SRRF-stream imaging of optogenetically controlled furrow formation shows localized and coordinated endocytosis and exocytosis mediating membrane remodeling

Abstract

Cleavage furrow formation during cytokinesis involves extensive membrane remodeling. In the absence of methods to exert dynamic control over these processes, it has been a challenge to examine the basis of this remodeling. Here we used a subcellular optogenetic approach to induce this at will and found that furrow formation is mediated by actomyosin contractility, retrograde plasma membrane flow, localized decrease in membrane tension and endocytosis. FRAP, 4-D imaging and inhibition or upregulation of endocytosis or exocytosis show that ARF6 and Exo70 dependent localized exocytosis supports a potential model for intercellular bridge elongation. TIRF and Super Resolution Radial Fluctuation (SRRF) stream microscopy show localized VAMP2-mediated exocytosis and incorporation of membrane lipids from vesicles into the plasma membrane at the front edge of the nascent daughter cell. Thus, spatially separated but coordinated plasma membrane depletion and addition are likely contributors to membrane remodeling during cytokinetic processes.

Graphical Abstract

Cytokinesis is the last phase of cell division, which starts with anaphase and ends with telophase, resulting in the separation of the parent cell into two daughter cells^1–4. It is highly conserved in animal cells and has long been observed and studied⁵. During cytokinesis the cell exhibits shape changes which require membrane remodeling, and it is known that the cell recruits key vesicular proteins to the middle region^{6, 7}. However, it is still unclear how the complex vesicular trafficking machinery works together and how the cell uses endocytosis and exocytosis to incorporate membrane in a dynamic way to ultimately separate the cell and organelles into two independent cells.

Cytokinesis involves activation of the small GTPase RhoA in the middle of the dividing cell, which drives actomyosin contractility, the formation of the contractile ring, and the cleavage furrow^8–10. The intercellular bridge subsequently forms between two daughter cells prior to separation^11–13. Cleavage furrow and bridge formation require extensive remodeling of the plasma membrane¹³, and it is not well understood how this remodeling is accomplished. There is evidence that vesicular trafficking takes place in the furrow region. Trafficking proteins like Rab11 and Arf6 have been reported in the furrow region, and genetic depletion of these proteins has been shown to interfere with furrow formation^14–17. Additionally, inhibition of the exocyst complex, which is responsible for the last steps of exocytosis, inhibits furrow formation^{18, 19}.

Despite the evidence suggesting a role for exocytosis and vesicle trafficking in cytokinesis, it has not been clear how these processes contribute to remodeling of membranes that underlies furrow formation, an essential step in cytokinesis. Recent studies showed that de novo synthesis of lipids from the Golgi is not required⁷, but it is still unclear where these vesicles originate and if they are trafficked from the plasma membrane. If endocytosis plays a role, it is not clear whether it is global or localized and precisely how it contributes to the membrane remodeling during furrow or bridge formation. The location of exocytosis and whether and how it works in conjunction with endocytosis is also unclear.

Experiments to answer these questions are challenging due to the lack of methods to rapidly induce furrow formation at will. While synchronizing cells at a specific stage of the cell cycle has facilitated the studies of cytokinesis, it has been difficult to perturb and observe dynamic processes in a live cell because events occur over a time scale of hours²⁰. Subcellular optogenetic approaches which allow signaling protein activity to be induced in a rapid, reversible and spatiotemporally precise manner are attractive alternatives²¹. They can be used to optically direct cell behaviors that rely on spatially localized signaling activity such as cell migration or cleavage furrow formation^21–24. They allow simultaneous imaging of the cellular response in real time and facilitate tracking cell membrane structures and protein activity across the space of the cell. Here we used this optogenetic strategy which has been used to identify mechanisms at the basis of cell migration²³ to activate RhoA in the middle of a macrophage cell. This induced furrow formation, consistent with a previous report using a different optogenetic strategy to achieve RhoA activation in fibroblasts and HeLa cells²⁴. We used further optical activation of the furrow region and found that this induces elongation resembling intercellular bridge formation. Total Internal Reflection Fluorescence (TIRF) microscopy and Super Resolution Ratio Fluctuation (SRRF) stream imaging enabled us to visualize membrane incorporation and localized exocytosis at the front of the nascent daughter cells. Additionally, we used genetic and pharmacological perturbation. The results uncover a possible role for spatially localized and coordinated endocytosis and exocytosis in cleavage furrow and potentially intercellular bridge formation. These findings suggest that the cell is able to remodel the plasma membrane to form a furrow through plasma membrane flow to the furrow region, consequent localized endocytosis at the furrow and backs of the daughter cells, directional vesicular trafficking, and localized exocytosis to facilitate the formation of daughter cells.

RESULTS AND DISCUSSION

RhoA photoactivation in the middle of the cell induces cleavage furrow formation

A light-inducible dimerization (iLID) system²⁵ has recently been used to optically recruit LARG, a guanine nucleotide exchange factor (GEF) for RhoA, to the plasma membrane of RAW 264.7 macrophage cells, activate RhoA locally, and direct cell migration²³. This optogenetic tool consists of a plasma membrane-targeted component, iLID-CaaX, and a cytosolic component consisting of LARG fused to the peptide SspB. Upon blue light exposure, iLID undergoes a conformational change, which exposes the peptide SsrA, which then dimerizes with SspB. This allows LARG to be recruited to any part of the plasma membrane exposed to blue light (Figure 1A). Here we introduce a model using this optogenetic tool to study the membrane mechanism in cleavage furrow and intercellular bridge formation.

Figure 1. Localized optical RhoA activation drives cleavage furrow formation.

(A) Diagram of optical activation of RhoA using the iLID system. Plasma membrane-associated iLID undergoes a conformational change upon illumination with blue light, exposing SspB. The dimerization of SspB with SsrA recruits LARG to the plasma membrane.

(B) Diagram for optical control of cleavage furrow elongation by RhoA activation using the iLID system. Maintaining two photoactivation regions, each at the edge of the nascent daughter cells proximal to the furrow, allows for the formation of an intercellular bridge-like structure.

(C) Cleavage furrow formation via optogenetic activation of RhoA. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, plasma membrane marker Venus-KRasCT (green), and nuclear marker H2B-mCh (red). Data is representative of 11 out of 13 cells. See also Movie 1.

(D) Unphotoactivated cells do not form the cleavage furrow. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, plasma membrane marker Venus-KRasCT (green), and nuclear marker H2B-mCh (red). Data is representative of 5 of 5 cells. See also Movie 1.

(E) Optogenetic RhoA activation. Cell is expressing LARG-mCH-SspB (red), iLID-CaaX, and RhoA activity sensor Venus-rGBD (green). Data is representative of 11 out of 11 cells. See also Movie 1.

(F) Intensity profile of Venus-rGBD distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 7 cells). Gray region is SEM.

(G) Intensity profile of Venus-KRasCT distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 9 cells). Gray region is SEM.

For this and all subsequent data shown, experiments were repeated on at least two separate days. Rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec. Fig. 2SA shows the profile line used for the intensity analysis.

First, we evaluated whether our tool was able to induce the cleavage furrow. RAW cells were transfected with iLID-CaaX, LARG-mTurq-SspB, nuclear marker mCh-H2B, and Venus-KRasCT, which attaches to the inner leaflet of the plasma membrane and thus serves as a membrane marker. A narrow rectangular region in the middle of the cell was stimulated with pulses of 445 nm light every 5 s to induce RhoA activation in that region (Figure 1A,C). We found that when cells containing two or more nuclei were photoactivated at a location between the nuclei, within minutes of photoactivation the cell membrane ingressed between the nuclei and formed a cleavage furrow (Figure 1A,C and Movie 1). Macrophage cells can contain multiple nuclei^{26, 27}, so the presence of more than one nucleus in RAW cells is not indicative of native cytokinesis. Indeed, RAW cells that were transfected with iLID-CaaX, LARG-mTurq-SspB, nuclear marker mCh-H2B, and Venus-KRasCT and imaged every 5 seconds without photoactivation never formed an incision or furrow, even in the case of binucleated cells (Figure 1D and Movie 1). This confirmed that furrow formation observed in photoactivated cells is not a spontaneous process. Because imaging mTurquoise results in global photoactivation of iLID, we did not image LARG-mTurq-SspB in this and all subsequent experiments where this construct was used.

We found that prolonged photoactivation of RhoA in the middle of the cell led to the formation of an elongated structure resembling the intercellular bridge. The addition of a second photoactivation region, with each region positioned at the edge of the newly-forming daughter cells closest to the middle, maintained this elongated structure (Figure 1B). Therefore, a second photoactivation region was added after the completion of furrow formation in all subsequent experiments. The intercellular bridge is the final link between daughter cells before abscission^{28, 29}. Intercellular bridges closely resembling the elongated structures observed here have been seen in native cytokinesis in multiple cell types, such as epithelial cells^{30, 31} and fibroblasts³². Therefore, by inducing elongation in the furrow region with our optogenetic tool, we can potentially uncover a possible mechanism behind intercellular bridge formation. Since the newly-forming daughter cells move apart as the furrow forms and elongates, we refer to the regions of the cells distal to the furrow as the “front” and the region proximal to the furrow as the “back” of the daughter cells.

We used the RhoA activity sensor rGBD to verify that furrow formation was due to optogenetically-induced RhoA activity. Cells were transfected with LARG-mCh-SspB, iLID-CaaX, and Venus-rGBD. Upon photoactivation, LARG-mCh-SspB rapidly translocated to the photoactivated region, and RhoA activity simultaneously increased in the region of LARG-mCh-SspB accumulation (Figure 1E,F and S2A). Furrow ingression coincided spatially and temporally with RhoA activity increase, demonstrating that furrow formation was directly due to optogenetically-induced RhoA activation.

Consistent with previous results showing that optogenetic activation of RhoA in the middle of the cell induces actomyosin contractility and ingression²⁴, we found that F-actin and myosin also accumulate in the cleavage furrow (Figure S1A,B and Movie S1). This is in agreement with evidence showing that RhoA activation increases actomyosin contractility during cytokinesis^{8, 33} and inhibition of actin turnover and myosin results in cytokinesis defects^{34, 35}. Additionally, we saw increased F-actin and myosin in the elongated bridge-like structure, which could suggest that actomyosin contraction mediates intercellular bridge formation. The finding that the furrow and a structure resembling the intercellular bridge can be induced through optogenetic RhoA activation in the middle of the cell demonstrates the effectiveness of this optogenetic tool at inducing processes which are known to occur during cytokinesis.

Plasma membrane flows to the cleavage furrow

In addition to actin and myosin, we found that the plasma membrane marker KRasCT also accumulates in the region of RhoA activation, showing that membrane lipids flow to the furrow (Figure 1C,G,S2A and Movie 1). As a complimentary method for visualizing plasma membrane flow, we next used mCh-GL-GPI, which consists of mCh fused to membrane inserting glycosylphosphatidylinositol (GPI) lipid containing a N-glycosylation (GL) site. This targets the mCh to the outer leaflet of the plasma membrane, serving as an additional distinct plasma membrane marker. Cells were transfected with LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT, and mCh-GL-GPI. As furrow formation was induced by RhoA photoactivation in the middle of the cell, GL-GPI, like KRasCT, accumulated at the furrow (Figure 2A, Movie 2). This result additionally supports the occurrence of retrograde membrane flow during furrow formation.

Figure 2. Plasma membrane flow during cleavage furrow formation.

(A) Plasma membrane lipid dynamics in a cell undergoing optogenetically-induced furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, and plasma membrane markers Venus-KRasCT (green) and mCh-GL-GPI (red). Data is representative of 4 out of 4 cells. See also Movie 2.

(B) Movement of plasma membrane functionalized beads during furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, and Venus-KRasCT (green). NileRed-containing beads (red) are attached to integrins on the membrane surface. Data is representative of 15 out of 15 cells. See also Movie 2.

(C) Movement of plasma membrane functionalized beads during furrow elongation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, and Venus-KRasCT (green). NileRed-containing beads (red) are attached to integrins on the membrane surface. Data is representative of 5 out of 5 cells. See also Movie 2.

Rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec.

Lastly, we visualized plasma membrane dynamics using integrin-bound microbeads. Carboxyl-functionalized polystyrene beads containing the fluorophore Nile Red were conjugated to GRDS peptide sequence, which serves as a ligand for integrins. Cells expressing LARG-mTurq-SspB, iLID-CaaX, and Venus-KRasCT were incubated with the beads to facilitate binding of beads to integrins on the cell surface. When furrow formation was induced, surface-bound beads near the front of the daughter cells rapidly moved to the newly-forming furrow (Figure 2B and Movie 2) further confirming that the plasma membrane flows forwards the furrow. Together with the KRasCT and GL-GPI data, the bead movement demonstrates that the plasma membrane flows towards the back of the daughter cells during RhoA-induced furrow formation. These results are consistent with previously reported activation of RhoA at one edge of a cell inducing actomyosin contraction and rearward plasma membrane flow during cell migration²³.

Importantly, we found that if we keep photoactivating the bead-bound cells after furrow formation to induce the intercellular bridge-like structure, the beads remain at the back of the daughter cells and do not move through or along the elongated region itself (Figure 2C and Movie 2). This could suggest that the intercellular bridge is built by new membrane incorporation and not because of membrane flow.

We also observed membrane dynamics during spontaneous cell division in cells transfected with mCh-H2B and Venus-KRasCT. Cells were synchronized by incubating in serum-free media for 4 hours and imaged every 5 min for a period of 18 hours. The plasma membrane accumulated in the furrow of dividing cells (Movie 2). This demonstrates that membrane flow occurs in native cleavage furrow formation and again confirms that our optogenetic model is effective at inducing naturally occurring phenomena. The intercellular bridges we observed in native cytokinesis were considerably shorter than the elongated structures induced with optogenetic stimulation, possibly because bridge elongation happens very rapidly along with the final cut. It is also possible that optical RhoA activation allows us to exaggerate the length of this structure by constantly activating RhoA to move the daughter cells apart. Nevertheless, the ability of our tool to induce and maintain a structure closely resembling the intercellular bridge allows us to study the possible membrane mechanisms behind native intercellular bridge formation.

Membrane tension decreases in the furrow and in the backs of the daughter cells

Previous studies have shown that vesicular trafficking is needed for furrow formation^{11, 15, 16}, but the precise location, the source of the vesicles, and their exact role in mediating membrane remodeling in furrow formation has not been clear. The optogenetic method used here, which allowed us to induce furrow ingression similar to the one in native cytokinesis, helped us directly observe the spatial dynamics of endocytosis and exocytosis during furrow formation, which has not been possible with traditional methods of protein knockdowns and imaging fixed cells.

One parameter known to influence endocytosis is plasma membrane tension, with low tension regions exhibiting higher endocytosis rates and high tension regions favoring exocytosis^{36, 37}. The accumulation of plasma membrane in the cleavage furrow led us to hypothesize that tension is lowered in the furrow. We used formin-binding protein 17 (FBP17), which binds to areas of low membrane tension³⁸ to visualize tension changes in a cell during furrow and bridge formation. Cells were transfected with LARG-mTurq-SspB, iLID-CaaX, mCh-KRasCT, and Venus-FBP17. Upon initiation of photoactivation, FBP17 localized to the cleavage furrow. As the furrow elongated, FBP17 was enhanced at the plasma membrane in the back of the daughter cells (Figure 3A,B,S2A and Movie 3), suggesting that the membrane tension is lower there.

Figure 3. Plasma membrane tension changes during cleavage furrow formation.

(A) Dynamics of low membrane tension sensor FBP17 during RhoA-mediated furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Venus-FBP17 (green), and mCh-KRasCT (red). White arrows denote regions of FBP17 accumulation. Data is representative of 7 out of 8 cells. See also Movie 3.

(B) Intensity profile of Venus-FBP17 distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 14). Gray region is SEM.

(C) Caveolin dynamics during RhoA-mediated furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT (green), and mApple-Caveolin1 (red). Data is representative of 14 out of 15 cells. See also Movie 3.

(D) Intensity profile of mApple-Caveolin1 distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 7). Gray region is SEM.

Rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec. Fig. 2SA shows the profile line used for the intensity analysis.

Caveolin is another protein that can be used to visualize membrane tension changes, since it localizes to membrane invaginations and undergoes internalization during vesicle fission under low tension conditions³⁹. We transfected cells with LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT, and mApple-caveolin1 and optically induced furrow formation. Caveolin exhibited some plasma membrane localization in the basal state. Upon photoactivation, it localized to the cleavage furrow (Figure 3C,D,S2A and Movie 3). Additionally, caveolin-containing vesicles appeared in the cell body near the furrow, suggesting that endocytosis occurs in that region. Together with the above results, these findings suggest that the flow of plasma membrane to the cleavage furrow results in decreased membrane tension in the furrow and the back of the daughter cells. Consequently, this induces endocytosis in these areas, consistent with previous evidence for lowered membrane tension supporting endocytosis.

Endocytosis is localized to the cleavage furrow and the back of the daughter cells

Like caveolin, clathrin also mediates endocytosis, but through a distinct pathway. It is largely cytosolic in the basal state and translocates to the plasma membrane to mediate internalization of phosphorylated transmembrane receptors in clathrin-coated vesicles^{40, 41}. We reasoned that clathrin-mediated endocytosis, like caveolin-mediated endocytosis, would also be enhanced in low membrane tension regions. Cells were transfected with LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT, and mCh-Clathrin and optical activation was induced. Clathrin-coated vesicles rapidly accumulated at the cleavage furrow, with the highest concentration being close to the plasma membrane. As the furrow elongated, clathrin was enhanced in the back of the daughter cells (Figure 4A,B,S2A and Movie 4). This is consistent with endocytosis being enhanced in regions of low membrane tension and the results above with caveolin.

Figure 4. Endocytosis dynamics during cleavage furrow formation.

(A) Clathrin dynamics during RhoA-mediated furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT (green), and mCh-Clathrin (red). Data is representative of 10 out of 12 cells. See also Movie 4.

(B) Intensity profile of the mCh-Clathrin distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 7). Gray region is SEM.

(C) Localization of transferrin vesicles during RhoA-mediated furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, and Venus-KRasCT (green) and was incubated with 0.01 μg Tf-AlexaFluor594 (red) for 5 min prior to imaging. Data is representative of 5 out of 5 cells. See also Movie 4.

(D) Intensity profile of the Tf-Alexa-594 distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 9). Gray region is SEM.

Rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec. Fig. 2SA shows the profile line used for the intensity analysis.

We next used fluorescent transferrin to visualize the uptake of the transferrin receptor. Transferrin binding to its receptor triggers receptor internalization, making fluorescent transferrin a convenient tool for visualizing endocytosis^{42, 43}. Cells transfected with LARG-mTurq-SspB, iLID-CaaX, and Venus-KRasCT were incubated with 0.01 μg of AlexaFluor594-tagged transferrin for 5 min prior to photoactivation. As the cleavage furrow was optically induced, vesicles containing transferrin receptor localized near the furrow (Figure 4C,D,S2A and Movie 4). This result provides another piece of evidence that endocytosis is localized at the furrow region, and it could be responsible for membrane remodeling at the back of the daughter cells during cytokinesis.

The presence of localized endocytosis in these regions during furrow formation raises the question of where these endocytic vesicles are trafficked. We first examined the distribution of Rab11, which is associated with recycling endosomes and has previously been shown to localize to the cleavage furrow^{11, 15, 16}. Cells were transfected with LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT, and mPlum-Rab11. Upon cleavage furrow formation, Rab11 localized near the furrow, suggesting the presence of recycling endosomes there. Upon furrow elongation, Rab11 localized to the backs of the daughter cells but was not present inside the bridge-like structure (Figure 5A,E,S2A and Movie 5). In addition to being consistent with previous studies which have shown that proteins associated with vesicle trafficking are present in the cleavage furrow^{6, 15, 17}, the finding that Rab11 localizes to the same areas of the dividing cell as the endocytosis markers suggests that the endocytosed vesicles may be transported to recycling endosomes. Cells expressing a dominant negative Rab11 mutant appeared to contract in the middle and exhibit some membrane flow to the middle, but were unable to form a complete cleavage furrow (Figure S1C and Movie 5). This shows that Rab11 activity is a requirement for furrow formation and not merely a result of it. Together, these findings suggest that, while cortical contraction alone is sufficient for minor ingression, trafficking to recycling endosomes is required for complete furrow formation.

Figure 5. Role of recycling and exocytosis during cleavage furrow formation.

(A) Rab11 localization during cleavage furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT (green), and mPlum-Rab11 (red). Data is representative of 7 out of 7 cells. See also Movie 5.

(B) Arf6 localization during cleavage furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, YFP-Arf6 (green), and mCh-KRasCT (red). Data is representative of 8 out of 9 cells. See also Movie 6.

(C) Exo70 localization during cleavage furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Exo70-GFP (green), and mCh-KRasCT (red). Data is representative of 15 out of 16 cells. See also Movie 7.

(D) Effect of Exo70 inhibition on furrow formation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT (green) and mCh-H2B (red) and was incubated with 40 μM of endosidin-2 for 1 hr prior to imaging. Data is representative of 6 out of 6 cells. See also Movie 7.

(E) Intensity profile of the mPlum-Rab11 distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 8). Gray region is SEM.

(F) Intensity profile of the YFP-ARF6 distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 8). Gray region is SEM.

(G) Intensity profile of the Exo70-GFP distribution in the cell after furrow formation. Black line is the mean intensity profile (n = 9). Gray region is SEM.

Rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec. Fig. 2SA shows the profile line used for the intensity analysis.

Proteins regulating exocytosis are essential for cleavage furrow formation

To better understand where endocytic vesicles are trafficked during furrow formation, we visualized the distribution of Arf6, a protein that is involved in recycling and exocytosis^{44, 45}. Although it has been reported to localize to the cleavage furrow and intercellular bridge^{16, 46–48}, its precise dynamics during furrow formation, as well as its role in intercellular bridge formation, have not been clear. We therefore sought to visualize Arf6 during furrow formation and elongation to better understand the dynamics of exocytosis during these processes. We transfected cells with LARG-mTurq-SspB, iLID-CaaX, mCh-KRasCT, and YFP-Arf6, and photoactivated RhoA at the midpoint to induce furrow formation. As anticipated for a protein partially associated with the plasma membrane, Arf6 localized to the cleavage furrow, consistent with membrane flow to the furrow. Additionally, it also localized in the elongated bridge-like structure, suggesting that Arf6-mediated exocytosis could potentially occur in the intercellular bridge during its formation (Figure 5B,F,S2A and Movie 6). Cells expressing a dominant negative mutant of Arf6 were unable to form the furrow (Figure S1D and Movie 6), showing that Arf6 presence in the bridge is not merely correlative, but that its activity is essential for furrow formation. Consistent with such an essential requirement, cells expressing a constitutively active Arf6 mutant were likewise unable to form a complete furrow (Figure S1E and Movie 6), further suggesting that the dynamic switching between active and inactive forms of Arf6 is required.

Exocyst complex-mediated exocytosis has been previously implicated as playing a role in cleavage furrow formation¹⁸. Exo70, a subunit of the exocyst, mediates the final steps of exocytosis by tethering vesicles to the plasma membrane via its interaction with PI(4,5)P₂^{49, 50}. We aimed to visualize the dynamic distribution of Exo70 during furrow formation and elongation to gain further evidence on the localization of exocytosis. We transfected cells with LARG-mTurq-SspB, iLID-CaaX, mCh-KRasCT, and Exo70-GFP. Exo70 was found at the plasma membrane. As we optically induced furrow formation, Exo70 localized to the cleavage furrow and remained there as the furrow elongated (Figure 5C,G,S2A and Movie 7). The localization of Exo70 at the bridge-like structure similar to the localization of Arf6 suggests that exocytosis is a potential mechanism behind intercellular bridge formation.

We then sought to determine whether Exo70 activity is required for cleavage furrow formation. Cells were incubated for 1 hour with 40 μM endosidin2, a small molecule that interacts with Exo70 and inhibits exocyst-mediated exocytosis⁵¹. In 6 out of 6 cells treated with endosidin2 and subsequently photoactivated, the cell displayed some contraction and membrane flow towards the middle, but complete furrow formation did not occur (Figure 5D and Movie 7). This result suggests that exocyst-mediated exocytosis is a requirement for furrow formation. Overall, these results suggest that for the furrow and potentially the intercellular bridge to form, exocytosis of vesicles is required in the middle of the dividing cell.

Upregulation of endocytosis and exocytosis leads to increased elongation in the furrow region

Since we found that inhibiting vesicular trafficking inhibits cleavage furrow formation and elongation, we reasoned that enhancing trafficking may enhance these processes. To test this hypothesis, we introduced fluorescent transferrin to cells while optically photoactivating the cell. Unlike the above-mentioned experiment involving incubation with transferrin, where the majority of transferrin endocytosis was complete by the time that furrow formation was induced (Figure 4C), the present experiment overactivated the endocytosis machinery specifically during furrow elongation. Cells were transfected with LARG-mTurq-SspB, iLID-CaaX, and Venus-KRasCT. The cell was photoactivated in the middle and 0.01 μg of Transferrin-AlexaFluor594 was added once the cleavage furrow formed. Photoactivation was maintained at the backs of the daughter cells in order to induce furrow elongation. Whereas in previous experiments prolonged activation had little effect on the length of the bridge-like structure (Figure 4C), in this case the length increased significantly compared to control cells, and the daughter cells moved farther apart (Figure 6A,D,E, and Movie 8). This enhanced elongation suggests that the amount of endocytosis directly influences the ability of the furrow to elongate, with increased endocytosis resulting in longer bridge-like structures. Uptake of fluorescent transferrin can be observed at the back of the daughter cells (Figure 6A and Movie 8), consistent with above findings that endocytosis and trafficking machinery localizes to the back of the daughter cells.

Figure 6. Effect of endocytosis and exocytosis upregulation on cleavage furrow elongation.

(A) Effect of endocytosis upregulation on furrow elongation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, and Venus-KRasCT (green). 0.01 μg of Tf-AlexaFluor594 was added to the cells at t=4:00. See also Movie 8.

(B) Effect of exocytosis upregulation through cytohesin-2 expression on furrow elongation. Cell is expressing LARG-mCh-SspB (red), iLID-CaaX, and Cytohesin2-Venus (green). See also Movie 8.

(C) Effect of exocytosis upregulation through septin-7 expression on furrow elongation. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Septin7-YFP (green) and mCh-KRasCT (red). See also Movie 9.

(D) Furrow elongation resulting from endocytosis and exocytosis upregulation. Length of the bridge-like structure in cells in A–C was compared to that of control cells expressing LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT and mCh-H2B.

(E) Box plot of the average final length analysis at t=11 min from cells analyzed in D. *p<0.05, paired t test.

Rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec.

The finding that increased endocytosis specifically during furrow elongation results in increased length of the bridge-like structure, along with our above finding that plasma membrane does not flow to the middle of the structure but rather accumulates at the back of the daughter cells (Figure 2C), suggests that the membrane required for elongation is trafficked from elsewhere in the cell. This supports a possible model for intercellular bridge building in which trafficking of previously-endocytosed plasma membrane lipids to the bridge and their exocytosis in the bridge contribute to bridge formation and elongation. It therefore follows that increasing exocytosis would also result in increased bridge length. Using our system as a model for bridge formation, we tested this hypothesis by overexpressing cytohesin-2, a GEF for Arf6, and optically inducing the bridge-like structure. Higher levels of activation of Arf6 in the structure where Arf6 is concentrated would be expected to increase exocytosis and enhance the length. Cells were transfected with LARG-mCh-SspB, iLID-CaaX, and Cytohesin2-Venus. As in the case of transferrin, prolonged photoactivation of cytohesin-expressing cells resulted in significantly longer bridge-like structures than in control cells (Figure 6B,D,E and Movie 8). This suggests that intercellular bridge formation may be mediated by localized exocytosis.

These results could help answer important questions regarding membrane dynamics during intercellular bridge building. Some studies have suggested that vesicle exocytosis in the bridge is required for abscission⁴⁶. However, the source of the vesicles, their dynamic distribution during various stages of bridge formation, and their precise role in building the intercellular bridge and the daughter cells prior to abscission have remained unclear. Our optogenetic approach allowed us to identify a potential model in which intercellular bridge elongation occurs through trafficking of membrane lipids from the nascent daughter cells to the newly-forming bridge, and the extent of trafficking influences the bridge length. These results also suggest that localized endocytosis at the back of the daughter cells and localized exocytosis at the bridge may work in a coordinated manner to promote bridge elongation.

Septin expression increases elongation of the furrow region

We then used our optogenetic tool to further understand the role of exocytosis in intercellular bridge building. To test our hypothesis that exocytosis specifically localized to the intercellular bridge contributes to its elongation, we expressed septin-7 in cells and examined the elongation after optical activation of RhoA. Septins are a family of proteins involved in intercellular bridge building by forming ring structures around actin filaments in the bridge. Septin-7 rings have been found in the midbody. Septin-7 depletion leads to cytokinesis defects, and septins can also function as cytokinesis markers^{32, 52, 53}. Recent studies have also shown that septins interact with exocytosis machinery, recruiting the exocyst complex to the intercellular bridge^{32, 54, 55} and can be involved in mediating vesicle exocytosis⁵⁵. We examined whether expressing septin-7 in cells would result in longer bridge-like structures. Cells were transfected with LARG-mTurq-SspB, iLID-CaaX, mCh-KRasCT, and septin7-YFP. Prolonged photoactivation resulted in the formation of narrow structures over twice the length of those in control cells (Figure 6C,D,E and Movie 9). The narrowness is consistent with a role of septin-7 in maintaining the three-dimensional structure of the bridge. These results suggest that septin-7 could promote exocytosis in the intercellular bridge, leading to its elongation. Additionally, these results suggest that RhoA activation is sufficient for recruiting septin7, a protein known to localize to native intercellular bridges, to the elongating furrow region, where it likely facilitates exocytosis through recruitment of the exocyst complex. Thus, the bridge-like structure we observe is not only morphologically, but also in some ways structurally similar to native intercellular bridges. This further validates our optogenetic tool as a model for studying the mechanisms behind intercellular bridge formation.

FRAP shows vesicle endocytosis at the furrow, trafficking to the elongating furrow, and exocytosis

To observe the appearance of vesicles and their trafficking patterns during cleavage furrow formation, we used fluorescence recovery after photobleaching (FRAP). FRAP experiments were performed on cells expressing LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT, and mApple-caveolin. mApple-caveolin was first photobleached at the midpoint of a cell (Figure 7A). RhoA was then photoactivated in the same region. 3D projections of the cells were obtained before and after photobleaching using z-stacking to observe recovery over the entire cell instead of a single plane as furrow formation was optically induced. This enabled us to better observe the dynamics of caveolin-containing vesicles throughout the cell during furrow formation. Optical activation in the photobleached region resulted in ingression and furrow formation. mApple-caveolin gradually accumulated in the photobleached region at the site of photoactivation (Figure 7A and Movie 10). This suggests that plasma membrane-associated caveolin flows to the furrow along with the membrane lipids, consistent with plasma membrane flow induced by RhoA activation and consequent actomyosin contractility. Additionally, caveolin-containing vesicles appeared at the furrow (Figure 7A, white arrows, and Movie 10) showing that endocytosis predominantly occurs in the furrow region where membrane tension is lowered. This confirms data for membrane flow and endocytosis at the furrow.

Figure 7. Dynamics of directional trafficking during furrow formation and elongation examined through FRAP.

(A) Visualization of localized endocytosis during furrow formation and elongation through FRAP. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Venus-KRasCT (green), and mApple-Caveolin1 (red). mCh-Cavoelin1 was photobleached (red rectangle) at t=0:14. Images are maximum intensity projections from 5-plane z-stacks. The middle of the cell is shown magnified in the top panel. White arrows denote location of vesicles. Data is representative of 8 out of 8 cells. See also Movie 10.

(B) Visualization of directional vesicle trafficking to the intercellular bridge-like structure through FRAP. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, Cytohesin2-Venus, and mCh-GL-GPI (red). mCh-GL-GPI was photobleached (red rectangle) at t=10:45. The photobleached region is shown magnified in the top panel. Brightness and contrast are increased in magnified images to better visualize fluorescence recovery. White arrows denote areas of fluorescence recovery at the plasma membrane. Data is representative of 17 out of 18 cells. See also Movie 10.

Yellow rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec.

The ability to generate long furrows resembling intercellular bridges by expressing cytohesin-2 in cells enabled us to observe vesicular trafficking in these regions, which in turn allowed us to further build a possible model for intercellular bridge formation. We performed FRAP experiments where fluorescent vesicles containing GL-GPI, which has previously been used to visualize exocytosis⁵⁶, were photobleached in select areas of the elongated furrow. We transfected cells with LARG-mTurq-SspB, iLID-CaaX, Cytohesin2-Venus, and mCh-GL-GPI and photoactivated the cell to induce furrow formation and elongation. GL-GPI was seen in vesicles which localize to the elongated structure (Figure 7B and Movie 10). While continuously photoactivating the backs of the daughter cells, thereby inducing further elongation, we photobleached mCh-GL-GPI in the middle with a pulse of 594 nm laser light. As the furrow continued to elongate, GL-GPI-containing vesicles from adjacent non-photobleached regions of the cytosol were seen moving into the bleached region (Figure 7B and Movie 10). Importantly, the recovery at the plasma membrane occurred in spots within the photobleached region (Figure 7B, arrows - magnified view), rather than starting at the edges of the interface of photobleached and unbleached region and moving inward. This suggests that the fluorescence recovery is not due to lipid diffusion, but rather is caused by vesicles fusing with the plasma membrane in the bleached region. These findings support a possible model in which intercellular bridge elongation is achieved by trafficking of endocytosed vesicles to the bridge from the backs of the daughter cells, and their subsequent exocytosis inside the bridge contributes to bridge elongation.

Two different localized exocytosis mechanisms are involved in membrane remodeling during furrow formation and elongation

Having observed vesicle movement and fusion in the bridge-like structure, we asked whether this was the only region of the cell where exocytosis is localized. In particular, since we have shown above that plasma membrane lipids flow rearward from the front of the daughter cells towards the furrow, we wondered how the plasma membrane at the front of the daughter cells is replenished and whether exocytosis could be involved. TIRF microscopy is ideal for visualizing events such as exocytosis which occur at or near the plasma membrane, around 100 nm above the substrate^{57, 58}. Here we used TIRF to visualize exocytosis events while optogenetically inducing furrow formation and elongation.

VAMP2 is a SNARE protein involved in membrane fusion^{7, 59}, one of the last steps of exocytosis. VAMP2 remains at the plasma membrane for a period of time after fusion takes place^60–62. It is therefore an attractive candidate for visualizing exocytic events and their localization using TIRF. We used VAMP2 fused to a pH sensitive protein, pHuji, the fluorescence of which exhibits more than 20-fold increase from pH 5.5 to 7.5. Therefore, it has low fluorescence in intracellular vesicles, which are acidic, and high fluorescence at the plasma membrane^{63, 64}. We transfected cells with LARG-mTurq-SspB, iLID-CaaX, YFP-clathrin, VAMP2-pHuji, and septin7 to induce furrow elongation. The cells were photoactivated to induce furrow formation and elongation, and the plasma membrane was imaged using TIRF. VAMP2-pHuji intensity at the membrane increased with furrow formation and elongation (Figure 8A and Movie 11). To quantify the VAMP2-pHuji intensity and distribution, we used a custom algorithm to generate the cell outline at each time point and segment the cell into equal height sections. The pHuji intensities in the sections distal to the middle corresponding to the front of the daughter cells were averaged to obtain the intensity at the front, and these values were compared to the intensity in the middle section of the cell (Figures 8B,E,S2B and S2C). We found that VAMP2 is recruited predominately to the front of the daughter cells and not to the middle, suggesting that there is localized VAMP2-mediated exocytosis at the front of the daughter cells.

Figure 8. TIRF microscopy shows exocytosis localization at the front of the daughter cells.

(A) VAMP2 localizes to the front of the daughter cells. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, septin7, VAMP2-pHuji (black). See also Movie 11.

(B) Fractional change in the localized intensity of VAMP2 over time in the front of the cell and in the middle of the cell. Error envelope is SEM.

(C) Exocytosis and secretion at the front of the daughter cells. Cell is expressing LARG-mTurq-SspB, iLID-CaaX, septin7, and NPY-pHtomato (black). Red squares show the exocytosis events detected with our algorithm. See also Movie 11.

(D) Localization of NPY secretion events. The number of NPY secretion events in the cell middle was subtracted from the number of NPY secretion events at the front regions for each time point. Error envelope is SEM. Correlation coefficient (R = 0.8896) implies strong positive correlation.

(E) Comparison of final VAMP2 intensity at the front of the daughter cells and in the cell middle. Intensity at t = 10 min was divided by intensity at t = 0 min to obtain fractional change in intensity. Dotted lines depict paired data. The fractional change is significantly greater at the front of the daughter cells than in the middle (*** p < 0.001,paired t-test).

(F) Localization of NPY exocytosis events. Number of events at the cell middle was subtracted from the number of events at the front regions at t = 0 min and t = 10 min. Dotted lines depict paired data. The difference in the number of events is significantly greater at t = 10 min than at t = 0 min (***p < 0.001, paired t-test).

Yellow rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec.

To visualize more clearly where individual exocytosis events take place, we imaged neuropeptide Y (NPY), a marker for the last step of exocytosis, secretion of molecules into the extracellular media^{65, 66}. Unlike VAMP2 and other exocytosis markers which remain at the plasma membrane after vesicle fusion, NPY is secreted and diffuses rapidly through the extracellular media^{63, 65, 67}. Thus, whereas VAMP2 allows us to quantify exocytosis polarization, NPY allows us to quantify the distribution of independent exocytosis events. We used NPY fused to pHtomato, another pH-sensitive protein that increases its fluorescence around pH 7. We transfected cells with LARG-mTurq-SspB, iLID-CaaX, septin7, YFP-clathrin, and NPY-pHtomato and photoactivated the cell to induce furrow formation and elongation while acquiring TIRF images. By expressing septin7 we were able to induce long bridge-like structures. In the basal state, there is negligible NPY-pHtomato fluorescence at the plasma membrane. After inducing furrow formation and elongation, punctate regions of NPY-pHtomato appear at the plasma membrane in brief flashes that correspond to individual exocytosis events. The NPY regions then rapidly disappear, consistent with NPY being secreted into the extracellular media where it rapidly diffuses (Figure 8C and Movie 11). Using intensity and size thresholding along with the segmentation process described above, we found that more independent exocytosis events occur at the front compared to the middle as the furrow elongates (Figures 8D,F and S2D), consistent with the VAMP2 results.

Together, these results suggest that two different localized exocytosis mechanisms are involved in the plasma membrane remodeling during furrow formation. The first mechanism, which involves Arf6, septin, and exocyst-mediated exocytosis to promote elongation in the furrow region, could explain how the intercellular bridge is formed and sustained. The second mechanism relies on SNARE proteins such as VAMP2 to deliver membrane to the front of the newly-forming daughter cells. This process could be used to maintain the shape of the daughter cells by replenishing the plasma membrane lipids which flow from the front towards the furrow. It could also help the daughter cells move apart by supplying new membrane to the advancing front.

SRRF-stream imaging shows that newly-incorporated membrane comes from vesicle pools

Having demonstrated that exocytosis is enhanced during furrow formation and VAMP2-mediated exocytosis is localized to the front regions of the nascent daughter cells, we hypothesized that exocytosis could be used to incorporate plasma membrane lipids at the front of the daughter cells to replenish the plasma membrane that flows from the front to the furrow. If this is the case, it remains unclear whether the incorporated lipids are coming from vesicles or from elsewhere in the cell. To answer these questions, we performed a set of experiments where we stained the lipids of vesicle membranes and visualized their movement with SRRF-stream TIRF microscopy. SRRF-stream is a super-resolution algorithm that uses multiple low-exposure images of the same field to analyze radial intensity gradients and come up with a point of maximal radial symmetry, corresponding to the location of the fluorescent protein. This method allowed us to obtain images with resolution below 150 nm in real time^{68, 69}.

The amphiphilic styryl dye FM4–64 is a red fluorescent marker that fluoresces only in hydrophobic environments like membranes and binds to the outer leaflet of the membrane bilayer^{70, 71}. We transfected cells with LARG-mTurq-SspB, iLID-CaaX, and YFP-clathrin and incubated with 4 μM of FM4–64 for at least 30 min to stain the plasma membrane lipids. FM4–64 undergoes internalization after 10 minutes, appearing in vesicles inside the cell. After 30 minutes of incubation to ensure complete internalization, we washed the cells to remove any remaining FM4–64 from the plasma membrane⁷². We then photoactivated the cell to induce furrow formation and performed TIRF and SRRF-stream imaging to look for new membrane incorporation (Figure 9A).

Figure 9. SRRF- stream imaging shows membrane incorporation at the front of the daughter cells.

(A) Schematic representation of the FM4–64 protocol used in the membrane incorporation assay (see Material and Methods).

(B) Cell is expressing LARG-mTurq-SspB, iLID-CaaX, and YFP-Clathrin (green). Cells were incubated with 4μM FM4–64 (black) for at least 30 min and washed with HBSS containing glucose to remove dye from the plasma membrane. See also Movie 12.

Yellow rectangles represent the area of photoactivation. Scale bar is 10 μm. Time is in min:sec.

The vesicles containing FM4–64 were initially localized in the middle of the cell near the furrow, and with time they began to move towards the front of the daughter cells. Eventually, a line of FM4–64 fluorescence appeared at the front edge of the cell (Figure 9B, red arrows, and Movie 12), indicating that vesicles are incorporated into the plasma membrane at the front. This is in contrast to YFP-clathrin, which appeared in vesicles but did not form an outline at the front edge. These result shows that exocytosis at the front of the daughter cells incorporates new lipids into the plasma membrane, and the source of these lipids are intracellular vesicles containing previously-endocytosed membrane, rather than lipids synthesized de novo. This membrane incorporation was not obvious with confocal or TIRF imaging alone. Thus, the super-resolution power of SRRF-stream imaging combined with the fine spatiotemporal control over furrow formation provided by optogenetics allowed us to visualize localized membrane incorporation during this process in real time.

Model for membrane remodeling during cleavage furrow and intercellular bridge formation

The optogenetic approach used here, combined with super-resolution live cell imaging, allowed us to identify a mechanism for membrane remodeling during cleavage furrow formation. Our results show that during cleavage furrow formation, plasma membrane flows to the furrow due to RhoA-mediated actomyosin contractility in that region. As a result, plasma membrane tension is lowered at the furrow, which induces localized endocytosis there. The endocytosed vesicles are recycled and undergo exocytosis in the furrow region, which could contribute to the formation of the intercellular bridge. Additionally, there is exocytosis of plasma membrane-containing vesicles at the front of the nascent daughter cells, which could serve to replace the membrane lipids that flow towards the furrow (Figure 10).

Figure 10. Proposed model for the role of localized endocytosis and exocytosis in cleavage furrow and intercellular bridge formation.

RhoA activation in the middle of the cell causes ingression and cleavage furrow formation due to actomyosin contraction and plasma membrane flow to the site of RhoA activation. This leads to decreased membrane tension in the furrow and increased endocytosis. As the intercellular bridge forms, endocytosis is localized to the backs of the nascent daughter cells. The endocytosed vesicles undergo exocytosis in two distinct regions using two different mechanisms. In the first mechanism, the vesicles are trafficked to recycling endosomes containing Rab11 and then to the intercellular bridge where they undergo Arf6- and Exo70-mediated exocytosis. This supplies membrane lipids needed for bridge elongation. In the second mechanism, the vesicles undergo exocytosis at the front of the daughter cells, mediated by SNARE proteins like VAMP2. This mechanism serves to replace the plasma membrane lipids at the front edges, allowing the daughter cells to maintain their shape as they move apart and elongate the bridge.

Our optogenetic tool also allowed us to identify potential mechanisms for membrane remodeling during intercellular bridge formation. Morphologically, the bridge-like structure we observed with prolonged RhoA activation in the middle of the cell closely resembles native intercellular bridges^30–32. It also contains septin, which is known to localize to intercellular bridges and stabilize their structure^{32, 52, 54, 55, 73}. Because of these structural and morphological similarities, we used this bridge-like structure as a model for native intercellular bridges. We propose a model in which, as the bridge forms and elongates, low membrane tension and endocytosis are maintained at the backs of the nascent daughter cells. The membrane-containing endocytosed vesicles are then trafficked to the bridge and undergo localized exocytosis there to promote bridge elongation. Septin works to recruit exocytosis machinery to the bridge. As the bridge elongates and the daughter cells move apart, vesicles containing previously-endocytosed membrane lipids continue to fuse to the plasma membrane at the front of the daughter cells. This could help the daughter cells maintain their shape as they move apart and continue to replenish membrane lipids which flow towards the furrow (Figure 10).

It is worth noting that although our tool is able to induce the cleavage furrow and a structure closely resembling native intercellular bridges, the cells never undergo abscission. This is likely because by directly activating RhoA in the cell middle, we are forgoing the steps which occur prior to RhoA activation in native cytokinesis, such as formation of the central spindle and recruitment of kinases required for abscission. These processes, along with the increased RhoA activation, are likely needed to recruit abscission machinery, such the ESCRT complex, to the furrow to facilitate the final cut⁷⁴. However, as validated here, the optogenetically controlled events here are a precise model system for furrow formation which is an important but not fully understood event during cytokinesis. It has allowed the membrane-based mechanisms underlying furrow and bridge formation to be identified.

Spatially localized endocytosis, vesicle trafficking, and exocytosis are known to support processes such as maintaining apical-basal polarity in epithelial cells⁷⁵ and delivering synaptic receptors in neurons⁷⁶. Amoeboid migration has been shown to utilize rearward plasma membrane flow to the cell rear^{23, 77}, accompanied by localized endocytosis at the back, directional trafficking of vesicles from back to front, and localized exocytosis at the front²³. It is remarkable that directional vesicle trafficking along with coordinated and spatially localized endocytosis and exocytosis has been recruited evolutionarily to support a variety of distinct and important cellular processes.

MATERIALS AND METHODS

DNA Constructs

The following constructs were purchased from Addgene: mApple-Caveolin (#54872, Michael Davidson), mCh-Clathrin (#55019, Michael Davidson), H2B-mCh (#20972, Robert Benezra), mPlum-Rab11 (#55998, Michael Davidson), Arf6-YFP (#11389, Joel Swanson), Arf6(Q67L)-CFP (#11387, Joel Swanson), EGFP-Exo70 (#53761, Channing Der), NPY-mHTomato (#83501, Sebastian Barg) and VAMP2-pHuji (#105289, Justin Taraska).

Venus-FBP17 was synthesized by ligating FBP17 (BglII and EcoR1), Venus (HindIII and BglII), and pcDNA3.1 (HindIII and EcoR1). Arf6(T27N)-Venus was synthesized by ligating Venus (BamH1 and XbaI), Arf6(T27N) (EcoRI and BglII), and pcDNA3.1 (EcoR1 and XbaI). Cytohesin2-Venus was made by ligating cytohesin2 (EcoR1 and NotI), Venus (NotI and XbaI), and pcDNA3.1 (EcoR1 and XbaI). LARG-mCh-SspB, LARG-mTurq-SspB, Venus-KRasCT, and mCh-KRasCT were synthesized as previously described (O’Neill et al., 2018). iLID-CaaX was synthesized as previously described (O’Neill et al., 2016).

GFP-Rab11(S25N) was kindly provided by Jennifer Stow (University of Queensland, Australia). Septin7-YFP was kindly provided by Manos Mavrakis (Institut Fresnel, France). mCh-GL-GPI was kindly provided by Valeria Caiolfa (Spanish National Center for Cardiovascular Research, Spain).

Cell Culture and Transfections

RAW 264.7 cells (ATCC, Manassas, VA) were cultured in high glucose DMEM (Sigma Aldrich, St. Louis, MO) containing 10% dialyzed FBS (Atlanta Biologicals) and 1% penicillin-streptomycin. Cells were grown at 37° C and 5% CO₂. Cells passaged between 3 and 15 times were used for experiments.

Cells were transiently transfected using Amaxa Nucleofector 2b electroporator following previously-described protocols⁷⁸. Around 2–3 million cells and 0.2–3 μg of the appropriate DNA were used for each transfection. Transfected cells were plated on glass-bottom dishes and imaged 4–10 hours after electroporation.

Reagents

Nile Red functionalized fluorescent beads containing carboxyl groups (Spherotech, Lake Forest, IL) were attached to integrin-binding peptide GRGDS (Sigma Aldrich, St. Louis, MO) as previously described²³. Cells were incubated with bead-peptide solution for 1 hour before imaging, followed by washing with media to remove unbound beads.

Stock solutions of endosidin2 were prepared in DMSO and stored at −20°C until use. AlexaFluor594-Transferrin (Molecular Probes, Eugene, OR) was dissolved in water and stored at 4°C until use. Stock solutions of FM4–64 (Thermo Fisher Scientific, Waltham, MA) were prepared in water and stored at −20°C until use. All dilutions were performed in Hanks’ Balanced Salt Solution (HBSS; Corning, Manassas, VA) containing 1 g/L of glucose.

Membrane Incorporation Assay

The cells were loaded with 4 μM FM4–64 to stain the plasma membrane lipids and incubated for 30 minutes at 37° C and 5% CO₂ to ensure internalization of FM4–64 in vesicles. Then the cells were washed three times with HBSS, to remove any remaining FM4–64 from the plasma membrane⁷². The cells were kept in HBSS during TIRF and SRRF-stream imaging.

Imaging

Imaging was performed on an Andor Revolution spinning disk confocal system consisting of a Leica DMI6000B microscope, Yokogawa CSU-X1 spinning disk, Andor iXon camera, Andor FRAPPA unit for photoactivation and photobleaching of select regions, and Andor iQ2 software. For optical activation of optogenetic constructs, 445 nm laser at 145 nW power was scanned across the designated region at 0.9 ms/μm². For imaging, 515 nm and 594 nm solid state lasers were used along with Venus 528/20 nm and mCherry 628/20 nm emission filters. Images were obtained with a Leica 63x, 1.4 NA oil immersion objective.

TIRF and SRRF-stream imaging was performed on an Andor Dragonfly spinning disk confocal system using a iXon Lite EMCCD camera and a Nikon Eclipse Ti2 inverted microscope. Localized photoactivation was performed with Andor Mosaic 3 DMD array using the CoolLED pE-4000 LED illumination system. A 460 nm wavelength LED was used for photoactivation. 514 nm and 561 nm solid state lasers were used for imaging. Images were obtained with a Nikon APO TIRF 60× 1.49 NA oil immersion objective. TIRF penetration was set to 100 nm. For SRRF-stream imaging, 30 frames per image were acquired.

All imaging was performed at 37°C and 5% CO₂.

Quantitative Analysis

We quantified the distribution of various markers across the space of the cell after furrow formation. At a time point after furrow formation was complete, we used Andor iQ to draw a linear region of interest from the front of one daughter cell to the front of the other daughter cell, with the furrow in the middle (Figure S2A). Intensity at each pixel along the line was measured with iQ and graphed as intensity versus position. Curves for at least 7 cells were averaged in OriginPro to generate the mean intensity and SEM.

To quantify the length of the furrow region in cells with cytohesin, transferrin, septin, and H2B, the edges of the daughter cells were tracked using the MTrackJ plugin in ImageJ. The distance between the two points was calculated to determine the length at each time point.

For analysis of VAMP2 and NPY data, an outline of the cell membrane at each time point was obtained from brightfield images in ImageJ using a custom algorithm. The outline was superimposed on the TIRF image and converted into an ROI selection. A bounding rectangle was created around the ROI and split into five segments of equal height. Calculating the intersection between the bounding rectangle and the ROI returned the cell outline split into five equal height sections. To quantify VAMP2 polarization, the fluorescence intensities in sections 1 and 5 were averaged to obtain VAMP2-pHuji intensity at the cell front, and the fluorescence intensity in section 3 was used to obtain VAMP2 intensity in the cell middle. The intensity at each time point was normalized to the initial intensity. To quantify the number of NPY fusion events at each time point, an intensity threshold equal to two standard deviations above the average pixel intensity was applied to the TIRF image at each time point. Then, the average size of an incorporation event was determined to act as a secondary filter against general noise. Each of the five regions of the cell was analyzed to find the number of events above the size and intensity thresholds. The average number of events in the cell middle (section 3) was subtracted from the average number of events in sections 1 and 5 (cell edges) for each time point.