Comments of Reviewer # 2 and response by the authors

 We would like to thank referees for taking the time to reviewour manuscript and thank the referee 2 for pointing out the potential problemsof three crystal structures. We have addressed your comments below and feelthat the work is greatly improved as a result of your input. In what followsthe referees’ comments are in black and the authors’ responses are in blue.

The structure of racemic 6 is more interesting. It seems torefine satisfactorily, with a conventional R of 6%, but the refinement wasconducted with nearly 5000 restraints! What is going on here?

 We understand the concern of the reviewer. The 5000 restraintsare primarily as a result of the applied default DELU and SIMU with no atoms specified(all non-hydrogen atoms are assumed) to facilitate the stable refinement of theanisotropic displacement parameters of the side chains. When the DELU and SIMUwere removed, displacement ellipsoids of some atoms on the side chains became somewhatelongated because of the dynamic disorders of these side chains. The list of“summary of restraints applied” was checked and did not show strong deviations,indicating these restraints were appropriately applied.

 There might be better restraint strategy which we do not foreseeto modeling atomic displacement parameters of the highly disordered side chains.

 When I examined the packing, it seemed to me that the entiremolecule was disordered across a mirror plane, but of course there is no mirrorplane in the chiral space group C2. I took the hkl file supplied in the CIF andran it through XPREP. There, the space group with the best figure of merit (byfar) is C2/m, which naturally contains the missing mirror plane. Structureslike this, where a molecule resides on a site with crystallographic symmetrythat the molecule itself cannot possess, are not at all uncommon. The trick isto carry out the refinement using a PART -1 command for the entire molecule,which is given half occupancy. The other half molecule is then generated bysymmetry. Such a refinement should eliminate the majority of the thousands ofrestraints used in the present work, even if the R is not lowered significantly.

After some consideration, the authors mayconclude that C2 is the correct space group for racemic 6 (it's notimpossible), but they need to justify this choice. Furthermore, the discussionof the crystal structures should at least mention the gross disorder present.

 Thank you for this suggestion. We agree that the space group C2/mfor this crystal structure is not impossible. However, our attempts tore-refine the structure using a PART -1 command in C2/m was unsatisfactory (R1= 12.61%). It is primarily the tert-butyl group that appears randomly orientedand unable to accommodate C2/m. Space group C2 describes the structure betterthan C2/m and the mirror symmetry is the pseudo-symmetry operator. The twomolecules are not quite related by a crystallographic mirror symmetry elements ofa higher-symmetric space group C2/m.