What does ZFC do that Peano arithmetic can't do?

More precisely, is there a "natural" statement (a statement that isn't deliberately constructed to be an example) that can be stated in PA, proved in ZFC, but not provable in PA?