Let me write it out more explicitly. Define a linear mapping $T : E \to \mathcal B(F, G)$ by
$$ x \mapsto a_x,$$
where $a_x \in \mathcal B(F, G)$ is defined by
$$ a_x(y) = a(x,y).$$
Notice that $a_x$ is indeed in $\mathcal B(F, G)$, since the mapping $y \mapsto a(x,y)$ is continuous for each $x \in E$.
Let $B(1) \subset E$ be the unit ball in $E$. We are going to show that $T(B(1))$ is a bounded subset of $\mathcal B(F, G)$.
By the uniform boundedness principle, it is only necessary to show that for every $y \in F$, the set
$$\{ T(x)( y ) : x \in B(1) \} = \{ a(x,y) : || x || \leq 1 \}$$ is a bounded subset in $G$. This statement is true, because the mapping $x \mapsto a(x,y)$ is continuous for every $y \in F$.
Having shown that $T(B(1))$ is bounded, we deduce that there exists a $c$ such that
$$ || x || \leq 1 \implies ||a_x || \leq c.$$
Therefore,
$$ || a_x || \leq c ||x ||$$
for all $x \in E$.
Finally,
$$ || a(x,y) || \leq ||a_x || ||y || \leq c || x || \ || y ||.$$