Novel vacuum stable ketone-based matrices for high spatial resolution MALDI imaging mass spectrometry

Article abstract of DOI:10.1002/jms.4277

We describe the use of aromatic ketones and cinnamyl ketones that have high vacuum stability for analyzing tissue sections using matrix-assisted laser desorption/ionization imaging mass spectrometry. Specifically, the matrix, (E)-4-(2,5-dihydroxyphenyl)but-3-en-2-one (2,5-cDHA) provides high sensitivity and high vacuum stability while producing small size crystals (1-2?μm). A high throughput and highly reproducible sample preparation method was developed for these matrices that first involves using an organic spray solution for small matrix crystal seeding followed by spraying of the matrix in a 30% acetonitrile/70% water solution on the tissue surface to obtain a homogeneous coating of small crystals, suitable for high spatial resolution imaging.

Full text of DOI:10.1002/jms.4277

Novel Vacuum Stable Ketone Based Ma-
trices for High Spatial Resolution
MALDI Imaging Mass Spectrometry
Junhai Yang†, Jeremy L. Norris† and Richard Caprioli†‡*
Mass Spectrometry Research Center, National Research Resource for Imaging Mass
Spectrometry and Departments of †Biochemistry, ‡Pharmacology, Medicine and Chem-
istry, Vanderbilt University, Nashville, TN.
ABSTRACT: We describe the use of aromatic ketones and cinnamyl ketones that have
high vacuum stability for analyzing tissue sections using MALDI imaging mass spec-
trometry. Specifically, the matrix, (E)-4-(2,5-dihydroxyphenyl)but-3-en-2-one (2,5-
cDHA) provides high sensitivity and high vacuum stability while producing small size
crystals (1-2 µm). A high throughput and highly reproducible sample preparation
method was developed for these matrices that first involves using an organic spray solu-
tion for small matrix crystal seeding followed by spraying of the matrix in a 30% ace-
tonitrile/70% water solution on the tissue surface to obtain a homogeneous coating of
small crystals, suitable for high spatial resolution imaging.
This article hasbeen accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/jms.4277
This article is protected by copyright. All rights reserved.

Matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry (IMS) has become a highly
effective technology to produce molecularimages of compoundscontained in biologicaltissuesections.1,2 MALDI
IMS has been used to map the localization of a broad spectrum of analytes including proteins, peptides, drugs
and theirmetabolites as well as endogenous cell metabolites and lipids3. Theability to correlatepotentially thou-
sands of molecules with histological tissue features observed using microscopy and IMS is vital to our under-
standing of complex biological samples.
For high spatial resolution MALDI IMS(<10 µm), itis importantto havea tightly focused laser beam4-6, a sam-
plepreparation method that is capableof attaining high sensitivity and an instrumentcapableof fastdata acqui-
sition. Thematrix coating should be homogeneous and the diameter of matrix crystals should bemuch smaller
than that of the laser spot diameter on target. High spatial resolution generally results in longanalysis time(>2
hours), and therefore, a vacuum stable MALDI matrix coating is required to ensure consistency of the result
throughout the experiment. Furthermore, for high throughput applications, the sample preparation protocol
mustbe simple, fast and highly reproducible.
The matrices commonly used forprotein analyses are sinapinic acid (SA), 2’,5’-dihydroxybenzoic acid (DHB),
ferulic acid (FA) and α-cyano-4-hydroxycinnamic acid (CHCA)7,8. These matrices perform well when imaging
with low spatial resolutions (>30 µm), butthesecompounds arenot acceptableat spatial resolutions of 5-10 µm
or less. Wehaverecently used 2’,5’-dihydroxyacetophenone(2,5-DHA) for theimageanalysis of proteins athigh
spatial resolutions (2.5 and 5 µm)6,9. This matrix provides greatersensitivityforproteins than SA, CHCA and DHB
and the crystal size of 2,5-DHA usingspray application is around 1-3 µm6. Unfortunately, 2,5-DHA is volatileand
readily sublimes from thetissueunderinstrumentvacuum conditions and as a result, can only beused forshort
acquisition times (<1 hour) in high vacuum ion sources. Although data processing methods such as total-ion-
current normalization of spectracan be used to mitigatethe fadingof the ion intensities thatoccurs through loss
of matrix, this “fix” is farfrom ideal and reduces detection of somechemical species.
In the work reported in the current paper, we describe a matrix that is vacuum stable over many hours, ame-
nable forimaging analytes at high spatial resolution, and is suitablefor use with robotic spraying. Weevaluated
matrices structurally similar to 2,5-DHA, and also tested cinnamyl ketones that are structural analogs of com-
monly used matrices forprotein imaging.
. EXPERIMENTAL
Materials and methods
This article is protected by copyright. All rights reserved.

Ethanol, acetone, chloroform, acetonitrile(ACN), toluene, acetic acid (AcOH), activated charcoal, trifluoracetic
acid (TFA) and 2’,5’-dihydroxyacetophenone(2,5-DHA), 2’,6’-dihydroxyacetophenone(2,6-DHA),2’,4’-dihydrox-
yacetophenone (2,4-DHA), (E)-4-(4-hydroxy-3-methoxyphenyl)but-3-en-2-one (HMP), 2’,6’-dihydroxybenzal-
dehyde, 2’,5’-dihydroxybenzaldehyde, ferulic acid ((E)-3-(4-hydroxy-3-methoxy-phenyl)prop-2-enoic acid, FA)
and syringaldehydewere purchased from Sigma-Aldrich (St. Louis, MO);2,5-DHA was purified by recrystalliza-
tion two times with 90% ACN at 70°C. (Acetylmethylene)triphenylphosphorane was obtained from TCI America;
Xylenewas purchased from Acros (Morris Plains,NJ). Conductiveindium tin oxide(ITO) coated microscopeglass
slides (2.5 x 7.5 x 1.1 mm, CG-81IN-S115) were purchased from DeltaTechnologies (Stillwater, MN). A sublima-
tion apparatus was constructed with a condenserand flask from Chemglass LifeSciences (Vineland, NJ). Robotic
spraying was performed using a TM-Sprayer™ (HTX Technologies, Carrboro, NC). MilliQ water was obtained
from a MilliQ AdvantageA10 system.
Carnoy’s fluid was prepared using a 6:3:1 ratio by volume of ethanol: chloroform: acetic acid.10 Fresh frozen
mousebrains were purchased from Pel-Freez Biologicals (Rogers, AZ); ratkidney, rat testis and rat brains were
purchased from Bioreclamation, LLC (Hicksville, NY).
Microscope images of matrix coated targets wereacquired usinga Nikon Eclipse90i microscope. SEM pictures
of matrix coated targets were acquired using a Zeiss Merlin instrument with a Gemini II Column from Vanderbilt
Institute of NanoscaleScienceand Engineering.
Synthesis of cinnamyl ketonematrices:
(E)-4-(2,5-dihydroxyphenyl)but-3-en-2-one (2,5-cDHA), (E)-4-(2,6-dihydroxyphenyl)but-3-en-2-one (2,6-
cDHA) and (E)-4-(4-hydroxy-3,5-dimethoxyphenyl)but-3-en-2-one(HDMP) wereprepared as described below.
In a 500 mL round bottom flask, 11.3 mmol of thecorrespondingaldehyde(2,5-dihydroxybenzaldehydefor2,5-
cDHA, 2,6-dihydroxybenzaldehyde for 2,6-cDHA, or syringaldehyde for HDMP) was mixed with 10.7 mmol
(8.30g) (acetylmethylene)triphenylphosphorane and 300 mL of anhydrous toluene. The mixture was agitated
using a magnetic stirring bar for 48 hours at room temperature. The target products (2,5-cDHA, 2,6-cDHA or
HDMP) were insoluble and precipitated from reaction solution by filtration under vacuum resulting in crude
products of ~70% yield. 2,5-cDHA was recrystallized using 8% ACN (detail procedure described below). 2,6-
cDHA was recrystallized with 20% ACN at60°C and HDMPwas recrystallized with 30% ACN at70°C usingcom-
mon organic recrystallization procedures (briefly, the mixture is heated on a hotplate until it becomes a clear
solution then is cooled to room temperature)11,12. NMR spectrawere acquired (data in supplementary material)
and confirmed with the literature values for2,5-cDHA13 and HDMP14, or predicted NMR for 2,6-cDHA usingAd-
vanced Chemistry Development, Inc. (ACD/Labs) SoftwareV11.01 from Scifinder.cas.org.
This article is protected by copyright. All rights reserved.

Recrystallization of 2,5-cDHA:
Briefly, 1.5 gof crude 2,5-cDHA was dissolved in 17 mL of 95% acetonitrileat 70°C and 500 mg of activechar-
coal was added to the solution. The mixture was stirred for another 5 min and then filtered under vacuum to
obtain a light-yellow solution which was poured into 200 mL of Milli-Q water under room temperature. The
solution was placed at 4°C overnightfor the crystallization of 2,5-cDHA.Thecrystals were then filtered, washed
with icecold Milli-Q water, and allowed to dry to obtain 1.1 g of purified matrix.
UV/vis spectraof matrices
Five mg of each matrix was dissolved in 1 mL of ethanol, and 10 µL of this solution was mixed with 1990 µL
ethanol to obtain 200 time dilution. OnemL of the diluted solution was used to measure the UV/vis absorbance
from 220 nm to 420 nm at 25 °Cusing a SpectraMax® M2 microplatereader (MolecularDevices). Theextinction
coefficients was calculated using the formula: A355=Ɛlc (A355, absorbance at 355 nm; Ɛ, extinction coefficients; l,
length of optical path=1 cm; c, molarconcentration of compound).
For measurementof the solid film,5 µL of each matrix solution (200 timediluted) was pipetted on one sideof
optical window of UV/vis cuvetteand left to dry. Thecuvette is a 1.5 mL semimicro methacrylatecuvettefrom
Fisher Scientific (14-955-128). The absorbance of the solid film was acquired as the same way as for ethanol
solution butusingair as reference.
Evaluation of matrices usingprotein standards:
All the matrix solutions wereprepared by dissolving15 mgmatrix in 800 µL of H2O, 200 µL acetonitrileand 5
µL of TFA. Then 5 µL of each matrix solution was mixed with 5 µL of protein standard solution. The protein
standard contains: oxidized insulin chain b from bovine pancreas (m/z: 3497), insulin from porcine pancreas
(m/z: 5779), ubiquitin from bovine erythrocytes (m/z: 8561), cytochrome c from equine heart (m/z: 12385),
apomyoglobin from equineskeletal muscle(m/z:16952) and trypsinogen from bovinepancreas (m/z: 23982).
Theprotein standard/matrix mixture (0.4 µL of each) was spotted on a BrukerMTPAnchorChip 384-well target
(Bruker Daltonics, Billerica, MA). The MALDI spectra were acquired on Bruker Autoflex Speed MALDI TOF in-
strument (Bruker Daltonics). The spectra of four locations (100 laser shots each location) having the highest
intensity were summed during acquisition and the resulting spectrum was baseline subtracted using Bruker
FlexAnalysis 3.3.
Matrix stability measurements:
Ketone matrices (2,5-DHA, 2,5-cDHA, 2,6-DHA, 2,6-cDHA, HMP, HDMP) that produced high quality spectra
with protein standards were coated on ITO slides by spraying usingacetonitrile as the solvent(5 mg/mL) with
the TM Sprayer™. The solutions were sprayed ata flow rate of 150 μL/min with nitrogen as the nebulization gas
This article is protected by copyright. All rights reserved.

(set to 10 psi). Thespray nozzlewas housed insideaheated block maintained at 30°C. Thespray was deposited
at a linear velocity of 1200 mm/min with an offset line spacing of 1.5 mm. Eight passes were performed. On
alternating passes, thespray pattern was changed by 90° to achieve a crossed pattern. The distanceof the spray
nozzle from the MALDI target was approximately 40 mm.
The weights of the slides were measured using a balance (Mettler Toledo MS105DU Semi-Micro Analytical
Balance) before matrix coating, after the matrix coating and after storage inside the instrument under vacuum
(4.4E-7 mbar) for various timeperiods.
Tissuesectioningand storage:
Frozen tissuesections weresectioned at 12 µm thickness usingaLeicaCM3050 cryostat(LeicaMicrosystems
GmbH, Wetzlar, Germany), thaw mounted on ITO-coated microscopeslides, and dried underambientconditions
for 10 minutes. The slides were placed in a plastic slide holder box (Electron microscope science, Hatfield, PA)
sealed with parafilm and stored at -80°C. When ready foranalysis, theslidebox was placed into a vacuum desic-
cator under ambientconditions for30 minutes to allow the slides to reach room temperature whilepreventing
water condensation on the sections which could causedelocalization of analytes.
Matrix deposition onto tissuesections by robotic spraying:
Tissue sections mounted on ITO slides were sequentially rinsed in glass petri dishes (100 x 15 mm, Ther-
moFisher) with aseries of 10 mL of thefollowingsolutionsusingamodified method described previously15:70%
ethanol for 30 s, 100% ethanol for 30 s, Carnoy’s fluid for2 min, 100% ethanol for30 s, 40% ethanol for30 s and
100% ethanol for30 s. This washing procedure removed excess salts and lipids. Theslides were allowed to dry
for 15 min in a hood prior to matrix application.
Spray application of 2,5-cDHA on tissuesections
Step 1: Thirty mg of 2,5-cDHA was dissolved in 5 mL ethyl acetate and 5 mL of toluene. This solution was
sprayed on the slides containing the washed tissue sections with the TM Sprayer™: 95°C nozzle temperature,
600 mm/min nozzlevelocity, 1.5 mm track spacing, 0.05 mL/min flow rate, and 4 passes.
Step 2: Ninety mg2,5-cDHA was dissolved in 3 mL ACN, 7 mL H2O, 100 µL TFAand 50 µL ammonium hydroxide
followed by robotic sprayingof the slides from step 1: 95°C nozzletemperature, 1100 mm/min nozzlevelocity,
1.5 mm track spacing, 0.05 mL/min flow rate and 8 passes.
Spray application of HMP, HDMP, and 2,6-cDHA on tissuesections
For these three matrices, the tissue was first seeded with dry powdered matrix. Briefly, 50 mg of the matrix
was ground by usinga mortar/pestle to obtain ~5 µm diameter crystals and the matrix powder was applied on
the surface of the tissuesection using a fine paint brush. Thebrush was loaded with matrix powder and shaken
This article is protected by copyright. All rights reserved.

abovethe surfaceof the tissuesection to release the powder until the surfaceof the tissues was totally covered.
The excess powder on the tissue was removed by blowing a stream of nitrogen across the surface. This step is
importantfor obtaininga consistentcoatingquality in terms of crystal size (10-20 µm) and coverageof thetissue
surface16. The dry powder “seeded” slides were then sprayed with HMP, HDMP, and 2,6-cDHA respectively by
the two-step method used for 2,5-cDHA.
SA coatingfor the serial sections of 2,5-cDHA coated sections forthecomparison:
Thesesections were rinsed usingthe sameprocedures as for the 2,5-cDHA samples and coated with SA using
sublimation (0.15mg/cm2), and then recrystallized with 1 mL of water and 100 µL TFA at 85 °C.15
Imaging Mass spectrometry
MALDI IMS experiments were performed using a Bruker Rapiflex Tissuetyper (Bruker Daltonics) equipped
with a Smartbeam 3D laser at 10 k repetition rate and FlexControl 3.3 software. Protein analysis was performed
in positiveion linearmode. Ion images wereprocessed usingFlexImaging5 and spectral analysis was performed
with FlexAnalysis 3.3. Thelaserbeam dimension was setto thesame value as thespatial resolution. Thenumber
of lasershots at each pixel was 600 when imaged at a 15 µm spatialresolutionand 300when imaged at 10 and 5 µm
spatialresolution.
Hematoxylin and eosin stain for imaged sections.
FollowingMALDI imaging analysis, theslides wererinsed with ethanol and acetone to removethe matrix. The
tissue sections were then subjected to a standard hematoxylin and eosin (H&E) staining procedure.17 H&E
stained sections werescanned usinga Leicaslidescanner(SCN 400) at 20 X magnification.
. RESULTS AND DISCUSSION
Ketone matrices
2,5-DHA was introduced as a matrix for protein detection two decades ago,18 enablinghighly sensitiveMALDI
analysis of peptides, proteins and glycoproteins.19 This compound provides higher sensitivity for protein anal-
yses than that of SA6. However, 2,5-DHA is volatileunderhigh vacuum, limitingits usage foracquiringimages for
less than 1 hour. We investigated higher molecular weight compounds that have similar structures to 2,5-DHA
that could provide high quality images buthave better vacuum stability. Foreaseof use, a new matrix should be
suitablefor automatic spraying for high throughput applications and to achievehigh reproducibility. Weinves-
tigated compounds based on the structure of 2,5-DHA (MW 152) such as the cinnamyl version, 2,5-cDHA (MW
178) (Figure 1A) that exhibits markedly increased vacuum stability.Wealso investigated several othercinnamyl
ketone containing compounds such as HDMP (based on SA), HMP (based on FA) and 2,6-cDHA (isomer of 2,5-
This article is protected by copyright. All rights reserved.

cDHA), respectively(Figure1A). Wealsoinvestigated fourotherisomers of 2,5-DHA:2,6-DHA,2,4-DHA, 3,5-DHA
and 2-chloro-3,4-DHA. Thesynthetic schemes for 2,5-cDHA,2,6-cDHAand HDMPareshown in Figure 1B.
This article is protected by copyright. All rights reserved.

The preparations of gram scale quantities of 2,5-cDHA, 2,6-cDHA and HDMP are straightforward using the
Wittig reaction20. Theseproducts precipitate from toluene, greatly simplifyingtheirpurification. Further purifi-
cation is obtained by recrystallization.
Performance with proteins standards and UV/vis absorbanceof matrices
We tested the performance of these ketone matrices using proteins standards (Figure 2). All the matrices
tested provided high quality spectra except 2,4-DHA, 3,5-DHA, and 2-Cl-3,4-DHA. Wealso acquired UV/vis spec-
tra of these matrices in ethanol solution or as a solid film (supplementary Figure 1). TheUV/vis spectraof these
matrices arevery similar, and thus theextinction coefficients at355 nm in ethanol can beused to quicklyevaluate
these matrices. In this regard, spectra of 2,4-DHA, 3,5-DHA and 2-Cl-3,4-DHA corelate with their low extinction
coefficients (in theorderof 102) whileall othermatrices haveextinction coefficients in theorderof 103 or higher
(Table 1). As a result, we focused on 2,5-DHA, 2,5-cDHA, 2,6-DHA, 2,6-cDHA, HMP and HDMP for evaluation of
stability undervacuum and suitability of thesematrices forimagingproteins in tissuesections.
Vacuum stability of ketonematrices
The vacuum stability of thematrices was evaluated by a gravimetric measurement after incubation under in-
strument high vacuum conditions at different times throughout the process. We assessed the stability in terms
of the percentage weightloss undervacuum as shown in Figure3. Under high vacuum conditions of 4.4E-7 mbar
along with the heat present in some MALDI sources (~35°C for Rapiflex), 2,6-DHA is completely sublimated in
30 min and 2,5-DHA in 2.5 hours. HMPloses 50%of its originalcoatingon theslidein 4 hours, HDMPloses about
20% in 21 hours, and 2,6-cDHA loses 15% in 24 hours. No significantchangeof 2,5-cDHA coatingwas measured
with ourmethod after 24 hours. Since 2,6-DHA and 2,5-DHA are extremely volatileunderhigh vacuum, they were
not investigated further.
Ketone matrices for imagingproteins in tissue:
To image proteins in tissue sections athigh spatial resolution, a uniform coatingof matrix with small crystals
is necessary. We observed that ketone matrices tend to form large needle-like crystals having a length ranging
from 30 to 200 µm when they are sprayed onto tissue in 30-70% acetonitrile solution and an oversaturated
solution accumulates on the surface in the form of liquid droplets. To obtain small matrix crystals (≤ 5 µm), we
applied a “seeding” approach used in spotting matrix previously16,21 and developed a new method for spraying.
With theseedingapproach, ahomogeneous and consistentcoatingcan beobtained (examples of seeded and non-
seeded coatings areshown in Supplementary Figure 2). Duringthe seedingstep, either by powderapplication or
by non-aqueous solution spraying, small crystals are formed leading to a thin layer matrix coating. The second
step of sprayingis madewith matrix solution with 30-70% waterto extractproteins. Theconcentration of matrix
This article is protected by copyright. All rights reserved.

solution is madeto closeto thesaturation pointof thematrix, so thatafterthe solution is sprayed fromtheheated
nozzle, thedroplets of the matrix solution becomesaturated. Wehypothesizethatthe aqueous componentin the
solution extracts proteins in the tissueand the extracted proteins co-crystallizewith the matrix. It is important
to havethe solution in thesecond step solution closeto thesaturation pointof thematrix to minimizedissolution
of the seeds.
Using this seeding approach, we found that the 2,5-cDHA coatingprovides superior performance when com-
pared to 2,6-cDHA, HMPand HDMP underthesameconditions. The sizeof the matrix crystalsof 2,5-cDHAranges
from 1 to 2 µm (Figure 4 A and B), while2,6-cDHA, HMPand HDMPall provide coatings with singlecrystals in
the 10 µm range or larger with large matrix clusters on the surfaceand with gaps between these matrix clusters
(Figure 4 A). As a result, 2,5-cDHA is the only matrix among those investigated that provided a coating suitable
for high spatial resolution imaging(<10 µm) for longacquisition times.
It should be noted that 2,5-cDHA can be coated on tissue sections without using the manual powder seeding
step, but this approach gives morereproduciblefinal images. ForHMP, HDMP, and 2,6-cDHA, non-aqueous solu-
tion seedingdoes not providehomogenous small crystalson thesurface of the tissue. Forthese three matrices,
dry powder seeding is necessary to apply before the non-aqueous solution spraying.
2,5-cDHA was compared with SA for the imageacquisition of protein IMS data. SA is the mostcommonly used
vacuum stable matrix forimaging proteins as previously described5,15. Experiments werecarried out by coating
the serial sections of different tissues (mouse brain, rat testis, rat kidney and rat liver) with 2,5-cDHA and SA.
These sections were then imaged under identical conditions. As shown in Figure 5 and supplementary Figures
3-5, 2,5-cDHA is greatly superior to SA in terms of sensitivity atspatial resolutions of 5 µm and 10 µm. Mostof
the peaks in the mass spectra obtained from the tissuecoated with 2,5-cDHA have 4 times or higher peak inten-
sity than that of the samepeaks from theserial section coated with SA. Importantly, atthe higher m/z range, 2,5-
cDHA shows better performance than that of SA, e.g., proteins in the m/z 22,000 range from mouse brain, rat
testis, rat kidney and rat liver were not detected in the SA coated sample. Thecomparison conditions werecho-
sen with the best protocols forboth matrices (sprayingfor 2,5-cDHA and sublimation/recrystallization forSA)
because we found out that under identical conditions (both matrices are sprayed), the SA coatingdoes not pro-
videas many peaks at5 µm spatial resolution (Supplemental Figure7), and itwas shown thatSA performs better
with the sublimation/recrystallization protocol rather than spraying/recrystallization (Supplementary Figure
8).
The performance of 2,5-cDHA as a MALDI matrix for imaging at high spatial resolution was further demon-
strated by imaginga mousebrain section at5 µm for 7 hours (Figure 6). As the previous data would suggest, the
This article is protected by copyright. All rights reserved.

performance of 2,5-cDHA under the demanding conditions of a high spatial resolution MALDI IMS experiment
demonstrates that this matrix can be used to acquire high quality protein results whilecircumventingthe nega-
tiveimpacts that poor vacuum stability has on imagequality. Even duringthe prolonged (>7 hours) exposure to
high vacuum, no changein ionization is observed.
CONCLUSIONS:
We successfully synthesized a new matrix, 2,5-cDHA, suitable for imaging proteins at high spatial resolution
in biological tissues where data acquisition took place over many hours. This is the result of its high vacuum
stability and propensity to form small crystals (1-2 µm) homogeneouslyoverthetissuesurface. In addition, 2,5-
cDHA performs well for lower molecular weight analytes in both positive and negative ionization modes (sup-
plementary Figure 9).
Although HMPappears to have great sensitivity forimaging protein amongthematrices screened, we found it
difficultto achievethe coatinghomogeneity that high spatial resolution imaging(<5-10 µm) demands. Further-
more, HMP is somewhatvolatileunder high vacuum, even though it is more stablethan 2,5-DHA and 2, 6-DHA.
With current sprayingmethod, HMP may be useful forimagingat medium spatial resolution (~50 µm) forshort
timeperiod acquisition (~ 2.5 hr.) (supplementary Figure6 B).
Two of the other matrices that were evaluated, 2,6-cDHA and HDMP, gave poorresult for imagingproteins in
tissuesections and are not recommended usingthe matrix application method described here (supplementary
Figure 6 A and C).
SUPPORTING INFORMATION
NMR data of 2,5-cDHA, HDMPand 2,6-cDHA;UV/vis spectraof matrices in solution orsolid phase(film);differ-
ence of the 2,5-cDHA coating with/without seeding step; comparison of imaging performance of 2,5-cDHA and
SA as matrix on the serial sections of a rat testis, a rat kidney, and a rat liver; Ion images from tissue sections
coated by HDMP, HMPand 2,6-cDHA;lipid ion images of rat brain by 2,5-cDHA;
. AUTHOR INFORMATION
Corresponding Author
This article is protected by copyright. All rights reserved.

* Corresponding Author: Richard M. Caprioli (r.caprioli@vanderbilt.edu). Mass Spec-
trometry Research Center, 9160 MRB III, Vanderbilt University, Nashville, TN, 37232-
8575, USA, Tel: (+1) 615 343 9207, Fax: (+1) 615 343 8372
Notes
The authors declare no competing financial interest.
. ACKNOWLEDGMENT
This project was supported by funding from NIH/NIGMS grant 5P41 GM103391-08. The
authors thank Martin Dufresne for the acquisition of SEM pictures.
This article is protected by copyright. All rights reserved.

REFERENCES
(1) Caprioli, R. M.; Farmer, T. B.; Gile, J. Anal Chem 1997, 69, 4751-4760.
(2) Norris, J. L.; Caprioli, R. M. Proteomics. Clinical applications 2013, 7, 733-738.
(3) Aichler, M.; Walch, A. Lab Invest 2015, 95, 422-431.
(4) Zavalin, A.; Todd, E. M.; Rawhouser, P. D.; Yang, J.; Norris, J. L.; Caprioli, R. M.
Journal of Mass Spectrometry 2012, 47, 1473-1481.
(5) Zavalin, A.; Yang, J.; Haase, A.; Holle, A.; Caprioli, R. Journal of The American Society
for Mass Spectrometry 2014, 25, 1079-1082.
(6) Zavalin, A.; Yang, J.; Hayden, K.; Vestal, M.; Caprioli, R. M. Anal Bioanal Chem 2015,
407, 2337-2342.
(7) Bahr, U.; Karas, M.; Hillenkamp, F. Fresen J Anal Chem 1994, 348, 783-791.
(8) Hillenkamp, F.; Karas, M.; Beavis, R. C.; Chait, B. T. Anal Chem 1991, 63, A1193-
A1202.
(9) Yang, J.; Zavalin, A.; Caprioli, R. In 62th ASMS Conference on Mass Spectrometry and
Allied Topics: Baltimore, United States, 2014.
(10) Carnoy, J. B. La Cellule RECUEIL DE CYTOLOGIE ET D'HISTOLOGIE GÉNÉRALE,
1887; Vol. 3.
(11) Armarego, W. L. F.; Chai, C. Purification of Laboratory Chemicals, 5th ed., 2003.
(12) Rohani, S.; Horne, S.; Murthy, K. Org Process Res Dev 2005, 9, 858-872.
(13) Kuo, P.-C.; Damu, A. G.; Cherng, C.-Y.; Jeng, J.-F.; Teng, C.-M.; Lee, E. J.; Wu, T.-S.
Arch. Pharmacal Res. 2005, 28, 518-528.
(14) Joubert, P.; Beaupère, D.; Wadouachi, A.; Chateau, S.; Sangwan, R. S.; Sangwan-
Norreel, B. S. Journal of Natural Products 2004, 67, 348-351.
(15) Yang, J. H.; Caprioli, R. M. Anal Chem 2011, 83, 5728-5734.
(16) Clerens, S.; Ceuppens, R.; Arckens, L. Rapid Communications in Mass Spectrometry
2006, 20, 3061-3066.
(17) Prophet, E. B.; Armed Forces Institute of, P. Laboratory methods in histotechnology;
American Registry of Pathology: Washington, D.C., 1992.
(18) Krause, J.; Stoeckli, M.; Schlunegger, U. P. Rapid Communications in Mass
Spectrometry 1996, 10, 1927-1933.
(19) Wenzel, T.; Sparbier, K.; Mieruch, T.; Kostrzewa, M. Rapid Communications in Mass
Spectrometry 2006, 20, 785-789.
(20) Wittig, G.; Schöllkopf, U. Chemische Berichte 1954, 87, 1318-1330.
(21) Sugiura, Y.; Shimma, S.; Setou, M. Anal Chem 2006, 78, 8227-8235.
This article is protected by copyright. All rights reserved.

Table 1 Extinction coefficients of different matrices at 355 nm in ethanol
This article is protected by copyright. All rights reserved.

Figure 1 A) Structure of aromatic ketones: 2,4-DHA, 2,5-DHA, 2,6-DHA, 3,5-DHA and 2-Cl-
3,4-DHA; cinnamyl ketones: 2,6-cDHA, 2,5-cDHA, HDMP, and HMP; B) Synthesis of cin-
namyl ketones
This article is protected by copyright. All rights reserved.

Figure 2 Performance of ketone matrices for the detection of standard proteins on Bruker Au-
toflex Speed in linear positive ion mode: insulin chain b oxidized from bovine pancreas (m/z:
3497), insulin from porcine pancreas (m/z: 5779), ubiquitin from bovine erythrocytes (m/z:
8561), cytochrome c from equine heart (m/z: 12385), apomyoglobin from equine skeletal mus-
cle (m/z: 16952) and trypsinogen from bovine pancreas (m/z: 23982). For this experiment, laser
energy, concentration of standards, and matrix concentration were identical for each matrix
tested.
This article is protected by copyright. All rights reserved.

Figure 3 Stability of ketone matrices under vacuum (4.4E-7 mbar) in term of remaining matrix
as a percentage of initial mass. 2,5-cDHA, 2,6-cDHA and HDMP appear to be stable more than
24 hours.
This article is protected by copyright. All rights reserved.

Figure 4 A) Microscope pictures of 2,5-cDHA, 2,6-cDHA, HDMP and HMP on rat brain at
different zoom. 2,5-cDHA coating provides smallest crystals and homogeneity; B) SEM pic-
tures of 2,5-cDHA on rat brain
This article is protected by copyright. All rights reserved.

Figure 5 comparison of imaging performance of 2,5-cDHA and SA as matrix on the serial sections
of a mouse brain imaged at 5 µm, scale bar is 100 µm; spectra were TIC normalized. For this ex-
periment, laser energy settings were the same for both matrices. The numbers under each ion images
are m/z for the ions.
This article is protected by copyright. All rights reserved.

Figure 6 ion images mouse brain at 5 µm spatial resolution with acquisition time of 7 hours, scale
bar is 500 µm.
This article is protected by copyright. All rights reserved.