www.lipidmaps.org

Other LIPID MAPS Sphingolipid Core members:Mass spectrometry Cell biology

Al Merrill Samuel Kelly Elaine WangChris Haynes Ying LiuRebecca Shaner Kacee SimsYanfeng Chen Amin Momin

LIPID MAPS Lipidomics WorkshopEB2009

April 19, 2009

Lipidomic analysis of sphingolipids(and precursor fatty acyl-CoA’s)

M. Cameron Sullards

Schools of Chemistry, Biochemistry and Biology &Petit Institute for Bioengineering and Bioscience

Georgia Institute of Technology

Outline:A. Brief introduction to the lipid class: nomenclature & range of

compounds to analyzeB. Sample preparation issues: solvents, chromatography, recovery,

reproducibilityC. Compound identification: Characteristic fragmentations; MS/MS

and MSn (LC for isomers and isobars, etc.)D. Quantitation: MRM, Internal standards, etc.E. Data analysis/visualization: LIMS, Website, otherF. Remaining challenges and opportunitiesG. Discoveries from sphingolipidomic analysis thus farH. Comparison of Lipid MAPS methods with others in the literature

Backbone variation

Sphingoid base:

Sphinganine (d18:0)

4-Hydroxysphinganine(phytosphingosine) (t18:0)

Ceramide:

Shown: N-palmitoylsphingosine (d18:1/16:0)Other fatty acids- typically C16-C26 0-1 double bond sometimes α- or ω-hydroxy

Lactosylceramide (LacCer)

GlcCer

Gal

GalNAc

NeuAc

GalGlc

Neu5Acα2-3(Galβ1-3GalNAcβ1-4)Galβ1-4Glc1-1CerII3Neu5AcGg4Cer

III

III

IV

GM1GM3

A. Brief introduction to the lipid class: nomenclature & range of compounds to analyze

D-erythro-sphingosine (d18:1)

Headgroup variation

Phosphosphingolipids:

-OP(O2-)O-choline, Sphingomyelin

Glycosphingolipids:

Glc, Gal, Lac, Sulfatides…>400; see www.lipidmap.org; www.sphingomap.org

Backbone variation

Sphingoid base:

Sphinganine (d18:0)

4-Hydroxysphinganine(phytosphingosine) (t18:0)

Ceramide:

Shown: N-palmitoylsphingosine (d18:1/16:0)Other fatty acids- typically C16-C26 0-1 double bond sometimes α- or ω-hydroxy

Lactosylceramide (LacCer)

GlcCer

Gal

GalNAc

NeuAc

GalGlc

Neu5Acα2-3(Galβ1-3GalNAcβ1-4)Galβ1-4Glc1-1CerII3Neu5AcGg4Cer

III

III

IV

GM1GM3

A. Brief introduction to the lipid class: nomenclature & range of compounds to analyze

D-erythro-sphingosine (d18:1)

Headgroup variation

Phosphosphingolipids:

-OP(O2-)O-choline, Sphingomyelin

Glycosphingolipids:

Glc, Gal, Lac, Sulfatides…>400; see www.lipidmap.org; www.sphingomap.org

Novel sphingoid bases

1-Deoxysphinganine

1-Desoxymethylsphinganine

Zitomer et al. JBC 284:4786-95 (2009)Pruett et al. J. Lipid Res. 49:1621-39 (2008).

SphinSphinGGOOMMAP AP ((www.sphingomap.org www.sphingomap.org & Merrill et al., & Merrill et al., Trends Biochem Sci., 2007))

Shaner et al. J. Lipid Res. (2009)

Validation of extraction recovery (Shaner et al., JLR 2009):

>95% recovered in Extract 1 and 2 (which are pooled for analysis)

Validation of extraction recovery (Shaner et al., JLR 2009):

>95% recovered in Extract 1 and 2 (which are pooled for analysis)

InletSystem

IonSource

Mass Analyzer(Q1, Q2, Q3 or LIT)

Detector

Data System forquantitation by MRM

(Multiple ReactionMonitoring) w/

appropriate internalstandards

Reverse phaseHPLC

Species analyzed to date (in mostmammalian sphingoid base & FA variants):

SM’s, GlcCer, GalCer, LacCer, Sulfatide(quanlitative analysis of globosides andgangliosides)Cer & Cer-PSo, Sa, 4-hydroxy-Sa, 1-DeoxySa, etc.So-P, Sa-P and other derivatives (lysoSM,psychosine, N-methyl-)Plus metabolites labeled with stableisotope precursors

A (Polar, e.g. So, Sa, 4-HO-Sa, SoP, So-Pcholine, Psychosine, Cer-P, etc.)

B (Nonpolar, e.g. Cer, Cer-P, SM, GlcCer, GalCer, LacCer, etc.)

Normal phase HPLC

Sullards, Merrill & coworkersSTKE (2001),

Methods (2005),Meth. Enzymol. (2008) &Shaner et al. JLR (2009)

(Dihydro)ceramides

OHO

OH

OH

O

O

OH

OH

O R'

HN R"

O

OH

Z0

Y0

Z1

Y1

B1

C1

B2C2

0

45

2

1

23

13

04

1,5X2

5

2,5A2

HO

HO

H+OH

NH

OR

O

m/z 184

OH

NH

OPhosphocholine

O

Sphingomyelins

Glycosphingolipids

m/z 264

(266, etc.)

m/z 264

(266, etc.)

Sphingolipid analysis by LC-MS/MS

*Sulfatide from cells treated w/KDO2 Lipid A

d18:0;C18:0Cer

d18:1;C18:0Cer

d18:1;C18:0GlcCer

d18:0;C16:0DHCer

d18:0;C16:0DHSM

d18:0;C16:0

DHCer-P

d18:0;C16:0

GalDHCer

d18:1;C16:0

SM

d18:0;C16:0

GlcDHCer

d18:1;C16:0Cer-P

d18:1;C16:0

1-O-AcylCer

d18:0;C16:0

1-O-AcylDHCer

d18:0;C20:0DHCer

d18:1;C20:0Cer

d18:0;C22:0DHCer

d18:1;C22:0Cer

d18:0;C24:1DHCer

d18:1;C24:1Cer

d18:0;C24:0DHCer

d18:1;C24:0Cer

d18:0;C26:1DHCer

d18:1;C26:1Cer

d18:0;C26:0DHCer

d18:1;C26:0Cer

d18:0;C2:0

DHCer

d18:1

d18:0

d18:1;C2:0Cer

All subspecies have been quantified for these chain lengths, too(only a few are shown; all are at www.lipidmaps.org)

d18:0-P

d18:1-P

Sphinganine (d18:0)

H CH2OHHO

NH3 (+)

CH3(CH2)10CH2

3-Ketosphinganine Sphinganine 1-phosphate (d18:0-P)

Ceramide synthases+ Fatty acyl-CoA

Serine +Palmitoyl-CoA

Approximate number of molecules per RAW cell (million)

d18:1;C16:0GalCer

<0.1

<0.1

<0.1

<0.1

SulfatideLacCer <0.1[<1]

d18:1;C16:0GlcCer

d18:1;C16:0Cer

<0.1

<0.1

11.7

13529.7 87 294 6.3 1.5<0.1<0.1<0.1<0.1<0.1<0.1

23.4 36.9 1.5 0.6

1.8 0.03<0.1

13.5

30.3

378

19.5

1.8

0.03

1023

23.4

0.1

12

<0.1

0.9

0.03

d18:0;C18:0Cer

d18:1;C18:0Cer

d18:1;C18:0GlcCer

d18:0;C16:0DHCer

d18:0;C16:0DHSM

d18:0;C16:0

DHCer-P

d18:0;C16:0

GlcDHCer

d18:1;C16:0Cer

d18:1;C16:0

SM

d18:0;C16:0

GalDHCer

d18:1;C16:0Cer-P

d18:1;C16:0

1-O-AcylCer

d18:0;C16:0

1-O-AcylDHCer

d18:0;C20:0DHCer

d18:1;C20:0Cer

d18:0;C22:0DHCer

d18:1;C22:0Cer

d18:0;C24:1DHCer

d18:1;C24:1Cer

d18:0;C24:0DHCer

d18:1;C24:0Cer

d18:0;C26:1DHCer

d18:1;C26:1Cer

d18:0;C26:0DHCer

d18:1;C26:0Cer

d18:0;C2:0

DHCer

d18:1

d18:0

d18:1;C2:0Cer

Variation in backbone, acyl chain & headgroup specificity

d18:0-P

d18:1-P

Sphinganine (d18:0)

H CH2OHHO

NH3 (+)

CH3(CH2)10CH2

3-Ketosphinganine Sphinganine 1-phosphate (d18:0-P)

Ceramide synthases+ Fatty acyl-CoA

Serine +Palmitoyl-CoA

d18:1;C16:0GalCer

d18:1;C16:0GlcCer

Sulfatide GM3 GM2 GM1 GD1a

Raw cells+KdO2Lipid A

Acidic SphingolipidsInfusion 0.6 mL/h (MeOH)

Prec 290 scan [MCA 60 Scans (3s/scan)]DP = -100 CE = -75 IS = -4500

Raw cells

Next category of compoundsbeing analyzed by “Inside-out”

sphingolipidomics:Gangliosides

Work-flow for analysis of new samples using thisLC-MS/MS Methodology

1.1. Identify structure specific dissociations unique to variousIdentify structure specific dissociations unique to variousclasses (e.g., SM, classes (e.g., SM, GlcCerGlcCer, , GalCerGalCer, , LacCerLacCer, etc.), etc.)

3.3. Optimize ionization and dissociation conditions for all Optimize ionization and dissociation conditions for all species.species.

2.2. Utilize precursor ion and neutral loss scans to identify Utilize precursor ion and neutral loss scans to identify individual headgroup, sphingoid base, and fatty acid individual headgroup, sphingoid base, and fatty acid combinations.combinations.

4.4. Optimize LC as required to minimize ionization suppressionOptimize LC as required to minimize ionization suppressioneffects, and interferences arising effects, and interferences arising from isobaric, isotopic,from isobaric, isotopic,and isomeric species (repeat #3 if necessary).and isomeric species (repeat #3 if necessary).

5.5. Optimize conditions for quantitation via ratio of peak areasOptimize conditions for quantitation via ratio of peak areasvs validated internal standards for all of the species vs validated internal standards for all of the species

present.present.

Rela

tive

Ion

Abu

ndan

ce

650 700 750 800 850

m/z

Q1

671.6

688.6

709.6

Rela

tive

Ion

Abu

ndan

ce

650 700 750 800 850

m/z

(SM)

C16:0

C18:0 C22:0C24:1

Pre 184

Example: Identification of sphingolipid subspecies viaNeutral Loss or Precursor Ion Scans

300 400 500 600 700

C2:0

C24:1

300 400 500 600 700

C2:0

C24:1

Precursor(25 eV)

Precursor(45 eV)

300 400 500 600 700

C2:0C24:1

MRMrela

tive

ion

abun

danc

e

Comparison of ion abundance forceramides of varying chain length

when analyzed under singleionization and dissociation

conditions vs optimized MRM

m/z

D. Quantitation of Sphingolipids

Criteria for selection of internal standards

1.1. Must have the same chemical and physical properties asMust have the same chemical and physical properties asthe the analyte analyte of interest, ideally stable isotope labeled of interest, ideally stable isotope labeled

analogs.analogs.2.2. Should be practical for Should be practical for ““omicomic”” analysis--ianalysis--i.e., cover as many.e., cover as many

subspecies as possible because adding an internal subspecies as possible because adding an internal standard for every standard for every analyte analyte would require 100s to 1000s of would require 100s to 1000s of molecules to be synthesized, added and analyzed, which is toomolecules to be synthesized, added and analyzed, which is tooexpensive, time consuming and possibly analytically impossible.expensive, time consuming and possibly analytically impossible.

LIPID MAPS Sphingolipidomics cocktail (available from AvantiLIPID MAPS Sphingolipidomics cocktail (available from AvantiPolar Lipids): 10 uncommon sphingolipid species that are used toPolar Lipids): 10 uncommon sphingolipid species that are used tospike samples prior to extraction (Walt Shaw)spike samples prior to extraction (Walt Shaw)

LIPID MAPS internal standard cocktail (cont.)For For sphingoid sphingoid bases: odd chain length variants that elute under similarbases: odd chain length variants that elute under similarconditions so there is little ionization or dissociation effects and precursor andconditions so there is little ionization or dissociation effects and precursor andproduct ion masses are slightly shifted.product ion masses are slightly shifted.

For complex sphingolipids: shorter fatty acid chain length variants (C12:0) thatFor complex sphingolipids: shorter fatty acid chain length variants (C12:0) thatco-elute with analytes of interest so there are no ionization effects, and haveco-elute with analytes of interest so there are no ionization effects, and havedifferent precursor ion masses but similar fragmentation when optimized.different precursor ion masses but similar fragmentation when optimized.

d17:1 d17:1 ““sphingosinesphingosine”” and and ““sphingosine 1-phosphatesphingosine 1-phosphate”” homologs homologsd17:0 d17:0 ““sphinganinesphinganine”” and and ““sphinganine 1-phosphatesphinganine 1-phosphate”” homologs homologs

C12-Cer, [C25-Cer], C12-Cer-1-P, C12-GlcCer, C12-LacCer, C12 SMC12-Cer, [C25-Cer], C12-Cer-1-P, C12-GlcCer, C12-LacCer, C12 SM

Also available: Also available: C12-SulfatideC12-SulfatideUnder development:Under development: C12-GM1 and other complex C12-GM1 and other complex glycosphingolipidsglycosphingolipids

C17:1 C17:1 sphingosylphosphocholinesphingosylphosphocholine; ; N-methyl-sphingoid N-methyl-sphingoid basesbases1-deoxysphinganine1-deoxysphinganine

2

4

6

4

6

8

10

2 4 6Sphingoid base or sphingoid base 1-phosphate (log fmol)

Inte

grat

ed a

rea

for a

nalyt

e (lo

g cp

s)

3 5

Shaner et al. JLR (2009)

API 3000

4000 QTrap

So-PC17 So-P

C17 SaC17 SoC17 Sa-PSaSoSa-P

C17 SaC17 SoC17 Sa-PSaSoSa-P

So-PC17 So-P

2 4 6

2

4

6

4

6

8

10

2 4 6

A. API 3000 (d18:1)

B. 4000 QTrap (d18:1) D. 4000 QTrap (d18:0)

C. API 3000 (d18:0)

C12:0C16:0C18:0C24:0C24:1C25:0

C18:0C24:0C24:1

C16:0

3 5 3 5Ceramide (d18:1) or dihydroceramide (d18:0) (log fmol)

Inte

grat

ed a

rea

for a

nalyt

e (lo

g cp

s)

Shaner et al. JLR (2009)

16:0 18:0 20:0 22:0 24:1 24:0 26:1 26:0 0 0.5 1 2 4 8 12 24

0

30

60

(pm

ol C

eram

ide/

ug D

NA

)

Example data from Lipid MAPS experiments:Changes in Cer in RAW 267 cells treated with Kdo2 Lipid A

(see www.lipidmaps.org)

Treatment Control

Timecourse (h)

N-acyl chainsubspecies

Basic LIPID MAPS Protocol

Plate cells

Pre-treatments (if any)

Add agent (LPS, Kdo2 Lipid A)

Incubation

GrowAnalyze

LIPID MAPS

TM

0.5 1 2 4 8 12 24

E. Data analysis/visualization:

C16DHCer

C16Cer

C16DHSM

C16DHGlcCer

C16SM

C16GlcCer

CORE I: Sphingolipids(only C16 shown)

KDO2

Control

Sphinganine

Sphingosine

Sphingosine 1-P

Sphinganine 1-P

C16Cer-P

Biological Replicates +/- SE3-Ketosphinganine (below LOD)

Serine + Palmitoyl-CoA

Data Sets Available at:www.lipidmaps.org

(Eoin Fahy)

F. Remaining challenges (and opportunities)

Discovery of new subspecies (and new functions for known subspecies)--such as 1-deoxy-sphingoid bases

Obtain standards for glycosphingolipids (and new sphingolipids)

Determine how to better visualize changes in abundances in multiple classes of sphingolipids over time.

Develop methods to differentiate appearance/disappearanceof particular subspecies via de novo biosynthesis vs turnover.

RAW 10% time coursew/[13C]palmitate

Ser + PalCoA

3-KetoSa

Sa

GlcDHCer DHCer DHSM

GlcCer Cer SM

C16 C18

C22

C24

C26 0h

r

0.5h

r 1hr 2h

r

4hr

8hr 12hr 24

hr

Con

trol 0

.5hr

Con

trol 1

hr

Con

trol 2

hr

Con

trol 4

hr

Con

trol 8

hr

Con

trol 1

2hr

Con

trol 2

4hr 0

10

20

30

40

50

(pm

ol/u

g D

NA

)

13C DUAL Ceramide (13C in sphingoid base and fatty acid)

C16 C18

C22 C24

C26

0hr

0.5h

r 1hr 2h

r 4hr 8h

r 12hr

24hr

Con

trol 0

.5hr

Con

trol 1

hr

Con

trol 2

hr

Con

trol 4

hr

Con

trol 8

hr

Con

trol 1

2hr

Con

trol 2

4hr 0

51015

20

25

(pm

ol/u

g D

NA

)

13C BASE Ceramide (labeled sphingoid base only)

0.1 mM [13C]palmitate

Analysis of sphingolipidbiosynthesis using stable

isotope labeled precursors

100 200 300 400 500 600 700 800 900 10000

50

100

136.0 261.1

397.0428.0

499.4

(M + H)+1006.4

m/z

NNN

N

NH2

O

OHOPO

OHHO

OP

OP

O

NH

NH

S

CH3H3COH

O

O

OH

O

OHO

O

499.4

261.1

136.0

428.0

To understand sphingolipid biosynthesis one must also know theavailability of the co-substrate fatty acyl-CoA’s.

C16:0-CoA (10 pmol/µL in methanol 10 mM triethylammonium acetate) was infused at 5 µL/min. 4000QTrapparameters: IS = 5400 V, Gs1 = 12 psi, CUR = 10 psi, DP = 180 V, CE = 52 V, CXP = 14.3 V, CAD = Med

404514

2624

286

6 3 8 18 13

187

967

1454

671

236

401

871

1620

2159

0

0.1

0.2

0.3

C14:0 C16:0 C18:1 C18:0 C20:0 C22:0 C24:1 C24:0 C26:0

Molecular Species

pm

ol p

er

10

*6 c

ells

(m

ea

n +

SE

, n

=7)

RAW

MCF7

Fatty acyl-CoA profile of RAW cells(and for comparison, MCF7 cells)

MS/MS ofC16:0-CoA

Haynes et al. J. Lipid Res. (2008)

16:0 18:0 20:0 22:0 24:1 24:0 26:1 26:0 0 0.5 1 2 4 8 12 24

0

30

60

(pm

ol C

eram

ide/

ug D

NA

)

de novo sphingolipid biosynthesis is induced by Kdo2Lipid A (has been correlated with gene array data

showing increases in SPT1 and SPT2 mRNA)

Treatment Control

Timecourse (h)

N-acyl chainsubspecies

Basic LIPID MAPS Protocol

Plate cells

Pre-treatments (if any)

Add agent (LPS, Kdo2 Lipid A)

Incubation

GrowAnalyze

LIPID MAPS

TM

0.5 1 2 4 8 12 24

G. Examples of discoveries from sphingolipidomic analysisthus far

de novo sphingolipid biosynthesis is induced by Kdo2Lipid A has been correlated with gene array data andfor other steps of the pathway, see www.lipidmaps.org)

G. Examples of discoveries from sphingolipidomic analysisthus far

G. Examples of additional (recent) discoveries from applicationof these sphingolipidomic methods

Ceramide synthase inhibition by fumonisin B1 causes accumulation of 1-deoxysphinganine: a novel category of bioactive 1-deoxysphingoid bases and 1-deoxydihydroceramides biosynthesized by mammalian cell lines and animals.(2009) Zitomer NC, Mitchell T, Voss KA, Bondy GS, Pruett ST, Garnier-Amblard EC,Liebeskind LS, Park H, Wang E, Sullards MC, Merrill AH Jr, Riley RT. J Biol Chem.284(8):4786-95.

Acid Sphingomyelinase Deficiency Prevents Diet-induced Hepatic TriacylglycerolAccumulation and Hyperglycemia in Mice (2009) Deevska GM, Rozenova KA,Giltiay NV, Chambers MA, White J, Boyanovsky BB, Wei J, Daugherty A, Smart EJ,Reid MB, Merrill AH Jr, Nikolova-Karakashian M. J Biol Chem. 284(13):8359-68.

Human cytomegalovirus regulates bioactive sphingolipids (2008) Machesky NJ, ZhangG, Raghavan B, Zimmerman P, Kelly SL, Merrill AH Jr, Waldman WJ, Van Brocklyn JR,Trgovcich J. J Biol Chem. 283(38):26148-60.

Imaging MALDI mass spectrometry using an oscillating capillary nebulizer matrixcoating system and its application to analysis of lipids in brain from a mousemodel of Tay-Sachs/Sandhoff disease (2008) Chen Y, Allegood J, Liu Y, Wang E,Cachón-Gonzalez B, Cox TM, Merrill AH Jr, Sullards MC. Anal Chem. 80(8):2780-8.

H. Comparison of these methods with other sphingolipidomictechniques in the literature

“Infusion" techniques: Use the same precursor ion / neutral loss scans - Great for profiling, not quantitation, suffer from ionization suppression, isotopic, isobaric, and isomeric

interferences especially without hydrolysis and extraction.

Nanospray ionization: Greatly improved sensitivity and reduced chemical noise: allows detection of low abundance species, detailed structural analyses on numerous species, chip-based systems can be coupled to LC, and fraction collection.

Ultra high resolution mass analysis: allows differentiation of isobaric / isotopic interferences and alternative fragmentation techniques.