yeast two-hybrid kits and screening services - (Dec/04/2008 )

I recently switched from a yeast molecular bio lab to a mammalian cell bio laboratory. The new lab is currently looking to do a yeast two-hybrid (Y2H) screen. Our protein of interest is a ~50 kd integral membrane protein but we are mostly interested in its ~5kd cytoplasmic domain (as this is the domain we think is mediating the interactions). Although I do have experience in yeast biology, I have never actually performed a two-hybrid screen myself.

I was wondering if anyone here has had experience in doing a Y2H screen and if they could advise me on which commercial kit (or what lab's vectors/strains) to use. Or if they think that I would be better off contacting a company to perform a custom screen and if they can advise me on which company?

Currently, I think Clontech's Matchmaker 3 and Invitrogen's ProQuest with Gateway technology are the two commercial kits that best meet the lab's needs regarding the protein we wish to screen.

I appreciate any input any one can give me on the Y2H screen, including any pitfalls I may encounter.

Thanks!

Really, all the commercial kits are the same. They all work on the same concept of an interaction bringing the DNA binding domain and transactivation domain of a transcription factor together close enough to allow transcription of a reporter cassette. Sure, they may have slight variations but essentially they are the same. I've worked with Matchmaker as well as the bacterial 2-hybrid from Stratagene. Bacterial is faster but you don't get post-translational modifications, which may be necessary for the interaction to occur. However, I did the bacterial screen using a 20-amino acid bait. We wanted to find interactions specifically mediated by this small domain. I then did a secondary screen with a mutant of this domain to eliminate non-specific interaction.

I just warn you that 2-hybrids are notoriously dirty. You must verify the interaction biochemically (IP and pulldown). I spent a full year chasing interactions from the 2-hybrid which turned out to be false-positives. A postdoc in my lab spent 2 years chasing interactions that were not true. I once gave a lab-talk entitled "friends don't let friends do two-hybrids". It did eventually pay-off though. I did isolate and publish a true interaction from the hybrid but it took quite awhile. You might be better off doing a huge co-IP followed by mass spec if you have the facilities available. This way, if it works, you know you have a legitimate interaction from the very beginning. This method,however, is more difficult and more expensive. If you do go the 2-hybrid route, don't worry. Decide on a kit and the instructions are easy to follow. Just be prepared to spend sometime verifying the interaction in alternative assays since no one will believe just the 2-hybrid alone.