Good question. N-glycosylation occurs in the ER lumen, so you'd need A) a signal-peptide-independent translocation into that compartment or B) a different method (e.g. dolichol-independent, etc.) of transferring glycans onto your asparagine residue. I am unaware of either, but would be extremely interested if there were!!
Just to add to the answer of Michael, there are cases of chloroplast-based N-glycosylated proteins (they probably ended there via Golgi-to-chloroplast transport). A useful review on evolution of protein N-glycosylation across major model organisms is a recent paper in Mol.Cell: http://www.cell.com/molecular-cell/abstract/S1097-2765%2812%2900355-3?script=true
They list there some "odd" cases, which perhaps will be useful to you.