In this paper, we characterize strategy-proof voting rules when the set of alternatives has a multi-issue structure, and the voters' preferences are represented by acyclic CP-nets that follow a common order over issues. Our main result is a simple full characterization of strategy-proof voting rules satisfying nonimposition for a very natural restriction on preferences in multi-issue domains: we show that if the preference domain is lexicographic, then a voting rule satisfying non-imposition is strategy-proof if and only if it can be decomposed into multiple strategy-proof local rules, one for each issue and each setting of the issues preceding it. We also obtain the following variant of Gibbard-Satterthwaite: when there are at least two issues and each of the issues can take at least two values, then there is no non-dictatorial strategy-proof voting rule that satisfies non-imposition, even when the domain of voters' preferences is restricted to linear orders that are consist...