Nitrogenase is one of the most fascinating enzymes in nature, being responsible for all biological nitrogen reduction. Despite decades of research, it is among the enzymes in bioinorganic chemistry whose mechanism is the most poorly understood. The MoFe protein of nitrogenase contains an iron-molybdenum-sulfur cluster, FeMoco, where N-2 reduction takes place. The resting state of FeMoco has been characterized by crystallography, multiple spectroscopic techniques, and theory (broken-symmetry density functional theory), and all heavy atoms are now characterized. The cofactor charge, however, has been controversial, the electronic structure has proved enigmatic, and little is known about the mechanism. While many computational studies have been performed on FeMoco, few have taken the protein environment properly into account. In this study, we put forward QM/MM models of the MoFe protein from Azotobacter vinelandii, centered on FeMoco. By a detailed analysis of the FeMoco geometry and comparison to the atomic resolution crystal structure, we conclude that only the [MoFe7S9C](1-) charge is a possible resting state charge. Further, we find that of the three lowest energy broken-symmetry solutions of FeMoco, the BS7-235 spin isomer (where 235 refers to Fe atoms that are "spin-down") is the only one that can be reconciled with experiment. This is revealed by a comparison of the metal-metal distances in the experimental crystal structure, a rare case of spin-coupling phenomena being visible through the molecular structure. This could be interpreted as the enzyme deliberately stabilizing a specific electronic state of the cofactor, possibly for tuning specific reactivity on specific metal atoms. Finally, we show that the alkoxide group on the Mo-bound homocitrate must be protonated under resting state conditions, the presence of which has implications regarding the nature of FeMoco redox states as well as for potential substrate reduction mechanisms.