Decoding the Complex Messages of Cells

he prospect of modeling the biochemical processes of living cells has gone from a vague hope to an attain- able goal of computational scientists with the help of a national initiative called the Alliance for Cellular Signaling (AfCS). Established in 2000 and led by Nobel laureate Alfred G. Gilman, the AfCS has brought together 52 scientists from diverse fields at 21 institutions across the nation. These scientists will research, model, and promote understanding of the ways in which cells communicate with one another to coordinate the functions of an organism. The AfCS has established seven laboratory centers across the nation, including its Bioinformatics Laboratory, led by Shankar Subramaniam, a professor of Bioengineering and of Chemistry and Biochemistry at UCSD and a member of the Whitaker Institute of Biomedical Engineering at the university. The SDSC laboratory will acquire, process, abstract, and disseminate data supplied by biologists and chemists, and will use this information to model the processes of living cells. The knowledge-management capabilities of the TeraGrid (see stories, p. 1 and 2) will help ensure the success of the effort.

Figure 1. Research Target The AfCS research program seeks to understand the signaling pathways of two types of cells, one of them a mouse cardiac myocyte such as this one.

Figure 2. Calcium Signaling in Cardiac Myocytes

Click Image to View

The AfCS is a 10-year, $100 million effort funded by the National Institute of General Medical Sciences under its Glue Grants program to "enable the solution of major problems in biomedical research and to facilitate the next evolutionary stage of integrative biomedical science."

EVOLUTION OF INTERCELLULAR SIGNALING

Single-celled organisms respond to their environments in basic ways: they move, change shape, ingest, divide. The cells of a multicellular creature exist within the environment of the organism, coordinating their myriad functions by altering and responding to changes in their chemical environments. In the course of a billion years, multicellular organisms have evolved complex intercellular signaling systems.

"Biological systems are very difficult to reduce to simple elements, because not one of the processes takes place in isolation," said Subramaniam. "Although biological processes are based in physics and chemistry, biological interactions occur on so many different levels that simply observing the individual processes in cells and organisms will never be sufficient to give a complete understanding of their collective interaction."

Gilman, chairman of the University of Texas Southwestern Medical Center’s Department of Pharmacology, shared the 1994 Nobel Prize in Medicine with Martin Rodbell for their discovery several years earlier of proteins that bind to guanosine triphosphate–G-proteins. Gilman and Rodbell found that G-proteins act as signal transducers, which transmit and modulate signals in cells. G-proteins have the ability to activate different cellular amplifier systems. They receive multiple signals from the exterior, integrate them, and interact with other functioning proteins to activate or deactivate them or to modify the magnitude of their activity. G-proteins control the fundamental life processes of cells, and disturbances in their function can lead to a variety of illnesses. For example, the gastrointestinal disease cholera results from a bacterial toxin that keeps a single G-protein switched on.

The AfCS is focusing on immune-system cells called B lymphocytes, and heart muscle cells called cardiac myocytes (Figure 1). B lymphocytes produce antibody proteins that help destroy foreign substances, and the cells migrate in response to chemical cues in a process called chemotaxis. Cardiac myocytes produce and coordinate heartbeat rhythms, and their contraction is modulated by chemical messages (Figure 2).

"Our primary goal is control and delineation of messages or signaling events within cells," Subraman-iam said. "These messages arise in the context of specific external or extracellular events. We will use the external event as a key to the intracellular response."

BIOINFORMATICS EXCELLENCE

AfCS laboratories will specialize in cell preparation and analysis, development of signaling assays, microscopy, molecular biology, protein chemistry, and antibody research. The Bioinformatics Laboratory at SDSC will acquire, process, and manage the other six laboratories’ data, and disseminate information, both in raw form and as processed data that is more readily understood by the community.

Subramaniam was a natural choice to lead the state-of-the-art informatics laboratory. He is a pioneer in bioinformatics and computational biology, and he is well-known as the developer of the Biology WorkBench, a Web-based analysis environment that enables biologists to search a variety of popular protein and nucleic acid sequence databases.

SDSC’s computing infrastructure contributed to the decision to establish the laboratory. SDSC’s diversified information management facilities include some of the fastest supercomputers and biggest information archives in the world. Its staff also includes several national leaders in data-intensive computing as well as in bioinformatics.

The National Science Foundation’s August 9 award for the $53 million TeraGrid means that even more sophisticated resources will be available to AfCS researchers in the future.

"The TeraGrid infrastructure at SDSC focuses on data-intensive computing, very large-scale distributed collections management, and data mining," said Mike Vildibill, SDSC’s deputy director for resources and leader of the local TeraGrid effort. "We lead the consortium in the support of data-management services, including collections management, collections replication, and the housing of specific data collections such as those of the Alliance for Cellular Signaling and the Protein Data Bank."

MOLECULE PAGES AND MODELS

Subramaniam’s group will develop definitions and relationships for creation and storage of information on protein sequence and functional annotation, cell preparation, gene array, proteomics, cell-based and molecular assays, and microscopy. The data will be linked to specific query environments in an object-relational database, to be called the Alliance Information Management System (AIMS). AfCS researchers will be able to deposit, query, and analyze both primary and processed data at the AIMS website.

AfCS member researchers are being assigned specific molecules to track in the scientific literature, and they will contribute detailed, standardized, and quality-controlled records extracted from reports and papers. Each of these "molecule pages" is intended to contain an exhaustive set of data for each of the signaling molecules, and will be maintained and continually updated by the assigned researchers over the course of the program.

"We anticipate that the AfCS website and the Alliance Information Management System will serve as a portal and resource for the entire signaling research community," Subramaniam said. Molecule pages will be reviewed by peer researchers nominated through an AfCS editorial committee. These pages will have provisions for Web-based interactive comments from other researchers.

The molecule pages will be linked to evolving maps of cellular signaling modules. The AfCS group at SDSC will undertake to represent the entire signaling system as a set of interacting processing modules that can be mathematically modeled.

"We plan to use the molecule pages and the Alliance experimental results to generate a signaling database," said Peter Morrison, project manager for the AfCS Bioinformatics Laboratory at SDSC. "Our construction of signaling pathways in combination with the experimental investigations will lead to identification of signaling modules. Molecule pages will be linked to evolving maps of cellular signaling modules. It is our intent to model the pathways in terms of the component modules." Each intracellular signaling pathway is comprised of interacting sets of these component modules. But the modules are regulated by their environmental context.

"The first stage of modeling will be concerned with topology–the basic flowchart of signal processing," Subramaniam said. "When quantitative data on the elements of the various reactions and processes inside the cells become available, we will be able to carry out stoichiometric analysis of the flow of signals from input to output.

"All of the physical and life sciences are becoming more driven by information than by startling experimental or field discoveries. Experimental data will always be essential, but our understanding depends on seeing how the individual pieces of the puzzle fit together. As our knowledge base grows and our ability to perform detailed simulations improves, the laboratory and the field are becoming places for testing and confirmation of the concepts developed by computational means." –MG