Understanding the chemical step in the catalytic reaction of DNA polymerases is essential for elucidating the molecular basis of the fidelity of DNA replication. The present work evaluates the free energy surface for the nucleotide transfer reaction of T7 polymerase by free energy perturbation/empirical valence bond (FEP/EVB) calculations. A key aspect of the enzyme simulation is a comparison of enzymatic free energy profiles with the corresponding reference reactions in water using the same computational methodology, thereby enabling a quantitative estimate for the free energy of the nucleotide insertion reaction. The reaction is driven by the FEP/EVB methodology between valence bond structures representing the reactant, pentacovalent intermediate, and the product states. This pathway corresponds to three microscopic chemical steps, deprotonation of the attacking group, a nucleophilic attack on the P-alpha atom of the dNTP substrate, and departure of the leaving group. Three different mechanisms for the first microscopic step, the generation of the RO- nucleophile from the 3'-OH hydroxyl of the primer, are examined: (i) proton transfer to the bulk solvent, (ii) proton transfer to one of the ionic oxygens of the P-alpha phosphate group, and (iii) proton transfer to the ionized Asp654 residue. The most favorable reaction mechanism in T7 pol is predicted to involve the proton transfer to Asp654. This finding sheds light on the long standing issue of the actual role of conserved aspartates. The structural preorganization that helps to catalyze the reaction is also considered and analyzed. The overall calculated mechanism consists of three subsequent steps with a similar activation free energy of about 12 kcal/mol. The similarity of the activation barriers of the three microscopic chemical steps indicates that the T7 polymerase may select against the incorrect dNTP substrate by raising any of these barriers. The relative height of these barriers comparing right and wrong dNTP substrates should therefore be a primary focus of future computational studies of the fidelity of DNA polymerases.