(vielleicht in den Titel des Threads ergänzen. "- Newsletter Volume 6" ?)
Vier Seiten Text-Information mit - zugegeben - schönen Bildern, aber 3,4 MB
The Personnel Touch
We?ll start with the bad news, and then move on to the good. Dr. Alex Gillet, a long-time member of both the Olson laboratory and the FightAIDS@Home team, has moved on to a new position at Illumina, a biotech company in San Diego. Alex joined the lab as an intern 7 years ago, arriving from France with a Masters degree in Bioinformatics. He stayed on originally as our computer systems administrator, but he showed great talent for programming molecular modeling code. He worked on a novel human computer interface for molecular investigations using computer auto-fabrication and augmented reality. For that work he earned his Ph.D. this year.
On the FightAIDS@Home project, Alex Gillet maintained our front-end servers and developed database software for organizing and delivering data to the World Community Grid. He also kept our websites uptodate, and he made sure that the computer systems in the lab ran smoothly. He will be deeply missed in the lab, but we wish him the greatest success and the best of luck in this next step in his career.
Volume 6: November 3, 2008
The Progress that You?ve Recently Helped us Make
After implementing the newest version of the AutoDock code on the World Community Grid, you have already helped us perform over 1,028,000 docking jobs for the FightAIDS@Home project. Each job corresponds to docking one compound against one conformation of the drug target called ?HIV protease.?
To see what this important drug target looks like, see the image below this paragraph. Being able to perform over a million docking jobs in only six months demonstrates the tremendous power of the unused computer cycles that you all donate to the World Community Grid. The FightAIDS@Home team wholeheartedly thanks all of you for your interest and for your dependable help.
We have been using these donated computer cycles to perform ?Relaxed Complex? experiments against HIV
protease. The Relaxed Complex method/scheme is a new approach to drug design that was created by Prof.
Andy McCammon?s lab at our neighbor, UCSD. The Relaxed Complex is able to incorporate the flexibilities of both the potential drugs and the viral proteins we are trying to inhibit. Including flexibility in the drug discovery and design process can make these virtual experiments more realistic and more accurate. For a detailed description of this method, please read our previous edition, ?Volume 5,? and/or read the bottom half of http://fightaidsathome.scripps.edu/discovery.html
Structurally diverse compounds can bind to the same site, and a single compound can display several different binding modes against one drug target. Some inhibitors, such as the piperine-based renin inhibitors, the allosteric inhibitors of beta-lactamase, and the neuraminidase inhibitor Oseltamivir carboxylate, bind to conformations of a receptor that were never previously observed in NMR or crystallographic studies of their molecular structure. Since HIV protease is an usually dynamic and flexible enzyme, incorporating the flexibility of this drug target is especially important when pursuing the discovery of novel classes of inhibitors. The Relaxed Complex method is a good approach for including the substantial dynamic flexibility that HIV protease displays in solution.
One Relaxed Complex experiment that you recently helped us complete examined the performance of nine ?false negatives? from the NCI Diversity Set against 2,000 different conformations of the wild type HIV-1b protease. These conformations, or snapshots, were harvested from 20 nanoseconds of Alex Perryman?s (ALP) previous Molecular Dynamics simulations. Previously these nine compounds had not performed well when they were docked against 77 different crystal structures of HIV protease in the first experiment on FightAIDS@Home, but they did display anti-HIV activity in a FRET-based protease inhibition assay. Our Relaxed Complex experiments of these nine false negative hits were surprisingly
successful?all 9 compounds scored better than the ?significance threshold? of -7.0 kcal/mol that was established in previous FightAIDS@Home experiments [and published in Max W. Chang, William ?Lindy? Lindstrom, Arthur J. Olson, and Richard K. Belew. ?Analysis of HIV Wild-Type and Mutant Structures via in Silico Docking Against Diverse Ligand Libraries.? Journal of Chemical Information and Modeling, 47(3): 1258-1262 (2007)]. Thus, applying the Relaxed Complex method allowed us to correctly classify all nine of these compounds as actual hits. ?Hits? are smaller compounds that display
some inhibitory activity but that bind too weakly or too non-specifically to become drugs. The compound that displayed the highest anti-HIV protease activity in the FRET-based assay is being studied in more detail. We will modify, combine, and ?chemically decorate? the best hits, to try to produce better protease binders. After testing the new compounds that we create on the computer, our collaborators at The Scripps Research Institute will synthesize the most promising candidates and assess their actual potencies. These results on the ?nine false negatives? are not yet published; thus, we are giving you a ?sneak peak.? But after we re-analyze the results a bit and polish a few figures in the next couple of months, we will write and submit a paper on these exciting new results. We?ll let you know as soon as it is published.
The unused computer cycles that you donated to FightAIDS@Home also helped us perform Relaxed Complex experiments involving several different reference compounds docking to an ensemble of 2,000 wild type conformations of HIV protease. The reference compounds we used included the FDA-approved drugs Amprenavir, Atazanavir, Darunavir, Indinavir, Lopinavir, Nelfinavir, Ritonavir, Saquinavir, and Tipranavir as well as compounds in development which include AB2, AB3, JE-2147, KNI-272, TL3, and TMC-126. Even though AB2 is only two-fold more potent than AB3, our Relaxed Complex results correctly reproduced the fact that AB2 has higher affinity against wild type HIV-1b protease. We are now investigating several different methods for sorting, ranking, and analyzing these Relaxed Complex results.
The protocol that best reproduces the experimental binding affinities of these reference compounds will be used in all of our subsequent experiments on FightAIDS@Home. To increase the efficiency of the virtual High-Throughput screens that we are performing with the Relaxed Complex method on FightAIDS@Home, we have been exploring different methods for selecting a small number of snapshots that still capture the overall binding variations that inhibitors display against the entire ensemble of conformations of the drug target. From Dr. Rommie Amarro?s research in Prof. Andrew J. McCammon?s lab at the University of California, San Diego, we learned about the utility of a new tool called ?Structure QR? or ?QR factorization.? By using the new ?Structure QR? method in VMD (from the Luthey-Schulten lab at the University of Illinois at Urbana-Champaign), we have generated a structurally-diverse, non-redundant set of conformations that characterizes the variations sampled within an MD simulation of HIV protease. This QR factorization approach has allowed us to reduce the number of conformations that we need to target in Relaxed Complex experiments by approximately 20-fold. This increase in efficiency will allow us to include a large panel of many different drug-resistant mutants of HIV protease in our future experiments on the World CommunityGrid.
A ?QR-selected? subset of conformations of one of the most drug-resistant ?superbugs? of HIV protease is being used in our virtual High-Throughput screening experiments on FightAIDS@Home. In one experiment that you are currently helping us perform on the WCG, our new version of the NCI?s ?DTP library of moderately active compounds? is being screened against the active site of 103 mutant conformations. Since we included the different protonation states and tautomers that the compounds in this DTP library can likely sample in water, the 2,000 compounds in this library produced approximately 6,000 different versions of these compounds. All 6,000 versions are being screened against the active site of this QRselected subset of mutant conformations. The active site (where the current anti-HIV protease drugs bind) is
the location where the ball-and-stick model of a compound binds in the image on the left.
In a different experiment that?s currently running on FightAIDS@Home, we are screening this DTP library against the putative exo site on the peripheral surface of these mutant snapshots. This experiment is searching for compounds that have the potential to be developed into ?flexibility wedges? (i.e., allosteric inhibitors that can disrupt the conformational changes that HIV protease must undergo during its catalytic cycle). If the computations running on your computers can help us discover and develop these flexibility wedges, they will represent a completely new class of anti-AIDS drugs.
Preparing for Future FightAIDS@Home Experiments To provide atomically-detailed information regarding the differences in both dynamic flexibility and in conformational preferences that different drug-resistant mutant ?superbugs? of HIV protease can display, we are now performing a large panel of new Molecular Dynamics (MD) simulations. The conformations we will harvest from these MD simulations will be used in new Relaxed Complex experiments on FightAIDS@Home with the clinically-approved HIV protease inhibitors, to enable us to study the mechanisms of multi-drug-resistance. The new MD simulations that we recently began include five of the seven ?spanning proteases? that were identified in previous experiments with FightAIDS@Home (see the publication cited on page 2). Each of these ?spanning proteases? present unique structural features that represent a distinct portion of the set of 71 different wild type and mutant proteases that were investigated.
In addition to trying to defeat the superbugs of ?HIV-1b? protease (i.e., the group of strains of HIV found in the U.S. and in Europe), we are also going to start performing research against the groups of strains of HIV found in Africa and in Asia (which are called ?HIV-2? and ?HIV-1c,? respectively). HIV-1b, HIV-1c, and HIV-2 protease display similar conformations of their backbones in the available crystal structures, but they have very different amino acid sequences. The current HIV protease drugs were all designed and optimized against the wild type HIV-1b protease, and some of them are not quite as
effective against the HIV-1c and HIV-2 subtypes. By studying the affinities of the clinically-approved HIV-1b protease inhibitors against the ensembles of conformations from our new MD simulations of HIV-1c and HIV-2 protease, we will try to help optimize the treatment guidelines for patients in Asia and Africa. In the second cycle of experiments that we plan to perform on the World Community Grid, we will investigate the structural features and modifications that can help improve the affinity of an active site protease inhibitor against HIV-1c, HIV-2, or both. By comparing the performance of the best compounds against HIV-1c and HIV-2 with their performance against conformations from HIV-1b protease, we should obtain insight regarding the structure-activity relationships that provide high affinity against all three subtypes of HIV. The strategies produced from this insight should also be quite useful when developing new inhibitors that maintain potency against the drug-resistant superbugs of HIV-1b.
We could not perform this research without your help. Thank you very much for helping us advance the fight against multi-drug-resistant ?super bugs? of HIV and for helping us improve the tools and techniques used in the entire field of structurebased drug design.
Prof. Arthur J. Olson
Dr. Alex L. Perryman
Dr. Stefano Forli
Dr. Garrett M. Morrishttp://fightaidsathome.scripps.edu/