good evening I’m Mary McGee I’m the
executive director of the forum and I am thrilled to welcome you to the forum
Columbia’s newest resource and the gateway to Columbia’s Manhattanville
campus we are thrilled also to be hosting tonight’s brain insight lecture
I just want to remind you to turn off all your electronic devices so you
really can enjoy and engage with tonight’s events I want to point out to
you the exits the four exits if you need to leave early I would suggest you take
one of the upstairs exits so you don’t need to enter exit past our speaker
bathrooms are out the doorways to your left and one other thing tonight’s
lecture will be recorded video recorded and broadcast so you should be aware of
that and with that said welcome and I will introduce you to mr. Costa the
executive director or the chief executive officer of the Zuckerman
Institute good evening everyone it’s really great
to see such a turnout at our first event at the Forum I hear there people still
wrapping around downstairs without coming as the lecture proceeds so it’s
really my pleasure to introduce you to the first event of the Stavros Niarchos
foundation brain inside lecture vizier and tonight’s event is gonna be special
and we’re gonna hear about how to find the barcodes in our brain and how a
hundred billion neurons can have identity and connect to each other these
wonderful series of lectures is made possible through the generosity of the
Stavros Niarchos foundation and we really are grateful to have them as a
partner in this in addition to the series of lectures lectures a lot of
these content is then transformed into content that goes out into the schools
into our education program and this teacher scholar program that’s
complementary is also sponsored by the Stavros Niarchos foundation so I would
like to thank the foundation for the support thank you this is really an exciting time for
neuroscience at Columbia and for the Zuckerman Institute who is the building
right next to us also designed by Renzo Piano this is an institute where people
interested in the brain but from very diverse disciplines are working together
to figure out simple things that puzzle must like for example how do we remember
what we like and we don’t like how do we love and hate how do we taste sweet how
do we smell how are we transforming the world making decisions and so I invite
you and those of you that are new to these to follow us to come to the
lectures there’s many other events and if you’re interested in partnering with
us and and support many of you here have done there thankfully you’re welcome
this is gonna be a great journey we don’t know when it’s gonna end but it’s
gonna be fantastic and we remain honored to march Tuckerman
for his vision in establishing and endowed this institute to the green
foundation for providing the wonderful building in which the Institute this
house and again for many donors here including the Stavros Niarchos
foundation for the help in supporting this endeavor
now don’t what matters it is really really a great honor and a pleasure and
a bit nervous to introduce professor Tom maniatis to you today
Taman Yaris is really one of the greatest molecular biologists of all
time he was the one that developed from fundamental recombinant DNA methods that
we all use in our lives today he pioneered the generation and screaming
of cDNA libraries and genomic libraries leading us to the discovery of many many
genes he also used these methods to identify and characterize the DNA and
RNA regulatory sequences so what’s happening there and also to find out how
things like RNA splicing works or how signal transduction pathways can
actually control gene regulation use this knowledge also to with his team
cloned the first human gene what’s extraordinary about Tom is that not only
he did all these two developments and discoveries but he didn’t keep them to
himself he broadcasting to society and to science in general and for example I
learned to work from the book he co-authored the molecular cloning math
manual which I think many of you that are in science or study here at Columbia
have used and he has used this knowledge not only to find fundamental mechanisms
really about biology like what we’re gonna hear today about protocol engines
but also to have insights into these like ALS so it’s it’s not a surprise
that tom has had a brilliant and diverse a career with positions at Cold Spring
Harbor at Caltech apes Harvard and luckily for us now at Columbia where
he’s the easy door atom and professor at the Department of Biochemistry and
molecular biophysics and also my upstairs neighbor at Zukerman Institute
he continues to put his knowledge at the service of the community and so he’s the
co-founder and chairman of the New York genome Center and he’s also the director
of the precision medicine initiative at the Columbia Medical Center he has
received many awards including the scientific achievement award from the
American Medical Association the Lasker Award and he was elected to the National
Academy of Sciences he’s a fellow of the American Academy of Arts and Sciences
and also a member of the National Institute of Medicine tom is really an
inspiration as a scientist and as a person and as you will hear tonight he
continues to push discovery every day we’ll hear about the important role of
protected hearings proteins in brain wiring which is a complex but
fascinating story and I hope you’ll appreciate the complexity but also
admire the capacity that tom has of making things so complex accessible to
everyone Tom thank you so much well Thank You Rui for that very nice
introduction I’d like to thank the Stavros and Rawkus foundation for
supporting this lecture and the many other things they do both in New York
City and and internationally I’d like also to thank the organizers for
inviting me to speak and to indicate at the outset that I’m going to try to
cover 20 years of work that’s very complicated in 50 minutes and for a
broad audience so you’ll please forgive me if I slip or I either say something
too complex or too simple so as it indicates in this slide so the original
title is finding the bar codes in our brains using genetics to provide an
identity for the brains 100 billion neurons I think that was done last May
and I would like to add an alternative or at the same time title and that is
how genetics and genomics can provide insights into the brain inter brain
wiring and neuro psychiatric disease and what this is really about is is
following mechanism and I should say that I began this interest on a very
simple way on a bacteriophage lamda working in the laboratory of mark Pesce
who’s here tonight and it was that work that really began to crystallize how you
look at a problem how you really identify the details for how a machine
works how in this case how gene expression works and so what you’ll see
I hope is that starting with the attempt to understand a gene cluster which I’ll
describe has led one step to another from gene regulations of protein
structure to Mouse genetics and so I hope I can make make it through this so
start off a various simple come a very simple concept and
that is I’m going to describe two principles of neural circuit assembly
and and it’s important if you see oops sorry it’s important that that you see
there are many others because it’s obviously a very complex problem of how
you assemble a hundred billion neurons into a functional brain and there are
all sorts of things about guidance how two neurons get to where they have to be
how they engage and interact with each other in a specific way and all those
things are being studied it’s a sacrament Institute and I’m gonna focus
just on these two problems the first one is called self avoidance and what that
is is that neurites axon and dendrites are the same neuron must avoid each
other to prevent clumping and non-uniform distribution so what you see
here these are the dendrites extending from from the main body of the neuron
and at the end of the axon these axon terminals and both of these have to
avoid clumping and getting in each other’s way so there has to be a
mechanism for doing that and also it’s important for individual neurons to tell
themselves from others they have to when their next next to another neuron they
have to know who is who in order to make the right connection in before on the
right circuits and this is this is illustrated here you can see that these
are two neurons interacting and you can see their interactions between axon and
dendrites with the cell body and in dendrites with dendrites and so all this
complex interaction really requires every neuron to understand who they are
and so the question is how does that happen
and the the simple description of this is shown here in the wiring of the
retina which the faculty works on here and that is there is a
series of neural circuits that extend from the neural from the photoreceptors
which detect light in color into the transmission of this information to the
brain and so among these cells are something called starburst amacrine
cells and they’re at a particular layer in the retina and they interact with
each other and with other neurons to form the appropriate circuit and this is
how they look in in a planar view and you can see each of these have this
they’re called starbursts because they are beautiful stars that emanate from
the from the cell body and extend and you see they’re all uniformly
distributed and moreover so that self avoidance say they don’t get tangled and
the dendrites don’t get tangled but they see their neighbors as well they touch
their neighbors and they don’t invade their neighbors space and in fact as
shown here they actually interact so this is called hetero neural or non-self
discrimination so this is just one small example of the kind of complexity that’s
that’s required to build a functional circuit and the the presence of this
these protocol hearings which I’ll discuss play an important role in that
so what’s shown here is the sort of sequence of events of their laying down
of this particular circuit so the cells start dividing they they first displayed
the self avoidance or the dendrites do not touch each other or clump
they then tile and then in this tiling engage their neighbors in a way that
leads to the assembly of a circuit so what are the underlying mechanisms for
self avoidance in tiling and as you’ll see today the ultimate answer lies in
our genes in the proteins they encode that is how how do the genes in the
regulation of their Russian impose on this incredibly
complex problem which you can see is it’s a equivalent to the branching of
arbors of a tree that do not that manage to take in the maximum amounts of light
in order to feed the tree so I have to start off with a very simple slide to
illustrate where we’re going with this and that is this is the way the nucleus
of a human cell is organized this is the nucleus the cytoplasm and in the nucleus
there are chromosomes this is a metaphase chromosome and each chromosome
has highly compacted DNA which is bound to proteins and if you unwind this you
see that there is something called chromatin and nucleosomes
these are histone proteins that bind to the DNA and organize and compact it and
ultimately you see naked DNA which is the double-stranded DNA which encodes
proteins so this is the the complex organization of the nucleus and an
underlying question related to what we’re going to talk to is how is the
genetic information read out at the right place the right time and that’s
what gene regulation is about and there’s been 50 years of research on
this to understand that we’re now making great progress in this understanding
well and I should say missed the point here is that the story that I’m about to
tell you started 1999 and it was just about the time that the human genome
project was ending in the determination of the nucleotide sequence of 3 billion
base pairs of DNA in this this is the science and nature magazines that
announced that breakthrough and while that was happening the data was made
available for everyone and that’s really a principle of genomics that were
to aspire to that making this open-source data and we were able to go
through and look for a particular gene that we are interested in I won’t go
into the reasons why we’re interested in but we were able to go on a database and
pull down bits of sequence and then assemble them into a long region and
this region is shown here so this is how metaphase chromosomes of a human cell
looked and you can see there are two copies of each and on chromosome five
there’s a locus which is called the clustered protic it here in jeans this
is a multi gene family and it is organized in a remarkable way what you
see is each of these colored boxes corresponds to a gene that encodes a
very particular protein and it has both variable exons and I’ll explain what
that means that encodes an isoform of a particular protein so an isoform is an
is a very similar protein that has some changes that make it distinct and unique
from the others and you’ll see why that’s important and this region spans a
million base pairs of DNA on chromosome 5 and so when this was discovered it it
was obviously remarkable and we had to begin thinking about what this
organization meant and what this gene cluster does I should say that tykishia
yogi’s lab in Japan had identified cDNA clones corresponding to this region and
it demonstrated that that they have common 3 prime ends the in the ends of
the RNA which is the constant region which which I’ll explain in a moment so
I’m going to take you through four steps the first is the generation of the cell
surface identity code which I’ll begin with the nature and reading of the code
which requires understanding where these proteins are made and where they
function the function of the code and that is in mice
they they’re responsible for self avoidance and tiling and then finally
end with the medical implications in neuropsychiatric diseases so just as a
brief primer so that we are all on the in the same phase this is how genes look
in human cells they’re not simple linear relationships they have interruptions in
them and these interruptions are called introns which is shown here so the
region that encodes our protein is an exon as you see here and it’s
interrupted by introns and there is a mechanism that that generates a
precursor RNA so it copies this entire region and then a process called
splicing which generates in this case alternatives of different combinations
of these exons as you see here so 1 2 3 4 5 1 2 4 5 so it’s skipped 1 and that
encodes then a series of proteins protein a B and C so this is a
remarkable mechanism that generates enormous diversity in the human genome
and is is required for most of the fundamental biological properties that
we know about so let’s just focus on one piece of this gene is of this cluster is
called the protocol here an alpha gene cluster and what you see here is that
this is the linear array of the gene these are the coding regions for the
protein and what we demonstrated many years ago that each one of these coding
regions has immediately upstream something called a promoter and the
promoter is where RNA transcription initiate sand reads through and and down
here there is a a three prime splice site so their
splice sites in this that are necessary to to drive this mechanism that I
mentioned before and splicing and so what happens as transcription begins
here it goes through all of these regions it happens to start at eight and
then there’s splicing between this piece here and this to generate a messenger
RNA which encodes the mature protein and that’s what’s shown here so this is the
messenger RNA it has this variable X on alpha eight and then these three
constant regions which are shown here and if we look at the protein sequence
what we see is this protein has an extracellular domain which has six of
these domains here a transmembrane domain in the constant region and this
piece is inside the cell is called an intracellular domain and so the question
is how are these promoters chosen in a in a in a particular neuron a nice
example of this is provided by the olfactory system which as you know was
discovered by Linda buck and Richard Axel which led to the Nobel Prize over
20 years ago and what this system does is it provides the diversity necessary
to detect odors but I show this because as you can see here at the epithelium in
the nose which is right here you generate a whole series of olfactory
neurons each of which have a different receptor they migrate into the brain and
they they form very specific regions called glomeruli which are sort of the
relay stations to the deep brain that really sorts out this information and to
four sensory perception so if we look at this by using a method called single
cell sequencing and so that’s a technology it’s really just been
developed recently it’s we’re able to sort these cells and
look at an individual cell and ask what are the messenger rna’s that are present
there and what we see is that these three neurons which were olfactory
sensory neurons which have different colours actually express a series of
alpha beta and gamma protocol here ins that are random so you can see five
seven one six five and in this red neuron this yellow neuron expresses
seven a eight and ten and in the corresponding ones here
so each neuron then each olfactory sensory neuron is expressing a random
set of these genes that are on the cell surface and so this and so the question
was how how is this generated and what does it do and so this is sort of an
overview of what happens here is the again an illustration of the genes
maternal and paternal chromosomes and you can see you see various promoters
firing in each of the two chromosomes in a completely random stochastic way they
give rise to these transcripts that make these proteins and you can see that
there are over 58 isoforms encode in about 15 random sets of these and
individual neurons so what so how does that happen how is this generated and I
need to show you another important fact and that is the DNA sequence regulatory
elements that are required to turn genes on and off and so what you see here is
an important principle which is a Illustrated very simpler simple here is
that the copying of the gene the transcription occurs at something called
a promoter which is indicated here and on the promoter there’s assembled a
large protein complex where in which the enzyme RNA polymerase that generates
the transcript is it’s shown here and the key is that they’re highly specific
DNA binding proteins called transcription factors that bind to these
sequences in thereby activate transcription and they do that by
recruiting the components necessary for this assemblage and so that’s the the
fundamental principle that was worked out in yeast actually by Mark patashnik
and has been it has been shown to be the case for higher the highest eukaryotes
including humans so with that in mind one of the first things we did was to
try to identify those enhancers and promoters within this gene cluster and
this is this summarizes several years of work and very deep experiments which
involve Mouse genetics and so on but the bottom line is that the is that
we identified these enhancers which are the blue diamonds here and there’s
spread throughout the locus and and the promoters and so we have a complete map
then of the regulatory sequences within within these clusters to begin to
understand how the gene is expressed so again a long story very short what
happens is that there is an enhancer happens to be called HS 5-1 it’s located
roughly 300,000 base pairs away from the promoters it activates and we’ve shown
through a whole series of experiments that that occurs the activation of these
genes occurs by DNA looping between the enhancer and for example in this alpha
12 in order to do that there are proteins that bind to the DNA and
they’re very specific proteins that do a special job they happen to be called HS
v 1 and ctcf and the purple ones here are the ctcf side
and these rings are are the cohesin which are proteins that wrap around DNA
and hold them together so how does this work the what we’re able to do by using
various methods and in genomics to show that ctcf and cohesin bind to the
enhancer of two sites they’re shown here one here and one here and they bind to
the promoter at two sites right here and right here and in between are some of
these nucleosomes and other transcription factors but this is
basically the activated state of the gene so how does this occur major
advances have been made only very recently in actually leonid Murray’s lab
at MIT which indicates how these this complex of ctcf and cohesin assembles
the appropriate enhancers and promoters to each other and that’s a very complex
process but it I hope to be able to explain how this works and that is that
when when you see that if there is a ctcf motif here the the the cohesin
binds to the DNA and then it extrudes DNA it moves it through these rings and
tell the ctcf binding sites come together and then it stops if there is a
ctcf protein bound there and so it’s a it’s a way of bringing the right
promoter with the right enhancer at the right time and so if I could have the
video all I’ll show how this actually works this is a video from the Marigny
lab may have the video please yeah okay do we have the sound okay so this is
chromatin these are the ctcf binding sites and you can see the cohesin and so
the the CTF sites are a barrier that basically determined when two pieces of
DNA connect and loop and so when cohesin is halted in both
directions a lasting loop is made so there’s this constant dance of these
regulatory elements that are searching each other out to do the right thing the
right time and if ctcf is not bound they they test they try and they move on to
the next so it’s the CTF bound to these sites
that really determines where this process ends and the loop is made
between the enhancer and the promoted so we can actually see this in cells now
using the latest technology we can see when DNA loops where it loops and we can
do that even in single cells so it really and I should say that this this
shows a artists view of cohesin bound to these two pieces of DNA so how does this
relate to stochastic expression of the protocol here in locus again this is
years of work that I’m distilling down to just a picture of how it works so
here’s the protocol here in alpha gene cluster and what you see here are the
ctcf binding sites would correspond to these two promoters and the black
circles indicate that the DNA is methylated and I didn’t tell you that
when the DNA is methylated ctcf cannot bind and so when cohesin does its thing
ctcf is not there so it doesn’t make the make this loop however if you see if
what we’ve been able to show by using very sensitive genomic methods is that
the first thing that happens is that this promoter which is actually within
the coding sequence is activated and makes an anti sense leak link RNA that
goes through the promoter upstream and as it does so it D methylates this
promoter it takes the methyl groups off so that ctcf can bind and it does so
through an enzyme called tet 3 which I can’t I don’t have time to explain but
now you have both of these sites that are unmethylated and so ctcf can bind
and when that happens it binds to both the enhancer and the promoter forms this
loop and activates transcription and the remarkable thing is that this is
propagated in all the subsequent generations so if this happens in a
early progenitor cell the the loop is made and then it survives DNA
replication until you have the post mitotic
and so it’s it’s a remarkable process in it and it eliminates any distance
dependent because it always is determined by where ctcf is bound so you
can see here the cohesin keeps moving until it hits the first ctcf bound and
so that that appears to be how this this random generation of protocol here
instance accomplished so so what does the code do the first thing we have to
understand is how its read I mean what does it do at the cell surface and this
involves a very productive and several-year collaboration with the
laboratories of larry shapiro and barry Hoenig here at at the zurka man
institute and i want to quickly again go through this it’s a very complicated
story but it really comes together in a beautiful and elegant way so the protein
that’s encoded by this cluster is looks like this I’ve shown you this before
there’s a part of the protein that sticks inside the cell is called the
intracellular domain and then all of these all of these other domains which
through biophysics through protein chemistry their function has been
determined and so this is the dimerization interface where where two
proteins come together so we discovered that they’re not monomers inside the
neuron they’re actually dimer so they’re two of them that stick together and that
what we discovered through a large series of of experiments is that these
are actually protein-protein interaction interfaces so they’re like they’re like
velcro they stick to each other but the important thing is that they’re very
specific so one protocol here and we’ll interact with only its identity and so
when it’s assembled at the surface it gives a very specific code and because
of this unusual structure which was worked out through biophysics
experiments or x-ray crystallography determining the atomic resolution and
atomic resolution the structure the proteins the what you see is that the
they interact with each other because to dimers can interact through these
interfaces they can do this so it makes assists trans tetra mer so now they’re
now four but most importantly they’re capable of extending that more or less
indefinitely to make a long lattice and so this this solved a big problem having
to do with the sufficient diversity because now you you generate these long
chains of just incredible diversity if they don’t fit as I’ll show you in a
moment they don’t work so this just shows you with the ability to make these
dimers you can make this huge collection of system errs each one is highly
specific we will only interact with itself so in
a very important point here which was established in invertebrates in
drosophila by Larry Zabur Skees lab at UCLA using a completely different
mechanism that I don’t have time to get into for generating diversity it’s
actually the most striking example of alternative splicing but this is the
important point i need to make here and that is that if neurons from the same
neuron touch each other they recognize through these protein protein
interactions and that leads to repulsion they push each other away we don’t
understand the mechanism that’s the next big step and understanding this problem
but it’s a fact when they touch and they engage their repulse and so if there are
neurites from different neurons as shown here they don’t form this engagement
there’s no repulsion so this homo philic interaction between distinct protein
isoforms results repulsion between the neurites and the lack of interactions
leads to crossover and clumpy so this is a very important experiment
but relatively simple so what we’re doing here is we’re taking the coding
sequence for these isoforms we’re hooking it to a plasmid and putting it
into a cell and we can lay label one set with a green label that’s fluorescent
and one with red and what you can see is that if you put in exactly the same
protocol here in coding sequences they’ll make mixed aggregates if they’re
perfectly matched between the two but if they’re different they make separate
aggregates so they can’t they can’t form multiple interactions so the experiment
was done was to test a series of combinations two isoforms three up until
five mix them and then ask whether they do this or this and then see what
happens when you have a single mismatch for the isoform and this is what happens
and this is really an important result so what you see here is that this is
what happens when a cell is expresses this alpha B 17 b6 and gamma C 5 they
form these mixed aggregates because they can all interact with each other however
if we have a single mismatch here the mismatches here in alpha 9 and here
invade 18 and so on you see that they never form mixed aggregates they form
separate and so what this says is that there is a mechanism that requires
perfection in the interaction at the cell surface it does not allow any
mistakes and that was a that was a puzzle but thanks to both theoretical
calculations and some biophysics a proposal was made that that what’s
happening here is that since they can form a lattice as shown here through
these dimers of dimers when they make a long chain you see an avoidance signal
and they work but if you have a this match the the lattice only extends
to a certain distance and because of that either because of the the affinity
or the intracellular activation of signaling pathways it doesn’t work and
so it’s a yes sort of this chain termination idea and so that was a that
was a proposal and additional experiments were done and this is a
beautiful experiment that Kari Gooden and a postdoc in the Shapiro and Hoenig
labs did she determined the three-dimensional structure of a cyst
trans tettemer and that’s shown here and what you can see that in theory it’s
capable of making an extraordinary lattice that could be indefinite and it
was this structure as you see here that then made it possible to test this idea
by putting these into cells and doing something called electron tomography
which I’ll show here and I don’t have time to go into detail in this but
what’s done is to is to express this this protein in Olympus ohm this is just
membranes and what you can see here is that when you express it and mix these
you can beautifully identify these lattices there you know you can see them
as it’s like scanning down through the sections you know I’ll play that again
and you can see that as you as you go through the lipice ohms you can see
these amazing structures forming like this and because they had the crystal
structure they know the exact dimensions and the parameters of the protein it was
possible then to as to take those two sets of data and show at a very high
level of resolution that in fact what happens they do form a lattice and this
shows in of how that conform so the the idea then
as I said earlier is that when you have the same neuron they make this long
lattice and with this intracellular signaling they avoid each other and for
those that are mismatched its below the threshold for signaling so this is
coupling in the protocol here in zipper two structured elements in the cytoplasm
and this mismatch interferes of the localization to contact and so the
signaling is not formed so you can see that this amazing set of very high-tech
very difficult experiments in structural biology and using the cry electron
microscope it was possible to really sort out how these how these molecules
work so how about the function of the code well in order to study that we had
to go to mice and that required some manipulations of their genomes and so
this is this is a again the description of the cluster as you can see here the
alpha beta and gamma and so what we did was to systematically delete each
cluster and and then attempt to demonstrate the function through the
effect on wiring in the brain we do that by crispers that I’m sure
everybody in this room knows about of using these amazing complexes that use a
guide RNA to specifically introduce a nucleus a very specific site in DNA and
by having to guide RNAs that recognize the boundaries of each cluster and of
the total cluster we’re able to cut them out they’re repaired in the mouse embryo
and result in a very precise targeted deletion and so this just shows what
happens you take the protein that’s essential nucleus that has these guide
RNAs you inject them into an reowww transfer the embryo into into a
mouse and then generate their progeny so you can very quickly construct a whole
series of deletions that you can then test so this is what happened this is
the whole cluster it was deleted the Alpha the beta and gamma separately or
the entire cluster is shown here I should say it took a very long time to
really begin to understand how they work and there are examples in which a
deletion of a single cluster has a clear phenotype but they’re examples in which
you see no phenotype with deleting any one of the clusters individually but
when you click when you delete the whole locus you see a profound difference so I
just want to quickly go through that this was a beautiful demonstration of
self avoidance this is the starburst amacrine cells again and what you can
see here is that these cells have this wonderful image of a star and that in a
mouse in which there was a deletion of the protocol here in gamma locus you can
see that everything clumps because they these these dendrites can no longer
recognize each other and so they just clump and so it’s a beautiful
illustration of the importance in this case of the gamma cluster there was a
very different thing seen in olfactory neurons which is shown here so we’re
looking at mice in which we see the olfactory bulb in the mouse and this is
how in a in a coronal section how these glomeruli look and again this is how it
looks in a normal Mouse these are the olfactory neurons that are converging to
form a glomerulus and this is the coronal section in which you can see
these beautiful globular structures that they’re called glomeruli and so what we
found is though if we delete any one of the single clusters alone we saw no
effect on olfactory wiring it was surprising but we were able to delete
the entire cluster and this is what we saw so what you can see here is that
this is the wild-type case this is a heterozygous deletion in which the
entire cluster on one of the two chromosomes were deleted you still see
these in highschool and marry lie but in the triple knockout you can see you no
longer see glomeruli they’re completely diffuse and so we wonder what was
happening there so we we we looked at neuron single neurons in using a method
although mention later to actually label a single neuron in the wild-type which
you can see here you see this beautiful extension into the glomerulus the
fingers are separated but in the Tri cluster deletion in the single neuron
you see it’s totally clumped so this is a an axonal self avoidance mechanism and
it shows I think quite beautifully the importance of the cluster in the in the
wiring of the olfactory system so this is the idea that normally in the
wild-type case you have this display of protocol hearings on the surface of each
of the of these extension and that when you delete them they clump and that’s
what we exactly what we saw the wild-type looks like this and the mutant
looks like this so another example of this and it’s a cluster that we see a
phenotype by itself and this work was also done by Yogi’s landed
about at the same time and what we see here is that if we delete only the
protocol here in alpha cluster and then we look into the hippocampus in a way
that we can detect certs and urging neurons and so serotonergic neurons
extend from a single site are spread through the brain and their function is
to release serotonin locally as a as a percent as a nervous zero ligand and so
what you see here is that this is the wild type case you see they’re sort of
randomly distributed within the hippocampus and there are certain
regions that can be looked at specifically but this is what happens in
the in the protocol here in Alpha knockout and you can see rather than
being diffusely spread they’re all clustered there clumped in one part of
the hippocampus which is indicated here so this is a phenomenon that I mentioned
earlier this is tiling so these are neurons of the same type that are
spreading into us in a region and they repulse each other in order to make this
uniform distribution so whole series of experiments was done and I won’t go
through this in detail this is what happens if you knock out the beta
cluster nothing happens the gamma cluster nothing happens deleting all
these alternate isoforms has no effect however when you delete these two C
types here you see the you know the the clumping and we actually were able to
show that a single gene protic it here in alpha c 2 is completely responsible
for this phenotype and if we look at the expression we see that in these neurons
really the only thing that’s expressed is protic it here in alpha c2 so there
are two separate genes within the cluster one type of gene these alphas
are involved in tiling and the alternate isoforms are involved in self avoidance
so fine oh and I should say a really important
point in this and I’m not going to go through the data but when we look at
these mines there are behavioral defects that we can see and mouth
neuroscientists and geneticists have involved a whole series of assays in
mice that mimic for example anxiety and depression and by that criteria these
mice are very fearful and very depressed and so they actually have the phenotype
that is characteristic of autism and so of course that suggests that maybe in
humans this gene cluster provides the same function and what you can see here
is that thanks to the Simons Foundation and all the work they’ve done they’ve
identified a whole series of genes 150 or so genes that by virtue of where
where the mutations occur within family organized cohorts a number of genes that
contain DNA sequence variants are associated with the disease and they
haven’t been shown to be causative yet but they clearly by all these criteria
are involved in the disease and what you see here among the various genes on the
cell surface are our protocol hearings the cluster protocol hearings and also
these non-clustered as well and and so this is from the Simons Foundation
website and if you can see on here these are all the protocol here in alpha so
this nineteen eight beta gamma all through the cluster they’re these
sequence variants that associate with the disease and most importantly you can
see right here protic a turn alpha c2 is as present so there’s a real connection
between the genetic studies that have been done on autism families the
function of this cluster and mice and the protocol here in locust so I
just want to make one final point is that very massive studies have been done
to look at genetic variation within large cohorts this particular one is is
in schizophrenia and I I won’t go into explaining this chart
but just to tell you that anything that is above this line here it is a possible
associated variant that’s causal and what you can see is that in all of these
cases there are sequence variants that correlate with the disease all through
the protocol here in cluster so the protocol here in gene cluster is
associated with both autism and schizophrenia and there are also some
variants in bipolar disease as well so this is one of many genes that are that
are shared in common in these kinds of studies between those three
neuropsychiatric diseases so I just finished by summarizing here protic it
here in cell surface code is generated by stochastic promoter choice through a
remarkable mechanism involved the anti-sense RNA transcription DNA
methylation chromosome looping and cohesin extrusion the structure and homo
philic interactions of protic you turn proteins at the cell surface of neurons
allows the identity code to be read through a highly specific interaction
and this is as I said was done really the atomic resolution and then finally
the protocol hinge gene cluster functions in normal circuit assembly and
behavior in mice as shown in this experiment here and that DNA sequence
variants in human DNA associate with neuropsychiatric diseases autism and
schizophrenia so we go from 20 years ago when we first discovered this cluster
through a whole series of basic research on what these genes do how they’re
expressed how they’re regulated and we end up here really having
some very important insights into the genetic basis of these human diseases so
I want to thank all the people who worked on this project over the years
this generating the protein here an identity code are shown here this is den
Daniella Kenzie uh Kenzie oh who now moved on to UCSF Chamaco no Kazi who’s
an MD PhD student in my lab Eliot coffee who is now a graduate student at MIT
sandy ruck amar who’s at UCSF and Sean O’Keefe and we collaborated with the
Matt Simon’s lab at Yale and in the work that I told you about the CTF cohesin we
had a very productive and and wonderful collaboration with Stavros Lombardi’s
lab Aden Horta Kevin Monahan and Rachel Duffy so this as I say this was years of
work but it required all these very devoted teams the structural work was
done as I said by Larry Shapiro and Barry Koenig’s lab and members of their
laboratory a two and maxime Chevy in my lab and importantly in the last work
that I showed on this crime tomography was led at the through Larry chaperone
the Hoenig labs fried Bridgette Gary gur and Clint Potter at the New York
structural biology Center the Simon’s electron microscopy centers and these
are the people who really did the work so this is an incredible collaboration
using all the latest technology to to bear on this problem
finally all the functional work was done by wei zheng Shan who’s now at
leveraging george Monto phoresis now at Caltech and Chim Chim maca and daniela
made important contributions along with Sean and we had a nice collaboration
with Frank poloz lab with you Sookie Kiribati who did the single cell
experiments so with that I would like to again thank the narcos foundation and
remind everyone that the next lecture is going to be given in this series by
Jeannette wing who is the director of the data science sensitive at Columbia
thank you thank you thank you for this fabulous
lecture I’m sure there are a lot of questions I have quite a few we’re gonna
have two microphones on the sides and as usual we encourage questions there can
be questions from a lot of you and we’ll stay here until there are questions so I
can start with one so Tom you found that one specific protocol here was
responsible for quite a defect in the hippocampus so although there’s this
amazing genetic code for all over the brain there there seems to be quite some
specificity for which are in which right now is there some meta regulation on
this well that is really where the research is going right now and that
from the few things we’ve looked at there really is cell specific
differential expression of both these alternate isoforms in the C type and
it’s true for alpha beta and gamma and we’re just beginning to look into that
and so for example dope nergic neurons make no alpha okay and certain nergic
have the expression that I showed so what is happening here is that
superimposed in all this complexity is the cell specific differences in the
differential use of clusters of alternate and the C type isoforms so
there’s a whole layer of regulatory complexity that’s superimposed on this
fantastic so any takers or is everyone wanting to go home and rest
as one first of all I’m gonna say that that was
incredibly fascinating I had no idea that all of that was going on in my in
my brain I’m a cognitive researcher and one of the things that you said in the
beginning was bringing different disciplines together so I’d like to
offer you my perspective not on the technical things that you talked about
because I can’t touch that I only have understood like 20% of it but when
you’re migrated and just get the frame then that’s where I would like to add my
comments I’d like you two to compare the human brain to a computer system all
right for this analogy your cognition is equivalent to an application program
like Microsoft Word you have a subconscious system like a subconscious
execution system semi conscious that’s equivalent to the Windows operating
system a hardware architecture of the computer is equivalent to the neuronal
processor that you were explaining as to how all of this was put together and
that’s your auroral processor your memory architecture and so forth all
right and then you have on the computer level
the off/on circuit control architecture like they’re talking about quantum
computers and so forth and that’s where you began to talk about neuronal
architecture of how they keep away from each other and don’t touch each other
and so forth okay and then you have the circuits atomic and subatomic structures
like the quantum bets and so forth and that’s where you have the genetics
the genomics now but schizophrenia before I say that if you have a a if
somebody hacks into your computer and you have a virus you’re not gonna look
at the subatomic level you’re gonna look at the application at
the operating system level schizophrenia is a content and effective content
disorder you can create schizophrenia I’m sure there are many hypnotherapists
in here you can create schizophrenia using hypnotherapy all right you can
create the older identities and create the voices and so forth all right now
from my perspective I’ll wrap it up where I where I do see this type of of
of choking down I into what goes on and we’re in the brain and looking for the
identity of the neurons and the neuronal processor is in you identifying where
for example voice based neuronal activation is occurring so with all the
information that you have if you can identify that location and then design a
form of pharmaceutical solution to stopping that kind of activation and you
can cure your schizophrenia as opposed to using the psychotherapy okay so mm-hmm
we’re going to let everyone rest thank you tom again for the wonderful lecture
and I hope to see you here at the School of Journalism thank you for your support
please partner with us and see you again soon

Articles, Blog Tags:

Leave a Reply

Your email address will not be published. Required fields are marked *