Label-Free Multimodal Protease Detection Based on

Protein/Perylene Dye Co-assembly and Enzyme-Triggered

Disassembly

Yiyang Lin, Robert Chapman, and Molly M. Stevens*

Department of Materials, Department of Bioengineering and

Institute for Biomedical Engineering, Imperial College London,

Exhibition Road, London SW7 2AZ, UK.

E-mail: m.stevens@imperial.ac.uk

Tel: +44 (0)207 594 6804

Fax: +44 (0)207 594 6757;

1

ABSTRACT

The development of novel assays for protease sensing plays an important role in

clinical diagnostics and therapeutics. Herein, we report a supramolecular platform for

label-free protease detection, based on protein/dye self-assembly and enzyme-

triggered disassembly. In a typical case, co-assembly of protamine sulfate and

perylene dye via electrostatic attractions and π-π interactions caused significant

colorimetric and fluorescent responses. Subsequent addition of trypsin was found to

cleave the amide bonds of protein, triggering the dissociation of protein/dye

aggregates and the release of perylene dyes. The enzyme-triggered disassembly was

transduced into multiple readouts including absorption, fluorescence, and

polarization, which were exploited for trypsin detection and inhibitor testing. This

assay was also used for turn-on fluorescence detection of Cathepsin B, an enzyme

known to be over-expressed in mammalian cancer cells. The integration of

supramolecular self-assembly into enzyme detection in this work has provided a novel

label-free biosensing platform which is highly sensitive with multimodal readouts.

The relative simplicity of the approach avoids the need for time-consuming substrate

synthesis, and is also amenable to naked eye detection.

2

Keywords: self-assembly, biosensor, protease, protein, perylene bisimide.

INTRODUCTION

Proteolytic enzymes or proteases play central roles in most biological processes

owing to their ability to regulate the physiological functions of many proteins through

initiating hydrolysis at the post-translational level. Aberrant activity of proteases is

also central to major human diseases such as cardiovascular, oncologic,

neurodegenerative, and inflammatory diseases.1-3 The screening of protease-targeted

inhibitors is thus also of high clinically relevance to identify new therapeutic

approaches for these diseases. For example, protease inhibitors against matrix

metalloproteinases (MMPs) have been proposed to control tumor growth.4

Over the past decade, proteases have mostly been detected using immunoassays

which involve antibody or affinity based detection5, have been mostly used to detect

proteases, owing to the high sensitivity and specificity of this approach. However, the

immunoassay format always requires a process of antibody

identification/generation/isolation, as well as potentially complicated bioconjugation

for antibody immobilization and fluorophore/chromophore labeling. Besides, readout

of direct enzyme activity cannot be achieved from the immunoassay. Another strategy

for protease sensing is based on hydrolysis of the peptide substrate by the target

protease. This approach can provide information about enzyme activity and is used in

diagnostics and drug discovery.6-8 Assaying proteolytic activity can be diversified by

3

combining short peptide substrates with different detection principles (e.g.

fluorescence resonance energy transfer, FRET) and signal-enhancing reporters (e.g.,

fluorophores and chromophores).9-12 However, this method always requires organic

and bioconjugation chemistry to synthesize specific peptide substrates and attach

molecular probes to the substrates. Therefore, the development of novel label-free

approaches for sensitive protease detection is of great interest.

To this end, we have designed a versatile biosensing platform for protease activity

detection based on the non-covalent co-assembly of protein and perylene bisimide.

Protein self-assembly is of great fundamental interest to understand the

physicochemical basis of protein-protein interactions and can be used to design new

to novel biomaterials for disease diagnostics and therapy.13 However, the concept of

integrating organic dyes into protein self-assembly has been rarely reported. On the

other hand, perylene bisimide (PBI) derivatives have been extensively used as

industrial pigments owing to photo/thermal stability, high fluorescence quantum yield,

and inert chemistry.14 More recently, research into PBI derivatives has emerged in

various optical and electronic applications including organic photovoltaic devices,

field-effect transistors, and n-type semiconductors.14,15 Owing to the hydrophobic

effect and π-π interactions, PBI derivatives are mostly water-insoluble and have a

strong tendency to aggregate,16 which results in fluorescence quenching in aqueous

solution. In the past, this has greatly restricted their applications in biosensing and

bioimaging.17-21

Here, a highly fluorescent water-soluble PBI dye was synthesized by attaching

4

multiple charge groups to perylene bisimide.22 Driven by electrostatic attractions and

π-π interactions, the protein/PBI dye system self-assembled into soft nanoparticles,

which led to a series of optical responses including aggregation-induced fluorescence

quenching, polarization, and decrease in UV-vis absorption. The presence of protease

hydrolyzed protein and dissociated the protein/dye self-assembly, which consequently

restored PBI absorption/fluorescence emission and decreased its fluorescence

polarization. In particular, we have used protamine and polylysine as enzyme

substrates to fabricate protease sensors for trypsin and Cathepsin B, respectively. In

comparison with common protease detection methods, our assay utilizes natural

proteins or peptides as substrates which avoids time-consuming substrate synthesis. In

addition, the optical probe is integrated into our assay platform through

supramolecular self-assembly or non-covalent interactions without the need for

complicated fluorophore labeling. The combination of the dynamic nature of non-

covalent self-assembly with the possibility to incorporate a variety of molecules that

can trigger the optical responses in the protein holds great promise for applications as

stimuli-responsive supramolecular systems (e.g., nanocarriers) and biosensors.

Experimental Section

Materials. Protamine sulfate from salmon, poly-L-lysine hydrobromide, trypsin from

bovine pancreas (TPCK-treated, essentially salt-free, lyophilized powder, 10,000≧

BAEE units/mg protein) and Cathepsin B were purchased from Sigma–Aldrich (UK).

GRP peptide was provided by GenScript (US) and used without further purification.

5

Water was purified using a Millipore filtration system.

Synthesis of water-soluble PBI derivative. Aspartic acid-functionalized water-

soluble perylene bisimide (PBI-Asp) was synthesized according to the literature.22 In

detail, 3,4,9,10-Perylenetetracarboxylicacid bisanhydride (196 mg, 0.5 mmol),

aspartic acid (146 mg, 1.1 mmol), zinc acetate (1.83 mg, 0.05 mmol), and 4.0 g of

imidazole were heated at 120 °C for 12 h with stirring under argon atmosphere. The

reaction mixture was cooled to 90 °C, and then poured into water in the presence of

argon. The mixture was filtered and the filtrate was acidified to pH 2-3 with hydrogen

chloride solution. The precipitate was filtered, washed with water and dried under

vacuum at 80 °C to give the product.1H NMR (400 MHz, d6-DMSO,) δ: 12.85 (-

COOH, 2H), 7.91 (-CH-, 4H), 7.75 (-CH-, 4H), 5.99 (-CH-, 2H), 3.38 (-CH 2-, 2H),

2.90 (-CH2-, 2H). FTIR v/cm−1: 3525, 2939, 1750, 1695, 1635, 1590, 1571, 1508,

1435, 1401, 1362, 1341, 1302, 1255, 1172, 1132, 992, 960, 854, 809, and 745.

Sample characterization. TEM micrographs were obtained with a JEOL 2000FX

(working voltage of 200 kV) by negative-staining method with uranyl acetate solution

(1.0 wt%) as the staining agent. One drop of the solution was placed onto a carbon

Formvar-coated copper grid for 3~5 min. The excess liquid was sucked away with

filter paper. After this, one drop of the staining agent was placed onto the copper grid

for 2~5 min. After removing the excess staining agent with filter paper, the sample

was dried in air before TEM observation. Fluorescence spectra were recorded on

Fluorolog®-3 spectrofluorometer. UV-vis absorbance was measured on a Perkin

6

Elmer Lambda 25 UV-vis spectrometer with the path length of 1.0 cm. Fluorescence

polarization was measured on SpectraMax M5 plate reader and was experimentally

calculated from the measurement of fluorescence intensity parallel to the plane of

linearly polarized excitation light (I ∥), and that perpendicular to the excitation plane (

I⊥), which was expressed as:P=I ∥−( G× I⊥ )I ∥+(G × I⊥ )

, wherein G was an instrumental factor.

Dynamic light scattering (DLS) and ζ-potential were measured on a Malvern

Zetasizer Nano ZS (Malvern, UK) with a backscattering detection at 173o, equipped

with a He-Ne laser (λ=632.8 nm).

Trypsin activity assay. Trypsin stock solutions were prepared in 1.0 mM hydrogen

chloride solution and stored at -20 oC. The trypsin activity test was conducted in 384-

well plates and monitored by plate reader. In detail, 10 μM PBI-Asp and 15 μg/mL

protamine sulfate were dissolved in 5 mM phosphate buffer (pH 8.5) and transferred

to 384-well plates. After that, 5 μL of trypsin stock solution were added to reach

required enzyme concentrations. The kinetics of enzymatic hydrolysis were monitored

by fluorescence intensity (λex=490 nm, λem=550 nm) using plate reader at room

temperature. After enzymatic digestion for 3 hours, the UV-vis absorbance,

fluorescence spectra, and polarization were recorded.

Trypsin inhibitor test. The trypsin inhibitor, benzamidine hydrochloride, with

varying concentrations was pre-incubated with trypsin at 25 oC for 15 min, and then

added into solutions of protamine/PBI-Asp mixture in phosphate buffer (5.0 mM, pH

7

8.5). The final concentration of PBI-Asp, protamine sulfate, and trypsin was 10 μM,

15 μg/mL, and 10 nM, respectively. The resulting solutions were incubated at room

temperature and the fluorescence intensity (λex=490 nm, λem=550 nm) was recorded

with time.

Cathepsin B assay. The stock solution of Cathepsin B (10 U/mL) was added to

activation buffer with a final concentration of 20 mM DTT, 10 mM EDTA and

incubated at 37 °C for 15 min. To the activated enzyme solution, reaction buffer (25

mM sodium acetate, 1 mM EDTA, pH 5.0, pre-warmed at 37 °C) and polylysine/PBI-

Asp solution (1.0 μg/mL and 1.0 μM) were added. The reaction solution was

incubated at 37 °C and the fluorescence emission from PBI-Asp (λex=490 nm, λem=550

nm) was recorded.

RESULTS AND DISCUSSION

1. Supramolecular self-assembly of protein/perylene dye

A water soluble perylene-derivative named PBI-Asp (Fig. 1a) was synthesized

through condensation reaction of 3,4,9,10-perylenetetracarboxylicacid bisanhydride

and L-aspartic acid. The attachment of aspartic acid to perylene introduced charge

repulsion between multiple carboxyl groups and endowed a high solubility in aqueous

conditions. Three well-resolved absorption peaks (534, 495, and 466 nm) and a weak

broad shoulder around 415 nm were observed in the UV-vis spectrum of PBI-Asp

(Fig. 1b), which were ascribed to characteristic S0→S1 transitions with different

vibronic structures. The ratio of 0−0 to 0−1 transition in absorption intensity was

8

calculated to be 1.56, being close to the normal Frank-Condom progressions

(A0−0/A0−1≈1.6) for non-aggregated PBI.23 This confirmed the existence of free dye

monomer in solution. Meanwhile, the monomeric PBI-Asp displayed a high

fluorescence quantum yield (~50%, Supporting Information) with two emission peaks

at 548 and 590 nm (Fig. 1c).

Figure 1. (a) Molecular structure of PBI-Asp; (b) UV-vis and (c) fluorescence spectra

of 1.0 μM PBI-Asp solution in 5.0 mM phosphate buffer (pH 8.5). The fluorescence

excitation wavelength was 490 nm.

Protamine sulfate is a small nuclear protein with an arginine-rich sequence and

has the ability to condense plasmid DNA to increase gene therapy transduction rates

by both viral and nonviral mediated delivery mechanisms.24 Herein, protamine was

chosen as a protease substrate because it is trypsin-degradable. Strong interactions

between PBI-Asp and protamine sulfate were demonstrated by multiple techniques.

As shown in Fig. 2a, the absorption intensity of PBI-Asp decreased with the addition

9

of protamine sulfate, suggesting the protein-induced aggregation of PBI-Asp. When

the protamine concentration reached 10 μg/mL, the absorption peaks at 534 and 495

nm disappeared completely and a new shoulder attributed to the extended oligomers

was observed at 562 nm. The isosbestic point at 552 nm indicated that two species

exist in equilibrium between monomers and oligomers. The spectral variations were

accompanied by notable changes in the appearances of the solution (Inset in Fig. 2a).

Simultaneously, when adding protamine sulfate, the fluorescence emission

decreased gradually due to the formation of non-emitting aggregates with forbidden

low-energy excitonic transitions (Fig. 2b).25 The fluorescence emission was largely

quenched (~99.8%) by 15 μg/mL of protamine sulfate (Fig. 2c), which was noticeable

with a conventional UV lamp (Inset in Fig. 2b). Interestingly, further addition of

protamine caused the fluorescence to increase (Inset in Fig. 2c), indicating a

weakening in the aggregation of PBI-Asp. This phenomenon is unexpected and will

be discussed in the next section.

10

Figure 2. (a) UV-vis and (b) fluorescence spectra of protamine/PBI-Asp. The

concentration of PBI-Asp is 10 μM while the protamine concentration is (from top to

bottom): 0, 2.0, 4.0, 6.0, 8.0, and 10 μg/mL. The inset in Fig. 2a shows PBI-Asp

solution without (left) and with (right) protamine; the inset in Fig. 2b shows the same

solutions upon UV irradiation (365 nm). (c) Fluorescence intensity of PBI-Asp

solution (10 μM) (λex= 490 nm, λem = 550 nm) with different amounts of protamine

sulfate. Inset: Intensity profile of PBI-Asp fluorescence when protamine concentration

varies from 15 to 120 μg/mL. (d) Fluorescence polarization of protamine/PBI-Asp

mixtures, indicating the variations of perylene rotation diffusion rate: φ1<φ3<φ2.

Fluorescence polarization was further employed to provide information into

protein/dye self-assembly at the molecular level, especially with regards to perylene

11

rotation. In principle, when a dye is excited with polarized light, it will emit

fluorescence with the same polarization assuming the dye doesn’t rotate during the

lifetime of the excited state.26 Depolarization occurs when the dye rotates during its

emission lifetime. The faster it rotates, the smaller the fluorescence polarization will

be. As shown in Fig. 2d, the addition of protamine (0~15 μg/mL) greatly enhanced the

fluorescence polarization. This suggests that the rotation diffusion of PBI-Asp was

restricted in the presence of protamine (φ1>φ2, Fig. 2d), owing to the protein-assisted

perylene aggregation. Further addition of protamine (15~200 μg/mL) unexpectedly

decreased the fluorescence polarization, implying a faster PBI-Asp rotation (φ3>φ2,

Fig. 2d). This phenomenon coincided with the fluorescence results in Fig. 2c, in

which a maximum fluorescence quenching effect was observed with 15 μg/mL

protamine.

We hypothesized that the optical responses (i.e., absorption decrease, emission

quenching, and fluorescence polarization) originated from protein-induced dye self-

assembly, which were studied by ζ-potential, dynamic light scattering (DLS), and

transmission electron microscopy (TEM) (Fig. 3). It is known that arginine and

aspartic acid can form a salt bridge through guanidinium–carboxylate interaction

which is a key stabilizing structural element in natural systems including RNA stem

loops25 and zinc finger/DNA complexes26. In this case, the guanidinium–carboxylate

interaction between protamine and PBI-Asp became the main driving force for

protein/dye self-assembly (Fig. 4). As shown in Fig. 3a, the ζ-potential decreased

from negative (-14 mV) to neutral as the protamine concentration increased,

12

indicating the incorporation of cationic protamine into the protamine/PBI-Asp

complex. Zero ζ-potential was achieved with ~10 μg/mL of protamine, consistent with

the results of UV-vis absorbance, fluorescence, and polarization (Fig. 2). These results

suggest the important role of guanidinium–carboxylate charge interaction in

protamine/PBI-Asp self-assembly. The formation of 100-200 nm nanoparticles in

protamine/PBI-Asp solution was demonstrated by TEM and DLS (Fig. 3b, 3c, and

S1). Owing to the protein-assisted aggregation, PBI-Asp was caged (Fig. 4) and its

absorption/fluorescence was strongly quenched owing to π-π stacking (Fig. 2a and

2b). Meanwhile, molecular aggregation prohibited the rotational diffusion of PBI-Asp

and hence intensified its fluorescence polarization (Fig. 2d). When the protein

concentration exceeded 10 μg/mL, the ζ-potential was reverted to positive (Fig. 3a),

suggesting the incorporation of excess protamine into the protein/dye complex. The

increased positive charges on protamine/PBI-Asp particles led to a loose molecular

packing (Fig. 4), which restored the absorption and fluorescence of PBI-Asp (Fig. 2c

and 2d). Owing to the less dense packing of PBI-Asp inside protamine/PBI-Asp

particles, the diffusion rotation of PBI-Asp was enhanced and the fluorescence

polarization decreased when the protamine concentration exceeded 10 μg/mL (Fig. 2c

and 2d).

13

Figure 3. (a) ζ-potential of protamine/PBI-Asp with varied protamine concentrations,

in which the PBI-Asp concentration is 10 μM. (b) DLS and (c) TEM image of

protamine/PBI-Asp with 10 μM of PBI-Asp and 15 μg/mL of protamine sulfate.

To further illustrate the factors that contribute to protein/dye self-assembly, we

investigated the co-assembly behavior of PBI-Asp and a protamine-analogous

peptide. To this end, a short arginine-rich peptide with the sequence of Gly-Arg-Pro-

Gly-Arg-Pro-Gly-Arg-Pro (GRP) was synthesized. It is interesting that no significant

fluorescence quenching or UV-vis absorption decrease was observed when adding

GRP peptide to PBI-Asp solutions (Fig. S2). Although GRP is highly positive and will

electrostatically interact with peryelene dyes, the short peptide chain makes it less

efficient at assembling perylene dyes. This indicates the polymer-analogue structure

of the protein is indispensible in promoting the self assembly of protamine/PBI-Asp

(Fig. 4).

14

Figure 4. Schematic illustration of guanidinium–carboxylate electrostatic interaction,

protamine/PBI-Asp self-assembly into positive and negative nanoparticles, and

GRP/PBI-Asp interactions.

2. Trypsin detection based on protein/dye disassembly

As discussed above, the polymer-analogous structure of protamine plays a crucial role

in protamine/PBI-Asp self-assembly, while small peptides are less efficient at

promoting the formation of perylene dye oligomers. It is therefore expected that if

protamine is enzymatically cleaved into short peptides, the self-assembled

protamine/PBI-Asp nanoparticles will disaggregate and the caged PBI-Asp be

released. Based on this hypothesis, protamine/PBI-Asp self-assembled complexes are

expected to be applicable for the detection of protamine-specific proteases.

15

Figure 5. (a) Real-time fluorescence enhancement of protamine/PBI-Asp in the

presence of 0-50 nM trypsin; (b) fluorescence spectra after 3 h enzyme digestion. The

protein/dye concentration in (a, b) is 15 μg/mL protamine and 10 μM PBI-Asp for 0-

10 nM trypsin. (c) Emission intensity of protamine/PBI-Asp after incubation with

different amounts of trypsin for 3 h (λex=490 nm, λem=550 nm). The protein/dye

concentrations in Fig. 3c are: (A) 15 μg/mL protamine and 10 μM PBI-Asp, (B) 2.0

μg/mL protamine and 1.0 μM PBI-Asp. (d-g) Appearance of protamine/PBI-Asp

mixtures before (d, f) and after (e, g) trypsin digestion.

Protamine is an ideal trypsin substrate owing to its arginine-rich structure. As the

most important digestive enzyme, trypsin is secreted by the pancreas and plays a key

role in controlling pancreatic exocrine function. It is involved in the digestive enzyme

16

activation cascade, which induces the transformation of other pancreatic proenzymes

into their active forms. Changes in trypsin levels are also closely related to the

presence of some pancreatic diseases.27,28 Here, the real-time detection of trypsin

activity was realized by tracking the fluorescence recovery of protamine/PBI-Asp

solution in the presence of trypsin (λex =490 nm, λem =550 nm, Fig. 5a). A gradual

fluorescence increase was observed dependent on the trypsin concentration, with

higher enzyme concentrations generating faster fluorescence increments. The

fluorescence spectra after 3 h incubation with trypsin were also recorded to evaluate

the enzyme activity (Fig. 5b). The limit of detection (LOD) defined as the lowest

assayed concentration of analyte that yields a signal higher than three times the

standard deviation of the background was determined to be 0.051 nM trypsin which

gave a notable fluorescence rise (Fig. 5b and 5c). Importantly, the enzyme activity

was detectable directly with the naked eye, both under sunlight or a UV lamp, which

is very useful for point-of-care enzyme detection.

Figure 6. (a) UV-vis spectra and (b) fluorescence polarization of protamine/PBI-Asp

solution after incubation with trypsin for 3 h. The results of fluorescence polarization

17

indicates the perylene rotation diffusion φ1<φ2. The PBI-Asp concentration is 10 μM

and the protamine concentration is 15 μg/mL.

The enzyme activity could also be measured by UV-vis absorbance and

fluorescence polarization. As shown in Fig. 6a, the absorption peaks at 495 nm and

534 nm that were suppressed by protamine sulfate were restored after enzyme

incubation. A noticeable increase in the UV-vis absorption was observed with a

minimal trypsin concentration of 0.2 nM. The trypsin activity could be also detected

by fluorescence polarization. As shown in Fig. 6b, trypsin-catalyzed protein digestion

caused fluorescence depolarization, which resulted from the enhanced rotational

diffusion of PBI-Asp. Similarly, the value of fluorescence polarization was found to

plateau with 0.2 nM trypsin. To exclude the contribution of non-specific interactions

between trypsin and protamine/PBI-Asp to fluorescence enhancement, control

experiments were conducted by replacing trypsin with different proteins (Fig. S3). No

significant fluorescence increase was noted in the protamine/PBI-Asp system after

incubating with streptavidin, histone H1, lysozyme, cytochrome c, bovine serum

albumin, and human serum albumin. In addition, trypsin cleavage-induced

disassembly of protamine/PBI-Asp complexes was demonstrated by TEM, wherein no

self-assembled particles were observed after trypsin digestion. The drop of light

scattering intensity from 162 to 2.80 kcps after trypsin incubation confirmed the

dissociation of large particles. The z-potential of protamine/PBI-Asp solution after

enzyme treatment was found to be -3.64 mV, also indicating the disassembly of

18

protamine/PBI-Asp particles.

The detection limit of trypsin by this assay can be further lowered to 0.006 nM

when the protamine/PBI-Asp concentration is reduced (Fig. 5c and S4). This

detection limit is much lower than that of commercial kits using FITC-Casein as a

substrate (~0.5 μg/mL). Commercial protease-sensing protocols are based on a dye-

labeled peptide or protein substrate, which means one amide bond cleavage

corresponds to one chromophore or fluorophore release. In our assay, the absorption

and fluorescence signal is quenched by the non-covalent self-assembly of

protein/perylene dye. This means that in principle the enzymatic digestion of an amide

bond is possible to uncage several perylene dyes, which is transduced through

multimodal signals. In addition, the high fluorescence quantum yield, large extinction,

and strong aggregation-induced quenching of the perylene dye also contribute to the

high sensitivity of our assay. Another advantage of this assay is its tunable dynamic

detection range. For example, the dynamic detection range shifted from 0.05-10 nM

to 0.005-2.0 nM when the PBI-Asp concentration is lowered from 10 μM to 1.0 μM

(Fig. 5c).

The development of rapid and simple methods for screening chemical libraries of

potential protease inhibitors is important in the pharmaceutical industry. For this

reason the developed protamine/PBI-Asp self-assembly system was also used to study

the inhibition of trypsin activity by inhibitors. In Fig. S6a, the degree of fluorescence

enhancement for the ensemble of protamine sulfate and PBI-Asp containing trypsin

was retarded by the trypsin inhibitor benzamidine hydrochloride. Based on the plot of

19

the inhibition efficiency vs. inhibitor concentration (Fig. S5), the IC50 value of

benzamidine hydrochloride toward trypsin was estimated to be 5.0 μM (Fig. S6b).

3. Label-free detection of Cathepsin B

Considering the universality of non-covalent self-assembly in protein/perylene dye

systems, we hypothesized that this strategy could be extended to detect different

proteases. Of particular interest are disease-related proteases, such as HIV or cancer

relevant enzymes. To this end, Cathepsin B (CTSB), which is frequently over-

expressed in premalignant lesions, was chosen as a target protease. Increased

expression of Cathepsin B in primary cancers, especially in preneoplastic lesions,

suggests that this enzyme might have pro-apoptotic features. For the sensing of

Cathepsin B, we replaced protamine with polylysine, a highly positive peptide that

has been demonstrated to be a CTSB substrate.29,30 Owing to the strong charge

interactions, electrostatic self-assembly of polylysine/PBI-Asp was expected. Indeed,

upon the addition of polylysine to PBI-Asp, self assembly was observed by UV-vis

absorbance and fluorescence. As before, the absorption peaks at 495 nm and 533 nm

gradually decreased, finally disappearing at 2.0 μg/mL of polylysine (Fig. 7a).

Significant fluorescence quenching (~98%) was observed in the presence of

polylysine (Fig. 7b). It was expected that enzymatic cleavage of polylysine amide

bonds would disrupt the polymer–analogue structure and reduce the positive charges

(via the formation of carboxyl groups), and would consequently trigger the

disassembly of polylysine/PBI-Asp (Fig. 7c). In our experiment, the enzyme-triggered

20

fluorescence recovery in polylysine/PBI-Asp was recorded after 6 h incubation with

CTSB and notable emission increases were observed at enzyme concentrations of 20

mU/mL (Fig. 7d). Unlike previous work, in which fluorophores modified polylysines

have been used for imaging CTSB, here we provide a simple and effective approach

to design a CTSB sensor via a non-covalent strategy without dye labeling.

Figure 7. (a) UV-vis absorbance and (b) fluorescence spectra of polylysine/PBI-Asp

solution (pH 5.0) for polylysine concentrations from 0-2 μg/mL. The inset in Fig. 7b

shows the fluorescence intensity (λex =490 nm, λem =550 nm) of PBI-Asp with

different amount of polylysine. (c) Cleavage of polylysine amide bond by CTSB. (d)

Fluorescence spectra of polylysine/PBI-Asp mixtures after incubating with Cathepsin

B for 6 h at 37 oC for CTSB concentrations of 1.0, 5.0, 10, 20, 50, and 200 mU/mL,

21

respectively. The inset in Fig. 7d gives the fluorescence intensity of polylysine/PBI-

Asp for PBI-Asp concentration of 1.0 μM and polylysine concentration of 1.0 μg/mL

after incubation with CTSB (λex =490 nm, λem =550 nm).

CONCLUSIONS

In this work, we have designed a versatile biosensing platform for label-free protease

detection through the non-covalent self-assembly of proteins/perylene dye. The

driving forces of protein/dye self-assembly were ascribed to electrostatic attractions

and π-π interactions, which led to the formation of activatable soft nanoparticles and

significant colorimetric and fluorometric responses. The presence of protease

specifically hydrolyzed the protein, resulting in the disassembly of supramolecular

nanoparticles as detectable by multiple optical readouts (i.e., UV-vis absorbance,

fluorescence, and depolarization). This assay was exploited for sensitively detecting

trypsin and Cathepsin B by using protamine and polylysine respectively as substrates.

Because of its inherent modularity and dynamic nature, the developed non-covalent

biosensing platform is highly sensitive, and can be further explored for the detection

of other proteases. We anticipate that the presented concept of integrating natural

proteins or polypeptides with synthetic dyes via supramolecular self-assembly will

open up a new avenue towards multifunctional biomaterials for smart nanocarriers

and biosensors.

22

AUTHOR INFORMATION

Corresponding Author

*E-mail: m.stevens@imperial.ac.uk. Phone: +44 (0)20 7584 6804.Notes

The authors declare no competing financial interest

ACKNOWLEDGMENTS

M.M.S. thanks the Engineering and Physical Sciences Research Council (EPSRC)

grant EP/K020641/1. Thanks to Dr. Roberto de la Rica for help in preparing this

manuscript.

ASSOCIATED CONTENT

Supporting Information

Measurement of fluorescence quantum yield, TEM image, effects of different proteins

to protamine/PBI-Asp fluorescence, comparison of GRP and protamine, and trypsin

inhibitor test. This material is available free of charge via the Internet at

http://pubs.acs.org.

REFERENCES

(1) Lopez-Otin, C.; Overall, C. M. Nat. Rev. Mol. Cell Biol. 2002, 3, 509-519.

(2) Mason, S. D.; Joyce, J. A. Trends Cell Biol. 2011, 21, 228-237.

(3) Neurath, H. J. Cell. Biochem. 1986, 32, 35-49.

(4) Lee, S.; Cha, E. J.; Park, K.; Lee, S. Y.; Hong, J. K.; Sun, I. C.; Kim, S. Y.; Choi, K.;

Kwon, I. C.; Kim, K.; Ahn, C. H. Angew. Chem. Int. Ed. 2008, 47, 2804-2807.

(5) Laack, E.; Kohler, A.; Kugler, C.; Dierlamm, T.; Knuffmann, C.; Vohwinkel, G.;

23

Niestroy, A.; Dahlmann, N.; Peters, A.; Berger, J.; Fiedler, W.; Hossfeld, D. K. Ann. Oncol.

2002, 13, 1550-1557.

(6) Schilling, O.; Overall, C. M. Nat. Biotechnol. 2008, 26, 685-694.

(7) Thomas, D. A.; Francis, P.; Smith, C.; Ratcliffe, S.; Ede, N. J.; Kay, C.; Wayne, G.;

Martin, S. L.; Moore, K.; Amour, A.; Hooper, N. M. Proteomics 2006, 6, 2112-2120.

(8) Matayoshi, E. D.; Wang, G. T.; Krafft, G. A.; Erickson, J. Science 1990, 247, 954-

958.

(9) Medintz, I. L.; Mattoussi, H. Phys. Chem. Chem. Phys. 2009, 11, 17-45.

(10) Lowe, S. B.; Dick, J. A. G.; Cohen, B. E.; Stevens, M. M. ACS Nano 2012, 6, 851-

857.

(11) Shi, L.; De Paoli, V.; Rosenzweig, N.; Rosenzweig, Z. J. Am. Chem. Soc. 2006, 128,

10378-10379.

(12) Yao, H.; Zhang, Y.; Xiao, F.; Xia, Z.; Rao, J. Angew. Chem. Int. Ed. 2007, 46, 4346-

4349.

(13) Rajagopal, K.; Schneider, J. P. Curr. Opin. Struct. Biol. 2004, 14, 480-486.

(14) Huang, C.; Barlow, S.; Marder, S. R. J. Org. Chem. 2011, 76, 2386-2407.

(15) Wuerthner, F.; Stolte, M. Chem. Commun. 2011, 47, 5109-5115.

(16) Goerl, D.; Zhang, X.; Wuerthner, F. Angew. Chem. Int. Ed. 2012, 51, 6328-6348.

(17) Wang, B.; Yu, C. Angew. Chem. Int. Ed. 2010, 49, 1485-1488.

(18) Tang, D.; Liao, D.; Zhu, Q.; Wang, F.; Jiao, H.; Zhang, Y.; Yu, C. Chem. Commun.

2011, 47, 5485-5487.

(19) Soh, N.; Ariyoshi, T.; Fukaminato, T.; Nakajima, H.; Nakano, K.; Imato, T. Org.

Biomol. Chem. 2007, 5, 3762-3768.

(20) Yang, X.; Pu, F.; Ren, J.; Qu, X. Chem. Commun. 2011, 47, 8133-8135.

(21) Liu, H.; Wang, Y.; Liu, C.; Li, H.; Gao, B.; Zhang, L.; Bo, F.; Bai, Q.; Ba, X. J.

Mater. Chem. 2012, 22, 6176-6181.

(22) Gao, B.; Xia, D.; Zhang, L.; Bai, Q.; Bai, L.; Yang, T.; Ba, X. J. Mater. Chem. 2011,

21, 15975-15980.

(23) Li, A. D. Q.; Wang, W.; Wang, L. Q. Chem.-Eur. J. 2003, 9, 4594-4601.

(24) Sorgi, F. L.; Bhattacharya, S.; Huang, L. Gene Ther. 1997, 4, 961-968.

24

(25) Ma, T.; Li, C.; Shi, G. Langmuir 2008, 24, 43-48.

(26) Jameson, D. M.; Ross, J. A. Chem. Rev. 2010, 110, 2685-2708.

(27) Byrne, M. F.; Mitchell, R. M.; Stiffler, H.; Jowell, P. S.; Branch, M. S.; Pappas, T.

N.; Tyler, D.; Baillie, J. Can. J. Gastroenterol. 2002, 16, 849-854.

(28) Hirota, M.; Ohmuraya, M.; Baba, H. J. Gastroenterol. 2006, 41, 832-836.

(29) Weissleder, R.; Tung, C. H.; Mahmood, U.; Bogdanov, A. Nat. Biotechnol. 1999, 17,

375-378.

(30) Wunder, A.; Tung, C. H.; Muller-Ladner, U.; Weissleder, R.; Mahmood, U. Arthritis

Rheum. 2004, 50, 2459-2465.

25

for TOC only

26