 There it goes. So thanks, everyone, for coming today. So we'll go ahead and get started. So welcome to the 2015 Everson Lecture. And this year's Everson Lecture is Squire Booker and Perry Fry will introduce Squire in just a minute. But I wanted to say a few words about Gladys Everson, who endowed this lectureship. So Mrs. Everson was a student in Harry Steambox Lab in 1942. She then went on to become the founder of the nutritional and home economics department at UC Davis, where very fittingly, she studied trace elements in nutrition, trace metals in nutrition. You'll see today that Squire is an expert in metal dependent enzymes. She then endowed a series of lectureships, which began in the 1970s. So they've been going on for about 40 years now in the biochemistry department. And the purpose of these lectureships are to bring back alumni of biochemistry, whether they were students or staff, to present their work and see what they've done since their time at UW. And so with that background, I'll turn the microphone over to Perry Fry, who will introduce Squire. Thank you for allowing me to introduce Squire. Squire and I have a lot in common. This lecture series is to support lectures by distinguished visiting scientists. Squire fits the bill. He is indeed. Squire was born and grew up in Beaumont, Texas, and went to Austin College and majored in chemistry, graduated with a bachelor's degree in chemistry there. From there, he went on to MIT, where he studied under Joanne Stubbe for the PhD degree in biochemistry, really the very chemical aspects of biochemistry. Later, he spent a year here at Madison in my laboratory. And in that experience, he studied acidenosilmethionine and the role of acidenosilmethionine in a radical reaction. There have been two penetrating experiments on the mechanism by which iron sulfur centers cleave acidenosilmethionine into a radical. Squire did the first one. And it was a beautiful experiment, and I love that experiment. I'm sure Squire does too. Since then, he's gone on to Penn State. And quite immediately after his work, one of our former graduate students, Heidi J. Sophia, she was a graduate student here in biochemistry, put together, discovered the radical Sam superfamily, of which the enzyme that Squire was working on is a founding member. And the radical Sam superfamily has now grown to more than 70 families of enzymes and more than 70,000 proteins in the known genomic database. All of these enzymes catalyze extremely chemically difficult reactions. And that's why Squire is studying. And he is certainly a leading figure, in my opinion, the leading figure in this field. These enzymes studied in Squire's lab, he began with lipoyl synthase, which inserts sulfur atoms into unreactive CH bonds. He extended that into thiomethylation of nucleic acids, inserting sulfur between CH in a heterocyclic base. He has continued and has studied carbon-carbon bond formation in post-translational modification of proteins. Carbon-carbon bond formation in post-transcriptional modification of nucleic acids. And that is the subject for today. Squire? Thank you so much, Hala. Who exactly it was? I've only flown into Madison twice in my life. Normally, I drive for some reason when I come to Madison. I've been to Madison a number of times. But the first time that I ever flew into Madison was 1995, when I left Paris, France, to come to University of Wisconsin to do a post-doc. And as you can imagine, after spending this wonderful year in Paris, I got off the plane and I was almost in tears, right? And the second time is to give this particular lecture. And I can't be more honored and more excited to be able to have been invited to give this particular lecture. While I was at Wisconsin, I just had a wonderful time. People were just so warm and friendly. And the science was just so amazing. And every time I come, I really feel like I'm amongst family here. So it's just a wonderful experience to be able to be here. So within the last five years or so, not too much more, have we begun to sort of understand how nature adds maple groups to unactivated atoms? And on this particular slide, what I'm showing you is a number of natural products that have maple groups added to either carbon atoms. The carbon atoms can be sp3 hybridized, as shown right here for phosphamycin, gentamycin, fortamycin, bacterial chlorophyll. Or in some cases, the carbon atoms can be sp2 hybridized, as shown here for chlorobiasin, tryptophane. And for these molecules right here, this is adenosine 2503 and bacterial ribosomal RNA. Also, carbons can be added to unactivated phosphorus atoms, as shown right here in this phosphinothraicin, which is continuing to be studied by Susan Wong, who was also a postdoc here at Wisconsin-Madison. And so it's only relatively recently that we've begun to understand how you add a methyl group to an unactivated carbon center or phosphorus center, although we thought we understood quite well how the universal methyl donor, S-adenosyl methionine, functions in general. And so what I'll talk about today are two modifications of an sp2 hybridized carbon center, C2 of adenosine A2503 of 23S ribosomal RNA, and C8 of the exact same nucleotide, A2503. So in almost all bacteria, it's known that carbon 2 of A2503 is methylated, thanks to this protein called RLMN. A2503 resides in the peptidyl transferase center of the ribosome near the entrance to the exit channel for the nascent polypeptide. This activity is not absolutely required, but it is believed to play a role in translational fidelity. In a number of other bacterium, C8 of that exact same nucleotide can be methylated thanks to a protein called CFR. CFR is evolutionarily related to RLMN, but it methylates at C8 preferentially than C2. RLMN will methylate only at C2. CFR methylates preferentially at C8. If C2 is open, it will methylate at that site as well. It's not clear whether C2 methylation requires prior C8 methylation, but it's clear that C8 methylation is much more preferred than C2 methylation. This post-transcriptional modification, like I said, plays a housekeeping role in translational fidelity. This post-transcriptional modification confers resistance to about five classes of antibiotics that target the ribosome. Here's a crystal structure of the ribosome. Here is A2503. The crystal structure is in the presence of this person-made antibiotic called Linazolid, which is completely synthetic. Here is C8 of A2503. Here is C2 of the exact same nucleotide. You can appreciate how incorporating a methyl group at this particular position can sterically impede the binding of Linazolid to its site. Many of these other classes of antibiotics have overlapping binding sites with Linazolid, and they, too, have weakened binding affinities when this methyl group is installed at this particular position. So the gene for CFR was only identified in about the year 2000 in Staphylococcus shiriii, which is an animal pathogen. In 2007, it was found in a hospital patient in Columbia who, unfortunately, was fatally infected with methicillin-resistant Staphylococcus aureus. And since that time, it has been observed in a number of other countries, including the United States, Mexico, Italy, Ireland, and Spain. And these are only Staphylococcus isolates from people who are infected. The gene has also been found in a number of animals that have been infected by a variety of species of Staphylococci. More recently, in 2012, the CFR gene was found on a plasmid in a patient in Thailand infected with interococcus vacalis. So it's clear that this has jumped on species. And there's a concern that there is basically a reservoir of CFR genes out there. And so there's a concern that this is going to spread. Now studies from Donica Fujimori's lab showed that the methyl group that is appended at C2 in the R-alumin reaction, or C8 in the CFR reaction, is actually derived from acidenosilmothionine, okay? And so the chemistry of acidenosilmothionine, we all, we felt like we knew quite well in terms of how acidenosilmothionine is used to methylate substrates. You need to generate a nucleophile. The nucleophile simply attacks the activated methyl group right there of acidenosilmothionine. We know something about the transition state from isotope effects, et cetera, it's supposedly tight. And then the transition state falls to give you your methylated species and your secondary product with acidenosil homocysteine. So key is you need to generate a nucleophile to make acidenosilmothionine-dependent reactions. Go. But if you look at C2 of adenosine, or C8 of adenosine, you'll notice, first of all, that those positions are not nucleophilic, they're electrophilic, and the protons associated with C8 and C2 are not sufficiently acidic to allow generation of a nucleophile by simple deprotonation. So, Houston, we have a problem in terms of how we understand SAM-dependent methylations to take place in these particular reactions. It turns out that Donica Fujimori's lab showed that these two enzymes are members of this radical acidenosilmothionine superfamily that Perry has already told you about. This superfamily now encompasses at least 70 different types of reactions and has over 113,000 individual sequences in it. They're involved in a diverse assortment of very interesting and mechanistically rich reactions that include cofactor biosynthesis, DNA biosynthesis and DNA repair, anaerobic metabolism, shown right here, the biosynthesis of a variety of different antibiotic and anti-tumor-natural products. Of course, I'm only showing you a few, a number of post-transcriptional and post-translational modifications and also in host defense and viral invasions. So this is a very important superfamily of enzymes. So if you look at the sequences of radical SEM enzymes, almost all, about 99% of them have this unique spacing of cysteines, cis, x, x, x, cis, x, x, cis, and those cysteines coordinate a required four-iron-4 sulfur cluster in catalytic. The fourth-iron ion has available coordination sites to which acidenosilmothionine binds in a bidentate fashion. And the general dogma, most of which was worked out here at University of Wisconsin in Perry's lab, is that upon reduction of this cluster to what we call a four-iron-4 sulfur plus one state, we do an inner-sphere electron transfer into the sulfonium of acidenosilmothionine to give this five-prime deoxyadenosil, five-prime radical. And that's this common intermediate in just about all of these 113,000 radical SEM enzymes. And the role of this five-prime deoxyadenosil radical is simply to rip off a hydrogen atom from some type of substrate, and that initiates the catalytic process. Now, based on that information, we can come up with a working hypothesis for how these reactions might take place. Here's acidenosilmothionine bound to that unique iron site, all right? We're gonna reduce that particular cluster, transfer an electron into the sulfonium. We're gonna cleave this carbon-sulfur bond right here to generate our five-prime deoxyadenosil radical. And what we postulated was that in CFR, which methylates at C8, this radical abstracts a hydrogen atom from C8 to give this particular radical. And by some act of God, we're gonna put a methyl group on that, basically a methyl radical, all right? And then in RLMN, this radical would be directed to C2, rips off a hydrogen from that position to give this C2 radical. And again, we're gonna somehow put a methyl radical on that, all right? Now it's somewhat embarrassing for me to show this initial, I never published it, but to show this initial proposal in front of my old mentor, I mean, it's a terrible mechanism for how you would do this, but in the absence of any other kind of data, it was all we had, right? And the main thing that we did, there were a number of things that we didn't like about it, but the main thing is that if you looked at the homolytic bond dissociation energies for cleaving a C, the CH bond at C8, it's 100,000, 113 Kcals per mole, all right? That would be better than Brian Fox and methane monoxygenase, right? And if we look at C2, we're about methane monoxygenase, about 105 Kcals per mole, so we never really liked this particular mechanism. But the good thing I'd say about the mechanism, as I learned when I was a student, is that it's experimentally testable, right? We can design experiments that will allow us to interrogate this particular mechanism. And so one prediction that this mechanism makes is that the isotopic distribution of the methyl group of SAM that we have in our assay should be reflected in the methylated product. In other words, if I have acidinosilmethionine with a CH3 group here, our product should have a CH3 methyl group. If I have acidinosilmethionine with a CD3 here, our product should be a CD3 methyl group because we know that the methyl group comes from acidinosilmethionine. That's the first prediction that our proposal makes. The second prediction that our proposal makes is that deuterium incorporation at C2 or C8 of the substrate should result in transfer of deuterium to 5-prondeoxyadenosine. In other words, if I synthesize substrates with deuterium at C8, then in the CFR reaction, if this guy abstracts that deuterium, I have stable incorporation of deuterium in 5-prondeoxyadenosine. And I can look at that by LCMS, all right? For the RLMN reaction, if I have deuterium at this particular position of our substrate, if that rips off the deuterium atom, again, I have stable incorporation of deuterium in 5-prondeoxyadenosine. Okay, so before we get to the mechanistic aspects of this particular project, let me just tell you that there are five cysteines that are absolutely conserved, okay? And whenever you have cysteines, they're absolutely conserved. They're typically doing something interesting, right? Three of those cysteines, the ones that are overmarked by this red bar are those that bind to the foreign for sulfur cluster that we've already talked about. But there are two remaining cysteines, one towards the interminable part of the protein and one towards the extreme C-terminal part of the protein. And I'll show you that these two cysteine residues play very intricate roles in this reaction mechanism. Okay, and then so lastly, let me say that this is what our substrate looks like, right? And as an enzymologist, you never want your substrate to get any way, right? You're trying to address the chemistry. But our substrate is 23S ribosomal RNA, which is 2004 nucleotide. So it's very much a substrate from hell, all right? Donica Fugimori showed that if you're able to synthesize just domain five, you only have to deal with 771 nucleotide, all right? So what a bargain, all right? Much less complex. Nevertheless, we initiated our studies with a small seven-mer nucleotide substrate that flanks the adenosine that's methylated, both by RLMN and CFR. And because this substrate was known to be incredibly poor, in fact, it was deemed not turned over, we ran these reactions under what are called single turnover conditions where we have enzyme in much greater excess than our substrate. We're just trying to look for anything to happen, okay? And so these are what these studies look like. If we take RLMN, unlabeled acidenosilmothionine, incubate them together under turnover conditions, we get this peak in our LC traces that comes off at about 6.9 minutes. If we analyze that peak by mass spec, we get an M over Z that is consistent with our expected product. If we leave out the enzyme, we don't see a peak at all. So that's our control. However, if we take the deuterated form of acidenosilmothionine, CD3 right here, this is what our expected product would be, all right? Here is the peak with unlabeled acidenosilmothionine. Here's the peak that we get off in our traces with labeled acidenosilmothionine. That's what we expected M over Z 285.1. What we observed was M over Z 282.1. So the product that we get is not reflected in the methyl donor that we add to our assay. So there's a problem there. If we turn to CFR, which methylates at C8, we see the same result. If we use a deuterated form of acidenosilmothionine, that's what our expected product would be. This is where the C8 product elutes in our LC trace. If we analyze that by mass spec, this is what we get. We get an unlabeled methyl group. So what's the problem? So we assume that, well, perhaps acidenosilmothionine binds tightly to the enzyme. And so when we isolate it, it's already bound. And so we take a good bunch of protein, 250 micromolar, and we simply denature it and analyze it by our standard HPLC procedures. This is where acidenosilmothionine would elute. As you can see, no acidenosilmothionine is present. So we have a clear problem. So at this point, we decided that perhaps the methyl group is already located somewhere on the protein. And so we did a proteomics analysis of a tryptic digest of RLMN. And so what I'm showing you are the Y-ion series, which have all of the peptides that have the C-terminal arginine associated with it and the B-ion series, which are all the peptide fragments that have the N-terminal aspartate associated with them. And the Y-ion series is particularly diagnostic. If you look at each of these fragments, Y2 would be DR, Y3 would be IDR, Y4, BIDR, et cetera, et cetera, they are exactly the size that they should be based on the sequence of the protein in the database, okay? All of that works until you get to the transition from Y10 to Y11. This is the Y10 fragment right there, which is exactly the size that it should be. But the Y11 fragment is exactly 14 mass units greater than it should be. And then 12, 13, 14, 15, 16, 17 are all 14 mass units greater than they should be. So this allows us to pinpoint this cysteine right there as the carrier of a methyl group, okay? So we can see that by mass spec. So about this time I had a high school student come to my office and say, can I do research with you? And I wanna be encouraging, but there are all sorts of other issues that I tend to deal with on a day-to-day basis. But of course we need structure. We need structure. All of these iron sulfur-containing enzymes are unstable in the presence of oxygen. And so when we do crystallizations, we have to do these inside of an anaerobic chamber. Grad students aren't paid enough to screen anaerobically. And so I said, yeah, I got a great project for you. You know? And I put this person in the glove box to generate conditions for screening for this particular enzyme. And miraculously, he was able to generate conditions under which the enzyme crystallized. We collaborated with Amy Rosenzweig at Northwestern and her incredible postdoc at the time who was now my colleague at Penn State, Amy Boll. And we solved the crystal structure of the protein. And so I won't get much into the crystal structure. It looks like a typical radical SIM enzyme in the sense that it has a derivative of a Tim barrel. In this particular case, it's alpha-6, beta-6. And so that's the core domain here. This domain right here is often found in proteins that bind to RNA. But there's this loop here that is only visible in the structure with acidenosilmothionine bound. In our structure that doesn't have acidenosilmothionine, you don't see this particular loop at all. On this loop is that cysteine, 355, the one that's at the very C-terminus of the protein, that we identify by mass spectrometry. And if we look closer at the electron density, you'll notice that the electron density fits quite well with a methylated cysteine residue. And if we tell the program like, no, it's really just cysteine and it tells you you're missing some electron density there. So at this point, we see this methylated cysteine by mass spectrometry and we see it also by x-ray crystallography, all right? So at this point, we know that the transferred methyl group derives from a methyl cis residue, but we don't know how it's transferred to the nucleotide substrate. And at this point, we're sort of confused because in acidenosilmothionine, the methyl group is activated, right? But in a thiolether, the methyl group is not activated for polar transport. So why would an enzyme deactivate this methyl group in the course of methylating its substrate? So at this point, we needed to generate some sort of a probe that will allow us to determine how you transfer a methyl group from a cysteine residue to the substrate. Okay, so the isotopic distribution of the methyl group of sandmatted to the assay should be reflected in the methylated product. We showed that that's not the case under single turnover conditions, okay? And so now we need to figure out how you go from a methylated cysteine residue to the final methylated product. And so to generate a probe, we use an oxytrope of methionine that won't grow unless we add methionine to it. And instead of normal natural abundance methionine, we add a methionine containing a CD3 methyl group that gets into the cell. And of course it has many fates. One of the fates is that it's used in protein synthesis. But one of the other fates is that it is acted upon by an enzyme called acidenosylmethionine synthetase, which takes methionine and ATP and makes acidenosylmethionine, okay? And so all the acidenosylmethionine in the cell now has this deuterated form of acidenosylmethionine. And so this is being transferred to all of the typical acceptors of methyl groups from acidenosylmethionine, including that cysteine residue in RLMN and CFR if we co-express them in the same cell, okay? And so we isolate our proteins. This is the unlabeled RLMN protein. If you use unlabeled RLMN, of course our expected product would be unlabeled 282.1. This is our labeled protein. This is the protein that we produce in this oxotrope in which we add deuterated methionine. And now our meth, our expected product under one turnover conditions, should be M over Z 285.1, okay? When we actually did the analysis, what we found is that we cleanly made a product with M over Z 284.1. So this is the second time when our initial hypothesis proved to be wrong. So in going from this form of the protein to the product, we always lose one hydrogen atom from the methyl group on the cysteine. Okay, so the transferred methyl group derives from a methyl cis residue, but loses one of its hydrogens in the process. Is the missing hydrogen removed by the five-pronged deoxyadenosyl radicals? So instead of abstracting a hydrogen atom directly from the nucleotide substrate, is it possible that now you abstract a hydrogen atom from the methyl group bound to the cysteine? And that was a very favorable idea to me due to the studies that Perry did on lysine 2,3-amutase with thiolysine, right? What happens is by having a carbon on a heteroatom, especially sulfur, you decrease significantly the homolytic bond association energy associated with breaking that carbon hydrogen bond. So is the missing hydrogen atom removed by the five-pronged deoxyadenosyl radical? Again, that's easy to test. The first thing that we did, of course, is synthesize substrates that had deuterium at C8 for CFR or C2 for R-alamin. And we showed in those particular cases that you don't see deuterium transferred to five-pronged deoxyadenosyl. So that tells you that this does not rip that off, that does not rip that off, okay? But if we use our deuterated form of either R-alamin or CFR and if we generate this five-pronged deoxyadenosyl radical, we find deuterium transferred to five-pronged deoxyadenosyl, which we can see by mass spec, as shown by an M over Z of 253.1. So that's this peak right here for R-alamin and that peak right there for CFR. So that information allows us now to start building a mechanism for how this reaction takes place. And so here's the mechanism that we have finally settled on. So Cysteine 355 is the Cysteine that gets methylated in the cell, okay? In the first step of the reaction, Sam binds to the protein, acidenosyl methionine, and a methyl group is transferred to this Cysteine via typical polar SN2 chemistry, no radicals involved in this particular process. That gives you this methyl Cysteine residue. Acidenosyl Homo Cysteine can now depart from the active site and another molecule of acidenosyl methionine binds in the exact same place, okay? Now, this molecule of acidenosyl methionine can't donate a methyl group because this sulfur is already blocked with a methyl group and so it sits there and waits for the cluster to be reduced by an electron. When the cluster gets reduced by electron, it gets fragmented to the five-prime deoxydenosyl radical. That radical does not abstract the hydrogen atom from the substrate, but does abstract the hydrogen atom from the methyl Cysteine residue. That gives this methylene radical which can add to the nucleotide substrate in this particular fashion. This is very reminiscent of what organic chemists have called the Menichi reaction, okay? That gives this particular intermediate wherein your protein shown here via this methyl Cysteine or methylene Cysteine and that Cysteine is cross-linked to your nucleic acid substrate, okay? This radical on your substrate decreases the pKa of the proton on the carbon adjacent to it so you remove that particular proton to give a radical anion that rewrite in this particular resonance form. That allows you to undergo a fragmentation to generate an enamine and a phyl radical. This enamine can tautomerize such that this picks up a proton from the Cysteine residue, so that becomes a general asset now, to give you your methylated product and a phyl radical and then lastly an electron donated by the cluster reduces the phyl radical to the phyl A, okay? So that's the mechanism that we've settled on and that we'll talk about in the next part of the lecture. Okay, so the next question that we wanna ask is that we know what happens, or we have a very good idea of what happens once the methyl group is already on the Cysteine, right? But we have no idea how the methyl group actually gets on that Cysteine. What we do know is that we can run this reaction in the absence of any other factors suggesting that the enzyme itself catalyzes methylation of the Cysteine and then also methylation of the product. If you look at the crystal structure that we solved, again this is in the absence of our RNA substrate, the methyl group of acidinosilimethionine bound to the iron sulfur cluster is only about four or five angstroms away from the sulfur of the Cysteine that gets methylated. And so we predicted that this is actually the source of the methyl group for the methyl Cysteine. So when acidinosilimethionine binds, this Cysteine just simply attacks that particular methyl group and that you generate acidinosilimethionic Cysteine and your methylated Cysteine residue. So the question is, so what this tells us is that the iron sulfur cluster itself is an important binding site for SAM that's used to methylate that Cysteine. So how can we test this experimentally? So what we did is we produced our element and we've done this experiment with CFR as well under conditions such that just before induction of expression, we throw in the iron chelator ortho-finanthraline. So ortho-finanthraline gets into the cell and it binds up all the iron. And then we induce with IPTG and we make our element in its APO form inside the cell. And then we purify, APO meaning no iron sulfur cluster. And so we purify our protein much like we purify our typical normal wild type protein, okay? On this particular slide, I'm showing you in this black curve the UV visible spectrum for APO-RLMN. And as you can see in this region from about 300 to 700 nanometers, you don't see very much absorption at all consistent with very little iron sulfur cluster associated with the protein. And if we take that particular form of the protein and treat it with acidinosilimethionine, that's these purple dots right here, nothing happens, okay? So if we take our APO-RLMN, no iron sulfur cluster and we treat it with acidinosilimethionine, nothing happens. Now when the iron sulfur community, there's this process that we call reconstitution where we can take iron and sulfide and reductant, mix them all together, stir them for a little while and amazingly you reconstitute an iron sulfur cluster on the protein. And this spectrum right here is consistent with an iron sulfur cluster being reconstituted on the protein. I turn to my colleague, Karsten, and I say, Karsten, is this really an iron sulfur cluster? And he says, why are about 80% of all of your iron is in the form of an iron sulfur cluster in that sample? And then if we take that particular sample and treat it with acidinosilimethionine, we see rapid formation of acidinosil homocysteine consistent with methyl transfer. And the amount of methyl transfer we get is equivalent to the amount of iron sulfur cluster that we have based on quantification by Moss-Bauer spectroscopy, all right? So again, this is consistent with the iron sulfur cluster being required for that methyl transfer. We can look at the other product which is the methylated cysteine and I won't walk you through every part of this particular slide. The take-home message is that under conditions in which the iron sulfur cluster is present, we generate the methylated cysteine residue and acidinosil homocysteine. Under conditions in which the iron sulfur cluster is not present, we don't generate a methylated cysteine and that cysteine then becomes trapped by iodacetamide before we do our mass spectromaric analysis, okay? So this is what's going on so far. This is the state of the protein in which acidinosil amethionine is bound to the iron sulfur cluster. This is the cysteine that undergoes methylation. This is that second cysteine that we'll talk about in just a minute. This is about five angstroms, four to five angstroms away. That sulfur and that methyl group of SAM bound to the cluster is about five angstroms away. That attacks the methyl group to give you your methylated cysteine acidinosil homocysteine. This then leaves the active site to give you this species. And so this species is the form of the protein that we typically isolate in all of our experiments. In the next step, another molecule of acidinosil amethionine binds to that same site, okay? You can't do another methyl transfer because you're blocked. You have a methyl group there. So this guy just sits and waits until that cluster gets an electron. When that cluster gets an electron, you do that reductive fragmentation that I told you about at the start of the lecture to generate a five prime deoxyadenosil radical that is basically juxtaposed against the methyl group from which it needs to abstract a hydrogen atom. So everything seems to be perfectly aligned for the chemistry that we're proposing. Okay. Can we see any intermediates in this particular reaction? You can postulate and you can speculate, but until you see via some sort of method, you're not really walking on solid ground. One of these intermediates is paramagnetic or several of these intermediates is paramagnetic, but this is a key intermediate because it is an intermediate that includes both the protein and the nucleic acid substrate via this methylene bridge. So the protein and nucleic acid substrate are cross linked and the intermediate has an unpaired electron. All right. It turns out, now I'm showing you the reaction mechanism for RLMN, the reaction mechanism for CFR is basically the exact same, okay. It turns out that for CFR, we can actually see that paramagnetic intermediate, okay. Now at this point in the lecture, we're no longer working with, or we left that quite a while ago, but we're no longer working with that small seven mer oligonucleotide. We have to roll up our sleeves and learn how to make RNA in a robust way. And we were able to quite easily make domain five, the 771 nucleotides and substrate quantities, but we also found that a small 155 mer oligarchs or mer oligomer also gave robust activity. So most of our studies from this point on are conducted with this 155 mer RNA oligomer. So if we take wild type CFR, the 155 mer RNA oligomer, an acidinosomathionine and we mix them all together and then we throw in dithionite and freeze rapidly in cryogenic ethane, this is what we observe. We see this beautiful pattern that looks like an unpaired electron that's being split by a single hydrogen nucleus, okay. In order to thoroughly simulate this lecture, we have to include these wings that you see out here, which come from the contribution of the nitrogen shown right here. So how do we know that this radical is associated with our nucleotide substrate? Well, we can use a substrate that's deuterium labeled and you see collapse of the splitting to give you this particular species shown here. How do we know that this radical involves a cross link to the protein? We can make our protein such that the methyl group is C13 labeled and when we generate the EPR sample with that particular protein using this particular background where you have deuterium on your substrate, you see the splitting come back due to the I is equal to one half nuclear sped of the C13. So this in combination with indoor experiments and density functional theory experiments lets us fairly thoroughly assign this particular radical in this reaction. And so most of the spin density is on the N7 nitrogen as we drawn it, but it's also delocalized throughout the nucleotide. Now as an enzymologist you can't get away with, hey, here's this EPR spectrum of what we say is an intermediate. You have to show that it's an intermediate by showing chemical and kinetic competence, right? And so we use a process that I think was initiated here at Penn State back with Binard and Ray and everybody, but rapid freeze quench, here at Wisconsin, I said Penn State, here at Wisconsin with Helmut Binard and his colleagues, Ray. But we use rapid freeze quench where we take CFR, the 155 RNA substrate, acidinosilmothione, we put it in one syringe, we have dithionite in a second syringe and we quickly mix them together on the millisecond time scale and the solution goes into this cryogenic ethane. We can blow off the ethane, take the crystals, pack them in an EPR tube and then measure the EPR signal as a function of time. And these are just some representative spectra, but this is what the time course looks like in red for formation and decay of that particular EPR signal. The black right here corresponds to formation of our product. The rate constants for formation and decay are consistent with the black line, which is formation of our product. So this radical that we see has all the hallmarks of a chemically and kinetically competent intermediate, okay? Now let's go back to this particular structure, okay? This structure tells us we need to figure out what this cysteine is doing and in our mechanism we're suggesting that the cysteine is acting as a base to remove the proton at this particular step in the reaction, okay? If that is true, then if we change this cysteine to an amino acid that doesn't have a compatible functionality, we might be able to see this particular spectrum build up. In fact, in RLMN, so what I showed you was the EPR of the CFR reaction. In RLMN under normal turnover conditions with wild type enzyme, the radical never builds up, indicating that decay is faster than formation, okay? However, if we change this cysteine, that's the second cysteine that I told you that's absolutely conserved to something else, we might be able to build up this particular radical because it can't decay without prior removal of that particular proton. And that's exactly what we see if we make the cysteine to serine or cysteine to alanine variant. We see EPR spectra that are absolutely consistent with this particular radical in RLMN. I won't go through all the assignments. We did that in the CFR case, but again, this is the exact same analogous radical in RLMN that we see in CFR. And if you look at the kinetics of it, these are just some representative spectra. You see the radical in RLMN accumulate, but never decays. This is the radical in CFR, wild type, where you see formation and then decay, but in RLMN variant, the radical doesn't decay. We see the same behavior for CFR if we make that particular variant in CFR, it would be C105S instead of having formation and decay, the radical forms and never decay. So that's consistent with that cysteine playing the role of gating that radical fragmentation process. Now the second prediction that our mechanism makes is that if you remember, in this particular intermediate, our protein is cross-linked to our nucleotide substrate. So if we take that same variant, cysteine to alanine or cysteine to serine, at this particular stage of the reaction, the nucleic acid should be cross-linked to the protein. When we isolate that particular variant from the cell, this is the RLMN C118A variant, you'll see that the UV visible spectrum shown in blue, I mean in black, is blue shifted from that shown in red. That's the typical wild type spectrum, indicating or suggesting that nucleic acid is bound to that particular protein. We only see this behavior in that particular mutant or variant, okay? If we take that protein and then chop it up with a variety of nucleases, we see all the hallmarks of RNA bound to the protein, consistent with the RNA bound to the protein covalently. Now at this particular point, we desperately needed a structure of the protein in complex with the RNA and for us to really get down to the details of the mechanism of this particular reaction. But everything we did to try to crystallize the protein either with the 771 mer or the 155 mer didn't work, okay? And so I have this incredible bold student, Tyler Grove, who just says, I'm gonna purify this out of the cell and then crystallize it, okay? And I said, I remember I said, Tyler, that's the stupidest experiment I have ever heard of. And the reason being is because we're thinking, we're pulling out 23S RNA, that's 2,904 nucleotides that's never been crystallized outside of the ribosome. I'm not a car carrying crystallographer. If Jennifer Downing can't do it, I certainly can't do it, right? And so you can't tell Tyler that he can't do something. And so what Tyler did was he yanked this right out of the cell and he set up crystallizations and he told me he got crystals. And of course, I assume that they were just crystals of the protein itself. He was mad at me for calling him stupid. And so I said, Tyler, come to Prague with me for this meeting and in Prague, he never showed up to the meeting because he was feeding electrical intensity data back in the room. And what he found actually was that, indeed he had RNA bound to the protein. But if you look closely at the RNA that he pulled out of the cell, it's tRNA rather than ribosomal RNA, all right? So all of my colleagues in the field say, Squire, you were right, that really was stupid. But the reason why it worked is that he pulled out tRNA rather than ribosomal RNA. And that has a very tight and rigid structure. There are a number of structures out there with tRNA bound to tRNA modifying enzymes, okay? And so another student of mine, so it turned out that two years ago, there was a paper by Alman God's Laboratory in Spain that showed that RLMN is a dual specificity methylase. It's responsible not only for methylating 23S ribosomal RNA, but it's also responsible for methylating adenosine 37 in about seven different E. coli tRNAs, okay? And so with that information, we can now appreciate why we pulled tRNA outside of the cell since, of course, it's much more abundant in the cell than the ribosomal RNA. And so these are the tRNAs that are modified by RLMN. Arginal, glutamyl, histidyl, spartal, glutaminyl in another glutaminyl form. And another student of mine, Erica Schwam, synthesized all of these tRNAs in vitro and used them as substrates for RLMN. And she found that tRNA arginine and tRNA glutamate were the best. And it's actually the tRNA glutamate that Tyler pulled out of the cell, all right? So those support it, the greatest activity. And so Erica then took tRNA glutamate, mixed it with RLMN under turnover conditions and generated the cross link in vitro and then resolved the structure to about two point four angstroms resolution, okay? So the question is, how is it that this protein can recognize both ribosomal RNA and seven different tRNAs? And one of the answers that we see from the structure is that there are very few sequence specific hydrogen bonds between amino acid side chains and nucleobases, for example. Most of the interactions come from either amide functionalities from the backbone and the phosphate of the RNA or amino acids side chains and the phosphate or the sugars of the RNA. But very few interactions are between the protein, between side chains and the nucleobases of the RNA. Along the part of the structure where the modification actually takes place, there are a few sequence specific interactions between side chains and nucleobases. One of them is with a conserved arginine residue, arginine 206, which forms a hydrogen bond to guanine 29. And in all of the tRNAs that are modified by RLMN, this guanine 29 is there, okay? This appears to be a specificity determinant for tRNA, this particular arginine. If Erica takes this arginine and changes it to a alanine, for example, the protein catalyzes methylation of 23S ribosomal RNA with wild type activity, but it doesn't touch tRNA at all. So this is a binding determinant for tRNA, okay? So lastly, I just want to show you what the active side of the protein looks like just before binding of your tRNA substrate. So this is your methylated cysteine residue. Here's acid dinosilma binding bound to the cluster. We're gonna undergo a fragmentation to generate a radical at this particular position. That is about four or five angstroms away from the methyl group of methylcysteine. We're gonna abstract a hydrogen atom to generate a radical at that particular position. And this is what the active site looks like in this addup that we made. The first thing that you'll notice is that methionine and five-prime deoxyadenosine are present. They're there both in the crosslink that we generate in vitro and the crosslink that we yank out of the cell. So that tells you that they're bound tightly and they're trapped in the active site by the RNA substrate. Down here, so you'll see here's the methionine still coordinated to the iron sulfur cluster. Here's the sulfur of methionine. I think Perry will appreciate this. That's also forming a coordination with the unique iron of the iron sulfur cluster as he and I and our collaborators showed via selenium exapse experiments with lysine two, three, I mean a mutase. Here's the five-prime deoxyadenosine which is the product formed after abstraction of a hydrogen from the methylcysteine by the five-prime deoxyadenosyl radical. And here's the addup that's actually formed and the electron density you can tell that this particular carbon is sp3 hybridized and so it's very much consistent with all of our EPR data. And so that's my story and I'm sticking with it for right now, okay? And lastly, I just wanna thank the people who did the work. This is Tyler, he moved on from my laboratory to do some more crystallography. Felon Lever crystallography is in Steve Almos' lab at Albert Einstein. This is Erica Schwarm who worked with my colleague, Amy Boll to sort of learn some crystallography and so she's worked with Amy to solve the structure of the in vitro cross-link. This is Alexei Silikov who's an expert EPR indoor ECM, any type of post-EPR technique person. Here's Matt Barley who's a graduate student in the lab as well who did some of the variant studies with RLMN. And lastly, I am truly honored to have been asked to do this lecture and I wanna thank you for attending. I wanna thank the Everson family for sponsoring the lectureship. And if you have any questions, I'm happy to answer. Then these are other people who have worked on the project in the past and this has been supported by a grant from the National Institutes of Health. George. Is the reduced cluster form? Well, you certainly don't need it, right? So when we do those assays, there's no reductant around. I mean, we don't even need DTT to do that. So we certainly don't need it. And this is one of these radical sample proteins where the cluster doesn't get fully reduced so we can't really do the experiment in the absence of non-reduced cluster. Yeah, I don't see why it wouldn't but if you have acidinosilmethionine there and you reduce the cluster, it might fragment anyway. But you don't require a reelectron to do that step. I can say that. Yeah, Brian. Right. So many of them outside of the active site, right? Most of those interactions, so it's recognizing shape, right? It's recognizing shape. And so many of the reactions are interactions are with the phosphate backbone. Yeah. Well, so there are other sequence specific interactions that are important as well, right? And so for example, in the ones that are modified, there's that guanine there that makes the hydrogen bond with the arginine. And if you don't have that guanine there, like I showed you, you don't get modification of at least tRNA glutamate, right? We haven't checked the other tRNAs yet but that appears to be one of the few sequence specific determinants of tRNA. There are probably other interactions as well. This is fairly new and so we haven't tested all of the other parts of the tRNA yeah, for their ability to support the reaction. Yeah, yes? Does the methyl group go to the same carbon in the tRNA that does it? Right, it's C2, it's C2 of adenosine 37 in all the tRNAs. So that's just the nucleotide right after the last codon and the enicodon, right? Does that get dilated later? In some tRNAs, there is a methyl-phio group there, right? But in the tRNAs where there's a methyl group there, there isn't a methyl-phio group. So in some of the other tRNAs, there's a methyl-phio group and that's put in by Mya B, which we study as well. Yeah, the sulfur in that case comes from an auxiliary iron sulfur cluster we believe and then you methylate that as well. So that's an important, that's an interesting reaction too. Yes, absolutely, absolutely. So that's a very good question. So if we, for example, the reason why we had to go to bigger, so if we try to do, so first of all we can do all of our experiments using an in vivo reductant, flavidoxin, right? But flavidoxin only works if you have a piece of RNA that's at least 155 mer in length, right? Shorter than that, it's clear that, you know, I'd imagine it's a reduction potential issue. But shorter than that, flavidoxin doesn't reduce the cluster, but we can use dithionite for the reaction to go in that particular case. So that's, you know, we've started to do some really nice studies with Sean Elliott at BU to look at redox potentials at different stages of the reaction, but you're absolutely right. The tRNA clearly gates modulates the reduction potential of that cluster, I think. Yeah. So does that particular region of the ribosomal RNA have a particularly tRNA like shape? Yeah, yeah, well, so what Tyler did is he was able to take the 155 mer and turn it on its side, right, and in that sort of configuration, what you see is that, you know, this part right here mimics one of those mini helices in the RNA. So, yeah, from a global sense, you can see perfectly how it can recognize both the tRNA and the ribosomal RNA. Now, you know, Tyler tried to get us to publish this sort of in silico model that he made, but once you get up into here, it doesn't quite, you know, there's clearly, I mean, we're taking the structure of 23S ribosomal RNA as it exists in the ribosome, right, that's all we got. And so, you know, there's clearly some distortion that would take place, but out here, yeah, you turn it on its side, you're making, you know, and some of the, you know, this only makes a couple of interactions with, you know, this part of the tRNA, and so we're thinking that, you know, this part may be more important for the 23S ribosomal RNA than it is for the tRNA, but one of the interesting things about the structure, if you've looked at structures of many tRNA modifying enzymes, I think this is the only one that I know of wherein the whole tRNA is recognized except for, like, tRNA synthetases, right? But if you look at enzymes that modify tRNA in the anticodon stem loop, like Arlamin, they bind, you know, you can use just the anticodon stem loop as a substrate for activity, and so this one clearly recognizes the entire length of the tRNA, and I think that's because it's a dual, it has to recognize 23S ribosomal RNA as well. That's what we think. Hi. Do you know wherein the ribosome assembly pathway that methylation occurs? I mean, it wouldn't... So, yes, based on work done by Donica, it would occur before you start to assemble, I would imagine, you know, Ryda, as you're coming off RNA polymerase or something like that, but I guess you have to have some sort of folding to have occurred, right? But she did some studies where she took the entire ribosome and showed that that's not methylated, and then she took the 50S subunit and the 30S subunit and showed that she couldn't get, you know, methylation. And so it looks like it's either before you start adding proteins to it or somewhere early in that pathway. Yeah. Thank you.