A -Galactosidase-Based Bacterial Two-Hybrid System To Assess Protein-Protein Interactions in the Correct Cellular Environment Jimmy Borloo,* Lina De Smet, Bjorn Vergauwen, Jozef J. Van Beeumen, and Bart Devreese Department of Biochemistry, Physiology and Microbiology. Ghent University. Laboratory for Protein Biochemistry and Protein Engineering. K.L. Ledeganckstraat 35, 9000 Ghent, Belgium Received January 23, 2007 The vast majority of proteins functions in complex with one or more of the same or other proteins, indicating that protein-protein interactions play crucial roles in biology. Here, we present a -galac- tosidase reconstitution-based bacterial two-hybrid system in which two proteins of interest are fused to two non-functional but complementing -galactosidase truncations (ΔR and Δω). The level of complemented -galactosidase activity, driven by the protein-protein recognition between both non- -galactosidase parts of the chimeras, reflects whether or not the proteins of interest interact. Our approach was validated by reconfirming some well-established Escherichia coli cytoplasmic and membranous interactions, including well-chosen mutants, and providing the first in vivo evidence for the transient periplasmic interaction between Rhodobacter capsulatus cytochrome c 2 and cytochrome c peroxidase. We demonstrated the major advantages of this in vivo two-hybrid technique: i) analyses of interactions are not limited to particular cellular compartments, ii) the potential of using the system in mutation-driven structure-function studies, and iii) the possibility of its application to transiently interacting proteins. These benefits demonstrate the relevance of the method as a powerful new tool in the broad spectrum of interaction assessment methods. Keywords: protein-protein interactions • bacterial two-hybrid system • -galactosidase Introduction The elucidation of protein-protein interactions is subject to intense research and currently constitutes one of the highest priorities in the attempt to understand some particular biologi- cal processes. Many techniques have been developed to identify these interactions; among them are in vitro techniques such as pull-down analyses and co-immunoprecipitation, as well as in vivo strategies such as fluorescence resonance energy transfer (FRET). 1 These techniques have proven to be successful and offer major advantages such as sensitivity in the case of FRET and a pragmatic character when considering pull-down analyses. These techniques also bear serious drawbacks; the characteristic high sensitivity of FRET, for instance, results in significant background signals, whereas co-immunoprecipita- tion and pull-down analyses require specific antibodies and harsh elution conditions, respectively. A popular technique, developed in the late 1980s and early 1990s and nowadays often used in so-called ‘functional ge- nomics’, is the yeast two-hybrid system. 2,3 This technique is based on the coexpression of two chimeras or hybrid proteins, each consisting of a protein of interest and either a DNA binding domain or a transcription activation domain. When the two proteins of interest associate with each other, a fully functional transcription factor is formed and a reporter gene is transcribed, leading to an easily detectable phenotype. Since its development, the technique itself has been extensively improved and diversified, giving rise to one- and three-hybrid systems 4,5 as well as to reverse n-hybrid systems. 6 The yeast two-hybrid system paved the road for functional genomics, that is, the screening of an entire library of proteins (preys), encoded by a given genome, that may form complexes with a particular protein (bait). 7,8 Notwithstanding its genius, the yeast two- hybrid system has the restriction that all interactions should occur in the nucleus, which limits its use to cytosolic proteins. An alternative method for detecting protein-protein interac- tions is the use of enzyme complementation, rather than transcriptional activation. Different systems, including -lac- tamase, -galactosidase, and mouse dihydrofolate reductase complementation have been described for use in eukaryotic cells. 9-11 Bacterial two-hybrid alternatives have been developed as well. These systems are mostly based on the repression or activation of a reporter gene. 12,13 Although a majority of the bacterial two-hybrid systems are still restricted to analyses of cytoplasmic interactions, some have been developed that allow the analysis of periplasmic and membranous protein-protein interactions, such as the Vibrio cholerae cytoplasmic membrane- localized transcriptional regulator ToxR, 14 the Bordetella per- tussis adenylate cyclase reconstitution system, 15 and, analogous to the eukaryotic variant, the -lactamase reconstitution based * Corresponding author. Laboratory for Protein Biochemistry and Protein Engineering, Ghent University, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium. Phone: +32 9 264 51 26. Fax: +32 9 264 53 38. E-mail: Jimmy. Borloo@ugent.be. 10.1021/pr070037j CCC: $37.00  2007 American Chemical Society Journal of Proteome Research 2007, 6, 2587-2595 2587 Published on Web 06/01/2007