A surjective map $\pi: X \rightarrow Y$ is a quotient map when a subset $V$ is open in $Y$ if and only if $\pi^{-1}(V)$ is open in $X$. It's well-known that the restriction of $\pi$ to a saturated open subset $S$ in $X$ is also a quotient map. By a saturated subset we mean $S=\pi^{-1}(\pi(S))$.
The proof can go as follows. First, let $V$ be a subset of $\pi(S)$ and $\pi|_S^{-1}(V)$ is open in $S$. Since $S$ is open, $\pi|_S^{-1}(V)$ is also open in $X$. Because $S$ is a saturated subset, we have $\pi|_S^{-1}(V) = \pi^{-1}(V)$. $\pi^{-1}(V)$ is open in $X$ implies $V$ is open in $Y$, thus open in $\pi(S)$. The other direction, that $V$ is open implies $\pi|_S^{-1}(V)$ is open, comes from the fact that $\pi|_S$ is continuous as a restriction of a continuous map.
We have used the condition that $S$ is open in the proof. My question is very simple, what if we don't assume $S$ is open? Can anyone kindly give an example that the restriction of a quotient map to a saturated subset is not a quotient map?