Well, I notice immediately without even reading the documentation very closely that you are not using the documented way to declare/initialize a GMP integer:
http://gmplib.org/manual/Integer-Functions.html#Integer-Functions
This says 'GMP integers are stored in objects of type mpz_t' -- not mpz_class which you have used in your code. Now, I don't know GMP, and for all I know this 'mpz_class' type may be a fine substitute, but it's not what the documentation says to do. If you are trying to do some unconventional (but okay) manner of utilizing this library then you might want to say a bit more about exactly how you're trying to use it.
http://gmplib.org/manual/Initializing-Integers.html#Initializing-Integers
This page says to use functions like mpz_init() for initializing, not the standard '=' operator.
http://gmplib.org/manual/Assigning-Integers.html#Assigning-Integers
If I understand this page right, it's saying that if you want to drop a massive value into your mpz_t integer, you use the function mpz_set_str and pass in a char array which contains a textual representation of the number you want to assign it to. It makes sense that you must do this, because the compiler cannot, on its own, grasp such a large number in true numerical form (if it could, there would be no need for this library).