1312: Hfq TB puzzle - monomer subunit

Closed since over 9 years ago

Summary

• Created: November 30, 2016
• Expires: December 07, 2016 at 23:00 UTC
• Max points: 100

Description

This is the monomor subunit for puzzle 1311. If puzzle 1311 is crashing for you, this puzzle can serve as a substitute for 1311. When we gather solutions, we will use both the solutions from this puzzle and from 1312.

This TB puzzle is summarized in http://fold.it/portal/node/2003049. It is a hexameric protein responsible for virulance factors in TB. We do have crystals and phase data, but none of our models work! Help us get a model to solve this structure.

VEVKIGITDSPRELVFSSAQTPSEVEELVSNALRDDSGLLTLTDERGRRFLIHTARIAYVEIGVADARRVGFGVGVDAAAGSAGKVATSG

Top groups

• 1. Anthropic Dreams 100 pts. 9,430
• 2. Contenders 78 pts. 9,407
• 3. Beta Folders 60 pts. 9,391
• 4. Gargleblasters 45 pts. 9,380
• 5. Go Science 33 pts. 9,379
• 6. L'Alliance Francophone 24 pts. 9,326
• 7. Russian team 17 pts. 9,270
• 8. Void Crushers 12 pts. 9,205
• 9. Deleted group pts. 9,154
• 10. xkcd 6 pts. 9,096

Top soloists

• 1. Susume 100 pts. 9,427
• 2. gitwut 98 pts. 9,407
• 3. johnmitch 95 pts. 9,404
• 4. Galaxie 93 pts. 9,396
• 5. reefyrob 90 pts. 9,391
• 6. bertro 87 pts. 9,389
• 7. actiasluna 85 pts. 9,380
• 8. Bruno Kestemont 83 pts. 9,379
• 9. LociOiling 80 pts. 9,376
• 10. tokens 78 pts. 9,370

Comments

[jeff101]

Is it possible for us to load solutions from puzzle 1312 into puzzle 1311?

[free_radical]

As of now, I am unsure. I will be looking into this possibility this weekend.

[jeff101]

The site below is calculating a Contact Map for
the 90-segment monomers in Puzzles 1311 & 1312:

http://raptorx2.uchicago.edu/ContactMap/myjobs/60338832_109300/

[retiredmichael]

good you are looking into it, as for now will take a break from folding. have a couple of versions in 1312 that i would like to test in 1311.
until that is possible i feel like i am a working in the dark.

[free_radical]

This is a post from a scientist in Jim's lab.

1. It is hard to pin-point the RNA interacting residues from the sequence alignments - sequence homology is too low, so you may end up being severely misled. On top of it, Mtb Hfq has extra strands on N-terminus (predicted to form beta-clasp), but it is not clear how it would fit into classic Hfq doughnut arrangement.
2. From studying solved Hfqs with RNA bound few potentially useful observations could be made:
   • RNA interacting spots located at the edges of secondary structure elements and/or in between them
   • Most common interactions with RNA include: Q and N - H-bonding with the base; F (or Y) - stacking with the base ring; Y - stacking or bonding with the sugar; H - H-bonding with the sugar O and/or with the phosphate group. Mtb Hfq has one of each for those within the useful positions (I would pair this with the secondary structure predictions):
     10 20 30 40 50 60 70
     | | | | | | |
     VEVKIGITDSPRELVFSSAQTPSEVEELVSNALRDDSGLLTLTDERGRRFLIHTARIAYVEIGVADARRV
     GFGVGVDAAAGSAGKVATSG
3. Yes, there are H-bonds with RNA (see above in (2). Structures were solved with polyU (pdb entry 4Y91)

[Bruno Kestemont]

Thanks free_radical I feel this is very useful info; please continue to post this kind of info.
Now I hope that one of our "almost-biochemist" players will be able to translate this for normal players.

Would we try to put the following inside of the ring ?
-the side of some sheet in order to try blue H bonds with the expected RNA?
-H-bonding sidechains ? (in CPK view: many red oxygen and blue N inside of the ring?)

Thus, the following AAs would be present in quantity on the inside of the ring?
serine(s) 1+-
threonine(t) 1+-
asparagine(n) 1+, 1-
glutamine(q) 1+, 1-
arginine(r) 3+
histidine(h) 1+, 1-
lysine(k) 1+
aspartic acid(d) 1+, 1-
glutamic acid(e) 1+, 1-
tryptophan(w) 1+
tyrosine(y) 1+-

Possible links for inspiration here: Hydrogen Bond Network Filter ?

Please correct me if I'm wrong.

[Bruno Kestemont]

This one here?

pdbid=4Y91

[Susume]

Thanks free_radical for getting us some more input from the scientists!

Bruno, my understanding from reading is that Hfq binds to many different RNA molecules, and it uses a few different methods to bind them. The Hfq hexamer has multiple sites where it can bind the various RNA. The scientist tells us these sites tend to be on the loops between sheets or helices. In solved Hfq (like the one you linked), the hexamer is like a flower with the 6 helices on the front or "proximal" face, and no helices on the back or "distal" face. One place it binds RNA is at the middle of the hexamer on the front face (the side with helixes). Another place it binds (a different kind of) RNA is the middle of the hexamer on the back face. So the middle of the hexamer would probably be a good place for loops, both on the front face and on the back face.

The amino acids the scientist recommends to put on or near these loops are:
Q, N, or H - to H-bond with the RNA
F or Y - to "stack" rings with rings on the RNA
I don't think you need very many bondable sidechains together in one place - just the few right ones - and it sounds like no one is sure which ones are right for this protein.

The article I found most helpful is: http://nar.oxfordjournals.org/content/early/2013/11/26/nar.gkt1171.full

This article also says that different species of Hfq have very different tails (the unstructured part after all the sheets), that these tails tend not to show up on ED because they are flexible, and that they may move around to contact different RNA types in different ways.

(Hopefully one of the scientists will jump in here if I got something wrong.)

[Susume]

Also the scientist says the two sheet strands before the helix are possibly a "beta clasp", which I had never heard of. As near as I can tell, that means a flap that can move to sometimes cover or bind to something and sometimes not.

[Bruno Kestemont]

It's facinating (and beautiful) to work on this hexamer 1311 with different potential binding sites on the loops.

I hope that all of us will soon get a workable hexamer to play with, using the current results. Nice to suspect that the helix would be on one side, the long loop on other side, and that connecting loops can be long in the water (I've solutions where I tried to short them or to make them joining helices far away each others - may be it was not necessary to complicate so much).