Matching the distributions of the marginals and the sums for the Meixner class

For a given set of independent random variables (r.v.'s) $X_1,\dots,X_d$ belonging to a given Meixner class, we seek r.v.'s $Y_1,\dots,Y_d$ such that the marginal laws and the laws of the sums match: $Y_i\stackrel{d}{=} X_i$ and $\sum_iY_i\stackrel{d}{=}\sum_iX_i$. We give a full characterization of the r.v.'s $Y_1,\dots,Y_d$ and propose extensions and practical constructions by means of finite mean square expansions.