1501: Revisiting Puzzle 58: Insulin Mutant

Closed since almost 8 years ago

Summary

• Created: March 25, 2018
• Expires: April 02, 2018 at 23:00 UTC
• Max points: 100

Description

This is a throwback puzzle to the early days of Foldit. These are the two chains of a bio-engineered variant of human insulin, which contains six cysteine residues that oxidize to form three disulfide bonds. We are revisiting old Foldit puzzles so we can see how useful the recent additions to the game have been.



Sequence:

GIVEQCCTSICSLYQLENYCN / FVNQHLCGSHLVEALELVCGERGGFYTPK

Top groups

• 1. Beta Folders 100 pts. 10,036
• 2. L'Alliance Francophone 74 pts. 10,023
• 3. Go Science 54 pts. 9,993
• 4. Void Crushers 38 pts. 9,936
• 5. Anthropic Dreams 27 pts. 9,931
• 6. Contenders 18 pts. 9,913
• 7. Gargleblasters 12 pts. 9,894
• 8. Marvin's bunch 8 pts. 9,780
• 9. FoldIt@Netherlands 5 pts. 9,463
• 10. HMT heritage 3 pts. 9,437

Top soloists

• 1. RockOn 100 pts. 10,391
• 2. eusair 97 pts. 10,339
• 3. retiredmichael 94 pts. 10,033
• 4. nicobul 91 pts. 10,023
• 5. LociOiling 87 pts. 10,006
• 6. ZeroLeak7 84 pts. 9,993
• 7. reefyrob 81 pts. 9,980
• 8. pauldunn 78 pts. 9,954
• 9. Timo van der Laan 76 pts. 9,936
• 10. Bruno Kestemont 73 pts. 9,932

Comments

[NinjaGreg]

Was this intended to be two separate lengths? The numbering is continuous across the gap, but it's not clear if the two gap ends are supposed to remain close to one another.

[bkoep]

This puzzle contains two distinct chains that can be moved independently of one another. The chains may interact with one another and form disulfides with one another, but there is no reason that the gap ends should remain close together.

Foldit automatically renumbers all the residues in a puzzle. If there are multiple chains in a puzzle, then Foldit renumbers the first chain (from the N- to C-terminus), and then continues this numbering at the next chain (from its N- to C-terminus). The main purpose of the Foldit numbering is so that the program can keep track of each residue. (Importantly, the Foldit numbering does not correspond to conventional numbering that we use to talk about proteins. So, for example, if you read something on Wikipedia about "position 22 of insulin," that is not the same as residue number 22 in this Foldit puzzle.)

The synthesis of insulin in the body is actually a multi-step process. It actually starts out as a single chain, forms disulfide bonds, and then the middle portion of the single chain is excised by proteases, resulting in two distinct protein chains connected by disulfide bonds.

[NinjaGreg]

Thanks for the explanation, very interesting. Fascinating how this all is so intricate!

(Side note: if you could delete the duplicate post below, I'd appreciate it. I hit save twice, apparently.)

[LociOiling]

Totally geeking out, but here goes.

This puzzle has two chains, so there are two C terminals. (Also, two N terminals.)

The main reason that matters is atom numbering - the C terminal has an extra oxygen backbone atom, so the beta carbon, the first atom of the sidechain, is atom 6 instead of atom 5.

This scrap of code bands the beta carbons of the C terminals in 1501 to demonstrate:

band.AddBetweenSegments ( 21, 50, 6, 6 )

Lots of banding recipes offer to band beta carbons, but most probably miss the target when there are multiple chains. They'll end up banding to the extra oxygen atom in every C terminal except the last one. (Missed it by this much!)

This protein is available in the PDB as 1A7F. The two chains are identified as A and B. The A chain is segments 1 - 21 in Foldit, and the B chain is segments 22 - 50. (In the PDB, each chain is numbered starting at residue 1.)

If you look at the PDB header for 1A7F, the SSBOND records give the disulfide bridges:

SSBOND   1 CYS A  6  CYS A 11   1555   1555  2.02  
SSBOND   2 CYS A  7  CYS B  7   1555   1555  2.02  
SSBOND   3 CYS A 20  CYS B 19   1555   1555  2.02  

(I removed some excess spaces there, Fortran fans beware. I'm actually not sure what the 1555s mean, but the 2.02s are the distance in Angstroms between the two sulfur atoms in the bridge.)

The first bridge is within the A chain, but the other two connect the A chain to the B chain. So "CYS A 6" and "CYS A 11" connects the cysteines at segments 6 and 11. Both are on the A chain, so the numbers are the same in Foldit.

The next bridge connects "CYS A 7" to "CYS B 7". Both are residue 7 in PDB terms, each in its own chain.

To convert the B chain number to the Foldit number, just add 21, the length of the A chain. So "CYS B 7" refers to Foldit segment 28. In the third bridge, "CYS B 19" becomes segment 40.

The bridges in bridge wiggle terms are:

6,11 7,28 20,40 

Some puzzles may have more than two chains. The recent 1448: Classroom Puzzle: Amyloid Protein is the only one I can think of, but there may have been others. Since Foldit doesn't have chain ids, the only way to detect N terminals and C terminals in the middle of the protein is to look at the atom count from structure.GetAtomCount and compare it to the normal value for the same amino acid in the middle of the chain.

The N terminals have two extra atoms, but they're both hydrogens, which get counted after other atoms. So the extra N terminal atoms won't throw off beta carbon banding recipes. The extra oxygen in the C terminal is what changes the number of the beta carbon from 5 to 6.

Enough already. One day print protein will identify chains, and maybe the some weak spots in the wiki will get patched up.

[frood66]

this has already been mentioned regarding 1498…same prob here?

[bkoep]

Are you referring to the erroneous high scores, or something else? Have you also experienced problems in this puzzle?