Dear Isa, Do you have any cysteines? I assume so given you mention class B metals... If so, and you have a ready supply of relatively isomorphous crystals, I would consider giving native SAD a go, collecting datasets from 5-10 (or however many you need) datasets at a reasonably low energy (say 7keV) and merging.
If you have any ordered phosphates they would help here too, of course. You could also try a brief soak in a solution where the PO4 is partially or completely replaced by WO4, which should give a whopping anomalous (and potentially isomorphous) signal. Otherwise yes, Hg or Pt reagents would probably be your best bet. Anionic Gd complexes with chelating ligands will probably not bind so well in the presence of a high concentration of phosphate. We have also had a bit of luck with iodination - I like to dissolve a few crystals of iodine in paraffin oil (gives a beautiful purple oil), and use that to introduce iodine by vapor diffusion into the drop containing the crystals (be careful with this - preferably manipulate in a fume hood, as iodine vapor I imagine is not very good for the lungs). Good luck! Best, Oliver.
