Published: February 03, 2011
r 2011 American Chemical Society 2005 dx.doi.org/10.1021/ac102762q |Anal. Chem. 2011, 83, 2005–2011
ARTICLE
pubs.acs.org/ac
Kinetics and Thermodynamics of Biotinylated Oligonucleotide Probe
Binding to Particle-Immobilized Avidin and Implications for
Multiplexing Applications
Graham R. Broder,† Rohan T. Ranasinghe,† Cameron Neylon,§ Hywel Morgan,‡ and Peter L. Roach*,†
†School of Chemistry and ‡School of Electronics and Computer Science, University of Southampton, Southampton SO17 1BJ, U.K.
§STFC Rutherford Appleton Laboratory, Harwell Science and Innovation Campus, Didcot OX11 0QX, U.K.
b
S Supporting Information
ABSTRACT: In this work, the kinetics and dissociation constant
for the binding of a biotin-modified oligonucleotide to micro-
particle-immobilized avidin were measured. Avidin has been
immobilized by both covalent coupling and bioaffinity capture
to a surface prefunctionalized with biotin. The measured rate
and equilibrium dissociation constants of avidin immobilized
by these different methods have been compared with those for
nonimmobilized avidin.We found that immobilization resulted
in both a decrease in the rate of binding and an increase in the
rate of dissociation leading to immobilized complexes having equilibriumdissociation constants of 7( 3
10-12M, higher than the value
measured for the complex between biotin-modified oligonucleotide and nonimmobilized avidin and approximately 4 orders of magnitude
larger than values for the wild-type avidin-biotin complex. Immobilized complex half-lives were found to be reduced to 5 days, which
resulted in biotin ligands migrating between protein attached to different particles. Different immobilization methods showed little
variation in complex stability but differed in total binding and nonspecific biotin-modified oligonucleotide binding. These findings are
critical for the design of multiplexed assays where probe molecules are immobilized to biosensors via the avidin-biotin interaction.
The high affinity of the tetrameric proteins avidin and strep-tavidin for the vitamin biotin, with equilibrium dissociation
constants (Kd) of 0.6
10-15 and 4
10-14 M, respectively,
have led to their use as capture elements in a multitude of bio-
technological applications.1,2 Proteins and peptides,3 nucleic
acids4-6 and aptamers7 have been modified with biotin and
immobilized to a (strept)avidin surface for use in biological
sensing and immunoassays.8-11
Typically, avidin or streptavidin are immobilized on surfaces via
electrostatic interactions,12,13 covalent coupling14-17 or bioaffinity
capture to a surface prefunctionalized with biotin.16,17 The most
common covalent coupling strategies involve amide bond forma-
tion using protein lysine residues, though streptavidin mutants
bearing a single cysteine residue have also been developed for
attachment to maleimide-functionalized solid surfaces.18 The wide
use of this technology for purification and immobilization of nucleic
acids has led to many commercial products including streptavidin-
coated magnetic microparticles (Dynabeads, Invitrogen),
(strept)avidin and biotin coated polystyrene particles
(Spherotech), phosphoramidite derivatives of biotin and desthio-
biotin for labeling of oligonucleotides in automated synthesis (Glen
Research),19, and forms of the proteins selected for low nonspecific
binding to nucleic acids (avidin DN, Vector Laboratories and
NeutrAvidin, Thermo Fisher Scientific).
The widespread use of (strept)avidin for the attachment of
biotinylated probes is based upon the great stability of the
unmodified complex, with a half-life (t1/2) of 200 days for the
avidin-biotin complex.1 However, studies have shown that
modification of both the protein and the ligand can reduce this
stability. Studies of the dissociation of succinoylavidin with
biotinylated insulin have highlighted the importance of a spacer
between the biotin and conjugated protein, with t1/2 of 2.6 h
exhibited without a spacer and 76 days with a six carbon spacer.20
An increased rate of dissociation was also observed when
polyethylene glycol (PEG) (3400/5000) spacers were used in
biotin-modified enzymes, reducing the avidin-biotin t1/2 to
3.5 days.21 Of more importance when the system is used in
sensors are the kinetics when either the ligand or the protein are
immobilized on a surface. Reported rate constants for the binding
of fluorescently modified avidin to surface-immobilized biotin
vary, though differences in the immobilization strategy used may
account for this variation. Zhao and Reichert measured the
binding of avidin to biotin-doped lipid films on a fiber-optic
sensor observing an association rate constant, kon of 1.2
105
M-1 s-1 and a dissociation rate constant, koff of 3.3
10-7 s-1
(t1/2 ≈ 24 days).22 Wayment and Harris studied the binding of
avidin to biotin immobilized via an amide linkage to amino-
modified glass, where faster kinetics were observed with kon =
Received: October 20, 2010
Accepted: January 4, 2011

2006 dx.doi.org/10.1021/ac102762q |Anal. Chem. 2011, 83, 2005–2011
Analytical Chemistry ARTICLE
2.1
108M-1 s-1 and koff = 3.8
10-4 s-1 (t1/2≈ 0.5 h).23 Both
studies resulted in a similar calculated equilibrium dissociation
constant, Kd ≈ 2.3
10-12 M. A direct measurement of the Kd
has been made by Graves and co-workers, using commercial
(Spherotech) polystyrene microparticles covalently functiona-
lized with biotin to a much higher density than the previously
noted studies, yielding Kd = 4
10-9 M.24 To the best of our
knowledge, a detailed kinetic analysis of biotin binding to surface
immobilized avidin has not been published, however three
previous studies disclose informative kinetic data by studying
the binding of various biotinylated proteins. Polzius and co-
workers adsorbed avidin onto tantalum oxide sensors and then
used this surface to measure the rate of specific biotin binding
(using various biotinylated proteins) where kon = 2.2
105 M-1
s-1.25 Bolivar and co-workers covalently immobilized avidin to
carboxymethylated dextran surface plasmon resonance (SPR)
sensor chips and determined the Kd to be less than 1
10-11
M,26 while Prosperi and co-workers measured it to be 2
10-10
M when studying binding to avidin adsorbed onto nano-
particles.12 It is important to note that the differing surface
materials, immobilization strategies, and modifications of the
binding species employed in these studies may significantly
affect the kinetics and thermodynamics of binding: the disparity
between the reported rate22,23,25 and equilibrium con-
stants24,26,12 emphasizes the requirement for a more detailed
study of binding in the immobilized system.
In the course of our research on the development of multi-
plexed hybridization assays on polymer microparticles,27 we have
studied the kinetics of binding between biotinylated oligonucleo-
tide probes and polymer microparticles bearing immobilized
avidin DN. In this study, wemeasured the rates of association and
dissociation of biotinylated oligonucleotides with avidin immo-
bilized onto microparticles. The effect on the kinetics of different
immobilization chemistries (Figure 1) has been compared with
values measured for the nonimmobilized avidin DN control. The
immobilization systems studied have all been widely employed to
attach protein receptors to surfaces.28 Measurements were also
made of the equilibrium dissociation constant for the immobi-
lized avidin-ligand complex. Furthermore, the differing methods
of protein immobilization have been evaluated for surface cover-
age, activity, and nonspecific binding. In this work, we present the
first detailed kinetic and thermodynamic analysis of the binding
between immobilized avidin DN and solution phase biotinylated
oligonucleotides. The results of this study define the practical
limits of stability for the protein:ligand complex, which are
important in the design of array based assays that exploit
biotin:avidin mediated immobilization.
’MATERIALS AND METHODS
Reagents andApparatus. Labeled and unlabeled biotinylated
oligonucleotides FP1: (50-Cy5-CTAGTTACTCTTGTTC-
biotin-30) and P2 (50-biotin-TTGTTATAGTTCTCTC) were
purchased fromATDBio (Southampton, UK) employing an 8 atom
spacer between the biotin and DNA. Epoxy SU-8-5 and micro-
fabrication reagents were purchased from Chestech Ltd. (Rugby,
UK). Avidin DN was purchased from Vector laboratories
(Burlingame, CA). BCA Assay Kit was obtained from Pierce Bio-
technology (Rockford, IL). Other reagents were obtained from
Sigma-Aldrich (UK), Alfa Aesar (UK), or ThermoFisher (UK).
Epoxy SU-8 monomer was exposed using an EVG 620 mask
aligner with photomask (Compugraphic, UK). Accuspin Micro
centrifuges (Fisher Scientific, UK) (r = 8.5 cm) were used for
particle sedimentation. Fluorescence and fluorescence anisotropy
of FP1 solutions was measured using a SafireII microplate reader
(Tecan, Switzerland). Particles were analyzed under a fluorescence
microscope (Carl Zeiss AG, Germany) or using a FACSaria flow
cytometer (BD Biosciences, U.S.). Kinetic and thermodynamic
constants were calculated by fitting data using commercially
available software (SigmaPlot 11.0, Systat Inc., Chicago, IL and
Origin 8.1, Northampton, MA). Goodness of fit measurements for
these analyses is given in Supporting Information (SI) Table S-1.
Figure 1. Three different routes to avidin-functionalized particles starting from amine functionalized particles. (a) succinic anhydride, DMAP, DIPEA,
DMF, (b) avidin DN, EDC.HCl, pH 7.0, (c) i. EDC.HCl, sulfo-NHS, pH 5.0, ii. avidin DN, pH 7.4, (d) biotin, DIC, DIPEA, DMF, DMSO, (e) avidin
DN, pH 7.0. Type 1 avidin DN particles.

2007 dx.doi.org/10.1021/ac102762q |Anal. Chem. 2011, 83, 2005–2011
Analytical Chemistry ARTICLE
Microparticle Fabrication. SU-8 microparticles were fabri-
cated from photoactive epoxy SU-8-5 with dimensions 5
10

20 μm as reported previously.27
Microparticle Functionalization. Amine functionality was
introduced to the microparticles by the ring-opening of residual
surface epoxide groups of the SU-8 with the bis-amine Jeffamine
ED-900, as described previously.29 The loading density of
primary amine was quantified by the Kaiser test,30 and was
typically between 25 and 35 μmol g-1 (4.4-6.2 nmol cm-2).
Carboxyl groups were introduced by the reaction between
succinic anhydride and amino particles. Avidin DN was covalently
immobilized to these acid functionalized particles either with the
coupling reagent, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide
(EDC), in situ (Type I) or stepwise by first preactivating the acid to
the corresponding N-hydroxysuccinimidyl (NHS) ester before
addition of protein (Type II) (Figure 1).
Biotin functionalization was achieved using carbodiimide-medi-
ated coupling to Jeffamine ED-900-functionalized particles. Avidin
DN was immobilized to biotin functionalized particles by affinity
capture by allowing both species to mix (in suspension, 2 h) (Type
III). Protein functionalized particles required gentle agitation in
SSPE (5
) storage buffer (0.75MNaCl, 50mMsodiumphosphate,
5 mM EDTA, pH 7.0) for 48 h to completely remove any physi-
sorbed protein. The particles were stored as 1mgmL-1 suspensions
in SSPE (5
) buffer with 0.02%Tween 20 and 0.01% sodium azide.
Immobilized protein was quantified using the BCA assay.31
Detailed protocols for all synthetic steps are available as SI.
Binding Studies. Theassociation anddissociationkinetics of the
labeled and biotinylated oligonucleotide, FP1, interaction with avidin
DNweremeasuredboth in solution and for all threemodes of protein
immobilization to microparticles. The protein-FP1 complex was
quantified in solution bymeasuring the anisotropy of the fluorescence
signal from the Cy-5 label on FP1. For solution experiments, it was
necessary to use an excess of unlabeled biotin in the reactionmixture,
since solutions containing only FP1 and avidin DN were found to
undergo self-quenching upon binding, likely due to the proximity of
four fluorophores to each other in the protein tetramer (SI Figure
S-1).32 The particle-immobilized complex was quantified using
fluorescence microscopy where a linear relationship between fluor-
escence intensity and amount of bound FP1 was assumed.
The association of biotin and avidin was measured throughout a
time course experiment and the resultant data analyzed to deter-
mine the association rate constant, kon. For the solution phase
reaction this was achieved with real-time monitoring, whereas the
binding to particleswas terminated at time-points by the addition of
an excess of biotin before themeasurement of particle fluorescence.
Dissociation kinetics were obtained by monitoring the decrease in
the fluorescence anisotropy of avidin DN-FP1 complex in solution
and the decrease in fluorescence of particles with the complex
immobilized when either was exposed to an excess of biotin to
impose effectively irreversible complex dissociation.
Equilibrium thermodynamics of binding on microparticles
were measured by titrating FP1 with immobilized avidin. The
concentration of immobilized complex was determined by
measuring particle fluorescence after binding equilibrium had
been established. For large sample sizes, particles were analyzed
individually using flow cytometry.
’RESULTS AND DISCUSSION
Protein Immobilization. The surface densities of immobilized
avidin and bound oligonucleotide FP1 were determined by BCA
protein-assay and fluorescencemicroscopy respectively (Table 1). All
avidin functionalized particle typeswere found to have a stablemono-
layer of protein; repeating the BCA assay after the particles had been
suspended in buffer for fourweeks indicatedno loss of avidin from the
surface. Furthermore, the suspension of particles functionalized with
fluorescently labeled avidin togetherwith particles functionalizedwith
unlabeled avidin for three days resulted in no change to the fluore-
scence observed for either particle type. This result confirmed the
stability of the immobilized avidin tetramer. Importantly, no protein
migration between particles with avidin immobilized by affinity
capture (Type III) was detected, likely due to each avidin tetramer
being linked to the particle via two biotin ligands.
Initial studies of immobilized complex dissociation using an
excess of biotin in a displacement assay format showed biphasic
dissociation kinetics made up of two distinguishable unbinding
processes (Figure 2). The slow dissociation rate represented
the specific protein-ligand complex while the fast rate was
assumed to be the release of nonspecifically bound ligand (SI
Figure S-2). This nonspecific binding was highest for Type I
particles where avidin had been immobilized with EDC in situ.
The protein on these particles also displayed the lowest activity
with only 1.3% of the total biotin binding sites capturing FP1,
compared with 6.5% for Type II particles. It should be noted
that greater than 50% binding would not be expected due to the
orientation of the avidin tetramer upon immobilization, result-
ing in ligand access to two of the four binding sites being
Table 1. Coverage and Activity of Avidin DN Immobilized to
Particles
FP1
binding capacity
avidin DN particle
immobilization
avidin
loading
(pmol cm-2)
avidin
layers
(number)
specific
binding
(pmol cm-2)
nonspecific
binding
(pmol cm-2)
type I 11.6( 4.4 1.4( 0.5 0.61( 0.02 0.95( 0.05
type II 6.7( 1.0 0.8( 0.1 1.73( 0.06 1.29 ( 0.06
type III 8.3( 2.3 1.0( 0.3 0.75( 0.02 0.36( 0.07
Figure 2. The displacement of biotinylated labeled-oligonucleotide
FP1 from immobilized avidin particles in an excess of biotin, displaying
biphasic dissociation due to the fast initial loss of nonspecifically bound
ligand followed by the loss of complexed ligand. Data are fitted to a
double exponential decay. Subsequent experiments used surfactant to
eliminate the nonspecific binding (detail in SI Figure S-3).

2008 dx.doi.org/10.1021/ac102762q |Anal. Chem. 2011, 83, 2005–2011
Analytical Chemistry ARTICLE
blocked by the particle surface.14 The lower activity observed
for Type I immobilized avidin may be due to cross-linking of the
protein in the presence of EDC, which could potentially result
in more of the binding sites being rendered inactive. Affinity
captured avidin (Type III) was also found to have low activity
(2.3% of binding sites active) but did display very little non-
specific binding toward FP1.
Nonspecific binding of FP1 was reduced to below detect-
able levels with little effect on the specific interaction by
using an increased concentration of surfactant in the buffer
[Tween-20 (1%) or sodium dodecyl sulfate (SDS, 0.2%)]
(SI Figure S-3).
Experimental Determination of Rate Constants for Asso-
ciation and Dissociation of Avidin-biotinylated Oligonu-
cleotide Complex Formation (kon and koff). The capture of
biotinylated oligonucleotide FP1 by avidin DN was modeled
according to a simple bimolecular association process as shown
in eq eq. 1, where [avidin DN], [FP1] and [avidin FP-1] are the
concentrations of free avidin, free biotinylated oligonucleotide
FP1 and avidin-FP1 complex, respectively.
½avidinDNþ½FP1s
F
R
kon
koff
½avidinDN - FP1 ðeq. 1Þ
Data sets measuring the binding in solution and on function-
alized microparticles were fitted to eq eq. 2 to determine the rate
constant, kon (Figure 3).
D½avidinDN - FP1
Dτ
¼ kon ½avidinDN½FP1 ðeq. 2Þ
The solution-phase association rate constant was measured as
1.22 ( 0.11
106 M-1 s-1, whereas this value was halved for
all three immobilized systems, being measured as 0.53( 0.06

106 M-1 s-1. These values were lower than that exhibited by the
wild-type avidin-biotin complex (Table 2). Solution-phase
diffusion limited rate constants (kdiff) were calculated using the
Smoluchowski equation (eq eq. 3, where NA is Avogadro’s
constant (mol-1), R
Rβ
the sum of the (hydrodynamic) radii
(m) of avidin DN and the biotin-modified oligonucleotide FP1
and D
Rβ
the sum of the diffusion coefficients (m2 s-1)). This
equation can be modified to allow calculation of the rate constant
for the immobilized system (eq eq. 4).33-35 Diffusion coeffi-
cients were based on values reported for similar binding species,
wild-type avidin (5.98
10-11 m2 s-1)36 and a 13-mer labeled
oligonucleotide similar in size to that used in our study (8.4

10-11 m2 s-1).37 Hydrodynamic radii were calculated using the
Einstein-Stokes equation (eq eq. 5, where kB is the Boltzmann
Figure 3. Kinetic studies of the Avidin DN FP1 complex. Association (top); with change in complex concentration with time (left) and linearized data
plotted as inverse noncomplexed avidin site concentration versus time (right). Dissociation (bottom); with normalized change in complex concentration
versus time when exposed to an excess of biotin, blocking FP1 rebinding (left) and linearized data plotted as the natural Log of concentration versus time
(right).

2009 dx.doi.org/10.1021/ac102762q |Anal. Chem. 2011, 83, 2005–2011
Analytical Chemistry ARTICLE
constant (J K-1), temperature is 293.15 K, and viscosity (η) is
1.098
10-3 Pa s).
kdif f ¼ 4πR
Rβ
D
Rβ
NA103 ðeq. 3Þ
kdif f ¼ 2πR
Rβ
D
β
NA103 ðeq. 4Þ
D ¼
kBT
6πηR
ðeq. 5Þ
kdiff was thus calculated as 6.09
109 and 1.78
109 M-1 s-1
for the solution phase and immobilized systems, respectively. kdiff
is much greater than the measured rate constant (kon) due to the
low steric factor (f) (where only a fraction of the surfaces of
avidin and biotin are involved in binding which results in the
probability of complex formation from a collision being <1).38
The steric factor can be described as the ratio between the
observed and diffusion limited rate constants and for the solution
phase binding discussed herein is calculated as f = 0.0002. The
difference in the measured association rate as caused by protein
immobilization is predicted by these calculations (eqs eq. 3 and
eq. 4), supporting our conclusion that binding is limited by
diffusion. It is important to note the association rate constant
between surface bound avidin DN and oligonucleotide FP1 is
decreased by over 2 orders of magnitude compared with the wild-
type avidin-biotin binding. This decrease in association rate
constant is caused by a composite of two effects: first, conjuga-
tion of biotin to the oligonucleotide probe results in a 98%
reduction of kon and second, the attachment of avidin to the
surface eliminates its diffusion, leading to a further 57% decrease
of kon.
Data sets measuring the irreversible dissociation of complex in
an excess of biotin were fitted to eq eq. 6 (Figure 3).
D½FP1
Dτ
¼ kof f ½avidinDN - FP1 ðeq. 6Þ
The avidin DN-FP1 dissociation rate constant in solution was
calculated from fluorescence anisotropymeasurements as 0.73(
0.06
10-6 s-1. This value is more than an order of magnitude
greater than reported for the wild-type system.1 Binding studies
have identified the urea moiety of biotin as having the strongest
interaction with the active-site of avidin and this remains
unaltered in our biotinylated probe.39 However, crystal struc-
tures have indicated that additional binding interactions between
the carboxylic acid moiety of biotin and residues in the binding
site of the protein further stabilize the complex.40 Other studies
using the biotin analogue desthiobiotin found that modification
of the acid to the correspondingmethyl ester resulted in an 8-fold
increase in the rate of dissociation.41 In many applications of the
biotin-avidin system for probe capture including the oligonu-
cleotide-modified ligand FP1 used in this work, biotin is con-
jugated to the probe via an amide linkage, replacing the terminal
carboxylic acid. Interactions of the carboxylic acid functional
group of biotin with the binding site are therefore replaced by an
amide which has significantly weaker binding and this weakened
interaction is observed as an increase in the dissociation rate
constant.
The immobilized complexes were found to have a faster rate of
dissociation compared with avidin DN-FP1 in solution, regard-
less of the immobilization method used. FP1 dissociation from
Type I immobilized avidin DN was six times faster than the
solution system, resulting in a complex half-life of less than two
days. Conformational changes in the protein have been shown to
occur upon biotin binding, most notably a loop between two β-
strands of each avidin tetramer subunit (residues 35-46, avidin
isolated from hen egg white) which moves to cap the entrance of
the binding pocket when occupied and contributes to the excep-
tionally high complex stability.42 The immobilization of avidin
and the steric restrictions imposed by close-packing of a surface
monolayer may restrict these conformational rearrangements
required for maximum complex stability. Avidin immobilized in
the presence of coupling reagent (Type I particles) may be
further restricted by protein-protein cross-linking which would
not form under either Type II/III immobilization conditions or if
avidin were spread more sparsely on the surface. Comparing the
different values of koff for the three immobilization methods,
Type I avidin particles are shown to form the least stable complex
with FP1.
Kd ¼
½avidinDN½FP1
½avidinDN - FP1
¼
kof f
kon
ðeq. 7Þ
FP1 titrations against immobilized avidin (SI Figure S-6) re-
sulted in experimentally determined equilibrium dissociation
constants when fitted to eq eq. 7, which correlated well with
and corroborated the values calculated from the individual
kinetic experiments (Table 2).
The Significance of Increased Complex Dissociation on
Multiplexed Bioassays. The increased rate of dissociation and
resulting decrease in half-life observed for the immobilized avidin
DN-FP1 complex has implications for any assay designed to have
discrete probes sharing the same analyte solution, such as
particle-based multiplexed assays and spotted microarrays. No-
tably, biotinylated probes could potentially dissociate and reas-
sociate to surface avidin forming a uniform coating of the probe
mixture where probe distribution would have previously been
patterned. This problem would bemagnified over extended assay
time scales and was investigated by mixing two discrete popula-
tions of avidin-functionalized particles, one complexed with an
unlabeled biotinylated oligonucleotide probe (P2) and the other
complexed with fluorescently labeled oligonucleotide probe
Table 2. Kinetic and Thermodynamic Data Comparing the Stability of the Avidin Biotin Complex with Avidin DN FP1 Complex
in Solution and When Particle-Immobilized
protein-ligand complex kon (M-1 s-1) (
106) koff (s-1) (
10-6) t1/2 (days) koff/kon (pM) Kd (pM)
free avidin-biotin (w.t.)1 70 0.04 200 0.00057
free avidin DN-FP1 1.22( 0.11 0.73( 0.06 11.0 0.60
type I avidin DN-FP1 0.54( 0.04 4.15( 0.29 1.9 7.73 10.03( 1.72
type II avidin DN-FP1 0.52( 0.03 1.57( 0.16 5.1 3.04 4.49( 0.85
type III avidin DN-FP1 0.52( 0.06 1.52( 0.08 5.3 2.91 4.05( 0.46

2010 dx.doi.org/10.1021/ac102762q |Anal. Chem. 2011, 83, 2005–2011
Analytical Chemistry ARTICLE
(FP1). Particle mixtures were analyzed by flow cytometry and
initially two particle populations could be readily distinguished
by the difference in their fluorescence intensity (Figure 3).
However, over time (up to 19 h), probe dissociation from and
reassociation to immobilized avidin merged these into a single
fluorescent population, ultimately resulting in a mixture of both
fluorescent and nonfluorescent probes on every particle
(Figure 4).
The migration of biotinylated probes highlights the impor-
tance of minimizing the time between initiation and sample
reading in this assay format. An example of this can be found in
our previous work where multiplexed encoded-microparticle
DNA hybridization assays were completed within 40 min to
successfully discriminate between sequences differing by a single
base27 The avidin-biotin capture system may not be ideally
suited to assays run over longer time scales, especially where the
detection of multiple analytes is required. An example of this may
be the immobilization of specific antibodies to quantum dots for
use in multiplexed biological imaging, with different cellular
targets for each antibody-quantum dot pair.43 When biotin disso-
ciation is significant, the time scale for any assays would be
limited. The rate constants for dissociation and association
therefore define the time scale limits for multiplexed experi-
ments. Further problems may be encountered when probes have
been immobilized and then stored as a mixture for long periods
in advance of their use in a multiplexed assay. Migration prob-
lems during long-term storage might be mitigated by storing dry
then rehydrating prior to use. This should work well for micro-
arrays, but aggregation may be a problem for some particle-based
systems.
’SUMMARY
Quantitative comparisons have been made between three
popular protein immobilization methods used to attach avidin
tomicroparticles. Stable proteinmonolayers were formed in each
case, however Type I protein coupling (EDC in situ) resulted in
both the highest levels of nonspecific and the lowest levels of
specific biotinylated-probe binding. Furthermore, complex sta-
bility was reduced with a half-life of less than two days. The
immobilization of avidin via activation with NHS or affinity
capture with immobilized biotin led to identical association
and dissociation rate constants and similar levels of nonspecific
binding.
The kinetic and thermodynamic properties of biotinylated-
oligonucleotide (FP1) binding to avidin DN immobilized on
polymer microparticles via different attachment chemistries have
been characterized. Immobilization was shown to have no effect
on the rate of association beyond that expected from the
reduction of the possible approach trajectories of biotin due to
surface attachment. Explicitly, the rate of association was de-
creased for the immobilized system compared with the free
system by the proportion predicted using equations derived from
collision theory (eqs eq. 3 and eq. 4). Rates of association of FP1
to avidin DN were found to be reduced relative to the binding of
free biotin (which is ∼60 times faster). This slower binding was
due to the decreased rate of diffusion as a consequence of the
conjugation of the (relatively) large oligonucleotide to biotin.
Rates of immobilized complex dissociation were increased com-
pared with the nonimmobilized avidin system, which in turn
displayed faster dissociation than the wild-type avidin-biotin
complex. These results showed that both the loss of the
carboxylic acid moiety of biotin (in formation of the amide bond
to DNA) and the immobilization of avidin have significant effects
on complex stability, leading to the system having a complex half-
life of only five days. The decreased association rate and increased
dissociation rate for modified biotin to avidin increased the
equilibrium dissociation constants to 0.6 pM (solution) and
4 pM (immobilized), a significant shift in equilibrium compared
with the tighter binding in the wild type system. This is important
as when avidin-biotin binding is used to immobilize or link
larger species, binding is often assumed to be as stable as that for
unmodified binding, which we have demonstrated is not always
the case. Furthermore the increased dissociation rate (and
reduced half-life) has implications for the use of avidin affinity
as a probe immobilization method in multiplexed arrays when
different probe types are to share the same sample solution. This
was exemplified when different probes were immobilized on
Figure 4. Flow cytometry data highlighting biotinylated-probe migra-
tion between Type II immobilized avidin DN particles. A sample
containing particles with the unlabeled biotinylated oligonucleotide
(P2) bound was mixed with a sample of an equal amount of particles
with fluorescently labeled biotinylated oligonucleotide (FP1) attached,
the suspension was analyzed by flow cytometry at time points (from top;
0, 143, 1114 min). Initially two discrete fluorescence populations are
discernible both visually (histogram and dot-plot) and by Gaussian
fitting (Origin 8.1, www.originlab.com). However within 19 h the
continued dissociation and reassociation of ligands results in indiscern-
ible populations with the surface of all particles tending toward a mix of
both biotinylated-probe types.

2011 dx.doi.org/10.1021/ac102762q |Anal. Chem. 2011, 83, 2005–2011
Analytical Chemistry ARTICLE
different particles, then mixed and probe migration was observed
with the previously heterogeneous particle sets being indistin-
guishable within one day.
These results highlight potential problems with the use of the
avidin-biotin system for the immobilization of sensor probes in
multiplexed biological assays. This need not preclude the use of
the system but may be important for researchers to consider.
With careful optimization of multiplex assay conditions these
problems can be circumvented, most notably by decreasing the
storage and assay times over which probes are mixed within the
same buffer solution.
’ASSOCIATED CONTENT
b
S Supporting Information. Figures and discussion regard-
ing fluorescence quenching of complexed avidin DN-FP1 in
solution and the determination of nonspecific binding between
FP1 and immobilized protein and the blocking of this binding. Also
included is a description of the conversion of experimental data into
kinetic and thermodynamic constants and equilibrium dissociation
constants, Kd, for particle-immobilized avidin DN. Types of line
fitting and goodness of fit measurements for graphical data is
included and detailed syntheticmethods formicroparticle function-
alization are also documented. This material is available free of
charge via the Internet at http://pubs.acs.org.
’AUTHOR INFORMATION
Corresponding Author
*Tel:þ44 2380 595919. E-mail: plr2@soton.ac.uk.
’ACKNOWLEDGMENT
This project was supported by Research Councils, U.K.,
through the Basic Technology Programme. We thank the clean-
room staff at EPFL, Switzerland, for their assistance with
fabrication and Andrew Whitton for technical assistance.
’REFERENCES
(1) Green, N. M. Methods Enzymol. 1990, 184, 51.
(2) Wilchek, M.; Bayer, E. A. Methods Enzymol. 1990, 184, 5.
(3) Bingham, J. P.; Bian, S.; Tan, Z. Y.; Takacs, Z.; Moczydlowski, E.
Bioconjugate Chem. 2006, 17, 689.
(4) Dupont-Filliard, A.; Roget, A.; Livache, T.; Billon, M. Anal.
Chim. Acta 2001, 449, 45.
(5) Fang, X.; Liu, X.; Schuster, S.; Tan, W. J. Am. Chem. Soc. 1999,
121, 2921.
(6) Huang, H.; Jie, G.; Cui, R.; Zhu, J. Electrochem. Commun. 2009,
11, 816.
(7) Rodriguez, M. C.; Kawde, A.; Wang., J. Chem. Commun. 2005,
34, 4267.
(8) Chunga, J. W.; Park, J. M.; Bernhardt, R.; Pyuna, J. C.
J. Biotechnol. 2006, 126, 325.
(9) Ouerghi, O.; Touhami, A.; Jaffrezic-Renault, N.; Martelet, C.;
Ben Ouada, H.; Cosnier, S. Bioelectrochem. 2002, 56, 131.
(10) Tombelli, S.; Mascini, M. Anal. Lett. 2000, 33, 2129.
(11) Haddour, N.; Chauvin, J.; Gondran, C.; Cosnier, S. J. Am. Chem.
Soc. 2006, 128, 9693.
(12) Prosperi, D.; Morasso, C.; Tortora, P.; Monti, D.; Bellini, T.
ChemBioChem. 2007, 8, 1021.
(13) Bengali, A. N.; Tessier, P. M. Biotechnol. Bioeng. 2009, 104, 240.
(14) Misawa, N.; Yamamura, S.; Yong-Hoon, K.; Tero, R.;
Nonogaki, Y.; Urisu, T. Chem. Phys. Lett. 2006, 419, 86.
(15) Meiser, F.; Cortez, C.; Caruso, F. Angew. Chem., Int. Ed. 2004,
43, 5954.
(16) Su, X.; Wu, Y.; Robelek, R.; Knoll, W. Langmuir 2005, 21, 348.
(17) Vermette, P.; Gengenbach, T.; Divisekera, U.; Kambouris,
P. A.; Griesser, H. J.; Meagher, L. J. Colloid Interface Sci. 2003, 259, 13.
(18) Reznik, G. O.; Vajda, S.; Cantor, C. R.; Sano, T. Bioconjugate
Chem. 2001, 12, 1000.
(19) Olejnik, J.; Krzymanska-Olejnik, E.; Rothschild, K. J. Nucleic
Acids Res. 1996, 24, 361.
(20) Finn, F. M.; Titus, G.; Hofmann, K. Biochemistry 1984, 23,
2554.
(21) Ke, S.; Wright, J. C.; Kwon, G. S. Bioconjugate Chem. 2007, 18,
1644.
(22) Zhao, S.; Reichert, W. M. Langmuir 1992, 8, 2785.
(23) Wayment, J. R.; Harris, J. M. Anal. Chem. 2009, 81, 336.
(24) Graves, S. W.; Woods, T. A.; Kim, H.; Nolan, J. P. Cytometry,
Part A 2005, 65A, 50.
(25) Polzius, R.; Diessel, E.; Bier, F. F.; Bilitewski, U. Anal. Biochem.
1997, 248, 269.
(26) Bolivar, J. G.; Soper, S. A.; McCarley, R. L. Anal. Chem. 2008,
80, 9336.
(27) Broder, G. R.; Ranasinghe, R. T.; She, J. K.; Banu, S.; Birtwell,
S.; Cavalli, G.; Galitonov, G.; Holmes, D.; Martins, H. F. P.; MacDonald,
K. F.; Neylon, C.; Zheludev, N.; Roach, P. L.; Morgan, H. Anal. Chem.
2008, 80, 1902.
(28) Rusmini, F.; Zhong, Z.; Feijen, J. Biomacromolecules. 2007, 8,
1775.
(29) Cavalli, G.; Banu, S.; Ranasinghe, R. T.; Broder, G. R.; Martins,
H. F. P.; Neylon, C.; Morgan, H.; Bradley, M.; Roach, P. L. J. Comb.
Chem. 2007, 9, 462.
(30) Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal.
Biochem. 1970, 34, 595.
(31) Smith, P. K.; Krohn, R. I.; Hermanson, G. T.; Mallia, A. K.;
Gartner, F. H.; Provenzano, M. D.; Fujimoto, E. K.; Goeke, N. M.;
Olson, B. J.; Klenk, D. C. Anal. Biochem. 1985, 150, 76.
(32) West, W.; Pearce, S. J. Phys. Chem. 1965, 69, 1894.
(33) Smoluchowski, M. V. Z. Phys. Chem. 1917, 92, 129.
(34) Berg, O. G.; von Hippel, P. H. Annu. Rev. Biophys. Chem. 1985,
14, 131.
(35) Rice, S. A. In Compnhensive Chemical Kinetics, Chapter 2;
Bramford, C. H., Tipper, C. F. H.; Compton, R. G., Eds; Elsevier:
Amsterdam, 1985; Vol. 25.
(36) Green, N. M. Biochem. J. 1964, 92, 16C.
(37) Zhang, C. Y. Analyst. 2006, 131, 484.
(38) Northrup, S. H.; Erickson, H. P. Proc. Natl. Acad. Sci. U. S. A.
1992, 89, 3338.
(39) Green, N. M. Biochem. J. 1963, 89, 599.
(40) Livnah, O.; Bayer, E. A.; Wilchek, M.; Sussman, J. L. Proc. Natl.
Acad. Sci. U. S. A. 1993, 90, 5076.
(41) Green, N. M. Adv. Protein Chem. 1975, 29, 85.
(42) Pugliese, L.; Coda, A.; Malcovati, M.; Bolognesi, M. J. Mol. Biol.
1993, 231, 698.
(43) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.;Mattoussi, H.Nat.
Mater. 2005, 4, 435.
’NOTE ADDED AFTER ASAP PUBLICATION
This paper was published on the Web on February 3, 2011.
Equations1 and 2 were corrected. The corrected version was
reposted on February 15, 2011.