Besides selecting for the presence of the two plasmids based on leucine and tryptophan selection, we employed a high stringency selection of positive protein interactions by using 3 more reporter genes to eliminate potential false positives: histidine, adenine, and Mel1, which encodes alpha-galactosidase for blue/white screening. The Catharanthine sulfate growth of colonies of the diploid yeast cells with a blue staining Butenafine hydrochloride within 3 weeks after plating was scored as a positive pairwise interaction between the two proteins tested while other results were scored as a negative interaction. There was no significant difference in the growth of the diploid yeast cells representing all the positive interactions. Two-hybrid screens may generate a significant number of false positives that represent random generation of histidine-positive colonies. This is possibly due to rearrangement and deletion of the DNA-binding domain plasmids, recombinational events between the DNA-binding and activation domain plasmids, or genomic rearrangement of the host strain. To exclude these possibilities, three sets of experiments were carried out. First, plasmid DNAs from the yeast transformants were recovered and analyzed. Our results indicated that these constructs were present in all transformants examined and exhibited similar restriction enzyme profiles as those in E. coli. Second, all screens were carried out in duplicate to determine whether the results were reproducible. Third, any protein that resulted in histidine-positive growth with the empty vector controls was classified as a false positive. The protein species, which migrated at approximately 82 kDa and was detected by the anti-HA antibody in the input protein samples of the cell lysate but not in the protein samples immunoprecipitated with either the anti-HA or anti-myc antibodies, may represent a cellular protein that non-specifically reacts with the anti-HA antibody. Furthermore, the potential interactions among 5 capsid proteins and 28 tegument proteins of HCMV have recently been investigated using the YTH approach. Although powerful in the amount of data generated by the many interactome studies, much work still remains to fully understand the biological significance of the identified interactions. Nonetheless, the interactome studies have proven to be very valuable starting points for predicting functions of unknown proteins when found to interact with known partners. In addition, they serve as building blocks for many systems-level studies. In this study, we used the yeast two-hybrid screen approach to study potential binary interactions among HCMVencoded virion proteins. Our YTH analysis revealed 79 potential interactions, 45 of which were also identified using co-IP experiments by expressing these proteins in human cells. It has been reported that two-hybrid screens may generate significant number of false positives. One type of false positives may represent random generation of histidine-positive colonies. This is possibly due to rearrangement and deletion of the DNA-binding domain plasmids, recombinational events between the DNA-binding and activation domain plasmids, and genomic rearrangement of the host strain. Other types of false positives may be due to either non-specific activation of different reporter systems or autoactivation, which is caused by activators of transcription with only the binding domain-fusion proteins. Similarly, false negatives may arise from the YTH screens. YTH systems that test for protein interactions in the nucleus are prone to false negatives due to the expression of proteins that are normally not found in the nuclear environment. Furthermore, this system may potentially miss interactions involving proteins with significant hydrophobic domains such as membrane proteins, which may not be folded correctly. To exclude these possibilities, five different sets of experiments have been carried out. First, plasmid DNA from the yeast transformants was recovered and analyzed.