There will be a shift in the potential energy resulting from the difference between the point charge (presumably this is what your model previously assumed) and the shell charge distribution.
Another hint:
And another:
I have a function z=f(xz+y) and I want to find the partial differential of z with respect to y (it's a general sort of question, I only need it in terms of the variables already given).
My answer would be just partial df/dy but this isn't the right answer. I'm not too hot on partial...