Features:

THE FUTURE OF ANTIBIOTIC AMMUNITION

Antibiotics are indispensable to the modern health care system, assisting and complementing our immune systems. Over the past 10 years, the rapid emergence of bacteria strains that are resistant to multiple drugs has heightened the need to develop new classes of antibiotics. A particularly promising class of these new antibiotics are called antimicrobial peptides (AMPs).

Hundreds of AMPs of vertebrate and invertebrate origin have been discovered in the past decade, but they have existed since prehistoric times. Gene-encoded antimicrobial peptides are now well known to be a pervasive component of the immune defense system throughout the animal kingdom. They work by attacking the bacterial cell membrane. This process is called cell lysis. Formation of pores in the protective layer causes cell death by allowing the flow of ions and molecules into and out of the cell non-selectively. Most other antibiotics attack bacteria by attacking specific molecules that are part of the bacterial cellular machinery, against which bacteria can develop resistance. Yet bacteria have not been able to develop a resistance to AMPs that have existed for millennia. In order to develop resistance to AMPs, the bacterium must reengineer its outer membrane, a very difficult and complex proposition.

Unfortunately, most naturally occurring AMPs are toxic. They cause hemolysis, or the premature breakdown of blood cells, and are thus inappropriate for therapeutic purposes. The exact mechanism behind AMPs bringing about cell lysis, whether in blood cells or targeted microbial cells, is not yet understood. Hence, further efforts are needed to understand and engineer AMPs that are less toxic and have improved therapeutic qualities.

Yiannis Kaznessis, who leads one such research project at the University of Minnesota, is focusing on a class of AMPs called cathelicidins. He and his graduate student, Himanshu Khandelia, are trying to understand how these antimicrobial peptides work at the molecular level. They are using computational modeling on NCSA's SGI Origin2000 and new IBM p690 supercomputing system known as Copper to quantify interactions with lipid molecules, which are an integral part of membranes in both pathogens and friendly cells. How exactly AMPs disrupt the membrane structure remains unclear. This lack of a molecular-level picture hampers the engineering of peptide antibiotics. However, the Kaznessis team's observations will help clarify this picture and enable the design of a novel class of antibiotics.

SMALL, FAST, AND LETHAL

Kaznessis has chosen cathelicidins as the focus of his research for many reasons, but perhaps most important is their pervasiveness. They are a major antimicrobial peptide family found in many mammals, including pigs, guinea pigs, mice, goats, cattle, sheep, rabbits, and monkeys. They are also one of the two major classes of AMPs produced in the human body. Cathelicidins typically are found in white blood cells produced in myeloid organs, but they also can be found in the spleen, stomach, airways, and intestines.

Cathelicidins are also characterized by their ease of synthesis, the speed at which they work, their small size, and their remarkable structural diversity. They are capable of destroying pathogens in minutes whereas current antibiotics may need days or even weeks to overcome an infection. Speed is particularly important for fighting viruses. Many viral infections mutate very rapidly. That's why a new strain of influenza emerges every year.

"AMPs will be very useful against viral infections once we come up with a good design of a peptide that can attack the outer viral envelope," Kaznessis adds. It is not as easy for viruses to mutate to stop an AMP attack on their lipid bilayers.

Peptides have alpha-helical, extended helical, or beta-sheet conformations. This structural diversity, coupled with their small size, make them a threat to a wide range of microbes. For example, bacteria, viruses, and fungi are often attacked with multiple peptides of different structural types. Therefore, the microbe cannot rest after surviving the attack of one or two peptides. This wide number of AMP sequences and structures make it difficult for pathogens to develop resistance.

PROMISING RESEARCH

Promising data about the activities and specifics of many of the cathelicidins are already available. Current experiments have quantified and analyzed the interaction of several AMPs with lipids. These experiments have shown that cathelicidins in particular are very active against a host of pathogenic organisms. They have also shown that mutations of at least one cathelicidin, known as ovispirin-1, can reduce harmful effects while preserving the peptide's antimicrobial effects.

As a result, the Kaznessis team is working to understand how this peptide, among others, causes remarkable alterations in lipid structure. They have already implicated specific electrostatic interactions between specific amino acids and lipid head groups. "We will help design better and more rational experiments from our knowledge and observations of molecular-scale phenomena," Kaznessis says.

The research team's initial results are also valuable because they corroborate experimental data. "We have seen two peptides disrupt model lipid monolayers and bilayers in our simulations," Kaznessis says. Simulations with other peptides in the next two years will lead to a more complete, global understanding of the nature of peptide-lipid interactions.

In order to create simulations of empirical research, the Kaznessis team is standardizing a procedure to carry out dynamic molecular simulations of peptide-membrane systems that will run the CHARMM (Chemistry at HARvard Molecular Mechanics) software package with the PARAM28b parameter set. This CHARMM package has allowed the team to model the lipid molecule system, water molecules, and peptides using an empirical potential field. This field is used to calculate forces on individual particles in the system. Newton's equations of motion are then solved for all atoms of the system to obtain their time- dependent trajectory in phase space. Thus, they are able to see how these molecules behave and interact in the presence of each other. Ultimately, these trajectories can be used to calculate several important properties like surface tension, lipid order parameters, and peptide dihedral angles, which can be correlated to experimentally observed quantities.

Kaznessis has simulated more than 30 systems and five different peptides on the SGI Origin2000 at NCSA. Because NCSA is phasing out this system, his team is moving to the new IBM p690, transitioning their version of CHARMM to the new architecture. Though they use the machines to run processor- and memory- intensive molecular dynamics simulations of large lipid-peptide systems, they also plan to simulate metabolic and gene networks, as well as protein folding.

Kaznessis' team has just begun to better understand how cathelicidins work. Ultimately, they hope to demonstrate systematically that "fine-tuning," or manipulating peptide sequences to achieve various results such as reduced toxicity, is possible. Their research will also enhance the understanding of how other AMPs work. A global understanding of the interaction of peptides with model lipid systems will lead to the design of some of the most effective and long-lasting antibiotic medicines ever known.