Discovery and Mechanistic Investigation of Antifungals

October 31, 2013

Ameeta Agarwal Senior Scientist National Center for Natural Products Research University of Mississippi

Overview

First Half:

Introduction to DNA microarray technology

Experimental considerations for using DNA microarray technology

Second Half: Example of how DNA microarray technology has been used in investigating the mechanism of action of an antifungal natural product compound

Limitations and future directions

How to Study the Effect of an Antifungal Agent at a Whole-Genome Level?

By Using DNA Microarray Technology

~7.5 cm ~2.5 cm

(~20,000 genes)

Treatment of a cell with a antifungal agent

Gene expression changes in response to the agent

Monitor gene expression changes using DNA microarrays

What is a DNA Microarray?

A solid support (glass microscope slide, nylon membrane, plastic slide) on which small DNA spots are deposited robotically and chemically fixed in place. The identity of the DNA spot (i.e., gene information) is pre-determined. The entire set of genes of an organism can be placed on a microarray.

~7.5 cm ~2.5 cm

(~20,000 genes)

What is a DNA Microarray? (contd.)

Types of DNA Microarrays
Spotted Arrays (DNA is spotted on a solid surface) Affymetrix GeneChips (DNA is synthesized on a silica surface)

~7.5 cm ~2.5 cm

~7.5 cm ~2.5 cm

(~20,000 genes)

(~20,000 genes)

cDNA Arrays
cDNA samples are spotted

Oligonucleotide Arrays
Oligonucleotide samples are spotted (consisting of a stretch of 60-70 nucleotides)

(up to ~40,000 genes, each represented 16 times)

High-density Oligonucleotide Arrays

Oligonucleotide samples are synthesized directly on the array (consisting of a stretch of 25 nucleotides)

Gene

Exon 1 Intron 1 Exon 2 Intron 2 Exon 3

Transcription mRNA
Exon 1 Exon 2 Exon 3

Gene

Exon 1 Intron 1 Exon 2 Intron 2 Exon 3

Either cDNA or an oligonucleotide is placed on an array: - cDNA is made from mRNA using the enzyme Reverse Transcriptase. - An oligonucleotide is synthesized chemically.

16 oligos covering different areas of the gene

Microarray Experiment Overview (Spotted Arrays)

Cy5

Control Cells (solvent-treated)

Isolate RNA

Prepare cDNA

Label

Cy3

Hybridize target samples to DNA probes on a microarray slide

Sample Cells (drug-treated)

Isolate RNA

Prepare cDNA

Label

Acquire fluorescent image Quantitate data Identify differentially expressed genes

Gene expression changes in response to drug treatment are reflected in the RNA populations The two fluorescent dyes allow simultaneous detection of gene expression in the two samples Amount of hybridization to the DNA microarray slide corresponds to level of gene expression Differentially expressed genes are identified using a database from the microarray manufacturer

Example of a Microarray Image (Spotted Arrays)

Raw Image of Cy3 Fluorescence (Control Cells) Raw Image of Cy5 Fluorescence (Compound-Treated Cells)

Data Analysis
Is the signal real? Is the data comparable across arrays (normalization)? Comparison between untreated and treated classes (modified t-tests) Data filtering (is useful data being compared?) - Remove genes with missing signal values - Remove genes with very low signal values - Remove genes with similar signal values across the arrays (unchanged genes)

Final result: Gene expression change = Signal in treated sample Signal in untreated sample

These parameters apply to both Spotted arrays and Affymetrix arrays

Normalization: Correcting Differences Between Cy3 and Cy5 Hybs

Untreated (Cy3)

Treated (Cy5)

Untreated (Cy3)

Treated (Cy5)

Treated (signal)

Treated (signal)

Genes with equal hybridization in the two samples

Untreated (signal)

Untreated (signal)

Normalization corrects for hybridization differences between Cy3 and Cy5 samples (majority of genes should give equal hybridization) Global normalization (total intensity normalization) Use of house-keeping genes Lowess normalization (based on linear regression)

Example: Microarray Data

Cy3 Gene Description Signal YGR226C:N/A:unknown 694.78 YGR226C:N/A:unknown 733.35 YJR035W:RAD26:putative helicase 495.80 YJR035W:RAD26:putative helicase 388.85 YJL196C:ELO1:fatty acid elongation protein 941.06 YJL196C:ELO1:fatty acid elongation protein 838.96 YJL132W:N/A:unknown; similar to phospholipase D 337.75 YJL132W:N/A:unknown; similar to phospholipase D 237.49 YJL068C:N/A:unknown; similar to human esterase D 1218.57 YJL068C:N/A:unknown; similar to human esterase D 978.54 YJL007C:N/A:unknown 348.14 YJL007C:N/A:unknown 320.10 YIL164C:NIT1:nitrilase 2319.78 YIL164C:NIT1:nitrilase 2212.60 YIL100W:N/A:unknown 226.69 YIL100W:N/A:unknown 297.08 YIL036W:N/A:unknown; similar to Mei4p and to cAMP response element 1078.21 YIL036W:N/A:unknown; similar to Mei4p and to cAMP response element 1207.44 YHR194W:N/A:unknown 1192.48 YHR194W:N/A:unknown 1163.26 YHR132C:ECM14:unknown 2402.31 YHR132C:ECM14:unknown 2446.35 YKL127W:PGM1:phosphoglucomutase 2663.07 YKL127W:PGM1:phosphoglucomutase 2895.34 YKL064W:MNR2:unknown 7763.88 YKL064W:MNR2:unknown 7305.94 Cy5 Signal 711.58 680.26 436.99 492.67 2444.75 2504.20 435.66 373.80 1112.03 866.67 344.73 374.78 2693.64 2515.26 367.39 236.89 1225.87 1139.12 1280.03 1367.75 2385.96 2124.38 2334.90 2843.34 6444.80 6060.51 Cy5:Cy3 Ratio 1.02 0.93 0.88 1.27 2.60 2.98 1.29 1.57 0.91 0.89 0.99 1.17 1.16 1.14 1.62 0.80 1.14 0.94 1.07 1.18 0.99 0.87 0.88 0.98 0.83 0.83

Majority of genes will have a Cy5:Cy3 ratio of 1.0

Example: Microarray Data (Contd.)

Expt#2 Expt#3 Cy5 Cy5 Cy3 Cy5 Cy3 Signal Signal Cy5:Cy3 Cy3 Signal Signal Cy5:Cy3 Signal Signal Cy5:Cy3 (Untreated) (Treated) Ratio (Untreated) (Treated) Ratio (Untreated (Treated) Ratio Gene Description YGR226C:N/A:unknown 694.8 711.6 1.0 781.2 802.6 1.0 431.5 510.7 1.2 YGR226C:N/A:unknown 733.4 680.3 0.9 1014.1 904.7 0.9 417.0 551.5 1.3 YJR035W:RAD26:putative helicase 495.8 437.0 0.9 607.2 491.1 0.8 316.2 407.3 1.3 YJR035W:RAD26:putative helicase 388.8 492.7 1.3 652.0 540.4 0.8 325.6 348.5 1.1 YJL196C:ELO1:fatty acid elongation protein 941.1 2444.8 2.6 1151.2 3447.7 3.0 634.3 1695.4 2.7 YJL196C:ELO1:fatty acid elongation protein 839.0 2504.2 3.0 953.4 3723.4 3.9 525.1 1597.1 3.0 YJL132W:N/A:unknown; similar to phospholipase D 337.7 435.7 1.3 420.2 552.5 1.3 173.8 342.7 2.0 YJL132W:N/A:unknown; similar to phospholipase D 237.5 373.8 1.6 352.1 326.1 0.9 176.8 216.9 1.2 YJL068C:N/A:unknown; similar to human esterase D 1218.6 1112.0 0.9 2025.5 1442.8 0.7 892.0 880.9 1.0 YJL068C:N/A:unknown; similar to human esterase D 978.5 866.7 0.9 1853.4 1765.4 1.0 753.7 587.4 0.8 YJL007C:N/A:unknown 348.1 344.7 1.0 692.0 477.4 0.7 234.2 305.6 1.3 YJL007C:N/A:unknown 320.1 374.8 1.2 681.5 464.6 0.7 211.6 235.4 1.1 YIL164C:NIT1:nitrilase 2319.8 2693.6 1.2 4195.6 4817.7 1.1 1516.2 1718.0 1.1 YIL164C:NIT1:nitrilase 2212.6 2515.3 1.1 4330.2 4769.1 1.1 1330.5 1614.9 1.2 YIL100W:N/A:unknown 226.7 367.4 1.6 492.1 330.0 0.7 151.1 169.7 1.1 YIL100W:N/A:unknown 297.1 236.9 0.8 403.8 327.7 0.8 152.8 175.5 1.1 YIL036W:N/A:unknown; similar to Mei4p and to cAMP response 1078.2 element 1225.9 1.1 2129.1 1809.3 0.8 803.8 732.2 0.9 YIL036W:N/A:unknown; similar to Mei4p and to cAMP response 1207.4 element 1139.1 0.9 2082.2 2127.3 1.0 879.3 908.4 1.0 YHR194W:N/A:unknown 1192.5 1280.0 1.1 1996.1 1985.2 1.0 775.9 824.1 1.1 YHR194W:N/A:unknown 1163.3 1367.8 1.2 2063.5 1946.5 0.9 788.4 799.6 1.0 YHR132C:ECM14:unknown 2402.3 2386.0 1.0 5312.1 4425.4 0.8 1666.0 1262.6 0.8 YHR132C:ECM14:unknown 2446.4 2124.4 0.9 5542.5 4392.5 0.8 1698.3 1435.6 0.8 YKL127W:PGM1:phosphoglucomutase 2663.1 2334.9 0.9 3311.2 2835.1 0.9 1619.4 1529.2 0.9 YKL127W:PGM1:phosphoglucomutase 2895.3 2843.3 1.0 3830.7 2711.4 0.7 2070.1 1679.5 0.8 YKL064W:MNR2:unknown 7763.9 6444.8 0.8 7952.6 7393.2 0.9 5128.1 4178.3 0.8 YKL064W:MNR2:unknown 7305.9 6060.5 0.8 8341.0 6702.2 0.8 4541.8 3753.5 0.8 Expt #1

Statistical Analysis of Data: Class comparison between untreated class and treated class (probability test for difference between the two classes, i.e., deviation from 1.0)

Microarray Experiment Overview (Affymetrix GeneChips)

B B
Control Cells (solvent-treated) Isolate RNA Prepare cDNA

B B
Hybridize target samples to DNA probes on a GeneChip

Biotin-labeled cRNA

B
Sample Cells
(compound-treated)

B B
Hybridize target samples to DNA probes on a GeneChip

B
Isolate RNA Prepare cDNA

Biotin-labeled cRNA

Wash chip Stain with fluorescent antibody Detect fluorescence signal Quantitate signals Identify differentially expressed genes

Affymetrix Image File: Example

Is the Signal Real? (Affymetrix Arrays)

Each gene is represented ~32 times on the array Some gene elements are designed to give greater specificity Some gene elements are designed to give weaker specificity The Affymetrix software (AGCC) performs a modified t-test to determine if the signal is specific Result: One computed signal value per gene P-value to indicate significance Detection call P=Present (very strong specific signal) M=Marginal (weaker signal) A=Absent (very weak non-specific signal) Genes with P calls, strong signal values and low p-values are considered useful

Statistical Analysis of Data: Class Comparison

Signal Signal Un-A Un-B (Chip 1) (Chip 2) Gene 1 Gene 2 Gene 3 Gene 4 Gene 5 Gene 6 Gene 7 Gene 8 Gene 9 Gene 10 Gene 11 Gene 12 Gene 13 Gene 14 Gene 15 Gene 16 Gene 17 Gene 18 Gene 19 Gene 20 Gene 21 Signal Un-C (Chip 3) Signal Tr- Signal Tr- Signal Tr- Mean Signal A B C (Chip 4) (Chip 5) (Chip 6) (Un) Mean Ratio p-value Signal (Tr/Un) (Tr)

Data Analysis: Class comparison between untreated & treated classes (probability test for difference between the two classes) Software used at NCNPR for statistical analysis of microarray data: BRBArrayTools

What is the Final Outcome of a DNA Microarray Experiment?

Initial data analysis identifies a list of genes that are differentially expressed under a given experimental condition, i.e., which genes are induced and which genes are repressed by a compound.

Further data analysis involves identifying the genes (gene annotation), determining what biochemical pathways they belong to, determining which biochemical pathways are over-represented in the data set, and interpreting the data.

Advanced data analysis involves comparing the data to previously described genomic profiles (searching databases, cluster analysis), and network analysis (is there cross-talk between the identified pathways).

Identification of MOA Using DNA Microarrays - Example

Squalene ERG1 ERG7 ERG11 ERG24 ERG25 ERG26 ERG6 ERG2 ERG3 ERG5 ERG4 Squalene expoxidase Lanosterol synthase Lanosterol C-14demethylase Sterol C-14reductase Sterol C-4methyloxidase Sterol C-3 dehydrogenease Sterol C-24 methyltransferase Sterol C-8 isomerase Sterol C-5 desaturase Sterol C-22 desaturase Sterol C-24 reductase

Gene Expression Response to Ketoconazole in Yeast Cells

Carbohydrate Cell stress, 7.7% metabolism, 1.5% Cell cycle Cell wall control, 1.5% Amino acid maintenance, 9.2% metabolism, 1.5%
Unknowns, 20.0% Transport, 7.7% Energy generation, 1.5% Lipid, fatty acid and sterol metabolism, 26.2% HES1 YSR3 CYB5 ERG28 ERG3 UPC2 ERG4 ERG1 ERG25 ERG5 ERG2 ERG24 ERG6 ERG26 43.5 (18.4, 68.6) 4.4 (4.3, 4.6) 4.1 (3.2, 4.9) 3.9 (3.7, 4.0) 3.8 (3.0, 4.6) 3.8 (4.3, 3.2) 3.4 (2.5, 4.3) 3.1 (2.6, 3.5) 2.8 (2.6, 3.0) 2.6 (2.6, 2.5) 2.5 (2.0, 3.0) 2.5 (2.3, 2.6) 2.4 (2.3, 2.5) 2.3 (2.1, 2.5) Differentiation, 1.5%

Signal transduction, 3.1% Protein translocation, 1.5% Transcription, 3.1% Other function, 10.8%

Ergosterol

Ketoconazole = Antifungal drug that inhibits ergosterol biosynthesis by targeting ERG11

Mating response, 3.1%

ERG11 2.2 (2.3, 2.1)

INO1 ELO1 -2.1 (-2.0, -2.1) -3.9 (-4.0, -3.7)

Classification of Compounds
Known MOA
Compound A Compound B

Unknown MOA
Compound X Compound Y Compound Z

(Membrane-disrupter) (DNA-damager)

DNA-damager

Experimental Considerations for using DNA Microarray Technology

Is DNA microarray technology the appropriate technology for my project?
- Does my project require the study of whole-genome effects? - Does my project require the study of a few genes?

Are the appropriate tools available or accessible to me for DNA microarray studies?

- Are DNA microarrays available for my target organism? Is a model organism more suitable? - Is the appropriate instrumentation available to me? Can I find a collaborator who has the instrumen - Do I have the funding to conduct DNA microarray studies?
Cost of a drug MOA study with DNA microarrays GeneChip Cost: $1,380 (One chip = $230, need 6 chips) Reagents for target preparation: $600 (One target = $100, need 6 targets) Reagents for hybridization, staining, etc : $600 ($100 per chip) Miscellaneous supplies: $300 (media, plasticware, RNA isolation kits) Total: ~$3,000

Statistical considerations
- At least triplicates - Do I have access to appropriate statistical analysis software? Do I need training? Collaborator? - Are there appropriate databases available for downstream analyses?

Example of how DNA microarray technology has been used by our group to identify the MOA of the antifungal compound sampangine
N

N O

References: Agarwal AK, Xu T, Jacob MR, Feng Q, Lorenz MC, Walker LA, Clark AM (2008) Role of heme in the antifungal activity of the azaoxoaporphine alkaloid sampangine. Eukaryot Cell. 7:387-400. Huang Z, Chen K, Xu T, Zhang J, Li Y, Li W, Agarwal AK, Clark AM, Phillips JD, Pan X (2011) Sampangine inhibits heme biosynthesis in both yeast and human. Eukaryot Cell. 10:1536-1544.

DNA Microarray Experiment with Sampangine using the Model Yeast Saccharomyces cerevisiae

DMSO (0.25%)

SMP (IC50 conc., 1.2 g/mL)

Expose for 1.5 doublings (~ 4.5 h)

Isolate RNA

Prepare biotinylated targets

Hybridize to Affymetrix Arrays

DNA Microarray Experiment with Sampangine - Results

Biological Process

Response to Stimulus

Metabolism Cellular Metabolism Lipid Metabolism

Cellular Process

Cell Organization and Biogenesis

Homeostasis

Response to Oxidative Stress

Transport Generation of Precursor Metabolites and Energy Ion Transport Cellular Respiration and Mitochondrial Electron Transport

Ion Homeostasis Organelle Organization and Biogenesis

Iron Ion Homeostasis

SRX1 (15.1) AAD4 (6.7) GPX2 (5.1) AAD6 (4.6) GTT2 (4.3) CCP1 (3.9) TSA2 (2.7) CTA1 (2.6) PRX1 (2.6)

IZH4 (14.4) HES1 (11.4) IZH2 (2.9) ATF2 (2.9) UPC2 (2.7) SFK1 (2.7) ATF1 (2.7) ARE1 (2.0) YMR210W (-2.0) ERG20 (-2.1) MCR1 (-2.4) ERG28 (-2.6) ERG5 (-2.8) CYB5 (-3.0)

Sterol Transport

Cell Wall Organization

Iron Ion Transport

DAN1 (67.5) AUS1 (5.6) SUT2 (3.2) SUT1 (2.3)

NDI1 (-2.2) COX15 (-2.2) COX6 (-2.3) MCR1 (-2.4) COX5B (-2.4) COX12 (-2.5) QCR7 (-2.6) DLD1 (-2.8) QCR9 (-2.8) CYC7 (-3.0) RIP1 (-3.1) COX13 (-3.4) QCR2 (-3.4) PET10 (-4.0) COX5A (-4.4) QCR10 (-5.2) NDE1 (-5.2) COX4 (-13.2) CYC1 (-13.3)

FRE1 (-2.2) ENB1 (-2.4) FRE4 (-2.6) FRE5 (-2.7) ARN1 (-3.3) SIT1 (-3.6) FIT3 (-5.2) FTR1 (-6.0) ARN2 (-9.0) FET3 (-11.0) FIT2 (-14.3)

DAN1 (67.5) TIR1 (9.9) DAN2 (3.7) TIR4 (3.0) DAN4 (2.5) TIR3 (2.5) DAN3 (2.5)

YLR126C (-3.3) SIT1 (-3.6) CCC2 (-5.0) HMX1 (-5.8) TIS11 (-8.6) ARN2 (-9.0) Other Processes

ANB1 (6.2) HEM13 (5.6) PAU1 (4.3) AAC3 (3.5) YLR413W (3.5) PAU6 (3.5) PAU4 (3.4) PAU3 (2.9) YGL039W (2.9) PAU7 (2.6) YJR116W (2.5) YGR131W (2.4) YLR437C (2.4) YMR002W (-2.0) YBR230C (-2.3) YOR215C (-2.5) YDL110C (-2.8) YHB1 (-3.7)

Transcript Profiling with Sampangine: Heme Deprivation Response in Yeast Cells

Upregulated genes
DAN1, DAN2, DAN4, TIR1, TIR3, TIR4

Cell wall mannoproteins Induced in anaerobic conditions Some are involved in sterol uptake
Induced in anaerobic conditions Involved in sterol uptake Transcription factor Regulates anaerobic genes Induced in anaerobic conditions

AUS1, SUT1, SUT2 UPC2

ANB1, PAU1, PAU3,PAU4, PAU5

HEM13

Involved in heme biosynthesis

Downregulated genes
COX12, COX13, COX15, COX4, COX5A, COX5B, COX6, CYC1, CYC7, QCR10, QCR2, QCR7, QCR9 ARN1, ARN2, FET3, FIT2, FIT3, FRE1, FRE4, FRE5, FTR1, HMX1, SIT1 Cytochrome c oxidases/reductases (mitochondrial electron transport) Iron uptake and homeostasis

Heme Biosynthesis Pathway in Yeast

Glycine + Succinyl CoA
Hem1

ALA
Hem2

- Does SMP inhibit the heme synthesis pathway? - Does this inhibition cause accumulation of pathway intermediates? - Does this inhibition cause reduction in heme production? - Will exogenous heme rescue cells from SMP sensitivity? - Will mutants with deletions in the pathway genes show hypersensitivity to SMP? - Does SMP interfere with the activity of a specific enzyme in the pathway?

Porphobilinogen
Hem3

Preurogen
Hem4

Urogen III
Hem12

Coprogen III
Hem13

Protogen IX
Hem14

Protoporphyrin
Hem15

Heme

Role of Heme in the Antifungal Activity of Sampangine

Mutant Analysis
-SMP
Wild-type

Checkerboard Assays

Biochemical Analysis

120

hem1

% Growth

100

+ 80 Heme
60 0.001 0.01 0.1 40 20 0 1 10

+SMP
Wild-type

15

7.5

SMP Conc. (g/ml)

-20

Amount total (% total porphryin) (% porphyrin) Amount

70 60 50 40 30 20 10 0

% Growth

hem1

120 100

+ SMP - SMP

+ Ergo
60 0.001 0.01 0.1 40 1 20 0 -20 10

80

Tr

Un

Uro

SMP Conc (g/ml)

Uro po rphyrin

Co pro po rphrin III P ro to po rphyrin IX

Copro III

Proto IX

SMP Reduces Heme Levels in Yeast and Human Cells

Yeast cells

Human cells

Xuewen Pan, Baylor College of Medicine (now at Novartis)

Effect of SMP on Heme Synthesis Pathway Enzymes

Glycine + Succinyl CoA
Hem1

SMP does not inhibit Hem12 activity SMP induces Hem3 and Hem4 activity

ALA
Hem2

Porphobilinogen
Hem3

Preurogen
Hem4 Hem3 Hem4 Hem12

John Phillips, University of Utah

Urogen III
Hem12

SMP

Overexpression of HEM4 reduces heme levels

Coprogen III
Hem13

Protogen IX
Hem14

Protoporphyrin
Hem15
Xuewen Pan, Baylor College of Medicine (now at Novartis)

Heme

Conclusion: SMP hyperactivates Hem4 resulting in disruption of heme synthesis

Limitations and Future Directions

DNA microarrays are a starting point in a MOA study. Followup studies are necessary to identify the precise MOA of a drug. DNA microarrays only measure mRNA abundance, and dont account for changes in protein levels or metabolite levels

Use of proteomics and metabolomics technologies allow the measurement of proteins and metabolites respectively
Combining genomics, proteomics and metabolomics would be a powerful way of determining the global effects of drugs