You are on page 1of 10

Ultra-performance liquid chromatography/tandem mass

spectrometric quantication of structurally diverse drug


mixtures using an ESI-APCI multimode ionization source
Kate Yu
1
*
, Li Di
2
, Edward Kerns
2
, Susan Q. Li
2
, Peter Alden
1
and Robert S. Plumb
1
1
Waters Corporation, Milford, MA, USA
2
Wyeth Research, Princeton, NJ, USA
Received 19 July 2006; Revised 3 January 2007; Accepted 4 January 2007
We report in this paper an ultra-performance liquid chromatography/tandem mass spectrometric
(UPLC
1
/MS/MS) method utilizing an ESI-APCI multimode ionization source to quantify structu-
rally diverse analytes. Eight commercial drugs were used as test compounds. Each LC injection was
completed in 1 min using a UPLC systemcoupled with MS/MS multiple reaction monitoring (MRM)
detection. Results fromthree separate sets of experiments are reported. In the rst set of experiments,
the eight test compounds were analyzed as a single mixture. The mass spectrometer was switching
rapidly among four ionization modes (ESIR, ESIS, APCIS, and APCIR) during an LC run.
Approximately 810 data points were collected across each LC peak. This was insufcient for a
quantitative analysis. In the second set of experiments, four compounds were analyzed as a single
mixture. The mass spectrometer was switching rapidly among four ionization modes during an LC
run. Approximately 15 data points were obtained for each LC peak. Quantication results were
obtained with a limit of detection (LOD) as low as 0.01 ng/mL. For the third set of experiments, the
eight test compounds were analyzed as a batch. During each LC injection, a single compound was
analyzed. The mass spectrometer was detecting at a particular ionization mode during each LC
injection. More than 20 data points were obtained for each LC peak. Quantication results were also
obtained. This single-compound analytical method was applied to a microsomal stability test.
Compared with a typical HPLC method currently used for the microsomal stability test, the
injection-to-injection cycle time was reduced to 1.5 min (UPLCmethod) from3.5 min (HPLCmethod).
The microsome stability results were comparable with those obtained by traditional HPLC/MS/MS.
Copyright # 2007 John Wiley & Sons, Ltd.
Over the past decade, the two major commercial atmospheric
pressure ionization (API) sources used for liquid chroma-
tography/mass spectrometric (LC/MS) coupling are the
electrospray ionization (ESI) source and the atmospheric
pressure chemical ionization (APCI) source.
1,2
Despite the
immense popularity, ESI and APCI each have their own
benets and limitations. ESI is especially well suited for
analytes that can be ionized in the liquid phase such as polar,
non-volatile, high molecular mass and thermally labile
compounds.
3
APCI is very effective in the analysis of
medium- and low-polarity compounds or when relative
non-polar solvents have to be used.
4
In recent years, there
have been an increasing number of reports for coupling LC
with atmospheric pressure photoionization (APPI) for the
analysis of non-polar compounds.
5
It is believed that more than 80% of compounds can be
ionized readily by a generic ESI method.
6
To analyze addi-
tional compounds, it may be necessary to choose a different
ionization mode, such as APCI or APPI. This typically
requires a physical source change and re-optimization of the
ion source parameters, which increases analytical cycle time
and makes automation impractical. To simplify the process
and enhance the analytical throughput, many attempts have
been made to analyze all target compounds by different
ionization modes during a single experiment. Examples of
such efforts include: using dual parallel mass spectrometers
to acquire ESI and APCI data simultaneously from a single
LCsystem;
7
using a combined APPI-APCI source;
8
and using
a combined ESI-APCI source,
9
etc.
The work reported in this paper is focused on the
application of a combined ESI-APCI source (ESCi
1
multi-
mode ionization source). This technique allows experiments
to be conducted simultaneously by ESI and APCI without
requiring a physical change of the ionization source. This is
accomplished by using a regular ESI probe with the APCI
corona needle placed in the source.
10
This technique has been
successfully employed by Siegel and Truebenbach
11
and
Gallagher et al.
12
in full scan mode for qualitative structural
RAPID COMMUNICATIONS IN MASS SPECTROMETRY
Rapid Commun. Mass Spectrom. 2007; 21: 893902
Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/rcm.2908
*Correspondence to: K. Yu, Waters Corporation, Milford, MA,
USA.
E-mail: Kate_yu@waters.com
Copyright # 2007 John Wiley & Sons, Ltd.
identications. No quantication studies using this tech-
nique have been reported so far.
The use of ESCi multimode ionization is helpful for
enhancing analytical throughput and for obtaining wider
chemical coverage. The ESCi source allows a mass spec-
trometer to collect data in four ionization modes (ESI, ESI,
APCI, APCI) during a single LC injection; it can
signicantly extend the MS data acquisition time. This
affects the nal analytical data quality due to reduced data
point coverage per LC peak.
13
To reduce the data acquisition
cycle time, the mass spectrometer must collect data with an
extremely high speed.
Some latest developments in LCpresent further challenges
for mass spectrometry in ESCi-MS applications. For example,
the LC throughput can be improved by employing sub-2 mm
particle chromatography.
14,15
These small particles give rise
to very high-resolution chromatography with peak widths
less than a second at half height. To obtain the same number
of data points as a high-performance liquid chromatography
(HPLC) system, the mass spectrometer needs to collect data
at even higher speed. It is obvious that there is a strong need
to perform an evaluative study of ultra-performance liquid
chromatography/tandem mass spectrometry (UPLC
1
-
tandem MS) using the ESCi multimode ionization source
for quantitative applications.
Currently, the LC/MS technique is applied in virtually
every stage of the drug discovery and development
process.
2,1618
One application example is the use of LC/
MS/MS for the metabolic stability assays in drug discov-
ery.
1921
High-throughput microsomal stability assays utilize
rapid quantication with LC/MS/MS, usually equipped
with an ESI source. When compounds fail to be detected by
ESI, the analysis is usually repeated with an APCI source.
In this paper, we report the development of methodology
using the ESCi source for quantitative analysis under current
state-of-the-art high-throughput separations (using UPLC).
We rst investigated whether an ESCi experiment can be run
in MS scan and product ion scan to give quality spectra. We
then studied the limits of applying the ESCi technique to the
quantitative experiments. In the end, a method was
developed using ESCi for the quantication of multiple
analytes in a single batch. The method was applied to
microsomal stability sample quantication. The results were
compared with those obtained from a conventional single
ionization mode ESI or APCI source.
EXPERIMENTAL
Chemicals and reagents
The eight drug standards (Fig. 1) were purchased individu-
ally from Sigma-Aldrich (St. Louis, MO, USA). Ammonium
acetate was purchased from Sigma-Aldrich. Acetonitrile and
methanol (Optima grade) were purchased from Fisher
Scientic (Fair Lawn, NJ, USA). Glacial acetic acid was
purchased from J.R. Baker (Phillipsburg, NJ, USA). Deio-
nized water puried with an Elix ultra-pure water system
from Millipore Corporation (Bedford, MA, USA) was used
throughout the project.
Figure 1. Chemical structures of the drug standards.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
894 K. Yu et al.
The Sprague-Dawley pooled male rat liver microsomes,
NADPH regenerating system solution A, and solution B
were purchased from BD Biosciences Discovery Labware
(Bedford, MA, USA). Potassium phosphate monobasic,
potassium phosphate dibasic, and dimethyl sulfoxide
(DMSO) were purchased from Sigma-Aldrich. The EDTA
solution (0.5 M) was purchased from Invitrogen (Carlbad,
CA, USA). The three standards zolpadem, midazolam, and
verapamil (used as standards to check activity of the rat liver
microsomes) were purchased from Sigma-Aldrich.
Instrumentation
The separation was performed on a Waters ACQUITY
Ultraperformance LC
1
system, which was equipped with a
binary solvent delivery manager, and a sample manager. The
MS detection was performed on a Waters Micromass
1
Quattro Premier
TM
tandem quadrupole mass spectrometer.
The LC/MS system was controlled by MassLynx
TM
4.0 with
QuanLynx
TM
Application Manager. The ESCi
1
multimode
ionization was enabled rst during the MassLynx installa-
tion, then from the tune page (Fig. 2).
UPLC conditions
The UPLC
1
separation was performed with a Waters
ACQUITY UPLC
1
BEH C
18
column (1.7 mm, 2.1 50 mm)
at 408C. A ow rate of 1.0 mL/min was employed. The ow
was split post-column so that approximately 0.4 mL/min
ow went into the mass spectrometer. A 10 mM ammonium
acetate buffer was used for the mobile phase. The
ammonium acetate stock solution at a concentration of
500 mM was prepared with deionized water. The pH of the
mobile phase was adjusted to 5.0 with glacial acetic acid. This
stock solution was diluted 50 times with acetonitrile/water
(10/90) to obtain mobile phase A, and with acetonitrile/
water (95/5) to obtain mobile phase B. Analytes were eluted
from the column with a linear gradient of 95%A to 5% in
0.4 min, then hold at 95%B for an additional 0.1 min before
returning to initial conditions. The run time was 1 min with
injection-to-injection cycle time at 1.5 min.
MS conditions
MS optimizations were performed in MS scan and in product
ion scan. Quantications were performed in multiple
reaction monitoring (MRM) mode.
During the optimization stage, each standard solution
(10 mg/mL) was individually infused into the mass spec-
trometer at a ow rate of 10 mL/min. To ensure that the
optimized tune page parameters were compatible with the
analytical LC ow, a 0.4 mL/min of LC ow with 50%B was
also introduced into the mass spectrometer by utilizing a T
unit (Upchurch Scientic, Oak Harbor, WA, USA). Each
analyte was tuned in four ionization modes (ESI, ESI,
APCI, and APCI) separately, and the ionization mode
that displayed highest ion intensity was chosen for
quantication experiments. The optimized tune page
settings were: 3.0 kV for capillary voltage in ESI, 4.0 mA for
corona current for APCI, 1308C for the source temperature,
Figure 2. Screen shot of the tune page with the multimode ionization enabled. Both ESI and
APCI-related parameters can be optimized from this single tune page.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
ESI-APCI multimode ionization source for UPLC/MS/MS 895
4208C for the desolvation temperature, 980 L/h for the
desolvation gas ow, and 50 L/h for the cone gas ow.
For multi-analyte quantications, data were collected in
four ionization modes during a single LC injection. The mass
spectrometer switched from one ionization mode to another
every 20 ms. Every data collection cycle included the switch
of ESI to ESI to APCI to APCI, then back to ESI. For
single-analyte quantications, the mass spectrometer col-
lected data in a specic ionization mode during each
injection. The quantitative analysis was completed for all
eight test compounds without interruption in all four
ionization modes in a single batch.
Quantications
A standard stock solution was prepared individually for
each compound at 1 mg/mL in methanol. The calibration
solutions and quality control (QC) samples were prepared by
diluting the stock solutions with deionized water to the
desired concentrations. For multi-analyte analysis, a single
set of calibration solutions at desired concentrations was
prepared, each solution contained multiple standards. For
single-analyte analysis, eight sets of calibration solutions
were prepared; each solution contained a single test
compound.
A complete quantication experiment included injections
of the calibration solutions and two levels of QC samples.
The injection sequence started with the injection of solvent
blank (three injections), followed by the injection of the
calibration solutions from low to high concentrations (0.01,
0.02, 0.05, 0.1, 0.2, 0.5, 1.0, 2.0, 5.0, 10.0, 20.0, 50.0, 100, 200, 500,
1000 ng/mL), followed by another three injections of solvent
blank. Two levels of QC samples (20 and 200 ng/mL) were
then injected in low to high concentration sequence. Six
replicates were injected at each QC level with three solvent
blank injections made between every QC level. For a
complete quantication protocol, the above injection
sequence was performed three times. This quantication
protocol was run once for multi-analyte quantications, and
was run eight times for single-analyte quantications, one
compound each time.
Data processing was accomplished by the use of
QuanLynx Application Manager in MassLynx 4.0. Peak
integration was based on the Apex Track
1
peak integration
algorithm. Calibration curves were constructed using
external calibration based on the peak areas using 1/x
weighting factor with linear regression.
Microsomal Incubation
A single-time-point (15 min) microsomal incubation pro-
cedure was used. The procedure has been previously
described
21
as the following: the test compound solution
was individually prepared in DMSO and later diluted with
1:1:2 DMSO/acetonitrile/water in a 96-deep-well plate to
obtain the 25 mM stock solution. Incubation was performed
by mixing10 mL of the stock solution with 10 mL of 1 mg/mL
(protein) rat liver microsomes, 10 mL of 1 mM NADPH, and
220 mL of 0.1 M phosphate buffer (pH 7.4). The t
0
(time
0 min) samples were obtained by immediately adding 250 mL
of acetonitrile at 08C. The t
15
(time 15 min) samples were
obtained by incubating the mixture for 15 min on a Boekel
Scientic Jitterbug microplate incubator shaker (Bridgeport,
NJ, USA) at 378C. The incubation was quenched by the
addition of 250 mL of acetonitrile at 08C, followed by
centrifugation for 20 min at 3000 rpm on an Eppendorf
5810Rcentrifuge (Hamburg, Germany). The supernatant was
transferred to fresh 96-well plates for instrumental analysis.
The incubation was performed in duplicate. Sample
concentration was at 1 mM.
RESULTS AND DISCUSSION
MS method optimization
Eight commercially available drugs were carefully selected
so that a wide range of polarity and chemical properties was
covered (Fig. 1). Generally speaking, compounds with strong
polarity can be easily ionized by ESI, and compounds with
medium to low polarity can be analyzed by APCI. On the
other hand, basic compounds show preference in positive
ionization, better sensitivity is obtained at lower pH; and
acidic compounds show preference in negative ionization,
better sensitivity is obtained at higher pH. In order to
simultaneously analyze the mixture of these compounds by
LC/MS/MS, the mobile phase pH needs to be adjusted at a
compromised level. As a result, the LC mobile phase pHwas
adjusted to 5.0.
13
As shown in Fig. 2, the parameters for both ESI and APCI
that can be optimized from the tune page include: capillary
voltage and cone voltage for ESI, corona current and cone
voltage for APCI. Both ESI and APCI parameters are
optimized simultaneously during the MS tuning process.
All other tune page parameter settings such as the source and
desolvation temperatures, the desolvation and cone gas
ows, the mass resolutions, and collision cell energies are the
same for both ESI and APCI. Apart from the capillary
voltage/cone voltage for ESI and corona current/cone
voltage for APCI, the most inuential parameters for signal
quality (especially for APCI) are the desolvation temperature
and the desolvation gas ow. When both ESI and APCI
modes are to be used during a single LC injection, it is
preferable to set the desolvation temperature and desolva-
tion gas ow value higher.
To obtain the optimum MS acquisition conditions for each
analyte, a typical tuning process for the tandem quadrupole
mass spectrometer includes two steps: the rst step is the MS
scan to determine the precursor ion and cone voltage; and the
second step is the product ion scan to determine the product
ion and collision energy. To tune an analyte using the ESCi
source, the tuning process can be simplied if the optimum
ionization mode can be determined during the initial MS
scans. For example, for the acidic analyte ibuprofen, signal
was only available in negative ion modes (both ESI and
APCI) during the MS scan; and the ion intensity in ESI was
signicantly stronger than in APCI. As a result, the product
ion scan was performed in one mode only: ESI. Another
compound, tolbutamide, showed signals with good intensity
in all four ionization modes during MS scans. A slight
preference was shown for ESI over APCI in both polarities. In
this case, the product ion scan was performed four times,
once for each ionization mode.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
896 K. Yu et al.
The complete MS optimization results and the nal
ionization mode chosen for each analyte for quantication
are summarized in Table 1. Even though tolbutamide
showed a slight preference for ESI, the negative APCI mode
was decided to be the nal analytical condition. The reason
was that in order to investigate the instrument feasibility of
running in all four ionization modes during a single LC
injection, an APCI- test compound was necessary, and
among all eight standards tested, tolbutamide showed the
strongest signal in APCI- mode.
An additional set of experiments was performed as the last
stage of the MS optimization by injecting each of the
individual standard solutions into the LC/MS/MS system
twice. Data collection was performed once with MS scan and
once with product ion scan by using the optimized
parameters chosen from the tuning stage. The purpose of
these two experiments was to conrm the optimization
results and to investigate the instrument feasibility for
running the LC/ESCi-MS scans. During these scan exper-
iments, for each LC injection, the mass spectrometer was
scanning one ionization mode at a time, and it was switching
between the four ionization modes rapidly (20 ms delay time
for each switch). LC/MS results were obtained for all four
ionization modes simultaneously during each injection.
Figure 3 shows the example total ion chromatograms
(TICs) of MS scans and product ion scans for two example
analytes (ibuprofen on the left, and tolbutamide on the right).
As shown in the gure, ESI was the best ionization mode
for ibuprofen, and tolbutamide was ionizable in all four
ionization modes.
ESCi MRM quantications
The multiple reaction monitoring (MRM) mode is the most
commonly used data collection method for LC/MS/MS
quantications. It is generally accepted that a chromato-
graphic peak should be dened by no less than 15 data points
to ensure the integration precision.
22
Each data point is
obtained when an MS data acquisition cycle is completed.
The time required for each individual acquisition function is
the sum of the MRM dwell times and the inter-channel delay
Figure 3. Total ion chromatograms (TICs) fromthe ESCi MSscan (top) and the ESCi product ion scan (bottom)
for ibuprofen (left) and tolbutamide (right).
Table 1. Summary of the ESCi MS optimization results for the eight analytes. If signal was observed in a particular ionization
mode, it is indicated by a sign. The stronger the signal, the more the number of s. If no signal was observed, it is indicated by a
sign. The column labeled Ionization mode lists the ionization mode chosen for each analyte in MRM quantications
ESI ESI APCI APCI Ionization mode MRM transition
Acetophenone APCI 120.72 >43.07
Corticosterone APCI 347.11 >310.87
Daspone ESI 248.97 >156.07
17a-Hydroxy progesterone APCI 331.18 >295.12
Ibuprofen ESI 205.29 >161.07
Nortriptyline ESI 264.25 >233.18
Sulfadimethoxine APCI 311.25 >155.93
Tolbutamide APCI 269.02 >170.03
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
ESI-APCI multimode ionization source for UPLC/MS/MS 897
times. The more MRM channels to be monitored in a
function, the longer it takes to complete a function. The more
acquisition functions dened by a method, the longer the
total cycle time will be.
The mass spectrometer used in this work has a Traveling
wave (T-Wave
TM
) collision cell which reduces the ion transit
time, and ensures that the collision cell is completely relled
before acquisition commences.
23,24
As a result, the mass
spectrometer can collect data with very short dwell time
(5 ms), inter-channel delay (5 ms) and inter-scan delay
(20 ms) without losing the signal intensity. To analyze
samples in all four ionization modes during a single
injection, the total inter-scan delay time per MS data
acquisition cycle is 80 ms.
During the evaluation stage of the project, three separate
sets of experiments were carried out. For the rst set of
experiments, during each LC injection, the eight test
compounds were analyzed simultaneously as a mixture.
The mass spectrometer was collecting data in MRM mode,
and it was switching among four ionization modes rapidly.
As a result, the MS MRM acquisition method has two MRM
channels for ESI (dwell time 5 ms each, inter-channel delay
was 5 ms), one MRM channel for ESI (dwell time 20 ms),
four MRM channels for APCI (dwell time 10 ms each,
inter-channel delay was 5 ms), and one MRM channel for
APCI (dwell time 10 ms). The inter-scan delay was 20 ms.
The total acquisition cycle time was 175 ms. Figure 4 shows
the MRM chromatograms obtained from this eight-
compound mixture analysis. The average data points per
peak were approximately 810, and the UPLC peak width at
base was averaged at 1.8 s. This coverage of data points was
marginally sufcient for semi-quantitative work.
A second set of experiments was performed by reducing
the number of analytes from eight to four, such that one
MRM channel was monitored for each ionization mode.
During each LC injection, the mass spectrometer was still
collecting data in MRM mode and was switching rapidly
among four ionization modes. As discussed above, the total
inter-scan delay time for running MRM in all four ionization
modes was 80 ms (20 ms per switch). The dwell times were
5 ms in ESI (Daspone), 20 ms in ESI (Ibuprofen), 20 ms in
APCI (tolbutamide), and 10 ms in APCI (sulfadimethox-
ine). As a result, the total cycle time was 125 ms. Figure 5
shows the MRM chromatograms obtained from this four-
compound mixture analysis. The average points per peak
were approximately 15 (peak width at base was 1.8 s).
A complete quantication was then performed using the
second set of experimental conditions; the calibration curves
are shown in Fig. 6 and the results are summarized in Table 2.
The linear ranges for these four analytes were from 3.5 to
4 orders of magnitude, the limit of detection (LOD) was as
low as 0.01 ng/mL, and the r
2
values were over 0.991 for
daspone, sulfadimethoxine and tolbutamide, and 0.984 for
ibuprofen. This is comparable to what is typically obtained
by traditional ESI or APCI.
For the third set of experiments, the eight test compounds
were analyzed as a single analytical batch. However, during
each LC injection, only a single analyte was injected. The
mass spectrometer was detecting a single MRM channel in a
particular ionization mode with its dwell time set at 20 ms.
Time
-Hydroxyprogesterone
0.50 0.40 0.30 0.20 0.10 1.00 0.90 0.80 0.70 0.60
0
100
0
100
0
100
0
100
0
100
0
100
0
100
0
100
APCI+: 331.18 > 295.12
3.95e4
0.430.44
ESI-: 204.96 > 160.79
3.91e4
0.42
APCI-: 269.02 > 170.03
3.12e5
0.37
ESI+: 264.25 > 233.18
1.68e6
0.36
APCI+: 120.72 > 43.07
1.51e4
0.36
APCI+: 347.23 > 311.12
1.12e5
0.35
APCI+: 311.25 > 155.93
4.52e6
0.31
ESI+: 248.93 > 156.07
4.78e6
0.27
Sulfadimethoxine
Corticosterone
Acetophenone
Tolbutamide
Nitriptyline
Ibuprofen
17
Daspone
Figure 4. The ESCi MRM chromatograms obtained from the eight-compound mixture analysis. These
eight MRM chromatograms were obtained from a single LC injection. During each LC injection, the mass
spectrometer was switching among the four ionization modes rapidly. Approximately 810 data points were
obtained across each peak.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
898 K. Yu et al.
The benet of using the ESCi source for this type of analysis is
that the analysis can be carried out in a single analytical batch
without interruption, regardless of the ionization mode
needed for each analyte.
For the quantication analysis using the third set of
experimental conditions, a single calibration curve was
developed upon completing a quantication sequence. The
quantication sequence was run eight times in order to
quantitatively analyze all eight analytes. As shown in Fig. 7,
the LC peaks were better dened with more than 20 data
points collected per peak. Table 3 lists the quantication
results, and the example calibration curves are shown in
Time
0.50 0.40 0.30 0.20 0.10 1.00 0.90 0.80 0.70 0.60
%
0
100
%
0
100
%
0
100
%
0
100
ESI-: 204.96 > 160.79
3.62e4
0.44
APCI-: 269.02 > 170.03
1.03e5
0.39
APCI+: 311.25 > 155.93
2.73e5
0.33
ESI+: 248.93 > 156.07
3.41e6
0.29
Sulfadimethoxine
Tolbutamide
Ibuprofen
Dapsone
Figure 5. The ESCi MRM chromatograms obtained from the four-compound mixture analysis.
These four MRM chromatograms were obtained from a single LC injection. During each LC
injection, the mass spectrometer was switching among the four ionization modes rapidly. Approxi-
mately 15 data points were obtained for each peak.

R2 = 0.992
1000 pg/ 0.1 L
Sulfadimethoxine, APCI+
pg/uL
800 600 400 200
R
e
s
p
o
n
s
e
0
10000
20000
30000

R2 = 0.991
pg/ 200 0.01 L
Dapsone, ESI+
pg/uL
180 140 100 60 20
R
e
s
p
o
n
s
e
0
20000
40000
60000

R2 = 0.984
pg/ 100 0.1 L
Ibuprofen, ESI-
pg/uL
800 600 400 200
R
e
s
p
o
n
s
e
0
1000
2000
3000

R2 = 0.991
1000 pg/ 0.1 L
Tolbutamide, APCI-
pg/uL
1000 800 600 400 200
R
e
s
p
o
n
s
e
0
2000
4000
6000
8000
10000
12000
Figure 6. The calibration curves obtained from the mixture analysis. Each mixture contained four
analytes. During each LC injection, the mass spectrometer was switching among the four ionization
modes rapidly. The quantication results are summarized in Table 2.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
ESI-APCI multimode ionization source for UPLC/MS/MS 899
Fig. 8. Linear dynamic range was up to 4 orders of
magnitude. The r
2
value was over 0.99 for all analytes
except nortriptyline (0.980). The LOD was as low as 0.01 ng/
mL. Compared with the results obtained from the four-
compound mixture analysis, we saw better linearity, wider
dynamic range and lower LODs.
UPLC/ESCi-MS/MS applied to a microsome
stability test in drug discovery
The UPLC/MS/MS method showed promise for the further
enhancement of the analytical throughput for the microsome
stability test. Currently typically used for the quantitative
determination of the microsomal stability is a fast HPLC/
Table 2. Summary of the UPLC/ESCi-MS/MS quantication results for the four-compound mixture analysis. During each
injection, the mass spectrometer was switching rapidly among four ionization modes (ESI, ESI, APCI, and APCI) so that
four MRM chromatograms were obtained at the end of each LC run, one chromatogram per ionization mode
Ionization
Mode
Linear
Range (mg/mL) r
2
LOD
(mg/mL)
QC RSD%
20 mg/mL, N18
QC RSD%
200 mg/mL, N18
Daspone ESI 0.01200 0.991 0.01 9.39 8.61
Ibuprofen ESI 0.1100 0.984 0.1 8.26 3.57
Sulfadimethoxine APCI 0.11000 0.992 0.1 7.30 3.44
Tolbutamide APCI 0.11000 0.992 0.1 7.74 4.71
Figure 7. ESCi MRM chromatograms obtained from the single-compound analysis. These four chro-
matograms were obtained from four separate LC injections. During each LC injection, the mass
spectrometer was detecting a single MRM channel at a specic ionization mode of choice with
20 ms dwell time. More than 20 data points per peak were obtained.
Table 3. Summary of the UPLC/ESCi-MS/MS quantication results for the single-analyte protocol. A single MRM channel in a
specic ionization mode was monitored during each injection. For the whole analytical batch, the mass spectrometer could detect
samples with a choice among any of the four ionization modes (ESI, ESI, APCI, and APCI)
Ionization
Mode
Linear
Range (mg/mL) r
2
LOD
(mg/mL)
QC RSD%
20 mg/mL, N18
QC RSD%
200 mg/mL, N18
Acetophenone APCI 5.01000 0.990 10.0 7.61 2.84
Corticosterone APCI 2.01000 0.991 2.0 6.65 4.58
Daspone ESI 0.11000 0.996 0.01 1.74 2.21
Hydroxyprogestrone APCI 0.2200 0.990 0.2 6.52 2.52
Ibuprofen ESI 0.1500 0.990 0.1 3.35 2.58
Nortriptyline ESI 1500 0.980 2.0 2.70 8.11
Sulfadimethoxine APCI 0.01100 0.995 0.01 10.23 6.48
Tolbutamide APCI 0.1100 0.991 0.1 6.92 6.70
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
900 K. Yu et al.
MS/MS method with 2.5 min run time. The injection-
to-injection cycle time is approximately 3.5 min.
19
The
UPLC/MS/MS method has a 1 min run time with an
injection-to-injection cycle time of approximately 1.5 min. As
a result, the analysis time can be reduced from 5.6 to 2.4 h for
the analysis of a 96-well plate by switching the separation
protocol from HPLC to UPLC.
The single-analyte UPLC/MS/MS protocol was used for
the microsome stability test. During the microsomal stability
screening, a large number of samples containing wide
chemical polarity and lipophilicity are encountered. For each
individual sample, the target is always an individual parent
analyte, not the metabolites. Each MS method requires a
single MRM channel in an optimized ionization mode.
However, for an entire batch of samples to be analyzed, there
may be a need to analyze compounds in different ionization
modes or polarity from one injection to another. The ESCi
multimode ionization source allowed multiple ionization
experiments to be carried in a single batch. This allows a
large variety of compounds to be analyzed by a single
ionization source regardless of the preferred ionization
mode, hence further increasing the analytical throughput.
It was previously demonstrated that for microsome
stability assays, a single-time-point assay is sufcient at
the screening stage to support drug discovery.
21
A moderate
incubation time (e.g. 20 min) is preferred since a short
incubation time is more suitable to give stability resolution
among unstable compounds, whereas a long incubation time
is more applicable to compounds with longer half-life.
A 96-well plate was used for the incubation and injection.
For each compound, duplicate tests were run, each in a
separate well. Typically, three standards covering a range of
metabolic rates were used as references. The reference percent
remaining values for these three compounds from the HPLC/
MS/MS protocol are: verapamil 39%, loperamide 1015%,
and zolpadem 4459%. Table 4 shows the results obtained
from UPLC/ESCi-MS/MS protocol with the in vitro metabolic
half-life calculated. The percent remaining values for all three

Dapsone, ESI+

R2 = 0.992
1000 pg/ 0.1 L
pg/uL
450 350 250 150 50
R
e
s
p
o
n
s
e
0
50000
150000
250000

R2 = 0.992
500 pg/ 0.1 L
Ibuprofen, ESI-
pg/uL
450 350 250 150 50
R
e
s
p
o
n
s
e
0
500
1000
1500
2000

R2 = 0.992
100 pg/ 0.01 L
Sulfadimethoxine, APCI+
pg/uL
90 70 50 30 10
R
e
s
p
o
n
s
e
0
2000
4000
6000
8000
10000

R2 = 0.995
1000 pg/ 0.1 L
Tolbutamide, APCI-
pg/uL
800 600 400 200
R
e
s
p
o
n
s
e
0
5000
10000
15000
20000
25000
Figure 8. The example calibration curves obtained from the single-compound
analysis. During each LC injection, the mass spectrometer was detecting in a
particular ionization mode. The quantication results of this analysis are summarized
in Table 3.
Table 4. The microsome stability analysis results obtained from the single-compound analysis method. Tolbutamide, verapamil
and zolpadem were used as the reference standards. The reference percent remaining values typically obtained from the HPLC/
MS/MS protocol are: 39% for verapamil, 1015% for loperamide, and 4459% for zolpadem
Compounds MW Ionization mode %Remaining
In vitro metabolic
half-life (min)
Acetophenone 120 APCI 55.60 18
Corticosterone 346 APCI 56.14 18
Daspone 248 ESI 48.19 14
17a-Hydroxyprogesterone 330 APCI 0.54 2
Ibuprofen 206 ESI 74.73 >30
Loperamide 476 ESI 11.21 4
Nortriptyline 263 ESI 0.59 2
Sulfadimethoxine 310 APCI 87.65 >30
Tolbutamide 270 ESI 81.88 >30
Verapamil 454 ESI 3.79 3
Zolpadem 307 ESI 47.12 14
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
ESI-APCI multimode ionization source for UPLC/MS/MS 901
reference compounds fall within the reference range. This was
an indication of a good correlation between the UPLC/
ESCi-MS/MS and the HPLC/MS/MS protocols.
CONCLUSIONS
The UPLC/ESCi-MS/MS technique offers time savings in
analysis and therefore can further improve analytical
throughput. The evaluation of the UPLC/MS/MS studies
indicated that 510 data points were obtained for an
eight-compound mixture analysis (insufcient for quanti-
tative analysis), and 15 data points were obtained for a
four-compound mixture analysis (sufcient for quantitative
analysis). In both cases, the mass spectrometer was detecting
in MRM mode and switching rapidly among the four
ionization modes.
An ultra-fast LC/ESCi-MS/MS quantication protocol for
the microsomal stability assay allows a single analyte to be
detected during each LC injection. The mass spectrometer
was detecting in a single MRM channel in a particular
ionization mode. However, for the entire analytical batch, the
mass spectrometer was able to detect compounds in all four
ionization modes of choice.
The test results showed good agreement with the pre-
viously acquired HPLC/MS/MS results. The UPLC separ-
ation was obtained with a 1.5 min injection-to-injection cycle
time. This allowed more than 3 h to be saved for the analysis
of one 96-well plate (compared with the 3.5 min injection-
to-injection cycle time in HPLC). Additional time saving was
also obtained by eliminating sample re-analysis using APCI
mode for the compounds that are not detected by ESI.
Acknowledgements
The authors wish to thank Mr. Michael Balogh from Waters
Corporation (Milford, MA, USA) for many helpful discussions.
REFERENCES
1. Abian J. J. Mass Spectrom. 1999; 34: 157.
2. Bruins AP. Mass Spectrom. Rev. 1991; 10: 53.
3. Bruins AP. Trends Anal. Chem. 1994; 13: 81.
4. Sakairi M, Kambara H. Anal. Chem. 1988; 60: 774.
5. Hanold K, Fischer S, Cormia P, Miller C, Syage J. Anal. Chem.
2004; 76: 2842.
6. Gallagher R, Davey P, Sinclair I, Balogh M, Jackson M,
Southern L. Proc. 50th ASMS Conf. Mass Spectrometry and
Allieds Topics, June 25, 2002.
7. Byrdwell W, Neff W. Rapid Commun. Mass Spectrom. 2002; 16:
300.
8. Syage J, Hanold K, Lynn T, Horner J, Thakur R.
J. Chromatogr. A 2004; 1050: 137.
9. Siegel M, Tabei K, Lambert F, Candela L, Zoltan B. J. Am. Soc.
Mass Spectrom. 1998; 9: 1196.
10. Balogh M. PCT International Patent, WO 2003102537 A2,
2003.
11. Siegel M, Truebenbach C. Proc. 50th ASMS Conf. Mass Spec-
trometry and Allied Topics, June 25, 2002.
12. Gallagher R, Balogh M, Davey P, Jackson M, Sinclair I,
Southern L. Anal. Chem. 2003; 75: 973.
13. Yu K, Little D, Plumb R, Smith B. Rapid Commun. Mass
Spectrom; 2006; 20: 544.
14. Ultra-Performance Liquid Chromatography
TM
Separation
Science Redened, Supplement to LC.GC North America, May,
2005.
15. Waters White Paper, Ultra Performance LC9UPLC): New
Boundaries for the Chromatography Laboratory,2004; May:
720000819EN.
16. Niessen WMA. J. Chromatogr. A 1999; 856: 176.
17. Ruffolo R Jr. Drug Discovery World 2005; Fall, 913.
18. Xu X, Lan J, Korfmacher W. Anal. Chem. 2005; 77: 389A.
19. Di L, Kerns E, Gao N, Li S, Huang Y, Bourassa J, Huryn D.
J. Pharmaceut. Sci. 2004; 93: 1537.
20. Kerns E. J. Pharmaceut. Sci. 2001; 90: 1838.
21. Waters Application Note 720000917EN, June, 2004.
22. Kerns E, Kleintop T, Little D, Tobien T, Mallis L, Di L,Mei H,
Yu Y, McConnel O. J. Pharmaceut. Biomed. Anal 2004;
34: 1.
23. Giles K, Pringle S, Worthington K, Bateman R. Proc. 51st
ASMS Conf. Mass Spectrometry and Allied Topics, Montreal,
Canada, 2003.
24. Kirchner N. US Patent, 5, 206, 506. April 27, 1993.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 893902
DOI: 10.1002/rcm
902 K. Yu et al.

You might also like