what is the difference between LES modelling of a compressible flow and an incompressible flow???

is it just the way we treat the isotropic part of the stress tensor???

Depends on the regime. For low mach numbers you can assume the pressure in the momentum equations is not thermodynamic, threfore uncopuling energy-momentum equations an absorbing into the pressure the isotropic part of the stress tensor. The models are essentially the same than incomrpessible flow although numerical probelms can arise if strong gradients of density can occur. For highest Mach numbers (not neccesarly > 1), the pressure is the thermodynamic pressure so you have to introduce a equation of state to link energy and momentum. The ispotropic part of the stress have to be model and you need to introduce futher models to close your subgrid contributions in the energy equations. As well, in te latter case the numerical treatment is different (the non-linearity of the equatiosn increases)