Protein Protocols Protein Protocols

620 Bonenfant, Mini, and Jenö
up to 6 M. However, due to the cleavage specificity of the enzyme, peptides containing
internal arginine residues are generated, which in some cases suppresses extensive fragmentation,
therefore making the assignment of the site of phosphorylation more difficult.
2. For interfacing peptide separation with an ESI source, the flow rates applied to the separation
system have to meet the requirements of the inlet system of the mass spectrometer.
HPLC can be carried out on commercially packed reverse-phase columns provided the
solvent delivery rate is within the limits of nebulization for the particular inlet system.
Otherwise, either post- or precolumn splitting (20) is required to reduce the flow rate. For
the capillary column system described in this chapter, a practical limitation in solvent
delivery is the ability to generate gradients at low flow rates (1–20 µL/min). Whereas
dual-syringe pumps capable of delivering gradients at 10 µL/min are appropriate for capillary
columns with 500 µm i.d., 100 µm i.d. columns require flow rates of 1 µL/min and
less. Newer models (e.g., Evolution 200, Prolab) have been introduced that are able to
deliver flows of 1 ml/min without the use of flow splitters.
3. Peptides can be lost during the chromatographic separation process. Therefore, it is mandatory
to analyze the complete LC/MS data set with respect to sequence coverage of the
protein. Particular attention should be paid to missing peptides containing
phosphorylatable residues (Ser, Thr, and Tyr). The most common reason for missing peptides
is that they are either too small (therefore often too hydrophilic) to bind to the reversephase
support, or too large (too hydrophobic) to be eluted from the column. These peptides
can often be recovered by selecting a different digestion strategy. The data set should also
be examined for incompletely cleaved peptides, as phosphorylation located close to a
cleavage site often reduces digestion efficiency. Large phosphopeptides can also be
subdigested with a different protease to yield smaller peptides with better elution and
fragmentation behavior.
4. The presence of a phosphorylation site can be observed when the mass of a peptide is
80 Da larger than predicted from the peptide sequence. Such candidates, however, have to
be checked rigorously for the presence of a phosphorylation site. Often, calculating the
masses of incompletely cleaved peptides with various combinations of missed cleavage
sites yield peptide masses identical to the phosphopeptide candidate. A simple test is to
dephosphorylate the digest with alkaline phosphatase and to repeat the analysis. If the
candidate peptide is phosphorylated, dephosphorylation leads to a reduction of its mass by
80 Da (or multiples thereof in the case of multiple phosphorylation). Enzymatic dephosphorylation
with (CIP) can be directly carried out in the digestion mix, as CIP is completely
resistant to proteolysis by enzymes such as trypsin, endoproteinase LysC, or
endoproteinase GluC, although urea tends to inactivate the enzyme. For protein digests in
the 1–500 picomol range, 1 U of CIP was found to be sufficient to lead to complete
dephosphorylation of phosphopeptides. Alternatively, peptides phosphorylated on
serine and threonine tend to undergo neutral loss of H 3PO 4 on fragmentation, while
phosphotyrosine shows no neutral loss. H 3PO 4 loss can be exploited to selectively screen
protein digests for the presence of phosphopeptides, by scanning for decreases in m/z of
all ions between two mass analyzers of a tandem instrument (21). Separate experiments
have to be performed to scan for mass decreases corresponding to each charged form (e.g.,
98 Da for the singly, 48 Da for the doubly, 33 Da for the triply charged ion). Also, by
raising the orifice potential during LC/MS, and scanning in negative ion mode for
the appearance of the PO 3 – (79 Da) and PO2 – (63 Da) fragment ions derived from
phosphoserine, phosphothreonine, or phosphotyrosine, selective phosphopeptide
detection is possible (11). However, the latter two methods can be as much as 10-fold less
sensitive than simple mass detection, and many peptides do not show the required
fragmentation loss.