Experiment 41: 36% Completed

Experiment 41 involves screening the Enamine library of 2.345 million compounds against the two allosteric sites on the surface of HIV protease. This experiment is similar to Exp. 36, but a different, much larger library of compounds is being screened, and some new targets are being included. The first part of this experiment involves docking compounds against the two allosteric-fragment-bound crystal structures presented in our recent article in Chemical Biology and Drug Design, vol. 75: 257-268 (March 2010). Three other, brand new crystal structures from Dave Stout's lab that involve allosteric fragments bound to these two sites on the surface of HIV protease are also being used as targets. Two of these new targets are presented in a new research manuscript from the Stout lab that is currently being peer-reviewed. When this paper is accepted, we will describe these targets in more detail and provide a link to this new paper.


This is by far the largest experiment we have submitted to FightAIDS@Home; it involves faah33,529 - faah45,253.

The results of these calculations started arriving at TSRI on 5/15/2012.

Experiment 40: 100% Completed

Experiment 40 involves screening the ChemBridge library of 1,013,483 models of compounds against the newly-discovered allosteric binding site on HIV-1 integrase. This new allosteric binding site (which we are also targeting in Experiments 38 and 39) was discovered by Professor John J. Deadman's group, and it was described in "Structural basis for a new mechanism of inhibition of HIV-1 integrase identified by fragment screening and structure-based design," by D.I. Rhodes, T.S. Peat, J.J. Deadman, et al., published in the journal Antiviral Chemistry and Chemotherapy, 21: 155-168 (2011). The new crystal structure from this paper that contains the atomically-detailed, 3-D data on this new allosteric site is called "3NF6.pdb". We are screening compounds against this allosteric site to try to discover new, larger, more potent allosteric inhibitors of HIV-1 integrase. It is hoped that these new allosteric inhibitors of integrase will be effective at disabling the current drug-resistant mutant superbugs of HIV integrase. For more information about this new allosteric site, see Volume 10 of the FightAIDS@Home newsletter or our recent World AIDS Day webinar (both are linked at the top of the homepage for this site).

Similar to Experiment 39, in Experiment 40 we are screening one million compounds against the new allosteric site on HIV-1 integrase using two slightly different docking approaches: in the first half of these calculations, we are using the smaller dimensions of the "grid box" (the region that the compounds are allowed to explore during the docking calculations) that produced the best results in the "positive control" docking calculations that reproduced the known binding mode of this new allosteric fragment (see the figures in Volume 10 of the FAAH Newsletter, page 8). In the second half of these calculations, we are using a larger grid box, to try to find even larger allosteric inhibitors that can bind strongly with both the allosteric site and other sub-pockets that are adjacent to it.


This experiment involves faah31,501 - faah33,528.

These results began arriving at TSRI on 3/24/2012, and the experiment finished 5/29/2012.

Experiment 39: 100% Completed

Experiment 39 involves screening the Enamine library of 2,345,014 compounds against the newly-discovered allosteric binding site on HIV-1 integrase. This new allosteric binding site (which we are also targeting in Experiment 38) was discovered by Professor John J. Deadman's group, and it was described in "Structural basis for a new mechanism of inhibition of HIV-1 integrase identified by fragment screening and structure-based design," by D.I. Rhodes, T.S. Peat, J.J. Deadman, et al., published in the journal Antiviral Chemistry and Chemotherapy, 21: 155-168 (2011). The new crystal structure from this paper that contains the atomically-detailed, 3-D data on this new allosteric site is called "3NF6.pdb". We are screening compounds against this allosteric site to try to discover new, larger, more potent allosteric inhibitors of HIV-1 integrase. It is hoped that these new allosteric inhibitors of integrase will be effective at disabling the current drug-resistant mutant superbugs of HIV integrase. For more information about this new allosteric site, see Volume 10 of the FightAIDS@Home newsletter or our recent World AIDS Day webinar (both are linked at the top of the homepage for this site).

In Experiment 39, we are screening these 2.3 million compounds against the new allosteric site on HIV-1 integrase using two slightly different docking approaches: in the first half of these calculations, we are using the smaller dimensions of the "grid box" (the region that the compounds are allowed to explore during the docking calculations) that produced the best results in the "positive control" docking calculations that reproduced the known binding mode of this new allosteric fragment (see the figures in Volume 10 of the FAAH Newsletter, page 8). In the second half of these calculations, we are using a larger grid box, to try to find even larger allosteric inhibitors that can bind strongly with both the allosteric site and other sub-pockets that are adjacent to it.


This experiment involves faah26,811 - faah31,500.

These calculations began (that is, the results started arriving at TSRI) on 12/17/2011. The last batch of results arrived 5/15/2012.

Experiment 38: 100% Completed

Experiment 38 involves screening the full NCI library of 316,179 compounds against the newly-discovered allosteric binding site on HIV-1 integrase. This new allosteric binding site was discovered by Professor John J. Deadman's group, and it was described in "Structural basis for a new mechanism of inhibition of HIV-1 integrase identified by fragment screening and structure-based design," by D.I. Rhodes, T.S. Peat, J.J. Deadman, et al., published in the journal Antiviral Chemistry and Chemotherapy, 21: 155-168 (2011). The new crystal structure from this paper that contains the atomically-detailed, 3-D data on this new allosteric site is called "3NF6.pdb". We are screening compounds against this allosteric site to try to discover new, larger, more potent allosteric inhibitors of HIV-1 integrase. It is hoped that these new allosteric inhibitors of integrase will be effective at disabling the current drug-resistant mutant superbugs of HIV integrase. For more information about this new allosteric site, see Volume 10 of the FightAIDS@Home newsletter (pages 7-8) or our recent World AIDS Day webinar (both are linked at the top of the homepage for this site).


This experiment involves faah26,179 - faah26,810.

These calculations began 12/06/2011 and ended 12/22/2011.

Experiment 37: 100% Completed

Experiment 37 involves screening the newly-updated version of the Asinex library of 360,00 compounds. 507,000 different models are used to represent these 360,000 compounds (due to the need to represent different protonation states and different tautomers that some of these compounds can form in solution). These compounds are being screened against the active site and the "eye site" of 6 different crystal structures of HIV protease. When the flaps have a semi-open conformation, then we can target both the "eye site" and the floor of the active site. But when the flaps have a closed conformation, the "eye site" is no longer accessible (which means that we will only target the traditional active site, which is where the current HIV protease drugs bind).

The 1st target is the crystal structure of the wild type HIV protease with 5-nitroindole bound in the eye site. This new crystal structure from Prof. C. David Stout's lab was presented in the Supporting Information for our recent article in Chemical Biology and Drug Design, vol. 75: 257-268 (March 2010). Since this crystallographic structure has semi-open flaps, we are screening these compounds against both the eye site and the active site of this target.

The 2nd target is the semi-open crystal structure of wild type HIV-1b protease from 1HHP.pdb. This crystal structure has been used in previous virtual screens that were performed by Prof. Heather Carlson's group, in which they did find a novel inhibitor of HIV protease activity. Thus, this particular crystal structure has already been proven to be useful for virtual screens that target the "eye site." The idea of targeting this eye site was first proposed by Prof. Heather Carlson's group. But in this experiment, we are screening different compounds against this crystal structure than the compounds that were used in previous screens against 1HHP.pdb.

The 3rd target has only been used in one previous FightAIDS@Home experiment (i.e., Experiment 35). It is a brand new crystal structure from Assoc. Prof. C. David Stout's lab of the chimeric "FIV 6s98S" protease, which was developed by our collaborators Ying-Chuan Lin, Prof. Bruce E. Torbett, and Prof. John H. Elder, and which has a closed conformation of the flaps. A paper on this new crystal structure of FIV 6s98S protease was recently accepted for publication in Acta Crystallographica and can be found at the above link. This protease enzyme is "chimeric," because it contains 5 residues from HIV protease that were substituted into the corresponding positions in FIV protease. The 6th residue was also substituted from HIV protease, but it changed into a different residue during serial passage experiments (i.e., during directed evolution studies). This 6s98S FIV protease has HIV-like drug sensitivity profiles and is a new model system for multi-drug-resistant HIV protease.

The 4th target is the crystal structure of wild type HIV-1b protease bound to the drug darunavir. This crystal structure from 2IEN.pdb has a closed conformation of the flaps. Whenever we target a crystal structure of HIV protease that has a compound bound to it, we delete that ligand before we prepare the model of the target for these docking studies (so that a new compound might be able to bind in its place).

The 5th target is the crystal structure of wild type HIV-1b protease bound to the compound TL-3 and to the allosteric fragment "4d9". This crystal structure, which has closed flaps, was also from the 2010 Chemical Biology & Drug Design paper cited above.

The 6th target is the crystal structure of the V82F/I84V multi-drug-resistant mutant (or "superbug") of HIV protease from 1MSN.pdb, which has closed flaps. The model of this target has one protonated aspartic acid 25 (i.e., one of the two catalytic residues), which should cause us to fish out slightly different types of ligands.

The 7th target is another version of the semi-open crystal structure of wild type HIV-1b protease from 1HHP.pdb. That is, this is the same molecule as the 2nd target in this experiment, but this time the model has one protonated aspartic acid 25.


This experiment involves faah22,630 - faah26,178.

These calculations began 5/12/2011 and ended 12/15/2011 (except for 3 batches, which were finished on 2/07/2012).

Experiment 36: 100% Completed

Experiment 36 involves screening the full NCI library of ~ 316,000 compounds against the two allosteric sites on the surface of HIV protease. This experiment is similar to Exp. 31, but a different library of compounds is being screened, and in this experiment we are only docking compounds against the two allosteric-fragment-bound crystal structures from our recent article in Chemical Biology and Drug Design, vol. 75: 257-268 (March 2010).

In addition, in this experiment we are investigating four slightly different docking protocols, to both advance the methods we use for drug discovery and to increase our probability of finding new "hits" against HIV protease. Since two different allosteric-fragment-induced structures are being used with four slightly different protocols, this experiment is composed of 8 parts, as follows:

1a) The crystal structure bound to allosteric fragment "1f1" is being screened using an "expanded grid box" that includes the 1f1 allosteric site, the 4d9 allosteric site (in the non-induced conformation = a decoy site), and several other decoy sites. The docking calculations will end according to the number of generations each compound explores in the genetic algorithm (that is, all 316,000 compounds will experience the same number of generations as they explore the target).
1b) The crystal structure bound to allosteric fragment "4d9" is being screened using an "expanded grid box" that includes the 4d9 allosteric site, the 1f1 allosteric site (in the non-induced conformation = a decoy site), and several other decoy sites. The docking calculations will end according to the number of generations each compound explores in the genetic algorithm (that is, all 316,000 ligands will experience the same number of generations as they explore the target).

2a) The crystal structure bound to allosteric fragment "1f1" is being screened using an "expanded grid box" that includes the 1f1 allosteric site, the 4d9 allosteric site (in the non-induced conformation), and several other decoy sites. The docking calculations will end according to the number of energy evaluations each ligand explores in the genetic algorithm (that is, all 316,000 compounds will experience the same number of energy evaluations as they explore the target).
2b) The crystal structure bound to allosteric fragment "4d9" is being screened using an "expanded grid box" that includes the 4d9 allosteric site, the 1f1 allosteric site (in the non-induced conformation), and several other decoy sites. The docking calculations will end according to the number of energy evaluations each compound explores in the genetic algorithm (that is, all 316,000 ligands will experience the same number of energy evaluations as they explore the target).

3a) The crystal structure bound to allosteric fragment "1f1" is being screened using a "focused grid box" that only includes the 1f1 allosteric site (that is, no decoy sites were included in the grid box that the compounds are allowed to explore). The docking calculations will end according to the number of energy evaluations each ligand explores in the genetic algorithm (that is, all 316,000 compounds will experience the same number of energy evaluations as they explore the target).
3b) The crystal structure bound to allosteric fragment "4d9" is being screened using a "focused grid box" that only includes the 4d9 allosteric site (that is, no decoy sites were included in the grid box that the compounds are allowed to explore). The docking calculations will end according to the number of energy evaluations each compound explores in the genetic algorithm (that is, all 316,000 ligands will experience the same number of energy evaluations as they explore the target).

4a) The crystal structure bound to allosteric fragment "1f1" is being screened using a "focused grid box" that only includes the 1f1 allosteric site (that is, no decoy sites were included in the grid box that the compounds are allowed to explore). The docking calculations will end according to the number of generations each ligand explores in the genetic algorithm (that is, all 316,000 compounds will experience the same number of generations as they explore the target).
4b) The crystal structure bound to allosteric fragment "4d9" is being screened using a "focused grid box" that only includes the 4d9 allosteric site (that is, no decoy sites were included in the grid box that the compounds are allowed to explore). The docking calculations will end according to the number of generations each ligand explores in the genetic algorithm (that is, all 316,000 ligands will experience the same number of generations as they explore the target).


This experiment involves faah20,093 - faah22,620.

These calculations began 4/16/2011 and ended 7/01/2011.

Experiment 35: 100% Completed

Experiment 35 involves screening the full NCI library of ~ 316,000 compounds against the active site of 8 different versions of HIV protease. Thus, this experiment is similar to Exp. 32, but a different library of compounds is being screened, and one new target has been added. All but two of these target conformations were generated by Dr. Alex L. Perryman's Molecular Dynamics (MD) simulations of 5 different variants of HIV protease. These 8 targets include 2 snapshots of the V82F/I84V mutant from ALP's 2004 paper in Protein Science. These 2 snapshots of a multi-drug-resistant "superbug" have semi-open conformations of the flaps, which makes these models good targets for the "eye site" that is located between the tip of a semi-open flap and the top of the wall of the active site. The 3rd target is the equilibration MD (EqMD) output for 1HSI.pdb, which is a semi-open conformation of HIV-2 protease. HIV-2 is the group of strains of HIV that are most common in Africa. We'll be targeting the "eye site" of 1HSI, as well. The 4th target is the EqMD output from 1MSN.pdb, which was created using a different crystal structure of the V82F/I84V superbug. This model has a closed conformation of the flaps, which means that we'll be targeting the floor of the active site. The 5th target also has a closed conformation of the flaps, but this EqMD output is from 2R5P.pdb, which is the wild type HIV-1c protease. HIV-1c is the group of strains of HIV that are most commonly found in Asia. The 6th target has semi-open flaps, and it is the EqMD output from 1TW7.pdb, which is a superbug with the mutations L10I/D25N/M36V/M46L/I54V/I62V/L63P/A71V/V82A/I84V/L90M. We'll be targeting the eye site of this superbug, too.

The 7th target is a crystal structure of the wild type HIV protease with 5-nitroindole bound in the eye site. This new crystal structure from Prof. C. David Stout's lab was presented in the Supporting Information for our recent article in Chemical Biology and Drug Design, vol. 75: 257-268 (March 2010). This new research article of ours was recently discussed in a press release on Science Daily and in a news story on KPBS-FM. This paper was recently listed as one of the "most read papers" from Chemical Biology and Drug Design in 2010! I deleted the 5-nitroindole fragment from this structure before generating the AutoDock input file for this target. We'll be screening new fragments against this crystal structure's eye site, as well.

The 8th target has never been used on FightAIDS@Home before. It is a brand new crystal structure from Assoc. Prof. C. David Stout's lab of the chimeric "FIV 6s98S" protease, which was developed by our collaborators Ying-Chuan Lin, Prof. Bruce E. Torbett, and Prof. John H. Elder. A paper on this new crystal structure of FIV 6s98S protease is currently being peer-reviewed. This protease enzyme is "chimeric," because it contains 5 residues from HIV protease that were substituted into the corresponding positions in FIV protease. The 6th residue was also substituted from HIV protease, but it changed into a different residue during serial passage experiments (i.e., during directed evolution studies performed with the presence of different HIV protease drugs). This 6s98S FIV protease has HIV-like drug sensitivity profiles and is a new model system for multi-drug-resistant HIV protease.


This experiment involves faah17,565 - faah20,092.

These calculations began 11/26/2010 and ended 4/25/2011.

Experiment 34: 100% Completed

Experiment 34 is our second FightAIDS@Home experiment that targets the HIV integrase system. See the description and the link listed under Experiment 33.

Experiment 34 involves screening the "full National Cancer Institute's (NCI) library" of over 315,000 different compounds against our new dynamic models of the wild type, the G140S/Q148H drug-resistant mutant, and the E92Q/N155H drug-resistant mutant of HIV-1 integrase. We are trying to discover compounds that can bind to and inhibit the active site of the wild type and these two drug-resistant mutants. Since these models all have two magnesium ions in the active site, we are searching for compounds that can inhibit HIV integrase's "strand transfer reaction," which is what the fairly new drug raltegravir (Isentress) does.

We are screening the NCI library of compounds against 4 different snapshots (that is, conformations or 3-D shapes) of HIV integrase: a) the conformation of wild type HIV integrase against which raltegravir docked the best (in the results published in our paper in the Journal of Molecular Biology, March 26, 2010), b) the snapshot of the G140S/Q148H drug-resistant mutant against which raltegravir docked the best in our previous studies, c) the most representative (that is, the most frequently observed) conformation of the E92Q/N155H drug-resistant mutant (according to the results of the QR Factorization method in VMD), and d) the 2nd most representative conformation of the E92Q/N155H drug-resistant mutant of HIV integrase's catalytic core domain.

The best compounds from this virtual screen will be assessed in test tubes, in "strand transfer" assays being developed by our collaborator Dr. Ying-Chuan Lin in Prof. John Elder's lab at TSRI.


This experiment involves faah16,199 - faah17,462.

An extension to this experiment in which the NCI Diversity Set II library of compounds is being screened against these targets involves faah17,463 - faah17,466.

A second extension to this experiment in which the "high pH" version of the full NCI library of compounds is being screened against these targets involves faah17,473 - faah17,564.

These calculations began 10/11/2010 and ended 1/16/2011.

Experiment 33: 100% Completed

Experiment 33 is our first FightAIDS@Home experiment that targets the HIV integrase system. While we are busy analyzing and extending the previous experiments you helped us perform against HIV protease, FAAH will now start screening compounds against the new dynamic models of wild type HIV integrase, the E92Q/N155H drug-resistant mutant, and the G140S/Q148H drug-resistant mutant that we recently created. We published these new models as the cover article for the March 26, 2010, issue of the Journal of Molecular Biology. An image of an ensemble of conformations (shapes) of the G140S/Q148H mutant is shown at the bottom of this page.

Experiment 33 involves screening the "Asinex library" of over 360,000 different fragments and slightly larger compounds against our new dynamic models of the E92Q/N155H drug-resistant mutant of HIV-1 integrase. We are trying to identify compounds that can attach to brand new binding sites on this mutant. Thus, we are both searching for new types of inhibitors and for new, non-active site regions to which inhibitors can bind. We are screening these compounds against 6 different snapshots of this mutant that had the most open conformations of the "140s loop" near the active site. This loop is known to be critical to the catalytic function of integrase. This 140s loop is located in the top, left corner of the image at the bottom of this page. Since the 140s loop likely has a closed conformation during catalysis, we are searching for fragments that can stabilize the open, likely inactive conformations of the 140s loop. In other words, we are searching for fragments that might act as allosteric inhibitors (flexibility wedges that alter the shapes and motions sampled by this very flexible loop). The best compounds from this virtual screen will be assessed in test tubes, in "3' processing" assays performed by our collaborator Dr. Ying-Chuan Lin in Prof. John Elder's lab at TSRI.


This experiment involves faah13,985 - faah16,198.

An extension to this experiment in which the NCI Diversity Set II library of compounds is being screened against these targets involves faah17,467 - faah17,472.

These calculations began 6/03/2010 and ended 12/01/2010.



Jan. 2011 update: 10 of the top-ranked compounds were purchased and are currently be tested by our collaborators at TSRI.

Experiment 32: 100% Completed

Experiment 32 involves screening the Otava library of approximately 335,000 "building blocks" (i.e., fragments) against the active site of 7 different versions of HIV protease. All but one of these target conformations were generated by Dr. Alex L. Perryman's Molecular Dynamics (MD) simulations of 5 different variants of HIV protease. These 7 targets include 2 snapshots of the V82F/I84V mutant from ALP's 2004 paper in Protein Science. These 2 snapshots of a multi-drug-resistant "superbug" have semi-open conformations of the flaps, which makes these models good targets for the "eye site" that is located between the tip of a semi-open flap and the top of the wall of the active site. The 3rd target is the equilibration MD (EqMD) output for 1HSI.pdb, which is a semi-open conformation of HIV-2 protease. HIV-2 is the group of strains of HIV that are most common in Africa. We'll be targeting the "eye site" of 1HSI, as well. The 4th target is the EqMD output from 1MSN.pdb, which was created using a different crystal structure of the V82F/I84V superbug. This model has a closed conformation of the flaps, which means that we'll be targeting the floor of the active site. The 5th target also has a closed conformation of the flaps, but this EqMD output is from 2R5P.pdb, which is the wild type HIV-1c protease. HIV-1c is the group of strains of HIV that are most commonly found in Asia. The 6th target has semi-open flaps, and it is the EqMD output from 1TW7.pdb, which is a superbug with the mutations L10I/D25N/M36V/M46L/I54V/I62V/L63P/A71V/V82A/I84V/L90M. We'll be targeting the eye site of this superbug, too.

The 7th target is a crystal structure of the wild type HIV protease with 5-nitroindole bound in the eye site. This new crystal structure from Prof. C. David Stout's lab was presented in the Supporting Information for our recent article in Chemical Biology and Drug Design, vol. 75: 257-268 (March 2010). This new research article of ours was recently discussed in a press release on Science Daily and in a news story on KPBS-FM. I deleted the 5-nitroindole fragment from this structure before generating the AutoDock input file for this target. We'll be screening new fragments against this crystal structure's eye site, as well.

The Protein Data Bank is located at http://www.rcsb.org/pdb/home/home.do. You can search the PDB using the 4 character codes listed above (i.e., before the .pdb suffix) to learn more about each of these targets.

This experiment involves faah11,647 - faah13,984.

These calculations began 4/12/2010, they were put "on hold" for a while, and they finished 3/06/2011.

Experiment 31: 100% Completed

Experiment 31 involves screening the Otava library of approximately 335,000 "building blocks" against the "exo sites" on the side surfaces of HIV protease (i.e., the allosteric sites). This experiment will dock these fragments against the exo sites of 8 different targets, which include 5 carefully-selected snapshots from previous Molecular Dynamics simulations of the V82F/I84V multi-drug-resistant mutant "super bug." The other targets correspond to the two new fragment-bound crystal structures of HIV protease that were produced by our collaborator, Prof. C. David Stout at TSRI. We are targeting both sides of one of these two new targets, which is why we have "8 different targets." These two new structures from Prof. Stout were also targeted in Experiments 25-28 and 30.

This experiment involves faah8972 - faah11,646.

These calculations began 12/07/2009 and ended 4/18/2010.

Experiment 30: 100% Completed

Like Experiment 29, Experiment 30 also uses the "Asinex" library of over 360,000 different compounds in a virtual screen against HIV protease. But in this experiment, the Asinex compounds are being docked against the allosteric inhibitor site (that is, the "exo site" on the sides of HIV protease). In addition, this experiment targets two new fragment-bound crystal structures of protease that were produced by our collaborator, Prof. C. David Stout at TSRI. These two new structures from Prof. Stout were also targeted in Experiments 25-28.

This experiment involves faah8234 - faah8971.

These calculations began 10/05/2009 and ended 12/07/2009.



Update May, 2012: from the initial analysis of the results against the "1F1" allosteric site on the outside/top of the flaps, 10 fragments were purchased (using some of the funds that IBM's "Watson" won on Jeopardy! and which the IBM International Foundation donated to the FightAIDS@Home project). These 10 fragments were analyzed by Max Chang in Prof. Bruce Torbett's lab in "DSF" assays (differential scanning fluorimetry), which measure the thermal stability (melting temperature) of HIV protease in solution. 5 of these 10 fragments caused a significant shift in HIV protease stability in solution, which indicates that these compounds are able to bind to HIV protease. These 5 fragments are now being investigated in further tests performed by the Finn lab, the Elder lab, and the Stout lab.

Experiment 29: 100% Completed

This experiment involves docking a huge library of compounds against the active site of six of our new models of HIV protease (which are a subset of the targets being used in Experiments 25-27). These six targets are the outputs of the equilibration phase of six different Molecular Dynamics simulations (hence the "mdEq" part of the work units' names). The six different types of HIV protease that we are docking compounds against in Exp. 29 include the "Model6Xapo," which is a drug-resistant "super bug" with 6 different mutations. Our collaborator, Prof. Dave Stout, figured out the structure of this 6X mutant. The "apo" part of the name indicates that this mutant protease molecule did not have a substrate or drug present when its structure was solved. This Model6Xapo has semi-open flaps (that is, the two double-arrows that point towards the center of the molecule and form a roof over the active site have opened up). We've been working with the IBM members of the FAAH team to update the graphics on your screen-savers. We've sent them new graphics to use, and they've already started testing them. Soon, you will be able to see exactly what we mean when we say "semi-open flaps."

Another new model that is being targeted in Exp. 29 is the wild type HIV protease from 1HHP.pdb. We consider this model to be interesting, because the flaps were fairly open, but then they closed again. But this time, they closed down in the opposite arrangement/they "switched handedness" (that is, the flap that is normally in the front is now in the back). Having this different conformation of the flaps might allow us to fish out new types of interesting compounds for subsequent examination in the "test tube."

We included models of a multi-drug-resistant "super bug" with mutations at V82F/I84V and another "super bug" with mutations at I62V/V82A/I84V/L90M.

We are targeting a model of the protease molecule from "HIV-1c," as well. HIV-1c is the subtype, or group of strains, that is most commonly found in Asia. We are also targeting a model of "HIV-2" protease with semi-open flaps. HIV2 is the group of strains that are most commonly found in Africa. The current anti-AIDS drugs were developed and optimized against "HIV-1b," which is the subtype most commonly found in Europe and the USA. But some of these anti-HIV protease drugs do not work as well against even the wild type strains that are found in other regions (let alone their "super bugs"). Since we are not controlled by the desire to make profit, we are devoting some of our research efforts to the groups of HIV strains that affect the often-neglected patients in Africa and Asia. In addition, studying these versions of HIV protease can also help us learn how to defeat the "super bugs" we find here in the USA.

This experiment is the first one in which we are using the "Asinex" library of over 360,000 different compounds in our virtual screens against HIV protease.

This experiment involves faah6022 - faah8233.

These calculations began 4/23/2009 and ended 10/20/2009.

Experiment 28: 100% Completed

This experiment utilizes the same set of brand new crystal structures and new models of HIV protease that are being used in Experiments 25-27. See the description of Experiment 25 for the details about these new targets. In experiment 28, we are AutoDocking these compounds against the active site of different variants of HIV protease.

This experiment incorporates a library of ligands that we just started using: the ChemBridge building blocks library. The library of "building blocks" from ChemBridge contains many small fragments that were derived from larger compounds. Using these small fragments, or building blocks, should help us cover a larger amount of structural diversity (i.e., of "chemical space") within each experiment. Thus, this library should help us find new hits in a more efficient way.

This experiment involves faah5710-6021.

These calculations began 4/05/2009 and finished 4/24/2009.



Feb. 2011 update: 10 of the top-ranked fragments were purchased and are currently being tested by our collaborators at TSRI. Preliminary data indicate that 2 of these 10 fragments are able to inhibit HIV protease, according to the standard FRET-based activity assay. More testing needs to be done before we can publish these results, but thus far it appears that we discovered 2 completely novel "hits" against the active site/"eye site" in the results of Experiment 28.



Extension to Experiment 28

736 compounds from the "ZINC" server that are somewhat similar to these 2 active fragments (discussed above) were hand-picked by Dr. Alex L. Perryman. These 736 compounds generate a "focused library" that we are screening against the active site and the "eye site" of a panel of 9 different variants of HIV protease. Since these 736 compounds are larger than the two active fragments discovered in Experiment 28, this extension of the experiment is searching for more potent compounds that are also able to target both the "eye site" and the floor of the active site.

This extension involves faah22,621 - faah22,629.

These calculations began 6/01/2011 and ended 6/04/2011.

Experiment 27: 100% Completed

This experiment utilizes the same set of brand new crystal structures and new models of HIV protease that are being used in Experiments 25 and 26. See the description of Experiment 25 for the details about these new targets. In experiment 27, we are targeting the exo site on the sides of HIV protease.

This experiment incorporates a library of ligands that we have never used before. We recently downloaded and reformatted the "ChemBridge building blocks library" of ~ 12,000 models of compounds from "ZINC," (which stands for Zinc Is Not Commercial). See the paragraph below for a few details about ZINC. The library of "building blocks" from ChemBridge contains many small fragments that were derived from larger compounds. Using these small fragments, or building blocks, should help us cover a larger amount of structural diversity (i.e., of "chemical space") within each experiment. Thus, this library should help us find new hits in a more efficient way.

This experiment involves faah5398 - faah5709.

These calculations began 3/26/2009 and finished 4/09/2009.

Experiment 26: 100% Completed

Similar to Experiment 25, this experiment also involves screening the NCI's "DTP library of moderately active compounds" against several brand new structures and new models of HIV protease. But in this experiment, we are docking the potential inhibitors against the active site, instead of the exo site. See the description of Experiment 25 for the details regarding the new structures and models that we are targeting.

This experiment involves faah5320 - faah5397.

These calculations began 3/16/2009 and ended on 3/31/2009.

Experiment 25: 100% Completed

This experiment involves screening the NCI's "DTP library of moderately active compounds" against the exo site of several brand new structures and new models of HIV-1b protease. (See the description of "Experiment 21" for more details about this DTP library.) This experiment targets the potential allosteric inhibitor site (i.e., the "exo site") on the sides of HIV protease. The new structures utilized in this experiment are based on brand new, currently unpublished x-ray crystallographic structures from our collaborator, Prof. David Stout.

The new models of HIV protease that are also included in this experiment were harvested from Molecular Dynamics simulations recently performed by Dr. Alex L. Perryman. We harvested the equilibrated structures from the beginnings of our new MD simulations on several different multi-drug-resistant mutants of HIV-1b protease (i.e., several different "super bugs" against which the current drugs no longer work well.) We also included models of HIV-1c protease ("1c" is the HIV subtype, or group of strains, that are most common in Asia) and HIV-2 protease ("HIV-2" is the predominant subtype in Africa). "HIV-1b" is the subtype of HIV most commonly found in the U.S. and in Europe. The current anti-AIDS drugs were all designed and optimized against HIV-1b, but the FightAIDS@Home project is devoted to trying to help all patients with HIV throughout the world.

This experiment involves faah5224 - faah5319.

These calculations began on 03/03/2009 and ended on 3/31/2009.

Experiment 24: 100% Completed

Similar to Experiment 12, this experiment performed by Dr. Ruth Huey involves HIV protease "cross-docking" (i.e., this is a test of the new AutoDock code and the new scoring function that involves docking all the known HIV protease inhibitors against 100 different crystal structures of HIV protease).

This experiment involved faah4998 - faah5017.

These calculations began 08/11/2008 and finished 08/31/2008.

Experiment 23: 100% Completed

Similar to Experiments 19 and 19a, this Relaxed Complex experiment involves docking the different FDA-approved HIV protease inhibitors (and a few compounds still in development) against the active site of 2,000 different snapshots of HIV-1b protease that were harvested from a Molecular Dynamics simulation. However, this experiment involves docking these reference compounds against conformations of the V82F/I84V multi-drug resistant "super bug" of HIV protease. We'll compare the performance of these compounds in this experiment versus their calculated affinities from Experiments 19 and 19a.

The reference compounds used in Experiments 19, 19a, and 23 include the FDA-approved drugs amprenavir, atazanavir, darunavir, indinavir, lopinavir, nelfinavir, ritonavir, saquinavir, and tipranavir, while the compounds in development include AB2, AB3, JE-2147, KNI-272, TL3, and TMC-126.

This experiment involves faah4726 - faah4791.

These calculations began 12/01/2008, and they finished 02/19/2009.

Experiment 22: 100% Completed

This Relaxed Complex experiment is very similar to Experiment 21, but this time the "DTP" library is being docked against the potential allosteric inhibitor site on the peripheral surface of HIV protease (i.e., the "exo" site), instead of docking them against the active site. These compounds are also being docked against the same "QR-selected" subset of conformations from the V82F/I84V multi-drug-resistant mutant of HIV protease. For a description of the QR method and a few citations, see the description of Experiment 21.

Experiment 22 involves faah4417 - faah4622 and faah5018 - faah5223.

These calculations began 10/18/2008, and they ended 03/05/2009.

Experiment 21: 100% Completed

This Relaxed Complex experiment is testing both a new library of ligands and a new method for selecting the snapshots of the target from MD against which to dock these compounds. The first 1/3 of ligands from the NCI's "DTP" library of compounds is being tested now, while we prepare the files that describe the other 2/3 of this library of compounds.

**Update 07/31/2008** All 3/3 of this library have now been prepared by Dr. Stefano Forli. The other 2/3 of this experiment have been submitted.

These compounds were "moderately active" in cell-based assays at the NCI, but noone knows which targets these bind to or how they are able to inhibit them. These compounds are being docked against the active site of a "QR-selected" subset of conformations harvested from Molecular Dynamics simulations of the V82F/I84V multi-drug-resistant mutant of HIV protease (i.e., a target from one of the most drug-resistant "super bugs" of HIV). The Structure QR method is a new tool for selecting a structurally diverse, non-redundant set of conformations from a group of different structures that have similar sequences.

We thank John Eargle of the Luthey-Schulten lab at UIUC for helping us learn how to apply this method. For more info. on QR, see P. O'Donoghue and Z. Luthey-Schulten; Evolutionary profiles derived from the QR factorization of multiple structural alignments gives an economy of information; J. Mol. Biol., 346, 875-894, (2005). See also MultiSeq of VMD: Elijah Roberts, John Eargle, Dan Wright, and Zaida Luthey-Schulten; MultiSeq: Unifying sequence and structure data for evolutionary analysis; BMC Bioinformatics, 7:382 (2006).

This experiment involves faah4314 - faah4416, faah4623 - faah4725, and faah4792 - faah4997.

These calculations began 08/28/2008, and they finished 2/19/2009.

Experiment 20: 100% Completed

This Relaxed Complex experiment is similar to Experiment 16, but different run parameters are being used during the docking and several different protocols for preparing the input files of these compounds are being tested (such as using different protocols to calculate the charges on the atoms within each ligand, using different "atom types" to describe the ligands, and using different protocols for minimizing the structures of the ligands). Thus, this experiment will also allow us to investigate the best way(s) for preparing ligands that will be used in subsequent Relaxed Complex experiments.

This experiment involves faah4202 - faah4313.

These calculations began 06/08/2008, and they ended on 9/30/2008.

Experiment 19a: 100% Completed

This Relaxed Complex experiment involves docking the different FDA-approved HIV protease inhibitors (and a few compounds still in development) against the active site of 2,000 different snapshots of the wild type HIV-1b protease that were harvested from the same Molecular Dynamics simulation discussed above in Exp. 19. These calculations will provide a base-line against which to compare the performance of the compounds used in Experiment 19. Different protocols for preparing the input files of these current drugs were used (such as using different protocols to calculate the charges on the atoms within each ligand, using different "atom types" to describe the ligands, and using different protocols for minimizing the structures of the ligands). Thus, this experiment will also allow us to investigate the best way(s) for preparing ligands that will be used in subsequent Relaxed Complex experiments. The reference compounds used in Experiment 19a include the FDA-approved drugs amprenavir, atazanavir, darunavir, indinavir, lopinavir, nelfinavir, ritonavir, saquinavir, and tipranavir, while the compounds in development include AB2, AB3, JE-2147, KNI-272, TL3, and TMC-126.

This experiment involved faah4070 - faah4201.

These calculations began 05/26/2008 and finished 08/19/2008.

Experiment 19: 100% Completed

This experiment is similar to Exp. 13, but different run parameters are being used (for example, 2 point cross-over is being used, while Exp. 13 used the arithmetic crossover protocol in the genetic algorithm used in the docking calculations; a new and improved version of the AutoDock code is being used, as well). Exp. 19 is a Relaxed Complex experiment of the "9 false negatives" from the NCI Diversity Set. These 9 compounds did not dock well in previous experiments (by Max Chang and Dr. Lindy Lindstrom) that targeted different crystal structures of HIV protease, but they did display some activity in an experimental assay against HIV protease. Different versions of these 9 ligands and of a few reference compounds are being docked against the active site in 2,000 different snapshots of the wild type protease that were harvested from Dr. Alex Perryman's previously-published Molecular Dynamics simulations (i.e., the cover article of the April, 2004, issue of Protein Science).

The reference compounds used in Experiment 19 include the FDA-approved drugs amprenavir, atazanavir, darunavir, indinavir, lopinavir, nelfinavir, ritonavir, saquinavir, and tipranavir, while the compounds in development include AB2, AB3, JE-2147, KNI-272, TL3, and TMC-126.

To view some of the recent results from the Relaxed Complex experiments on these reference compounds, see the graph with part of Indinavir's Relaxed Complex "trajectory" or the comparison of AB2's versus AB3's RC "trajectories" at the bottom of this page.

This experiment involved faah4000 - faah4069.

These calculations began 05/02/2008 and finished 06/14/2008.