zinctest.mol has weird charge on nitrogen: the nitrogen is charged, but lacks an explicit hydrogen. Moreover, the geometry is planer and the nitrogen participates in a aromatic ring system. Quite unreasonable I think.
I suggest removal of the positive charge, or replacement by a newer entry from ZINC.
BTW, this should affect the CDK-ZINC test set too, from which it is taken.