Cell Cycle Conference October 2022

Video Transcript

• 00:00 - 00:13:  So, welcome to the last Cell Cycle Club meeting of 2022.
• 00:13 - 00:17:  We'd like to thank you all for attending these meetings, and we'll continue those.
• 00:17 - 00:21:  There's going to be three more meetings in the next year.
• 00:21 - 00:24:  One in March or April, which is going to be a virtual one.
• 00:24 - 00:27:  The one in May, June is going to be a hybrid one.
• 00:27 - 00:31:  And then the one in October in 2023 is going to be a virtual one again.
• 00:31 - 00:35:  So, we have a very exciting program today.
• 00:35 - 00:39:  We'd like to thank everybody for submitting all the abstracts.
• 00:39 - 00:41:  We've got many excellent abstracts.
• 00:41 - 00:44:  It was very hard to select three of them.
• 00:44 - 00:48:  But we also have 14 exciting talks now in the breakout sessions where you can choose
• 00:48 - 00:53:  one of seven in two breakout sessions during the breakout.
• 00:53 - 00:58:  As already been said, if you have a question, please put it in the chat.
• 00:58 - 01:02:  And then at the end of the talk, we will select you in the order they're being put in.
• 01:02 - 01:03:  We will call your name.
• 01:03 - 01:05:  They will unmute you and you can ask your own question.
• 01:05 - 01:08:  If you don't want to ask your own question, just put it in anonymously or just put in
• 01:08 - 01:13:  the chat that the chair can read out your question.
• 01:13 - 01:19:  So, also a reminder that we select the talks that are largely unpublished work.
• 01:19 - 01:22:  So, please treat it as personal communication.
• 01:22 - 01:25:  It's like a big lab meeting where we're having here, and we're hoping that it's going to be
• 01:25 - 01:29:  very interactive, as interactive it can be virtually.
• 01:29 - 01:35:  And then, before we go to our first speaker, if you hear the fire alarm, it's either your
• 01:35 - 01:38:  fire alarm or the speaker's fire alarm.
• 01:38 - 01:43:  And just assume it's not the speaker's fire alarm, so then get out of your house.
• 01:43 - 01:48:  So, then we can start with the first talk, which is from Ryan.
• 01:48 - 01:54:  If you can start sharing your screen, Ryan Murray from North Carolina from Mike's lab,
• 01:54 - 02:02:  Mike Emanuel's lab, and he's going to talk about a degradation, a G2M specific degradation
• 02:02 - 02:05:  program that depends on PLK1.
• 02:05 - 02:09:  So, Mike, we can see your, yeah, first slide.
• 02:09 - 02:12:  So, the stage is yours.
• 02:12 - 02:13:  Okay.
• 02:13 - 02:18:  Well, good morning or afternoon or evening, everyone, wherever you're watching.
• 02:18 - 02:21:  I'm excited to be able to talk today, and so, of course, thank you to the organizers
• 02:21 - 02:24:  for giving me this opportunity.
• 02:24 - 02:29:  As Rob mentioned, my name is Ryan and I work in Mike Emanuel's lab at UNC Chapel Hill.
• 02:29 - 02:32:  And so, I'm going to talk to you today about some of my thesis work that's really starting
• 02:32 - 02:38:  to uncover the role that coded protein degradation plays as cells pass the G2M transition point,
• 02:38 - 02:44:  and particularly the dependency that this degradation has on the mitotic kinase PLK1.
• 02:44 - 02:48:  And so, of course, this group needs no introduction to the events of the cell cycle, but just
• 02:48 - 02:53:  to kind of set the stage, we know that two of the most fundamental features of the cell
• 02:53 - 02:59:  cycle are first going to be DNA replication, and this, of course, occurs in S phase, and
• 02:59 - 03:03:  then ultimately to segregate that genetic information into two identical daughter cells
• 03:03 - 03:06:  during mitosis.
• 03:06 - 03:10:  And so, we know that a hallmark feature of the cell cycle is going to be the irreversible
• 03:10 - 03:14:  commitment to completing a round of cell cycle once the cells enter S phase, and this is
• 03:14 - 03:18:  to prevent genome amplification.
• 03:18 - 03:23:  And so, one of the major features that the cell uses to accomplish this is coded protein
• 03:23 - 03:24:  degradation.
• 03:24 - 03:29:  And then the cell cycle of this system is regulated by a network known as the ubiquitin
• 03:29 - 03:31:  proteasome system.
• 03:31 - 03:34:  And so, I'm sure many people are familiar with this, but this involves an enzymatic
• 03:34 - 03:36:  cascade.
• 03:36 - 03:42:  And so, in the system, a molecule of ubiquitin is first activated by an E1 enzyme.
• 03:42 - 03:47:  This molecule of ubiquitin can get passed on to an E2 ubiquitin-conjugating enzyme,
• 03:47 - 03:52:  and then the E2 enzyme, in collaboration with an E3 ubiquitin ligase, which will recognize
• 03:52 - 03:58:  the target protein, can attach that molecule of ubiquitin to the target protein, and we
• 03:58 - 04:03:  can have multiple rounds of this process to generate what are known as polyubiquitin chains.
• 04:03 - 04:07:  And although this can lead to a diverse network of signaling outcomes, today we'll think about
• 04:07 - 04:11:  the most canonical outcome, which is protein degradation.
• 04:11 - 04:17:  And so, again, we know that protein degradation is particularly useful at key transition points
• 04:17 - 04:19:  between the cell cycle phases.
• 04:19 - 04:26:  And so, perhaps most appreciated is a large-scale wave of degradation that is mediated by an
• 04:26 - 04:31:  E3 ubiquitin ligase, known as the anaphase-promoting complex, or cyclosome, or the APC.
• 04:31 - 04:37:  So, the APC is going to drive the degradation of a multitude of substrates, which will complete
• 04:37 - 04:42:  the process of mitosis and drive cells into the following G1 phase.
• 04:42 - 04:48:  The APC continues this degradation across G1 to prevent entry into the next S phase.
• 04:48 - 04:54:  Eventually, the APC becomes inactivated, and we pass the wall of degradation onto a second
• 04:54 - 04:57:  E3 ligase, known as the SCF complex.
• 04:57 - 05:01:  This is named because of its components, so this uses a cullin-1 backbone with an adaptor
• 05:01 - 05:06:  protein known as SCF1, and then the complex uses an interchangeable substrate receptor
• 05:06 - 05:08:  known as an F-box protein.
• 05:08 - 05:13:  And here, I'm just showing an example of an F-box protein known as SCF2, which degrades
• 05:13 - 05:18:  the CDK inhibitors P21 and P27 to promote S phase entry.
• 05:18 - 05:22:  And so, comparatively to these two transition points, we feel that there is less known about
• 05:22 - 05:29:  how globally protein regulation is controlling the cell cycle at the G2M transition point.
• 05:29 - 05:33:  And of course, we do know that there is some degradation happening that is reported in
• 05:33 - 05:34:  the literature.
• 05:34 - 05:40:  For example, with the SCF complex, with the second substrate receptor known as beta-TRCP,
• 05:40 - 05:45:  it can drive the degradation of proteins such as EMI1 and WE1.
• 05:45 - 05:50:  And so, particularly, we became interested in the subset of proteins whose degradation
• 05:50 - 05:55:  turns out to be dependent on the activity of PLK1.
• 05:55 - 06:02:  And the reason that these proteins are dependent on PLK1 is because PLK1 is going to create
• 06:02 - 06:07:  a phosphodegron within these substrates to allow the binding of the E3 ubiquitin ligase.
• 06:07 - 06:11:  So, there are four examples here: EMI1 and WE1, which I already mentioned, as well as
• 06:11 - 06:13:  the proteins claspin and BORA.
• 06:13 - 06:18:  And so, how this typically works is an upstream kinase will first phosphorylate the substrates.
• 06:18 - 06:23:  And this is because PLK1 typically has to bind prime to a pre-phosphorylated substrate.
• 06:23 - 06:28:  And so, that primary kinase can add the first phosphate, which allows PLK1 to bind to a
• 06:28 - 06:31:  domain known as the polo box domain.
• 06:32 - 06:38:  And then, PLK1 can add a second phosphate, which allows beta-TRCP in coordination with
• 06:38 - 06:43:  the rest of the SCF complex to bind to the substrate, and promoting ubiquitination and
• 06:43 - 06:44:  degradation of these substrates.
• 06:44 - 06:53:  And so, really, the goal of my project was to understand, so we knew about this handful
• 06:53 - 06:58:  of substrates, but we really wanted to understand whether these substrates were kind of the
• 06:58 - 07:04:  bulk of what we knew about, or whether additional substrates existed.
• 07:04 - 07:07:  And so, to get at this question, we took a fairly simple approach.
• 07:07 - 07:14:  We took HCT116 cells, and we trapped them in mitosis using nocodazole, and then we
• 07:14 - 07:18:  co-treated these cells with DMSO as our negative control with two structurally
• 07:18 - 07:22:  distinct PLK1 inhibitors.
• 07:22 - 07:26:  We then collected whole cell extracts from these cells, and we used label-free DIA mass
• 07:26 - 07:29:  spectrometry to quantify the whole proteome.
• 07:29 - 07:37:  And so, we also had an asynchronous condition for comparison of the mitotic proteome independent
• 07:37 - 07:38:  of PLK1.
• 07:38 - 07:44:  Today, I will mostly talk about the dependency on how PLK1 shapes the mitotic proteome now.
• 07:44 - 07:51:  And so, also, just a quick note here, we did use two drugs here, or two small molecule
• 07:51 - 07:52:  inhibitors.
• 07:52 - 07:57:  If we plot the results against each other of the fold change of our first drug against
• 07:57 - 08:02:  the control, and then on the y-axis, we used drug number two against the control, we see
• 08:02 - 08:07:  that out of nearly 6,900 proteins that the data correlates pretty well with each other,
• 08:07 - 08:10:  suggesting that we have high confidence that these are on-target effects of inhibiting
• 08:10 - 08:16:  PLK1 and not off-target effects of the small molecule inhibitors.
• 08:16 - 08:20:  And so, now, looking at a volcano plot for just one of the two drugs, so here on the
• 08:20 - 08:25:  x-axis, we're now looking at the log-2 ratio of our PLK1-inhibited cells over our control
• 08:25 - 08:26:  mitotic cells.
• 08:26 - 08:30:  And on the y-axis, we have the negative log-10 of our p-value.
• 08:30 - 08:37:  So, anything to the right of this dashed line was higher in the cells that had PLK1 inhibited.
• 08:37 - 08:43:  And we can see that some of our top hits, we have WE1, BORA, Claspin, and EpiXO5, which
• 08:43 - 08:44:  is also EMI1.
• 08:45 - 08:50:  And so, these are all known substrates of the PLK1 beta-TRCP axis, as I mentioned, and
• 08:50 - 08:55:  so we felt pretty good about identifying our positive controls here.
• 08:55 - 08:59:  And so, if we use a p-value of 0.01 as our threshold for significance, we can see that
• 08:59 - 09:05:  we identified roughly 217 proteins that were higher in mitosis when we inhibited PLK1.
• 09:05 - 09:09:  And so, we became really interested in what these proteins were and assessing them as
• 09:09 - 09:16:  potentially novel proteasomal targets that are regulated in a PLK1-dependent manner.
• 09:16 - 09:20:  And so, when we look at the processes that these proteins are largely playing, we see
• 09:20 - 09:25:  processes that we would expect, such as regulation of the cell cycle process, cell division,
• 09:25 - 09:27:  and regulation of DNA replication.
• 09:27 - 09:33:  And so, again, based on the roles that we know PLK1 plays and normally regulates, these
• 09:33 - 09:36:  processes make sense and would be what we would expect to see.
• 09:37 - 09:41:  And so, I'm going to tell you just briefly two key findings that we found out of this
• 09:41 - 09:43:  dataset.
• 09:43 - 09:48:  And so, first, I want to talk about a group of proteins that we became interested in as
• 09:48 - 09:50:  a collective whole.
• 09:50 - 09:55:  And so, it's not exactly important what each one of these proteins does, but the reason
• 09:55 - 09:59:  that we were interested in this collection of proteins is because these are known substrates
• 09:59 - 10:05:  of another E3 ligase, which also works in the SCF complex, but this time the substrate
• 10:05 - 10:08:  receptor is a protein known as cyclin F.
• 10:08 - 10:14:  And so, it's been shown that the E2F activators 1, 2, and 3, as well as the atypical repressor
• 10:14 - 10:20:  7, 8, RM2, CB110, and some of the other proteins I showed on the previous slide are all known
• 10:20 - 10:26:  substrates of cyclin F. And we know that this process occurs as cells approach the G2M transition.
• 10:26 - 10:32:  However, up to this point, the dependency on PLK1 for their degradation was unknown.
• 10:32 - 10:37:  And so, again, this collection of proteins stuck out to us, and we wanted to validate
• 10:37 - 10:42:  whether this, in fact, was true out of our mass spec experiment.
• 10:42 - 10:48:  And so, we performed the same experiment and validated the results first by Western blot.
• 10:48 - 10:52:  And so, here, first, I'm just showing one substrate of cyclin F, which is CB110.
• 10:52 - 10:58:  And so, again, we can see first that in an untreated condition, we see CB110, and this
• 10:58 - 11:00:  abundance goes down when you put cells in the control.
• 11:01 - 11:04:  And we see that we can rescue it with both of our PLK1 inhibitors.
• 11:06 - 11:11:  We can also see this for essentially every cyclin F substrate we tested.
• 11:11 - 11:16:  So, the same pattern holds true for RM2, CPC6, C2F1, and SOVP.
• 11:17 - 11:21:  And then, here, we have Claspin, which is a known beta-TRCP substrate as a positive
• 11:21 - 11:23:  control for how this would look in this assay.
• 11:24 - 11:31:  And then, we also have cell cycle markers to show both the PLK1 levels were equal, as
• 11:31 - 11:35:  well as phosphorylated 3 in cyclin B, suggesting that these cells were equally arrested in
• 11:35 - 11:36:  mitosis.
• 11:37 - 11:41:  And this phospho-TCTP marker is going to be in a lot of my experiments here.
• 11:41 - 11:47:  PLK1 phosphorylates TCTP at this site, and so this is just a marker that we use to verify
• 11:47 - 11:49:  that we, in fact, inactivated PLK1.
• 11:49 - 11:54:  And so, of course, we're really excited that all these substrates validated in this
• 11:54 - 11:54:  assay.
• 11:55 - 12:00:  However, this, of course, doesn't necessarily mean that the changing of these proteins was
• 12:00 - 12:05:  dependent on cyclin F. It could be that PLK1 somehow independently regulated all of these
• 12:05 - 12:06:  proteins.
• 12:06 - 12:08:  And so, we wanted to address this question.
• 12:08 - 12:13:  And to do this, we utilized cyclin F knockout cells in HeLa cells that we have in the lab.
• 12:13 - 12:15:  And then, we repeated the same experiment.
• 12:16 - 12:18:  And so, here, again, we're looking at CB110.
• 12:18 - 12:22:  And if we're looking at the left four lanes, we can see that, essentially, this is the
• 12:22 - 12:28:  same in the HCT cells, where we first see a downregulation of CB110 levels, and then
• 12:28 - 12:31:  we see an increased abundance when we inhibit PLK1.
• 12:31 - 12:35:  However, if we no longer have cyclin F, we can see that there's no difference in the
• 12:35 - 12:39:  mitotic cells between our negative control and the cells that had PLK1 inhibited.
• 12:40 - 12:44:  And again, I'll just flash this up, but we can see that this is true for all the cyclin
• 12:44 - 12:46:  F substrates that we tested here as well.
• 12:47 - 12:54:  And again, we have our cell cycle controls.
• 12:55 - 13:00:  So, this really starts to kind of suggest a model where PLK1 sits atop the regulation
• 13:00 - 13:04:  of cyclin F, and particularly sitting atop cyclin F substrate degradation.
• 13:05 - 13:07:  And so, you can see here that we have a dashed line.
• 13:07 - 13:12:  So, we're still kind of working out the mechanisms to exactly how this regulation happens.
• 13:12 - 13:17:  Unlike beta-TRCP, whose substrates are typically phosphorylated at the substrate level to create
• 13:17 - 13:22:  a phosphodegron, we typically don't think in the cyclin F field as cyclin F substrates
• 13:22 - 13:25:  needing themselves to be phosphorylated.
• 13:25 - 13:30:  And so, we're really curious as to how PLK1 was regulating all of the cyclin F substrates.
• 13:32 - 13:36:  And so, what we were able to find, essentially, is that PLK1 activity can contribute to the
• 13:36 - 13:39:  abundance and stability of the ligase of cyclin F itself.
• 13:40 - 13:45:  And so, here we can see that across six different cell lines, if we perform an acute treatment
• 13:45 - 13:52:  for four hours with two molecules targeting PLK1, we can see that we lose cyclin F abundance
• 13:52 - 13:53:  in all six cell lines.
• 13:56 - 14:00:  And so, we were also interested in testing, of course, this could happen at the transcriptional
• 14:00 - 14:03:  level, or it could happen as a post-translation regulation mechanism.
• 14:04 - 14:09:  And so, to kind of get at this question, we performed a cycloheximide chase analysis
• 14:09 - 14:14:  to inhibit new protein translation in the presence of either DMSO as a control or the
• 14:14 - 14:16:  two PLK1 inhibitors.
• 14:17 - 14:22:  And so, we can see here by blot, we can also quantify the blot, that if we perform this
• 14:22 - 14:26:  analysis, we can see that cyclin F becomes a less stable protein when you do not have
• 14:26 - 14:28:  PLK1 activity.
• 14:29 - 14:33:  And so, this really is starting to suggest that this is a post-translational level of
• 14:33 - 14:34:  regulation.
• 14:34 - 14:41:  And so, we suspect that in response to PLK1, there is some sort of E3 ligase that is normally
• 14:41 - 14:47:  regulating cyclin F, that when PLK1 presumably phosphorylates cyclin F, now that degradation
• 14:47 - 14:48:  can occur.
• 14:49 - 14:55:  You stabilize cyclin F, allowing for its degradation of its substrates as cells approach G2M.
• 14:55 - 14:57:  And so, we're still working out some of the mechanism here.
• 14:58 - 15:07:  But what this really does show is kind of globally that PLK1 now is sitting atop a larger
• 15:07 - 15:12:  role of protein degradation at G2M that maybe we did not appreciate up to this point.
• 15:12 - 15:18:  So, we kind of already knew about this side of the equation, how PLK1 regulates beta-TRCP
• 15:18 - 15:18:  substrates.
• 15:19 - 15:24:  And now, on this side, we're adding the regulation of the substrates of a second E3
• 15:24 - 15:27:  ligase via a distinct mechanism of regulating the ligase.
• 15:28 - 15:29:  Rather than the substrates.
• 15:31 - 15:36:  And so, really thinking about this, you know, global role that PLK1 might play, we became
• 15:36 - 15:40:  interested in really, you know, validating additional substrates within this population
• 15:40 - 15:44:  of proteins that we identified by the proteomics.
• 15:44 - 15:48:  And so, particularly, we're looking at substrates that show a general pattern of both being
• 15:48 - 15:51:  low in mitosis compared to an asynchronous population.
• 15:52 - 15:56:  This suggests that they might normally be regulated by proteasomal degradation.
• 15:56 - 16:00:  And then we can see for many of the substrates that are validated, such as the beta-TRCP
• 16:00 - 16:05:  or selenium substrates, we can see that in the two drugs, compared to control mitotic
• 16:05 - 16:07:  cells, we see an upregulation.
• 16:08 - 16:12:  And so, we identified a handful of proteins that also kind of followed the same pattern.
• 16:13 - 16:16:  And today, briefly, I'll just show you one of our top hits that we're following up on,
• 16:16 - 16:20:  which is known as A-kinase-anchoring protein 2, or ACAP2.
• 16:20 - 16:26:  And so, for ACAP2 specifically, there's very, very little known about ACAP2.
• 16:26 - 16:28:  You will find hardly any papers on it.
• 16:29 - 16:34:  And so, it certainly has no reported functions in controlling cell cycle or cell proliferation.
• 16:35 - 16:40:  But more generally, as the name implies, what ACAP proteins do as a family, they are
• 16:40 - 16:48:  able to bind to protein kinase A, as well as then scaffold them to different subcellular
• 16:48 - 16:49:  localizations.
• 16:49 - 16:54:  So, here, I'm just depicting ACAP scaffolding it to the plasma membrane, but this can also
• 16:54 - 16:59:  occur at the mitochondria or various other subcellular localizations.
• 17:00 - 17:06:  And so, it serves kind of as a scaffolding unit to coordinate PKA signaling networks.
• 17:08 - 17:14:  And so, first, again, we know that a lot of proteins that control cellular proliferation
• 17:14 - 17:19:  are themselves regulated both at the transcript and protein levels across the cell cycle.
• 17:20 - 17:25:  And so, here, if we look at all the proteins we identified by mass spec, so, on the x-axis,
• 17:25 - 17:27:  each dot is an individual protein.
• 17:27 - 17:31:  And here, now, we're just looking at our mitotic cells versus an asynchronous population.
• 17:32 - 17:36:  And we can see, by and large, most of the proteome remains unchanged or only mildly
• 17:36 - 17:37:  changed.
• 17:37 - 17:41:  And we see that at the tails here, we see a subset of proteins that are upregulated in
• 17:41 - 17:44:  mitosis, as well as downregulated.
• 17:44 - 17:50:  And we can see known markers of cell cycle regulation, such as cyclin E and EMI1, that
• 17:50 - 17:52:  are downregulated in mitosis.
• 17:52 - 17:57:  And we can see that ACAP2 was one of our top hits as being down in mitotic-arrested cells.
• 17:57 - 18:01:  And so, this really suggests that this could be a novel cell cycle-regulated protein.
• 18:03 - 18:08:  And so, first, we just validated this by trapping cells in mitosis using nocodazole.
• 18:08 - 18:12:  We can see that, in fact, ACAP2 is strongly downregulated.
• 18:14 - 18:18:  We can also perform cell cycle analysis using a double-thymidine block and release to start
• 18:18 - 18:22:  them at the G1/S transition point and then release them back into the cell cycle.
• 18:22 - 18:27:  And we can see that around 10 to 12 hours after release from the thymidine block, we
• 18:27 - 18:30:  start to see a decrease in ACAP2 levels.
• 18:30 - 18:33:  This kind of mirrors what we start to see for Claspin here.
• 18:34 - 18:42:  And interestingly, this also starts to peak or is in collaboration with the peak of PLK1
• 18:42 - 18:48:  activity, as shown by this marker here, which is a marker of PLK1 activation loop, as well
• 18:48 - 18:50:  as the TCTP marker I mentioned.
• 18:52 - 18:55:  And again, we can also look at thalidomide block and release.
• 18:55 - 18:59:  And we can see, compared to an asynchronous sample, we see that we're already starting
• 18:59 - 19:00:  relatively low.
• 19:00 - 19:06:  ACAP2 is degraded for a little bit after release from mitosis and starts to come back as cells
• 19:06 - 19:08:  reach the next G1/S transition point.
• 19:08 - 19:13:  So, this really suggests that ACAP2 is, in fact, a cell cycle-regulated protein that
• 19:13 - 19:16:  gets degraded at some point in mitosis and peaks around G1/S.
• 19:18 - 19:22:  And so, if we go back to the volcano plot of our original proteomics, we can see that
• 19:22 - 19:27:  ACAP2, in fact, was one of our strongest hits here in terms of both statistical significance
• 19:27 - 19:30:  as well as the fold change in the PLK1-inhibited cells.
• 19:32 - 19:36:  And so, we can go back to our original assay, and we can validate that ACAP2 is, in fact,
• 19:36 - 19:39:  strongly upregulated in PLK1-inhibited cells.
• 19:40 - 19:46:  We can also use a complementary approach using siRNA knockdown of PLK1 to collaborate these
• 19:46 - 19:51:  findings, showing that ACAP2 is higher if we knock down PLK1.
• 19:51 - 19:57:  Further showing that PLK1 is able to control the degradation of ACAP2, we can perform co-expression
• 19:57 - 19:58:  assays on PLK1.
• 20:00 - 20:10:  Using exogenous proteins and 293T cells. Here, I'm expressing ACAP2 alone. ACAP2 co-expressed
• 20:10 - 20:16:  with either a wild-type version of PLK1 or a mutant version of PLK1 that is inactive
• 20:16 - 20:22:  catalytically. Again, we can see that ACAP2 goes down when expressing the wild-type PLK1,
• 20:22 - 20:30:  and this is dependent on the catalytic activity. Additionally, so I mentioned that PLK1 typically
• 20:30 - 20:35:  binds to pre-phosphorylated substrates by recognizing the phosphopeptide in the substrate.
• 20:36 - 20:40:  And so it's been shown that there are two residues within PLK1's polo box domain that,
• 20:40 - 20:46:  if you mutate them, should abrogate binding to the substrate. And so we can perform co-IP analysis
• 20:46 - 20:53:  between PLK1 and ACAP2 here, and we can see that PLK1 is able to strongly bind, sorry, PLK1
• 20:53 - 20:58:  wild-type strongly binds to ACAP2. However, if we mutate those residues in the substrate binding
• 20:58 - 21:05:  domain, we can see a strong reduction in the ability to bind to ACAP2. So altogether, this
• 21:05 - 21:12:  suggests that PLK1 is able to regulate the mitotic degradation of ACAP2. And so from the first half
• 21:12 - 21:19:  of my talk, this suggests that this could occur by either beta-TRCP or, presumably, cyclin F.
• 21:19 - 21:23:  And so we were interested in understanding the ligase that might be contributing to ACAP2
• 21:23 - 21:30:  degradation. And so to do this, we again used our nocodazole-arrested cells, and we knocked down
• 21:30 - 21:37:  either BioFly or Luciferase as a negative control, a Cullin-1, which should inhibit all of the Cullin-1
• 21:38 - 21:42:  ligases, so all the F-box proteins that can work with Cullin-1. And then we knocked down either
• 21:42 - 21:49:  beta-TRCP or cyclin F. And we can see, again, this strong down-regulation in mitotic cells,
• 21:49 - 21:54:  and we see that the degradation of ACAP2 was rescued if we knocked down either Cullin-1
• 21:54 - 21:59:  or beta-TRCP, but not cyclin F. So this suggests that this could potentially be a novel
• 21:59 - 22:07:  beta-TRCP substrate. We can also perform co-IP analysis with ACAP2 and beta-TRCP.
• 22:07 - 22:14:  So we can see that ACAP2 is able to bind by co-IP to both beta-TRCP1 and 2.
• 22:15 - 22:19:  And again, if we perform co-expression assays, we can see that this also drives
• 22:19 - 22:26:  the degradation of ACAP2, both by co-expression of beta-TRCP1 and beta-TRCP2.
• 22:26 - 22:35:  And so, finally, so beta-TRCP, the consensus motif here is a DSG, and then either two or three
• 22:35 - 22:40:  amino acids that can be anything, followed by another S. And the two serines here are typically
• 22:40 - 22:46:  phosphorylated. So ACAP2 and its sequence does not have any that fit perfectly into this
• 22:46 - 22:52:  beta-TRCP consensus motif, but it does have three DSG motifs, and there are numerous
• 22:52 - 22:58:  examples in the literature of beta-TRCP substrates that don't fit perfectly or conform
• 22:58 - 23:05:  perfectly to this motif. So we went ahead and mutated the three DSG motifs in ACAP2.
• 23:05 - 23:11:  to triple alanine mutations, and then we performed co-IP analysis again. And what we show is that,
• 23:11 - 23:18:  again, the wild type ACAP protein is able to bind to beta-T or Cp. And the first and third motifs
• 23:19 - 23:25:  the mutation did not affect binding. However, this second motif in the substrate, we can see
• 23:25 - 23:29:  that there's no longer binding, suggesting that this is probably the functional dead
• 23:29 - 23:35:  ground that beta-T or Cp uses to regulate ACAP2. And so we look at some conservation
• 23:35 - 23:40:  of this sequence. We can see across different species of ACAP2 that this DSG sequence is fairly
• 23:40 - 23:47:  conserved in the ACAP2 protein. And so currently, what we're at is that we're looking at a
• 23:47 - 23:52:  And so currently, what we're at is we're going to hope to generate stable cell lines expressing
• 23:52 - 24:00:  either wild type or ACAP2, DSG2 mutant. So we can generate a non-degradable mutant of this protein
• 24:00 - 24:05:  and assess the consequences that it might have on contributing to cellular proliferation.
• 24:06 - 24:10:  And so, again, we can hypothesize that there's something going on with PKA signaling here,
• 24:10 - 24:19:  although PKA does have some roles reported in proliferation, although this is mostly in
• 24:19 - 24:26:  meiotic cells in model systems. And in the mitotic cell cycle, there's less known about how PKA might
• 24:27 - 24:32:  contribute to cellular proliferation. So just wrapping up again, we are hoping to kind of
• 24:32 - 24:38:  keep elucidating the mechanism as to how PRK1 promotes NF abundance and stability. So
• 24:39 - 24:45:  hoping to find maybe where in select enough PRK1 phosphorylates it, as well as either an E3 or a
• 24:45 - 24:53:  DUB that might recognize and regulate select enough in response to PRK1 activity. We hope to continue
• 24:55 - 24:58:  characterizing this axis with this new cell cycle-related protein,
• 24:58 - 25:04:  ACAP2. And then finally, I think an interesting thing to think about. So PRK1 is commonly
• 25:04 - 25:09:  overexpressed in a multitude of different cancer types. And this has led to the efforts to develop
• 25:09 - 25:15:  small molecule inhibitors for clinical use in these cancer types. And so now that my data
• 25:15 - 25:22:  suggests or our data suggests that PRK1 is largely remodeling the proteome, it's interesting to think
• 25:23 - 25:28:  after treatment with these PRK1 inhibitors, what the proteome of these tumors might look like,
• 25:28 - 25:35:  and how understanding this might help us to better administer how these PRK1-targeted therapies are
• 25:35 - 25:41:  actually given in the clinic. And so with that, I'd like to thank everyone in the Emanuel Lab,
• 25:41 - 25:46:  Mike, for being a wonderful mentor and helping me with this project. I'd like to particularly
• 25:46 - 25:50:  thank Carolyn Shi, who is a wonderful undergraduate that I mentor in the lab,
• 25:51 - 25:56:  and who has helped with some of the experiments in this project. I'd like to thank Laura and
• 25:56 - 26:01:  Allie in the UNC Proteomics Core for helping with the initial proteomics experiment, as well
• 26:01 - 26:07:  as the analysis of the proteomics data. I'd like to thank my thesis committee and funding.
• 26:07 - 26:09:  And with that, I'd be happy to take anyone's questions.
• 26:12 - 26:17:  There's me clapping, right? Everybody else is clapping. You just can't hear them.
• 26:17 - 26:22:  Thank you very much. It was a great presentation. I'll start with a question.
• 26:23 - 26:30:  Did you try—if you overexpress cyclin F, do you rescue the sensitivity to PLK
• 26:32 - 26:35:  inhibitors of the proteins that are targeted with cyclin F?
• 26:37 - 26:42:  So, we actually—we haven't tried that, but we have talked about, like, cyclin F expression levels
• 26:42 - 26:48:  and whether that might contribute to the sensitivity to PLK1 inhibitors. And so,
• 26:48 - 26:53:  we have thought about maybe using the knockout cells we have and treating with PLK1 inhibitors.
• 26:55 - 27:00:  So, you might think, with all the changes that are happening, that it could affect the sensitivity.
• 27:00 - 27:04:  So, it's definitely an interesting question, just something that we haven't gotten to
• 27:04 - 27:05:  in the study yet.
• 27:06 - 27:13:  And are there cell lines that develop resistance to PLK1 inhibitors? And could
• 27:13 - 27:16:  you test whether cyclin F is overexpressed in those?
• 27:18 - 27:22:  I actually don't know that off the top of my head. It's a good question if there's resistance.
• 27:23 - 27:28:  I do know, just generally, that PLK1 inhibitors have been a challenge in the clinic just because
• 27:28 - 27:33:  of their toxicity. So, a lot of them have not actually made it through the way to the clinic.
• 27:34 - 27:38:  But in terms of resistance mechanisms, off the top of my head, I actually don't know that,
• 27:38 - 27:40:  but that's something that's interesting to look into.
• 27:41 - 27:47:  Okay. We have a question from Peter. Peter, if you could unmute yourself and show your—
• 27:47 - 27:53:  Really nice talk. What proportion of the proteins that had altered levels after you
• 27:53 - 27:57:  changed PLK activity were known PLK phospho targets?
• 28:00 - 28:08:  I mean, so there's four, so Claspin, EMI1, RE1, and Bora, the four that are maybe most
• 28:08 - 28:12:  appreciated. There are a couple others, and they were, again, beta-TRCP substrates that we just
• 28:12 - 28:19:  didn't identify by mass spec, but there are a couple other known. But other than that, in terms
• 28:19 - 28:23:  of like known proteasomal targets, I think just maybe those four or five are the ones that we
• 28:23 - 28:30:  already knew about. I'm just sort of thinking whether, you know, PLK is mostly affecting
• 28:30 - 28:35:  E3 ligases directly and therefore affecting many targets, or whether it's affecting some of the
• 28:35 - 28:39:  targets as well, because it could work both ways, right? PLK could phosphorylate a target,
• 28:39 - 28:46:  preventing the E3 from getting it. Yeah, that's a good point. This is, yeah, I mean, this is why
• 28:46 - 28:51:  we think it's, like, interesting between the cyclin F and the beta-TRCP side is because we
• 28:51 - 28:57:  think of the beta-TRCP substrates as getting phosphorylated at the substrate level, whereas
• 28:57 - 29:01:  we haven't exactly worked it all out, but we think on the cyclin F side that it's probably not at the
• 29:01 - 29:07:  substrate level, but rather of controlling the ligase itself. So, it's kind of interesting that
• 29:07 - 29:14:  it's doing this, but by potentially two different mechanisms. Cool, thank you. And you could
• 29:14 - 29:19:  potentially dissect those two mechanisms to see what you're saying, that the inhibitors are toxic,
• 29:19 - 29:24:  to see what the toxic effect is. Is that through cyclin F or is it through the other
• 29:24 - 29:30:  degradation pathway, by overexpressing cyclin? Yeah, yeah, and the hope eventually is to be
• 29:30 - 29:35:  able to identify the sites in cyclin F that gets phosphorylated, so hopefully we can generate
• 29:35 - 29:41:  a mutant that is resistant to degradation in response to the PRK1 inhibitors, which would
• 29:41 - 29:48:  be a pretty powerful tool to assess why this is all happening. Yeah, very cool. If there are no
• 29:48 - 29:52:  further questions, we'll move on. Thank you very much.
• 29:54 - 30:01:  Stanford from Kim Pritchard's lab, Berlin's lab, and is going to talk about R-loops and how they
• 30:01 - 30:09:  can cause RNA-DNA hybrids in the cytoplasm, I assume, to activate the immune system. Breaking
• 30:09 - 30:18:  bad, great talk title. Okay, share your screen, because, oh, we lost Maya.
• 30:23 - 30:28:  Sorry, I'm just, I'm not able to start my video for some reason. I don't know if someone can...
• 30:29 - 30:32:  Yeah, I mean... Okay.
• 30:38 - 30:40:  We practiced it twice, and of course...
• 30:46 - 30:51:  Okay, I'll just carry on without the video, because I'm not, I don't seem to have permission
• 30:52 - 30:58:  to activate that, or maybe it's on now. Okay, go ahead.
• 30:58 - 31:05:  Yeah, thanks. So, I actually spoke at the Cell Cycle Club, I think, six or seven years ago,
• 31:05 - 31:10:  when I was just finishing my PhD and about to move to California. So, it's great that I can
• 31:10 - 31:21:  be here again and share with you my postdoc work from Colleen Simprich's lab. And this audience,
• 31:21 - 31:26:  of course, all knows about the importance of safeguarding the integrity of our genome,
• 31:26 - 31:32:  and what happens when this goes awry, and the consequences of genome instability.
• 31:32 - 31:39:  And this includes mutations that affect cellular fitness and can contribute to aging and various
• 31:39 - 31:46:  degenerative diseases and cancer. And one of the angles that we're studying this in the Simprich
• 31:46 - 31:53:  lab is how the act of transcription itself contributes to genome instability, and
• 31:53 - 31:59:  specifically through the formation of these structures called R-loops, which, as I mentioned,
• 31:59 - 32:07:  form co-transcriptionally, and they consist of an RNA-DNA hybrid and a displaced DNA strand.
• 32:09 - 32:14:  And R-loops can be thought of as a double-edged sword for genome stability because they have
• 32:14 - 32:21:  important regulatory roles that I've outlined some of these here in green. But when they're
• 32:21 - 32:27:  deregulated on the genome, or their levels are elevated, for example, through oncogene or growth
• 32:27 - 32:36:  factor activation, they can lead to DNA damage, they can block progression of the replisome,
• 32:36 - 32:44:  induce some replication stress, and basically threaten genome stability. And elevated R-loops
• 32:44 - 32:50:  and associated DNA damage have been linked to diseases such as cancer and neurodegeneration.
• 32:51 - 32:58:  And cells, therefore, have many mechanisms to prevent or suppress R-loops, basically to keep
• 32:58 - 33:06:  their levels in check. And these include helicases, which unwind the hybrid, such as senataxin or
• 33:06 - 33:12:  Cetex. Then there are various replication and repair proteins that resolve R-loops, including
• 33:12 - 33:18:  BRCA1. There's the RNase H family of enzymes that can specifically degrade the RNA-DNA
• 33:18 - 33:27:  in a hybrid. And also, previous work in the lab identified endonucleases XPG and XPF,
• 33:27 - 33:35:  which act on the R-loop to generate damage. And I had a really intriguing initial observation
• 33:35 - 33:40:  that triggered much of the work that I'm going to tell you about today. And that was, I was looking
• 33:40 - 33:46:  at the consequences of inducing unscheduled R-loop formation by loss of R-loop resolution
• 33:46 - 33:55:  factors, senataxin and BRCA1. And so I made an imaging probe to look at hybrid levels in cells.
• 33:55 - 34:02:  And this is a fusion of GFP to catalytically dead RNase H1, and I can apply this to fixed cells.
• 34:03 - 34:08:  And what I found is that upon loss of senataxin and BRCA1, you can see there's an increase in
• 34:08 - 34:15:  hybrid signal in the nucleus of cells, as would be expected. But what I was really intrigued by
• 34:16 - 34:24:  was this strong increase in signal also in the cytoplasm. And these signals are reduced
• 34:24 - 34:30:  by pretreatment with RNase H, indicating that they're coming from hybrids. And, you know,
• 34:30 - 34:37:  this really triggered a lot of kind of questions for me, like where are these hybrids coming from
• 34:37 - 34:45:  that are in the cytoplasm? And so I'll tell you about two questions today. So where do these come
• 34:45 - 34:50:  from and how do they get there? And how do these species more broadly affect the cell?
• 34:52 - 34:57:  To answer these questions, I needed to come up with good tools to detect cytoplasmic hybrids,
• 34:58 - 35:03:  because these had not been studied before. And so the tools and approaches that I've developed
• 35:04 - 35:11:  include imaging, using the probe that I just showed you, also biochemistry, a biochemical
• 35:11 - 35:19:  approach to specifically isolate and sensitively detect cytoplasmic hybrids, and also sequencing
• 35:19 - 35:29:  to really characterize the genomic origins of these molecules. So one of the first
• 35:29 - 35:34:  hypotheses that I had in my mind, and this is, of course, using previous work that was done in the
• 35:34 - 35:42:  lab, and that was kind of rattling around my head, and I thought, well, what if these endonucleases
• 35:42 - 35:48:  like XPG were acting on the R-loop in a way that was releasing the hybrid, and this could then be
• 35:48 - 35:55:  accumulating in the cytoplasm? So to test this, I developed this biochemical approach that we
• 35:55 - 36:02:  call cytodrip to isolate hybrids in the cytoplasm. I then removed bulk protein and RNA
• 36:02 - 36:09:  and pulled down the hybrids with this S9.6 hybrids antibody. And by end labeling and
• 36:09 - 36:15:  running on a gel, I can then blot for the hybrid signal. And using this approach,
• 36:17 - 36:22:  you can see, first of all, that this is the kind of gel, the kind of data that I get from this.
• 36:23 - 36:32:  And firstly, and consistently with the imaging data, if I deplete senataxin or BRCA1 in cells,
• 36:32 - 36:41:  I see this increased smear of hybrid signal on these gels. And in cells where I've tagged or
• 36:41 - 36:48:  where I've made XPG degrons so that I can rapidly degrade XPG in these cells, you can see that loss
• 36:48 - 36:56:  of XPG rescues this effect, indeed suggesting that the activity of this nuclease is required
• 36:56 - 37:04:  to generate these cytoplasmic hybrids. And the data I'm showing you here is from HeLa cells
• 37:04 - 37:12:  specifically. However, I've shown this phenomenon of the hybrids accumulating and being XPG dependent
• 37:12 - 37:19:  now in many different cell lines, cancer and also normal cells. So we think this is something
• 37:19 - 37:26:  that's widespread and conserved. And then a final point on this slide is I want to point out that
• 37:26 - 37:32:  these are native conditions. So I can determine the size distribution of the hybrids. And you
• 37:32 - 37:40:  can see that they range from hundreds of base pairs up to several KBs. And this is actually consistent
• 37:40 - 37:45:  with known nuclear R-loop sizes that people have determined by footprinting.
• 37:47 - 37:54:  So after these initial observations, I wanted to further probe the dynamics of cytoplasmic hybrids
• 37:55 - 38:02:  and to really try and flesh out potentially some of this model. And first of all, if XPG is indeed
• 38:02 - 38:08:  acting on R-loops in the nucleus, I would expect to also be able to see release in soluble hybrids
• 38:08 - 38:16:  in the nucleus. And then a question would be how quickly are these shuttled or escape out into the
• 38:16 - 38:22:  cytoplasm? And does this involve active export? And once in the cytoplasm, how stable are the
• 38:22 - 38:30:  hybrids? And finally, is there a cell cycle regulation of R-loop processing? So to address
• 38:30 - 38:39:  these questions of dynamics, I used splicing inhibition with Pladienolide B as a way to
• 38:39 - 38:48:  rapidly deregulate R-loop levels. So for the first question of hybrid formation and export,
• 38:49 - 38:57:  I performed a time course of PlaB treatment in cells. And you can see that I then extracted
• 38:57 - 39:03:  hybrid levels from cytoplasm and nucleoplasm. And it's interesting that the hybrids appeared
• 39:03 - 39:10:  very rapidly within 30 minutes of PlaB treatment. And these levels accumulated over time.
• 39:11 - 39:16:  And given that this is such a rapid process, I asked if the hybrids were being actively exported.
• 39:17 - 39:23:  And for this, I used leptomycin B to inhibit export in one and then induced
• 39:24 - 39:29:  R-loop deregulation with PlaB. And you can see that indeed,
• 39:31 - 39:38:  addition of leptomycin B partially reduces the levels of cytoplasmic hybrids, whilst at the
• 39:38 - 39:45:  same time increasing the levels within the nucleoplasm, suggesting that export of one has a
• 39:46 - 39:52:  partial role in mediating the localization or export of hybrids from the nucleus to the
• 39:52 - 39:59:  cytoplasm. So then how stable are these hybrids once they're shuttled?
• 40:00 - 40:11:  And for this, to address this question, I added PlaB to cells, and then I found that this was rapidly reversible.
• 40:11 - 40:20:  So I washed this out and then traced the levels of hybrids in the cytoplasm for up to 24 hours afterwards.
• 40:21 - 40:30:  And what I saw is that three hours after washout of PlaB, there was a peak level of hybrids in the cytoplasm,
• 40:30 - 40:36:  and this probably reflects the time needed to export them and also to de-repress splicing.
• 40:36 - 40:43:  And four hours after this peak, I already saw a significant reduction in hybrid levels,
• 40:43 - 40:48:  and they were basically nearly all cleared after 24 hours of washout.
• 40:49 - 40:55:  So I conclude that the hybrids have a half-life of about four hours in the cytoplasm.
• 40:56 - 41:02:  And the last part of the dynamics question, is R-loop processing cell cycle regulated?
• 41:02 - 41:11:  And specifically what I asked is, are cytoplasmic hybrids produced in dividing as well as non-dividing cells?
• 41:11 - 41:20:  So I serum starved MCF10A cells and then induced hybrids with PlaB.
• 41:20 - 41:26:  And you can see that in both asynchronous and non-dividing serum starved cells,
• 41:26 - 41:32:  I was indeed able to see accumulation of hybrids in the cytoplasm.
• 41:32 - 41:37:  And I repeated this in other cell lines using serum starvation as well.
• 41:38 - 41:46:  So to summarize these first parts, I've managed to flesh out some of this model
• 41:46 - 41:55:  and show that I do indeed see release of soluble nuclear hybrids that are then rapidly exported to the cytoplasm,
• 41:55 - 41:58:  and that this is partially mediated by exportin 1.
• 41:58 - 42:03:  These hybrids have a half-life of about four hours in the cytoplasm,
• 42:03 - 42:08:  and this process occurs in dividing as well as non-dividing cells,
• 42:08 - 42:13:  suggesting it does not need or require DNA replication.
• 42:15 - 42:19:  So then, of course, the next big question for me was, well, where did these come from?
• 42:19 - 42:24:  I've shown the dynamics, but that big question still remained.
• 42:25 - 42:36:  And to address this, I modified and adapted my approach to be able to sequence the hybrids that I had isolated from the cytoplasm,
• 42:36 - 42:41:  and specifically by sequencing the single-strand DNA portion of the hybrid.
• 42:42 - 42:50:  And I had two main hypotheses that I entertained for how the sequencing results might look like.
• 42:50 - 43:00:  The first is shown in this cartoon here, and that is perhaps most R-loops within the nucleus are processed in a stochastic manner,
• 43:00 - 43:07:  such that if I compare my R-loop mapping with the cytoplasmic hybrid sequencing,
• 43:07 - 43:11:  I might expect most regions are represented in the cytoplasm.
• 43:12 - 43:23:  And secondly, and perhaps more interestingly, maybe only a subset of R-loops are acted on and therefore enriched in the cytoplasm,
• 43:23 - 43:32:  and studying these would be able to give me mechanistic insight into R-loop processing and why it even occurs.
• 43:33 - 43:40:  So it turns out that the data really point to the second of these hypotheses,
• 43:40 - 43:48:  and I'll take you through on this slide an example of the sequencing data at one site on the genome.
• 43:48 - 43:54:  So going from top down, the tracks are an IgG control,
• 43:54 - 44:01:  and then hybrids sequencing from the cytoplasm from control and senataxin-depleted cells.
• 44:02 - 44:09:  And on the bottom track here is total nuclear R-loops from regular nuclear DRIP-seq.
• 44:11 - 44:17:  And first of all, I want to point out that consistent with the previous data,
• 44:17 - 44:22:  loss of senataxin induces an increase in signal at this site.
• 44:22 - 44:31:  And also, gratifyingly, the cytoplasmic hybrids do indeed overlap with sites of nuclear R-loop formation.
• 44:32 - 44:38:  But what's interesting is whereas nuclear R-loops map to about 11% of the genome in our hands,
• 44:38 - 44:45:  you can see that the cytoplasmic hybrids map to just a small fraction of these.
• 44:46 - 44:53:  And finally, I want to point out that I developed these approaches to be high resolution
• 44:53 - 44:56:  and also to give strand-specific information.
• 44:56 - 45:04:  So you can see here that there's this interesting feature of signal coming from both plus and minus strands in this overlapping manner.
• 45:04 - 45:09:  And I'm going to come back to that and what it means mechanistically in a couple of slides.
• 45:10 - 45:16:  So where do these hybrid sites map to that are derived from the cytoplasm?
• 45:16 - 45:24:  And they come from both genic regions, so mostly within gene bodies, but also in intergenic sites.
• 45:24 - 45:32:  And those intergenic sites are enriched, and they're shown in blue here for the cytoplasmic hybrids.
• 45:32 - 45:37:  They're enriched at enhancers and to a lesser extent insulator regions.
• 45:37 - 45:45:  And I'm not showing the data here, but I also see an enrichment of the cytoplasmic hybrids at certain repetitive DNA elements.
• 45:46 - 45:51:  So back to that overlapping stranded signal that I mentioned.
• 45:51 - 45:59:  And here I'm showing an aggregate plot where I've oriented the hybrid signal with respect to transcription.
• 46:00 - 46:10:  And you can see that there's this peak in blue, or sense signal, followed by this shifted peak in red, or antisense signal.
• 46:11 - 46:16:  And thinking about the molecular models of what this signal would mean,
• 46:16 - 46:21:  it's actually consistent with the formation of converging R-loops,
• 46:21 - 46:30:  whereby you'd get this sense signal followed by this antisense signal coming from this antisense R-loop.
• 46:31 - 46:42:  And I'm not showing the data, but I confirmed by various nascent transcription data, for example, ProSeq and GroSeq,
• 46:42 - 46:50:  that these indeed are sites where we see signal from both sense and antisense strands.
• 46:50 - 46:56:  Suggesting that this is a site where converging transcription and converging R-loops could form.
• 46:58 - 47:05:  So then I asked, do these cytoplasmic hybrid sites have a particular nucleotide signature?
• 47:05 - 47:07:  And indeed they do.
• 47:07 - 47:16:  And that is that they're sites that have these shifts in nucleotide skew, and specifically purine skew.
• 47:16 - 47:22:  So this is defined as an asymmetry of purine on one strand compared to the other strand.
• 47:23 - 47:29:  So this is, again, an aggregate plot showing the skew across all of the cytodrip peaks.
• 47:29 - 47:36:  And you can see that right at the center, there's a shift from high purine skew, or purine richness,
• 47:36 - 47:43:  down to a lack of purines, where pyrimidines are instead enriched on the coding strand.
• 47:44 - 47:49:  And by doing a heat map of this signal, I can see that this shift in skew,
• 47:49 - 47:55:  so you can see from blue to red here, occurs at most of the peaks.
• 47:55 - 48:01:  So this aggregate signal isn't generated from just one or two really strongly skewed sites.
• 48:03 - 48:07:  And putting this onto our model then of the converging R-loops,
• 48:07 - 48:14:  what this means is that both of these hybrids would have a purine-rich RNA strand.
• 48:14 - 48:19:  And digging around in a lot of classical biochemistry papers,
• 48:19 - 48:25:  I found that this combination in a hybrid of having a purine-rich RNA strand
• 48:25 - 48:31:  actually has been associated with highly stable hybrids thermodynamically.
• 48:32 - 48:41:  And I validated this experimentally, showing that the cytodrip R-loops have very long half-lives on the genome.
• 48:41 - 48:47:  They persist for several hours, whereas a typical R-loop is degraded within 10 minutes.
• 48:47 - 48:50:  And they're even partially resistant to RNase H,
• 48:50 - 48:56:  suggesting these are really highly stable R-loops that form on the genome.
• 48:57 - 49:00:  So putting these sequencing data together then,
• 49:00 - 49:09:  our model is that senataxin resolves particularly stable converging R-loops on the genome.
• 49:09 - 49:18:  And when this is lost from cells, XPG instead can act on these generating hybrids that accumulate in the cytoplasm.
• 49:19 - 49:21:  The key question here is why.
• 49:21 - 49:27:  Why would cells cut out potentially big chunks of the genome to do this?
• 49:27 - 49:34:  And what I think is, of course, this does lead to DNA damage,
• 49:34 - 49:40:  but this could be a way of ultimately resolving these problematic R-loops
• 49:40 - 49:47:  that otherwise could interfere with transcription that may not be able to be resolved.
• 49:48 - 49:50:  Apart from through this backup pathway.
• 49:52 - 49:56:  Okay, so now I'll move to the second question and thinking more broadly,
• 49:56 - 50:00:  how does this R-loop processing affect the cell?
• 50:00 - 50:04:  And so one thing that really came to my mind several years ago
• 50:04 - 50:08:  when I was looking at this buildup of these hybrids in the cytoplasm was,
• 50:08 - 50:11:  what if this was triggering innate immune responses?
• 50:12 - 50:21:  And I was inspired by some findings at the time that had reported links to synthetic RNA-DNA hybrids
• 50:21 - 50:29:  and innate immunity through activation of these immune receptors in the cytoplasm that are shown in red here.
• 50:29 - 50:37:  And that included the activity of cGAS and TLR9 by synthetic hybrids.
• 50:37 - 50:48:  And then since those earlier studies, there have been links between deregulation of R-loops, DNA damage, and inflammation.
• 50:48 - 50:55:  However, whether endogenous hybrids coming from R-loops could activate these responses had not been reported.
• 50:56 - 51:05:  And I want to point out that downstream of these pattern receptors is IRF3,
• 51:05 - 51:15:  Interferon Response Factor 3, and this becomes phosphorylated, dimerizes and translocates into the nucleus where it activates interferon genes.
• 51:15 - 51:24:  And so we use the levels of phosphorylated IRF3 as a marker of activation of the innate immune response.
• 51:25 - 51:34:  And so, first of all, we were able to show that under conditions of both senataxin and BRCA1 depletion,
• 51:34 - 51:44:  we did see induction in phosphorylated IRF3, and this was abrogated by loss of XPG.
• 51:44 - 51:56:  So this is consistent with this model then that the hybrids building up in the cytoplasm are leading to innate immune activation and upregulation of phosphorylated IRF3.
• 51:57 - 52:06:  But we also know that R-loop deregulation can cause DNA damage and so potentially there could be other nucleic acids that were triggering this response.
• 52:06 - 52:18:  And to really definitively show the contribution of hybrids, we developed this cytoplasmically expressed RNase H construct and we expressed this in cells.
• 52:19 - 52:26:  And I was able to show that this RNase H was able to degrade hybrids in the cytoplasm.
• 52:27 - 52:39:  And at the same time, it partially reduced phosphorylated IRF3 levels, suggesting that whilst other nucleic acids could be contributing in these contexts,
• 52:39 - 52:47:  the cytoplasmic hybrids themselves directly were leading to immune activation in these scenarios.
• 52:48 - 52:52:  We then asked which immune receptors were sensing the hybrids.
• 52:52 - 53:00:  And so we depleted or knocked out these various pattern recognition receptors shown in red here.
• 53:00 - 53:09:  And as expected from the synthetic hybrids data that I mentioned, we did find that there was a dependence on cGAS.
• 53:10 - 53:20:  So here you can see when I inhibit cGAS, and we also found this for knockouts or knockdowns, we're able to suppress the levels of phosphorylated IRF3.
• 53:20 - 53:25:  But to our surprise, we also found that TLR3 was involved.
• 53:25 - 53:29:  And this is a canonical double-strand RNA sensor.
• 53:30 - 53:37:  And moreover, the cGAS and TLR3 activate IRF3 in this cooperative manner.
• 53:39 - 53:49:  And we were further able to show that cGAS and TLR3 bind to hybrids in the cytoplasm as well as in vitro in EMSA assays.
• 53:50 - 54:00:  So finally, my last data slide, we wanted to really apply our findings to a more disease-relevant model.
• 54:00 - 54:14:  And so in this case, we used ataxia patient-derived cells, specifically this AOA2 form of ataxia that has a loss-of-function mutation in senataxin.
• 54:15 - 54:25:  And comparing these AOA2 cells with control fibroblasts, I showed that these cells have increased hybrid levels in the cytoplasm.
• 54:25 - 54:34:  And at the same time, they also have upregulated interferon and cytokine expression shown in the red bars.
• 54:35 - 54:44:  But interestingly, depleting XPG in these cells rescues both the hybrid levels and immune activation,
• 54:44 - 54:51:  suggesting that these cells have elevated ALU processing and associated innate immune activation.
• 54:52 - 54:56:  So I'll then summarize all of our findings in this model.
• 54:57 - 55:12:  We find that when R-loops are deregulated, nucleases, including XPG, process these R-loops, generating hybrid fragments that accumulate in the cytoplasm.
• 55:13 - 55:22:  These are sensed by cGAS and TLR3 receptors, leading to an IRF3-mediated innate immune response.
• 55:23 - 55:36:  And we've been able to detect this in patient-derived cells with senataxin mutations and also in data that I didn't have time to show you in BRCA1-mutated cancer cells.
• 55:37 - 55:51:  So more broadly then, I'd like to suggest and propose cytoplasmic hybrids as really a previously unknown activator of the innate immune response and a pro-inflammatory factor.
• 55:52 - 56:00:  And you can imagine that at a low level, R-loop processing is beneficial in cells to remove the really stable R-loops.
• 56:00 - 56:03:  And this happens below the threshold of immune activation.
• 56:03 - 56:13:  However, upon deregulation of R-loops and increased R-loop processing, there are more cytoplasmic hybrids released, leading to inflammation.
• 56:14 - 56:27:  And this could occur in many diseases where R-loops are deregulated by certain mutations, including cancer, but also autoimmunity and neurodegenerative diseases.
• 56:28 - 56:37:  So with that, I'd like to thank members of the Simprich Lab, Carleen, and the other people shown in bold here that contributed to this work.
• 56:38 - 56:42:  In particular, Chenlin Song, another postdoc who worked on some of this with me.
• 56:43 - 56:46:  And I'd be very happy to take your questions.
• 56:46 - 57:01:  Excellent talk. Very interesting. Let's get out of the way with the questions because there's not a lot of time left. We'll start with Wei Ting. Could you show yourself and could Abcam unmute Wei Ting?
• 57:01 - 57:23:  Hello, a very interesting talk. I'm just wondering, basically, two questions. So the first thing is about damage specificity of R-loops. So several years ago, I think some lab actually shows that IRF-specific R-loop exists after damage, and it's dependent on Dicer-Drosha.
• 57:24 - 57:36:  So I'm wondering whether you see some sort of like, well, given the role of Dicer-Drosha in exporting small RNA, do you see a dependency on those specific R-loops?
• 57:37 - 58:03:  And the second one is about location. So I was, in my past life, when I was working on R-loop, I was always surprised by the intensity of S9.6 towards nucleolus and R-loop formation around there. So do you see more cytoplasmic R-loop coming from our ribosomal RNA?
• 58:04 - 58:21:  Okay, so yeah, I think I can answer all of those questions. So firstly, we haven't checked for Dicer-Drosha. There's a postdoc now in the lab working to do a screen to look more unbiasedly on export factors for hybrids. So we don't know that.
• 58:21 - 58:44:  We haven't specifically tested, like say, you know, IRF or, you know, perhaps Galer-Legube system of inducing breaks site specifically in seeing whether hybrids are released as a result of DNA damage repair. I think that's, you know, we've discussed that. And I think that would be an interesting thing. Maybe other people are even, you know, following up on.
• 58:45 - 59:05:  And yes, I do, I didn't have time to show you, but I do see an enrichment of cytoplasmic hybrids at rDNA more than you'd expect from averaging out the signal from nuclear R-loops. So yeah, I think this is happening also at ribosomal DNA.
• 59:06 - 59:20:  So it's a real pleasure to welcome Sophie Martin. We've been trying to get Sophie to come for a year or so now, but one thing or another conspired against it. So it's fantastic that she's been able to come.
• 59:20 - 59:36:  For those of you who don't know Sophie, she started out in Switzerland, working, doing her degree there and partly working, part of the work there was with Susan Gasser, who was a previous keynote speaker at this meeting.
• 59:37 - 59:54:  She then did her PhD at the Gurdon Institute with Daniel St. Johnston, and then a postdoc at Columbia University in New York with Fred Chang when he was there, and we must have overlapped partially, I was upstairs in Rodney's lab.
• 59:54 - 59:59:  And it's interesting that, you know, Sophie was working very much on polarization.
• 60:00 - 60:06:  In Fred's lab, that was in Pombe, which is a model system that she's continued with.
• 60:06 - 60:12:  And at the same time, I went to Rodney's lab just upstairs, swearing I would never work on budding yeast.
• 60:12 - 60:21:  I would only work on mice and ended up also converting to yeast.
• 60:21 - 60:37: She started her own lab at the University of Lausanne and has done tremendously well, looking both at cell polarization, but I think more broadly at the spatial organization of cells.
• 60:37 - 60:55: And really it's that incredible spatial organization which really sets Sophie's work apart and thinking carefully about the way that cells are organized in spatial context that really appeals to me.
• 60:55 - 60:57: She's a full professor there.
• 60:57 - 61:15: She has been director of the Fundamental Microbiology there during the pandemic, which must have been interesting. You hate to think she's won many awards, she's an EMBO gold member, gold medal winner, EMBO Young Investigator Award winner, I'm not going to mention them all,
• 61:15 - 61:26: but an EMBO member, tremendously successful. So it's a real pleasure to have Sophie here today and we greatly look forward to your talk.
• 61:26 - 61:30: Well, thanks a lot. Can you hear me?
• 61:30 - 61:35: We can. Okay, good. Well thanks a lot for the invitation.
• 61:35 - 61:41: I'm happy to be here, although I guess I'm still just in my office.
• 61:42 - 61:47: And thank you for the introduction and it's going to be hard to follow up from that.
• 61:47 - 62:03: So, I first have to apologize I'm actually not really going to talk about the cell cycle so I've racked my brain a little bit about how to link it to the cell cycle but I think I'll start with a link and then it's going to diverge a little away.
• 62:03 - 62:10: Because what I want to tell you about is some of our recent work on the process of cell fusion.
• 62:10 - 62:35: And of course, cell fusion can be viewed in the context of the general life cycle of organisms, most of which have a sexual reproduction cycle, where they go through a haploid phase and then a diploid phase, and these two phases must be coordinated so that there's
• 62:35 - 62:55: a balance of haploid and diploid phases. And, of course, the going from diploid to haploid involves the meiotic program and going from haploid to diploid involves fertilization or cell fusion, where two haploid gametes will merge together to form the
• 62:55 - 62:57: diploid zygotes.
• 62:57 - 63:01: So this is a step that we're very interested in.
• 63:01 - 63:16: And as Peter said, we are using fission yeast as our workforce to understand some of the basic principles by which cells organize during their life cycle.
• 63:16 - 63:38: So, of course, these cells are mostly in a haploid state, and they can cycle in a vegetative growth cycle, which has been extremely useful to study the basics of the cell cycle, but they also can go through sexual reproduction, where two different
• 63:39 - 63:59: types are called simply plus and minus, differentiate when they are faced with nitrogen deprivation to form gametes that will secrete pheromones to signal to each other, and therefore induce the growth of one to the other, and eventually fuse together to form
• 64:00 - 64:12: haploid zygotes and fusion in the diploid state is typically unstable and will immediately return to the haploid state by meiosis forming four haploid spore progeny.
• 64:12 - 64:32: So, we're interested in this stage and our general mode of study of this is through a lot of live-cell imaging and I'm showing you here a very simple movie of this where the plus cells are labeled with a cytosolic signal, and the minus cells are not
• 64:32 - 64:46: and you can see the cells growing towards each other and eventually merging together where the cytosolic fluorescence is now spreading through the cytoplasm.
• 64:46 - 65:02: So, as I said, the cells signal to each other and there's been extensive work in fission yeast and also in budding yeast, where there's a related signaling pathway to understand how the cells talk to each other.
• 65:02 - 65:12: So this is a bit scary and the point is not that you remember any of these gene names, but I just want to point out some of the basics of how this comes about.
• 65:12 - 65:32: So, in fission yeast, this is triggered by nitrogen starvation, and the nitrogen starvation will have two main effects: one is to promote the arrest of the cell in G1, and G1 is the only phase that is permissive for mating and sexual reproduction, and the nitrogen
• 65:32 - 65:54: starvation will also start to lead to the initiation of expression of a transcription factor that will initiate a feedback positive feedback mechanism that leads to the transcription of lots of genes involved in pheromone production and communication, including
• 65:54 - 66:17: the cells themselves, and then these proteins will feel partners and elicit very typical MAP kinase signaling pathways that feed back into this transcriptional regulation, therefore locking the cells into the differentiated states.
• 66:18 - 66:30: This pathway also activates through ways connections that are not all well characterized, cell polarization that ultimately will lead to the cell fusion process.
• 66:31 - 66:43: And coming back to this question of alternation between haploid and diploid states. And, in fact, it's possible to find conditions where this doesn't go correctly.
• 66:43 - 66:49: So, a few years back when Alex was in my lab.
• 66:49 - 67:08: He found a mutant that had problems in the alternation between haploid and diploid state with, for example, the formation of triploid zygotes, so this is an example here, where two gametes fuse to form a zygote, but this zygote instead of stopping there and going to my
• 67:08 - 67:24: other gametes goes on to fuse with a third gamete, therefore, forming a triploid cell that will nevertheless attempt some sort of meiosis and probably these spores are not very viable.
• 67:24 - 67:38: So, this, Alex found depends on the specific expression in the two gametes in the two cell types of a half of a transcription factor.
• 67:39 - 67:58: And these two halves are called Mi, M-gamete, and Pi, P-gamete, and these two halves are inactive by themselves but when they come together, they form a homeobox transcription factor that now binds the DNA and will lead in the zygote
• 67:59 - 68:10: to the transcription of a few genes, the major one called mei3, but it doesn't really matter I will not talk about it much further.
• 68:10 - 68:24: And this, we found leads to not only the initiation of meiosis—this was known previously—but also to blocking further mating attempts and thus fertilization events.
• 68:25 - 68:39: In follow-up work, Alex also established that this pathway, by forming this MIT complex, that leads to the mei3 transcription, leads to the activation of this pathway that was well established previously, and actually feeds into the regulation, not only of the entry into meiosis, but also directly into the core of the cell cycle by ways we don't yet understand, but we know that
• 69:00 - 69:15: both of these genes, mei2 and mei3, will feed into the cycling CDK regulation to promote the G1 to S transition.
• 69:16 - 69:38: And in fact, if we block zygotes in G1 by ways that have nothing to do with this pathway. And the way we had used was to delete all the non-essential cyclins in fission yeast and the cells are viable they can live on the sole CDC13 cyclin.
• 69:39 - 69:57: They have a very extended G1 phase, and when they mate, they also occasionally have problems in stopping the mating, and therefore, you can see cases where again there are two cells here that make together to form initial zygote and decide
• 69:58 - 70:14: to stop, goes on and fuses with the third cell here and forms this triplet because normally this pathway ensures that the cells go quickly into the cell cycle and therefore block the possibility for the cycles to
• 70:14 - 70:20: continue the mating process.
• 70:20 - 70:41: So he started his lab and he's now working on these problems and studying these further, and I will then shift my attention to the work we've been concentrating on mostly, which is really to understand this transition between the gamete
• 70:42 - 70:48: and the zygotic fate, which happens at cell fusion and try to understand really how cell fusion happens.
• 70:48 - 71:11: So, cell fusion is necessary for fertilization, as I've already mentioned, but it's also quite ubiquitous in the development of various organisms—in our own bodies, we have quite a few tissues and organs that are formed by processes of cell fusion.
• 71:11 - 71:19: The most important ones are the muscles, the bones, and the placenta.
• 71:19 - 71:39: In other organisms, there are full programmed processes of cell fusion during development. For example, in C. elegans, in fungi, cell fusion happens also in somatic cells in filamentous fungi.
• 71:39 - 71:47: So these are very closely fused together, and this is thought to allow long-range communication within the mycelium.
• 71:47 - 71:51: Okay, so focusing on this now.
• 71:51 - 72:03: I also want to highlight one important point that is specific to fungi—or I should say specific to any organism that has a cell wall.
• 72:04 - 72:20: These cells are surrounded not only by the plasma membrane but also by a cell wall, and the cell wall offers a very strong protection against external insults, but also against the internal turgor pressure that the cells are under.
• 72:20 - 72:37: And it's important to think about this in trying to understand the fusion process. I just want to put in your head the picture that the measured turgor pressure in the cells is equivalent to what you would find in the tire of a racing bike.
• 72:37 - 72:50: So you can imagine that a cell typically wants to keep its cell wall intact, otherwise, it will simply burst and that will lead to cell lysis.
• 72:50 - 73:03: But during fusion, the cell has to break open its cell wall, and therefore the cell fusion in wall cells is a very dangerous process.
• 73:04 - 73:22: It has to be very carefully controlled. So, we can represent schematically the process of cell fusion by representing a few different steps. So first the cells have to communicate with each other and pair together—I've mentioned this involves pheromone signaling,
• 73:22 - 73:25: and I will not talk very much about this further.
• 73:25 - 73:36: Once the cells have paired, now they need to start digesting the cell wall and they have to do that very precisely. And that's what I will focus mostly on.
• 73:37 - 73:47: And then, this relies on the reorganization of the actin cytoskeleton, and in particular this protein, Fus1, which is a protein that nucleates linear actin filaments.
• 73:47 - 74:00: And then once the cell wall has been remodeled, the plasma membranes have to fuse together. And in fungi, the protein machinery that may be involved in that is unknown.
• 74:00 - 74:12: It's supposed to be better known in other phyla, but we're still missing a lot of information about how plasma membranes are fusing.
• 74:12 - 74:32: So, let me tell you a little bit about Fus1. So, Fus1 was actually identified in the 90s and studied by by Ann-Petterson. Initially, this mutant is strictly incapable of fusing, and what happens is that the cells can pair together as you can
• 74:32 - 74:48: see here. They form lots of nice pairs; the cells keep growing against each other, but they never digest the cell wall. And this is in contrast to wild-type that will go through and then form a nice tetrad that's the result of the meiotic process.
• 74:48 - 75:04: And what we found is that this Fus1 is required for the tight organization of the actin cytoskeleton, which in wild-type concentrates at the point of contact between two partner cells, but it's much more distributed in the Fus1 mutant, and
• 75:04 - 75:23: this concentrated structure we've called the actin fusion focus, and we found it serves to concentrate through transport by type five myosin, a number of cargoes to the type five myosin itself is also highly concentrated at this position, and in the absence of Fus1,
• 75:23 - 75:27: it's much more broadly distributed.
• 75:27 - 75:48: So, I will just summarize what we initially described, which is that the formation of the actin fusion focus serves to concentrate secretory vesicles transported by the myosin in the secretory vesicles contain several digestive enzymes, glucanases,
• 75:49 - 76:12: released on the outside of the cell. And what we suggested was that this very tight localization of release is changing the distribution of where the vesicles are released, allowing to concentrate cell wall digestion activity to the center of the contact region, while the
• 76:12 - 76:23: membrane remains intact, and therefore form just a little hole between the two cells, so the membranes can come into contact and eventually merge.
• 76:24 - 76:37: So we're very interested in trying to understand how this happens. So we can follow these events in live cells here I'm showing you an example where the myosin is labeled in different colors and the two partner cells.
• 76:37 - 76:55: And in this reaction, we express a cytosolic marker in one of the two cells, so we can monitor when fusion happens. And you can see that the structure forms, concentrates, and then stabilizes a bit before fusion. It may be easier to see that in a few
• 76:56 - 77:12: minutes from this time-lapse. We see the structure initially is quite broad and kind of moves around, and then one focus forms in each of the two cells, and then the two foci move closer and closer together, we can measure that very precisely until the point
• 77:12 - 77:18: where we are expecting when there's an abrupt movement of the two together.
• 77:18 - 77:24: And within the next 10-15 minutes then the structure disappears.
• 77:25 - 77:47: So, this structure, as I said, contains these actin and secretory vesicles, but in fact it contains a lot of other elements. And in particular, we found a few years back that it contains all the elements of the communication between the two cells.
• 77:48 - 77:50: So here are some examples.
• 77:50 - 78:07: The pheromone receptor accumulates in this location, the downstream signaling cascade, the active RAS protein accumulates there, as well as all three kinases of the MAP kinase cascade—I'm showing you just one of them here.
• 78:07 - 78:26: And this led to the question of why should the signaling be accumulated there and does it have any sort of function. And one way we addressed this question was by finding conditions in which this accumulation was perturbed.
• 78:26 - 78:44: So in one case where this accumulation is perturbed is in a mutant of the pheromone receptor that cannot be endocytosed; therefore, it is secreted, distributed at the plasma membrane, but then it diffuses laterally away from the point of secretion at the
• 78:45 - 78:56: focal point, and ends up not being accumulated, although the fusion focus is still present as you can see by the red localization here.
• 78:56 - 79:13: And interestingly in this mutant, what we found is that even though the focus forms, it is unable to stabilize relative to the partner that has a wild-type pheromone receptor and forms apparently correctly.
• 79:13 - 79:33: And as a consequence of that, these cells are almost completely unable to fuse, and the thought there is that the enzymes are no longer positioned at the center of the contact zone, but over time they're distributed over the whole region, and
• 79:33 - 79:43: there's not enough concentration for cell wall digestion; instead, the cell keeps growing as it was doing before when it had distributed secretion.
• 79:43 - 80:02: And so, with this, we made the reflection maybe we can re-target the receptor at the right position and see what happens. And so we did that by simply hooking to the myosin a GFP binding protein which will be
• 80:00 - 80:04: bind to the GFP moiety with nanomolar affinity.
• 80:04 - 80:07: And the result of that is that we indeed managed
• 80:07 - 80:12: to concentrate again the receptor at the focus.
• 80:12 - 80:15: You can see the yellow point here coming,
• 80:15 - 80:17: and you can also see that in the examples
• 80:17 - 80:20: I'm showing you here, which doesn't want to loop,
• 80:20 - 80:25: unfortunately, the cells actually managed to fuse.
• 80:26 - 80:28: And in fact, we found that in this condition,
• 80:28 - 80:31: we get up to about 50% of the pairs
• 80:31 - 80:33: that managed to fuse correctly.
• 80:33 - 80:37: So this suggested that indeed this site,
• 80:37 - 80:41: this fusion focus is also a site of local pheromone
• 80:41 - 80:44: and MAP kinase signaling, and that this has a purpose.
• 80:44 - 80:46: It serves to stabilize the structure
• 80:46 - 80:48: and promote the fusion process.
• 80:49 - 80:52: So now we don't know how it does that.
• 80:52 - 80:54: We have started working towards this,
• 80:54 - 80:58: and this is the work of Melvin, a current PhD student
• 80:58 - 81:03: who decided to take a proteomic or phosphoproteomic approach
• 81:05 - 81:10: to try and find targets of the MAP kinase cascade
• 81:10 - 81:13: on the focus or at the time of fusion.
• 81:13 - 81:16: And one of the challenges for that he faced
• 81:16 - 81:19: was to synchronize the fusion process
• 81:19 - 81:21: so he could take populations of cells
• 81:22 - 81:23: at specific time points.
• 81:24 - 81:28: Unfortunately, the fusion process is very asynchronous
• 81:28 - 81:30: in fission yeast.
• 81:30 - 81:34: And so he developed a way that involves splitting
• 81:34 - 81:36: the formin in two parts,
• 81:36 - 81:39: each of the parts being non-functional.
• 81:39 - 81:43: And you can see that in this example here,
• 81:43 - 81:47: the cells form pairs, but they don't fuse.
• 81:47 - 81:51: There's fluorescence only in one of the halves.
• 81:51 - 81:53: And then he hooked each of the two halves
• 81:53 - 81:58: to an optogenetic system that allows to reform
• 81:59 - 82:03: full formin in the presence of light.
• 82:03 - 82:07: And amazingly, that works remarkably well.
• 82:07 - 82:11: And upon a few pulses of light,
• 82:11 - 82:14: he gets fusion to happen within about 20 minutes
• 82:14 - 82:16: in the whole population.
• 82:16 - 82:21: And so that allowed him to perform phosphoproteomics.
• 82:21 - 82:24: And I'm just giving you this as a teaser
• 82:24 - 82:26: because we don't—we're still analyzing the data
• 82:26 - 82:30: and we don't have yet an understanding of what happens,
• 82:30 - 82:32: but we found robust phosphorylation
• 82:32 - 82:34: in a large number of genes
• 82:34 - 82:37: by doing both a mating time course and a fusion time course
• 82:37 - 82:41: and found a number of very interesting candidates,
• 82:41 - 82:43: including formin itself.
• 82:43 - 82:47: So we are currently characterizing this.
• 82:48 - 82:51: Okay, so I've told you the fusion focus
• 82:51 - 82:56: is this structure that has actin, secretory vesicles,
• 82:56 - 82:58: and a lot of signaling components.
• 82:58 - 83:01: So can we see it in more detail?
• 83:01 - 83:05: To do that, we took an ultra-structural approach
• 83:05 - 83:10: and used—actually went to Van der Krupelski's lab.
• 83:11 - 83:14: It's now a number of years ago to learn from her
• 83:14 - 83:18: how to do correlative light electron microscopy
• 83:18 - 83:20: where we can use the information
• 83:20 - 83:25: of the fluorescence accumulation of the formin or the myosin
• 83:26 - 83:30: to tell us these cells are now in the process of fusion
• 83:30 - 83:32: and identify these pairs
• 83:32 - 83:34: and now go back to the electron microscope
• 83:34 - 83:36: and identify this pair
• 83:36 - 83:41: and find how the ultra-structure of the fusion site is.
• 83:42 - 83:47: And by doing that, we get extremely detailed information.
• 83:48 - 83:51: And I'm showing you here one example—in green.
• 83:51 - 83:52: I've just highlighted the cell wall
• 83:52 - 83:54: in magenta is the plasma membrane,
• 83:54 - 83:58: and you can see this strong accumulation of vesicles
• 83:58 - 84:01: in both of the cells here.
• 84:01 - 84:05: And I'll point out as well that you see a lot of
• 84:06 - 84:09: ribosomes at lots of other locations.
• 84:09 - 84:12: And of course you see lots of other organelles.
• 84:12 - 84:14: But the key point is that we see
• 84:14 - 84:16: this strong accumulation of vesicles,
• 84:16 - 84:19: which we determined are indeed secretory vesicles.
• 84:19 - 84:23: And we can look at the depth
• 84:23 - 84:28: because we acquire tomographic information.
• 84:28 - 84:33: So we have about 300 nanometer depth of vision.
• 84:33 - 84:36: And as I'm going to animate going up
• 84:36 - 84:38: and down one of these tomograms,
• 84:38 - 84:40: and you can see the ultra-structure
• 84:40 - 84:45: and I have also segmented the key structures
• 84:46 - 84:47: to make it easier.
• 84:47 - 84:49: So in yellow is all the filaments,
• 84:49 - 84:53: the actin filaments that we can see
• 84:53 - 84:58: and in pink and magenta, the vesicles.
• 84:59 - 85:02: And you see in blue, the plasma membrane.
• 85:03 - 85:05: So we observed what we expected to observe,
• 85:05 - 85:06: which is a good thing.
• 85:06 - 85:10: A lot of actin, linear actin filament, a lot of vesicles.
• 85:10 - 85:14: We also made observations we did not expect at all.
• 85:14 - 85:16: And one of these is already visible
• 85:16 - 85:19: on this reconstruction here,
• 85:19 - 85:21: where you can see that one of the cells
• 85:21 - 85:24: has a more pointy appearance.
• 85:24 - 85:27: It seems to be protruding forward.
• 85:27 - 85:32: While the other one has a flatter surface
• 85:32 - 85:36: and also the plasma membrane is quite wavy in appearance.
• 85:36 - 85:40: It's not as tense as the other one.
• 85:40 - 85:44: In some cases, we see a much more prominent feature
• 85:44 - 85:46: like that, much more exaggerated features
• 85:46 - 85:50: where one cell appears to be really almost protruding
• 85:50 - 85:52: into its partner cell.
• 85:53 - 85:58: And this asymmetric organization is very unexpected
• 85:58 - 86:02: because these cells do not differ much at all.
• 86:02 - 86:05: We found that actually this is largely dictated
• 86:05 - 86:08: by the mating type with the minus cell
• 86:08 - 86:11: being typically the protruding one.
• 86:11 - 86:16: It's not 100% of the cases, it's more like 90%.
• 86:16 - 86:19: And this suggests that there may be some asymmetry
• 86:19 - 86:22: that's important in the fusion event as well.
• 86:22 - 86:27: So, so far what we established
• 86:27 - 86:29: is that there are indeed differences
• 86:29 - 86:32: between the minus and the plus cell.
• 86:32 - 86:35: The minus—sorry, the plus cell
• 86:35 - 86:40: seems to have a higher amount of exocytosis
• 86:40 - 86:43: relative to endocytosis compared to the minus cell.
• 86:43 - 86:46: And we imagine this may cause these waves
• 86:46 - 86:48: that we see at the plasma membrane.
• 86:48 - 86:53: Maybe we imagine it a bit like a membrane traffic jam
• 86:53 - 86:56: where so much membrane arrives and is secreted at the location
• 86:57 - 86:59: that endocytosis has difficulty to catch up
• 86:59 - 87:01: and therefore there's essentially
• 87:01 - 87:03: an excess of plasma membrane.
• 87:04 - 87:09: We also found that if we reduce the turgor pressure
• 87:11 - 87:13: in the h minus cell,
• 87:15 - 87:19: we lose this asymmetric appearance
• 87:20 - 87:24: and the cells take longer to fuse.
• 87:24 - 87:28: And so this suggests that there's some contribution
• 87:28 - 87:33: of turgor pressure in suddenly generating this asymmetry
• 87:33 - 87:37: and also in promoting somehow the fusion process.
• 87:38 - 87:42: But it is not at this point clear how much this contributes
• 87:42 - 87:46: and what the effect really of the turgor pressure maybe.
• 87:48 - 87:53: Okay, so I'm not going to go into more detail in that part,
• 87:54 - 87:59: but I want to point out that in this ultrastructural work,
• 87:59 - 88:01: we also made one interesting observation,
• 88:01 - 88:03: which is that this whole region
• 88:03 - 88:06: that has all these secretory vesicles
• 88:06 - 88:08: is actually devoid of ribosomes.
• 88:08 - 88:12: And I'm highlighting here in pale blue,
• 88:12 - 88:14: this whole region that doesn't have ribosomes.
• 88:14 - 88:19: The ribosomes are these dark spots you see everywhere.
• 88:19 - 88:22: The yeast cytosol is full of ribosomes.
• 88:22 - 88:24: And so when a region is devoid of ribosomes,
• 88:24 - 88:28: it typically indicates that ribosomes are inhibited
• 88:28 - 88:30: from penetrating there.
• 88:30 - 88:32: That's something that's denser
• 88:32 - 88:37: that prevents entry or diffusion of the ribosomes in.
• 88:38 - 88:40: And so this is the wild-type situation.
• 88:40 - 88:43: We also found in some mutant situations
• 88:43 - 88:45: that this could be exacerbated.
• 88:45 - 88:48: So for example, we have a mutant,
• 88:48 - 88:50: I can answer a question about this,
• 88:50 - 88:54: but the important thing is that in this particular mutant,
• 88:54 - 88:58: there's an excess of formin and actin at the fusion site,
• 88:58 - 89:02: but very few secretory vesicles.
• 89:02 - 89:06: But this mutant also forms this very large region,
• 89:06 - 89:08: again highlighted here in blue,
• 89:08 - 89:10: that is devoid of ribosomes.
• 89:10 - 89:12: So it is very dense in something.
• 89:14 - 89:16: And so the last part of my talk,
• 89:16 - 89:20: I want to talk about what is this structure.
• 89:20 - 89:22: This suggests this fusion focus
• 89:22 - 89:26: is essentially a membrane-less organelle.
• 89:28 - 89:31: And it has some sort of integrity
• 89:31 - 89:36: that prevents entry of large macromolecules.
• 89:36 - 89:40: So I need to introduce a little bit more about formins.
• 89:40 - 89:42: So I've told you formin,
• 89:42 - 89:47: formin is a key component and regulator of this.
• 89:47 - 89:51: Formins are well understood proteins.
• 89:51 - 89:54: They consist of the characteristic
• 89:54 - 89:57: formin homology domains one and two.
• 89:57 - 90:02: And these domains together serve to nucleate the formation
• 90:03 - 90:07: and then promote the elongation of linear actin filaments.
• 90:07 - 90:09: And this is extremely well understood.
• 90:09 - 90:14: What is a bit less characterized and more diverse
• 90:14 - 90:16: in lots of different formins
• 90:16 - 90:18: is the other parts of the protein,
• 90:18 - 90:22: in particular, the typically very large N-terminal region,
• 90:22 - 90:26: which is regulatory and typically regulates localization.
• 90:28 - 90:30: In fission yeast, there are three formins,
• 90:30 - 90:34: and fission yeast has been a very good model
• 90:34 - 90:36: to study the actin cytoskeleton.
• 90:37 - 90:39: For one of the reasons for that is that
• 90:39 - 90:41: each of the nucleators,
• 90:41 - 90:44: nucleates a distinct actin structure.
• 90:44 - 90:47: And so we will focus on the formin, Fus1,
• 90:47 - 90:50: but the two others nucleate distinct structures
• 90:50 - 90:55: for division and for polarization of mitotic cells
• 90:55 - 90:57: that have different appearances.
• 90:58 - 91:02: And so we started to understand better
• 91:02 - 91:03: about the fusion focus.
• 91:03 - 91:06: We started by asking the question
• 91:06 - 91:10: of how much specificity is there in fus1.
• 91:10 - 91:12: And so the rest of my talk,
• 91:12 - 91:15: I will present the work of Ingrid Biot-Chamartin,
• 91:15 - 91:19: who just finished her PhD in my lab.
• 91:19 - 91:23: And so she started by simply trying to replace formin1
• 91:23 - 91:24: with the two other formins
• 91:24 - 91:27: and testing whether the cells were able to fuse.
• 91:27 - 91:28: And that didn't work.
• 91:28 - 91:33: And so then she constructed more subtle chimeric formins
• 91:35 - 91:37: where she replaced only parts of Fus1
• 91:37 - 91:39: with the corresponding part of the two others.
• 91:39 - 91:42: And she did that with various parts of the protein,
• 91:42 - 91:46: but I will focus on the replacement of the N-terminus.
• 91:46 - 91:49: And she found that if she replaced formin1 N-terminus
• 91:49 - 91:51: with either of the two other formins,
• 91:51 - 91:53: again, that was completely non-functional.
• 91:55 - 91:57: In the case of the replacement with the cdc12,
• 91:57 - 92:00: this is perhaps not surprising.
• 92:00 - 92:02: This chimeric formin is unable to localize
• 92:02 - 92:04: at the fusion site.
• 92:04 - 92:07: But in the case of the formin3 replacement,
• 92:07 - 92:10: actually the chimeric formin did localize correctly.
• 92:10 - 92:12:  or largely correctly.
• 92:12 - 92:14:  And so this suggested there's something else
• 92:14 - 92:16:  in this N-terminal region.
• 92:17 - 92:20:  So what we did then is to see
• 92:20 - 92:21:  whether we could learn something
• 92:21 - 92:26:  by expressing the protein ectopically in mitotic cells.
• 92:26 - 92:29:  So I failed to inform you that first one
• 92:29 - 92:32:  is normally expressed only during mating.
• 92:32 - 92:37:  It's expressed and is induced upon sexual differentiation,
• 92:38 - 92:41:  and it's not normally expressed in mitotic cells.
• 92:41 - 92:43:  But if we express it ectopically,
• 92:43 - 92:46:  what we found is that it's also able to form
• 92:46 - 92:50:  a fairly large focus in this mitotic cell
• 92:50 - 92:53:  that recruits also the type 5 myosin
• 92:53 - 92:55:  as shown here by the little arrows.
• 92:56 - 93:00:  And this focus of the formant is very resistant
• 93:00 - 93:02:  to various treatments.
• 93:02 - 93:04:  So if we treat with latrunculin A,
• 93:04 - 93:07:  that depolymerizes the actin cytoskeleton,
• 93:07 - 93:09:  the focus still forms,
• 93:09 - 93:13:  although it's not able to recruit the myosin anymore.
• 93:13 - 93:17:  And also if we treat cells with 1,6-hexanediol,
• 93:17 - 93:20:  which is a drug that's been used a lot
• 93:20 - 93:23:  in the phase separation field
• 93:23 - 93:27:  to disrupt weak interactions,
• 93:27 - 93:31:  we find that actually the focus remains
• 93:31 - 93:34:  and even some of the myosin 5 remains in there.
• 93:35 - 93:39:  So in mitotic cells, it forms something quite solid,
• 93:39 - 93:42:  and we can actually see it
• 93:42 - 93:44:  by correlative light electron microscopy.
• 93:44 - 93:48:  And again, focusing on the fluorescence part,
• 93:48 - 93:51:  we can see it forms a zone that excludes ribosomes
• 93:51 - 93:54:  and can recruit some vesicles,
• 93:54 - 93:56:  although these vesicles are typically
• 93:56 - 93:58:  on the periphery of this zone.
• 94:00 - 94:04:  We did then the same with just the N-terminus of first one
• 94:04 - 94:05:  and found that this N-terminus
• 94:05 - 94:08:  carries a number of information.
• 94:08 - 94:10:  When we express it in mitotic cells,
• 94:10 - 94:12:  it localizes to the poles of the cells,
• 94:12 - 94:17:  and it also forms these little foci clusters
• 94:18 - 94:20:  inside the cytosol.
• 94:20 - 94:22:  These clusters, again, are not perturbed
• 94:22 - 94:24:  by treatment with latrunculin A,
• 94:24 - 94:27:  nor is the pole localization.
• 94:27 - 94:29:  And when we treat with 1,6-hexanediol,
• 94:29 - 94:32:  we diminish them, but they're not gone completely,
• 94:32 - 94:35:  although we lose the pole localization
• 94:35 - 94:37:  and we perturb the cells; they basically die
• 94:37 - 94:39:  after a few minutes of this treatment.
• 94:40 - 94:42:  Again, we can look at them
• 94:42 - 94:44:  by correlative light electron microscopy
• 94:44 - 94:48:  and find that the zones that contain the fluorescence
• 94:48 - 94:52:  form a region that is devoid of ribosomes.
• 94:52 - 94:55:  And so, again, this forms a fairly dense structure.
• 94:55 - 94:57:  So that suggests that first one
• 94:57 - 95:00:  has probably the potential to self-interact
• 95:00 - 95:01:  and form some aggregates
• 95:01 - 95:05:  or something that we think in the mitotic cells
• 95:05 - 95:09:  is not really normal in its state.
• 95:11 - 95:13:  Okay, so what is in this N-terminus?
• 95:13 - 95:16:  So this N-terminus, if we look at the sequence,
• 95:16 - 95:20:  has a predicted domain
• 95:20 - 95:25:  that previous work showed likely negates localization.
• 95:25 - 95:28:  And then it has a large predicted disordered region.
• 95:29 - 95:33:  Interestingly, if we truncate the disordered region,
• 95:33 - 95:38:  then we keep the localization of the forming fragments
• 95:40 - 95:41:  at the poles of the cells,
• 95:41 - 95:46:  but we lose the cytosolic aggregates.
• 95:46 - 95:48:  If we truncate from the N-terminus,
• 95:48 - 95:52:  encroaching into this GbDFH3 domain,
• 95:52 - 95:56:  we lose the polar localization,
• 95:56 - 95:58:  but we keep the cytosolic aggregates.
• 95:58 - 96:02:  So it seems that we can distinguish these two localizations
• 96:02 - 96:03:  and what we'll focus on
• 96:03 - 96:05:  is particularly this disordered region.
• 96:06 - 96:08:  This is while treating mitotic cells.
• 96:08 - 96:10:  Now what happens in mating cells?
• 96:11 - 96:15:  So in mating, we can do exactly the same experiments,
• 96:15 - 96:19:  take the FOS1 N-terminus and express it in the mating cells.
• 96:19 - 96:20:  Now it's a little more complicated
• 96:20 - 96:24:  because the mating cells have endogenous Fus1.
• 96:24 - 96:27:  And so if we express these fragments in wild type cells,
• 96:27 - 96:30:  we find that it localizes just like the endogenous form
• 96:30 - 96:33:  and it forms a focus that's very concentrated
• 96:33 - 96:36:  at the point of contact between the cells.
• 96:36 - 96:38:  But if we express it in a mutant
• 96:38 - 96:41:  that lacks the endogenous Fus1,
• 96:41 - 96:44:  we see that it can still localize,
• 96:44 - 96:47:  but it decorates a broader zone
• 96:47 - 96:52:  at the contact site between the cells.
• 96:52 - 96:55:  And we can see this distribution quantified here.
• 96:56 - 96:58:  And now, like we did in mitotic cells,
• 96:58 - 97:01:  we can truncate from the N-terminus.
• 97:01 - 97:04:  We lose the localization as we would have predicted.
• 97:04 - 97:05:  But interestingly,
• 97:05 - 97:09:  if we have the endogenous Fus1 still present,
• 97:09 - 97:12:  we keep the ability of this fragment
• 97:12 - 97:14:  to be recruited to the focus,
• 97:14 - 97:16:  suggesting it is interacting with something
• 97:16 - 97:17:  that's present on the focus.
• 97:18 - 97:23:  And now if we truncate from the disordered region site,
• 97:24 - 97:26:  now we don't perturb localization.
• 97:26 - 97:28:  It's perfectly fine.
• 97:28 - 97:31:  But in the wild type situation,
• 97:31 - 97:33:  this fragment fails to form a focus.
• 97:34 - 97:38:  It is still distributed broadly over the contact zone.
• 97:38 - 97:40:  And so it suggests it has lost the ability
• 97:40 - 97:42:  to be recruited to the fusion focus.
• 97:43 - 97:48:  So this region is required for self-interaction.
• 97:48 - 97:51:  And so we then ask whether it is important
• 97:51 - 97:53:  for the function of FOS1.
• 97:53 - 97:55:  And so we simply truncated it
• 97:55 - 97:58:  now in the context of the full-length protein
• 97:58 - 98:00:  expressed at the endogenous locus.
• 98:00 - 98:05:  And this truncated version lacking the IDR,
• 98:06 - 98:08:  localized correctly as we would have predicted,
• 98:08 - 98:13:  but distributed broadly as would also be predicted
• 98:13 - 98:16:  if this is required for self-interaction.
• 98:16 - 98:19:  And so we can see that this mutant form
• 98:19 - 98:22:  has a much broader distribution than the wild type.
• 98:24 - 98:28:  We played a little bit more with deletions in this domain.
• 98:28 - 98:31:  Oh, I'm getting ahead of myself.
• 98:31 - 98:33:  So one thing I should say that's important
• 98:33 - 98:37:  is that this mutant is now completely unable to fuse.
• 98:37 - 98:41:  These cells form pairs, but they can't fuse.
• 98:41 - 98:46:  In fact, a large fraction of them undergo cell lysis.
• 98:46 - 98:49:  And what we think happens is that the cell
• 98:49 - 98:52:  still tries to bring the vesicles
• 98:52 - 98:56:  that carry the digestive enzymes to this location,
• 98:56 - 98:58:  but not precisely enough,
• 98:58 - 99:00:  so that instead of being concentrated
• 99:00 - 99:02:  in the center of the zone,
• 99:02 - 99:05:  they may be a little bit off target.
• 99:05 - 99:07:  And therefore, as the cell is digested,
• 99:07 - 99:13:  it is not protected, and the cells lyse.
• 99:13 - 99:18:  So we dug a little further into this disordered region
• 99:18 - 99:24:  and created smaller deletions in there.
• 99:24 - 99:26:  And the details don't really matter,
• 99:26 - 99:27:  but what is interesting
• 99:27 - 99:30:  is that we obtained an intermediate phenotype.
• 99:30 - 99:34:  So there's an intermediate distribution
• 99:34 - 99:39:  in the width of the zone.
• 99:39 - 99:41:  Some of them, actually some of the deletions
• 99:41 - 99:44:  have almost no phenotype whatsoever,
• 99:44 - 99:47:  and these have very small phenotype in terms of fusion.
• 99:47 - 99:50:  And then if we combine by making longer deletions
• 99:50 - 99:53:  that combine two of these deletions together,
• 99:53 - 99:57:  we get something intermediate again in the distribution
• 99:57 - 99:59:  and in the fusion ability.
• 100:00 - 100:19:  So this suggests that this region probably serves to form multivalent interactions, each of which contributes a little bit, so when you remove one and not much happens but you remove more than one or all of them in the complete deletion, then it's a broad
• 100:19 - 100:22:  distribution and failure to fuse.
• 100:22 - 100:35:  So, this is all very nice; we can remove this region and have a fusion failure. But what we wanted to do is to see can we now restore the function that we've removed.
• 100:35 - 100:53:  Can we have Fus1 concentrate in some other ways and show that this is really the function that's missing. And so what we thought is that we could replace this disordered region with other domains, known to self-interact.
• 100:53 - 101:09:  And what we ended with was to take this optogenetic protein, this light-sensitive protein that comes from Arabidopsis called CRY2, which is known to oligomerize upon blue light illumination.
• 101:09 - 101:28:  And we simply created mutants, lacking a disordered region instead having CRY2. And remarkably, what we found is that when we expose this mutant to light, they are now better able to fuse than in the complete absence of the disordered region.
• 101:28 - 101:33:  And this is specific to light; in the dark, the fusion does not happen.
• 101:34 - 101:40:  So this tells us that indeed, the self-interaction is something critical in this region.
• 101:40 - 101:52:  But it's also a little frustrating because it's only 50% in there. So, depending on whether you like to see the glass half full or half empty, you can be a bit frustrated with this.
• 101:52 - 102:13:  So we looked a little deeper at what happens there. And what we found is that in these conditions here, in fact, we have an excess of first one, relative to the wild type, but these cells are not able to recruit either actin or the myosin
• 102:14 - 102:16:  efficiently at all.
• 102:16 - 102:30:  And in fact, we also found that if we use a mutant version of CRY2 that has been previously shown to oligomerize more tightly, the phenotype is worsened.
• 102:30 - 102:45:  And maybe this replacement created some sort of first one that oligomerizes too much, and therefore is not very functional. And so we decided, oh, and again I'm getting ahead of myself.
• 102:45 - 103:04:  And another indication toward this is what happens when we observe the cells under the microscope. So I'm going to show you the wild type first, and you will see these cells, forming a fusion focus; you're seeing first one here accumulating
• 103:04 - 103:08:  and disappearing, accumulating, and disappearing.
• 103:08 - 103:13:  And over time you will see the spores forming in these cells.
• 103:13 - 103:26:  When we have the CRY2 replacement, what happens is that you see the focus appear here; it appears, looks quite normal, and then suddenly doesn't look normal at all.
• 103:26 - 103:40:  And then as the cells are fusing and disappearing, it looks like there's more and more first one that appears, and then eventually this seems to form such a big structure that the structure kind of detaches, and the cells haven't managed to fuse and they
• 103:40 - 103:44:  each have a floating structure around.
• 103:44 - 103:52:  And if we do a FRAP experiment we find that the turnover in these structures here is much slower than in the wild type.
• 103:52 - 104:06:  Okay, so all of this suggested that maybe by doing this replacement we haven't provided the right affinity or we've provided too much binding of first one with itself.
• 104:06 - 104:22:  So we decided to perform the same type of experiment, but with another domain. And this time we took the low complexity region of a protein that's very well characterized in the phase separation field.
• 104:22 - 104:30:  This protein is called FUS, which has nothing to do with Fus1; it stands for fused in sarcoma.
• 104:30 - 104:38:  And if you take just the low complexity region in vitro, it forms this de-mixed liquid droplets.
• 104:38 - 104:54:  And so we did the same experiment, replacing the disordered region with the low complexity of FUS. And remarkably, this worked absolutely perfectly. So it restored a fusion focus that has normal dimensions.
• 104:55 - 105:09:  And these cells were able to fuse to nearly 100% indistinguishable from wild type. And also we measured the kinetics of fusion and it happened at the same rate as the wild type.
• 105:09 - 105:22:  So this indeed confirms that just having self-interaction in this region, and probably weak binding, is what this region normally confers to Fus1.
• 105:23 - 105:41:  I'll show you just images here showing you that this replacement, indeed, is able to now have actin assembled at this location, similar to wild type and for comparison, these were the less functional CRY2 replacement and recruit the myosin like wild type.
• 105:41 - 105:45:  And these were the less functional CRY2.
• 105:45 - 106:00:  And the other experiment I want to show you is that we then took the opportunity that there's a lot of work done on this FUS protein, and in particular the alleles that have been described to have more liquid or more solid properties.
• 106:00 - 106:21:  And we used them to see whether this would modify the function. And what we found is that the more liquid version had no change; it worked just as well as the initial FUS low complexity we used, but when we used this mutant that has been shown to lead to enhanced
• 106:22 - 106:36:  gelation and solidification of these droplets, we found that even though the cells fused, still all of them managed to fuse, that they required almost twice as much time to do that.
• 106:36 - 106:48:  And so the process was not as efficient, suggesting again that that needs to be quite, not very tight for it to work.
• 106:49 - 107:12:  This leads me to my conclusion: we think now this actin fusion focus is a membrane-less organelle that is organized by Fus1 self-interaction, and that this needs to be of a quality that is permeable to various proteins that I've mentioned—assembly
• 107:12 - 107:25:  factors, of course localization factors that will link it to the right place of the plasma membrane. It needs to be permeable to vesicles transported by type 5 myosin.
• 107:25 - 107:33:  But it's sufficient to exclude ribosomes, and this is probably contributed a lot by the actin assembly itself.
• 107:34 - 107:55:  Our model is that the quality here needs to be, you know, fluid enough to be functional to allow the entry of all these factors and also the pheromone signaling machinery, and likely in the CRY2 situation, we made it too tight, too solid, and therefore,
• 107:55 - 108:04:  all of these, or these factors were not able to penetrate as efficiently, and this did not allow good enough fusion.
• 108:04 - 108:07:  Okay, so I'll stop there.
• 108:08 - 108:31:  I will finish by thanking all the people who have contributed. I've mentioned the work of Ingrid, Olivia, Valentin, Melvin, and of course, foundational work on this was done by Maya, Alex, and Laura, who all have left my lab, and I will take any questions.
• 108:31 - 108:35:  Thank you.
• 108:35 - 108:40:  Thank you very much, Sophie; that was absolutely fantastic.
• 108:40 - 108:57:  So if people want to put questions into the chat, I will invite you to talk, but as I'm chairing, I get to ask the first question. And so, I wonder this sort of disordered domain, have you tried popping that into other proteins? Is it sufficient to call self
• 108:57 - 109:03:  aggregation or does it require… maybe I missed that? Yeah, we haven't, we haven't tried that.
• 109:03 - 109:24:  We have done limited attempts to purify the N-terminus of first one, which so far was not very successful, and therefore did not encourage us much further, but we definitely have to do more in trying to go in this direction to understand really what happens.
• 109:24 - 109:40:  So, if we express just this region, it makes an unstable protein in the yeast cells, and therefore we've not been able to see whether it was by itself sufficient to form aggregates.
• 109:41 - 109:46:  I actually suspect that the very N-terminus also contributes.
• 109:46 - 109:48:  But, I, yeah, this…
• 109:48 - 109:50:  That's where we are.
• 109:50 - 109:56:  Okay, let's move on to Professor Olafranko, who has a question.
• 109:56 - 109:58:  Hello. Hello.
• 109:58 - 110:03:  That was lovely. So I was thinking about the architecture.
• 110:04 - 110:12:  that would demand such an architecture, why not concentrated at the plasma membrane?
• 110:12 - 110:24:  You know, I suppose what I don't understand is that, why do you need to make a ball in which movement of things will become more difficult, rather than make something linear.
• 110:25 - 110:38:  Do you know what I mean? Something, something quite obviously oriented. And also, I suppose another question is, how do you imagine, like in your model, you drew a cross-section of the ball.
• 110:38 - 110:43:  And you have actin inside. Is that what's happening? I mean, yes, actin outside; I mean, what is the architecture of that?
• 110:52 - 111:12:  Okay, so why questions are always difficult to answer. All I can say is, that's what we observe. And this is similar to, you know, the organization of the Spitzenkörper in filamentous fungi, which is this big assembly of vesicles that drives
• 111:12 - 111:15:  the growth of the hyphal tip.
• 111:15 - 111:28:  It's also similar to what happens in synapses, in neuronal synapses where there's a lot of presynaptic vesicles that typically accumulate.
• 111:28 - 111:38:  And there's some work that has shown that disordered proteins contribute to their hooking them together.
• 111:39 - 111:44:  I think one shouldn't view that as a solid ball.
• 111:44 - 111:54:  It's a very fluid, you know, things are coming in and out all the time. That's how I view it. Of course, it's very hard to visualize that.
• 111:55 - 112:10:  But the way that actin is in or how it's organized, definitely from the correlative light electron microscopy, when we've been able to see actin filaments, they permeate through; they're not on the edge.
• 112:11 - 112:26:  The only case where I think we've seen vesicles on the outside is when we express first one in mitotic cells, which I think is, you know, it doesn't make a physiological focus in that case.
• 112:28 - 112:34:  But if it is a porous ball, then why do ribosomes not go in? I mean, that's smaller than vesicles, right?
• 112:35 - 112:53:  Yeah, this is a good question. I think the reason actually for ribosomes to be excluded is probably mostly the actin. Similar to the actin patches or the lamellipodium, it's also excluding ribosomes; it's very dense in actin.
• 112:53 - 112:58:  But I think we have to do more work to really understand this.
• 112:58 - 113:07:  It's a good question also whether there's a purpose for excluding ribosomes. I don't know. I, at first glance, I don't think so but who knows.
• 113:07 - 113:12:  I have another question if there are no… there's no one ready to ask.
• 113:12 - 113:18:  I think Jamie's question, which is in the chat, may have just been answered, which was…
• 113:18 - 113:21:  Is that right, Jamie?
• 113:22 - 113:29:  Yeah, I was wondering about the size difference of vesicles and ribosomes and kind of exclusive entry and just your thoughts on that.
• 113:29 - 113:41:  Like, do you think this myosin-driven force is able to overcome the exclusive nature of this, or kind of how do you envision this happening on the force level?
• 113:42 - 113:52:  Yeah, I mean, I guess that the force conferred by myosin allows it to penetrate through.
• 113:53 - 114:04:  You know, we don't know exactly how the structure is made out there, like are there holes somewhere to allow the entry of vesicles to get in.
• 114:04 - 114:13:  It's an obvious question but we don't have an answer. Other than thinking the myosin provides force. Yeah, cool. Thank you.
• 114:13 - 114:27:  So, it's almost a technical question but kind of interesting, I think, how you said that you modulated the target pressure in H minus cells, and you got, you know, changes in the shape.
• 114:27 - 114:42:  How, how did you modulate target pressure? Well, we used the trick that Nicola Mark had used, which is to delete one of the glycerol biosynthetic enzymes.
• 114:42 - 114:50:  Okay, so it is a GPD1, but yeah, but then you don't really know because GPD1, you know, you'll probably have a whole energetic balance.
• 114:50 - 114:55:  You know, of course—of course it's very indirect. Okay, but it's GPD1. Okay.
• 114:55 - 114:58:  Thank you.
• 114:58 - 115:07:  So, I'll sort of ask another question if I may, which is, you talked earlier about the MAP kinase pathways; what other MAP kinases in this structure?
• 115:07 - 115:20:  So the good question; we haven't looked at others. We've looked only at the cascade, the pheromone signaling MAP kinase cascade.
• 115:20 - 115:34:  So, I wouldn't be surprised if those involved in the cell wall integrity pathway are also there, but we haven't looked. I mean, definitely Rho activity can be seen there.
• 115:34 - 115:39:  And I think the PKC kinases are there too.
• 115:39 - 115:54:  So one of the advantages of MAP kinases, right, is they seem to do so many different things, seemingly sometimes opposite things in opposite places so that, you know, kinases are one of these things where we do know a little bit about cell cycle kinases, for example,
• 115:54 - 116:15:  which can regulate things, but MAP kinases, it's still thought of just as cytoplasmic action. It's, people are still thinking about it like a test tube that there are substrates and kinases, and if you have these structures which could potentially demarcate essentially
• 116:16 - 116:33:  membrane-bound organelles, you could have kinases doing different things in different places. Yeah, I mean I, I'm quite convinced that the concentration of the MAP kinases in there has a functional output.
• 116:33 - 116:51:  And, you know, I don't know whether it's simply to put the kinase close to the substrates it needs to phosphorylate, or whether it also serves to concentrate the kinase beyond a certain point changes the threshold.
• 116:51 - 117:07:  So, for example, I mean, we don't know for the kinases but for the form itself. It's actually quite interesting; we had introduced point mutants that are very well characterized to, you know, knockout activity in vitro completely.
• 117:07 - 117:16:  And some of these point mutants were still able to fuse; no, not a hundred percent, but like 60 to 90%.
• 117:16 - 117:25:  And I think it's because they are present in such high concentrations in this structure that it is not the concentration that's been tested in vitro.
• 117:25 - 117:38:  But, you know, it works well enough. So I could well imagine that this leads to changes of concentration of kinase substrate and changes in what is phosphorylated.
• 117:38 - 117:51:  Can you get abhorrent fusion if you put these things in the wrong place? I'm thinking of putting it in the mid-zone or something like this, and because they're always at the tip, you rescue but it's fusion at the tip, correct?
• 117:52 - 118:10:  Yes, so we haven't tried that; the only thing we have done is at some point we created auto-crime cells, so we created cells that respond to their own pheromone by transplanting the receptor.
• 118:11 - 118:13:  And these cells…
• 118:13 - 118:16:  Well actually the H minus cell.
• 118:16 - 118:19:  When you do that in the H minus cell,
• 118:19 - 118:29:  they go wild, and they form a focus that kind of moves around and eventually the cell goes and lyses.
• 118:30 - 118:40:  Not at the—I mean not absolutely at the tip, somewhere. Okay, so that essentially it can happen so that it doesn't need to be at the tip.
• 118:40 - 118:51:  Yeah, I mean, the pairing of cells does not always happen at the tip either. They're much more flexible in where they can be polarized and then during my dirty work.
• 118:51 - 118:55:  So the contact is more important. Yeah. Yeah.
• 118:55 - 119:00:  Okay, fantastic. Are there any other questions? I'm looking in the chat.
• 119:00 - 119:02:  Don't see any.
• 119:02 - 119:17:  In which case, it just remains for me to thank all the speakers from today, both the invited talks and the breakout sessions. It was a super competitive set of abstracts we received this time and thanks very much to Sophie.
• 119:18 - 119:20:  It was a really nice talk.
• 119:20 - 119:42:  We need, as always, to thank Abcam; Abcam have sponsored this meeting since the very beginning, 11 years ago, and you might think that the organizers do some organizing but actually Cheska, Robin, and I do remarkably little and Lana and Polly and all the others,
• 119:42 - 119:50:  do a huge amount of work behind the scenes to organize this meeting. So, thanks to the events team for all their hard work.
• 119:50 - 119:54:  Rob, is there anything else I'm forgetting?
• 119:54 - 120:00:  There's just going to be a survey very quickly before it finishes, so please complete that before you leave.
• 120:00 - 120:12:  Yeah, and of course, thank you all for attending, and the next one is going to be in February or March next year and it's going to be virtual, and we'll let you know if you're on the mailing list, we'll let you know about that one.