Nonacid cleavable detergents applied to MALDI mass spectrometry profiling of whole cells

Article abstract of DOI:10.1002/jms.914

Although cleavable detergents were first synthesized a number of years ago, they have only recently been successfully applied to problems involving biological molecules. Recent reports have demonstrated that these compounds are useful for applications inv

Full text of DOI:10.1002/jms.914

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2005; 40: 1319–1326
Published online 11 October 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.914
Nonacid cleavable detergents applied to MALDI mass
spectrometry profiling of whole cells
Jeremy L. Norris, Matthew J. Hangauer, Ned A. Porter and Richard M. Caprioli∗
Departments of Chemistry and Department of Biochemistry, Vanderbilt University, Nashville, TN, USA
Received 13 May 2005; Accepted 24 July 2005
Although cleavable detergents were first synthesized a number of years ago, they have only recently been
successfully applied to problems involving biological molecules. Recent reports have demonstrated that
these compounds are useful for applications involving both 2D PAGE and mass spectrometry. However,
most cleavable surfactants have utilized acid-labile functional groups to affect cleavage. In applications
where extreme pH is required, acid cleavable detergents have limited usefulness. We report the synthesis
of fluoride cleavable silane compounds and photolabile cinnamate esters as cleavable detergents having
alternative cleavage chemistries than previously reported cleavable detergents. These compounds were
applied to whole cell analysis using MALDI mass spectrometry, and it was demonstrated that their use
results in an increase in the number of proteins analyzed by increasing protein solubility. Copyright 
2005 John Wiley & Sons, Ltd.
KEYWORDS: MALDI; mass spectrometry; cleavable detergent; molecular profiling; membrane protein
Although cleavable detergents have properties similar
to conventional detergents,10–12 they can undergo specific
cleavage to yield products with reduced surface activity.
Although the concept of cleavable detergents was introduced
many years ago, few applications have focused on their
potential use in biological systems. Attempts have been made
to synthesize cleavable detergents suitable for the analysis
of proteins using two-dimensional gel electrophoresis.13–15
These were largely ineffective until the recent introduction
of an acid-labile surfactant (ALS)15 based on a structure
first synthesized by Jaeger et al.16 These compounds enable
two-dimensional gel analyses of comparable quality to
that of traditional SDS-PAGE while allowing subsequent
protein identification by chromatography and tandem mass
spectrometry.15 Acid cleavable detergents have also been
demonstrated to be useful for the mass spectrometric analysis
of intact membrane proteins and analyses requiring the
direct profiling of cells and tissue with matrix-assisted
laser desorption ionization mass spectrometry (MALDI-
MS).17,18 These applications demonstrate that the increased
solubility induced by the presence of cleavable detergents
results in an increased signal-to-noise ratio and a greater
number of proteins observed in a given MALDI-MS profile.
The benefits of increased solubility are realized without
the corresponding loss in mass spectral quality normally
associated with the addition of conventional detergents.
Currently, all reported cleavable detergents used for the
mass spectrometric analysis of proteins have been acid-labile.
However, applications of mass spectrometry preceded by
isolation steps that require extreme pH, such as acid or
base extraction to remove posttranslational modifications,
hydrogen–deuterium exchange, and separation techniques
such as isoelectric focusing that expose the sample to a
INTRODUCTION
Detergents are critical additives necessary for the isolation
and solubilization of hydrophobic biomolecules. Because of
their amphiphilic nature, detergents enhance the solubility
of hydrophobic molecules in aqueous solution by shield-
ing hydrophobic solutes from unfavorable interactions with
water. Detergents are an effective substitute for the lipids
constituting the natural environment of hydrophobic pro-
teins. This allows proteins to be isolated in a functional
state, enabling many biochemical experiments that would
otherwise be impossible.1 Nevertheless, it is not always
advantageous to use detergents for protein solubilization
because they aggregate to form noncovalent complexes with
the desired analyte and thereby interfere with analytical char-
acterization of the biomolecules. One specific result of this
phenomenon is that detergents exacerbate the ion suppres-
sion that occurs in mass spectrometry of mixtures containing
proteins and detergents.2–7 Previous research suggests that
there are no inherent properties of membrane proteins that
prevent analysis using mass spectrometry provided that the
solubility problems can be overcome.2,3,8,9 Thus, reagents
that allow the extraction and solubilization of hydrophobic
proteins while allowing for the analysis of the extracted pro-
tein using mass spectrometry would be of great value to the
field.
ŁCorrespondence to: Richard M. Caprioli, Mass Spectrometry
Research Center, Vanderbilt University, Room 9160 MRB3, 465 21st
Avenue South, Nashville, TN 37232-8575, USA.
E-mail: R.Caprioli@vanderbilt.edu
Contract/grant sponsor: NIH/NIGMS; Contract/grant numbers:
GM 58008-05; NCI CA 86243-02; NSF 4-20-430-3381.
Contract/grant sponsor: NIH Training Grant in Molecular
Biophysics; Contract/grant number: NIH T32 GM08320.
Copyright  2005 John Wiley & Sons, Ltd.

1320
J. L. Norris et al.
(t, 3H; OCH2CH3), υ D 0.10 (t, 2H; SiCH2CH2), υ D 0.07 (s,
6H; SiMe2). 13C NMR (300 MHz, CDCl3): υ D 172.8, 59.6, 33.4,
31.8, 29.3, 25.4, 23.2, 22.6, 18.3, 15.2, 14.3, 14.0 and ꢀ3.3.
Figure 1. Structure of fluoride cleavable detergents used in
MALDI-MS analyses.
2-(Dimethyloctylsilanyl)ethanol (6)
Preparation of this compound was accomplished using
previously described methodology.19,20 An amount of 2.05 g
(7.75 mmol) of compound 5 was refluxed with 0.55 g
(14.3 mmol) of lithium aluminum hydride in 50 ml of ether
for 1 h. After cooling to room temperature, 0.55 ml of water,
0.55 ml of 15% NaOH, and 1.65 ml of water was added
sequentially with stirring. The precipitates were removed
by filtration through celite. The alcohol was produced in an
Figure 2. Photolabile detergent, (E)-3-(2,4,6-trihydroxyphenyl)
acrylic acid octyl ester.
range of pH environments, will not be feasible using acid-
labile detergents. Additionally, if the analyte of interest is
sensitive to pH changes (e.g. easily hydrolyzed) or if it must
be purified in an extreme low or high pH environment
because of its limited solubility, acid cleavable detergents
will be of limited utility. This report evaluates two different
classes of compounds as alternatives to acid-labile detergents
and reports the application of these compounds to protein
analysis using MALDI-MS. Two types of cleavable functional
groups were used to design these new reagents, fluoride
reactive silane compounds (Fig. 1) and photolabile cinnamate
esters (Fig. 2). These reagents are stable in a wide variety of
solution conditions and can be selectively cleaved by the
introduction of fluoride or light, respectively. MALDI-MS
analysis of cell lysates prepared with these reagents show
improvement in sensitivity and spectral quality.
1
85.7% yield (1.47 g). H NMR (300 MHz, CDCl3): υ D 3.71
(t, 2H; CH2OH), υ D 1.24 (m, H; alkyl), υ D 0.96 (t, 2H;
SiCH2CH2OH), υ D 0.86 (t, 3H; CH2CH2CH3), υ D 0.49 (m,
2H; SiCH2CH2), υ D 0.06 (t, 1H; CH2CH2OH), υ D ꢀ0.02 (s,
6H; SiCH3). 13C NMR (300 MHz, CDCl3): υ D 60.2, 33.6, 31.9,
29.3, 29.2, 23.8, 22.7, 20.9, 15.4, 14.1 and ꢀ3.1.
Bromoacetic acid 2-(dimethyloctylsilanyl)ethyl ester (8)
A volume of 0.658 ml (3.05 mmol) of 2-(dimethyloctylsi-
lanyl)ethanol (6), 0.267 ml (3.3 mmol) of pyridine, 0.122 g
(1 mmol) of N,N-dimethylaminopyridine, and 20 ml of
methylene chloride were placed in a 50-ml round-bottom
flask. The reaction mixture was placed under inert atmo-
sphere. A volume of 0.290 ml bromoacetyl bromide (7)
was added drop wise to the reaction mixture. The reac-
tion proceeded for 2 h at room temperature. The reaction
mixture was washed twice with 1 M HCl followed by a
wash with saturated NaCl. The organic layer was dried over
MgSO4. A quantitative yield of 1.01 g of bromoacetic acid 2-
(dimethyloctylsilanyl)ethyl ester (8) was obtained. 1H NMR
(300 MHz, CDCl3): υ D 4.24 (t, 2H, CH2OC(O)), υ D 3.78
(s, 2H, C(O)CH2Br), υ D 1.23 (m, 12H, alkyl), υ D 1.01 (t,
2H, SiCH2CH2O), υ D 0.85 (t, 3H, CH2CH2Me), υ D 0.55
(m, 2H, CH2CH2CH2Si), υ D ꢀ0.01 (s, 6H, SiMe2). 13C NMR
(300 MHz, CDCl3): υ D 167.3, 64.8, 33.5, 31.9, 29.2, 29.2, 26.1,
23.6, 22.6, 15.9, 15.2, 14.1 and ꢀ3.3.
EXPERIMENTAL
Synthesis
All compounds were synthesized according to the methods
outlined in the following text. All reagents were purchased
from Sigma-Aldrich (St. Louis, MO) and used as indicated
unless stated otherwise.
(Dimethyloctylsilanyl)acetic acid ethyl ester (5)
Preparation of this compound was accomplished using
previously described methodology.19,20 All reagents were
distilled prior to use. A dry 250 ml flask was fitted with
a condenser and an addition funnel. An amount of 2.03 g
(31.0 mmol) of powdered zinc was placed in the flask under
inert atmosphere. To a mixture of 65 ml benzene and 15 ml
diethyl ether, 3.46 g (16.77 mmol) of chlorodimethyloctyl
silane (3) and 2.4 ml (21.6 mmol) of bromoacetic acid ethyl
ester (4) were added and the reagents were placed in an
addition funnel. Approximately, 10 ml of the reagent mixture
was added to the zinc; initiation of the reaction was evident
after approximately 5 min. The additional reagents were
added drop wise over 30 min. The reaction was allowed to
proceed for 20 h at room temperature before quenching with
40 ml of 1 M HCl. The organic layer was further washed with 1
M HCl, water, saturated sodium bicarbonate, and water, and
dried over magnesium sulfate. Compound 5 was produced
in 82.5% yield (3.57 g). 1H NMR (300 MHz, CDCl3): υ D 4.07
(t, 2H; OCH2CH3), υ D 1.86 (s, 2H; C(O)CH2), υ D 1.22 (m,
10H; Alk), υ D 0.86 (t, 5H; SiCH2CH2, CH2CH2CH3), υ D 0.57
[2-(Dimethyloctylsilanyl)ethoxycarbonylmethyl]-
trimethylammonium bromide (1a)
Trimethyl amine (9a) was condensed over 0.88 g of com-
pound 8 in a pressure tube. The tube was sealed and the
reaction was stirred overnight at room temperature. A brown
precipitant formed within an hour. The tube was cooled to
°
ꢀ78 C, opened and the reaction was allowed to return to
room temperature. When the trimethyl amine had evap-
orated, 0.89 g (2.24 mmol) of product (1a) remained as a
1
brown solid (86.1%). H NMR (300 MHz, DMSO): υ D 4.40
(s, 2H; CH2NMe3), υ D 4.25 (t, 2H; CH2OC(O)), υ D 3.13
(s, 9H; NMe3), υ D 1.25 (m, 3H; Alkyl), υ D 0.99 (t, 2 H;
SiCH2CH2O), υ D 0.82 (t, 3H; CH2CH2Me), υ D 0.51 (t, 2H;
CH2CH2Si), υ D 0.01 (s, 6H; SiCH3). 13C NMR (300 MHz,
DMSO): υ D 164.8, 64.19, 62.6, 53.1, 32.9, 31.3, 28.7, 28.6,
23.2, 22.1, 15.5, 14.6, 14.0 and ꢀ3.3. HRMS (ESI) calcd for
C17H38NO2Si [MC], 316.2666; found, 316.2645.
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 1319–1326

Nonacid cleavable detergents for MALDI-MS 1321
131–132 C. ꢀmax(MeOH) D 329 nm.1 H NMR (d4-methanol)
υ D 0.869–0.891 (br s, 3H), 1.300–1.349 (br s, 10H),
1.631–1.699 (br s, 2H), 4.099–4.143 (t, 2H), 5.871 (s, 2H),
6.694–6.748 (d, 1H, J D 16.2 Hz), 8.061–8.115 (d, 1H,
J D 16.2 Hz); 13C NMR (d4-methanol) υ D 14.43, 23.70, 27.15,
30.36, 30.38, 32.97, 65.16, 95.43, 104.16, 115.37, 139.00, 161.35,
162.08 and 171.82; HRMS (ESI) calcd for C17H24O5 [M C Na]C,
331.1516; found, 331.1521.
°
[2-(Dimethyloctylsilanyl)ethoxycarbonylmethyl]-
trimethylphosphonium bromide (1b)
A volume of 2.11 ml trimethyl phosphine (1 M in THF)
(9b) and 7 ml benzene was added to 0.647 g (1.92 mmol) of
compound 8 in a round-bottom flask. The reaction was stirred
for 48 h at room temperature. The trimethyl phosphine
and benzene were removed in vacuo. A weight of 0.60 g
(1.76 mmol) of product (1b) remained as a light brown solid
(92.3%).1H NMR (300 MHz, CDCl3): υ D 3.71 (t, 2H; CH2OH),
υ D 1.24 (m, H; alkyl), υ D 0.96 (t, 2H; SiCH2CH2OH), υ D 0.86
(t, 3H; CH2CH2CH3), υ D 0.49 (m, 2H; SiCH2CH2), υ D 0.06
(t, 1H; CH2CH2OH), υ D ꢀ0.02 (s, 6H; SiCH3). 13C NMR
(300 MHz, CDCl3): υ D 60.2, 33.6, 31.9, 29.3, 29.2, 23.8, 22.7,
20.9, 15.4, 14.1 and ꢀ3.1. HRMS (ESI) calcd. for C17H38O2PSi
[MC], 333.2373; found, 333.2361.
5,7-Dihydroxycoumarin (19)
A 13 mM solution of (E)-3-(2,4,6-trihydroxyphenyl)acrylic
acid octyl ester (2) in CD3OD was photolyzed at 330 nm
in a NMR tube by an Oriel Spectral Energy GM 252
Spectrophotometer for 30 h after which the tube contained
5,7-dihydroxycoumarin (19) and 1-octanol. The n-octanol
was removed by high vacuum with a dry ice/acetone trap
over 24 h. The 5,7-dihydroxycoumarin (19) NMR spectra
corresponded with literature spectra.21
Octyloxycarbonylmethyl-triphenyl-phosphonium bromide
(13)
n-Octanol (7.0 g, 53.9 mmol), pyridine (4.78 ml, 59.3 mmol)
and N,N-dimethylaminopyridine (DMAP) (3.28 g, 26.9
Determination of critical micelle concentration
Critical micelle concentrations (CMCs) were determined by
the DuNo¨y ring method using a manual tensiometer. All
CMC measurements were made in water.
°
mmol) were stirred in CH2Cl2 (100 ml) at 0 C. Bromo-
acetylbromide (7) (5.2 ml, 59.3 mmol) was added drop wise,
°
after which the mixture was stirred for 30 min at 0 C. The
yellow suspension was diluted with additional CH2Cl2 and
extracted with water, 1.0 M aqueous HCl, saturated aqueous
NaHCO3 and finally water. The organic layer was dried with
MgSO4, filtered and concentrated to give crude bromoester
(11) (12.2 g, 90%), which was added to EtOAc (150 ml)
with Ph3P (12) (19.5 g, 74.13 mmol) at room temperature
for two days. Excess ether was then slowly added to induce
precipitation. The salt was filtered and dried affording a
white solid (23.27 g, 84% based on n-octanol). An analytical
sample was recrystallized from EtOAc to afford a white
Fluoride cleavable detergents
Solutions of protein solubilized in fluoride cleavable deter-
gent were prepared for MALDI analysis using Slide-A-
Lyzer MINI (Pierce) dialysis units with a molecular weight
cut-off (MWCO) of 3500. A volume of 20 µl of each sam-
ple was dialyzed against 0.5
M
potassium fluoride in
33% isopropanol for 18 h. The potassium fluoride was
removed from the sample by further dialysis against 3 : 2 : 1
water/isopropanol/formic acid for 1 h. A 0.25 µl aliquot is
removed for MALDI analysis. This sample is cocrystallized
on target with an equal volume of sinapinic acid (20 mg/ml
in 50 : 50 acetonitrile/0.1% TFA). To determine whether the
cleavage is complete, aliquots were taken periodically and
tested in order to determine the extent of cleavage. Aliquots
were prepared for analysis using the dried-droplet method.
Excess potassium fluoride was removed by washing the
MALDI spot with 0.1% TFA after matrix crystals were formed
on the surface. The intact detergent was monitored using
MALDI-MS.
1
°
solid. m.p. 122 C; H NMR (CDCl3, 300 MHz) υ D 7.86–7.61
(m, 15H, ꢀPPh3), 5.52 (d, 2H, PCH2, 13.8 Hz), 3.87 (t, 2H,
OCH2, J D 6.7 Hz), 1.39 (p, 2H, OCH2CH2, J D 7.1 Hz),
1.15 (m, 10H, alkyl), 0.85 (t, 3H, Me, J D 6.7 Hz); 13C NMR
(CDCl3, 300 MHz) υ D 164.8, 135.6, 134.2, 133.6, 130.7, 118.2
(d, J D 384 Hz), 67.3, 31.9, 29.3, 28.4, 25.9, 22.9, and 14.4; 31
P
NMR (CDCl3, 300 MHz) υ D 22.11; HRMS (ESI) calcd for
C28H34BrO2P [M C Na]C, 433.2291; found, 433.2286.
(E)-3-(2,4,6-Trihydroxyphenyl)acrylic acid octyl ester (2)
An aliquot of 1.887 g of octyloxycarbonylmethyl-triphenyl-
phosphonium (13) in dichloromethane was washed (3ð)
with saturated aqueous NaHCO3. The product was dried
over MgSO4 and filtered, and the organic layer was con-
centrated in vacuo to yield 1.880 g (7.203 mmol) of the ylide,
(triphenyl-ꢀ5-phosphanylidene)acetic acid octyl ester, as a
yellow oil. Immediately thereafter, 875.3 mg (5.684 mmol,
1 equiv) of 2,4,6-trihydroxybenzaldehyde (14) (Aldrich) was
added to the 1.880 g of (triphenyl-ꢀ5-phosphanylidene)acetic
acid octyl ester (7.203 mmol, 1.27 equiv), and the Wittig reac-
tion was stirred in DMF for 50 h in the dark. The DMF
was removed by high vacuum with a dry ice/acetone
trap over 24 h. The resultant product mixture was a dark
brown/purple oil. This oil was chromatographed with silica
gel using 6% methanol in dichloromethane to isolate 1.753 g
(5.688 mmol) of pure (E)-3-(2,4,6-trihydroxyphenyl)acrylic
acid octyl ester (2), an orange solid in 54% yield. m.p.
Photolabile detergents
A solution of protein solubilized in the photolabile detergent
was prepared for MALDI analysis using the dried-droplet
method. A volume of 0.5 µl of each detergent was deposited
on the MALDI target along with 0.5 µl of 20 mg/ml of
sinapinic acid in 50% acetonitrile. The detergent can be
cleaved by incubation of the spot under an ultraviolet lamp
during sample preparation or sufficient exposure to ambient
room light. Samples containing the photolabile detergent
are prepared for analysis using the dried-droplet method.22
Confirmation of the mechanism of cleavage of the photolabile
detergent was determined using NMR. A 0.5 ml solution of
a 13 mM solution of the photolabile detergent was prepared
in d4-methanol and placed in an NMR tube. A proton NMR
confirmed the structure of the unreacted cinnamate. The
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 1319–1326

1322
J. L. Norris et al.
sample was removed and irradiated at 330 nm in a NMR
tube by an Oriel Spectral Energy GM 252 Spectrophotometer.
The sample was analyzed by NMR at 1, 3, 5, 8, 21, 24, and
30 h. The identity of 5,7-dihydroxycoumarin is confirmed
by comparison with the literature values (see the preceding
text). Additionally, detergent cleavage is confirmed using
MALDI-MS.
Protein mass spectrometry
Samples were analyzed using an Applied Biosystems
Voyager DE STR (Framingham, MA) equipped with a
nitrogen laser (ꢀ D 337 nm) operating at 20 Hz. Twenty
shots were averaged per spectrum, and these spectra were
accumulated for a total of 300 shots per spectrum. Samples
were measured using a linear mode method optimized for
best resolution at m/z 12 000.
Preparation of cell lysates
RESULTS AND DISCUSSION
The human colorectal cell line, RKO, was grown in DMEM
Hi glucose (Gibco BRL, Gaithersburg, MD) medium sup-
plemented with 10% fetal calf serum (Gemini Bio-Products,
Inc., Calabasas, CA) and 1% penicillin/streptomycin (Sigma,
St. Louis, MO). Three cell extracts containing each cleavable
detergent were prepared along with three separate extracts
containing commercial detergents. Control cell extracts were
lysed in RIPA lysis buffer (50 mM Tris (pH 7.4), 150 mM NaCl,
1% (v/v) Nonidet P-40, 0.1% (v/v) SDS, 0.5% (w/v) sodium
deoxycholate, 50 mM NaF, 0.1 mM NaV, 1 mM dithiothreitol,
and the protease inhibitors antipain (10 µg/ml), leupeptin
(10 µg/ml), pepstatin A (10 µg/ml), chymostatin (10 µg/ml),
phenylmethylsulfoyl fluoride (50 µg/ml) (Sigma), and 4-
(2-aminoethyl)benzenesulfonylfluoride (200 µg/ml) (Calbio-
chem-Novabiochem Corp.)). Cell extracts containing cleav-
able detergents were prepared using a modified version of
RIPA buffer that excludes the detergent components SDS,
Nonidet P-40, and sodium deoxycholate. Experiments with
detergent 1 were prepared without sodium fluoride to avoid
premature detergent cleavage. Solutions containing the pho-
tolabile detergent 2 were prepared in the dark. Additional
controls were also prepared excluding all detergents. Cells
were incubated for 1 h in the presence of lysis buffer followed
by homogenization. Cell debris was removed by centrifuga-
tion at 3000 ð g for 30 min. For the fluoride labile detergent
(1b), solutions containing 1ð, 2ð, 5ð, 10ð, and 15ð CMCs
were prepared. For the photolabile detergent (2), solutions
containing 0.1, 1, and 2 mM were prepared. Solubility of
compound 2 prevents testing at higher concentrations. The
supernatant was prepared as described in the previous sec-
tions and was analyzed using MALDI-MS.
Fluoride cleavable detergents were designed in accor-
dance with the general structure in Fig. 1. Compounds
based on a silyl ethyl ester group are known to undergo
elimination when they react with fluoride ions result-
ing in degradation of the detergent. This cleavable linker
is based on
a commonly used protecting group, 2-
trimethylsilanyl ethanol. A hydrophobic chain of eight
carbons was chosen as the hydrophobic tail. Synthe-
sis of compound 1 was accomplished using the proce-
dure outlined in Scheme 1. The synthesis of compound
1 was accomplished in ca 78% overall yield in four
steps. The first step is a Reformatsky coupling19,20 of
chlorodimethyloctyl silane (3) with bromoacetic acid ethyl
ester (4) utilizing a zinc catalyst to form (dimethyloctylsi-
lanyl)acetic acid ethyl ester (5). Compound 8 is reduced
to 2-(dimethyloctylsilanyl)ethanol (6) using
a LiAlH4
reduction.19,20 The product 2-(dimethyloctylsilanyl)ethanol
(6) was converted to bromoacetic acid 2-(dimethyloctyl-
silanyl)ethyl ester (8) by reaction with bromoacetyl bromide
(7). Reaction of trimethyl amine (9a) or trimethyl phosphine
(9b) completes the synthesis of [2-(dimethyloctylsilanyl)-
ethoxycarbonylmethyl]trimethylammonium bromide (1a)
and
[2-(dimethyloctylsilanyl)ethoxycarbonylmethyl]tri-
methylphosphonium bromide (1b) respectively.
The photolabile compound, (E)-3-(2,4,6-trihydroxy-
phenyl)acrylic acid octyl ester (2), was designed to have a
high degree of structural homology with that of sinapinic acid
when cleaved. Toward that end, a phenolic compound was
chosen to comprise the hydrophilic head of the detergent.
Photochemically reactive molecules of this type have been
Scheme 1. Synthesis of fluoride cleavable detergents. (a) Zn, benzene/diethylether, RT, 20 h; (b) LiAlH4, ether, reflux, 1 h; (c) DMAP,
pyridine, CH2Cl2, 2 h.
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 1319–1326

Nonacid cleavable detergents for MALDI-MS 1323
Scheme 2. Synthesis of the photolabile detergent, (E)-3-(2,4,6-trihydroxyphenyl)acrylic acid octyl ester. (a) DMAP, pyridine, CH2Cl2,
°
0 C, 30 min; (b) EtAc, 48 h; (c) triethyl amine, DMF, 48 h.
previously described.23 The synthesis of compound 2 was
performed according to Scheme 2. Octyloxycarbonylmethyl-
triphenyl-phosphonium bromide (13) was synthesized by
the reaction of bromoacetyl bromide (7) with n-octanol (10)
to form bromoacetic acid octyl ester (11). Compound 11
is further reacted with triphenyl phosphine (12), yield-
ing octyloxycarbonylmethyl-triphenyl-phosphonium bro-
mide (13). The phosphonium salt undergoes a Wittig
reaction with 2,4,6-trihydroxybenzaldehyde (14) to yield
the final product, (E)-3-(2,4,6-trihydroxyphenyl)acrylic acid
octyl ester (2) in 45% overall yield.
To characterize the detergent properties of the novel
cleavable compounds synthesized in this study, the CMCs
were measured. Although the CMC for a given molecule does
not necessarily correlate with detergent efficacy, it is useful
for estimating the effective concentration of the detergent. It
is commonly regarded as good practice to use a concentration
that is at or near the CMC for the extraction and solubilization
of proteins.24 The CMC was computed using surface
tension measurements made using the DuNo¨y ring method.
Compound 1 has a high degree of structural homology
with the commercial detergent cetyltetramethyl-ammonium
bromide (CTAB) (CMC = 1.0 mM).25 This structural similarity
is also reflected by the similarity of surface-active properties.
The CMC of these compounds were measured to be 1.17 and
1.30 mM for compounds 1a and 1b respectively. Compound
2 has no measurable CMC; this compound has limited
solubility in water. The solubility limit was reached before
an inflection in the surface tension plot could be observed.
Compounds 1a and 1b are sensitive to reaction
with fluoride ions. The resulting cleavage products are
alkyldimethylsilyl fluoride (15) and ethylene (16), which are
sufficiently volatile to evaporate under a high vacuum envi-
ronment prior to introduction into the mass spectrometer.
These components are not present during the final analysis
at levels sufficient to cause interference with MALDI. The
polar head of the molecule (17) is a water soluble organic
salt that remains in the sample unless an additional cleanup
step is added. The head of the detergent is easily removed
using dialysis or by adding a washing step to the MALDI
sample preparation procedure. The mechanism of cleavage
of compound 1 is shown in Scheme 3. Cleavage of the pro-
tecting group after which these compounds were modeled,
2-(trimethylsilyl)ethyl ester, is very fast (<1 h) and conver-
sion is quantitative. However, cleavage of compounds 1a
and 1b in aqueous solutions is markedly slower. A 1ð
CMC solution of compounds 1a and 1b failed to be com-
pletely cleaved at 24 h in 0.5 M potassium fluoride. This
reduced reaction rate relative to the 2-(trimethylsilyl)ethyl
ester can be attributed to two phenomena. First, the steric
hindrance provided by the octyl chain provides a barrier
to the reaction of fluoride ion with the silane. Second, in
aqueous solution the detergent forms micellar structures
that shield the hydrophobic part of the detergent from inter-
actions with water. Since the silane is in the interior of the
micelle, it becomes unlikely that a fluoride ion present in the
micelle exterior would be able to react with the silane. This
phenomenon has been reported previously for acid cleav-
able surfactants.10–12 Phase transfer catalysts such as crown
ethers and tetrabutylammonium fluoride might be used to
increase the speed of cleavage by transporting fluoride ions
into the micelle, but these compounds are detrimental to
MALDI sample preparation. In spite of the slow cleavage
of compounds 1a and 1b, they can be easily adapted to
applications in MALDI-MS. Protein samples are prepared
for MALDI analysis by dialyzing the sample against 0.5 M
potassium fluoride for 18 h. Figure 3 shows the results of
using this sample preparation procedure. The top spectrum
shows that there is no detectable amount of the detergent, [2-
(dimethyloctylsilanyl)ethoxycarbonylmethyl]trimethyl-
phosphonium bromide (1b) (MC D 333.8), remaining in the
sample using this method. When this result is contrasted with
Scheme 3. Cleavage of fluoride cleavable detergents.
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 1319–1326

1324
J. L. Norris et al.
100
80
60
40
20
0
KF 18 hours
% Trans Cinnamate
% Cis Cinnamate
% Coumarin
O
+
Si
P
O
Control
800 1000
0
200
400
600
0
5
10
15
20
25
30
35
Mass (m/z)
Time (hours)
Figure 3. Fluoride cleavable detergent (a) with and (b) without
treatment with KF. The detergent solutions were dialyzed
against these respective solutions for 18 h.
Figure 4. Photolysis of (E)-3-(2,4,6-trihydroxyphenyl)acrylic
acid octyl ester (2): reaction monitored using 1H-NMR during
photolysis (330 nm). Conversion of the trans-cinnamate to the
cis-cinnamate (18) is observed, followed by conversion of the
cis-cinnamate to coumarin (19).
the same sample dialyzed against water for 18 h, much of the
detergent remains associated with the protein. Although pre-
vious experiments indicate that the cleavage is not complete
at 18 h, the combination of detergent cleavage and dilution
of the detergent lowers the concentration below the CMC.
At these concentrations, aggregation and micellization are
eliminated, allowing the detergent to be readily removed.
The photolabile detergent, (E)-3-(2,4,6-trihydroxy-
phenyl)acrylic acid octyl ester (2), cleaves by the mech-
anism shown in Scheme 4. Compounds analogous to 2
have been shown to cleave by the mechanism shown.23
To determine if this compound cleaves through the pro-
posed mechanism, the photolysis reaction in d4-methanol
was monitored using NMR spectroscopy. Compound 2 was
photolyzed in an NMR tube at 330 nm using a photochem-
ical bench. NMR spectra were obtained at different time
intervals. The relative abundance of each species is plot-
ted in Fig. 4, showing the progress of the reaction. Figure 4
demonstrates the progress of the reaction from (E)-3-(2,4,6-
trihydroxyphenyl)acrylic acid octyl ester (2) to the interme-
diate, (Z)-3-(2,4,6-trihydroxyphenyl)acrylic acid octyl ester
(18), to the formation of 5,7-dihydroxycoumarin (19).
detergent is not detected. The product coumarin is observed
in the low-mass MALDI spectra of all samples, both
irradiated and nonirradiated. Preparation of the sample
under an ultraviolet lamp (ꢀ D 254 nm) ensures that the
detergent cleaves completely; however, little difference is
seen when comparing with a sample prepared in ambient
room light. Therefore, to simplify sample preparation, the
samples were prepared using the dried-droplet method in
ambient light.
The primary application for which the cleavable deter-
gents in this study were designed is to improve the analysis
of complex protein mixtures using direct cell and tissue
profiling by MALDI-MS. The expectations for overall perfor-
mance were similar to previous studies using acid cleavable
compounds;17,18 enhanced signal-to-noise ratio, and the addi-
tion of new peaks in the MALDI mass spectrometric profile
+
Cinnamate (M+H) = 309
(a)
(b)
The extent of the cleavage can also be observed in the
MALDI mass spectrum; however, it is difficult to determine
the exact state of the detergent on the MALDI plate prior to
irradiation because the N2 laser (ꢀ D 337 nm) causes partial
cleavage of the detergent during desorption/ionization.
Figure 5 confirms that when a sample is prepared under
ultraviolet light (ꢀ D 254 nm), the molecular ion of the
300
310
320 300
310
320
Figure 5. MALDI mass spectrum of (E)-3-(2,4,6-
trihydroxyphenyl)acrylic acid octyl ester (2) (a) with irradiation
(ꢀ D 254 nm) and (b) without irradiation.
Scheme 4. Mechanism of cleavage of (E)-3-(2,4,6-trihydroxyphenyl)acrylic acid octyl ester (13).
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 1319–1326

Nonacid cleavable detergents for MALDI-MS 1325
should be observed because of increased protein solubil-
ity and extraction from the cell. The effect of the cleavable
detergents on the analysis of whole cell extracts of the human
colorectal cell line RKO is shown in Figs 6 and 7. These figures
highlight the strengths of each class of cleavable detergent by
focusing on the mass range in which the greatest advantages
were observed. An expanded view of the high molecular
weight regions is also shown for clarity in Fig. 6.
Examination of the protein profiles in Fig. 6 reveals
that the signal-to-noise ratio of the proteins found in
the sample increases dramatically when compound 1 is
used as a detergent relative to the control experiment
where no detergent was used. This effect is attributed
to an increase in protein solubility upon the addition of
the cleavable detergent to the lysis buffer. In addition to
the increase in signal-to-noise ratio, additional peaks are
observed in the spectrum. In the low molecular weight
range (ca 5–10 kDa), the greatest benefits are observed when
the detergent concentrations are kept at or close to the
CMC for the detergent used. Unique ions are observed
in the mass range after detergent treatment; however,
as the detergent concentration increases, these peaks are
suppressed. At higher detergent concentrations (CMC ½ 5),
a number of unique high molecular weight proteins can
be observed above 20 kDa. These proteins contribute to
the ion suppression of lower molecular weight proteins.
Therefore, the concentration of compound1 must be carefully
considered depending on the range of molecular weights one
is interested in profiling. Similar results are also observed for
the photolabile compound 2. Figure 7 shows the resulting
profiles. The high detergent concentration in the control
sample extracted using RIPA severely suppresses the signals
in the MALDI spectrum. Additional peaks are observed
in the extract prepared without detergent; however, many
proteins are absent from this analysis because they failed
to be extracted without the use of detergents. Applying the
photolabile detergent to extract proteins from RKO cells
provides the necessary surfactant properties to aid in the
extraction and solubilization of proteins while reducing
the influence of the detergent on the MALDI process. The
primary effects observed are that the overall signal intensity
of the peaks common to both the photolabile detergent
extract and the control extracts are enhanced, and a number
of peaks that were previously unseen are observed while
5x
15x CMC
10x CMC
5x CMC
2x CMC
1x CMC
No Detergents
SDS/NP-40/SD
3000
15400
27800
40200
52600
65000
Mass (m/z)
Figure 6. MALDI-MS analysis of RKO cell lysates prepared using compound 1b: cell lysates prepared using various concentrations
of the Fꢀ cleavable detergent, 1b, were analyzed using MALDI-MS and compared to control extracts prepared using RIPA
(containing detergents SDS, NP-40, sodium deoxycholate)without detergent.
1 mM
Photolabile
Detergent
No Detergents
SDS/NP-40/SD
4000
6400
8800
11200
13600
16000
Mass-to-charge
Figure 7. MALDI-MS analysis of RKO cell lysates using photolabile detergents. Cell lysates prepared using the photolabile
detergent, 2, were analyzed using MALDI-MS and compared to control extracts prepared using RIPA (containing detergents SDS,
NP-40, sodium deoxycholate)without detergent.
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 1319–1326

1326
J. L. Norris et al.
5. Jeannot MA, Zheng J, Li L. Observation of sodium gel-induced
protein modifications in dodecylsulfate polyacrylamide gel
electrophoresis and its implications for accurate molecular
weight determination of gel-separated proteins by matrix-
assisted laser desorption ionization time-of-flight mass
spectrometry. J. Am. Soc. Mass Spectrom. 1999; 10: 512.
6. Zhang N, Li L. Ammonium dodecyl sulfate and an alternative
to sodium dodecyl sulfate for protein sample preparation with
improved performance in MALDI mass spectrometry. Anal.
Chem. 2002; 74: 1729.
7. Puchades M, Westman A, Blennow K, Davidsson P. Removal of
sodium dodecyl sulfate from protein samples prior to matrix-
assisted laser desorption/ionization mass spectrometry. Rapid
Commun. Mass Spectrom. 1999; 13: 344.
8. Ghaim JB, Tsatsos PH, Katsonouri A, Mitchell DM, Salcedo-
Hernandez R, Gennis RB. Matrix-assisted laser desorption
ionization mass spectrometry of membrane proteins:
using cleavable detergents. Higher concentrations of 2 could
not be tested because of limited solubility.
This study demonstrates two different strategies that can
be used to make cleavable detergents. These detergents dif-
fer from the existing reagents in that they are not based on
acid-labile functional groups to carry out cleavage of the
detergent. The reagents reported here provide a viable alter-
native to acid cleavable detergents, and the conditions of their
use provide results that are directly comparable to other acid
cleavable detergents.17,18 The criteria that must be considered
when selecting a cleavable detergent are the nature of the
solute, the effect of the cleavage reagents and products on the
MALDI process, the required stability of the cleavable deter-
gent, and the time required for processing. Though effective
for the extraction, solubilization, and analysis of proteins
using MALDI-MS, acid-labile detergents have only limited
stability when pH differs significantly from physiological
conditions. The fluoride cleavable reagents, compound 1,
described in this paper are more stable to pH extremes
than acid cleavable detergents; however, the time required
to degrade and remove the sample from solution may be
prohibitive to some applications. The photolabile detergent,
compound 2, has the advantage that no chemical reagent is
required for cleavage, reducing potential signal suppression
that could be problematic when using detergents cleaved
by acid or fluoride. The photolabile detergent described,
however, is a less effective detergent than the other deter-
gents described. The most important consideration one must
have is that the proposed analyte must be solubilized by
the detergent chosen, so regardless of the time required, a
more effective detergent might be necessary. When choos-
ing to use a cleavable detergent, it is important to choose a
reagent that is appropriate to the analyte properties and the
analytical task to be accomplished. As detergents continue
to be adapted to new experiments, the number of limitations
imposed on researchers will be reduced enabling a variety
of new experiments to be performed.
demonstration of
a simple method to determine subunit
molecular weights of hydrophobic subunits. Biochim. Biophys.
Acta 1997; 1330: 113.
9. Cadene M, Chait BT. A robust, detergent-friendly method for
mass spectrometric analysis of integral membrane proteins. Anal.
Chem. 2000; 72: 5655.
10. Hellberg P-E, Bergstrom K, Holmberg K. Cleavable surfactants.
J. Surf. and Deterg. 2000; 3: 81.
11. Holmberg K. Cleavable surfactants. In Novel Surfactants:
Preparation, Applications, and Biodegradability, Holmberg K (ed).
Marcel Dekker: New York, 1998; 333.
12. Jaeger DA. Cleavable Surfactants. Supramol. Chem. 1995; 5: 27.
13. Epstein WW, Jones DS, Bruenger E, Rilling HC. The synthesis of
a detergent and its use in the isolation and characterization of
protein. Anal. Biochem. 1982; 119: 304.
14. Dunkin IR, Gittinger A, Sherrington DC, Whittaker P.
A
photodestructible surfactant. J. Chem. Soc. Chem. Commun. 1994;
19: 2245.
15. Meng F, Cargile BJ, Patrie SM, Johnson JR, McLoughlin SM,
Kelleher NL. Processing complex mixtures of intact proteins
for direct analysis by mass spectrometry. Anal. Chem. 2002; 74:
2923.
16. Jaeger DA, Jamrozik J, Golich TG, Clennan MW, Mohebalian J.
Preparation and characterization of glycerol-based cleavable
surfactants and derived vesicles. J. Am. Chem. Soc. 1989; 111:
3001.
17. Norris JL, Porter NA, Caprioli RM. Mass spectrometry of
intracellular and membrane proteins using cleavable detergents.
Anal. Chem. 2003; 75: 6642.
Acknowledgements
18. Norris JL, Porter NA, Caprioli RM. Combination detergent/
MALDI matrix: functional cleavable detergents for mass
spectrometry. Anal. Chem. 2005; 77: 5036.
The authors would like to acknowledge Lucy Tang and Jennifer
Pietenpol for providing the RKO cells. This work was supported
by NIH/NIGMS GM 58008-05, NCI CA 86243-02, and NSF 4-20-
430-3381. JLN acknowledges support from NIH Training Grant in
Molecular Biophysics, NIH T32 GM08320.
19. Stangier P, Palcic MM, Bundle DR. 2-(Dimethyloctylsilanyl)-
ethyl lactoside:
a versatile intermediate for chemical and
enzymatic ganglioside synthesis. Carbohydr. Res. 1995; 267: 153.
20. Ardah AA, Aisa AMA, Richter H. Darstellung von langkettigen
2-(Dimethylalkylsilyl)ethylsacchariden. Pharmazie 2000; 55: 741.
21. Hirakura K, Saida I, Fukai T, Nomura T. Mulberroside-B, a new
C-glucosylcoumarin from the cultivated Mu’berry tree (Morus-
Lhou koidz) 22:2,5-dihydroxybenzoic acid- a new matrix for laser
desorption ionization mass-spectrometry. Heterocycles. 1985; 23:
2239.
REFERENCES
1. Stryer L. In Biochemistry, 4th ed. W. H. Freeman: New York, 1995;
1064.
2. Schnaible V, Michels J, Zeth K, Freinang J, Welte W, Bu¨hler S,
Glocker MO, Przybylski M. Approaches to the characterization
of membrane channel proteins (porins) by UV MALDI-MS. Int.
J. Mass Spectrum. Ion Process. 1997; 169/170: 165.
22. Strupat K, Karas M, Hillenkamp F. Int. J. Mass Spectrom. Ion
Process. 1991; 111: 89.
23. Turner AD, Pizzo SV, Rozakis G, Porter NA. Photochemical
activation of acylated ˛-thrombin. J. Am. Chem. Soc. 1988; 109:
1274.
24. Womack MD, Kendall DA, MacDonald RC. Detergent effects on
enzyme activity and solubilization of lipid bilayer membranes.
Biochim. Biophys. Acta 1983; 210: 210.
3. Rosinke B, Strupat K, Hillenkamp F, Rosenbusch J, Dencher N,
Kru¨ger U, Galla H-J. Matrix-Assisted Laser Desorption/Ioni-
zation Mass Spectrometry (MALDI-MS) of membrane proteins
and non-covalent complexes. Int. J. Mass Spectrom. 1995; 30: 1462.
4. Bo¨rnsen KO, Gass MAS, Bruin GJM, Adrichem JHMv, Biro MC,
Kresbach GM, Ehrat M. Influence of solvents and detergents on
matrix-assisted laser desorption/ionization mass spectrometry
measurements of proteins and oligonucleotides. Rapid Commun.
Mass Spectrom. 1997; 11: 603.
25. Biological detergents table of physical properties, 2003; Sigma-
Aldrich,
http://www.sigmaaldrich.com/img/assets/4202/
Detergent table2edited.pdf.
ꢀ
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 1319–1326