Barth Syndrome and "Neuwald's Hypothesis": From Computer to Clinic
 SHL scientist Andy Neuwald uses neither cells, nor beakers, nor pipettes, nor Petri dishes, let alone patients in his research. So how has he recently cracked one area of medical research wide open? With a computer program he wrote called PROBE.
"I'm quite happy about it. It was kind of like lightning striking, but in a way that I hope to repeat!" says Andy, referring to how PROBE revealed the probable cause of Barth syndrome, a heritable genetic disease that is often fatal in childhood.
Ansy is one of several "bioinformatics" researchers at CSHL. Rather than pipetting their way toward greater understanding of how cells work, Andy and his colleagues program computers to analyze biological data in the digital realm.
Consider the following twenty-letter DNA sequence: TCAAAGTGTACTTACCTCGC. No human being can look at that sequence and tell what it means. But plug it into a suitable bioinformatics computer program and watch what happens: It is sent into cyberspace and compared with millions of other sequences in existing databases. In a few seconds, a message returns indicating that this twenty-letter sequence is part of a chicken gene that encodes the protein ovalbumin, better known as egg white.
To try this yourself, point your web browser to www.ncbi.nlm.nih.gov (the homepage of Andy's postdoctoral stomping grounds at the National Center for Biotechnology Information in Bethesda, Maryland) and click on BLAST. Better yet, use Andy's new program, PROBE. Why? "Because PROBE can detect subtle similarities among genes that programs like BLAST may miss," says Andy.
To picture how PROBE works, imagine that a master carpet weaver makes hundreds of very different rugs, but incorporates a subtle signature element-say four small green squares-in every rug. Similarly, proteins that carry out related functions may appear very different in their overall pattern of amino acid building blocks ("hundreds of very different rugs"), but will contain signature elements of amino acid similarity ("four small green squares") that are key to how the proteins work.
PROBE can detect such signature elements, or "motifs," among otherwise very weakly related proteins. The program does this by comparing thousands of related protein sequences and "teaching itself" to recognize patterns as it speeds along. "PROBE keeps learning as it realigns these proteins in various ways, and it modifies its own statistics accordingly, until subtle but real patterns begin to emerge," says Andy. He credits his colleagues Jun Liu of Harvard University and Charles Lawrence of the New York State Department of Health for their help with designing PROBE.
Remarkably, bioinformatics programs like PROBE reveal that at the level of genes and proteins, we have a lot in common not only with the birds and the bees, but also with plants, worms, molds, and even bacteria. We carry unmistakable traces of our distant evolutionary kinship with bacteria and all other organisms in our genes. The presence of these ancient, shared gene sequences in modern day organisms is the key to Andy's discovery about Barth syndrome.
In 1997, Andy was using PROBE to hunt for new families of related proteins. The hunt provided Andy and his colleagues a great deal of material for future research. Within one of several protein families identified by PROBE, a protein called "tafazzin" stood out.
The tafazzin gene had been shown to be mutated in patients afflicted with Barth syndrome. The disease is characterized by short stature and poor muscle tone, including potentially fatal defects in heart muscle. Muscles, especially the heart, require a great of energy to function properly. Heart and other muscle defects in Barth syndrome patients had been traced to abnormal mitochondria, the sausage-shaped energy-producing power plants of the cell. But how tafazzin gene mutations might cause defects in mitochondria was a mystery. No bioinformatics program had ever detected motifs in the tafazzin protein that provided unambiguous clues to its function. Until PROBE.
PROBE placed tafazzin squarely in an ancient family of proteins found in all organisms called "acyltransferases." THis discovery enabled Andy to propose a disease mechanism underlying Barth syndrome. Acyltransferases are enzymes that control the composition of cell membranes, the sac-like structures that surround cells and things within cells like the nucleus and mitochondria. In addition to their outer membranes, which can be likened to sausage casings, mitochondria have convoluted inner membranes. The inner mitochondrial membrane is the site of energy production. Andy proposed that mutations in the tafazzin gene alter the composition of mitochondrial membranes, thus causing defects in energy production and Barth syndrome.
In 1999, the eponymous Dutch scientist who first characterized Barth syndrome dubbed Andy's proposal, "Neuwald's hypothesis." Last year, Barth's laboratory used patient's cells to confirm that Andy's theory is correct. Confirmation of the disease mechanism proposed by Andy has enabled clinicians to consider a new treatment for Barth syndrome patients, which is now being tested in clinical trials.
"Things don't often work out this nicely," says Andy. "Even though it is relatively easy to make computer-based biological discoveries these days, they seldom lead to medical progress of this kind." He predicts that as the field of bioinformatics matures, discoveries like these will become much more common.
|
