|

faculty of mathematicsand natural sciences

Generation and use of metabolome data

Matthias HeinemannMatthias HeinemannMolecular Systems Biology lab

2

Where is Groningen...?

Groningen

3

• PhD student? Postdoc?• How many years into your PhD?• Bio-background? Computational background?• Working with yeast?• Signaling? Metabolism? Regulation?• Systems Biology?• Academic career? Industrial career?

Who are you?

4

Who am I?University of Stuttgart, Germany

Diploma in Environmental Engineering

University of Western Ontario, CanadaExchange year

Diploma thesis, Bremen, Germany

RWTH Aachen University, Germany, PhD in Biochemical Engineering

ETH Zurich, SwitzerlandPostdoc in Bioprocess Lab

ETH Zurich, SwitzerlandGroup leader at the Institute of Molecular Systems Biology

University of Groningen, NetherlandsProfessor for Molecular Systems Biology

5

Systems biology

Biological knowledge/insight (components, interactions, …)

Insights from molecular biology/biochemistry research

Transcriptomics Proteomics Metabolomics FluxomicsGenome-wideexperimental

data

Systemunderstanding

Computationaltools

Mathematicalmodels

Top-

dow

n br

anch

Bot

tom

-up

bran

ch

environment phenotype

6

Metabolism!

7

Metabolism?

• How can metabolite levels be measured?• What can we learn from metabolite levels?• Where is metabolism in the cellular

hierarchy?

8

The plan for the next hours1st hour: How to generate metabolome data?

2nd hour: How to usemetabolome data?

What is the metabolome?

11

What is the metabolome?

Gene

mRNA

protein structure

regulationDire

ct li

nks

DNA: passive library

RNA: ‘passive’ intermediate

Proteins: active workhorses

substrate product

metabolism

Metabolome

Fluxome

= state

= activity

Metabolites (sugars, hormones, vitamins, amino acids .....)No direct link to the genome

Genome

Transcriptome

Proteome

13

What‘s in a metabolome?

What’s in a metabolome?• Microbes ≈ 1000 species• Single plant species ≈ 5000 in Arabidopsis · • All plant species together ≈ 90,000-200,000• humans 2180 endogenous metabolites from text mining …… ! Wishart et al 2007 Nucl Acid Res 35: D521

15

Metabolomics applications= technique to analyze/determine metabolites

In the following, we will only look into the targeted approach!

Profiling:Screening of all detectable by a selected analytical techniqueTypically >50% of peaks have no name

Targeted analysis:Identification and quantification of predefined compounds

Application: - plant traints- disease state, classification (biomarkers)- mutant, condition discrimination

Application:- Systems biology, modeling

17

Typical metabolomics workflow

? ? ? ??

interpretation

G6P/F6PF16PDHAP/G3P2PG/3PGPEPPYRR5P/RU5P/X5P6PGLATPADP

Measuredconcentrations (mM)

0.481.110.674.041.781.930.150.631.100.78

19

Challenges for metabolomics in general Metabolic turnover rates on the

order of seconds requires fast cell processing

A B C

20

Challenges for metabolomics in general

t=0.1-1min t ≈ minutest≤1s t≈ hourst≤1min

-Nucleotides-Redox

cofactors

-Central carbon metabolites

-Secondary metabolites

-Free aminoacids

-Storage carbohydrates

-Proteins

Turnover times of metabolic pools in the cell

21

Challenges for metabolomics in general Metabolic turnover rates on the

order of seconds requires fast cell processing

Metabolome is complicated to analyze number of compounds chemical diversity (+ similarity) small molecular weights large concentration differences A B C

22

Challenge: Diversity & Similarity

From ionic inorganic species to hydrophilic carbohydrates and sophisticated secondary natural products to hydrophobic lipids.

In each class, highly similar compounds exist (diastereoisomers)

25

Typical metabolomics workflow

? ? ? ??

G6P/F6PF16PDHAP/G3P2PG/3PGPEPPYRR5P/RU5P/X5P6PGLATPADP

Measuredconcentrations (mM)

0.481.110.674.041.781.930.150.631.100.78

26

QuenchingGoal: Block metabolism before further biotic changes occur

Villa

s-B

oas

2005

MS

Rev

24:

613

Quenching

Heating 80-100 °C protein denaturation

instantaneous arrest of any

metabolic activity

29

Extraction

Goals• Extraction metabolites from cells with ideally 100% recovery• Remove disturbing agents (to avoid interference with separation (salts,

solvents,...) and with detection (lipids, proteins,...)

Technique• Multiple rounds of solvent extraction (boiling ethanol)

Has to be adapted to cells, analytes, and matrix etc.Important: validation for losses and uneven extraction efficiency

Extraction

breaking open the cells, denaturing

protein and extracting the metabolites

30

Analysis with enzymatic assays

Quenching Extraction

Bergmeyer et al. (eds.), Methods of enzymatic analysis. Verlag Chemie, Weinheim

enzyme assays metabolite concentrations

Disadvantages:- Tedious experimental procedure for each metabolite- You only ‚find‘ the metabolites you are looking for

Calls for genericmetabolite detection method!

G6P/F6PF16PDHAP/G3P2PG/3PGPEPPYRR5P/RU5P/X5P6PGLATPADP

Measuredconcentrations (mM)

0.481.110.674.041.781.930.150.631.100.78

31

Which detection method?

Quenching Extraction

32

Solution ...

Separation DetectionQuenching Extraction

chromatography mass spectrometryUV, NMR, ...

33

SeparationGoals• discriminate between analytes based on

physico-chemical properties > resolve in time

• reduce mix complexity before detection > decrease interferences> increase sensitivity

3 major alternatives• Gas chromatography (GC)• Liquid chromatography (LC)• Capillary electrophoresis (CE)

Separation depends on detection and biological question!

There is not a one-fit-all separation method!

Separation

34

Chromatographic separation → characteristic retention time

Separation - chromatogram Separation

36

Detection - principles

• Spectrum– UV – VIS – IR absorption,

fluorescence

• Molecular weight (mass)• Charge, redox potential• Nuclear magnetic resonance• Raman spectroscopy

Main strength of:MS: Sensitivity, speed, flexible separation, cost (per sample), compound identificationNMR: Quantitative, compound identification, non-destructive

Detection relies on physico-chemical properties of analytes, e.g.

Hollywood et al 2006 Proteomics 6: 4716

In metabolomics, detection must contribute to separation.Very often, broad-scope method are very sensitive to the environment/matrix.

Detection

38

Mass spec in a nutshell Detection

Quadrupole

Time-of-flight

Electrosprayionization

ion with instable

trajectory

ion with stable

trajectory

to detect

or

ion beam

39

Mass spec in a nutshell Detection

Tandem (MS-MS) mass spectrometers are instruments that have more than one analyser and so can be used for structural and sequencing studies.

Two, three and four analysers have all been incorporated into commercially available tandem instruments, and the analysers do not necessarily have to be of the same type, in which case the instrument is a hybrid one.

More popular tandem mass spectrometers include those of the quadrupole-quadrupole, magnetic sector-quadrupole , and more recently, the quadrupole-time-of-flight geometries.

43

masses

time

intensity

chromatogram

time

Combining chromatography and mass spec2 dimensional separation

54

Is there a one-fits-all method?Platform

Method met

hoxy

+ T

MS

met

hoxy

+ T

BD

MS

coat

pol

yE32

3co

at p

olyb

rene

bare

d / F

Aba

red

/ PA

RP

-C18

IP-R

P (D

AP

)IP

-RP

(TB

A)

HIL

IC (Z

WIT

TER

ION

IC)

HIL

IC (A

MIN

O)

NP

DIA

MO

ND

Suitable for Profiling Sugars Sugar-P (glycolysis + PPP) Amino acids Organic acids (TCA)

CoA derivatives Nucleotides (Energy) NAD(P)X (Redox)

Purines Pyrimidines

CEGC LCSystematic test- Mix of ~100 analytes (90% anions)- Calibration curves in water and with

13C-yeast extract

Qualify:CoverageChromatographyProblems with biomass Chromatographic problems arisingDetectionsSuppression/amplificationMultimers/adductsRobustnessThroughput

Büscher et al.

55

...

From spectra to results

2D-Spectra[time x mass x intensity] Results

Feature findingPeak findingDeisotopingDeclustering

IntegrationDenoisingSmoothingBaseline subtractionIntegrationNormalization (time, int.)

IdentificationPeak deconvolutionComponent recognition Database search (MS, MSn)Alignment of several runs

QuantitationIS, QCCalibration with standards

Compound XYConcentration

Peak YZArea

time

mol

ecul

ar w

eigh

tDetection

56

LC or GCChromatogram(50-1000 species)

Measured mass spectrum

Library Match: Serine

Calibration for Serine 362

peak

are

a

concentration

Absolute [c] of serine

Sample

Measured mass spectrum

No match in spectral library!

Relative peak area ofunknown compound

measured

library

Quantification

Quantification

Separation

Detection

Quantification

62

Overview: Typical metabolomics workflow

Quenching Extraction Separation QuantificationDetection

gas-chromatographycapillary-electrophoresisliquid-chromatography

mass-spectrometry

instantaneous arrest of any

metabolic activity

interpretation

breaking open the cells,

denaturing protein and

extracting the metabolites

68

Is it possible to measure everything?Microbial metabolomics: Toward a platform with full metabolome coverage, Mariët J. van der Werf, Karin M. Overkamp, Bas Muilwijk, Leon Coulier and Thomas Hankemeier, Anal Biochem 2007

Achieving metabolome data with satisfactory coverage is a formidable challenge in metabolomicsbecause metabolites are a chemically highly diverse group of compounds. Here we present a strategy for the development of an advanced analytical platform that allows the comprehensive analysis of microbial metabolomes. Our approach started with in silico metabolome information from three microorganisms—Escherichia coli, Bacillus subtilis, and Saccharomyces cerevisiae—and resulted in a list of 905 different metabolites. Subsequently, these metabolites were classified based on their physicochemical properties, followed by the development of complementary gas chromatography–mass spectrometry and liquid chromatography–mass spectrometry methods, each of which analyzes different metabolite classes. This metabolomicsplatform, consisting of six different analytical methods, was applied for the analysis of the metabolites for which commercial standards could be purchased (399 compounds). Of these 399 metabolites, 380 could be analyzed with the platform. To demonstrate the potential of this metabolomics platform, we report on its application to the analysis of the metabolomecomposition of mid-logarithmic E. coli cells grown on a mineral salts medium using glucose as the carbon source. Of the 431 peaks detected, 235 (=176 unique metabolites) could be identified. These include 61 metabolites that were not previously identified or annotated in existing E. coli databases.

Out of the 176 detectable/identifiable, 90% can be quantified with 2 (3) platforms only.

72

Ongoing research and outlook

• Improve sensitivity• Improve coverage (separation, extraction)• Improve quantitation• Improve precision• Improve identification of unknown compounds• Improve throughput

79

Cell populations are often heterogeneousSources for heterogeneity Genetic differences Different microenvironments Different cell cycle stages ....

80

Metabolic changes during cell cycle

MurrayDB, BeckmannM, KitanoH: Regulation of yeast oscillatory dynamics. Proc NatlAcadSciUSA 2007, 104:2241-2246.

81

Metabolic changes during cell age

Lesur I,Campbell JL: The transcriptome of prematurely aging yeast cells is similar to that of telomerase-deficient cells. Mol Biol Cell 2004, 15:1297-1312.

Lin SS,Manchester JK,Gordon JI: Enhanced gluconeogenesis and increased energy storage as hallmarks of aging in Saccharomyces cerevisiae. J BiolChem 2001, 276:36000-36007.

Koc A,Gasch AP,Rutherford JC,Kim HY,Gladyshev VN: Methionine sulfoxide reductase regulation of yeast lifespan reveals reactive oxygen species-dependent and-independent components of aging. Proc NatlAcadSciUSA 2004, 101:7999-8004.

Replicative aging

82

Cell populations are often heterogeneousSources for heterogeneity Genetic differences Different microenvironments Different cell cycle stages Different cell ages ....

Another source of heterogeneity:stochasticity-induced phenotypic heterogeneity

low copy numbers

Stochastic fluctuation of intracellularmolecules (‘noise‘)

Positive feedback circuits

yx

Kussell & Leibler, Science, 2005 - Balaban et al., Science, 2004 -Ozbudak et al., Nature, 2004 Elowitz et al., Science, 2002 -Thattai & van Oudenaarden, PNAS, 2001

Single-cell-analysis techniques required for Systems Biology(Most current “omics“ techniques average over populations)

86

Single-cell measurement on different levels

DNA

mRNA

proteins enzymes

metabolite metabolite

Single-cell DNA analysis / genomics(lysis, dielectrophores. (DEP), chromatogr., electrophoresis, PCR, sequencing)Quake group, Stanford (E. coli, microbes), Sturm group, Princeton (E.coli)

Single-cell transcription analysis / transcriptomics(lysis, separation and isolation (e.g. beads), cDNA, fluorescence)Quake group, Stanford (mouse fibrobl., hum. stem cells), Potier group, Paris (mouse neurons)

Single-cell protein analysis / proteomics(lysis, electrophoresis, fluorescence) Zare group, Stanford (insect cells, bacteria), Dovichi group, U Washington (cancer cells), Xie group, Harvard (E. coli), Sweedler group, Urbana, IL (neurons, MS) , Ramsey group, UNC, microfluidics/MS,Toner group, MIT (lymphoblasts), Cooper group, Glasgow, UK (cancer), Anselmetti group, Bielefeld (insects, sf9), Single Cell Proteomics and Lipidomics project, UCL, London (started 2006)

Single-cell metabolite analysis / metabolomics(electrophoresis, partly fluorescence, MS)Masujima group, Hiroshima University, mast cells,pipette, suck out cell, ESI-MS, Fang group, Northeastern Univ. Shenjang, China (oxygen, glutathione), Netherlands Metabolomics Center (SCM projected)

87

Single-cell metabolomics

Single-cell metabolomics challenges large number of compounds chemical diversity (+ similarity) low metabolite quantities low molecular weights: fluorescent

labeling practically impossible fast metabolic turnover rates

(< 1 sec)

Single-cell metabolite analysis / metabolomics(electrophoresis, partly fluoresence) Masujima group, Hiroshima University, mast cells, pipette, suck out cell, ESI-MS, Fang group, Northeastern Univ. Shenjang, China (oxygen, glutathione), Netherlands Metabolomics Center (SCM projected)

DNA

mRNA

proteins enzymes

metabolite metabolite

88

Challenge of low metabolite abundance

Challenge: sampling and processing of single cells without loss

Challenge: transfer to detector without loss

1 attomol = 10-18 mol

attomoles number of molecules

attomoles number of molecules

ATP 1 600'000 140 85'000'000

F16BP 2 1'300'000 10 7'500'000G6P 2 1'300'000 200 120'000'000PEP 1.5 900'000 90 50'000'000PYR 1.5 900'000 5 3'000'000

E. coli S. cerevisiae

89

Measurement of neurotransmitters and amino acids from single neurons

R.T. Kennedy et al., Science 246 (1989) 57.

Voltammetric detection (neurotransmitters)

Previous work: Kennedy

Fluorescence detection (labelled amino acids)

Separation by capillary electrophoresis

90

Single cell metabolomics (SCM)

Cur

rent

Opi

nion

in B

iote

chno

logy

, 201

0, 2

2: 2

6-32

91

Mass-spec based technologies for SCM

92

M. Shimizu et al., Anal. Sci. 19 (2003) 49.

Aspiration of a single mast cells(mast cells = rich in histamine and heparin)

Embedded in DHB MALDI matrix

Development of histaminesignal during maturation ofa mast cell

Previous work: Masujima

93

Mass-spec based technologies for SCM

94

Microfluidics + MALDI-MS

Cooling

Cell suspensionCell

sizingCell

focusingCelllysis

Cellfragment

separation

Chip-to-MS

interface

Massspectrometer

Cell culture Microfluidic cell handling Chip-to-MS MS

highly efficient interface to mass-spectrometer

microfluidic platform for processing single cells

highly sensitive mass-spectrometric method

97

Microfluidic platformsSet-up for prototyping:Three-layerglass/PDMS hybridmicrodevice forprototyping

Set-up for final chip:Glass/SU-8/glass sandwichwith opposing electrodes

98

Functions of the microfluidic device

differentialimpedance

measurement

electricfield

pulses

MS

99

Cell sizing: differential impedance measurementDifferential impedance measurement with opposing electrodes to achieve evenly distributed E-field

Amplitude and phase can be correlated with cell size.

electronics and software for signal acquisition, amplification and analysis

peak

am

plitu

de a

.u.

velocity of the beads

4 µm

6 µm

8 µm polystyrene beads

101

Highly sensitive mass spectrometric method

Sample+Matrix

Premix

SampleMatrix

Layered

MALDI-TOF (negative mode) with 9-aminoacridine matrix thin layer preparation: analyte deposited on top of a thin matrix layer deposition of picoliter quantities with a piezoelectric dispenser

Development path with many dead-ends:

Mass spectrometry or alternative detection?Negative or positive mode?MALDI or ESI?Regular MALDI or matrix-free system?Optimized or all-purpose matrix?Layered sample deposition or sample & matrix pre-mixed?

102

Analyte is co-crystallized with an excess of a solid matrix material. This mixture isdeposited on a substrate. Laser irradiation evaporates material and creates a plumeconsisting of 99.9% neutrals and only 0.1% ions, both positive and negative.

Substrate SubstrateSubstrate

Analyte

MatrixLaser

++

-

-

Matrix-Assisted Laser Desorption/ Ionization (MALDI)

106

Single cell quantities of cell extract

4 nL ~ 6 cells

0.39 nL ~ 0.6 cells

A. Amantonico, J. Y. Oh, J. Sobek, M. Heinemann, R. Zenobi, Angewandte Chemie Int. Ed., 2008

Strategy to analyze

metabolites contained in single yeast

cells

140 attomol ATP

107

Interface between microfluidic chip and MS

x/y-stage with linear motors forcontrolled sample deposition

"Capillary-writing“ sample deposition

chip with outlet needle

plate with MALDI matrix stripes

108

Interface between microfluidic chip and MS"Capillary-writing“ sample deposition

109

Mass-spec based technologies for SCM

110

Micro-arrays for SCM

Urban et. al. Lab Chip. 2010 Dec 7;10(23):3206-9.