Enzyme solid-state support assays: a surface plasmon resonance and mass spectrometry coupled study of immobilized insulin degrading enzyme.

Abstract

Solid-support based assays offer several advantages that are not normally available in solution. Enzymes that are anchored on gold surfaces can interact with several different molecules, opening the way to high throughput array format based assays. In this scenario, surface plasmon resonance (SPR) and mass spectrometry (MS) investigations have often been applied to analyze the interaction between immobilized enzyme and its substrate molecules in a tag-free environment. Here, we propose a SPR-MS combined experimental approach aimed at studying insulin degrading enzyme (IDE) immobilized onto gold surfaces and its ability to interact with insulin. The latter is delivered by a microfluidic system to the IDE functionalized surface and the activity of the immobilized enzyme is verified by atmospheric pressure/matrix assisted laser desorption ionization (AP/MALDI) MS analysis. The SPR experiments allow the calculation of the kinetic constants involved for the interaction between immobilized IDE and insulin molecules and evidence of IDE conformational change upon insulin binding is also obtained.

Full text

Page 1

ORIGINAL PAPER
Enzyme solid-state support assays: a surface plasmon resonance
and mass spectrometry coupled study of immobilized insulin
degrading enzyme
Giuseppe Grasso Æ Æ Ashley I. Bush Æ Æ Roberta D’Agata Æ Æ
Enrico Rizzarelli Æ Æ Giuseppe Spoto
Received: 23 July 2008/Revised: 6 November 2008/Accepted: 11 November 2008
? European Biophysical Societies’ Association 2008
Abstract
advantages that are not normally available in solution.
Enzymes that are anchored on gold surfaces can interact
with several different molecules, opening the way to high
throughput array format based assays. In this scenario,
surface plasmon resonance (SPR) and mass spectrometry
(MS) investigations have often been applied to analyze the
interaction between immobilized enzyme and its substrate
molecules in a tag-free environment. Here, we propose
a SPR-MS combined experimental approach aimed at
studying insulin degrading enzyme (IDE) immobilized
onto gold surfaces and its ability to interact with insulin.
The latter is delivered by a microfluidic system to the IDE
functionalized surface and the activity of the immobilized
enzyme is verified by atmospheric pressure/matrix assisted
laser desorption ionization (AP/MALDI) MS analysis. The
SPR experiments allow the calculation of the kinetic con-
stants involved for the interaction between immobilized
IDE and insulin molecules and evidence of IDE confor-
mational change upon insulin binding is also obtained.
Solid-support based assaysofferseveral
Keywords
Surface plasmon resonance ? Mass spectrometry ?
Insulin degrading enzyme ? Conformational change
Solid-state assay ?
Introduction
Insulin degrading enzyme (IDE) (Farris et al. 2003) is a zinc
metalloprotease, able to degrade several different substrates
besides insulin (e.g., b-amyloid) that are involved in many
pathological conditions such as Alzheimer’s disease (AD)
and Parkinson’s disease (PD) (Vepsa ¨la ¨inen et al. 2007;
Kurochkin and Goto 1994; Blomqvist et al. 2004). Very
recently, the structures of human IDE in complex with four
different substrates have been reported and some insights
into the interaction mechanism have also been given (Shen
et al. 2006). IDE has a buried catalytic site in the structure
and access to this chamber is kinetically controlled by a
closed–open conformational switch, so IDE probably con-
forms to a complex kinetic model where catalysis does not
lead automatically to product release. Instead, an additional
step is required in which the protease opens up to allow the
products to escape and a ‘‘latch’’ system has been proposed
in order to explain the experimental results obtained so far
for this enzyme (Leissring and Selkoe 2006). Recent
developments on IDE conformational changes suggest that,
because of the extensive interaction between N- and C-
terminal domains, IDE could exist in its closed conforma-
tion without requiring the binding energy contributed by its
substrate (Im et al. 2007). Moreover, it has been reported
that in solution IDE exists as a mixture of monomers,
dimers, and tetramers and the equilibrium between the
different forms is concentration-dependent, with the dimer
the more active form (Song et al. 2003). Very recently,
(unpublished data) we have also found that the various IDE
G. Grasso (&) ? R. D’Agata ? E. Rizzarelli ? G. Spoto
Dipartimento Scienze Chimiche, Universita ` di Catania,
v.le A. Doria 6, 95125 Catania, Italy
e-mail: grassog@unict.it
A. I. Bush
Department of Psychiatry, Massachusetts General Hospital,
Charlestone, MA, USA
A. I. Bush
Department of Pathology,
Mental Health Research Institute of Victoria,
The University of Melbourne, Parkville, Australia
G. Spoto
Istituto di Biostrutture e Bioimmagini, CNR,
v.le A. Doria 6, 95125 Catania, Italy
123
Eur Biophys J
DOI 10.1007/s00249-008-0384-y

Page 2

oligomeric forms produce different insulin fragmentation
patterns. The latter can also be altered by the presence of
other molecules such as ubiquitin (Grasso et al. 2008).
In thisscenario, although the exact molecular mechanism
of the interaction between IDE and insulin is still unclear, it
is likely that IDE has an open conformation in the active
state and a closed one when is inactive or is bound to its
substrates. Similar behaviors have been observed in the past
for other enzymes which undergo a hinge twist motion
betweentwoseparatedomains(Sharffetal.1992).Although
optical and spectroscopic techniques, like CD, fluorescence,
NMR, and X-ray scattering, are routinelyusedtoinvestigate
the conformational state of proteins in solution or crystal
(Drobny et al. 2003; Andrade et al. 2004), a valid contri-
bution toward a better understanding of IDE-substrates
interaction mechanism is also expected from alternative
experimental approaches such as solid-state assays. The
development and optimization of an immunocapture-based
assay for the specific measurement of IDE activity in brain
tissuehomogenateshasalreadybeendescribed(Minersetal.
2008). However, such method requires a fluorescent tag to
be present in the substrate and the investigation of the pro-
teolytic action of the enzyme is limited to the only cleavage
site where the fluorescent tag is attached.
Surface plasmon resonance (SPR) is able to detect var-
iation in physical parameters (for instance, dielectric
constant) caused by volume changes of surface bound
proteins and it is usually applied to study biomolecular
interactions in real time to obtain kinetics parameters
(Homola 2006). In 1998, the first investigation of pH-
induced structural transitions of immobilized proteins by
SPR was reported (Sota et al. 1998). This was the first
description of a correlation between resonance angle shifts
and conformational changes of immobilized proteins,
opening the way to several indirect observations of protein
conformational changes using SPR (Kim et al. 2005; Kang
et al. 2006; Geitmann and Danielson 2004).
Our group has already shown the advantages offered by
coupling SPR and mass spectrometry (MS) for studying a
certain class of enzymes (Grasso et al. 2005, 2007a) in a
solid-state format. However, in the latter cases no insights
into the molecular mechanism of the interaction were given
and the analysis was limited to the calculation of enzyme
activity or the identification of the immobilized biomole-
cules (Grasso et al. 2006).
In order to overcome the above mentioned limitations, in
this paper we propose an SPR-MS combined experimental
approachthatgivesaninsightintothemolecularmechanism
of the interaction between an enzyme and one of its sub-
strate. The proposed method is able to detect if the
interacting immobilized enzyme undergoes a conforma-
tional change upon substrate binding. For this reason, the
interaction between IDE and insulin molecules is here
scrutinized. We immobilized IDE onto a gold substrate by
theaminocouplingapproachandusedatmosphericpressure/
matrix assisted laser desorption ionization (AP/MALDI)-
MS to monitor the activity of the anchored biomolecules.
The latter are arrayed in a spatially resolved manner by
coupling the SPR imaging (SPRI) technique with a home-
made microfluidic system, allowing the calculation of the
kinetic constants involved for the IDE-insulin interaction.
The proposed SPR experimental approach produces evi-
dences of the enzyme conformational changes upon insulin
binding, in accordance with the view that active substrate-
free IDE is in its open conformation (Im et al. 2007). In this
way, the study of the monomeric form of the wild type
enzyme interacting with one of its natural substrates is fea-
sible and insight into kinetic details of the interaction
between immobilized IDE and insulin is provided.
Materials and methods
Reagents
IDE, his-tag, rat, and recombinant from Spodoptera
frugiperda was purchased from Calbiochem. Insulin
from bovine pancreas, phosphate buffer solution (PBS),
a-cyano-4-hydroxycinnamic acid (CHCA), trifluoro acetic
acid (TFA), acetonitrile (C2H3N), ethanol solution, etha-
nolamine-HCl 1M, guanidine-HCl 8M, sucrose and
dithiobis (N) succinimidyl propionate (Lomant’s reagent)
were all purchased from Sigma-Aldrich, while ZipTipSCX
pipette tips were from Millipore, and dithiol tethers SPT-
0013 and SPT-0014C were purchased from Sensopath.
Gold substrates (GWC Technologies, USA) were obtained
by thermally evaporating a gold layer (450 A˚) onto SF-10
glass slides (Schott, USA). Chromium (50 A˚) was used as
the adhesion layer.
Surface plasmon resonance
Two different immobilization procedures were scrutinized
for IDE and positive results were obtained in both cases.
Specifically, similar SPRI signals (see Fig. 1) were indeed
registered after a 40 min injection at 5 ll min-1of a
36 nM IDE solution into a microchannel in contact with a
gold surface previously functionalized with:
1.
2.
Lomant’s reagent (Grasso et al. 2005);
Dithiol tethers (SPT-0013:SPT-0014C = 10:1 mixed
ethanol solution; Lahiri et al. 1999).
We found that the pH of the PBS buffer used for sample
dilution is crucial for a positive result of the activity mea-
surements and a pH of 7.4 was chosen for all experiments.
Ethanolamine-HCl 1M was used for deactivation of the
Eur Biophys J
123

Page 3

unreacted NHS groups, while 5 min injection at 5 ll min-1
of guanidine-HCl 8M was used for denaturation of IDE. In
order to estimate the number of IDE molecules anchored on
1 cm2of gold surface, we followed the method previously
described (Grasso et al. 2005). Briefly, the adlayer thickness
(d)isobtainedfromthemeasuredSPRresponse(R),whichis
the shift in wavenumber of the SPR minimum in reflected
light intensity associated with changes in the index of
refractionofthe mediumincontactwith themetalsurfaceof
the SPR device, Dg. Once a calibration curve that correlates
RwiththechangeinrefractiveindexDgisobtained(sucrose
solutions were used for this purpose), it is possible to esti-
mate the adlayer thickness (d) from the following equation:
d ¼ ðld=2Þ ? ðR=RmaxÞ
where
ld
electromagnetic field into the specific medium, that is
usually 25–50% of the wavelength of the light. Rmaxis the
SPR response that is the maximum response that would be
measured for an infinitely thick adlayer. In the special case
where d is very small compared to ld, Eq. 1 reduces to:
ð1Þ
is thedecay lengthof theevanescent
d ¼ ðld=2Þ ? fR=½mðga? gsÞ?g
where gais the refraction index of the adsorbate (the pure
IDE protein in our case), while gsis the refraction index of
the buffer. The SPR shift due to the anchoring of IDE was
ð2Þ
determined (Fig. 1) and a value for the adlayer thickness
d = 0.9 nm was obtained, using Eq. 2. It is then
straightforward (Darnell et al. 1990; Leslie and Lilley
1985) to convert the adlayer thickness into the surface
concentration, h, in molecules per cm2:
hðmolecules=cm2Þ ¼ dðcmÞ ? Nðmolecules=cm3Þ
where N is the bulk number density of the adsorbate and
can be estimated from the bulk density of the adsorbate, q,
in units of g cm-3, just by dividing by the molecular
weight and multiplying by the Avogadro’s number. In our
case, q = 1.30 g cm-3was obtained from literature (Jung
et al. 1998) and h % 6 9 1011(molecules cm-2) was
calculated, using Eq. 3.
The SPRI apparatus (GWC Technologies, USA) was the
same as reported in some of our previous works (Grasso
et al. 2005). SPR images were analyzed, using the V??
software (version 4.0, Digital Optics Limited, New Zea-
land) and the software package Image J 1.32j (National
Institutes of Health, USA). SPRI provides data as pixel
intensity units (0–255 scale). Data were converted in per-
centage of reflectivity (%R), using the formula:
ð3Þ
%R ¼ 100 ? ð0:85Ip=IsÞ
where Ip and Is refer to the reflected light intensity detected
using p- and s-polarized light, respectively. The experi-
ments were carried out by sequentially acquiring 15 frames
averaged SPR images with 5 s time delay between them.
Kinetics data were obtained by plotting the difference in
percent reflectivity (%R) from selected regions of interest
(ROIs) of the SPR images as a function of time. All the
SPRI experiments were carried out at room temperature.
The rate constants reported in Table 1 were calculated
by fitting adsorption/desorption kinetics data through
numerical integration analysis (Myszka and Morton 1998).
After insulin interaction immobilized enzyme could be
recycled for other SPRI analyses by flowing buffered
solution onto the surface for about 20 min in order to allow
complete insulin dissociation and restoring of the SPRI
baseline. As long as the functionalized surface was kept in
contact with the buffered solution, immobilized enzyme
Fig. 1 SPRI response registered for the immobilization of IDE
obtained by injecting a 36 nM IDE solution in a microchannel in
contact with the pre-functionalized gold surface. The reported curves
refer to IDE immobilization on a gold surface pretreated with
Lomant’s reagent (black line) or dithiol tethers (gray line) and to
reference (dotted line) have been obtained by mediating the SPRI
response of five different channels in two separate experiments. The
double arrow indicates the amount of immobilized enzyme that
differs slightly for the two experimental procedures (about 10%).
However, the observed difference is within the reproducibility range
(12% is the largest calculated standard deviation of the SPRI signal
for the immobilization) and no appreciable difference was detected
for the two cases in the following SPRI interaction experiments
Table 1 Kinetics parameters obtained by fitting the sensorgrams
recorded for the interaction between immobilized IDE and insulin
according to the conformational model described in the text (see
Fig. 5)
Rate constants
ka1
kd1
kr
k-r
1.0 (±0.3) 9 103M-1s-1
1.3 (±0.2) 9 10-2s-1
2.4 (±0.2) 9 10-3s-1
1.5 (±0.7) 9 10-3s-1
The calculated standard deviations are indicated. The low reported
value of ka1is discussed in the text
Eur Biophys J
123

Page 4

degradation was observed as an altered SPRI baseline only
after several hours (data not shown).
PDMS microfluidic channels fabrication
Microfluidic channels were fabricated in poly(dimethylsi-
loxane) (PDMS) polymer as described elsewhere (Grasso
et al. 2007a). Briefly, PDMS microchannels having a vol-
ume of about 0.7 ll were created by replication from
masters in polyvinyl chloride (PVC), with a pattern of
parallel microchannels (80 lm depth, 1.4 cm length,
400 lm width), featuring circular reservoirs (diameter
400 lm) at both ends of each channel. A six microchannels
microfluidic device was used in this case for following the
interaction between IDE and insulin at different concen-
trations (see Fig. 2). PEEK tubes (Upchurch Scientific)
were inserted in such reservoirs in order to connect the
PDMS microfluidic cell to a Masterflex L/S (Cole-Parmer,
USA) peristaltic pump, operating at 100 ll min-1. Repli-
cas were formed from a 1:10 mixture of PDMS curing
agent and prepolymer (Sylgard 184, Dow Corning, USA).
The mixture was degassed under vacuum and then poured
onto the master in order to create a layer with a thickness of
about 3–4 mm. The PDMS was then cured for at least 2 h
at 60?C before it was removed from the masters.
The evaluation of the kinetic constants for biomolecular
interactions by SPR experiments can be affected by dif-
fusion problems (D’Agata et al. 2008). In order to estimate
the contribution of diffusion to the kinetic parameters
obtained, kmis normally used to describe the diffusion of
the protein to the surface and it can be related to the dif-
fusion coefficient of the protein. The dimensions of the
above described microchannels ensured that, at the flow
rate used in our SPRI experiments for the interaction
between IDE and insulin (100 ll min-1), the value of km
describing the diffusion of insulin to the gold surface was
1.4 9 108M-1s-1. The latter is well above the rate con-
stants values reported in Table 1 and therefore the results
obtained are not affected by diffusion problems.
Mass spectrometry
AP/MALDI-MS experiments were carried out, using a
Finnigan LCQ Deca XP PLUS (Thermo Electron Corpo-
ration, USA) ion trap spectrometer which was fitted with a
MassTech Inc. (USA) AP/MALDI PDF-source. 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 (wave-
length 337 nm, pulse width 4 ns, pulse energy 300 lJ,
repetition rate up to 10 Hz) and a pulsed dynamic focusing
(PDF) module that imposes a delay of 25 ls between the
laser pulse and the application of the high voltage to the
AP/MALDI target plate. Laser power was attenuated to
about 55%. The target plate voltage was 1.8 kV. The ion
trap inlet capillary temperature was 220?C. Capillary and
tube lens offset voltages of 30 and 15 V, respectively, were
applied. Other mass spectrometer parameters were as fol-
lows: multipole 1 offset at -3.75 V, multipole 2 offset at
-9.50 V, multipole RF amplitude 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 mi-
croscans. Although there is the risk of loosing resolution,
the latter experimental conditions were chosen as sensi-
tivity was the main goal in all MS experiments carried out
on the immobilized enzyme.
Results and discussion
Immobilization ofbiomoleculesonsolidsupportsanduseof
microfluidic systems have shown to offer several advanta-
ges not available in the standard solution assays as
mentioned above. In this work, MS and SPRI experiments
on immobilized IDE were performed. Particularly, the
activity of the immobilized IDE was verified by AP/
MALDI-MS on the functionalized gold surface by sampling
Fig. 2 SPR image of the gold surface arrayed with six microchan-
nels. While one of the latter was used for referencing, five insulin
solutions at different concentrations could be monitored in real
time simultaneously. Dimension of microchannels were 0.2 lm
(width) 9 80 lm (depth) 9 1.6 cm (length)
Eur Biophys J
123

Page 5

the supernatant solution in contact with the functionalized
surface as previously described for other enzymes (Grasso
et al. 2005), and representative spectra are reported in
Fig. 3. Fig. 3a shows the spectrum of the doubly charged
insulin molecular peak which is detected at m/z 2866.5
together with the associated sodiated peaks (presence of
sodium salt is necessary in order to maintain enzyme
activity). Fig. 3bshows the mass spectrumregistered for the
supernatant solution that had been in contact with the IDE
functionalized surface for about 20 min. In this case, the
molecular peak at m/z 2866.5 disappeared, while new peaks
at m/z 2312.4 and 2475.3 assigned to the insulin fragments
reported in Table 2 appeared together with associated so-
diated peaks at m/z 2334.4, 2355.5, and 2497.4. From these
data, we concluded that the immobilized IDE is able to
cleave insulin molecules, producing some of the expected
insulin fragments (Grasso et al. 2007b). It has already been
found by our group that it is possible to have information on
the IDE oligomeric forms distribution from the produced
insulin fragmentation patterns (unpublished data). Particu-
larly, the peaks observed in the case of immobilized IDE are
mainly produced by the monomeric form of the enzyme.
Therefore it is possible to conclude that, if the immobili-
zation ofIDE iscarriedoutaccordingtothe above described
experimental protocol, anchored IDE exists mainly as a
monomer and it is able to cleave insulin molecules.
In Fig. 4 the change in percent reflectivity over time
obtained for the interaction between immobilized IDE and a
9 lM insulin solution (black line) is shown together with
the SPRI curve obtained by putting the same solution in
contact with denatured IDE (gray line). IDE denaturation
wasachieved byinjecting
5 ll min-1for 5 min (Di Venere et al. 2000; Yamaguchi
et al. 2003). Although some non-specific interactions
between the insulin molecules and the functionalized sur-
faces were observed, a large difference was recorded
between active and denatured IDE. Particularly, no shift in
the SPRI response was recorded when the insulin solution
(experiments were carried out up to 80 lM insulin con-
centration) flowed into the microchannel in contact with
denatured IDE functionalized surface, indicating that insu-
lin molecules did not interact with the immobilized
denatured enzyme. Furthermore, non-specific insulin inter-
actions were not observed in this case, probably because
unfolded denatured IDE molecules completely blocked the
gold surface that is sampled by SPRI (Jung et al. 1998). This
last observation reinforces the conclusion that the SPRI data
recorded in the case of active IDE-insulin is due to a
physiological interaction between the two biomolecules.
guanidine-HCl8 M at
Fig. 3 AP/MALDI-MS data
recorded for a supernatant
insulin solution (18 lM)
sampled from a clean gold
surface (a) and from IDE
functionalized surface (b). The
insulin doubly charged
molecular peak at m/z 2866.5 in
(a) is substituted with peaks
attributed to insulin fragments
in (b)
Table 2 Insulin fragments detected from the supernatant solution in
contact with the IDE functionalized surface
Fragments
combinations
Calculated
peaks (m/z)
Experimental
peaks (m/z)
A(1–13) ? B(1–9) 2313.02312.4
A(1–14) ? B(1–9)2476.02475.3
For explanation/discussion see text
Eur Biophys J
123

Page 6

Therefore, immobilized IDE interacts with insulin
molecules only in its active conformation. In order to
elucidate the mechanism of such an interaction a common
procedure is to propose a plausible model and to fit the
experimental kinetic data accordingly (Honjo et al. 2002;
Yowler and Schengrund 2004). For this reason, SPRI
kinetic curves obtained with four insulin solutions at
different concentrations were fitted, assuming that IDE
undergoes a conformational change upon substrate binding
(see Fig. 5). Kinetic parameters obtained from the fitting of
the experimental data are shown in Table 1, while the
reactions describing the model used in order to fit the
kinetic data are:
L þ A? !
ka1
?
kd1
LA? !
?
kr
k?r
ðLAÞr
where L represents the immobilized IDE, A is the free
insulin molecules in solution and LA is the IDE-insulin
complex before and after the conformational change (LA)r.
The poor results obtained by fitting data assuming pseudo
first-order kinetics are evident from Fig. 5 (dotted black
line). Analogously, other interaction models (surface het-
erogeneity, etc.) were used but the best fit of the kinetic
data was obtained by applying a conformational change
interaction model. The calculated value of ka1reported in
Table 1 is low if compared to the values normally reported
in the literature for various biomolecular interactions (Ji
et al. 2003). In order to interpret this result correctly, it is
important to consider that the values reported in Table 1
are meaningful only for the interaction between immobi-
lized IDE and insulin. In fact, although immobilized
enzymes have often similar kinetic parameters as in solu-
tion (Grasso et al. 2007a), in the case of IDE the interaction
mechanism is more complicated as it involves different
oligomeric forms of the enzyme (Shen et al. 2006; Im et al.
2007; Song et al. 2003). AP/MALDI-MS results show that
IDE is mainly immobilized as a monomer and therefore it
is expected that the kinetic parameters for the interaction
with insulin are somehow different from the ones referring
to the solution state where all the possible oligomeric forms
of the enzyme are present (the ka1value is expected to be
higher in solution). Therefore, our results serve to charac-
terize the interaction between the monomeric immobilized
form of the enzyme and insulin, opening the way to high
throughput solid-state assays. Moreover, in our 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 (Manno et al. 2007). However, the interaction
between IDE and insulin does not seem to be affected by
the substrate oligomeric distribution, as revealed by the
same insulin fragmentation patterns detected by MS
(Grasso et al. 2007b).
A further SPRI experiment was performed in order to
support the hypothesis that the conformational change
model is correct to describe the IDE-insulin interaction.
The experiment consisted of varying association times and
analyte concentrations in order to have always the same
amount of analyte interacting with the immobilized mole-
cules (Myszka et al. 1999; Aguilar and Small 2005). In
fact, the amount of insulin bound to the same IDE surface
Fig. 4 SPRI response registered when a 9 lM insulin solution is
flowed for 9 min into two differently modified microchannels:
immobilized active (black line) and denatured (gray line) IDE. The
latter functionalized surface was achieved by injecting guanidine-HCl
8 M at 5 ll min-1for 5 min onto the active immobilized IDE
molecules
Fig. 5 SPRI sensorgrams (solid lines) obtained by flowing insulin
solutions for 550 s at the following different concentrations onto
active immobilized IDE: 4 lM (dark gray), 6 lM (light gray), 9 lM
(gray), and 18 lM (black). Fitted lines according to the conforma-
tional change model are shown in dashed lines, while only in the case
of the 18 lM insulin solution the line fitted according to the simple
biomolecular model is also shown (dotted line)
Eur Biophys J
123

Page 7

is dependent on both association time and insulin concen-
tration. The %R variation detected by SPRI is proportional
to the amount of bound insulin, thus different combinations
of association time and insulin concentration were used to
reach the same %R value (0.53 ± 0.02). Results from the
experiments are reported in Fig. 6 and they have been
interpreted as follows. In the case of high concentration of
analyte molecules and short contact time (25 lM and 44 s),
it is insufficient for immobilized IDE molecules to change
their conformation. Therefore, the insulin molecules exhi-
bit mainly non-specific interactions with the IDE surface
and give rise to a rapid decrease of the SPRI response in the
dissociation phase (see Fig. 6, dark gray line). In contrast,
when a low concentration solution is flowed into the
microchannel for a prolonged period (6.3 lM and 325 s),
there is sufficient time for the immobilized IDE molecules
to change their conformation and the dissociation phase is
characterized by a slower dissociation rate (see Fig. 6, light
gray line). In fact, in the latter case, IDE has to change its
conformation back to the open state in order to release
bound insulin molecules, a process requiring more time
than a simple surface desorption.
The above described SPRI approach is applied to verify
the conformational change of immobilized enzyme mole-
cules occurring upon substrate binding, giving an insight
into the IDE-insulin interaction mechanism. In fact, it is
important to highlight that the MS results showed that the
enzyme is mainly immobilized on the surface as a mono-
mer and therefore the conformational change observed by
applying our SPRI approach refers only to this specific
oligomeric form of IDE, conferring an advantage to the use
of the solid-supported coupled SPRI-MS approach over
other solution-based investigation tools. Particularly, we
confirmed a model where, in the absence of substrate, the
inactive resting state of monomeric IDE is in a closed
conformation, since substrates cannot gain access to the
catalytic chamber when the protease is in this state. Only
when the enzyme switches to its open conformation the
substrate molecules can properly interact with IDE (see
Table 1 for the rate constants values). Thus, the closed
conformation is critical for regulating the catalytic cycle of
monomeric IDE while a conformational switch to the open
state needs to occur if the enzyme is to carry out its cata-
lytic activity.
Conclusions
A solid-support based assay that allows a multiplexed
approach to study IDE-insulin interaction was proposed.
We immobilized IDE molecules on gold surfaces and
monitored their activity by AP/MALDI-MS, demonstrating
the ability of the anchored monomeric form of the enzyme
to cleave insulin molecules. SPRI experiments subse-
quently were carried out on the immobilized enzyme
molecules, producing evidence of a conformational change
of IDE upon insulin binding. Kinetic parameters were
calculated and SPRI experiments based on different ratios
of insulin concentrations and contact times have also been
carried out to give an insight into the IDE-insulin interac-
tion. Particularly, the results confirmed a model where the
closed conformation is the inactive resting state of mono-
meric IDE while the open conformation is its active state
that switches to a closed state upon substrate binding.
According to this model IDE should shrink its hydro-
dynamic radius upon insulin binding.
Acknowledgment
RBIN04L28Y) and ‘‘EURAMY: Systemic Amyloidoses in Europe’’,
037525 (LSHM-CT-2006-037525) for partial financial support.
We thankMIUR(FIRB RBNE03PX83,
References
Aguilar M-I, Small DH (2005) Surface plasmon resonance for the
analysis of b-amyloid interactions and fibril formation in
Alzheimer’s disease research. Neurotox Res 7:17–27
AndradeSM,CarvalhoTI,ViseuMI,CostaSM(2004)Conformational
changes of beta-lactoglobulin in sodium bis(2-ethylhexyl) sulfo-
succinate reverse micelles. A fluorescence and CD study. Eur J
Biochem 271:734–744. doi:10.1111/j.1432-1033.2004.03977.x
Blomqvist ME-L, Silburn PA, Buchanan DD, Andreasen N, Blennow
K, Pedersen NL, Brookes AJ, Mellick GD, Prince JA (2004)
Sequence variation in the proximity of IDE may impact age at
onset of both Parkinson disease and Alzheimer disease. Neur-
ogenetics 5:115–119. doi:10.1007/s10048-004-0173-4
D’Agata R, Grasso G, Spoto G (2008) Real-time binding kinetics
monitored with surface plasmon resonance imaging in a
Fig. 6 Varying contact time and insulin concentration: 6.3 lM and
325 s (light gray line); 15.6 lM and 182 s (black line); and 25 lM
and 44 s (dark gray line), respectively. The concentration and contact
time of insulin solutions flowing into the microchannels were varied
with the intent of keeping the overall amount of insulin bound to the
immobilized IDE constant at %R of 0.53 ± 0.02 at the beginning of
the dissociation phase
Eur Biophys J
123

Page 8

diffusion-free environment. Open Spectrosc J 2:1–9. doi:
10.2174/1874383800802010001
Darnell JE, Lodish H, Baltimore D (1990) Molecular cell biology.
Scientific American Books, New York
Di Venere A, Rossi A, De Matteis F, Rosato N, Finazzi Agro ` A, Mei
G (2000) Opposite effects of Ca2? and GTP binding on tissue
transglutaminase tertiary structure. J Biol Chem 275:3915–3921
Drobny GP, Long JR, Karlsson T, Shaw W, Popham J, Oyler N,
Bower P, Stringer JD, Mehta GM, Stayton PS (2003) Structural
studies of biomaterials using double-quantum solid-state NMR
spectroscopy. Annu Rev Phys Chem 54:531–571. doi:10.1146/
annurev.physchem.54.011002.103903
Farris W, Mansourian S, Chang Y, Lindsley L, Eckman EA, Frosch
MP, Eckman CB, Tanzi RE, Selkoe DJ, Gue ´nette S (2003)
Insulin-degrading enzyme regulates the levels of insulin, amy-
loid b-protein, and the b-amyloid precursor protein intracellular
domain in vivo. Proc Natl Acad Sci USA 100:4162–4167. doi:
10.1073/pnas.0230450100
Geitmann M, Danielson UH (2004) Studies of substrate-induced
conformational changes in human cytomegalovirus protease
using optical biosensor technology. Anal Biochem 332:203–214.
doi:10.1016/j.ab.2004.06.008
Grasso G, D’Agata R, Rizzarelli E, Spoto G, D’Andrea L, Pedone C,
Picardi A, Romanelli A, Fragai M, Yeo KJ (2005) Activity of
anchored human matrix metalloproteinase-1 catalytic domain on
Au (111) surfaces monitored by ESI-MS. J Mass Spectrom
40:1565–1571. doi:10.1002/jms.929
Grasso G, Fragai M, Rizzarelli E, Spoto G, Yeo KJ (2006) In situ AP-
MALDI characterization of anchored MMPs. J Mass Spectrom
41:1561–1569. doi:10.1002/jms.1126
Grasso G, Fragai M, Rizzarelli E, Spoto G, Yeo KJ (2007a) A new
methodology for monitoring the activity of cdMMP-12 anchored
and freeze-dried on Au (111). J Am Soc Mass Spectrom 18:961–
969. doi:10.1016/j.jasms.2007.02.003
Grasso G, Rizzarelli E, Spoto G (2007b) AP/MALDI-MS complete
characterization of the proteolytic fragments produced by the
interaction of insulin degrading enzyme with bovine insulin.
J Mass Spectrom 42:1590–1598. doi:10.1002/jms.1348
Grasso G, Rizzarelli E, Spoto G (2008) How the binding and
degrading capabilities of insulin degrading enzyme are affected
by ubiquitin. Biochim Biophys Acta 1784:1122–1126
Homola J (ed) (2006) Surface plasmon resonance based sensors.
Springer, Berlin
Honjo E, Watanabe K, Tsukamoto T (2002) Real-time kinetic
analyses of the interaction of ricin toxin A-chain with ribosomes
prove a conformational change involved in complex formation.
J Biochem 131:267–275
Im H, Manolopoulou M, Malito E, Shen Y, Zhao J, Neant-Fery M,
Sun C-Y, Meredith SC, Sisodia SS, Leissring M, Tang W-J
(2007) Structure of substrate-free human insulin-degrading
enzyme (IDE) and biophysical analysis of ATP-induced confor-
mational switch of IDE. J Biol Chem 282:25453–25463. doi:
10.1074/jbc.M701590200
Ji ZL, Chen X, Zhen CJ, Yao LX, Han LY, Yeo WK, Chung PC, Puy
HS, Tay YT, Muhammad A, Chen YZ (2003) KDBI: kinetic data
of bio-molecular interactions database. Nucleic Acids Res
31:255–257. doi:10.1093/nar/gkg067
Jung LS, Campbell CT, Chinowsky TM, Mar MN, Yee SS (1998)
Quantitative interpretation of the response of surface plasmon
resonance sensors to adsorbed films. Langmuir 14:5636–5648.
doi:10.1021/la971228b
Kang T, Hong S, Choi I, Sung JJ, Kim Y, Hahn J-S, Yi J (2006)
Reversible pH-driven conformational switching of tethered
superoxide dismutase with gold nanoparticle enhanced surface
plasmon resonance spectroscopy. J Am Chem Soc 128:12870–
12878. doi:10.1021/ja0632198
Kim M, Jung SO, Park K, Jeong E-J, Joung H-A, Kim T-H, Seol D-W,
Chung BH (2005) Detection of bax protein conformational change
using a surface plasmon resonance imaging-based antibody chip.
Biochem Biophys Res Commun 338:1834–1838. doi:10.1016/
j.bbrc.2005.10.155
Kurochkin IV, Goto S (1994) Alzheimer’s b-amyloid peptide specif-
icallyinteractswith andisdegradedbyinsulindegrading enzyme.
FEBS Lett 345:33–37. doi:10.1016/0014-5793(94)00387-4
Lahiri J, Isaacs L, Tien J, Whitesides GM (1999) A strategy for the
generation of surfaces presenting ligands for studies of binding
based on an active ester as a common reactive intermediate: a
surface plasmon resonance study. Anal Chem 71:777–790. doi:
10.1021/ac980959t
Leissring MA, Selkoe DJ (2006) Structural biology: enzyme target to
latch on to. Nature 443:761–762. doi:10.1038/nature05210
Leslie TE, Lilley TH (1985) Aqueous solutions containing amino
acids and peptides. Part 20. Volumetric behavior of some
terminally substituted amino acids and peptides at 298.15 K.
Biopolymers 24:695–710. doi:10.1002/bip.360240409
Manno M, Craparo EF, Podesta A, Bulone D, Carrotta R, Martorana
V, Tiana G, San Biagio PL (2007) Kinetics of different processes
in human insulin amyloid formation. J Mol Biol 366:258–274.
doi:10.1016/j.jmb.2006.11.008
Miners JS, Kehoe PG, Love S (2008) Immunocapture-based fluoro-
metric assay for the measurement of insulin-degrading enzyme
activity in brain tissue homogenates. J Neurosci Methods
169:177–181. doi:10.1016/j.jneumeth.2007.12.003
Myszka DG, Morton TA (1998) CLAMP: a biosensor kinetic data
analysis program. Trends Biochem Sci 23:149–150. doi:10.1016/
S0968-0004(98)01183-9
Myszka DG, Wood SJ, Biere AL (1999) Analysis of fibril elongation
using surface plasmon resonance biosensors. Methods Enzymol
309:386–402. doi:10.1016/S0076-6879(99)09027-8
Sharff AJ, Rodseth LE, Spurlino JC, Quiocho FA (1992) Crystallo-
graphic evidence of a large ligand-induced hinge-twist motion
between the two domains of the maltodextrin binding protein
involved in active transport and chemotaxis. Biochemistry
31:10657–10663. doi:10.1021/bi00159a003
Shen Y, Joachimiak A, Rosner MR, Tang W-J (2006) Structures
of human insulin-degrading enzyme reveal a new substrate
recognition mechanism. Nature 443:870–874. doi:10.1038/
nature05143
Song ES, Juliano MA, Juliano L, Hersh LB (2003) Substrate
activation of insulin degrading enzyme (insulysin), a potential
target for drug development. J Biol Chem 278:49789–49794.
doi:10.1074/jbc.M308983200
Sota H, Hasegawa Y, Iwakura M (1998) Detection of conformational
changes in an immobilized protein using surface plasmon
resonance. Anal Chem 70:2019–2024. doi:10.1021/ac9713666
Vepsa ¨la ¨inen S, Parkinson M, Helisalmi S, Mannermaa A, Soininen H,
Tanzi RE, Bertram L, Hiltunen M (2007) Insulin-degrading
enzyme is genetically associated with Alzheimer’s disease in the
Finnish population. J Med Genet 44:606–608. doi:10.1136/jmg.
2006.048470
Yamaguchi S, Mannen T, Zako T, Kamiya N, Nagamune T (2003)
Measuring adsorption of a hydrophobic probe with a surface
plasmon resonance sensor to monitor conformational changes in
immobilized proteins. Biotechnol Prog 19:1348–1354. doi:
10.1021/bp034015n
Yowler BC, Schengrund C-L (2004) Botulinum neurotoxin A changes
conformation upon binding to ganglioside GT1b. Biochemistry
43:9725–9731. doi:10.1021/bi0494673
Eur Biophys J
123