 All right, so I think I'll go ahead and get started. So thank you all for being here today for the second lecture in the Steenbach lecture series. So just to remind everyone, this lectureship was made possible through the work of Harry Steenbach and a gift from his wife by Evelyn Steenbach to the university to endow this lectureship. And for those of you who attended yesterday's lectures on radical chemistry in ribonucleotide reductase, I have an interesting development, and that's that's gone so well that the Madison Radicals, which is a professional Frisbee team here in Madison, has offered to recruit Joanne, so they sent over a Madison Radicals Frisbee T-shirt for Joanne, so she may be coming back to Madison as a professional Frisbee player pretty soon. So with that, I'll turn over the stage to Joanne. Thank you. All right, can everybody hear me? Okay, whoops, that's not the first slide. Okay, so today I'm gonna continue in another sub-ery or ribonucleotide reductases that we worked on, and specifically what I'm gonna talk to you about today is identification of the long-proposed but elusive Dynemanganese tyrosyl radical cofactor in the class one B-ribonucleotide reductases, and at the end of my talk, I'm going to try to link the discovery of this cofactor to pathogenesis of many bacteria. So the problem is outlined on the first slide. We can start out with a protein that has no metals in it, and the question is, how do you put a metal cluster into the 1A enzyme that has a di-ferric tyrosyl radical cofactor, which is essential for conversion of nucleotide to deoxynucleotide, and in the case of the 1B enzyme, which is the focus of today's talk, you're gonna have a dimanganese in both in the plus three oxidation state with a tyrosyl radical cofactor. So in my opinion, a major unsolved problem as biology is how you biosynthesize metallocofactors, and it turns out 30 to 35% of all of our proteins have metals, and metals such as redox active metals simply don't float around in the cell because there's this issue of generating what biologists call reactive oxygen species. And so we believe that there are biosynthetic pathways involved in the generation of these cofactors, and in many cases these cofactors can be damaged, and we believe that there are biosynthetic pathways involved in the repair of these cofactors. So today what I'm gonna do is talk about the 1B enzyme, but the 1A and the 1B enzymes are structurally homologous to each other. They basically have identical coordination around the metal site. So the question is how do you avoid mismetallation inside the cell? So if you put manganese into the 1A enzyme, the enzyme is inactive. So that's what I'm gonna focus on today, what we've learned about this system, and I'm going to, after I give you a little bit of background, I'll then divide the talk into three parts, which I will outline where we're gonna be going. So this is the system, those of you who were in the audience yesterday will recall that we've been working on for many years with the arrival of nucleotide reductases. These enzymes catalyze the conversion of nucleotides to deoxynucleotides in all organisms. It's the only way to make deoxynucleotides to NOVO. Consequently, this enzyme is largely responsible for supplying the Mymeric Building Blocks for DNA replication and repair, and also controlling the relative ratio and absolute amounts of the deoxynucleotide pools, which are essential for the fidelity of these processes. And again, targeting this enzyme has been highly successful to develop new cancer chemotherapeutics because of the central role in nucleic acid metabolism. So as I showed you yesterday, the enzyme catalyzes what appears to be a very simple reaction, Cleveland of a carbon hydroxyl bond, formation of a carbon hydrogen bond. This is the reduction reactions. There are two cysteines in the active site that supply the reducing equivalents form a disulfide. And to get multiple turnovers, one has enzymatic systems to re-reduce the disulfide back down to the two files. And in the class Ia enzymes, they use enzymes called thyrodoxins, thyrodoxin reductases, where the ultimate source of electrons come from NADPH. But in the class Ib enzymes that I'm going to be talking about today, they use a cofactor called neuroH and we'll come back to that later on. So these reactions involve complex free radical chemistry and the process is initiated by transient formation of a bio radical, which then removes a hydrogen atom from the substrate to trigger off this complex reduction process. So as I told you yesterday, and we discussed in some detail, there are many mechanisms by which reductases can catalyze the oxidation of an SH to a thio radical, which is essential for nucleotide reductase. And today, nucleotide reduction. And today what I'm going to specifically focus on is the class I enzymes. And yesterday I didn't tell you, although it was on the slide, that these enzymes have now been subdivided based on the metallocofactors. So the one we were talking about yesterday was the dipheric tyrosyl radical cofactor. And you will remember hopefully that I told you that the function of this cofactor is to oxidize a tyrosine to a tyrosyl radical. And this tyrosyl radical, which in bacterial systems has lifetime of four days, is essential for catalysis. If you reduce it, you lose all catalytic activity. So how this cluster catalyzes this oxidation is important to understand. Now within the last five years, two additional metallocofactors have been discovered. Katruver in my lab for the first time identified what had been postulated for many years by biologists, a dimanganese tyrosyl radical cofactor. These two proteins are structurally homologous to each other that even contain identical coordination environments around their metal. And so why has this cofactor been so elusive? And it turns out there's a biosynthetic pathway that plays a key role in putting the manganese two into the right oxidation state. And that's what I'm gonna be focusing on today. Now in fact, in the last five years, an additional ribonucleotide reductase has been discovered and these studies have been done in vitro and in vivo. This has only been done in vitro. But what you see in this cofactor system, and this is found in a human intracellular parasite, you don't have a tyrosyl radical, but you have a phenylalanine. And what you see the active oxidant of assisting to a thiol radical is, is a manganese four iron three. That kind of a cofactor had never been observed before. So this immediately raises the issue. If you were inside the cell, how do you put the manganese in the iron into the appropriate site to form the active cofactor? And that's what I'm gonna really be addressing today. Okay, so here's the issue we're gonna be talking about. And these are the only acronyms you need to remember in my talk. Nured B is the beta two subunit that we talked about yesterday, which has two irons. And the function of these irons is to oxidize in some way. And I'll show you what our model is, the tyrosine to this essential tyrosyl radical. The class one B, which is Nured F. And so this will be the major focus of today's talk. Can assemble both a manganese cluster and can self-assemble, we'll talk a little bit about that, a dye iron cluster that's identical to the 1A. And it turns out, I will show you that both the manganese cluster and the iron cluster can be active. So the question is, which one is found in vivo? And is the manganese cluster really active in vivo? And is it involved in keeping the organisms that have 1B alive? And again, the 1C, which I'm not gonna talk about further, has a manganese and an iron present. Okay, so the system we've worked on two systems, actually three systems for many years. We work on organisms where we can do simple genetics to study these systems. So we work on E. coli, we work on S. cerevisiae. And more recently, our findings in S. cerevisiae have led us into homologs and humans. Today I'm gonna focus on the bacteria and I'm gonna focus initially on the bacteria E. coli and then I'm gonna branch into additional bacteria. So what do we know about E. coli and what our studies on assembly of the dye iron tyrosyl radical cofactor in the 1A enzyme was the motivating factor that allowed us to discover the 1B enzyme. So the 1A enzyme has this cofactor and we were asking the question, where does the iron come from? Okay, so you can do things like knock out five iron two transporters that E. coli gets and ask the self the question, can you see iron move around in the cell? And when we did that, what we realized is that this enzyme was no longer being expressed. But what was being expressed in E. coli is the 1B enzyme. And the 1B enzyme, we still don't really understand that much about the regulation but it's expressed under limiting iron and under conditions of oxidative stress, okay? Now I wanna make another point. Most microorganisms do not contain a 1A enzyme at all and the 1A and E. coli is involved in DNA replication. Most organisms only contain a 1B enzyme and that is the enzyme involved in DNA replication. In addition, one also has to have, and this will become important at the very end of my talk, a class III enzyme that I talked about last time which can only be expressed under completely anaerobic conditions and uses a glycol radical for catalysis. So today what I'm gonna do is focus on the 1B enzyme and I'm gonna give you a little bit of background of how we got involved in chasing around this 1B system by giving you some background on the 1A system, okay? So the 1A protein, you express it, it can come out APO. The question is, and this is an issue that I think makes studying cofactor assembly of many different kinds of metallocofactors challenging is that maybe as remains of the prebiotic world, many of these clusters can self-assemble, okay? So that always give you some kind of background. Now the efficiency of the self-assembly is highly variable, okay? But the 1A enzyme self-assembles actually very efficiently and you can simply throw in two iron twos and oxygen is gonna turn out to be the oxidant in this reaction. And the tyrosine 122 is oxidized to a tyrosyl radical. So the oxygen then can pick up two electrons from iron two oxidation to iron three. One electron and lots of a proton to oxidize tyrosine to a tyrosyl radical and it needs an additional electron to do a four electron reduction of oxygen to water. So this is the stoichiometry of the reaction. The self-assembly process was originally identified in 1974 by Atkins Laboratory. And after we did the stoichiometry and many laboratories have contributed over the years to our understanding of how this assembly works, we have the following model. So this is complicated, but it's not really that complicated. And it sets the stage for the kinds of experiments I wanna talk about today. Okay, so where are we? We have apoprotein, that's the nerd B protein. Somehow we have to deliver iron two to the protein. Our hypothesis is that iron two comes from an iron sulfur cluster. So we have iron two into the protein. We then need as an oxidant, and this is gonna be distinct from the one B oxygen gas. And so oxygen gas can bind to the iron two cluster. You dump two electrons onto the oxygen to form an iron peroxide. We then showed using a lot of time-resolved physical biochemical experiments, which I'm not gonna discuss today at all, that one needs to do reductive cleavage of the OO bond. And this is catalyzed by a tryptophan oxidation to a tryptophan radical. You probably don't remember it yesterday, but that tryptophan radical was in the pathway that I talked about yesterday. And we can show that this happens in a kinetically competent fashion. So what you generate is this intermediate X, which we identified many years ago, which is an iron three, iron four species. And in fact, this was the first non-hemion four species that had been identified in biology. Now, to get recycling, you wanna have a way of re-reducing the tryptophan cation radical back down to a tryptophan. And there's a protein which we identified that's found in the operon of what most 1A ribonucleotide reductases that codes for a two iron, two sulfur cluster ferrodoxin. Those are proteins that donate one electron at a time by doing redox chemistry with the iron sulfur cluster. So this gets us back and resets the rheostat. And you get to this intermediate three four and you're gonna see three fours running throughout my talk. People haven't seen these much in biology, but this is the oxidant that converts tyrosine into a tyrosyl radical in the active dipheric cluster. And to do this oxidation, you need to use proton coupled electron transfer, not the long non-range I talked about yesterday, but here we have it constrained in a little area. Okay, so we then get to active protein. Now we're ready to do nucleotide reduction. But we know, and many of you in the laboratory have worked on cell cycle regulation have probably stopped cell cycle by inhibiting DNA replication, throwing in something like hydroxyurea. Hydroxyurea catalyzes the reduction of the tyrosyl radical back to tyrosine. And the tyrosine is completely inactive because you need the tyrosyl radical to initiate phyl radical formation. And so the question is, is this the end of it? Or is there a way you can reactivate your iron cluster to reform an active tyrosyl radical? So what we believe happens is the same YFAAE, which is a one electron reductant, can donate an electron to reduce this iron three to iron two, this iron three to iron two, iron two, it's recycled by reductase, regenerating the reduced state and then the reaction recycles. And so this served as a model for us thinking about, and we're still working on this very actively in both bacteria and yeast, for thinking about how this might happen inside the cell and provided a framework for what I'm gonna tell you about today, okay? So I'm now gonna focus the rest of my talk on the one B enzyme giving with the information I just gave you as being the background. So if you look in all the operons and bacteria that contain one B enzymes, what you see is nerdy and nerdeff, okay? And that's the alpha and the beta subunit is required for nucleotide reduction. We're gonna be focusing on nerdeff, which has the metal. What you also see is nerdy H. In many cases, nerdy H is a thyroid oxin-like molecule that can recycle your di-cell-phide that forms to re-reduce to get multiple turnovers. And the main focus of today's talk really is gonna be this protein nerdy. And nerdy at the time we began working on this was annotated as a flavodoxin, okay? So when we thought about this in E. coli, the proteins expressed under iron-limiting conditions, you might not have a two-iron, two-sulfur cluster ferrodoxin to donate the electrons required for cluster assembly. And in fact, we initially thought this flavodoxin might also serve as a one-electron reductin avoiding the problem of needing iron, okay? And I'll tell you that that turns out to be wrong, but that was our working hypothesis. So in the first part of my talk, I'm gonna tell you about the discovery of the dimanganese tyrosyl radical cofactor in its unusual properties that we've determined crystallographically. And then in the second part of my talk, I'm gonna focus on studies. We've just very recently published on the B-subtlis enzyme, which is better behaved than the E. coli enzyme and has allowed us to study how cluster assembly actually works and tells us what the oxidant is gonna be to convert the manganese two-manganese two into the active cofactor. It's not oxygen. And at the very end of my talk, I'm gonna come back and talk about streptococcal sanguinis, which is the causative agent of infective endocarditis. I collaborate with a dentist and I will show you some animal studies that make us think that this dimanganese cofactor plays an important role in keeping pathogenic bacteria alive. Okay, so what I wanna do now is start and look at the flavodoxin. Okay, so we know a lot, not really a lot, but we know something about flavodoxins and generic flavodoxins can take the reduced state the hydroquinones, so you'll see this before. This is colorless. You can do a one electron oxidation to form the semiquinone state. It's EPR active. It's colored with absorption at 610. You can do another one electron oxidation which gets us into the really colored state of a flavin in the oxidized state. And if you look at the redox potentials of normal flavodoxins, these two half reactions are quite distinct. Okay, so the first thing we did was try to look at the redox potentials of our nerd eye and what we found was they were very different from the redox potentials of all the reported flavodoxins. So what you see is the two half reactions have about the same redox potential. So that immediately suggested to me that this nerd eye wasn't a one electron donor. It suggested to me that this might be the long elusive biosynthetic pathway member that is gonna supply the oxidant to help assemble a diamanganese cluster that people hadn't been able to find in the past 30 years. So the other thing that you need to remember is that you can do one electron reduction, one electron oxidation and oxygen can get reduced to superoxide. That could be the oxidant. You could do two rapid one electron oxidations to supply hydrogen peroxide. That might be the oxidant. And we're gonna address that question in the second part, the middle of my talk today. Okay, so I just wanna say a few things about the nerd eyes. One can make a phylogenetic tree and what we found with the E. coli enzyme which is the one I'm gonna be talking about initially. When we started looking at nerd eye and nerd F, again, that's the one that has the metallocofactor. We could co-purify them so that the two proteins have affinity for each other. The KD is less than 0.05 micromole and that's our lower limit of detection. If you turn to the B. subtilis enzyme which I'll be talking about in the middle of my talk, you can see the KD is quite distinct. It's 10-fold higher. And we think we understand that. The 50s loop which is over the flavin which is going to interact with the nerd F protein is distinct in all these systems. And then if you go to the S. sanguina spectral system, the KD is even higher. So that causes different kinds of issues we have to deal with in setting the biochemistry of these systems. So nerd eye can co-purify with nerd F in the case of E. coli and they really have, despite the fact that it's a small little protein, they have a lot of distinctions between branches of this tree. Okay, so this is the first experiment that Trover in my lab did now in 2009. And we were incredibly excited by this result. So all we're doing is we're taking nerd F, dumping in manganese too, okay? We do this in an anaerobic box. We make the flavin, nerd eye in the hydroquinone form, the completely reduced form, which likes to get rapidly oxidized. So it needs to be in an anaerobic box. You mix the things together, you take them out of the box, you let oxygen in, you take a spectrum. And the spectrum we saw is this big black thing. And what we were excited about is this guy, okay? So only somebody that works on tyrosyl radicals would get excited about that little blip. Why? Because that's the hallmark of a tyrosyl radical. So most of the absorption you see is reduced flavin, rapidly gets oxidized to oxidized flavin, which has a huge extinction coefficient is bright yellow. And so if you subtract the dotted line, which is the oxidized flavin, what you get left is this guy, the sharp feature at 410 with shoulder 390, which is the hallmark of a tyrosyl radical. In addition, we see broad sort of featureless region, which is consistent with some kind of an oxidized manganese state in our cofactors. So we thought we had discovered for the first time this protein that people have been chasing around. And in fact, we had. So the one issue we had with this system, and I'm gonna tell you a little bit more about this system, but then move on, is that in all reductases, we in general see one tyrosyl radical per homodimer, per beta 2, and here we only saw 0.25. So if you wanna study the chemistry of what's going on in these reactions, you've got an issue in terms of sensitivity. And so we still haven't studied this reaction in any chemical detail. Okay, so what do we do in this system? We set up a collaboration immediately with Amy Boll, who just took a faculty job at Penn State. And Amy Rosenzweig, who I've collaborated with many times over the years. And we were able to get a structure of the neurodepth protein with both manganese two and iron two loaded, okay? And I wanna just point out a few differences between these systems. So in the case of the iron two system, you see that the tyrosine is hydrogen bonded to an aspartate, which is a ligand to this iron, which we'll call iron one, Mrs. Iron Two, and they're always organized in the same way in all my slides. So we have a hydrogen bonding interaction. What we will also see, and this will be a focus, and this is the key to catalysis in almost all these dye iron clusters, is this particular glutamate is by dentate to the iron two. And you have no waters bound to the iron twos, that that's what we observe. Okay, if you look at the diamanganese cluster, it's really strikingly different. Again, they have the same, it's the same protein. So now what you can see is this aspartate is directly hydrogen bonded to the tyrosine that's gonna be oxidized to a tyrosyl radical. In addition, this residue, this carboxylate is the same carboxylate that's located specifically to the iron two is now by dentate. We probably have 50 structures of dye iron clusters and no one has ever seen this coordination before. And again, the essence of catalysis and looking at metal cluster doing chemistry is these ligand reorganizations, which are extremely challenging to look at if you don't have good crystallography. In addition, what I wanna point out is that you have two waters. Okay, and so I'm gonna show you that the key to delivering the oxidant is going to be this metal at the two site and the place where whatever the oxidant is would come in from this direction. We're gonna look at this in a minute. But in assembly of the dye iron cluster and this enzyme can self-assemble a dye iron cluster to give you tyrosyl radical, the oxygen comes in from this backside. Oxygen is very hydrophobic and so you have hydrophobic tunnels which allow things to come in and out of the active site. And I'm gonna show you that there is a tunnel that comes into this active site that's much more hydrophilic suggesting a difference in the oxidants. So one question we asked and many people had tried to put manganese into the class one A enzyme and oxidize it to a tyrosyl radical. Many people had tried to do this in the one B enzyme and they were all unsuccessful. So we repeated those early experiments. The way we did the experiments, the only difference between our experiments and the earlier experiments is that we had nerd eye in there but we had nerd eye in the oxidized state so can't do any redox chemistry. Thinking maybe it's a conformational it somehow changes the conformation to allow the chemistry to happen. So we have manganese two loaded protein, nerd eye in the oxidized state and then what we did is added the oxidant. So one oxidant you might try is superoxide. So we generated superoxide enzymatically with xanthine, xanthine, xanthine oxidase whether the nerd eye was present or not we got no cluster assembly, no tyrosyl radical. We then did an experiment with hydrogen peroxide. Again, plus or minus the oxidized form of the flavor protein. Again, we got no cluster assembly. So that's basically what people had reported before. We also tried hypochloric acid. Again, we got no cluster assembly. So then we did yet an additional experiment. What we did was we put in the manganese two. We now use the reduced form of the flavored oxen so it can generate whatever the oxidant is. And then in addition to that we had a protein and solution called superoxide dismutase. So that would take superoxide and dismutate it rapidly. So if nerd eye produced the oxidant and solution it would scarf it up the way we designed the experiment. Or we had catalase in there. Catalase works on hydrogen peroxide. So again, it would scarf up hydrogen peroxide. But despite having large amounts of these two proteins around we can still assemble the cluster very efficiently. So that suggested to us that this interaction between nerd eye and nerd F was critical and that what was happening is we never are going to lose whatever the oxidant is into solution. What's going to channel between the two proteins into the metal two site. So before I go on and show you the data that supports that hypothesis, let me also say, and I think this is from a chemical point of view this is extremely interesting. You can self assemble. You don't need anything else in there except iron two, oxygen as the oxidant and a source of electrons. And iron can function as a source of electrons. And you can make 0.7 tyrosyl radicals per beta two when you have activity. So the dye iron tyrosyl radical can be made without any protein factors just like the 1A enzyme can be made. And the manganese enzyme can also be made and we haven't optimized any of these conditions but what you see here we have 0.2 to 0.3 radicals per beta two and you can see the activity is much higher. And I'm going to show you, we've now looked at many proteins. The activity of the manganese is always five to 10 fold higher than the activity of the iron loaded protein. Okay, so we were extremely lucky and I think the goddess of crystallography was shining down on us in the case of this problem. We can crystallize these proteins in a period of two hours and get very high diffracting resolution structures. And what you're looking at here in green and blue is the dye manganese cluster. And what you're looking at here in yellow and pink is the flavin. So this is the flavin backbone and here's the manganese cluster. Okay, so we form a one to one complex with the flavin, with the nerd eye and the nerd F in the manganese two state. So what we then did was we were able to co-crystallize in an anaerobic box the reduced state, the nerd eye state. And we can look at both the reduced and the oxidized state and this is just a cartoon summary of what we reported now about a year and a half ago. So here's the flavin and this is the nerd eye protein. Here's the manganese two. Here's that unusual carboxylate that was bridging both manganeses. And what do you see from the flavin into the active site? You see a string of waters. So that is very unusual to have an oxidant that is in a water channel that can access the manganese two site. And that's suggested to us that we might have something like superoxide, hydrogen peroxide, peroxide, radical that could function as the oxidant. So if you look at this a little closer, one can see and use one of these programs that allows you to look at tunnels. What you can see is that we have a channel directly from the flavin into the active site to form the oxidant. So we have been able to tell with this particular enzyme that we can generate an active dimanganese tyrosyl radical cofactor and we believe that the function of the nerd eye is to deliver the oxidant directly into the active site. Okay, so what I told you in that case is the amount of radical we saw was low. And if you wanna do time-resolved physical methods, you need to have as much sensitivity as you can get. So we turned to additional organisms for two reasons. Number one, we wanted to know, is the 1B an E. coli, which isn't the major reductase an E. coli, is the 1B and all other organisms also a dimanganese cluster. And it turns out, B. subtilists only have the 1B. That's their only ribonucleotide reductase. And so it turns out that we have been able to clone and overproduce all the enzymes in the operon. So nerd E and nerd F are the reductases. Nerd I, again, is the flavodox and we'll be talking about this. Another open reading frame that's essential. We don't know what its function is. And so what I wanna do now is discuss with you how we've used this system to be able to watch and determine number one, what is the oxidant of manganese two to the oxidized state that then oxidizes tyrosine to the tyrosine radical. And can we see any novel intermediates in this overall process? Okay, so again, in collaboration with Bolan Rosenzweig, we've been able to get a structure of the B. subtilist nerd F. Over in this slide, you're looking at the E. coli nerd F I showed you before. Here's that unusual carboxylate in the water channel. The flavin is sitting over here. But you can see in the case of the B. subtilist enzyme, this carboxylate is no longer liganded between the two metals, but is in a different position. It's in the position we normally see that carboxylate. Furthermore, what one sees is one doesn't see a water channel all the way into the active site. And in this region where we have a methionine versus an isoleucine, we have a distinct number of waters, which is probably accounting for the differences between the two systems. So the question is, is what we learned from the B. subtilist enzyme going to be extrapolatable to the E. coli enzyme? I don't know, I believe that for you to judge. But what I'm going to be talking about now for the rest of the time is the B, for the next few minutes is the B. subtilist system. So again, here is our system where we have this carboxylate with E. coli and the waters. In this system, that same carboxylate, sorry, the same carboxylate is over here on this iron. And again, we have the differences in coordination between this system and the tyrosyl radical in these two clusters. So the proteins are slightly different, at least in this crystal form. Okay, so I know most of you don't think very much about mechanisms. So what I'm going to do is walk you through the mechanism and I will then give you a few pieces of data to support the mechanism. So let me say at the outset, it's complicated. And why is it complicated? Because nobody has ever studied a flavodoxin in the reduced state with oxygen. And any of you who know anything about flavin chemistry, you can do a lot of things. So you're going to have to trust me. The paper is online in Jack's if you want to read the 150 pages of supplementary information that says we carried out all the controls correctly. Okay, so let me just walk you through the model. Okay, so our goal is to start out with two manganese twos in these sites and the hydroquinone form of the flavodoxin. And in the end, what we want to be able to do is generate this dimanganese three tyrosyl radical cofactor. That's our goal. So the nerd eye hydroquinone binds. It can channel the oxidant. It reacts with oxygen gas to form the oxidant. And I'm going to show you we believe the oxidant is superoxide. Okay, that's distinct from the model we proposed a couple of years ago in the literature. So down that channel, we believe superoxide is going to tunnel in. Why is that interesting? It's interesting because in the iron cluster, we know we need oxygen and an electron. Here we put the two together and avoided a lot of issues by directly channeling superoxide into the active site. We then proposed the kinds of chemistry that we would thought would happen in this reaction and let me also point out here that the rate constant for semiquinone formation, which we can see by visible spectroscopy is 40 to 50 per second, so that's pretty fast. We then wanted to think about what could be the intermediates along this reaction pathway and measure the rate constants for the formation. And the next rate constant we could measure is around two to three per second. We don't see the manganese two, manganese three per oxide intermediate. So we believe it's formed but that it's rapidly converted into a manganese three, manganese four site. And therefore the rate limiting step is a conformational change. Is that surprising? Maybe not because again, where is the water and what do you have to do to the metal site to get the oxidant in the right region to be able to reorganize itself to do the chemistry. So we believe we form a manganese three, four site. I'm gonna show you that that turns out to be the case. We believe that the semiquinone formed over here can then react additionally. Maybe it falls off the enzyme, maybe it doesn't. We don't really know. And generate more superoxide and generate oxidized flavin. Again, oxidized flavin is bright yellow so that clouds up your spectrum. So it's something we have to contend with. And so what's particularly appealing to my lab is that this three, four species, we believe is the oxidant that converts tyrosine to a tyrosyl radical by proton coupled electron transfer. This is strikingly similar to the iron three, iron four species we identified some 20 years ago that is the key oxidant in tyrosyl radical formation in the class one B enzyme. So how do you study this? So there are two ways you can study this. All of these things happen on the millisecond time scale. So you have two kinds of probes. I'm gonna show you that they undergo color changes. And the other kind of probe is the paramagnetic so you can use EPR. If you look at these cartoons of the colors that we can be looking at, I think here is the flavin, what the flavin looks like. So this bright, this feature here, this big black feature you saw before, it's bright yellow. That's an oxidized flavin. The reduced flavin in the 400 region has no absorption. It's non-planar. And in this region out at 610, which is where we're gonna monitor everything, is where the semi-quinone form, indicative of superoxide formation absorbs. You can see this region, it's a mess. And that's where a lot of things absorb like a manganese cluster is absorbed. So it will be really challenging to see something in that region, okay? If you look at the nerd F in its active form, here you see again the spectacular spectrum of a tyrosyl radical with a sharp feature at 410 in a shoulder. And once also sees this blobby region, which is indicative of a manganese-free system. And we and other labs have now convincingly demonstrated that that's the oxidation state. Okay. So let me just give you a few pieces of data. Hopefully you can follow what I'm telling you as I picked out. I've simplified everything, so only focus on things that I think are hopefully easy to understand. Okay, so the first thing we do is we take the flavidoxin in the reduced state and add oxygen. So you can watch semi-quinone form. And so that's in black. You see semi-quinone form. We then repeat the experiment by putting in the nerd F. Remember it can form a complex with nerd I. It's in the APO form. And what you see is this red curve. Okay, so it's faster. Semi-quinone formation is faster. Rate of disappearance is, it turns out oxidation to the oxidized flavin is the same. Now when you put in the manganese two and the nerd F, so that's the active form we believe that's involved in the channeling. Now what you see is more rapid formation where the rate is increased to a factor of 25 fold. And then at the end what you see is the semi-quinone disappear where it's oxidized to the complete flavin. Okay, so the experiments we're doing are simple, but complicated. And the reason they're complicated is because the interaction between the two proteins and the B-subtleis enzyme only has a Kd of 0.6 micro molar. Okay, so you don't want nerd I floating around in solution that can react with oxygen because it will overwhelm everything in your setup. So the way we did these experiments is we use the manganese two form in excess over the reduced form to avoid cluttering up our spectra with flavents which have high extinction coefficients. So we mix these things in syringe. The second syringe has oxygen and buffer. You mix them, you watch on a millisecond timescale either by UV visible or I'll show you by EPR by freeze quenching EPR and taking doing time courses. Okay, so here's a spectrum. Remember now, remember that this process isn't so complicated. You have something that happens 40 to 50 per second. You have something else that happens that only two to three per second. So they're different by a factor of 10. It's almost like you're looking at different reactions. And then you have something that's still slower, 0.1 to 0.2 per second. So you have a chance of seeing intermediates build up because of the kinetics of this system. So if you look from zero to 83 milliseconds, what you see is this feature at 610 grows in. That's a semi-quinone. That's indicative of superoxide formation. Then if you look from 83 milliseconds to three seconds, that's sort of the second regime, you see the semi-quinone goes away and you see flavin oxidized growing in. Okay, and I'll show you the model again in a minute. And you see at the very end of the spectrum, you see this little tiny blip. That's the tyrosyl radical I showed you at the very beginning. So now if you go down between three seconds and 60 seconds, what you see is you can now see the tyrosyl radical grow in. And furthermore, this region, which is a mess up here, is starting to clean up. And this is what we think we're looking at the manganese three, four oxidizing the tyrosine to a tyrosyl radical. Okay, so those are the spectra. Here's an experiment where we combine the EPR and the visible spectrum. This is the best fit we've gotten because of complications I'm not going into. But if you fit the rate constants for the visible spectra, you see fast formation of the semi-quinone indicative of superoxide formation and then slow disappearance of the semi-quinone at 0.8 per second, which is indicative of oxidizing the semi-quinone flavin to the oxidized flavin. Okay, so let me say that again. The fast form, 40 to 50 per second, we see the semi-quinone, which we believe is indicative of superoxide formation. And over here, we have the semi-quinone. We believe that it falls off nerd F and can further be oxidized to the flavin oxidized, which gives us that second rate constant. That's the interpretation of our data. Okay, so now what I want to do is look at EPR, okay? And EPR is also complicated in this system, so I'll point out why and then I'll show you what we've concluded from our data. The manganese two, manganese two site is EPR active, okay? Manganese two, manganese three is potentially EPR active. Manganese three, manganese four is potentially EPR active. Semi-quinone is EPR active of the flavin, but the relaxation properties are very different between the metals in the organic radical. The tyrosyl radical is EPR active. So this makes interpretation of the data complex. So this is just a spectrum, this is for George. We collaborate with Dave Ritzlab at UC Davis. We still haven't been able to simulate the spectra of these things. This is the E. coli system of the manganese two, manganese two, and here is the B. subtilis enzyme which always has a little bit of manganese two in solution. It's distinct between these two systems. So we have a spectrum and why is this important? It's important because the way we design our experiment is we have an excess of the manganese two, manganese two protein over the flavidoxin protein so that we don't worry about the flavin. That means we're always gonna have manganese two, manganese two in there that we need to subtract. So being able to simulate these spectra are important and we haven't succeeded in doing that very well. On the other hand, because the kinetics are, we lucked out with the kinetics, they were simple enough to look at this. What we have observed in the second regime and the regime of two, two, three per second is this feature. In this feature, if you look in the region of organic radicals is where the tyrosyl radical and the semiquinone radical absorbs, but you can see in the low field and the high field sides you have features which have allowed us to actually quantitate production of this new intermediate and we're convinced that we don't have a manganese two, manganese three species that's produced during this process. So what is this the spectrum of? And people have been extremely interested in manganese three, manganese four species. Why? Because all of the oxygen in the world is produced by a manganese machine, a four manganese cluster with calcium that involves three, four oxidation states. So there are huge numbers of model systems where people make the three, four states. And also, it turns out to be a manganese, dimanganese catalase system where people put it in an artificial three, four state. And if you look at the spectra, they look convincingly similar to one another. And so our interpretation is that this is a three, four manganese intermediate. And so one can look at the kinetics of formation by EPR and disappearance and you can fit the rate constants. This is three per second. And then it disappears at 0.12 per second. And this rate constant would, disappearance of the three, four would be correlated with tyrosyl radical oxidation. So we look at 410, which is where the tyrosyl radical absorbs. The flavin also absorbs there, okay? So the first part of our fit, so it's multi-phasic, biphasic. The first part of the fit, which is 0.8 per second, I showed you before, that's the oxidized flavin being formed. And the slower phase is almost the same as this phase, suggesting that the manganese three, four is generating the tyrosyl radical during this transformation. So this is our working hypothesis now. We are convinced from this data and we convinced the reviewers this was true, that the function, at least in the B-subtle-less case of the flavidoxin is to generate this reactive oxygen species, superoxide, which is delivered directly into the active site. We do chemistry, which we would love to know, but we think it's associated with conformational changes. That's not unusual because the essence of catalysis and diametral clusters is ligand reorganization. We generate a three, four species, which we believe we can see. No one has ever seen a three, four active species in a protein before. And it's kinetically competent to oxidize tyrosine to a tyrosyl radical. So I think what's particularly interesting about this is remember in the very beginning, I told you it was 1A, 1B, 1C, iron, iron, manganese, manganese, iron, manganese, they all have very similar centers. And what we now think from looking at all of these things in details I'm not gonna go through today, is that in all cases, to go here from the tyrosine to the tyrosyl radical, the electron comes in to the iron two. Here, the electron, or the oxidant comes in and delivers the reducing equivalents to the metal two, the manganese two. And in this case, in the 1C, you go from a four, four to a four, three. Or again, the active species gets reduced in the metal center. So the question then is raised, and this is what I wanna talk about for the last few minutes, is how do you control ventilation of these systems? Okay, so first of all, I've done all this stuff in vitro. Is it really true in vivo? Do you have any manganese clusters in vivo? And so now from three different organisms, my lab and about the same time, the Oling and Lubitz lab have purified this protein from endogenous sources. So in the case of CO-oneogenes, the enzyme was overproduced in this organism about 70-fold, and they were able to pull it out. In the case of the E. coli system, we didn't overproduce it at all. We had to start with five kilograms of protein to get out a tiny amount to be able to get a spectrum. In the case of B. subtilis, we integrated the whole operon into the genome. It was overproduced based on antibody analysis, maybe 15-fold. And in all cases, this is at 77K. You can see the spectrum looks pretty much the same, and it doesn't look anything like an iron tyrosyl radical. So basically we believe from these studies that the 1B inside the cells of these organisms, we think it's gonna be inside the cells of all 1B enzymes, is a manganese, dimanganese cluster. Okay, so now, in the last few minutes, I'm gonna turn to the kinds of studies we've tried to do to answer the following questions. So what I've told you is that we can assemble in vitro a dimanganese cluster. We can self-assemble in vitro a di-iron cluster. The question is, which one happens in vivo and can they both happen in vivo? And could, under the circumstances in which this protein finds itself, maybe kept alive a little bit by the iron cluster, and then when the circums growth conditions change, they can replace the iron cluster with the manganese cluster. And so you really need to go to an animal model to be able to look at something like that. And to me, the more burning question is, where does the manganese and the iron come from? So if you put manganese into the 1A enzyme, it's inactive. Okay, you cannot oxidize it with nerd eye to the tyrosyl radical. So if you have manganese in balance, you're in trouble. Okay, so how do you control all of that? So this is data from my lab, the top three in, very recently from the Schuyberg lab, a lot of people have now jumped into this area again. And what I wanted to show you is, we can measure the activity of the iron in the manganese-loaded protein. We've spent a lot of time looking at endogenous reductants in the case of these two organisms, B-suttleus and S-sanguinus, and you can see the difference is in activity of five to 10-fold, okay? Any coli, we haven't optimized anything, it's still five-fold. In the case of B-anthracis, the difference in activity is 10-fold, but they really haven't optimized to figure out how to assemble a cluster yet. Okay, so in the last few slides, I wanna present to you our recent studies on this organism, streptococcal sanguinus. And it turns out, as soon as our paper became impressed with manganese, I was called by a dentist. And this dentist works at the University of Virginia Commonwealth, where he has been interested in streptococcal sanguinus as the causative ages of effective endocarditis. Many times, if you've ever had surgery in your mouth, what happens is you take antibiotics to avoid colonization of these bacteria in the heart, which result in so many people dying per year. I don't know what the number is. But this kitten, I'll show you a picture of him at the very end, had sequence the whole genome, knocked out all the genes, was looking for factors that called virulence. And the one factor that he found that was far more important than any other was a manganese transporter. And so then the question is, why would this organism need manganese? Of course, everybody thinks of manganese superoxide dismutase. But he did a bunch of experiments that suggested that that wasn't the interesting reagent. So he immediately called us when he saw a manganese ribonucleotide reductase. And that's how we get into this project. Again, here's the NERD, the NERD-F. Here's the NERD-H, which catalyzes nucleotide reduction. And the NERD-I part of this protein is found. There were three NERD-Is annotated. We cloned and overproduced all of them, characterized all of them. Only one of them is essential. And that essential one is the one that forms the dimanganese tyrosyl radical cofactor. In addition, I think we spent a lot of time on this little protein, thinking that it might be a manganese chaperone that could deliver manganese into the active site. Okay, so we've done just a couple of simple experiments thus far. This organism, Esanguinis, can only grow into anaerobic conditions, only grow into anaerobic conditions if you take out the 1B enzyme because it needs to have a ribonucleotide reductase to stay alive. That's the class three enzyme. So if you grow the organism and then put it into oxygen, and then you take the wild-type strain and knock out the NERD gene, you can keep it alive by growing it under anaerobic conditions. And then you can ask the question, if you grow this organism in the presence of iron, for example, can you put iron into the NERDF enough to keep this organism alive under oxygen conditions? And the answer is, under these sets of growth conditions, we see no growth when we knock out the NERDi and the wild-type or the wild-type that's growing in the presence of iron. And we've recently done experiments to look in the animal model that Kitten uses, which for effective endocarditis, you catheterize the rabbit. You can probably see the catheter there. And then what they do is they typically inject the wild-type strain, which has some kind of resistance marker in it. And then they do the knockout and use another resistance marker in the strain. You mix them in equal amounts. You inject them into the heart valve. You sacrifice the animal and you plate the systems. And the bottom line is the enzyme. You cannot replace it with iron to keep the organism alive. It's completely lethal when you knock out the NERDi. So this is our preliminary data. And I think it's very interesting. What it suggests to me is that pathogenic organisms may have in fact evolved away to selectively put manganese into their active site. How do they do that? Where does iron come from? We've been working on this for 10 years. So have many other people in iron sulfur clusters and heme clusters. Does it come from the transporter? Does it come from an iron storage protein? Does it come from a label iron pool? How do you control all of this? Pretty darn important to avoid oxidative stress. The answer is we don't know. Okay, so we're pretty excited about the fact that this manganese system appears to be quite important. And I think it suggests, given the success in targeting reductases for many different diseases, we think it's possible that we could target assembly of this cluster to fight some of these pathogenic organisms. So let me close by thanking the people that have done this work. Really, this project was done by Joey. He's now in Chris Chang's work. I mean, this is conceptually all him. This is when being at MIT makes it spectacular. I mean, he designed all the experiments, executed all the experiments. I hope I taught him how to write a little bit better. But the fact is that he's done spectacular science. Amy Rosenswag and Amy Bohl. Amy Rosenswag is at Northwestern. I've collaborated with her for many years. Did all the x-ray structures I showed you. Amy Bohl is a new faculty member at Penn State. Manganese EPR is something I have not become intellectually engaged in. It's way too complicated for me. And so we have engaged a collaborator, Dave Britt, who spent his whole life working on the oxygen-evolving cluster and he along with a few other people in the world know more about manganese than anybody else in the world. And this is Todd Kitten, who I haven't ever met, but he's the dentist that called me up and got me excited about thinking about something practical finally in my life. So I'd like to thank the NIH for supporting all this work and thank you for your attention. Yeah. Well, okay, so we really haven't studied that. In the iron systems, for example, the E. coli enzyme, the half-life is four days and E. coli and pseudomonas, it's 10 minutes and humans, it's 20 minutes at 37 degrees. So it's highly variable dependent on how open the active site of the beta-2 subunit is. In this case, all I can tell you is, you know, we can freeze and thaw this many times and get the same specific activity back, but we really haven't studied that yet. We've been focused on trying to understand how the cluster was assembled then and looking at a few in vivo things to see if this was really interesting. So I don't know, but it's stable. I mean, we're not talking about seconds at all, we're talking about hours. Yeah. That was pretty, yeah, that's pretty stable. I mean, if you look at the rate of, rate constant for disappearance, it's 0.12 per second. So it's not as hot and oxidant, for example, as iron III, iron IV, and intermediate X and the 1A assembly. And it's certainly not as hot as iron IV, iron IV, right? So, but if you leave, you know, we've had, in the iron III, iron IV system, we've had samples in Brian Hoffman's liquid nitrogen chamber for 15 years. And we can start seeing radical, we can start seeing disappearance of the iron III, iron IV, and we can see radicals. What happens is, depending on the protein, you get electron transfer. If you have a hot oxidant, depending on the pathways, and you start generating radicals all over your protein, like I talked about yesterday, you want to avoid, that nature figures out how to avoid that. But that's what happens if you make a hot oxidant. And in this case, we don't know. Yeah. Yeah, George. What did you say, I can't hear you? Oh, I see. Yeah, so the difference in concentrations of, you know, manganese between the heart and the saliva in your mouth is, I don't know, it's 10 to the fourth or something in terms of concentrations. So that's another reason why they think the manganese is interesting. But, you know, the streptococcal pneumoniae, which is a pathogenic agent that's probably much more prevalent than streptococcus endocarditis also have the same issue, and they don't colonize in the same places. So I think nature figures out we had to colonize to survive. And I don't know why this organism goes to the heart, but that's where it finds itself. Yeah. Yeah. Why did you go one at the same time? I don't know, I'm not God. You know, iron is much more prevalent than manganese and iron, I guess I don't really know, but I think, depending on how this evolved, so in the anaerobic world, I think you had a lot of iron II around. Perry and George have written an article on this, which how many billions of years ago when you were anaerobic, we used to use iron II for everything, and then we moved into the oxygen-dependent world. There was a lot of soluble iron around, and that's maybe when you changed from the earlier reductases into the class I reductases. I don't know what the availability of manganese is, and I can't remember the solubility products, but I can tell you where you can find the solubility products. So, you know, the issue with manganese, I think, is it can substitute for magnesium and a lot of enzymes, but there aren't that many enzymes where it's been really clearly demonstrated that manganese is essential in catalytic activity, and part of that is challenging because the manganese II doesn't bind that tightly. Okay, and that's the issue with iron as well. It doesn't bind that tightly. So, figuring out what the correct medallation state is in these systems is really challenging. Yeah, yeah, Brian? No, no. So, you know, so, no, I would say, you know, there are reports of chaperone proteins in the human and yeast systems that by Carolyn Philpott, I don't know whether I believe it or not. I think, you know, our hypothesis is we publish a paper, we collaborate with a yeast geneticist, Lil, in Germany, and we have data that, I don't know if you know anything about Dray II, GRX-34, TAH-18, so Dray II, you know, can donate electrons. TAH-18 is a flavodox and GRX-34 have these unusual iron clusters with iron sulfur and glutathione centers. We have genetic evidence and some biochemical evidence that GRX-34 can donate the iron, and we had some data early on in my lab about YFAE. We haven't published any of this, so you can see that it's still not, I think it is a universal system. Actually, Dray II, TAH-18, and GRX-34 I think are involved in all iron homeostasis and cytosol. Nature has been amazingly conserved in iron sulfur clusters, and I think the experiments we did with Lil demonstrate that iron sulfur cluster biosynthetic pathways are also involved in heme and non-heme iron insertion of iron, and I think that was the most important thing of the paper actually, whether it be mononuclear or dinuclear centers, but proving that this is true, I mean a lot of people with iron sulfur clusters, if you read the literature, it makes a chemist's hair stand on end, and they're extremely challenging problems because of the low ability of the clusters. Yeah, so there's no analog, because I don't think you have to deliver oxygen. So I think oxygen goes down, I mean we don't really have a good channel in our protein, that's why I didn't show you, but we have a hypothesis for how oxygen gets in. We think it's to the iron II state based on all the crystallographic data. So there's a hydrophobic tunnel, and people in other di-iron clusters have putative channels to get in and out where they trap scene on and the structures. So, well you know, I mean again, we don't know, I think it's gonna be, if we find something I think it will end up being universal, I think we haven't gotten to the stage where we publish anything that means that this is still very much a work and progress, which I hope I can finish before I retire. Okay, so if you wanna read about that, we published an article in annual reviews of biochemistry and it outlines all the gory details. We know a lot about the structures and how to think about the redox potentials based on work that people have done on flavodoxins. And so based on what we know about the structures, we come up with a model for why that might be true. Yeah.