The results presented by Paris et al are very interesting, and add knowledge to the molecular basis of some pathogenic processes. The discovery of a bound flavin cofactor is truly novel, and the ability to produce the protein as recombinant protein is highly valuable, also for future studies.
The authors should consider to test some characteristics of the purified protein. It is indicated that NADH/NADPH is not bound to the protein. Yet, the measurement/data shown for this is not convincing (Fig 6E only shows a possible binding of NADH and NADP, and binding is just measured based on fluorescence) when one would like to study the dependence on nicotinamide cofactors. The authors should test whether the FAD cofactor is reduced by NADH or NADPH, using anaerobic conditions. The homologous enzymes (GGR and PHBH) all use NAD(P)H as reductant to generate a reduced FAD. It would be interesting to see whether the isolated holo protein can react with one of the cofactors.
Also, the figures suggest that the docked FAD binds in such a way that an arginine (R329) interacts with the benzylic part of the isoalloxazine ring. This is rather unusual for flavins in proteins. It should be checked whether flipping the isoalloxazine ring (rotate along N5-N10 axis) makes more sense.

Minor comments:
- correct co-factor > cofactor
- explain "no addressing sequence", unclear
- make clear whether the His-tagged protein was investigated concerning cytosolic location. It is unclear at the moment. The His0tag may influence solubility.
- "4-D architecture" should not be used
- "flavin-derived cofactor" is not correct, just use "flavin cofactor" (it is not a derivative)
- riboflavin is never used as cofactor (p. 8)
- "As redox conditions are important parameters": what does this mean??
- "participate in the joint chelation of an internal ion." rewrite, a flavin is not a metal