Lis1 co-localizes with actin in the Phagocytic cup and regulates Phagocytosis

Abstract

Phagocytosis, the ingestion of solid particles by cells is essential for nutrient uptake, innate immune response, antigen presentation and organelle homeostasis. Here we show that Lissencephaly-1 (Lis1), a well-known regulator of the microtubule motor Dynein, co-localises with actin at the phagocytic cup in the early stages of phagocytosis. Both knockdown and overexpression of Lis1 perturb phagocytosis, suggesting that an optimum level of Lis1 is required to regulate actin dynamics within the phagocytic cup during particle engulfment. This requirement of Lis1 is replicated in mouse macrophage cells as well as in the amoeba Dictyostelium, indicating an evolutionarily conserved role for Lis1 in phagocytosis. In support of these findings, a role for Lis1 in regulating actin dynamics is suggested by observing defective migration of Dictyostelium cells overexpressing Lis1. Taken together, Lis1 localises to the phagocytic cup and influences actin dynamics in a manner that appears important for the uptake of solid particles in cells.

Introduction

Phagocytosis involves ingestion of foreign particles larger than ~500nm by cells. Phagocytosis is strongly dependent on the actin cytoskeleton, wherein actin acts as a scaffold to form membrane protrusions around the particle at the cell cortex (reviewed in [1]). Lissencephaly-1 (Lis1) is a member of the conserved WD (tryptophan-aspartic acid) family of proteins classically known to interact with and modulate the function of microtubule-based dynein motors [2][3][4]. However, outside this classical dynein-dependent pathway, Lis1 also acts as a scaffold for actin-nucleating machinery at the leading edge of neuronal growth cones [5][6]. Lis1 also cross-talks with the extracellular adhesion complex [7] and is necessary for generating force against the substratum for cell migration. It is also known that knockdown of Lis1 reduces F-actin concentration at the leading edge of neurons [6]. These observations suggest an additional role for Lis1 in actin-dependent remodeling of membranous structures. Actin driven remodeling of the plasma membrane also happens, albeit on a localized scale, during phagocytosis(reviewed in [1]). Actin nucleation at the site of phagocytosis generates the force to form protrusions (phagocytic cup) around the particle. Upon completion of engulfment, the phagocytosed particle is enclosed in an actin-enriched bilayer membrane [8].

To probe the potential importance of Lis1 in phagocytosis we investigated the connection between Lis1, actin and phagocytosis in mouse macrophage cells and in the social amoeba Dictyostelium. These two cell types were chosen to demonstrate the evolutionarily conserved role for Lis1 in phagocytosis. Dictyostelium cells are highly phagocytic, and are a good model system for this study because Dictyostelium Lis1 (DdLis1) and human Lis1 are 53% identical with similar domain structure. Further, a role for Lis1 in controlling actin dynamics is known in Dictyostelium [9]. It was proposed that impairment of DdLis1 reduces F-actin content to alter actin dynamics, and a direct interaction of DdLis1 with DdRac1 (a Rho-GTPase) was also shown, suggesting that DdLis1 promotes actin polymerization via the well-known Rho-GTPase pathway [9]. DdRac1 delivery to the cell cortex may also require dynein, and indeed it was suggested that DdLis1 affects actin dynamics via its effect on dynein function [9]. Lis1 could therefore regulate the actin machinery via a direct actin-dependent mechanism, or an indirect dynein dependent pathway.

Here we show that Lis1 co-localises with actin in the phagocytic cup during the early stages of particle capture and engulfment. Interestingly, Lis1 is reduced as the phagosome undergoes maturation inside cells, suggesting a specific role for Lis1 during early events in the phagocytic pathway. Knockdown and overexpression of Lis1 leads to defects in phagocytosis, including reduced phagocytic efficiency and incomplete formation of the phagosomal cup. Such perturbations also lead to defects in cell motility, suggesting that Lis1 alters actin dynamics to cause defects in phagocytosis. Taken together, a specific role for Lis1 is suggested in actin-dependent remodelling of membranes during early stages of the phagocytic pathway.

Results and Discussion

Effects of Lis1 Knockdown and Overexpression on Phagocytosis

We have reported detailed protocols for the purification of latex bead phagosomes from Dictyostelium and mouse macrophage cells [10][11][12]. Latex bead phagosomes are amenable to precise biophysical force measurements for studying the activity of motor proteins inside cells [13] and in in-vitro assays [10]. Beads are added (pulse) to cell culture medium, and then “chased” inside cells for defined durations, allowing phagosome properties to be investigated at well-defined stages during the phagocytic pathway. Latex beads impart buoyancy to the phagosomes, allowing their purification by centrifugation in large quantities for biochemical studies. Further, all latex bead phagosomes have a spherical shape with the same size, allowing us to compare phagosome cup formation quantitatively across individual phagosomes in cells subjected to different conditions. These unique advantages of latex bead phagosomes will be used here to probe the possible role of Lis1 in phagocytosis.

We fed silica beads (D = 2.5μm diameter) to RAW 264.7 macrophages. The cells were previously transfected (Methods) with either scrambled shRNA or Lis1 specific shRNA by electroporation [14]. To identify transfected cells, a GFP plasmid was co-transfected with the Lis1 shRNA plasmids. In western blots of cell lysate using antibody against Lis1, the band intensity in Lis1 shRNA cells was reduced by ~50% (Supplementary Fig 1A). Such conditions of mild Lis1 knockdown were deliberately chosen to minimize possible off-target effects as a result of Lis1 knockdown. Beads were incubated with RAW 264.7 macrophages for 5 minutes at a density of 100 beads/cell. Cells were then fixed and imaged on a confocal microscope. Beads appeared as dark circles on a green background in such images (Fig 1A), where a single representative confocal section of RAW cells is shown. To estimate the total number of beads/cell we summed up all beads across the depth of a cell (right panel of Fig 1A). The mean number of beads/cell across many cells (at least 20 cells for each condition) was taken as a measure of phagocytic efficiency. Lis1 shRNA cells exhibited reduced phagocytic efficiency compared to scrambled shRNA (Fig 1A) suggesting that Lis1 is required for efficient phagocytosis.

Fig. 1. Optimal levels of Lis1 are required for early events in phagocytosis. Similarly, Optimal Actin balance is necessary for phagocytosis.

A. Confocal sections of RAW 264.7 macrophage cells co-expressing GFP plasmid with indicated shRNA. Scatter plot shows quantifications of beads ingested per cell across all the z-stacks imaged. Each dot represents the number of phagosomes in one cell. Between 20 to 30 cells were counted for each set across 2-3 independent experiments. Student’s t-test was performed to obtain the mentioned P values. Mean values (horizontal lines) and SEM (whiskers) are indicated.

B. Confocal section of RAW 264.7 macrophage cells expressing indicated plasmids. Scatter plot shows quantifications of beads ingested per cell across all the z-stacks imaged. Student’s t-test was performed to obtain the mentioned P values. Between 22 to 41 cells were counted for each set across 2-3 independent experiments. Mean values (horizontal lines) and SEM (whiskers) are indicated.

C. DIC images of either WT or GFP overexpressing Dictyostelium cells treated with 750nm polystyrene beads (ratio of 100 beads per cell) in presence of DMSO. Polystyrene beads appear as bright circles. Scale Bar = 5 microns in all cases.

D. DIC images of Dictyostelium cells overexpressing control GFP construct treated with 750nm polystyrene beads in presence of Jasplakinolide or LatrunculinA.

E. DIC images of Dictyostelium cells overexpressing GFP-Lis1 construct treated with 750nm polystyrene beads in presence of Vehicle (DMSO) or LatrunculinA.

F. Scatter plots showing phagocytic index of cells as mentioned. P-values are obtained using Student’s t-test. Between 32 to 80 cells were counted for each set across 2-3 independent experiments. Mean values (horizontal lines) and SEM (whiskers) are indicated.

G. Magnified to show the difference in phagocytic efficiency between DMSO and LatA treated cells.

Next, we overexpressed GFP-Lis1 in RAW cells. GFP cDNA was transfected as a control. Because Lis1 knockdown inhibited phagocytosis (Fig 1A), we expected to observe enhanced phagocytosis after Lis1 overexpression. Contrary to our expectations, overexpression of GFP-Lis1 also reduced bead uptake (Fig 1B). As discussed earlier, Fig 1B (left panel) shows a single Z-section, but the total number of beads/cell (at least 30 cells for each condition) was obtained after counting beads across all Z-sections (Fig 1B, right panel).

Effects of Actin stabilization and depolymerization on Phagocytosis in Dictyostelium

The above observations suggested that an optimal level of Lis1 is required, possibly to regulate actin dynamics, during phagocytic uptake of particles. To investigate further we used pharmacological agents that destabilize or stabilize F-actin in Dictyostelium cells. As explained earlier, Dictyostelium was chosen to investigate a possible evolutionarily conserved function of Lis1 in phagocytosis, and also because Lis1 is reported to control actin dynamics in Dictyostelium [9]. We standardized a latex bead assay to measure phagocytic efficiency where Dictyostelium cells in suspension culture were fed with beads for a specified period, fixed with paraformaldehyde, and then flattened on a coverslip by overlaying an agar sheet on the cells before imaging to count the number of ingested beads inside individual cells. We first verified that Dictyostelium cells overexpressing the GFP-only vector (control) exhibited robust phagocytosis that was comparable to untransfected cells in this latex bead assay (Fig 1C, 1F). We next treated GFP-vector cells with LatrunculinA (LatA), a drug that sequesters cellular G-actin leading to depolymerisation of actin filaments [15]. Bead uptake was inhibited in cells treated with 0.3 μM LatA (Fig 1D, 1F), thus verifying a role for the actin cytoskeleton in phagocytosis in our experimental system. Next, we treated cells with jasplakinolide, a drug that has the opposite effect of inducing actin polymerization and/or stabilising pre-existing F-actin filaments [16]. Phagocytosis was again inhibited in cells treated with jasplakinolide (Fig 1D, 1F). This data again suggests that an optimal balance of F and G-actin is necessary for phagocytosis in Dictyostelium and sets the stage for the next set of experiments.

Because both knockdown and overexpression of Lis1 had the same effect of reducing phagocytosis in RAW 264.7 cells (Figs 1A, 1B), we hypothesized that Lis1 helps maintain a balance between F and G-actin, and this balance ensures optimal actin dynamics leading to phagocytic cup formation. Although a role for Lis1 in phagocytosis has not been discussed previously, our hypothesis derives strength from the results of genetic perturbation where the actin cytoskeleton was necessary for particle uptake by phagocytosis [17][1]. Stabilizing, as well as destabilizing F-actin cytoskeleton by genetically disrupting its modulators caused defects in phagocytic uptake [18][19]. Further, Lis1 +/- heterozygous neurons show a reduction in F-actin content [5], implying that Lis1 promotes the formation of F-actin in cells. Therefore, Lis1 overexpression may increase F-actin to inhibit the dynamic re-organization of actin necessary for cup formation around phagosomes. Supplementary Fig 1B shows a typical result, where the intensity of GFP-Lis1 band is ~1.8-fold more than endogenous Lis1 in His-GFP-Lis1 (hereforth, GFP-Lis1) overexpressing cells. Such conditions of mild overexpression of Lis1 were deliberately chosen to minimize possible off-target effects of Lis1 overexpression. If Lis1 promotes F-actin formation, then mild depolymerisation of actin with LatA in Lis1-overexpressing cells should rescue phagocytosis. Indeed, when Dictyostelium cells overexpressing GFP-Lis1 were treated with 0.3 μM LatA, a rescue of the phagocytic index was observed (Fig 1E, 1F; also see the magnified panel in Fig 1G).

Notably, we did not observe any difference in the F/G actin ratio by using an actin-pelleting assay on Dictyostelium cells (Fig 2A; also see Methods). This suggests that the effect of GFP-Lis1 overexpression on actin dynamics in our experiments is local (and not global), perhaps kicking in when actin reorganization is required during phagocytic cup formation. Our results differ from that of ref. [9], where the over-expression of maltose binding protein (MBP)-tagged Lis1 reduced F-actin in an actin pelleting assay, thus suggesting global perturbations in actin. A dominant negative mutant of Lis1 is known to reduce F-actin concentration at the leading edge of neurons [6], and the MBP-tagged Lis1 overexpressed in Dictyostelium was inferred to act in dominant negative manner with the effect that radially organized microtubules were disrupted and the nucleus-to-centrosome distance was increased [9]. These phenotypes could have arisen in reference [9] because the dominant negative MBP-Lis1 was expressed ~10 fold in excess compared to endogenous Lis1 in Dictyostelium cells. In contrast to that work, we believe that the overexpressed GFP-Lis1 in our case does not function in a dominant negative manner. This is because GFP-Lis1 could pull down dynein from Dictyostelium cell lysate (Fig 2B), suggesting that the overexpressed Lis1 is functioning “normally” (i.e. it can interact with dynein). Further, unlike the results in [9], we did not find any gross defect in the organization of microtubules in GFP-Lis1 overexpressing cells, and the distance between nucleus and centrosome was also unchanged (Fig 2C). As noted earlier, unlike the effect of MBP-Lis1 overexpression in ref. [9], there was no change in F/G actin ratio after GFP-Lis1 overexpression under our experimental conditions. Lastly, the overexpressed GFP-Lis1 was observed to bind microtubules in a microtubule pelleting assay (Fig 2D). Overall, these findings suggest that the mild GFP-Lis1 overexpression in our studies functions like endogenous (normal) Lis1, and does not cause a dominant negative effect like the MBP tagged Lis1 reported in Ref. [9].

Fig. 2. Total cellular Actin content, microtubule organization and centrosome-nucleus distance is not altered upon Lis1 overexpression. DdGFP-Lis1 interacts with Dynein and microtubules.

A. Western Blot and densitometric quantitation of actin pelleting assay. Relative actin content in the pellet and supernatant did not change upon Lis1 overexpression (right panel). Result is a representative of 3 independent experiments. P-value as obtained from Student’s t-test.

B. Pull down of Dictyostelium cell lysates shows interaction of overexpressed Lis1 with endogenous dynein. Cells used were overexpressing His-GFP-Lis1 or His-GFP (negative control). Dynein heavy chain is detected only in pulldown fraction of His-GFP-Lis1. Lis1 antiserum was used to detect the Lis1 protein. Equal protein amount was used for both samples.

C. Representative confocal images of Dictyostelium cells overexpressing GFP or GFP-Lis1 constructs. Radial organization of microtubules is maintained in both cases. Images are maximum intensity projections of z-stacks taken through the thickness of cells. Scale Bar = 5μm. Scatter plots for the distance between centrosomes (white arrowhead) and nuclei of cells. n = 32 cells across 3 independent experiments in each case. P value as obtained by Student’s t-test. Horizontal line is the mean and whiskers show SEM.

D. Microtubule pull down assay with microtubules prepared from goat brain and cell lysate from Dictyostelium cells expressing GFP-only or GFP-Lis1. The pelleted microtubule samples were probed with antiserum against Lis1. Both GFP-Lis1 (higher mol. wt.) and endogenous Lis1 are found in the microtubule pellet.

Perturbations in Lis1 cause incomplete phagocytic cup formation

We next asked whether phagocytic cup formation is visibly different in cells with Lis1 over-expression. Dictyostelium cells can be incubated with beads for a short period (5 minutes; called pulse) to prepare “early phagosomes”, or pulsed followed by a chase period (45 minutes) to prepare “late phagosomes” as described by us [10][11]. This allowed us to assay for actin dynamics at a defined (early) stage of phagocytic uptake in Dicytostelium, for example, 125 seconds after addition of beads (see next). Rhodamine-phalloidin was used (see Methods) to visualize actin in the phagocytic cup around phagosomes [17]. Formation of actin around a bead is shown schematically in Fig 3A. The use of latex bead phagosomes of defined size allowed us to quantify the extent of cup formation around a single bead. To do this we measured the arc length of the actin cup that was not engulfed (= d ; see Fig 3A) around phagocytosed beads. Fig 3B shows images of phagocytosed beads in cells. Actin is stained using rhodaminephalloidin, and beads are visualized using differential interference contrast (DIC) microscopy. The insets in Fig 3B show the actin cup around individual phagosomes (inverted for clarity). The arc was linearized as a function of the angle (θ) around a circle centred at the bead by converting the image into polar coordinates (Fig 3C). The line intensity of this linearized arc was plotted as a function of θ (Fig 3D). We then measured the percentage of the circumference covered by the actin cup for individual beads [= 100 − (d/πD*100)]. Here D is the diameter and πD the circumference of the bead. We observed a significant reduction in phagocytic cup engulfment after Lis1 overexpression (Fig 3E), showing that cup formation is defective and/or delayed in Lis1 overexpressing cells.

Fig. 3. Overexpression of Lis1 causes defect in phagocytic cup progression.

A. Schematic showing progression of the actin-rich phagocytic cup. d = length of arc not engulfed at given a time-point; D = diameter of bead. % bead engulfed = 100-(d × 100/πD). When d = 0, bead is completely engulfed.

B. Representative confocal images of Dictyostelium cells expressing GFP-only (control) or Lis1-GFP plasmids, counterstained with rhodamine-phalloidin. Inset: LUT inverted hi-mag images of partially engulfed 2.5μm silica beads. Cells were fixed and stained at 125 seconds post pulse (see main text).

C. ROI of either GFP only or GFP-Lis1 cups along with their polar transforms. Rhodamine labelled arcs (inverted LUTs) are represented by line segment when radial segment (r) sweeps the angular function (Ɵ).

D. Line intensity plots of red dotted lines shown in A. Notice the shorter FWHM for GFP-only cells.

E. Bar graphs showing lengths of arcs (as % of total bead circumference) engulfed by cell. n = 45 beads across 3 independent experiments in each case. Student’s t-test used to obtain P value.

Lis1 localizes to the Actin cup on Phagosomes

Lis1 has been shown to interact with actin in-vitro[5][9]. To test if endogenous Lis1 is localized to the phagocytic cup, we performed dSTORM super-resolution imaging of phagocytic cups in RAW cells using an anti-Lis1 antibody and alexa488-phalloidin to visualise actin (Methods). We expect the optical resolution in these measurements to be ~20nm, as estimated using the gattaquant nanoruler (Methods). Fig 4A shows the localisation of Lis1 and actin in RAW cells treated with scrambled shRNA, and the localisation of actin in RAW cells treated with Lis1 shRNA. The insets (labelled “1” and “2”) show staining for Lis1 and actin around individual phagosomes.

Fig. 4. Lis1 localizes to the phagocytic cup, but thickness of phagocytic cup is not perturbed by Lis1 Knockdown.

A. Representative dSTORM images of RAW cells on scrambled shRNA or Lis1 shRNA background, fed with 2.5um silica beads. Boxed region encloses two phagosomes. Scale bars = 5μm. Insets: Magnified images of phagocytic cups with representative yellow lines along which fluorescence intensity was measured (see next).

B. Intensity traces along yellow lines (see insets in Fig 4A) across the cortical membrane of phagosomes. Error bars are SEM. Total of 7 beads used across 2 independent experiments for creating the averaged traces (see main text for statistical significance).

The upper panel of Fig 4B shows the fluorescence intensity along radial direction across the surface of a phagosome (see yellow line in insets) averaged over multiple phagosomes. Intensity was measured along the yellow lines by using the ImageJ “profile” function. Lis1 and actin both showed a sharp peak exactly at 1.25 μm (= bead radius; marked with green dashed line) in cells treated with scrambled shRNA. This suggests that both proteins localize to the surface of the early phagosome. The width (= full width at half maximum; FWHM) of this peak was 112±12 nm, in agreement with reports of the thickness of cortical actin [20][21]. Fig 4B (lower panel) shows the averaged peak for actin observed in Lis1 shRNA knockdown cells (see inset of Fig 4A). The peak for actin in Lis1 shRNA cells has a FWHM of 116±14 nm, statistically same as the width in scrambled shRNA knockdown cells (P = 0.86; 7 beads were analysed in 4 cells in each set across 2 independent experiments). These findings suggest that there is no gross perturbation in the thickness of the actin cup surrounding the phagosome after Lis1 knockdown. Rather, the (lateral) dynamics of actin on the phagosome surface may be perturbed because the actin cups were found unable to extend around the phagosome in absence of Lis1 (Fig 3E).

As noted earlier, GFP-Lis1 overexpression did not cause any change in the bulk F/G actin ratio (Fig 2A), but the overexpressed Lis1 may exert more localized effects on actin dynamics resulting in incomplete phagocytic cup formation (Fig 3E). Migration of Dictyostelium cells depends on actin dynamics [21][22]. It is possible therefore that GFP-Lis1 overexpression can perturb local actin dynamics (e.g. within pseudopods) to inhibit cell migration. To investigate this, we quantified the migration of Dictyostelium cells in a chemotaxis assay (Fig 5A; Methods)[24]. Indeed, the overexpression of Lis1 reduced migration velocity and caused cells to take more tortuous paths (Fig 5B). Our results agree with [25], where radial migration of neurons in developing mouse brain was delayed upon expression of Lis1. As mentioned earlier, it is possible that Lis1 affects actin dynamics locally (e.g. within pseudopods during cell migration), and therefore varying Lis1 (e.g. by knockdown or overexpression) caused no visible difference in the F/G actin ratio (see Fig 2A).

Fig. 5. Lis1 overexpression causes cell motility defects. Lis1 is enriched on Early phagosomes (EPs).

A. DIC images of Dictyostelium cells overexpressing GFP only or GFP-Lis1 constructs. Individual cells were tracked using the manual tracker plugin of ImageJ to extract speed of movement. Overlay shows tracks for individual cells that were tracked over 40mins. An under-agarose chemotaxis protocol for Dictyostelium was used (Methods)

B. Speed and Tortuosity measurements of migration. Individual cells were tracked using the manual tracker plugin of ImageJ to extract speed of movement. Schematic describes the meaning of tortuosity. The persistence of motion during a defined time interval was estimated from the tortuosity (τ). The meaning of τ is illustrated in the inset for a period of motion starting at (x1,y1) and ending at (x2,y2). τ = [Integrated distance travelled along entire path / Net displacement] = [Distance along red arrows / Distance along dotted red line]. The denominator is the square root of [(x2−x1)²−(y2−y1)²]. Lis1 overexpression leads to reduction in speed and increase in tortuosity of migration. n = 30 cells in each set across 3 independent experiments. Error bars are SEM. P-values are calculated using a Student’s t-test.

C. Representative immuno-blots showing Lis1 and Dynein on early phagosomes (EPs) or late phagosomes (LPs). Total protein extracted from equal number of EPs and LPs (~ 5×10¹⁰) was loaded in each lane. Dynein is a loading control (see main text for justification). Lis1 antiserum was generated and used for this experiment (Methods and Supplementary Fig 2). Densitometric quantification shows that Lis1 is ~2 fold higher on EPs while dynein is equal. Error bars indicated are SEM from 3 independent experiments. P-values are calculated using a Student’s t-test.

Lis1 levels are reduced with maturation of Phagosomes

We previously reported[10][12] significant changes in the lipid composition of a phagosome membrane as it matures from an “early phagosome” (EP) to “late phagosome” (LP). Our data suggests a specific role for Lis1 in actin cup formation at the early stage of phagocytosis. If this is true, Lis1 levels may be higher on EPs as compared to LPs. To test this, we next purified EPs and LPs from cells according to a protocol described by us [11]. Indeed, western blotting of purified EPs and LPs revealed that Lis1 was ~2-fold elevated on EPs (Fig 5C). The EP and LP samples were normalized by using their optical density, which also yielded equal protein content on both samples [11]. Because dynein levels are known to be unchanged across the EP-to-LP transition [10], dynein was used as a loading control to verify the normalization of EP and LP samples (Fig 5C). This observation supports a more prominent role for Lis1 during the early stages of phagosome formation, for example, during extension of the phagocytic cup.

To conclude, we have shown that both knockdown and overexpression of Lis1 causes defects in phagocytosis. This is likely due to perturbed actin dynamics because membrane protrusions around particles are defective upon Lis1 perturbation (Fig 3E). We provide evidence that the over expressed GFP-Lis1 in our studies is active. It is possible that when overexpressed, Lis1 recruits excess actin nucleating machinery locally to the site of phagocytosis, thereby rendering the actin cup less dynamic. We favour this possibility because stabilizing F-actin by drug treatment also inhibited phagocytosis (Fig 1F). Notably, a phenomenon of cortical actin balance has been reported using atomic force microscopy and computer simulations [20]. Both reducing and increasing cortical tension from its endogenous levels leads to perturbation in cell division. While we have demonstrated a role of Lis1 in phagocytosis, we do not have structural and mechanistic insight into how Lis1 scaffolds the actin nucleation machinery. A role for Lis1 in modulating dynein function is well accepted [2][3][25][26], but less is known about Lis1 function in controlling the actin machinery. Targeted mutagenesis in Lis1 coupled with the phagocytosis assay described here may help identify Lis1 mutations relevant to phagocytosis [28]. Screening point mutants of Lis1 for phagocytosis and migration defects might also help in identifying the domain of Lis1 which scaffolds the actin nucleation machinery.

Materials and Methods

Plasmids and cloning

Mammalian Lis1 shRNA construct was a kind gift from Prof. Subba Rao Gangi Shetty (IISc, Bangalore) and mammalian Lis1-GFP construct was a gift from Prof. Deanna Smith (University of South Carolina). For cloning Dictyostelium Lis1 (Dd Lis1), total RNA was isolated from Dictyostelium AX2 cells in exponential phase. cDNA was prepared from total RNA and Lis1 cDNA was PCR amplified Forward primer: GTCTAGATTATTGTAATTTCCAAAC

Reverse primer: GTGGATCCATGGTATTAACTTCA Lis1 was cloned after the C terminus of His-GFP tag in pTX-GFP vector between BamHI and XbaI sites.

Antibodies

Mammalian Lis1(ab2607), Kinesin-2(ab11259) antibodies were obtained from Abcam. GAPDH antibody was obtained from AbClonal Inc. The antibodies were used at 1:3000, 1:2000 and 1:2000 working dilutions respectively. The source and protocol for generation of Dictyostelium dynein antibody has been described in detail[29]. For Dictyostelium Lis1 antibody, DdLis1 was cloned in the bacterial vector pET28a between NdeI and HindIII sites. Purification of Dd Lis1 protein was done by expressing this construct in BL21 cells followed by affinity chromatography using Ni-NTA beads. Final dialysis of purified protein was performed in PBS. Bacterially purified Dd Lis1 was used as the antigen for Lis1 antibody generation in mice. 150 μg purified Dd Lis1 protein (100μl) was resuspended in 100ul of Freunds complete adjuvant and was injected intra-peritoneally in each of the two Swiss mice. Pre-immune serum was collected from these animals before injection of antigen. Primary injection was followed by three booster injections of 100μg Dd Lis1 in incomplete adjuvant at regular intervals of one month. After one month of the third booster, animals were sacrificed to obtain blood. Serum was separated from total blood by letting it stand at room temperature for 2 hours followed by a low speed spin at 3000rpm at 4 °C. The clarified serum obtained as supernatant was used as Lis1 antibody. 1 in 2000 dilution of serum in milk was used for detecting Lis1 band by western blotting. Lis1 serum identified Lis1 protein at the correct size from Dictyostelium lysates and purified protein whereas no band was obtained with control pre-immune serum. Dynein antiserum was also used as 1 in 2000 diluted in milk.

Cell culture

Dictyostelium

AX2 cells were maintained in HL5 media (Formedium, UK) on 10cm cell culture dishes containing 100 μg/mL Penicillin/Streptomycin, at 22°C. Cells were induced for development by plating them with Klebsiella aerogens on SM/5 agar plates. Developed spore bodies were picked and frozen in phosphate buffered glycerol as aliquots and stored at -80°C for long-term storage.

RAW 264.7 Macrophages: Mouse macrophage cells (hereafter RAW cells) were cultured in DMEM-High glucose (Sigma) supplied with 10% (v/v) heat inactivated foetal bovine serum (Himedia), under standard culture conditions (37°C, 5% CO₂, Humidified incubator.)

Cell Transduction

Lis1 shRNA Lentiviral particles were prepared in HEK293T cells and RAW 264.7 macrophages were transduced with these particles using protocol as recommended by Addgene. Transductant cells were selected on Puromycin (2.5 μg/mL). Stocks were frozen in liquid nitrogen. Lis1 knockdown was validated by Western Blotting. Lis1 band was identified at correct size in scrambled shRNA cells, but not in Lis1 shRNA cells.

Transfection

Dictyostelium

AX2 cells were electroporated with either empty pTX-GFP or pTX-GFP-Lis1 plasmid in 0.1cm gap electrode cuvettes using square wave protocol of BioRad Gene Pulser XCell (2 pulses each of 650V, 1 ms duration, 1500 ms interval). Transformed cells were cultured on HL5 medium containing 10μg/mL G418 as selectable resistance marker, sub-cultured from clonal colonies at 22°C and stored at -80°C as mentioned before.

RAW 264.7 Macrophages

RAW cells were electroporated as described in [14]. Briefly, 7.5 X 10⁶ cells were electroporated with plasmids of interest in 0.4cm gap electrode cuvettes using the exponential protocol of BioRad Gene Pulser XCell (250V, 960μF capacitance). Transfected cells were washed once in ice-cold PBS (pH 7.4) by centrifugation (200g for 5mins, 4°C) and plated on coverslips (CORNING no. 1.5, 22mm X 22mm) overnight before experiment.

Phagocytic Index Assay

Dictyostelium

750nm diameter latex beads (PolySciences Inc.) were washed thrice with Sorenson’s phosphate buffer and sonicated briefly. Cells were scraped from cultured dishes, washed with Sorenson’s buffer twice, and counted in a hemocytometer. Cell density was adjusted to 6 X 10⁶ /mL by diluting the suspension in HL5 medium. Cells were temporarily stored as 50μL suspension aliquots at 22°C. LatrunculinA and Jasplakinolide were diluted in DMSO at stock concentrations of 1mM each. All the reagents were obtained from Sigma. For initiating phagocytosis, a diluted bead suspension was directly added to one aliquot of cells, at 100:1 cell to bead ratio. Specified drug was diluted appropriately during experiment. For control experiment, equal volume of DMSO was added as vehicle. Cells were thoroughly mixed with beads by gentle tapping, and phagocytosis was allowed to progress for 5mins. Cells were flattened on coverslips (CORNING no.1.5, 40mm X 22mm) using 0.8% phosphate agar pieces (1cm X 1cm). Excess bead solution was gently aspirated out using wicks. Snapshots of individual cells were acquired on NIKON TE2000 DIC microscope, 100X oil objective N.A. 1.4, with a Cohu 4910 camera. Phagocytic Index was defined as the number of beads phagocytosed per cell.

RAW macrophages: 2.5μm diameter silica beads (PolySciences Inc.) were washed thrice with DMEM (Sigma) by centrifugation (200g for 5mins) and briefly sonicated. Transfected (or control) RAW cells cultured overnight on coverslips, as mentioned before, and were washed once with PBS, and re-supplied with culture medium. Phagocytosis was initiated by directly adding the bead suspension to coverslips. Cells were incubated in humidified incubator at 37°C and 5% CO₂ for 5 mins. Post bead pulse, cells were washed twice with ice cold PBS to remove non-phagocytosed beads, and fixed in 4% buffered paraformaldehyde (PFA), and further processed for fluorescence staining protocol.

Staining for fluorescence microscopy was performed by permeabilizing cells in PBS containing 0.5% v/v Triton X-100 (PBS-T), following blocking in 5% (w/v) BSA solution prepared in PBS-T (Abdil). Post blocking, cells were stained with Rhodamine-phalloidin at 1:400 dilution in Abdil for 15mins at room temperature. Coverslips were mounted on microscopy slides in VectaShield mounting medium (VectorLabs). Images were acquired on Olympus FV1200 CLSM under 100X Oil objective (N.A. 1.4). Entire cell volumes of individual cells were acquired as Z-stacks. Image analysis was performed by manually counting beads per cell in ImageJ. Beads appeared as hollow spheres on green (GFP) background.

Fractionation of Actin

The protocol was adapted from ref. [9]. In brief, GFP or GFP-Lis1 expressing Dictyostelium cells were harvested from culture dishes, pelleted at 900g, and washed thrice with PBS. Cells were lysed in HEPES buffer containing 0.1% Triton X-100. After estimating and normalization of total protein from both samples, the triton insoluble fraction was pelleted by centrifugation of lysates at 1,80,000g for 15mins. The pellet was resuspended in equal volumes. Supernatant, pellet were processed for SDS-PAGE and western blotting analysis. Total lysate was used as loading controls.

Measurement of Phagocytic Cup

Dictyostelium cells cultured on coverslips were challenged with 2.5μm diameter silica beads as described before, with a minor modification. Instead of 5min pulse, the cells were incubated with beads for only 2.5 mins to let phagocytosis complete partially. The cells were stained, mounted and imaged as described before. The imaging conditions were kept consistent across samples and experiment sets. Arc-like rhodamine-positive membrane protrusion surrounding the beads was chosen as ROI for analysis. The length of cup (arc of a circle) was calculated using polar transformer plugin of ImageJ (See Results).

Preparation of cell lysate and western blotting

Dictyostelium cells either expressing pTX-GFP (empty vector control) or pTX-GFP-Lis1 (Test) were scraped from 10cm confluent dishes, washed once by centrifugation, and resuspended in phosphate buffered 0.1% Triton X-100 solution, to obtain total cell lysates.

RAW cells expressing either scrambled or Lis1 shRNA were scraped from confluent T25 flasks, washed once in PBS by centrifugation, and resuspended in 0.1% Triton X-100 solution of phosphate buffered saline to obtain total cell lysate.

For immunoblotting, the blot was blocked in 5% (w/v) non-fat dry milk prepared in Tris-buffered 0.1% (v/v) Tween-20 (TBS-T) for 2 hours. It was then probed with primary antibodies made in 5% (w/v) TBS-T solution of BSA (Abdil) for 1.5 hours. After thorough washing with TBS-T, Horse-radish peroxidase conjugated donkey secondary antibody (Sigma; made in 5% TBS-T solution of milk) was added to the blot (1:10000 dilution) and incubated for 1.5 hours. All the incubations were carried out at room temperature. Blots were developed using chemi-luminescence substrate (Millipore, Merck) on AI600 blot developer (GE).

Pulldown of HisGFP Lis1 from Dictyostelium cell lysate

1mg total protein from lysates from HisGFP and HisGFPLis1 cells were incubated with Ni-NTA beads, previously blocked with 5% milk for 1 hour. Incubation of HisGFP and HisGFP Lis1 lysates with blocked beads was carried out at 4 degrees for 3 hours for His pull down. Beads were then extensively washed three times in Tris buffer pH 8 containing 1M NaCl. His tagged complexes were finally eluted from beads by adding Tris buffer pH 8 containing 400mM imidazole.

Microtubule pulldown of Lis1 from Dictyostelium lysate

800 μg total protein lysates were extracted from GFP-only and GFP-Lis1 cells. Each of these lysates were incubated with microtubules polymerized from 140 μg tubulin in the presence of 1mM GTP and 20μM taxol for 30 mins at 37 degrees. ATP in the lysate was depleted by addition of 3mM glucose, 1.3 Units of hexokinase and 4mM MgCl₂. Incubation of lysate with microtubules was carried out for 20 mins on ice following which microtubules were pelleted by centrifuging the samples at 1,00,000 g for 20 mins. Endogenous and overexpressed Lis1 co-pelleted with microtubules were probed with Lis1 antiserum using western blotting.

Cell Migration assay

The protocol was adapted from ref [24]. Plates were prepared by cutting out Agarose troughs from 1% phosphate agar. One trough was filled with buffered 2mM folate solution (pH 7.4). The plate was incubated at 22°C for 10 mins to let the folate form a chemotactic gradient in the Agarose. For the experiment, approximately 10⁴ cells either expressing pTX-GFP or pTX-GFP-Lis1 were added to empty trough. The cell motion was then recorded using time lapse microscopy using NIKON TE2000 microscope (10X objective) for 40 minutes. Cells were tracked using manual tracker plugin of ImageJ. Migration speed and persistence (tortuosity) was calculated for each cell type using pair-wise distance formula of each cell’s co-ordinate obtained.

Super-Resolution Microscopy (dSTORM)

For dSTORM imaging, RAW cells were cultured on coverslips expressing either Lis1 shRNA or scrambled shRNA. Cells were incubated with 2.5μm diameter silica beads for 2.5 minutes at 37°C. Cells were then fixed and incubated with Rabbit anti-Lis1 antibody (Abcam) at 1:300 dilution, and further stained with Alexa555-conjugated anti-rabbit secondary antibody (Invitrogen) at 1:500 dilution. Cells were counterstained with Alexa488-phalloidin (Invitrogen) at 1:250 dilution. Samples were mounted in STORM buffer as specified by the manufacturer.

Data was acquired on the NanoImager S Mark II from ONI (Oxford NanoImaging) with the lasers 405nm/150mW, 473nm/1W, 561nm/1W, 640nm/1W and dual emission channels split at 560nm. 12500 frames were acquired sequentially at 100% laser power in both samples for each channel. Image reconstruction and post-acquisition processing for background filtering was done on NimOS (Version 1) from ONI. The optical resolution under our conditions of imaging is estimated to be ~20nm, as measured using gattaquant 20 nm nanoruler (as per manufacturer guidelines). For co-localization, we performed channel mapping of two channels using tetra spec beads with an expected ~10nm standard error of mean. Localization precision filter of <50nm was applied for post processing and background correction of the images. For analysis, a line was drawn radially from the bead centre into the cell, and line intensities were plotted for both channels using ImageJ.