You'll find many references in pubmed: e.g.: http://www.ncbi.nlm.nih.gov/pubmed/15064475

1. Methods Mol Biol. 2004;261:445-68.

Computational prediction of protein-protein interactions.

Obenauer JC(1), Yaffe MB.

Author information: (1)Center for Cancer Research, Massachusetts Institute of Technology, Cambridge, MA, USA.

(..) A computational approach to predict protein interactions on a proteome-wide basis would therefore consist of identifying modular domains and sequence motifs from protein primary sequence data, creating sequence specificity-based algorithms to connect a domain in one protein with a motif in another in "interaction space," and then graphically constructing possible interaction networks

Using CO EVOLUTION is one way of predicting PPI, without use of any structures.

Have a look at http://www.ncbi.nlm.nih.gov/pubmed/21877284

You might give a try to one of the following web-servers:

• iFrag
• Struct2Net