Some kind of better oxygen-transport protein or something… Can you help me find it?

Some kind of better oxygen-transport protein or something… Can you help me find it?

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

I remember reading about artificial or man-made, oxygen-transport protein that is somehow an improvement on hemoglobin, and that it is possibly immune to sickle-cell anemia, or something… But I forgot where I found it and what exactly it is…

Can you guys help me find it?

Lecture 19: Cell Trafficking and Protein Localization

Professor Imperiali talks about trafficking, or how things get to where they need to be within a cell. She will discuss the mechanisms by which proteins are coded very early on in their biogenesis, in order to go to certain locales in or out of the cell.

Instructor: Barbara Imperiali

Lecture 1: Welcome Introdu.

Lecture 2: Chemical Bonding.

Lecture 3: Structures of Am.

Lecture 4: Enzymes and Meta.

Lecture 5: Carbohydrates an.

Lecture 9: Chromatin Remode.

Lecture 11:Cells, The Simpl.

Lecture 16: Recombinant DNA.

Lecture 17: Genomes and DNA.

Lecture 18: SNPs and Human .

Lecture 19: Cell Traffickin.

Lecture 20: Cell Signaling .

Lecture 21: Cell Signaling .

Lecture 22: Neurons, Action.

Lecture 23: Cell Cycle and .

Lecture 24: Stem Cells, Apo.

Lecture 27: Visualizing Lif.

Lecture 28: Visualizing Lif.

Lecture 29: Cell Imaging Te.

Lecture 32: Infectious Dise.

Lecture 33: Bacteria and An.

Lecture 34: Viruses and Ant.

Lecture 35: Reproductive Cl.

BARBARA IMPERIALI: I always like to just remind you that the sixth -- it's kind of an assignment, but the numbers-- we're going to do this news brief project, where it's a teamwork project if you choose. If you take a look at the piece that you have in your hands now, it asks you for a little bit of information on that, who you're going to be working with, if you choose to work with someone. Or you can work on your own. That's fine.

And we're looking to get a news brief that's of significance to research going on in the life sciences. And I've given you-- there are a couple of links in the sidebar of the website, so good places where you can find interesting material. What I'm super interested in for you, as a group where many of you are in the engineering fields, is to find something really cool at the interface between the life sciences and engineering, where engineering has a huge impact on the life sciences.

You have alternatives. You can download the coordinates of a protein and print it on a 3D printer and give us a summary of what the protein is, what it does, and submit your 3D print. I'll give it back to you afterwards, once we've had a look at it. But actually submit the 3D print.

And then the other opportunity is-- I think you'll remember back to when we were talking about molecular biology of the cell. I did kind of a clunky demo at the front of the class, nothing like Professor Martin's demos at all. This was me with the ethernet cables showing you what topoisomerase did. But in my demo, I didn't show you how topo also cuts a strand of DNA, holds it while the supercoiling unwinds, and then stitches it together. So I thought some of the engineers might be able to come up was something that was really better than that for me to use, for us to use, next year in class. So I'm really laying down the challenge there.

So I always like things in the news. I thought this was kind of interesting that the first vertebrates evolved in shallow waters. I thought those were really cool first vertebrates. I'd love to get one of them in a fish tank and keep it. But anyway, that's that.

It's truly amazing what you can see in the science reports, news briefs. I look at them whenever they come in. I get the posts every two or three days. And I'm kind of pleased to see that there's a lot of things that are in those news briefs that I feel that we're enabling you to read with some appreciation because of what we're covering in the class.

So what we're doing now is we're really taking a leap forward here into cells and organisms, with respect to understanding how structure and function of individual macromolecules, proteins, nucleic acids, sugars, determine life, determine the dynamics of life that are necessary for an organism to really go through a life cycle, divide, have cells divide, go forward, have cells move. So what we're going to be talking about in the next lectures, which is section 6, is cellular trafficking and signaling. And so for the first lecture, which is 19 that we're on now-- so we're past the midway mark-- I'm going to be talking about trafficking. And that is how, within a cell, things get to where they need to be, or they get exported from a cell.

Because all of the actions of a cell-- I really like thinking about the cell as a circuit board, where there's a receiver that gets information. And then the complex circuitry determines what outcome you get at the end of the day. So many of the proteins that we've talked about need to be in specific places for the cell to function.

We have to have DNA polymerase in the nucleus. It's not going to be useful in the cytoplasm. We have to have a transcription factor that helps transcription go to the nucleus at the right time for transcription to occur. But we don't want it there all the time, because otherwise you'd have the light switch on the entire time. That wouldn't be useful. So we need to regulate where certain macromolecules are. We need to have the receivers on the surface of the cell to receive signals from outside.

This is not just pertinent for multicellular organisms. It's pertinent for unicellular organisms, for them to sense their environment, know what's going on around them. Is the salt concentration changing? Is it getting very hot? Is it getting cold? Is there enough oxygen? Even unicellular organisms need to receive signals and respond to them.

Multicellular organisms are way more complicated. Because you need to establish organs and different parts of a multicellular organism that have specialized function. So trafficking really is about what happens after you've made a replicated DNA in the nucleus, transcribed it, made a mature messenger that goes out to the cytoplasm in most cases. We'll talk about the exceptions to that case.

And then in the cytoplasm, when proteins are expressed, all the different things that happen that guarantee that the protein gets to a proper destination for function. And some of those are quite complicated. Because remember, if I'm going to park a receiver in the cellular membrane with the signals being captured from outside, I've got to get from the cytoplasm out there in a reliable way.

In lectures 20 and 21, I'll talk to you about cellular signaling with a focus on mammalian cells and the sorts of signaling processes that may go awry in cells, for example, proliferating cells. And then Professor Martin will really focus in on neuronal cells, optogenetics in lecture 22. So this bundle really allows you to call in the things that you've learned until now and apply them into much more intriguing and complex situations.

So here's a wonderful, sort of silly drawing of a triangular cell. There's always a joke in cell biologists, when they're trying to talk to mathematicians and mathematicians want to simplify everything. And so everything gets-- imagine a cell, and there's this box shows up on a screen. Well, we all know that cells aren't triangular or box-shaped. But nevertheless, I thought this one was particularly cool.

And so trafficking, the process of trafficking, is really all about, where is the information encoded into the protein that ensures that the protein is where it needs to be for the dynamics that we observe in living system? We've talked a lot about static things. We make the protein. Here's the protein. The protein folds.

We've talked a lot about things that are kind of fixed in time and space. But what we want to do is understand what makes a cell programmed to undergo a new function. For example, something as simple as cell division, we have to orchestrate a huge variety of activities in order for the cell division process to start to occur. Something as really simple a cell mobility, think about, how do cells move? They're not moving all the time, but sometimes they will move towards a signal. What triggers that kind of interactions?

So in looking at the cell, these are some of the older images, where certain organelles, for example, are stained so that you can see them. So peroxisomes are where degradation happens. The golgi and the ER are a part of what's known as the endomembrane system. You'll see a lot about this towards the later part of the class, where we talk about how things get outside the cell through the endomembrane system.

There's the surface plasma membrane. The cytoplasm is this sort of not really aqueous-- it's an open space. But it really isn't open. It's highly congested with all kinds of molecules, all kinds of structural proteins and so on. So don't think of the cytoplasm as a solution, but think of it as a much more gel-like structure with a lot of things happening in it.

The nucleus itself is also surrounded by a membrane, as is the endomembrane system. So this would be the nuclear envelope. Within the nucleus, you have a structure called the nucleolus, where aspects of the nucleic acids necessary for protein biosynthesis are made. Then there are structural proteins like microtubules and actin.

But now, in this day and age, we don't have to deal with these vanilla images. We can actually use the methods that you've learned about in the last section, recombinant biology, to create new versions of proteins that have along with their sequence a marker that gives them a fluorescence-colored marker. So we are, later on in the semester, going to spend three lectures on fluorescence and cellular imaging, where you'll learn more about these fabulous proteins beyond just saying we've got a green one and a red one. We're going to give you all the background on the protein engineering that enabled those to become tools for biology.

But for now, I'm just going to show you how much more interesting the images of the subcellular structures are when you've labeled, for example, a particular protein that goes exclusively to the nucleolus with a blue fluorescent protein, or to the mitochondria. Remember, Professor Martin told you we always think of these as-- and I'm not going to do the push-up. I'm just going to say it, powerhouse of the cell. I'm not doing-- [LAUGHS] I'm not great with push-ups, to be honest.

But you see these sort of more tangled, extended structures. Vimentin is more of a structural protein. Here are the golgi, the endoplasmic reticulum, and the nucleus.

So the colored fluorophore proteins, or the fluorescent fluorophore proteins, actually allow us, in real time, to observe dynamics. Once a protein is made, where does it go? If we add a trigger to the cell to cause an interaction, can we observe that protein, for example, migrating to the plasma membrane. Can we watch proteins being made through the ER? A variety of different things that allow us in modern biology to really look at dynamics, not just static information.

And so what I'm going to talk to you about is the ways in which proteins are coded very early on in their genesis, in their biogenesis, in order to go to certain locales within the cell. So let me just give you a bit of a road map here with a protein. And where things may start-- so we have some options. Do we want to send the protein outside the cell or keep it inside the cell?

Obviously, two big default differences, if you're going to go to a particular venue inside the cell. Are we going to just stay in the cytosol? That's a sort of simple-- actually, that is the default position. Because you want to remember that most proteins are made on ribosomes in the cytosol of the cell.

But the statistics are that about 50% of proteins end up somewhere else than the cytoplasm. They may end up in an organelle, back in the nucleus on the surface, or secreted. So there's a lot-- so it's a good, solid 50% that don't end up staying in the cytosol, where they were originally made.

Their alternative is to go to organelles. And if you're going to an organelle, remember, the ribosome is not membrane. It doesn't have a membrane perimeter. But many of the organelles do have membrane perimeters.

So we're talking here about the mitochondria. That is far too long of a word. The nucleus-- so I'm going to abbreviate things like peroxisomes, or various membrane-bordered organelles, where we're going to have to figure out, if something is made in the cytoplasm, how does it get into those organelles?

Now we've spoken a little bit about the fact that some proteins are made in the mitochondria. I'm going to get back to that in a moment. But all the proteins in the mitochondria are not made in the mitochondria. Some of them are shipped in.

Remember the thing the endosymbiont theory, where we said that mitochondria may have originated from bacteria and been engulfed into cells. Those bacteria obviously were originally self-sufficient. But a lot of the proteins that were expressed in the mitochondria were dispensed with, and mitochondria now use proteins that are encoded by the nuclear DNA rather than the mitochondrial. But to this day, some proteins remain encoded within the mitochondria.

So these are opportunities for where that may be. And I'm going to talk very specifically about signals that can get proteins into the mitochondria and into the nucleus. And it turns out that the barriers around those organelles are pretty different. I'll come back to that in a second when we get on the next slide.

With respect to going outside the cell, there are two options. One option is for the protein to remain in the plasma membrane but with part of its structure outside the cell. So the other option is for the protein actually to be spit out of the cell as a soluble entity that can travel around an organism, for example, in the bloodstream and go to a remote site. And that becomes very important in signaling. So we would call those proteins secreted and soluble.

So these would be membrane-bound. These would end up being soluble proteins. Let's take a look at the structure of the cell and look at where these various components are.

So if you see these dots, those are free ribosomes in the cytoplasm. They would start to express different proteins. A lot of proteins are expressed in the ribosome. But in some cases, proteins become expressed on ribosomes that are associated with the endoplasmic reticulum. And therefore, you start a process whereby proteins end up being shipped to the outside of the cell.

So where you see the speckles here, the free ribosome, and then the ribosomes bound to the rough endoplasmic reticulum, here, your destinies are on the right-hand side of that picture. And here, the destiny of these proteins ends up on the left-hand side of this sort of family tree that I'm showing you. There's obviously one more place where proteins are made, and that's in the mitochondria.

And if you remember the first question on your exam, it described the DNA that's in the mitochondria. Going back to the endosymbiont theory, that's a circular piece of DNA. And it sets it apart. And the ribosomes in the mitochondria look more like bacterial ribosomes than you eukaryotic ribosomes. So remember, all along, we're going to try in the second half of the course to bring back knowledge we've taught you, but sort of, in a sense, endlessly remind you to keep the big picture in mind. Because we've already spoken to you about it.

So this now is a nice pictorial vision of what I've just described to you. And I'm going to first of all talk about proteins that are made in the cytoplasm and may be shipped to various organelles, and how that's accomplished. And then in the second part of the class, I'll talk about how proteins are shipped to cell surface, or through expulsion from the cell.

So the key mechanisms whereby proteins are trafficked to new locations are first of all using targeting sequences that are part of the protein sequence. And this is a very common way in which proteins are trafficked. They are part of the sequence. They may be at the amino or the carboxy terminus. But they are woven into the structure of your protein.

So your protein comes along with a barcode saying where it's going to necessarily end up. And for the nucleus mitochondria and peroxisomes, for example, people have done extensive work with bioinformatics to basically look up protein sequences and find common themes of particular sequences that may be common to where a set of proteins may end up. Sometimes those sequences may not be easy to see just at first glance. But now there are websites that you can very, very readily put your protein sequence into the web site, and it will say, it's got a nuclear localization sequence, or a mitochondrial-targeting sequence. So we can either do this by eye or we can use informatics analysis.

Informatics analysis is very valuable because sometimes information may be a bit more encrypted. And it may be a real struggle to slog through a lot of sequences. So you can really find out about the targeting sequences through bioinformatics. Because nowadays, the genomes of dozens and thousands of organisms are available readily online. And you can literally parse out information from the genomic information that gives you the proteomic information.

So that's one way, so with sequences that are targeted. In some cases, those targeting sequences remain part of the protein. But in other cases, in order to ensure that the protein stays put, the targeting sequences are removed. So that's another important point. You may keep the targeting sequence, or you may lose it through the action of another enzyme that cuts off the targeting sequence when destination has been reached.

Now, there's a second way that we can program where a protein may go. And these are rather useful transformations that make things even more dynamic. So let me walk you through a concept.

If you think of a protein that's made on the ribosome, it's got a targeting sequence. In order to get that protein to destination, you've got to make a new batch of protein that's going to go to its destination. It's going to end up in the mitochondria. You've got to make the protein de novo.

Sometimes when we need to have the action of a cell we can't wait that long. We can do things quickly and expect the cell to suddenly change what it's doing. Because we're sitting around waiting for the ribosome to make new copies of the protein. So the second way in which proteins are targeted to new destinations is through what's known as post-translational modifications.

This is so unfair, Adam. I saw you using the middle boards, but it looked so much easier.

So the second way to target a protein to a destination is using post-translational modification. What does this mean? What it means is that the protein is made. It's ready. It's waiting. But we haven't engaged its final destiny. We haven't triggered it to go where it needs to be. But we're waiting for an enzyme to just carry out a seemingly minor modification of that protein. And then the protein will go to its destiny.

And I've shown you here examples of three types of modifications. One we will talk about today, because it's very simple to understand, lipidation. And then the other two, we'll talk about next time, phosphorylation and ubiquitination.

And these are all what are known as PTMs, Post-Translational Modifications. And they are changes that occur to an amino acid side chain within an already made protein to alter its destiny. And I'd like to talk about lipidation first, because I get to remind you about cellular membranes. So remember, we've talked about these semipermeable barriers that are around organelles and around cells.

And let's say that this is a membrane-- I've got to put my-- that exists between the cytoplasm and the outside of a cell. And let's say I have a protein lurking around in the cytoplasm, but I need it at the membrane. I need it to get involved in a signaling process. And I need it now to be there.

If I have a soluble protein, it's not associated with the membrane. But I can use another enzyme to attach a hydrophobic, greasy tail to that protein. So what it really wants to do is to get to the hydrophobic membrane. Lipidation is such a modification.

It's just the modification with a long-chain, often C16, C18, fatty acid that then renders the protein lipophilic and makes it want to move, and insert this lipophilic tail into the membrane, and part the protein of the plasma membrane. So the information is still, though, encoded within the protein.

How could that happen? How could I have made that information be in the protein? What might be the strategy there? It's still encoded, but it's secret. It's cryptic. Any ideas?

So I'm not going to just glom this group onto a protein. I'm going to put it somewhere specific. And so oftentimes, lipidation reactions occur site-specifically at particular sites within a sequence, and an enzyme recognizes that site and transfers the lipidic molecule to it.

So lipidation actually may occur, for example, of the amino terminus of a protein. But if there are certain features within that protein, you may then attach the lipidic group. So once again, using bioinformatics, you can look at the target protein of interest and predict that it's the target of a post-translational modification reaction.

So once again, the information is programmed into the sequence, but it's quite cryptic. It could be within the middle of the sequence. There could only maybe be a couple of clues. But the clues are there nonetheless that can be parsed out using computer learning and screening of sequences to say that is a target for lipidation, or phosphorylation or such.

Is that clear to people? Does that make sense? The information is encoded, but you can't see that it's there. But the advantage of the post-translational modifications is that they occur on demand, as opposed to making a new protein de novo, and then having it go to a particular cellular location.

Later on, when we talk about phosphorylation, you will see that phosphorylation is the bread and butter of cellular signaling. It's the light switch in every room in the cell that turns on and off in order to make functions happen within the cell. And that's a really major, dynamic post-translational modification that has significant meaning.

So the reason on this little image-- I just wanted to show you the membrane and just remind you that the membrane is a supramolecular structure that's assembled with a hydrophobic core and polar head groups on both faces, as I've sort of indicated in this cartoon. So let's start with sequences that might take us to the nucleus.

Now, the nuclear membrane is rather a strange entity. Because the nuclear membrane isn't a simple membrane like the plasma membrane. It's actually a double-layered membrane. So if you look at a nuclear membrane-- and I'm just going to do a job of showing a portion of the nuclear membrane.

Within the nuclear membrane, there are pores, quite launch openings. And the membrane is actually a double membrane, where all of these lipid bilayers. So it's not a single membrane. It's a double membrane with large openings.

And you might say, well, that's no use. There's just these great big, gaping holes in the nucleus. Anything can come and go if it wants. But the nuclear pores are kind of a special structure. Because they have a protein that's kind of disordered, that creates a tangled network. That means that that pore isn't totally open, but there's some stuff that something's got to get through to get from one side to the other.

And my colleague Thomas Schwartz in biology works on the macromolecular structure of nuclear pores to understand these structures. Because these are also made through the auspices of having a lot of proteins that help create this structure. Otherwise, that membrane wouldn't stay in its proper format.

So in order for a protein to get into the nucleus, if it needs to, or leave the nucleus, it has to have some kind of mechanism to get through this structure that's plugging the nuclear pore. So this would be the inside of the nucleus. And this would be the cytoplasm.

So as shown on this slide, the nucleus, there's a particular protein sequence that's appended to a protein. That's known as the Nuclear Localization Sequence, or NLS. And what an NLS sequence is, it's a short sequence of amino acids that enables a protein to get to its proper destination.

And these sequences are quite well recognized. They may end up being highly basic sequences. So an example of an NLS would be Lys-- it's not very exciting, but it just goes on, Lys, Lys, Lys, arginine, lysine. And it may be bounded by hydrophobic residues or other types. So that would be a typical NLS sequence that's in a protein.

And I want to remind you that lysine and arginine all have side chains that at physiological pH are positively charged. So the nuclear localization sequence is something that's easily recognized because of this sort of short sequence that may be at the N- or C-terminus. I think there's either possibility. But it's a very clear sequence. You could look at your protein sequence and say, there's an NLS on that sequence.

And it's the NLS sequence alone that's responsible for getting the proteins in and out of the nuclear pore. Let's mostly focus on getting into the nucleus. Basically, you have a protein structure that has an NLS sequence at one terminus. And that NLS sequence binds to another protein.

Creatively, you had a little bit of chance to give proteins names in the exam. It's called importin. So it's an import protein that binds to the NLS, and as a consequence of that, will carry cargo. It will escort cargo into the nucleus of the cell. And it sends it through this meshwork of proteins.

That's a very loose mesh work of proteins. And they're not ordered proteins. They're highly disordered proteins. So they make more of a filter than a plug. But they are definitely something that doesn't allow any old protein to go through that nuclear pore.

NLS tags are very easy to recognize, once again, through bioinformatics analysis. And what's really cool is that you can reprogram a protein to be where you want by manipulating the NLS. So this is rather a nice set of experiments.

Let's say we have a protein that we're going to micro-inject into the cytoplasm of the cell. And we want to program it to either go into the nucleus or stay outside the nucleus. That can be done readily by attaching a nuclear localization sequence to a protein along with a fluorophore dye or fluorescent protein that will allow you to observe that experiment. If you micro-inject into the cytoplasm, that protein that's got an NLS will get run into the nucleus through association of the NLS with importin. But if you chop that NLS, the protein the stuck, remains out in the cytoplasm.

Let's say you want to study a new protein. I just want to show you that these NLS sequence are totally independent of the cargo they carry. You can just stick an NLS on your favorite protein who you want to interrogate.

Let's take pyruvate kinase. It doesn't have anything to do with specific transport to the nucleus. But nevertheless, if you put-- if it doesn't have an NLS, it's fluorescently labeled, it stays outside in the cytoplasm. But if you put an analysis on it, you concentrate into that region of the cell.

So these experiments show you that what we know about these targeting sequences can be manipulated and used to enable you to move things around in the cell. So that's one particular type of mechanism. The next mechanism I want to describe to you is the mechanism that's used for mitochondrial transports. And it's a little bit different in its strategy.

So to get into the mitochondria, there is, again, a recognition sequence, in this case, a mitochondrial localization sequence that has particular characteristics. In this case, the mitochondrial localization sequence, let's say it's at the N-terminus of your protein. And it would be something that might be a mix of charges. Some Arg, Glu, Arg, Glu. So that's a typical MLS sequence.

And in this case, the charge at physiological pH is different from the nuclear localization sequence, because it's an alternating positive and negative charge. So this is pretty different from this. It doesn't say bioinformatics to figure that one out. So you can then pick out mitochondrial localization sequences.

And so in this case, remember, mitochondria make some of their own proteins on their circular DNA. But they've abandoned expressing all the proteins that are needed in the mitochondria. And some proteins are transported into the mitochondria using these types of sequences. But the approach, the strategy, is different from getting into the nucleus.

In this case, the MLS sequence associates with a protein channel that is in a closed state. So here's a membrane. Here's the makings of a channel. But it's in a closed state.

But once the protein with the NLS sequence binds to that, that channel opens. It's triggered by the binding of that sequence to a portion of the protein that's outside that membrane. And that then allows the protein to be unfolded and transported into the mitochondria, where that sequence may be removed. And then protein refolds in the mitochondria.

So it's a very different strategy for that and the nuclear localization sequence. So you'll find, for many different organelles in the cell, there might be very specific localization sequences that you could look up and learn about. But one thing I want to mention to you is that these localization details are very important. And many diseases in cells are a consequence of proteins not being localized to the right place.

If you're not in the right place at the right time, then things will start to go wrong with the signaling or the processes of the cell. So diseases are frequently associated with mislocalization. So now what we're going to do is basically say, we've taken care of understanding things made in the cell. They either stay in the cytosol or they'll go to organelles based on particular types of strategies that are largely dependent on short tagging sequences, but in other cases, may be dependent on post translational modification.

All right. So here is a cartoon. But actually, I want to do something slightly different if it doesn't take too long.

Now, when we first talked about translation on the ribosome, what you see there in green and yellow is the ribosome. The dark band is a messenger RNA. The dark blue are transfer RNAs that are being helped with elongation factors to get to the ribosome. But what I want to point out here is the emerging sequence of polypeptide coming out through a tunnel on the ribosome.

Now, if a protein is going to be destined outside the cell, it is expressed with what's known as a signal sequence. It's about a 20-amino acid residue sequence that is recognized by the signal recognition particle. And then translation slows down and clamps the ribosome on the endoplasmic reticulum membrane so that the new peptide starts being threaded into the endoplasmic reticulum through what's known as the translocon. So you're now not sending the protein out to the cytoplasm, but you're rather sending the protein into the endoplasmic reticulum. And you're also sending it down this branch of the protein biosynthesis pathway.

You see this piece of protein emerging. This hatched portion is the cytoplasm. The gray portion is the endoplasmic reticulum. So there is a complex machinery at play that enables proteins to be made in the cytoplasm but now targeted to a completely new location. And these are the proteins that are going to be destined to either stay in the plasma membrane or be secreted from the cell.

And this view here gives you a little bit more than the cartoon. So ribosome-- a signal peptide is made that is a green peptide sequence that's about 20 amino acids long. That is actually called a signal peptide. It's signaling for synthesis through the endomembrane network.

That causes the ribosomes to dock down on the cytosol ER membrane and keep on being synthesized so that proteins are made into that endomembrane system. And you can think of this cavernous endomembrane system as your tunnels out of a cell for either display on the surface of the cell or for secretion entirely in vesicles. So let's take a look at how that occurs.

When you make a protein in that way, see the dark dots, the rough ER? These are ribosomes that are attached to the membrane. Proteins are made into the membrane. And then the endomembrane system is not really just a tunnel or a labyrinth. But actually, each of those layers spits off vesicles that fuse with next layers to gradually make their way outside of the cells.

So here you see there are vesicles. You're always keeping proteins associated with membrane as you go through the endomembrane system. And here is a vesicle that's got protein in it. It may either release it to the outside of the cell, or the protein may be associated with the membrane of the vesicle and stay parked in the plasma membrane.

And so I just want to give you one final slide where I talk about the biogenesis of membrane proteins. Now, this is pretty complicated stuff. Because you have to remember what's inside and out. So I spent more time than I should have on this cartoon to show you which end of the protein ends up outside the cell and which inside the cell, and how you make multi-membrane-spanning proteins.

So let's take a look at this in detail now, looking-- here's the ribosome. Here's the protein emerging. If there's signal sequence there, that ribosome docks down on the membrane and starts translating the protein, amino terminus first, into the endoplasmic reticulum. We'll all OK with that.

As synthesis continues, we may reach the stop codon on the messenger RNA. And what may happen is that the protein may remain associated with membrane. The amino terminus will be in the ER. And the C-terminus will remain on the other side. There are a number of different configurations.

But if we want to start to transport this protein to the surface of the cell, that will then stay associated with membrane but not in the form of the flat membrane that it was delivered into. But that membrane may pinch off into a spherical vesicle. But you still have the C-terminus outside and the N-terminus inside. That will then work its way through the endomembrane system, and ultimately, fuse with the cytosol. This is the really fun part.

And then, once it's fused with the cytosol, it has the option to be displayed on the outside of the cell. Why? You have a protein. The N-terminus is on the outside. The C-terminus is on the inside.

So that shows you the biogenesis of the cell surface protein that's stuck in the membrane through its membrane-associated domain. If you're not going to stay with the membrane, you can actually also simply release this into the vesicle for release of a soluble protein. I will not go through this. But there are miraculous steps that end up in the biogenesis of multi-transmembrane proteins.

Because each of those transmembrane domains gets made in the translocon and gets shuttled sideways. And you start piling up transmembrane domains that span the membrane. And in the next class, we're going to see how useful these proteins are in cellular signaling. So those are very important proteins to think about.

One last thing-- so let's think about this. For either configuration, either post-translational modification or using targeting sequences, when do we define where the protein's going to end up? Where's the information first defined? Anyone want to answer me and explain why? Yes?

AUDIENCE: Would it be B, the mRNA sequence, because that would have a significant portion of the splicing?

BARBARA IMPERIALI: It's a good try. But you want to remember, yes, splicing is important. But when was the sequence actually in the entire pre-mRNA? When would that have been defined? Yeah? Sorry. Carmen?

AUDIENCE: Is it in the genomic DNA sequence?

BARBARA IMPERIALI: Yes. Because you never have information in the RNA that wasn't in the DNA. So the DNA has got the information there. Yeah, it may need a bit of splicing to put things in the right place. But the information is there in the DNA.

So you want to remember, for all of this targeting information, it's in the genomic information most commonly. It's the genomic information that has the patterns of sequences for post-translational modification. It's the genomic information that has things like NLSes and MLSes. They're already there.

But they are often encrypted. And there was a very nice point there, though. If you want to send to make a single chunk of a genome that encodes either a protein that's going to be exported through the secretory pathway or stay in the cytosol, you might splice in or out a signal sequence. So that's a really good way, using the same original DNA sequence, to actually get to proteins that fulfill different final destinies within the cell. So next time, we're going to talk about signaling. It's going to be a blast.

Chronic Fatigue 'Brain Fog' Clues in Spinal Fluid

TUESDAY, March 31, 2015 (HealthDay News) -- People with chronic fatigue syndrome show a distinct pattern of immune system proteins in their spinal fluid -- a finding that could shed light on the "brain fog" that marks the condition, researchers say.

The new study found that, compared with healthy people, those with chronic fatigue syndrome had lower levels of certain immune-system proteins called cytokines in the fluid that bathes the spinal cord and brain.

The exception was one particular cytokine, which was elevated in not only people with chronic fatigue, but also those with multiple sclerosis.

The finding could offer clues as to why people with chronic fatigue syndrome typically have problems with memory, concentration and thinking, said lead researcher Dr. Mady Hornig, a professor at Columbia University's Mailman School of Public Health in New York City.

The study also bolsters evidence that some type of immune dysfunction underlies the puzzling disorder, Hornig said.

Chronic fatigue syndrome is known medically as myalgic encephalomyelitis/chronic fatigue syndrome, or ME/CFS. In the United States, it affects up to 2.5 million people, according to the Institute of Medicine, a scientific panel that advises the federal government.


In February, the IOM released a report affirming that chronic fatigue syndrome is a legitimate medical condition that many health professionals still misunderstand -- or even dismiss as a figment of patients' imagination.

The term "chronic fatigue syndrome" was coined back in 1988, and in hindsight, it was a "lousy" choice, said Suzanne Vernon, a virologist and scientific director of the Solve ME/CFS Initiative, based in Los Angeles.

"People hear it and think, 'Oh, you're tired. I'm tired, too,'" said Vernon, who was not involved in the study. "But this is debilitating fatigue. It's like having a case of the flu that never goes away."

Plus, symptoms go beyond fatigue, and include what's been dubbed "brain fog" -- a collection of thinking-related problems such as confusion and difficulty with concentration and short-term memory.

For the new study, reported March 31 in the journal Molecular Psychiatry, Hornig's team studied spinal-fluid samples from 32 people with chronic fatigue syndrome, 40 with multiple sclerosis, and 19 healthy people.


Overall, the researchers found reduced levels of most cytokines in chronic fatigue syndrome patients' spinal fluid, versus the two other groups. But one cytokine, eotaxin, was elevated in people with chronic fatigue syndrome and those with multiple sclerosis.

The significance of that finding is not clear yet, Hornig said. But she said eotaxin is involved in allergy-like immune responses.

To Vernon, the findings offer "additional evidence of clear [biological] markers in ME/CFS."

"These biomarkers are indications of some kind of disease process," Vernon said. In other words, she added, chronic fatigue syndrome is "not made up."

Why did the study include people with multiple sclerosis? There are some similarities between MS and chronic fatigue syndrome, Hornig explained. MS patients suffer fatigue, and the disease is believed to be caused by an abnormal immune reaction -- in this case, against the body's own nerve tissue.

The precise cause of chronic fatigue syndrome is far from clear, but in general, it's thought to involve some type of immune system dysfunction, Hornig explained.


In a recent study, her team found that in people who've had chronic fatigue syndrome for a relatively short time -- fewer than three years -- cytokine levels in the blood were actually elevated. They dropped again, though, in people who'd had the disease for a longer time.

People in the current study had had chronic fatigue syndrome for about seven years. So the relatively low cytokine levels in their spinal fluid "parallel" what was seen in the earlier study, Hornig said.

"I think what we're seeing is an immune system exhaustion over time," Hornig speculated.

The theory is that the immune system may initially go into overdrive against an invader -- like a virus -- and then be unable to dial itself down, Hornig explained. That could account for the high cytokine levels in people who've had chronic fatigue syndrome for a short time.

Over time, though, the immune system may essentially wear itself down, leading to weak responses to mild infections that a healthy immune system would readily handle, Hornig suggested.

One hope, Hornig said, is that these findings could lead to objective tests that can diagnose chronic fatigue syndrome early.


An objective test, such as a blood test measuring cytokines, would be welcome, Vernon said. Right now, she noted, people often wait for years for a diagnosis, which is based on symptoms.

Understanding the biology of the disease could also lead to treatments, Hornig said.

"We can't promise this will translate into treatments around the corner," she said. "But we hope to start giving doctors some tools."

DeepMind "solves" protein folding November 30, 2020 8:55 AM Subscribe

My love affair with DeepMind (many previous posts on MeFi) continues.

One side effect of this breakthrough is likely that foldit is dead in the water.

As always with this kind of thing, "wait for the paper" applies. Having said that, DeepMind's record on this stuff a la AlphaGo is pretty solid. They don't seem in the habit of overhyping early results for press releases.
posted by lazaruslong at 9:11 AM on November 30, 2020 [2 favorites]

As always with this kind of thing, "wait for the paper" applies.

This is a little bit future-tech, jet-packs for all my friends! level news. I mean, this doesn't sound like . in five years we'll have fusion reactors powering our toaster-ovens.

If this is actually functional this is a follow-up to CRISPR : or puts it in a new category of usefulness, doesn't it?
posted by From Bklyn at 9:15 AM on November 30, 2020

What a beautiful contribution to humanity from Google!

Sorry, apparently this is made by Alphabet whoever that is
posted by East Manitoba Regional Junior Kabaddi Champion '94 at 9:19 AM on November 30, 2020 [4 favorites]

Protein folding has been a pretty heavy-duty problem. One of the PIs behind Foldit gave a talk a few years back, noting that the algorithms hit a wall (at that time) and no amount of raw computation power in our lifetimes would really help on getting an accurate answer within a usable timeframe, unless there were improvements there. Drug and catalyst designs could be helped immensely with a fast, accurate system.

As with our CASP13 AlphaFold system, we are preparing a paper on our system to submit to a peer-reviewed journal in due course.

That's going to be the real test, I think. Publication and reproducing results are going to be what meet the actual gold standard. And as much as Google relies on public datasets (publicly-funded datasets) to train their networks, they do not so far have a good track record on reproducibility, where machine learning intersects with academia:

I work in pharma, with computers, and I get e-mails from vendors trying to convince me to be really excited about AI advances, with the implication that I will be missing out on things if I don't pay them to guide me through the thicket.

So I got an article much like these about a year ago from a sales side guy, which included some professor making dismissive quotes about the dinosaurs in industry. He had accidentally included the internal e-mail chain though, so I saw that their technical guy had already succinctly made the point that, while impressive, protein folding is not really the limiting step in drug discovery.

And for your next Nobel Prize, go the other direction: given a target structure (e.g. a receptor), predict the amino acid sequence that results in a protein that binds to that target. That would practically solve biologic drug design.

The article says you can get as a good as cryo EM or x-ray. It's worth remembering that you can't predict anything about potency from that--the difference between nanomolar and micromolar is sub-angstrom and can't be deduced from looking at a structure. (It can sometimes be modeled when you have a structure, but you're still talking 10x or more uncertainties.)

I don't mean there is zero chance. There are certainly people who think this sort of thing can be applied to ligand/protein binding and do better than, or meaningfully improve, the QM approach.
posted by mark k at 9:38 AM on November 30, 2020 [16 favorites]

A few decades ago you could get your PhD by solving one protein structure.

protein folding is not really the limiting step in drug discovery

. this is also extremely true.
posted by aramaic at 9:58 AM on November 30, 2020 [4 favorites]

As mark k notes, static protein folding is not the limiting step because the difference between "lab curiosity" and "drug lead" let alone an actual drug is due to the dynamic behaviour of proteins. Still very cool though and maybe this method could be a promising new avenue for solving the latter? No idea as I don't know much about protein-drug interaction modelling but still cool.
posted by atrazine at 10:30 AM on November 30, 2020 [3 favorites]

I'm amused by the idea that the best extant approach to the protein folding problem involves evolving an algorithm that's probably about as analysis-resistant as the protein folding it simulates.

Correct me if I'm wrong, but isn't it the case that nobody, not even the people responsible for winding these things up and letting them rip, really has any idea how any of the Alpha* family actually works? I mean, we can clearly see what they do, but is there anything to be learned by inspecting the innards of a successfully trained Alpha* other than a huge and completely inscrutable table of mysterious numeric weightings?
posted by flabdablet at 10:55 AM on November 30, 2020 [4 favorites]

And for your next Nobel Prize, go the other direction: given a target structure (e.g. a receptor), predict the amino acid sequence that results in a protein that binds to that target. That would practically solve biologic drug design.

Target binding is NOT the rate-limiting step in drug discovery. We're really good at finding binders the vital work of binders that have no offtarget effects, the right ADMET, and years of in-vivo testing are usually much bigger problems and timesinks. The data sets for those are much noisier and will be much less tractable to "solution"* by deep learning compare to folding, IMHO.

Paraphrasing Derek Lowe author of inthepipeline, "if we could discover selective binders in hours rather than weeks, it's still the equivalent of saving 5 minutes off a taxi ride to the airport for a multi-leg transnational flight"

* deep learning will affect all of those topics and speed them up, but protein folding takes a pretty well graphed and well defined problem - orientation of atoms in space, whereas e.g. toxicity/off target effects: the literature and graph is so.much.bigger and noisier.
posted by lalochezia at 11:09 AM on November 30, 2020 [8 favorites]

DeepMind has also been working at throwing deep neural networks at quantum chemistry (very readable and interesting blog post).

None of this is really "AI". it's more like, modern AI required getting really good at solving complicated optimization problems, and it turns out you can use these new (or old with new tweaks) techniques on optimization problems from physics and chemistry too.
posted by vogon_poet at 11:15 AM on November 30, 2020 [1 favorite]

Correct me if I'm wrong, but isn't it the case that nobody, not even the people responsible for winding these things up and letting them rip, really has any idea how any of the Alpha* family actually works?

I can't speak to this model specifically, but the notion that deep learning models are total black boxes hasn't been true for a few years now. There's still a lot to be done in the field of explainable AI, but there has been progress.
posted by jedicus at 11:19 AM on November 30, 2020 [8 favorites]

I'm seeing a lot of comments second guessing the importance of the results (especially on the orange website). It's worth keeping in mind that these kind of results are similar to turning a trip to the library into a google/wikipedia search: sure, all the same knowledge might be available, but shaving orders of magnitude off the latency can unlock a lot of questions that you might not have even bothered trying to answer before. I'm really excited to see where this leads!

My understanding is that the computational accuracies here are about equal to the accuracy of the experimental outcomes. It would be super interesting to know if the 'misses' are uncorrelated. If they're uncorrelated, you get even better results by following up with experiments, and if they're heavily correlated you've got a really interesting next problem to solve.
posted by kaibutsu at 11:52 AM on November 30, 2020 [4 favorites]

For some of these problems, pragmatically, it doesn't matter so much if a human cannot understand and follow exactly what "moves" were made by a black-box optimiser to arrive at a good solution, provided the solution itself can be externally verified without depending on that black-box, and the solution can be measured to be superior when compared to solutions produced by competing approaches.

You can have a similar philosophy for solving other combinatorial optimisation problems that pop up in real life. For a completely arbitrary example that isn't protein folding, consider mega-bakery supply chain optimisation:

> [Pasco Shikishima Corporation]'s supply chain involves 15 factories in Japan, each one with several production lines, and more than 100 distribution centers. Pasco's catalog contains more than 1,000 products. 900,000 orders have to be executed each day in Pasco's factories. For each order, Pasco has to decide where and when to produce it. Moreover, Pasco has to decide where to source raw materials and which routes to deliver distribution centers. The goal is to minimize production and distribution costs over several days of horizon, while respecting production and distribution capacities.

Does it matter exactly how they figure out which 900,000 orders to execute each day? If a human does it with chalk on a blackboard or if some software black box does it? Not really, provided they have a repeatable process for doing so, since they need to keep generating new plans. Does it matter how good the plan of 900,000 orders is? (e.g. how accurate it is at reflecting the constraints of the real situation and maximising profit or what-not). Very much so.
posted by are-coral-made at 11:56 AM on November 30, 2020 [1 favorite]

Does it matter exactly how they figure out which 900,000 orders to execute each day? If a human does it with chalk on a blackboard or if some software black box does it? Not really, provided they have a repeatable process for doing so, since they need to keep generating new plans.

IME the problem with AI is that you can't explain why it does what it does, without resorting to math talk that pointy haired bosses like even less than hand waving about how it's 90% accurate (but 10% of the time makes bizarro world errors), but don't ask me why. The problem isn't measuring accuracy, it's selling other people on employing AI. As jedicus points out there are efforts being made in Explainable AI but IMHO they're too nascent. "I built another equally-unexplainable AI to help me explain the first one" isn't really a great answer. Certainly the type of people that make non-academic, real-world decisions in many arenas aren't going to accept that.
posted by axiom at 12:26 PM on November 30, 2020 [2 favorites]

the notion that deep learning models are total black boxes hasn't been true for a few years now

That article doesn't really bear that out, it's more like "yo dawg I heard you like opaque black boxes". Apparently they're using neural networks to generate plausible explanations about what other neural networks are doing?
posted by echo target at 12:26 PM on November 30, 2020 [2 favorites]

the notion that deep learning models are total black boxes hasn't been true for a few years now.

Still pretty fucking dark charcoal grey though.
posted by flabdablet at 12:29 PM on November 30, 2020 [7 favorites]

using neural networks to generate plausible explanations about what other neural networks are doing

does sound like a pretty plausible explanation for an awful lot of human "reasoning".

That said, I'm not at all sure that I want to be driven along a highway three metres behind another car at 100km/h by something deliberately engineered to implement Artificial Cognitive Distortions.
posted by flabdablet at 12:31 PM on November 30, 2020

And new drugs, and new drugs, and new drugs for problems caused by the problems we have created, and new drugs.

to be sure we've created some problems. and the usual disclaimers about capitalism, colonialism and power pertain.

to characterize drug discovery as only responding to the modern condition is to dismiss the unnecessary suffering and early death of billions of people over the last few thousand years.

fundamentally, bodies develop disease, and NO value system or culture has addressed disease as well as modern science. sorry. but it's true.

- modern medicine - of which drug discovery is a part - has helped more people than any. social. movement. you could name. for most of those people there was NO other way to help them.

we have words for people who want to stop people minimizing suffering.
posted by lalochezia at 1:49 PM on November 30, 2020 [36 favorites]

Does it matter how good the plan of 900,000 orders is?

If there's a better plan, yes. As a network gets increasingly complex, a loss function can be rife with local minima (some visually extreme examples), so you are certainly not guaranteed a globally optimal solution.

That's one reason people want to understand how the larger network is built: where it fails can lead to catastrophic outcomes in edge cases a network cannot handle (e.g., self-driving cars, AI medicine or guided cancer diagnoses, etc.).

Delivering baked goods is not a life or death matter, granted, but for most companies, finding efficiencies is all about getting from one local minima to a deeper and more profitable one, whether it is by using neural networks or some other process/model.
posted by They sucked his brains out! at 2:41 PM on November 30, 2020

is there anything to be learned by inspecting the innards of a successfully trained Alpha* other than a huge and completely inscrutable table of mysterious numeric weightings?

Probably not, but maybe it's a sufficiently complex problem that there's no aesthetically pleasing solution. Just like there's no music of the spheres, if you want to predict the planets' motion it's just a bunch of boring numerical integration.
posted by RobotVoodooPower at 6:06 PM on November 30, 2020 [2 favorites]

is set up that this actually makes a case for big bad corporations to be making the breakthroughs now.

Historically this was true in many cases. E.g. IMB and transistors, Exxon in soft matter, etc. I would be super interested to see a real history, but the influence of companies in scientific progress is longstanding. It went through a lull because of profit-driven cuts to basic research.
posted by lab.beetle at 7:15 PM on November 30, 2020

The experience of the chess community with DeepMind and AlphaZero was very mixed and may be a cautionary tale for researchers in protein folding. To recap they claimed that their software was able to beat Stockfish — the most powerful chess program at the time. People were very excited that a significant improvement in chess programs would would help bring new insights into the game that top players would be able to bring to their games.

At the top level modern classical slow chess (as opposed to the rapid and blitz game that has become popular during covid) has become very driven by computer analysis of opponents games and various opening lines in advance of the match. So the initial hope around AlphaZero was that top players would be able to use this tool to get a flood of new ideas and insights into positions.

Yet when the details were released it was clear to most involved in the Stockfish open source project that Google had crippled Stockfish and configured it to get an outcome that would make a good press release. Then they never made it available for any kind of real external scrutiny . The hired a couple of well known chess authors to review some of its games and put out a book — but that seemed like more google /Deep Mind PR.

The Chess Engine Championship TCEC is the most important open competition for chess programs and AlphaZero has never entered. However open source developers have attempted to user the techniques described in the DeepMind research papers to make an open source engine called “Lc0 /Leila”.

Lc0 has quickly become a top program but unlike in Google’s paper it has not shown a clear leap over Stockfish and generally has finished second. Stockfish itself recently shifted to something called NNUE (Efficient Updatable Neural Networks) as a model for a deep learning approach. That seems to have shown much more promise than Lc0 and has resulted in some impressive gains in their performance.
posted by interogative mood at 9:42 PM on November 30, 2020 [5 favorites]

There's a interesting question I got from reading Hacker News (adding grain of salt), of what this means for academia vs Google, is there something wrong about how academic research (in protein folding or in general) is set up that this actually makes a case for big bad corporations to be making the breakthroughs now. Will Google get the Nobel next year? Should academics do some soul searching, between chess and now molecular biology? I'd like to see a journalist or expert write about this social aspect of doing science.

It's a very superficial analysis to split the people doing this kind of work into "academia" vs "private sector". What's really important is the conditions under which people are working. DeepMind is part of Google, sure, so "private sector" whatever that means to you, but Google itself if an incredibly profitable monopoly able to buy and then fund DeepMind to do this kind of work on the basis that they might make money from it somehow. As a result, it is very well funded and they are under no pressure to produce or sell a product the way most of the "private sector" is nor under pressure to publish papers (regardless of actual merit) nor to write grant proposals the way that people in academia have to.

If anything, this is an example of the effectiveness of the XeroxParc / Bell Labs approach of just giving a bunch of smart people money to work on things with a very vague steer on what that should be over either the way academia currently works or the way that most of the private sector approaches r&d spending. Google of course is able to produce things where they only capture a small fraction of the public good generated through the incremental knowledge because of their monopoly position.
posted by atrazine at 4:53 AM on December 1, 2020 [4 favorites]

Former computational biologist here who worked on protein sequence alignment, it's impressive work but not even close to being a solved problem. One of my best friends has spent the last 15 years working on protein folding, so I picked up a bit from him too.

Someone who once interviewed me for a grant laid out some of it pretty well today.

Basically (combining the opinions of me and the people mentioned):
- the advance in average accuracy is impressive
- some existing programs do better on certain types of protein
- the way proteins fold depend on conditions in the cell, the most basic example being many proteins need other proteins present (chaperone proteins) to fold correctly, so it's not necessarily deterministic
- some proteins are like dry tagliatelle nests and some are like nests where bits have been cooked and can move around, in fact in general their function is to move around - what does it even mean to "solve their folding"
- Google have computational resources (good) and a PR department (ehhh) that the average research group can only dream about
posted by kersplunk at 6:51 AM on December 1, 2020 [2 favorites]

Google of course is able to produce things where they only capture a small fraction of the public good generated through the incremental knowledge because of their monopoly position.

Not sure if I read this correctly, but if a big bad capitalist corporation is appropriating the public good, i.e. the fruits or giant shoulders of academia, it nevertheless is a problem of academia vs Google leading the research breakthroughs in a given field, only that the assessment is Google's approach is fundamentally parasitic or at best dependent on academic basic science. And still the problem remains that academics were in the CASP race too, so does that not beg some kind of reflection on academia's own structures regardless of how Google's R&D functions? (Example, just pay professors and grad students Google quantities more money is the condition, but that cannot happen independently of reorganizing academia and its sociological context.)
posted by polymodus at 1:09 PM on December 1, 2020

Not sure if I read this correctly, but if a big bad capitalist corporation is appropriating the public good, i.e. the fruits or giant shoulders of academia, it nevertheless is a problem of academia vs Google leading the research breakthroughs in a given field, only that the assessment is Google's approach is fundamentally parasitic or at best dependent on academic basic science. And still the problem remains that academics were in the CASP race too, so does that not beg some kind of reflection on academia's own structures regardless of how Google's R&D functions? (Example, just pay professors and grad students Google quantities more money is the condition, but that cannot happen independently of reorganizing academia and its sociological context.)

Not quite what I meant. Research has value. Companies try and capture the value of innovation they pay for but the nature of pure research in particular is that it isn't possible to capture it through patents and trade secrets sufficiently and therefore "normal" private sector companies don't do it.

Society benefits from research (even if you just look at the vulgar economic side) and benefits more than it costs to do it. A rational society therefore should spend lots on research. If you leave research entirely to the competitive, normal parts of the private sector, very little would ever get done. That's because private sector companies will only fund research where they can capture enough of the value to cover their costs and profit and most basic research generates most of its value over a long period of time and very broadly which means that the private sector tends to concentrate its R&D in very narrow late stage development spending which has a short pay-back and is easily capturable through IP laws.

Monopolies have historically spent on R&D almost as if they were governments and DeepMind is analogous to the 20th century Bell Labs, money is being to spent do relatively basic research into deep learning with the sort of vague directional idea that Google might make money from it but they're not developing products or anything like that. A non-monopoly would be unlikely to act this way because it's not economically rational. unless you're able to control such a massive slice of the economy or a sector of it that you can profit from general growth. Bell for instance was confident that improvements in communications technology would come back to them in the form of profits somehow or another because of their ability to extract monopoly rents. Google has invested heavily in their machine learning hardware and cloud business and a breakthrough in machine learning, even if they own literally none of the IP and/or give it away will make them a huge amount of money.

The model of Bell Labs was effectively to pay researchers, including early career researchers quite well and to let them do essentially what they wanted to do with very light supervision. Academia might want to reflect on how much more productive researchers would be if they were able to purely pursue ideas in a very different sociological context as you say. (Individual academics generally agree with that, nobody likes writing grant proposals, but there's a whole complex context around them).

My original point was that taking this as an example of "the private sector" being good at innovation is misguided since this kind of innovation mostly comes from a very unusual part of the private sector which has little in common economically with most business.
posted by atrazine at 1:34 AM on December 2, 2020 [2 favorites]

« Older How old, ambient Japanese music became a smash hit. | Sweet Land of Liberty Newer »

Exercise Precautions

You'll probably be able to exercise with no problem if your doctor says you can and if you pay attention to how you're feeling while you're working out.

Stop and get immediate medical help if you have pain or pressure in your chest or the upper part of your body, break out in a cold sweat, have trouble breathing, have a very fast or uneven heart rate, or feel dizzy, lightheaded, or very tired.

It's normal for your muscles to be mildly sore for a day or two after your workout when you're new to exercise. That fades as your body gets used to it. Soon, you might be surprised to find that you like how you feel when you're done.


National Heart, Lung, and Blood Institute: "Your Guide to Physical Activity and Your Heart."

How to make ice cream base using evaporated milk easier to work with?

I sometimes try to make “chewy” ice cream using evaporated milk or dulce de leche made using the “boil a can of sweetened condensed milk” method in the base.

For the batch I’m wrangling now, I used a 14oz can of dulce de leche, 3 cups of heavy cream, a uhh. lot(?) of cocoa powder (I just added more until the dairy looked like it’s not getting darker anymore) and a 4oz bar of dark chocolate. (The dulce de leche adding around 220g of sugar.)

When chilled, this thickened into a puddingy paste that barely flows without coaxing, and my (cheap) ice cream machine started being unable to move it in about 10-15 minutes. (Although a thermometer registered the mixture as below freezing near the middle-ish.)

I usually just pry this out of the machine and into the freezer because what else am I going to do, and end up with fairly tough, albeit not icy icecream.

Is there anything that would make this base easier to work with without significantly changing the ingredients? Recipes for this style of ice cream usually have even more evaporated milk compared to the rest in them, sometimes even adding egg yolks which would likely thicken this even more. (Based on my experiences making chocolate+caramel icecream, replacing sugar in the recipe with caramel made out of it seems to make it so that using egg yolks is the difference between having the base churnable or not.)

I’ve been debating whipping the liquid up (or maybe just tossing it into a blender) before and after chilling to aerate it before it goes into the churn - after all it’s mostly cream and should be around 25% butterfat, but I have no idea if this would be helpful, and maybe there’s better tricks.

I don't care — it just works

There are some Wittgensteinians who think Kripke misses the point. They argue that when you are part of a form of life, or when you wholesale accept a system of rules, part of doing that means that you don't question it. When you play chess you don't spend all your time asking, "But why do the knights move this way? It makes no sense!" You just play the game.

Likewise, when we speak to each other, we're not crippled by doubt that we might be choosing the wrong word. We just assume that we're right and get on with it. So, too, with Kripke's "rule following considerations." To understand a rule is to accept it, not to doubt it. Addition is no different.

But, that being true, it's still an interesting thought: How do you know that you're doing anything properly? We all think that we're competent and intelligent, but what if we're just monumentally lucky? What if one day, we're exposed as poseurs?

Jonny Thomson teaches philosophy in Oxford. He runs a popular Instagram account called Mini Philosophy (@philosophyminis). His first book is Mini Philosophy: A Small Book of Big Ideas

Part 2: Multiphase Liquid Behavior of the Nucleus

00:00:15.09 Hi.
00:00:16.21 My name is Cliff Brangwynne from Princeton University and HHMI,
00:00:19.20 and I'm happy to tell you about
00:00:22.24 some work on the multiphase liquid behavior
00:00:25.18 of the nucleolus.
00:00:27.09 So, we've been discussing these membrane-less nuclear bodies
00:00:32.04 in the first lecture,
00:00:33.21 and I'm particularly interested in these kinds of structures,
00:00:37.13 these condensates within the context of the nucleus of cells.
00:00:43.13 It's important to keep in mind that
00:00:45.22 the nucleus of the. of the cell.
00:00:47.10 that's the seat of the genome,
00:00:49.11 and all the organization that takes place
00:00:51.20 within the nucleus is entirely in the absence
00:00:54.09 of any membrane-bound vesicle-like organization.
00:00:59.17 So, these membrane-less nuclear bodies
00:01:02.18 really are key structures
00:01:05.26 for organizing the contents of the nucleus
00:01:07.19 and organizing the genome and gene expression.
00:01:09.18 They include structures like nucleoli that'll be the focus of this talk
00:01:13.07 other things like transcriptional.
00:01:15.29 these transcriptional factories that are regulating
00:01:18.21 expression of individual genes
00:01:21.15 these PML bodies, which are also playing roles
00:01:23.29 in the flow of genetic information
00:01:25.18 or Cajal bodies and Snurposomes,
00:01:28.11 involved in splicing, again, in gene regulation.
00:01:30.27 So, we would like to try to understand
00:01:33.26 how these structures form and what role they're playing
00:01:36.28 in gene expression and organizing the genome.
00:01:40.26 So, nucleoli, or a nucleolus.
00:01:44.24 and its plural is nucleoli.
00:01:47.12 these are really fascinating structures.
00:01:48.26 They've been known for over 150 years.
00:01:51.26 They're one of the first things that the early microscopists saw
00:01:54.18 when they looked at cells,
00:01:56.23 for example human cells like these HeLa cells.
00:02:01.24 So, the nucleoli are the dark.
00:02:04.12 large dark occlusions within the nucleus
00:02:06.21 of these individual cells.
00:02:09.01 They're really interesting in thinking about
00:02:12.29 this flow of genetic information from DNA to RNA to protein
00:02:16.21 because they're sitting on sites
00:02:19.18 where there's active transcription of the ribosomal RNA genes.
00:02:23.19 So, they're really important for the transcription of ribosomal RNA,
00:02:27.15 and the ribosome is the machine that actually makes protein,
00:02:30.05 ultimately, in the cytoplasm,
00:02:32.06 so they're also important for this step from RNA to protein.
00:02:35.17 And so, the way we think about the nucleolus
00:02:38.08 is it's this membrane-less condensate
00:02:40.29 that is helping to facilitate
00:02:43.22 the numerous reactions that are required
00:02:46.18 for processing these ribosomal RNA transcripts
00:02:49.16 to ultimately form these mature preribosomal particles
00:02:54.23 and ultimately the ribosome, this protein translational machine.
00:02:59.13 So, we started thinking about the nucleolus, you know,
00:03:04.07 right after the initial studies we did on P granules
00:03:06.09 that I told you about in the last talk.
00:03:08.05 The nucleolus.
00:03:11.11 you know, as this large membrane-less condensate,
00:03:13.03 we were wondering if it.
00:03:15.26 if it, you know, could be thought of as a liquid phase-separated assembly
00:03:18.25 in cells.
00:03:20.23 So, a postdoc in my lab, Steph Weber,
00:03:22.21 who's now a faculty member at McGill University,
00:03:25.20 started to tackle this question using C elegans as a model system.
00:03:28.21 And this movie shows cycles of assembly and disassembly
00:03:32.28 of the individual nuclei
00:03:36.04 within the developing C elegans embryo.
00:03:39.07 And what you see is that the nucleolar proteins
00:03:42.06 are condensing out of the nucleoplasm
00:03:44.21 into many of these droplets,
00:03:46.25 and then resolving ultimately around the two sites in C elegans
00:03:50.15 where ribosomal RNA is being actively transcribed.
00:03:53.28 And so, work from Steph
00:03:56.19 as well as our collaborators, Mikko Haataja and Joel Berry,
00:03:59.22 led to a mapping where we were able to show that
00:04:05.22 the dynamics of the assembly of the nucleolus are.
00:04:08.22 can be well described using
00:04:13.00 classical theories of phase separation,
00:04:15.09 in particular this Cahn-Hilliard formalism,
00:04:17.16 which is a kind of a. a way of describing,
00:04:20.19 what is the uphill diffusion that is.
00:04:22.26 that is required for phase separation
00:04:25.04 to sort of overcome that entropic effect that we talked about
00:04:28.12 in the last lecture.
00:04:29.24 So, nucleoli really are, you know,
00:04:32.27 a type of liquid condensate.
00:04:38.03 And actually, the first study that we did was in this frog oocyte system
00:04:39.28 to look at these nucleoli
00:04:42.04 and to start to ask these kinds of questions
00:04:44.19 about what they are as biophysical objects
00:04:46.24 and how to think about their assembly and structure and so forth.
00:04:51.01 Xenopus laevis is a really powerful system.
00:04:55.01 It's a. it's a. it forms an oocyte, or an egg,
00:04:58.28 that is ready to be fertilized. when.
00:05:02.25 it's very large. So, it's about a millimeter in size.
00:05:06.12 It contains a single large nucleus that's also quite large --
00:05:10.15 it's about 600 microns in diameter.
00:05:12.22 And inside this large nucleus,
00:05:15.15 there are numerous
00:05:17.19 -- actually hundreds or up to a thousand --
00:05:19.15 nucleoli, which you can see in these individual droplets.
00:05:22.12 droplet-looking structures within.
00:05:25.09 And so, this was a powerful system for us
00:05:28.03 because we could really start to interrogate these biophysical properties
00:05:31.28 within this large nucleus that contains many nucleoli.
00:05:36.03 And what we showed was that these nucleoli.
00:05:38.04 when we push them together, they fuse
00:05:40.09 if we start to pull them apart,
00:05:42.16 they'll actually undergo these liquid bridge rupture events,
00:05:44.14 which you see in this movie.
00:05:46.07 And this allowed us to start to measure
00:05:49.03 the properties, the viscosity and surface tension,
00:05:51.24 of these structures to really show that they were liquid-like,
00:05:55.11 but there's an interesting twist to this,
00:05:57.14 which is that this liquidity, this fluidity,
00:06:00.14 is dependent on non-equilibrium biological processes
00:06:05.19 that are occurring throughout the cell.
00:06:08.29 And so, this is showing that if we.
00:06:11.13 if we deplete the cell of ATP,
00:06:14.02 which is kind of the battery of the cell,
00:06:16.01 then the apparent viscosity of these structures
00:06:18.25 goes up significantly.
00:06:20.26 So, there seem to be features of non-equilibrium dynamics
00:06:23.24 that are regulating the fluidity of these structures,
00:06:26.06 and we. that's why we refer to them as active liquids,
00:06:28.16 a kind of active liquid condensate.
00:06:32.03 In the spirit of asking the very simplest question
00:06:35.02 that we can think of in starting out our research,
00:06:39.05 here we asked the question,
00:06:41.00 if these are really liquids.
00:06:45.00 if these are liquid condensates within the nucleus of this frog.
00:06:50.08 of this frog oocyte,
00:06:53.23 then why don't they all fuse into a single larger structure?
00:06:57.06 Why are they remaining as distinct droplets,
00:06:59.23 when I showed that if. you know, if we push them together,
00:07:02.12 they seem to readily fuse?
00:07:04.00 So, why aren't they all fusing into one large droplet?
00:07:06.08 And that's a. that's a. clearly, a very simple question,
00:07:10.14 and I think it's powerful, and it has led us down some interesting roads.
00:07:14.15 So, in trying to answer that question,
00:07:18.09 I want to introduce you to some very fundamental ideas
00:07:20.24 that are important in many areas of cell biology.
00:07:24.08 They're important to think about.
00:07:27.02 So, Brownian motion was first described
00:07:30.24 by Robert Brown in this really classic
00:07:33.20 and very interesting, highly recommended paper
00:07:36.09 from 1828.
00:07:38.08 So, in this paper he describes the random motion of pollen particles
00:07:44.05 and other kinds of particles inside of a cell,
00:07:46.11 and the way in which they undergo a random walk
00:07:49.29 when viewed in the microscope.
00:07:52.24 The paper is really fascinating for a number of reasons.
00:07:55.05 One of them is he's clearly very disappointed
00:07:57.13 when he discovers that the motion he sees is.
00:08:02.07 actually doesn't have anything to do with life,
00:08:04.28 in the sense that even, you know,
00:08:07.13 ground-up bits of quartz and other dead materials
00:08:10.21 undergo Brownian motion.
00:08:12.19 And so, he was pretty upset about that.
00:08:14.11 Of course, it's laughable now,
00:08:17.03 because we now recognize this as an absolutely fundamental concept
00:08:19.21 that's important not only for nonliving materials
00:08:22.02 but also within the cell.
00:08:25.09 And so, these are.
00:08:28.03 this is a movie of particles that are undergoing this Brownian motion,
00:08:32.01 these random thermal fluctuations
00:08:34.16 within a solution of water.
00:08:37.27 And so, the idea here is that the random energy is.
00:08:43.00 one can think about it as.
00:08:45.21 kT of energy, that thermal energy scale that we talked about before,
00:08:48.06 is kicking around these particles
00:08:51.04 and causing them to undergo a random. a random walk.
00:08:53.28 Now, there's some interesting mathematical features
00:08:56.24 of the way in which this works
00:08:59.02 that I want to draw to your attention.
00:09:00.25 First off, Brownian motion is different than directed motion.
00:09:05.01 So, directed motion is. if you're walking in a straight line
00:09:08.02 or if you're driving,
00:09:10.23 you know, on a straight highway
00:09:12.29 and you're just driving along.
00:09:15.01 So, in directed motion, we can think about.
00:09:18.00 if we're moving at a steady speed or velocity,
00:09:20.24 if we. if we go for twice the amount of time,
00:09:25.29 then we have traveled twice the distance.
00:09:28.15 That sort of makes sense --
00:09:31.01 you know, you sit in the car twice as long, you move at the same speed,
00:09:33.20 you've gone twice as far.
00:09:35.14 And so, if I were to ask you about the square of this quantity
00:09:38.11 -- what is the square of the displacement? --
00:09:41.04 then I would say, okay, well I just square both sides of this equation,
00:09:45.02 and I have some v-squared prefactor
00:09:47.02 and then I have the time squared,
00:09:49.00 and that makes perfect sense.
00:09:50.25 So, if I, you know, wait twice as long,
00:09:53.04 then the square of the distance should be.
00:09:55.06 you know, it should be four times as large.
00:09:58.23 Now, what's interesting is for diffusive motion
00:10:01.07 the square of the distance is.
00:10:03.04 does not go like the square of the time,
00:10:05.04 but it only goes linearly with time.
00:10:07.23 So, that's pretty interesting, then.
00:10:10.03 If I wait twice as long, I don't.
00:10:12.01 you know, this quantity, this delta-x-squared,
00:10:15.00 isn't a factor of four larger.
00:10:17.10 It's only a factor of two larger.
00:10:19.22 And so, the prefactor here is called the diffusion coefficient.
00:10:22.21 It's something like the velocity, but a little bit different.
00:10:27.04 It has different units.
00:10:30.01 In general, this idea, though,
00:10:32.08 that directed motion has the. the means.
00:10:35.16 the mean-squared displacement, delta-x-squared going like time-squared,
00:10:38.28 versus diffusive motion, where it just goes like time.
00:10:42.14 you can have something in between in some cases.
00:10:44.12 So, in general we would write the square of the displacement
00:10:48.08 goes like time to some power, alpha,
00:10:50.22 which we'll call the diffusive exponent.
00:10:53.07 So, what we often do is take movies
00:10:55.23 like the one you saw here with the beads
00:10:57.21 and then plot the square of the displacement --
00:11:00.01 mean squared displacement, or MSD --
00:11:02.13 on this log plot. log-log plot,
00:11:05.19 and the slope, then, is this diffusive exponent.
00:11:08.21 So, you can see. this is actually data from a movie
00:11:11.16 just like that one here,
00:11:13.16 and you can see that the mean squared displacement
00:11:15.24 is linear as a function of time on this log-log plot,
00:11:18.10 and that says that the diffusive exponent is 1,
00:11:20.27 which is exactly what we expect for the.
00:11:22.25 you know, for the simple description of Brownian diffusion.
00:11:26.01 So, it's important to keep these ideas in mind
00:11:28.04 in what I'm going to tell you next
00:11:30.25 because what we tried to do in this system
00:11:33.21 in addressing the question
00:11:37.01 of why these droplets all don't fuse with one another
00:11:40.02 is we asked, could we introduce probe particles
00:11:43.16 into the nucleus of this cell and ask,
00:11:46.04 are they free to move around?
00:11:48.00 So, as little probes of the environment,
00:11:50.08 something like the droplets,
00:11:52.22 and asking whether they're free to move around.
00:11:55.26 This we call microrheology.
00:11:57.14 That's a term. rheology is the study of the flow and deformation behavior in materials.
00:12:03.21 Microrheology is that, just on a microscopic scale.
00:12:05.25 So, a really talented graduate student in my lab,
00:12:07.24 who's now postdoc at NIH, Marina Feric,
00:12:10.13 injected small probe particles into the nucleus.
00:12:14.14 So, these are inert little beads,
00:12:17.15 microscopic beads,
00:12:19.14 she injected into the nucleus,
00:12:21.28 and asked, are those beads free to diffuse around or not?
00:12:23.24 Are they constrained in some way?
00:12:25.26 And so, what Marina found is.
00:12:28.07 you know, in the initial studies, she said,
00:12:30.23 well, they don't actually look particularly constrained.
00:12:32.19 If I put in really small particles.
00:12:34.11 these are 0.2 micron/200 nanometer beads
00:12:37.11 injected into this nucleus, this GV, or germinal vesicle,
00:12:40.10 what she found is that the beads actually seem to be
00:12:44.08 dancing around and undergoing what really looks like Brownian motion.
00:12:47.02 So, they seem pretty free to move around.
00:12:48.28 So, that at first was surprising.
00:12:51.06 We said, well, if that's happening, then why aren't these.
00:12:53.25 you know, these liquid condensates diffusing around as well?
00:12:57.15 And we said, well, let's look at different sized beads
00:13:00.08 and ask what happens.
00:13:02.23 So, again, at the.
00:13:05.12 if I plot the mean squared displacement of these very small beads,
00:13:07.10 just like what I showed you before,
00:13:09.16 you see that there's this linear dependence on the time,
00:13:12.05 so the diffusive exponent here I would call 1.
00:13:14.17 But as we go to larger and larger bead sizes,
00:13:18.11 what you see is that exponent goes down.
00:13:20.07 So, the slope of the curve comes down and down and down.
00:13:24.22 And so, there seems to be increasing constraint
00:13:27.16 on the motion of the particles
00:13:30.04 as we get to larger and larger particle sizes.
00:13:31.26 And you can see that here in the position traces of the beads.
00:13:35.29 The small beads really kind of just undergo Brownian motion,
00:13:39.00 diffusing around.
00:13:40.14 And you get to the larger particles,
00:13:42.01 we see this intermittency, where they seem to be hopping,
00:13:43.24 maybe between pores, if you like,
00:13:45.14 and then the largest particles really are starting to be highly constrained
00:13:49.19 within the nucleus.
00:13:51.16 And so, if I plot the slope of those curves,
00:13:54.15 that diffusive exponent,
00:13:56.27 as a function of the size of the particles,
00:13:58.25 I see this very clear trail off.
00:14:03.08 So, it goes from a diffusive exponent close to 1,
00:14:05.06 close to simple diffusion, for small particles,
00:14:08.07 but for larger particles it comes down, so.
00:14:11.10 again, consistent with some kind of constraint on the motion of the particles
00:14:16.23 once they are larger than a certain size,
00:14:17.16 which, you know, is something like a couple hundred nanometers/0.2 microns.
00:14:23.01 So, we started thinking about this.
00:14:26.09 So, it looks like the particles are constrained
00:14:29.00 by some network that has a size scale
00:14:32.15 that's a few hundred nanometers.
00:14:34.20 In fact, the data that we got in this.
00:14:37.05 in this system looks a lot like some data
00:14:39.17 that other folks had seen in looking at similar bead motion
00:14:43.27 in purified networks
00:14:46.24 that are reconstituted from the protein actin
00:14:50.09 and the filaments that actin forms.
00:14:52.01 So, actin forms these beautiful filamentous networks.
00:14:54.19 And if you put particles in there, you saw really the same type of thing,
00:14:58.04 where if the particles are large compared
00:15:02.05 to the average spacing between the filaments,
00:15:05.15 then the motion is highly constrained
00:15:07.11 -- you can see that here in this in this black.
00:15:09.23 this black curve on the bottom --
00:15:11.17 and if the if the particles are much smaller
00:15:14.03 than the average mesh size of the network,
00:15:16.09 then you have something that looks like diffusive motion,
00:15:18.22 as you can see in this top curve.
00:15:23.05 We also see in these data the intermittency
00:15:25.23 of hopping between the different states.
00:15:27.24 So, we started to wonder, well,
00:15:30.11 maybe there's a cytoskeletal network like actin
00:15:34.14 that's inside the nucleus.
00:15:36.04 And that would be a little bit surprising, though,
00:15:38.29 because actin is well known in the cytoplasm
00:15:41.05 but usually not thought of as being so important structurally inside the nucleus.
00:15:45.21 And so, we started to wonder about that.
00:15:47.28 Is it possible that the motion of these beads
00:15:49.19 is constrained by some kind of a network
00:15:53.14 -- maybe it's even a nuclear actin network --
00:15:56.08 within this. within this nucleus?
00:15:58.15 And so, the picture, then,
00:16:01.02 would be that the small beads are able to kind of move around
00:16:04.02 in the interstices, but the big beads are sort of trapped within this network.
00:16:07.05 So, we tried to test that idea.
00:16:09.24 And the key experiment was then,
00:16:12.18 well, let's try to bust up the actin network
00:16:15.12 and see what happens to the motion of these beads.
00:16:17.26 And so, we did that experiment,
00:16:20.08 and if this were a live audience I might ask the audience,
00:16:23.14 you know, what you'd expect.
00:16:25.28 So, what do you expect? You can ponder for a moment.
00:16:29.04 If you've been listening to the last few slides,
00:16:31.01 then you would say, well, the diffusive exponent
00:16:33.16 should be 1 if it's simple diffusion.
00:16:35.22 So, if I. if it's an actin network
00:16:37.27 and I were to bust up the actin network,
00:16:40.07 then this curve should basically all go up to 1.
00:16:43.02 All those points should be up around a diffusive exponent of 1.
00:16:46.06 So, we did this experiment.
00:16:49.09 It turns out there's a number of ways we can disrupt
00:16:52.20 an actin network, and we tried all of them.
00:16:55.23 And what we saw very consistently was in all cases
00:17:00.03 the bead motion now exhibited a diffusive exponent
00:17:03.16 that was consistent with simple diffusion
00:17:06.03 in the absence of any elastic constraint.
00:17:09.18 So, all the beads, now,
00:17:10.21 were able to diffuse around freely within this nucleus.
00:17:14.22 So, it really does look like there's a nuclear actin network
00:17:19.07 that's forming a scaffold within this nucleus.
00:17:22.27 Now, this was all to study the mechanics and structure
00:17:27.19 using these probe particles.
00:17:29.15 The real key question at this point is,
00:17:31.08 what about the embedded RNA-protein droplets?
00:17:34.09 What about these nuclear condensates
00:17:39.16 that are sitting inside this actin network?
00:17:41.15 And by the way, this is a picture that gives you a nice visual of the network.
00:17:43.21 You might say, well, why didn't we just look at that picture
00:17:45.29 in the first place?
00:17:47.19 And the reason for that is because there's some controversy
00:17:49.23 around how one visualizes this,
00:17:52.03 and is this really representative of the network.
00:17:55.29 But this is. this is what the network looks like,
00:17:58.11 and that's consistent with all the data I just showed you.
00:18:00.21 So, we have these droplets, these nucleoli,
00:18:03.01 the red droplets that are sitting within this network,
00:18:07.06 and. you know.
00:18:10.29 and apparently constrained within the network.
00:18:13.21 So, the question that we then had was,
00:18:16.07 well, what happens to these structures
00:18:18.23 when we disrupt the actin?
00:18:20.27 So, what happens to the nucleoli?
00:18:22.28 And so, we did an experiment where we disrupted
00:18:27.05 this actin network and then looked, not at the beads,
00:18:29.27 but now at the nucleoli, which are labeled in green here.
00:18:32.03 And something I think pretty remarkable happened,
00:18:34.17 that we. that we were really fascinated by.
00:18:36.12 So, we often were looking at these samples
00:18:38.26 in the microscope,
00:18:40.26 where we're looking down and showing movies looking down.
00:18:44.20 the camera is projecting through a plane like this.
00:18:48.05 But what we decided to do was look from the side,
00:18:50.06 so we were able to make stacks of these images
00:18:53.04 and look from the side.
00:18:55.10 And this is a movie, then, looking at.
00:18:57.02 from the side at what happens
00:18:59.07 when we disrupt this actin network.
00:19:01.04 And I'll show you that it's pretty remarkable.
00:19:03.15 What happens is the nucleoli within this nucleus
00:19:08.03 all come crashing down to the bottom of the nucleus,
00:19:10.13 apparently sedimenting under gravitational forces.
00:19:15.02 So, that's really surprising, then, and interesting,
00:19:17.29 because we usually think about gravity as being negligible
00:19:20.18 within living cells.
00:19:22.23 But in this case, apparently that's not true.
00:19:25.05 So, the actin network seems to be holding these nucleoli in place,
00:19:30.03 and when we get rid of it they all come crashing down to the bottom.
00:19:32.19 So, that was genuinely surprising to us.
00:19:34.22 And so, it's also. it's also interesting.
00:19:39.11 if you. if you let the system sit for long enough,
00:19:42.12 these nucleoli, when they do come crashing down to the bottom,
00:19:44.27 they all coalesce into what I.
00:19:48.00 what may very well be the world's largest nucleolus.
00:19:50.21 This structure is now about, you know,
00:19:54.11 100+ microns in diameter.
00:19:57.12 It's larger than many individual cells.
00:20:00.03 And so, all of the nucleolar droplets have coalesced at the bottom
00:20:03.14 when we disrupted the actin network.
00:20:05.27 So, why is gravity important?
00:20:09.03 You know, why do we usually ignore gravity in cells,
00:20:11.21 but in this case it seems to be important?
00:20:14.22 So, it turns out this physics of this
00:20:18.09 is really interesting.
00:20:20.01 So, there's something called the gravitational length scale.
00:20:22.05 It goes by different names
00:20:25.00 some people refer to as the gravitational Péclet number.
00:20:27.21 It reflects the competition between gravity and the random thermal fluctuations
00:20:30.20 that want to. want to randomize the positions of these particles.
00:20:35.27 That's an entropic effect,
00:20:38.14 which you'll remember from the first lecture,
00:20:40.08 where entropy.
00:20:42.13 you know, this thermal energy scale, kT,
00:20:46.08 basically kicks everything around and wants it well mixed.
00:20:48.15 So, that would tend to have particles that would fluff up
00:20:50.15 and distribute evenly.
00:20:52.23 But gravity, if each of these particles has a mass,
00:20:54.28 wants to pull the particles down to the surface.
00:20:57.15 So, there's a competition between those two effects,
00:21:01.02 and that gives rise to this.
00:21:03.15 what I'm calling the l-sub-gravity,
00:21:06.10 the gravitational length scale.
00:21:08.02 So, kT is an energy scale
00:21:10.12 it has units of energy.
00:21:12.21 And mg, mass times the gravitational acceleration,
00:21:15.08 that's a force.
00:21:16.25 So, an energy divided by a force is a length scale.
00:21:20.07 And so, this length scale is basically the length scale
00:21:22.19 over which the concentration profile decays
00:21:26.25 as you get up higher and higher.
00:21:29.06 Now, the physics of this that I'm describing
00:21:33.09 is exactly why the atmosphere
00:21:36.14 thins out at high altitude.
00:21:38.01 So, you know, if one is on
00:21:41.27 the 3rd, 4th, or 25th story of a building,
00:21:45.00 you usually don't notice that the atmosphere is any thinner.
00:21:47.12 And so, why is that?
00:21:49.20 Well, that's because. you know, if I think about it,
00:21:52.03 if I'm in the 2nd story of a house or, you know, even a tall building,
00:21:55.10 the size of that. the height of that building
00:21:57.29 is smaller than this gravitational length scale
00:22:00.01 for something like oxygen.
00:22:02.06 So, I could put in the mass of oxygen
00:22:04.06 and these other things. you know, room temperature and so forth.
00:22:07.29 and I'll get a gravitational length scale that's something like a mile-plus.
00:22:11.05 So, I don't notice the atmosphere
00:22:14.10 thinning out at high altitude.
00:22:16.06 But I will notice if I go climb a tall mountain
00:22:18.25 or, you know, fly into La Paz, Bolivia for example.
00:22:22.01 I will notice that the atmosphere is much thinner,
00:22:25.21 and that's because now we're starting
00:22:28.19 to approach, or even exceed, potentially,
00:22:30.25 this length scale.
00:22:32.17 So, that same physics, remarkably,
00:22:34.29 is at play and relevant within these cells.
00:22:39.13 So, we think we usually ignore gravity within cells
00:22:42.17 because the idea is that cells,
00:22:45.01 in general, are smaller than this gravitational length scale
00:22:47.01 for any of the structures that we'd be considering.
00:22:49.21 The one change we would make to this equation
00:22:51.19 is instead of mass we would put a buoyant..
00:22:53.26 a buoyant mass, so some volume times the density difference.
00:22:56.25 In the. these very large frog oocytes,
00:23:01.26 what seems to be happening is the cell has gotten so large
00:23:05.11 that it's now exceeding this gravitational length scale
00:23:08.26 for these structures,
00:23:10.18 and now gravity really becomes important.
00:23:12.07 We can actually quantify this
00:23:15.22 and, with all the measurements we've done,
00:23:17.12 we can determine the gravitational length scales
00:23:21.01 and figure all this out, and make what I'll call a state diagram for this.
00:23:24.07 And what this shows is that for large cells
00:23:27.22 -- here, we're plotting it as a function of the relevant compartment,
00:23:31.21 which is the nucleus, but we could also put it as a function of the cells --
00:23:35.17 gravity starts to become important
00:23:38.16 as you get larger and larger.
00:23:40.19 And so, this is kind of, I think, a really interesting system,
00:23:43.21 and it's a place where there's some fun physics at play,
00:23:47.14 which was sort of unexpected
00:23:50.17 until we started making these observations
00:23:53.17 and this discovery.
00:23:56.09 Now, let me shift gears a little bit
00:23:59.04 and tell you about some other really interesting observations
00:24:02.02 that we made in this system.
00:24:06.01 Now, the graduate student, Marina Feric,
00:24:09.02 that I mentioned earlier.
00:24:10.17 in the course of these experiments where she was disrupting the actin network
00:24:13.01 and looking at how the nucleoli all coalesce with one another,
00:24:17.06 you know, forming this kind of world-record-sized nucleolus,
00:24:21.12 what she noticed is that the nucleoli coalesce in a very interesting way.
00:24:27.09 if we start to label. fluorescently label
00:24:30.22 the different sub-parts of the nucleolus.
00:24:34.02 So, this is a movie that I'll show you
00:24:37.06 where we've labeled one set of proteins in green
00:24:39.03 -- that's this protein fibrillarin,
00:24:41.07 which is a protein associated with the inner core of the nucleolus --
00:24:43.29 and another set of proteins.
00:24:47.03 I'm labeling here nucleophosmin,
00:24:49.23 which is associated with this outer layer of nucleolar proteins.
00:24:53.29 And what Marina noticed is that when she disrupts the actin network
00:24:58.17 and lets these nucleoli all coalesce with one another,
00:25:00.17 they indeed form droplets that get larger and larger,
00:25:03.10 but the cores, as well,
00:25:08.23 this fibrillarin-rich core within the nucleolus,
00:25:10.22 also coalesce with one another.
00:25:13.27 So, it's as if we have a liquid within a liquid.
00:25:16.09 And you can see that even better in these images.
00:25:19.03 These are higher-resolution images within.
00:25:22.14 within the cell after we've left them all coalesce.
00:25:25.29 You really have this sense that the fibrillarin-rich proteins
00:25:30.09 form a core, a liquid within the nucleophosmin-rich outer liquid.
00:25:36.10 And so, this is kind of a really interesting idea,
00:25:39.23 and we started to think about what it means
00:25:41.25 and how to understand it
00:25:45.13 from a molecular biophysical perspective.
00:25:46.29 Now, some other work we'd been doing.
00:25:51.24 we had been able to show that when we take fibrillarin,
00:25:55.09 this core-associated protein, and purify it,
00:25:57.27 it phase separates into these beautiful liquid droplets.
00:26:00.21 And that may be not so surprising
00:26:03.00 based on what I told you in the last lecture,
00:26:04.25 because fibrillarin has significant stretch.
00:26:11.03 which is a conformational heterogeneous. this intrinsically disordered regions,
00:26:15.01 which we think are really driving phase separation.
00:26:17.09 And so, we'd shown that it forms these nice liquid droplets.
00:26:19.07 What about the other proteins?
00:26:21.15 So, for that work we collaborated
00:26:23.23 with Richard Kriwacki and Diana Mitrea at St. Jude,
00:26:26.23 who had been working with nucleophosmin in vitro,
00:26:30.09 with a purified system.
00:26:32.04 And they had shown that nucleophosmin phase separates
00:26:34.14 into these nice liquid droplets in vitro.
00:26:36.06 And so, we said, hey, let's just do
00:26:39.02 what is a simple and obvious experiment,
00:26:41.09 and take these two samples and mix them
00:26:44.10 and ask what happens.
00:26:46.11 And when we did that, quite remarkably I think,
00:26:48.23 what we found is that the fibrillarin-rich droplets
00:26:50.26 and the nucleophosmin-rich droplets
00:26:53.04 are relatively immiscible with one another --
00:26:54.28 they do not mix with one another.
00:26:56.23 And instead we get this.
00:26:58.25 what I would call a three-phase system.
00:27:00.13 There's a. the dark region here
00:27:02.22 would be a low concentration phase of.
00:27:04.15 it has some nucleophosmin and some fibrillarin.
00:27:06.00 The green punctae inside here
00:27:08.14 are a fibrillarin-rich liquid phase,
00:27:11.08 and the red or sort of orange-colored outer phase
00:27:14.02 is a nucleophosmin-rich phase.
00:27:17.05 So, this is really, I think, quite remarkable,
00:27:19.27 and it's remarkable for a few reasons.
00:27:21.14 One of them is that this purified protein system
00:27:25.03 essentially completely recapitulates
00:27:28.02 the in vivo organization of these structures.
00:27:32.22 So, in vivo, recall that we have fibrillarin-rich liquid-like droplets
00:27:37.18 within a nucleophosmin-rich outer layer,
00:27:41.07 and that's exactly what the in vitro purified system
00:27:43.20 looks like.
00:27:45.11 So, we thought that was. that was pretty incredible.
00:27:47.09 With a very simple system of a small number of protein
00:27:51.10 and RNA components,
00:27:53.20 we've been able to recapitulate this core shell architecture
00:27:56.07 of the nucleolus.
00:27:59.29 Now, the idea of liquid immiscibility
00:28:03.21 -- having multiple immiscible liquid phases
00:28:06.15 or non-mixable liquid phases --
00:28:08.18 is itself something that is well known in chemical engineering
00:28:12.05 and physical chemistry and soft matter communities.
00:28:15.15 This is just an example of some immiscible liquids
00:28:19.00 that one can form. can form from, you know,
00:28:22.03 nonliving organic solvents and water and oils,
00:28:24.18 where you can get these things
00:28:28.05 that do not mix with one another.
00:28:31.04 So, we can have phase separation, then,
00:28:33.11 not just of forming two phases, but actually forming many phases,
00:28:39.15 and the way in which those phases interact is quite interesting
00:28:41.14 and sort of a rich set of physics
00:28:43.26 that is yet to be completely understood.
00:28:46.23 One of the questions that we asked in doing these experiments is,
00:28:49.12 why would fibrillarin droplets be on the inside?
00:28:51.19 So, in other words,
00:28:53.23 why is the green inside the red?
00:28:55.16 Why isn't the red inside the green, for example?
00:28:58.12 That would. that would be seemingly just as valid.
00:29:03.11 You'd have immiscibility
00:29:05.19 they're not mixing with one another.
00:29:07.29 From the biology perspective,
00:29:10.14 that would probably be a problem,
00:29:13.19 because the processes that are associated with fibrillarin
00:29:16.25 and the fibrillarin-dense. condensed state take place first,
00:29:21.26 and the idea is that those RNA transcripts
00:29:24.17 are processed in a sequential fashion inside to outside.
00:29:27.16 So, if the order weren't right, then that would probably cause
00:29:31.27 problems for the biology.
00:29:34.11 So, why is it that fibrillarin knows
00:29:37.06 that it's supposed to be on the inside,
00:29:39.03 and nucleophosmin and those components that
00:29:41.17 know that they're supposed to be on the outside?
00:29:43.25 Well, it turns out that a key to answering that question
00:29:45.28 comes from surface tension.
00:29:47.28 So, as the name implies,
00:29:50.23 surface tension is a kind of tension associated with surfaces
00:29:54.26 or interfaces between two different types of phases.
00:29:56.29 In particular, it's really.
00:29:59.10 it's an energy cost associated with having an interface,
00:30:02.05 and it actually has units of energy per unit area.
00:30:07.02 You can think about this in a simple.
00:30:09.14 the simple schematic as reflecting the fact that
00:30:14.23 molecules within a bulk phase are sort of experiencing
00:30:17.06 a homogeneous environment surrounded by.
00:30:20.20 by, you know, a particular set of molecules,
00:30:25.21 but at the interface there's this kind of.
00:30:27.18 you know, on one side, they're seeing the homogeneous phase
00:30:29.27 on the other side, they're seeing this external environment.
00:30:32.14 And that sets up an energy cost.
00:30:35.10 The manifestation of surface tension
00:30:38.22 is something that we're all familiar with
00:30:41.03 by looking at things like water-walking bugs
00:30:43.28 or the fact that we can, in some cases,
00:30:46.06 get paper clips to actually float on water
00:30:49.04 because of these surface tension effects,
00:30:51.15 or if you were to wash your car and put some wax on the car,
00:30:55.11 for example, and then watch,
00:30:58.19 the water droplets would bead up on that surface.
00:31:01.01 Those are all surface tension effects --
00:31:03.06 the energy of interfaces between different phases.
00:31:06.05 And it turns out that surface tension
00:31:08.17 is really well known to be key in structuring
00:31:13.13 multi-phase liquids.
00:31:15.00 So, in particular, for multi-phase systems
00:31:18.11 from nonliving matter,
00:31:20.03 we know that that to have this kind of a core shell organization
00:31:25.17 of a three-phase system,
00:31:27.22 the surface tension between phase 3,
00:31:29.27 this green phase,
00:31:32.05 and phase 1, the black phase,
00:31:34.21 would have to be large. So, if that surface tension,
00:31:36.15 the energy associated with that interface, is very large,
00:31:39.21 then the system can minimize the energy
00:31:43.15 by making that interface go away --
00:31:45.15 in other words, by having the green phase within the red phase.
00:31:49.17 And then there's no longer any interface between, you know, 1 and 3.
00:31:53.08 You can do a very simple experiment.
00:31:56.24 So, Marina actually did this really simple sort of experiment
00:31:58.25 you can do in your kitchen
00:32:01.06 by taking water, Crisco oil, and silicone oil
00:32:03.25 and showing that because the water-silicone oil interface.
00:32:06.28 the tension. the surface tension is large,
00:32:09.24 larger than that between the water and the Crisco oil,
00:32:12.18 the red stuff,
00:32:14.25 then the silicone oil gets embedded within the Crisco oil.
00:32:18.05 So, we had a prediction that the core shell architecture,
00:32:21.27 the core shell organization we see within the nucleolus,
00:32:24.26 was reflecting these differential surface tensions.
00:32:27.24 To try to test that idea,
00:32:31.07 we took the purified proteins that form these liquids
00:32:33.10 and asked, how do they interact with surfaces
00:32:36.06 of different hydrophobicity?,
00:32:38.00 os that we could try to understand the relative strength
00:32:41.02 of their interactions,
00:32:45.22 the favorability for interactions with water versus with oil.
00:32:49.04 And so, we took a surface that's relatively hydrophobic --
00:32:51.21 this is not a strongly hydrophobic surface, but relatively hydrophobic
00:32:52.19 -- and we looked from the side,
00:32:56.00 again kind of visualizing from the side,
00:32:58.00 at how these droplets interact with the surface.
00:33:00.10 What we found is that the nucleophosmin droplets
00:33:03.20 tend to behave like water,
00:33:08.10 in that they tend to bead up on these relatively hydrophobic surfaces,
00:33:10.27 where the fibrillarin-rich droplets tend to better wet
00:33:14.14 these hydrophobic surfaces,
00:33:16.26 and that's all consistent with the idea
00:33:18.26 that the surface tension between fibrillarin,
00:33:21.02 the green stuff,
00:33:23.12 and water is larger than that between nucleophosmin,
00:33:26.05 the red stuff, and water.
00:33:29.01 And that's why the green droplets are embedded within the red,
00:33:34.00 just like you see in this schematic here.
00:33:36.02 And so, this idea of surface tension and surface tension structuring
00:33:39.08 of multi-phase liquids has implications
00:33:42.23 I think much beyond the nucleolus,
00:33:44.26 which is the system, you know,
00:33:46.22 we've been studying to elucidate these ideas.
00:33:48.25 In particular, if the interface
00:33:52.00 and the surface tension between two different phases..
00:33:55.01 . let's say between phase 3 and phase 2,
00:33:57.23 the red and green.
00:34:00.03 if that's really energetically costly --
00:34:02.24 in other words, anytime red is next to green, there's a huge energy cost --
00:34:05.05 then what happens is these two types of liquids
00:34:08.05 just will never interact.
00:34:11.06 They won't touch each other at all.
00:34:12.28 And so, that's probably why many of the condensates in cells
00:34:16.11 actually don't interact at all.
00:34:17.16 They're sort of not sticking to one another, not wetting one another,
00:34:19.28 not engulfing one another and so forth.
00:34:21.23 In other cases, though, there can be partial wetting,
00:34:25.06 where the surface tensions are roughly of the same magnitude,
00:34:30.01 and then you can have droplets
00:34:32.08 that are not fully engulfing one another,
00:34:34.08 but they're interacting,
00:34:36.25 and you can have morphologies that look like this.
00:34:38.14 We think this is important because
00:34:41.02 there are many structures in the cell
00:34:43.13 where there's partial interaction between these condensates,
00:34:46.07 for example in this beautiful micrograph from Joe Gall
00:34:49.08 showing Cajal bodies and Snurposomes and the sort of partial wetting
00:34:51.29 -- these kind of well-defined contact angles and so forth --
00:34:56.00 that one can start to use to measure the surface tensions between these droplets,
00:35:01.17 and between the droplets and the surrounding nucleoplasm, in this case.
00:35:06.22 So, in this talk, I've tried to convey to you some of the excitement
00:35:10.14 about the nucleolus, and more generally
00:35:13.11 the idea that these condensed states of biomolecular matter
00:35:17.20 are important within the nucleus.
00:35:20.28 The nucleolus is just one type of these structures.
00:35:24.15 We focused on it quite a bit
00:35:27.13 because it's the largest of the nuclear bodies,
00:35:30.00 but there are really dozens of different types of these things inside of the nucleus.
00:35:36.25 And in fact, we think that the nucleolus probably
00:35:38.28 is a hypertrophied example of the kind of phase separation
00:35:45.01 and the way in which the condensates that form are impacting the genome,
00:35:47.10 and potentially playing a really important role
00:35:49.27 in the flow of genetic information.
00:35:52.07 We can think about these condensate as something
00:35:55.21 like the condensation of water on this.
00:35:58.22 on this spider web, that can deform the spider web and.
00:36:04.14 and actually restructure it in some way.
00:36:06.11 And so, those are the kinds of questions that we're starting to address,
00:36:09.00 and there's a lot of excitement in this field
00:36:12.00 in thinking about these ideas moving forward.
00:36:14.22 So, how do liquid condensates impact genome architecture and activity?
00:36:19.00 So, how might they be playing a key role
00:36:21.18 in regulating the expression of genes
00:36:23.15 that give rise to the traits of an organism?
00:36:26.16 The, you know, hair color and eye color, height, weight,
00:36:30.27 number of arms and legs and toes,
00:36:33.21 and all those sorts of things that really make biology
00:36:38.28 so interesting, that we can go from genes to traits.
00:36:42.12 How. do these transitions.
00:36:44.26 so the. you know, I've talked a lot about the.
00:36:46.29 in the last talk as well. about the transitions between liquid states
00:36:50.05 and solid states,
00:36:52.01 these liquids and gels and amyloids.
00:36:53.26 Do transitions between those different states of biomolecular matter
00:36:56.27 play a role in gene regulation?
00:36:59.18 So, it's really a fascinating question.
00:37:01.17 We have very little knowledge about that,
00:37:03.05 and I think that's going to be one of the key questions
00:37:05.11 going forward in this area,
00:37:09.26 at the interface of soft matter physics, c
00:37:12.22 ell biology, genetics,
00:37:16.14 and probably a number of other fields.
00:37:20.14 So, there's a lot of excitement right now,
00:37:22.22 and we're gonna be happy to see this move forward.
00:37:25.29 I want to thank you for your attention.
00:37:28.07 It's been really a pleasure to work with some of the folks.
00:37:33.04 I mentioned Marina Feric, a former grad student.
00:37:35.11 We've got a really wonderful set of people in the lab,
00:37:37.01 and I've been fortunate to interact w
00:37:39.27 ith a lot of really brilliant collaborators.
00:37:42.14 I'm very grateful for that and the funding
00:37:45.09 that has helped make some of the work we're doing in our lab possible.
00:37:48.10 Thank you so much for your attention.


NSR: Many of your designed proteins are formed of alpha helixes. Why?

Baker: I think we started with alpha helixes because the interactions stabilizing the helixes are between amino acids closer in the primary sequence. It is more complicated to figure out how beta strands come together to form beta sheets.

But now our understanding of the principles for beta-sheet proteins has improved a lot, and it is no longer a problem. We are now routinely making proteins out of only beta sheets. For example, we designed beta-barrel proteins that go through the membrane and let ions through.

NSR: One of your designed barrels shows selectivity for potassium-ion transfer. Is it designed to be so?

Baker: No. We were just trying to design proteins with pores of different sizes and we were surprised that it was selective for potassium. People have been arguing about how potassium channels work and our work showed that maybe it is just the right size for potassium ions with the water molecules being stripped off. So that is an example of how protein design can help to better understand how natural proteins work.

NSR: You designed a protein rotator assembled with two independently designed parts. So is protein design becoming modular?

Baker: That's right. We are making a common set of building blocks that can be easily put together in different ways for de novo protein design.

NSR: DNA/RNA design is also developing fast. Will protein design combine with the design of DNA/RNA or other materials?

Baker: Yes, it will be very interesting to combine protein design with DNA origami. We are also designing proteins interacting with inorganic compounds. For example, we are trying to control inorganic crystal growth with proteins. This may be helpful for applications including semiconductor production, etc.

NSR: Are your designs beginning to be applied in translational science?

Baker: Yes. We are starting several companies to commercialize these protein designs. There is a company called Neoleukin Therapeutics, which is trying to commercialize the designed protein mimicking interleukin-2 for cancer treatment. We also have a company that is developing vaccines for various types of viruses.

NSR: What are the proteins you are working on now?

Baker: We have quite a few interesting directions. We are working on transmembrane pores, which may have many applications such as filtering and nanopore DNA sequencing. We are also trying to design molecular machines, new types of drug molecules and different types of materials. In addition, we want to design proteins that are able to regulate cell behaviors such as cell growth, cell differentiation and cell fate.

NSR: To many people, work from your lab seems like magic.

Baker: Sometimes it seems like magic to me, too. I like doing things that seem like magic. But I should point out some kind of dishonest or misleading impression here. We show them the successful magic-like designs but seldom talk about all the steps and the things that did not work.

We show them the successful magic-like designs but seldom talk about all the steps and the things that did not work.

—David Baker

Why is the mRNA technology of the Pfizer/BioNTech Covid-19 vaccine such a big deal? And what kind of role can bioinformaticians play in future mRNA vaccine developments?

Hello, I'm a math/CS person who's recently been interested in bioinformatics and I'm curious to know why the recent development of the Pfizer / BioNTech vaccine is considered a big deal in terms of the mRNA tech being used. Obvsiously, I understand the importance of a vaccine for a pandemic, but why is the mRNA part such a breakthrough?

I had trouble understanding the basics but I just watched a video that I think explains it, could someone tell me if the following explanation is accurate?

There is a protein that is part of the virus that we would like our antibodies to target.

The vaccine introduces mRNA into our bodies that codes for that protein.

Our body's ribosomes read off the mRNA and create that protein as they would for any other mRNA.

The body's immune response then targets this foreign protein.

When live virus is introduced, the body's antibody response knows how to target this foreign protein.

It is less dangerous than attenuated virus because the full virus is never created -- just the single protein.

It also doesn't require any sort of genetic manipulation because there is no transcription or reverse transcription, mRNA is introduced as already transcribed.

Yes, that's a great explanation.

[disclaimer: I don't make vaccines, I'm a theoretical geneticist I work in an immunology lab where vaccines are tested]

The bits not mentioned here are the challenges that need to be overcome before it works:

Transporting the mRNA from production facilities (possibly something like yeast fermentation vats) to people without degrading it

getting the mRNA into the cell in a consistent, reliable fashion

getting the ribosomes to latch onto the mRNA in a consistent, reliable fashion (i.e. making sure all the right initiating complexes bind in all the right places)

These aren't particularly hard problems when compared to the challenges of other vaccine production, but it's a new technology, so it may take a while for a standardised process to be developed. I consider things like the -80°C storage requirement to be teething problems we'll work out a better way in the near future.

If the body's immune system is capable of recognizing the virus based on its spike proteins, why would we need to synthetically create them?

The body would recognize, combat, and destroy the virus in any case.

If the body isn't capable of doing that, what good would these synthetic spike proteins do?

Just theoretically, is it known what will happen to the foreign protein and how much of it will be translated? I'm assuming it will be ubiqinated and destroyed but is there any chance that in the absence of the other viral proteins, it may start to accumulate in cells and form plaques or aggregates?

Can u give us a link to that video?

Apparently it strips u of ur conscious thus leaving the human body just a “living entity “

A few details are missing though. mRNA cannot penetrate a cell membrane. You need a vector virus(or other method) to get it in the cell. This is still somewhat experimental. We have our first world wide medical experiment! This is not in line with medical ethics. You don't use the world population as a guinea pig for an idea that's been around for decades.

I’m not an immunology expert but to my understanding it’s because it can be created in a lab setting using a DNA template instead of being grown in a host mammalian cell, the mRNA can generate a more reliable immune response, and because it’s not an “inactive” part of a dangerous virus - there’s no danger to it

Great answer. It’s also a big deal I believe because the question of how to get mRNA across a plasma membrane has been stumping biochemists for years. It will be interesting to see how Pfizer accomplished it with this vaccine. Someone please correct me if I’m wrong, but I don’t believe Pfizer has released the details on how they accomplished this.

The TLDR metaphor for this kick-ass non-technical summary of how the mRNA Pfizer/BioNTech vaccine works is: the vaccine helps your body develop it’s own equipment to train to fight and defeat real coronavirus if you get infected. The vaccine is Mr Miyagi [paraphrase]

Here's my attempt at explaining how much of a game changer this technology is:

These mRNA vaccines are an amazing technology [paywalled, but source-code readable] for vaccine production: really fast production ramping capability, simple / cheap production requirements, and directly tapping into the things that the viruses already do (so it’s more likely to elicit a specific and effective response).

However, in addition to all that amazing stuff, a demonstration of an effective mRNA-introduction-into-cell-produces-protein is in itself going to be a huge game-changer across medicine.

I hear it occasionally mentioned that of the hundreds of thousands of proteins (and modifications) our cells can create, only a small number are actual druggable targets (a few hundred, if my memory serves me correctly and wikipedia seems to agree). I first remember hearing about this little fact while attending a public talk at the University of Oxford, and the statement blew my mind. All of our medical practise of prescribing drugs to people in an attempt to cure them of particular ailments is built on modifying a few hundred proteins. These are the proteins that we have a known mechanism for, and we can create a synthetic drug that sufficiently mimics the protein (or things it interacts with) in order to alter the function of that protein.

But with mRNA, we can make proteins. If we can introduce the mRNA into the right part of the cell, and get it to connect in with the right proteins, DNA, RNA, and other metabolites (a big if, but the vaccine success demonstrates it works at least in some situations), then the cells will do the rest. They’re already designed to do the rest.

That means, potentially, no more need for insulin, where simple insulin deficiency is the only issue. Feed the relevant cells with the right mRNA recipe for making insulin, and they’ll make it, with no / fewer issues about rejection from forein contaminants.

I mention insulin because it’s small. The mRNA recipe for it will be small as well, which should make it more likely to be successful.

Are cancer cells spreading because they’re not producing a particular protein that allows them to be targeted for destruction? Feed them some mRNA to kick those protein production factories into action, so that the immune system can once again deal with the cancer.

All those non-druggable proteins now become potentially copyable proteins. If a protein is manufactured by the body (or the body of other people) in sufficiently large quantities, we can isolate it, fragment it, and sequence it. We can then fish through the DNA to find the sequence(s) that encode the protein. We can identify the RNA sequences that dock into the ribosomes and create the proteins. And given those sequences, we can construct synthetic plasmids in yeast and bacteria that generate billions of copies of the target RNA molecules.

As technology progresses, the effort required to sequence proteins will go down. Maybe we’ll be able to reduce the amount of protein required to isolate and sequence. Maybe we’ll work out a generic way to sequence all proteins directly without fragmentation.

I expect there will still be non-copyable proteins ones that need the mRNA at a particular pH, embedded in a particular membrane, with just the right concentration of other molecules. An exquisitely balanced system would be required so that the protein is folded in just the right way to create the desired function.

… but for everything else, there’s in-situ protein fabrication.

My prediction is that this concept is not going to get legs in the usual pharmaceutical circles precisely because it’s tapping into the things that our body already does so well. As such, getting monopolised money from this technology will be challenging, and it will displace a lot of existing profitable medicine. Patents might come out testing whether we can create a protein with mRNA that’s close to something already created, but just different enough that it passes the “non-natural” threshold. I could imagine legal challenges even for that. Is an RNA sequence patentable? Given that the general method is obvious (to me), wherein lies the novelty? And trade secrets for protein production probably won’t last long, especially if we can get sequencing down to a single-molecule level.

Watch the video: Η σημασία του πρωινού στη ζωή μας! Efi Koloverou Dietitian (July 2022).


  1. Tighearnach

    no words! just wow! ..

  2. Boniface

    I confirm. I agree with all of the above. We can communicate on this theme. Here or at PM.

  3. Curtiss

    How should you assess your question?

  4. Mezilkis

    I apologize, but this option was not suitable for me.

Write a message