**4. Conclusions**

Our analysis, carried out with CCCPP program on several crystal structures of CYP3A4, enabled identifying relevant ingress/egress channels, with their lining heavy atoms and residues. Our calculations support the hypothesis of channels 2a and 2f as major channels of substrate/product egress in CYP3A4, plus two secondary ones, 2e and S. We propose potential pathways of the ligands, inside these channels, toward the distal face of the heme together with information on the movements of the residues associated to the opening/closing of the channels.

Our analysis suggests that block 1 anchored in the membrane opens at ligand entrance and channel 2a is the only ingress/egress channel. Then, in the case of one large ligand or two ligands, block 2 opens. Smaller channels 2e and S could be involved in the egress of the metabolites, either

by enlarging or by use for circulation of water or dioxygen. We did not consider proximal channels smaller than distal channels.

Channels 2a and 2f are occupied by ligands that are not yet oxidized. Either channel 2a enlarges to accept the ligand, or a new path (channel 2f) opens due to the nature and/or the large size of the ligand [60]. Residues obstructing the channel create a bottleneck. These changes are influenced by the location of the channels with respect to the protein topology and the protein's overall global motion [37,38]. Channel 2e exits near the cytosol (see Figure 6 in [27]). Channel 2a is opened in the apo structure, although channel 2f is not. At input of the ligand, channel 2a enlarges until reaching a critical value, above which other paths are needed to accept the entrance of more ligands. When the second ligand enters in channel 2f, its orientation is the opposite of the one of the first ligand, thus fitting the weaker hydrophobicity of channel 2f which starts at the cytosol/membrane interface. These movements are located at flexible secondary structures such as B-C and F-F' loops and C-terminal loop [64]. Molecular dynamics simulations could help to know if channel 2f is closed in the apo form then opens to accept the ligand, or if it is already opened. The F-G block acts as a multi-hinged lid on the distal side of the protein and many channels border it, permitting an opening at the membrane's interface or on the cytosol.

The major channel lining residues mentioned in our study, which we suggested to have a role in channel opening, were proved to be involved in ligand binding, in the activity or cooperativity of CYP3A4 and to be key residues governing allosteric processes in P450 catalyzed substrate oxidations [92]. Another important structural element is the B-C loop, which borders channels 2a and 2e.

Although molecular dynamics simulations were able to exhibit channels not visible by a rough examination of crystallographic structures [94], it was possible to detect such channels by geometric methods because the ligands were indeed present there. A major use of our results could be to provide pertinent starting points for molecular dynamics computations to observe the opening and closing of the channels.

**Author Contributions:** The main results came from the doctoral work of L.B., realized under the co-supervision of F.M. and M.P., with the collaboration of F.A. and B.T.; L.B. and M.P. wrote the original draft; M.P. produced the software CCCPP; and F.A. and G.M. critically reviewed the manuscript. All authors approved the final version of the text.

**Funding:** The early stages of this work were funded by the IBC company (Integrative BioComputing, Rennes, France). The final part of this work was funded by the Maghreb edition of the programme *L'Oréal-UNESCO For Women in Science*, which awarded L.B. in October 2016.

**Acknowledgments:** L.B. gratefully acknowledges Pridi Siregar, founder and manager of the IBC company, for supplying financial support.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study, in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.
