The latter approach has as advantages that precise calibration of protein concentrations in the two samples is not required, no long interscan delays are needed to ensure equilibrium, and no Lorentzian line shapes are required. Precise treatment of the intermolecular PRE effect as distance restraints necessitates knowledge of the exchange kinetics between free and bound states that averages the PRE effect. In addition, the tag might need to be explicitly modeled in the docking

process and its flexibility accounted for, either by increasing the error bounds, or, more properly, by ensemble averaging [38] and [39]. Alternatively, intermolecular PREs can be used in a more qualitative manner to map the binding interface [40]. An alternative method without the need for covalent attachment see more of the paramagnetic center is to use solvent PREs. Here, HIF inhibitor chemically

inert paramagnetic probes are added as co-solvents and cause relaxation and thus signal attenuation of solvent accessible protons [41]. Applied to protein complexes, solvent PREs can be used to quantitatively describe the distance of the observed nucleus to the molecular surface of the complex [42]. Paramagnetic lanthanide ions attached to a protein can also give rise to chemical shift changes, the so-called pseudocontact shifts (PCS). These depend on both the distance and relative orientation to the unpaired electron and may give long-range information up to 40 Å from the paramagnetic center [43]. Using a rigid, two-point anchored lanthanide tag, the possibility of obtaining not both distance and angular information between subunits has been shown to allow for efficient docking [44], [45] and [46]. Information on the relative orientation of subunits can also be obtained from residual dipolar couplings (RDCs) caused by incomplete averaging of dipolar interactions in anisotropic conditions [47]. Finally, cross-saturation methods can effectively be used to map binding interfaces by saturating protons in one subunit and observing the transfer of saturation

to non-overlapping protons in the deuterated observed subunit [48]. Here, we have focused on methods that provide information on the intermolecular interface within large complexes. It should be noted that complementary information on the bound-state conformation of the subunits may also be acquired using either backbone chemical shift prediction of dihedral angles [49], transferred NOEs [50] or cross-correlation experiments [51] and [52]. Overall, NMR provides the experimentalist with many options to obtain site-specific data, either at the atom or residue specific level, on the binding interfaces and structure of a complex. Other biophysical or biochemical sources of structural information that the experimentalist may turn to are listed in Table 3.