The Hollenberg mouse embryo cDNA library (Behrens et al., 1996) was amplified to gain enough plasmid for yeast transformation. E.coli library culture was thawed on ice, and 10³ and 10⁶ dilutions were made in LB-ampicillin selective medium. One μl of the 10³ dilution, and 50-100 μl of the 10⁶ dilution was mixed with LB-ampicillin medium, spread onto LB-ampicillin plates and incubated at 37°C overnight. On the next day, colonies were counted to calculate the library titer (number of colony forming units per ml of bacterial suspension). A volume of this suspension containing twice as much cells as the number of independent clones of the library (5x10⁶) was diluted in LB-ampicillin, plated at 200,000 colonies per 15-cm dish (nearly confluent) and incubated at 37°C overnight. Colonies were scraped into 1.5-liter selective medium; this suspension was incubated at 37°C with shaking for a further 2 h, bacteria were pelleted and large-scale plasmid isolation was performed using Megaplasmid columns 2500 (Qiagen).

Bait proteins for the yeast two-hybrid system are encoded in the pBTM116 vector (Bartel and Fields, 1995). This vector carries the TRP1 gene and has a polylinker downstream of LexA coding sequences. The E. coli repressor LexA consists of two domains: a C-terminal domain, which is responsible for dimerization prior to DNA binding, and an N-terminal domain responsible for specific binding to a palindromic operator containing the CTGTNNNN consensus half-site. The Hollenberg library is inserted in the VP16 vector, which carries the LEU2 gene and contains a nuclear localization signal-VP16 activation domain sequence upstream of the multiple cloning site. Both pBTM116 and VP16 yeast expression vectors bear a bacterial origin of replication and an ampicillin resistance gene, which allow plasmid amplification in bacteria.

To generate TprMet-ErbB2 yeast baits, a cDNA fragment encoding amino acids 1005-1260 of rodent ErbB-2 harbouring a Y1253F mutation was inserted into the Not I/Sal I sites of pSK+ (Stratagene) This region stretches over Y1028 (Y1), Y1144 (Y2), 1201 (Y3) and 1226/7 (Y4) of ErbB-2. A cDNA sequence coding for dimerization and kinase domains of TprMet (amino acids 1-481; Park 86 cell) was generated by PCR; this fragment, flanked by 5´ Not I/EcoR I and 3´ Not I sites, was inserted into the Not I site at 5´ end of ErbB-2 in pSK+, thus rendering the hybrid cDNA for the TprMet-ErbB2(Y1-4) bait. The complete cDNA sequence was finally subcloned downstream of LexA into EcoR I/Sal I sites of pBTM116 yeast expression vector. A new TprMet hybrid protein was constructed to include Y1253 (Y5) of ErbB-2 in the bait. An ErbB-2 sequence encoding amino acids 1197-1260 was fused to TprMet as above to generate the TprMet-ErbB2(Y3-5) bait; thus, Y1201 (Y3), 1226/1227 (Y4) and 1253 (Y5) of ErbB-2 were present in this bait. A BTM-TprMet-Δtail control plasmid lacking the Met C-terminal multiple docking site was constructed by inserting a PCR fragment encoding amino acids 1-479 of TprMet into the EcoR I site of pBTM116. To generate kinase-deficient bait proteins, the wild type sequences of TprMet were excised by EcoR I/Not I and replaced by a PCR fragment containing the inactivating mutation K243A (Rodrigues and Park, 1993).

Both TprMet-ErbB2(Y1-4) and TprMet-ErbB2(Y3-5) bait proteins were used to screen the Hollenberg cDNA library from E10.5 mouse embryos. Saccharomyces cerevisiae strain L40 was used as host. This yeast strain carries two reporter genes, HIS3 and lacZ, whose expression is driven by minimal GAL1 promoters fused to multimerized LexA binding sites; therefore, yeast expressing LexA activators can be detected as histidine prototrophs or by measurable β-galactosidase activity. A pBTM116-lamin plasmid was used as a further control to eliminate false positive clones. Reagents and methods for yeast two-hybrid analyses were adapted from MATCHMAKER^TM Handbook (PT1265-1, Clontech).

Bait plasmids and library were sequentially introduced into host L40 yeast strain to improve the efficiency of transformation. Briefly, yeast was initially transformed with bait plasmid in a small-scale procedure, and then library screens were performed (Fields and Song, 1989). To prepare competent yeast, a single colony of L40 yeast was inoculated into 20 ml of YPD (10 g/l yeast extract, Difco; 20 g/l peptone, Difco; 2% dextrose) medium and incubated at 30°C overnight. On the next day, culture was diluted 10-fold and further incubated until OD₆₀₀ was 0.5. Cells were pelleted, washed with water and resuspended in 1.5 ml of sterile 1X TE/LiAc (10 mM Tris-HCl pH 7.5, 1 mM EDTA, 100 mM lithium acetate). Competent L40 yeast was then transformed with bait plasmid; 0.5 μg of plasmid DNA was mixed with 50 μg of salmon sperm carrier DNA, 50 μl of competent yeast and 300 μl PEG/TE/LiAc (40% polyethylene glycol, 1X TE/LiAc), vortexed and incubated at 30°C for 30 min. DMSO was added to 10%, and heat-shock transformation was performed at 42°C for 15 min. Mixture was chilled on ice, cells were pelleted, resuspended in 250 μl water and spread onto selection agar plates lacking tryptophan. Colonies appear after 2 or 3 days. For the library screen, one single colony of the bait transformants was grown to obtain 200 ml of a saturated cell suspension (OD₆₀₀ greater than 1) in medium without tryptophan. This culture was added to 800 ml YPD medium to prepare 20 ml of competent yeast as above. Yeast suspension was incubated at room temperature for 10 min, and 10 mg carrier DNA, 250 μg library DNA and 140 ml PEG/TE/LiAc was addded. Mixture was incubated at 30°C for 30 min, 17.6 ml of DMSO was added, and heat-shock transformation was performed as above. Co-transformed yeast cells were resuspended in 1 liter YPD and incubated at 30°C for 1 h. Cells were pelleted, resuspended in selection medium lacking tryptophan and leucine and further incubated at 30°C for 8 h. Finally, cells were resuspended in 10 ml water and 200 μl of a dilution series was plated onto 50 selection agar plates without tryptophan, leucine and histidine in the presence of 20 mM 3-aminotriazole and incubated at 30°C. Double transformants that express interacting proteins rendered colonies within 8-10 days, which were re-plated onto fresh selection agar plates for further analysis.

Interaction between bait and preys was confirmed by β-galactosidase activity filter assay and by re-transfection of bait and prey plasmids back into yeast. For filter assay, colonies from selection plates (without tryptophan and leucine) were replica-transferred onto a Whatman Nr.1 filter. The replica filter was submerged in liquid nitrogen to permeabilize cells, and then allowed to thaw. The filter was placed (colonies facing up) onto another filter presoaked in Z buffer/X-gal solution (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, pH 7.0, 0.27% β-mercaptoethanol and 0.334 g/l X-gal) and incubated at 30°C. Colonies expressing β-galactosidase appeared blue within 1-12 h.

For yeast re-transformation, library plasmids encoding interacting proteins were isolated from individual positive colonies. To remove the bait plasmid from double transformant yeast, colonies were inoculated into medium lacking leucine and cultured at 30°C for 1 day. Growth in the absence of tryptophan selection allows survival of yeast segregants without bait plasmid, which is randomly lost. To isolate the library plasmid from yeast segregants, cells from 1 ml of the above culture were pelleted, lysed in 200 μl of lysis buffer (2% Triton X-100, 1% SDS, 100 mM NaCl, 1 mM EDTA, 10 mM Tris pH 8.0) and disrupted in the presence of 200 μl phenol/chloroform/isoamyl alcohol (25:24:1) and 0.3 g of acid-washed glass beads by vortexing for 2 min; suspension was clarified by centrifugation and plasmid was recovered from the supernatant by standard ethanol precipitation. Library plasmid was amplified in E. coli HB101 strain, which is leucine auxotroph due to a leuB mutation. HB101 cells were electroporated with the isolated library plasmid and plated onto selection agar plates without leucine; therefore, transformants can be selected by complementation with the yeast LEU2 gene from the VP16 library plasmid. Bait and prey plasmids were re-transformed into yeast according to the small-scale transformation protocol previously described. True positive interacting clones grew again on selective agar plates without tryptophan, leucine and histidine and were positive in β-galactosidase assays.

Mutational analysis of the bait proteins was performed to characterize the ErbB-2 binding sites for each interacting protein found in the screen. Deletion mutants bearing single or tandem tyrosine residues of the ErbB-2 multiple docking region were generated by standard PCR. The PCR fragments were flanked by 5´ Not I and 3´ Sal I sites, and were used to replace the ErbB-2 wild type sequence of the hybrid TprMet-ErbB2 baits. The ErbB-2 deletion mutants fused to TprMet were: Y1 (amino acids 1014-1081) includes Y1028; Y1-2 (amino acids 1014-1160) includes Y1028 and Y1144; Y2 (amino acids 1138-1160) includes Y1144; Y2-3 (amino acids 1138-1221) includes Y1144 and Y1201; Y2F-3 is a variant of Y2-3 in which Y2 was mutated to F using a commercial kit (Clontech); Y3 (amino acids 1194-1221) includes Y1201; Y3-4 (amino acids 1194-1244) includes Y1201 and Y1226/1227; Y4 (amino acids 1220-1244) includes Y1226/1227; Y4-5 (amino acids 1220-1260) includes Y1226/1227 and Y1253; Y5 (amino acids 1248-1260) includes Y1253.

Expression of the various TprMet-ErbB2 proteins in the yeast was checked by Western blot analysis using anti-LexA antibodies (Clontech). L40 yeast transformants were grown in selective medium without tryptophan at 30°C. During exponential growth phase, OD₆₀₀ was measured for 1 ml suspension; cells were pelleted from an aliquot corresponding to OD₆₀₀= 0.5 and resuspended in 100 μl 2X denaturing buffer for SDS-PAGE (100 mM Tris base, pH 6.8, 20% glycerol, 4% SDS, 2% β-mercaptoethanol, 0.2% bromphenol blue). Twenty μl of this suspension was loaded per slot onto an 8% polyacrylamide gel, proteins were resolved, blotted in a PVDF membrane and Western Blotting was performed according to standard protocols.