Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2002 Nov;12(11):1773-84.
doi: 10.1101/gr.406902.

Protein-protein Interactions Between Large Proteins: Two-Hybrid Screening Using a Functionally Classified Library Composed of Long cDNAs

Affiliations
Free PMC article

Protein-protein Interactions Between Large Proteins: Two-Hybrid Screening Using a Functionally Classified Library Composed of Long cDNAs

Manabu Nakayama et al. Genome Res. .
Free PMC article

Abstract

Large proteins have multiple domains that are potentially capable of binding many kinds of partners. It is conceivable, therefore, that such proteins could function as an intricate framework of assembly protein complexes. To comprehensively study protein-protein interactions between large KIAA proteins, we have constructed a library composed of 1087 KIAA cDNA clones based on prior functional classifications done in silico. We were guided by two principles that raise the success rate for detecting interactions per tested combination: we avoided testing low-probability combinations, and reduced the number of potential false negatives that arise from the fact that large proteins cannot reliably be expressed in yeast. The latter was addressed by constructing a cDNA library comprised of random fragments encoding large proteins. Cytoplasmic domains of KIAA transmembrane proteins (>1000 amino acids) were used as bait for yeast two-hybrid screening. Our analyses reveal that several KIAA proteins bearing a transmembrane region have the capability of binding to other KIAA proteins containing domains (e.g., PDZ, SH3, rhoGEF, and spectrin) known to be localized to highly specialized submembranous sites, indicating that they participate in cellular junction formation, receptor or channel clustering, and intracellular signaling events. Our representative library should be a very useful resource for detecting previously unidentified interactions because it complements conventional expression libraries, which seldom contain large cDNAs.

Figures

Figure 1
Figure 1
Diagram of our strategy for constructing a functionally classified library composed of long cDNAs. (A) Illustration of different combinations according to functionally classified order. Because there are ∼30,000 human proteins, our ultimate aim would be to prepare a complete protein–protein interaction map by assessing interactions between ∼30,000 × ∼30,000 protein combinations. One dot represents the interaction between one vertical-ordered protein (i.e., left, ordinate) and one horizontal-ordered protein (i.e., left, abscissa). The set of possible interactions to be assayed is extremely large. If proteins are, instead, first categorized according to their location and function, as shown in the right panel, interaction dots are clustered to smaller regions, and thus the set of possible interactions to be investigated is reduced considerably. (B) Schematic diagram of library construction. Black and white bars represent open reading and untranslated regions, respectively. Triangle indicates the first methionine codon. After functional classification in silico, randomly fragmented cDNAs were introduced into vectors for yeast two-hybrid screening using the Gateway system.
Figure 2
Figure 2
Yeast two-hybrid screening and subsequent determination of interacting domains. The domain organization for bait (left) and prey (right) KIAA proteins are shown. Blue bars indicate independent clones (one bar represents one prey). The length of the blue bar indicates the sequence of the KIAA protein contained within an isolated prey. The interacting domains located in the bait and preys are indicated by pink (bait) and green (prey) brackets. In the case of KIAA1634, we estimated that KIAA1132 can bind two different domains: the first and second PDZ domain, as shown.
Figure 3
Figure 3
Reassessment of interactions after exchanging cDNAs in baits and preys via the Gateway system. Image of yeast two hybrids grown on selection media. After exchanging the plasmids contained in the original bait and prey using the Gateway system, new bait plasmids (KIAA0777, KIAA1256, KIAA1296, or KIAA1496), new prey plasmid (KIAA1368), and an empty vector (pACTGW-attR) were transformed into yeast AH109. The viability of the transformants was assessed on SD/−Leu/−-Trp and SD/−Ade/−His/−Leu/−Trp plates. The + symbol located in the right-hand column indicates that the MEL1 gene product, α-galactosidase, is actively expressed. Blue-colored growth on nutritional selection plates in the presence of X-α-Gal indicates active expression of α-galactosidase.

Similar articles

See all similar articles

Cited by 40 articles

See all "Cited by" articles

Publication types

MeSH terms

LinkOut - more resources

Feedback