It is difficult to understand whether the siRNA experiment is poorly explained in this thread or poorly conceived. In general, GEFs and GAPs are upstream of small GTPases. Therefore, knockdown of the GEF using siRNA and measuring Rac1 or cdc42 activation makes sense.

But the converse experiment in which you knockdown the GTPases is problematic. If you think about it, cells deficient in Rac1 will not make lamellipodia, and therefore you won't observe anything localizing to this type of structure, because it no longer exists. So you don't really learn anything meaningful about a direct protein-protein interaction between the GEF and the GTPase from this type of experiment. I think this is the reason that both David and Yan Cai were critical.

You can perform a GTP-Rac1 assay and GTP-Cdc42 assay with or without your GEF. If you see increase of Rac1 activity (as revealed by GTP-bound Rac1), then you can claim it's a GEF for Rac1. You can do the same for Cdc42. It's a pulldown assay with WB analysis. You can either purifiy the GTP-bound active domain by yourself or buy the kit available. See this link: http://www.cytoskeleton.com/bk035

Thanks for you answer. If I co-transfected the cells with  Rac1 siRNA + GFP-GEF1, I observe an effect on GFP-GEF1 localization. Can I conclude that GFP-GEF-1 activates Rac1? Rac1 can be activated by TIAM1, VAV2, and other GEFs.

Greeting

Since Rac1 and Cdc42 participate in a positive feedback loop (at least in neurons), the effect you described on GEF localization could imply that a Rac1 effector controls "GEF1" localization, not necessarily that your GEF acts on Rac1. I believe you should try the GST pull-down assay that Jacque suggested and also check if your GEF physically interacts with dominant negative Rac1 or Cdc42.  

Regards

Which FRET assay did you use? Is it reporting on GTPase conformation or simply nearby proximity of the GTPase to your GEF?

In general, for GTPase activation, pull-down assays, or Cytoskeleton's G-LIZA assays are the gold standard. Other surrogates are to look for phenotypes in your GEF overexpressing cells that are consistent with activated GTPases, for example lammelipodia formation for Rac1, and filopodial formation by cdc42. In addition, PAK activation can be readily assays by Western blotting, and often correlates with Rac1 activation.

We used Raichu Rac1, cdc42 FRET biosensor. I think it is simply nearby proximity of the GTPase to my GEF because I used rapamycin treatment to recruit my GEF to plasma membrane. I also did TIRF image to confirm the activation of GEF with Rac1 by observing PAK1 fluorescene level. How about Cdc42, can I do with WASP to make sure my GEF activates cdc42? I tried pulldown assay but I could not generate purified protein because GEF is located in the pellet.

The Raichu FRET construct reports directly on protein conformation. If you observed a robust change in FRET, I think you can draw a solid conclusion that your GEF affects Rac1 and/or cdc42. A pull-down assay to confirm physical interaction of the GEF and a target GTPase is non-trivial, because these interactions can be quite transient, and as you mention, GEF-GTPase binding may occur in a subcellular fraction that is not readily solubilized. The pull-down assay that both Jacques and I endorsed is to measure the total amount of active cdc42 or active Rac, not the GEF-bound fraction.

Do you have reason to expect that the GEF is selective for Rac vs cdc42? It would not be surprising if it regulated both.

You should read&think more about the experiments you are doing, and try to understand the principles. I really don't understand why you did the siRNA experiments.

Thanks Jacques and Daniel for your advices. I will try pull down assay to measure the total amount of active cdc42/Rac. The GEF I am working on localizing in the leading edge so I expect this GEF activates both Rac1 and cdc42.

@Yan cui: I do siRNA is just to confirm that GEF-1 activates Rac1 and cd42. It is not a good experiment? This experiment could not prove anything?

@Jacques and Daniel: Do you agree with Yan cui that I should not use siRNA?

Moreover, David menttioned that if I use siRNA Rac1/Cdc42, it only imply that the effectors downstream of Rac1/Cdc42 control GEF. From David and Daniel suggestions, I also test with PAK1 and WASP to confirm that GEF-1 activates Rac1/ cdc42

It is difficult to understand whether the siRNA experiment is poorly explained in this thread or poorly conceived. In general, GEFs and GAPs are upstream of small GTPases. Therefore, knockdown of the GEF using siRNA and measuring Rac1 or cdc42 activation makes sense.

But the converse experiment in which you knockdown the GTPases is problematic. If you think about it, cells deficient in Rac1 will not make lamellipodia, and therefore you won't observe anything localizing to this type of structure, because it no longer exists. So you don't really learn anything meaningful about a direct protein-protein interaction between the GEF and the GTPase from this type of experiment. I think this is the reason that both David and Yan Cai were critical.

Thank you Daniel~

If you want to identify the GTPase(s) which are directly activated by your new GEF, one way will be to measure nucleotide exchange activity for each GTPase. These are in vitro assays were you mix your GEF + Rac1 (or RhoA or Cdc42) and you follow the exchange of GDP for GTP. You can use some commercial kit from cytoskeleton (same company which was suggested by Jacque for the GTPase activity pulddown assay): http://www.cytoskeleton.com/bk100

Good luck with your experiments!

@Christophe: I did the experiment as you mentioned. We called MANT GTP/GDP essay but it is very difficult for us to do this experiment. We tried to purify our protein but the protein locates mostly in the pellet. We induced our GEF in 20 Liters of medium to get enough protein. Then, we tried to test one time with MANT GDP/GTP but it didnt work.

Most of the GEFs are at an auto-inhibited conformation in the full-length proteins. Thus you may try to purify the DH-PH domain of your GEF, then use the DH-PH domain to do the MANT GTP/GDP assay.

It may be much easier to purify the DH-PH domain from Ecoli than the full-length protein.

For the in vivo experiments by the active GTPase pulldown strategy, you can also compare the full-length GEF and the DH-PH domain.

My suggestion to you to over-express the GEF in mammalian cells and perform GST-effector pull down assay using the kits that available from many companies such as ThermoScientific, Cytoskeleton etc [http://www.jbc.org/content/289/48/33378.long]. Then you can confirm the result by assessing the effect of siRNA-mediated down-regulation of the GEF on Rac/Rho/CDC42 activation.

I also think that the easiest solution is performing the GST pull-down assay. The GEF binding domain of Rac/Cdc2 should be rather constant and allow interaction with other possibly unknown GEFs. This in association with other evidence should provide a  reliable proof that Rac and CDC42 are the downstream effectors of your GEF.

@Antimo Migliaccio： The GEFs/small GTPases interaction is rather weak and transient, it may be not possible to detect such interaction by pulldown approach.

Venkateswarlu Kanamarlapudi's suggestion is more realistic.

If you want to go further to prove the direct interaction, more sensitive biochemical approach is required to measure the affinity, or measure the in vitro GEF activity of the DH-PH domain. 

Of course you need to make some trials to set up the best condition, but it can work

@Antimo Migliaccio： I am not very sure I understand your post correctly. Just clarify my opinion:

It is easy to demonstrate the interaction between Active small GTPases and the GTPase binding domain of their effectors, e.g., Rac1.GTP and PAK1-PBD.

It is difficult to demonstrate the interaction between small GTPases and their GEFs by pulldown.

You're right, but I didn't say that it is easy but it think that it is possible and should be tried, and demonstrating interaction between a GEF and small G-protein would strongly suggest that that G-protein is an effector of the GEF under investigation.

@Antimo Migliaccio：To do that, you need find the right buffer condition (very gentle not to disrupt the interaction, and efficient to reduce non-specific binding), also you may have to use dominant negative version of the small GTPases which are able to sequester the GEF to stablize the interaction, like Ras S17N. 

Hi there, I would certainly check out the known phenotypes of those small gtpases by knocking down or over expressing the GEF that you found. 

Adam