1. Antibody screening and website constructionFor better quantitative analysis, specificity of antibody is a crucial criterion. However, each antibody's performance is different depending on what type of samples used for testing. Protein phosphorylation, for example, can be different within a cell line depending on treatment or time-points. We have developed a web compatible relational database which currently includes >200 antibodies, and >700 screenings. Approximately 50% of antibodies produced single predominant band in the right samples. It will soon be uploaded in the web using Asp.net-MySQL scripting language along with the 'reverse-phase' protein lysate microarray page.2. Application of reverse-phase array for surgically removed specimensSurgically removed specimens have considerable heterogeniety and limited amount of target component, which make difficult to interpret any data output. Using frozen biopsy materials from lymphoma tissue, we were able to capture analyzable signals using lysate array technique. However, we also found that the noise is dominant when using extremely small amount of starting material. We are trying to make large scale arrays to answer such fundamental questions with most updated technical improvements.3. Quantitative proteomic network monitoring in response to DNA damageWe acquired multiple dimensions of information on protein dynamics in response to DNA damage (x-ray ionizing radiation, ultraviolet light, and adriamycin) using RPAs. Our initial quantitative results revealed that: (i) there is clear stress dose-response effect in p53 protein; and (ii) a comparison of the rates of increase of p21 and Cyclin D3 in response to DNA damage is predictive of the phenotypic fate; i.e., the rate of increase in p21 expression is greater than the rate of increase in Cyclin D3 expression at 0.5uM of adriamycin exposure, whereas Cyclin D3 showed a greater rate of increase than p21 at 1.0uM, which were well associated with G2 arrest and apoptosis at 72h after the damage, respectively.4. Evaluation of mathematical modeling on the p53-Mdm2 feedback loopTo evaluate the mathematical models experimentally, we used RPAs on which x-ray irradiated cell lysates of an isogenic pair of HCT116 colon cancer cell line with wild-type p53; and its knock-out strain, were printed. Both proteins including phosphorylated ones showed expected pulses in the wild-type, whereas no pulses were seen in the knockout. Interestingly, we have found different behaviors of phosphorylated proteins for both p53 and Mdm2: (i) A rise in p53-Ser15 occurred more rapidly within the first hour after stress when compared to total p53; and (ii) Mdm2-Ser166 stayed at the initial level within the first 3 hours before the oscillations, suggesting that phosphorylation at this site could stabilize the molecule from rapid degradation. Based on these experimental observations, we determined the model parameters and then generated an in silico 'knockout'. We were able to produce in silico 'knockout' simulations that resembled our experiment.
