A master chemist used lessons from nature’s molecules to design anti-HIV agents.  Now one of his creations is helping tens of thousands of HIV-positive children and adults.

The HIV protease inhibitor (PI) pipeline was just starting to trickle.  In 1995 saquinavir became the first PI to earn FDA approval, and it was co-credited for the rapid turnaround in America’s rising AIDS death toll.1 But saquinavir is vulnerable to changes in the enzyme that it plugs up; alterations to the binding site that weaken its interaction can appear with repeated use and cause drug resistance.  It also has poor pharmacokinetics, due in part to its size and peptide-like structure.2 Four years and three other PIs later, amprenavir was approved and it addressed some of saquinavir’s shortcomings, being smaller and less peptide-like.  But even with amprenavir, the available PIs were overmatched by emerging multidrug-resistant proteases.


By this time Professor Arun Ghosh had nearly two decades of experience synthesizing natural products and designing anti-HIV agents, including PIs.3 He planned to graft cycloethers like those found in natural products onto the amprenavir scaffold, which could increase drug-enzyme interactions and its antiviral effects.  Ghosh hypothesized that these changes would yield improved second-generation PIs.2

In this first example, the spirocyclic system (red) in bacterial antibiotic monensin A inspired a simplified spiro system with repositioned oxygens,2 which was substituted for the cycloether (THF; green) in amprenavir.  The oxygens and their spatial configurations were critical:  replacing the THF (5-membered ring) oxygen with methylene or inverting stereochemistry at the spirocenter significantly worsened potency.4 Although Compound 14’s inhibitory potency versus isolated HIV protease was in the nanomolar range (Ki = 20 ± 3 nM), its potency in a cell-based viral replication and infection assay was two orders of magnitude weaker (ID50 = 2.4 µM).4 So further substrate-mimicking structures needed to be tried.


A second natural product inspired Ghosh to evaluate fused bicyclic moieties in place of amprenavir’s cycloether.  He was familiar with ginkgolide B, having completed its total synthesis as a postdoctoral researcher.5 Now nearly 20 years later, he replicated its bis-THF (blue) on the amprenavir scaffold—said another way, he fused a second THF onto amprenavir’s THF.6  The resulting compound UIC-94017 has improved potency versus the first analogue:  its enzyme inhibitory potency is ten-fold better (Ki = 2.1 nM) and its cellular potency is nearly a thousand-fold better (ID50 = 4.5 nM).6 Remarkably, the new molecule thrives against at least a dozen drug-resistant HIV strains, against which other PIs are ineffective at concentrations below 5 μM.  At nanomolar concentrations (IC50 = 3 to 29 nM), UIC-94017 blocks infectivity and replication of HIV strains that are resistant to PIs saquinavir, indinavir, nelfinavir, and ritonavir.  It is even active against multidrug-resistant HIV isolated from patients for whom no other PI would work.7


Why should the addition of another THF be so beneficial?  For one, UIC-94017’s binding affinity for wildtype protease is two orders of magnitude better than amprenavir’s (Kd = 4.5 × 10−12 M versus 3.9 × 10−10 M).8 Also, its binding interactions are mostly with the enzyme backbone, whose shape remains stable from wildtype to mutant variants, rather than with the more variable side chains.  This may explain UIC-94017’s ability to override enzyme mutations that deactivate other PIs.2


UIC-94017’s interactions with the binding site can be seen above, as well as clean views of the ligand from two angles.  The models were obtained from X-ray co-crystal structures of the ligand with HIV protease.7,9

UIC-94017 is now better known as darunavir (brand name Prezista), a 2006 FDA-approved oral drug that is used to treat children and adults living with HIV-1.  Darunavir is a part of drug regimens for tens of thousands10 of treatment-naïve and treatment-experienced patients due to its extraordinary resistance profile.11 It works through a dual mechanism of action:  it not only competes with viral substrates for the protease active site, it also prevents dimerization of the subunits that come together to form competent protease.7 These molecule-level actions translate to decreased production of mature, infectious virus and reduced viral load.  Darunavir’s pharmacokinetic profile is another advantage.  For example, its bioavailability is 82% and its serum half-life is 15 hours with ritonavir booster, while saquinavir’s bioavailability is 4% with a serum half-life of 1–2 hours.12 Darunavir must be taken with food to increase oral absorption, but the type of food does not affect it, unlike first-generation PI indinavir.12


HIV rapidly mutates and resistance to darunavir, although uncommon, can happen with at least eight protease mutations.13 So Ghosh’s laboratory built and tested darunavir-like molecules because they anticipated darunavir resistance.  One successful direction was to fuse a third THF to bis-THF, producing an unprecedented tris-THF structure.14 GRL-0519, like darunavir, is effective at nanomolar concentrations versus multidrug-resistant HIV strains—in fact, it is 10-fold more potent than darunavir across different strains (IC50 = 0.6 to 4.3 nM) and has 10-fold stronger enzyme dimerization inhibitory activity.7


Although GRL-0519 is not being studied in humans,16 it or another analog may be in the future.  PIs are still common in treatment regimens because they help manage HIV infection as a chronic illness.  But future PIs will need to address continuing problems like toxicity, multidrug resistance, and latent HIV reservoirs, all of which contribute to the diminished life expectancy and health of HIV-infected patients.17

Complex natural products from a bacterium and the ginkgo tree inspired Ghosh’s design of novel HIV protease inhibitors, including a frontline therapeutic for HIV infection.  This strategy of mimicking nature’s molecules continues to deliver many new medicines.18


Please note that this is not medical advice.


1. Centers for Disease Control and Prevention, Morbidity and Mortality Weekly Report, June 3, 2011, Vol. 60, No. 21. http://www.cdc.gov/mmwr/PDF/wk/mm6021.pdf

2. Bioorg. Med. Chem. 2007, 15, 7576–7580.

3. https://www.chem.purdue.edu/ghosh/allpublications.htm (Accessed 12/11/2016)

4. Bioorg. Med. Chem. Lett. 1998, 8, 979–982.

5. J. Am. Chem. Soc. 1988, 110, 649–651.

6. Bioorg. Med. Chem. Lett. 1998, 8, 687–690.

7. https://www.chem.purdue.edu/ghosh/research-hiv.htm (Accessed 12/11/2016)

8. J. Virol. 2004, 78, 12012–12021.

9. PDB code 3D1Z (DOI: 10.2210/pdb3d1z/pdb)

10. http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/PediatricAdvisoryCommittee/UCM234430.pdf (Accessed 12/12/2016)

11. http://www.fda.gov/ForPatients/Illness/HIVAIDS/History/ucm151081.htm (Accessed 12/11/2016)

12. Pharmacology for Nursing Care. 8th Ed. pp1186–1187. Elsevier. Richard A. Lehne.

13. http://www.accessdata.fda.gov/drugsatfda_docs/label/2012/021976s021lbl.pdf (Accessed 12/11/2016)

14. ChemMedChem. 2010, 5, 1850–1854.

15. PDB code 3OK9 (DOI: 10.2210/pdb3ok9/pdb)

16. Clinicaltrials.gov (Searched terms: “hiv protease inhibitor” | Recruiting | “Anti-HIV Agents” | Studies received from 01/01/2005 to 12/11/2016; Accessed 12/11/2016)

17. The Lancet. 2008, 372, 293–299.

18. J. Nat. Prod. 2016, 79, 629–661.