JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2007; 42: 1590–1598
Published online in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/jms.1348
AP/MALDI-MS complete characterization of the
proteolytic fragments produced by the interaction of
insulin degrading enzyme with bovine insulin
Giuseppe Grasso,1∗ Enrico Rizzarelli1,2 and Giuseppe Spoto1,2
1
2
Dipartimento di Scienze Chimiche, Università di Catania, Viale Andrea Doria 6, 95125, Catania, Italy
Istituto Biostrutture e Bioimmagini, CNR, Viale A. Doria 6, Catania, Italy
Received 4 May 2007; Accepted 12 October 2007
The prominent role that insulin degrading enzyme (IDE) has in the clearance of insulin as well as of
other molecules such as amyloid-b has recently drawn much interest in the scientific community toward
this protease. In order to give an insight into the manner of interaction of IDE with its substrates, several
papers have focused on the structure of the IDE/insulin complex. In this scenario, although the cleavage
sites involved in the interaction of insulin with IDE are known, a convenient experimental method that
is able to identify in a complete and unambiguous way, all the peptide fragments generated by such
interaction has yet to be found. MS-based experiments have often represented to be invaluable tools for
the assessment of the cleavage sites, but the reported MS-spectra always show a partial coverage of all the
peptide fragments generated by the enzyme interaction, lacking a complete characterization.
In this work, we report a new experimental procedure by which an unambiguous as well as complete
assignment of all the peptide fragments generated by the interaction of insulin with IDE is described.
Atmospheric pressure/matrix-assisted laser desorption ionization (AP/MALDI) mass spectra are reported
and the data recorded, together with the introduction of a reduction/alkylation step, allows us to fully
characterize the cleavage sites of the bovine insulin interacting with IDE. Different experimental conditions
are screened and some insights into the IDE/insulin system regarding preference of the cleavage and its
dependence on particular experimental conditions used are also given. Investigation on the tendency that
different insulin fragments have toward aggregation is also carried out.
Good reproducibility, global and unambiguous assignment, low time-consuming experimental
procedure, and requirements of enzyme in small amounts are some of the advantages of the proposed
AP/MALDI based approach. Copyright  2007 John Wiley & Sons, Ltd.
KEYWORDS: insulin; IDE; alkylation; reduction; MALDI
INTRODUCTION
Proteases play a variety of roles in many physiological
processes including regulation of general protein turnover,
elimination of abnormal proteins, zymogen activation,
digestion, and hormone processing.1 – 3 Insulin degrading
enzyme (IDE)4 is a zinc metalloprotease that belongs to
a newly identified superfamily of metalloendopeptidases
containing an unusual active site having the consensus
sequence HXXEH5 (where X is any amino acid, H is histidine,
and E is glutamate). It has been shown that its activity is
inhibited both by sulfhydryl reagents and by chelators6 so
that they can decrease insulin degradation in intact cells,7
while, conversely, overexpression of IDE increases the rate
of insulin degradation several fold.8
The interest in studying IDE has been boosted by
recent findings that showed how this enzyme is capable of
Ł Correspondence to: Giuseppe Grasso, Dipartimento di Scienze
Chimiche, Università di Catania, Viale Andrea Doria 6, 95125,
Catania, Italy. E-mail: [email protected]
Copyright  2007 John Wiley & Sons, Ltd.
degrading many other substrates: ˇ-amyloid peptide among
the others.9 It is well known10 that a wrong regulation of
the latter can cause several pathological conditions such as
Alzheimer’s disease (AD) and Parkinson’s disease (PD), and
support for the role of IDE in amyloid diseases also comes
from recent genetic studies, showing a significant association
between the gene locus encoding IDE on chromosome 10q
and AD.11 Also, the gene encoding IDE has been identified
as a susceptibility gene in type 2 diabetes.12
It has been shown13 that the common feature shared by
IDE substrates is the ability to form amyloid fibrils14 under
certain physiological conditions. It is therefore tempting
to speculate that IDE selects its substrates on the basis of
their potential ability to adopt a ˇ-sheet conformation upon
interaction with the enzyme.9 Very recently, the structure
of human IDE in complex with four different substrates has
been reported and conformation changes of IDE substrates as
well as cleavage sites have been highlighted.15 Confirmation
of the cleavage sites involved in the interaction of IDE
with insulin have been searched since the discovery of
AP/MALDI-MS studies of insulin–IDE interaction
such interaction and, overall, several different experimental
approaches16 – 18 have been applied in order to identify them.
Mass spectrometry (MS), often coupled to high-performance
liquid chromatography (HPLC), has demonstrated to be one
of the most valuable tool in this scenario.19,20 Nevertheless,
so far, some cleavage sites involved in the interaction of
IDE with its substrates were elusive to the MS investigation
and in order to have a global and complete assessment
of the peptide fragments generated by the interaction of
IDE with insulin, a recourse to different approaches21 or to
complicated and time-consuming experimental procedures
utilizing [125I]iodoinsulin were necessary.22 Additionally,
although the nature of the species present in solution when
bovine insulin undergoes self-assembly, aggregation and
amyloid fibril formation, has been widely investigated,23,24
the complex equilibrium between monomeric and oligomeric
forms of the protein that exists in insulin solutions can
contribute toward complicating the MS-spectra further.25
Finally, the presence of disulfide bonds between cysteine
residues of the insulin chains, as well as buffers and salts
in the insulin solution, causes additional problems that
makes a complete MS-based cleavage sites assessment a
very challenging task.18
Atmospheric pressure/matrix assisted laser desorption
ionization (AP/MALDI) MS has already proved to be a very
valuable tool in a number of important applications, including the identification of proteins,26 the structural analysis of
oligosaccharides,27 the characterization of enzymes immobilized on solid support,28 and the study of phosphopeptides,29
due to its ability to detect intact molecular masses up to
several 1000 Da.30
In this work, the use of AP/MALDI-MS together with
the introduction of a reduction/alkylation step allowed
us to produce MS data that confirmed all the cleavage
sites involved in the IDE/insulin interaction at once.
Confirmation of a complete assignment was doubly checked
by both chemical modification of the peptide fragments
(reduction/alkylation step) and MS/MS patterns of some
ambiguous peaks. Different experimental conditions were
screened and some insights into the IDE/insulin system
regarding preferentiality of the cleavage and its dependence
on particular experimental conditions used are also given.
Good reproducibility, global and unambiguous assignment, low time-consuming experimental procedure, and
requirements of enzyme in low amounts are some of the
advantages that the new AP/MALDI MS-based approach
has in the study of enzyme/substrate interactions as is
demonstrated by the experimental results obtained in the
case of the IDE/insulin system that are reported below.
EXPERIMENTAL
Materials
IDE expressed in Spodoptera frugiperda was purchased
from CALBIOCHEM. Insulin from bovine pancreas and
all the other reagents used [phosphate buffer solution
(PBS), ˛-cyano-4-hydroxycinnamic acid (CHCA), trifluoro
acetic acid (TFA), acetonitrile (C2 H3 N), ethanol solution,
ammonium bicarbonate (NH4 HCO3 ) and iodoacetamide]
Copyright  2007 John Wiley & Sons, Ltd.
were all purchased from SIGMA-ALDRICH, while ZipTipSCX
pipette tips were from MILLIPORE. DTT (1,4-dithio-DLthreitol) was purchased from Fluka.
AP/MALDI-MS experimental setup
All the AP/MALDI-MS experiments were carried out
using a Finnigan LCQ Deca XP PLUS (Thermo Electron
Corporation, USA) ion trap spectrometer that was fitted with
a MassTech Inc. (USA) AP/MALDI pulsed dynamic focusing
(PDF)-source.31 The latter consists of a flange containing
a computer-controlled X-Y positioning stage and a digital
camera, and is powered by a control unit that includes a
pulsed nitrogen laser (wavelength 337 nm, pulse width 4 ns,
pulse energy 300 µJ, repetition rate upto 10 Hz) and a PDF
module that imposes a delay of 25 µs between the laser pulse
and the application of the high voltage to the AP/MALDI
target plate. PDF has been shown to improve S/N ratios
in the AP-MALDI spectra.32 Laser power was attenuated
to about 55%. The target plate voltage was 1.8 kV. The
ion trap inlet capillary temperature was 230 ° C. Capillary
and tube lens offset voltages of 30 and 15 V, respectively,
were applied. Other mass spectrometer parameters were as
follows: multipole 1 offset at 3.75 V, multipole 2 offset at
9.50 V, multipole RF amplitude at 400 V, lens at 24.0 V
and entrance lens at 88.0 V. Automatic Gain Control
(AGC) was turned off and instead, the scan time was
fixed by setting the injection time to 220 ms and using
five microscans. Although there was the risk of losing
resolution, the latter experimental conditions were chosen
as sensitivity was the main goal in most experiments, where
the presence of all the insulin fragments had to be verified at
different experimental conditions. For the same reason, even
if about 1 min acquisition per sample was usually performed,
acquisition up to 5 min was in some cases necessary due
to the low signal/noise level spectra recorded under the
particular experimental conditions used, as it is outlined
below. Experiments were reproduced three to five times for
each experimental protocol applied.
Spectra of the insulin/IDE solutions were acquired in a
data-dependent fashion by first acquiring full extended mass
range from m/z 200 to 4000 (standard MS range from m/z
50 to 2000 was not informative for our purpose) followed
by MS/MS scans of the most intense ions of the previous
full MS scan. Applying a source fragmentation having low
collision energy (100 V) improved the spectral appearance
by removing some of the numerous peaks due to matrix and
salt clusters. MS/MS scans were acquired using an isolation
width of 5 m/z, activation qz of 0.250, activation time of 30 ms,
and normalized collision energy (NCE) in the range 30–40%
depending on the ion [NCE is the amplitude of the resonance
excitation RF voltage scaled to the precursor mass based on
the formula: RF amplitude D [NCE%/30%] (precursor ion
mass ð tick amp slope C tick amp intercept), where tick amp
slope and tick amp intercept are instrument-specific values.
For our LCQ Deca, 35% NCE for m/z 1000 D 1.8 V].
Several experimental parameters were investigated over
a wide range of values and the best MS-spectra quality
was reached when the following experimental conditions
were matched: [insulin] D 18 µM, [IDE] D 36 nM, time of
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms
1591
1592
G. Grasso, E. Rizzarelli and G. Spoto
digestion at 37 ° C D 60 min, use of ZipTipSCX pipette tips,
[PBS] D 150 mM, pH D 7.4; matrix solution: [CHCA] D 5 mM
in 30% TFA (0.1%) and 70% CH3 CN (CHCA was purified
by re-crystallization from ethanol solution). Any further step
of washing the obtained matrix crystals did not seem to
improve the mass spectra quality. The solution obtained
after the action of IDE contained different insulin fragments
according to the time of digestion applied, as any variation
of the experimental parameters listed above caused changes
in the mass spectrum recorded that have been rationalized,
as it is described in the next paragraph. In order to better
assign the insulin fragments, a reduction/alkylation step was
carried out in the solution containing the insulin fragments
without any purification step, according to the following
procedure: 10 µl of insulin/IDE solution was added to 20 µl
of 100 mM NH4 HCO3 . Then, 1 µl of 50 mM DTT in water was
added to the resulting solution and incubated for 15 min at
60 ° C. The solution was allowed to cool and then, 5 µl of
22 mM iodoacetamide in water was added and incubated for
25 min at room temperature in the dark. As much as 1.4 µl
of 5 mM DTT was finally added and incubated for 25 min at
room temperature in the dark. A solution of 1 µl obtained
was then spotted on the AP/MALDI plate by ZipTipSCX
pipette tips and dried. CHCA of 0.5 µl was layered on top
and the AP/MALDI mass spectrum was acquired when the
spot was completely dry. It is important to note that while
the reduction/alkylation step is usually carried out before
the enzymatic digestion, in our experimental protocol the
order of the two processes is inverted and some advantages
derive from such approach: (1) absence of purification step
(no proteolytic enzyme would work in the presence of DTT),
(2) possibility to monitor by MS if there are any changes in
the way IDE degrades whole insulin (instead of separate
insulin chains) when the experimental digestion conditions
are altered (time of reaction, pH, presence of metal ions, etc.).
RESULTS AND DISCUSSION
Although insulin is one of the most widely studied protein
hormones due to its use in the treatment of insulin-dependent
diabetes, it still remains one of the most problematic proteins
to be studied experimentally.33 MS investigations of insulin
solutions have been usually carried out by electrospray
ionization (ESI),34 nanoflow ES,25,35 or MALDI-time of flight
(TOF),36 while in the past the interaction of the hormone
with IDE was addressed by fast atom bombardment (FAB)20
or ESI coupled to HPLC.19 Insulin structure consists of two
chains, usually named A and B, linked by disulfide bonds
between the cysteines at positions A7 and B7 and between
positions A20 and B19 (Fig. 1). An intrachain disulfide bond
Figure 1. Primary structure of bovine insulin showing the
cystine bridges (lines) and the bonds broken by IDE (arrows).
Copyright  2007 John Wiley & Sons, Ltd.
Table 1. Complete assignment of the peaks generating from
the insulin fragments produced by IDE on an 18 µM insulin
solution at 37 ° C. The time of reaction was 1 h and all the
possible fragment combinations were detected experimentally
except for the peaks indicated with ND (not detected). For
explanation/discussion see text
Fragment combinations
A1–13 C B1–9
A1–13 C B1–10
A1–13 C B1–13
A1–13 C B1–14
A1–13 C B1–16
A1–14 C B1–9
A1–14 C B1–10
A1–14 C B1–13
A1–14 C B1–14
A1–14 C B1–16
A14–21 C B10–24
A14–21 C B10–25
A14–21 C B10–30
A14–21 C B11–24
A14–21 C B11–25
A14–21 C B11–30
A14–21 C B14–24
A14–21 C B14–25
A14–21 C B14–30
A14–21 C B15–24
A14–21 C B15–25
A14–21 C B15–30
A14–21 C B17–24
A14–21 C B17–25
A14–21 C B17–30
A15–21 C B10–24
A15–21 C B10–25
A15–21 C B10–30
A15–21 C B11–24
A15–21 C B11–25
A15–21 C B11–30
A15–21 C B14–24
A15–21 C B14–25
A15–21 C B14–30
A15–21 C B15–24
A15–21 C B15–25
A15–21 C B15–30
A15–21 C B17–24
A15–21 C B17–25
A15–21 C B17–30
Calculated
peaks (m/z)
Experimental
peaks (m/z)
2313.0
2450.1
2791.3
2862.3
3138.5
2476.0
2613.1
2954.3
3025.3
3301.5
2749.2
2896.3
3456.6
2612.2
2759.3
3319.6
2271.0
2418.1
2978.4
2200.0
2347.0
2907.3
1923.8
2070.9
2631.2
2586.1
2733.2
3293.5
2449.1
2596.2
3156.5
2107.9
2255.0
2815.3
2036.9
2183.9
2744.2
1760.7
1907.8
2468.1
2312.2
2450.0
ND
ND
ND
2475.0
2613.1
ND
ND
ND
2749.6
2896.2
3456.9
2612.4
2759.7
3320.4
2271.5
2418.5
2978.6
2203.3
2347.4
2907.3
1923.9
2070.5
2632.8
2586.8
2733.1
3294.1
2449.4
2596.7
3157.0
2108.3
2255.0
2815.4
2036.5
2184.3
2744.5
1760.5
1907.7
2469.7
occurs also between the cysteines at A6 and A11. In Fig. 2
the AP/MALDI mass spectrum obtained by spotting 1 µl of
18 µM solution of insulin in PBS in the range m/z 2100–3800
is shown (ZipTipSCX pipette tips were used as described in
the previous section). It is evident that the soft ionization
method used does not generate any insulin fragment in the
experimental conditions used, so only the molecular peak
at m/z 2866.4 (doubly charged, molecular weight of bovine
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms
AP/MALDI-MS studies of insulin–IDE interaction
insulin is 5733.5 Da) is detected together with the sodiated
ones at m/z 2877.4 and 2888.7. The latter are present because
of the high salt content of the buffered solution, a necessary
condition when enzyme activity is under investigation,
explaining why our MS approach without any purification
steps can be a very challenging task for a buffered solution.
Other peaks due to the two separated insulin chains (m/z
2309.7, 2333.3, and 2357.8 for chain A; m/z 3399.1 for chain B)
are barely visible, not being able to properly stick out from
the average noise level. It is worth noting that impossibility to
scan at m/z > 4000 due to a limited mass range of our ion trap
spectrometer does not hinder the search for possible insulin
fragments as none of the generated peptides are expected to
fall at m/z > 3500 (Table 1). Moreover, the doubly charged
insulin molecular ion (Fig. 2) is a good indication of the
presence or absence of intact insulin molecules, which is a
good indicator of the completion of the IDE/insulin reaction.
On the contrary, doubly charged insulin fragment peaks
(all of which should fall within the standard mass range
m/z 50–2000) are not easily detected in our experimental
conditions. Although we can only make hypotheses about the
possible explanations (differences in ionization properties
between insulin molecules and its fragments, etc.), the
absence of doubly charged insulin fragment peaks does
not represent a problem for our study, as the latter is
mainly aimed at a qualitative investigation of the IDE/insulin
interaction.
In Fig. 3 the mass spectrum detected for the same
insulin solution after reaction with IDE, according to the
experimental procedure described in the previous section,
is reported. It is clear how, this time, the spectrum is dense
with peaks attributed to insulin fragments generated by the
interaction with IDE, while the insulin molecular peak has
disappeared. Because of the multiple cleavage sites and the
presence of cystine bridges between the two insulin chains,
Figure 2. AP/MALDI mass spectrum obtained from 1 µl of
18 µM insulin in PBS solution (range m/z 2100–3800). The main
peak at m/z 2866.4 is due to the doubly positive-charged
insulin molecule, while the mono- and doubly sodiated peaks
are also detectable due to the high salt content of the solution.
Although, very close to the limit of detection, peaks originated
from the two separate insulin chains are also highlighted.
Copyright  2007 John Wiley & Sons, Ltd.
Figure 3. AP/MALDI mass spectrum obtained from 1 µl of
18 µM insulin in PBS solution (range m/z 2020–3500) after
interaction with IDE as described in the text. The high number
of detected peaks is due to the many possible combinations of
insulin fragments originating from the two different chains of
the molecule and the presence of sodiated peaks. A complete
assignment is reported in Table 1. The spectrum obtained by
MS/MS experiments for the peak at m/z 2418.6 is reported in
the insert. Owing to the presence of cystine bridges between
the insulin chains, it is very difficult and laborious to assign all
the peaks as it is discussed in the text.
the spectrum is quite crowded with peaks coming from all
the possible combinations between the A-chain and the Bchain fragments (Table 1). In order to assign all the peaks
present in the spectrum of Fig. 3, some MS/MS experiments
were also carried out and all the nine cleavage sites reported
in the literature were confirmed21 (see insert of Fig. 3 as an
example). It is important to notice from Table 1 that, although
fragments that involve cleavage sites (B 13–14), (B 14–15),
and (B 16–17) are detected in the spectrum, unambiguously
demonstrating the action of IDE at these peptide bonds, some
possible calculated peaks are missing from the experimental
spectrum, precisely the ones involving fragments B (1–13), B
(1–14), and B (1–16). Although any quantitative conclusions
drawn by these arguments should consider the difference in
the ionization process for the different peptide fragments,37,38
these results confirm that cleavage site (B 9–10) is one of
the most preferred by IDE.19 However, presence of peaks
at m/z 2450.0 and 2613.1, involving fragment B (1–10),
excludes the possibility of a 100% rate of cleavage action
by IDE even for this peptide bond. Nevertheless, within
the limit mentioned above, this MS-based approach, by
giving information about all the cleavage sites at once, has
the possibility to give indications on the overall interaction
process. Moreover, it is worthwhile to note that while the
standard enzyme activity measurements usually deal with
the kinetic parameter values of enzyme/substrate systems
in different experimental conditions (inhibitors, activators,
etc.),39 this new approach focuses on the molecular processes
involved, being able to investigate and screen all the peptide
fragments generated by such interactions.
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms
1593
1594
G. Grasso, E. Rizzarelli and G. Spoto
Figure 5. AP/MALDI MS/MS fragmentation pattern for the m/z
1431.7 peak, confirming the assignment for the alkylated
peptide fragment B (14–25), having the sequence
ALYLVCGERGFF. Note how the absence of cystine bridges
allows a straightforward prediction of the fragmentation
pattern.
Figure 4. AP/MALDI mass spectrum obtained from 1 µl of
18 µM insulin in PBS solution in the range m/z 100–2010 (a) and
2200–2750 (b) after the reduction/alkylation step. Peaks are
very well resolved and their complete assignment is reported in
Table 2. Sodiated peaks are also present, while peaks
detected at m/z < 1000 are mainly attributed to matrix clusters.
The insert in (a) shows a very good accordance between the
experimental line shape (solid line) and the calculated isotopic
distribution (dashed line) for the peak at m/z 1431.7.
Although the assignment reported in Table 1 was already
complete, it is very laborious and difficult to verify such
assignment by MS/MS experiments of every single peak, as
the fragmentation pattern for peptide fragments that involve
both insulin chains is difficult to be unambiguously assigned
(see insert of Fig. 3). For this reason, in order to doublecheck the results and to have a further confirmation of
cleavage sites, we performed a reduction/alkylation step
of the insulin/IDE solution as it was described in the
previous section. The spectrum obtained for the two different
m/z range values of relevance (100–2010 and 2200–2750)
is reported in Fig. 4, while in Table 2 the assignment of
the experimental peaks is also reported, and this time
MS/MS experiments produced fragmentation patterns that
were easily and unambiguously assigned (for example, see
Fig. 5). Supporting again the presence of a strong action
by IDE at cleavage site (B 9–10), all the peaks involving
fragments B (1 X, X > 10) are missing. Nevertheless,
Copyright  2007 John Wiley & Sons, Ltd.
fragment B (1–10) produces again a detectable peak, ruling
out the possibility of a 100% rate of cleavage action by
IDE for this peptide bond. Therefore, the MS data after the
reduction/alkylation step perfectly confirmed the previous
assignment and allowed MS/MS fragmentation patterns to
be easily assigned, proving to be a very good chemical tool for
the spectrometric investigation of cleavage sites originating
from a general protein/protease interaction where cystine
bridges are involved.
A drawback of this approach is that some expected peaks
at m/z < 1000 are missing and this can be confidently
attributed to the significant interference of matrix and buffer
clusters at this mass range.40 Nevertheless, although in this
case the whole spectrum was shifted to lower mass range, this
problem could be solved by alkylating the peptide residues
with alkyl groups having larger masses, shifting all the
detected peaks toward higher mass ranges and allowing the
detection of species that would be otherwise covered by
peaks originating from matrix clusters.
Despite the above-mentioned drawback, the reduction/alkylation step offers several advantages that are not
available in the standard approach. First, by lowering the
mass range, a gain in the mass resolution (the dimensionless
ratio of the mass of the peak divided by its width) is achieved
so that in most cases assignment can be further verified by
isotopic distribution (for example, see insert in Fig. 4(a)).
Second, in our experimental conditions, sensitivity is also
increased as demonstrated by the mass spectra obtained for
a solution where the action of IDE on insulin molecules was
stopped after 30 min (Table 3). In the latter case, while some
cleavage sites could not be identified in the insulin/IDE
solution because of the short time of reaction (top part of
Table 3), all but one of the cleavage sites could be identified
after the reduction/alkylation step (bottom part of Table 3).
Missing fragments B (1–10) and B (11 X, X D 13, 14, 16,
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms
AP/MALDI-MS studies of insulin–IDE interaction
Table 2. Complete assignment of peaks generating from the
insulin/IDE solution (time of reaction D 1 h) after the
reduction/alkylation step. Note that fragments could contain up
to four cysteine residues and this was considered for the
computation of the expected m/z values. All the possible
alkylated fragments were detected experimentally except for
the peaks indicated with ND (not detected). For
explanation/discussion see text
Fragments
Calculated
mass (Da)
Alkylated
peaks (m/z)
Experimental
peaks (m/z)
A (1–13)
A (1–14)
A (1–21)
A (14–21)
A (15–21)
B (1–9)
B (1–10)
B (1–13)
B (1–14)
B (1–16)
B (1–24)
B (1–25)
B (1–30)
B (10–13)
B (10–14)
B (10–16)
B (10–24)
B (10–25)
B (10–30)
B (11–13)
B (11–14)
B (11–16)
B (11–24)
B (11–25)
B (11–30)
B (14–16)
B (14–24)
B (14–25)
B (14–30)
B (15–16)
B (15–24)
B (15–25)
B (15–30)
B (17–24)
B (17–25)
B (17–30)
B (25–30)
B (26–30)
1310.6
1473.6
2337.9
1045.4
882.3
1003.4
1140.5
1481.7
1552.7
1828.9
2690.3
2837.4
3397.7
496.3
567.3
843.4
1704.8
1851.9
2412.2
359.2
430.2
706.4
1567.8
1714.9
2275.2
365.2
1226.6
1373.7
1934.0
294.2
1155.6
1302.6
1862.9
879.4
1026.5
1586.8
725.4
578.3
1482.6
1645.6
2566.9
1103.4
940.3
1061.4
1198.5
1539.7
1610.7
1886.9
2805.3
2952.4
3512.7
497.3
568.3
844.4
1762.8
1909.9
2470.2
360.2
431.2
707.4
1625.8
1772.9
2333.2
366.2
1284.6
1431.7
1992.0
295.2
1213.6
1360.6
1920.9
937.4
1084.5
1644.8
726.4
579.3
1482.6
1645.7
2566.8
1103.6
ND
1061.7
1198.8
ND
ND
ND
ND
ND
ND
ND
568.1
845.1
1762.8
1909.9
2470.6
ND
ND
ND
1625.9
1772.7
2333.5
ND
1284.9
1431.7
1991.9
ND
1213.7
1360.8
1920.8
ND
1084.7
1644.8
ND
ND
24, 25, 30) proved also that (B 10–11) is a less favored cleavage site for IDE. Therefore the potential of this MS-based
approach must be highlighted, as any changes in the experimental conditions that could alter the IDE ability to cleave
insulin in all or only some of the cleavage sites reported
could be easily monitored by changes in the mass spectra
of Figs 3 and 4. In order to prove this point, we carried
out the same experiments on aged insulin solutions. In our
Copyright  2007 John Wiley & Sons, Ltd.
Table 3. Assignment of peaks generated from the insulin/IDE
solution for a reaction time of 30 min before (top) and after
(bottom) the reduction/alkylation step. Although, in the latter
case some expected peaks appearing in Table 2 are still
missing (bold), the detected peaks allowed a more complete
assignment of the cleavage sites. Particularly, it is impossible
to detect cleavage sites (B 10–11), (B 13–14), (B 14–15), and
(B 24–25) from the insulin/IDE solution, while, after the
reduction/alkylation step, fragments B (10–14), B (14–24), B
(14–25), and B (15–25) confirmed three of those ((B 13–14), (B
14–15), and (B 24–25)). For explanation/discussion see text
Fragment combinations
A1–13 C B1–9
A1–14 C B1–9
A14–21 C B17–25
Experimental peaks (m/z)
2312.2
2475.0
2070.5
Fragments
Experimental peaks (m/z)
A (1–13)
A (1–14)
A (1–21)
A (14–21)
B (1–9)
B (1–10)
B (10–14)
B (10–16)
B (10–24)
B (10–25)
B (10–30)
B (11–24)
B (11–25)
B (11–30)
B (14–24)
B (14–25)
B (14–30)
B (15–16)
B (15–24)
B (15–25)
B (15–30)
B (17–25)
B (17–30)
1482.6
1645.7
2566.8
1103.6
1060.9
ND
567.9
ND
ND
ND
2470.6
ND
ND
ND
1284.9
1431.6
ND
ND
ND
1360.8
ND
1084.7
1644.8
working buffer (PBS at pH 7.4) insulin exists in equilibrium
as a mixture of monomers, dimers, hexamers, and possibly
higher oligomeric species, and we refer to the abundance
of work in the literature addressed to study such equilibria
in solution.41 – 43 For our purposes, it is sufficient to mention
that aged insulin solutions have a tendency toward fibrillation that is higher for lower pH and higher temperature at
which aging occurs.44,45 The effect that 48 h aging at 37 ° C
of a 18 µM insulin in PBS solution has on the mass spectral
appearance (Fig. 2) is a drastic decrease of the overall intensity. Moreover, the relative intensity of the detected peaks is
also changed as the molecular peak at m/z 2866.4 (doubly
charged) almost disappears, while there is a steep increase
of the relative intensity of the chain A peaks at m/z 2309.7,
2333.3, and 2357.8 (data not shown), revealing a loss of intact
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms
1595
1596
G. Grasso, E. Rizzarelli and G. Spoto
insulin monomers available in solution (the same result was
confirmed also after the reduction/alkylation step, data not
shown). The action of IDE on such aged insulin solution was
investigated by the MS approach proposed above. IDE was
added to the aged insulin solution (48 h aging at 37 ° C, 18 µM
in PBS) and the mass spectrum was recorded after a reaction
time of 1 h at 37 ° C. Even in this case, the introduction of the
reduction/alkylation step increased the sensitivity, allowing
the detection of cleavage site (B 16–17) that could not be
assigned from the standard mass spectrum (Table 4). From
these results, it is evident that although the aging of the
insulin solution drastically lowers the amount of hormone
fragments generated by IDE, cleavage sites remain the same,
demonstrating that IDE is not able to cleave the fibrillose
forms of insulin formed because of the aging process.
Finally, previous studies have shown that various insulin
degradation products have biological effects including both
early degradative products that consist of partially degraded
B-chain attached to intact A-chain by disulfide bonds as well
as small fragments of the B-chain.46 Interesting information
about the tendency that insulin fragments produced by
IDE have toward fibrillation could be gathered by keeping
freshly prepared insulin solution in contact with IDE for long
time intervals. Changes in the mass spectrum of Fig. 3 were
monitored for different times and the gradual disappearing
of the detected peaks was interpreted as a proof of their
tendency toward alteration (aggregation/fibrillation). In
Fig. 6 the mass spectrum of IDE/insulin solution that was
left at 37 ° C for 24 h (a) and 48 h (b) is reported together
with the mass spectrum of the same solution as in (b) but
after the reduction/alkylation step (c). It is possible to
see that although all the fragments tend to gradually
decrease their intensity, some of them never disappear
from the spectrum even after very long reaction time
(Table 5). These experimental results give an insight into the
different attitudes that the insulin fragments have toward
aggregation and therefore, they are very different from the
ones obtained in the case of adding IDE to the aged insulin
solution (Table 4). The conclusion that, at these experimental
conditions, fragments [A14–21 C B17–25], [A15–21 C
B14–24], [A15–21 C B15–25], [A15–21 C B14–25],
A (1–21), [A15–21 C B15–24], and [A1–13 C B1–13]
are the least keen to form fibrils, shows the potentiality
of this MS approach for studying similar biomolecular
Figure 6. AP/MALDI mass spectrum of IDE/insulin solution left at 37 ° C for 24 h (a) and 48 h (b). The reduction/alkylation step was
carried out in the 48-h solution and the AP/MALDI mass spectrum is reported in (c). It is possible to see that although all the
fragments tend to gradually decrease their intensity, some of them never disappear from the spectrum even after a very long
reaction time (Table 5).
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms
AP/MALDI-MS studies of insulin–IDE interaction
Table 4. Assignment of peaks generating from the insulin
aged/IDE solution for a reaction time of 1 h before (top) and
after (bottom) the reduction/alkylation step. Only the detected
peaks are listed from which it is possible to infer all the
expected cleavage sites. (B 16–17) could be inferred only by
the presence of the peak at m/z 1644.8 in the mass spectrum
obtained from the insulin/IDE solution that underwent the
reduction/alkylation step
Fragment combinations
A1–13 C B1–9
A1–14 C B1–9
A14–21 C B11–25
A14–21 C B14–24
A14–21 C B15–25
Experimental peaks (m/z)
2313.1
2475.2
2759.5
2271.5
2347.4
Fragments
Experimental peaks (m/z)
A (1–13)
A (1–14)
A (1–21)
A (14–21)
B (1–9)
B (1–13)
B (14–24)
B (17–30)
1482.5
1645.8
2566.8
1105.2
1061.3
1537.9
1284.5
1644.8
interactions. Sampling the same solution at different times is
like taking a global picture of the entire system at different
stages of the interaction and this can give an insight not
only on the way the interaction takes place but also on
the molecular mechanism involved as well as the different
chemical attitudes of the biomolecular products formed.
Table 5. Assignment of peaks generating from the insulin/IDE
solution for a reaction time of 48 h. Although loss of water
molecules has to be considered in order to identify some of the
fragment combinations reported (top), the reduction/alkylation
step allowed once again, a confirmation of such assignment
(bottom)
Fragment combinations
A14–21 C B17–25 H2 O
A14–21 C B17–25
A15–21 C B14–24 H2 O
A15–21 C B14–24
A15–21 C B15–25 H2 O
A15–21 C B15–25
A15–21 C B14–25 H2 O
A15–21 C B15–24 H2 O
A (1–21)
A1–13 C B1–13 H2 O
Experimental peaks (m/z)
2052.5
2070.5
2090.8
2108.3
2165.9
2184.3
2238.8
2022.5
2335.3
2775.0
Fragments
Experimental peaks (m/z)
A (1–13)
A (1–14)
A (1–21)
A (14–21)
B (1–9)
B (1–13)
B (14–24)
B (15–24)
B (15–25)
B (17–25)
B (17–30)
1482.6
1645.7
2566.8
1103.6
1059.7
1539.0
1284.9
1213.7
1360.8
1084.7
1644.8
REFERENCES
CONCLUSIONS
An experimental procedure that is low time-consuming
(absence of cleaning and washing steps in the experimental
protocol) and that requires small amount of samples, yielding
reproducible mass spectra that allow a global and unambiguous assignment of all the cleavage sites involved for the
insulin/IDE interaction, is proposed. AP/MALDI MS technique proved to be a powerful tool for the generation of intact
molecular fragments that are representative for the peptidase
action, and the introduction of a reduction/alkylation step
confirmed the results obtained, simplifying the spectrum and
proving to be a very good chemical tool for the spectrometric
investigation of cleavage sites originating from a general protein/protease interaction where cystine bridges are involved.
Different experimental conditions were screened and the
mass spectra obtained in each case have been discussed,
showing the potential of the approach for monitoring the
IDE/insulin interaction. Differences on the preferences of
cleavage sites and on the tendency of insulin fragments
toward aggregation have also been highlighted.
Acknowledgements
We thank MIUR (FIRB RBNE03PX83, RBIN04L28Y, and PRIN 2005
n 2005038704 004) for partial financial support.
Copyright  2007 John Wiley & Sons, Ltd.
1. Turk B. Targeting proteases: successes, failures and future
prospects. Nature Reviews Drug Discovery 2006; 5: 785.
2. Puente XS, Sanchez LM, Overall CM, Lopez-Otin C. Human and
mouse proteases: a comparative genomic approach. Nature
Reviews Genetics 2003; 4: 544.
3. Neurath H. Proteolytic processing and physiological regulation.
Trends in Biochemical Sciences 1989; 14: 268.
4. Farris W, Mansourian S, Chang Y, Lindsley L, Eckman EA,
Frosch MP, Eckman CB, Tanzi RE, Selkoe DJ, Guénette S.
Insulin-degrading enzyme regulates the levels of insulin,
amyloid -protein, and the -amyloid precursor protein
intracellular domain in vivo. Proceedings of the National Academy
of Sciences 2003; 100: 4162.
5. Becker AB, Roth RA. An unusual active site identified in a
family of zinc metalloendopeptidases. Proceedings of the National
Academy of Sciences 1992; 89: 3835.
6. Duckworth WC. Insulin degradation: mechanisms, products,
and significance. Endocrine Reviews 1988; 9: 319.
7. Gehm BD, Rosner MR. Regulation of insulin, epidermal growth
factor, and transforming growth factor-alpha levels by growth
factor-degrading enzymes. Endocrinology 1991; 128: 1603.
8. Kuo WL, Gehm BD, Rosner MR. Regulation of insulin
degradation: expression of an evolutionarily conserved insulindegrading enzyme increases degradation via an intracellular
pathway. Molecular Endocrinology 1991; 5: 1467.
9. Kurochkin IV. Insulin-degrading enzyme: embarking on
amyloid destruction. Trends in Biochemical Sciences 2001; 26: 421.
10. Blomqvist ME-L, Silburn PA, Buchanan DD, Andreasen N,
Blennow K, Pedersen NL, Brookes AJ, Mellick GD, Prince JA.
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms
1597
1598
G. Grasso, E. Rizzarelli and G. Spoto
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Sequence variation in the proximity of IDE may impact age
at onset of both Parkinson disease and Alzheimer disease.
Neurogenetics 2004; 5: 115.
Bertram L, Blacker D, Mullin K, Keeney D, Jones J, Basu S,
Yhu S, McInnis MG, Go RCP,
Vekrellis K,
Selkoe DJ,
Saunders AJ, Tanzi RE. Evidence for genetic linkage of
Alzheimer’s disease to chromosome 10q. Science 2000; 290: 2302.
Fakhrai-Rad H, Nikoshkov A, Kamel A, Fernström M, Zierath
JR, Norgren S, Luthman H, Galli J. Insulin-degrading enzyme
identified as a candidate diabetes susceptibility gene in GK rats.
Human Molecular Genetics 2000; 9: 2149.
Kurochkin IV, Goto S. Alzheimer’s ˇ-amyloid peptide
specifically interacts with and is degraded by insulin degrading
enzyme. FEBS Letters 1994; 345: 33.
Kurochkin IV. Amyloidogenic determinant as a substrate
recognition motif of insulin-degrading enzyme. FEBS Letters
1998; 427: 153.
Shen Y, Joachimiak A, Rosner MR, Tang W-J. Structures of
human insulin-degrading enzyme reveal a new substrate
recognition mechanism. Nature Letters 2006; 443: 870.
Muir AV, Offord RE, Davies JG. The identification of a major
product of the degradation of insulin by ‘insulin proteinase’.
Biochemical Journal 1986; 237: 631.
Davies JG, Muir AV, Offord RE. Identification of some cleavage
sites of insulin by insulin proteinase. Biochemical Journal 1986;
240: 609.
Davies JG, Muir AV, Rose K, Offord RE. Identification of
radioactive insulin fragments liberated by insulin proteinase
during the degradation of semisynthetic [[3H]GlyA1]insulin
and [[3H]PheB1]insulin. Biochemical Journal 1988; 249: 209.
Vu L, Stöcklin R, Rose K, Offord RE. Facile identification by
electrospray mass spectrometry of the insulin fragment A14 – 21 B17 – 30 produced by insulin proteinase. Rapid Communications in
Mass Spectrometry 1993; 7: 1048.
Savoy LA, Jones RML, Pochon S, Davies G, Muir AV, Offord RE,
Rose KB. Identification by fast atom bombardment mass
spectrometry of insulin fragments produced by insulin
proteinase. Biochemical Journal 1988; 249: 215.
Duckworth WC, Bennett RG, Hamel FG. Insulin degradation:
progress and potential. Endocrine Reviews 1998; 19: 608.
Duckworth WC, Garcia JV, Liepnieks JJ, Hamel FG, Hermodson MA, Frank BH, Rosner MR. Drosophila insulin degrading
enzyme and rat skeletal muscle insulin protease cleave insulin
at similar sites. Biochemistry 1989; 28: 2471.
Manno M, Craparo EF, Martorana V, Bulone D, San Biagio PL.
Kinetics of insulin aggregation: disentanglement of amyloid
fibrillation from large-size cluster formation. Biophysical Journal
2006; 90: 4585.
Hong D-P, Ahmad A, Fink AL. Fibrillation of human insulin A
and B chains. Biochemistry 2006; 45: 9342.
Nettleton EJ, Tito P, Sunde M, Bouchard M, Dobson CM,
Robinson CV. Characterization of the oligomeric states of
insulin in self-assembly and amyloid fibril formation by mass
spectrometry. Biophysical Journal 2000; 79: 1053.
Mehl JT, Cummings JJ, Rohde E, Yates NN. Automated protein
identification using atmospheric-pressure matrix-assisted
laser desorption/ionization. Rapid Communications in Mass
Spectrometry 2003; 17: 1600.
Creaser CS, Reynolds JC, Harvey DJ. Structural analysis of
oligosaccharides by atmospheric pressure matrix-assisted laser
desorption/ionisation quadrupole ion trap mass spectrometry.
Rapid Communications in Mass Spectrometry 2002; 16: 176.
Grasso G, Fragai M, Rizzarelli E, Spoto G, Kwon JY. In situ
AP/MALDI-MS characterization of anchored matrix
metalloproteinases. Journal of Mass Spectrometry 2006; 41: 1561.
Moyer SC, Cotter RJ, Woods AS. Fragmentation of phosphopeptides by atmospheric pressure MALDI and ESI/ion trap mass
Copyright  2007 John Wiley & Sons, Ltd.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
spectrometry. Journal of the American Society for Mass Spectrometry
2002; 13: 274.
Mayrhofer C, Krieger S, Raptakis E, Allmaier G. Comparison of
vacuum matrix-assisted laser desorption/ionization (MALDI)
and atmospheric pressure MALDI (AP-MALDI) tandem mass
spectrometry of 2-dimensional separated and trypsin-digested
glomerular proteins for database search derived identification.
Journal of Proteome Research 2006; 5: 1967.
Tan PV, Laiko VV, Doroshenko VM. Atmospheric pressure
MALDI with pulsed dynamic focusing for high-efficiency
transmission of ions into a mass spectrometer. Analytical
Chemistry 2004; 76: 2462.
Schneider BB, Lock C, Covey TR. AP and vacuum MALDI on
a QqLIT instrument. Journal of the American Society for Mass
Spectrometry 2005; 16: 176.
Olsen HB, Ludvigsen S, Kaarsholm NC. Solution structure of an
engineered insulin monomer at neutral pH. Biochemistry 1996;
35: 8836.
Ramanathan R, Gross ML. Monitoring recombinant protein
drugs: a study of insulin by H/D exchange and electrospray
ionization mass spectrometry. Analytical Chemistry 1997; 69: 5142.
Tito P, Nettleton EJ, Robinson CV. Dissecting the hydrogen
exchange properties of insulin under amyloid fibril forming
conditions: a site-specific investigation by mass spectrometry.
Journal of Molecular Biology 2000; 303: 267.
Farah MA, Bose S, Lee J-H, Jung H-C, Kim Y. Analysis of
glycated insulin by MALDI-TOF mass spectrometry. Biochimica
et Biophysica Acta 2005; 1725: 269.
Le Bihan T, Robinson MD, Stewart II, Figeys D. Definition and
characterization of a ‘‘trypsinosome’’ from specific peptide
characteristics by nano-HPLC-MS/MS and in silico analysis
of complex protein mixtures. Journal of Proteome Research 2004; 3:
1138.
Shevchenko A, Shevchenko A. Evaluation of the efficiency of
in-gel digestion of proteins by peptide isotopic labeling and
MALDI mass spectrometry. Analytical Biochemistry 2001; 296:
279.
Song ES, Juliano MA, Juliano L, Hersh LB. Substrate activation
of insulin-degrading enzyme (insulysin): a potential target for
drug development. The Journal of Biological Chemistry 2003; 278:
49789.
Amini A, Dormady SJ, Riggs L, Regnier FE. The impact
of buffers and surfactants from micellar electrokinetic
chromatography on matrix-assisted laser desorption ionization
(MALDI) mass spectrometry of peptides: effect of buffer type
and concentration on mass determination by MALDI-time-offlight mass spectrometry. Journal of Chromatography A 2000; 894:
345.
Brange J, Andersen L, Laursen ED, Meyn G, Rasmussen E.
Toward understanding insulin fibrillation. Journal of
Pharmaceutical Sciences 1997; 86: 517.
Ahmad A, Millett IS, Doniach S, Uversky VN, Fink AL. Partially
folded intermediates in insulin fibrillation. Biochemistry 2003; 42:
11404.
Manno M, Craparo EF, Podesta A, Bulone D, Carrotta R,
Martorana V, Tiana G, San Biagio PL. Kinetics of different
processes in human insulin amyloid formation. Journal of
Molecular Biology 2007; 366: 258.
Ahmad A, Uversky VN, Hong D, Fink AL. Early events in
the fibrillation of monomeric insulin. The Journal of Biological
Chemistry 2005; 280: 42669.
Nielsen L, Khurana R, Coats A, Frokjaer S, Brange J, Vyas S,
Uversky VN, Fink AL. Effect of environmental factors on the
kinetics of insulin fibril formation: elucidation of the molecular
mechanism. Biochemistry 2001; 40: 6036.
Duckworth WC, Fawcett J, Tsui BT, Bennett RG, Hamel FG.
Biological activity of a fragment of insulin. Biochemical and
Biophysical Research Communications 2004; 318: 1019.
J. Mass Spectrom. 2007; 42: 1590–1598
DOI: 10.1002/jms