热搜第一，“清华女神”颜宁宣布离美归国！不满30岁成为清华最年轻教授，还曾当选美国科学院外籍院士！（附视频&演讲稿）

Hi. I'm Nieng Yan. I'm a professor in the School of Medicine at Tsinghua University.

I'm also an investigator of the Center for Structural Biology at Tsinghua University, Beijing, China.

Welcome to my iBiology seminar series.

In the first part, I'd like to give you a brief introduction to membrane transport proteins.

We all know that, on Earth, all life forms share one common feature -- that is, we are made of cells.

A cell is the most fundamental unit of life.

And despite different sizes, shapes, functions, all kinds of cells also share one common feature -- that is, they are enclosed by bio-membranes, mostly of lipid bilayers.

Not only cells, in higher organisms, within the cells, we're also compartmentalized by membranes.

Therefore, we have these organelles -- mitochondria, chloroplast, endoplasmic reticulum, the Golgi apparatus, etc.

The presence of a membrane defines the boundary of a cell or an organelle.

It protects the contents from the environment and it allows the simultaneous occurring of different chemical reactions, so they help life... they keep life in order.

So, we all know that bio-membranes are not only just lipid bilayers.

There are embedded proteins, which are also generally modified by polysaccharides.

However, even in the absence of the proteins, the lipid bilayer itself is complex and dynamic.

It is known that there are more than 1000 lipid species within a cell.

Therefore, a lipid bilayer may have distinct compositions.

And they are highly dynamic -- the lipids can freely diffuse laterally, and they can undergo flip and flop, and they can go in and out of the lipid bilayer.

While this lipid bilayer defines the cell, protects the cell, they also set up a barrier to prevent the free exchange of chemicals and information and energy in and out of the cell.

But we may not want the free exchange; life wants everything under control.

So in this lipid bilayer, it actually provides the perfect opportunity to set one tier of regulation.

However, for life to undergo metabolism, we have to undergo constant exchange of chemicals and information and energy...

again, chemicals, information, and energy.

So, by definition, by the chemical and physical properties of lipids, some chemicals can actually penetrate the bio-membrane.

Therefore, bio-membranes are known to be semi-permeable.

Some small molecules, such as water, glycerol, gas molecules, or small hydrophobic molecules, they are allowed to go freely across this boundary, so this is known as simple diffusion.

But for the large majority of chemicals, exemplified here by glucose, amino acids, and nucleosides, especially those charged ions, they cannot penetrate this bilayer.

Therefore, they necessitate specific transporter mechanisms.

So, before I go to my favorite topic -- transport proteins -- let me remind you that there's one way of group translocation known as vesicular transport.

Shown here, it's a simplified cartoon of endocytosis and exocytosis.

In addition, there's translocation between different compartments within the cell.

It's vesicular transport that allows the exchange of chemicals, in large quantities, with low selectivity, between different compartments.

Okay, in addition to simple diffusion and vesicular transport, there is a third and major mechanism for membrane transport.

That is mediated by the transport proteins.

By mentioning membrane transport proteins, I actually refer to two major classes, as shown here.

Transporters, at the top, and channels, at the bottom.

Okay, from these two simplified cartoons, what differences can you tell? Think about it...

now, I will walk you through to see the differences and the common features of channels and transporters.

So, for channels, this animation is my favorite one, because it reveals two fundamental features of a channel.

First, the substrate selectivity...

shown here are two ion channels, the sodium channel and the potassium channel.

If it is a highly selective channel, it only allows the penetration of its target substrate -- a sodium channel can usually not allow the permeation of potassium ions, and vice versa.

The second feature, look at... here...

is the gating mechanism.

So, a channel is a pore within the membrane, but life can never be out of control.

That's why this pore has to be gated.

Only under certain circumstances can it open, and this opening of the channel allows the exchange, allows the translocation of a chemical that brings a new signal.

It responds to certain stimuli and converts these stimuli to another sort of signal.

Okay, for a channel, the fundamental feature of a channel that discriminates the channel from a transporter is once the channel is open, once one gate or multiple gates are open, this transport path opens to both sides of the membrane simultaneously.

By doing this, it actually lost control of its substrate.

Therefore, the substrate can only move down its electrochemical gradient.

It cannot go the opposite way.

And I have to remind you that the permeation rate of a channel can be fast or slow.

Therefore, the permeation rate cannot be used to distinguish a channel or a transporter.

Okay, how to classify a channel? As I told you, here, we can do so either by the substrate selectivity...

for example, there are selective transporters versus nonselective.

There's nonselective cation channels that can actually permeate different species of ions.

And, more interestingly, we mentioned a lot of ion channels, but as long as this transport protein can allow the penetration of the substrate along its translocation path, and when it does so it opens to both sides simultaneously, it's called a channel.

So, there are not only ion channels, there are also the water channels, glycerol channels, and even protein channels.

Over here, for water channels, remember that these small chemicals, they can actually penetrate the membrane freely, so the discovery of the presence of water channels, also known as aquaporins, was actually very amazing.

Therefore, a Nobel Prize was awarded to Peter Agre for his discovery of the first water channel.

And for the cation channels, they are more famous or better studied than anion channels.

Shown here are the representative cations, like sodium, potassium cations, magnesium...

sorry, calcium, magnesium, etc.

And only recently did one channel for fluoride...

was one channel for fluoride discovered by Chris Miller.

So, probably, there are more channels to be discovered for different types of ions.

Okay, now let's talk about gating mechanisms.

So, a porin, by definition, is a pore within a membrane.

It seems there's no gating mechanism, but that may not be true.

Even for the water channel, aquaporins, gates have been identified for some types of the water channels.

So, whether there's any channel that has absolutely no control or there are actually gating mechanisms that await to be characterized, this is an interesting question.

And for the known gating mechanism, based on different signals, we can divide them into the ligand gated channels -- basically, they respond to the change in chemicals like the ATP-gated channels, calcium-gated channels, etc.

And another major type of channels is called the voltage-gated channels.

As the name indicates, they respond to the change of membrane potential to open or close the channel, so this channel plays an important role in neuronal signal transmission and muscle contraction, which I'll come back to later.

And, light-gated channels...

have you heard of optogenetics, and you must be aware that the primary tool for optogenetics is the channel rhodopsin, which is a channel gated by light.

And recently more channels have been identified responding to different types of stimuli.

For example, we can all sense the temperature change -- cold, warm, or hot -- this is owing to the presence of temperature-sensing channels, exemplified by the Trp channels.

And there is also another type of channel known as mechanosensing channels.

They can sense the difference between osmolarity, or the change of the membrane surface tension, or, you know, we can tell the difference between pet or punch, so, probably, this sensing is all attributed to the mechanosensing channels.

Unfortunately, we know very little about them.

Only recently were some channels identified, such as the Piezo channels.

So, this is really an open field, whether there are other means of gating mechanisms.

Think about it...

so, life's organisms have evolved to respond to all kinds of signals, stimuli, stress in nature...

probably, whatever signal you can sense, there are corresponding biomolecules, probably, most likely, channels there to sense these stimuli.

And by opening the channel, allowing the penetration of certain chemicals, mostly ions, they convert this kind of stimuli to the intracellular signal for the cell to process, or the downstream cells, neurons, to further process.

So, channels are in a way the information converter, an information transducer.

Okay, so we can classify channels based on their substrate selectivity or gating mechanism.

We can also classify them based on their cellular localization, either plasma membrane channels or intracellular channels.

Okay, now let's talk about transporters.

Shown here are two general mechanisms of the transporter mechanism of transporters.

So, one is the alternating access, the other is the so-called elevator mechanism.

So, why do I present them here? Because, in history, transporters were originally proposed to translocate the substrate through a so-called solute carrier mechanism.

In the 1950s, Widdas proposed this solute carrier mechanism, according to which a transporter may function like a carrier, or a little boat, if you think about it.

It uploads the substrate from one side of the membrane and then it swings across the membrane to release the substrate on the other side.

That's why it's called a solute carrier.

However, if you think about it, if the substrates are hydrophilic, then most probably the surface of this carrier is also hydrophilic...

then, how can this hydrophilic carrier go through this hydrophobic lipid bilayer? There is an energy barrier, therefore, this model was gradually abandoned and replaced by another prevailing model called alternating access.

However, in recent years, structural studies identified this kind of elevator mechanism.

You can see, there's a scaffold that remains almost static and then the substrate binding domain undergoes this translocation across the membrane.

In a way, this substrate binding domain functions like the solute carrier.

But this model is now known as the elevator mechanism, but it's reminiscent of the solute carrier mechanism.

And then, what about alternating access? This predicts that this transporter harbors the substrate binding site within its interior, so it opens to one side of the membrane to expose the substrate binding site, to upload the substrate, and then it undergoes conformational change to expose the substrate to the other side of the membrane to finish the release.

This is a transport cycle.

So, because of the alternative exposure of the substrate to either side of the membrane, it's called alternating access.

And if you think about it, even for this elevator mechanism, the bound substrate actually is accessible from either side of the membrane, so the elevator mechanism, in a sense, is a specific type of alternating access.

Therefore, the alternating access is a well-accepted, prevailing mechanism for all of the transporters, as of today.

Shown here, remember, in this cartoon, is the alternating access.

So, the fundamental feature is it has at least two gates, one of the extracellular side, the other to the intracellular side.

And the bound substrate can never be exposed to both sides of the membrane simultaneously.

So, at any time, at least one gate is closed.

And by doing so, actually, it creates a coupling mechanism that allows the transporter to catalyze the uphill movement of a specific substrate across the membrane.

But by doing so, it has to harness another form of energy to compensate this electrochemical potential change.

So, this process is called a coupling mechanism.

In a way, a transporter is a miniature machinery that can complete this energy conversion through this coupling mechanism.

So, depending on the different types of energy used to pump the substrate...

to pump, the active transport, transporters can be divided into primary active transporters or secondary active transporters.

So, if the energy source is light or energy released from chemical reactions such as ATP hydrolysis, it's regarded as the primary active transporters.

In this slide, I listed three types of representative primary active transporters.

On top is the very famous sodium-potassium pump.

Basically, they harness the energy released from ATP hydrolysis to catalyze the exchange of potassium and sodium against their concentration gradients -- it is a change against both of their concentration gradients at fixed stoichiometry.

So, this sodium-potassium pump is extremely important for maintaining the asymmetric distribution of sodium and potassium, hence creating the membrane potential, which is important for the neural signaling or muscle contraction.

And, at the bottom, ABC transporters, known as the ATP-binding cassette transporters.

Basically, they have this ATP-binding domain, shown at the bottom...

so, the ATP binding, ATP hydrolysis, or dissociation of ADP and phosphate all may cause conformational change of this transporter, of the transmembrane region, to complete the alternating access, that's why they are called ABC transporters.

And shown, here, are...

do they look familiar? Yes.

They are the complexes involved in the electron transport chain in mitochondria or E. coli... or bacteria, but in essence complexes 1, 3, and 4, they are actually the proton pumps.

So, they harness the energy released from the electron transfer, or the redox reactions, to pump protons from the matrix to the intermembrane space of mitochondria.

And the maintenance of this proton gradient across the inner membrane of mitochondria is extremely important, because this is the direct energy to drive the synthesis of ATP.

So this transmembrane proton gradient is called the proton motive force.

So, remember, these complexes, in nature, they are primary active transporters.

Alright.

So, if the energy is not from light, not directly from chemical reactions, some transporters, they can exploit another form of electrochemical potential, such as the one I told you, the proton transmembrane gradient, or the gradient of another ion or other chemicals, these transporters are classified as the secondary active transporters.

So, basically, they convert one type of electrochemical potential to another type.

They exploit this energy to drive, again, the uphill translocation of specific substrates, like sugars or nucleosides or amino acids.

There are many different such secondary active transporters.

And there is also a third type of transporter known as a facilitator.

So, basically, they catalyze this diffusion of the substrate down its concentration gradient, like glucose, it itself already has a transmembrane gradient, but it cannot penetrate through the membrane, so it requires this type of transporter.

It's known as a facilitator.

But what's the difference between a facilitator and a channel? Remember, the answer is in the gate.

So, for a channel, once the gates are open, the channel, the transporter path opens to both sides of the membrane simultaneously.

For a transporter, even if it's not active transport, even if it facilitates diffusion, it has two gates.

At any time, the substrate is isolated from one side of the membrane, so the transporter always undergoes alternating access.

Alright.

For the transporters, we can also classify them through another mechanism, that is, the transport orientations of the substrates.

So, for a facilitator, by catalyzing the diffusion of one type of substrate, it's called uniporter, but for active transporters, especially for the secondary active transporters, by definition they must transport at least two substrates.

So, depending on the orientations of these two or even more substrates, they can be called either symporter, if the two substrates go along the same direction, or antiporters, if the two substrates go oppositely.

So, for transporters, there are uniporters, symporters, and antiporters.

Okay, to summarize what I have told you so far...

there are different mechanisms of membrane transport, there are different transporters to mediate the active transport -- primary active transporters and secondary active transporters -- and also there are membrane proteins to facilitate diffusion, including channels and facilitators, but in addition, remember, we also see the simple diffusion for certain chemicals and the vesicular transport for a large number of substrates with low selectivity.

In terms of the physiological, pathophysiological significance, I cannot tell you how important the membrane transport proteins are, just think about it...

almost in each and every process, there are membrane transport proteins, from photosynthesis, respiration, nutrient uptake, metabolite extrusion or drug resistance...

and remember, each chemical itself is a signaling molecule.

And the translocation of this chemical from one side to the other of the membrane, itself carries a signal, carries information.

Therefore, the transport proteins are involved in signaling, as I just told you, especially for channels.

And for energy conversion, we already learned the secondary and primary active transporters, in a way, they are the machineries that convert energy to different forms.

And because of this fundamental significance, malfunction or misregulation of the transport proteins are always involved, or even the direct cause, of many debilitating diseases.

Therefore, they are also important drug targets.

Now, let's take a break to watch a little movie generated in my lab, to try to understand this very simple but fundamental process of glucose uptake.

So, when the starch is digested to glucose, they have...

these glucose molecules, which are highly hydrophilic, have to pass the barrier of the lipid bilayer, and this process is mediated by glucose transporters.

Shown here is GLUT1, whose structure was determined by us a couple of years ago, and GLUT1 undergoes typical alternating access to complete the translocation of glucose from outside of the cell to the inside of the cell.

This is fundamental and it's mediated by a transporter.

And shown here is another example of the physiological function of channels.

So, from the lectures of Ron Vale and others, you have learned the cytoskeleton and the motor proteins, which are important for muscle contraction, but remember, prior to this process, there is a communication between the motor neuron and muscle cells, and this communication is mediated by different types of transporters and channels.

Briefly, when the signal arrives at the muscle cells, first, the sodium channels, here, they are voltage-gated sodium channels that are activated.

Voltage-gated sodium channels are responsible for the initiation and propagation of the action potential.

And when the action potential is relayed to a specified membrane structure known as a T-tubule, or transverse-tubule, where the voltage-gated calcium channel resides, those channels are activated and undergo a conformational change that is further linked to the downstream RyR membrane receptor, which is a calcium channel.

So, activation of RyR allows rapid release of calcium ions from the sarcoplasmic reticulum to the cytoplasm, where the calciums activate the proteins required for muscle contraction.

The function and mechanism, and the structure of these channels, have represented important targets for structural biology, and also, because of their important role and because they reside at the interface between the cell and the environment, membrane transport proteins represent major drug targets.

Shown here is a statistic made about a decade ago.

At that time, about one quarter of FDA-approved drugs target ligand-gated ion channels, voltage-gated ion channels, and the transporters.

And, now, more and more drugs are being developed, targeting these channels and transporters.

Because structural details are important for the development and optimization of the lead compounds, structures of membrane proteins have been pursued by many labs throughout the years.

However, due to the technical difficulties, structural biology of membrane proteins, particularly of membrane transport proteins, have been much slower than the other macro-biomolecules.

For example, shown here is a brief history of structural biology of proteins.

The first structures of hemoglobin and myoglobin were determined in the late 1950's by John Kendrew and Max Perutz, and then three...

almost three decades later, the first structure of a membrane protein was determined by Hartmut Michel, Johann Deisenhofer, and Bob Huber.

So, that structure was of the photosynthetic reaction center from a bacteria.

That's in 1985.

And then the first structure of a transport protein, namely the potassium channel, appeared more than a decade later.

What about the transporters? Actually, the first structures of transporters appeared in 2002, entering the 21st century.

Why is that? Because these transport proteins, they are not abundant in nature.

Most of the time, you have to overexpress them through recombinant expression and they are highly mobile, which made them even more difficult for structural elucidation.

And in the past, the major approach for structural determination of membrane proteins was X-ray crystallography.

As shown here, in order to crystallize your target protein, you have to have a large amount of purified proteins.

For membrane proteins, they are expressed in lipid bilayers, so you have to extract them out using detergents, and you have to select the most optimal detergent to protect your protein, to make them happy.

And not too happy, so that they can still pack together to form a crystal.

Therefore, expression, extraction, and crystallization...

each step represents a major challenge for structure determination of membrane proteins.

Fortunately, in recent years we have realized, basically, a revolution in structural biology, due to the technical advances of electron micrography.

In recent years, more and more structures of membrane proteins have been elucidated using cryo-EM, single-particle cryo-EM.

The technical advances are mainly attributed to, first, the development of direct electron detection and, second, the development of algorithms for data collection from imaging to classification of the images and also for structural reconstruction.

So, these technical advances, together, led to the rapid boom in the cryo-EM structures of membrane proteins.

So, if you are interested in this technique, please refer to...

I would like to recommend a three minute, brief introduction of transmission electron microscopy prepared by Gab Lander several years ago.

So, if you are interested in the structures of membrane proteins in general, you know, their identities, their classifications, the approach of structural determination, you can visit the website of membrane proteins of known structures maintained by Steve White at UC Irvine.

And, above all, it seems I have told you quite a lot about transporters and channels.

However, we have to admit that we know very little about membrane transport proteins.

So, it's estimated that approximately 10% of the human genome encodes for transport-related functions.

That means they encode for transporters and channels.

Yet, for most of them, the substrates and the physiological functions remain unknown.

The analogy to GPCRs (G protein-coupled receptors) whose ligands are unknown, are called orphan GPCRs.

We can see there are a lot of orphan transporters and channels there.

We need to de-orphanize these transporter s and channels.

Therefore, a group of scientists recently published a preview...

a Perspective in Cell calling for systematic research on solute carriers.

Not only solute carriers, remember, there are also channels whose functions remain unknown.

This is a very exciting field. Hi.

I'm Nieng Yan.

I'm a professor in the School of Medicine, Tsinghua University, Beijing, China.

Welcome to iBiology seminar series.

In part 2, I'd like to share with you one major research interest in my lab.

That is, the structure elucidation of one very fundamental physiological process , the cellular uptake of glucose.

We all know glucose is the primary energy source to most of the lives on Earth.

From the textbook of biochemistry or cellular biology, you all learned how glucose is burned to release energy to support life.

We know through glycolysis, one glucose molecule is split to two pyruvate molecules, and during this process two ATP molecules are generated.

And in the aerobic conditions, the pyruvate molecules are further burned through the TCA cycle, or the citric acid cycle, and the electron transport chain, to generate carbon dioxide.

And during this process...

I mean, if it's complete metabolism, then one glucose can be used to produce over 30 ATP molecules.

That is the energy currency for all life.

However, before the metabolism of glucose, there is also one critical step -- that is to take the glucose into the cell.

From part 1, I already told you that glucose is highly hydrophilic, that means, they are water soluble.

However, the cell is surrounded by the hydrophobic lipid bilayer.

So, glucose cannot enter the cell through free diffusion.

There must be different proteins to mediate this process.

These proteins are called glucose transporters.

So, as we see here, glucose transporter is important, is essential for cellular uptake of glucose.

And, throughout the years, we have identified different types of glucose transporters, and more glucose transporters are being identified, but among all of those, the most rigorously characterized ones are called GLUTs, as shown here, G-L-U-T, glucose transporters.

So, in human bodies, there is a huge family called major facilitator superfamily, and the GLUTs belong to this family.

Even within the GLUT family, there are 14 different isoforms that exhibit tissue specificity and substrate specificity.

As summarized here, for example, GLUT1 functions in brain and red blood cells, and GLUT2 is for liver.

GLUT3 is also called neuronal glucose transporter, indicating that it functions in neurons.

And GLUT4 is very famous -- it take glucose into adipocytes and muscle cells.

So, these are the four most famous GLUTs -- GLUT 1, 2, 3, 4.

And for the other 10 different isoforms, unfortunately, for some of them, their substrates remain uncharacterized.

Oh, besides, for these glucose transporters, despite their sequence similarity with each other, they actually may have different binding affinities for glucose and for other similar sugars, and they have different turnover rates.

For example, GLUT1 can take up to 1200 glucose per second, but that's... yeah, that's very fast...

however, GLUT3...

for GLUT3, the number is 6000...

it's five times faster than GLUT1, and this is amazing.

Because of their fundamental significance in physiology, you can imagine, malfunction or misregulation of these proteins are associated with various diseases.

For example, GLUT1 deficiency syndrome is actually a rare genetic disease manifested by early onset seizure or retarded development.

And GLUT2, because it's associated with the liver...

so, mutations of GLUT2 are associated with a type of disease called Fanconi-Bickel syndrome.

And more and more evidence shows that GLUT1 and GLUT3 are overexpressed in cancer cells, especially solid tumor cells, because of the so-called Warburg effect.

I just told you, without oxygen, one glucose can be converted to pyruvate.

During this process, two ATP molecules are generated.

However, in the presence of oxygen, that is, aromatic...

or, sorry, aerobic conditions, about... I mean, over 30 ATP molecules can be generated.

For solid tumors, it's usually under hypoxic conditions.

That's... you know, that means it can only...

one glucose can only generate two ATPs.

Consequently, more glucose transporters have to be expressed to take more sugar to compensate for this amounts.

And for GLUT4, it's very famous because of its association with type 2 diabetes mellitus and obesity.

So, as I just mentioned, glucose transporters belong to the so-called major facilitator superfamily.

As a matter of fact, they are the prototypes of this largest secondary active transporter family.

And for members in this family, actually, they are widespread across species, from bacteria to human beings.

And members in this family have a very broad spectrum of substrates, from ions, sugars, amino acids, or even peptides.

And in terms of transport mechanisms, if you watched part 1 already, actually, members in this family can be uniporters, symporters, or antiporters.

That's in terms of the orientations of the transport.

And, as I also told you, a general alternating access model or mechanism has been proposed to account for all the secondary transporters.

Especially for MFS members, this works very well.

And we thought that because GLUTs are the prototypes in the understanding of this family, so structural and biochemical characterization of GLUTs may also shed light on the understanding of other members of this largest family.

Okay, why is it a prototype, especially GLUT1? Because it was one of the first transporters to be cloned and characterized.

So, let me bring you to the history.

Actually, the characterization of glucose uptake into our blood cells can be dated back to about a century ago.

And at that time it was already discovered that the uptake rate or the "diffusion"...

at that time people didn't know it's active transport, so they still called it diffusion...

but one PhD student found that the diffusion coefficient is actually concentration dependent, suggesting it was not free diffusion.

In 1948, LeFevre, in one paper, speculated on the active transport component, although he didn't specify whether it was protein or something else, but he just speculated there would be an active transport mechanism.

And in the 1950s, Widdas, in his very famous paper, proposed a so-called mobile solute carrier mechanism.

As a matter of fact, this mechanism was so famous that all the secondary transporters in humans are named after SLC.

So, for example, GLUT1 is actually...

the gene name for GLUT1 is SLC2A1, but don't call it slack, because scientists don't like that name, so it's SLC.

And then...

and so far it's all about these components.

And in 1977, these scientists actually were able to purify the protein component from red blood cells and reconstitute them into a liposome, and they reconstituted the uptake of glucose.

So, they named this protein component GLUT1.

And then, in 1985, Harvey Lodish's lab cloned GLUT1, and when the sequence was available it was clear that this protein contains 12 transmembrane helices.

And in the 1990s, the study efforts were shifted to the pathophysiological investigations, as well as structural characterizations, because we would like to understand their structure, to see their structure, so as to understand its functional mechanism and disease mechanism.

However, 30 years...

almost 30 years passed...

so what we learned from the textbook about the structure of GLUT1 was still this, the one published by Harvey Lodish in 1985.

This is the topological structure.

Alright, umm...

so, I started my lab in 2007 and we were very interested in the structure of GLUTs because we thought it could help address a lot of interesting questions, as listed here.

Of course, the first thing is you try to see the architecture of GLUTs, that's the most direct, but superficial, purpose.

And with the structure we might be able to reveal the molecular basis underlying the substrate selectivity, why it selects glucose, but not, for example, maltose.

And we...

because we understand that these transporters follow this alternating access cycle, so we'd like to reveal the conformational changes during the transport cycle, to understand their functional mechanism.

And we also hope to provide a molecular interpretation for all these disease-related mutations.

And, for my own research, I'm also very interested in the difference, the mechanistic difference, between symporters, particularly proton symporters, and facilitators, but I may not have time to go into the details of this part.

And finally, because membrane proteins are embedded in the lipid bilayer, we would really like to understand how they are modulated by lipids, and there are more and more questions...

they just emerge during your research.

So, to address these questions, we started not with a glucose transporter, but with their relatives, their relatives from E. coli, which are technically easier than the human protein.

So we determined the two structures of E. coli proton-sugar symporters, FucP and XylE.

So, as the name indicated, they are proton symporters, meaning they exploit this transmembrane proton gradient to drive the uptake of the substrates, either L-fucose or D-xylose, from a low concentration environment to the high concentration interior of the cell.

In the past three years, we were very lucky.

We were finally able to determine the crystal structures of GLUT1, and its closely related GLUT3, in three different conformations.

That means they adopt different states during a transport cycle, as shown here.

So, all the way from outward-open, occluded, and inward-open.

When I say outward or inward, that refers to the substrate binding site, that is... remember, for the alternating access, that is... the substrate binding site can never be exposed to both sides of the membrane, so it's always open to one side, the substrate comes, and this protein undergoes conformational change to expose the substrate to the other side.

This is called alternating access.

So, with these three structures, we have a relatively better understanding of this transport cycle of GLUTs.

Alright, first thing.

To address the question of the architecture...

but, before that, I know many people are interested in the crystallization of membrane proteins and GLUT1 has been a target for several decades.

Why were we able to crystallize and determine the structure of this very intriguing protein? In retrospect, there are three key elements that contributed to the crystallization of GLUT1 and gave us the diffracting crystals.

First, we actually introduced point mutations...

first is to eliminate glycosylation, which really represents major troubles for crystallization.

And the other point mutation, glutamate-329 to glutamine, this one is a disease-related mutation originally identified in GLUT4, and it was suggested to l ock the protein in the inward-open conformation, which was exactly the case, as seen in our structure.

And second, on the detergent we used for crystallization is nonyl-glucoside.

I will come back later with why this was important.

And third, you know, for glucose transporters, they are highly mobile, so we would try to slow them down, to lock them at certain conformations, so we did all the experiments at low temperature, at 4 degrees Celsius -- that helped a lot.

And to cut a long story short, one particular day my student showed me these crystals, these tiny crystals.

I thought, probably, they were contaminations from insect cells, however, you know, it doesn't hurt to send them to the synchrotron for data collection and we sent this single crystal to the Shanghai synchrotron, and several hours later we solved the structure that was exactly our target, GLUT1.

As shown here, this structure exhibits a very typical MFS fold, remember, major facilitator superfamily.

It contains 12 transmembrane helices, with the first 6 named the N-domain or the N-terminal domains, shown in silver, and the C-terminal one in blue.

And very unexpectedly, we also see an intracellular helical domain, we named it ICH.

Actually, this domain...

this little domain harbors a lot of serine or threonine or lysine, so these residues are probably important for their post-translation modification.

Now, with this structure, we really can provide the answer to many questions.

So, as I asked at the beginning -- so, what is the mechanism of substrate selectivity? For this purpose, we actually examined, through a biochemical approach, the sugar selectivity by GLUT1 and GLUT3, shown here are the results for GLUT3.

As you can see, indeed, this protein has kind of a stringent selectivity, and one...

wherever you see these lower, these shorter bars, that means these sugars can inhibit the uptake of glucose, meaning that they can be recognized by GLUT3, to compete for glucose binding.

And when we examined these chemicals, very interestingly we found one common feature, that is, their C3 hydroxyl group all points to one orientation, so that means that C3 hydroxyl group is important.

That's the conclusion from biochemistry, from our biochemical characterizations.

Then, how is one sugar molecule recognized by the protein? So, the answer is from the very high resolution structure of GLUT3.

So, we determined GLUT3 in complex with its substrate, D-glucose, at 1.5 Angstrom resolution, and shown here is the omit electron density map.

As you can see, it's beautiful.

To our surprise, we identified...

although we just, you know, add glucose to the protein and we identified two different anomeric forms of glucose, simply by the electron density map.

As you can see, both alpha and beta glucose are present in the structure...

I mean, I have to clarify...

so, one protein can only bind to one glucose, but for crystallography, you know, this is the average of many billions of molecules, so you know some proteins bind to the alpha form, some bind to the beta form.

And this observation, this structural observation actually settled down one long-term controversy, that is, whether glucose transporters can recognize the alpha form of glucose, because we know the beta form is the prevailing one, the dominant form in solution.

And this observation shows, yes, GLUT1 or GLUT3, they can bind and transport both anomeric forms of D-glucose.

Alright.

Another interesting discovery is that, as I told you, glucose transporters have two distinct domains, N-domain and C-domain.

However, in the structure of GLUT3 in complex with glucose, as well as in the structure of GLUT1 in the presence of this detergent molecule NG...

what is NG? It is actually a derivative of glucose, so that's why NG is important for us to capture the structure of GLUT1 -- it mimics the substrate binding.

And if you compare these two structures, a common feature is the C-domain provides the primary accommodation site for glucose, so the C-domain is the primary substrate binding site.

Then, what does the N-domain do? Alright.

So, before that, you know, we tried to complete the alternating access cycle by, you know...

in the attempt to capturing another conformation, that is, the outward-open, because now we have GLUT3 in complex with glucose in the occluded conformation, that is, the substrate is trapped in the center of the transporter and isolated from either side of the membrane.

And GLUT1 is open to the inside of the cell, so it's called inward-open.

Now, we still need this outward-open conformation.

In order to capture the outward-open structure, we really had some rational thinking.

So, people always say that crystallization is an art, it seems like you have to do a lot of screening, but in this case we really did some rational thinking.

That is, when we obtained the structure of GLUT1, I told you NG is important, right? So we introduced several factors, like the mutation E329Q, that is, to lock the inward-open conformation, and then when we see the binding of NG to the protein, as you can see on the tail, the aliphatic tail of this detergent, it actually is lining down the intracellular vestibule, when the sugar moeity is specifically coordinated by the C-terminal domain.

So, along...

so, basically, the presence of this aliphatic tail precludes the closure of these two domains on the intracellular side, that is, to stabilize this inward-open conformation -- with this aliphatic tail, it cannot close, right? So, along this line of thinking, we thought, if we can find a chemical, a glucose derivative, that has some chemical groups on the other side, on the upper side of the sugar ring, probably that can preclude the closure of the protein on the extracellular side, that is, to capture an outward-open conformation.

Do we have these kind of chemicals? Yes, we have a lot of disaccharides that are derivatives of glucose.

As shown here, we selected a few and we examined their ability to inhibit glucose uptake.

As shown here, it turns out that maltose was a potent inhibitor, and when we checked the literature this was really consistent, because maltose was regarded...

was suggested as the exofacial inhibitor, that means it can inhibit glucose uptake from the extracellular side.

To cut a long story short, in the presence of maltose we actually crystallized the protein using lipidic cubic phase.

It gave us two different structures.

One is almost identical to...

shown on the left, it's almost identical to the glucose-bound GLUT3, and is occluded from...

so, maltose is bound in the center, occluded from either side of the membrane.

But the other crystal form gives us this outward-open conformation, so this was really serendipity, I mean, we just mixed them together and it gave us two different crystal structures.

So, I will focus on the illustration of this outward-open conformation, with comparison of the inward-open GLUT1 and the occluded GLUT3.

So, now we have these three conformations I showed before.

We could generate a morph that illustrates the whole transport process.

As you see here, outward-open, arrival of glucose, and it undergoes this alternating access by the relative rotation of these two domains, and the substrate is released into the inside of the cell.

And, very interestingly, remember this small domain, shown in yellow, is the ICH, intracellular helical domain, and during the conformation change, we can see it also undergoes interdomain rearrangement.

In a way, it restrains the N- and the C-domains from opening too much, so this ICH domain, we named it the latch, to secure this intracellular gate.

Alright.

From the movie you might think, hmm...

these two domains undergo a rigid body rotation, but close examination of the structures of the outward-open and occluded GLUT3 suggest, no, it's not rigid body.

Actually, we can see very sophisticated local rearrangement of the C-domain elements.

As shown here, the one shown in cyan is the C-domain and the one in green is the N-domain.

Please pay attention to this TM7 motif.

You can see it undergoes a bending, a bending.

Right? This is TM7.

Not only a bending...

so, when the sidechains are shown, you will see it actually also undergoes a rotation, so this TM7 undergoes very complicated local rearrangement by bending...

the combination of bending and rotation.

So, whether this is induced by substrate binding or this is the so-called dynamic equilibrium, remains to be further characterized, and our preliminary MD simulations suggest that this is dynamic equilibrium.

even in the absence of substrate, you can see this kind of conformational change of TM7.

Now, here's the question...

why the C-domain, shown in cyan, is so flexible, whereas the N-domain is just so rigid, as a stone? And when we examine the interior of these two domains, the answer is really clear.

So, as shown here, the red dashes represent hydrogen bonds.

As you can see, the interior of the N-domain is really hydrophilic, so the high-resolution structure of GLUT3 allowed us to identify seven water molecules within the N-domain of GLUT3, and these water molecules, together, interact with a set of groups of many polar residues as a strip of hydrogen bonds, and this stabilizes the N-domain, so it makes it very rigid during conformation change.

In contrast, the interior of the C-domain is highly hydrophobic, as shown here, so these hydrophobic residues, they just contact each other through Van der Waals interactions, so they make the interior relatively greasy, and that's easier for bending and rotation.

So, the structural analysis really provides a good answer to account for the distinct features of the N-domain and the C-domain during the alternating access cycle.

Alright.

Now, I'll...

shown here is the very simple diagram of alternating access.

With our structures, the three structures, we are able to update this model with more sophisticated features.

As you can see, TM7 and also TM10, they undergo local conformational change, and the overall relative rotation of the N- and the C-domains results in the alternating exposure of the substrate to either side of the membrane.

And, besides, please pay attention to these yellow bars, they are the intracellular ICH domain, we call them the latch, the intracellular latch.

Okay, now with the structure.

We were able to map the disease-related mutations.

Shown her is an example of the mutations identified in patients with the so-called GLUT-1 deficiency syndrome.

So, in total, more than 40 mutations were identified.

So, when we mapped them onto the structure of GLUT1, very interestingly we realized that they clustered to three areas, as shown here.

So, area 1 is really involved in substrate binding and it's easy to understand how mutations of this cluster would affect substrate recognition or substrate binding, hence compromising the transport activity.

And cluster 2, as shown here, highlighted by this cyan circle...

the cyan circle...

so, basically they mapped to the interface between ICH, N-domain, and C-domain, and they together constitute the intracellular gate.

And, not surprisingly, cluster 3 maps to the extracellular gate.

So, the structure really provides a beautiful answer to understand most of these disease-associated mutations, so they either affect substrate binding or the two gates, hence affecting the alternating access cycle of the protein.

Alright.

Now, with these results, we can address the questions asked at the very beginning, right? So, we know the architecture and we provide a basis to see the substrate selection and we revealed three conformations of GLUT1 and GLUT3 during the transport cycle, the alternating access cycle, and we provided some answers to the disease-related mutations.

And, with regard to the mechanistic difference between symporters and facilitators, we are now doing some MD simulations and biochemical characterizations, and we have some tentative clues, but this really requires further characterizations.

And now our focus has shifted to the modulation of the transport activity by lipids, as well as, you know, the kinetic study of transport cycles.

And finally, we are very interested in structure-based ligand design, because these proteins are important drug targets.

So, with this, I would like to conclude my talk by acknowledging the people who made this work possible.

So, Dong, he was my postdoc who has been the primary driver of this project, he was leading this team of Tsinghua undergraduate students and graduate students to elucidate the structures of both GLUT1 and GLUT3, and he's now a professor in Tsinghua University.

And this work was in collaboration with many colleagues, in Tsinghua or in the US, as shown here.

And I'd like to thank you for watching this online seminar.