Under reducing conditions interactions between two polypeptides is disrupted. However, under non reducing conditions, the interactions are preserved. Thus, two conditions allow us to identify protein-protein interactions. In the paper, this is what authors tried to determine.

Reducing conditions in this context means using beta-mercaptoethanol (2-ME) or dithiothreitol (DTT) to reduce disulphide bridges in proteins so that they can adopt random coil conformation and better separate by size in SDS-PAGE gels. This is a frequently used approach in protein electrophoresis for their detection by Western blotting. However, certain antibodies cannot detect proteins in their reduced forms and in that case one uses non-reducing conditions, which means  not using 2-ME or DTT in a gel loading buffer.

Thank you!

Any why is like denaturation. 

Then we can say that in reducing condition the protein contains 3 sub-units (51Kda).

In non reducing condition, the have the form active of the protein withe 150Kda  

Is this true?  

Correct. If  disulfide bond tethers together two polypeptide chains (e.g. in insulin), the reducing conditions allow subunit separation. Disulfide bonds can also form within single polypeptide chain (e.g. ribonuclease A) - in that case the reducing conditions allow protein to achieve unfolded structure when denatured by heat in the presence of SDS.

Yes indeed, Basically the disulfide bridge is not being disturbed. Thereby allowing intermolecular interactions.  between two monomers. Hence, your protein would be comprised by two monomers e.g.(dimer)

ْThank you all for the best information.