Chemotherapy For Colon Cancer

>> it's my pleasure to welcome to you this year's g. burroughs mider lecture. our lecturer today is weu yang, senior investigateories at niddk, honoring the memory of g g. burroughs mider. he was the n.i.h. director of laboratories and clinics from

1960 to 1968 under the leadership much james shannon, legendary n.i.h. director. there was an expansion in the size of the intramural research program and in the 1960s were if four nobel laureates working at the n.i.h. at the same time. it was an amazing period.

the lecture recognizes a scientist who simply nic exemplifies excellence. they are vetted by the board of directors and moved forward for the director to consider. this year's lecturer is wei yang, deputy chief in the laboratory of molecular biology

in niddk. she's used creative structural approaches to revolutionize her understanding of the interaction of dna with proteins during replication, repair and recombination. she received a b i fro b.a. from stoney book, ph.d. from columbia,

and post doc with tom stites at isle. she's been iat -- at yale. dr. yang has appreciated the need for detailed crystal graphic studies of the way in which dna and enzymes involved in replication repair and recombination interact.

this work is s. ha has provided new init into mutl and mutl, polymerase bound to substraits, specific recombination complex and structure of rnh bound to rna, characterized by a elegantion of el against chemistry to gain understanding of biological function.

her talks as you'll see and those who have heard her talks before will understand this, are models of clear communication and use as they aesthetically beautiful structures to clarify new biological concepts. dr. yang's work wack

recognizes -- was recognized by the hodgkin award. the talk is molecular mechanisms of dna repair. wei? >> thank you, mike. thank you for introducing me. i was very surprised when dr. gotsman told me the news i was

selected to give this year's as one of my professors said many years ago, the best award we scientists work for is peer recognition, and i feel very honored and humbled to receive such distinction by my peers at nih and the directors. and when i'm preparing this

lecture, i couldn't stop asking myself what i have done truly matters to medical research, and a couple days ago i was alerted by mike liston that my talk announced on the nih website at 3:00 a.m., february 22. [ laughter ]

it dawned on me what really is that we are willingling to to work hard to make a difference. i'm going to tell but something related to dna replication, essential for cell proliferation. for every cycle of cells to

divide before that dna has to be -- joh genomic dna has to be replicated, using checkpoints before going to the replication cell division. is has to complete the replication to survive. it's critical. an associate with dna

replication, we study two events, translesion isn't situation is necessary because all dna is not perfect. exposed to sun, it can be cross linked if we eat too much barbecue food, we may ingest carcinogennened at the dna base gets damaged.

it blocks normal dna preliminaries. on the other side of the coin, dna dna is made by klan preliminaries, it's known for low errors. in a genome of 3 billion base pairs, even errors in the ratio of one in ten million base

copied, wei going t we are going to have tens to hundreds left in the dna preliminaries, and those are taken care of by mismatch repair. to bypass the dna blocking replication blocking lesions, to repair errors made by dna preliminaries, are two sides of

the same coin. the purpose is to keep the genome at minimal mutation, as everyone knows cancer rises because of mutations, the pathways replicate on mutation rate and cancer incidence. i want to start my talk telling you what we do in the lab,

using x-ray crystallography, starting from a gene identified that's causing human disease, and if wei luck we are lucky we have wonderful people that discovered no pathways and proteins before the publish. i have two very good colleagues stopping me on mismatch repair

at nih, roger wingate and peggy see. the ones we identify, we generate enough protein to make crystals, they may be beautiful, maybe not so pretty, but the important thing is that we can deflect x-ray and record those spots on film or on

various recording devices nowadays, it's very advanced. and then using that information we can generate and build structure model and start to interpret what the model means and what the structure tells us how the molecule works. with advanced technology,

growing crystals back becomes easier. we have radiation source in chicago, we don't even need to go there. tonight wei goin we are going to collect data in the office and ship the crystal by fed ex and spend more time to understand

structure. it's pretty but doesn't tell us. when you see the structure, you can't understand how it works. we spend a lot of time trying to understand that, using biochemical cell-based tools, and the other thing, advanced

technology allows us to instead of studying one single small protein, now we can study large protein complexes, made of more proteins because no protein works alone. so the first part i'm going to focus on it mismatch repair, repairs the mismatching repairs

by dna preliminaries. here is story from a family. it's inheritable. one in 20 colon cancer patients diagnosed in america carry mutations deficient in in mismatch repair. according to the american cancer society a million of

americans carry one of those defective mismatch repair genes and those people have 80% of a chance to develop cancer in their lifetime. it affects many of us. the two genes or four encoding two special proteins called mutl and muts.

it's easier to handle, mutl and muts are preserved from bacteria to human. here is a demonstration of black slime, the re line, the red line. the muts recognizes the single mas mismatch and recognizes mutl. the nucleus will make an

incision in unmethylated strands at the site. the onconce made, it unwinds the dna, because the sequence can be in the genome and usually 100 base pair away from the mis mismatch. the whole strand will be resynthesized in that patch of 100.

as you can see, the process is extremely expensive. this process is conserved on humans as well as e. coli and mutl is playing the intention same role, in addition the process actually contains three atpa's. iatpa's.

interestingly, muts is known for abc transporter, mutl is a ghl, a difference atpa, and u.v. is a bone a sid bona fide protein. people say atp can do many things, but you ask most people what you think about that, people say it's a motor protein.

what i'm going to show you is mutl and muts are doing something different. we call it a sinker, mutl we think is a talker, recruiting protein to come to do the only the uvrd is a true motor i'll start from the simplest, that's the function everyone

understands. motor protein is burning atp, the energy is used to organize the motion to directional walking. it's not enough to make directional movement. if you drop a person into a virgin forest without any past,

the person can walk in the forest for days, never get out, because it doesn't know where gps.lk to unless he has a if there's a path one can walk on the path and get out of the forest. and the motor proteins are like humans.

in order for them to walk in certain directions they need a track, so everyone knows the filament, acting filament is the track. crystallography.o large for here is a picture using atm and video recording in japan. that's the state of art right

now, you can watch a motor move. you can see it's hard to really see the detail. one walks quickly, one gets stuck. you don't know what happens there. it's just like using a watch,

if the watch is telling you the perfect time, you're satisfied. but if the watch is telling you wrong time, or doesn't work, you need to look a little closer. open the watch, look at the detail. to really understand how the

protein works we need a closer look. however, the time resolution and spatial resolution really is going against one another. if you want to see the real time, you have to drop the resolution. if you want to see the high

resolution like x-ray crystallography we never can record a dynamic process because we process one at a time. that incompatibility it be resolved, using the example by edward maybridge, probably many people have seen this movie, it

looks beautiful. it was recorded. way before there was video camera. it was done in 1878. and the way he made this movie, the frames per second are fewer. he made it by placing cameras

around the racetrack and those pictures, he may have sorted them and put them together to make a stop action movie but that gives the idea crystallography can use the same way to make movies. we may not be able to make a panel of the pictures but maybe

two or three very essential stages of the molecule, how that works. and then splice them together to make a movie. that's exactly what we did. so we crystallized uvrd with dna and crystallized with dna and with a nonhydrolzable.

hydrolysis., a high droll sis if you couldn't the base par from this point to that point, they are different. binding adp and not binding adp. i will string it together as a movie. the work was done really

thoroughly by a post doc, j allen lee, who identified the three states. now you can watch. atp comes in and induces emotion between this portion of the protein to that portion of the protein. it's a rotation.

dna in the screw motion comes down and pushes against this, hairhe base closer to the pair pin gets separated. the adpi release motion. we break down the atpa cycle, binding induced the unscrewing of the double strand and the one base gets impaired.

and after atp hig hydrolysis, it allows single strand location to complete the cycle by making the three-state structure observation and put them together, we can answer with confidence, and we know where the protein dna occurs. some day we'll be able to use

it to really help the human understand how it works in the human body. the work we did leer is u here is up to date. probably it's still the best structure description of how it works. now i will switch gears to the

second atp, mutl, it's a dimer. 350 residue, very conserved. we said let's crystallize it and see how it works. the post doc, my first post doc did the work. he solved the structure. actually after the structure

was solved we were quite disappointed. the reason is that out of 350 residues that's crystallized we only can see 280. 70 were disordered. many are conserved. so we had to leave the structure on the shelf for a

while, true to understand what the structure means. while we were busy doing biochemistry we stumbled on that, this protein is an adpa, in the lit a fuel i literature it was labeled with no enzym enzymeatic activity. there were two powers, one from

nih, identified the mutl sequence hemology. the dna gyrase was discovered on this campus by my colleagues. so we went ahead and recrystallized a protein in the presence of atp. now the structure looks

different from without atp. the one showing orange, the 50, 60 residues now upon atp binding become ordered. some of those second structure formed interface for the protein to dimerize. this scheme of change actually is applicable to the whole

family of this and t-shock protein 90. so i'm just recapitulating this movement of atp induced change in the movie showing you again the protein starting looking pathetic, it's missing a lot of things. once atp bonds, you see induced

a huge amount of confirmational change, secondary structure formation, the protein goes from monimer to dimer. it was difficult to identify this is an atpa, protein 90 was not known, denied as an atpa for years. the reason the atp turnover

rate is slow, mutile turnover atp once per five minutes. what is it doing in five minutes? it turns over so fully. we were looking for what can control the speed of atp hydrolysis. we made a bunch of mutations

along the pathway, blocked from atp binding, blocked dimerization, atp hydrolysis. when the mutiles bind it's the only state it can activate. the dna has single strand exposed, speeding up this portion and reduced time here, increased time in the free

form. and the whole protein in the whole cycle of ath hydrolysis is essential for activating. it needs a signal and needs to be health to do it in hundred based pairs. here is a model we think how it it's turning over quickly and

unwinding hundreds per minute. the mutiles holding it together are dna, and itself turning over once in a few minutes. it's only when the dna becomes tangled becaus, reactivated if needed again. so those two stories are relatively still straight

forward, changing structure, servicing an energy source to allow a motor to work. i'm going to the most intriguing atpa of the family, the abc family. it's recognizing dna mismatch. there are two types of areas that dna preliminaries make.

one is missing the base against the template. microans we have mike roy satellite, it may skip forward or backward resulting in loops in dna. thosthe miss match inwe need to find them. the ratio of occurrence, frequence of occurrence of the

mismatch is about one in ten million base, you think about it. the proteins search through the genome, once finding it, one in ten million base pairs, to find the mismatch, which look very different on normal pairs, it's not gc based, it's a tremendous

task, how to find this one mismatch buried in tens of millions of normal base pairs. it's very reasonable to liken it to finding a needle in a haystack. that's intriguing, the mismatch has different shapes. i was approached, wonderful

protein, heat resistant, her post doc even better. while we were sidetracked by some crystal that's not deflecting well, galia slowed a crystal and we solved the we were the first one that solved ever the test of the dna together, it's a fairly large

protein, two subunits, green, blue. blue is the one showing interacting with the mismatch, because even though it's a homo-dimer, the structure are different because there's only one dna strand that contains the mismatch.

the blue subunit is interacting with the mismatch here, but each subunit has two domains, one on top, one below holding the dna together. the dna is quite bent. and we determined the different mismatch, we published together bindingb

mismatch. it look just the same. seven years later, human protein that caused colon cancer, it looked similar. in humans blue and green are sequence related but not exactly the same, the blue subunit again recognized the

and after solving the structure, of course, i will tell you the molecular member mechanism. dna is bent. the second thing is that how does it increase the specificity, because the different to bind in normal

dna, it's only hundred-fold, not enough to explain finding a mismatch in 10,000 of the normal base pairs. turns out it's an abc atp family. you would think maybe adding atp then the protein would bind to the mismatch more tightly

but exactly the opposite. there was all kinds of proposals that muts is the mode of protein, it moves away to finding new cleavage sides or doesn't hydrolize. the seamless structure that dna is bent, we don't think it can slide so easily, also if it

leaves mismatched what will happen to the mismatch is there a second coming, a huge mess there? and why shouldn't it stay? so the post doc designed this experiment to test whether muts needs mismatch to activate the repair process.

normally the dna are connected. there's a mismatch at the cleavage site, if the protein has to travel along dna hydrolyzing to come to the side to activate nuclease, enabled activating mismatch repair. and as a result, afte it requires bridge it.git, suggesting

later on john did a different experiment using human protein. he blocked the dna path using evidence and showed again that muts can activate in the presence of a blocker. the kinetics of activation is the same. from the high energy co-factor

used in protein synthesis, using gtp to increase the specificity of amino selection. so the idea actually was -- came out when i was talking about this problem. so we proposed this model that muts binds to normal protein and atp helps it release so it

has a chance to get the once there, it can recruit the two together now in the presence of atp can bind dna this data is generated by many labs. the model we proposed is still in the testing stage, and we still cannot exactly tell why

it banded to mismatched and is different. when the two together being shown by many groups can activate in the nuk nucleus and by us in either case. it was a positiveel for more than ten years.zzle for more a post doc in marty's lab and

worked with me as well solved the structure of human muts beta, recognizing dna that's not a simple missed pair but has large based loops that have no partner. and she work the diligently for four years and solved the muts complex with dna

contained two, three, four, and six, four crystal structures, the reason we solved many structures, sometimes crystal structures are not as unbiased, because of interaction, having poor structure we have a complete view. here is the original structure

of bacteria in humans, muts recognizes missed pair or single impaired base. this. the structure looks like7] the first view doesn't look that different except at the top like a corkscrew, now it's tighter, twined between the green and blue subunit.

this portion is is conserving all muts protein except when we were studying bacteria protein we removed it for easy crystallization, and in the muts structure this gets disordered. the real difference, i'll show you in the other slide when we

zoom in, here it doesn't really show. so there are a different number of base impaired in the loop. wei showing you two and four and six. what you see is when the loop size changes, the dna angle is and the angle is proportion to

the loop size. all the different sizes of loop are recognized in the repairs. exactly the dna shape may not be important, how much it's bent is not important. the second thing we learned from here is that muts even gives two base pair impaired,

same dna substrate, to impair the base also flipped out in opposite directions. so exactly where the mismatch locates relative to the protein is not even important, because they are repaired by different muts. what does it recognize?

clearly one common feature you see is that the dna is bent. it's bent by different angles. that comes to an analogy, maybe a good dna is just a good spring perfect. the lesion which is true for any dna lesion, when there's a lesion, the spring is no longer

as rigid. it welcomes more bendable. how do you spell the spring is perfect, that you can stretch or go back, if you stretch and broke it it won't be back. the best way is jumping out, give it energy to see if it can resist and bounce back.

if so, that's what we think muts is doing. we know what to look for, we find the difference. there are two subunits. on the right it's multi-color but remembers the blue sub-unit, dna is containing a on the left, it's the green

sub-unit. the difference is in the dna binding domain. you see that the one, this domain, blue, is way down. this subunit is not involving direct reading of mismatch. it's way up. it's like a crane.

straight standing up, the other leg is bent. the standing one is looking for bent, thedna can be wept and other is flexible. what's really interesting is this sub-unit in present in both humans muts, alpha and beta.

they have different confirmations. it's free to move without directly interacting with the able to move like this, that this affects hydrolyzed atp. this subunit is pulled down, the abc family is occluded for htp binding leads to the

explanation why the kinetics is we were one of the groups, we never published it because we were behind, but it's always a puzzle that muts will hydrolyze atp well. with normal dna, hydrolyze this. if you present a mismatched

dna, the birth plac first phase is eliminated, any step before hydroolysis is slow. we have an answer. it's binding to the normal dna, the dna is like a good screen, resists being bent. this domain is trying to bounce on it, and the dna wouldn't

budge. it's sliding along, the two atp hydrolysis goes on strongly. once it binds to a mismatch, it gets pulled down and it cannot bind atp well. the green sub-unit can bounce is slower.rolysis is joe we have a model how muts

identifies mismatch, because using the atp as a proof reader, it binds normal dna, tries to bend it, and dna is very strong, resists to bending, allows atp hydro lysis to go on strongly, sliding along the dna. and if it reaches the mismatch,

by direct association or by sliding, because the mismatch, the dna can be intent, the blue finds easy footing. once it gets stuck on this, the atp is inhibited, the hydrolysis is not enough to disassociate it, at least for a time duration, and by green

sub-unit cap capable of hydrolyzing the two recruit the next protein forming the pair complex. we are doing experiments, a lot of people are out in the world doing experiments, either to prove or disprove it, but so far it's satisfying green with

kinetic data and structural data. now i'm going to return to the point when we started, the cancer. how it related colon cancer and how it's helping knowing the molecular mechanisms to help understand the cancer

so it was really nice to have the whole pool of data, hundreds of thousands of data points, mutations found in the cancer patient. so once we have structure we can project them where they are. and wherever there's colon

cancer, that's functionally important. it wasn't that we always can go to the right place. we first found it, immediately it's noticed the atp site is because a lot of mutations cause cancer, and then this is a helix link to dna binding,

also caught our attention. only recently we noticed this bouncing effect of the dna binding. we appreciate this interdomain connection interface are so important for communicating the atp activity to dna binding, and allow the atp energy used

to test whether the dna is containing a mismatch or not. and furthermore, we know that now, that between the communication between two subunits, the ente interface is we can only propose maybe it's important for binding to the mutl or other factors in

that's how we use the data, in terms of helping, it's not a good drug target because it's malfunctioning, that's why we have cancer. to the structure of the molecular mechanism, now if we find a mutation, we can help to identify whether the mutation

actually is truly causing cancer or maybe it's just a polymorphism. it heap helping diagnosing. wei a long way from why certain mutations cause colon cancer, in women it's ovarian or brain there's a long way to go. understanding molecular

mechanisms empowers us to fight the war against cancer. i switch gears now for the next ten or fifteen minutes on the dna preliminaries of translesion synthesis. dna are not always perfect. you need special preliminaries to b bypass road blocks.

when roger woodgate found a new one, he approached me. we talk solved the structure. wei telling you a good preliminary set, makes very few errors and makes many errors. the difference is in the active site. here you can see an active site

represented by two blue series, here you cannot see it because the good preliminaries is hidden inside. it's much more enclosed and very much more strongly discriminating against the round base or not good template.

this is not exactly telling us how humans work. in humans, people who have mutations in that gene have high sensitivity to uv light. and normally you know we go out, enjoy the sun, because it helps us strengthen the bones, converting cholesterol to

vitamin d. but those people who have deficiency to repair uv lesion, they are sensitive and have high incidence of skin cancer. what uv does to us, ultra violet, the adjacent base points, eachthree months, can serve as a template for dna

they get cross linked, two bases linked together like a siamese twin. 20% of two bases become 90 degrees to each other, six-four product. because this is a more different from normal, actually the repair of excision, a bert

substrate to be khraoef cleaved off. it's harder to be removed by dna repair, the moving crew. there are eight genetic groups that mutations cause people to be over sensitive to sunlight. seven of them are involved in this repair pathway to remove the lesion, and the disease,

abcdefg, all greens involved, synthesizing.d the roll of that, in the sell cycle, during the synthesized time, it's very dangerous, cutting the dna to pieces. so we need a special preliminary in humans to bypass one, two, three, four steps,

insertion based, extending it. it's been shown the most faithful and most efficient in crossing the barrier of the there are a lot of preliminaries, they need signals to recruit the cells, special preliminaries to the replication forecast.

it's important for pcna when cells encounter sustained uv lesions, and pcna, it can recruit the special preliminaries to bypass this road block. and it goes through four steps of synthesis and hands over the fine replication fork back to

the about primaries. so we were interested. because human pro teen i protein is more difficult to geta post doc and they spentrn, mark, he they two on or three months. the protein dna in traction wasn't the native enter action because it distortedtion

interaction, catalytics. i'll show you the difference between not so perfect dna complex and perfect dna the protein moves subtly. but to get the right structure we spend another two years and that's a team of a post doc and student, years out.

they spent two years, engineering new crystal ladders so we can see when it's in the catalytic mold, showing how it accommodates the dimer in the active site, not making any errors. normal preliminaries wouldn't take two base pair into an

active site. by having this it won't be able to close. and they solved three more structures of the lesion with dna, each step of bypass, crossing this phase, the next phase, the next phase. i won't bore with you the

structures ba because they look quite similar. iswing you here how the dna up superimposed on normal dna, they look very normal. about five or ten minutes ago i was telling you a good dna is like a good spring, the lesion is a broken spring, the repair

protein here what the broken spring look likes. with a uv lesion being repaired by the repair protein, it will exploit this flexibility and the in bent dna. somehow when it's complex with this lesion bypass dna, the dna actually was able to keep dna

straight. the reason that if you are building a wall you want the bricks laying scarily on top of another, if they are leaning over you can't build more. they are using extensive protein and dna interactions to keep this dna normally that's

bent straight for the chemistry to take place. we liken this as a molecular splint. the protein preliminaries will keep the dna straight in order for dna isn't 'tis t synthesis to occur. at least five mutations that mpv patient, the pigmentosut

patientses, the mutations affect the sturdiness. this supports this protein backbone. in the last couple slides i will turn the attention to the application of this study of this special polymerase. it's good to help us to guard

against the lesion caused by uv, but it plays a negative role in treatment of cancer patients, it's broadly used chemotherapy agent. it cross-links, the thymine dimers here, it cross-links to adjacent, and kills cancer cells because so many of the

lesions here, the cell will be stopped in replication, fast-dividing cancer cells. it turns out that the study based on cancer-treated patients shows if you have high levels of expression level, shown here, this curve, the surviving time is much shorter

than if you have low expression level because it probably bypasses, causing the lesions, like uv lesions. in stage three disease the difference is even more severe. if you have lower, patients respond much better to the treatment and live longer.

and biochemically, other groups as well as us have shown it can bypass lesion efficiencily inserting against 80% or 85% of inserting against normal it drops a little when it goes on the next base and drops much further when it's on the third and fourth.

the first two are the hardest ones to get by themselves. so we were looking for ways to inhibit it and combine platinum-based treatment, the patient will probably respond better with lower dosage and survive longer. and it was all in the talk,

people in rockville, the drug screening group and also roger wingate talking about it, but we find something in the crystal structure, so we were crystallizing with pretty high concentration incoming, and we found a seconde second binding site away from the active site.

we were kind of bothered by this density, how to get rid of if we stop the crystal with no incoming nuk, and what's more interesting is here, we find normally when you increase the substrate concentration, the activity goes up. with this, when we increase 50

micro molar, i'm getting inhibitus. inhibition can be removed with the base of the incoming, if we mutate the tryptophan, not conserved in the y family polymerase, now the activity we think because the interaction with dna is affected, but it's

no longer being inhibited. so this may be a good target for finding new drugs to inhibit and use for cancer therapy. i'll conclude my talk leer, so i showed you about dna lesions which can be caused by external effects or environmental

agents, or just because they make an error leading to increased flexibility in dna. it's used by both repair enzymes, assisted by atp for repair, or fortified by special preliminaries to carry out translesion dna synthesis. in both process, the outcome is

prevent the mutation and reduce the cancer incidents. the site may be leading us to think about developing drugs for inhibition and helping effectiveness of cancer treatment. so i've acknowledged, i talk about the work and here is the

current group. we have pictures taken together with marty's group, we have extended labs, a lot of people's work, i'm not mentioning here because they are still in the working with some larger complex. i mentioned by former people,

also former post doc students, collaborators, and we collaborated with' in japan, a lot of wonderful clack lators collaborators, many you recognize from nih, supported by intramural funding. lastly but not least i want to acknowledge laboratory of

molecular biology at niddk, many are here. i feel sort of highly pressured and nervous because i'm just representing our group, because all of them, everyone of them have a wonderful story to tell, it's a great group to long to. the senior member inspires me

by doing the signs the way they do, the action, they show what's important, and just great fun to be a member of this family, and also belong to other groups, very supportive, i have a lot of positive and interesting interactions with members and also the society of

chinese scientists in america, nih, we have sporadic interaction and very supportive in forming a lot of conversations and hopefully productive collaborations in the future. thank you all. [applause]

>> questions? >> yes. great talk. my question is whe how it interacts with muts? >> you mean the lesion search? >> right. it cannot be sort of sliding on dna forever.

it's only sliding on very short distance. it has to be -- it's a combined 3d, 2d search, the association is important. and then we'll hit another spot. >> thank you. >> yes, usually combined 1d and

3d movements in the nucleus. >> all right. have the structure given you any hint as to how the complex can tell the old from the new dna strand? in bacteria, e. coli, in some it's clear because the template in e. coli is methylated, the

daughter stand is not methylated. the nucleus will be cutting the methylated strand. any other bac bacteria it's not clear. in humans the consensus, moves, com polymerase doesn't fast, the tren strand is recognized.

this is still -- the jury is still out because we haven't found the nucleus cleaving the daughter strand specifically, it's more recent that mutl can caught daughter strand, you have a three-way interaction, the pcna, the loading clamp, will be loaded in certain ways

relative to the primary strand. i need helpers to differentiate two-strand. >> so i'm afraid this is going to be a rather arcane question. but it comes from my interest in heteroduplex dna, recognizing a cc mismatch is one of the most destabilizing

of the dna duplex. yet it is the mismatch that is not recognized. >> not recognized well by muts, you know why? >> i don't know why. that's why i'm asking. >> the answer is two sides of the coin.

polymerase doesn't make c.c. mismatch, because cc is so unstable, the polymerase doesn't make it much. it's the same side. two sides of the coin. it doesn't make much. it's human made. >> but in terms of the

molecular mechanism, in terms of what you said in terms of probing the dna structure for flexibility, i would have expected it to be the most flexible. >> it's some what different. no, it's different. >> okay.

>> yeah, so normally the base would be stacked weaker, like doing this. but it's so bad most likely they are pushed out, and adjacent base becomes stacked. >> that becomes more stable. and because the base pushed out, not well recognize the by

muts but but other pathways. >> that was beautiful. so with the story where you show the splint straightening the dna, but before pol-a get there, the dna is bent >bent. >> the dna is not always bent. the increased flexibility means

it isn't -- normally the dna -- if it's half a lesion, it does it could be straight. it's very little. so the polymerase captured it, went straight and captured it that's most of the same. probably 95% of the time it's not induced to fit.

the confirmation is ronald there it's not stable and gets trapped. >> before i thank wei, i wanted to make a couple announcements while i have the captive audience. the first is that the wednesday afternoon lecture next week is

on thursday. so change your calendars accordingly. secondly, we had another interesting event here at nih, which was sponsored by the journal of science, a web that so called super-resolution imaging, not the resolution of

x-ray crystallography but pretty good. nih today recorded discussion with harry shroff, and claire waterman, and that webinar will take place officially on wednesday, february 29th at noon. so if you're interested in

checking in, i think you might find that an interesting event. so i just wanted to say that in discussing this lecture before wei got on the podium, she told me she felt the burden of representing the laboratory of molecular biology which is one of the gems o of the nih, the

niddk and women scientists, chinese-american scientists, and she was worried she would acquit her in the best possible way and i can eye sur assure you, you've done a magnificent job. thank you very much.