Attempting to Use EOM with Ig Fragments

Linear (OLIGOMER), and non-linear (MIXTURE) analysis, singular value decomposition (SVDPLOT), addition of missing fragments (BUNCH, CORAL), analysis of flexible systems (EOM/RANCH & GAJOE), flexible refinement of high-resolution models (SREFLEX)
Post Reply
Message
Author
ANelson
Guest
Posts: 1
Joined: 2015.04.15 06:00

Attempting to Use EOM with Ig Fragments

#1 Post by ANelson » 2015.04.15 06:18

I have SAXS data for a Fab (antigen binding fragment) that is made of of ~90% beta sheet structure. The Fab is heterodimer of two chains commonly referred to as heavy and light chains covalently linked by a disulfide residue. Each chain contains two immunoglobulin (Ig) folds which contain stabilizing intrastrand disulfide bonds. I have the sequence for my Fab, no crystal structure, but I do know that it is an IgG2ak antibody and there is a published crystal of an IgG2a antibody (PDB id: 1IGT). The constant domains of the Fab are highly conservative but the variable domains are least conservative. The major sites of variability lie within the complimentary determining regions (aka hypervariable regions). I know which cysteine residues from each chain are involved in interstrand disulfide bonding between the heavy and light chains. I would like to generate a library of Fab conformers to use in EOM to generate an ensemble that fits my SAXS data.

Does anyone have suggestions on how to impose a disulfide link between two separate protein chains (heavy and light chain)?

Thanks-ANelson

ckerr
Active member
Posts: 86
Joined: 2015.03.25 09:03
Location: EMBL Hamburg

Re: Attempting to Use EOM with Ig Fragments

#2 Post by ckerr » 2015.04.17 14:48

A contact file might do what you want. The documentation only talks about symmetric proteins but I think it should also work for heterodimers.

http://www.embl-hamburg.de/biosaxs/manu ... ontactFile

Alex
Active member
Posts: 692
Joined: 2007.08.09 21:10
Location: Planet Earth

Re: Attempting to Use EOM with Ig Fragments

#3 Post by Alex » 2015.04.17 17:51

if i understand you right, you have a light chain (A-A') and a heavy chain (B-B'), connected by -S-S- bond. do you have structure of A and B, whereas variable regions A' and B' need to be modeled? If you want to model everything from scratch - it does not make sense. If I am right and you have partial structures, note, that EOM does not support hetero oligomers (there are tricks however to overcome it).

Kento Yonezawa
Member
Posts: 3
Joined: 2017.03.30 12:00

Re: Attempting to Use EOM with Ig Fragments

#4 Post by Kento Yonezawa » 2017.03.31 04:30

I am also just trying to trace the EOM project of IgG (PDBID: 5DK3) like(Tian X et al. IUCrJ. (2015)).
However the EOM does not work well. Maybe it needs some tricky techniques.
This is a homo-dimer ,but asymmetric structure.
The structure can be described as follows.

(Fab domain)--Flexible region--CPPC--Flexible region-- (Fc domain)
| | ||
(Fab domain)--Flexible region--CPPC--Flexible region-- (Fc domain)

We would like to know the fractions and the basic structures in the solution.
CPPC sequence constructs dislfide bonds on both Cystein residues.
Fc domains form dimer.
And I have a monomer sequence file as follows.

(Fab domain)--Flexible region--CPPC--Flexible region-- (Fc1)

My questions are
1)Should I treat this structure as homo-dimer? If I treat this structure as dimer, it needs to input P2 symmetry.
However this structure is asymetric form. how should I input?
2) I would like to input Fc dimer as one pdb file. But maybe it is difficult, because sequence is not continus between two Fc domains.
3) Should I edit this sequence file?

Thank you

Regards,

Kento

Alex
Active member
Posts: 692
Joined: 2007.08.09 21:10
Location: Planet Earth

Re: Attempting to Use EOM with Ig Fragments

#5 Post by Alex » 2017.03.31 11:39

Hi Kento
1)Should I treat this structure as homo-dimer? If I treat this structure as dimer, it needs to input P2 symmetry.
you can try doing that, but then you need to align Fab domain dimer to Z-axis (using alpraxin program), then also specify,
the contact conditions (which i think also should be possible for disulfide bonds).

Second scenario is to change the sequence such that one Fc domain is at N-term and the other is at C-term and create
a single chain where you would fix FAB and FC domains and use contact restraints. Of course this is just going to be
some approximation of the system since your inverse the the sequence for half of it.

Third scenario can be used if you know position of Fab in respect to Fc, so that you can first model one chain independently,
then second chain, then combined them together in all possible ways and filter obviously wrong (clashed) models.

HTH, Alex

Kento Yonezawa
Member
Posts: 3
Joined: 2017.03.30 12:00

Re: Attempting to Use EOM with Ig Fragments

#6 Post by Kento Yonezawa » 2017.04.03 03:26

Thank you for your reply!
I'll try these senarios!

Regards,

Kento

Alex
Active member
Posts: 692
Joined: 2007.08.09 21:10
Location: Planet Earth

Re: Attempting to Use EOM with Ig Fragments

#7 Post by Alex » 2017.04.03 17:38

Please, do update this thread with your successful "recipe".

Kento Yonezawa
Member
Posts: 3
Joined: 2017.03.30 12:00

Re: Attempting to Use EOM with Ig Fragments

#8 Post by Kento Yonezawa » 2017.05.15 03:36

I tried 1st and 2nd suggestion. In the case of second suggestion, I could get much reasonable results.
Although I performed the first and program was run, the contact file does not work.

It was carried out as p2 semmetry and monomer sequence.
The model structure is also mononer which is shown as follow.

(Fab domain)----Flexible region---CPPC---Flexible region---(Fc domain)

In this homodimer, Fc domain constitutes a dimer with the other Fc domain.
So I needed to make contact file.
At first I performed alpraxin program for Fc domain dimer in order to set contact condition.
Then I removed one Fc structure. Second, I prepared the contact file as follow.

dist 5
2 225 229
dist 5
3 234 440

This contact file includes CPPC disulfide (first line) and Fc contact (second line).
(The residue number from N terminal of Fc mononer to C terminal of Fc monomer was input in order to
connect both Fc domains.)

But, results of structures show that Fc domain in some of generated structures did not adhere with the other Fc domain.
So, I would like to know more information about EOM contact in order to analyze correctly.

(1) Is it possible to set contact condition when I would like to use P1 condition?
Because command line does not include contact input when I choose P1.

(2) For example, If there are four domains which are connected with long flexible linkers,
is it possible to make contact file which 4 domains are close to each other as follows?

(B)--(C)
| |
| |
(A) (D)

(3) I checked some of the structures generated 10000, but some of them formed Fc contact.
I would like to know whether the contact condition adjust to all generated 10000 structure or not.

(4) I also used "Keep the subunit in the original PDB coordinates?" (yes). It work correctly when alpratxin is performed.
Does this command also work like contact?

Thank you.

Regards,

Kento

Post Reply