Luận án tiến sĩ: A chemical genomic approach towards determining the molecular target of anti-proliferative compounds

The initial transcription profiles showed an increase in the transcript ratios of genes involved in amino acid biosynthesis upon treatment with borrelidin.. The application of chemical g

GRADUATE SCHOOL OF ARTS AND SCIENCES

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy

2007

Copyright 2006 by Eastwood, Erin L

All rights reserved

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy submitted Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted Also, if unauthorized copyright material had to be removed, a note will indicate the deletion

®

UMI

UMI Microform 3246602 Copyright 2007 by ProQuest Information and Learning Company

All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code

ProQuest Information and Learning Company

300 North Zeeb Road P.O Box 1346 Ann Arbor, MI 48106-1346

2006

I would like to thank the Boston University Department of Chemistry and the Center for Chemical Methodology and Library Development for this opportunity and for providing financial support Dr Mark W Grinstaff, Dr James S Panek, Dr Sean J Elliott, and Dr Pinghua Liu, thank you for serving on my committee and for your guidance I would like to thank my research advisor, Dr Scott E Schaus, for giving me the chance and the confidence to grow as a scientist Thank you for your support over the last five years

I would also like to thank the past and present Schaus group members including Josh Giguere, John Westbrook, Nolan McDougal, Melissa Dominguez, Sha Lou, Stacy Rodgen, Whitney Trevellini, Christiane Bode, Amal Ting, Jen Goss, Josh Bishop, Phil Moquist, Allison Wensley, Elise Birkett, Laura Kliman, Andrew Wojtovich, and Valerie

Curtis for providing laughs as well as support in and outside the lab You made me smile

on my worst days I could not have asked for a better group of people to work with I wish you all the best

I would like to thank our collaborators; Sarah Chobot and Mike Hamil! from the Elhott lab in the BU Chemistry Department for their help with the thioredoxin assays; Melissa Landon and Dr Sandor Vajda from the BU Bioinformatics Department for the GOLD studies; the Gardner and Collins labs at the Center for BioDynamics and

iv

using the MNI; and Dr John Tullai from the Cooper lab in the Department of Biology at

BU for his help with the western blot analysis

I would like to thank the office staff in the chemistry department for all of their hard work A special thank you goes to Katinka Csigi for her help in my job search as well as her suggestions on my presentations and résumé You are a wonderful lady I am very thankful to the CMLD faculty and staff, especially Paul Ferrari and Aruna Jain Aaron Beeler, Dayle Acquilano, and Chris Singleton, thank you for the help with my library synthesis and data analysis

To my friends and family, I am so lucky to have such a strong support system Stacy, Christiane, Melissa, Jen, and Allison, I am grateful to have such a great set of friends that are also my colleagues SuzAnn, we have been through a lot together I do not know how I would have made it through those first years without you Thank you for being the sister that I never had

My family has always been an important part of my life I would not be who I am without their love and support Mom and Dad, words can not explain how much you both mean to me I love you both very much Thank you for letting me find my own way and making me believe in myself The support of my family extends beyond my parents I am lucky to have a wonderful extended family, including my aunts, uncles, cousins, and

much You have showed me what really matters in life

Matt, thank you for being so supportive, and understanding about my work schedule It really made me happy to share my free time with you I have found the chemical reaction that I was looking for, and I look forward to what the future may hold

vi

MOLECULAR TARGET OF ANTI-PROLIFERATIVE COMPOUNDS

ERIN L EASTWOOD Boston University Graduate School of Arts and Sciences, 2007

Major Professor: Scott E Schaus, Assistant Professor of Chemistry and Pharmacology

ABSTRACT Drug target identification is a time consuming stage of the drug discovery process Chemical genomics offers a solution to this hurdle In chemical genomics, a target specific chemical ligand is applied on a genomic scale This technique was used to identify the molecular target of anti-proliferative agents using changes in mRNA transcript levels upon treatment Whole-genome transcription profiling experiments employed the eukaryotic model organism Saccharomyces cerevisiae for small-molecule perturbation experiments in addition to traditional genetics

Chemical genomics was used to examine the molecular target of borrelidin, a macrolide with conflicting published biological activities The initial transcription profiles showed an increase in the transcript ratios of genes involved in amino acid biosynthesis upon treatment with borrelidin In yeast, the GCN4 pathway regulates general amino acid control The accumulation of uncharged tRNA activates Gen2p which

vil

Gen4p, which results in the transcription of over 30 genes involved in amino acid biosynthesis Experiments using GCN2 and GCN4 gene deletions determined that borrelidin targets the amino acid biosynthetic pathway through GCN4p The profiling data indicates that an alternative mechanism exists for the translational regulation of Gen4p other than through Gen2p, which was confirmed using immunoblot analysis with elF2 « and phosphorylated elF2 o antibodies

In the second application of chemical genomics, a diverse collection of synthetic compounds was evaluated in a cell-based toxicity assay The screen revealed a subset of cyclic sulfones that inhibited growth of A549, human small lung carcinoma, cells Within this subset, 4-(1-phenyl-1H-tetrazole-5-sulfonyl)-butyronitrile (PTSB) was the most active compound PTSB was shown to inhibit growth of both wild-type S cerevisiae and A549 Whole-genome transcription profiling experiments in S cerevisiae indicated that PTSB is involved in the cellular response to oxidative stress Analysis of the profiling data using systems biology predicted the thioredoxin pathway as the target Biochemical assays with thioredoxin (Trx) and thioredoxin reductase (TrxR) validated that PTSB inhibits TrxR The structure of PTSB suggests a novel mechanism of inhibition

This research illustrates the significance of applying chemical genomics to the

target validation stage of drug discovery

vin

The application of chemical genomics to examine the molecular

target of an anti-proliferative compound, borrelidin

A chemical genomics approach towards the target identification

of a novel anti-proliferative compound, PTSB

XI

XV xvi

Fold-change of several repressed genes upon treatment with borrelidin

Fold changes of the transcript ratios for genes involved in amino acid biosynthesis in the drug sensitive (at 30, 60, and

90 minutes), gcn4A, and gcn2A strains upon treatment with borrelidin

ICso values of lead compounds in growth inhibition assays

in A549 and HeLa-S3 cell lines The top five gene ontology predictions for PTSB using the MNI algorithm

Glso values for gene deletion and heterozygous strains of Trx and TrxR in yeast

Drug target identification methods

An overview of chemical genetics

A chemical genomic approach towards drug target identification using whole-genome transcription profiling experiments

The chemical structures of borrelidin, 3-amino-1,2,4- triazole (3AT), and several anti-cancer agents

Color display plot of the expression ratios of genes involved

in amino acid biosynthesis in the wild-type, gcn4A, and gcn2A strains as a result of treatment with 100 uM borrelidin Color display plot of the expression ratios of genes involved

in amino acid biosynthesis in the wild-type, gcen4A, and gen2A strains as a result of treatment with 100 uM borrelidin and

100 mM 3-aminotriazole (3AT), and wt treated with borrelidin and threonine

Color display plot of the expression ratios of genes involved

in amino acid biosynthesis in the wild-type, gcen4A, gcen2A, gcn1A, and gcn20A strains as a result of treatment with 100

Borrelidin growth inhibition of the gcn4A strain Borrelidin growth inhibition of gen2A strain Borrelidin growth inhibition of the GCD1::gcd1 heterozygous strain

Borrelidin growth inhibition of hom3A strain Growth of CCY333, DS’strain, in the presence of 400 nM borrelidin at various concentrations of threonine

Representative members from a diverse library Representative members of the cyclic sulfone library Color display plot of expression ratios of genes involved in oxidative stress, ribosome biogenesis, and rRNA processing upon treatment with PTSB

(A) DTNB assay for Trx/TrxR activity (B) DTNB assay for TrxR activity

DTNB assay for thioredoxin/thioredoxin reductase activity in the presence of 0, 5, and 50 uM PTSB

DTNB assay for thioredoxin reductase activity in the presence

of 0, 12.5, and 25 uM PTSB Apparent K,, versus [I] for PTSB High throughput screen for thioredoxin reductase activity

of 12.54M PTSB varying the concentration of NADPH from

50 uM and 200 uM DTNB assay for thioredoxin reductase activity in the presence

of 0, 12.5, 25 and 50 uM PTSB Apparent K,, versus [I] for PTSB Percent growth inhibition of A549 by PTSB Percent growth inhibition of HeLa-S3 cells by PTSB Percent growth inhibition of A549 by 4-[1-(2-Bromo-4-fluoro- phenyl)-1H-tetrazole-5-sulfonyl]-butyronitrile

Percent growth inhibition of HeLa-S3 cells by 4-[1-(2-Bromo- 4-fluoro-phenyl)-1H-tetrazole-5-sulfonyl]-butyronitrile

Percent growth inhibition of A549 by 4-[1-(4-Bromo-phenyl)- 1H-tetrazole-5-sulfonyl]-butyronitrile

Percent growth inhibition of HeLa-S3 cells by 4-[1-(4-Bromo- phenyl)-1H-tetrazole-5-sulfonyl]-butyronitrile

Percent growth inhibition of A549 by 3-(benzothiazole-2- sulfonylmethyl)-benzonitrile

Percent growth inhibition of HeLa-S3 cells by 3- (benzothiazole-2-sulfonylmethyl)-benzonitrile Percent growth inhibition of A549 cells by 4-(1-Phenyl-1H- tetrazole-5-sulfonylmethyl)-benzonitrile

Proposed mechanism for TxR inactivation by an unsaturated Mannich base in Plasmodium falciparum

3-aminopyridine-adenine dinucleotide phosphate absorbance

apparent Km

arginine antisense RNA adenosine triphosphate 4-[1-(2-bromo-4-fluoro-phenyl)-/H-tetrazole-5-sulfonyl]- butyronitrile

bovine serum albumin wild-type yeast, MATa his3A1 leu2A0 met15A0 ura3A0 wild-type yeast, MATa his3A1 leu2A0 lys2A0 ura3A0 BY4741 x BY4742

degree centigrade calcium

Xvi

cyclin-dependent kinase complementary deoxyribonucleic acid Chinese hamster ovary

carbon dioxide cysteine doublet 2’-deoxyadenosine 5’-triphosphate 2’-deoxycytidine 5’-triphosphate doublet of doublets

2’-deoxyguanosine 5’-triphosphate Dulbecco’s modified Eagle’s media dimethyl] sulfoxide

deoxyribonucleic acid drug sensitive strain of yeast (CCY333), MATa ura3-1 leu2-3,11

his3-11 trp1-1 can1-100 ade2-1 barl-1 erg64::TRP1 pdr34::HISS

pari4::KAN 5,5'-dithio-bis(2-nitrobenzoic acid), Ellman’s reagent

Xvli

Escherichia coli ethylenediaminetetraacetic acid eukaryotic initiation factor eukaryotic initiation factor phosphorylated at serine 51

ERG6 yeast deletion strain, (MATa his3Al leu2A0 met15A0 ura3A0 AERG6)

ethanol example flavin adenine dinucleotide fetal bovine serum

FK506 binding protein FK506 binding protein Rapamycin Associated Protein grams

Geneticin GCD1 heterozygous yeast strain, (MATa/MATalpha his3Al/his3 Al leu2A0Neu2A0 lys2A0/+ met15A0/+

ura3A0/ura3A0 AGCD1!) GCN1 yeast deletion strain, (MATa his3 Al leu2A0 met15A0 ura3A0 AGCN1)

xvin

Genetic Optimisation for Ligand Docking glutathione-S-transferase

hour proton

tritium hydrochloric acid human cervical cancer cells histidine

HOMS3 yeast deletion strain, (MATa his3 Al leu2A0 met15A0 ura3A40 AHOM3)

XIX

Interleukin-2, controls the growth and function of many types of cells

isopropanol potassium chloride inhibition constant concentration of substrate that leads to half-maximal velocity multiplet

molar, moles per liter milligram per liter minute

mega hertz Munich Information Center for Protein Sequencing millimolar

magnesium monoperoxyphthalate mode-of-action by network identification messenger ribonucleic acid

XX

sodium sulfite Transcriptional activator of genes involved in nitrogen catabolite repression, member of the GATA family of DNA binding proteins; activity and localization regulated by nitrogen limitation and Ure2p nanomolar

4-(1-phenyl-/H-tetrazole-5-sulfonyl)-butyronitrile

quartet ribonucleic acid interfering ribonucleic acid reactive oxygen species revolutions per minute

XXI

sodium dodecyl sulfate polyacrylamide gel electrophoresis sodium chloride sodium citrate buffer

triplet tetrahydrofuran threonyl-tRNA synthetase potassium trimethylsilanolate target of rapamycin 1

target of rapamycin 2 transfer ribonucleic acid thioredoxin reductase 1 heterozygous yeast strain, MATa/MATalpha his3Al/his3Al leu2A/leu2A lys2A0/met15A0/+ ura3A0/ura340 ATRRI1

thioredoxin reductase 2 yeast deletion strain, (MATalpha his3 Al leu2A0 lys2A0 ura3A ATRR2)

microliter micromolar enzyme units per milliliter Nitrogen catabolite repression regulator that acts by inhibition of GLN3 transcription in good nitrogen source

URE2 yeast deletion strain ultraviolet

initial velocity vascular endothelial growth factor volume to volume

wild-type times gravity yeast extract bacto-peptone dextrose media yeast protein database

zinc chloride

XXII

Introduction

Drug target validation is a time consuming stage in the drug discovery process The potential number of drug candidates and protein targets has increased significantly through advances in genomic sciences, combinatorial chemistry, and automated high- throughput screening Combining chemical genomics and systems biology provides a practical solution to screening the growing number of drug candidates and the target validation process

In the initial stages of drug discovery, natural products and small-molecule libraries are assayed for activity in target-directed or phenotype-based screens.’ Small molecule libraries are generated using target-directed and diversity-oriented synthesis.” In target-directed synthesis, the compounds are designed using a retrosynthetic route from a particular target.” In one such example, 70 bishydroquinone derivatives of a natural product, saframycin A (Figure 1.1), were synthesized and screened for anti-proliferative activity in A375 melanoma and A549 lung carcinoma cell lines.” Half of the derivatives showed greater activity than saframycin A.° Salicylamide and quinoline-2-carboxylic acid amide derivatives were the most potent.’ In addition to growth inhibition, target- directed screens also implement binding and enzymatic response to detect inhibition or activation of a protein target.' One of the downfalls of these screens is that the assays require purified enzymes, which entail protein expression, isolation, and then purification In addition, many new targets require assay development

are employed in phenotype-based screens to explore new targets and examine biological pathways.” In these screens, a phenotype is linked to a particular response in an organism, and analysis includes marker or reporter genes, functional assays, and cell imaging.’ There are many examples in the literature in which diversity-oriented libraries are screened for biological activity In one such screen, over 3,000 1,3-dioxanes were

synthesized without bias towards a particular protein target.’ Then the diversity-oriented library was arrayed on glass slides and probed with fluorescently labeled Ure2p, which is involved in glucose signaling and represses the transcription factors Gln3p and Nillp.* Several compounds, including uretupamine, were identified to bind to Ure2p4 In addition, this example is highlighted as whole-genome transcription profiling experiments were used to further identify the target Uretupamine was shown to activate a glucose-sensitive pathway downstream of Ure2p.* As with uretupamine, phenotype- screens often require additional biochemical assays to further investigate the mechanism

of action of the drug

After library synthesis and biological screening, the next step in the drug discovery process is the identification and validation of the drug target Few methods for efficient drug target identification exist The most common method is affinity chromatography for direct identification which requires immobilization of the compound

on to a resin.° Cellular extracts are then applied to the resin, and the proteins that bind the small molecules are isolated and sequenced to determine the target.!

The target validation of FK506 (Figure 1.1) employed affinity chromatography FK506 is a macrolide that was identified in a screen for immunosuppressive activity.”

The target of FK506 was elucidated using target-oriented synthesis.® A FK506 affinity matrix was synthesized using an amino derivative of FK506 A drawback to this method

is that synthesis was required to access the target Cellular extract was added to the affinity matrix, and the target protein, FK506 binding protein (FKBP), was eluted using FK506 and sequenced A summary of the target identification is described in Figure

1.2.8

Two complications encountered in affinity chromatography are the immobilization cannot alter the activity of the compound and the extract must contain the target in the active form.’ Additional direct methods include expression or display cloning, yeast three-hybrid systems, and protein microarrays.”

Other methods for target identification are more indirect, focusing on inference, and examine cellular changes to elucidate a protein target Monastrol, an inhibitor of mitotic kinesin Eg5, a motor protein required for spindle bipolarity, was identified using

a combination of two phenotype-based screens.’ A library of over 16,000 small- molecules was screened in a whole-cell mitotic arrest assay (Figure 1.2.C) In this assay, the library was screened for an increase in phosphorylation of nucleolin Nucleolin is a nucleolar protein that is phosphorylated in cells entering mitosis Increased phosphorylation was observed for 139 compounds.”

Compounds that showed increased phosphorylation were subjected to a phenotype assay to remove the compounds that target tubulin, which is a known target in mitotic arrest In this assay, the compounds were analyzed for inhibition of tubulin polymerization The compounds that inhibited this process were eliminated from the

Direct Methods Indirect Methods

Affinity chromatography Phenotypic assays

Yeast 3-hybrid Drug induced DNA/protein expression profiles

Proteomics B) Example of Direct Drug Target Identification

CHO lol ° ne boda? -

1 Library sereen ` ° „" Targetorlented -, Ps Add cell extract,

2 Identify lead Zoo ° synthesis sự eo” “4

C) Example of Indirect Drug Target Identification

ene, Whole-cell mitotic

| arrestassay | Phenotypic Assay Phenotypic Assay

139 leads Posttranslatonal @ visualizing

modification microtubules and

Identify lead y„ '-⁄ o for Eg5-driven Inhibits mitotic

HN Sooty microtubule '” kinesin Eg5

of microtubules, actin and chromatin using fluorescence microscopy.” Five compounds affected only mitosis One of these compounds, monastrol, showed a different

microtubule and chromosome arrangement, in which a monoastral microtubule array was surrounded by a ring of chromosomes.’

Mitotic kinesins are a class of proteins that are involved in the assembly and maintenance of the mitotic spindle Inhibition of kinesin Eg5 induced monoasters similar

to those resulting from treatment with monastrol It was proposed that Eg5 was the target

of monastrol The target was confirmed using an in vitro assay for Eg5-driven microtubule motility Monastrol inhibits Eg5 activity.” Another indirect approach to drug target identification includes the comparison of compound-induced DNA or protein expression profiles to those obtained from known drug targets.”

The target identification process for cyclosporin A and FK506 emphasizes the complexity of the validation process and demonstrates many of the previously mentioned methods of target identification Cyclosporin A (Figure 1.1) is a cyclic peptide with immunosuppressant activity which is used to prevent organ rejection in kidney and liver transplants.'° It inhibits the initial steps of T-lymphocyte activation and the production of interleukin-2.'!'? The protein target of cyclosporin A, cyclophilin, was identified using a radioactive binding assay with [°H] labeled cyclosporin A.'? Cyclophilin is a pepidyl- prolyl cis-trans isomerase.'“!> Cyclosporin A inhibits the rotamase activity of cyclophilin '!

Similar to cyclophilin, FKBP, the protein target of FK506, catalyzes the cis-trans isomerization of the proline amide in a peptide bond.® Rapamycin (Figure 1.1), another

macrolide with immunosuppressant activity, displaces FKBP from a FK506 affinity matrix.®'> There was no cross displacement with either FK506-cyclophilin or cyclosporin A-FKBP.° FK506 is 10 to 100 fold more potent in its ability to inhibit IL-2 mRNA synthesis than cyclosporin A, but an equally specific anti-T-cell agent, “l8

Rapamycin is a macrolide with a structure similar to FK506 although it has a different mechanism of action.'® The macrolide does not inhibit IL-2 transcription, but rather blocks later events for T-cell activation.'” Rapamycin penetrates the cell membrane and binds to FKBP.”° This complex then binds to FRAP (FKBP Rapamycin Associated Protein) which inhibits the T-cell response to IL-2, stopping the cell cycle in the G1-S transition.”° A genetic screen in S cerevisiae for rapamycin-resistant mutants identified two genes, TORJand TOR? (target of rapamycin 1 and 2).”' The yeast homolog of FRAP are Tor proteins, phosphatidylinositol kinase-relate kinases.”” Whole-genome profiling data in S cerevisiae reveals shift to low-quality carbon and nitrogen sources upon treatment with rapamycin, which suggests the Tor proteins are sensors for carbon and nitrogen.”>**

The similarities in the biological functions of cyclosporin A and FK506 suggest their associated complexes have a common biological target In order to determine the target, the proteins, FKBP and cyclophilin, were fused separately to glutathione S-

*> Calcium and the transferase, and purified using a glutathione affinity column

corresponding peptide ligands were added to the GST-fused proteins The tissue extract was incubated on the column, and the bound proteins were eluted using glutathione Calmodulin, a calcium binding protein, and a heterodimer of calcineurin (A and B) were isolated from the affinity column.’> Ca**- and calmodulin- dependent serine/threonine

FK506 complexes.”°

In mammals and budding yeast, FK506 and cyclosporin A inhibit calcineurin.?6?7 Immunosuppressant drugs that effect calcineurin are mediated through receptors called immunophilins.Š Chemical genetics was used to examine the effects of cyclosporin A and FK506 in the presence and absence of calcineurin in yeast.”®

Chemical genetics is the study of gene-protein function at the cellular level using chemical ligands.’ This technique is a practical tool to help alleviate the slow stage of the drug target identification process In a traditional genetic approach, mutations and gene deletions are used to identify genes and proteins that regulate biological processes The forward genetic approach begins with random mutagenesis, followed by biological screening for a desired phenotype.” The mutations resulting in this phenotype are then identified In the reverse approach, a single mutation is introduced, and the resulting phenotype is observed This approach is limited by the growth rate and large genomes of some model systems.! In addition, mutations are not conditional and some are lethal

Chemical genomics is the use of chemical ligands to perturb the entire genome of

a biological system and overcomes some of the obstacles encountered with traditional genetics (Figure 1.3) In the forward approach, chemical ligands, or mutation equivalents, are screened for a desired phenotype and then the genetic or protein target is identified.’

In this approach, the mechanism of the small molecule is often unknown.”’ In the reverse approach, a protein of interest is overexpressed Then ligands are screened for binding with the protein of interest, and confirmed using a phenotypic assay The chemical genetic approach is conditional and applicable to more complex systems.’ In addition to

Target directed

libraries

Leads Target Identification

Screen library & ee

B) Reverse chemical-genetic approach

Screen library a a _ Leads „ Assay for phenotype

for protein ligand

the biological screens, a drawback to this method is the synthetic effort required to prepare the ligands

Thirty percent of the genes associated with human disease have yeast homologs ”>? Saccharomyces cerevisiae is often used as a model system to examine eukaryotic diseases and pathways in chemical genomic experiments In addition, signaling pathways and cellular processes, like macromolecule synthesis, chromosome replication, and chromosome segregation are generally conserved between yeast and higher eukaryotes.?”” The sequencing of the yeast genome and creation of public yeast databases generates access to a plethora of genetic information.***° Yeast provide several experimental advantages to other model systems Yeast exist as stable haploid and diploid cells Gene deletion strains and heterozygous strains are commercially available.°” The Saccharomyces Genome Deletion Project has deleted 95% of the approximately

6200 open reading frames.°’ Heterozygous strains provide insight into deletions that may otherwise be lethal and sensitize a diploid cell to a drug, which is referred to as drug- induced haploinsufficiency.”Š Compared to other models, yeast offers the additional desired properties of rapid growth, clonability, and ease of culture maintenance.””

Whole-genome microarray experiments measure changes in mRNA transcript levels of an entire genome in response to a small molecule (Figure 1.4) Total RNA is separated from treated and control cultures using a phenol chloroform extraction, followed by mRNA isolation using a poly T resin Complimentary DNA is synthesized

by reverse transcription using amino-allyl dUTP, dATP, dGTP, dCTP, and dTTP After reverse transcription, the control and treated cDNA are coupled with fluorescent probes, cyanine-3 and cyanine-5 derivatives of N-hydroxysuccinimidyl esters, respectively The

amino-allyl dUTP reacts with dyes, labeling the samples The labeled probes are combined and then competitively hybridized to DNA microarrays consisting of 70mer oligonucleotide probes for 6300 yeast genes The levels of hybridization are determined

by simultaneous laser scanning at 532 and 635 nm for the two fluorescent probes The transcript ratios are determined for each gene by comparing the relative intensities of the treated channel (635 nm) to the control channel (532 nm)

Transcription profiles are used to measure small molecule specificity, analyze pathways, predict biological responses, and as a screening tool.“ Microarray analysis was used in the chemical genomics experiments with FK506 and cyclosporin A.”* Wild-type yeast were treated with FK506 and cyclosporin A, independently.”® Yeast strains with the calcineurin subunits deleted were grown in parallel without the drugs.”* The expression profiles for each experimental condition were generated The profiles of the calcineurin deletion strain and the drug treated stains were highly correlated.”® The results indicate that cellular response to treatment with FK506 and cyclosporin A is similar to the effects

Scan slide and >

analyze data using GenePix Pro

and pathways involved in diauxic shift, the shift in yeast from fermentation to respiration using a non-fermentable carbon source, ethanol.*'

Expression data is also used to predict biological responses and as screening tool The National Cancer Institute uses 60 human cancer cell lines in the drug screening process."? The microarray profiling data of these cell lines upon treatment with 14,000 drugs was combined into a gene expression database.” The selected drugs were in clinical use for the treatment of cancer.” The gene expression and drug activity were correlated and clustered using bioinformatics and cheminformatic tools.*” The database

of expression data is a predictive tool for gene-gene, gene-drug and drug-drug relationships.”

The following chapters describe the application of chemical genomics to explore the molecular target of an anti-proliferative compound, borrelidin, and to identify the target of a tetrazole-containing compound, PTSB, with novel biological activity In addition, this work illustrates the importance of combining transcription profiling experiments with biochemical experiments and systems biology in the overall drug discovery process

Chapter 2- The application of chemical genomics to examine the molecular

target of an anti-proliferative compound, borrelidin

Introduction

Borrelidin (Figure 2.1) was first isolated in 1949 from Streptomyces rochei and shown to have antibiotic activity.’ In more recent studies, borrelidin displayed antimalarial activity, as well as inhibited angiogenesis, cyclin-dependent kinase (CDK),

and threonyl-tRNA synthetase (ThrRS) activity.“**’ The reported ICso values (Table 2.1)

suggest an undescribed molecular target As described in Chapter 1, chemical genomics provides a useful tool to examine biological activity by measuring changes in mRNA transcript levels upon treatment with a small molecule Chemical genomics was used to explore the molecular target of borrelidin as the literature presented conflicting reports

Borrelidin inhibits capillary tube formation in a rat aorta culture model, and protein and DNA synthesis in Human umbilical vein endothelial cells (HUVEC) with ICao values of 0.8 nM and 20 nM resepectively.”° In addition, the macrolide inhibits cell proliferation in a rat aorta matrix in which growth has been induced with vascular endothelial growth factor (VEGF).”°

In addition, borrelidin inhibits cyclin-dependent kinase (CDK) in Saccharomyces cerevisiae and selectively inhibits E coli threonyl-tRNA synthetase activity.“°*’ The ICso

of borrelidin for Cde28/Cln2, a cyclin-dependent kinase in yeast, is 24 nM.“ In

mammalian cells, cyclin dependent kinases phosphorylate the retinoblastoma protein, which inhibits the protein from blocking the expression of genes required to enter the S phase of the cell cycle, and prevents cell cycle progression.”

i O UCN-01 OH O flavopiridol

Figure 2.1 The chemical structures of borrelidin, 3-amino-1,2,4-triazole (3AT), and several other anti-cancer agents

A) Biological activity

inhibits Cdc28/Cin2 in Saccharomyces cerevisiae %8

inhibits protein and DNA synthesis in HUVEC 45 inhibits capillary tube formation in rat aorta mode | 45

inhibits H460 human large-cell lung cancer cells 45

inhibits WiDR human colon cancer cells 45

B) yeast strain

BY4741

BY4743 CCY333, DS*

erg64A

CDC28::cdc28

gcn4A

gen2A GCD1::gced1 hom3 A

ICso (u M) —

24

0.020 0.0008 0.18 0.10

1Cso (AM) -

25.1+0.2 21.6+6.5 0.021+0.007 0.41+0.2 7.2218 28.2£1.2 7,542.7 13.9+1.8 113+1.8

Table 2.1: A) Published ICzo values for borrelidin B) Experimentally determined ICso values for wild-type, deletion and heterozygous yeast strains BY4741 and BY4743 are the wild-type strains CCY333 and erg6A are dmg sensitive deletion strains CDC28::cdce28 is the heterozygous strain coding for a cyclin-dependent kinase (CDK) The gene products of GCN4 and GCN2 are a transcriptional activator of amino acid biosynthetic genes, and the eukaryotic initiation factor 2 a (eI[F2a) kinase, respectively GCD1 is a negative regulator in the general control of amino acid biosynthesis, and the gene product is the gamma subunit of the translation initiation factor eIF2B The gene product of HOM3 is aspartate kinase which catalyzes the first step in the common pathway for methionine and threonine biosynthesis The expression of HOM3 is

regulated by Gen4p