BioWorld International Correspondent

LONDON - If scientists knew exactly how pathogenic bacteria manage to stick to human tissues, they might be able to design new therapies for infections caused by the bacteria. One more step toward this goal has been taken, with the publication of a detailed account of how streptococcal and staphylococcal bacteria bind to proteins on human cells.

Streptococci, such as Streptococcus pyogenes, and staphylococci, such as Staphylococcus aureus, are important causes of human infections, including, for example, throat infections, skin infections and infections of bones, joints and the heart valves.

Both bacteria bind to the host protein called fibronectin, which allows them to stick to human cells. Recently, it has also been shown that, by binding to fibronectin, bacteria can enter endothelial and epithelial cells - cells that do not normally carry out phagocytosis. This observation fits with the ability of some bacteria to invade other body tissues by spreading through the bloodstream. It may also play a role in allowing bacteria to hide from antibiotics, some of which cannot cross cell membranes, thus allowing them to establish chronic infections.

Jennifer Potts, British Heart Foundation basic science lecturer at the University of Oxford in the UK, told BioWorld International, "Our study allows a new understanding at the molecular level, and even at the level of individual amino acids, of this particular interaction between host and pathogen. Our model will help researchers to design experiments to gain a deeper understanding of the mechanism of uptake of bacteria into host cells. Obviously, by understanding these processes, eventually that may allow us to develop strategies to block uptake or adhesion of bacteria."

An account of the study appears in the May 8, 2003, Nature in an article titled "Pathogenic bacteria attach to human fibronectin through a tandem b-zipper."

Potts, together with Ulrich Schwarz-Linek and others at the University of Oxford, and a team of collaborators at the Center for Extracellular Matrix Biology at the Institute of Biosciences and Technology in Houston, investigated the interaction between fibronectin and fibronectin-binding proteins of bacteria. They were interested in studying that interaction partly because of the potential practical applications, and also because the section of the bacterial protein involved in the binding does not have a clearly defined 3-dimensional structure - a feature that, for structural biologists, makes the interaction even more interesting.

They used nuclear magnetic resonance spectroscopy to reveal how a peptide from the fibronectin-binding protein of Streptococcus dysgalactiae (which is primarily a bovine pathogen) interacts with part of the fibronectin molecule that contains a series of five domains, called F1 modules.

Their results comprise the first 3-dimensional data on that interaction. Potts told BioWorld International, "It turns out that what we can see is a protein-protein interaction that we called a tandem b-zipper, which as far as we are aware has never been reported before."

The structure showed that short regions of the bacterial peptide bind specifically to the first two F1 modules of fibronectin, largely by extending the sheet structure of each F1 module. The unstructured bacterial protein therefore lines up next to the F1 modules of fibronectin, forming an extra strand next to them, just as pulling on a zip fastener draws the other half of the zip into line.

Other researchers already had shown that there was some homology between the genes encoding the fibronectin-binding protein of S. dysgalactiae and those encoding this protein in S. aureus and S. pyogenes. Further analysis carried out by Potts and her colleagues showed that, in the human pathogens, there is a repeating pattern of protein sections that would bind the F1 modules of fibronectin. "We identified longer strings of short peptide motifs that bind strings of four or five F1 modules," Potts said. The team correctly predicted binding between short bacterial peptides and specific F1 modules.

"We also hypothesised that each motif would bind its specific F1 module, in a similar manner to the way in which the peptide bound the F1 modules in the structure that we determined," Potts added. "Our model provides an explanation for the previous observation that multiple copies of fibronectin can bind to individual bacterial proteins."

Next, the group is planning to obtain structural data of the type they already have for S. dysgalactiae, S. pyogenes and S. aureus. "We want to confirm that these peptides do bind in the same way, by forming this antiparallel tandem b-zipper. Our strategy will be to try to understand these interactions further at the amino acid level, because this is the sort of data that could one day help us to design strategies for blocking uptake or adhesion of bacterial cells," Potts said.