Comparing Mass Defect Filtering and Accurate Mass Profile Extracted Ion
Chromatogram (AMPXIC) for Metabolism Studies
Don Kuehl, Ming Gu and Yongdong Wang, Cerno Bioscience, 14 Commerce Drive, Danbury, CT 06810
¾ The MS scans acquired from the loop injection are used
Verapamil MS Scan in Profile Mode
6
Counts
RAW
455.5376
455.2910
1.5
Verapmil001M003, Bile
Verapmil001M002, Clean
Verapmil001M004, Urine
5
0
5
Counts
0.5
5
10
x 10
15
20
25
30
35
Min
Mass Defect Filtering [0.241,0.341]Da
40
45
50
10
15
20
25
Min
30
35
40
45
50
40
45
50
AMPXIC: demethylation,-CH2,441.2753Da
5
4
x 10
3
5
2
1
0
454.5
455
455.5
456
456.5
Correct for peak
asymmetry and
mass errors
457.5
458
458.5
Demethylated
Verapamil
2
Verapamil MS Scan in Profile Mode
x 10
1
15
0.5
0
532
533
534
535
455.2880
531
Calibrated Spectrum
Cal
ibr
ate
d
10
Counts
• Apply parameter-free peak picker to
detect peaks with high mass accuracy to
create AMXIC and MDF.
• Create AMPXIC directly from calibrated
continuum MS spectrum.
Apply correction for
peak shapes and
mass errors
454.5
455
455.5
456
456.5
457
457.5
458
458.5
459
m/z
Counts
4
With the high mass accuracy available after
calibration, it is now possible to perform AMXIC on
unit mass resolution data in an attempt to reduce or
eliminate the contribution of matrix ions to the
extracted ion chromatogram.
3
2
1
0
0
5
10
15
20
25
30
35
40
45
50
Min
16
x 10
XIC: 455.291±0.500 Da
7
14
Mass Spectrum: Scan #1215-1215, 30.453-30.453Min
6
x 10
12
10
7
8
6
Counts
6
4
5
2
4
-2
0
0
10
3
20
30
Min
40
50
60
50
60
AMXIC: 455.291±0.015 Da
6
6
2
x 10
5
1
4
0
440
442
444
446
448
450
452
3
454
2
m/z
1
0
-1
0
10
20
6
5
30
Min
40
454.2981
3
2.5
Verapamil Exact Mass: 455.2910Da
455.2943
Counts
XIC: 440.8-441.8Da
1
Verapmil001M004, Urine
Verapmil001M002, Clean
Verapmil001M003, Bile
456.3360
x 10
4
0.5
0
453
3
453.5
454
454.5
455
m/z
455.5
456
456.5
457
2
1
0
0
10
20
30
Min
40
50
60
While still useful in revealing ions of interest in the
presence of moderately complex urine sample, XIC now
contains false positives already. For the most bile
sample, the contamination becomes so severe that it is
hard to discern the ions of interest amongst all the false
positives, especially at lower concentration levels.
5
441.2692
441.2885
2
441
460
480
500
m/z
520
540
560
Due to the high mass accuracy available through
MassWorks calibration, MDF, developed specifically
for high resolution systems, can now be applied to the
three LC/MS runs with unit mass resolution in a
±50mDa mass defect window.
For the clean
incubation, very little difference is seen between TIC
and MDF Ion Chromatogram (MDFIC) in terms of
peak patterns, though the overall background level is
reduced by ~10x. Towards the end of the LC run
where most of the matrix ions appear, significant
portions of these matrix ions have their mass defects
within the given window, resulting in a rise in baseline
much like that in the TIC. The benefits of MDF is
most clearly seen in the urine sample where significant
improvement in peak patterns is achieved through
filtering of the background ions prior to RT=30min,
resulting in MDFIC very similar to that from the clean
sample. When applied to the bile sample, however,
MDF does not yield much more information than the
TIC, due possibly to the level and variety of the matrix
ions involved, as evidenced in the filtered mass
spectrum at the bottom.
441.5
442
442.5
443
443.5
m/z
Different from MDFIC, AMPXIC requires a list of all
possible biotransformations which can then be
searched throughout the whole calibrated LC/MS data
set to create a filtered ion chromatogram matching the
exact mass and isotope distribution of the postulated
ion. As seen from the above graph for the example of
demethylation metabolite, XIC from the clean sample
offers a clear picture of the few demethylation
metabolites while XICs from the urine and bile
samples are clearly contaminated.
With mass
accurately calibrated and MS peak shape
mathematically matched, AMPXICs for all three
samples, including the most challenging bile sample,
show the same clear chromatographic peak patterns,
subject only to retention time shifts amongst the three
LC/MS runs, demonstrating highly selective filtering
power of AMPXIC processing. Furthermore, the
corresponding mass spectrum after the calibration can
be reliably analyzed for the purpose of peak detection
and peak integration, resulting in accurate masses and
unbiased peak area integration for confirmation and
quantitative analysis.
4
453.2664
5
441.2725
3.5
2
¾ The metabolites were generated by incubation of rat
15
20
25
30
35
Min
Mass Spectrum: Scan #1154-1154, 28.923-28.923Min
5
x 10
10
Background?
Verapamil
10
50
Mass Spectrum: Scan #1850-1850, 46.382-46.382Min
x 10
1.5
7
45
4.5
Counts
For clean incubation, TIC shows clearly the ions of
interest as chromatographic peaks. For a more complex
with urine background, its TIC becomes too complex to
be useful. For the most complex bile sample, all ions of
interest becomes invisible in its TIC, in spite of the high
concentrations used.
15
20
25
30
35
40
Min
Filtered Mass Spectrum: Scan #1221-1221, 30.603-30.603Min
0
Total Ion Chromatogram
Verapmil001M003, Bile
Verapmil001M002, Clean
Verapmil001M004, Urine
x 10
440
Results and Discussion
Methods
10
0
5
x 10
5
6
4
Counts
5
1.5
530
5
0
459
m/z
2.5
Statistically
examine peak
shapes and mass
errors
Run external
calibration
standard
457
Mass Spectrum
Counts
0
x 106
0
microsomes with verapamil and separated by C18 column
with a gradient and acquired on AB/Sciex 4000 QTRAP
instrument, a triple stage quadrupole with the 2nd stage
quadrupole operating as a linear ion trap.
¾ The incubations were analyzed with LC/MS as is, after
addition of rat bile or urine matrix, resulting in a total of three
LC/MS runs to demonstrate and compare the various
approaches for processing ion chromatograms with varying
background levels.
¾ Prior to the LC/MS analysis, a standard mixture containing
8 different drugs covering 100-800Da mass range is injected
through a loop with a mobile phase and flow rate similar to
those for the LC/MS analysis.
x 10
10
2
10
3
9
Experiments and Data Processing
Verapmil001M004, Urine
Verapmil001M003, Bile
Verapmil001M002, Clean
7
Run LC/MS of
Verapamil sample
5
For biofluid samples with many background ions, the ions of
interest will typically be obscured in the TIC. Even when
processed into XIC, some of these background ions and their
isotopes would happen to have their nominal masses fall
within the XIC integration window of ±0.5Da, resulting in a
rather complex XIC. With the high mass accuracy capability
of MassWorks™ software, a more selective Accurate Mass
XIC (AMXIC™) can be generated in a tight accurate mass
window of ±15mDa even on a unit mass resolution MS
system. AMXIC, however, does not take into consideration
of the isotope clusters and their isotope profiles, resulting in a
sub-optimal tradeoff between sentitivity and selectivity. A
novel filter based on both accurate mass and isotope profile,
AMPXIC, is implemented and tested along with Mass Defect
Filtering (MDF) on Verapamil incubation samples with
varying degrees of background ions from clean incubation to
bile matrix.
x 10
4
0
1
MassWorks Calibration & Data Processing
XIC: 440.8-441.8Da
7
15
Total Ion Chromatogram
9
6
2
Counts
x 10
Counts
2.5
Counts
for a comprehensive mass spectral calibration involving
both the mass axis and MS peak shape using the
MassWorks software.
¾ Each MS scan in an LC/MS run would then be
calibrated using this comprehensive MS calibration prior
to ion extraction analysis via AMIC, MDF, and AMPXIC.
Intensity
Introduction
In LC/MS experiments, Total Ion Chromatogram (TIC)
provides a good summary of ions being separated/detected.
While usually very informative, TIC works best when the
analytes are dominating the ion signals in any given retention
time window. In the presence of many other ions of similar
abundance levels, however, TIC may become too complex to
be informative, giving rise to eXtracted Ion Chromatogram
(XIC), which only integrates ion intensities falling within a
given mass window (±0.5Da). For more complex samples
such as those collected with bile matrix, more selective filter
is needed for effective metabolite identification. Three such
filters will be studies and compared in this paper: Accurate
Mass eXtracted Ion Chromatogram (AMXIC), Mass Defect
Filtering (MDF, Ref 1) and Accurate Mass Profile eXtracted
Ion Chromatogram (AMPXIC™, Ref 2), all using MS data
at unit mass resolution with the high mass accuracy
calibration available through MassWorks calibration.
It is seen that AMXIC can indeed simplify the
extracted ion chromatogram due to the more
accurate mass window used. When applied to the
most complex bile sample, it is also clear that other
matrix ions still make into the extracted ion
chromatograms, including the example shown above
when the M+1 isotope of a matrix ion makes into
the AMXIC as the determined accurate mass for
M+1 cluster happens to fall within the accurate mass
window.
Conclusions
Acknowledgement
¾ TIC is only useful in revealing ions of interest when
there is no significant background.
¾ XIC suffers from both chemical noise as well as
isotope interferences and could not reveal ions of
interest in complex matrices.
¾ AMXIC is significantly more selective than XIC but
still suffers from significant false positives in
complex matrices, especially towards the end of
LC/MS run.
¾ MDF is useful for clean and moderately complex
samples but becomes as featureless as TIC for a
truly complex sample.
¾ AMPXIC is able to handle various types of matrices
but does require a list of possible biotransformations a priori.
The authors would like to thank Dr. Xian-guo Zhao
and Dr. Zhe-ming Gu from Xenobiotic Laboratories,
Inc., for acquiring the LC/MS data used in this study.
References
1. Haiying Zhang, J. Mass Spectrom. 2003; 38:11101112.
2. Ming Gu et al, Rapid. Commun. Mass Spectrom.
2006:20:764-770.