Research Topics

General Aspects

Dr. Floriano’s research involves the development and application of computational tools to simulate biologically relevant molecular interactions. This includes the development of computer programs to predict three-dimensional structure from genetic data, find potential binding sites in proteins, and scan virtual libraries of chemicals for potential drugs or modulators of biological activity.

Using computational techniques, we can study correlations between chemical structure and biological activity from both a ligand-only and a binding to biological target perspectives. Computer assisted molecular design (CAMD) techniques can aid the development of new and modified chemical entities, such as new medicinal drugs and taste modifiers. Computational tools are also useful to process proteomic data, understand the molecular basis of drug action, and the molecular basis of individual differences in taste and odor preferences.

One of the biological fields Dr. Floriano is currently active is the study of the chemical senses (taste and smell). The study of taste and olfaction from a molecular point of view has become increasingly important with the availability of the receptors repertoire from mouse and human genomes. Research in taste and olfaction has potential application in various industries. Examples of applications include the design of taste modifiers (new artificial sweeteners, blockers for bitter taste in food and medicines, enhancers of unpleasant taste response for treatment of food-related disorders such as obesity and addictions to alcohol and tobacco), and the development of odor sensors for use in health (disease identification using characteristic smells present in body fluids), food and perfume manufacturing, and defense (identification of explosives, drugs and chemical weapons agents).

Scientific Approach

We use computer assisted molecular design (CAMD) techniques in our research. These techniques rely on a three-dimensional (3D) model of the ligand binding pocket into a biologically relevant protein. Different conformations of the ligand in the binding site of the protein are generated and scored by a mathematical equation. The scoring functions used to estimate the binding affinity between ligand and receptor are based on calculated free energies of binding (energy-based functions) or derived from empirical data (knowledge-based functions).

The 3D structure of the receptor is generally obtained experimentally through different techniques (x-ray diffraction of crystals, nuclear magnetic resonance, neutron diffraction). However, in cases were experimentally determined structures are not yet available or are unlikely to be obtained, computational techniques can be used to generate 3D models of the receptor. There are two approaches for computationally predicting 3D structures of proteins: 1) homology modeling, in which 3D models are generated using sequence homology to other proteins with experimentally determined structures, and 2) ab-initio modeling that uses computational techniques to predict secondary structure and tertiary folding (structure prediction).

We use a technique called Virtual Ligand Screening (VLS) to simulate the docking of small molecules to proteins. In VLS, small chemical compounds (ligands) are docked computationally into proteins. The interaction energies between ligands and protein are calculated and used to sort ligands according to their predicted affinity for the target protein. The ligands with higher affinities are more likely to be biologically active than the ones with lower affinity. This creates a list of chemical compounds prioritized in terms of their potential to be biologically active.

Ongoing Research Projects

Selective Ligands for Fast Identification of Botulinum Neurotoxin Types A and B. This project has been funded by NIH for two years, starting September 1st, 2006. The objective of this project is to identify chemical compounds that can be used for fast, reliable and low-cost diagnosis and identification of botulism toxin exposure. I am using structure-based computational techniques I have developed while working at Caltech under the supervision of Dr. William Goddard. These techniques are used to (a) identify potential binding sites in proteins (ScanBindSite? ) and (b) screen large chemical compound libraries using a hierarchical high throughput virtual ligand screening approach (HierVLS? ). In this project, they will be further developed and used to identify high affinity and high selectivity chemical compounds that bind to the biologically active and also to alternative binding sites in the three-dimensional structures of C. botulinum toxin types A and B. The expected outcome is a new chemical assay for identification and fast clinical diagnosis of exposure to the botulism toxin.

Sweet Receptor Gene Variation and Aspartame Blindness in Primates and Other Species.

Collaborators: Xia Li, Alexander A Bachmanov, Danielle R Reed (Monell Chemical Senses Center, Philadelphia PA)

The main objective of this project is to understand the molecular basis of species differences in taste response to aspartame, an artificial sweetener that appears to elicit sweet response only in humans, apes and Old World monkeys. The Monell group is collecting and analyzing DNA sequence data for the sweet receptor genes T1R2 and T1R3 from different species. My participation in this project includes the development of computer-generated models of sweet receptors and their interactions with natural and artificial sweeteners, and the comparative analysis of sequences and modeled structures for receptors from different species. We hope these models will help us to understand the molecular basis for species differences in sweet taste sensitivity.

Molecular Basis of Umami Taste Response. Our objective in this project is to investigate the synergy between different receptors believed to participate in umami taste response. 3D structures for the human receptors taste metabotropic glutamate receptor subtype 4, T1R1 and T1R3 are modeled with various amino acids, sweet tastants, and other molecules of interest bound to them. The predicted binding profiles and 3D structures of protein-ligand complex are cross-compared and analyzed using available experimental data.

Computer-aided Molecular Design of Bitter Taste blockers The objective of this project is to identify/design inhibitors of bitter taste. Bitter taste often elicits aversion behavior in humans and it has been linked to resistance in intake of medicines and foods such as vegetables from the Brassicaceae family (broccoli, Brussels sprouts, cauliflower) especially in children and elderly. Using computational tools, we can model the structures of the 25 functional bitter taste receptors identified in the human genome, and use these structures in virtual ligand screening studies aimed at identifying small commercially available chemicals that may function as antagonists of bitter receptors.

Virtual Ligand Screening of Olfactory Receptors for Design of Biosensor Arrays. The general objective of this project is the design of sensor arrays based on mammalian olfactory receptors (ORs) for recognition of chemicals associated with particular diseases (disease biomarkers). Students involved in this project work on biological aspects of OR classification, modeling of OR family head(s), and virtual screening of odorant libraries for identification of potential agonists.

Recognition Site Classification of GPCRs and the Genseq Web Server. The general objective of this project is the development of a method to classify G protein coupled receptors (GPCRs) into families based on recognition sites. The development of this classification method focuses on olfactory receptors (ORs) because they are the largest GPCR subfamily in the human genome, and recognition sites for them have previously been suggested by me and collaborators, and other research groups. These recognition sites are amino acid positions believed to play an essential role in ligand binding and activation of the GPCRs. The expected outcome of this project is a binding site classification method for GPCRs that can be used for de-orphanizing receptors, for improving specificity of biologically active compounds, and for checking cross-reactivity potential of novel compounds. Students in this project work on algorithm development and the programming of computational tools to implement the classification method. These tools include a web server (Genseq) that will run the Recognition Site Classification (Resicla) algorithm. This algorithm will be implemented into the open source sequence alignment program Clustal (Thompson et al., 1994), which is accessible through genseq. Although genseq is being primarily developed for olfactory receptors, its extension to include other classes of G protein coupled receptors is straight-forward. A prototype web server that acquires genetic sequences from external databases (GenBank? , ExPASy? , etc), aligns the sequences using Clustal (Thompson et al., 1994), and identifies user-specified patterns in the output alignment is already available. This prototype has all its basic functions operational and it is ready for the implementation of the classification algorithm.

Design of Lead Compounds for Transcriptional Regulation of Glucocorticoid Responsive Elements The objective of this project is to identify/design potential lead compounds for transcriptional control of glucocorticoid response elements (GREs) using first principles computational approaches. This project has a number of specific aims and, within these aims, specific tasks that need to be performed. This is a continuation of a previous project developed in collaboration with Dr. W. A. Goddard III (Caltech), Dr. N. Vaidehi (City of Hope), and Dr. Eugene Roberts (former City of Hope).

Student Training

The work in my lab is heavily computational. Students either

• apply computational tools to find biologically active molecules and to gain knowledge about biological processes through the computer simulation of molecular interactions, or
• are involved in the development of computational tools.

This type of work requires computer skills (i.e., the ability to comfortably work with computers and the willingness to learn how to use new computer programs, operating system (Linux), and programming languages).

Graduate and undergraduate students starting research projects in my lab need to be trained on basic aspects of bioinformatics and computational biology research:

1.Use of Linux/Unix computer environment and high-performance computer clusters

2.Use of web-based and local bioinformatics tools for data mining and data archiving

3.Use of locally-installed software for sequence alignment/phylogeny, molecular structure visualization, small molecules manipulation, homology modeling, protein structure health assessment, molecular mechanics and dynamics calculations, docking and virtual screening

They are also trained on general aspects of scientific research such as design, execution and reporting of scientific work.