A powerful method to detect protein interactions in mammalian cells
Tanya Hosfield Cathy Chang
With Stratagenes mammalian two-hybrid assay kit, hybrid genes are used to detect protein-protein interactions via the activation of reporter gene expression. This expression occurs as a result of reconstitution of a functional transcription factor caused by the association of two hybrid proteins. It allows researchers to study interactions in vivo between mammalian proteins that require posttranslational modification or external stimulation not present in lower eukaryotes.
Protein-protein interactions in mammalian cells are the basis of virtually every cellular process, such as DNA replication, transcription, translation, splicing, secretion, cell cycle control, signal transductions, and intermediate metabolism.1 As efforts to sequence whole genomes result in data for vast numbers of new proteins, there is a corresponding need to identify interactions among these proteins. Methods used to discover these interactions will be extremely helpful in elucidating the in vivo function of newly discovered genes and, in turn, will aid in further understanding known protein functions.
One of the most powerful methods for studying protein-protein interactions is the yeast two-hybrid system developed by Dr. Stanley Fields and colleagues.2 This system has been highly useful for detecting and identifying protein-protein interactions in vivo. The Fields two-hybrid system, and the many second-generation versions currently used,2,3,4,5 exploit the modular nature of a transcriptional activator. Transcriptional activators, such as the GAL4 protein of yeast, contain a DNA binding domain and an activation domain that can be separated. If these domains are then reconstituted in trans, their acti vity is restored. Therefore, by creating genetic fusions between the gene of interest to one domain and its interacting partner to the other domain, interaction of the two proteins results in restoration of the activation of the transcriptional activator. Genetic selection for this interaction is provided by an amino acid biosynthetic enzyme that is absent in the yeast host and whose expression becomes induced only when the transcription activator is reconstituted. This two-hybrid system has been widely used.
Despite these successes, the Fields yeast two-hybrid system does not represent a universal system for gene discovery by protein-protein interactions. There are significant limitations to the Fields method that make research into developing experimental alternatives to this procedure a high priority. A potentially significant limitation is the host organism, yeast, that is used for identifying protein-protein interactions. Although the yeast is a eukaryote, it is far removed from human, mammalian, and higher eukaryotic organisms. For example, yeast does not carry out some of the posttranslational modifications (e.g., most tyrosine-phosphorylation events) that are responsible for important protein interactions in mammalian cells.5,6 In addition, the environment of a yeast cell is so different from higher eukaryotes that uncertainties about the proper folding of these proteins are valid.7 Finally, a number of proteins of interest have been discovered to be toxic when expressed in yeast.7 This further underscores the limitations of using a single host organism to study eukaryotic proteins. Clearly, developing a system to study protein-protein interactions using a mammalian host organism would obviate many of these concerns and provide an attractive method for uncovering the biological role of many gene products that cannot be studied in the yeast. The mammalian two-hybrid syst em first developed by Dr. Dang and colleagues8 provides an excellent alternative to the yeast-based system because this assay is performed in mammalian cells. Since mammalian proteins are more likely to retain their native conformation in a mammalian host, the results would more likely represent biologically significant interactions.
To determine the capability of the mammalian two-hybrid assay kit for studying protein-protein interactions, we constructed a set of user-friendly GAL4 BD (binding domain)-bait and NF-kB AD (activation domain)-target plasmids. These vectors were used to validate the system with known pairs of interacting proteins in various mammalian cell lines. The working principle of the mammalian two-hybrid assay kit is similar to that used in Fields yeast two-hybrid system, except that the assay detects transcriptional activation of a reporter gene (such as luciferase, SEAP, or b-galactosidase) located downstream of the multiple GAL4 DNA-binding sites. The GAL4 BD-bait, NF-kB AD-target, and a reporter plasmid are cotransfected together into a mammalian cell. The protein of interest, or bait, is expressed as a fusion protein with the GAL4 DNA binding domain, and the target protein is expressed with the activation domain. If the bait and target proteins interact, the activation domain is brought together with the DNA binding domain and together they activate transcription of the reporter gene (Figure 1).
The pCMV-BD and pCMV-AD vectors are designed for the construction and expression of gene fusions with the GAL4 DNA binding domain and the NF-kB transcriptional activation domain, respectively (Figure 2). The pCMV-BD vector is designed to construct bait plasmids. This vector contains DNA encoding amino acids 1 to 147 of the GAL4 gene (DNA binding domain) and unique 3 cloning sites. The pCMV-AD vector is used to construct target plasmids. This vector contains a DNA encoding nuclear localization sequence from the SV40 large T-antigen (amino acid PKKKRKV), amino acids 364 to 550 of the mouse NF-kB gene (transcriptional activation domain), and unique 3 cloning sites. Both the pCMV-BD and pCMV-AD vectors contain the ColE1 origin for replication in E. coli. The pCMV-BD vector contains the kanamycin-resistant gene and the pCMV-AD vector contains the ampicillin-resistant gene. The CMV promoter### governs the expression of the bait and target proteins in both the pCMV-BD and pCMV-AD vectors. The SV40 poly (A) signal provides the signals necessary for transcriptional termination and polyadenylation of the bait and target genes in mammalian cells.
Stratagene offers this system as a vector kit containing both the bait vector, target (prey) vector, the pFR-luc reporter vector, and three control vectors. All plasmid vectors are available separately.
The transcription activation domain we used for pCMV-AD is derived from the carboxyl terminus of the mouse NF-kB p65 protein.9 The p65 activation domain has been reported to be significantly more potent than that derived fr om the herpes virus VP16 protein.10 In addition, whereas overexpression of transcription factors containing the activation domain from VP16 can be toxic,11 overexpression of the activation domain from NF-kB p65 has not been reported to be toxic.10,12
It was demonstrated that the transcriptional-activation domain of NF-kB p65 is located in its carboxyl terminus.9 To determine the region of p65 that has the highest transcription activity, various segments of the mouse NF-kB p65 protein were PCR amplified. The PCR products of the NF-kB p65 segments (amino acids 283 to 550, 364 to 550, and 519 to 550) were subcloned to make pGAL4 BD-NF-kB 478, pGAL4 BD-NF-kB 476, and pGAL4 BD-NF-kB 479, respectively (Figure 3, Panel A).
To map domains required for maximum transactivation, we constructed chimaeric p65 proteins that contain at their N-termini the DNA binding domain of the GAL4 protein (Figure 3, Panel A). The transactivating potential of the GAL4 BD-p65 fusion proteins was tested by cotransfection with a luciferase reporter construct that contains the luciferase gene downstream of five GAL4 binding sites (pFR-Luc: PathDetect reporter plasmid). Our results indicated that sequences of p65 between amino acids 364 and 550 gave the highest luciferase activity and were even higher than the activity obtained with the GAL4 BD-VP16-fusion protein (Figure 3, Panel B). Therefore, this segment (amino acids 364 to 550) of NF-kB p65 was used as the activation domain in pAD.
The ability of the mammalian two-hybrid assay kit to detect protein-protein interactions was verified by the expression of p53 and SV40 large T-antigen. Amino acids 72 to 390 of murine p53 were fused in frame to the carboxyl (C) terminus of the GAL4-DNA binding domain (pBD-53), and amino acids 84 and 708 of the SV40 large T-antigen were fused to the C-terminus of NF-kB p65 activation domain (pAD-SV40T). As negative controls, amino acids 218 to 277 of CD40, 13 which does not interact with SV40T, were fused to the C-terminus of GAL4 DNA binding domain (pBD-CD40), and amino acids 297 to 503 of TRAF-2,13 which does not interact with p53, were fused to the C-terminus of NF-kB p65 activation domain (pAD-TRAF). These pairs of plasmids were cotransfected with the reporter plasmid pFR-Luc into various cell lines, and their luciferase activities were monitored. The cell lines used were CHO, COS, HeLa, 293, and HLR (HeLa luciferase reporter). Figure 4 shows data for luciferase expression in the various cell lines; transfection with the combination of plasmid pBD-53 and pAD-SV40T, whose expressed fusion proteins interact in vivo, resulted in strong activation of luciferase gene expression. Conversely, transfections with the combination of plasmids pBD-CD40 and pAD-SV40T or pBD-53 and pAD-TRAF2, whose expressed fusion proteins do not interact in vivo, displayed low background luciferase activity. Only combinations of pBD-53 and pAD-SV40T dramatically activated luciferase gene transcription. Expression of the GAL4 BD-p53GAL4 BD-CD40 fusion proteins and the NF-kB AD-SV40TNF-kB AD-TRAF2 fusion proteins were confirmed by Western blot analysis using anti-GAL4 DNA-binding domain monoclonal antibody (Santa Cruz Biotech) and anti-NF-kB polyclonal antibody (Santa Cruz Biotech), respectively (data not shown). These results demonstrated that the mammalian two-hybrid assay kit detects known interaction between p53 and SV40 large T-antigen.
In addition to the reporter plasmid pFR-Luc, other PathDetect reporter plasmids such as pFR-bGal (b-galactosidase) and pFR-SEAP (secreted alkaline phosphatase) were also used in the mammalian two-hybrid assays. Pairs of plasmids were cotransfected with reporter plasmids pFR-bGal or pFR-SEAP into CHO cells. Forty-eight hours after transfection, cells lysates were collected and assayed for reporter gene expression (b-galactosidase and secreted alkaline phosphatase).
We demonstrated our systems ability to detect known protein-protein interactions, such as p53 and SV40 large T-antigen in vivo in mammalian cells. In addition, we showed that various reporter genes, such as luciferase, b-galactosidase, and secreted alkaline phosphatase can be used to detect interactions. The mammalian two-hybrid assay kit is especially useful in confirming suspected interactions between two proteins or interactions identified by yeast two-hybrid screens. This system is the most practical method for studying interactions between mammalian proteins in their native environment because they are more likely to represent biologically significant interactions than interactions observed in a foreign environment. The confirmation eliminates the possibility of a false positive that is an artifact from working in yeast cells or in vitro. In addition, once an interaction between two proteins has been detected, the mammalian two-hybrid assay can be used with deletions or site-directed mutagenesis to identify domains and amino acids involved in specific protein-protein interactions.
The authors thank Ning Jiang, Xu Li, Lisa Hexdall, Katherine Felts, Chao-Feng Zheng, Peter Vaillancourt, Carsten Carstens, Alan Greener, Mary Buchanan, and John Bauer of the Genetic Systems group at Stratagene for reagents, discussions, suggestions, and excellent jokes.
Phizicky, E.M. and Fields, S. (1995) Microbiological Review 59: 94-123.
Fields, S. and Song, O.-K. (1989) Nature (London) 340: 245-246.
Chien, C.-T., et al. (1991) Proc. Natl. Acad. Sci. USA 88: 9578-9582.
Fields, S. and Sterngianz, R. (1994) Trends Genetics 10: 286-292.
Osborne, M.A., et al. (1997) In P.L Bartel and S. Fields (Eds.) The Yeast Two-Hybrid System. Oxford University Press. New York, New York.
Buckholz, R.G. and Gleeson, M.A.G. (1991) Biotechnology 9: 1067.
Bartel, P.L. and Fields, S. (1995) Methods in Enzymology 254: 241-263.
Dang, C.V., et al. (1991) Mol. Cell. Biol. 11: 954-962.
Schmitz, M.L. and Baeurle, P.A. (1991) EMBO J. 10: 3805-3817.
Rivera, V.M. (1998) Methods: A Companion to Methods in Enzymology 14: 421-429.
Shocke tt, P., et al (1995) Proc. Natl. Acad. Sci. USA 92: 6522-6526.
Rivera, V.M. (1996) Nature Med. 2: 1028-1032.
Mullinax, R. and Sorge, J. (1995) Strategies 9: 81-83.
Yang, T.T., et al. (1996) Nucleic Acids Res. 24: 4529-4593.
Cormack, B.P., et al. (1996) Gene 173: 33-38.