Inhibiting gene mutation

eric lander: good morning. good morning. so last time, we raninto a problem. we had mendel, my hero mendel,this mit-like mathematical, physical monk, had developedthis gorgeous theory of particles of inheritance. he didn't use theword "gene" yet. gene doesn't get inventedfor much longer. for every trait, you hadtwo such particles.

you gave one to youroffspring. each of the parents gaveone to their offspring. and that's how each offspringgets two of them. that choice of which of thetwo alleles to transmit to your offspring isa random draw. and that explains beautifully,for example, the 3-to-1 segregation patternthat mendel saw-- gorgeous. we put that model, which wasan ex post facto model, a

model made after the data wereavailable, to a test. you guys insisted we had totest it before you would publish it. and it holds up pretty well,making pretty surprising predictions that youwould otherwise not have ever expected. like amongst that 3 to 1, thethrees are not all the same. some of those round peaswere homozygous for the big-r allele.

and in selfing, they'll neverproduce wrinkled peas. but 2/3 of those round peaswere heterozygous. and when you self them, youget 1/4 wrinkled peas. so the wacky prediction that1/3 of the rounds will give rise to no wrinkleds, and 2/3of the rounds will give rise to 1/4 wrinkleds is a surprisingprediction and, therefore, something that bearsstating as a scientific prediction. and so all those kind ofpredictions can emerge.

then we turned to the questionof two factors. i'm practicing usingour words here-- homozygous, heterozygous,alleles, et cetera. we now turn to two traits,two phenotypes. we have two phenotypessegregating. we had round and wrinkled, andwe had green and yellow. and mendel determined, ratherbrilliantly, that they segregated, were transmittedindependently of each other. there was no correlation betweenwhich alleles you got

at round and which allelesyou got at wrinkled. that was pretty cool. let's just go over that, becausethere's a real tension to be resolved because we havemendel's second law versus the chromosome theory. so mendel's second law-- let's just go back to it-- inthe f0 generation, we had our round green peas, genotype bigr, big r, big g, big g. we had our wrinkled yellow peas,genotype little r, little r,

little g, little g. we cross them together, we getf1 double heterozygotes, big r, little r, big g, little g. we then perform a back crossor test cross to the doubly homozygous parent with the tworecessive phenotypes-- we're practicingour words here. and what do we get? well, we get certain options. as we said, the gametes thatemerge from this parent on the

left could be of thefollowing types. the gametes from theparent on the right are all those alleles-- are those the recessivealleles? no. you'll forget this over time,but just to make me comfortable-- they're actually the allelesassociated with the recessive phenotype that we'retalking about.

because you know-- but willforget, i assure you, because all my colleagues in thebiology department have forgotten-- that they could also controlmultiple other phenotypes, some of which couldbe dominant. but that's ok. i forgive you in advancethat you'll call them recessive alleles. anyway, you get this.

and then these should occur atequal frequency of one to one to one to one. all right. now, we had the chromosometheory. the chromosome theory, theobservation of the choreography of chromosomesduring meiosis-- chromosomes in meiosis-- looks like this. we have chromosomes liningup in pairs.

now, i'm not drawingit terribly well. but the two members of eachpair are of the same size. but this pair could bebigger, and this pair could be smaller. it looks like they're reallyfinding each other. these chromosomes are visiblydifferent in shape. and so maybe i'll make this guya little longer, just to indicate that, thatit actually knows. now an explanation, we said, forhow mendel's second law of

independent assortment of twodifferent phenotypes could occur is that, forexample - round. it could be that the gene forroundness is located on chromosome number one. these are two copies ofchromosome one that have been duplicated, each of whichhas the big-r allele. next to it, two copies ofchromosome one that have been duplicated, each of them havingthe little-r allele. over here on chromosomenumber two.

let's say, lies the genefor green or yellow. and in this picture here, thebig g's are on the two copies of this chromosome numbertwo that are here. the little g's are on the twocopies of chromosome number two that are here. when the cell undergoes meiosisone, we get to the situation where we have big g,big g; little g, little g. we've got big r, big r;little r, little r. could it have been the casethat the big g was on the

right and the littleg was on the left? yeah, of course. it's totally independentwhich way. i happened to drawit this way. but with probability 50%,it's the other way. that's why they're independentof each other. and then when it undergoesmeiosis two that looks just like mitosis, we end up withour four gametes here. sorry, our four gametes with bigg, big g; little g, little

g; big r, big r; littler, little r. and that accounts for the big g,big g; big r, big g; little r, little g gametes. and then when they went theother way, the organism would make a set of gametes that hadthe other combination of big r's with little g's. so that's perfectly fine. and because the secondchromosome is independently ordered compared to the firstchromosome-- when they line up

at the midline, theydon't really care which way they are-- that'll account for oneto one to one to one. it's so straightforward. but what happens if, instead,both the roundness gene and the greenness gene live onthe same chromosome? we'll have little r,little r there. we'll have big g, big g here;little g, little g. and then when they split,we'll end up with--

now, chromosome two hasnothing that we care about on it. it has a lot of genes. in fact, there could bechromosomes three, four, five. i'm just not drawing them. but we're really goingto focus on chromosome number one here. and you'll notice that here onchromosome one, the big g's or the little g's are coupled,physically coupled to each

other, the bigs withthe bigs and the smalls with the smalls. so now the kind of gametes thatcan emerge from this will only be big r, big g type or-- big g type-- or they'll be littler, little g type. we can't get the reversecombination. we can't get bigs and littles. so this, because these areindependent of each other,

will get us one to oneto one to one. these are the big, little,and then these other combinations like that. this will get us only, if welook at the big r, big g; little r, little g; big r,little g; little r, big g; will get us one to oneto zero to zero. let's give a name to this type,the big r's and the little g's. let's call that a recombinanttype, ok?

i'm just going to use thatword for the moment. the recombinant types-- thebigs and the littles, and littles and the bigs-- we're not going tosee any of them. this is a very strong differencebetween mendel's second law and the chromosometheory. if the chromosome theory isright, and these chromosomes are physical entities that haveintegrity, they can't both be right.

so that's a great thing inscience, when you have two different models and they can'tboth be right, because you learn things then. you can test them. now, mendel tested this withoutactually knowing the and he always got one to oneto one to one for the seven traits he looked at. was he just lucky that theyhappened to lie on the seven different chromosomes?

or is there some problemwith this chromosome theory, or what? it took a while. and then, of course, everybodyforgot about mendel from 1865 until the year 1900. in the year 1900, people beginto rediscover mendel. cytology has come along. in january of the year 1900,plant breeders start rediscovering mendel.

sorry. thank you. i see already. i could tell by the look on yourface that something must be wrong there. good. plant breeders startrediscovering mendel. and in january of 1900, threedifferent groups say, you know, we found these laws.

and they're just likemendel's laws-- which now everybody startspaying attention to. but plant breeding-- and peopletried to do mice, and people tried to do rats. what turned out to be thewinner, the place to really study genetics, wasthe fruit fly. thomas hunt morgan at columbiauniversity decided, after he was really frustrated wastingyears breeding mice and rats that just took too long, around,oh, i don't know, 1906

or something like that, began tostart breeding fruit flies. fruit flies are the teeny littleflies that when you open a banana or fruitsor other kinds of things, you'll see them. and studying drosophila gave usthe answer to this question of what the problem is, howcan it be that either it's independent or totallydependent? so we're going to talk aboutdrosophila melanogaster, the fruit fly, and the discoveryof recombination.

so now, morgan-- i'm going to now start usingfruit fly notation to give you some practice with fruitfly notation. we're not going to use bigg's and little g's. they like to refer tothe normal allele and the mutant allele. the normally allele is plus. the mutant allele getssome kind of a name. and so he had a female fly thatwas normal and normal at

two different loci, twodifferent genes. i haven't told you whatthe genes are. and the male fly had ablack-colored body-- black across its whole body-- and its wings wereshriveled and, therefore, called vestigial. so the phenotype here is blackand vestigial, black body and vestigial wings. and this wasn't the normalappearance of a fly.

so he took these females bythese males, and he crossed them together. and he got an f1. and the f1 was black overvestigial, plus over plus. and what was their phenotype? were they normal appearance,which is kind of a sandy-colored bodyand normal wings? or were they this all blackbody and vestigial wings? turns out they were normal.

from that, what do we inferabout these two traits, black body and vestigial wings? they're recessive traits. so here, the phenotype wasnormal appearance. so then he crosses them back,doing a test cross. let's say he'll take males hereand females here, but it actually works either way. and what he does is justlike we did there. he could get gametes thatwere black, vestigial.

he could get gametes that wereblack, plus; plus, vestigial; or plus, plus. those are the four possibilitiesthat come out. so when he does it--let's keep score. i'm going to write them now-- plus, plus; black, vestigial. and from the other parent,you got this-- black, plus; black,vestigial; plus, vestigial; black, vestigial.

those are the fourpossibilities. and if this was just likemendel's traits, it would be one to one to one to one. these were the parentaltypes that went in. plus, plus went in. and black, vestigial went in. those were the combinationsof traits here. these were new combinations. what he observed--

let's see. if mendel's right, it'll beone to one to one to one . if the chromosome theory'sright, it'll be one to one to zero to zero. and who was right? student: no one. eric lander: no one. the answer was 965 to944 to 206 to 185. neither model was right.

neither model's right. the new combinations, therecombinant combinations, the non-parental combinations-- we can call these recombinantcombinations. these were recombinant. they recombined in some way. they were a new combination,or they were the non-parental types. we use both of thosewords frequently--

were neither equal nor werethey completely absent. they occurred, but ata lower frequency. what was the frequency? well, we could just add it up. the frequency of recombinanttypes, of new types of were different than theparents', is 17%. what's going on? now, maybe this is some magicnumber like the 3-to-1 ratio. and you should look at that 17%and say, ah, this is some

constant of the universe, thatwhen you put in traits you get 17% percent of recombinanttypes. but it takes a little judgmentto say, i don't think so. and he actually triedother traits. and sometimes he got oneto one to one to one. but very often he gotsome funny number-- 6%, 28%, 1%. there was some funnybusiness going on. recombination.

that's what's goingon is there is recombination occurring. what do we mean byrecombination? recombination is very importantstuff, by the way. at some point, i will tellyou that understanding recombination was actuallythe origin of the human genome project. and it traces back toa good mit story. but that will be for a littlelater in the semester.

so what do we think'shappening here? i'm now going to draw a close-upof that chromosome. and here's another chromosome,the other pair. and what we think here might behappening is that you might have plus and black;black and plus-- oh, sorry. black, right? black, black; plus, plus. this would be plus.

this is the normal chromosome. this is plus. so this chromosome herecarries the pluses. this guy here carries thosealleles black and vestigial. well, what happens, the ideawas that somehow these chromosomes exchangedmaterial here. and the chromosome that wasplus, plus; plus, plus now somehow acquires that bit,and this chromosome somehow gets that bit.

and we end up instead with apicture like this, where some of this came from here. and those two loci, black andvestigial, were separated from each other such that the blackallele moves over to that chromosome. is that clear? that's the notion. why did they thinkthis was true? well, it turns out that in fruitflies, you can actually

look at eggs underthe microscope. and you can look atthe chromosomes. and if you take if you takethe cells, and you take a cover slip, and you squash itdown with your finger, you can actually see chromosomes lyingright over each other, making little crosses like i drewthere, up there, called chiasmata, whichmeans crosses. and so people said, see, in themicroscope, you can see they're lying on topof each other.

are you impressed by thatpiece of evidence? you took the cover slip,you squished it down with your fingers. so they're lying ontop of each other? big deal. i'm not going to be impressed. and calling it chiasmata doesn'tmake me any more impressed, right? although it's always good tocall things greek names,

because people think they'remore sophisticated if you call them greek names. but this was the notionpeople had. and the frequency 17% wouldindicate how often these crossovers occur. but if i were in the situationwe talked about with mendel, and i wrote this up, and i said,see, it's 17% sometimes, 6% sometimes, 28% sometimes;and when i look in the microscope, they lie on top ofeach other; therefore, it's

recombination-- there were actually other ideasfloating around too. maybe it has something to dowith developmental biology. it was a puzzle. when you have a really deeppuzzle, the most important thing in science is tofind a young person. because young people comewithout prejudice. they say, let me justlook at all of this. stand back.

i don't come withany prejudice. so at mit, what is the solutionwhen you have a problem like this? a urop. you want a urop. so even 1911, that was thesolution at columbia. thomas hunt morgan got a urop. i'm serious. he was called alfredsturtevant.

alfred sturtevant was asophomore at columbia. everybody else was busy findingthis recombination data, how often this recombinedwith this, this with this, this with this. sturtevant-- he's a sophomore-- he says, god, this stuff'sinteresting! professor morgan, could ihave all the data and try to look at it?

sturtevant took it homeand actually pulled an all-nighter, blew offhis homework-- it actually says so inhis autobiography. he says, i blew off my homeworkand pulled an all-nighter-- essentially in those words, hesays. "to the detriment of my undergraduate homework"is the way he puts it. but in any case, genetic mapsand sturtevant's all-nighter. what sturtevant didwas the following.

he said, how are we going toprove the chromosome theory? i like this idea thatrecombination is about distance on the chromosomes. i like the concept thatmaybe 17% is how often these things recombine. why would things only recombine1% of the time? what would that mean? they've got to be pretty closetogether so that a crossover between them happensnot so often.

and what if thingswere far apart? well, it could be more likely. so he likes the ideathat recombination frequency means distance. but how are you goingto prove that? it could mean a zillionother things. it could mean biochemicalpathways, developmental biology. how are you going to prove thatrecombination frequency

means distance? you've got to make predictions,right? the only to do it would beto make predictions. so sturtevant takes the data,and he starts making predictions. i think, says sturtevant,these things are alleles living at genes with locationson the chromosome-- black, vestigial. how often do blackand vestigial

recombine with each other? what is the frequencyof recombinant non-parental types? 17%. so sturtevant goes through thedata and he says, what about other crosses peopledid in the lab? well, it turns out people didcrosses with another mutant that produces a funny eye colorcalled cinnabar, cn. so cinnabar.

it turns out that therecombination frequency between cinnabar tovestigial was 8%. vestigial, cinnabar,8% recombination. if this chromosome businessis right, where should i put cinnabar? sorry? where do you want it? student: [inaudible]. eric lander: somewherein between.

you'd like me toput that there. student: on the other side. eric lander: oh, ok. wait a second. on the other side. ok, which is it? how many vote for the left? how many vote for the right? how many conscientiousabstainers are there?

do we know? student: no. eric lander: no. there are two possibilities. it could be 8% this way, orit could be 8% that way. how are we going to know? yes? eric lander: black. what if we knew the answerbetween cinnabar and black?

that would constrainthe problem. can you give me two predictionsfor what the answer might be? student: either 9% or-- eric lander: either 9% or 25%. so now we have a prediction. we don't know wherecinnabar is. but the answer could be thatblack to vestigial should either be about 9%--

that's what thatwould be here-- or about 25%. the answer? about 9%. that's what sturtevant found. that's a predictionand kind of cool. and you can imagine taking thedata home and it's probably 9 o'clock, and you'venow realized, wow, freaky, it's 9%.

so then he looked at themutation "lobe." lobe was another mutation. the lobe mutation showed 5%recombination from vestigial. where should we put it? left or right, well, let'smake some predictions. suppose it's over here. will it be very closeto cinnabar? will it be closer to black? but what if it was over here?

well, it would be further. so let's put lobe in. and suppose we knowthat lob is 5%. then what's the predictionfor black to lobe? 22%. answer, according to thenotebooks, 21%, pretty close. you'd like it tobe exactly 22. but life doesn't alwaysturn out that way. 21's pretty close to 22.

what other predictionscould we make? if this is cinnabar here, couldyou give me a prediction for cinnabar to lobe? 13%. yep, that works. curved wing. recombination distancefrom lobed, 3%. so now you have somepredictions. you have this prediction here.

you predict 8%. answer, about 8%. over here you predict 16%. answer, about 16%. bingo. sturtivant says, ifthese genes-- we call them loci, often. i'll use the word locussynonymous with gene. locus means a place.

and geneticists think aboutgenes as a place on a if these loci-- the plural of locus-- if theseloci were really arrayed along a linear structure, then itwould have to be the case that they would have certainadditive relationships between them. and the chance that they wouldhave these additive relationships if they weren'tpart of a line is pretty implausible.

that's a real prediction, a veryremarkable prediction. and it holds up with the data. sturtivant pulls theall-nighter. by the time the sun comes up atcolumbia university-- this is morningside heights-- he's got the wholething worked out. yes, this chromosome theorymust be right. it fits beautifullyall of these data. pretty cool.

you are all authorized to blowoff your homework anytime you make a discovery like that. [laughter] that's a course rule. any homework willbe forgiven for discoveries of that magnitude. tell your tas. so now, what does itreally tell us? it tells us that if genes arevery close together, the

recombination frequency, rrecombination frequency, or rf, might be very little. they could be as low as almostzero, which means you never see a recombinant becausethey're right next to each other. or it could be that they'refurther apart. it might be 1%. it might be 10%. it could keep growing.

it might be 30%. suppose it's way, way, way,way, way far away. what's the largest itever gets to be? well, if they were on differentchromosomes-- suppose there were totallyindependent segregation, different chromosomes-- what would they be? student: [inaudible] eric lander: no, it's 50%because remember, one to one

to one to one says thatthe recombinant-- well, actually, this is aninteresting question. let's come to that 100. on different chromosomes, oneto one to one to one means that it's 50%, half of themare recombinant types. it turns out on the samechromosome, as you get farther and farther and farther, youmight say there's going to be 100% chance of a crossover. and you might say therecombination frequency could

keep growing past 50%. it turns out it doesn't. the reason is that multiplecrossovers can happen. so mathematical interjectionhere, if here's my gene and here's my gene, there couldbe one crossover. there could be two crossovers. there could be threecrossovers. and so in fact, it turns out, asyou get very far away, you have to start paying attentionto the probabilities of double

crossovers and triplecrossovers. and so it turns out that it'sa poisson process of the number of crossovers thatoccurs, give or take. and you see a recombinationif there's an odd number. and for that reason, itnever gets above 50%. so as the distance gets furtherand further and further, it goes from zero to50, which is the same number you get for separatechromosomes. otherwise, you might think thatif there was just one

crossover, it gets to 100%probability of recombination. but it never does, becausethere are doubles. and you can actually observe,if you make a cross that has three different genessegregating in it, you can actually see the doublecrossover type. so you can see very nicelythat if you make a cross involving black, cinnabar, andvestigial over plus, plus, plus, you can see that cinnabaris right in the middle because thisrecombination happens at a

pretty good frequency. this recombination happens ata good frequency, giving you plus, cinnabar, vestigialor black, plus, plus. sometimes you will getthat recombination. you'll get out gametes that areblack, plus, vestigial, but at a much lower frequencybecause they take two crossovers. what will be the probabilityof seeing a black, cinnabar, vestigial?

well, we said that thisone was about 9%, this one was about 8%. what's the product of a 9%chance and an 8% chance? it's about a 1% chance, a littleless than a 1% chance. that how frequently you seeblack, plus, vestigial. you can even predict theprobability of a double crossover event by multiplyingthe two events that have to happen. so your bottom-line rules hereis that because of these

double crossovers ourrecombination frequency can go from zero to about 50%. this is independentassortment. and it either occurs if you'reon different chromosomes or, if you're very far away on thesame chromosome, they behave as if they're independentof each other. any questions aboutany of that? student: why didn't mendelever see recombination? eric lander: why didn't mendelever see recombination?

it turns out that with sevenchromosomes, and they're biggish in length, he neveractually ran into two loci that were close enough,to notice it. that's why. flies actually only have threemajor chromosomes. there's a fourth, but it'sa puny little thing. and because they were much moreintensively collecting mutations in morgan's fly room,they began having a lot of them, and theyhad to bump into

recombination pretty early. mendel simply didn't have enoughthat were close enough. think about what would havehappened if just by chance, somebody were selling a strainof peas in the market which had a mutation in a locus thathad 10% recombination distance, and it screwed upmendel's law of independent assortment of two loci? mendel might not havepublished the paper. sometimes in science it'sactually valuable to first get

the oversimplification out ,like his second law, and then deal with the complexitythat sits on top of the oversimplification. it's kind of lucky that mendeldidn't have enough of them to be bothered, in thefirst paper. so in fact, all of mendel'sseven loci have now been mapped. most have been clonedmolecularly. and so we actually know wherethey are, et cetera.

it's a really good question. that's sort of why he didn't.