[gmx-users] how modify the forcefield amino acid definitions for non-naturally occuring amino acids?

[Justin A. Lemkul]

Kirill Bessonov wrote:
> Dear Gmx users,
> 
> I had been using Gromacs for couple of years and always found these 
> forums quite helpful. Right now again I have another question to ask 
> that I was not able to figure out yet myself and reading previous posts.
> 
> I need to the ffG53a6 forcefield to be able to recognize the 
> non-standard amino acid like AZE (Azetidine-2-carboxylic acid) which 
> closely resembles Proline, but lacks CH2 group in the ring of the 
> side-group, making the R-group a 4 member ring instead of 5. Thus, this 
> amino acid has very close chemical structure to the Proline which is 
> properly defined in the *.rtp file as below
> 
> [ PRO ]
>  [ atoms ]
>     N     N     0.00000     0
>    CA   CH1     0.00000     1
>    CB  CH2R     0.00000     1
>    CG  CH2R     0.00000     2
>    CD  CH2R     0.00000     2
>     C     C       0.450     3
>     O     O      -0.450     3
>  [ bonds ]
>     N    CA    gb_21   
>     N    CD    gb_21   
>    CA    CB    gb_27   
>    CA     C    gb_27   
>    CB    CG    gb_27   
>    CG    CD    gb_27   
>     C     O    gb_5    
>     C    +N    gb_10   
>  [ angles ]
> ;  ai    aj    ak   gromos type
>    -C     N    CA     ga_31   
>    -C     N    CD     ga_31   
>    CA     N    CD     ga_21   
>     N    CA    CB     ga_13   
>     N    CA     C     ga_13   
>    CB    CA     C     ga_13   
>    CA    CB    CG     ga_13   
>    CB    CG    CD     ga_13   
>     N    CD    CG     ga_13   
>    CA     C     O     ga_30   
>    CA     C    +N     ga_19   
>     O     C    +N     ga_33   
>  [ impropers ]
> ;  ai    aj    ak    al   gromos type
>     N    -C    CA    CD     gi_1    
>    CA     N     C    CB     gi_2    
>     C    CA    +N     O     gi_1    
>  [ dihedrals ]
> ;  ai    aj    ak    al   gromos type
>   -CA    -C     N    CA     gd_14   
>    -C     N    CA     C     gd_39   
>    CA     N    CD    CG     gd_39   
>     N    CA    CB    CG     gd_34   
>     N    CA     C    +N     gd_40   
>    CA    CB    CG    CD     gd_34   
>    CB    CG    CD     N     gd_34   
> 
> 
> The question is how can I derive from this Proline definition topology 
> the corresponding topology for AZE? The problem is that I do know how to 

You can use PRO as a framework, but you'll likely have to define completely new 
angle and dihedral terms for the four-membered ring.  No such group exists in 
normal biomolecules, so likely the force field cannot handle it without 
modification.

> interpret the each column such as the "gromos definition" column gd_14, 
> etc, as well as other columns..... Where can I get information or 
> tutorials about this important issue on how to create own topology 
> files? Maybe someone can explain those columns in easy terms? What are 

The terms are defined in ffbonded.itp:

; Dihedral-angle type code
; Force constant
; Phase shift
; Multiplicity
; Example of usage in terms of non-bonded atom types
;
;
; ICP(H)[N]  CP[N] PD[N] NP[N]
;
#define gd_1    180.000       2.67          1
; CHn-CHn-CHn-OA (sugar)  0.6

Therefore, gd_1 is the type code, 180.000 is the force constant, 2.67 is the 
phase shift, and 1 is the multiplicity.  The comment line below this definition 
is its example usage in terms of standard atom types/functional groups.

> corresponding atoms referred by atom names in the final structure as the 
> there is not ASCI diagram of those as I've seen in AMBER tutorials? 
> 

The atom names used are standard nomenclature for amino acids.

> Also I am interested in knowing on how the topology files are created 
> from scratch?
> 

Sorry, I don't understand what you mean here.

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================