WEBVTT

00:00:01.000 --> 00:00:05.000
Good morning.
Good morning.

00:00:05.000 --> 00:00:10.000
I don't know about you, but I can't
take too many more nights like this.

00:00:10.000 --> 00:00:15.000
I confess, I haven't gotten a thing
done for so many nights in a row now,

00:00:15.000 --> 00:00:20.000
but what a game! How many of
you saw the game? Excellent.

00:00:20.000 --> 00:00:25.000
Very good, very good. You have
your priorities straight in

00:00:25.000 --> 00:00:30.000
the world. Very good. Well,
if it's possible to get your

00:00:30.000 --> 00:00:35.000
minds off Curt Schilling last night,
and off more importantly tonight.

00:00:35.000 --> 00:00:40.000
Perhaps we can spend a bit of time
this morning in the meanwhile with

00:00:40.000 --> 00:00:45.000
whatever spare neurons you have
talking about recombinant DNA for a

00:00:45.000 --> 00:00:50.000
bit, OK? What we talked about last
time was different ways to clone

00:00:50.000 --> 00:00:55.000
your gene based on its properties.
We started off with cloning by

00:00:55.000 --> 00:01:00.000
complementation, right,
the idea that if you took a

00:01:00.000 --> 00:01:05.000
library of clones, you
would be able to put it into

00:01:05.000 --> 00:01:10.000
bacteria and select a bacterium
whose phenotype had been restored by

00:01:10.000 --> 00:01:14.000
virtue of having the plasmid.
You would complement the defect.

00:01:14.000 --> 00:01:18.000
You'd find the clone you wanted
because it complemented the defect.

00:01:18.000 --> 00:01:22.000
That's great if you can put it
into an organism that has a defect.

00:01:22.000 --> 00:01:25.000
You can do it with bacteria.
You can do that with yeast.

00:01:25.000 --> 00:01:29.000
It's harder to do with large
organisms because you can't inject

00:01:29.000 --> 00:01:33.000
enough of them with different clones
to be able to make that practical

00:01:33.000 --> 00:01:37.000
unless you're working in cell
culture or some very small,

00:01:37.000 --> 00:01:41.000
fast growing organism. We
talked about being able to use a

00:01:41.000 --> 00:01:45.000
protein sequence, reverse
translating that protein

00:01:45.000 --> 00:01:50.000
sequence in the computer from amino
acid sequence to nucleotide sequence,

00:01:50.000 --> 00:01:54.000
and using the nucleotide sequence to
design a probe to hybridize back to

00:01:54.000 --> 00:01:59.000
the genome. That works fine
if you have a protein sequence.

00:01:59.000 --> 00:02:02.000
But the last topic we talked about
that I wanted to just touch on again

00:02:02.000 --> 00:02:05.000
this morning was suppose you were
trying to clone the gene that causes

00:02:05.000 --> 00:02:08.000
a certain human disease,
and you have no idea what the

00:02:08.000 --> 00:02:11.000
protein was. Then, you
can't use its amino acid

00:02:11.000 --> 00:02:15.000
sequence because you don't have the
protein. What can you possibly do

00:02:15.000 --> 00:02:18.000
when all you know is that you have
a gene which causes a genetic defect

00:02:18.000 --> 00:02:21.000
that causes a disease? And I
said you could clone it using

00:02:21.000 --> 00:02:24.000
the ideas of genetic mapping,
position, the things that Sturtevant

00:02:24.000 --> 00:02:28.000
developed. And, I
touched on it briefly,

00:02:28.000 --> 00:02:31.000
and I want to just touch on it a
bit more because some people had some

00:02:31.000 --> 00:02:34.000
questions about it. And
I've set up a very simple

00:02:34.000 --> 00:02:38.000
example to show you. Suppose
that, to make it easy,

00:02:38.000 --> 00:02:41.000
we're working in a fruit fly
first. We're working drosophila, and

00:02:41.000 --> 00:02:44.000
suppose that the true picture of the
underlying chromosome is like this.

00:02:44.000 --> 00:02:48.000
There's a locus that could either
have a mutant allele M or the wild

00:02:48.000 --> 00:02:51.000
type allele plus. There's
a bunch of other loci along

00:02:51.000 --> 00:02:54.000
the chromosome. And, let's
suppose we know all of

00:02:54.000 --> 00:02:58.000
where they are and all that.
And, they have two alternative

00:02:58.000 --> 00:03:01.000
alleles. At this locus
the alleles are orange

00:03:01.000 --> 00:03:05.000
or pink. At this locus I'll
call the alleles orange or pink.

00:03:05.000 --> 00:03:08.000
Now, these are different loci.
These are different alleles. I've

00:03:08.000 --> 00:03:11.000
just called them orange and pink in
both cases so I don't have a rainbow

00:03:11.000 --> 00:03:15.000
of colors up here to confuse
us. But all I mean is there's two

00:03:15.000 --> 00:03:18.000
possible alleles here, two
alleles here, two alleles here.

00:03:18.000 --> 00:03:21.000
This is the diseased
gene we're interested in,

00:03:21.000 --> 00:03:25.000
and these are passive markers.
These are other markers along the

00:03:25.000 --> 00:03:28.000
chromosome. If we were
to set up a cross between

00:03:28.000 --> 00:03:32.000
heterozygotes, a
heterozygote here,

00:03:32.000 --> 00:03:36.000
and a heterozygote here, and
it were the case that on the

00:03:36.000 --> 00:03:40.000
chromosome bearing the mutant allele,
it happened that at these three

00:03:40.000 --> 00:03:44.000
markers we had orange alleles.
I don't know what they are, but

00:03:44.000 --> 00:03:48.000
whatever these orange alleles are,
they might be a visible phenotype,

00:03:48.000 --> 00:03:52.000
forked or yellow or bristled.
They could be a DNA sequence

00:03:52.000 --> 00:03:56.000
difference. They could be whatever
you want, but let's suppose the M

00:03:56.000 --> 00:04:00.000
chromosome has a set of alleles that
are different in each location than

00:04:00.000 --> 00:04:03.000
the plus chromosome. Then,
when we look at the offspring

00:04:03.000 --> 00:04:07.000
that come out of this cross,
let's only, for the sake of

00:04:07.000 --> 00:04:11.000
simplicity, look at those offspring
who are homozygous mutants.

00:04:11.000 --> 00:04:15.000
Well, in general, if there's
been no crossover here,

00:04:15.000 --> 00:04:18.000
then the M chromosome
will have orange, orange,

00:04:18.000 --> 00:04:22.000
orange, orange, orange, orange.
If there's been a crossover,

00:04:22.000 --> 00:04:26.000
however, it could go orange,
orange, pink on one of those

00:04:26.000 --> 00:04:30.000
chromosomes. Or if there's
been crossovers like this,

00:04:30.000 --> 00:04:34.000
it could go orange, orange, pink
on one chromosome, and orange,

00:04:34.000 --> 00:04:37.000
pink, pink on the other
chromosome. It could even,

00:04:37.000 --> 00:04:41.000
in the extreme, have had
crossovers very close to

00:04:41.000 --> 00:04:44.000
the gene maybe here, and even
maybe here. And you've got

00:04:44.000 --> 00:04:47.000
orange, pink, pink,
and pink, pink, pink.

00:04:47.000 --> 00:04:51.000
But if we look at the many
segregates, you know from genetic

00:04:51.000 --> 00:04:54.000
mapping that the closer the
locus is to the disease gene,

00:04:54.000 --> 00:04:57.000
the more strongly correlated
the inheritance will be,

00:04:57.000 --> 00:05:01.000
the tighter the linkage will be.
This is nothing more than linkage

00:05:01.000 --> 00:05:04.000
mapping. But now,
suppose we were doing

00:05:04.000 --> 00:05:08.000
linkage mapping, but for
the sake of argument the

00:05:08.000 --> 00:05:12.000
whole genome had already been
sequenced. Suppose the genome had

00:05:12.000 --> 00:05:15.000
been sequenced in a cross, and
the whole genome of the fruit

00:05:15.000 --> 00:05:19.000
fly had been sequenced
which it has been sequenced.

00:05:19.000 --> 00:05:22.000
And, we looked at a cross
and we looked at the mutants.

00:05:22.000 --> 00:05:26.000
And what we did was we tried
different positions along the genome.

00:05:26.000 --> 00:05:30.000
And at each position, we
had some genetic marker.

00:05:30.000 --> 00:05:34.000
And that genetic marker might be
as simple as the fact that at that

00:05:34.000 --> 00:05:38.000
position, maybe there is an A
in the DNA sequence on one of the

00:05:38.000 --> 00:05:42.000
chromosomes, and maybe I don't
know a G in the other sequence.

00:05:42.000 --> 00:05:47.000
And over here, this marker might
be, there's a T in some particular

00:05:47.000 --> 00:05:51.000
position, and there's a C
in some particular position.

00:05:51.000 --> 00:05:55.000
If we could assay that, if
we could tell, we could look

00:05:55.000 --> 00:06:00.000
whether this spelling variation is
closely correlated with the mutant.

00:06:00.000 --> 00:06:03.000
And this spelling variation
is closely correlated with the

00:06:03.000 --> 00:06:07.000
inheritance of the mutant allele.
And we could just try up and down

00:06:07.000 --> 00:06:11.000
the genome, different sites of
spelling difference as if they were

00:06:11.000 --> 00:06:15.000
genetic markers in our cross because
they are genetic markers in our

00:06:15.000 --> 00:06:18.000
cross, and see which one
is most tightly correlated.

00:06:18.000 --> 00:06:22.000
The minute we get any
genetic sequence difference,

00:06:22.000 --> 00:06:26.000
that shows co-inheritance linkage
in this cross, we know that this spot

00:06:26.000 --> 00:06:30.000
in the genome must be
nearby our mutation.

00:06:30.000 --> 00:06:33.000
So, we'll try one closer, and
we'll try one on the other side.

00:06:33.000 --> 00:06:37.000
And, what you do is you test
sites of genetic variation,

00:06:37.000 --> 00:06:40.000
first to find one that
shows any co-inheritance.

00:06:40.000 --> 00:06:44.000
And once you've got that, you
try ones closer, and closer,

00:06:44.000 --> 00:06:48.000
and closer. Last time I
talked about the process of,

00:06:48.000 --> 00:06:51.000
if you had one of those markers
you could use it to isolate the next

00:06:51.000 --> 00:06:55.000
clone and the next clone and the
next clone. But you know what I

00:06:55.000 --> 00:06:59.000
realized? That's so old fashioned.
We might as well deal with the fact

00:06:59.000 --> 00:07:02.000
we have a sequence of the genome.
No more would you ever isolate the

00:07:02.000 --> 00:07:05.000
next clone and the next
clone and the next clone.

00:07:05.000 --> 00:07:08.000
You just look it up in the computer.
So, even if you have the whole

00:07:08.000 --> 00:07:11.000
sequence of the genome, we
have to figure out what part of

00:07:11.000 --> 00:07:14.000
it was co-inherited along with this
disease, and that's the way you do

00:07:14.000 --> 00:07:17.000
it, OK? Genetic mapping, ust
as Sturtevant invented it,

00:07:17.000 --> 00:07:20.000
can be applied if you have a
whole sequence of the genome,

00:07:20.000 --> 00:07:23.000
and enough sites of variation.
And, I've drawn it for a fruit fly

00:07:23.000 --> 00:07:26.000
cross, but this could equally
well be cystic fibrosis.

00:07:26.000 --> 00:07:29.000
The only difference is if we're
doing this in human families and

00:07:29.000 --> 00:07:33.000
it's cystic fibrosis we
don't have as many offspring.

00:07:33.000 --> 00:07:36.000
So, we have to pool data
from many families. And,

00:07:36.000 --> 00:07:40.000
we can't arrange it so that every
family has exactly the same orange

00:07:40.000 --> 00:07:43.000
alleles up here and pink alleles
down there, but computers can deal

00:07:43.000 --> 00:07:47.000
with that. They can still figure
out the correlation across many

00:07:47.000 --> 00:07:50.000
families, and you find the spot
in the genome where for many,

00:07:50.000 --> 00:07:54.000
many, many families the kids who
all got the disease show correlated

00:07:54.000 --> 00:07:57.000
inheritance with this marker.
And that eventually pins you down

00:07:57.000 --> 00:08:01.000
to a region of the genome. It
pins you down to those genetic

00:08:01.000 --> 00:08:04.000
markers that show the
absolute tightest correlation,

00:08:04.000 --> 00:08:08.000
tight correlation, and
that's where you look.

00:08:08.000 --> 00:08:12.000
And in that fashion, people
went being able to map the

00:08:12.000 --> 00:08:16.000
location of Huntington's
Disease in 1984 to, by now,

00:08:16.000 --> 00:08:20.000
mapping the locations of more than
1, 00 different human genetic diseases

00:08:20.000 --> 00:08:25.000
where people didn't know the protein
in advance. They did it entirely

00:08:25.000 --> 00:08:29.000
based on this positional mapping.
So, Sturtevant's idea, which I like

00:08:29.000 --> 00:08:33.000
so much, has played itself out
so beautifully now in the area of

00:08:33.000 --> 00:08:38.000
modern molecular
medicine. OK. So, onward.

00:08:38.000 --> 00:08:43.000
I want to talk about a few other
variations on the theme rather

00:08:43.000 --> 00:08:48.000
quickly, and then I think I want
to talk about how you analyze your

00:08:48.000 --> 00:08:53.000
clones. First,
variations on cloning,

00:08:53.000 --> 00:08:58.000
I should just at least mention
it. We talked about cloning in an

00:08:58.000 --> 00:09:04.000
autonomously replicating
plasmid in a bacteria.

00:09:04.000 --> 00:09:07.000
So, you go to a bacteria.
They have some autonomously

00:09:07.000 --> 00:09:10.000
replicated pieces of DNA.
There are circles. You can clone

00:09:10.000 --> 00:09:13.000
in them, and you can typically,
these things are on the order of,

00:09:13.000 --> 00:09:17.000
I don't know, 1,000 to 2,000 to 5,
00 bases can be readily cloned in

00:09:17.000 --> 00:09:20.000
these plasmids.
You can do more,

00:09:20.000 --> 00:09:23.000
but that's a typical kind
of number is the insert size,

00:09:23.000 --> 00:09:27.000
typically. But we in the lab
go up to much higher numbers

00:09:27.000 --> 00:09:31.000
like 10,000 sometimes. You can
also, if you wanted to study

00:09:31.000 --> 00:09:36.000
yeast, it turns out yeast
happily have plasmids as well,

00:09:36.000 --> 00:09:41.000
and you can do a similar
sort of thing for yeast.

00:09:41.000 --> 00:09:47.000
It turns out that instead of using
plasmids, you can use bacterial

00:09:47.000 --> 00:09:52.000
viruses. These bacterial viruses
have all different shapes as we've

00:09:52.000 --> 00:09:57.000
talked about, circular or linear,
and they can typically hold, oh, 15,

00:09:57.000 --> 00:10:02.000
00-40,000. Some of these
viruses are quite big.

00:10:02.000 --> 00:10:06.000
The bacteriophage lambda
tends to carry a lot of stuff.

00:10:06.000 --> 00:10:11.000
And, it can replicate. So,
you could do the same thing to

00:10:11.000 --> 00:10:15.000
that. You can even use viruses that
infect mammalian cells and there are

00:10:15.000 --> 00:10:19.000
all sorts of viruses now
that people clone in again,

00:10:19.000 --> 00:10:24.000
linear or circular. I don't
know, for mammalian cells,

00:10:24.000 --> 00:10:28.000
you often, the viruses like 1,000-5,
00. You can even make artificial

00:10:28.000 --> 00:10:33.000
whole chromosomes now.
You can do this in yeast.

00:10:33.000 --> 00:10:39.000
Artificial chromosomes are called
YACs. They have all the little

00:10:39.000 --> 00:10:44.000
machinery, little telomeres on
them, little centromeres. They have a

00:10:44.000 --> 00:10:50.000
selectable marker, and then
you can clone into it your

00:10:50.000 --> 00:10:55.000
piece of DNA. And these can take
up to a million bases of DNA.

00:10:55.000 --> 00:11:01.000
So, if you wanted, there are
bacterial artificial chromosomes.

00:11:01.000 --> 00:11:05.000
They're called BACs if they're
in bacteria. And recently,

00:11:05.000 --> 00:11:09.000
people have developed artificial
chromosome systems for mammalian

00:11:09.000 --> 00:11:13.000
cells, and specifically human cells.
And they're called unfortunately

00:11:13.000 --> 00:11:17.000
MACs and HACs and things like that.
Basically, any molecule that can

00:11:17.000 --> 00:11:21.000
replicate in any system, some
smart molecular biologist will

00:11:21.000 --> 00:11:25.000
come along and say, how do
I use that for my purpose,

00:11:25.000 --> 00:11:30.000
to stick my DNA in it, and get
it to replicate in this organism?

00:11:30.000 --> 00:11:36.000
And so, if something's not
on this list, it will be soon,

00:11:36.000 --> 00:11:43.000
OK? Now, here's another thing.
This is cloning chunks of DNA.

00:11:43.000 --> 00:11:50.000
Just to have the piece of DNA in a
library, but suppose we want to do

00:11:50.000 --> 00:11:57.000
more than just have the DNA
sitting there in the bacterium,

00:11:57.000 --> 00:12:04.000
suppose what I'd really like
to do is take a bacterium,

00:12:04.000 --> 00:12:10.000
E coli, and put it to work for us.
Maybe what I'd like to do is take a

00:12:10.000 --> 00:12:14.000
plasmid and insert in that
plasmid the gene for human insulin.

00:12:14.000 --> 00:12:19.000
So, I'm going to take the DNA locus
corresponding to human insulin,

00:12:19.000 --> 00:12:23.000
clone it into my plasmid. Maybe
I'll have isolated it from my

00:12:23.000 --> 00:12:28.000
library because, let's
see, insulin's protein

00:12:28.000 --> 00:12:32.000
sequence is known so I could
reverse translate it to a nucleotide

00:12:32.000 --> 00:12:36.000
sequence. So, I
could probe a library.

00:12:36.000 --> 00:12:40.000
So, I could find the clone that has
insulin. Now what I'd like to do is

00:12:40.000 --> 00:12:43.000
persuade this bacteria not just to
carry the DNA but to make insulin

00:12:43.000 --> 00:12:47.000
for me. Would that be useful?
Yeah, how did people used to get

00:12:47.000 --> 00:12:51.000
insulin? Cadavers, dead
bodies; it would be much easier

00:12:51.000 --> 00:12:54.000
to get them from a fermenter,
right, to get insulin from a

00:12:54.000 --> 00:12:58.000
fermenter, if you could
just ask E coli to make it.

00:12:58.000 --> 00:13:02.000
So, if we put it into E coli,
will it make insulin for us?

00:13:02.000 --> 00:13:09.000
Here's the human locus, DNA
for insulin. Will it make

00:13:09.000 --> 00:13:17.000
insulin? Let's see, how
do you make a protein?

00:13:17.000 --> 00:13:24.000
You've got to start by making RNA,
right? You've got to transcribe the

00:13:24.000 --> 00:13:32.000
gene. Will E coli
transcribe this gene?

00:13:32.000 --> 00:13:36.000
Well, why? It's got a promoter,
right? It's got the insulin

00:13:36.000 --> 00:13:41.000
promoter. There we go. The
insulin promoter is here.

00:13:41.000 --> 00:13:45.000
So, E coli will come along to the
insulin promoter and start making

00:13:45.000 --> 00:13:50.000
RNA? No, it turns out that
promoters in humans and promoters in

00:13:50.000 --> 00:13:55.000
bacteria are sufficiently different.
They don't work across species.

00:13:55.000 --> 00:14:00.000
They won't recognize the human
promoter. Too bad. Any ideas?

00:14:00.000 --> 00:14:05.000
Yep? Stick a bacterial promoter
there. Good, you're acting like a

00:14:05.000 --> 00:14:10.000
good molecular biology designer here.
Let's put a bacterial promoter here.

00:14:10.000 --> 00:14:15.000
It will recognize its own promoter.
That's great. Then, let's put the

00:14:15.000 --> 00:14:21.000
DNA for the human insulin gene here.
And now, maybe we'll put the Lac

00:14:21.000 --> 00:14:26.000
operon, and when it has lactose
it'll start making RNA from the

00:14:26.000 --> 00:14:32.000
human insulin gene. And
it'll start translating it.

00:14:32.000 --> 00:14:38.000
And, we get insulin.
Any problems? Well,

00:14:38.000 --> 00:14:44.000
will it make any, for starters?
What's another aspect of mammalian

00:14:44.000 --> 00:14:50.000
genes that's different
from bacterial genes?

00:14:50.000 --> 00:14:56.000
Processing, what kind of processing
with the RNA? And the splicing,

00:14:56.000 --> 00:15:02.000
ooh, the insulin gene has introns
that have to be spliced out.

00:15:02.000 --> 00:15:06.000
So, this is going to make some
RNA, insulin RNA, and it needs to be

00:15:06.000 --> 00:15:10.000
processed like this. Will
bacteria carry on our splicing

00:15:10.000 --> 00:15:14.000
for us? They don't do splicing.
Yep? Well, that's a very

00:15:14.000 --> 00:15:18.000
interesting question
because we haven't. But,

00:15:18.000 --> 00:15:22.000
what do you propose? You see,
I've just taken a piece of

00:15:22.000 --> 00:15:26.000
human DNA from the human genome,
which encodes the introns and the

00:15:26.000 --> 00:15:30.000
exons. But, you seem to have
a solution to our problem,

00:15:30.000 --> 00:15:35.000
and what would that be? So,
instead of making a library of

00:15:35.000 --> 00:15:42.000
genomic DNA, what you're
suggesting is a radical idea.

00:15:42.000 --> 00:15:50.000
Let's instead take human RNA.
Here's some human RNA, lots of

00:15:50.000 --> 00:15:57.000
human RNA, a big collection of
human RNA. What was at the end of the

00:15:57.000 --> 00:16:02.000
human RNA: a poly(A) tail.
And what I understand you to be

00:16:02.000 --> 00:16:06.000
suggesting is if we take human
mRNAs, a whole collection of them,

00:16:06.000 --> 00:16:10.000
you want me to turn these mRNAs back
into DNA and clone them instead of

00:16:10.000 --> 00:16:14.000
using the chromosomal DNA. How
do I turn an RNA back to DNA?

00:16:14.000 --> 00:16:18.000
Is that possible? What do you
use: reverse transcriptase.

00:16:18.000 --> 00:16:22.000
We have to give it a primer.
So remember, five prime to three

00:16:22.000 --> 00:16:26.000
prime, we'd like to put
a primer going over here.

00:16:26.000 --> 00:16:30.000
Any ideas for a good primer?
Poly(T), isn't that convenient?

00:16:30.000 --> 00:16:35.000
One of the reasons that mammalian
messages have poly(A) tails is so

00:16:35.000 --> 00:16:41.000
that we are able to reverse
transcribe them using poly(T)

00:16:41.000 --> 00:16:46.000
primers. No, that's actually
not true. So, we use reverse

00:16:46.000 --> 00:16:52.000
transcriptase. And what
we can do is we'll copy

00:16:52.000 --> 00:16:58.000
this RNA into a strand
of DNA. There we go.

00:16:58.000 --> 00:17:03.000
Then what we'll do, next
step, is we'll take the DNA,

00:17:03.000 --> 00:17:09.000
and we'll copy back into
a second strand of DNA.

00:17:09.000 --> 00:17:15.000
And now, we have double-stranded
DNA whose sequence matches the

00:17:15.000 --> 00:17:21.000
already-processed mRNAs.
Sorry? So, the sequences would

00:17:21.000 --> 00:17:27.000
match the mRNAs. So what
you could do is instead of

00:17:27.000 --> 00:17:32.000
taking human DNA from the
nucleus, you could take RNAs,

00:17:32.000 --> 00:17:38.000
turn them back into DNA
by reverse transcriptase,

00:17:38.000 --> 00:17:43.000
and make a library now that
consists of zillions of inserts,

00:17:43.000 --> 00:17:49.000
each of which has what's
called a cDNA, a copied DNA,

00:17:49.000 --> 00:17:54.000
copied back from the RNA. The
great advantage of this is that

00:17:54.000 --> 00:18:00.000
the human cell has already
done the splicing, and so there

00:18:00.000 --> 00:18:05.000
are no introns left. Now,
when you stick it in a

00:18:05.000 --> 00:18:09.000
bacterium, the bacterium is
able to express this. It's able,

00:18:09.000 --> 00:18:13.000
if you give it its own bacterial
promoter, to make an RNA.

00:18:13.000 --> 00:18:17.000
And if you don't ask the
bacteria to have to splice,

00:18:17.000 --> 00:18:21.000
if you just give it a pre-spliced
piece of DNA that doesn't need

00:18:21.000 --> 00:18:25.000
splicing, it can translate that
DNA. Now, notice we used all of our

00:18:25.000 --> 00:18:29.000
tricks. You had to know
about reverse transcriptase,

00:18:29.000 --> 00:18:34.000
poly(A) tails, structures of genes,
introns, exons, yes, question?

00:18:34.000 --> 00:18:38.000
It doesn't. You do this in the test
tube. You purify human mRNA in the

00:18:38.000 --> 00:18:42.000
test tube. You take that mRNA in a
test tube, add reverse transcriptase,

00:18:42.000 --> 00:18:47.000
add poly(T), make this reaction of
RNA to DNA in the test tube go back.

00:18:47.000 --> 00:18:51.000
Where does it come from? Viruses
that copy themselves back for a

00:18:51.000 --> 00:18:56.000
living, right? So, again,
every single thing we're

00:18:56.000 --> 00:19:00.000
using comes from some living
organism that does this

00:19:00.000 --> 00:19:04.000
kind of stuff. And, when
I teach you about the

00:19:04.000 --> 00:19:08.000
facts of how viruses replicate or
what the structure of mRNAs look

00:19:08.000 --> 00:19:11.000
like or whatever, it's
because every bit of knowledge

00:19:11.000 --> 00:19:14.000
we get about the way biology works
turns into an incredibly powerful

00:19:14.000 --> 00:19:18.000
tool as it's turning out for us to
actually be able to further study

00:19:18.000 --> 00:19:21.000
biology. So, great.
So, where does reverse

00:19:21.000 --> 00:19:24.000
transcriptase come from now?
Originally they come from viruses

00:19:24.000 --> 00:19:28.000
that turn themselves back from RNA
to DNA. Now, how do you get reverse

00:19:28.000 --> 00:19:32.000
transcriptase?
Catalog, right,

00:19:32.000 --> 00:19:38.000
very good. All right, so
this is called, finally,

00:19:38.000 --> 00:19:44.000
a cDNA library. And, if
you had made a cDNA library,

00:19:44.000 --> 00:19:49.000
you would be able to screen the cDNA
library to find the gene for insulin.

00:19:49.000 --> 00:19:55.000
Is this useful?
This happens to be,

00:19:55.000 --> 00:20:01.000
for example, one of the consequences
of this was the biotechnology

00:20:01.000 --> 00:20:06.000
industry. OK, so if you
have any doubts about

00:20:06.000 --> 00:20:10.000
the usefulness of understanding
these abstract things about E coli

00:20:10.000 --> 00:20:14.000
and bacteria and stuff like
that, one of the consequences was

00:20:14.000 --> 00:20:18.000
Genentech, Biogen, and
Amgen, and if you just simply

00:20:18.000 --> 00:20:22.000
walk around Kendall Square, within
a mile of this place you will

00:20:22.000 --> 00:20:26.000
see laid out before you the
consequences of this ability,

00:20:26.000 --> 00:20:30.000
OK? It's transforming
Cambridge. Yes?

00:20:30.000 --> 00:20:36.000
And the world.
Yeah. Indeed.

00:20:36.000 --> 00:20:43.000
It might be that producing large
amounts of insulin was bad for the

00:20:43.000 --> 00:20:50.000
bacteria because there would be so
much protein it would clump and kill

00:20:50.000 --> 00:20:57.000
the bacteria. It might be that
insulin, for various reasons,

00:20:57.000 --> 00:21:04.000
might not fold appropriately
in the bacterial environment.

00:21:04.000 --> 00:21:07.000
And, this is why the biotechnology
industry has lots of smart people

00:21:07.000 --> 00:21:10.000
working in it because you're totally,
100% right. You might decide that

00:21:10.000 --> 00:21:13.000
instead of cloning it in bacteria
it's better to clone it in some

00:21:13.000 --> 00:21:16.000
insect cell in culture which, in
fact, people like to work with,

00:21:16.000 --> 00:21:19.000
or some other cell, or
a mammalian cell. And so,

00:21:19.000 --> 00:21:23.000
I simplify by saying put it in
coli, but in fact that might test six

00:21:23.000 --> 00:21:26.000
different cell lines, six
different host possibilities.

00:21:26.000 --> 00:21:29.000
They might have to take the
insulin out and refold it in vitro

00:21:29.000 --> 00:21:33.000
and things like that.
You're totally right.

00:21:33.000 --> 00:21:37.000
This is actually something that
requires work to do it right,

00:21:37.000 --> 00:21:42.000
just like building an
airplane requires work.

00:21:42.000 --> 00:21:47.000
I could tell you Bernoulli's
principles, but then Boeing does

00:21:47.000 --> 00:21:51.000
more than just writes down
Bernoulli's principles.

00:21:51.000 --> 00:21:56.000
OK, so onward. Now, I'd like to
turn next to analyzing your clone.

00:21:56.000 --> 00:22:00.000
Analyzing the clone, so suppose
we have, maybe it's by positional

00:22:00.000 --> 00:22:05.000
cloning, maybe it's by cDNA cloning,
but one way or the other we've got

00:22:05.000 --> 00:22:10.000
us a clone that we're
very interested in.

00:22:10.000 --> 00:22:14.000
Maybe it has the insulin gene.
Maybe it has the Huntington's

00:22:14.000 --> 00:22:18.000
disease gene. Whatever it is,
we're going to want to study it.

00:22:18.000 --> 00:22:22.000
And at the moment, I haven't told
you how I would even read its DNA

00:22:22.000 --> 00:22:26.000
sequence or analyze its DNA.
So, the first step is, of course,

00:22:26.000 --> 00:22:31.000
I have to purify the plasmid. And,
it turns out that that can be done.

00:22:31.000 --> 00:22:34.000
There are simple biochemical
techniques, as I mentioned in a

00:22:34.000 --> 00:22:37.000
previous lecture, that
allow you to grow up a lot of

00:22:37.000 --> 00:22:40.000
the bacteria, crack them open,
and the plasmid being a little

00:22:40.000 --> 00:22:43.000
circle, and being a little more
tightly super-coiled and wound up

00:22:43.000 --> 00:22:46.000
has somewhat different physical
properties. And you can use those

00:22:46.000 --> 00:22:50.000
to purify the plasmid. So,
plasmid preps are not hard to

00:22:50.000 --> 00:22:53.000
do. You can get a fairly pure
collection of the plasmid.

00:22:53.000 --> 00:22:56.000
Now, suppose I've done this for,
oh, I don't know, let's take my

00:22:56.000 --> 00:23:00.000
first example, orange mutants.
Suppose I tried to rescue bacteria

00:23:00.000 --> 00:23:04.000
that were orange minus,
and suppose I found that 50

00:23:04.000 --> 00:23:08.000
different plasmids rescued my orange
mutant because I transformed a lot

00:23:08.000 --> 00:23:12.000
of plasmids in, I plated
it, and 50 colonies grew up.

00:23:12.000 --> 00:23:16.000
Are they all the same
thing or are they different?

00:23:16.000 --> 00:23:20.000
Is there any quickie way to take a
look at these 50 plasmids and see if

00:23:20.000 --> 00:23:24.000
they're identical or fairly close,
or obviously different? Well, I'd

00:23:24.000 --> 00:23:28.000
like to take some way to take the
DNA from the plasmid and analyze it

00:23:28.000 --> 00:23:32.000
kind of easily. I
might want to see,

00:23:32.000 --> 00:23:37.000
like, how big is the insert?
Right, that'd be one way,

00:23:37.000 --> 00:23:43.000
if they had different sized inserts
so they couldn't be the same thing.

00:23:43.000 --> 00:23:49.000
So, maybe what I could do is how do
I clone this? I used EcoRI sites I

00:23:49.000 --> 00:23:55.000
recall. So, I have EcoRI sites here.
Suppose I were to take this DNA,

00:23:55.000 --> 00:24:02.000
and I were to now cut the DNA
from the plasmid with EcoRI.

00:24:02.000 --> 00:24:09.000
Then, what I would get
is two separate molecules.

00:24:09.000 --> 00:24:16.000
I would get the vector and the
insert. How could I see how big

00:24:16.000 --> 00:24:24.000
they were? Gels, gel
electrophoresis is the way to do

00:24:24.000 --> 00:24:29.000
that. So, I take a gel. A
gel is a slab of gelatin,

00:24:29.000 --> 00:24:33.000
Jell-O, OK, and normally it's
laid flat, but I'm going to do it

00:24:33.000 --> 00:24:37.000
vertically here. I load
into the top of it here a

00:24:37.000 --> 00:24:41.000
little bit of my DNA, this
whole mixture. I take the

00:24:41.000 --> 00:24:45.000
plasmid. I cut it. I put
it in here. DNA's positive

00:24:45.000 --> 00:24:49.000
charge or negative charge?
Negative. So, where should I put

00:24:49.000 --> 00:24:53.000
the positive pull? On
the bottom, well done.

00:24:53.000 --> 00:24:57.000
That's often not done, and to
the detriment of the experiment.

00:24:57.000 --> 00:25:01.000
If you put the positive pull here,
it goes the wrong way, and everybody

00:25:01.000 --> 00:25:05.000
has to do that at least once.
So, what'll happen is the DNA

00:25:05.000 --> 00:25:11.000
fragments move through, and
the smaller fragments move

00:25:11.000 --> 00:25:16.000
faster than the big fragments,
right? If something's little, it'll

00:25:16.000 --> 00:25:22.000
move fast. If something's big,
it moves slowly: little, big.

00:25:22.000 --> 00:25:27.000
Smaller moves faster because it
wiggles through the little pores in

00:25:27.000 --> 00:25:33.000
the gel better. So, suppose
I were to do this for a

00:25:33.000 --> 00:25:39.000
bunch of plasmids, and
what I saw was this.

00:25:39.000 --> 00:25:47.000
First order, what do you guess?
Sorry? Top road's probably the

00:25:47.000 --> 00:25:55.000
plasmid vector. This
is probably the vector,

00:25:55.000 --> 00:26:03.000
and what do I know about the
inserts? At least two inserts,

00:26:03.000 --> 00:26:09.000
at least two distinct inserts.
Now, if I wanted to be sure that was

00:26:09.000 --> 00:26:13.000
the vector, maybe what I
could do is take another row,

00:26:13.000 --> 00:26:17.000
and run a known amount of the vector,
take the vector alone and I could

00:26:17.000 --> 00:26:21.000
check that the vector alone runs
over here. And maybe I might take

00:26:21.000 --> 00:26:25.000
some other known molecules.
These would be called molecular

00:26:25.000 --> 00:26:29.000
weight standards. So, if
I run some knowns in one of

00:26:29.000 --> 00:26:33.000
the lanes of the gel, I
can even measure and say,

00:26:33.000 --> 00:26:37.000
ah-ha, the insert is somewhere
between the size of this one and the

00:26:37.000 --> 00:26:40.000
size of that one. And so,
I get a little ruler that I

00:26:40.000 --> 00:26:43.000
can put on the gel. So, in
fact, that's the first thing

00:26:43.000 --> 00:26:46.000
you would do is you
digest your clone that way.

00:26:46.000 --> 00:26:49.000
Now, does the fact that these
guys have exactly the same,

00:26:49.000 --> 00:26:52.000
apparently, size on the gel mean
that they're the exact same piece of

00:26:52.000 --> 00:26:55.000
DNA? No, because you can't even
actually tell it's exactly the same.

00:26:55.000 --> 00:26:59.000
There's a limit to how
precisely you can measure it.

00:26:59.000 --> 00:27:04.000
So, what else could you do? You
could try another restriction

00:27:04.000 --> 00:27:10.000
enzyme. It turns out that since
there are so many restriction

00:27:10.000 --> 00:27:15.000
enzymes in the catalog,
if I take a piece of DNA,

00:27:15.000 --> 00:27:21.000
maybe that Eco fragment, I could
try cutting it with HinDIII.

00:27:21.000 --> 00:27:26.000
And when I cut it with HinDIII,
I'm going to get three distinct

00:27:26.000 --> 00:27:32.000
lengths. I could try cutting
it with, oh, I don't know,

00:27:32.000 --> 00:27:37.000
pick another enzyme, BamHI.
When I cut it with BamHI,

00:27:37.000 --> 00:27:43.000
I'll get some other lengths.
And, how to get these lengths by

00:27:43.000 --> 00:27:48.000
adding these, by running them out
on a gel and looking at their sizes.

00:27:48.000 --> 00:27:54.000
What if I added both HinDIII
and BamHI to my test tube?

00:27:54.000 --> 00:28:00.000
I'd cut at both sites.
So, I'd cut here, here,

00:28:00.000 --> 00:28:06.000
here, here, here. So,
this is cut with HinDIII,

00:28:06.000 --> 00:28:12.000
here cut with BamHI, here cut
with both and I could measure these

00:28:12.000 --> 00:28:19.000
lengths. So, suppose I gave
you this as a computer problem,

00:28:19.000 --> 00:28:25.000
I have a string and it's an unknown
string, and I cut it at two places

00:28:25.000 --> 00:28:31.000
and I get these lengths, X1, X2, X3.
And then I take that same string and

00:28:31.000 --> 00:28:35.000
I cut it at other positions, Y1,
Y2, and Y3 are the lengths that

00:28:35.000 --> 00:28:39.000
result. And then suppose I now
cut it at both of the sites,

00:28:39.000 --> 00:28:43.000
and I measure it, and I get Z1,
Z2, Z3, Z4, Z5. If I gave you all

00:28:43.000 --> 00:28:48.000
those numbers, could you
figure out where the sites

00:28:48.000 --> 00:28:52.000
must be? Probably. It
turns out to be a reasonably

00:28:52.000 --> 00:28:56.000
doable computer problem, although
it can get a little hard in

00:28:56.000 --> 00:29:00.000
places. And you could
try a third enzyme and

00:29:00.000 --> 00:29:03.000
a fourth enzyme, and it's
a cute exercise to write

00:29:03.000 --> 00:29:07.000
yourself a little piece of code that
will figure out where the sites are

00:29:07.000 --> 00:29:10.000
based on the lengths. The
reason it occasionally gets

00:29:10.000 --> 00:29:13.000
funny what if Z3 and Z4 are exactly
the same length and they run on top

00:29:13.000 --> 00:29:16.000
of each other in the gel,
and there are special cases.

00:29:16.000 --> 00:29:20.000
But you can kind of reconstruct
where those restriction sites must

00:29:20.000 --> 00:29:23.000
be just by writing a good piece
of code that'll put these pieces

00:29:23.000 --> 00:29:26.000
together. This is called
restriction mapping,

00:29:26.000 --> 00:29:30.000
and it's great fun. Everybody
likes to do this once.

00:29:30.000 --> 00:29:33.000
But, it's only a limited
amount of information, right,

00:29:33.000 --> 00:29:36.000
because you get where the sites are,
and I guess if I gave you ten clones

00:29:36.000 --> 00:29:40.000
and they all had exactly
the same restriction maps,

00:29:40.000 --> 00:29:43.000
the exact same positions
of these restriction sites,

00:29:43.000 --> 00:29:47.000
you'd feel pretty confident
they were the same clone.

00:29:47.000 --> 00:29:50.000
But you still wouldn't really know
much about the clone other than it

00:29:50.000 --> 00:29:53.000
had two HinDIII sites and two BamHI
sites, and here's where they were.

00:29:53.000 --> 00:29:57.000
What do you really want to
know about this clone? It's

00:29:57.000 --> 00:30:02.000
DNA sequence, right? Let's
not settle for anything less

00:30:02.000 --> 00:30:10.000
than the exact nucleotide
sequence of the clone. So,

00:30:10.000 --> 00:30:18.000
that's really the last key
topic is sequencing DNA.

00:30:18.000 --> 00:30:26.000
How are you going to sequence
DNA? Well, suppose I give you some

00:30:26.000 --> 00:30:34.000
double strand of DNA,
five prime to three prime,

00:30:34.000 --> 00:30:42.000
five prime, three prime,
double stranded DNA.

00:30:42.000 --> 00:30:47.000
Let me heat it up. What
happens when I heat up DNA?

00:30:47.000 --> 00:30:52.000
It melts the hydrogen bonds, the
non-covalent hydrogen bonds here

00:30:52.000 --> 00:30:57.000
break, and I got my two
strands separated. Now,

00:30:57.000 --> 00:31:02.000
what I'd like to do is I want to
start reading out this DNA sequence.

00:31:02.000 --> 00:31:08.000
So, I'm going to make me a primer.
Now, golly, here's a primer.

00:31:08.000 --> 00:31:14.000
You're going to ask me, how
did I even know what primer to

00:31:14.000 --> 00:31:21.000
use if I don't know the DNA
sequence? How can I make a primer?

00:31:21.000 --> 00:31:27.000
Hold that question. Make sure I
remember to come back and answer

00:31:27.000 --> 00:31:34.000
that, OK? But for the moment,
grant me that I have a primer here.

00:31:34.000 --> 00:31:39.000
What I'd like to do is
add DNA polymerase. So,

00:31:39.000 --> 00:31:45.000
let's add some DNA polymerase.
And, I'd like to add nucleotide

00:31:45.000 --> 00:31:50.000
triphosphates,
dNTPs, the dATP,

00:31:50.000 --> 00:31:56.000
dCTP, the dGTP, dTTP, and
if I add DNA polymerase and I

00:31:56.000 --> 00:32:02.000
add my nucleotides, what
does Arthur Kornberg tell us

00:32:02.000 --> 00:32:07.000
will happen? It'll
start polymerizing,

00:32:07.000 --> 00:32:13.000
right? And, it'll stop there. So,
the polymerase knows the bases,

00:32:13.000 --> 00:32:18.000
right? It knows what base to put
in because polymerase is very smart.

00:32:18.000 --> 00:32:24.000
So, the bases get put in correctly.
The only problem is, how do we get

00:32:24.000 --> 00:32:30.000
polymerase to tell
us what it just did?

00:32:30.000 --> 00:32:37.000
Here's a cute trick.
This is, by the way,

00:32:37.000 --> 00:32:44.000
a cute trick that won the Nobel
Prize. So, suppose my primer is

00:32:44.000 --> 00:32:51.000
like this: five prime, T,
A, A, T, T, C, T, and the

00:32:51.000 --> 00:32:58.000
template strand here, A, T,
T, A, A, G, A, now let's keep

00:32:58.000 --> 00:33:05.000
going, A, T, G,
C, C, A, A, T, G,

00:33:05.000 --> 00:33:14.000
G, A, T, T, A, five prime.
So, there's my primer.

00:33:14.000 --> 00:33:26.000
There's my template. I'm going
to start adding. Well, let's

00:33:26.000 --> 00:33:37.000
add our polymerase.
Let's add our dNTPs,

00:33:37.000 --> 00:33:47.000
polymerase, dATP, dCTP, dTTP,
dGTP, and then I want to add a

00:33:47.000 --> 00:33:58.000
special extra good old
ingredient into this.

00:33:58.000 --> 00:34:07.000
The special extra ingredient
I want to add is a defective T,

00:34:07.000 --> 00:34:17.000
a defective dTTP. What do I mean
by defective? I mean chemically

00:34:17.000 --> 00:34:27.000
modified in such a way that
it can't be extended, that you

00:34:27.000 --> 00:34:36.000
can't extend past it. So
now, let's follow my reaction.

00:34:36.000 --> 00:34:44.000
I'm going to start with, I'm just
going to write them down here,

00:34:44.000 --> 00:34:52.000
T, A, A, T, T, C, T. What's the
next base I'm going to put in?

00:34:52.000 --> 00:35:00.000
T, OK? Is that a defective
T or a good T? I don't know.

00:35:00.000 --> 00:35:05.000
It could be. Maybe it's a defective
T, which I'll put a little star

00:35:05.000 --> 00:35:11.000
there, OK? If so, what
happens to my polymerase?

00:35:11.000 --> 00:35:17.000
It stops. It can't go any further.
It can't go any further because the

00:35:17.000 --> 00:35:22.000
T's defective. But what
if it wasn't a defective T?

00:35:22.000 --> 00:35:28.000
What if it was a good T? Then
what goes on? The polymerase

00:35:28.000 --> 00:35:34.000
will put in, keep going guys. A,
C, G, G, and what does it put in

00:35:34.000 --> 00:35:39.000
now? T,
right? Now,

00:35:39.000 --> 00:35:45.000
is that a defective T?
Maybe. We don't know.

00:35:45.000 --> 00:35:50.000
If it is a defective T,
it stops there. Otherwise,

00:35:50.000 --> 00:35:56.000
polymerase goes here, and
the next space is what?

00:35:56.000 --> 00:36:02.000
T, and is that a
defective T? Maybe.

00:36:02.000 --> 00:36:09.000
And, if it's not a defective T,
then polymerase goes on, puts in an

00:36:09.000 --> 00:36:16.000
A, puts in a G, a
C, C, and then a T.

00:36:16.000 --> 00:36:23.000
And maybe that's defective.
All right, which of these

00:36:23.000 --> 00:36:31.000
possibilities is what polymerase
does when I throw it in?

00:36:31.000 --> 00:36:35.000
Well, all of them. There's
a lot of molecules there.

00:36:35.000 --> 00:36:39.000
Some of the molecules, by chance,
happen to install a defective T,

00:36:39.000 --> 00:36:43.000
and they grind to a halt here.
Sometimes, a good T's put in and the

00:36:43.000 --> 00:36:47.000
molecules stop here.
Sometimes they stop here,

00:36:47.000 --> 00:36:51.000
and if I start with a big collection
of primers in a lot of my template

00:36:51.000 --> 00:36:55.000
DNA, I'm going to get this whole
collection of different molecules of

00:36:55.000 --> 00:36:59.000
different lengths.
What lengths do I get?

00:36:59.000 --> 00:37:04.000
The lengths correspond precisely
to the positions of the Ts.

00:37:04.000 --> 00:37:10.000
I get a series of molecules
whose lengths perfectly match the

00:37:10.000 --> 00:37:16.000
positions of Ts.
Well, first off,

00:37:16.000 --> 00:37:23.000
how do I measure their lengths?
Run a gel, bingo, run a gel.

00:37:23.000 --> 00:37:30.000
So, if I could run a gel that could
separate nucleotides based on length

00:37:30.000 --> 00:37:37.000
that two next to each other,
another one up there, I'd see a

00:37:37.000 --> 00:37:44.000
small molecule,
length one, two,

00:37:44.000 --> 00:37:51.000
three, four, five, three,
six, eight, I'd see one of

00:37:51.000 --> 00:37:58.000
length eight. I'd see one of
length, what's the next one,

00:37:58.000 --> 00:38:05.000
13, eight, nine, ten,
13, 14, so eight, nine,

00:38:05.000 --> 00:38:12.000
ten, 11, 12, 13, 14, 15,
what's that, 13, 14, 15, 16,

00:38:12.000 --> 00:38:18.000
17, 18. OK, those would
be the positions at

00:38:18.000 --> 00:38:22.000
which I would see this T. So,
I'd need to have a special kind

00:38:22.000 --> 00:38:26.000
of gel that's so accurate that it
can separate single nucleotides,

00:38:26.000 --> 00:38:31.000
right, that the lengths,
but that can be done.

00:38:31.000 --> 00:38:36.000
There's acrylamide, the
polymer that will do that.

00:38:36.000 --> 00:38:41.000
That'll tell me the exact lengths
of the T's. What else do I do?

00:38:41.000 --> 00:38:46.000
Well, let's obviously do
it from the other bases.

00:38:46.000 --> 00:38:52.000
Let's try defective A,
defective C, defective G.

00:38:52.000 --> 00:38:57.000
Let's see, if I got it right,
which we'll try, it ought to end up

00:38:57.000 --> 00:39:02.000
looking something like that.
And if not, you get the picture,

00:39:02.000 --> 00:39:07.000
that this ought to match up as to
which columns have which lengths.

00:39:07.000 --> 00:39:13.000
OK, I think I got it right.
That tells me the lengths of the

00:39:13.000 --> 00:39:18.000
molecules. So, I could
read off at sequence.

00:39:18.000 --> 00:39:23.000
The sequence of that molecule
ought to be, starting over there,

00:39:23.000 --> 00:39:29.000
the sequence of what I've added in,
ought to be something like T, A, C,

00:39:29.000 --> 00:39:34.000
G, G, T, T, A, C,
C, T, yep, it worked.

00:39:34.000 --> 00:39:39.000
It's exactly right. Bingo.
I can now read the sequence.

00:39:39.000 --> 00:39:44.000
Fred Sanger, a brilliant scientist,
thought up this method of just

00:39:44.000 --> 00:39:49.000
exploiting E coli's own polymerase
or other organism's own polymerases.

00:39:49.000 --> 00:39:54.000
So, copying and all the chemistry
that had to be done was thinking up

00:39:54.000 --> 00:40:00.000
a defective nucleotide
that could not be extended.

00:40:00.000 --> 00:40:10.000
It could obviously be inserted.
It can't be extended. So, one

00:40:10.000 --> 00:40:20.000
question is, what's a
defective nucleotide? Well,

00:40:20.000 --> 00:40:30.000
you will recall that our nucleotide
in the sugar phosphate chain

00:40:30.000 --> 00:40:37.000
is sitting like this. Let's
see, hanging off the one prime

00:40:37.000 --> 00:40:42.000
carbon is the base. This
is the one prime carbon,

00:40:42.000 --> 00:40:47.000
the two prime carbon, the three
prime carbon, the four prime carbon,

00:40:47.000 --> 00:40:52.000
the five prime carbon. What do we
know in DNA at the two prime carbon?

00:40:52.000 --> 00:40:57.000
Normally in ribose there
would be a hydroxyl here,

00:40:57.000 --> 00:41:02.000
right? But in deoxyribose,
there's just a hydrogen.

00:41:02.000 --> 00:41:11.000
So, if this is deoxyribose, so
a dNTP really means a two prime

00:41:11.000 --> 00:41:20.000
deoxyribose, where do I now attach
my next base in the sugar phosphate

00:41:20.000 --> 00:41:30.000
train? Three prime ends, and
what do I attach it to: the OH.

00:41:30.000 --> 00:41:35.000
What do you think would
happen if there's no OH there?

00:41:35.000 --> 00:41:40.000
You're stuck. All you've got
to do is take off that hydroxyl.

00:41:40.000 --> 00:41:45.000
No hydroxyl group. If you made
nucleotides that don't have that

00:41:45.000 --> 00:41:50.000
hydroxyl group, they
can't be extended.

00:41:50.000 --> 00:41:55.000
So, instead of these being just
deoxy at the two prime position,

00:41:55.000 --> 00:42:00.000
they are dideoxy,
deoxy at two positions.

00:42:00.000 --> 00:42:04.000
They are two prime, three
prime, dideoxynucleotides.

00:42:04.000 --> 00:42:09.000
That's it. Now, if you needed
to get two prime three prime

00:42:09.000 --> 00:42:13.000
dideoxynucleotides, they're
in the catalogue of course,

00:42:13.000 --> 00:42:18.000
right, because Fred Sanger had
to make them himself and all that,

00:42:18.000 --> 00:42:23.000
but you can just buy them now.
And so, you can do the sequence.

00:42:23.000 --> 00:42:28.000
A few other little
details here, though, guys.

00:42:28.000 --> 00:42:32.000
How do we see the DNA and the gel?
One possibility would be staining

00:42:32.000 --> 00:42:37.000
it. There are some dies
like ethidium bromide,

00:42:37.000 --> 00:42:42.000
and for doing your restriction
mapping, using a dye that sticks to

00:42:42.000 --> 00:42:47.000
DNA like ethidium bromide does is
pretty good. And then you put it

00:42:47.000 --> 00:42:52.000
under fluorescent light and
you look. For sequencing,

00:42:52.000 --> 00:42:57.000
the amount of DNA is so little that
it's hard to see with a dye by the

00:42:57.000 --> 00:43:02.000
naked eye, which is what you do
with restriction map. So, sorry?

00:43:02.000 --> 00:43:06.000
So, the first thing people did was
radioactive. What they did was they

00:43:06.000 --> 00:43:10.000
took a primer,
made it radioactive,

00:43:10.000 --> 00:43:14.000
and you did this whole sequencing
reaction with radioactive primer.

00:43:14.000 --> 00:43:18.000
Then, when you run the gel, you
take your gel and you expose it for

00:43:18.000 --> 00:43:22.000
some number of hours, eight
hours maybe, a piece of x-ray

00:43:22.000 --> 00:43:26.000
film, develop the x-ray film,
and you'll see that picture. So,

00:43:26.000 --> 00:43:30.000
one solution that you could do
to visualize is using radioactive

00:43:30.000 --> 00:43:37.000
nucleotides. So, we got
the defective nucleotide.

00:43:37.000 --> 00:43:45.000
We now need to visualize our
DNA. Let's visualize the sequence.

00:43:45.000 --> 00:43:54.000
One possibility: radioactive.
The second possibility, someone

00:43:54.000 --> 00:44:03.000
already mentioned
it, a fluorescent dye.

00:44:03.000 --> 00:44:10.000
Now, here, a fluorescent dye could
be put on, and you can't read it

00:44:10.000 --> 00:44:17.000
with your eye, but lasers
are very good at reading.

00:44:17.000 --> 00:44:24.000
So, you might run a whole gel
here and have lasers scan it.

00:44:24.000 --> 00:44:31.000
But, you can actually do better
than that. Suppose I put my

00:44:31.000 --> 00:44:39.000
fluorescent dye on
my dideoxynucleotides.

00:44:39.000 --> 00:44:45.000
Suppose I put it on
my dideoxynucleotides,

00:44:45.000 --> 00:44:52.000
and suppose I even had enough
chemistry at my disposal that I

00:44:52.000 --> 00:44:59.000
could put a different color
of fluorescent dye on each

00:44:59.000 --> 00:45:05.000
of my nucleotides. Then,
whenever the dideoxy is put in

00:45:05.000 --> 00:45:09.000
to terminate the chain, it
carries with it its own color.

00:45:09.000 --> 00:45:14.000
Wouldn't that be cool? And, that's
what's done. Not just can you buy

00:45:14.000 --> 00:45:18.000
dideoxynucleotides now, but
you can buy the four different

00:45:18.000 --> 00:45:23.000
dideoxynucleotides each with
its own dye attached to it.

00:45:23.000 --> 00:45:27.000
So, there are di-dideoxies I guess,
sorry, but it's different di's,

00:45:27.000 --> 00:45:32.000
right? They're dye-dideoxies.
So, you could do that.

00:45:32.000 --> 00:45:36.000
And then what you get would be that
in this column you get this color.

00:45:36.000 --> 00:45:40.000
And in this column, you'd get this
color. And in this column you'd get

00:45:40.000 --> 00:45:44.000
this color, etc. I'm not
worrying about where they

00:45:44.000 --> 00:45:48.000
are here. And they'd all be
different colors and it would be

00:45:48.000 --> 00:45:52.000
very pretty. You know what?
Why do we need to run separate

00:45:52.000 --> 00:45:56.000
lanes anymore?
If we got a laser,

00:45:56.000 --> 00:46:00.000
we can tell the laser scan
it to tell it different. Stick

00:46:00.000 --> 00:46:05.000
it in one way. In fact,
what's done is stick it in

00:46:05.000 --> 00:46:13.000
a capillary tube, throw in
all four at the same time

00:46:13.000 --> 00:46:20.000
now, and as these fragments come by,
each has its own color. And all we

00:46:20.000 --> 00:46:28.000
need is a laser scanner capable
of sitting right here. Here's

00:46:28.000 --> 00:46:34.000
my laser scanner.
And the laser scanner,

00:46:34.000 --> 00:46:40.000
positive here, negative here,
as the DNA flows by through this

00:46:40.000 --> 00:46:46.000
polymer, the laser scanner reads
off which colors just went by.

00:46:46.000 --> 00:46:51.000
And it goes A color, C color,
T color, G color. That's it. So,

00:46:51.000 --> 00:46:57.000
there are actually machines now
that have 96 different capillaries.

00:46:57.000 --> 00:47:04.000
These are called capillary tubes.
And you can have 96 of them with

00:47:04.000 --> 00:47:12.000
laser scanning across,
and in each column now,

00:47:12.000 --> 00:47:20.000
it turns out that you can
read almost 1,000 letters,

00:47:20.000 --> 00:47:28.000
1,000 bases per column per capillary
times about 100 capillaries.

00:47:28.000 --> 00:47:36.000
Or in other words, you can read
out about 10^5 bases of information.

00:47:36.000 --> 00:47:40.000
You can read out 10^5 bases of
information in about two hours.

00:47:40.000 --> 00:47:45.000
Of course, you can do
that ten times a day. So,

00:47:45.000 --> 00:47:50.000
you can actually read out 10^6 or
about a million bases of information

00:47:50.000 --> 00:47:55.000
per machine. And here at MIT, we
have 100 of these machines. So,

00:47:55.000 --> 00:48:00.000
we actually can read out a little
shy of 100 million letters of DNA

00:48:00.000 --> 00:48:05.000
sequence per day,
which I mean is a lot.

00:48:05.000 --> 00:48:11.000
We read about 40 billion
letters per year here at MIT,

00:48:11.000 --> 00:48:17.000
and this is how we do it.
How much does a machine cost?

00:48:17.000 --> 00:48:23.000
List, or do you want a
deal? They list for $300,

00:48:23.000 --> 00:48:29.000
00, but if you buy in bulk, I
can do better. [LAUGHTER] We buy

00:48:29.000 --> 00:48:34.000
it in bulk, by the way. So
now, how are we going to get our

00:48:34.000 --> 00:48:38.000
primer there? That was the only
little bit we were missing is where

00:48:38.000 --> 00:48:42.000
did our primer come from? The
last little detail: here's my

00:48:42.000 --> 00:48:47.000
vector, remember, and I
want to sequence this insert.

00:48:47.000 --> 00:48:51.000
How am I going to get a primer in
the insert? I don't know what its

00:48:51.000 --> 00:48:56.000
sequence is. How do I even
start this? Sorry? Well,

00:48:56.000 --> 00:49:00.000
but that won't tell me what
the sequence is that I have to,

00:49:00.000 --> 00:49:05.000
I mean, I was looking to try to get
a primer that matches the insert.

00:49:05.000 --> 00:49:09.000
And I don't know what the insert is.
So, how am I going to get a primer?

00:49:09.000 --> 00:49:13.000
Oh, I know the vector.
The vector is well known.

00:49:13.000 --> 00:49:17.000
It sequence is in the catalog.
Let me instead just use a primer

00:49:17.000 --> 00:49:21.000
that happens to sit in the vector,
and I'll match to a known sequence

00:49:21.000 --> 00:49:25.000
to start with, and then
I'll sequence into my

00:49:25.000 --> 00:49:29.000
unknown territory. So, this
is how you get the initial

00:49:29.000 --> 00:49:33.000
primer was you arrange that your
initial primer is sitting in known

00:49:33.000 --> 00:49:37.000
vector sequence. All right,
so you can now sequence

00:49:37.000 --> 00:49:40.000
DNA. I've got to say, I've
taught this course for a little

00:49:40.000 --> 00:49:44.000
more than a decade,
and being able to say,

00:49:44.000 --> 00:49:47.000
now we can routinely sequence
about a million letters per machine,

00:49:47.000 --> 00:49:50.000
and 100 million letters per
day, and things like this was not

00:49:50.000 --> 00:49:53.000
routinely the case. When
we started teaching this

00:49:53.000 --> 00:49:58.000
course, I was
describing what we di