NMR

1
NMR MetabolomicsThe Possibilities The
Limitations

• David Wishart
• Depts. Comp. Sci and Bio. Sci.
• University of Alberta NINT
• david.wishart_at_ualberta.ca

2
The (Human) Pyramid of Life
Metabolomics Proteomics Genomics
1400 Chemicals
3000 Enzymes
30,000 Genes
3
The (Bacterial) Pyramid of Life
Metabolomics Proteomics Genomics
761 Chemicals
1152 Enzymes
4269 Genes
4
Why Measure Metabolites?
Metabolites are the Canaries of the Genome
5
Metabonomics Metabolomics

• MetabonomicsThe quantitative measurement of the
  time-related total metabolic response of
  vertebrates to pathophysiological (nutritional,
  xenobiotic or toxic) stimuli
• MetabolomicsThe quantitative measurement of
  invertebrate metabolic profiles to characterize
  their phenotype or phenotypic response to genetic
  or nutritional perturbations

MetaboXomics
6
Metabolomics Allows One to..

• Generate metabolic signatures
• Monitor/measure metabolite flux
• Monitor enzyme/pathway kinetics
• Assess/identify phenotypes
• Monitor gene/environment interactions
• Track effects from toxins/perturbants
• Monitor consequences from gene KOs
• Identify functions of unknown genes

7
Traditional Metabolite Analysis
HPLC, GC, CE, MS
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
Problems with Traditional Methods

• Requires separation followed by identification
  (coupled methodology)
• Requires optimization of separation conditions
  each time
• Often requires multiple separations
• Slow (up to 72 hours per sample)
• Manually intensive (constant supervision, high
  skill, tedious)

23
New Approaches
24
Advantages

• Measure multiple (10s to 100s) of metabolites
  at once no separation!!
• Allows metabolic profiles or fingerprints to be
  generated
• Mostly automated, relatively little sample
  preparation or derivitization
• Can be quantitative (esp. NMR)
• Analysis results in lt 60 s

25
Why NMR?
Mixture separation by HPLC (followed by ID via
Mass Spec)
Mixture separation by NMR (simultaneous separation
ID)
Chemical Shift Chromatography
26
Why NMR?

• 1H NMR
• Rapid metabolite identification and
  quantification
• Monitoring flux/kinetics in real time
• 13C NMR
• Metabolite sources/sinks, chemistry
• 31P NMR
• ATP/ADP ratios, energy balance, cAMP

27
Metabolomics and NMR
Principle Component Analysis
28
Functional Proteomics via Metabolic Profiling
Forster, J. et al., (2002) Biotechnol. Bioeng.
Vol. 79, 703-712
29
Detecting Silent Mutations in Yeast
Nature Biotechnology, vol. 19, pg. 45-50 (2001)
30
Is There A Better Way?

• Why not try to identify the individual peaks in
  an NMR spectrum automatically?
• Use software to deconvolute individual spectral
  signatures using a database of known compounds
• Use relative peak intensity to quantify
• Gives identification and quantitation
• Possibility of chemical shift microscopy

31
Spectral Deconvolution of a Mixture Containing
Compounds A, B and C
32
Data Analysis (Principles)
Glucose Fit
Extract Spectrum
Fructose Fit
33
Data Analysis (Principles)
Fructose Fit
Extract Spectrum
Glucose Fit
34
Data Analysis (Principles)
Glucose and Fructose Fit
Extract Spectrum
Fructose Glucose
Glucose Fit
35
Data Analysis

• Fitting 5-10 rounded peaks is trivial, fitting
  1000 sharp peaks is not, i.e. dense matrix
  problem with very high probability of cumulative
  rounding errors and singularities(LLSOL -
  Stanford)
• Peak positions shapes dependent on salt, pH,
  temperature, ligands, ligand/ion interactions,
  shimming, signal-to-noise digital resolution,
  phasing, field strength, etc. etc.
• Requires special databases - key innovation
• Requires intelligent data preprocessing -
  (ditto)
• Partnered with Chenomx/Varian ? Eclipse

36
Fitting NMR Spectra with Eclipse
37
Current Compound List

• L-Isoleucine
• L-Lactic Acid
• L-Lysine
• L-Methionine
• L-phenylalanine
• L-Serine
• L-Threonine
• L-Valine
• Malonic Acid
• Methylamine
• Mono-methylmalonate
• N,N-dimethylglycine
• N-Butyric Acid
• Pimelic Acid
• Propionic Acid
• Pyruvic Acid
• Salicylic acid
• Sarcosine

• ()-(-)-Methylsuccinic Acid
• 2,5-Dihydroxyphenylacetic Acid
• 2-hydroxy-3-methylbutyric acid
• 2-Oxoglutaric acid
• 3-Hydroxy-3-methylglutaric acid
• 3-Indoxyl Sulfate
• 5-Hydroxyindole-3-acetic Acid
• Acetamide
• Acetic Acid
• Acetoacetic Acid
• Acetone
• Acetyl-L-carnitine
• Alpha-Glucose
• Alpha-ketoisocaproic acid
• Benzoic Acid
• Betaine
• Beta-Lactose
• Citric Acid
• Creatine

• DL-Carnitine
• DL-Citrulline
• DL-Malic Acid
• Ethanol
• Formic Acid
• Fumaric Acid
• Gamma-Amino-N-Butyric Acid
• Gamma-Hydroxybutyric Acid
• Gentisic Acid
• Glutaric acid
• Glycerol
• Glycine
• Glycolic Acid
• Hippuric acid
• Homovanillic acid
• Hypoxanthine
• Imidazole
• Inositol
• isovaleric acid

38
Why Not Apply It to Bacterial ID?
39
Metabolomics E. coli

• 25-50 mL shake flasks or large 0.5L flask
• Remove aliquots at regular intervals
• Analyze cells or media or both?
• Cells
• Lysis protocol, sonication? freeze-thaw? soluble
  fraction? lipid fraction? protein rmvl
• Media
• Rich (LB)? Defined? MOPS? M9? glucose?

40
M9-Glucose
MOPS
41
A Toy Problem

• Succinate Dehydrogenase is a key enzyme in the
  aerobic TCA cycle (converts succinate to
  fumarate)
• Fumarate Reductase is responsible for converting
  fumarate to succinate under anaerobic conditions
• Fumarate Reductase Succinate Dehydrogenase
  share 60 sequence identity

42
The TCA Cycle
Acetate
Acetyl-CoA
Glycerol
Pyruvate
Oxaloacetate
Citrate
Isocitrate
L-Malate
?-Ketoglutarate
Fumarate
2
1
Succinate dehydrogenase
Succinate
Succinyl-CoA
43
Questions

• Can Fumarate Reductase (FumR) substitute for
  Succinate Dehydrogenase (SucD)?
• Can we detect any phenotypic differences between
  WT vs. SucD- vs. SucD-/FumR-?
• Can NMR-based metabolomics be used to detect
  mutations/characterize phenotypes?

44
Methods

• Obtain 3 E. coli strains as shown below

45
Methods

• Grow Cells on Glycerol Minimal Media at 37
  degrees
• Collect 3 mL aliquots every hour for 30 hours or
  until cells die
• Monitor pH and OD600
• Spin down cells, retain supernatant
• Lyse cells with chloroform/water, spin down
• Analyze cell extracts supernatant by NMR

46
Typical Eclipse Readout (Concentration in mM)
47
Glycerol Consumption
48
Acetate Production
49
Succinate Production
50
Metabolic Responses
Acetate Glycerol Pyruvate
Acetate Glycerol Pyruvate
Succinate
Succinate
51
Results Interpretation

• DW35 (the double knockout) demonstrates a clear
  increase in succinate over time
• DW35 complemented with FumR (on the PH3 plasmid)
  appears to be capable of metabolizing succinate
• Analysis of growth media via NMR was sufficient
  to distinguish the two strains and to ID the gene
  knockout

52
Interpretation via SimCell
53
Cellular Automata

• Computer modelling method that uses lattices and
  discrete state rules to model time dependent
  processes a way to animate things
• No differential equations to solve, easy to
  calculate, more phenomenological
• Simple unit behavior -gt complex group behavior
• Used to model fluid flow, percolation, reaction
  diffusion, traffic flow, pheromone tracking,
  predator-prey models, ecology, social nets
• Scales from 10-12 to 1012

54
Cellular Automata
Can be extended to 3D lattice
55
Succinate Production
Observed Predicted (SimCell)
56
Glycerol Consumption
Observed Predicted (SimCell)
57
Metabolic Profiling The Possibilities

• Genetic Disease Tests
• Nutritional Analysis
• Clinical Blood Analysis
• Clinical Urinalysis
• Cholesterol Testing
• Drug Compliance
• Dialysis Monitoring
• MRS and fMRI

• Toxicology Testing
• Clinical Trial Testing
• Fermentation Monitoring
• Food Beverage Tests
• Nutraceutical Analysis
• Drug Phenotyping
• Water Quality Testing
• Petrochemical Analysis

58
Metabolic Profiling and Drug Toxicology
Principal Component Analysis
59
Genetic Disease Testing
60
140 Detectable Conditions

• Adenine Phosphoribosyltransferase Deficency
• Adenylosuccinase Deficiency
• Alcaptonuria
• a-Aminoadipic Aciduria
• b-Aminoisobutyric Aciduria
• a-Aminoketoadipic Aciduria
• Anorexia Nervosa
• Argininemia
• Argininosuccinic Aciduria
• Aspartylglycosaminuria
• Asphyxia
• Biopterin Disorders
• Biotin-responsive Multiple Carboxylase Deficiency
• Canavans Disease
• Carcinoid Syndrome
• Carnosinemia
• Cerebrotendinous Xanthomatosis/sterol
  27-hydroxylaseDeficiency
• Citrullinemia
• Cystathioninemia

• Dicarboxylic Aminoaciduria
• Dichloromethane Ingestion
• Dihydrolipoyl Dehydrogenase Deficiency
• Dihydropyrimidine Dehydrogenase Deficiency
• Dimethylglycine Dehydrogenase Deficiency
• Essential Fructosuria
• Ethanolaminosis
• Ethylmalonic Aciduria
• Familial Iminoglycinuria
• Fanconis Syndrome
• Folate Disorder
• Fructose Intolerance
• Fulminant Hepatitis
• Fumarase Deficiency
• Galactosemia
• Glucoglycinuria
• Glutaric Aciduria Types 1 2
• Glutathionuria
• Glyceroluria (GKD)

• Histidinemia
• Histidinuria
• Homocystinsufonuria
• Homocystinuria
• 4-Hydroxybutyric Aciduria
• 2-Hydroxyglutaric Aciduria
• Hydroxykynureninuria
• Hydroxylysinemia
• Hydroxylysinuria
• 3-Hydroxy-3-methylglutaric Aciduria
• 3-Hydroxy-3-methylglutaryl-Co A Lyase Deficiency
• Hydroxyprolinemia
• Hyperalaninemia
• Hyperargininemia (Argininemia)
• Hyperglycinuria
• Hyperleucine-Isoleucinemia
• Hyperlysinemia
• Hyperornithinemia
• Hyperornithinemia-Hyperammonemia-Homocitrullinuria
  Syndrome (HHH)

61
Applications in Clinical Analysis

• 96 sensitivity and 100 specificity in ID of
  abnormal from normal by metabolite concentrations
• 95.5 sensitivity and 92.4 specificity in ID of
  disease or condition by characteristic metabolite
  concentrations
• 120 sec per sample

• 14 propionic acidemia
• 11 methylmalonic aciduria
• 11 cystinuria
• 6 alkaptonuria
• 4 glutaric aciduria I
• 3 pyruvate decarboxylase deficiency
• 3 ketosis
• 3 Hartnup disorder
• 3 cystinosis
• 3 neuroblastoma
• 3 phenylketonuria
• 3 ethanol toxicity
• 3 glycerol kinase deficiency
• 3 HMG CoA lyase deficiency
• 2 carbamoyl PO4 synthetase deficiency

Clinical Chemistry 47, 1918-1921 (2001).
62
Applications in Cancer
Acetic Acid Betaine Carnitine Citric
Acid Creatinine Dimethylglycine Dimethylamine Hipp
ulric Acid Lactic Acid Succinic
Acid Trimethylamine Trimn-N-Oxide Urea Lactose Sub
eric Acid Sebacic Acid Homovanillic
Acid Threonine Alanine Glycine Glucose
Normal Below Normal Above Norrmal Absent
Patient 1 Patient 2 Patient 3 Patient 4 Patient
5 Patient 6 Patient 7 Patient 8 Patient 9 Patient
10 Patient 11 Patient 12 Patient 13 Patient
14 Patient 15
Metabolic Microarray - 35 min.
63
Applications in Screening Plant Metabolism
AgriGenomics
64
NMR Metabolomics

• Rapid, robust, largely automatic
• Allows real time monitoring of metabolite fluxes
  as low as 1 uM
• Allows rapid ID of common and unusual metabolites
• Can be applied to chemical shift imaging
• But
• Is it sensitive enough?
• Need to expand metabolite database

65
Improving Sensitivity

• Higher fields (2x)
• Selective NOE enhancement (1.5x)
• Selective decoupling (3x)
• 2-3X longer acquisition (1.6x)
• 5-10X larger volume (5x)
• Chili-probe technology (3x)
• Advanced signal processing (2x?)

66
Expanding the Metabolomic Database

• Human Metabolome Project
• 7.25 million Genome Canada project officially
  launched Dec. 1
• 2 Key outputs
• Electronic database (HMD) of metabolites, chem.
  properties, spectra and pathways
• Freezer full of 1400 metabolites that are either
  isolated, synthesized or purchased

67
Future Challenges

• Completing the human metabolome and expanding the
  library of cmpds so that the spectral ID software
  (Eclipse) is more robust and more widely
  applicable
• Developing improved software or techniques to
  interpret both NMR and MS (or MS/MS) data for
  metabolite ID and classification

68
Future Challenges

• Developing software tools to interpret, visualize
  and predict metabolic outcomes due to genetic
  perturbations
• Developing more robust spectral deconvolution
  software that handles baseline distortion and
  peak position variability
• Developing software to interpret metabolic
  microarray data in terms of disease or phenotype
  identification

69
Thanks to...