Would it be a good idea to include the presence of Met in the heuristic rules for the exclusion of non-suitable peptides? I would be much more concerned about Met than about 3+ charged peptides (assuming that I can control for the latter through the inclusion of internal standards).
Another heuristic rule that I would find very sensible would be to exclude N-glycosylation sites NX[ST] - their presence would potentially mess up any SRM-based assay.