Combinatorial Chemistry & High Throughput Screening - Volume 19, Issue 5, 2016

Herein, we report the first 48-membered, dansyl-based, combinatorial fluorescent library. From the electronic and structural properties of the probes, we analyzed their optical properties and chemical yields, with an average of 49 %. The molecules were examined for their pH responses, and DS-2 and DS-45 showed blue-shifts, whereas DS-7 and DS-40 showed red-shifts in wavelength with increasing pH. Finally, cell permeability was investigated by treating SNU-2292 cells. Our results demonstrate the potential application of this library in biosensors, bio-imaging and pH indicators.

Many inflammatory processes are associated with an increase in the production of reactive oxygen species (ROS). Chemical probes that specifically detect ROS are potentially useful tools for the early diagnosis of inflammatory diseases as well as cancer. Herein we have developed a library of coumarin hybrids by condensation of various heterocyclic quaternary salts to a 7-hydroxycoumarin scaffold. From our library we identified one benzothiazole-coumarin hybrid as a red-fluorescent compound with emission maxima around 620 nm and a strong fluorogenic response. Furthermore, we proved that this scaffold is suitable for the preparation of activatable probes, such as by modification with a boronate group for selective sensing of hydrogen peroxide (H2O2). In vitro assays confirmed the reactivity and subsequent emission of our probe upon incubation with H2O2 with good selectivity over different ROS and reactive nitrogen species (RNS) as well as minimal toxicity in cells. Finally cell imaging experiments were performed in murine macrophages and validated the utility of the activatable probe for the detection of H2O2 in living cells.

Many applications in biomedical research require the long-term identification and tracking of cells over time. In previous work we have demonstrated that by sequentially dosing a cell population with different emission wavelength nanoparticles it is possible to use the random number of nanoparticle loaded vesicles generated by the cells as a barcode for individual cells within the population. In this paper we develop a simple model to describe the number of codes that can be generated using this sequential loading protocol. The methodology is validated by comparison with experiment and subsequently used to predict the effect of varying the number of colors used to encode the cells and also to assess the effect of misreading the cellular code due to errors in imaging the vesicles.

We previously utilized an in vivo peptide phage display selection technique, which included the use of detergent elution of phage from excised tumor, to obtain tumor-targeting phage with the ability to extravasate the vasculature and bind directly to prostate tumor tissue. It is hypothesized that this same in vivo phage selection technique can be used to functionally select for molecules that not only bind to cancer cells but also kill them. Here we analyzed two different in vivo phage display selected phage clones, G1 and H5, retrieved from PC-3 human prostate carcinoma xenografted tumors. First, cell de-attachment as an endpoint criterion for apoptosis and cell cycle was examined. After 2.5 hours incubation with G1 phage, PC-3 cell attachment was reduced by 23.8% and the percent of cell population in M phase reduced by 32.1%. In comparison, PC-3 cells incubated with H5 phage had a reduction of 25.0% cell attachment and 33.6% of cell population in M phase. These changes in combination with elevated caspase activation within cells in M phase, and no significant changes to G1/G0 or S phase cell populations suggest that the cytotoxic phages are targeting actively dividing PC-3 cells. Microscopic studies were also performed to further analyze the nature of cytotoxicity of these two phage clones. It was found that G1 phage induced and co- localized with tubulin based projections within apoptotic cells, while H5 phage did not. These phage may form the foundation for a new class of targeted prostate cancer therapeutic agents.

Control of fluorescent dye localization in live cells is crucial for fluorescence imaging. Here, we describe quantitative structure activity relation (QSAR) models for predicting intracellular localization of fluorescent dyes. For generating the QSAR models, electric charge (Z) calculated by pKa, conjugated bond number (CBN), the largest conjugated fragment (LCF), molecular weight (MW) and log P were used as parameters. We identified the intracellular localization of 119 BODIPY dyes in live NIH3T3 cells, and assessed the accuracy of our models by comparing their predictions with the observed dye localizations. As predicted by the models, no BODIPY dyes localized in nuclei or plasma membranes. The accuracy of the model for localization in fat droplets was 92%, with the models for cytosol and lysosomes showing poorer agreement with observed dye localization, albeit well above chance levels. Overall therefore the utility of QSAR models for predicting dye localization in live cells was clearly demonstrated.

Formazan-based colorimetric cytotoxicity assays, such as the MTT assay, are typically used to assess cell viability with only metabolically active cells reducing tetrazolium salts into the formazans, which is then quantified by absorbance. Fluorescence offers several advantages compared to colorimetric assays and would enable techniques such as flow cytometry and confocal microscopy to be used for analysis. Here, fluorescent formazans 10, 11 and 12, and their corresponding tetrazolium salts 13, 16 and 24, respectively, were synthesised by incorporation of a known fluorophore backbone (coumarin, fluorescein and rhodol) with disruption of the conjugated system preventing or reducing fluorescence of the tetrazolium salts. These tetrazolium salts were successfully reduced to the fluorescent formazans with cells and offer a step forward in the development of fluorescent cytotoxicity assays.

We describe a novel biosensor system for reporting proximity between cell surface proteins in live cultured cells. The biosensor takes advantage of recently developed fluorogen-activating proteins (FAPs) that display fluorescence only when bound to otherwise-nonfluorescent fluorogen molecules. To demonstrate feasibility for the approach, two recombinant rapamycin-binding proteins were expressed as single-pass plasma membrane proteins in HeLa cells; one of the proteins (scAvd- FRB) carried an extracellular avidin tag; the other (HL1-TO1-FKBP) carried an extracellular FAP. Cells were incubated with a membrane-impermeable bivalent ligand (biotin-PEG2000-DIR) consisting of biotin joined to a dimethyl-indole red (DIR) fluorogen by a polyethylene glycol linker, thus tethering the fluorogen to the scAvd-FRB fusion protein. Addition of rapamycin, which promotes FKBP-FRB dimerization and thereby brings the FAP in close proximity to the tethered fluorogen, led to a significant increase in DIR fluorescence. We call the new proximity assay TEFLA, for tethered fluorogen assay.

EGFR (ErbB-1/HER1) kinase plays an important role in cancer therapy. Two classification models were established to predict whether a compound is an inhibitor or a decoy of human EGFR (ErbR-1) by using Kohonen’s self-organizing map (SOM) and support vector machine (SVM). A dataset containing 1248 ATP binding site inhibitors and 3090 decoys was collected and randomly divided into a training set (831 inhibitors and 2064 decoys) and a test set (417 inhibitors and 1029 decoys). The descriptors that represent molecular structures were calculated by software ADRIANA.Code. Thirteen significant descriptors including five global descriptors and eight 2D property autocorrelation descriptors were selected by Pearson correlation analysis and stepwise analysis. The prediction accuracies on training set and test set are 98.5% and 96.3% for SOM model, 99.0% and 97.0% for SVM model, respectively. Both of these two classification models have good performance on distinguishing EGFR inhibitors from decoys.

Cancer cell-specific diagnostic or therapeutic tools are commonly believed to significantly increase the success rate of cancer diagnosis and targeted therapies. To extend the repertoire of available cancer cell-specific phage fusion proteins and study their efficacy as navigating moieties, we used two landscape phage display libraries f8/8 and f8/9 displaying an 8- or 9-mer random peptide fusion to identify a panel of novel peptide families that are specific to Calu-3 cells. Using a phage capture assay, we showed that two of the selected phage clones, ANGRPSMT and VNGRAEAP (phage and their recombinant proteins are named by the sequence of the fusion peptide), are selective for the Calu-3 cell line in comparison to phenotypically normal lung epithelial cells and distribute into unique subcellular fractions.

Decision trees are renowned in the computational chemistry and machine learning communities for their interpretability. Their capacity and usage are somewhat limited by the fact that they normally work on categorical data. Improvements to known decision tree algorithms are usually carried out by increasing and tweaking parameters, as well as the post-processing of the class assignment. In this work we attempted to tackle both these issues. Firstly, conditional mutual information was used as the criterion for selecting the attribute on which to split instances. The algorithm performance was compared with the results of C4.5 (WEKA’s J48) using default parameters and no restrictions. Two datasets were used for this purpose, DrugBank compounds for HRH1 binding prediction and Traditional Chinese Medicine formulation predicted bioactivities for therapeutic class annotation. Secondly, an automated binning method for continuous data was evaluated, namely Scott’s normal reference rule, in order to allow any decision tree to easily handle continuous data. This was applied to all approved drugs in DrugBank for predicting the RDKit SLogP property, using the remaining RDKit physicochemical attributes as input.