Wholesale biochemical re-engineering - Not too far off now

[jseah]

This is an incredibly recent paper that got published in Nature Chemical Biology. While I understand many people will not have access to scientific articles, you should be able to read the abstract, not that it will help you if you aren't familiar with biochemistry though...

http://www.nature.com/nchembio/journal/ ... o.921.html

The layman's summary is this:
This research group just managed to get bacteria to put human-like sugars on their proteins (the process is called glycosylation). This is harder than it sounds since bacteria don't do this at all, except for a few rare strains which have a simple version.

The biologists among us may feel free to squee now.

Essentially, they made bacteria do an incredibly large and complex thing that eukaryotes do. Golden rice was an achievement (they moved vitamin A synthesis into rice), this... this is insane. XD
While it does not have any immediate practical use, if they manage to optimize the system and make it more human-like, which could take anywhere up to 20-30 years, antibody medicines could see their production price drop by 20 to 100 fold.
Plus, things we could only make in human cell culture could be done in bacteria. That is a speed up by at least three times.

[lutzj]

This is super cool. I wonder about the day we can just dream up and manufacture new proteins.

[CorruptUser]

So... how long until I can do shots of pure adrenaline? Because I make my espresso using Redbull instead of water, and it just doesn't give me that kick anymore.

[Iulus Cofield]

If you really wanted to you could probably achieve that dream today if you had a corrupt friend in the right place and a few thousand dollars.

Then again a shot is about 45 milliliters and I'm fairly sure the usual highest dose is 1 milliliter, so, unless you've consumed a huge amount of beta blockers beforehand, I'm not sure you'll ever be able to take a whole shot at once without dying.

[lutzj]

If you really wanted to you could probably achieve that dream today if you had a corrupt friend in the right place and a few thousand dollars.

Then again a shot is about 45 milliliters and I'm fairly sure the usual highest dose is 1 milliliter, so, unless you've consumed a huge amount of beta blockers beforehand, I'm not sure you'll ever be able to take a whole shot at once without dying.

On the other hand, you'll probably see a very diminished effect if you take it orally.

[sourmìlk]

Do they cancel each other out like that? For example, one time Homer Simpson took a whole bottle of caffeine pills by accident, and then tried to counter it with a whole bottle of sleeping pills, and I'm pretty sure that wouldn't have any effect other than killing you.

[Iulus Cofield]

Do they cancel each other out like that? For example, one time Homer Simpson took a whole bottle of caffeine pills by accident, and then tried to counter it with a whole bottle of sleeping pills, and I'm pretty sure that wouldn't have any effect other than killing you.

Beta blockers work by preventing the body from using adrenaline. Sleeping pills and caffeine don't work against each, AFAIK.

On the other hand, you'll probably see a very diminished effect if you take it orally.

Derp. I should put a nickle in a jar every time I forget the most obvious thing in a situation, so that I can vacation in Hawaii before the end of the year.

[sourmìlk]

Do they cancel each other out like that? For example, one time Homer Simpson took a whole bottle of caffeine pills by accident, and then tried to counter it with a whole bottle of sleeping pills, and I'm pretty sure that wouldn't have any effect other than killing you.

Beta blockers work by preventing the body from using adrenaline. Sleeping pills and caffeine don't work against each, AFAIK.

If beta blockers only have the effect of preventing the body from using adrenaline, then sure, they cancel out. Do beta blockers have no effects that aren't countered by increased adrenaline?

[jseah]

Adrenaline isn't a protein. Now, the enzymes that make adrenaline...

That's the beauty of it. Each advance opens new opportunities for more advances.

Enzymes that don't work in bacteria due to a lack of sugar chains can be made to work. This lets us produce more things with bacteria, which is much cheaper than GM-cow serum or GM-plants.

So yes, not just antibodies. Anything that is blocked by the lack of sugar chains in bacteria, could potentially be unblocked. (I don't say all because nature has a way of screwing with you)

[CorruptUser]

I don't say all because nature has a way of screwing with you

Considering that civilization is basically Humanity saying "screw you, nature!", I say let the screwing begin!

[Izawwlgood]

Maybe I'm missing something here, but it seems like their accomplishment was inducing an artificial pathway to glycosylate certain targets. That's pretty cool; it's easy to get bacteria to synthesize a protein of interest; it's a lot harder to induce bacteria to include an entire biochemical pathway into their metabolism.

That said, I'm not seeing how this means we're nearing designer proteins.

While it does not have any immediate practical use, if they manage to optimize the system and make it more human-like, which could take anywhere up to 20-30 years, antibody medicines could see their production price drop by 20 to 100 fold.
Plus, things we could only make in human cell culture could be done in bacteria. That is a speed up by at least three times.

How do you figure this is going to facilitate antibody medicines? Antibodies aren't glycosylated to any extent that I'm aware of? And what do you think we do in human cell culture that could be transitioned to bacteria? Insulin, for example, is presently largely produced in yeast.

I'm not saying it's not cool, I'm just curious to hear what you think it's applications are.

[sourmìlk]

I don't like the idea of making bacteria manufacture proteins. It's too complex a system: there are things in the bacteria completely unrelated to manufacturing the proteins we want manufactured. Is there a simpler way to do it that doesn't involve unnecessary mechanisms?

[CorruptUser]

I don't like the idea of making bacteria manufacture proteins. It's too complex a system: there are things in the bacteria completely unrelated to manufacturing the proteins we want manufactured. Is there a simpler way to do it that doesn't involve unnecessary mechanisms?

AFAIK, when people develop ways to synthetically create relatively simple chemicals, it's worth a Nobel prize or so. So for something like protein chains, the simplest solution for large-scale product is often to harvest them from something that produces it. So the chemists figure out what they want, and the biologists figure out how to grow it.

If it helps you sleep at night, think of the modified bacteria as man-made microscopic robots that replicate themselves from materials in their surrounding environment, materials including those that comprise the human body, and are invisible to the naked eye so they could be anywhere, including under your fingernails right now. Pleasant dreams!

[poxic]

If it helps you sleep at night, think of the modified bacteria as man-made microscopic robots that replicate themselves from materials in their surrounding environment, materials including those that comprise the human body, and are invisible to the naked eye so they could be anywhere, including under your fingernails right now. Pleasant dreams!

+1 Would freak again.

[Izawwlgood]

I don't like the idea of making bacteria manufacture proteins. It's too complex a system: there are things in the bacteria completely unrelated to manufacturing the proteins we want manufactured. Is there a simpler way to do it that doesn't involve unnecessary mechanisms?

What is your contention with this? Bacteria are basically the most simplified system for producing these proteins effectively. There's a reason PCR is done with polymerase, not, say, some chemical compound that affixes the next base in a chain... Oh wait... that's what it does!

I'm trying to think of analogy that might make sense for you... You're a programmer right? It's akin to me asking why you trust a compiler to run your code, instead of relying on a pencil and paper to do it yourself.

[sourmìlk]

I am approaching this from a programming perspective, yeah. So, a compiler is specifically designed just to compile code and do nothing else. Bacteria do tons of shit, part of which is making the proteins we want. Is it not possible to make something that does nothing but make the proteins we want? That would be less complex.

[Izawwlgood]

Yes, but it's also easy to have them, say, attach an extra part to the end of the protein you want them to make (called a tag), and then purify protein from a culture based on that tag. You end up with the protein you want, with extremely high purity, and can then remove the tag with a chemical agent.

Is it not possible to make something that does nothing but make the proteins we want? That would be less complex.

No. Bacteria are basically doing nothing but growing and making the proteins we want. It really doesn't get much less complex than that. What you're asking for is effectively 'throw a bunch of amino acids into a tube, and get your protein'. The way that works in reality is 'give the directions to make a protein to a bacteria, and it makes it for you'. Because the 'get your protein' and 'make a protein' step is actually fairly complicated, and not surprisingly, life does it best.

The compiler analogy works because 'a compiler can interpret code in all kinds of ways!', just like you're worrying that a bacteria can do all kinds of things to a protein. Both the compiler, and the bacteria, are selected for the task because they do exactly what you want to them to do the way you want them to do it, and are the easiest way to do it.

[sourmìlk]

Okay, I was hoping we could find a way to do it better.

When you say "Bacteria are basically doing nothing but growing and making the proteins we want", I don't like the growing bit because it adds unnecessary complexity to a system that we just care about part of. It's like if a computer also made mint jelly, or if a universal translator made glow-in-the-dark noses. I'd want the mint jelly / glow-in-the-dark nose mechanisms removed. If we can't do that, then okay, we can't do that.

My problem isn't actually that a bacteria can do all kinds of things to a protein, it's that it does things that aren't doing things to that protein.

[lutzj]

My problem isn't actually that a bacteria can do all kinds of things to a protein, it's that it does things that aren't doing things to that protein.

Yeah, but those "other functions" are almost entirely reproduction and self-repair. Engineers would kill for more machines like bacteria.

[sourmìlk]

Sure, extra features sound nice until you realize it means more complexity.

[lutzj]

Sure, extra features sound nice until you realize it means more complexity.

In this case, the extra features mean much less complexity, since we don't have to worry about those functions.

[sourmìlk]

So we can tell bacteria to make any arbitrary protein without worrying about the effect on the bacteria?

[Angua]

Generally, yes. Seriously though - look up how chaperone proteins function and help the new proteins fold into the right shape, plus all the post-translational modification (it's not just glycosylation) that goes on. Bacteria are easy to grow, easy to put genes into them, and it's easy to purify your desired protein out.

Anyway, this is good as we can get some mammalian glycosylation and hopefully get even better results than what we're making at the moment (if you need glycosylation at the moment, you make things in yeast and hope for the best).

[sourmìlk]

Generally, yes. Seriously though - look up how chaperone proteins function and help the new proteins fold into the right shape, plus all the post-translational modification (it's not just glycosylation) that goes on. Bacteria are easy to grow, easy to put genes into them, and it's easy to purify your desired protein out.

That's quite cool. I suppose if the extra complexity of the bacteria never has an effect on the protein synthesizing then the complexity doesn't matter (although it still makes me uncomfortable on principle.)

[Angua]

I suppose you're waiting for the day where all food is made from their starting elements in a factory, as that would be less complicated too. Digging things out of the ground? Having trees where anything could be falling on them? Preposterous! Having to peel, and cook food!

Also, never eat cashews - even the 'raw' ones have been roasted, otherwise they'd be extremely nasty to eat (this is assuming you buy them in most places, however at home there was a time when they were selling truly raw cashews, as only an idiot would eat them without roasting them first because everyone knows you can't do that, until it was pointed at that tourists don't count as 'everyone').

[Angua]

So... how long until I can do shots of pure adrenaline? Because I make my espresso using Redbull instead of water, and it just doesn't give me that kick anymore.

Adrenaline = epinephrine. It's not that hard to find epinephrine.

[sourmìlk]

I suppose you're waiting for the day where all food is made from their starting elements in a factory, as that would be less complicated too. Digging things out of the ground? Having trees where anything could be falling on them? Preposterous! Having to peel, and cook food!

Actually yes, I would very much like this. It is less complicated (in terms of the number of moving parts), it's cleaner, and (with meat) it's less ethically ambiguous.

[Angua]

You have a very strange idea of less complicated.

[sourmìlk]

I want to say that it's a perfectly normal idea among programmers, but I'm probably just nuts.

[Ulc]

I want to say that it's a perfectly normal idea among programmers, but I'm probably just nuts.

Yep, just nuts.

What you're talking about here, is in programmers terms the equivalent of saying "man, computers are too complicated, can't we just run calculations without circuits?". Yes, it might involve more moving parts than a "pick up individual atoms and assemble" method out of your dreams - but newsflash here, we can't pick up single atoms, we're actually limited how the universe works.

And one of the things about proteins are that they are horrible complicated - we're still decades away from a proper understanding of how proteins actually basically work, the notion that dynamics (any other biochemists around, please, lets not get into a religious war on this subject!) are functionally important something that has only accepted in the last decade!

[sourmìlk]

I'm fine with circuits because the only purpose they serve in a computer is to compute. Bacteria as protein manufacturers sort of bother me because there are mechanisms there not related to protein manufacturing.

Also: can we predict, given a chain of amino acids, into what shape it will form?

[semicharmed]

I'm fine with circuits because the only purpose they serve in a computer is to compute. Bacteria as protein manufacturers sort of bother me because there are mechanisms there not related to protein manufacturing.

Also: can we predict, given a chain of amino acids, into what shape it will form?

First point: bacterial cells are basically self-replicating protein factories; 3.5 billions years of evolution has made them astonishingly efficient. I couldn't find hard numbers, but the bottom of this page has some hypotheticals.

You're also looking at your computer anology wrong. Sure, the circuits in a computer are there to computer, but they're capable of much more computing than just compiling a program. If you're equating protein synthesis == compiling code, then there's a lot less 'extra' in a bacterial cell than in a computer.

Second point: not really. The science behind folding@home is a good overview; like ulc said

[sourmìlk]

I'm fine with circuits because the only purpose they serve in a computer is to compute. Bacteria as protein manufacturers sort of bother me because there are mechanisms there not related to protein manufacturing.

Also: can we predict, given a chain of amino acids, into what shape it will form?

First point: bacterial cells are basically self-replicating protein factories; 3.5 billions years of evolution has made them astonishingly efficient. I couldn't find hard numbers, but the bottom of this page has some hypotheticals.

You're also looking at your computer anology wrong. Sure, the circuits in a computer are there to computer, but they're capable of much more computing than just compiling a program. If you're equating protein synthesis == compiling code, then there's a lot less 'extra' in a bacterial cell than in a computer.

I'm more comparing computing with making proteins. A computer's sole purpose is to compute and it does nothing else. A bacterium's sole purpose (for our use) is to make proteins, but it does other things. This bothers me on a philosophical level. If it's not something that can be worked around (or if the alternative solutions are impractical) so be it.

Second point: not really. The science behind folding@home is a good overview; like ulc said

What I gathered there is that it's not a theoretical problem, but a computer resources problem. Is this correct?

[Ulc]

Second point: not really. The science behind folding@home is a good overview; like ulc said

What I gathered there is that it's not a theoretical problem, but a computer resources problem. Is this correct?

No.

Garbage in = garbage out. We don't know the exact parameters to put into the algorithms - or even if exact universal parameters exist at all*!

We can somewhat reliable predict whether a protein will be folded at all, which is not a given at all. Estimations vary, but around 25-40% of all proteins are predicted to be at least partially natively unstructured. We can even predict with a fair degree of accuracy what type of secondary structure will be formed - in conventional proteins, for proteins that have a very unusual sequence, it's still pretty much a coin-toss.

Folding at Home is still working with ideal proteins. Small, rigidly folded and fast folding proteins.

We are making headway, but it's a long and bumpy road yet before understand what the heck is going on in regards to proteins - and when (if) we ever got a solid grasp on that, we'll have to start looking at it in the near infinitely complex and crowded system that is even a single cell!

As we type here, a paradigm shift is happening, we're slowly realizing that the fundamental assumption in biological (that protein function depends on a exact, and rigid structure) might be wrong - which is the basis for my comment about fellow biochemists, the debate is rather.. heated.

But the idea of being able to solve protein structure based on nothing more than the sequence is something that we'd really like to be capable of. It's called Ab inito prediction, and it's a huge field within protein chemistry.

Edit: I'm probably not the most neutral person in the "structure ? function" debate - my thesis is basically one big assumption that dynamics are the basis function, and from my position, it just seems so obvious that the "structure = function" paradigm is wrong that I can't even wrap my head around the opposition here. Which is probably why the debate is so heated.

*Oversimplification, of course exact parameters exist, but not necessarily on a useful level. Solving the structure of proteins through quantum physics are not really viable due to the whole "several thousand atoms in each protein", and solving the Schrödinger equation exactly for such a system is pretty much at the "never gonna happen, no way, no how" level.

[Angua]

Also, does anyone have any convenient links to the research (which I think was reported at one point on the bbc) about scientists using artificial selection to get bacteria which are able to catalyse new types of reactions, as it can be more efficient than chemists trying to do the same thing in a lab?

Having trouble figuring out search terms.

[gmalivuk]

A computer's sole purpose is to compute and it does nothing else.

Nonsense. My computer also produces heat, and runs a fan to dissipate that heat, and flashes some LEDs to tell me things, and runs a display, and takes input from a keyboard and a mouse, and sends and receives information wirelessly, and tons of other things that are not computing. Many of these things are absolutely essential to a machine that you want a human being to be able to actually use to do computations.

Sure, you could argue that there are some small *parts* of a computer that only do computing, but the same is true for bacteria and protein synthesis.

[Izawwlgood]

My computer is currently running 58 background processes. Therefor, anything I do on it I cannot trust to be 'less complex' than if I conjured the, say, text document, from thin air.

sourmilk, the comparison is valid; a bacteria producing a protein you want is as simplified as a computer compiling a program, or as holding a text document for you. You don't have to worry about other modifications or improper interpretation or abundant complexity, because the system has been selected and indeed, engineered, for maximal simplicity. Another thing to consider is that in protein synthesis, bacteria are often induced to produce the protein of interest via a selective promoter. So when the bacteria are actually producing the protein you want, they typically aren't even growing or dividing. They are devoting ALL their resources to producing the protein of interest. Now, granted, bacteria are complicated machines; but so is your computer. How come pushing the H key makes an H character appear where my cursor is? It seems like there's an awful lot of complexity between me physically pressing that key and getting the character on the screen, complexity that is quite analogous to feeding a bacteria a plasmid, letting it grow, and then inducing the bacteria to synthesize the protein encoded on the plasmid with a promoter.

Also, Angua, I'm not sure if this is what you're talking about, but I thought this was pretty rad.

[sourmìlk]

gmalivuk, all those things are directly related to computation. They are either necessary for it or byproducts of it.

Izawlgood, I rather like it when the bacteria devote all their resources to making a protein, because that means it's the simplest way to do it in the sense that there are no extra bits. Also, whereas we know exactly what happens between pressing the H key and making an H character appear where your cursor is, Ulc just explained that there's a lot we don't know about protein folding.

Ulc, why can we not solve the Schrodinger equation for each particle? Is it just too much computation with all the atoms in a protein?

[lutzj]

gmalivuk, all those things are directly related to computation. They are either necessary for it or byproducts of it.

Not really. Do those fans need to be spinning? I could just blow air across the motherboard to cool it, and then we'd eliminate an unnecessary, complicated extra function of a computer.

[sourmìlk]

That would just displace complexity, it wouldn't remove it.