescence image of the microtubules.

Movement analysis

Successive phase-contrast images were overlaid with the fluorescent image of the microtubules and the endocytic compartments, or beads were tracked on the overlay using metamorph software. Movement was considered as bidirectional if the vesicle moved back and forth on microtubules during at least 3 seconds (three frames); movement was considered as oscillatory if the displacement was inferior to 0.14 μm after 1 second. Tracks were processed to exclude the points corresponding to oscillatory motion prior to determining the velocity. Instantaneous velocity corresponds to the velocity of the endosomes between two frames. The average velocity corresponds to the mean of the instantaneous velocities.

Internalization of BODIPY-TR pepstatin A and in vivo video microscopy

Pepstatin A-BODIPY-TR cadaverine hydrochloride (BODIPY-TR Pepstatin A) was synthesized and internalized as described by Cordonnier et al.(30). Briefly, cells growing on coverslips were incubated for 20 min at 37°C in serum-free medium and then for 1 h 30 min at 37°C in complete medium supplemented with 1 μm BODIPY-TR pepstatin A. The coverslips were washed three times in complete medium and mounted in a small open chamber that was introduced into a chamber on the microscope stage regulated at 37°C and 5% CO2. A 20-min chase in culture medium was performed before starting image acquisition.

Time-lapse microscopy was performed as described by Cordonnier et al.(30) by using the same microscope as for in vitro motility assays. The movements of pepstatin-A-filled compartments were monitored at a rate of one image per second.

Wild-type HeLa cells and cells transfected with plasmids encoding green fluorescent protein (GFP) and kif3-C2–GFP were filmed during 4 min. Twenty BODIPY-TR pepstatin-A-filled compartments were chosen at random in each filmed cell and tracked using metamorph software. Their MSD was computed as a function of time: <r2(t)> =kΔtα, where brackets denote averaging from the coordinates (37,40). Cells transfected with siRNA were filmed during 4 min. Ten BODIPY-TR pepstatin A-filled compartments were chosen at random in each filmed cell and tracked using metamorph software. DYNC1H1–siRNA-transfected cells were easily identified as they showed accumulation of BODIPY-TR pepstatin-A-labeled compartments at the cell periphery. The average of the centrifugal (unambiguously directed toward the plasma membrane) and centripetal (unambiguously directed toward the perinuclear region) instantaneous velocities were calculated as the mean of all the instantaneous velocities measured for centrifugal and centripetal movements between two frames for each compartments. Only movements longer than 0.2 μm were considered to avoid to take into account Brownian movements. Centripetal and centrifugal distances covered by each compartment were calculated by adding up all the centripetal and centrifugal distances covered between two frames.

Statistical analysis

All statistical analyses were performed using the Student’s t-test with the excel software.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

We are grateful to Professor K. Kaibuchi for his generous gift of the recombinant plasmid encoding GFP–kif3-C2. We also thank Dr M. Chabrillat for helping us to set up the microtubule motility assay, Dr C. Leduc for providing us with biotinylated kinesin-1-coated polystyrene beads and Dr M. Morgan and C. Almeida for critical reading of the manuscript. This work was supported by a grant to E. C. and C. W. from ‘Action Concertée Interface physique-chimie-biologie’.

References

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information
  • 1
    Rink J, Ghigo E, Kalaidzidis Y, Zerial M. Rab conversion as a mechanism of progression from early to late endosomes. Cell 2005;122:735749.
  • 2
    Burkhardt JK, Echeverri CJ, Nilsson T, Vallee RB. Overexpression of the dynamitin (p50) subunit of the dynactin complex disrupts dynein-dependent maintenance of membrane organelle distribution. J Cell Biol 1997;139:469484.
  • 3
    Harada A, Takei Y, Kanai Y, Tanaka Y, Nonaka S, Hirokawa N. Golgi vesiculation and lysosome dispersion in cells lacking cytoplasmic dynein. J Cell Biol 1998;141:5159.
  • 4
    Hollenbeck PJ, Swanson JA. Radial extension of macrophage tubular lysosomes supported by kinesin. Nature 1990;346:864866.
  • 5
    King SJ, Brown CL, Maier KC, Quintyne NJ, Schroer TA. Analysis of the dynein-dynactin interaction in vitro and in vivo. Mol Biol Cell 2003;14:50895097.
  • 6
    Lebrand C, Corti M, Goodson H, Cosson P, Cavalli V, Mayran N, Faure J, Gruenberg J. Late endosome motility depends on lipids via the small GTPase Rab7. EMBO J 2002;21:12891300.
  • 7
    Lin SX, Gundersen GG, Maxfield FR. Export from pericentriolar endocytic recycling compartment to cell surface depends on stable, detyrosinated (glu) microtubules and kinesin. Mol Biol Cell 2002;13:96109.
  • 8
    Nakata T, Hirokawa N. Point mutation of adenosine triphosphate-binding motif generated rigor kinesin that selectively blocks anterograde lysosome membrane transport. J Cell Biol 1995;131:10391053.
  • 9
    Valetti C, Wetzel DM, Schrader M, Hasbani MJ, Gill SR, Kreis TE, Schroer TA. Role of dynactin in endocytic traffic: effects of dynamitin overexpression and colocalization with CLIP-170. Mol Biol Cell 1999;10:41074120.
  • 10
    Brown CL, Maier KC, Stauber T, Ginkel LM, Wordeman L, Vernos I, Schroer TA. Kinesin-2 is a motor for late endosomes and lysosomes. Traffic 2005;6:11141124.
  • 11
    Tanaka Y, Kanai Y, Okada Y, Nonaka S, Takeda S, Harada A, Hirokawa N. Targeted disruption of mouse conventional kinesin heavy chain, kif5B, results in abnormal perinuclear clustering of mitochondria. Cell 1998;93:11471158.
  • 12
    Lane JD, Allan VJ. Microtubule-based endoplasmic reticulum motility in Xenopus laevis: activation of membrane-associated kinesin during development. Mol Biol Cell 1999;10:19091922.
  • 13
    Lippincott-Schwartz J, Cole NB, Marotta A, Conrad PA, Bloom GS. Kinesin is the motor for microtubule-mediated Golgi-to-ER membrane traffic. J Cell Biol 1995;128:293306.
  • 14
    Vaughan KT. Microtubule plus ends, motors, and traffic of Golgi membranes. Biochim Biophys Acta 2005;1744:316324.
  • 15
    Bannai H, Inoue T, Nakayama T, Hattori M, Mikoshiba K. Kinesin dependent, rapid, bi-directional transport of ER sub-compartment in dendrites of hippocampal neurons. J Cell Sci 2004;117:163175.
  • 16
    Toyoshima I, Yu H, Steuer ER, Sheetz MP. Kinectin, a major kinesin-binding protein on ER. J Cell Biol 1992;118:11211131.
  • 17
    Chuang JZ, Yeh TY, Bollati F, Conde C, Canavosio F, Caceres A, Caceres A, Sung CH. The dynein light chain Tctex-1 has a dynein-independent role in actin remodeling during neurite outgrowth. Dev Cell 2005;9:7586.
  • 18
    Ghosh-Roy A, Desai BS, Ray K. Dynein light chain 1 regulates dynamin-mediated F-actin assembly during sperm individualization in Drosophila. Mol Biol Cell 2005;16:31073116.
  • 19
    Wang T, Hong W. Interorganellar regulation of lysosome positioning by the Golgi apparatus through Rab34 interaction with Rab-interacting lysosomal protein. Mol Biol Cell 2002;13:43174332.
  • 20
    Bananis E, Murray JW, Stockert RJ, Satir P, Wolkoff AW. Microtubule and motor-dependent endocytic vesicle sorting in vitro. J Cell Biol 2000;151:179186.
  • 21
    Bananis E, Murray JW, Stockert RJ, Satir P, Wolkoff AW. Regulation of early endocytic vesicle motility and fission in a reconstituted system. J Cell Sci 2003;116:27492761.
  • 22
    Bananis E, Nath S, Gordon K, Satir P, Stockert RJ, Murray JW, Wolkoff AW. Microtubule-dependent movement of late endocytic vesicles in vitro: requirements for dynein and kinesin. Mol Biol Cell 2004;15:36883697.
  • 23
    Nath S, Bananis E, Sarkar S, Stockert RJ, Sperry AO, Murray JW, Wolkoff AW. Kif5B and Kifc1 interact and are required for motility and fission of early endocytic vesicles in mouse liver. Mol Biol Cell 2007;18:18391849.
  • 24
    Murray JW, Bananis E, Wolkoff AW. Reconstitution of ATP-dependent movement of endocytic vesicles along microtubules in vitro: an oscillatory bidirectional process. Mol Biol Cell 2000;11:419433.
  • 25
    Nielsen E, Severin F, Backer JM, Hyman AA, Zerial M. Rab5 regulates motility of early endosomes on microtubules. Nat Cell Biol 1999;1:376382.
  • 26
    Tourinho F, Franck R, Massart R. Aqueous ferrofluids based on manganese and cobalt ferrites. J Mater Sci 1990;25:3249.
  • 27
    Wilhelm C, Billotey C, Roger J, Pons JN, Bacri JC, Gazeau F. Intracellular uptake of anionic superparamagnetic nanoparticles as a function of their surface coating. Biomaterials 2003;24:100111.
  • 28
    Leduc C, Campas O, Zeldovich KB, Roux A, Jolimaitre P, Bourel-Bonnet L, Goud B, Joanny JF, Bassereau P, Prost J. Cooperative extraction of membrane nanotubes by molecular motors. Proc Natl Acad Sci U S A 2004;101:1709617101.
  • 29
    Hoepfner S, Severin F, Cabezas A, Habermann B, Runge A, Gillooly D, Stenmark H. Degradation by the endosomal kinesin KIF16B. Cell 2005;121:437450.
  • 30
    Cordonnier MN, Dauzonne D, Louvard D, Coudrier E. Actin filaments and myosin I alpha cooperate with microtubules for the movement of lysosomes. Mol Biol Cell 2001;12:40134029.
  • 31
    Ingold AL, Cohn SA, Scholey JM. Inhibition of kinesin-driven microtubule motility by monoclonal antibodies to kinesin heavy chains. J Cell Biol 1988;107:26572667.
  • 32
    Steffen W, Karki S, Vaughan KT, Vallee RB, Holzbaur EL, Weiss DG, Kuznetsov SA. The involvement of the intermediate chain of cytoplasmic dynein in binding the motor complex to membranous organelles of Xenopus oocytes. Mol Biol Cell 1997;8:20772088.
  • 33
    He Y, Francis F, Myers KA, Yu W, Black MM, Baas PW. Role of cytoplasmic dynein in the axonal transport of microtubules and neurofilaments. J Cell Biol 2005;168:697703.
  • 34
    Durrbach A, Louvard D, Coudrier E. Actin filaments facilitate two steps of endocytosis. J Cell Sci 1996;109:457465.
  • 35
    Lyubimova A, Bershadsky AD, Ben-Ze’ev A. Autoregulation of actin synthesis responds to monomeric actin levels. J Cell Biochem 1997;65:469478.
  • 36
    Nishimura T, Kato K, Yamaguchi T, Fukata Y, Ohno S, Kaibuchi K. Role of the PAR-3-KIF3 complex in the establishment of neuronal polarity. Nat Cell Biol 2004;6:328334.
  • 37
    Abney JR, Meliza CD, Cutler B, Kingma M, Lochner JE, Scalettar BA. Real-time imaging of the dynamics of secretory granules in growth cones. Biophys J 1999;77:28872895.
  • 38
    Amblard F, Maggs AC, Yurke B, Pargellis A, Leibler S. Subdiffusion and anomalous local viscoelasticity in actin networks. Phys Rev Lett 1996;77:44704473.
  • 39
    Caspi A, Granek R, Elbaum M. Diffusion and directed motion in cellular transport. Phys Rev E Stat Nonlin Soft Matter Phys 2002;66:011916.
  • 40
    Chabrillat ML, Wilhelm C, Wasmeier C, Sviderskaya EV, Louvard D, Coudrier E. Rab8 regulates the actin-based movement of melanosomes. Mol Biol Cell 2005;16:16401650.
  • 41
    Deacon SW, Serpinskaya AS, Vaughan PS, Lopez Fanarraga M, Vernos I, Vaughan KT, Gelfand VI. Dynactin is required for bidirectional organelle transport. J Cell Biol 2003;160:297301.
  • 42
    Gotthardt D, Warnatz HJ, Henschel O, Bruckert F, Schleicher M, Soldati T. High-resolution dissection of phagosome maturation reveals distinct membrane trafficking phases. Mol Biol Cell 2002;13:35083520.
  • 43
    Li HS, Stolz DB, Romero G. Characterization of endocytic vesicles using magnetic microbeads coated with signalling ligands. Traffic 2005;6:324334.
  • 44
    Perrin-Cocon LA, Marche PN, Villiers CL. Purification of intracellular compartments involved in antigen processing: a new method based on magnetic sorting. Biochem J 1999;338:123130.
  • 45
    Driskell OJ, Mironov A, Allan VJ, Woodman PG. Dynein is required for receptor sorting and the morphogenesis of early endosomes. Nat Cell Biol 2007;9:113120.
  • 46
    Lenz JH, Schuchardt I, Straube A, Steinberg G. A dynein loading zone for retrograde endosome motility at microtubule plus-ends. EMBO J 2006;25:22752286.
  • 47
    Wedlich-Soldner R, Straube A, Friedrich MW, Steinberg G. A balance of KIF1A-like kinesin and dynein organizes early endosomes in the fungus Ustilago maydis. EMBO J 2002;21:29462957.
  • 48
    Habermann A, Schroer TA, Griffiths G, Burkhardt JK. Immunolocalization of cytoplasmic dynein and dynactin subunits in cultured macrophages: enrichment on early endocytic organelles. J Cell Sci 2001;114:229240.
  • 49
    Kural C, Kim H, Syed S, Goshima G, Gelfand VI, Selvin PR. Kinesin and dynein move a peroxisome in vivo: a tug-of-war or coordinated movement? Science 2005;308:14691472.
  • 50
    Raposo G, Cordonnier MN, Tenza D, Menichi B, Durrbach A, Louvard D, Coudrier E. Association of myosin I alpha with endosomes and lysosomes in mammalian cells. Mol Biol Cell 1999;10:14771494.
  • 51
    Jasmin BJ, Cartaud J, Bornens M, Changeux JP. Golgi apparatus in chick skeletal muscle: changes in its distribution during end plate development and after denervation. Proc Natl Acad Sci U S A 1989;86:72187222.
  • 52
    Tai AW, Chuang JZ, Sung CH. Localization of Tctex-1, a cytoplasmic dynein light chain, to the Golgi apparatus and evidence for dynein complex heterogeneity. J Biol Chem 1998;273:1963919649.
  • 53
    Niclas J, Navone F, Hom-Booher N, Vale RD. Cloning and localization of a conventional kinesin motor expressed exclusively in neurons. Neuron 1994;12:10591072.
  • 54
    Louvard D, Reggio H, Warren G. Antibodies to the Golgi complex and the rough endoplasmic reticulum. J Cell Biol 1982;92:92107.
  • 55
    Rios RM, Sanchis A, Tassin AM, Fedriani C, Bornens M. GMAP-210 recruits gamma-tubulin complexes to cis-Golgi membranes and is required for Golgi ribbon formation. Cell 2004;118:323335.
  • 56
    Fauconnier RN, Bee A, Roger J, Pons JN. Adsorption of gluconic and citric acids on maghemite particles in aqueous medium. Prog Colloid Polym Sci 1996;100:212216.
  • 57
    Plastino J, Lelidis I, Prost J, Sykes C. The effect of diffusion, depolymerization and nucleation promoting factors on actin gel growth. Eur Biophys J 2004;33:310320.

Supporting Information

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Figure S1: The magnetic fractions isolated after 30 and 90 min internalization of the magnetic nanoparticles have been observed by phase-contrast microscopy (A, C, E and G) or labeled with anti-Rab6 (B and F) or anti-ER antibodies (E and H) after injection in a flow cell coated with polylysine. Arrowheads mark isolated endosomes and arrows mark aggregates. Note that none of the isolated endosomes are labeled with anti-ER or anti-Golgi antibodies. Bars represent 8.7 μm.

Movie S1: Magnetic late endosomes undergo bidirectional movements on microtubules in vitro. Phase-contrast images of a magnetic late endosome were overlaid with the fluorescent image of the microtubules. The movie is accelerated 10-fold and represents 116 images shot at one image per second.

Movie S2: Magnetic early endosomes undergo bidirectional movements on microtubules in vitro. Phase-contrast images of a magnetic early endosome were overlaid with the fluorescent image of the microtubules. The movie is accelerated 10-fold and represents 111 images shot at one image per second.

Movie S3: Magnetic endosomes sometimes oscillate on microtubules in vitro. Phase-contrast images of a magnetic late endosome were overlaid with the fluorescent image of the microtubules. The movie is accelerated 10-fold and represents 122 images shot at one image per second.

Movie S4: HeLa cells were imaged after transfection with DYNC1H1 siRNA, internalization of BODIPY-TR pepstatin A as described in the Materials and Methods. The movie is accelerated 25-fold and represents 241 images shot at one image per second.

Movie S5: HeLa cells were imaged after transfection with DYNC1H1 siRNA, internalization of BODIPY-TR pepstatin A as described in the Materials and Methods and treatment with latrunculin A (1 μm during 10 min). This treatment was optimized in order to depolymerize actin filaments without perturbing the microtubules. This latrunculin A treatment is shorter than the treatment used for the distribution of the compartments after cell fixation (Figure 7F–H), which explains that some lysosomes are still present at the periphery of the cell. After latrunculin A treatment, lysosomes still exhibited fast bidirectional motion. The movie is accelerated 25-fold and represents 241 images shot at one image per second.

Movie S6: HeLa cells were imaged after transfection with DYNC1H1 siRNA, internalization of BODIPY-TR pepstatin A as described in the Materials and Methods and treatment with nocodazole (10 μm during 1 h). This treatment was optimized in order to depolymerize microtubules without perturbing the actin filaments. Nocodazole treatment strongly inhibited lysosomal motion. The movie is accelerated 25-fold and represents 237 images shot at one image per second.

Supplemental materials are available as part of the online article at http://www.blackwell-synergy.com

FilenameFormatSizeDescription
TRA_704_sm_FigureS1.jpg218KSupporting info item
TRA_704_sm_MovieS1.avi1339KSupporting info item
TRA_704_sm_MovieS2.avi860KSupporting info item
TRA_704_sm_MovieS3.avi783KSupporting info item
TRA_704_sm_MovieS4.mov1923KSupporting info item
TRA_704_sm_MovieS5.mov4028KSupporting info item
TRA_704_sm_MovieS6.mov4480KSupporting info item
TRA_figures1.jpg218KSupporting info item
TRA_movies1.avi1339KSupporting info item
TRA_movies2.avi860KSupporting info item
TRA_movies3.avi783KSupporting info item
TRA_movies4.mov1923KSupporting info item
TRA_movies5.mov4028KSupporting info item
TRA_movies6.mov4480KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.