Issue while using different scoring metric in Gridsearchcv sklearn

Question

I am doing elastic-net regression and trying to estimate the best hyper-perimeter using Gridsearchcv. But when I change scoring in Gridsearchcv from default r2 to mean_squared_e