Iron cluster topology and parameter question

1

I am trying to dock two proteins. Just using one chain from the first protein and the all chains of the second protein. The first protein has a lot of iron-sulfur clusters, including an Fe-Mo cluster with 7 irons, 9 sulfurs, and a Mo. But, we don’t expect that to participate in the docking. So, I am using the chain (does not include the Fe-Mo cluster) that participates directly in the docking with the second protein. This chain does, however, include a [2Fe-2S] cluster and the second protein has a [4Fe-4S] and a [3Fe-4s] cluster. The second protein is ferredoxin. Another paper already used HADDOCK with ferredoxin. They said “Topologies and parameters for the [4Fe–4S] and [2Fe–2S] clusters were added to the CNS parameter and topology files used by HADDOCK.” I can’t find those topologies and parameters. But, they were able to dock ferredoxin with another protein, despite these metal centers.

The paper is: Structural Insight into the Complex of Ferredoxin and [FeFe] Hydrogenase from Chlamydomonas reinhardtii. Wolfgan Lubitz et al.

  1. Why does it complain about FES ([2Fe-2S]) if that was added to the CNS topology and parameter files? Or, were these parameters not officially integrated into the CNS files, but into the authors local copy of HADDOCK?

  2. If there are no topologies and parameters for certain groups like metal clusters, what is a good practice to still use HADDOCK for docking but maybe not including that metal cluster? For example, if I want to use the full protein instead of just a chain, but it has a metal cluster for which topologies and parameters don’t exist. I’ve done parameterization before, but since this cluster isn’t involved, what could be done? Dummy atoms?

*Additional somewhat related question
3) If I need to use more than one chain in the first protein, but don’t want to use the full protein, is there a way to request, for example, chain F and chain G? The webserver dropdown only allows for selection of one chain or all.

2

the 2Fe-2S cluster is supported by needs to be defined/named properly for HADDOCK to recognise it.

See: https://rascar.science.uu.nl/haddock2.4/library

  • CFE: CYS with an iron sulfur cluster.
    HADDOCK will automatically search for closeby CYF residues and created bonds with the iron cluster
    Atoms:
    N,HN,CA,HA,CB,HB1,HB2,SG,FE1,SF1,FE2,SF2,C,O

The second cysteine to which it is attached should be named CYF

And it should be defined as ATOM and not HETATM

3

Well the [2Fe-2S] is not bound to a cysteine. I used CFE for that part of the first protein. I changed the HETATM to ATOM, but should they stay HETATM?

The [2Fe-2S] is like this in the PDB now:

ATOM 42123 SF1 CFE G2798 145.545 151.864 146.314 1.00 34.09 S
ATOM 42124 SF2 CFE G2798 147.790 149.497 145.104 1.00 31.21 S
ATOM 42125 FE1 CFE G2798 147.189 150.543 146.946 1.00 57.96 FE
ATOM 42126 FE2 CFE G2798 146.264 150.946 144.447 1.00 60.78 FE

But for the the second protein being docked (ferredoxin), there is a [4Fe-4S] and a [3Fe-4S]. Ferredoxin has been docked using HADDOCK previously, but the papers only mentioned adding parameters for [2Fe-2S] and [4Fe-4S], nothing about [3Fe-4S]. They are also not bound to cysteines. If I can get those last two defined, then HADDOCK should do what I need.

4

Well I am afraid we don’t have the params for the 3FE cluster.

If prodrg can not generate those than I am afraid the server won’t handle those.

But the question you should ask yourself is: Are those clusters at the interface of the complex? If not then you could model the complex without them and add them back later.

5

Ok, last question. What is the best practice for “you could model the complex without them”? Just removing them entirely or replacing them with something else? Thanks for the help!

6

If not part of the interface they can be simply removed.