In vivo NMR

1
The Nuts and Bolts of NMR-Based Metabolomics
CEHS Conference 25th August 2003
Dr. Mark R. Viant Department of Environmental
Toxicology University of California, Davis
2
What is metabolomics and why do we care?
3
Post-genomic Era of Biology
Genome
Gene expression (mRNA)
Metabolism
Proteins
4
Post-genomic Era of Biology
Genome
Transcriptomics (Microarrays)
Metabolomics
Gene expression (mRNA)
Genomics
Metabolism
Proteins
Proteomics
5
Post-genomic Era of Biology
Genotype
Genome
Transcriptomics (Microarrays)
Metabolomics
Gene expression (mRNA)
Genomics
Metabolism
Proteins
FunctionalMolecular Phenotype
Proteomics
6
Post-genomic Era of Biology
Genotype
Genome
Transcriptomics (Microarrays)
Metabolomics
Gene expression (mRNA)
Genomics
Metabolism
Proteins
Environmental stressors
FunctionalMolecular Phenotype
Proteomics
7

• Metabolomics is a powerful approach that takes a
  snap-shot of the hundreds of endogenous organic
  molecules that comprise the bodys metabolic
  system.
• Enables changes in metabolism to be monitored
• throughout exposure to drugs and toxicants
• during progression of a disease
• throughout organism development
• following genetic modification
• during nutritional intervention.
• Metabolomics metabonomics

8
Number of metabo_omics publications
70
60
50
40
Number of publications
30
20
10
0
1998
1999
2000
2001
2002
2003
Half yearly periods
9
Overview of experimental approach
10
Tissue or biofluid sample
11
Tissue or biofluid sample
1. Mass spectrometry 2. 1H NMR spectroscopy
Bioanalytical tools
Measure the metabolite profile
12
Tissue or biofluid sample
1. Mass spectrometry 2. 1H NMR spectroscopy
Bioanalytical tools
Measure the metabolite profile
Statistical bioinformatic tools
Treat profile as fingerprint for classification
purposes
(applied/clinical)
13
Tissue or biofluid sample
1. Mass spectrometry 2. 1H NMR spectroscopy
Bioanalytical tools
Measure the metabolite profile
Statistical bioinformatic tools
Treat profile as fingerprint for classification
purposes
Explore profile to gain mechanistic insight into
the biological response
(applied/clinical)
(basic research)
14
Overview of Presentation

• Red abalone and withering syndrome
• See poster
• Embryogenesis and toxicology of Japanese medaka
• Illustrate all aspects of basic NMR approach
• Concept of metabolic trajectories
• NMR methods for reducing spectral congestion
• See poster
• Nutritional studies using plasma samples
• Sample preparation
• Genetic algorithms for correcting pH shift
  artifacts

15
Embryogenesis and toxicology of medaka
Chris Pincetich, Dr. Ron Tjeerdema Department of
Environmental Toxicology, UC Davis
16
Embryogenesis study design

• We hypothesized that NMR-based metabolomics will
  provide a powerful, rapid and inexpensive method
  for characterizing toxicant-induced metabolic
  perturbations during embryogenesis.
• Initial aim Characterize the metabolic changes
  occurring throughout the 8 days of normal
  embryogenesis in Japanese medaka.

17
Experimental approach

• Medaka embryos developed at 25 C
• Froze groups of 200 eggs on each day of
  development

• Minimal sample preparation
• Perchloric acid extractions of whole eggs

1H NMR spectroscopy
Spectral preprocessing
Multivariate statistical analysis
18
NMR Hardware and Experiments
500 MHz and 600 MHz NMR spectrometers available
at UCD NMR Facility
1-D NMR methods - for rapid analysis of
metabolite profiles Single pulse 1H NMR
sequence CPMG 1H spin-echo sequence
projections from J-resolved spectra
2-D NMR methods - for confirmation of peak
assignments 1H-1H homonuclear correlation
spectroscopy (COSY) 1H-13C heteronuclear single
quantum coherence (HSQC)
19
Which metabolites can be observed by NMR?

• Low molecular weight organic metabolites
• Amino acids
• Organic acids and bases
• Nucleotides
• Carbohydrates
• Osmolytes
• Lipids (broad non-specific resonances)

20
1H NMR spectrum of medaka embryo extracts
Organic acids,
Carbohydrates,
Nucleotides,
e.g. succinate
e.g. ribosyl moiety
e.g. ATP
Amino acids,
e.g. tyrosine
1
2
3
4
5
6
7
8
9
chemical shift (ppm)
21
Statistical bioinformatics(Parul Purohits talk)

• Spectral preprocessing
• Transform NMR data into format for analysis
• Custom written MATLAB software
• Multivariate statistical analyses
• Summarize the similarities and differences
  between the metabolic fingerprints of the samples
• Identify potential metabolic biomarkers
• Principal components analysis (PCA)

22
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
Samples with similar metabolite profiles group
together
PC2 score
Day 1
Fertilization
PC1 score
23
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
PC2 score
2
Day 1
Fertilization
PC1 score
24
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
3
PC2 score
2
Day 1
Fertilization
PC1 score
25
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
4
3
PC2 score
2
Day 1
Fertilization
PC1 score
26
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
5
4
3
PC2 score
2
Day 1
Fertilization
PC1 score
27
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
5
4
6
3
PC2 score
2
Day 1
Fertilization
PC1 score
28
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
5
4
6
3
7
PC2 score
2
Day 1
Fertilization
PC1 score
29
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
5
4
6
3
7
PC2 score
2
8
Day 1
Hatch
Fertilization
PC1 score
30
PCA scores plot Summarizes changes in
NMR-visible metabolome throughout embryogenesis
5
4
6
3
7
PC2 score
2
8
Developmental trajectory
Day 1
Fertilization
Hatch
PC1 score
31
PCA loads plot Identifies specific metabolites
that change during embryogenesis
0.4
Tyrosine
Creatine
Histidine
Alanine
Lactate
ATP
0.2
PC 1 loadings
0.0
-0.2
Leucine
Citrate
-0.4
1
2
3
4
5
6
7
8
9
10
Chemical shift (ppm)
32
Developmental toxicity of trichloroethylene (TCE)
in Japanese medaka

• Expose medaka embryos to TCE throughout
  embryogenesis.
• Preserved replicates of 100 eggs on day 7 of
  development.

33
PCA scores plot Normal embryogenesis
5
4
6
3
7
PC2 score
2
8
Day 1
PC1 score
34
PCA scores plot Dose-dependent effects of TCE on
medaka metabolome
5
4
6
3
7
PC2 score
2
Day 7 controls
8
Day 1
PC1 score
35
PCA scores plot Dose-dependent effects of TCE on
medaka metabolome
5
4
6
3
7
PC2 score
2
3 ppm TCE
Day 7 controls
8
Day 1
PC1 score
36
PCA scores plot Dose-dependent effects of TCE on
medaka metabolome
5
4
6
3
7
PC2 score
2
Trajectory?
3 ppm TCE
46 ppm TCE
Day 7 controls
8
Day 1
PC1 score
37
PCA loads plot Identifies metabolic biomarker
profile of TCE toxicity
Hydrophobic amino acids
0.4
Lactate
Glucose
0.2
PC1 - PC2 loadings
0.0
ATP
-0.2
Glutamate
Creatine
-0.4
1
2
3
4
5
6
7
8
9
Chemical shift (ppm)
38
Normal developmental trajectory
Normal development
PC2 score
Hatch
Fertilization
PC1 score
39
Perturbations to normal developmental trajectory
Normal development
PC2 score
stage specific toxicity identified for targeted
gene expression studies
Permanent toxicant-induced perturbation
PC1 score
40
Perturbations to normal developmental trajectory
Normal development
Transienttoxicant-induced perturbation
PC2 score
PC1 score
41
Application to nutritional assessment
Normal development
Onset of obesity?
PC2 score
Targeted and personalized nutritional intervention
PC1 score
42
NMR methods for reducing spectral congestion
Single pulse 1-D 1H experiment
Skyline projection of 2-D J-resolved
spectrum (p-JRES)
43
PCA scores plot of medaka embryogenesis from
analysis of 1-D 1H spectra
Scores plot of medaka embryogenesis from analysis
of p-JRES spectrum
44
PCA loads plot from analysis of 1-D 1H spectra
Loads plot from analysis of p-JRES spectra
Metabolic heat map showing changes in metabolite
levels during embryogenesis
45
NMR analysis of plasma samples Zn nutritional
fortification of Peruvian children
Dr. Ken Brown et al. Program in International
Nutrition, UC Davis
46
Sample preparation for NMR analysis

• Biofluids (plasma, urine)
• 200 mL sample add 2H2O/phosphate
  centrifuge NMR
• buffer/TMSP std
• Simple, rapid, inexpensive
• More susceptible to pH variations than tissue
  extracts?

47
NMR peaks susceptible to pH-induced frequency
shifts
48
Before
After
49
Before
After
50
Experimental design
Zn fortification
Peruvian children
no fortification (control)
Plasma samples obtained at t 0, 18 days, 8 weeks
51
PCA scores plot of all 81 plasma samples
1
383
0.8
321
112
671
0.6
443
092
563
181
182
462
441
0.4
093
491
351
Scores on PC 2 (24.08)
261
511
0.2
371
291
263
672
322
042
471
051
113
303
151
301
152
251
212
293
292
0
352
041
591
372
562
373
513
091
381
593
402
431
141
053
213
253
422
211
473
052
442
492
401
592
423
403
323
-0.2
353
382
461
183
561
512
Real world complexity!
302
673
421
111
432
252
243
143
242
463
153
142
493
-0.4
433
-0.6
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Scores on PC 1 (34.87)
Bioinformatic challenge to extract useful
metabolic data
52
Strengths of NMR-based metabolomic approach

• Minimal sample preparation.
• High throughput analysis (potentially 200
  samples/day).
• Inexpensive per-sample cost.
• Robust, semi-quantitative (fully?) analysis.
• Non-destructive analysis.
• Unbiased identification of 1H-containing
  metabolites.
• 2-D NMR methods for metabolite identification.
• Ideal for screening samples, followed by more
  sensitive MS analysis of most interesting
  specimens.

53
Acknowledgments
Statistical Bioinformatics David Rocke David
Woodruff Parul Purohit Jinjin Liang
Aquatic Toxicology Group Chris Pincetich Eric
Rosenblum Ron Tjeerdema
Nutritional Studies Ken Brown Marjorie Haskell
NMR Spectroscopy Jake Bundy Jeff de Ropp
Funding UC Toxic Substances Research and
Teaching Program (Associate Directors
Discretionary Support from Marion Miller) UC
Davis NMR Facility Award UC Davis Clinical
Nutrition Research Unit UC Davis Center for
Environmental Health Sciences