 Well, good morning. Thank you, everyone, for joining us for this special event. I'm Dave Paglarini, and I have the pleasure of introducing Mike Viling today, who is here to defend his thesis in partial fulfillment of the requirements of earning a PhD in biochemistry at the University of Wisconsin-Madison. Mike is from outside of Boston, and he went to college at the University of Massachusetts Amherst, where he was really a star student. He did a lot of research in biochemistry and genetics and plants, earning him a Goldwater, a very Goldwater scholarship. So in 2013, he came to Wisconsin. And after rotations, we were lucky enough to recruit Mike into our laboratory. And since that time, I think it's fair to say Mike has changed the lab forever. He is, it's been, I would say, interesting, having Mike in the lab. I thought about this, and I was telling people, it's difficult to capture the Viling gestalt in a short introduction. He, Mike somehow, blends characteristics in a unique way, a combination of a considerable geekiness at times with a sweetness and a swagger, which is not an interesting combo. The man has survived primarily, as far as I've been able to observe, on peanut butter and jelly since he's arrived. And I mean, every day. It's not an exaggeration to say, thousands of peanut butter and jelly sandwiches have been consumed at our lunch table since I've known Mike. And it's really an art form. That's to give a seminar on the construction of a peanut butter and jelly sandwich. Maybe you will, but I don't know what you're going to talk about. Mike is not short on confidence or opinions. He recently explained to me that the reason for turning down a certain job offer was that, well, they kind of did stuff the way we do stuff. But he wanted to do something really cool. Fair enough. He's been overheard making fun of Zach Kimmerer, an accomplished athlete, for falling on his American Ninja Warrior competition. He's also claimed he could defeat Zach in a certain dance, dance, revolution competition. I don't even know what that really is. I would say that Mike is adorably dyslexic. Is that maybe a fair term? And when combined with his confidence, has ended up in some interesting situations, like Mike providing me a draft letter of recommendation for himself from me in glowing terms, but with my name badly misspelled. There have been times I will say when I've wanted to kill Mike. There was a period where emails from Mike to outside members of our community had to be vetted through me to avoid any further deterioration of collegial interactions. But I say all of this with a smile, because everyone here knows that Mike is a lab favorite, and everyone loves him. And he really has been a very important addition to our laboratory. He is undeniably sharp. Mike's got some real intellectual horsepower. In this business, though, it's not enough. Mike, though, is incredibly enthusiastic. And reliable. He gets things done, and he will surprise you with his talents and his efforts consistently. He's very generous with his time and talent. Everybody goes to Mike for certain things, and he's quick to help. And he's navigated some really powerful collaborations to some great ends. Mike has been funded by a competitive NIH training grant. He is an NSF graduate fellow. He will leave with three lead or co-lead publications in wonderful places. And additionally, I think it's worth noting that Mike has been very coachable. He is someone who has always taken constructive criticism and advice very well. And that's not always the case. And it bodes very well for his career, I believe. It's also worth mentioning that I think Mike has navigated the trials and tribulations of graduate school with a real grace and fortitude, especially because he's had some difficult moments. Mike lost his father early in graduate school. Mike has also navigated a long distance marriage to Macy who was here. So it's not, you know, these things are, grad school is hard enough. And Mike did that with and persisted admirably. And he is now heading back to Boston. He'll be joining Pam Silver's laboratory in the Department of Systems Biology at Harvard Medical School in the Dease Institute, where I'm sure he'll do really cool science and be closer to his family. But I'm glad we have one more chance to hear Mike deliver his work. And for me to say Mike, it has been a real honor and a privilege to have you in our lab. So take it away. All right, sound good. Everybody can hear me? All right, thank you so much, Dave, for that wonderful introduction. And I hope to tell you about some really cool science that I have done with Dave in the lab. And specifically I'm gonna tell you about how I used bioinformatics and biochemistry to assign function to uncharacterized mitochondrial proteins. And I'm gonna do that with two stories. First, I'm gonna tell you a story about how I used a protein-protein interaction network to identify a novel complex one assembly factor and how through a collaboration, we were able to link that to patients and show that mutations in that gene caused Lee syndrome. Second, I will switch ears and I will talk about a multi-only protease profiling study that we did and how it identified a new processing event required for coenzyme Q biosynthesis. So before I get into any of that, though, I wanna just start with a pretty broad background on everybody's favorite organelle, this kidney bean-shaped guy known as the mitochondria. Now, the mitochondria get their fame as the powerhouse of the cell. And this is it for good reason. Basically, ATP is produced in the mitochondria. Most of the ATP that we use is produced in the mitochondria. And it's produced through this big machine right here known as ATP synthase. Now, I'll be talking a lot about these complexes throughout my discussion as we did this big protein-protein interaction screen. And it's these big machines that really drive the function of the mitochondria. So for example, ATP synthase uses the protomotive force where you have a high concentration of protons outside of the mitochondria and a low concentration inside to drive this motor here that spins to produce ATP. Now, there are numerous other complexes in the mitochondria, notably this complex one right here, which is the largest of the respiratory complex members with about 45 different subunits. And basically, this machine right here takes reducing power in the form of NADH that is generated from the breakdown of food to pump protons across the membrane to set up this disequilibrium to produce ATP. There are other complexes that do the same thing, but this is the one that I will focus on in most detail. In addition, the mitochondria do a lot of other things, notably lipid biosynthesis. Here is a very important redox active lipid known as coenzyme Q. This lipid was identified right here, actually down the street a little bit at the enzyme institute. And it fulfilled a missing piece in the electron transport chain. For a while, we were wondering how do electrons pass through this chain? How do they pass through the membrane? And this lipid right here allowed us to really isolate complex one from the rest of the complexes and understand how the electrons got passed through one to three, two to three. And it's been widely appreciated as an important lipid for energy production. Now, in addition to energy production, mitochondria are a hub of numerous signals. So I was recent, well, not that recently anymore, but it was identified that mitochondria were a key cellular process known as apoptosis is controlled. Now, apoptosis is a key anti-cancer process, which basically allows cells to essentially commit suicide and stop as they're proliferating. So this is a very important function of the mitochondria. Additionally, there are numerous other processes that go into the mitochondria. There's a lot of heme biosynthesis here and many other processes. But I don't have time to go into detail about all of them. Now, in addition to all of these known proteins and known machines that do a known function, there are a lot of other machines that we know end up in the mitochondria, but we really don't have a good sense of what they're doing there. And we refer to these as mitochondrally uncharacterized proteins or MXPs. Now, interestingly, if there are mutations in some of these MXPs, even though we don't necessarily know where they're going, they can cause disease. They can cause what is known as inborn errors of metabolism or mitochondrial dysfunction when these machines are broken. So these diseases, as you can expect, are pretty broad spectrum. Every single cell in our body has mitochondria, so it makes sense that these diseases are pretty strong, pretty systemic. So, for example, mitochondria, mutations in mitochondrial genes end up leading to things like hypotonia at a very young age, where small children and babies kind of can't hold themselves up. If children are able to progress through that, they can lead to things like ataxia, where it's a neurological disorder and they're unable to coordinate their movements. And it's usually fatal at a pretty young age. So understanding diseases like this has a key interest in our lab. Additionally, if we look at patients that have inborn errors and metabolism like this, about 45% of the time, we can't even identify what gene was mutated that caused this dysfunction. Now this sets up a pretty big biomedical bottleneck. So if about half the time we can't even say what's wrong, how are we going to hope to treat that in the long run? So what we're trying to do in the Pagrini lab is understand uncharacterized mitochondrial proteins, look at patient populations like this with undiagnosed mitochondrial dysfunction, and just understand mitochondria more generally. So what I've done here is just drawn a simple diagram. So here is a pie chart representing every single protein that localizes to the mitochondria. And about 25% of them are these MXPs or uncharacterized mitochondrial proteins. At the same time, we can look at a lot of the pathways in the mitochondria and we can notice that there are certain gaps in our understanding. So what we're hoping to do through our work in the Pagrini lab is come up with some sort of systematic functional annotation to take these MXPs and slot them into where they go. So given that background, I'm gonna tell you two stories about how I helped to identify these mitochondrial uncharacterized proteins. And I'm gonna start with a story about a protein-protein interaction network that I did in collaboration with a lot of people. Notably, Brendan Floyd, who's a former MD-PhD iPib student here, was the real driver and the initiator of this project and really kind of helped push it through. In addition, we collaborated with two very talented mass spectrometrists in the Coon lab to do a lot of the mass spec analysis here. So how did this process work? Well, let me give you a little bit of background first. So as I mentioned in the beginning, there are lots of mitochondrial complexes and these complexes drive mitochondrial function. So this motivated our work for understanding the protein-protein interaction network in mitochondria because we knew that these complexes played important roles. So we wanted to see if we could discover new regulators or to maybe even discover totally new complexes. And this has been known for a very long time. This was a paper from right here in Madison back in the 1960s and we can even see a lot of these complexes by electron microscopy. As you can see the stem and bud structure of ATP synthase back in 1964. So basically what we did for this analysis is Brendan went through one of Dave's seminal papers where he essentially listed every single protein in the mitochondria and tried to identify key targets that were ripe for characterization. Through this process he identified 50 of these MXPs that we targeted for in IP mass spec type of analysis to identify what proteins interact with it. In addition to 27 sort of well characterized proteins that we could see if anything else interacted with. Now both of those played an important role as I will show a little later. Then we worked with the center for eukaryotic structural genomics to clone all targets into an expression vector. We then expressed them in HEC 293 and HEPG2 cells and specifically we noticed that all of the targets that we tried they properly expressed and they properly localized to the mitochondria. Then we took all of those samples and we did an immunoprecipitation to kind of pull out the protein that we were interested in. Then we took our samples over to the Coon Lab and they did a lot of really great mass spec analysis to identify which proteins were interacting. So when the data came back they looked something like this. Now this is a pretty complicated graph so let me just kind of walk through it. What I'm drawing here is along the x-axis is every single mitochondrial bait protein that I used. So just like fishing, the bait is the thing that you throw in to see what comes along. So this would be the targets that we studied, one of those 77 targets. Now on y-axis here, these are all the prey proteins. These are the things that we pulled back out of the cell and theoretically interact with our bait. And what you'll notice is that along the top here there are a lot of these really yellow or very strong interactors that seem to kind of interact with everything no matter what bait you throw into there. Now these are things that are generally sticky like heat shock proteins and they kind of interact with everything. So the specific interaction might not necessarily be that meaningful. Now at the bottom you can see more specific interactions. Now these are things that maybe only come down with one or two of the baits. Now these are probably, we're more excited about these because they're not just sort of generally sticky. But we needed a way to sort of do this systematically so that we could tackle this kind of middle ground where there are things that are maybe somewhat specific but maybe there's a really strong interaction with one. So we went back into literature. Oh, sorry, let me just kind of demonstrate this in a slightly different way. So just to make this very clear, what you can do is you can take a particular column and you can look at all of the different interactions that a particular bait protein has. So for example the C17 protein, when we throw it into the cell we notice over a thousand different interactors. Now we can't kind of go through those individually and test them. So we needed a way to really enrich our data for the biologically relevant interactions. And the way we did that was with this WD scoring algorithm which is from Wade Harper and Steve Gigi's group at Harvard. And basically what our goal here is to identify, obviously we want to keep the specific prey. Those are good interactors that we strongly believe in. And we kind of want to get rid of these non-specific prey that kind of come down all the time. But then in the middle there are some cases that are a little bit more complex. So basically we use this algorithm right here, this equation right here to convert our raw data into a more clean form where we took into account the specificity of the interaction, the reproducibility of the interaction and the intensity of the particular interaction. So given all of these data, we could eliminate the non-specific prey. We could enrich for specific targets in the kind of gray area. And obviously we could maintain our specific prey. So if we were to apply this to our entire data set, what you can see is that it significantly cleans up the data. And additionally, if we're looking at our individual protein right here, C17, you'll notice that we went from over 1,000 interactions to just 10. And I'm gonna talk in a little bit more detail about one of those interactions between C17 and this protein known as NDUF AF5. So let me just pull that out right here. So NDUF basically means that this is a subunit of complex one. NDUF stands for NADH dehydrogenase ubiquinone alpha because F is alpha, obviously. And it's assembly factor number five. So what this means is that it's important in supporting complex one, but it is not actually a subunit of mitochondrial complex one. Instead it is an assembly factor. And we noticed that it sort of reciprocally interacted with the C17 or 59 protein. And at the time, really that's all we knew about it. We knew that it was on chromosome 17 and we knew that it was open reading frame number 89. But seeing that they interacted, this sort of brought up a pretty simple hypothesis. Is C17 a new complex one assembly factor as NDUF AF5 was? So to test that, we started with a simple Western blot. Basically, we just took TEC-293 cells and we knocked down C17, as well as many other known complex one assembly factors. And we just blotted for complex one. And as you would expect, if we knocked down a complex one assembly factor, we lose the levels of complex one. We reduce how much complex one is available. Interestingly, C17 also shows that same effect. So this suggests that C17 might be a complex one assembly factor. Interestingly, we can look at AF5, the interacting partner, and we can notice obviously when you knock down AF5, you lose AF5. But interestingly, if you knock down C17, you also lose AF5. So what this suggests is that C17 is required for the stability of AF5. And that may be the mechanism by which C17 supports complex one activity. We can then take a look at the other complexes and you'll notice no real difference across the board. This is consistent with a defect in specifically complex one and not sort of a generic general mitochondrial deficiency. So okay, another thing that we can do to test this hypothesis a little further is we can look at complex one activity. So what you notice is that in C17 knockdown lines, you have complex one activity proportional to how much you've knocked it down. This sort of makes sense because we see less complex one in the cell lines where we knock down C17 and in your controls, you don't knock it down. Now, we can then take these cell lines and we can rescue them by re-expressing a healthy version of C17 in the knockdown cell lines to increase our complex one activity back up. So given all of these data, we published our first paper and we actually renamed C17 to NDUFAF8. So this is AFA, assembly factor number eight. But just for consistency, I will continue to call it C17 through the remainder of the talk. So this was pretty cool. It suggested that we had a new complex one assembly factor. But remember that the real reason we're interested in a lot of these assembly factors is because mutations in the assembly factors have a tendency to cause mitochondrial deficiency. Now, remember that about 45% of patients that come into the clinic with such a deficiency don't have an identified gene mutation. So seeing that we identified a new complex one assembly factor, it's possible that maybe some of this 45% of the pie, there are patients there that have mutations in this gene. So we just wanted to ask a simple question. Is this protein important in humans? And to answer this question, we collaborated with two excellent scientists over at the Welcome Trust Center for mitochondrial research in Newcastle in the UK, Rob Taylor and his student Charlotte, who have access to patient samples from this sort of 45% of the pie of unidentified mitochondrial deficiencies. And through some detailed searching, they were able to identify three patients that all seem to have mutations in C17 and also had a specific complex one defect. So what you'll notice is that two of the patients have passed away. And you can even, you can see the particular mutations. This patient had a homozygous inherited mutation of just a single point mutation that changed the phenylalanine to a leucine there. This particular patient had a start codon mutation that basically blocked all protein production. But interestingly, there was also the centronic mutation. The centronic mutation was replicated in a different line as well. And this is sort of strange because we typically wouldn't think of an entronic mutation as necessarily causative for complex one deficiency. But because of the work that we did, we looked into this particular mutation in more detail and we've been able to show that this mutation blocks mRNA production. Now this last patient right here, you'll notice has this blocking mRNA production, but also has a 10 base pair insertion that leads to a frame shift and an early stop codon. So this patient is essentially a knockout, but luckily was able to survive and has been willing to work with us. And he actually provided us with a skin sample or fiber blasts. And our collaborators then passed that off to me. And we performed a similar experiment as we performed in the HEC-293 cells by providing his fiber blasts with a healthy version of C17. So what we did is we either provided an empty vector or we provided a healthy version of C17. And what you'll notice is that in his primary fiber blast cell line that when you provide an empty vector, AF5 isn't really present at all. But really giving back C17 causes AF5 to be re-expressed or re-stabilized. Now we can also look at complex one. Complex one seems to be stabilized. And the remainder of the complexes are largely unaffected as we saw previously in the HEC-293 cells. We can then finally end this with a look at complex one activity. And what you can see is that complex one activity is significantly increased when C17 is reintroduced. So these data suggest that mutations in C17 underlie causes or underlying causes of complex one deficiency in human cells. And we've identified several specific mutations that could potentially be screened in future patients. So this is, I think, a really cool story. And I'm gonna now transition to the second half of my talk where I will tell you another really cool story about how I used a multi-omic protease profiling technique to identify a novel processing event required for coenzyme Q biosynthesis. So the way that this project worked is there's a lot of different proteases in the mitochondria that I'm showing by these little Pac-Man guys. These proteases have a number of functions, sort of traditionally they're thought of as just proteins that degrade unfolded or unused proteins, but they also have important regulatory roles as well. So the story that I will tell shows that a particular processing event is required to stabilize and activate protein involved in coenzyme Q biosynthesis. But in order to get to that point, we needed to understand what these individual, what these proteases were doing and how perturbing them affected the yeast system as a whole. So essentially, we played the role of this little ghosty guy right here and we went and we deleted each Pac-Man individually to try to understand what was going on in the yeast. And I say, we generously, it's really more the work of this graduate student right here, Andrew Reidenbach, who graduated recently and has moved on to Boston as well. So how did he go about doing this? Well, first of all, he took a good hard look at the yeast mitochondria and he identified 19 proteases that would be good to study. So basically all of these, most of these localized to the yeast mitochondria and they all play kind of different roles exemplified by their different localization because of the different substrates that they have access to. Now, after identifying all of these, we went back into the yeast deletion strain, the yeast deletion strain library and we pulled out a single yeast deletion. So basically every single one of those proteases was deleted in a different strain. So it's not a strain that had all of them deleted. These are all different strains that all had an individual mitochondrial protease deleted. We also had a wild type strain as well as sort of a bench, a sort of a standard for our data. Then what we did is we took these strains and we grew them in two different conditions. First, we grew them in a fermentative condition where we provided the yeast with an excess of glucose and this excess of glucose essentially means that they can just use glycolysis in order to produce all of the ATP that they need. So they can kind of grow on their merry way even if their mitochondria are messed up. Second, we wanted to compare this condition to a condition that really required them to use their mitochondria, this respiration addition. Here, we provide them with just 0.1% glucose, a little bit of glucose to get them started, but then we provide them also with 3% glycerol which allows them to enter the second phase if their mitochondria are working properly. And this allows us to really see a key moment as the yeast sort of transitioned from fermentative to respirative group. And then we took these yeast and we performed proteomics, metabolomics and lipidomics in collaboration with the Coon lab to identify systematically what is going on when you're deleting these individual proteases. And as you might expect, because there's a lot of data, the data can be somewhat complicated to analyze. So I just wanna quickly go through a key tool that we use to look at the data. This is known as a volcano plot. And in a volcano plot, we're looking at the levels of a particular protein, lipid or metabolite in a deletion strain, say PIM1, relative to wild type. So if you've got a point over here, that means that particular molecule goes up, but if you're seeing a point over here, that molecule goes down. We then plot that against the p-value of that change. So the more significant, the higher up. And then we can plot our data and they kind of look like a volcano spewing out. So this is why we call it a volcano plot. And you'll notice that when you delete PIM1, there are a lot of changes that happen in the mitochondrial proteome. But specifically, I'd just like to draw your attention to two molecules, ISU1 and ISA1. So these are both iron sulfur scaffolding proteins and you'll notice that they go up. Now this makes sense because ISU1 is a known substrate of PIM1. So if we delete PIM1, then we can no longer turn over ISU1 so its levels go way up in the Delta PIM1 strain. All of that seems to make sense. Now ISA1 is a protein involved in a similar pathway but wasn't shown to be necessarily a substrate of PIM1. So all of this suggests that hey, maybe ISA1 is a new substrate. But we can do a little bit better than that. We can push the data a little bit further and we can ask another question. Is this specific? So remember, we're just deleting a protease. So maybe deleting any protease is good enough to cause an increase in ISU1. So what we can do is we can take the volcano plot and we can kind of flip it on its head. And we can do what is known as an outlier analysis and we can see that here where instead of plotting all the different molecules for a particular deletion strain, I'm plotting just ISU1 across all the protease deletion strains. And you'll see that when you delete a mitochondrial protease you do get maybe a slight increase in ISU1 but that's nowhere near as strong as the increase that you see specifically with PIM1. And the same is true for ISA1. So this suggests that this relationship is specific. And all of you can do that with your favorite molecules as well because we've made our data publicly available as I will discuss a little later. Now, what we wanted to do was we wanted to study something a little bit more novel than this, right? So this is something that is well characterized and well known. So in the Pagliani lab, we're very interested in Cohen's MQ biosynthesis. So just to remind you a little bit, Cohen's MQ is a redox active lipid that is involved in the electron transport chain that generates the ATP that our body uses. Now, without this lipid right here, you can't continue the path of electrons from complex one and two to three. So essentially this lipid is required in order for a respiratory growth or a growth requiring mitochondrial energy production. So because we're interested in Cohen's MQ and its biosynthesis, we just wanted to take our multiomic profiles and look at the biosynthesis pathway. And so what you can see is that the biosynthesis of Cohen's MQ essentially has two major precursors. First is tyrosine. This tyrosine acts as the head group precursor. It's modified to this 4-HB molecule where it's then brought into the mitochondria. Then you have this long isoprene tail that is added onto the end. So the isoprene tail comes from the mevalonate pathway, which is primarily known to produce cholesterol, but some of those isoprene units are brought into the mitochondria, polymerized, and then attached to this head group. Then the head group is modified by many different enzymes to lead to your final Cohen's MQ here. So if there was a defect in the biosynthesis pathway in any of our protease strains, we would expect to see a reduction in the full CoQ. And that is exactly what we see. So what you'll notice is that if we delete a mitochondrial protease, yeah, okay, you lose a little bit of Q, which kind of makes sense, but there's a much stronger effect in specifically the delta-octon strain. So this suggests a relationship between octon and Cohen's MQ. We can push this again a little further. We can look at a metabolite precursor of Cohen's MQ. And you can notice that that goes way up. This is also consistent with the defect in the biosynthesis pathway. We're essentially blocking off here, and then we get a buildup earlier in the pathway. We can look at a different lipid, also consistent, and that seems to build up. And then interestingly, we can also look at this particular complex, which is the complex that helps to modify the head group and a particular protein here, CoQ5, that protein seems to go way down. So there seems to be a destabilization of at least CoQ5. Now, this is cool, but at this point, we're really wondering what is octon and why does knocking it out affect Q? So to answer that question, let me sort of just go through quickly what octon does. Octon is a protease involved in mitochondrial import of proteins into mitochondria. Now, basically, most mitochondrial proteins are synthesized in the cytoplasm, and they have this mitochondrial targeting sequence that is essentially a zip code saying bring me to the mitochondria. This targeting sequence is recognized and the protein is brought in to the mitochondria. And then typically at this point, the targeting sequence is unnecessary. So there's a complex of two proteins that essentially cleaves off this mitochondrial targeting sequence, and then for the most part, proteins go on their merry way and they're ready to their functional. However, in a small subset of cases, 14 and now 15 cases, proteins require a second step of processing where you need to remove eight residues from the end terminus, and then you have a fully functional protein. So even this, we just wanted to ask a simple question, can octon act on any of the proteins known to be involved in CoQ? So we just did a quick literature search and we came up with a list of all of the known substrates of octon. And you'll notice that they play a lot of different rules. So we've got some mitochondrial ribosome proteins. We've got iron sulfur cluster. We've got complex two. It seems like there's a lot of very core functionality that is processed by octon. But none of these really explain the specific CoQ defect that we see. So what we wanted to do was we wanted to take advantage of our deletion strain to identify new substrates of octon, maybe involved in CoQ. So as you can imagine, in a wild type strain, an octon substrate would be fully processed and you'd get a protein looking like this. However, if we genetically deleted octon, you would end up with the buildup of a protein that had eight more residues on the end terminus there, this octopeptide. So what we can do is we can look for this by expressing CoQ one through 10, the known proteins involved in cube biosynthesis, in either wild type yeast or in yeast lacking octon. We can do an immunoprecipitation to pull out the protein, we can run it on a gel, and then we can cut out the bands and we can do a really cool old school technique, notice Edmund degradation, to sequence the end terminus and to determine if there's a difference between the two proteins, purified either wild type or delta octon. So when we do this, what we notice is a specific difference in CoQ five when purified from wild type or delta octon. So in a wild type background, you get an end terminus that starts right here, KEE. However, in the delta octon one strain, you get an end terminus starting here, FTQAH. And what you'll notice is that if you just take those peptides and you align them, that the delta octon one has a protein sequence that is exactly eight residues longer. And the protein sequence seems to align with the previously identified substrates of octon with this key arginine right at this location, important for our recognition and cleavage by the MPP to remove that mitochondrial targeting sequence. And then also you got this key phenylalanine that's then recognized by octon to remove the octopeptide. So okay, these data suggests that CoQ five is a substrate of octon, but we're not really convinced at this point because we wanted to know that this processing event was important for CoN's MQ biosynthesis because you can imagine, this is just a eight residue extension on the end terminus. How important can that really be? So what we did is we found a mutation in CoQ five that specifically blocked the processing of octon of only CoQ five. And I'll skip over a lot of details about how we picked that one, but essentially we took phenylalanine and we knocked it back to an alanine. And what that does is it doesn't really affect localization of the protein very much. It also doesn't affect processing by MPP very much, but it does strongly block the processing by octon. So next we wanted to see how this particular mutation affects yeast growth. So let me just sort of remind everybody how yeast grow, at least when in our particular conditions we provide the yeast with 0.1% glucose which they burn through relatively quickly in this first fermentative phase of growth. And essentially remember in the fermentative phase they don't require their mitochondria to grow. And then second they have to hit this point known as the dioxic shift where they switch their metabolism and then they start to grow using glycerol as the primary carbon source. So if we start up to lead CoQ five, you'll notice that you kind of get a flat line in the second phase of growth. This is because they have non-functional mitochondria and they can't use glycerol to produce energy. Now interestingly, if we then rescue this with a version of CoQ five, that is unprocessable by oct one, you might see a little bit of respiratory growth, but overall it seems like there's a strong defect in respiratory growth when you have an unprocessable version of CoQ five. So we can then look at levels of coenzyme Q to really show that this is due to, this causes a Q defect and you can see that you get significantly less coenzyme Q when you have a version of CoQ five that cannot be processed. So the final question that we have here is why, right? So remember, this is just eight residues hanging off the internus, maybe not that important, but it seems to cause a big defect, right? Why is that? So if you look into the literature, what you'll find is the general answer to this question is well, proteins that are oct one substrates are not as stable when they're not processed and we can show that actually using what I refer to as a poor man's pulse chase. Basically we just take yeast and we treat the yeast with cyclohexamide that blocks protein synthesis and we can then watch the, we can watch our target CoQ five degrade over time and then we can calculate a half life just based on a quantification from western one. And you can see that in a wild type background, you get a half life of about 300 minutes. Now, if you delete oct one, you get a much faster half life of about 100 minutes. So this suggests that yes, when CoQ five can't be processed, you get a much faster turnover and therefore you end up degrading the protein more quickly. And we can even see this in our study state levels that the level of CoQ five in the delta oct one strain goes way down, remember from the math spec data. So all of this suggests that essentially with this oct peptide on the end, the protein is less stable and therefore CoQ five can't do its function and more or less produce coenzyme Q. So this was I think a pretty cool story, but in addition to this cool story, there are several other possible cool stories that could come out from our data. So I talked about the PEM one thing at the beginning. We also noticed interesting connections between this PRD one strain and metal ion homeostasis proteins. We noticed a connection between Wyoming one and complex four subunits and complex four assembly factors. So there's a lot of other stories that could potentially come out of our data. And so in order to make sure that future scientists can access our work, we generated a website. And again, I use we pretty generously. This was really the work of a talented graduate student from the Coon lab known as Nick Quichen. He really set up a lot of beautiful websites over there. And basically to just run through a simple demonstration of how something like this works. Obviously, if you just wanna take our data and use Python analyze it yourself, you can do that. Or you can use some of the tools that I've shown you so far. So for example, we have this outlier analysis suite where you can take your favorite molecule and you can see if there are any pretty solution strains where your molecule is strongly perturbed. So our favorite molecule is coenzyme Q. And you can see that specifically coenzyme Q goes way down in only this Delta Oculin strain. The rest of them do have an effect, but not as significant. Additionally, if you're interested in a specific protease, you can then just look at a volcano plot of sort of all of the data. And each of the points you can just kind of mouse over to know what's going on with that particular point. And then finally, if you're just kind of more generically interested in a molecule and how it changes across the deletion strains, you can enter it in here in a bar chart form. So Coach U5, for example, and you can see that it goes down again specifically in the octone strain. So feel free to take a look at the website if you're interested. And obviously you can ask me questions and we'd be happy to work with you if you find anything interesting. So with that, I'd just like to recap the stories that I told today. First, I told you about a protein-protein interaction network that identified a new complex bone assembly factor. Second, I told you a story about a multnomic profile that identified a key processing event in Coen's MQ biosynthesis. So with that, I would have a lot of people to thank for all of the work that I've done. First of all, I would like to thank my group of friends that I made here in Wisconsin from my class. We had this sort of tradition of going to Buffalo Wild Wings on Thursdays. So it's been great getting to know you guys and really having some other scientists to talk to about all of the work that we're doing here. I would also like to thank my friends back home, some of whom are watching on the live stream, who have really stuck with me despite the distance. And we often play games as a nice relaxing time. We also have a tradition of going to First Night in Boston every year. So this was, I think, just from this last year here. I would also like to thank the lab as a whole who are really supporting my work and being a great lab environment to work in. So one thing that I would say that I really enjoyed about working with Dave and being in Dave's lab is seeing the development all the way from when I started until now. So when I started, Dave didn't have tenure yet. And it was a very different place back then. I still remember Amelia right here, her defense. She was Dave's first graduate student. And through the years we've developed, we've grown and we've changed. We moved across the street and many new people have come, many very talented people have moved on with their life. And I'm just very happy with the way that everything has gone. I'd also like to thank my mom for all of her support. She's been very strong all the way from her baby picture up there to coming to Madison. This is a picture of us and Dottie's dumpling stourie here. And also her sister Kathy as well, my aunt Kathy, has been pretty supportive. I'd also like to thank my dad for the support that he's given me and the support that he gave my mom throughout the time that he was able to support her. He was really the one who pushed me to go as far as I did with my education. And even when I first moved here to Madison, this is a picture of us back in the moving truck. It was a pretty powerful day when I left the nest for the first time, really. Finally, I would like to thank my wife, Macy, for being with me despite the distance all the way to Michigan. Here's us on our graduation day from UMass when we were starting to part ways. She was going to Michigan. I was coming here to Wisconsin. But despite that, we eventually got married and we've had lots of good times together since then. So thank you for being with me and thank you for supporting me through all of the years here. And with that, I'll stop. I'd also like to thank my committee and all of my funding sources, all my collaborators and everybody who's helped me along the way. So thank you very much and I'll take any questions. Yeah, so I'll just have to say I'm not very familiar with the way that complex one deficiency is generally treated. I know that there is some really exciting work out of Dave's old lab where they've shown that maybe reducing oxygen can actually help with complex one deficiency or well with mitochondrial deficiency in general. I think generally complex one deficiency is something that's pretty difficult to treat and is generally fatal at a pretty young age. I know that, I mean, in terms of helping that particular individual, there's sort of limited options that we have. I mean, really the only thing that I feel like I can truly offer here was peace of mind. We now know sort of what's wrong. And the other thing is, I mean, this is also a tangential benefit. If that family wanted to have more children, then we now know sort of the mutations that need to be avoided in those children, so, you know. So that particular protease, yeah. So there was a study that came out during my PhD that showed that the human equivalent of OCT1, this MyPAP protein, there are mutations in that protein that underlie mitochondrial deficiency. Now we weren't really able to necessarily show that sort of this was a QD fact and they didn't really look for QD facts in that particular individual. But it's possible. We did just kind of a brief study for one of the response to reviewers where we noticed that knocking down, MyPAP didn't seem to affect CoQ, but it's a little bit hard to say because MyPAP is a really important protease and it's sort of, we didn't know if that was really a specific effect or if that was just sort of a general effect of making mitochondria sick at that point, so. But there's a vast disparity in the attention given to individual genes. Some things we know pretty well and things not at all. Even for an organism like yeast. So many proteins, these functions are not known. Why are there so many uncharacterized functions? Is it hard biology? Is it, what do you think? Yeah, I would say that science is hard and because science is hard, a lot of times it's easier to characterize something that's already known. If once you kind of have a grab on something, once you kind of know what's going on with it, the science becomes a lot easier. I think your work with the ADCK proteins has really demonstrated that, because at the beginning, we really didn't know what any of those proteins were doing and now we've got better tools to understand what they're doing and hopefully that will improve our understanding going forward. I really think that it's sort of a little bit of safety. It's safer to publish on something that we know a little bit more about than it is to really venture into the totally unknown areas. So that's why in Mortgridge, we do fearless science, so. You know. We like to make fun of this. Yeah. But even for a C17 or a C89, never published, right? So now we know it's part of this really complicated bacteria get away with 14 subunits and it's hardly any of these. Why is it so complicated? So why have humans complicated things in the process of making complex one? I don't know if I have a great answer for that. I mean, obviously we are multicellular. We're large and there's a lot of more coordination. Maybe that needs to take place for humans versus bacteria, but yeah, it's sort of curious because it does seem like sort of the core function of complex one can be simplified down to just 14 subunits instead of the 45 that humans have. So why have all of those things been tacked on? I'm not sure about that, really. Yeah, yeah. Let's see. So yes, yes I did that. I'm not gonna pull up the slide right now, but basically what I noticed is that when you do that, you lose mitochondrial localization. So essentially deleting that octopeptide leads to loss of mitochondrial localization for somewhat unknown reason, but essentially what that says is that this octopeptide has something to do with localization. So in addition to that, I sort of replaced that targeting sequence with a generic targeting sequence that'll get it to the mitochondria whether the octopeptide is there or not. Now that targeting sequence leaves three residues post cleavage, right? So I tried to also delete those three residues. And what I noticed is that the protein got to the mitochondria still, but processing didn't happen anymore. So it seems like basically MPP leaves a little bit of schmutz on the end of your protein, right? So what you need is you kinda need an adapter. If you wanna have a particular end terminus as well as mitochondrial localization, so you gotta have that octopeptide there. I'm not the first person to come up with that idea, but it's sort of a hard thing to prove solidly, right?