Robert Tjian (Berkeley/HHMI) Part 1: Gene regulation: An introduction

- My name is Bob Tjian,
I'm a Professor of MCB

at the University of
California at Berkeley,

and I'm also serving as the
President of the Howard Hughes.

I'm going to spend the
next 25 or 30 minutes

telling you about some fundamentals

of one of the most important
molecular processes

in living cells, which is
the expression of genes

through a process called transcription.

Now, first, to understand
what gene expression means,

you have to have a sense
of what we tend to refer to

in the field as a central
dogma of molecular biology.

Another way to think about this

is the flow of biological
information from DNA,

in other words, our chromosomes,

which every cell has its compliment,

to be transcribed into a
sister molecule called RNA.

So, this process of
converting DNA into RNA

is called transcription,

and that is the topic of this lecture.

This process is very complicated,

as you'll see by the
end of my two lectures,

and it is very important

for many, many fundamental
processes in biology.

So, what I'm gonna
spend today's lecture on

is the discovery of a large family

of transcription proteins.

These are factors we call
them that are key molecules

that regulate the use
of genetic information

that has been encoded in the genome.

Now, transcription factors or proteins

are involved in many
fundamental aspects of biology,

including embryonic development,

cellular differentiation, and cell fate.

In other words, pretty much
what your cells are doing,

how a tissue works,

and how an organism
survives and reproduces

is dependent on the
process of gene expression.

And the first step in this
process is transcription.

Now, there are many other reasons

why a large group of people and scientists

are interested in transcription,

and another reason is that understanding

the fundamental molecular mechanisms

that controls transcription in humans

or in any other organism

can inform us and teach
us about what happens

when something goes wrong,
for example, in diseases.

And I list here just a few diseases

that we could study as a result

of understanding the
structure and function

of these transcription factor proteins

that I'm going to be telling you about.

And of course, the hope
is that in understanding

the molecular underpinnings of
complex diseases like cancer,

diabetes, Parkinson's, and so forth,

that we will be able to develop and use

better, more specific therapeutic drugs

and also to develop more accurate

and rapid diagnostic tools.

So, those are a couple of the reasons

why many of us have spent,

in my case, over 30 years

studying this process of
transcriptional regulation.

Now, to get the whole thing started,

I have to give you a sense

of what the magnitude of the problem is.

So, imagine that one would
really like to understand

how this process of decoding
the genome happens in humans.

So, as you may know,

the human genome has
some 3 billion base pairs

or bits of genetic information,

and that encodes roughly 22,000 genes.

These are stretches of DNA sequence

that encode ultimately a product

that is a protein which actually
makes the cells function.

So, as I already explained to you,

there's this flow of
biological information

where you have to extract the
information buried in DNA,

convert it into RNA.

And what I'm not gonna
tell you about today

is the process of going
from RNA to protein,

which is a reaction called
a translational reaction.

I'm going to instead just
focus on the first step

of converting DNA into RNA,

which is the process of transcription.

Now, one of the most amazing results

that we got over the last decade or so

was when the human genome
was entirely sequenced,

the first few that were sequenced,

we realized that actually
the number of genes in humans

is not vastly different
from many other organisms,

even simple organisms like little worms

or fruit flies and so forth.

That is roughly 22 to 25,000 genes

is all the number of genes

that all of these
different organisms have.

And yet, anybody looking at us

versus a little roundworm
in the soil or a fruit fly

can tell that we're a
much more complex organism

with a much bigger brain,

much more complex behavior, and so forth.

So, how does this happen?

Part of the answer to this
very interesting mystery

or paradox lies in the way
that genes are organized

and how they're regulated.

And one of the most striking results

of the genome sequencing
project was to realize

that a vast, vast majority
of the DNA in our chromosomes

is actually not coding for
specific gene products,

and that only roughly 3% of
the DNA is actually encoding

let's call those little arrows

that I show you on this purple DNA

are the gene coding regions.

So, you'll notice that there's
a lot of non-arrow sequences,

which I'll show you in
this next slide as green.

These are non-coding regions.

So, the vast majority, 97%
or greater is non-coding,

so what are these other sequences doing?

And of course, it turns
out that these sequences

carry very important
little fragments of DNA,

which we call regulatory sequences.

And these are the sequences
that actually control

whether a gene gets turned on or not.

And I'll be spending much
of the next 20 minutes

telling you about how
this process all works

and what these little
bits of DNA sequences

actually function to
control gene expression.

Now, the other thing that I
have to bring you up to date on

is this mysterious process
we're calling transcription,

which reads double-stranded DNA

and then makes a related molecule,

which is a single-stranded RNA molecule,

which is a informational molecule.

That reaction is catalyzed
by a very complex

multi-subunit enzyme
called RNA polymerase II.

Now, there's the roman
numeral II at the end of this

because there were actually
three enzymes in most mammals,

at least three enzymes that
carry out different processes

and different types of RNA production.

But I'm only gonna tell you

about the ones that make
the classical messenger RNA,

which then ultimately becomes proteins.

So, now one of the things that we learned

early on in the study of mammalian

or other multicellular organism
transcription processes

is that despite the fact that this enzyme

is quite complex in its structure,

it turns out to be an
enzyme that's nevertheless

needs a lot of help to do its job.

So, on its own, this RNA polymerase II

cannot tell the difference
between the non-coding regions

of the genome and places where
it's supposed to be coding

or reading to make the
appropriate messenger RNAs.

So, this sort of leads you
to think that there must be

a number of other factors
that somehow direct

RNA polymerase to the right
place at the right time

in the genome of every cell in your body

so that the right products get made

so each cell in your body
is functioning properly.

And this is where things
get really interesting

is some 25, 30 years ago,

a number of laboratories took on the job

of hunting for these elusive
and, as it turned out,

a specialized protein factors

that recognize these little
stretches of DNA sequences

that I've been telling you about

that make up the vast majority

of the non-coding part of the genome.

And how these proteins then can recognize

and ultimately physically interact

with these little bits
of genetic information

to then turn genes on or off.

Now, in this lecture,

I can't go into all the details
of the types of experiments

or the ranges of experiments
that many, many laboratories

have done over the last two decades

to finally work out this molecular puzzle

of how transcription works.

But I can tell you that
there are fundamentally

two major approaches that have been taken

over the last few decades
to kind of get a parts list

of the machinery that decodes the genome

and carries out the
process of transcription.

One is kind of the old style,

I'll call it bucket biochemistry

or take a live cell, crush it up,

spread out all of its parts
and then try to figure out

how to put it back together again,

that's what I call in vitro biochemistry.

And the other one is in vivo genetics

where you effectively use genetic tools,

mutagenesis to go in there
and selectively remove

or knock down or knock out
certain genes and gene products

and then ask what is the
consequence on that cell

or that organism?

Both of these technologies
are very powerful

and highly complementary,
and they continue to be used.

Today, I will focus primarily

on the in vitro biochemical techniques

which led us to the discovery
of the first few classes

of transcription factors.

And in subsequent lectures,

we'll go to more recent technologies

that allows us to sort of
speed up this whole process

of identifying key regulatory molecules

and how they work.

So, let's go back to the
sort of the basic unit

of gene expression, which is a gene,

here shown in the orange arrow,

and the non-coding
sequences surrounding it.

And you'll see that now I've
added a few more elements

to this purple DNA.

You see some symbols, a blue square,

a round circle that's pink,
and then a yellow triangle.

Those are just a way for
me to graphically represent

the little bits of DNA sequences

that I told you about that
are the regulatory sequences.

So, the little round one
happens to very GC-rich,

the triangle one is a classical element

that's called a TATA box,

I'll tell you about a little bit later.

And the blue one is yet
another recognition element.

So, why are we so interested
in these little stretches

of nucleic acid sequence in the genome

when it's buried amongst
billions of other sequences?

Well, these individual little sequences

turn out to be very important
because of where they sit,

you'll notice they're sitting
near the top of the arrow,

and they are recognized
by very special proteins

which are the transcription factors.

So, now I've showing you some symbols

with little cutouts which
fit into either the square,

the circle, or the triangle.

So, transcription factors,

at least one major family
of transcription factors,

are proteins whose
three-dimensional structure

is folded into a shape that
allows them to recognize

these short stretches
of double-stranded DNA.

In fact, largely through interactions

with the major group of DNA,

and I'll show you a structure
of one in a little bit.

So, now it turns out
that there are probably

thousands of these transcription factors

because the number of genes
that we have to control,

as I showed you, is in the
order of 20 or 25,000 genes.

And so, it turns out that you
need a pretty large percentage

of the genome devoted to encoding
these regulatory proteins

in order for a complex organism
like ourselves to survive.

Then the other component of this,

let's call it the
transcriptional apparatus,

is, of course, the enzyme
that catalyzes RNA.

And I already told you
that this enzyme on its own

can't tell the difference
between random DNA sequence

and a gene or a promoter.

These other sequence-specific
DNA-binding proteins

are the ones that must recruit

or otherwise direct RNA polymerase

to essentially land on the right
place and at the right time

in the genome to turn on
a certain subset of genes

that are specifically required
in a specialized cell type,

whatever cell you happen to be looking at.

So, that is kind of the
first level of complexity

of sort of informational interactions

between the transcription factors

and the more ubiquitous,

and I would call it promiscuous
RNA polymerase II enzyme.

Well, as it turns out,

it took several decades to work out

most if not all of the components

of this so-called
transcriptional machinery.

And it turns out in this
slide I'm showing you

things are already starting
to get more complicated.

So, not only do you have RNA polymerase,

but you have a bunch of other
proteins that go by names

like TFIIA, B,

you know, D, E, H, F, and so forth.

So, it looks like there
are going to be many,

many proteins that are necessary

to form the transcriptional apparatus.

And then on top of that

you need sequence-specific
DNA-binding proteins

which are already described to you

to further inform or
otherwise regulate the process

of when a particular
RNA polymerase molecule

should be binding to a particular gene.

So, that's the sort of overview,

now let me get into the specifics

and how did we actually discover
these family of proteins.

And it'll be interesting for you to see

how science in this field evolved.

Now, as is often the case

when you first try to tackle
a very complex problem,

and, of course, we didn't
really know how complex it was

when we began these studies,

but we assumed it might be complicated,

certainly would be more
complicated than systems

that we had already had some idea about,

for example, in bacteria
or in bacteriophages.

We took a lesson from our
studies of bacteriophages

and decided that to begin to dissect

the molecular complexities

of the transcription
process in animal cells,

we should start with viruses

because we knew that viruses
will enter these host cells,

these complex cells that
we ultimately want to study

and have to use the
same molecular machinery

to transcribe their genes

as the host mammalian cell would do.

So, this was kind of a trick

or a way to look at a molecular window

into a complex system
and try to simplify it.

And in our case,

the early studies of the
late '70s and early '80s

involved very simple,

one of these simplest
double-stranded DNA viruses

called Simian virus 40.

And Simian virus 40, of
course, is a monkey virus,

which was nice because
it's very close to humans

and many things that we could learn

about the way this virus uses its host,

which are monkey cells, to replicate

and to express their RNAs and genes

would be applicable to our
studies of humans, as you'll see.

And this virus was one of the first

whose DNA, its double-stranded
DNA of about 5,000 base pairs

was fully sequenced.

This was long before a
rapid modern day sequencing

was available, so this gave
us a very powerful tool.

It basically allowed us to
look at the entire genome

of this virus, which
was tiny by comparison,

only 5,243 base pairs.

But just that information
was already very important

'cause it very quickly allowed us,

for example, to map where the genes are.

And one of the genes encoded a protein

called a tumor antigen,

which turns out to be
a transcription factor.

This then allowed us to get our hands

basically to do biochemistry and genetics

on the very first eukaryotic
transcription factor,

which in this case
happens to be a represser.

That is a protein that
when it binds the DNA

just the same way as I showed
you for the the model case,

it binds through specific
protein DNA interactions.

But in this case, actually
shuts transcription down

rather than turn it up.

In the process of studying
the way that this little virus

when it infects a mammalian cell

uses proteins like T-antigen

to regulate its gene expression,

it became clear that it had
to use the host machinery

to do the process.

And that meant that there
must be monkey proteins

that are also involved in activating

or repressing genes of this virus.

And this then led us to
the most important step,

which is to transfer the
technology we learned about viruses

and how to work with the
virus transcription factor

like T-antigen to the cellular ones.

And I'm gonna give you just one example

of how the simple jump into the host cell

allowed us to discover the first
human transcription factor.

So, the question that we then asked

back in the early 1980s
was what host molecule

is regulating the expression
of transcription of this virus

when the virus is in the host?

And we knew from the DNA
sequence of the virus

that there were these six
very GC-rich snippets of DNA

that were regulatory
'cause if we deleted them,

the virus no longer would
express the gene of interest.

So, we knew that something
was probably responsible

for recognizing these GC boxes,

and we knew that it wasn't
a virally encoded gene

because we had tested
all of the viral genes

of which there were
only six to begin with.

So, we knew it had to be a host gene

and that led us to a whole, I would say,

family of experiments
that led to the discovery

of sequence-specific mammalian
transcription factors.

And as I said, we could have
taken multiple approaches

to try to address this complicated issue.

I'll just give you one example

of using in vitro biochemistry

to finally get our hands
on this key sequence

specific human transcription factor,

which, of course, has a
homologue in the monkey.

And the way we did it was very interesting

and simple in retrospect,

and that is recognizing the fact

that whatever this protein was,

it had to have the property of recognizing

those GC boxes that were sitting
next to the the viral gene.

We assume that it must
be a sequence-specific

DNA binding-protein, so all we had to do

was figure out a way to extract proteins

from human cells or monkey cells

and then try to fish out
those specific proteins

out of the many thousands
of different proteins

that were in this gamish
of cellular extract

that would be responsible
for discriminating

between random DNA sequences
and the specific GC box.

And I'll quickly run through
sort of the logic behind this.

So, what I'm showing you
here is a solid surface

with DNA coupled to it
that is highly enriched

for the recognition element, the GC box,

which should be the sequence

recognized by the protein of interest.

Now, we had no idea what this
protein was gonna look like,

how many proteins there
were gonna be, and so forth,

but we knew it had to
recognize the GC box.

So, we're gonna try to
fish this out of a pool

of many thousands of other proteins.

Now, the the key trick here

was that because all cell extracts

contain not only one DNA binding protein,

but, as I told you, thousands of different

DNA binding proteins.

But most of them, or in fact in our case,

none of the other of several
hundred to a thousand proteins

that could bind DNA actually
happen to recognize the GC box,

they just bind other DNA sequences.

So, to kind of favor our protein

being able to bind to our GC box

and not have to compete
with all the other proteins,

what we did was to add non-specific DNA

and mask stoichiometric excess

so that all the other proteins
that wouldn't recognize

the GC box would still have
some partner to hang onto.

And this trick worked very well.

So, having the specific
DNA on the solid resin

and the non-specific DNA
flowing all over the place,

we could capture selectively
the pink molecules here,

which are the GC box recognition ones,

and the blue-green molecules,

of course, predominantly
bind to non-specific DNA.

I show you one little
blue one on the column

because nothing works
perfectly in real science

and tells you that we have
to go through this process

iteratively to actually
finally obtain a preparation

that's purely pink molecules
with no green-blue ones.

Well, that turned out
to work very, very well.

And that whole process of
biochemical fractionation

followed by a direct affinity
sequence-specific DNA resin

gave us the ability to perform
a biochemical purification

followed by a molecular cloning
of the transcription factor

that encodes the protein SP1.

And then we carried out
a bunch of experiments,

which I'll tell you next,

to show that this protein

actually does activate transcription.

And of course, we went back and
we proved that this protein,

which turned out to be a
rather large polypeptide,

can indeed recognize the GC box.

And it doesn't matter if it's
a GC box from the SV 0 genome

or any other GC box that we
could find in the human genome,

it would find that sequence and bind to it

and then it would generally
activate transcription.

So, this led to the discovery of the first

of a very large family

of sequence-specific DNA-binding proteins.

Now, I told you that
the way these proteins

tend to recognize short DNA sequences

is to interact with DNA
through the major groove.

And here's a perfect example.

So, the thick blue model there

shows the actual three structures

that are called zinc fingers.

And the reason they're called zinc fingers

is because there are amino
acids that are organized

around a center that
contains a zinc molecule

which holds the three-dimensional
shape of the polypeptide

in a position just right

for fitting into the
major groove of the DNA.

And the DNA here is shown in pink,

and you can see that that blue outline

fits right into the
major groove of the DNA,

but not to the minor groove.

And one of the most important findings

was not only the discovery

of the first human transcription factor,

but the realization that most
if not all sequence-specific

DNA-binding transcription factors

have a similar structural motif.

That is to say some structure
is built to recognize

sequences in the major groove of DNA.

And these three-dimensional motifs

are recognizable as amino
acid sequences in the genome.

So, we can now much more
quickly scan the entire sequence

of a genome and identify genes

that are likely to be DNA-binding proteins

as a result of understanding
the structure-function

relationships of these DNA-binding
motifs like zinc fingers.

So, what I'd like to show you now

is that I've only
introduced you to one class

of transcription factors,

which are the sequence-specific-DNA
binding proteins.

Well, I think I gave you a little taste

of the level of complexity

that's probably going to be needed

to be able to build the machine

that's ultimately going
to be able to allow you

to transcribe every gene in
every cell of a human body.

So, that turns out to be a
much more elaborated machine

than what I just showed you.

So, I wanna show you now

what is sort of our
state-of-the-art thinking

about what is actually
needed to build the machinery

at a gene to allow it to be
expressed and transcribed.

And the term I want to introduce you to

is the pre-initiation complex.

And it's pretty much what it says.

It's the complex of multiple subunits

that has to essentially land
on the promoter of a gene

which will be designated
for later expression.

And this is a process that
is probably quite orderly,

that is there's an order
of events that happens,

which we, by the way,
are not entirely sure

exactly what the order
is or even if the order

is the same from one gene to the next,

but we can kind of see where
it starts and where it ends up.

And the pathway in between,

I would say is still a little bit murky.

And the story here again starts
with a little snippet of DNA

called the TATA box,

which I already introduced you to briefly.

It's an AT-rich sequence which
sits at the five prime end

or the beginning of many
genes, but not all genes,

maybe 20% of the genes might
contain this AT-rich region.

And that AT sequence is the signal

or a landmark, if you like,

for a particular protein to bind to it.

And that protein is called,

not surprisingly, the TATA-binding protein

'cause it's the TATA sequence.

And so, this represents a second class

of transcription factors.

These are not the type that
I just introduced you to,

which are gonna be
different for every gene,

the TATA sequence is present

in a very large number of genes,

so it can't be gene specific,

but it turns out to be very crucial

for our understanding of
how gene regulation works.

So, so you start with
a TATA-binding protein

finding a TATA box.

We later found out that
the TATA-binding protein

rarely functions on its own
and has a bunch of friends

that we call TAFs or
TBP associated factors.

And now you're talking about an assembly

of multi-subunit complex of
almost a million daltons.

There are somewhere
between 12 to 15 subunits

in addition to the TATA-binding protein

that make up this little
complex of proteins

that kind of travels around together.

And this is found in most cell types,

and later on I'll show you
in a subsequent lecture

that not every cell type

might have exactly the same
compliment of these subunits,

but many of them have
this prototypic complex.

Is this enough for building
the pre-initiation complex?

Unfortunately not.

It turns out that there
are a host of other,

I'll call them ancillary factors

in addition to the multi-subunit
RNA polymerase itself

that are necessary for you
to build up an ensemble

that is necessary to form an active

ready to activate transcriptional
pre-initiation complex

or the PIC.

And this is kind of the
picture we're getting to,

and even this picture
with many, many colors

and many, many different polypeptides,

you know, that adds up to probably greater

than 85 individual proteins

that all have to kind of fit
together like a jigsaw puzzle.

It's probably not even the whole story,

you'll notice I still have one
big red question mark there

because I think as we begin
to study specific cell types

and specific processes
like embryonic development

or germ layer formation,

additional components
that are not present here

in this prototypic pre-initiation complex

will come into play,

and that's a subject
of subsequent lecture.

But already you can tell that
the transcriptional machinery

is anything but simple.

So, can we get a better
idea of what transcription

might actually look like
and what's happening

when a transcription process takes place?

So, let me first of all say
that I'm gonna finish my lecture

now with a little cartoon,

which is our attempt to imagine

the events that take place

when you form a pre-initiation complex,

you bring regulatory proteins
to the activated gene

and what happens during this process.

Now, keep in mind that
this is at this point

mostly a cartoon that
is in our imagination

and only parts or if any
of this is probably real,

but it gives you a sense of the complexity

of the transactions
that have to take place

just for one gene to
transcribe and express itself.

So, let me show you the movie,

and then we'll finish
just by keeping in mind

that there's much to be learned.

And in my next lecture,
we'll go into the selectivity

of this process in specialized cell types.

So, now let's see what
this sort of this cartoon

of transcription looks like.

So, we start off with DNA

with some preassembled TFIID molecule,

and along comes this other green molecule,

which is actually a co-factor,

which then forms this very large complex

with RNA polymerase.

And then a distal
activator protein came in

and activated the process.

And this molecule, this bluish
molecule that's moved away

from the complex is
actually the RNA polymerase.

And that little yellow
sort of bead on a string

is actually the RNA product.

So, that gives you a sense of
things have to happen quickly

and yet it involves many, many molecules

having to assemble and then disassemble

to give you this reaction to happen.

And in my next lecture,

we'll go into more specific
aspects of this reaction,

and particularly during
embryonic development

and tissue-specific gene expression.