This is not related to sheafififcation. The sheaf $\mathbb{C}^{*}$  of locally constant functions on $M$ is already a sheaf,  so sheafification will not change it. 

This  sequence is not an exact sequence of complexes but it is an _exact triangle_ of complexes. That is - it is an exact sequence of complexes, up to quasi-isomorphism.  The obvious short  exact sequence of complexes is 
`\[
0 \to \mathbb{C}^{*} \to \left[\begin{array}{c} \underline{\mathbb{C}}^{*} \\ \downarrow \\ A^{1}_{M} \end{array}
\right] \to \left[ \begin{array}{c} A^{1}_{M,cl} \\ \downarrow \\ A^{1}_{M} \end{array}\right] \to 0 .
\]`
Now note that the last complex has an obvious surjective map 
`\[
\left[ \begin{array}{c} A^{1}_{M,cl} \\ \downarrow \\ A^{1}_{M} \end{array}\right] \to \left[ \begin{array}{c} 0 \\ \downarrow \\ A^{2}_{M,cl} \end{array} \right]
\]`
and that this surjective map is a quasi-isomorphism. So up to a quasi-isomorphism, you can replace the last term in the short exact sequence with the complex you wanted.