The electron density is disconnected for the chlorophenyl group of AR-R17477 in complex with nNOSoxy and iNOSoxy. Therefore, several factors had to be taken into account to place the chlorophenyl group in a consistent way. Here are these factors:

1. The electron density is well defined up to the amino group of the inhibitor in nNOSoxy complex and the aliphatic link between the phenyl and the amino groups in iNOSoxy complex. Therefore, the conformational space for the chlorophenyl containing part of the structure can be reduced to 2 and 3 rotational degrees of freedom, respectively. This conformational space was investigated to reveal the rotation ranges that do not produce steric hindrance with other parts of the protein and the inhibitor structure. The final conformations of the chlorophenyl containing part of AR-R17477 fulfill this criterion.

2. There are relatively strong peaks of electron density (4s in case of nNOSoxy and 5s for iNOSoxy on the F_o - F_c map) that can be assigned to the chlorine atom because these peaks are approximately twice as high as the water peaks (number of electrons in the chlorine is 17 versus 8 electrons in oxygen). Some of these peaks are shown in Fig. 6 illustrating the electron density for the chlorophenyl group.

3. Other peaks of the electron density next to the chlorine ones have protruding shape and if a water molecule is placed in these peaks, then its distance to the chlorine atom happens to be too short (up to 2.5 Å). On the other hand, going down to s values, the shape of these peaks becomes more like the density for part of the phenyl ring (see Fig. 6 illustrating the electron density for the chlorophenyl group). However, at lower s , the map is getting more noisy obscuring other important details; this is why we do not use lower s map in Fig. 1.

4. There are no other electron density peaks in the substrate access channel fitting the whole or at least a part of the chlorophenyl group.

5. On the other hand, placing the chlorine atom in the strong peaks of electron density allows to obtain a good description of the neighboring protruding peaks automatically. These peaks nicely fit the part of the chlorophenyl group, and the resulting conformations make no steric hindrance with other atoms of the complex structure. Besides, these conformations are fully consistent with chemical criteria and make no interactions that would be energetically unfavorable.

6. Another evidence supporting our interpretation is the results of several rounds of x ray crystallographic refinement involving molecular dynamics procedures that always converged to the conformations presented in Figs. 1, 2, and 4.

Thus, careful consideration of the above arguments has lead to the current model of the chlorophenyl group of AR-R17477 in the nNOSoxy and iNOSoxy complexes, which (as we mentioned in the text) corresponds to the equilibrium states of the inhibitor. We think that the coordinate error for each conformation of the chlorophenyl group in iNOSoxy complex is not significantly higher than 0.4 Å, the average coordinate error of the complex structure calculated from the Luzzati plot. For the atoms connecting the amino group and the chlorophenyl group (including the amino group itself in iNOSoxy complex), where no electron density is observed, the coordinate error is higher.