Exo/Endocytosis

 

Fertilization exocytosis basics

 

To understand this work, a little background information on fertilization is in order.  In organisms ranging from certain algae to mammals, fertilization triggers an increase in intracellular free calcium.  This increase often travels through the egg as a wave, and it triggers (among other things) a wave of exoctyosis of secretory compartments referred to as "cortical granules".  As the cortical granules fuse with the plasma membrane, they release their contents outside the cell (Exo/endocytosis figure 1).  The biology behind this event is extremely important:  the cortical granule contents modify the extracellular matrix, which makes it impermeable to other sperm.  This is one of the basic mechanisms which ensure that only a single sperm fuses with the egg.  When the wave of exocytosis fails, multiple sperm fuse with the egg, leading to failed development of the embryo.  In addition to its importance for successful reproduction, the wave of exocytosis triggered by fertilization is also an important model for understanding regulated exocytosis in general, as it shares many features in common with exocytosis in other cell types such as neurons and neuroendocrine cells.

 

Exo/endocytosis figure 1--A schematic diagram representing the process of exocytosis during fertilization. The squiggly thing on the left represents the sperm, the large round thing in the middle represents the egg, and the round things inside the egg represent the cortical granules.  The cortical granules that are white have not undergone exocytosis; those that are beige or darker brown are in the process of undergoing exocytosis and have fused with the egg plasma membrane (and are supposed to be releasing their contents) and those that are empty have already released their contents.  Note that the pattern of exocytosis is wave-like, with the wave starting at the point of sperm-egg contact.

 

 

 

 

 

 

The discovery of actin "coats" on exocytosing cortical granules

 

We initially started studying the process of exocytosis in Xenopus eggs using fluorescent dextran as a marker for exocytosis.  The trick is simple:  the cell is bathed in fluorescent dextran and imaged at a level just below the plasma membrane (inside the cell), providing a "top" or "facing" view, or from the side, providing a cross sectional view.  When a cortical granule fuses with the plasma membrane and begins to release its contents, the dextran outside the cell rushes into the space formerly occupied by the cortical granule contents (Exo/endocytosis figure 2; see also movies below).

Exo/endocytosis figure 2--A schematic diagram representing the process of dextran uptake during cortical granule exocytosis.  In this example, the cell is bathed in green fluorescent dextran, and the cortical granules are shown as red discs.  The leftmost part of the figure shows the situation before cortical granule exocytosis--none of the granules have fused with the plasma membrane and released their contents.  The middle part of the figure shows the situation as exocytosis is occurring--two of the cortical granules have released their contents and taken up the green dextran, one has not, and one is in the process of releasing its contents.  The rightmost part of the figure shows the situation after all of the cortical granules have undergone exocytosis--they have all released their contents and taken up the external dextran.

 

In an early study, we found that the dextran-containing compartments only remain for a short period of time and then disappear.  We also found that the formation of the dextran containing compartments was perturbed by a drug called latrunculin, which inhibits actin assembly (Bement et al., 2000. J. Exp. Zoo. 286:767-775).  Anna Sokac decided to explore the role of actin filaments in this process and started by imaging actin filaments while triggering exocytosis of the cortical granules using photolysis of caged IP3 (this is a fancy way of saying that she used UV light to stimulate a calcium increase in the cells).  Anna found that following exocytosis, cortical granules become surrounded by actin filaments (Exo/endocytosis movie 1).  Anna went on to demonstrate that the actin forms a "coat" that completely encloses the exocytosing cortical granule, and that the coat then compresses (ie closes inward), which explains why the dextran disappears (Exo/endocytosis movie 2).  Coat compression is apparently needed to retrieve the cortical granule membranes after exocyosis and my also help release of cortical granule contents by squeezing them out of the cortical granule.

In collaboration with Jack Taunton's lab (UCSF) Anna showed that the actin coat forms due to activation of Cdc42 (like Rho, a G-protein) on exocytosing cortical granules (Exo/endocytosis movie 3) and the resultant recruitment of the actin regulator, N-WASP.  This study, which was published in Nature Cell Biology (http://www.nature.com/ncb/journal/v5/n8/full/ ncb1025.html) also revealed that the coats formed as a result of new actin assembly rather than recruitment of pre-existing actin filaments to the exocytosing cortical granules.  

Exo/Endocytosis movie 1--Confocal imaging of actin filaments (green) and fluorescent dextran (red) during cortical granule exocytosis in Xenopus eggs, with the focal plane of imaging just below the plasma membrane (top view).  As cortical granules undergo exocytosis, the red dextran rushes into them, appearing as red discs.  Subsequently, actin filaments accumulate around the exocytosing cortical granule, forming an actin "coat".  The coats close inward over time.  See also http://www.nature.com/ncb/journal/v5/n8/full/ ncb1025.html for more details.

 

 

 

Exo/endocytosis movie 2--4D Confocal imaging of actin filaments (green) and fluorescent dextran (red) during cortical granule exocytosis in Xenopus eggs.  The experiment is similar to that in movie 1, except that multiple focal planes are imaged at each time point, which allows the process to be viewed in 3 dimensions over time.  In this movie, the process of actin coating is viewed as a side cut-away view, which makes it clear that the coat completely surrounds the exocytosing cortical granule.  See also http://www.nature.com/ncb/journal/v5/n8/full/ ncb1025.html for more details.

 

 

Exo/endocytosis movie 3--Confocal imaging of actin filaments (green) and fluorescent dextran (red) during cortical granule exocytosis in Xenopus eggs following inhibition of Cdc42.  Exocytosis occurs, but the actin coat fails to form and the cortical granule membranes simply collapse into the plasma membrane.  Similar results are obtained when actin is directly perturbed with latrunculin.  See also http://www.nature.com/ncb/journal/v5/n8/full/ ncb1025.html for more details.

 

 

What causes compression of the actin coat?  Part of the answer came from a project Anna conducted in parallel to the one described above.  That is, she characterized the expression patterns of several different unconventional myosins during Xenopus oogenesis.  Most of the myosins showed fairly pedestrian expression patters, mimicking that of actin.  However, myosin-1C was quite different--its mRNA was upregulated during meiotic maturation, the stage when the developing oocyte takes the final steps needed for fertilization.  Part of this analysis is shown below in Exo/Endocytosis figure 3.  The figure shows that the mRNA encoding myosin-1C, but not the mRNAs encoding either actin or myosin-2, is polyadenylated during meiotic maturation.  This modification is often associated with an increase in expression and, indeed, Anna found that the protein level of myosin-1C increased sharply during maturation.

 

Exo/endocytosis figure 3--Upregulation of myosin-1C during meiotic maturation.  The top, left part of the figure shows an RNA blot demonstrating that meiotic maturation (the transition from stage 6 oocyte [6] to egg [E]) is accompanied by an increase in polyadenylation of myosin-1C mRNA (Xl myo1C) but not the mRNAs encoding myosin-2 (Xlmyo2A) or actin.  The top, right part of the figure shows that the total mRNA concentration remains the same for all three mRNAs.  The bottom part of the figure is an immunoblot that compares the protein levels of myosin-1C, myosin-2, and actin.  Note that only the myosin-1C shows an increase in abundance after meiotic maturation (compare the "6' sample to the "E" sample).  For more detail, see Sokac et al., Dev. Cell 2006. 11:629-640; http://www.sciencedirect.com/science/article/pii/S1534580706003959).

 

These results prompted Anna to undertake a detailed characterization of myosin-1C function in Xenopus eggs (see her paper in Developmental Cell from 2006:  http://www.sciencedirect.com/science/article/pii/S1534580706003959).  In collaboration with Cameron Gundersen's lab at UCLA, Anna showed that myosin-1C is recruited to exocytosing cortical granules and that disruption of this recruitment resulted in suppression of cortical granule content release.  When Anna analyzed the behavior of actin coats following disruption of myosin-1C recruitment, she found something quite remarkable:  the actin coats still form around exocytosing cortical granules, but they fail to remain tightly associated with the granules and instead the assembling actin polymerizes in all directions.  This results in the exocytosing cortical granule remaining trapped beneath the plasma membrane for far longer than usual (Exo/Endocytosis movie 4).  Anna hypothesized that the role of the myosin-1C is to link assembling actin to the cortical granule membrane so that the actin can compress the cortical granule as it assembles.  This hypothesis received support in another study (see below) while the results demonstrated that myosin-1C and the actin coat are needed to help retrieve the cortical granule membrane and, possibly, to promote expulsion of cortical granule contents from the exocytosing granules.  

 

Exo/endocytosis movie 4--Disruption of myosin-1c recruitment to exocytosing cortical granules results in loss of actin coat tethering to the cortical granule membranes.  This is a 4D movie, made the same way as movie 3 above except that myosin-1c has been experimentally prevented from undergoing recruitment to the exocytosing cortical granules.  Note that the actin filaments (green) of the actin coats only partially engulf the exocytosing cortical granules (red) and instead tend to shoot off into space (well, into the cytoplasm at any rate).  Note also that the cortical granule compartments are not retrieved but remain below the plasma membrane for a prolonged period of time.  See Sokac et al., Dev. Cell 2006. 11:629-640; http://www.sciencedirect.com/science/ article/pii/S1534580706003959 for more details. 

 

 

 

 

What triggers formation of the actin coats?  Based on several considerations, Anna and Bill propoosed the "compartment mixing hypothesis"  (see Exo/endocytosis figure 4 below and the paper in Molecular Biology of the Cell:  http://www.molbiolcell.org/content/17/4/1495.long).  The notion is that the combination of the material in or associated with the plasma membrane and the material in or associated with the cortical granule membrane can promote actin assembly when the two are combined, but not when they are separate.  Thus, membrane fusion itself has the potential to initiate the signaling events that lead to actin coat formation.

 

Exo/endocytosis figure 4--A schematic diagram representing the compartment mixing hypothesis.  The blue line represents the plasma membrane (and any proteins, lipids, or other factors associated with it), the red circle represents the cortical granule membrane (and any proteins, lipids, or other factors associated with it) and the light purple lines represent actin filaments.  The plasma membrane and the cortical granule membrane are each proposed to have a component that forms part of a signal transduction pathway leading to actin assembly, but as long as the two compartments are separate, the complete pathway is not formed.  Following fusion of the the cortical granule with the plasma membrane, the components of the plasma membrane and cortical granule membrane mix, as represented by red dots (cortical granule membrane components) moving in the plasma membrane and blue dots (plasma membrane components) moving into the cortical granule membrane.  This mixing unifies the previously separate parts of the hypothetical signal transduction pathway that leads to actin assembly.  Thus, actin begins to assemble first at the site of cortical granule-plasma membrane fusion and then, as mixing continues, spreads farther and farther down the surface of the cortical granule, creating the coat.

 

 

 

The compartment mixing hypothesis was tested by Elsie Yu.  Elsie found that upon calcium elevation, the trigger for exocytosis, the lipid, diacylglycerol (aka DAG) is generated in the plasma membrane.  When the cortical granule fuses with the plasma membrane, the diacylglcerol from the plasma membrane mixes with the cortical granule membrane and does so before the actin coat begins to assemble, as expected if the DAG is part of a signal transduction pathway leading to actin assembly.  This is shown below in Exo/endocytosis figure 5; see also Elsie's paper in Nature Cell Biology (http://www.nature.com/ncb/journal/v9/n2/full/ncb1527.html).

 

Exo/endocytosis figure 5--Diacylglycerol (DAG) encorporates into the cortical granule membrane shortly after exocytosis and before actin coating.  Top view of dextran (blue, top row) which provides a marker for exocytosis as described above, DAG is red in the triple label (top row) and white in the single label (middle row), and actin is green in the triple label (top row) and white in the single label (bottom row).  The time is in seconds, with minus indicating the number of seconds before exocytosis, zero indicating when exocytosis was first apparent and the other numbers indicating the time after exocytosis.

 

 

How do we know that the DAG is generated in the plasma membrane and then incorporates into the cortical granule membranes as a result of exocytosis?  After all, is it not possible that the DAG is made separately in the cortical granule membrane independently of fusion?  To distinguish between these possibilities, Elsie compared DAG dynamics in control cells and cells in which cortical granule exocytosis was blocked.  Some of the results of these experiments are shown below in Exo/endocytosis figure 6, and they demonstrate that following calcium elevation, DAG increases in the plasma membrane but does not appear in cortical granule membranes until after exocytosis.  When exocytosis in inhibited, DAG still increases in the plasma membrane, but never finds its way into the cortical granule membranes.  

Exo/endocytosis figure 6--Diacylglycerol (DAG) levels increase in the plasma membrane and DAG incorporates into cortical granule membranes only if exocytosis occurs.  This view is different than the one in figure 5--here we are looking at cross section of the plasma membrane and the underlying cortex of the cell.  Thus, the outside of the cell is toward the top of each figure, where the dextran concentration is very high and the inside is toward the bottom of each figure, where there is initially no dextran.   Another difference is that in this figure, the time in seconds refers to when calcium was elevated rather than when exocytosis occurs.  Note that in the control, following calcium elevation, DAG levels in the plasma membrane increase (compare -2s to 14s) and this precedes exocytosis.  Note also that after exocytosis (indicated by the arrowhead and the incorporation of dextran into a cortical granule) the DAG begins to accumulate in the cortical granule membrane.  In contrast, in the cell where exocytosis has been inhibited, DAG still increases in the plasma membrane, but is never seen on the cell interior.  Note that the last time point in the experimental sample is longer than that in the control (31 versus 28 s) and it really doesn't matter how long you wait--no exocytosis, no DAG in the cortical granule membranes.  Further, if exocytosis is inhibited, the actin coats do not form (not shown; but see http://www.nature.com/ncb/journal/v9/n2/full/ncb1527.html).

 

 

How do we know that the DAG is an important signal for actin coating?  Because experimental elevation of DAG increases the extent of coating, and experimental reduction of DAG reduces the extent of coating (http://www.nature.com/ncb/journal/v9/n2/full/ncb1527.html).  The next question, then, is what is the target of DAG?  That is, how is incorporation of DAG linked to actin coating?  Part of this answer is:  the serine threonine kinase, protein kinase C-beta (PKCb).  PKCb is activated by interaction with calcium and DAG which collectively target it to the plasma membrane.  Elsie found that PKCb is targeted to and remains on cortical granule membranes that undergo exocytosis prior to actin coating (Exo/endocytosis figure 7).  Further, elevation of PKCb levels increases actin coating while inhibition of PKCb reduces actin coating (http://www.nature.com/ncb/journal/v9/n2/full/ncb1527.html).  

Exo/endocytosis figure 7--PKCb is recruited to exocytosing cortical granules prior to actin coating.  Top view of dextran (blue, top row) which provides a marker for exocytosis as described above, PKCb is green in the triple label (top row) and white in the single label (middle row), and actin is red in the triple label (top row) and white in the single label (bottom row).  The time is in seconds, with minus indicating the number of seconds before exocytosis, zero indicating when exocytosis was first apparent and the other numbers indicating the time after exocytosis.

 

 

 

Thus, the plasma membrane provides DAG (which leads to PKCb activation) during compartment mixing.  But what does the cortical granule membrane provide?  At least one of the key ingredients is the small GTPase, Cdc42.  Elsie figured this out based on Anna's previous demonstration that actin coating is Cdc42 dependent (see above and below) and based on the finding that inactive Cdc42 localizes to cortical granules prior to the induction of exocytosis (Exo/endocytosis figure 8).

 

Exo/endocytosis figure 8--Cdc42 on cortical granules.   This is another cross sectional view, like figure 6.  The top part of the figure shows a cell prior to the induction of exocytosis.  On the far left is a triple label, with the plasma membrane (PM) labeled in blue, the cortical granule (CG) in red, and Cdc42 (which is not yet active) in green.   "Out" indicates where the outside of the cell is and "In".  To the right of the triple label, each of the individual labels is shown separately in white.  Note that the inactive Cdc42 is concentrated on the cortical granules, but not on the plasma membrane.   The bottom part of the figure shows the process of Cdc42 activation following the induction of exocytosis.  The view is also cross sectional, and dextran is red and active Cdc42 is green.  It is apparent that prior to exocytosis there is relatively little active Cdc42 on the cortical granule but that the level increases at increasing times after exocytosis (0s marks the onset of exocytosis).  See also http://www.nature.com/ncb/journal/v9/n2/full/ncb1527.html.

 

 

 

These results support the compartment mixing hypothesis and place some detailed "meat" on its bones:  prior to exocytosis, DAG is in the plasma membrane, and Cdc42 is on the cortical granules.  Thus, at least two key components in the signal transduction pathway leading to actin coating are kept separate from each other.  Upon exocytosis, the cortical granule membrane and the plasma membrane become unified, allowing the DAG and the Cdc42 to come together.  The DAG (with help from calcium) recruits PKCb which triggers Cdc42 activation.  In a sense, the whole process is like epoxy which is a sort of glue that is stored as two separate components that, when combined, form a glue but are nongluey as long as kept separate.  One of the more interesting features of this process is that there are many examples of cellular events wherein exocytosis is associated with actin assembly, including cytokinesis and single cell wound repair.

 

Now let us return to a question posed above, namely, "What makes the actin coats compress?" Based on Anna's work with myosin-1C, we had hypothesized that the compression is powered by actin assembly with myosin-1C functioning to link assembling actin to the cortical granule membrane.  In a paper published in Molecular Biology of the Cell (see http://www.molbiolcell.org/content/18/10/4096/F1.expansion.html) Elsie tested this hypothesis by triggering exocytosis in the presence of a very low dose of cytochalasin, a manipulation that prevented new actin filament growth without disrupting the preexisting actin network.  This produced a striking phenotype:  the cortical granules underwent exocytosis but failed to assemble actin coats, and instead remained trapped beneath the plasma membrane for as long as she cared to image (Exo/endocytosis figure 9)

 

Exo/endocytosis figure 9--Inhibition of acitn assembly prevents compression of cortical granule compression after exocytosis.  Top view, red is dextran, green is actin.  In controls, exocytosing cortical granules (arrows) become coated with actin (arrowheads), undergo compression, and disappear.  In samples treated with low concentrations of cytochalasin, to limit actin assembly, exocytosing cortical granules (arrows) can recruit some actin arrow heads), but the actin coats fail to compress, such that the cortical granule membranes, filled with dextran, never disappear.

 

These results are consistent with the hypothesis that actin assembly itself provides the force needed to compress the cortical granules after exocytosis.  Elsie also sought to characterize potential roles for other myosins, besides myosin-1C.  She showed that both myosin-2 and myosin-1e are recruited to cortical granules, but with differing temporal profiles:  myosin-1e is immediately recruited to all cortical granules upon calcium elevation but only retained on those cortical granules that undergo exocytosis, while myosin-2 is recruited somewhat later (Exo/endocytosis figure 10; see also see http://www.molbiolcell.org/content/18/10/4096/F1.expansion.html).

Exo/endocytosis figure 10--Recruitment of myosin-1e and myosin-2 to exocytosing cortical granules.  The top part of the figure shows the recruitment of myosin-1e (green) to cortical granules that have (indicated by arrows and the presence of red dextran) and have not (indicated by arrowheads) undergone exocytosis.  However, the myosin is only retained on those cortical granules that have undergone exocytosis nhibition of acitn assembly prevents compression of cortical granule compression after exocytosis.  The bottom part of the figure shows the recruitment of myosin-2 to an exocytosing cortical granule (revealed by the blue dextran in the top row).  In contrast to myosins-1c and 1e, myosin-2 (red in top row; white in middle row) is recruited relatively late, after the actin coat  (green in top row; white in bottom row) has already formed.

 

 

 

What do these myosins do during exocytosis and actin coat formation and compression?  Suppression of myosin-1e recruitment to cortical granule membranes results in a peculiar phenotype in which the actin coat initially forms, but then becomes concentrated on the basal portion of the cortical granule membrane, as if this myosin functions as a distributor that ensures that the assembling actin is spread uniformally over the granule surface (Exo/endocytosis figure 11).  Suppression of myosin-2 motor activity slows, but does not prevent the late stages of coat compression, and results in thicker coats (Exo/endocytosis figure 12), as if this myosin may be involved in disassembling the coat.

 

Exo/endocytosis figure 11--Inhibition of myosin-1e results in accumulation of actin on the basal side of the cortical granule membrane.  The top part of the figure shows a side view of actin coating in controls, this time using a marker for the membrane (green in top row, white in bottom row) in conjunction with an actin probe (red in the top row, white in the middle row) to allow visualization of the cortical granule membrane.  Note that the actin more-or-less uniformly coats the cortical granule membrane.  The bottom part of the figure shows what happens when myosin-1e function is compromised:  the actin, instead of uniformly coating the cortical granule membrane, ends up concentrated on the basal (bottom) side of the membrane.  (The time intervals are the same in both the control and the myosin-1e inhibited samples and 0s marks the beginning of exocytosis.)

 

 

 

Exo/endocytosis figure 12--Inhibition of myosin-2 results in slowing of the late stages of coat compression and disassembly.  Just actin is shown (white) with the controls in the top row and the myosin-2 inhibited samples in the bottom row.  At times when compression is essentially complete and the coat is disassembling in controls, compression is not yet complete and the actin coat appears far thicker than normal in myosin-2 inhibited sample.  (Time intervals are the same for controls and myosin-2 inhibited sample; ) 0s marks the onset of coat formation; see see http://www.molbiolcell.org/content/18/10/4096/F1.expansion.html for more details.)

 

 

Whither the exo/endocytosis project?  We are in the process of trying to figure out if compartment mixing functions to send signals controlling actin assembly (and other cytoskeletal events) during processes such as cytokinesis and wound healing.