Design and synthesis of HIV-1 protease inhibitors

(1)

Comprehensive Summaries of Uppsala Dissertations

from the Faculty of Pharmacy 245

_____________________________ _____________________________

Design and Synthesis of HIV-1

Protease Inhibitors

BY

MATHIAS ALTERMAN

ACTA UNIVERSITATIS UPSALIENSIS UPPSALA 2001

(2)

Dissertation for the Degree of Doctor of Philosophy (Faculty of Pharmacy) in Organic Pharmaceutical Chemistry presented at Uppsala University in 2001

ABSTRACT

Alterman, M. 2001. Design and Synthesis of HIV-1 Protease Inhibitors. Acta

Universitatis Upsaliensis. Comprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy 245. 70 pp. Uppsala. ISBN 91-554-4906-9.

Human Immunodeficiency Virus (HIV) is the causative agent of Acquired Immune Deficiency Syndrome (AIDS). The C2-symmetric HIV-1 protease is one of the prime

targets for chemotherapy in the treatment of the HIV infection. Inhibition of HIV-1 protease leads to immature and non-infectious viral particles. The design and synthesis of a number of C2-symmetrical C-terminal duplicated HIV-1 protease inhibitors and

subsequent biological evaluation is presented in this thesis.

A versatile three step synthetic route has been developed using a carbohydrate as an inexpensive chiral starting material thus allowing inhibitors with the desired

stereochemistry to be obtained. By this efficient method a series of tailor-made P2/P2' modified inhibitors were synthesized, and these were evaluated on purified HIV-1 protease and in HIV-1 infected cell assays. Highly active HIV-1 protease inhibitors were identified among the tested compounds. Analyses of the X-ray crystal structures of two of the most active compounds, as complexes with the protease, guided the further design of P1/P1' elongated inhibitors. Substitutions in the para-position of the P1/P1' benzyl groups were promoted efficiently by microwave-irradiated of palladium-catalyzed reactions. Particular modifications in the P1/P1' region of the inhibitors resulted in a 40-fold increase of the anti-viral activity on HIV-1 infected cells. Furthermore, a fast, efficient, and general one-pot microwave enhanced synthesis protocol for transformations of organo-bromides to tetrazoles was developed and applied on the inhibitor scaffold. Attachment of linker molecules to the P1/P1' benzyl groups of one inhibitor was used to develop of sensitivity enhancer tools in surface plasmon resonance biosensor assays. These new assays enable the evaluation of low-molecular weight compounds as HIV-1 protease inhibitors.

Mathias Alterman, Organic Pharmaceutical Chemistry, Department of Pharmaceutical Chemistry, Uppsala University, Box 574, SE-751 23 Uppsala Sweden

(3)

PAPERS DISCUSSED

This thesis is based on the following papers.

I. Alterman, M.; Björsne, M.; Mühlman, A.; Classon, B.; Kvarnström, I.; Danielson, H.; Markgren, P. O.; Nillroth, U.; Unge, T.; Hallberg, A.; Samuelsson, B. Design and Synthesis of New Potent C2-Symmetric HIV-1

Protease Inhibitors. Use of L-Mannaric Acid as a Peptidomimetic Scaffold. J. Med. Chem. 1998, 41, 3782-3792.

II. Alterman, M.; Andersson, H. O.; Garg, N.; Ahlsén, G.; Lövgren, S.; Classon, B.; Danielson, U. H.; Kvarnström, I.; Vrang, L.; Unge, T.; Samuelsson, B.; Hallberg, A. Design and Fast Synthesis of C-Terminal Duplicated Potent C2

-Symmetric P1/P1'-Modified HIV-1 Protease Inhibitors. J. Med. Chem. 1999, 42, 3835-3844.

III. Alterman, M.; Hallberg, A. Fast Microwave-Assisted Preparation of Aryl and Vinyl Nitriles and the Corresponding Tetrazoles from Organo-halides. J. Org. Chem. 2000, 65, 7984-7989.

IV. Alterman, M.; Sjöbom, H.; Säfsten, P.; Markgren, P. O.; Danielson, U. H.; Hämäläinen, M.; Löfås, S.; Hultén, J.; Classon, B.; Samuelsson, B.; Hallberg, A. P1/P1' Modified HIV Protease Inhibitors as Tools in Two New Sensitive Surface Plasmon Resonance Biosensor Screening Assays. Eur. J. Pharm. Sci.

2001, Accepted.

(4)

(5)

Contents

ABBREVIATIONS

Ac acetyl

Arg arginine Asp aspartic acid

9-BBN 9-borabicyclo[3.3.1]nonane Cbz carbobenyloxy

CD4 receptor on the surface of cells with in the immune system CSA (±)-camphorsulfonic acid

DCM dichloromethane

DDQ 2,3-dichloro-5,6-dicyano-1,4-benzoquinone

DIEA N,N-diisopropylethylamine

DNA deoxyribonucleic acid DMAP 4-(dimethylamino)pyridine DMF dimethylformamide

DSC N,N'-disuccinimidyl carbonate

ED50 50% inhibitory concentration in cell-assay

EDC 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride FDA US food and drug administration

Gag polyprotein containing structural proteins gag-gene genome for Gag polyprotein

Gln glutamine Gly glycine

gp120,41 glycoprotein 41 and 120 HOBT 1-hydroxybenzotriazole hydrate HOAc acetic acid

IC50 concentration of the inhibitor resulting in 50% inhibition

Ile isoleucine IN integrase

Ki inhibitory constant

M46I methionine in position 46 of the protease is mutated to isoleucin MT-4 CD4+ lymphoblastoid cells

NHS-LC-Biotin succinimidyl-6-(biotinamido)hexanoate NNRTI non-nucleoside reverse transcriptase inhibitor NRTI nucleoside reverse transcriptase inhibitor p7,17,24 protein 7, 17, and 24

Phe phenylalanine

Pol polyprotein containing functional enzymes Pol-gene genome for Pol polyprotein

PR HIV protease Rf retardation factor

RNA ribonucleic acid RT reverse transcriptase

RU resonance unit, arbitrary unit in SPR measurement SPR surface plasmon resonance

TBDMSCl tert-butyldimethylsilyl chloride

TEMPO 2,2,6,6-tetramathyl-1-piperidyloxy, free radical TFA triflouroacetic acid

THF tetrahydrofuran Thr threonine

(7)

Introduction

1. INTRODUCTION

1.1 Acquired Immune Deficiency Syndrome (AIDS)

In 1981 an increased occurrence of unusual cases of Pneumocystis carinii pneumonia and Kaposi´s cancer together with other opportunistic infections, was observed among previously healthy homosexual men and intravenous drug abusers in the USA.1,2_An

underlying immunosuppression was found to be promoter of these rare diseases. This syndrome became known as Acquired Immune Deficiency Syndrome (AIDS).3

In 1983 the causative agent of AIDS was identified as a human retrovirus, first isolated in France from a patient with multiple lymphadenopathies,4_{a condition linked to AIDS,}

and subsequently in 1984, from AIDS patients.5,6_{Initially, three different names were}

given to the virus isolated from AIDS patients; human T lymphotropic virus III (HTLV-III),5_{lymphadenopathy-associated virus (LAV),}7_{and AIDS-associated retrovirus}

(ARV).6_{Eventually the AIDS-causing virus was in 1986 given an alternative name,}

human immunodeficiency virus (HIV).8

A few years later a second similar virus, HIV-2, was isolated from patients in West Africa.9_{Both HIV subtypes can lead to AIDS, although the pathogenic course with}

HIV-2 might be longer. The genome homology of HIV-1 and HIV-2 are approximately 40%.10

Retrospective studies indicate that the first documented case of AIDS occurred in Central Africa in 1959 and the source of the virus is proposed to come from the same geographic area.11_{The origin of the two viruses has now been shown to be derived}

from two African monkeys, the chimpanzee (Pan troglodytes troglodytes) for HIV-112

and the sooty mangabay (Cercocebus atys) for HIV-2.13

A striking and somewhat unique feature of HIV is that the virus infects the helper T-lymphocytes, which exert a central role in the regulating of the immune response. 5,6,14-16_{Since HIV infection causes depletion of helper T-lymphocytes, AIDS patients}

demonstrate a weakened immune system. Thus, the gradual depletion of these cells makes the patient increasingly susceptible to opportunistic infections of bacterial, viral or fungal origin and to certain cancers, which are key features of the final stage of the HIV infection, i.e. AIDS.17

The helper T-lymphocytes were the first cell types to be identified as targets for HIV. Viral infections in macrophages and monocytes were recognized shortly after.

Moreover, it was discovered that the helper T-lymphocyte surface marker, CD4, was the receptor for the HIV viral surface glycoprotein gp120.18,19_{Since then, however, many}

(8)

Introduction

other cell types have been shown to be infected by the virus, including the cells in the brain, and the nervous system. The HIV virus enters the central nervous system at an early stage of the infection and forms a reservoir in the brain as evidenced by the presence of large quantities of unintegrated viral DNA in the brains of HIV infected individuals.20

The course of the HIV infection is reflected by the concentration of the CD4+ helper T-lymphocytes in the blood (Figure 1). Normally the concentration is 800-1200 cells/mm3 but in the final phase of the disease the number is < 200 cells/mm3 of blood.21

Figure 1. The natural history of HIV infection.

The primary infection, defined as the period during which the HIV infection is

established in the host, is characterized in 30 – 70% of the patients by a systemic illness including fever, headache, rash, pharyngitis, gastrointestinal disturbance, and

lymphadenophathy.22,23_{Subsequently, an asymptomatic phase follows that varies in}

length but has a mean of 10 years, prior to the final phase of the disease, when the individuals become highly vulnerable to infections by other viruses and microbial organisms (opportunistic infections).24_{The HIV infection in muscles and the central}

nervous system results in muscular wastage and AIDS-related dementia.25,26_The

average life expectancy without therapy from the appearance of AIDS is 1-2 years in developed countries.24

A decade ago, HIV/AIDS was primarily regarded as a serious health crisis. It was predicted in 1991, that in sub-Sahara Africa, by the end of the decade, nine million people would be infected and five million would die. This was later shown to be a threefold underestimate. Today, AIDS is undoubtedly a development crisis, and in some

AIDS CD4+ cells

Plasma viral load

(9)

Introduction

infected under 1999. Altogether more than 18 million people have died since the beginning of the epidemics.27

Figure 2. Number of people living with HIV/AIDS – total: 34.3 million (December 1999).

1.2 Viruses Microorganisms

The unicellular microorganism (like bacteria) however small and simple, are cells. Unicellular microorganisms always contain DNA as the repository of genetic information, and in addition, also RNA. These cells have their own machinery for producing energy and macromolecules. Thus, the unicellular microorganisms grow by synthesizing their own macromolecular constituents (nucleic acids, proteins,

carbohydrates, and lipids), and in the majority of cases, multiply by binary fission.21

Viruses

Viruses, on the other hand, are not cells. They are completely dependent on their cellular hosts for the machinery of energy production and synthesis of macromolecules. The virus particle contains only one type of nucleic acid, either DNA or RNA, never both, and differs from non-viral organisms by having two clearly defined phases in their life cycle. In the first phase (the transmission phase) outside a susceptible cell, the virus particle is metabolically inert. In the second phase (the reproductive phase) inside the cell, the viral genome exploits the metabolic pathways of the host to produce progeny genomes and viral proteins that assemble to form new infectious virus particles called virions.21

The primary criteria for the delineation of virus families are; (1) the kind of nucleic acid that constitutes the genome, either DNA or RNA, with DNA viruses the predominant group, (2) the mechanism of viral replication, (3) and the morphology of the virion.21

North America 900 000 Caribbean 360 000 Latin America 1.3 million Western Europe 520 000 North Africa & Middle East

220 000 sub-Saharan

Africa 24.5 million

Eastern Europe & Central Asia

420 000

East Asia & Pacific 530 000 South &

South-East Asia 5.6 million

Australia & New Zealand

(10)

Introduction

1.3 Human Immunodeficiency Virus (HIV)

HIV-1 and HIV-2 are RNA viruses and belong to the family of retroviruses, Retroviridae (retro, backwards). The genome of retroviruses consists of duplicate copies of positive single-stranded RNA. Once a cell has become infected with a retrovirus the viral genetic information will be transformed from RNA to DNA, catalyzed by viral enzyme reverse transcriptase. The name retrovirus is derived from this unique event, which is completely opposite to the normal process where RNA is transcribed from DNA. Retroviruses are divided into seven genera, where the genus Lentivirus (lenti, slow), is characterized by the slow development of disease after infection. HIV is a typical lentivirus, since it usually has a disease latency of several years.21

A schematic drawing of the mature HIV virion is shown in Figure 3. The virion is almost spherical and is about one ten-thousandth of a millimeter across (ca. 100 nm).28

The virus is enveloped by a lipid bilayer that is derived from the infected host cell. The outer surface is studded with surface glycoproteins (gp120) that are anchored to the virus via interactions with the transmembrane protein (gp41). These surface proteins play a crucial role when HIV binds to and enters the host cells. A shell of the matrix protein (p17) lines the inner surface of the viral membrane, and a conical capsid core particle constructed out of the capsid protein (p24) is located in the center of the virus. The capsid particle encapsulates two copies of the viral genome, stabilized by the nucleocapsid protein (p7), and also contains three essential virally encoded enzymes: protease (PR), reverse transcriptase (RT), and integrase (IN).29

RNA PR IN RT gp120 gp41 Lipid bilayer p17 p24 p7

(11)

Introduction

1.4 Replicative Cycle of HIV

The attachment of the viral surface protein (gp 120) to the CD4-receptor, located on various cells within the immune system, initiates the replicative cycle of HIV (Figure 4).30_{Attached virions utilize several additional cell-surface proteins to promote the}

fusion of the viral and host cellular membranes.31-34_{Membrane fusion is followed by a}

poorly understood uncoating event of the capsid that allows the release of the viral content into the host-cell cytosol. The single-stranded viral RNA complexes with reverse transcriptase, which catalyses the reverse transcription to yield a double-stranded DNA molecule.35-37_{The double-stranded viral DNA is then transported into}

the cell nucleus and is permanently integrated into the host genome by the catalytic activity of the viral integrase. The integrated viral DNA is designated provirus.38

Figure 4. Schematic drawing of the replicative cycle: 1. Attachment to the host cell CD4-receptor, 2.

Viral fusion and uncoating, 3. Reverse transcription, 4. Integration of viral DNA to the host genome, 5. Translation, 6. Viral budding, 7. Maturation via protease activities.

By an unknown activation process the cell initiates the transcription of the proviral DNA by the host cellular RNA polymerase II. Initially, short spliced RNA species that encode the regulatory proteins Tat, Rev, and Nef are synthesized. Tat acts as a

stimulator of the transcription of the proviral DNA to enhance the production of viral RNA.39-41_{Full-length and singly spliced RNA is needed in the cytoplasm for the}

synthesis of Gag and Gag-Pol polyproteins, and for packing into new virions. Rev binds to the full length and singly spliced RNA in the nucleus and protects it from further splicing and actively transports it to the cytosol. In this manner, Rev acts as a switch between the early synthesis of highly spliced RNAs and the later synthesis of unspliced

1 2 3 4 5 6 7

(12)

Introduction

and singly spliced RNAs.42-44_{Nef acts as a down-regulator of the number of CD4}

receptors on the surface of the infected cell.45

Translations of the unspliced RNA by the ribosomes produce the polyproteins Gag and Gag-Pol. These polyproteins are transported to the plasma membrane with two

molecules of viral RNA. They assemble together with the envelope protein to form an immature virus particle that is released from the cell by budding from the cell surface. To become infectious, the virion has to pass through a maturation process where the enzyme HIV protease cleaves the polyproteins into functional enzymes and structural proteins. The mature HIV virion is now ready to infect a new cell and start a new replicative cycle.29,38

1.5 Targets for Anti-HIV Chemotheraphy

In principal, every step in the HIV replication cycle can be considered as a potential target for anti-viral chemotherapy. However, the number of practical targets for drug interventions is reduced due to the fact that the virus is an intracellular parasite, which relies on the metabolic pathways of the host cell. Hence, most agents that block the replication of the virus are also lethal to the host cell. The key in selective anti-viral therapy is therefore to identify any process that is essential for the replication of the virus, but not for the survival of the cell.21,46_{The gained knowledge about the}

replicative cycle of the HIV-virus has led to the extraction of virus-specific processes. Predominantly, scientists have focused their attentions on the following processes: a) viral binding to target cells, b) virus cell fusion, c) virus uncoating, d) reverse

transcription of genomic RNA, e) viral integration, f) gene expression, and g) protease activity. So far, the two strategies d and g have been proven to be the most successful in the search for drugs that can be used for treatment of AIDS.47-50_{These two targets will}

be described briefly in the following sections.

1.6 Reverse Transcriptase Inhibitors

Virally encoded HIV reverse transcriptase (RT) catalyses the replication of single-stranded viral RNA to a double-single-stranded DNA. Inhibition of RT prevents the formation of this double-stranded DNA that can be integrated in the host DNA. Reverse

transcriptase inhibitors can be divided into two categories, nucleoside (NRTI) and non-nucleoside reverse transcriptase inhibitors (NNRTI).51

(13)

Introduction

(3'-azido-3'-deoxythymidine, AZT) was found to be active against HIV replication in cell culture and was approved for treatment of HIV.54_{Today, there are six NRTI}

approved (Figure 5). Unfortunately, all of these substances are associated with side effects, such as bone marrow suppression, peripheral neuropathy and acute

pancreatitis.55,56_{In addition, prolonged treatment with these compounds gives rise to}

clinical resistance.47,57

Figure 5. Nucleotide Reverse Transcriptase Inhibitors (NRTI) approved by the FDA.

The NNRTIs are a diverse group of compounds, which non-competitively interact with an allosteric site of HIV-1 RT and thereby inactivate the enzyme without need for pre-activation of the drugs. NNRTIs bind in a highly hydrophobic pocket of the enzyme and exhibit grater affinity for the enzyme-substrate complex than for the free enzyme.46_The

hydrophobic allosteric site is unique to HIV-1 RT and is not found in other RTs or DNA polymerases. This results in a high selectivity index and a low toxicity of the NNRTs. However, rapid eliciting resistance is a major problem with this type of inhibitor as well.51_{Currently, three NNRTIs are used in clinic for the treatment of HIV infection}

(Figure 6).

Figure 6. Non Nucleotide Reverse Transcription Inhibitors (NNRTI) approved by the FDA. O N HO N3 NH O O O N HO N N NH O O HO N N NH2 O S O HO N N NH2 O O N HO NH O O O HO N N N N HN NH2

Zidovudine58 Zalcitabine59 Didanosine60

Stavudine61 Lamivudine62 Abacavir63

NH N N N O N H O Cl O F3C N H N O N N HN HN S O O

(14)

Introduction

1.7 HIV Protease Inhibitors

HIV protease was first suggested as a potential target for AIDS therapy by Kramer et al. after it was shown that a frameshift mutation in the protease region of the pol-gene prevented cleavage of the Gag polyprotein precursor, which is essential for the

maturation of the HIV particles.67_{Blockage of HIV protease leads to the formation of}

immature non-infectious virions.68_{Compounds, having the ability to inhibit this}

protease have been studied intensively during the last decade and numerous reports of potent HIV-1 protease inhibitors have been published.69-74

Figure 7. Protease Inhibitors approved by the FDA.

Saquinavir was the first approved protease inhibitor and has been in clinical use since 1995.75_{Presently, there are six clinically approved protease inhibitors (Figure 7).}

Although the inhibitors on the market are highly selective they induce side effects such as lipodystrophy, hyperlipidaemia, insulin resistance,76-82_{and emergence of resistant}

mutants upon prolonged use.83-86_{Therefore there will probably be a constant demand}

N H H N N O OH O NH2 O N H H CONH-t-Bu O H N N H H N N O OH O N S O N S N H N OH O HO _N CON-t-Bu N H N N OH S H H CONH-t-Bu O HO O N H N S OH O O NH2 O O N H N N H O OH O O HN O Saquinavir75 Ritonavir87 Indinavir88 Nelfinavir89 Amprenavir90 Lopinavir91

(15)

Introduction

1.8 HIV-1 Protease

Enzymes, such as HIV protease, are nature’s own catalysts. Proteases are a diverse class of enzymes that catalyze the cleavage of peptides or proteins. The interaction of a substrate (peptide or protein) with the protease lowers the activation energy of the peptide bond cleavage reaction by stabilizing the transition state, which is the molecular arrangement with the highest free energy.92_{The enzymes thus enable fast reactions at}

room temperature that otherwise would have required extended reactions times. The half-life of a peptide bond at room temperature at neutral pH has been estimated to be about 500 years.93

Based on the presence of the characteristic signature amino acid sequence, Asp-Thr-Gly, it was suggested by Toh et al. in 1985 that the protease of HIV might belong to the family of aspartic proteases.94_{This was confirmed through pepstatine A inhibition, an}

aspartic protease selective inhibitor,95_{and by site-directed mutagenesis of the active site}

Asp 25, which led to abolition of the catalytic activity.96_{The aspartic proteases are}

well-characterized group of enzymes that can be found in vertebrates, plants, in addition to in fungi. Examples of proteases from the aspartic protease class are pepsin, cathepsin D, renin, chymosin, penicillopepsin, and Rhizopus pepsin, which all are two-domain enzymes with more that 300 residues in length and contain the Asp-Thr-Gly sequence in each domain that forms the active site, which effectuate the cleavage reaction.97,98

Since the HIV protease sequence is no more than 99 amino acids and contains only one of the required triad Asp-Thr-Gly it was suggested that the active form of the HIV protease was a homo-dimer of 198 amino acids.98_{This hypothesis was later confirmed}

by X-ray crystallographic determinations.99-101

The function of the dimeric structure of the protease is probably more sophisticated than simply to enabling the virus to be parsimonious in its genetic baggage.102_{A regulatory}

mechanism that controls activation of the enzyme is derived during the dimerization process, since it is reversible. In a concentrated solution the protease is activated (in the budded viral particle), and in a highly diluted solution the protease is inactivated (in the host cell). This regulatory mechanism seems to be important for the virus in order to prevent premature breakdown of the polyproteins and to minimize damage of the host-cellular proteins (Figure 8).69

(16)

Introduction

Figure 8. Dimerisation of the HIV protease.

Some general feature of the HIV-1 protease structure can be described (Figure 9): (i) The two monomers are identical and form a C2-symmetric elliptical-shaped enzyme. (ii)

The N- and C- termini of each monomer are juxtaposed in a four-stranded β-sheet that serves to hold the dimer together (the dimer interface). (iii) Each monomer has a hydrophobic core consisting of two loops, one of which includes the active site aspartic acid. (iv) The dimers come together to create an extended substrate-binding cleft capable of interacting with a minimum of seven consecutive amino acids in the substrate. (v) Each monomer contributes a flexible flap that folds down to make important contacts with the bound substrate.103

Figure 9. Structure of native HIV-1 protease.99

The HIV-1 protease processes the Gag and Gag-Pol polyproteins proteolytically at specific cleavage sites as shown in Figure 10.73,104_{The HIV-1 protease is specific for}

Highly diluted In the host cell Monomers

INACTIVE

Concentrated In the viral particle

Dimer ACTIVE Flap Flap Dimer interface Active-site

(17)

Introduction

Figure 10. Cleavage sites of HIV protease in the Gag and Pol polyproteins.

Several studies, experimental and ab initio calculations, of the protein cleavage mechanism have been performed. A schematic representation of the mechanism is outlined in Figure 11.105-109_{Hydration of the amide carbonyl group, with a water}

molecule accommodated between the two side-chains of the aspartic acid residues 25/125, gives a putative tetrahedral intermediate that is suggested to be an approximate representation of the transition state of the proteolytic reaction.

Figure 11. Schematic representation of the HIV protease cleavage mechanism.

PR RT/RNaseH IN p17 p24 p7 p6 Gag Pol 1 2 3 4 5 6 7 8 Site 1 2 3 4 5 6 7 8 Sequence -Ser-Gln-Asn-Tyr Pro-Ile-Val-Gln--Ala-Arg-Val-Leu Ala-Glu-Ala-Met--Ala-Thr-Ile-Met Met-Gln-Arg-Gly--Pro-Gly-Asn-Phe Leu-Gln-Ser-Arg--Ser-Phe-Asn-Phe Pro-Gln-Ile-Thr--Thr-Leu-Asn-Phe Pro-Ile-Ser-Pro--Ala-Glu-Thr-Tyr Phe-Val-Asp-Gly--Arg-Lys-Ile-Leu Phe-Leu-Asp-Gly-N H H N O R2 R1 O O O H O O O H H N H H N R2 R1 O O O O O O O H H H N H H N R2 R1 O O O O O O O H H N H H2N R2 R1 O O O O O H H OH O

Asp 25 Asp 125 Asp 25 Asp 125

Substrate Tetrahedral Intermediate

(18)

Introduction

Following the unambiguous determination of the native enzyme structure, several structures of various inhibitor complexes of the protease have been reported.110-115

Upon binding of an inhibitor, the protease undergoes significant structural changes. The most dramatic changes are observed in the flap region. The two flaps fold over the inhibitor to form a tunnel-shaped active site, which runs diagonally across the dimer interface. The flaps are held in this closed position by hydrogen bonding from the flap residues Ile 50/150 to a water molecule, which in turn is hydrogen bonded to two carbonyls in the inhibitor (Figure 12).

Figure 12. Structure of HIV-1 protease with bound inhibitor (Stereoview).

Starting from the central aspartates, in the active site tunnel, there are distinct subsites named S1, S2, S3, and S4, with corresponding S1', S2', S3', and S4' subsites according to the convention of Schechter and Burger.116_{The corresponding side-chains of the}

substrate or of the inhibitor are named P1 to Pn outwards from the scissile peptide bound toward the amino terminus and P1' to Pn' towards the carboxyl terminus (Figure 13). All subsites in the HIV-1 protease are bounded by mostly aliphatic side-chains and have hydrophobic character, with the exception of S4/S4', which are exposed to

water.117

Figure 13. Nomenclature of subsites in the enzyme (Sn, Sn') and of the substrate (Pn, Pn') according to

Schechter and Burger.

H2N H N N H H N N H H N O P1' O P2' O P3' OH O P1 O P2 O P3 S3 S1 S2 S2' S1' S3'

(19)

Introduction

1.9 Paradigms for Drug Discovery

A molecule that binds with high affinity to HIV protease is not necessarily suitable as a drug for the treatment of HIV infection. A drug must be endowed with a number of qualities that are not easily accomplished. For example, a high specificity of the drug is crucial, since significant inhibition of other host aspartyl proteases may lead to toxicity. A drug must be able to conquer a large set of viral genotypes, otherwise insurgence of drug resistance might appear. A drug should work in synergy with other chemical entities used for HIV treatment and be compatible with other therapies against opportunistic infections. The concentrations of the drug in the cells and in the circulation must be able to remain at levels far above the inhibitory constant (Ki) in order secure an effective treatment. Oral bioavailability is highly desirable to enable easy use of the drug. To reach the virus hiding in the brain the drug should be able to penetrate the blood-brain barrier. All of these criteria can be met by manipulating the structure of the molecules, but do not always go hand in hand, which makes the drug discovery process a laborious but challenging task. Finally, HIV infection requires a lifetime of treatment with a combination of several drugs, and therefore the drug must be inexpensive in order to be available to all those millions of people in need of therapy.69

(20)

Aims of the Present Study

2. AIMS OF THE PRESENT STUDY

This investigation is part of a research project aimed at the discovery of novel, selective HIV-1 protease inhibitors, with the ultimate objective of developing simple and cost-efficient, generic synthetic routes. The specific objectives of this study have been:

(i) To design a novel HIV-1 protease inhibitor scaffold and to develop a short, diverse, and cost-efficient synthetic route to inhibitors, using readily available carbohydrates as chiral starting materials.

(ii) To utilize fast microwave promoted palladium-catalyzed reactions for the optimization of identified leads.

(iii) To establish structure-activity relationships and to utilize X-ray structural determinations to guide further inhibitor design by an iterative process, with the proviso that target compounds of acceptable biological activity would be essential to enable this.

During the course of these studies, two new objectives arose:

(iv) To develop a fast microwave promoted method for the transformation of organo-bromides to tetrazoles.

(v) To modify the HIV-1 protease inhibitors into tools for surface plasmon resonance biosensor assays thus enhancing assay sensitivity.

(21)

Design of HIV Protease Inhibitors

3. DESIGN OF HIV PROTEASE INHIBITORS

In order for a molecule to inhibit HIV protease and thereby prevent cleavage of

polyproteins, it must possess binding properties and be stable to proteolytic processing. Strategies required to find such molecules could be random, semi-random or by rational design. The most obvious strategy, besides random high throughput screening, is to utilize the peptidic substrate as a template, with the scissile peptide bond substituted with a non-cleavable bond. An example of a non-cleavable peptide bond mimic is the reduced amide introduced into a peptide and used in the earliest inhibitors of HIV protease.118_{Only modest potency was achieved with most of the reduced amide}

compounds. Nevertheless, MVT-101 constituted the first inhibitor to be co-crystallized with the HIV-1 protease and provided valuable three-dimensional structural information of the HIV protease in its inhibited form.110

Figure 14. Example of a reduced amid inhibitor.

The transition-state of the substrate often possesses thousand-folds higher affinity for the enzyme than the substrate.92_{Since the cleavage mechanism of the enzyme was}

known and a tetrahedral intermediate was proposed as being closely related to the transition-state, the design was shifted toward the preparation of transition state analogues. Vast knowledge of transition-state analogue inhibitors was gained through many years of research of the human aspartic protease renin, which is involved in the regulation of high blood pressure.119-121_{Some examples of transition state analogues}

are depicted in Figure 15.

Figure 15. Examples of transition state analogues. Ile N H Gln O Arg NH2 Thr Ac MVT-101 N H R₁ OH O N H R₁ OH O N H _OH _R 2 R1 O N H _OH _R 2 R1 OH O N H R1 O O X _X N H R1 O R2 X X O N H N H OH R1 R2 O Statine Norstatine

Hydroxyethylene Dihydroxyethylene Hydroxyethylamine

Statone Difluorostatone X = H X = F Ketone DifluoroKetone X = H X = F

(22)

The statine skeleton, comprising five atoms, was employed as a dipeptide (six atoms) mimic and was incorporated in peptides to serve as inhibitors of HIV protease.122-124

The mediocre potency that was achieved with this fragment was considered to be attributable to the lack of a P1' side-chain.71_{However, inhibitors with the norstatine}

moiety, which more closely resembles a single amino acid residue carrying an extra carbon, provided increased activity against HIV protease when incorporated into inhibitors.125-127_{Hydroxyethylene isosteres constituted the basis for highly potent}

inhibitors and provided the first inhibitors to have activity toward HIV in cell-based assays.128-130_{The improvement achieved with hydroxyethylene isosteres encouraged}

investigation of dihydroxyethylene isoteres. It was found that the contribution from the second hydroxyl group was negligible, as suggested by similar binding affinities of hydroxyethylene and dihydroxyethylene based transition state analogues.71_The

hydroxyethylamine isosteres are perhaps most successful amongst the previously known transition state analogues used in HIV protease inhibitors. This transition state analogue was utilized in the first clinically approved HIV protease inhibitor, saquinavir (Figure 7).75

Figure 16. N- or C- terminal duplication strategy.

C2-Symmetric inhibitors

The C2-symmetric nature of the HIV protease homodimer provides a unique property of

the enzyme that could be exploited in the design process. X-ray data revealed that both the N- and C-termini of the asymmetric substrates and the inhibitors bind to identical subsites. The design of N- or C-duplicated inhibitors therefore was thought to be

beneficial in terms of novelty, potency, and selectivity (Figure 16). Abbot employed the N-terminal duplication strategy at an early stage and presented the pseudosymmetric diaminoalcohol core unit of A-74704 (Figure 17).131_{A highly symmetric binding of}

this inhibitor was confirmed by X-ray crystallographic studies.113_{The concept of}

N-terminal duplication was further developed by Abbot into the diaminodiol core, exemplified by A-77003, which was studied clinically as an intravenous agent.132

Systematic improvements of this class of compounds resulted in the non-symmetric

H2N H N OH O P1' O P1 H2N NH₂ P1' P1 OH P1' O P1 HO O

(23)

Figure 17. N-Terminal duplicated inhibitors.

Examples of C2-symmetric inhibitors based on C-terminal duplication have also been

reported, although this strategy has been studied less extensively.72_{Bone et al. were the}

first to report a high-affinity inhibitor, L-700,417, exploiting this concept (Figure 18).137_{A more peptide like inhibitor, which also exhibited a relatively high potency,}

was designed and synthesized by Babine et al. independently.138_{Analysis of the crystal}

structure of L-700,417 reveled highly symmetric binding.137

Figure 18. C-Terminal duplicated inhibitors.

3.1 Design of a New C-Duplicated Scaffold

We wanted to explore the concept of C-terminal duplication further. Computer aided molecular modeling was used as a tool in the search for potential scaffolds. A large number of models of potential inhibitors were built, geometrically optimized in silico and compared to reported inhibitor/HIV protease X-ray data for the evaluation of distances and positions of the different side-chains, hydroxyl groups and scaffold backbone. Our hypothesis was to introduce a dihydroxyethylene C-terminal duplicated core structure, that had been described in literature previously as an unpublished result by Norbeck et al.72_{Among our modeled substances evaluated, a scaffold}

accommodating a 1,6-retro amide motif with a six-carbon chain emerged as an attractive core structure (Figure 19).

H N H N OH Val Val Cbz Cbz _H N N H OH OH Val Val N N O O N N A-74704 A-77003 H N H N OH O O HO OH Ile Ile OH O O H N H N N HN NH N L-700,417 _{Babine et al.}

(24)

Figure 19. 1,6-Retro amide scaffold.

In order to synthesize our modeled in silico inhibitors with the desired stereochemistry, a chiral scaffold was needed. Carbohydrates have been utilized previously, by us139-141

and others groups,134,142-146_{as a chiral pool for the synthesis of inhibitors with desired}

stereochemistry. Hexoses provide a commercially available source of six carbon scaffolds with a defined stereochemistry. The modeled stereochemistry and

conformation needed for complementary interaction with HIV protease was suggested to be 2R, 3R, 4R, and 5R, a stereo configuration that could be derived from L-mannitol.

The distance between the two carbonyl groups in the 1,6-retro amide, designed to interact with the structural water that is hydrogen bonded to the flap residues Ile 50/150, is notably only six atoms compared to seven atoms in the natural substrates. The

suggested synthesis would deliver elongated the P1/P1' arms as a result of the insertion of oxygen into the benzyl side-chains designed to mimic the Phe amino acid.

N H H N OH OH O O O O R R

(25)

Synthesis of the 1,6-Retro Amide

4. SYNTHESIS OF THE 1,6-RETRO AMIDE

L-Mannitol, which is expensive compared to the naturally occurring D-mannitol, could be obtained by reduction of the less expensive oxidized form, L-mannonic-γ-lactone, by

reduction with lithium borohydride.146_{To enable introduction of the P1/P1'}

substituents, protection of the hydroxyl groups at the 1,3,4, and 6 positions was necessary (Scheme 1).

Scheme 1.

L-mannitol was first reacted with 2,2-dimethoxypropane to achieve the triacetonide 1,

which was subsequently treated with 70% acetic acid to deprotect the terminal

isopropylidene groups to furnish the 3,4-monoacetonide 2 in good yield.146_{A selective}

silylation of the primary hydroxyl groups was performed using t-butylchlorodimethyl silane in pyridine, in the presence of a catalytic amount of dimethylaminopyridine.147

No product derived from silylation of the secondary alcohol was detected in the reaction mixture. The 2- and 5- hydroxyl groups were now available for the introduction of the P1/P1' side-chains. Benzylation of the 2,5-diol in tetrahydrofuran using benzyl bromide, sodium hydride and a catalytic amount of tetrabutylammonium iodide gave the fully protected compound 4 in 76% yield. Desilylation of the primary alcohols was

performed with tetrabutylammonium fluoride in tetrahydrofuran to give the compound

5 in 98% yield. O O O O O O HO OH O O OH OH O O O O OH OH Si Si O O O O O O Si Si HO OH O O O O O _O OH HO HO HO Br HO OH OH OH OH OH TBDMSCl O O LiBH4, MeOH CSA, Acetone 78% 70% HOAc 2 DMAP Pyridine 96% NaH, THF 76% N(Bu)4F, THF 1 64% 3 4 5 98%

(26)

In order to complete the synthesis of the inhibitor scaffold the primary hydroxyl groups had to be oxidized. There are not many examples of diol oxidations, and most methods rely on the use of potassium permanganate or nitric acid.148-150_TEMPO

(2,2,6,6-tetramethyl-1-piperidyloxy, free radical) with sodium hypochlorite has been used for the direct oxidation of alcohols to corresponding acids under two-phase conditions.151

Oxidation of 1,4- and 1,5-diols with this system was reported to give the lactones in good yields while oxidation of the 1,6-diol was discouraging and delivered a mixture of polymeric compounds.152_{The same reagents had also been used previously for the}

synthesis of hydroxyaldehydes from 1,3-, 1,4-, and 1,6-diols.153

We first tried to oxidize the diol 5 with pyridinium dichromate-acetic anhydride, but the desired compound was not obtained.154_{More successfully, we applied 5 mol% of}

TEMPO and sodium hypochlorite at 0 °C, which resulted in the dicarboxylic acid 6 in 65-72% yield within 1.5 hours (Scheme 2). Attempts to optimize the reaction by

varying the amount of TEMPO or the temperature in the reaction only resulted in lower yields. However, an experiment performed by us more recently, with twice the amount of solvent resulted in 92% yield of the dicarboxylic acid 6. Apparently, the

conformational stabilization by the 3,4-O-isopropylidene group might be crucial for the formation of the 1,6-diacid, since when we performed a control experiment with non-cyclic protecting groups on the central hydroxyl groups, the desired product was not obtained.

Scheme 2.

To introduce the P2/P2' substituents, the diacid was preactivated, using the reagent system EDC, HOBT and triethylamine in dichloromethane and tetrahydrofuran.155_To

react the diacid with N,N'-disuccinimidyl carbonate to give disuccinimidyl ester 7 provided an alternative route.156_{The active ester 7 that was isolated in 88% yield could}

be stored prior to use. The desired amine was then added to form the compounds 8-12 in moderate to good yields (Scheme 3).

HO OH O O O O 5 HO OH O O O O O O 6 N O NaOCl, KBr N(Bu)4Br NaHCO2(sat.) DCM, 0 °C 67%

(27)

mixtures. More successfully, hydrochloric acid in dry methanol was used, which gave the desired products in about 70% yields. An alternative method was to use 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) in a 9:1 mixture of

acetonitrile/water.157

Scheme 3.

4.1 A New Three-Step Synthesis

Encouraged by the biological results (vide infra) and since we anticipated that a very large number of compounds would have to be synthesized during the lead optimization process, we felt it essential to improve this nine-step synthesis. A literature procedure describing the direct oxidation of D-mannonic-γ-lactone to the bicyclic D -mannaro-1,4:3,6-dilactone using nitric acid, caught our interest.158_{This reaction would furnish a}

compound with; protected/activated 1.6-diacids, protected 3,4-hydroxyl groups, and free 2,5-diols, which serve as handles for the direct introduction of P1/P1' substituents, and this in only one step from the starting material. Thus, we envisioned that this new strategy would significantly reduce the number of synthetic steps needed and provide a versatile route for the introduction of the P1/P1' and P2/P2' substituents.

Hence, oxidizing the L-mannonic-γ-lactone to the L-mannaro-1,4:3,6-dilactone was

easily performed by heating to reflux in nitric acid for 16 hours, and resulted in a 60%

HO OH O O O O O O N H H N O O O O O O R R O O O O O O O O N N O O O O N H H N OH OH O O O O R R RNH2 MeOH RNH2 CH2ClCH2Cl 6 HOBT,EDC N(Et)3 THF-DCM 61-73% DSC Pyridine MeCN 88% 8-12 7 4% HCl 15-18, 28-33

(28)

yield. Several alkaline reagents were used in combination with benzyl bromide for the introduction of the benzyl substituents, all with discouraging results, since the dilactone was subsequently proven to be alkali labile. Meanwhile, a serendipitous experiment with a related molecule resulted in the desired dilactone under acidic conditions. We therefore tried the alkylation using alternative acidic conditions, with benzyl-2,2,2-trichloroacetimidate as the benzyl source.159_{Reacting the}_L_{-mannaro-1,4:3,6-dilactone}

with benzyl-2,2,2-trichloroacetimidate in dry dioxane with trifluoromethylsulfonic acid as catalyst gave the 2,5-di-O-benzyl-L-mannaro-1,4:3,6-dilactone in 72% yield (Scheme

4). Scheme 4.

Thus far, the planned synthetic route had worked satisfactory. The next step would be to open the lactones with an appropriate nucleophile to afford the target compounds. Aware of the instability of the dilactone 14 to alkalis, the conditions had to be carefully tuned in order to minimize the elimination products formed due to the basicity of the applied amine. By adding six equivalents of the selected amines in dichloromethane (acetonitrile or chloroform) to the dilactone 14 and refluxing overnight, the target compounds formed in moderate to good yields. In more polar solvents the elimination reaction was predominant. To confirm the structure of the elimination product the dilactone 14 was treated with sodium hydroxide. This gave, as presumed, the α,β -unsaturated lactone 39 in high yield.

Scheme 5. N H H N OH OH O O O O R R O _O OH HO HO HO HNO3 O O O O HO OH O CCl3 NH O O O O O O RNH2 DCM reflux 60% CFDioxane3SO3H 72% 13 14 19, 20-26, 34-28 O O

(29)

4.2 Structure-Activity Relationship of P2/P2' Modifications

Purified HIV-1 protease was then used in a standardized spectrophotometric assay to determine the IC50 values (concentration of the inhibitor resulting in 50% inhibition) of

the tested compounds.160_{A more sensitive fluorometric assay was also used to}

determine the Ki values (inhibitory constant).161

Table 1. Structures, yields, and enzyme inhibition of the protease inhibitors.

R R OH OH O O O O O N H O O N H O O N H O Ph O N H O H N N H O H N N H O H N N H O H N N H O H N N H O F H N N H O HO H N N H O HO H N N H O S 15 16 17 18 19 20 21 22 23 24 25 26

Cmpd. no. Yield (%) Ki (nM)a Cmpd. no. Yield (%) Ki (nM)a

69 70 71 73 70 76 46 31 20 60 61 73 5000b ni ni ni 0.4 140 0.9 2.3 81 5000 500-1500 660 R-group R-group a ni = no inhibition at 10 µM. b IC50 value.

(30)

The first set of compounds to be evaluated, 15-18, all contained lipophilic amino acid ester side-chains in the P2/P2' position (Table 1). None of the compounds exhibited any activity for the HIV-1 protease, with the exception of 15, containing valine residues, that showed modest activity (IC50 = 5 µM). The above observation was not so surprising

considering that many of the inhibitors reported contain a valine moiety in P2/P2'.71

More interestingly, exchanging the ester functionality to an amide resulted in gained activity. A 300-fold activity increase was observed going from compound 15 to 19. Table 2. Structures, yields, and enzyme inhibition of the protease inhibitors.

R R OH OH O O O O HO N H N H N H N H N H N H N H N H HO N H N H 27 28 29 30 31 32 33 34 35 36 37

Cmpd. no. Yield (%) Ki (nM)a Cmpd. no. Yield (%) Ki (nM)a

74 29b 30b 31b 25b 32b 38b 59 65 49 35 22 ni ni ni ni ni ni 2000 7100 ni 0.2 ni R-group R-group N N N HO HO Cl F N S OH OH

(31)

A series of methyl amide compounds with different properties was therefore

synthesized and evaluated (20-26). The closely related isoleucine containing compound

21 showed a high inhibitory activity (Ki = 0.9 nM) and the slightly larger compound 22 also exhibited also good potency (Ki = 2,3 nM). Introduction of hydrogen bond

accepting or donating groups resulted in decreased activity (23-26). In order to reduce the peptidic character of our inhibitors a different series of compounds was synthesized (27-36, Table 2). Mediocre potency was achieved with compounds 34 and 35 but the majority of the compounds were found to be inactive. 1(S)-amino-2(R)-indanol is a moiety that has been used in many reported inhibitors including the clinically approved indinavir. Introduction of this substituent to our scaffold gave a highly potent inhibitor,

37 (Ki = 0.2 nM), while applying the 1(R)-amino-2(S)-indanol afforded an inactive compound.

For the determination of the in vitro anti-HIV activity of the high affinity inhibitors 19,

21, and 37, an assay in MT-4 cells was used. The cytopathogenic effect was quantified

using the vital dye XTT.161,162_{From the cytoprotection of the tested compounds the}

50% inhibitory concentration (ED50) was calculated. All three compounds exhibited

anti-HIV activity, although compound 37 (ED50 = 0.09 µM) showed 17 times higher

potency as compared to 19 (ED50 = 1.5 µM) and 21 (ED50 = 1.6 µM). Noteworthy,

compound 37 exhibited similar potency to the clinically approved ritonavir (ED50 = 0.06

µM) and indinavir (ED50 = 0.06 µM) which were evaluated in the same assay. At this

point we found it essential to obtain 3D structure data to guide the further design of our inhibitor scaffold.

4.3 X-Ray Crystallographic Data

Compounds 21 and 37 complexed with HIV-1 protease were crystallized and the X-ray structures were determined. Analyses of the structural data reveled that both of the inhibitors bound to the protease in a similar fashion (Figure 20). The central diol-moiety, designed to mimic the transition-state, was bound with one of the hydroxyls pointing toward the active site Asp 25/125 residues and formed hydrogen bonds to both the carboxyl oxygens. The second hydroxyl group points away from the active site, but was close enough to form a hydrogen bond to one of the active site Asp residues. The P1/P1' substituents, elongated with one oxygen atom, were nicely positioned in the S1/S1' pockets of the enzyme. In a distorted tetrahedral arrangement a structural water was hydrogen bonded to the amide nitrogens of the residues 50/150, positioned in the flap regions of the enzyme, and to the carbonyl oxygens of the mannaric acid scaffold, even though the scaffold is one atom shorter than the natural substrate. The X-ray data of compound 21 revealed that the isobutyl side-chains in the P2/P2' position did not completely fill the S2/S2' subsites, although compounds comprising larger side-chains in this position gave lower activity in the HIV-1 protease assay. Apparently, a larger

(32)

group must be structurally constrained to fit into the S2 pocket, since the larger aromatic part of the aminoindanole in compound 37 was accommodated nicely into the S2 pocket as shown by the X-ray data.

Figure 20. X-ray data for compound 21 and 37 (Steroview).

The next step in the iterative process of refining this class of compounds was to study the X-ray structures to decide which positions to modify next. In the extension of the para-position of the P1/P1' benzyl groups a tunnel-shaped cavity reaching towards the surface of the enzyme could be observed. There are reports describing the concept of P1 or P1' elongation and also some examples where the P1' substituent has been connected to the P3' side-chain.10,114,115,163-174_{Elongation of the para-position in the P1/P1'}

substituents often gave inhibitors where no improvement of affinity for the HIV-1 protease was observed. However, elongation seems to provide a tool for manipulation of the anti-viral activity in cells. Our strategy was to modify the P1/P1' substituents with the aim of improving the antiviral activity.

(33)

Synthesis of P1/P1' Substituted Inhibitors

5. SYNTHESIS OF P1/P1' SUBSTITUTED INHIBITORS

For the synthesis of the P1/P1' elongated inhibitors we wanted to utilize the short synthetic route via the dilactone. For the introduction of the para-benzyl substituents two synthetic strategies were considered: 1) Convert para-substituted benzyl alcohols into trichloroacetimidates, and react these with the L-mannaro-1,4:3,6-dilactone. 2)

Utilize synthetic-handles in the para-position of the P1/P1' benzyl groups, for

subsequent introduction of substituents. The second strategy was more appealing, since the substituents could be introduced in the last step of the synthesis. The synthetic-handle of our choice was a bromide, which would be applicable to our synthetic route and which could be substituted easily with a plethora of groups through palladium-catalyzed couplings without interference of the other inhibitor functionalities. Palladium-catalyzed reactions

Palladium-catalyzed reactions provide facile, catalytic, chemoselective, and non-toxic ways of forming carbon-carbon bonds. The Heck,175_Suzuki,176_{and Stille}177_reactions

are of particular importance to medicinal chemistry since a very large variety of functional groups can be introduced smoothly to suitable precursor fragments. Microwave enhanced reactions

For a long time microwave ovens have been standard equipment in most of our kitchens to heat and cook food quickly. In preparative organic chemistry the development has proceeded more slowly and the first reports of microwave assisted reactions were in 1986.178,179_{Initially, major problems were encountered with this technique due to lack}

of control and reaction reproducibility. The development of controllable single-mode microwave cavities brought the technique into another level of usefulness. Nowadays, the technique is gaining popularity among organic chemist and several reports have been published the last few years,180-182_{including microwave assisted}

palladium-catalyzed Heck, Suzuki, and Stille reactions.183

Synthesis

To form the (4-bromobenzyl)-2,2,2-trichloroacetimidate (40), 4-bromobenzylalcohol was deprotonated and reacted with trichloroacetonitrile.159_{Compound 40 was thereafter}

reacted with the L-mannaro-1,4:3,6-dilactone, synthesized as previously described (vide

supra), to give the 2,5-bis-O-(4-bromobenzyl)-L-mannaro-1,4:3,6-dilactone, 41, in 88%

yield (Scheme 6). The bulk of the product precipitated in the reaction vessel and could be isolated from the reaction mixture by filtration.

(34)

Scheme 6.

The dilactone 41 was subsequently reacted with valine methylamide in

1,2-dichloroethane to give the dibrominated inhibitor 42, which also precipitated from the reaction mixture and was collected by filtration in 64% yield. A common precursor was thereby obtained, which could be utilized for palladium-catalyzed coupling reactions. Substitutes were introduced in the para-position of the P1/P1' benzyl groups.

Nevertheless, the precursor compound 42 was proven to be a good inhibitor, Ki = 0.3 nM, ED50 = 0.8 µM.

The dibromo-compound 42 was reacted with aryl- and heteroarylboronic acids with sodium carbonate as base in the presence of palladium tetrakis(triphenylphosphine)184

as catalyst (Scheme 7). The reaction vessels were sealed with teflon-septa as a pressure control device and microwave irradiated at 45 W for 4 min to furnish the Suzuki coupled products 43-46 in high yields (Table 3). To introduce an ethyl spaced phenyl group we utilized a 9-BBN coupling.185_{An in situ hydroboration reaction of styrene}

with 9-BBN gave the phenylethyl-9-BBN, which was reacted with the bromo precursor

42, palladium tetrakis(triphenylphosphine), and sodium carbonate under microwave

irradiation (60 W) for 2 min. No Heck coupling derived product (stilbene derivative) was detected in the reaction mixture and the compound 47 was isolated in a moderate yield (38 %). Scheme 7. N H H N OH OH O O O O Br Br H N N H O O NH2 H N O O O O O HO OH O CCl3 NH Br CF3SO3H O O O O O O Br Br Dioxane 88% 13 41 42 + 40 CH2ClCH2Cl 45 °C, 64% N H H N OH O O Br H N N H O N H H N OH O O R H N N H O Reactant Reaction

(35)

The introduction of the 2-, 3-pyridyl or 2-thiazolyl substituents was performed with the trimethyl- or tributylheteroaryl tin under Stille reaction conditions, with

dimethylformamide as solvent and palladium tetrakis(triphenylphosphine) as catalyst. To improve the reactions of the pyridyl couplings, cupric oxide was used as additive and in the case of thiazolyl coupling, silver(I) oxide was employed.186_{The reaction}

mixtures were conducted at 60 W for 2 min to give the target compounds 48-50 in moderate yields (Table 3). To simplify the purification procedure, the compounds were dissolved in acetonitrile and washed with isohexane to minimize the contamination from tin residues.

Both of the Heck reactions were conducted under microwave irradiation for 2 min (60 W). The products derived from reaction with methyl acrylate and

1,2-cyclohexandione187_{both exhibit the same R}_f_{-value as the starting material 42. With}

these experiments we felt prompted to force the reactions to full completion. This was easy accomplished with methyl acrylate, which gave compound 51 in 76 % yield. With the 1,2-cyclohexandione every attempt to achieve full conversion of the starting

material e.g. employing increased power, prolonged reaction time, or adding additional portions of the catalyst, only resulted in increased formation of degradation products. We therefore performed the reaction under traditional thermal heating conditions at 100

°C for 48 h and were able to isolate pure compound 52 in moderate yield (Table 3).

5.1 Structure-Activity Relationships of P1/P1' Modifications

All of the tested compounds, 42-52, proved to have Ki values in the nanomolar range (determined as described earlier, vide supra), in agreement with our molecular modeling and with previously reported data. Remarkably, only minor deviations in Ki values were observed with this series of inhibitors (Table 3). The variance between the compound with the highest (51) and the lowest (47) affinity for HIV-1 protease are only a factor of 40. This should be compared to the sensitivity in P2/P2', where an exchange of a methyl group in compound 19 to a hydroxyl in compound 25 resulted in a 2500 times lower affinity for the protease. The HIV-1 protease seems be able to accommodate a variety of different substituents in the extension of the P1/P1', with a slight preference for small substituents containing hydrogen bond accepting groups.

As the next step antiviral effects of substituents in the para-position of the P1/P1' benzyl groups were considered (Table 3). We found that substitution of the para-benzyl

hydrogen in the parent compound 19 with phenyl groups led to a 40-fold increase of anti-viral activity (19 ED50 = 1.5 µM, 43 ED50 = 0.04 µM). Thus compound 43

exhibited ED50 values comparable to the clinically approved inhibitors tested in the

(36)

Table 3. Reagents, time/effect, structures, yields, enzyme inhibitions, and anti-viral activities in cell

cultures of the protease inhibitors.

N B S S NO2 N N S O O O HO Reactant R-group (HO)2B (HO)2B NO2 (HO)2B S (HO)2B S N (Me)3Sn N (Me)3Sn (Bu)3Sn N S O O O HO Cmpd. no. Yield (%) Ki (nM) 19 42 43 44 45 46 47 48 49 50 51 52 Time (min)/ Effect (W) ED50a (µM) ED50hsc (µM) ED50mutantsb (µM) 4/45 4/45 4/45 4/45 2/60 2/60 2/60 2/60 2/60 48 h/100°Cf 0.4 0.3 0.7 93 96 86 85 38 54 50 53 76 52 1.2 1.2 1.4 3.8 0.6 0.3 0.6 0.09 0.3 1.5 0.8 0.04 0.04 0.04 0.05 2.1 0.8 2.5 0.3 0.8 3.7 ndd, 20e 1.2d, 1.0e 0.4d, 0.7e 0.3d, 0.2e 0.3d, 0.2e 1.1d, 0.04e nd nd >3d, >3e >3d, >3e nd nd 20 2.1 0.8 0.8 1.0 1.1 nd 3.6 20 0.4 5.7 20

(37)

Similar effects were also encountered with the two electron-rich thienyl containing compounds 44, 45 and the nitro containing compound 46, despite the fact that they exhibited only half the affinity for the HIV-1 protease as compared to 43. This observed gain in anti-HIV activity might be attributable to the increased lipophility of the

compounds, which could enhance the cell-membrane penetration. Contrary to this, the more lipophilic compound 47 (ED50 = 2.1 µM) showed a 50-fold decrease in anti-HIV

activity compared to 43. Introduction of the electron-deficient, weakly basic 2- and 3-pyridyl heterocycles was predicted to preserve the positive effects obtained with the phenyl derivative 43 whilst in addition improving the solubility of the compounds. Unfortunately, both of the compounds exhibited lower anti-HIV activity (48 ED50 = 2.5

µM; 49 ED50 = 0.8 µM) as compared to 43. Similar anti-HIV activities were achieved

with the thiazolyl compound 50 (ED50 = 0.3 µM) and with the high affinity methyl

acrylate compound 51 (ED50 = 0.8 µM). The poorest anti-HIV activity was observed

with the acidic 1,2-cyclohexandione containing compound 52 (ED50 = 3.7 µM).

Mutant HIV-1 protease

Viral resistance to HIV protease inhibitors is a major problem due to the error-prone nature of reverse transcriptase and the lack of proof reading functions in the virus, which causes mutations in the viral proteins and enzymes.188-190_{At first this was}

thought to constitute a minor issue in the case of the HIV protease since the size of the enzyme is small (198 residues). However, mutations were indeed observed in many positions of the enzymes. Some of the most common mutations of the HIV protease are highlighted in Figure 21.191_{To evaluate our compounds against a mutant virus,}

MT4-cell cultures with mutant viruses were used. The mutants were selected by growth with increased concentrations of ritonavir. The first assay contained HIV-1 protease with the mutations V32I, M46I, and V82A and in the second assay the mutations were M46I, V82F, and V84I. Compound 46 was the single exception in the series of the highly active compounds 43-46 that also retained antiviral activity against one of the viral mutants (Table 3). The other three compounds exhibited a 10-fold decrease in antiviral activity in both assays. This loss of activity can in fact be seen as a 20-fold gain of activity compared to the parent compound 19 tested against one of the mutants. The compounds 42, 49, and 50 were also evaluated against both of the two mutant viruses. The bromo compound 42 was able to maintain antiviral activity against both of the mutant viruses, but for the pyridyl 49 and the thiazolyl 50 compounds the antiviral activity against the mutants was low. In summary, although 46 exhibited a favorable inhibitory effect in one of the assay systems, and the nitro-group is bound to the same area affected by the mutation (V82F), it is difficult to draw any conclusions on a molecular level since the 3D structure of 46 complexed with the mutant protease is missing.

(38)

Figure 21. Common mutations in the HIV-1 protease.

Non-specific protein binding

The blood contains a large amount of proteins. Consequently, non-specific protein binding of the HIV protease inhibitors can occur and as a result the inhibitor might not reach the therapeutic concentration needed. To evaluate the non-specific binding of our compounds, a cell-based assay containing 50% of human serum was used. A 20-fold drop in antiviral activity was encountered for the four most active inhibitors (43-46, Table 3). The other compounds (48-52) were also affected but to a lesser extent. The thiazole derivative 50 was the only compound that was able to maintain antiviral activity in the presence of human serum.

Oral absorption

Until this point the evaluations of the inhibitors were performed entirely in vitro. We therefore were interested in investigating whether or not the carbohydrate-based inhibitors exhibited bioavailability after oral administration. The inhibitors 43-45, and

49 were orally administrated to rats at a concentration of 30-40 mg/kg. Unfortunately,

all of the compounds failed to attain measurable blood levels in the rats (<0.02 µg/mL). It is not clear whether poor absorption or fast metabolism accounts for the low

bioavailability.

5.2 X-Ray Crystallographic Data

The 3D-structures of the 3-thienyl and 3-pyridyl compounds (44 and 49) co-crystallized with HIV-1 protease were determined with X-ray crystallography. Analyses of these two new structures in comparison to the old X-ray data from compound 21 revealed an almost identical mode of binding of the backbone structure of the inhibitors (Figure 22), in agreement with our modeled structures. The P1/P1' aromatic side-chains of

(39)

interactions between the protease and the elongated P1/P1' arms, the entrance was slightly contracted around the inhibitor. In the complex of compound 21, four water molecules were bound and linked together in a hydrogen bond network also involving the protease residue Arg8/108 at the common entrance of S1/S1' and S3/S3'. Upon binding of the larger compounds 44 and 49 the thienyl and pyridyl groups displaced three of these water molecules, but in the case of the thienyl compound two new water molecules bound to the entrance in slightly different positions. The similarity in Ki values of the three compounds suggest that net enthalpy and entropy, from binding the larger compounds 44 and 49 with the displacement of the hydrogen bound water as a consequence, balance the net effect of the affinity to the HIV-1 protease.192,193

Figure 22.X-ray data for compound 21, 44 and 49.

So far, an increased antiviral activity was obtained with the P1/P1' elongated C-terminal duplicated scaffold but still the problem with oral bioavailability had to be solved. Guided by the acceptance of acidic structures in the elongation of P1/P1', as seen in compound 52 and from published data,10,163,165_{we wanted to explore the possibility of}

introducing a different carboxylic acid bioisostere to increase the water solubility and hopefully also the oral availability. The strategy was to use the dibromo compound 42 as starting material for the introduction of the commonly used carboxyl bioisostere, tetrazole and to explore the possibility of promoting the reactions by microwave irradiation.

(40)

Microwave Promoted Preparation of Organo-Nitriles and the Corresponding Tetrazoles

6. MICROWAVE PROMOTED PREPARATION OF

ORGANO-NITRILES AND THE CORRESPONDING TETRAZOLES

In 1885 J. A. Bladin, a Swedish chemist at the University of Uppsala, investigated the reactions of dicyanophenylhydrazine, the condensation product of cyanogen and phenylhydrazine. He observed that the action of nitrous acid on

dicyanophenylhydrazine led to the formation of a compound, C8H5N5, to which he

ascribed the structure depicted in Figure 23.194_{Bladin proposed the name tetrazole for}

the new ring structure.195_{The tetrazole moiety is now found in a large variety of}

compounds spanning from explosives196_{to drugs}197_{(e.g. Losartan}198_{). In medicinal}

chemistry the tetrazole moiety is a commonly used bioisostere for the carboxyl group. The metabolically stable tetrazole mimics the carboxyl group in acidity (pKa ∼ 5) and size, and when used as a bioisostere a retained pharmacological effect and a more favorable pharmacokinetic profile is often observe.197_{There are several synthetic}

approaches to tetrazoles, although in general the most commonly used method is to react a nitrile with an azide source (e.g. hydrogen azide or trialkyltin azide).199

Figure 23. Structure proposed by Bladin.

Nitriles, which are used as precursors to tetrazoles, also constitute valuable synthetic intermediates for further transformations into a variety of functionalities, e.g., thiazoles, oxazolidones, triazoles, and amines.200-205_{Many reports have been published on the}

direct transition metal-catalyzed conversion of aryl halides to aryl nitriles,206-215_{and in}

addition several other methods are available for the preparation of nitriles.216-219

To obtain the desired tetrazole decorated inhibitor we needed a mild procedure that allowed nitrile introduction by displacement of the bromo atoms of 42. Tschaen et al. has reported an improvement of the palladium-catalyzed cyanation of aryl bromides.209

The success of the previously described palladium-catalyzed reaction on the inhibitor scaffold guided the selection of this procedure. The reaction times of the cyanation were reported to be 5-7 h and the subsequent transformation to tetrazoles also needed several hours for completion. We thought it likely that reduction of the reaction times might be

N N N N NC

Design and synthesis of HIV-1 protease inhibitors

Comprehensive Summaries of Uppsala Dissertations

from the Faculty of Pharmacy 245

Design and Synthesis of HIV-1

Protease Inhibitors

MATHIAS ALTERMAN

ABSTRACT

PAPERS DISCUSSED

CONTENTS

ABBREVIATIONS

1.

INTRODUCTION

2.

AIMS OF THE PRESENT STUDY

3.

DESIGN OF HIV PROTEASE INHIBITORS

4.

SYNTHESIS OF THE 1,6-RETRO AMIDE

5.

SYNTHESIS OF P1/P1' SUBSTITUTED INHIBITORS

6.

MICROWAVE PROMOTED PREPARATION OF

ORGANO-NITRILES AND THE CORRESPONDING TETRAZOLES