www.sciencemag.org/cgi/content/full/science.aaa1051/DC1

Supplementary Materials for

Mass spectrometry imaging with laser-induced postionization Jens Soltwisch, Hans Kettling, Simeon Vens-Cappell, Marcel Wiegelmann, Johannes

Müthing, Klaus Dreisewerd*

*Corresponding author. E-mail: klaus.dreisewerd@uni-muenster.de

Published 5 March 2015 on Science Express DOI: 10.1126/science.aaa1051

This PDF file includes:

Materials and Methods Figs. S1 to S9 Tables S1 to S3 References

2

Materials and Methods

Chemicals

All chemicals were from Sigma-Aldrich (Steinheim, Germany).

Preparation of Tissue Slices

Mouse organs were dissected from 13-14 week old female C57BL/6 mice according

to standard protocols and approval by the local ethical commission. Testes were dissected

from white Lewis rats (LEW/Crl; Charles River Laboratories, Sulzfeld, Germany).

Whole organs were snap-frozen in lq. N2 and stored at –75 °C before further use. Single

organs were embedded in 2-hydroxyethylcellulose (MWav, 1,500,000 mol/g; 10 g/L).

Braeburn apples were purchased from a local grocery store. A 1 cm3-cube was cut out of

the outmost region containing the skin, embedded in 2-hydroxyethylcellulose and snap-

frozen in lq. N2. Approximately 15 µm-thick tissue slices were prepared with a

cryomicrotome (Jung Frigocut 2800-E, Leica, Wetzlar, Germany) at –20 °C and thaw-

mounted on standard (non-coated) histological glass slides.

Matrix Coating of Animal and Plant Tissue Slices

A sublimation/recrystallization protocol adopted from (30) was used. 2,5-

dihydroxybenzoic acid (DHB) was sublimated for 8 min at 125 °C and 5 × 10-5

mbar

onto a water-cooled sample, mounted 8 cm above the sublimation site. Norharmane, a

powerful matrix for negative ion mode measurements (31), was sublimated for 8 min at

135 °C under otherwise identical conditions. Subsequently, the matrices were allowed to

recrystallize at 75 °C for 2.5 min in a saturated atmosphere of H2O:methanol (1000:5,

v/v).

Preparation of Homogenized Liver Tissue

Pig liver was purchased from a local butchery, filled in a plastic tube and snap-

frozen in lq. N2. 7 g of homogenized tissue was mixed with 3 g of aqueous 2-

hydroxyethylcellulose; ~15 µm-thick tissue slices were prepared with the cryotome.

Matrix spray coating was achieved by use of an airbrush with a 150 µm-nozzle (infinity

solo, Harder&Steenbeck, Norderstedt, Germany). DHB was dissolved to 30 mg/mL in

acetone/water (1:1, v/v). The distance between the airbrush outlet and the sample was set

to 25 cm. Matrix was sprayed in 40 bouts of 1 s duration, followed by 30 s drying

intervals. A back pressure of 3 bar of N2 was used and the gas flow maintained during

drying cycles to support solvent evaporation.

Sample Preparation for Standards

Synthetic compounds were prepared by mixing 9 parts of dissolved DHB matrix (10

mg/mL in 70% acetonitrile) with 1 part of analyte solution (1 mg/mL in

chloroform:methanol (1:1, v:v). 1 µL of the mixture was applied to a non-coated

histological glass slide and allowed to dry.

3

Spectrophotometry

Optical absorption spectra of DHB (dissolved in H2O) and norharmane [dissolved in

chloroform:methanol (1:1, v:v)] were recorded with a dual beam spectrophotometer (UV-

2102PC, Shimadzu, Duisburg, Germany).

Mass Spectrometry

Mass spectra were recorded with a hybrid, orthogonal-extracting time-of-flight

(QTOF) mass spectrometer (MALDI Synapt G2-S HDMS; Waters/Micromass,

Manchester, UK). The decoupling of the ion generation and mass analysis in the QTOF

geometry provided a constant mass accuracy, independent of the position of ion

generation in the ion source. Identical ions are thus always detected with their correct

mass regardless whether they were produced by regular MALDI or by MALDI-2. The

resolution mode, providing a mass resolution R = m/∆m (fwhm) of ~20,000, was used in

all experiments, except for the measurements in which “comprehensive” lipid profiles

were generated as listed in tables S1 and S2; here the high resolution mode (R~40,000)

was applied. While the high resolution mode assists in differentiating lipid species with

similar m/z values, it has the disadvantage of an about 5-times lower ion transmission

(23). All lipid species for which MS images are presented were sufficiently differentiated

also in the resolution mode. To corroborate the proposed structure of the compounds that

were selected for the MS images in Fig. 2 and 3, low-energy collision-induced

dissociation (CID) tandem mass spectrometry was performed using Ar as collision gas

and collision energies Elab between 15 and 75 eV.

Adjustment of the Cooling Gas Pressure in the Ion Source

For adjustment of the buffer gas pressure p, the standard MALDI ion source of the

MALDI Synapt mass spectrometer was previously modified (23) by adding a gas supply

line that is directly connected to the region of ion generation, a shielding ring aperture

between the ion source region and compartment housing of the transfer hexapole (Fig. 1)

for enhanced confinement of the gas fluxes, and miniaturized pressure gauges. Using

solenoid valves, control electronics, and custom-made software, the gas pressure could be

varied continuously between ~0.03 to 3 mbar. In the described experiments, a range of

0.5 to 3 mbar was investigated. To prevent sparking between the rods of the hexapole ion

guide upon increasing the gas pressure, the RF voltage driving the hexapole was reduced

from its default value of 450 V to 200 V.

Lasers

A N2 laser (MNL 100-LD; LTB Lasertechnik Berlin, Germany) was used as the

conventional MALDI-laser. This laser emits pulses of 3 ns duration (fwhm) at a

wavelength of 337.1 nm. The pulse-to-pulse energy stability was better than 5% (standard

deviation). In addition to the laser beam shaping described in (23), a plano-convex lens

with a focal length of 60 mm, mounted inside the MALDI ion source (Fig. 1), replaced

the “external” focusing lens (mounted outside of the vacuum housing), in order to further

reduce the focal spot size. The angle of incidence of the laser beam onto the sample was

45°. The effective laser spot diameter (area of visible material ejection) was determined

by inspection of the material ablation craters. For the MSI experiments, an effective laser

spot size (area of visible material ejection) of ~5 µm was determined (fig. S1). For the

4

spray-coated liver tissue samples, the effective focal diameter was ~9 µm, due to the

morphology-dependent threshold laser fluences. The laser fluences (pulse energy/area)

were adjusted to typical MALDI values, a factor of about 2 above the ion detection

threshold. For example, for DHB this corresponds to pulse energies of about 500 nJ and

1.2 µJ, respectively, for the two preparation protocols.

An optical parametric oscillator (OPO) laser (versaScan/170/MB-ULD, GWU-

Lasertechnik, Erftstadt, Germany), equipped with a second harmonic generator (SHG),

was used to initiate the postionization processes. The OPO was pumped by the

frequency-tripled beam (λ = 355 nm) of a Q-switched Nd:YAG-laser (Surelite II-20,

Continuum/Quantronix, Planegg, Germany). The output pulse of the OPO has a temporal

width of ~7 ns (fwhm). The overall tuning range of the OPO laser system is 213–2550

nm. The beam of the PI laser was steered using Al surface mirrors. A fused silica plano-

convex lens (f = 200 mm) served for focusing of the PI laser beam. This lens was

mounted close to a vacuum port on the rear side of the MALDI-ion source housing. In

this study a spare MALDI source (generous gift of Waters/Micromass) was used which is

containing three vacuum ports on the back side, of which one was previously unused. In

contrast, the regular ion source for the MALDI Synapt contains only two ports that are

both used for placing illuminating LEDs inside the source. One LED could be omitted

and the free port then used to adopt the laser beam. Two Al surface mirrors, one of which

was mounted inside the vacuum housing, were used for obtaining a parallel alignment of

the focal beam waist relative to the sample plate surface. During all experiments reported

here, the distance ∆z of the PI laser beam to the sample surface was set to ~0.5 mm (Fig.

1). The approximate width of the focal beam diameter of the PI laser was determined by

placing a piece of paper at the focal position and inspection of the laser-generated hole to

~0.1 mm.

A custom-made trigger unit was used to adjust the delay between the two laser

pulses. The timing sequence was controlled using two photodiodes and a fast digital

oscilloscope. Pulse energies for both lasers were adjusted with gradient density filters and

measured with a laser pulse energy meter.

Data Acquisition and Processing of Mass Spectra

MALDI-MS images were recorded by irradiating each ~5 µm-wide pixel with 20

laser pulses (at 20 Hz). The pitch size (pixel-to-pixel distance) was set to 15 µm, except

for the measurements on rat testis where a smaller pitch size of 11.25 µm was applied to

increase the lateral resolution. MassLynx software (Waters) was used for acquisition of

mass spectra. Mass calibration was achieved with red phosphorus cluster ions (23), which

provided an initial mass accuracy of ≤ 1 ppm. Care was taken to maintain the temperature

in the laboratory as constant as possible during the MALDI-MSI runs (e.g., by

equilibration of the room temperature by triggering the PI laser for a certain time period

before starting the MALDI-MSI experiment). However, a small temperature drift within

1 °C was generally not avoidable during hours-long MSI runs. Even such a small

temperature shift affected the calibration in the high-resolution mode. The MassLynx

software currently does not provide a correction for this “long-term” drift. Therefore, all

MS images shown were recorded in the resolution mode, where the small temperature

drift was tolerable within the 0.02 Da bin width applied for image generation.

5

All color-coded MS images were produced using BioMAP software (v3.8.0.4,

Novartis, Basel, Switzerland). HDImaging software (Waters) was applied for initial file

conversion to enable subsequent processing with BioMAP. All presented MALDI-MS

images and mass spectra display raw data without smoothing or interpolation.

Identification of ion species (tables S1 and S2) was based on comparison of

experimental and calculated m/z values, using Lipid Maps data base

(http://www.lipidmaps.org/), and comparison with literature (32-34). The identity of all

compounds for which MS images are presented, was further corroborated by tandem MS.

The negative ion mode data typically also enabled assigning the overall composition of

the fatty acyl (or ether-linked) side chains; for the positive ion mode, general only the

combined composition could be derived. For example, PE(40:6) refers to a species that

contains 40 carbon atoms and 6 double bonds in the two fatty acyl chains. For cholesterol

and all liposoluble vitamins, MS/MS spectra of synthetic compounds were generated for

comparison.

Hematoxylin and Eosin (H&E) Staining

H&E stains were obtained after washing off the matrix with chloroform/methanol

(2:1, v/v) after the MSI runs. Damage to the tissue slice that potentially might have been

caused by the focused laser beam was not notable. Presumably, the laser light is only

scattered in the tissue rather than being absorbed; at the wavelength of the N2 laser of 337

nm, the investigated tissues can be assumed to be essentially opaque.

Identification of Mouse Brain Tissue

Mouse brain tissue areas were identified by comparison with a stereotaxic brain atlas

(35).

Safety Hazard Notes

The employed lasers are of laser safety classes 3B (N2 laser) and 4 (OPO system).

Due safety precautions have to be taken upon working with free beams of these lasers

(e.g., by wearing protective goggles). Airbrush sample preparations should be handled in

a fume hood.

6

Figure S1

Fig. S1. Morphologies of matrix-coated tissue slices and laser ablation sites

(A) Image from scanning electron microscopy (SEM) showing the morphology of the

sublimated/recrystallized DHB matrix coating on a mouse brain slice. The optical images

in (B and C) show the ablation craters that were produced by the primary laser during the

MALDI-2-MS imaging runs. (D) SEM image of sublimated/recrystallized norharmane

matrix coating on a mouse brain slice. In this case, the laser ablation sites were produced

on a different tissue slice than used for MS imaging and buy using a slightly higher pulse

energy. This increased the diameter of the ablation spot slightly above 5 µm. All MSI

runs were performed with a 5 µm spot as corroborated by inspection of the slices after the

MSI run with an optical microscope.

7

Figure S2

100 200 300 400 500 600 700

0.0

0.2

0.4

0.9

1.0

313.2

7

413.2

7

684.5

5

561.5

3

399

.29

702

.54

364.2

7

403.3

8

578.5

3604.5

4

100 150 200 250 300 350

0.0

0.2

0.8

1.0

105.07

259.24

175.15

36

9.3

5

161.13

243.21

287.27

215.18

147.12

135.11

100 200 300 400 500 600 700 800

0.0

0.2

0.4

0.8

1.0

577

.53

340

.36

78

7.4

8

553.5

1

497.1

9

606.6

2

36

9.3

5

76

9.6

5

624.6

2

28

4.2

926

4.2

7

100 200 300 400 500 600 700 800

0.0

0.8

1.0

264.2

6

643.5

2

36

9.3

5

631

.56

76

7.5

0

341

.31

627.5

5

NH2

OP

O

HO

OOR

O H

O

R'

H+

631.56

[PA(38:2)+K]+

m/z: 767.50

[PE(O-34:2)+H]+

m/z: 702.54

[cholesterol-H20+H]+

m/z: 369.35

577.53

[PG(34:1)+K]+

m/z: 787.45

inte

nsity

/ a

.u.

inte

nsity

/ a

.u.

D

B

C

A

m/zm/z

m/zm/z

HO

H

H H

H

H

561.53

OP

O

HO

OOR

O

O H

O

R'

OH

HO H

K+

100 200 300 400 500 600 700 800

0.0

0.1

0.8

1.0

55

1.5

1

264

.27

61

2.6

1

70

1.5

1

630.6

2

810

.68

648

.62

79

2.6

7

751

.53

57

7.5

2

100 200 300 400 500 600 700 800

0.0

0.2

0.4

0.6

0.8

1.0

21

2.1

5

60

3.5

5

71

2.5

6

651.5

4

83

6.5

5

682

.57

695

.50

577

.52

100 200 300 400 500 600 700 800

0.0

0.2

0.4

0.8

1.0

26

4.2

7

61

5.4

7

75

5.4

9

577.5

2

798.5

4

656

.58

73

9.4

6

78

0.6

8

55

1.5

1

100 200 300 400 500 600 700 800

0.00

0.02

0.04

0.8

1.0

651.5

3

341

.31

792

.56

502

.98

354

.28

77

4.6

7

36

6.9

63

85

.27 6

23.5

0

NH2

OP

O

HO

OOR

O

O H

O

R'

H+

630.62

E

651.53

[GalCer(d18:1/C24:1)+H]+

m/z: 810.68

[PC(34:1)+K]+

m/z: 798.54

651.53 577.53

[PS(40:6)+H]+

m/z: 836.54

[PE(40:6)+H]+

m/z: 792.55

F

HG

inte

nsity

/ a

.u.

inte

nsity

/ a

.u.

m/zm/z

O

OH

OH

HO

OH

O

H OH

HNHR'

O

R

H+

8

Fig. S2. Tandem mass spectra of compounds presented in the MSI images of Fig. 2

(positive ion mode)

MALDI-2 CID tandem mass spectra generated from mouse cerebellum in the positive ion

mode and fragmentation schemes. A precursor ion selection window of approximately

1 u was applied. The MS/MS spectra of cholesterol and the PL and GL species (A to H)

were obtained from the same tissue slice as used for generating Fig. 2. MS/MS spectra of

the vitamins (I to L) were recorded from parallel slices. The m/z values of the targeted

precursor ions are underlined, diagnostic fragment ions are printed in bold face, or where

derived by comparison with MS/MS data of synthetic compounds (in particular for

cholesterol and all vitamins, which produce more complex fragmentation patterns) in

bold italics. Together with the exact mass of the precursor ions, the characteristic

fragment ions reveal the lipid class and overall composition of the two hydrocarbon

chains for PLs and GLs. The presence of further isobaric ions cannot be excluded.

224 225 226

0.000

0.005

0.010

0.015

0.8

1.0

225.09

186 187 188

0.000

0.005

0.010

0.015

0.8

1.0

187.07

443 444 445

0.0

0.2

0.8

1.0

444

.28

100 200 300 400

0.0

0.2

0.4

0.6

0.8

1.0

109.28

205.12

165.09

431

.39

185 186 187

0.00

0.01

0.8

1.0

186.14

255 256 257

0.00

0.01

0.8

1.0

256.21

100 200 300 400

0.0

0.2

0.4

0.6

0.8

1.0

145.11

159.12

131.09

367.34

38

5.3

5

187.15

286 287 288

0.0

0.2

0.8

1.0

28

7.2

4

m/z

I[vitamin D3+H]+

m/z: 383.35

[vitamin K2]●+

m/z: 444.29

J

LK

inte

nsity

/ a

.u.

inte

nsity

/ a

.u.

m/z

[vitamin A1+H]+

m/z: 287.24

[vitamin E+H]+

m/z: 431.38

9

700 800 9000.0

5.0x103

5E3

0in

tensity

/ a.u

.

m/z

single pixel

Ø ≈ 5 µm

►

�

�► �

��

▲

�

▲

♠

Fig. S3. Representative MALDI-2 mass spectrum acquired in the negative ion mode

from a single pixel

The same data set as used for production of Fig. 3 (norhamane matrix-coated mouse

cerebellum) was evaluated. �: [PA - H]-; �: [PE - H]

-; �: [PE-P - H]

-; ▲: [PS - H]

-; ♠:

[PI - H]-.

10

Figure S4

437.27

283.26

419.26

100 200 300 400 500 600 700 800

0.0

0.2

0.40.8

1.0

437

.27

283

.26

331

.26

751

.53

83

8.5

5

419

.26

670

.52

100 200 300 400 500 600 700 800 900 1000 1100 1200 1300

0.0

0.2

0.8

1.0

750.5

5

916.5

7

706

.24

36

4.1

2

72

6.6

0 88

8.6

4

10

91.7

3

303.2

3

12

53.7

8

1073

.71

1091.73

100 200 300 400 500 600 700 800 900

0.0

0.2

0.8

1.0

96.9

6

255

.23

88

8.6

2

283.2

7

100 200 300 400 500 600 700 800 900

0.0

0.2

0.40.8

1.0

241

.01

25

5.2

3

716.5

2

303

.23

28

3.2

6

59

9.3

2

885

.55

58

1.3

1

419

.26

241.01F[PI(18:0/20:4)-H]-

m/z: 885.55

599.32

581.31

303.23

m/z

E

GA1(d18:1/C18:0)-H]-

m/z: 1253.79

HG

inte

nsity

/ a.u

.in

tensity

/ a

.u.

m/z

[ST(d18:1/C24:1)-H]-

m/z: 888.62

364.12

[PS(18:0/22:4)-H]-

m/z: 838.56

751.53283.26

331.26

96.96726.60

706.24

888.64

283.26

100 200 300 400 500 600 700 800

0.0

0.2

0.4

0.6

0.8

1.0

48

0.3

2

28

3.2

7

46

2.3

0

766

.54

30

3.2

4

167

.06

100 200 300 400 500 600 700 800

0.0

0.2

0.4

0.6

0.8

1.0

43

6.2

8

283

.25

74

7.5

2

32

7.2

4

42

0.2

3

100 200 300 400 500 600 700 800

0.0

0.2

0.6

0.8

1.02

81

.25

167

.06

255

.23

43

5.2

5

699

.50

53

1.1

6

41

7.2

4

327.25

100 200 300 400 500 600 700 800

0.0

0.2

0.6

0.8

1.0

44

6.3

0

255

.22

28

1.2

5

464

.31

72

8.5

6

309

.28

39

1.2

2

435.25

m/z

A

[PE(18:0/20:4)-H]-

m/z: 766.54

B

DC

inte

nsity

/ a.u

.in

tensity

/ a.u

.

m/z

[PE(P-18:0/18:1)-H]-

m/z: 728.56

[LBPA(16:1/18:0)-H]-

m/z: 747.52

303.23

480.32

[PA(18:1/18:1)-H]-

m/z: 699.50

281.25

435.25

436.26

446.30

281.25

464.31

462.30

11

Fig. S4. Tandem mass spectra of compounds presented in the MSI images of Fig. 3

(negative ion mode) MALDI-2 CID tandem mass spectra generated from mouse cerebellum in the negative

ion mode and fragmentation schemes. The MS/MS spectra were obtained from an

adjacent tissue slice as used for Fig. 3. The m/z values of the targeted precursor ions are

underlined, diagnostic fragment ions are printed in bold face. Together with the exact

mass of the precursors these reveal the lipid class and overall composition of the

hydrocarbon chains for PLs and GLs. In most cases, also the composition of the

individual R and R’ residues could be derived. The presence of further isobaric ions

cannot be excluded. Due to the ~1 u wide precursor ion selection window, generally

several precursor ions are fragmented simultaneously, which results in a multitude of

additional fragment ion species.

12

Fig. S5. MALDI-2-MS images of rat testis

(A to D) MALDI-2-MS (left) and MALDI-MS (right) images showing the distribution of

different TG species in a cross section of rat testis. The data reflect the typical overall TG

composition of rodent testis (34). However, differential distributions in distinct

seminiferous tubules are visualized by the MS images, which could reflect testicular

maturation (36). Atypical for conventional MALDI, with MALDI-2 the TGs are

abundantly detected as (relatively labile (37)) [M + H]+ species. This indicates the

particular softness of the method. The signal intensities of alkali metal adducts of the TGs

are more moderately increased by the postionization step. The example of [TG(54:5)+K]+

is shown in (C). Only the overall composition of the three fatty acyl chains can be

derived from the MS1

data. (E to G) Additional information and differentiation of the

lumen of tubule, the smooth muscle layer surrounding the seminiferous tubules and the

interstitial space is obtained by the distribution of PA and PC species. In form of their

alkali metal adducts, these PLs are detected sensitively by both MALDI-2 and MALDI-

MS. (H) H&E stain of the tissue slice obtained after the MS analysis. The MS data were

recorded in the positive ion mode with a sublimated/recrystallized DHB matrix; λPI=280

nm, τ=9 µs, p=2.5 mbar; 20 laser pulses were applied per pixel and a pitch size of 11.25

µm was used for imaging. The displayed mass spectra were accumulated over 10 pixels.

919 920-2.0x10

3

0.0

2.0x103

907 908-2.0x10

3

0.0

2.0x103

881 882-2.0x10

3

0.0

2.0x103

855 856

-2.0x103

0.0

2.0x103

855.76 881.78 907.80

2E3

00

2E3

0

2E3

[TG(54:5)+H]+ [TG(56:6)+H]+[TG(52:4)+H]+

MALDI-2 MALDI

m/z m/zm/z

E F G

C

[PC(38:5)+K]+

4E3

0

[TG(54:5)+K]+

919.75

846.55

[PA(36:3)+K]+

737.45

[PC(34:1)+K]+

798.55

B DA

m/z

max0intensity / a.u.

500 µm

MALDI

MALDI-2

H&E stain

500 µm

13

Fig. S6. MALDI-2-MS images of plant tissue

(A and B) MALDI- (left) and MALDI-2-MS (right) images showing the distributions of

dihexose (Hex2) and the phenolic glycoside quercitin pentoside (querc-5S), respectively,

in a slice of a Braeburn apple coated with norharmane matrix. The images were recorded

in the negative ion mode with λPI=260 nm, τ=10 µs, p=2.0 mbar; 20 laser pulses were

applied per pixel and a pitch size of 15 µm used. The corresponding mass spectra

(accumulated over 400 pixels) show that upon using the PI laser a gain by about 2 orders

of magnitude is obtained for the [M – H]- signals of the disaccharide; note that the

MALDI-MS image shown in (A) has been scaled by a factor of 50 relatively to the

MALDI-2-data shown at the right side. Similar signal gains were obtained for mono- as

well as higher oligosaccharides. The querc-5S shows a more complex behavior. This

“antioxidant” is detected in the MALDI-2 spectra in form of multiplet signals, including

the displayed molecular [M – H]- and M

•- ions. Possibly, the radical ion species were

generated by the capture of free electrons in the MALDI-2 plume or by further yet

unknown processes that were enabled by a direct photoexcitation of the highly absorbing

polyphenol. (C) Comparison of the MS data with an optical image, obtained from an

adjacent slice, shows that the quercetin pentoside is enriched in the skin region, in line

with the findings made in a previous MALDI-MSI study with Golden Delicious apples

(38). Although dihexose is more evenly distributed throughout the hypanthium, fine

features are notable, which for example reflect the finer cell structures that are found

directly beneath the apple skin (this area is containing less dihexose). Further features,

such as the slightly inhomogeneous distribution of Hex2 in the central part of the

hypanthium or the “sandwich-type appearance” of the querc-5S distribution in the skin

region may (partially) have been caused by sample preparation artefacts. Likely, some

leakage of the carbohydrates during the embedding step in the 2-hydroxyethylcellulose

prior to the cryotome cutting also caused the low signal background around the apple

tissue.

341.0 341.5

-1x105

0

433 434

-4.0x103

0.0

0.0

2.0x105

[querc-5S]●-

MALDI-2MALDI

CA

[Hex2 - H]- [querc-5S - H]-

433.08 434.08341.11

hy

pa

nth

ium

skin

0.0

8.0x106

2E5

0

4E3

m/z

8E6

0

1E5

3300

0

inte

nsity

/ a.u

.

500 µm

341.0 341.5

MALDI

MALDI-2

B

66

0

max

0

inte

nsity

/ a.u

.

m/z

14

^

Fig. S7. Parameters affecting the MALDI-2 ion yields (additional data to Fig. 4).

Heat maps (generated from spray-coated homogenized liver tissue slices) illustrating the

effect of cooling gas pressure p and delay time τ (top row), and those of the PI laser

wavelength λ and pulse energy EPI (bottom row) on the signal intensities of different ion

species. The data were generated using the DHB matrix in the positive ion mode. The

signal intensities are color-coded. (A) Total ion count (TIC). (B) [M – H2O + H]+ ion

signals of cholesterol. (C) Protonated ion signals of PC(34:1). (D) Sodium adduct ion

signals of PC(34:1). The same data set used to generate Fig. 4A and Fig. 4B of the main

text was evaluated. Each data point (marked as black dots) was recorded by applying 600

laser pulses onto approximately 50 positions on the tissue slice. The solid white line

reflects the solution phase mean penetration depth (α-1) of DHB, the dashed vertical lines

in the bottom row graphs the two photon ionization threshold of DHB (27).

220 240 260 280 300 320 340

200

400

600

800

0.0

0.2

0.4

0.6

0.8

1.0

220 240 260 280 300 320 3400

200

400

600

800

0.0

0.2

0.4

0.6

0.8

1.0

0.5 1.0 1.5 2.0 2.5 3.00

5

10

15

20

25

0.5 1.0 1.5 2.0 2.5 3.00

5

10

15

20

25

0.5 1.0 1.5 2.0 2.5 3.00

5

10

15

20

25

0.0E+00

6.7E+07

1.3E+08

2.0E+08

2.7E+08

3.4E+08

TIC

0.5 1.0 1.5 2.0 2.5 3.00

5

10

15

20

25

220 240 260 280 300 320 3400

200

400

600

800

0.0

0.2

0.4

0.6

0.8

1.0

220 240 260 280 300 320 3400

200

400

600

800

0.0

0.2

0.4

0.6

0.8

1.0

de

lay

/ µ

sla

se

rp

uls

e e

ne

rgy / µ

J

pressure / mbarpressure / mbarpressure / mbarpressure / mbar

wavelength / nmwavelength / nmwavelength / nmwavelength / nm

inte

nsity

/ a

.u.

0

max

B C DA

α-1

mean

penetr

ation

depthα

-1/ a.u

.

[cholesterol-H20+H]+ [PC(34:1)+Na]+[PC(34:1)+H]+

2-p

hoto

n th

reshold

15

Fig. S8. Adduct formation between analyte and matrix

(A) MALDI-2 and (B) conventional MALDI mass spectra acquired from liver

homogenate in the positive ion mode with spray-coated DHB matrix. A few PE-matrix

adduct ions can be discerned in the MALDI-2 spectra, providing further evidence for the

occurrence of biomolecular gas phase collisions (39). In the conventional MALDI mass

spectra, recorded under otherwise identical conditions, sizable signal intensities of such

adducts are notable only for the most abundant PC ions.

860 880 900 9200

2000

860 880 900 9200

20000

40000

m/z

4E4MALDI[PE(O-36:5)+(m-H2O)+H]+

0

inte

nsity

/ a.

u.

m/z

MALDI-2

2E3

0

2E4

BA

[PE(38:4)+(m-H2O)+H]+[PC(34:1)+(m-H2O)+H]+

[PC(34:2)+(m-H2O)+H]+

16

430 431 432-1.0x10

2

-5.0x101

0.0

0.0

2.0x102

1200 1400 1600

-2.0x103

0.0

430 431 432

-1x105

0

0.0

2.0x107

4.0x107

0

1x105

600 650 700 750-1x10

6

0

1x106

2x106

3x106

3E6

1E6

1E6

0

[M+

H]+

[M+

Na]+

[M-H

G+

H]+

������������ �����

���������� ��� ��= 1.34

������������ �����

���������� ��� ��= 1.45

MALDI-2

MALDI

B

2E3

0

A

MALDI-2

MALDI

1E5

[GM1-H]–

[GA1-H]–

���������� ���� ��

���������� ��� �� = 6.7

���������� ���� ��

���������� ��� ��= 0.1

m/z

4E7

[M+H]+

M●+

0

1E5

MALDI-2

MALDI

D

m/z

m/z

[M+H]+

M●+

2E3

50

0

C

m/z

MALDI-2

MALDI

Fig. S9. Mass spectra of purified/synthetic compounds with different absorption

properties

(A) MALDI-2 and MALDI mass spectra of PE(36:0)—which is non-absorbing at the

postionization laser wavelength of 280 nm—show that the characteristic [M – HG + H]+

fragment signal, which arises from the loss of the PE head group (HG), does not increase

sizably despite of orders of magnitude higher [M + H]+ ion signals. (B) In contrast, the

MALDI-2-MS analysis of a mixture of GM1 (monosialotetrahexosylganglioside) and

STs, fractionated from mouse brain extract, shows a depletion of the molecular GM1 ion

by the loss of the weakly bound sialic acid (N-acetylneuraminic acid, Neu5Ac) residue,

giving rise to the asialoganglioside GA1. The dissociation could be triggered by the direct

photoexcitation of the Neu5Ac residue(s) at λPI = 260 nm. Examination of the isotopic

distribution of the intact molecular ion signals shows that next to the [M – H]- species a

small abundance of GM1 M•- is found. A loss of other, non-absorbing and stronger bound

carbohydrate units is not observed. (C) If analyzed from a concentrated sample with a

molar analyte-to-matrix (A/M) ratio of a few hundred, α-tocopherol (vitamin E) is

detected predominantly as M+•

and only to a lower degree as [M + H]+ ion species. Black

lines in the mass spectra denote the theoretical isotopic distribution of M+•

; the 13

C

isotope of this ion is overlapping with the 12

C isotope of the [M + H]+ ion. The radical

ions are presumably produced by direct two-photon ionization of α-tocopherol, which in

solution exhibits a high molar decadic extinction coefficient ε280nm of ~3 × 103 L mol-1

cm-1. (D) If desorbed from a complex tissue matrix, more elevated [M + H]+ abundances

are detected. This could be caused by a lower A/M-ratio, increasing the number of gas

phase reactions between matrix ions and neutral analyte molecules, or also by a depletion

of the reactive M+•

ions by gas phase reactions with other acceptor molecules in the

MALDI plume.

17

Table S1.

Experimental and calculated m/z values, lipid class, tentatively proposed structures, and

adduct ion types of lipids detected by conventional MALDI and MALDI-2 mass

spectrometry imaging from mouse cerebellum in the positive ion mode with a DHB

matrix

Exp.

m/z

Calc.

m/z

Tentatively

proposed identitya) MALDI

b) MALDI-2

b)

287.24 287.24 [vitamin A1 (retinol)+H]+ - +

369.35 369.35 [cholesterol-H2O+H]+ 0 +++

385.34 385.34 [vitamin D3+H]+ 0 ++

387.36 387.36 [cholesterol+H]+ - +

431.38 431.38 [vitamin E+H]+ - +

444.29 444.31 [vitamin K2 (MK-4)]•+ - +

672.53 672.54 [GalCer(d18:1/C14:0)+H]+ - +

675.50 675.50 [PA(34:1)+H]+ /

[PC(32:0)-N(CH3)3+H]+ - ++

697.48 697.48 [PA(34:1)+Na]+ /

[PC(32:0)-N(CH3)3+Na]+ - +

700.53 700.53 [PE(O-34:3)+H]+ + +

701.51 701.51 [PA(36:2)+H]+ /

[PC(34:1)-N(CH3)3+H]+ - +

702.51 702.51 [PC(30:2)+H]+ - ++

702.54 702.54 [PE(O-34:2)+H]+ - +++

704.53 704.52 [PC(30:1)+H]+ - ++

704.56 704.56 [PE(O-34:1)+H]+ - ++

710.49 710.48 [PE(32:2)+Na]+ - +++

713.45 713.45 [PA(34:1)+K]+ /

[PC(32:0)-N(CH3)3+K]+ ++ ++

718.53 718.54 [PE(34:1)+H]+ +++ +++

720.55 720.55 [PE(34:0)+H]+ - ++

721.48 721.48 [PA(38:6)+H]+ /

[PC(36:5)-N(CH3)3+H]+ - +

723.49 723.49 [PA(36:2)+Na]+ /

[PC(34:1)-N(CH3)3+Na]+ ++ ++

724.50 724.49 [PC(30:2)+Na]+ 0 ++

725.50 725.51 [PA(36:1)+Na]+ /

[PC(34:0)-N(CH3)3+Na]+ + ++

726.52 726.50 [PC(30:1)+Na]+ - +

726.54 726.54 [PE(O-36:4)+H]+ - ++

726.60 726.59 [GalCer(d18:1/C18:1)+H]+ - ++

728.47 728.46 [PE(32:1)+K]+ - 0

728.56 728.56 [PE(O-36:3)+H]+ - +++

728.60 728.60 [GalCer(d18:1/C18:0)+H]+ - ++

730.53 730.54 [PC(36:2)+H]+ - ++

730.57 730.57 [PE(O-36:2)+H]+ 0 +++

730.62 730.62 [GalCer(d18:0/C18:0)+H]+ - ++

731.61 731.61 [SM(d18:1/C18:0)+H]+ - +

732.55 732.55 [PC(32:1)+H]+ - +

734.57 734.57 [PC(32:0)+H]+ + ++

739.47 739.47 [PA(36:2)+K]+ /

[PC(34:1)-N(CH3)3+K]+ +++ +++

739.53 739.52 [SM(d18:1/C16:1)+K]+ - +

740.53 740.53 [PE(36:4)+H]+ - +

741.48 741.48 [PA(36:1)+K]+ ++ ++

742.54 741.53 [PE(36:3)+H]+ - +

744.55 744.55 [PE(36:2)+H]+ - +

744.60 744.59 [PC(O-36:2)+H]+ - ++

746.57 746.57 [PE(36:1)+H]+ - +++

747.50 747.50 [PA(38:4)+Na]+ /

[PC(38:6)-N(CH3)3+H]+ - ++

748.50 748.49 [PC(32:4)+Na]+ - +

748.53 748.53 [PE(O-34:4)+Na]+ - +++

748.58 748.59 [PE(36:0)+H]+ 0 +

750.54 750.54 [PE(O-36:3)+Na]+ - ++

18

751.53 751.53 [PA(40:5)+H]+ /

[PC(36:1)-N(CH3)3+Na]+ + ++

752.53 752.52 [PC(32:2)+Na] 0 ++

752.56 752.56 [PE(O-38:5)+H]+ 0 +++

754.56 754.54 [PC(34:4)+H]+ - ++

754.57 754.57 [PE(O-38:4)+H]+ - +

754.62 754.62 [GalCer(d18:1/C20:1)+H]+ 0 ++

756.55 756.55 [PC(32:0)+Na]+ ++ ++

756.59 756.59 [PE-O(38:3)+H]+ + +++

758.50 758.49 [PS(32:0)+Na]+ 0 +

758.56 758.55 [PC(O-22:0)+K]+ - +

758.60 758.61 [PE(O-38:2)+H]+ - ++

760.58 760.59 [PC(34:1)+H]+ + +++

761.45 761.45 [PA(38:5)+K]+ /

[PC(36:4)-N(CH3)3+K]+ + +

762.59 762.60 [PC(34:0)+H]+ - +

764.52 764.52 [PE(38:6)+H]+ - +++

765.48 765.48 [PA(38:3)+K]+ /

[PC(36:1)-N(CH3)3+K]+ + +

766.54 766.54 [PE(36:2)+Na]+ + ++

767.49 767.50 [PA(38:2)+K]+ ++ ++

768.50 768.50 [PC(32:2)+K]+ ++ ++

768.56 768.55 [PE(38:4)+H]+ + ++++

768.60 768.59 [PC(O-34:1)+Na]+ 0 0

770.51 770.51 [PC(32:1)+K]+ + +

770.56 770.57 [PE(36:0)+Na]+ + ++

771.49 771.49 [PA(40:6)+Na]+ 0 +

772.52 772.53 [PC(32:0)+K]+ +++ +++

772.58 772.59 [PE(38:2)+H]+ 0 ++

772.63 772.63 [GalCer(t18:1/C20:0)+H]+ - ++

774.54 774.54 [PE(O-38:5)+Na]+ + ++

774.60 774.60 [PE(38:1)+H]+ 0 ++

776.55 776.56 [PE(O-38:4)+Na]+ - +++

776.60 776.62 [PE(38:0)+H]+ - +

777.41 777.41 [PG(34:6)+K]+ 0 0

778.57 778.57 [PE(O-38:3)+Na]+ - ++

779.43 779.43 [PG(34:5)+K]+ 0 0

780.44 780.45 [PE(38:9)+Na]+ 0 0

780.56 780.55 [PC(36:5)+H]+ - ++

780.60 780.59 [PE(O-38:2)+Na]+ - +++

782.50 782.49 [PS(34:2)+Na]+ 0 +

782.56 782.60 [PE(O-38:1)+Na]+ + ++

782.57 782.57 [PC(34:1)+Na]+ + ++

782.65 782.65 [GalCer(d18:1/C22:1)+H]+ - +++

784.52 784.51 [PS(36:4)+H]+ 0 ++

784.58 784.58 [PC(34:0)+Na]+ + +

784.62 784.62 [PE(O-40:3)+H]+ - ++

784.66 784.67 [GalCer(d18:1/C22:0)+H]+ - ++

785.45 785.45 [PA(40:7)+K]+ /

[PC(38:6)-N(CH3)3+K]+ + +

785.48 785.47 [PG(34:2)+K]+ ++ ++

786.54 786.54 [PE(36:0)+K]+ - +

787.48 787.49 [PG(34:1)+K]+ + ++

788.47 788.46 [PC(34:6)+K]+ 0 +

788.54 788.54 [PS(36:2)+H]+ - ++

788.62 788.62 [PC(36:1)+H]+ 0 +

789.48 789.48 [PA(40:5)+K]+ /

[PC(38:4)-N(CH3)3+K]+ + +

790.54 790.54 [PE(38:4)+Na]+ + ++

791.49 791.50 [PA(40:4)+K]+ - 0

792.55 792.55 [PE(40:6)+H]+ 0 ++++

796.52 796.53 [PC(34:2)+K]+ 0 0

796.58 796.59 [PE(40:4)+H]+ 0 +

796.63 796.62 [PC(O-38:4)+H]+ - 0

797.60 797.59 [SM(d18:1/C20:0)+K]+ - +

798.50 798.50 [PC(36:7)+Na]+ + +

798.54 798.54 [PC(34:1)+K]+ +++ +++

19

798.64 798.64 [PC(O-36:0)+Na]+ - +

800.66 800.65 [PC(O-38:2)+H]+ - +++

802.48 802.48 [PE(38:6)+K]+ + +

806.51 806.51 [PE(38:4)+K]+ + ++

806.57 806.57 [PC(38:6)+H]+ - +

806.63 806.65 [GalCer(d18:1/C24:3)+H]+ - +

808.59 806.48 [PC(36:2)+Na]+ - +

808.66 808.66 [GalCer(d18:1/C24:2)+H]+ - +++

810.50 810.50 [PE(40:8)+Na]+ - +

810.54 810.54 [PE(38:2)+K]+ 0 ++

810.60 810.60 [PC(36:1)+Na]+ ++ ++

810.68 810.68 [GalCer(d18:1/C24:1)+H]+ 0 ++++

812.55 812.54 [PS(38:4)+H]+ 0 ++

812.69 812.70 [GalCer(d18:1/C24:0)+H]+ - +++

813.48 813.48 [PA(42:7)+K]+ ++ ++

814.67 814.67 [PE(O-44:2)+H]+ - ++

820.52 820.53 [PC(36:4)+K]+ + +

820.59 820.58 [PE(40:3)+Na]+ - +

824.55 824.56 [PC(36:2)+K]+ 0 +

826.57 826.57 [PC(36:1)+K]+ ++ ++

826.63 826.63 [PE(40:0)+Na]+ + +

826.67 826.67 [PC(O-40:3)+H]+ 0 +++

828.51 828.52 [PS(36:1)+K]+ 0 +

828.58 828.59 [PC(36:0)+K]+ + +

828.69 828.68 [PC(O-40:2)+H]+ - ++++

830.51 830.51 [PE(40:6)+K]+ ++ ++

830.70 830.70 [PC(O-40:1)+H]+ - +++

832.58 832.58 [PC(38:4)+Na]+ 0 +

834.60 834.60 [PC(38:3)+Na]+ 0 +

836.54 836.54 [PS(40:6)+H]+ - +++

838.53 838.54 [PE(42:8)+Na]+ - ++

838.57 838.57 [PE(O-34:2)+(DHB-H2O)+H]+ - ++

838.62 838.61 [PC(O-38:2)+K]+ ++ ++

844.52 844.53 [PC(38:6)+K]+ ++ ++

845.53 845.53 [PG(40:6)+Na]+ + +

848.56 848.56 [PC(38:4)+K]+ 0 0

848.64 848.64 [PS(40:0)+H]+ + ++

850.65 850.65 [GalCer(d18:1/C24:0)+K]+ ++ ++ a) the presence of isobaric compounds is possible; for sphingolipids, the most likely composition of the sphingosine

(most typically d18:1) is shown. b) 0: intensity per pixel too low for MS imaging; ion is detected in sum spectra recorded from 144 pixels

+: detected with low signal intensity per pixel; meaningful MS images obtained

++: good MS images obtained +++ / ++++: excellent/outstanding MS images with high S/N ratios and image contrast obtained

20

Table S2.

Experimental and calculated m/z values, lipid class, tentatively proposed structures, and

adduct ion types of lipids detected by conventional MALDI and MALDI-2 mass

spectrometry imaging from mouse cerebellum in the negative ion mode with a

norhamane matrix

Exp.

m/z

Calc.

m/z

Tentatively

proposed identitya) MALDI

b) MALDI-2

b)

391.23 391.23 [cLPA(16:0)-H]- c) 0 ++

464.32 464.32 [PE(O-18:1)-H]- - +++

507.28 507.27 [PG(18:2)-H]- - +++

599.33 599.32 [PI(18:0)-H]- 0 +

644.51 644.50 [Cer(d18:1/24:0)-H]- 0 ++

647.47 647.47 [PA(32:0)-H]- - ++

673.49 673.48 [PA(34:1)-H]- 0 +++

699.50 699.50 [PA(36:2)-H]- + +++

700.53 700.53 [PE(O-34:1)-H]- 0 ++

701.52 701.51 [PA(36:1)-H]- 0 ++

715.58 715.58 [PE-Cer(d18:1/C20:0)-H]- 0 ++

716.53 716.52 [PE(34:1)-H]- - ++

718.55 718.54 [PE(34:0)-H]- 0 ++

719.49 719.49 [PG(32:1)-H]- - +

721.50 721.50 [PG(32:0)-H]- - 0

722.52 722.51 [PE(O-36:5)-H]- 0 +

723.50 723.50 [PA(38:4)-H]- 0 +

726.55 726.54 [PE(O-36:3)-H]- 0 ++

728.56 728.56 [PE(P-36:1)-H]- + +++

742.55 742.54 [PE(36:2)-H]- - +

744.56 744.56 [PE(36:1)-H]- 0 ++

746.52 746.53 [PE(P-38:6)-H]- 0 ++

747.50 747.50 [PA(40:6)-H]- + ++

747.52 747.52 [LBPA(34:1)-H]- d) 0 ++

748.53 748.53 [PE(O-38:6)-H]- 0 ++

750.55 750.54 [PE(O-38:5)-H]- 0 ++

754.58 754.58 [PE(O-38:3)-H]- 0 ++

756.60 756.59 [PE(O-38:2)-H]- 0 ++

760.53 760.51 [PS(34:1)-H]- - +

762.51 762.51 [PE(38:6)-H]- + +++

764.53 764.52 [PE(38:5)-H]- 0 ++

766.54 766.54 [PE(38:4)-H]- 0 ++++

770.58 770.57 [PE(38:2)-H]- 0 ++

772.59 772.59 [PE(38:1)-H]- 0 ++

774.55 774.57 [PE(P-40:6)-H]- 0 ++++

778.58 778.58 [PE(O-40:5)-H]- 0 ++

786.53 786.53 [PS(36:2)-H]- 0 ++

788.55 788.5 [PS(36:1)-H]- 0 ++

790.55 790.54 [PE(40:6)-H]- + ++++

794.57 794.57 [PE(40:4)-H]- 0 ++

802.58 802.58 [PE(O-42:6)-H]- - 0

806.56 806.55 [ST(d18:1/C18:0)-H]- + +

808.52 808.51 [PS(38:5)-H]- 0 +

808.68 808.67 [GalCer(d18:1/C24:1)-H]- - ++

810.53 810.53 [PS(38:4)-H]- 0 +

810.70 810.69 [GalCer(d18:1/C24:0)-H]- - ++

814.55 814.56 [PS(38:2)-H]- - 0

816.56 816.58 [PS(38:1)-H]- - +

818.58 818.59 [PS(38:0)-H]- - +

822.55 822.54 [ST-OH(d18:1/C18:0)-H]- 0 0

833.52 833.52 [PI(34:2)-H]- - 0

834.54 834.53 [PS(40:6)-H]- + +++

838.55 838.56 [PS(40:4)-H]- 0 ++

853.50 853.49 [PI(36:6)-H]- - 0

21

857.53 857.52 [PI(36:4)-H]- 0 +

862.62 862.61 [ST(d18:1/C22:0)-H]- + +

863.54 863.57 [PI(36:1)-H]- - 0

865.56 865.58 [PI(36:0)-H]- - 0

881.53 881.52 [PI(38:6)-H]- 0 +

883.55 883.54 [PI(38:5)-H]- 0 +

885.56 885.56 [PI(38:4)-H]- ++ +++

888.64 888.62 [ST(d18:1/C24:1)-H]- +++ +++

890.65 890.64 [ST(d18:1/C24:0)-H]- + +

894.62 894.62 [PS(44:4)-H]- 0 +

896.64 896.64 [PS(44:3)-H]- 0 ++

904.63 904.62 [ST-OH(d18:1/C24:1)-H]- ++ ++

905.53 905.52 [PI(40:8)-H]- - +

906.65 906.64 [ST-OH(d18:1/C24:0)-H]- + +

907.55 907.53 [PI(40:7)-H]- - +

908.66 908.66 [PI-Cer(t18:0/C24:0-H]- + +

909.56 909.55 [PI(40:6)-H]- 0 +

948.72 948.73 [PI-Cer(t20:0/C26:0)-H]- - +

916.68 916.69 LacCer(d18:1/C20:0)-H]- e) - 0

1091.74 1091.74 GA2(d18:1/C18:0)-H]- e) - +

1.179.75 1.179.73 [GM3(d18:1/C18:0)-H]- e) 0 +

1.207.80 1.207.78 [GM3(d18:1/C20:0)-H]- + +

1.253.77 1.253.79 [GA1(d18:1/C20:0)-H]- e) - ++

1.281.83 1.281.81 [GA1(d18:1/C20:0)-H]- - +

1.382.84 1.382.82 [GM2(d18:1/C18:0)-H]- e) - 0

1.544.89 1.544.87 [GM1(d18:1/C18:0)-H]- e) 0 0

1572.93 1572.90 [GM1(d18:1/C20:0)-H]- 0 0

1857.98 1857.97 [GD1(d18:1/C18:0)+Na-2H]- e) 0 0

1873.96 1873.95 [GD1(d18:1/C18:0)+K-2H]- 0 0

1885.99 1885.98 [GD1(d18:1/C20:0)+Na-2H]- 0 0

1901.98 1901.96 [GD1(d18:1/C20:0)+K-2H]- 0 0 a) the presence of isobaric compounds is possible; for sphingolipids, the most likely composition of the sphingosine

(most typically d18:1) is shown. b) 0: intensity per pixel too low for MS imaging; ion is detected in sum spectra recorded from 300 pixels

+: detected with low signal intensity per pixel; meaningful MS images obtained

++: good MS images obtained

+++ / ++++: excellent/outstanding MS images with high S/N ratios and image contrast obtained c) cLPA: cyclic lysophosphatidic acid d) LBPA: lysobisphosphatidic acid e) GM1: Neu5Acα2-3(Galβ1-3GalNAcβ1-4)Galβ1-4Glcβ1Cer)

GM2: GalNAcβ1-4Gal(Neu5Acα2-3)1-4Glcβ1Cer

GM3: Neu5Acα2-3Galβ1-4Glcβ1Cer

GA1: asialo-GM1 (presumably produced by the MALDI-2-MS analysis)

GA2: asialo-GM2 (presumably produced by the MALDI-2-MS analysis) LacCer (lactosylceramide) asialo-GM3 (presumably produced by the MALDI-2-MS analysis)

GD1: either GD1a, GD1b or GD1c

GD1a: Neu5Acα2-3Galβ1-3GalNAcβ1-4(NeuAcα2-3)Galβ1-4Glcβ1Cer

GD1b: Galβ1-3GalNAcβ1-4(Neu5Acα2-8Neu5Acα2-3)Galβ1-4GlcβCer

GD1c Neu5Acα2-8Neu5Acα2-3Galβ1-3GalNAcβ1-4Galβ1-4GlcβCer

22

Table S3.

Experimental and calculated m/z values, proposed structure, type of ionization and signal

intensity of abundant matrix signals detected by conventional MALDI and MALDI-2

mass spectrometry in (A) the positive and (B) the negative ion mode using the DHB and

norharmane matrices, respectively. The spectra were accumulated by applying 20 laser

shots/pixel on 144 and 300 adjacent positions, respectively, on matrix-coated cerebellum.

A DHB, positive ion mode

Exp.

m/z

Calc.

m/z

Tentatively

proposed identity

Ion counts

MALDI MALDI-2

109.03 109.03 [m-COOH]+• 200 2,400

110.04 110.04 [m-COOH+H]+ 9 8,600

136.02 136.02 [m-H2O]+• 82 95,000

137.02 137.02 [m-H2O+H]+ 4,300 683,000

154.03 154.03 m+• 21 21,000

155.03 155.03 [m+H]+ 92 27,000

272.03 272.03 [2(m-H2O)]+• 17 3,000

273.04 273.04 [2(m-H2O)+H]+ 3,150 112,000

290.04 290.04 [m+(m-H2O)]+• n/d 7,600

291.05 291.05 [m+(m-H2O)+H]+ n/d 5,200

409.06 409.06 [3(m-H2O)+H]+ 740 9,000

Exp.

m/z

Calc.

m/z

Tentatively

proposed identitya)

Ion counts

MALDI MALDI-2

165.08 165.05 [m-3H]- 2,400 710,000

167.06 167.06 [m-H]- 48,000 4,400,000

168.07 168.07 m•- 17,000 1,500,000

331.10 331.10 [2m-5H]- 550 11,000

333.12 333.11 [2m-3H]- 7,200 400,000

499.17 499.17 [3m-5H- 130 2,600

501.18 501.18 [3m-3H]- 100 3,500

503.20 503.20 [3m-H]- 100 3,600 a) only the most abundant signals of an ion group (e.g. [3m-H]-, [3m-H]- and [3m-5H]- for trimeric ions) are listed.

B Norhamane, negative ion mode

n/d: not detected

23

References and Notes 1. F. Hillenkamp, J. Peter-Katalinić, Eds., MALDI MS—A Practical Guide to

Instrumentation, Methods and Applications (Wiley, Weinheim, Germany, ed. 2, 2013).

2. K. Dreisewerd, The desorption process in MALDI. Chem. Rev. 103, 395–426 (2003). Medline doi:10.1021/cr010375i

3. R. Knochenmuss, R. Zenobi, MALDI ionization: The role of in-plume processes. Chem. Rev. 103, 441–452 (2003). Medline doi:10.1021/cr0103773

4. T. W. Jaskolla, M. Karas, Compelling evidence for Lucky Survivor and gas phase protonation: The unified MALDI analyte protonation mechanism. J. Am. Soc. Mass Spectrom. 22, 976–988 (2011). Medline doi:10.1007/s13361-011-0093-0

5. M. Stoeckli, P. Chaurand, D. E. Hallahan, R. M. Caprioli, Imaging mass spectrometry: A new technology for the analysis of protein expression in mammalian tissues. Nat. Med. 7, 493–496 (2001). Medline doi:10.1038/86573

6. K. Chughtai, R. M. A. Heeren, Mass spectrometric imaging for biomedical tissue analysis. Chem. Rev. 110, 3237–3277 (2010). Medline doi:10.1021/cr100012c

7. N. Goto-Inoue, T. Hayasaka, N. Zaima, M. Setou, Imaging mass spectrometry for lipidomics. Biochim. Biophys. Acta 1811, 961–969 (2011). Medline doi:10.1016/j.bbalip.2011.03.004

8. S. Guo, Y. Wang, D. Zhou, Z. Li, Significantly increased monounsaturated lipids relative to polyunsaturated lipids in six types of cancer microenvironment are observed by mass spectrometry imaging. Sci. Rep. 4, 5959 (2014). Medline

9. K. A. Zemski Berry, J. A. Hankin, R. M. Barkley, J. M. Spraggins, R. M. Caprioli, R. C. Murphy, MALDI imaging of lipid biochemistry in tissues by mass spectrometry. Chem. Rev. 111, 6491–6512 (2011). Medline doi:10.1021/cr200280p

10. A. Römpp, S. Guenther, Y. Schober, O. Schulz, Z. Takats, W. Kummer, B. Spengler, Histology by mass spectrometry: Label-free tissue characterization obtained from high-accuracy bioanalytical imaging. Angew. Chem. Int. Ed. 49, 3834–3838 (2010). Medline doi:10.1002/anie.200905559

11. A. Zavalin, E. M. Todd, P. D. Rawhouser, J. Yang, J. L. Norris, R. M. Caprioli, Direct imaging of single cells and tissue at sub-cellular spatial resolution using transmission geometry MALDI MS. J. Mass Spectrom. 47, 1473–1481 (2012). Medline doi:10.1002/jms.3108

12. J. L. Norris, R. M. Caprioli, Analysis of tissue specimens by matrix-assisted laser desorption/ionization imaging mass spectrometry in biological and clinical research. Chem. Rev. 113, 2309–2342 (2013). Medline doi:10.1021/cr3004295

24

13. E. J. Lanni, S. S. Rubakhin, J. V. Sweedler, Mass spectrometry imaging and profiling of single cells. J. Proteomics 75, 5036–5051 (2012). Medline doi:10.1016/j.jprot.2012.03.017

14. B. Spengler, Mass spectrometry imaging of biomolecular information. Anal. Chem. 87, 64–82 (2015). Medline doi:10.1021/ac504543v

15. R. Knochenmuss, L. V. Zhigilei, What determines MALDI ion yields? A molecular dynamics study of ion loss mechanisms. Anal. Bioanal. Chem. 402, 2511–2519 (2012). Medline doi:10.1007/s00216-011-5194-x

16. K. Dreisewerd, M. Schürenberg, M. Karas, F. Hillenkamp, Influence of the laser intensity and spot size on the desorption of molecules and ions in matrix-assisted laser desorption/ionization with a uniform beam profile. Int. J. Mass Spectrom. Ion Process. 141, 127–148 (1995). doi:10.1016/0168-1176(94)04108-J

17. J. Schiller, R. Süss, J. Arnhold, B. Fuchs, J. Lessig, M. Müller, M. Petković, H. Spalteholz, O. Zschörnig, K. Arnold, Matrix-assisted laser desorption and ionization time-of-flight (MALDI-TOF) mass spectrometry in lipid and phospholipid research. Prog. Lipid Res. 43, 449–488 (2004). Medline doi:10.1016/j.plipres.2004.08.001

18. K. Dreisewerd, Recent methodological advances in MALDI mass spectrometry. Anal. Bioanal. Chem. 406, 2261–2278 (2014). Medline doi:10.1007/s00216-014-7646-6

19. L. Hanley, R. Zimmermann, Light and molecular ions: The emergence of vacuum UV single-photon ionization in MS. Anal. Chem. 81, 4174–4182 (2009). Medline doi:10.1021/ac8013675

20. C. H. Beckercor, K. J. Wufn, On the photoionization of large molecules. J. Am. Soc. Mass Spectrom. 6, 883–888 (1995). Medline doi:10.1016/1044-0305(95)00472-P

21. P. Nemes, A. Vertes, Laser ablation electrospray ionization for atmospheric pressure, in vivo, and imaging mass spectrometry. Anal. Chem. 79, 8098–8106 (2007). Medline doi:10.1021/ac071181r

22. J. Soltwisch, J. Souady, S. Berkenkamp, K. Dreisewerd, Effect of gas pressure and gas type on the fragmentation of peptide and oligosaccharide ions generated in an elevated pressure UV/IR-MALDI ion source coupled to an orthogonal time-of-flight mass spectrometer. Anal. Chem. 81, 2921–2934 (2009). Medline doi:10.1021/ac802301s

23. H. Kettling, S. Vens-Cappell, J. Soltwisch, A. Pirkl, J. Haier, J. Müthing, K. Dreisewerd, MALDI mass spectrometry imaging of bioactive lipids in mouse brain with a Synapt G2-S mass spectrometer operated at elevated pressure: Improving the analytical sensitivity and the lateral resolution to ten micrometers. Anal. Chem. 86, 7798–7805 (2014). Medline doi:10.1021/ac5017248

24. A. Rohlfing, A. Leisner, F. Hillenkamp, K. Dreisewerd, Investigation of the desorption process in UV matrix-assisted laser desorption/ionization with a liquid 3-nitrobenzyl alcohol matrix by photoacoustic analysis, fast-flash imaging, and UV-

25

laser post-ionization. J. Chem. Phys. C 114, 5367–5381 (2010). doi:10.1021/jp905251r

25. Further experimental details can be found in the supplementary online materials.

26. J. Y. Zhang, D. S. Nagra, L. Li, Measurement of gas-phase ultraviolet-visible absorption spectra of thermally labile molecules with a pulsed rapid heating technique for sample vaporization. Anal. Chem. 63, 2995–2999 (1991). doi:10.1021/ac00024a039

27. Q. Lin, R. Knochenmuss, Two-photon ionization thresholds of matrix-assisted laser desorption/ionization matrix clusters. Rapid Commun. Mass Spectrom. 15, 1422–1426 (2001). Medline doi:10.1002/rcm.380

28. J. Soltwisch, T. W. Jaskolla, F. Hillenkamp, M. Karas, K. Dreisewerd, Ion yields in UV-MALDI mass spectrometry as a function of excitation laser wavelength and optical and physico-chemical properties of classical and halogen-substituted MALDI matrixes. Anal. Chem. 84, 6567–6576 (2012). Medline doi:10.1021/ac3008434

29. A. N. Krutchinsky, B. T. Chait, On the nature of the chemical noise in MALDI mass spectra. J. Am. Soc. Mass Spectrom. 13, 129–134 (2002). Medline doi:10.1016/S1044-0305(01)00336-1

30. J. Yang, R. M. Caprioli, Matrix sublimation/recrystallization for imaging proteins by mass spectrometry at high spatial resolution. Anal. Chem. 83, 5728–5734 (2011). Medline doi:10.1021/ac200998a

31. B. Tissot, N. Gasiunas, A. K. Powell, Y. Ahmed, Z. L. Zhi, S. M. Haslam, H. R. Morris, J. E. Turnbull, J. T. Gallagher, A. Dell, Towards GAG glycomics: Analysis of highly sulfated heparins by MALDI-TOF mass spectrometry. Glycobiology 17, 972–982 (2007). Medline doi:10.1093/glycob/cwm072

32. R. Taguchi, M. Ishikawa, Precise and global identification of phospholipid molecular species by an Orbitrap mass spectrometer and automated search engine Lipid Search. J. Chromatogr. A 1217, 4229–4239 (2010). Medline doi:10.1016/j.chroma.2010.04.034

33. K. Dreisewerd, R. Lemaire, G. Pohlentz, M. Salzet, M. Wisztorski, S. Berkenkamp, I. Fournier, Molecular profiling of native and matrix-coated tissue slices from rat brain by infrared and ultraviolet laser desorption/ionization orthogonal time-of-flight mass spectrometry. Anal. Chem. 79, 2463–2471 (2007). Medline doi:10.1021/ac061768q

34. M. Jain, S. Ngoy, S. A. Sheth, R. A. Swanson, E. P. Rhee, R. Liao, C. B. Clish, V. K. Mootha, R. Nilsson, A systematic survey of lipids across mouse tissues. Am. J. Physiol. Endocrinol. Metab. 306, E854–E868 (2014). Medline doi:10.1152/ajpendo.00371.2013

35. G. Paxinos, K. B. J. Franklin, The Mouse Brain in Stereotaxic Coordinates (Elsevier, ed. 4, 2013).

26

36. N. Goto-Inoue, T. Hayasaka, N. Zaima, M. Setou, The specific localization of seminolipid molecular species on mouse testis during testicular maturation revealed by imaging mass spectrometry. Glycobiology 19, 950–957 (2009). Medline doi:10.1093/glycob/cwp089

37. T. W. Jaskolla, K. Onischke, J. Schiller, 2,5-Dihydroxybenzoic acid salts for matrix-assisted laser desorption/ionization time-of-flight mass spectrometric lipid analysis: Simplified spectra interpretation and insights into gas-phase fragmentation. Rapid Commun. Mass Spectrom. 28, 1353–1363 (2014). Medline doi:10.1002/rcm.6910

38. P. Franceschi, Y. Dong, K. Strupat, U. Vrhovsek, F. Mattivi, Combining intensity correlation analysis and MALDI imaging to study the distribution of flavonols and dihydrochalcones in Golden Delicious apples. J. Exp. Bot. 63, 1123–1133 (2012). Medline doi:10.1093/jxb/err327

39. A. V. Loboda, I. V. Chernushevich, Investigation of the mechanism of matrix adduct formation in MALDI at elevated pressure. Int. J. Mass Spectrom. 240, 101–105 (2005). doi:10.1016/j.ijms.2004.10.011