The idea is to express the uniqueness condition by means of identity between entities to which the property in question applies.
This amounts to stating that "$x$ has the property $P$, and all entities which also have property $P$ must be identical to that very $x$" or, slightly differently worded, "$x$ has the property $P$, and there is no other entity which also has the property $P$ and is different from $x$".
Formally, this can be expressed as
$\exists x (P(x) \land \forall y (P(y) \to x = y))$
or
$\exists x (P(x) \land \neg \exists y (P(y) \land x \neq y))$
It can easily be shown that the two formulations are equivalent in their truth conditions.