Identification of protein phosphorylation sites by advanced LC-ESI-MS/MS methods.
ABSTRACT Phosphorylation, the process by which a phosphate group is attached to a pre-existing protein, is an evolutionarily and metabolically cheap way to change the protein's surface and properties. It is presumably for that reason that it is the most wide-spread protein modification: an estimated 10-30% of all proteins are subject to phosphorylation.MS-based methods are the methods of choice for the identification of phosphorylation sites, however biochemical prefractionation and enrichment protocols will be needed to produce suitable samples in the case of low-stoichiometry phosphorylation. Using emerging MS-based technology, the elucidation of the "phosphoproteome," a comprehensive inventory of phosphorylation sites, will become a realistic goal. However, validating these findings in a cellular context and defining their biological meaning remains a daunting task, which will inevitably require extensive and time-consuming additional biological research.
Identification of Protein Phosphorylation Sites
by Advanced LC-ESI-MS/MS Methods
Christoph Weise and Christof Lenz
Summary Phosphorylation, the process by which a phosphate group is attached to
a pre-existing protein, is an evolutionarily and metabolically cheap way to change
the protein’s surface and properties. It is presumably for that reason that it is the
most wide-spread protein modification: an estimated 10–30% of all proteins are
subject to phosphorylation.
MS-based methods are the methods of choice for the identification of
phosphorylation sites, however biochemical prefractionation and enrichment pro-
tocols will be needed to produce suitable samples in the case of low-stoichiometry
phosphorylation. Using emerging MS-based technology, the elucidation of the
“phosphoproteome,” a comprehensive inventory of phosphorylation sites, will
become a realistic goal. However, validating these findings in a cellular context
and defining their biological meaning remains a daunting task, which will inevita-
bly require extensive and time-consuming additional biological research.
Key Words Phosphorylation; LC-ESI-MS/MS; phosphoproteome.
Phosphorylation, the process by which a phosphate group is attached to a pre-exist-
ing protein, is an evolutionarily and metabolically cheap way to change the pro-
tein’s surface and properties. It is presumably for that reason that it is the most
wide-spread protein modification: an estimated 10–30% of all proteins are subject
to phosphorylation. The reaction is catalyzed by a set of enzymes called kinases
that form one of the largest protein families of all. As the reaction is readily
reversed by another group of enzymes, called phosphatases, phosphorylation turns
out to be a pivotal regulatory mechanism that plays critical roles in the regulation
of many metabolic pathways and cellular processes, including cell cycle, growth or
differentiation (1). The determination of phosphorylation sites is the basis for a
deeper understanding of cellular regulation and will allow conclusions about the
enzymes involved in specific regulatory pathways. Because aberrant phosphorylation
From: Post-translational Modifications of Proteins.
Methods in Molecular Biology, Vol. 446.
Edited by: C. Kannicht © Humana Press, Totowa, NJ
34 C. Weise and C. Lenz
events are known to occur in many diseases, including various types of cancer, this
holds huge promise for the definition of new drug targets. Phosphorylation—from
the history of its discovery to methodological advances and biological aspects—is
covered by a number of excellent and exhaustive reviews (1–3).
Important recent technological advances have made mass spectrometry (MS) the
method of choice for protein analysis and proteome research over the last decade,
but despite the huge interest in protein phosphorylation, the determination of phos-
phorylation sites has remained analytically challenging. The classical chemical
sequencing approach (Edman degradation) was hampered mainly by the insolubil-
ity of the phosphoamino acid products and the necessity to obtain highly purified
phosphopeptides. Mass spectrometry, on the other hand, is well suited to deal even
with complex peptide mixtures, but still suffers from the often low stoichiometry of
phosphorylation leading to low signal intensities that tend to disappear into the
background. The often cited low ionization efficiency of phosphopeptides relative
to their nonmodified counterparts, though, appears to be a generalization that is not
supported by experimental evidence (4).
Several approaches have been reported to deal with the stoichiometry challenge (3):
– chemical replacement of the phosphate group by other functionalities that
enhance ionization efficiency and MS/MS fragmentation behavior, e.g., by β-
elimination and subsequent Michael addition. Because of incomplete reaction
and purification of the products this usually requires an increased amount of
– affinity enrichment of phosphorylated species, e.g., by immobilized metal-affin-
ity chromatography (IMAC) on Fe3+ or Ga3+ matrices. ZrO2 or TiO2 have also
been successfully used for this purpose, however the enrichment is rarely spe-
cific and acidic peptides are likely to be enriched as well.
– alternatively, peptides phosphorylated on tyrosine can be purified using anti-
P-Tyr-antibodies. No antibodies with good specificity for P-Ser and
P-Thr are available, however, although these form the bulk of cellular phos-
The phosphorylation-specific analytical method would have to introduce some
sort of filter that will allow systematic screening for phosphorylated compounds.
In mass-spectrometric analysis, precursor ion scanning can be used to identify
compounds from mixtures, such as proteolytic digests, that result in a common
product ion. In this experiment a first mass analyzer is set to scan the entire mass
range of possible precursor (peptide) ions. Through collision-induced dissociation
(CID) these are fragmented to produce a marker product ion, which is selectively
monitored using a second mass analyzer fixed on the m/z value of the marker ion.
Under CID conditions in negative-ion mode phosphopeptides produce distinct
marker ions at m/z 79 (PO3
detection (5–7). Various phosphorylation-specific aspects of precursor-ion-scanning
methods are discussed in the Notes (Section 4).
In Section 2.3 we describe a state-of-the-art LC-MS/MS method (direct coupling
of a liquid-chromatography system to a mass spectrometer) for determining
phosphorylation sites from a peptide mixture generated by in-gel digestion.
-) and m/z 63 (PO2
-), which can be used for their selective
3 Identification of Protein Phosphorylation Sites 35
Phosphopeptides selectively detected by precursor ion scanning are subsequently
fragmented by collision cell CID in a product-ion experiment to establish their
sequence and the site of phosphorylation. Using this set-up, amounts of phosphor-
ylated peptide as low as 5 fmol can be detected.
To summarize, MS-based methods are the methods of choice for the identification
of phosphorylation sites, however biochemical prefractionation and enrichment
protocols will be needed to produce suitable samples in the case of low-stoichiom-
etry phosphorylation. Using emerging MS-based technology, the elucidation of the
“phosphoproteome,” a comprehensive inventory of phosphorylation sites, will become
a realistic goal. However, validating these findings in a cellular context and defining
their biological meaning remains a daunting task, which will inevitably require
extensive and time-consuming additional biological research.
2.1 In-Gel Reduction, Alkylation and Tryptic Digestion
1. Gel Washing Solution: 50 mM ammonium bicarbonate in water.
2. Coomassie Destaining Solution: 50 mM ammonium bicarbonate in acetonitrile/
water (1:1, v:v).
3. Gel Dehydration Solution: acetonitrile.
4. Cystine Reduction Solution: 100 mM dithiothreitol in 100 mM aqueous ammo-
5. Cysteine Alkylation Solution: 55 mM iodoacetamide in 100 mM aqueous
6. Modified Sequencing Grade Porcine Trypsin (Promega): 10 µg/mL in 25 mM
aqueous ammonium bicarbonate.
7. Gel Extraction Solvent: 0.5% formic acid in acetonitrile:water (2:8, v:v).
8. Acetonitrile and water, HPLC grade.
9. 0.5-mL Eppendorf tubes.
10. A temperature-controllable heater/shaker.
11. A SpeedVac concentrator.
The protocol supplied here describes the in-gel tryptic digestion of a protein
detected by Coomassie staining on an SDS gel. One major challenge in the analysis
of protein phosphorylation is the substoichiometric degree of this modification
(3,4). The methodology described here is capable of detecting amounts as low as
5 fmol of phosphopeptide total. Assuming a 1% degree of phosphorylation of the
protein and a digestion/extraction efficiency of 50%, this translates into an amount
of 1 pmol protein loaded onto the gel. Protein amounts in the low picomole range
are usually detectable by Coomassie staining.
Another challenge lies in the choice of the proper endopeptidase for digestion.
The actual site of phosphorylation may lie in a region of the sequence where too
36 C. Weise and C. Lenz
many or too little trypsin cleavage sites are located, resulting in peptides not
suitable for the LC-MS/MS analysis (MW range 700–3,000 Da). For a comprehensive
phosphorylation analysis of unknown proteins additional analyses using digest
agents with different specificities should be used, such as endopeptidase GluC or
2.2 LC-ESI-MS/MS Analysis of Tryptic Digests
1. Loading Solvent: 0.5% formic acid in acetonitrile:water (2:98, v:v).
2. Mobile Phase A: 0.1% formic acid in acetonitrile:water (5:95, v:v).
3. Mobile Phase B: 0.1% formic acid in acetonitrile:water (95:5, v:v).
4. Make-up Solvent: 0.1% formic acid in acetonitrile:2-propanol:water (1:8:1, v:v:v).
5. A low-dead-volume T-junction (Upchurch Micro-Tee P775, Upchurch, Oak
Harbor, WA/US) with fused silica capillary (20 µm ID) and Teflon sleeves for
6. A hybrid triple-quadrupole/linear-ion-trap mass spectrometer (4000 Q TRAP
LC-MS/MS system, Applied Biosystems, Foster City, CA/US) coupled on-line
with nanoflow HPLC (Ultimate with Famos autosampler and Switchos column
switching module, all Dionex, Idstein, Germany). A Micro-ion spray head
(Applied Biosystems) fitted with a fused silica tapered tip sprayer needle (FS360-
20-10-N, New Objective Inc, Woburn, MA/US) and zero grade air or nitrogen as
Sheath Gas. A 75 µm × 15 cm PepMap RP-C18 column (3-µm particle size, 100-
Å pore size) and a 300 µm × 5 mm precolumn of the same material (Dionex).
7. An additional pump capable of generating nanoliter flow rates (Harvard
Apparatus Model 11 Pico Plus syringe Pump, Harvard Apparatus, Holliston,
MA/US) with a 100-µL glass syringe (Hamilton).
The LC-MS/MS analysis of phosphopeptides consists of 2 steps: A) detection of
phosphorylated peptides by a selective scan function, i.e., a precursor ion scan (Fig.
3.1) for m/z 79 (PO3
tide sequence in positive mode (see Notes 1 and 2). On hybrid triple quadrupole/lin-
ear ion trap mass spectrometers these 2 steps can be carried out in a single integrated
experiment (9–11). If such an instrument is not available though, the 2 parts can also
be carried out independently on other equipment. Both triple-quadrupole and hybrid-
triple-quadrupole/time-of-flight mass spectrometers are capable of precursor ion scan
experiments. The subsequent high-sensitivity MS/MS analysis can also be performed
on e.g., hybrid triple quadrupole/time-of-flight or conventional ion trap mass spec-
trometers. The sample will need to be split for the 2 experimental steps (5,6).
Special consideration should be given to the choice of solvents and organic
modifiers that are used for the reverse-phase separation and LC-MS/MS analysis.
For a single experiment, analysis conditions are needed that allow both for negative
and positive mode electrospray ionization. A weak acidic modifier has proven to
enable electrospray analysis of phosphopeptides in both polarities at comparable
sensitivity. From a chromatographic point of view, trifluoroacetic (TFA) acid is the
-) in negative mode, and B) MS/MS analysis of the phosphopep-
3 Identification of Protein Phosphorylation Sites 37
organic modifier of choice as it provides the best separation because of its strong
ion-pairing properties. As TFA in practice is severely detrimental to negative-mode
electrospray ionisation, formic acid is usually chosen as organic modifier.
Another challenge implied with negative-mode electrospray ionization is the
possibility of high-voltage corona discharge, or “arching,” leading to corrosion of
the sprayer needle and irreproducible ionisation conditions. The post-column
addition of isopropanol as a “make-up solvent” via a T-piece reduces the voltage
necessary to achieve ionisation in negative mode, thus significantly reducing the
danger of corona discharge. Isopropanol can be premixed with organic modifier
and acetonitrile to achieve constant modifier concentration and reduce the other-
wise high back-pressure of the more viscous isopropanol.
2.3 Data Interpretation
1. The amino-acid sequence of the protein.
2. Software for fragment mass matching of LC-ESI-MS/MS data against theoretical
fragment patterns obtained from in silico digestion of protein sequences
(MASCOT V2.1, Matrixscience Ltd., London, UK)
3. Software for annotation of raw ESI-MS/MS data with fragments generated from
hypothetical sequences including modified residues (Bioanalyst 1.4, Applied
Using fragment mass matching software like MASCOT provides a good first
screen for the detection of phosphorylated peptides from the LC-MS/MS data set.
The probability score values obtained for phosphorylated peptides are generally
lower compared to those obtained for nonmodified peptides though, resulting in a lower
chance of detecting their presence. A manual evaluation of the data using mass lists
of theoretically plausible phosphopeptides calculated from the protein sequence is
therefore highly recommended.
Assignment by the algorithm of the site of phosphorylation to an individual
S/T/Y residue in a phosphopeptide sequence is also frequently observed to be
Fig. 3.1 Principle of a precursor ion scan experiment on a triple-quadrupole mass spectrometer.
The first quadrupole Q1 scans the m/z range of possible intact phosphopeptide precursors; the
second quadrupole Q2 serves as a collision cell where precursors are fragmented by collisionally
induced dissociation (CID); Q3 is set to continually monitor production of the marker fragment
m/z 79 (PO3
38 C. Weise and C. Lenz
incorrect. The assignment suggested by the software should therefore be validated
by annotating raw data MS/MS spectra with sets of theoretical fragments generated
from putative sequences.
3.1 In-Gel Reduction, Alkylation and Tryptic Digestion
1. After SDS-PAGE separation the Coomassie-stained gel band is excised and cut into
smaller pieces using a scalpel. The pieces are transferred to a 0.6-mL Eppendorf
tube, and washed twice with 0.5 mL of Gel Washing Solution for 10 min.
2. To destain the gel pieces 0.5 mL of the Destaining solution is added for 30 min,
with occasional vortexing. The solution is then discarded. This procedure can be
repeated until the gel piece is completely destained.
3. The gel pieces are dried by adding 0.1 mL of the Gel Dehydration solution. After
5 min the gel pieces shrink and turn white. The Gel Dehydration solution is
pipeted off and discarded. Residual solvent is removed in a SpeedVac concentra-
tor for 10 min.
4. Cystine bridges are reduced by adding 0.03 mL of the Cystine Reduction solution
at 56°C for 30 min. Free cysteines are then alkylated by adding 0.03 mL of the
Cysteine Alkylating Solution at room temperature in the dark for 20 min. The
solvents are then pipeted off and discarded.
5. To remove residual reduction/alkylation agent, the gel piece is washed again with
0.5 mL of Gel Washing Solution for 10 min. Step 3 is then repeated to remove
6. Between 0.003 and 0.03 mL of trypsin solution are carefully added to the dried gel
pieces until they are fully rehydrated. When the gel pieces do not take up any addi-
tional solution, 0.05 mL of gel washing solution is added, the solvents quickly spun
down in a microcentrifuge, and the tube closed and sealed with Parafilm. Digestion
is achieved by placing the tube in a heater/shaker combination at 37°C overnight.
7. The tube is removed from the heater/shaker and let cool to room temperature.
The supernatant is pipeted off and set aside. To extract the majority of peptides
0.03 mL of the Gel Extraction Solvent is added to the gel piece, and the contents
of the tube are sonicated twice for 15 min. The Extraction Solvent is now pipeted
off, combined with the digestion supernatant in a 0.6-mL Eppendorf tube and
dried down in a Speedvac concentrator.
3.2 LC-ESI-MS/MS Analysis of Tryptic Digests
1. Prepare the solvents and equilibrate the nanoflow LC system on 5% solvent B.
3 Identification of Protein Phosphorylation Sites 39
2. Set up a micro-T junction to split in the make-up solvent post-column at a flow
ratio of 300:100 (LC eluent:make-up, nl/min). Measure the flow rates before and
after the micro-T to ensure proper set-up of the junction.
3. Optimise source conditions for both negative mode and positive mode operation,
and for rapid two-way switching between polarities.
4. Set up an LC-MS/MS acquisition method consisting of a precursor ion scan for
m/z 79 in negative mode, a high resolution MS scan of detected signals in posi-
tive or negative mode, and of up to 3 MS/MS experiments of accepted precursors
in positive mode (Fig. 3.2). The precursor ion scan should be set up using isola-
tion widths (FWHH) of 2 Th in Q1 and 0.6 in Q3 to accommodate the isotopic
patterns of the phosphopeptide precursors and the m/z 79 fragment ion, respec-
tively. The chromatography method should include pre-column concentration
and desalting. A linear gradient of 5–40% B across 45 min is often used.
5. Inject 100 fmol of a known phosphopeptide standard to evaluate the performance
of the system with regard to stable polarity switching, sensitivity and MS/MS
results. A tryptic digest of bovine casein a diluted down from a stock solution is
usually used for this purpose (Fig. 3.3).
Fig. 3.2 Scan cycle for the selective detection and sequence determination of phosphopeptides on
hybrid triple quadrupole/linear ion trap mass spectrometers. If a precursor has been selected for
MS/MS in 2–3 consecutive cycles, it is excluded from further selection to allow the analysis to
focus on other, lower-abundance precursors (dynamic exclusion)
40 C. Weise and C. Lenz
6. Once system performance has been established, inject approximately 1 pmol of
the digested protein sample dissolved in 10 µL of the Loading solvent.
A hybrid triple quadrupole/linear ion trap mass spectrometer (4000 Q TRAP
LC-MS/MS System, Applied Biosystems) coupled on-line with a nanoflow HPLC
is used for the simultaneous detection and sequencing of phosphopeptides in com-
Protein phosphorylation is usually substoichiometric. As a consequence an
endopeptidase digest of a phosphoprotein will often contain only minor amounts of
phosphopeptides. To increase the odds of detecting these compounds, an LC sepa-
ration is employed to reduce the complexity of the mixture presented to the mass
spectrometer at any given point in time. An additional simplification is achieved by
the selective precursor ion scan for the phosphopeptide-specific fragment m/z 79
As the precursor ion scan is a low-resolution experiment, it is usually not possi-
ble to determine the charge state of a detected peptide as its isotope pattern will not
be resolved. The precursor ion scan is therefore followed by a higher-resolution ion
trap MS experiment that allows for determination of the charge state of the precur-
sor ion. From the m/z and the charge state the molecular weight of the phosphopep-
tide can be determined to an accuracy of 0.3 Da or better. This scan is usually
carried out in negative mode (see Note 3).
If the two first scan events show one or multiple precursors that meet specific
criteria (minimum signal intensity, MW range, charge state), these are selected for
-) in negative mode (Fig. 3.1).
Fig. 3.3 Evaluation of the analytical system using a standard trypsin digest of bovine casein a at
an amount of 100 fmol injected on column. The traces show the different experiments at the reten-
tion time of the phosphopeptide TVDMEpSTEVFTK (MW 1465.61 Da): (A) Precursor ion scan
for m/z 79 in negative mode; (B) High resolution ion trap MS Scan of the detected [M-2H]2- in
negative mode; (C) MS/MS spectrum of [M+2H]2+ in positive mode
3 Identification of Protein Phosphorylation Sites 41
MS/MS analysis to establish their sequence. After 2–3 occurrences, these precursor
m/z values are excluded from selection for 60 s to focus on other, less abundant
precursors eluting at the same time (Dynamic Exclusion, Fig. 3.2).
As MS/MS fragmentation of peptides in positive mode is better understood than
in negative mode, the polarity has to be switched twice during the LC-MS/MS
cycle: from negative to positive after precursors have been selected for MS/MS, and
back to negative for the precursor ion scan of the following experiment. Switching
polarity during a nanoLC-MS/MS experiment requires careful optimisation of the
ion source parameters for both positive and negative mode. The performance of this
experiment should always be tested using a known phosphopeptide standard before
the first unknown sample is injected. Figure 3.3 shows a standard analysis of a
100 fmol injection of a casein α tryptic digest. At a retention time of 36.2 min, the
precursor ion scan shows a single signal at m/z 732.0 in negative mode (Fig. 3.3A).
The higher-resolution ion trap scan in negative mode (Fig. 3.3B) shows the isotope
pattern of a doubly charged peptide precursor [M-2H]2-, with the monoisotopic
peak at m/z 731.8 indicating a molecular weight of 1,465.6 Da. The corresponding
[M+2H]2+ at m/z 733.8 is then selected for MS/MS in positive mode (Fig. 3.3C).
Both the molecular weight and the MS/MS data unambiguously identify the pep-
tide sequence as TVDMEpSTEVFTK from bovine casein a.
The results obtained on the standard sample should be examined for the following
criteria: (i) intensity and signal-to-noise ratio in precursor ion scan mode, (ii) correct
determination of the charge state from the linear ion trap scan, and (iii) high-quality
MS/MS spectra that allow for the correct sequence assignment of the phosphopep-
tides at reasonably low amounts on column, preferably 100 fmol or less.
3.3 Data Interpretation
1. Open the LC-MS/MS data file generated from the sample. Select the Precursor
Ion Scan data. Extract and print out a base peak chromatogram (BPC) annotated
with retention time. Peaks in the precursor ion scan BPC will represent possible
2. Extract peak lists for all MS/MS spectra generated in the run and submit them
for a combined database search against a protein database. If possible use a small
dedicated database that just contains the protein sequence(s) of interest. Adjust
mass tolerances for MS and MS/MS to reflect values typically achieved in rou-
tine analysis, e.g., 0.3 Da for the linear ion trap instrument used. Use phosphor-
ylation of and neutral loss from S, T, and Y residues as variable modification
filters. As phosphorylation of residues close to arginine or lysine frequently
causes miscleavage of trypsin, allow for at least two missed cleavages in the
3. Closely examine the database search results. Validate every tentative phosphopep-
tide assignment by matching theoretically calculated peptide fragment patterns to
the raw MS/MS data. In case there are multiple possible phosphorylation sites in
42 C. Weise and C. Lenz
a peptide sequence, compare the patterns for every possible site population as the
database search engine can misassign the residue (see Note 4).
4. Manually look through the data to find high-quality MS/MS spectra that are not
explained by the database search results. Look for MS/MS spectra that show
neutral loss peaks at a distance of −98/z (z = charge state) from the precursor.
Compare the precursor m/z values of these MS/MS to a list of theoretically gen-
erated phosphopeptide m/z values. In case of a close match between theoretical
and observed precursor m/z values, annotate the spectrum with the fragment pat-
tern generated for this putative phosphopeptide sequence.
Figure 3.4 shows the sequence of the intracellular loop of acetylcholine receptor
(δ-subunit from Torpedo californica) heterologously expressed in Escherichia coli
(12). The protein phosphorylated in vitro using protein kinase A was kindly provided
by Dr.Viktoria Kukhtina, Berlin. After phoshorylation the molecular weight was
determined by MALDI-ToF to be 17,633 Da (theor. 17,552 Da), indicating within
experimental error at least one phosphorylation event (∆ 81 Da, theor. 80 Da).
After digestion and LC-MS/MS analysis with precursor ion scan detection, the
base peak chromatogram depicted in Figure 3.5 (bottom trace) was obtained. A
Mascot search against the SwissProt database indicated the presence of 5 phos-
phopeptides A–E, corresponding to 2 adjacent phosphorylation sites of serine resi-
dues 50 and 51, respectively. The higher number of peptides results from the
observation of singly and doubly phosphorylated peptides, as well as missed cleav-
ages from the trypsin digestion.
Figure 3.6 shows the MS/MS spectrum of the precursor m/z 1,119.02+, in chro-
matography peak C, corresponding to the sequence RSSSVGYISKAQEYFNIK
doubly phosphorylated. In addition to fragment ions that can be assigned to neutral
loss of the 2 phosphogroups ([M+2H-H3PO4]2+ m/z 1070.3, [M+2H-2H3PO4]2+ m/z
1021.6), a significant number of sequence-specific fragment ions is observed. The
spectrum is labeled with the sequence RSpSpSVGYISKAQEYFNIK (aa 48–65)
that gives the most comprehensive explanation of the fragments observed. While
the b2 ion indicates that Serine 49 is not phosphorylated, both the b3 and b4 ions
are observed in their phosphorylated and dephosphorylated states, indicating that
the protein is indeed phosphorylated at residues 50 and 51. Alternative sequences
like RpSSpSVGYISKAQEYFNIK or RpSpSSVGYISKAQEYFNIK do not explain
the experimental data as consistently.
Fig. 3.4 Amino-acid sequence of the protein analyzed here as an example (Figs. 3.5 and 3.6),
the heterologously expressed intracellular loop of the δ-subunit of the acetylcholine receptor from
Torpedo californica (12)
3 Identification of Protein Phosphorylation Sites 43
Fig. 3.5 The Precursor Ion Scan, a highly selective filter for the detection of phosphopeptides—
comparison of basepeak chromatograms using regular linear ion trap MS detection (top trace) and
precursor ion scan m/z 79 detection (bottom trace) of a tryptic digest of the acetylcholine receptor
intracellular loop. The precursor ion scan base peak chromatogram is annotated with the phos-
phopeptide sequences identified by database searching
Fig. 3.6 MS/MS spectrum of the phosphopeptide precursor m/z 1,119.02+ at a retention time of
41.1 min (Fig. 3.5, peak C). The spectrum is labeled with fragments calculated for the assigned
sequence RSpSpSVGYISKAQEYFNIK. The ions of the lower b series indicate phosphorylation
on S50 and S51, but not S49
44 C. Weise and C. Lenz
The method described here is a significant improvement in the analysis of
protein phosphorylation. It should be noted, though, that it will not always yield
comprehensive results on all sample types. There are still multiple stages where
problems can occur: (i) incomplete digestion of the protein around the site of modi-
fication, (ii) poor extraction of the phosphopeptide from the gel, (iii) loss of e.g.,
highly polar phosphopeptides during reverse-phase chromatography, (iv) failure of
the phosphopeptide to produce sufficient signal response in MS and (v) inability of
the database searching algorithm to assign the sequence and site of phosphorylation
because of nonconclusive fragmentation. As a consequence, the method should be
complemented by other strategies such as phosphopeptide enrichment (3), off-line
analysis or targeted LC-MS/MS analysis (13,14) wherever possible.
1. The method described here uses precursor ion scanning in negative-ion mode for
the selective detection of phosphopeptides. Other methods have been described
that use the neutral loss of phosphoric acid (H3PO4, −98 Da) in positive mode
for this purpose (15). Although, e.g., Constant Neutral Loss Scans on a triple
quadrupole instrument in positive mode are easier to perform experimentally,
they suffer from several shortcomings:
– not all phosphopeptides exhibit a strong neutral loss fragmentation when
collisionally activated in positive mode (16). As a consequence, especially
peptides containing phosphotyrosine are usually not detectable by neutral
– as the mass spectrometer analyses m/z (not MW), the actual neutral loss
observed is dependent on the charge state, e.g., 98.0/2 = 49.0 or 98.0/3 =
32.7. This has to be accounted for when setting up the analysis.
– Constant Neutral Loss experiments for, e.g., m/z 49 exhibit a high level of
false positives, e.g., from sulfopeptides, iodoacetamide-methionine con-
taining peptides (17) and to some degree from random tryptic peptides that
show singly charged fragments at an m/z of [Prec-98/z]z+.
Another approach for the analysis of tyrosine phosphorylation is a precursor ion
scan experiment targeted at the detection of the phosphotyrosine immonium ion at
m/z 216 (18). Although excellent selectivity can be achieved using high mass accu-
racy on the fragment ion, this experiment does not detect phosphoserine- and phos-
It should be noted in this context that the Neutral Loss/MS3-based methods fre-
quently described for ion trap mass spectrometers (19) are not selective experi-
ments at all. In this approach peptides that are detected by regular MS and exhibit
a strong neutral loss fragment at 98/z in MS/MS are selected for further analysis by
MS3. If the peptide has not been selected for MS/MS in the first place, however, the
neutral loss will go undetected.
3 Identification of Protein Phosphorylation Sites 45
As a consequence, the experimentally more demanding Precursor Ion Scan
method in negative-ion mode is the only approach that currently offers generic
detection of different types of unknown phosphopeptides.
2. A more promising approach that utilizes neutral loss fragmentations in positive
mode is the recently published MIDAS (MRM-initiated detection and sequenc-
ing) workflow (13,14). As a targeted approach, however, it relies on accurate
information about the protein sequence and the quality of the protein digestion.
3. On specific peptide sequences charge-state shifts have been observed between
positive and negative mode, i.e., a phosphopeptide could have 2- as the most abun-
dant charge state in negative mode, but 3+ as the most abundant charge state in
positive mode (4). This shift is dependent on the peptide sequence (number of
basic residues), chromatography conditions (pH value) and instrument parame-
ters (e.g., the declustering potential adjusted on the interface skimmer). As a
consequence, it is advisable to perform the resolving ion trap MS scan in nega-
tive mode. Even if the positive mode MS/MS does not give conclusive results,
the negative mode MS will at least allow to accurately determine the peptide’s
charge state and molecular weight. A second experiment targeted at different
charge states of this peptide can then be used to obtain conclusive results.
4. A peptide sequence containing multiple possible sites of phosphorylation gener-
ates a set of multiple possible phosphopeptides that differ only in the residue
actually modified. These regioisomers possess very similar physicochemical
properties and are often not separated by reverse-phase chromatography. As a
consequence, one should always account for the possibility of mixed MS/MS
spectra, where the regiosisomer precursors are isolated and fragmented together
as they do not separate in m/z or retention time. Each MS/MS spectrum should
be carefully examined and all possible sequence hypotheses tested by annotating
raw data to make a confident assignment of the residue actually modified.
1. Hunter, T. (1995) Protein kinases and phosphatases: the yin and yang of protein phosphoryla-
tion and signaling. Cell 80, 225–236.
2. Cohen, P. (2002) The origins of protein phosphorylation. Nat Cell Biol. 4, E127–30.
3. Reinders, J. and Sickmann, A. (2005) State-of-the-art in phosphoproteomics. Proteomics 5,
4. Steen, H., Jebanathirajah, J. A., Rush, J., Morrice, N. and Kirschner, M. W. (2006)
Phosphorylation analysis by mass spectrometry: myths, facts and the consequences for quali-
tative and quantitative measurements. Mol. Cell. Proteomics 5, 172–181.
5. Annan, R. S., Huddleston, M. J., Verma, R., Deshaies, R. J. and Carr, S. A. (2001) A multidi-
mensional electrospray ms-based approach to phosphopeptide mapping. Anal. Chem. 73,
6. Zappacosta, F., Huddleston, M. J., Karcher, R. L., Gelfand, V. I., Carr, S. A. and Annan, R. S.
(2002) Improved sensitivity for phosphopeptide mapping using capillary column hplc and
microionspray mass spectrometry: comparative phosphorylation site mapping from gel-derived
proteins. Anal. Chem. 74, 3221–3231.
46 C. Weise and C. Lenz
7. Steen, H., Küster, B. and Mann, M. (2001) Quadrupole time-of-flight versus triple-quadrupole
mass spectrometry for the determination of phosphopeptides by precursor ion scanning.
J. Mass Spectrom. 36, 782–790.
8. Schlosser, A., Pipkorn, R., Bossemeyer, D. and Lehmann, W. D. (2001) Analysis of protein
phosphorylation by a combination of elastase digestion and neutral loss tandem mass spec-
trometry. Anal. Chem. 73, 170–176.
9. Hager, J. W. (2002) A new linear ion trap mass spectrometer. Rapid Commun. Mass Spectrom.
10. Le Blanc, J. C. Y, Hager, J. W., Illisiu, A. M. P., Hunter, C., Zhong, F. and Chu, I. (2003)
Unique scanning capabilities of a new hybrid linear ion trap mass spectrometer (Q TRAP)
used for high sensitivity proteomics applications. Proteomics 3, 859–869.
11. Williamson, B. F., Marchese, J. and Morrice, N. A., (2006) Automated identification and
quantification of protein phosphorylation sites by lc/ms on a hybrid triple quadrupole linear
ion trap mass spectrometer. Mol. Cell. Proteomics 5, 337–346.
12. Kottwitz, D., Kukhtina, V., Dergousova, N., Alexeev, T., Utkin, Y., Tsetln, V. and Hucho, F.
(2004) Intracellular domains of the δ-subunits of Torpedo and rat acetylcholine receptors—
expression, purification, and characterization. Protein Expr. Purif. 38, 237–247.
13. Cox, D. M., Zhong, F., Du, M., Duchoslav, E., Sakuma, T. and McDermott, J. C. (2005)
Multiple reaction monitoring as a method for identifying protein posttranslational modifica-
tions. J. Biomol. Tech. 16, 83–90.
14. Unwin, R. D., Griffiths, J. R., Leverentz, M. K. Grallert, A., Hagan, I. M. and Whetton, A. D.
(2005) Multiple reaction monitoring to identify sites of protein phosphorylation with high
sensitivity. Mol. Cell. Proteomics 4, 1134–1144.
15. Covey, T., Shushan, B., Bonner, R., Schröder, W. and Hucho, F. (1991) In Methods in Protein
Sequence Analysis (Jörnvall, H., Höög, J.-O. and Gustavsson, A.-M., eds.) Birkhäuser Verlag,
Basel, Switzerland, pp. 249–256.
16. DeGnore, J. P. and Qin, J. (1998) Fragmentation of phosphopeptides in an ion trap mass spec-
trometer. J. Am. Chem. Soc. Mass Spectrom. 9, 1175–1188.
17. Krüger, R., Hung, Ch.-W., Edelson-Averbukh, M. and Lehmann, W. D. (2005) Iodoacetamide-
alkylated methionine can mimic neutral loss of phosphoric acid from phosphopeptides as
exemplified by nano-electrospray ionisation quadrupole time-of-flight parent ion scanning.
Rapid Commun. Mass Spectrom. 19, 1709–1716.
18. Steen, H., Küster, B, Fernandez, M., Pandey, A. and Mann, M. (2001). Detection of tyrosine
phosphorylated peptides by precursor ion scanning quadrupole tof mass spectrometry in posi-
tive ion mode. Anal. Chem. 73, 1440–1448.
19. Schroeder, M. J., Shabanowitz, J., Schwartz, J. C., Hunt, D. F. and Coon, J. J. (2004) A neutral
loss activation method for improved phosphopeptide sequence analysis by quadrupole ion trap
mass spectrometry. Anal. Chem. 76, 3590–3598.