(This is based on Heller's simple proof, according to Rotman)
$A, B$ are a complex and its subcomplex of boundaries. Assume all the modules in $A, B$ are flat.
We have an exact sequence of complexes: $0\to Z \to A \to B[-1] \to 0$ and after tensoring $C$ where $C$ is any complex. $0\to Z\otimes C \to A\otimes C \to B[-1] \otimes C \to 0$.
The corresponding long exact sequence is:
$H_{n+1}(B[-1]\otimes C) \to H_n(Z\otimes C) \to H_n(A\otimes C)\to H_{n}(B[-1]\otimes C) \to ...$
Then $H_{n+1}(B[-1]\otimes C) = (B \otimes C)_n$ and $H_n(Z \otimes C) = (Z \otimes C)_n$ since $Z, B[-1]$ have zero differentials. Thus, we may rewrite the long exact sequence as:
$(B \otimes C)_n \to (Z \otimes C)_n \to H_n(A\otimes C)\to (B \otimes C)_{n-1} \to ...$
I got a bit confused. I think as a total complex, the differential of $Z \otimes C$ should be of the form $a_{p}\otimes b_{q} \mapsto \Delta'a_{p}\otimes b_{q} + (-1)^p a_{p}\otimes \Delta''_{q}b_{q}$. So $\Delta'=0$ but we still have $\Delta''$ here, so it should not be a zero-map...
Any help would be appreicated!