Author Topic: Built-in Enzyme Regulation  (Read 4358 times)

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« on: March 23, 2005, 05:27:50 PM »
So far all I can find is allosteric regulation, but there are others.  Here's a way to introduce enzyme regulation to the enzymes themselves.  Any system that works like this should be overridable by the DNA for a certain nominal cost.

All enzyme complexes are built of activation sites.  We already (hopefully) understand how this works.  On top of this, we can add regulatory sites.  These are part of the enzyme complex in exactly the same way activation sites are, as bit patterns.

However, for this one the bit pattern has some extras.  One each side of it the 3 bits that used to code for efficiency now code for threshold values.  If the amount of stuff in the cell is above or below (I'm not sure yet) this threshold it turns the enzyme complex off.

I'm still working on the details of how it would work, but this is the skeleton of it.  This would help remove some regulation from the DNA when the DNA doesn't want to bother.

Offline shvarz

  • Bot God
  • *****
  • Posts: 1341
    • View Profile
Built-in Enzyme Regulation
« Reply #1 on: March 23, 2005, 07:02:12 PM »
This is a good idea.  The range should be very broad, though...


You know, with all these patterns for enzymes...  Maybe we can organize it logically and linearly.  Like this:

first N bits define molecule on which enzyme works
next N bits refer to a particular reaction that molecule is involved in
next N bits refer to threshhold shutdown
"Never underestimate the power of stupid things in big numbers" - Serious Sam

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« Reply #2 on: March 23, 2005, 08:51:03 PM »
That's possible, but it makes the enzymes very linear.  I was hoping to keep it as much like a string of amino acids as possible.  Some spots have Junk DNA, there really isn't an up or down, etc.  Whatever you happen to find in the string of bits is what the enzyme does.

Like someone giving you a box of junk and saying 'make something'.  A more linear system would mean that you have 'here's a part, it has to be used for X'.

Offline shvarz

  • Bot God
  • *****
  • Posts: 1341
    • View Profile
Built-in Enzyme Regulation
« Reply #3 on: March 23, 2005, 10:46:57 PM »
No-no, still same system, just more organized.  You were proposing to assign random sequences of bits to individual enzymes and I am saying let's have some kind of system.  It would still be a string of bits, just the activation sequences would be organized.  And it actually makes more sense, because that way an enzyme converting A>B has a higher probability to mutate into enzyme that converts A>C than into some random enzyme X>Z.
"Never underestimate the power of stupid things in big numbers" - Serious Sam

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« Reply #4 on: March 23, 2005, 11:51:29 PM »
Okay, I see what you're getting at.  Not a bad idea.

The only problem is that if I define the molecule on which the enzyme works as citric acid, and I define the reaction as glycolisis, that's rather silly, isn't it?  You just need one bit pattern that defines the substrates and reaction that takes place.

Other than that I think it would work well.

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« Reply #5 on: March 23, 2005, 11:55:11 PM »
Basically, any regulation of the enzyme has to be for the entire enzyme complex, not just for individual activation sites.

Remember that regulation sites deform the entire enzyme complex, not just individual parts.

So what we really need are activation sites and regulation sites, each defined similarly.

Activation site: efficiency bits - activation code - efficiency bits
Regulation sites: threshold bits - regulation code - threshold bits
« Last Edit: March 23, 2005, 11:58:37 PM by Numsgil »

Offline shvarz

  • Bot God
  • *****
  • Posts: 1341
    • View Profile
Built-in Enzyme Regulation
« Reply #6 on: March 24, 2005, 12:47:04 AM »
Nah, too complicated.

Here, I'll explain on example.  Let's say we want to have a potential for 4000 molecules, that is 2^12.  So a glucose can be defined as 111000111000.

Then we want to allow glucose to be involved in 16 different pathways: 8 degradational, 8 generational.  That is 2^4.  So we say conversion to pyruvate is defined as 1010

Then the activation site for enzyme converting glucose to pyruvate is:
111000111000 1010.

This is the bare minimum to activate enzyme: 16-bit string.  All enzymes working on glucose will have tha same first 12 bits, so when enzyme mutates it is very likely to turn into enzyme working on glucose.  We can even have sugars designed to be similar in sequence to glucose so that the mutated enzyme will be more likely to keep working on carbohydrates than to suddenly start splitting proteins.

After the first 16 bits of activation sequence we will have regulatory bits.  Say we want to define threshhold of inhibition of enzyme.  Maybe make it a log scale with 8 possibilities: 0 (no inhibition), 1,  5, 25, 125, 625, 3125, 15625.  This adds another three bits. And the sequence like 111000111000 1010 000 will mean "enzyme converting glucose to pyruvate with no feedback inhibition".

Then we can add more and more regulatory bits.  But their order should be of decreasing importance, allowing finer and finer tuning of enzyme's functionality.
« Last Edit: March 24, 2005, 12:49:44 AM by shvarz »
"Never underestimate the power of stupid things in big numbers" - Serious Sam

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« Reply #7 on: March 24, 2005, 12:54:03 AM »
I agree that coding patterns for like reactions should be very similar.  That's not difficult at all to acomplish.  But remember that alot of reactions aren't just A->B but A+B-> C.  As far as enzymes are concerned, I think it's the actual reaction more than the substrates that define it.

Also remember that sometimes you don't want to regulate a reaction with it's direct byproduct.  if A->B->C you might want to regualte A with C instead.  So you need a molecule code and a threshold value.

Then, on top of that, you have which direction the threshold works.  Does the protein turn on if the value is greater than or less than the threshold?

THEN, on top of all that, you need the ability to have multiple regulatory sites.  If A->B->C and A->B->D then you might want to regulate A with C and D.

Offline shvarz

  • Bot God
  • *****
  • Posts: 1341
    • View Profile
Built-in Enzyme Regulation
« Reply #8 on: March 24, 2005, 01:06:58 AM »
So, what is your solution?  I did not quite understand the idea of regulation codes.  Sounds like these bit-strings are starting to get huge.

I would solve it like this:

I will find what regulates what in nature and we just assign the regulatory molecule ourselves.  Just one.  And not even for all enzymes.  Yes, it is not flexible.  But it would be just one way to regulate them.  We'll have more.
"Never underestimate the power of stupid things in big numbers" - Serious Sam

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« Reply #9 on: March 24, 2005, 01:20:32 AM »
Okay, here's what I think I'm saying:

Imagine your long string of bits.  This is a complex.  Inside this string are several activation sites or enzymes.  Now, in addition to these activation sites are regulatory sites that work for all enzymes in this complex.

Liek this:

01010110101010101010101010101010101010101010100 -> complex
....|---enzyme1---|.|--enzyme 2|.....|---regulation site---|

The regulation site turns on/off both enzyme 1 and enzyme 2 depending.

Depending on what you ask?

Well, a regulatory site works like this:

3 bits for threshold value - molecule code (should be fairly long to prevent lots of regulation sites from cropping up in large complexes) - 1 bit that defines greater than or less than - 2 bits for threshold value

Threshold value works logarithmically as you suggest, with 0 being unregulated and 32 being regulated at 32000.

Maybe 20 is like ~9000.

The molecule code will need to be quite lengthy to prevent lots of regulation sites cropping up in a complex.  Longer complexes, though, are increasingly likely to posses one.  I haven't tried to figure out the statistical test for percentages yet.
« Last Edit: March 24, 2005, 01:20:45 AM by Numsgil »

Offline shvarz

  • Bot God
  • *****
  • Posts: 1341
    • View Profile
Built-in Enzyme Regulation
« Reply #10 on: March 24, 2005, 12:09:14 PM »
I like this idea.  It provides another benefit for keeping enzymes that work in the same chain in a single complex.  And it encourages moving non-related enzymes away from each other (you don't want to stop protein degradation simply because there is too much fat-CoA...

What I don't like is the fact that more precise regulation will favor disassembling of complexes into single-standing enzymes (because then you can fine-tune each enzyme individually).  But I guess we can live with it...

I would still prefer to keep enzymes sorted by molecules they work on.  In cases where its A+B>C, we can just pick one of the substrates.  Or we can have two versions.

P.S:  I don't see a need for 32000 threshhold value, since it is maximal amount anyway, this just means there is no regulation.  Is 12-bit molecule code long enough?
"Never underestimate the power of stupid things in big numbers" - Serious Sam

Offline Zelos

  • Bot Overlord
  • ****
  • Posts: 707
    • View Profile
Built-in Enzyme Regulation
« Reply #11 on: March 24, 2005, 12:15:32 PM »
soudsn good to me
When I have the eclipse cannon under my control there is nothing that can stop me from ruling the world. And I wont stop there. I will never stop conquering worlds through the universe. All the worlds in the universe will belong to me. All the species in on them will be my slaves. THE ENIRE UNIVERSE WILL BELONG TO ME AND EVERYTHING IN IT :evil: AND THERE IS NOTHING ANYONE OF you CAN DO TO STOP ME. HAHAHAHAHAHAHAHA

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« Reply #12 on: March 24, 2005, 12:18:35 PM »
Quote
What I don't like is the fact that more precise regulation will favor disassembling of complexes into single-standing enzymes (because then you can fine-tune each enzyme individually).  But I guess we can live with it...

We can offset this by rewarding reactions that take place on the same complex, as you suggested earlier.  Then it's a balancing act.

Oooo, should make for some interesting enzymes.

Quote
P.S:  I don't see a need for 32000 threshhold value, since it is maximal amount anyway, this just means there is no regulation.  Is 12-bit molecule code long enough?

32000 just seemed a logical point, but you're right, that's the same as no regulation.

So maybe we make 32000 no regulation and 0 always off.

I'm not sure how long everything needs to be.  Too short and you have 17 regualation sites on the same complex that's only 20 bits long.  Too long and you never develop them.

Keep in mind that 4 bits is a hexedecimal letter.  So 12 bits sounds like alot, but it's only 3 digits in hex.  I think we're underestimating what the length needs to be.  Real complexes are very long.  I don't want to go quite that long for obvious reasions, but there is certainly reason enough not to have the bare minimum.

Offline PurpleYouko

  • Bot God
  • *****
  • Posts: 2554
    • View Profile
    • http://kuro-tejina.com
Built-in Enzyme Regulation
« Reply #13 on: March 24, 2005, 12:49:02 PM »
Here's a thought.

Why leave it in hex.

We could use the entire alphabet, even differentiate between cases.
I have a code system in one of my sample log spreadsheets that uses base 36 but that was just limited because people were unable to right it accurately if I went to base 64 or even higher.

If you want complete incomprehensibility to programmers then use a super condensed number base system.
Instead of 0 through 9 then A through f, we could have 0 - 9, a-z, A-Z for base 62. Add a period or a comma or some other symbols to round it up to base 64 and you can quickly and easily convert back and forth to binary.
Means much shorter words in the DNA too.

In this system

z=35
Z=63
ZZ=4095 or 111111111111 in binary (a 12 bit word)
(This is FFF in Hex)

whereas in Hex, the biggest you can get with two characters is

FF = 255 or 11111111 in binary (an 8 bit word)

Just a thought. Kind of interesting and extremely (deliberately) confusing as hell to programmers.

 :D  PY  :D
There are 10 kinds of people in the world

Those who understand binary.

and those who don't

:D PY :D

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7714
    • View Profile
Built-in Enzyme Regulation
« Reply #14 on: March 24, 2005, 01:58:51 PM »
If we really want to ahve fun we could pick a non power of 2 base.  Then the resultant bit string doesn't divide right into the storage letters.

Base 12 was popular because it was divisible by 2,3,4 and 6.  Base 60 was really fun because it's divisible by 2,3,4,5,6,12,15, and alot of other numbers (this was back when people hated fractions and decimals weren't invented yet).