Simulations of non-globular proteins

GROMACS version: 2023.3
GROMACS modification: No
There is a protein nearly 1000 amino acids. Crystallographic structure has a gap of 150 missing amino acids in the internal part of the protein. Missing residues are difficult to model in silico with high quality.
Is it appropriate to divide such a protein into two separate chains at the site of the gap and simulate it as a protein-protein complex?
What is the better way to simulate such protein in Gromacs?