The CHARMM36 force fields have all been developed for use with the additiveTIP3P model, with fixed atomic partial charges. Substitutions are not recommended without careful testing, esp. for the lipid force field. Most alternate water models are more expensive to use, and have not shown marked improvement for solute properties.
CHARMM has a polarizable Drude force field, which has been implemented in OpenMM and NAMD, but which Gromacs does not support.