PROC GLIMMIX will accept only a single variable as a dependent variable in the model statement. It does not recognize the 'hyphen' syntax used in PROC GLM.
If you truly wish to do a multivariate response analysis, first check out Example 44.5. Joint Modeling of Binary and Count Data in the SAS/STAT13.2 documentation (it has a different example number, but the same content in earlier versions). Then read the posts by in this forum regarding multiple response variables. It is possible to do it, and to account for the correlations between the response variables. It is just not easy. For now, try fitting each separately.
Steve Denham