Building backbone with modified proline

GROMACS version:2021.4
GROMACS modification: No
Dear Gromacs Users/Developers,
I am trying to include Fluoroproline (PRF) in my peptide. I got the parameters in amber FRCMOD format and have used them to modify aminoacids.rtp file. The relevant section looks like this.

[ PRF ]
 [ atoms ]
   N    N            -0.146327     1
   CD   CT           -0.189338     2
   HD1  H1            0.094775     3
   HD2  H1            0.094775     4
   CG   CT            0.420275     5
   HF   Hf            0.017040     6
   F1   f            -0.297174     7
   CB   CT           -0.226092     8
   HB1  HC            0.091203     9
   HB2  HC            0.091203    10
   CA   CT           -0.087012    11
   HA   H1            0.102478    12
   C    C             0.540229    13
   O    O            -0.506134    14
 [ bonds ]
     N    CD
     N    CA
    CD   HD1
    CD   HD2
    CD    CG
    CG   HF
    CG   F1
    CG    CB
    CB   HB1
    CB   HB2
    CB    CA
    CA    HA
    CA     C
     C     O
   [ impropers ]
    CA    +N     C     O
    -C    CD     N    CA

I modified the ffbonded.itp, ffnonbonded.itp and atomtypes.atp as well. I also added the new residue to residuetypes.dat as Protein as well.

PRF     Protein

I am able to generate the system including solvation and ionization. I can also run simulations. The problem is that the backbone peptide bond between PRF and the preceding residue is not generated by pdb2gmx in the file. The peptide bond between PRF and the next residue is formed.
What could be the problem? Which files should I change so that pdb2gmx makes a bond between the modified proline and the preceding residue.
Thanks in advance.

Your [bonds] directive does not specify any bonds to the flanking residues. You should include N -C and/or C +N in the list.

Thanks. I copied the [bonds] section from a proline and somehow missed the last line.
Works perfectly now.
Thanks again.