Highly Adaptable and Sensitive Protease Assay Based on

Sep 5, 2011 - ... and can be easily adapted to act as a versatile tool for the sensitive detection of proteases. ... Seongwon Seo , Jongho Kim , Geuns...
0 downloads 0 Views 3MB Size
Download PDF
ARTICLE pubs.acs.org/ac

Highly Adaptable and Sensitive Protease Assay Based on Fluorescence Resonance Energy Transfer Thomas Zauner, Renate Berger-Hoffmann, Katrin M€uller, Ralf Hoffmann, and Thole Zuchner* Institute of Bioanalytical Chemistry, Faculty of Chemistry and Mineralogy, Center for Biotechnology and Biomedicine, Leipzig University, Deutscher Platz 5, 04103 Leipzig, Germany

bS Supporting Information ABSTRACT: Proteases are widely used in analytical sciences and play a central role in several widespread diseases. Thus, there is an immense need for highly adaptable and sensitive assays for the detection and monitoring of various proteolytic enzymes. We established a simple protease fluorescence resonance energy transfer (pro-FRET) assay for the determination of protease activities, which could in principle be adapted for the detection of all proteases. As proof of principle, we demonstrated the potential of our method using trypsin and enteropeptidase in complex biological mixtures. Briefly, the assay is based on the cleavage of a FRET peptide substrate, which results in a dramatic increase of the donor fluorescence. The assay was highly sensitive and fast for both proteases. The detection limits for trypsin and enteropeptidase in Escherichia coli lysate were 100 and 10 amol, respectively. The improved sensitivity for enteropeptidase was due to the application of an enzyme cascade, which leads to signal amplification. The pro-FRET assay is highly specific as even high concentrations of other proteases did not result in significant background signals. In conclusion, this sensitive and simple assay can be performed in complex biological mixtures and can be easily adapted to act as a versatile tool for the sensitive detection of proteases.

P

roteases (or peptidases) are enzymes which catalyze the hydrolysis of peptide bonds in proteins and are widely used in analytical sciences today.1 3 They have been extensively studied since the 19th century,4 and bioinformatic data reveal that genes encoding proteases represent as much as 2% of the human genome.1,5 Furthermore, proteases have been identified in almost every organism, and they play a key role in many biological pathways.4,6 Misregulated proteolytic enzymes are associated with several important and widespread diseases including cancer, degenerative diseases, coagulopathies, inflammation, and infectious diseases.4,7 For example, it has been shown that the overexpression of kallikrein-related peptidases (KLK) results in ovarian cancer, a leading cause of death among the female population.8 Another protease-related disease is rheumatoid arthritis, which leads to the destruction of bone and cartilage.9 It has been reported that metalloproteases and cysteine proteases are involved in this inflammatory disease.9 Due to their central role in analytical sciences and important biological functions, proteolytic enzymes are a major target for the development of simple and sensitive analytical tools.4,5,10,11 Several protease detection techniques exist today.12 One of the most common ways to detect protease activity is based on highly specific protease substrates (chromophores or fluorophores) which upon cleavage by the protease of interest lead to a detectable signal. Although the majority of tests are not very sensitive (picomole to femtomole range),13 a few highly specialized substrates exist which are suitable for sensitive r 2011 American Chemical Society

detection.12,14 17 It has been reported that time-resolved fluorescence resonance energy transfer (FRET) approaches using europium chelate-modified substrates are capable of delivering moderate sensitivities for protease detection.18,19 However, these substrates are highly specialized for a single protease and new substrates must be developed for every new protease of interest.20 Other detection methods which are compatible for the detection of proteases exist, including gel electrophoresis and even mass spectrometry.21 However, neither method can be easily applied to the quantification of protease activities, and they require either expensive instrumentation or show a lack of sensitivity. Enzyme-linked immunosorbent assay (ELISA)-based approaches offer a high flexibility but are also restricted in their detection limit and therefore not suitable for detecting enzymes.22,23 Concluding the above, a universal assay for measuring protease activities does not exist today. Such an assay would need to be highly sensitive, as the biomaterial available for medical diagnostics is generally limited and differences between the normal physiological and pathological concentration of a protease are often minute. Furthermore, biological samples are usually present as complex mixtures, which complicates the analysis dramatically.24 27 Received: May 19, 2011 Accepted: September 5, 2011 Published: September 05, 2011 7356

dx.doi.org/10.1021/ac201274f | Anal. Chem. 2011, 83, 7356–7363

Analytical Chemistry

ARTICLE

and B). For both enzymes, the assay was performed in buffer and in Escherichia coli lysate (E. coli lysate) and exhibited excellent limits of detection (LOD) as well as fast detection times. Importantly, the principle of our pro-FRET assay can be easily adapted to any other protease by using an appropriate peptide, which makes it an ideal and highly versatile tool for the early detection of any protease-related disease.

’ EXPERIMENTAL SECTION

Figure 1. Assay principle based on FRET and used for determination of protease activity (D = donor, A = acceptor, EP = enteropeptidase, Tinact. = trypsinogen, T = trypsin); cleavage of the FRET peptide leads to an increase of donor emission. The digest of the FRET substrate can be initiated by (A) direct addition of trypsin or (B) by applying an enzyme cascade which involves the conversion of inactive trypsinogen into trypsin by enteropeptidase.

Trypsin has been intensively studied due its high availability from animal pancreas extractions. It belongs to the serine protease family S1 and plays a major role in food digestion. Furthermore, it acts as a trigger enzyme and activates all other zymogens, including its own precursor trypsinogen.15,17 Trypsinogen is also activated by enteropeptidase to form trypsin.14,28 In particular, autoactivation is increased at higher pH (up to pH 9) and in the presence of Ca2+, due to the formation of a highly active calcium complex.15 Enteropeptidase is a highly specific serine protease of the intestinal brush border, converting trypsinogen into active trypsin by cleavage after the N-terminal peptide sequence DDDDK.6,14,28 Congenital deficiency of enteropeptidase results in fatal intestinal malabsorption.14 Furthermore, the protein shows high activity over a wide pH and temperature range.28 In the field of biotechnology this enzyme is used for the highly specific cleavage of fusion proteins.14 Here we present a highly sensitive and simple assay based on FRET to determine the activity of proteolytic enzymes (proFRET assay). The principle of our approach is based on a FRET peptide substrate. The substrate is labeled with two common and inexpensive fluorophores (fluorescein/TAMRA) and is cleaved by the protease of interest between the donor/acceptor position. Upon cleavage the FRET is abolished due to the larger donor acceptor distance, thereby increasing the donor fluorescence intensity (Figure 1A). In a second approach, the assay signal was amplified by an enzyme cascade to further improve the sensitivity 29,30 (Figure 1B). The FRET peptides were optimized for a fast and specific digest, to provide a strong signal response. As proof of principle, we determined the protease activity for trypsin as a specific protease in terms of a single residue (K, R) and enteropeptidase as representative for a protease, which is specific regarding a sequence (DDDDK) (Figure 1, parts A

Chemicals. All fluorenylmethoxycarbonyl (Fmoc) amino acid derivatives were purchased from Orpegen Pharma (Heidelberg, Germany). Polystyrene-based Rink amide (MBHA) resin (loading capacity 0.67 mmol/g) was from MultiSynTech GmbH (Witten, Germany). Acetonitrile (HPLC grade) and diethyl ether were obtained from VWR (Darmstadt, Germany). N,N-Dimethylformamide (DMF) and piperidine were from Biosolve (Valkenswaard, The Netherlands), and dichloromethane (DCM) was from Roth (Karlsruhe, Germany). 5(6)-Carboxytetramethylrhodamine (TAMRA), diisopropylcarbodiimide (DIC), trifluoroacetic acid (TFA), thioanisole, m-cresol, 1,2-ethanedithiol, and imidazole were purchased from Fluka (Steinheim, Germany). 5-Iodoacetamidofluorescein, bovine pancreatic trypsin, bovine pancreatic trypsinogen, porcine enteropeptidase, bovine thrombin, bovine α-chymotrypsin and α-chymotrypsinogen type II, 1-hydroxybenzotriazole (HOBt), and trifluoroacetic acid (TFA; g 99% GC for UV spectroscopy) were obtained from SigmaAldrich (Steinheim, Germany). Water was purified in house with a PURELAB ultraanalytic system (Elga, Berkefeld GmbH, Ransbach, Germany, 18.2 MΩ cm). Protease activity was determined in assay buffer, which consisted of Tris HCl (AppliChem, Darmstadt, Germany; 100 mmol/L, pH 8) and calcium chloride (Fluka; 0.5 mmol/L). Tween 20 was purchased from Serva (Heidelberg, Germany). All chemicals were obtained in the highest purity available. Synthesis of Peptide Substrates. All peptides were synthesized on MBHA resin using Fmoc/tBu-chemistry and a Syro2000 multiple peptide synthesizer (MultiSynTech) on a 25 μmol scale. Coupling cycles consisted of the appropriate Fmoc-amino acid (8 equiv), HOBt (8 eq, 0.5 mol/L in DMF), and DIC (8 equiv) in DMF. After complete synthesis, the N-terminal Fmoc group was cleaved with piperidine (40% in DMF, v/v) and the peptide resin was washed six times with DMF and DCM. TAMRA (4 equiv) was activated with HOBt (200 μL, 0.5 mol/L, 4 equiv) and DIC (15 μL, 4 equiv) in DMF in the dark at RT. After 4 h the peptide resin was washed six times with DMF and DCM and dried. Peptides were cleaved from the resin (in the dark, RT, shaking) using a cleavage solution consisting of 87.5% (v/v) TFA, 3.5% (v/v) water, 3.5% (v/v) thioanisole, 3.5% (v/v) m-cresol, and 2% (v/v) 1,2-ethanedithiol. After 4 h the peptide was precipitated with cold diethyl ether, washed twice with diethyl ether, air-dried, and dissolved in 0.1% (v/v) aqueous TFA for purification. The crude peptide mixture was purified € kta purifier high-performance liquid chromatograph using an A (HPLC) (GE Healthcare GmbH, Freiburg, Germany) with a Jupiter C18-column (internal diameter (i.d.) 21.2 mm, length 250 mm, particle size 15 μm, pore size 30 nm, Phenomenex Inc., Aschaffenburg, Germany) applying a linear aqueous acetonitrile gradient (0.6% acetonitrile per minute) in the presence of 0.1% (v/v) TFA and a flow rate of 10 mL/min. Purified peptides were characterized by matrix-assisted laser desorption ionization mass 7357

dx.doi.org/10.1021/ac201274f |Anal. Chem. 2011, 83, 7356–7363

Analytical Chemistry spectrometry (MALDI-MS) using a 4700 Proteomics Analyzer (Applied Biosystems, Darmstadt, Germany). Three equivalents of 5-iodoacetamidofluorescein were dissolved in a 1:1 mixture (v/v) of DMF in PBS buffer (500 μL, pH 7.4), added to the lyophilized peptides, and incubated overnight at 40 °C in the dark. The reaction mixture was diluted 10-fold in 0.1% (v/v) aqueous TFA (g99% GC for UV spectroscopy) and € kta purifier system (GE Healthpurified by RP-HPLC on an A care GmbH) using a Jupiter C18-column (i.d. 4.6 mm, length 150 mm, particle size 5 μm, pore size 30 nm, Phenomenex Inc.) and a linear acetonitrile gradient (1.1% acetonitrile per minute) in the presence of 0.1% TFA and a flow rate of 1 mL/min. The peptide identity and purity were determined by MALDI-MS. The concentration of the lyophilized peptides was estimated by aligning the extinction at 490 nm with a standard curve of pure fluorescein in a mixture of DMF (35%, v/v) and imidazole buffer (50 mmol/L, pH 8). The peptides were aliquoted to a final peptide concentration of 0.1 mmol/L in imidazole buffer (50 mmol/L, pH 8) containing 1% (v/v) DMF to improve the solubility and stored at 20 °C. Tryptic Digestion of the FRET Peptides. Tryptic digests were performed in assay buffer (100 mmol/L Tris buffer, 0.5 mmol/L CaCl2, pH 8) in a total volume of 50 μL per well at 37 °C. The peptide concentration per well was 2 μmol/L. Before the measurement was started, the microtiter plate (black polypropylene 384 well plates, Greiner, Frickenhausen, Germany) was blocked with 1% Tween 20 (v/v) in assay buffer (1 h, RT) and washed three times with assay buffer. The peptide solution was added into the well, and the digest was initiated by the addition of trypsin (10 pmol). The increase of the fluorescence (λex = 485 ( 10 nm, λem = 535 ( 12.5 nm) was monitored using a Paradigm detection platform (Molecular Devices, Ismaning, Germany). The maximum quenching rate QRmax was determined for each FRET peptide. QRmax was defined as the average percentage fluorescence increase per min. QRmax was chosen as a measurement parameter because it takes not only the signal increase into account but also how fast these signals are generated. The quenching efficiency E is defined as the percentage of the original fluorescence intensity quenched by the system, where 100% would equal a complete quenching. To confirm the complete cleavage of the FRET substrates, peptides 7, 8, 11, and 12 were incubated with 10 pmol of trypsin as described above and subjected to an HPLC analysis to confirm complete cleavage. In detail, the peptides were digested with 10 pmol of trypsin for 0, 15, or 45 min and afterward analyzed on a System Gold HPLC system (Molecular Devices Ismaning, Germany) with a Jupiter C18-column (i.d. 2 mm, length 150 mm, particle size 5 μm, pore size 30 nm, Phenomenex Inc.) and a linear acetonitrile gradient (2.7% acetonitrile per minute) in the presence of 0.1% TFA and a flow rate of 0.2 mL/min. As a control 10 pmol of trypsin was incubated under the same conditions as described above for 15 and 45 min, and the mixture was also used for HPLC characterization using the above-described conditions. After the confirmation of the complete cleavage of the peptides, the highest possible fluorescence signal was determined. For this, the fluorescence increase of the same peptide digest conditions was monitored for 45 min (λex = 485 ( 10 nm, λem = 535 ( 12.5 nm) using the Paradigm detection platform (Molecular Devices Ismaning, Germany). LOD Determination of Trypsin and Enteropeptidase in Tris Buffer and in E. coli Lysate. The experiments were

ARTICLE

performed in assay buffer using black polypropylene 384 well plates according to the procedure described above. Protease activities were determined in buffer or E. coli lysate (produced inhouse from E. coli BL21AI and tested for protease activity according to Knappe et al.).31 In all experiments the increase of the fluorescence (λex = 490 nm, λem = 514 nm) was monitored using a Paradigm detection platform (Molecular Devices) due to FRET peptide cleavage. In order to determine the trypsin LOD, the peptide solutions were placed in the wells and spiked with trypsin in a dilution series ranging from 10 pmol to 1 amol. As a negative control, the peptide was incubated in Tris buffer or lysate. Enteropeptidase was determined by testing a dilution series of enteropeptidase ranging from 10 fmol to 100 zmol. The assay was performed as follows: peptide solutions were spiked with the different amounts of enteropeptidase and the experiment was initiated by the addition of 10 pmol of trypsinogen per well. In order to determine the background signal (negative control) the peptide was incubated with 10 pmol of trypsinogen in absence of enteropeptidase. For the determination of assay specificity, the peptide solution in buffer (in case of the specificity regarding trypsin) or the peptide solution in buffer and 10 pmol trypsinogen (in case of the specificity of enteropeptidase) was spiked with control proteases (chymotrypsin, chymotrypsinogen, or thrombin) instead of the highest tested amount of trypsin (10 pmol) or enteropeptidase (10 fmol) and the fluorescence increase was measured. Additionally, the peptide solution was incubated with 10 fmol of enteropeptidase in the absence of trypsinogen to test the stability of the substrates against enteropeptidase. To determine the LOD, QRmax of each protease concentration together with the peptide was compared with QRmax of the respective negative control (only peptide for trypsin or peptide and trypsinogen in case of enteropeptidase). All experiments were performed with three to five replicates. The LOD was defined as the minimum signal still exceeding the negative control plus 3 times the standard deviation. The limit of quantitation (LOQ) was defined as the concentration in the linear range of the assay which was at least 10 times the standard deviation of the blank plus the blank signal or the lowest concentration. To determine outliers, three independent tests (Dixon, Nalimov, and Hampel)32 34 were used, each providing a 95% safety. Only values which were confirmed as outliers by all three tests were excluded from the analysis.

’ RESULTS For the protease assay, a FRET-based approach was chosen, which should in principle provide high sensitivities and specificities (Figure 1, parts A and B), and was optimized for trypsin and enteropeptidase. According to the assay design (Figure 1A), the digestion of the FRET peptide is directly related to the activity and concentration of the protease (trypsin). For determination of the second protease (enteropeptidase), an enzyme cascade involving the activation of trypsinogen into trypsin by enteropeptidase was chosen (Figure 1B). This system enables signal amplification, which should ideally result in an improved sensitivity.29,30 The key factor in achieving good signal responses and fast assay times arises from the substrate, which should (i) show a high FRET efficiency, (ii) be well-suited for the protease of interest, and (iii) not be cleaved by other proteases present in the biological sample. Thus, 12 different FRET 7358

dx.doi.org/10.1021/ac201274f |Anal. Chem. 2011, 83, 7356–7363

Analytical Chemistry

ARTICLE

Figure 2. Maximum quenching rates QRmax for trypsin in buffer using peptide 8 after 5.5 h of assay time (nPeptide = 100 pmol, bars represent the standard error): (A) values for the maximum quenching rates QRmax of different trypsin amounts (10 pmol to 1 amol) and the control (100 pmol of peptide) (dotted line corresponds to the signal of the background and 3σ); (B) linear range of 1.7 orders of magnitude from 10 to 500 fmol with R2 = 0.9954; (C) assay specificity, QRmax of the control proteases (10 pmol) in comparison to trypsin (10 pmol). n = 3.

peptides (Table S1, Supporting Information) were synthesized and tested. The peptides differed in their donor acceptor distance as well as in the position of the protease cleavage site within the peptide sequence. We used a well-characterized FRET pair, with fluorescein as donor and the rhodamine-based dye TAMRA as acceptor, which provide a good spectral overlap.20,35,36 Synthesis and Application of the FRET Peptides in a Tryptic Digest. TAMRA was coupled to the N-terminus of the peptides and fluorescein to the thiol group of the C-terminal cysteine (Table S1, Supporting Information), as confirmed by RP-HPLC and MALDI-MS. A representative chromatogram and mass spectrum of the optimal peptide substrate (peptide 7) are shown in Figure S1 (Supporting Information), indicating a purity of more than 95%. The two signals in the chromatogram are due to the fact that the TAMRA was used as a mixture of two different positional isomers (5- and 6-position). The chromatographic peak areas regarding the two isomers differ among the four substrates due to a purification step to obtain a high purity (Figures S1 S4, Supporting Information). However, we presume that the two isomers do not influence the assay. The broad signal at approximately 4 min corresponds to DMF (Figure S1A, Supporting Information). Furthermore, the characterization of peptides 8, 11, and 12 is provided (Figures S2 S4, Supporting Information). The purified FRET peptides were incubated with trypsin to determine the optimal substrates (Figure S5A, Supporting Information). Peptides 7, 8, 11, and 12 showed the highest quenching rate QR max , thereby providing a good signal response and short assay time (Figure S5A, Supporting Information). The quenching efficiency E was >95% for peptide 11 and even more than 97% for peptides 7, 8, and 12 (Figure S5B, Supporting Information). These four substrates reached the maximum signal within 15 min. The quenching efficiencies for all other peptides were between 33% and 93% (Figure S6, Supporting Information). Hence, the optimal four peptides (peptides 7, 8, 11, and 12) were tested in further experiments to determine the protease activity for trypsin and enteropeptidase.

Furthermore, the incubation of peptides 7, 8, 11, and 12 with trypsin resulted in a complete cleavage after 15 min for all four substrates (Figure S7, Supporting Information). This is indicated by the fact that the plateau of the fluorescence increase is reached after already 15 min for each peptide (Figure S7A, Supporting Information). Additionally, the HPLC analysis at different time points of the tryptic digest confirmed the complete cleavage of the peptides as no signal of the nondigested peptide substrate was observable after an incubation time of 15 and 45 min (Figure S7B F, Supporting Information). Determination of Protease Activity of Trypsin and Enteropeptidase in Buffer. The assays were characterized in terms of LOD and assay time for the determination of protease activity. This included the measurement of various enzyme amounts ranging from 10 pmol (200 nmol/L) to 100 zmol (2 fmol/L). Trypsin cleavage of peptide 8 in buffer yielded an LOD of 100 amol (2 pmol/L) after 5.5 h (Figure 2A). Furthermore, a linear range of about factor 50 was observed from 0.01 to 0.5 pmol (0.2 10 nmol/L) (Figure 2B). In order to test the specificity of the assay three control proteases were chosen, namely, chymotrypsin, which contains a similar active site to trypsin, the trypsin-like protease thrombin, and chymotrypsinogen as a precursor protease.37 39 These proteases showed a minor influence regarding the assay, even when tested at a concentration of 10 pmol (equivalent to the highest concentration of trypsin used) (Figure 2C). This indicated a high assay specificity. Additionally, the average intraday variance of the assay within the range of 10 pmol to 100 amol (200 nmol/L to 2 pmol/L) was 4% (Table S2, Supporting Information). The average interday variance for the same range was 19% (Table S3, Supporting Information). As a second enzyme, enteropeptidase was chosen to demonstrate the general applicability of the assay for a cascade system. Therefore, different dilutions of enteropeptidase (10 fmol to 100 zmol [200 pmol/L to 2 fmol/L]) were tested, and the reaction was started by the addition of 10 pmol trypsinogen. In buffer, a very low LOD of 10 amol (200 fmol/L) after 2.5 h was obtained for peptide 7 (Figure 3A). Furthermore, a linear range of 1.7 orders of magnitude was observed (from 500 to 10 000 amol 7359

dx.doi.org/10.1021/ac201274f |Anal. Chem. 2011, 83, 7356–7363

Analytical Chemistry

ARTICLE

Figure 3. Maximum quenching rates QRmax for enteropeptidase in buffer using peptide 7 and the enzyme cascade after 2.5 h of assay time (nPeptide = 100 pmol, nTrypsinogen = 10 pmol, bars represent the standard error): (A) values for the maximum quenching rates QRmax of different enteropeptidase amounts (10 fmol to 100 zmol) and the control (100 pmol of peptide and 10 pmol of trypsinogen) (dotted line corresponds to the signal of the background and 3σ); (B) linear range of 1.7 orders of magnitude from 500 to 10 000 amol with R2 = 0.9894; (C) assay specificity, QRmax of the control proteases (nControlprotease = 10 fmol and nTrypsinogen = 10 pmol) and enteropeptidase (10 fmol) in the absence of trypsinogen in comparison to the cascade (nTrypsinogen = 10 pmol, nEnteropeptidase = 10 fmol), control (100 pmol of peptide and 10 of pmol trypsinogen). n = 3.

Table 1. Summary of the Protease Detection Limits in Buffer and in E. coli Lysate Using Peptides 7, 8, 11, and 12 trypsin

sample experiment in buffer

experiment in E. coli lysate

peptide peptide 7 peptide 8

LOD (fmol) 10 0.1

enteropeptidase

assay time (min)

LOD (fmol)

assay time (min)

300

0.01

155

330

0.1

480

peptide 11

10

150

1

330

peptide 12

10

40

1

105

15

0.01

180

120

0.01

270

peptide 7

0.1

peptide 8

1

[10 200 pmol/L]) (Figure 3B). A good assay specificity was observed, as the incubation of the peptide solution and trypsinogen together with the control proteases (chymotrypsin, chymotrypsinogen, thrombin, and enteropeptidase alone) resulted in weak signal responses, even at the highest enteropeptidase concentration (10 fmol) tested (Figure 3C). Furthermore, the average intraday variance for the activity of the enteropeptidase within the range of 10 fmol to 10 amol (200 pmol/L to 200 fmol/L) was 11%, and the average interday variance for the same range was 18% (Tables S4, and S5, Supporting Information). When protease activities were determined in buffer, peptide 8 provided the best LOD in combination with a fast assay time for trypsin determination and peptide 7 showed the best LOD for the determination of enteropeptidase (Table 1). Determination of Protease Activity of Trypsin and Enteropeptidase in E. coli Lysate. To test the compatibility of the assay system with complex biological matrixes, experiments using the optimal substrate peptides 7 and 8 were performed in E. coli lysate (concentration of 0.1 mg/mL). For trypsin the peptide solutions in E. coli lysate were spiked with a dilution series of

Figure 4. Maximum quenching rates QRmax for trypsin in E. coli lysate using peptide 7 after 15 min of assay time (nPeptide = 100 pmol, cLysate = 0.1 mg/mL, bars represent the standard error): (A) values for the maximum quenching rates QRmax of different trypsin amounts (10 pmol to 1 amol) and the control (100 pmol of peptide in E. coli lysate) (dotted line corresponds to the signal of the background and 3σ); (B) linear range of 2 orders of magnitude from 100 to 10 000 fmol with R2 = 0.9850. n = 3.

10 pmol to 1 amol (200 nmol/L to 20 fmol/L) of protease, and the fluorescence increase was monitored. The LOD for trypsin was determined as described above and resulted in an improved LOD of 100 amol (2 pmol/L) after only 15 min (Figure 4A). A linear range over 2 orders of magnitude from 0.1 to 10 pmol (2 200 nmol/L) could be observed (Figure 4B). 7360

dx.doi.org/10.1021/ac201274f |Anal. Chem. 2011, 83, 7356–7363

Analytical Chemistry

ARTICLE

Figure 5. Maximum quenching rates QRmax for enteropeptidase in E. coli lysate using peptide 7 after 3 h of assay time (nPeptide = 100 pmol, nTrypsinogen = 10 pmol, cLysate = 0.1 mg/mL, bars represent the standard error): (A and B) values for the maximum quenching rates QRmax of different enteropeptidase amounts (10 fmol to 1 amol) and the control (100 pmol of peptide and 10 pmol of trypsinogen in E. coli lysate) (dotted line corresponds to the signal of the background and 3σ); (C) linear range of 1.7 orders of magnitude from 100 to 5000 amol with R2 = 0.99998. n = 3.

The LOD for enteropeptidase in E. coli lysate was determined as 10 amol (0.2 pmol/L) after an assay time of 3 h (Figure 5, parts A and B). Additionally, a linear range over 1.7 orders of magnitude from 100 to 5000 amol (2 100 pmol/L) was obtained (Figure 5C). The results in the complex biological mixture are summarized in Table 1 and demonstrate high sensitivities and fast assay times for both substrates. Peptide 7 was found to be the best substrate for both enzymes, leading to improved LODs and faster assay times (Table 1). Thus, peptide 7 is a powerful substrate for the determination of protease activity of both trypsin and enteropeptidase in buffer as well as in a complex biological mixture such as E. coli lysate.

’ DISCUSSION Due to the important role of proteases in analytical chemistry as well as in several widespread diseases, there is an immense need for highly adaptable and sensitive assays to detect protease activities.4,5,10,11 We have established a simple FRET-based approach, which can in principle be applied to the detection of every existing protease. This approach led to improved LOD values compared to the literature and fast assay times for trypsin and enteropeptidase in buffer as well as in E. coli lysate. Four different FRET substrates (peptides 7, 8, 11, and 12) were rapidly cleaved by trypsin and led to an increase in fluorescence signal. All four substrates provided access for the protease to the cleavage site of the substrate. At the same time, the donor/acceptor distance within these substrates was small enough to ensure efficient energy transfer and therefore high signal-to-noise ratios (Figure S5, Supporting Information). These four peptides were used in the assay to determine protease activity and showed good responses. However, some differences in both the LOD and the required assay time were observed between E. coli lysate and Tris buffer (Table 1). Numerous factors could be responsible for the observed differences. Among them, possible interactions of the peptide substrates with the

biological matrix or dynamic quenching of the substrates could play a role. Furthermore, we cannot exclude the possibility that components present in the E. coli lysate may act as potential inhibitors of the trypsinogen autoactivation and therefore contribute to the lower background signal observed in the E. coli enteropeptidase assay. However, the observed interday variances are relatively high, meaning that a standard curve has to be obtained with each assay performed. Among the four substrates tested, peptide 7 was found to be optimal for the reported assay and resulted in a good LOD. Our reported linear ranges cover 1.7 2 orders of magnitude. Radioimmunoassays or approaches based on chromophoric substrates like α-N-benzoyl-DL-arginine-p-nitroanilide (BAPNA) are capable of detecting 20 pmol/L of trypsin in 5 60 min in complex biological mixtures such as serum.40,41 Lefkowitz et al. have described a polyanionic focusing gel electrophoresis-based technique for detection of 10 pmol/L trypsin after 70 min in whole human blood.42 Also, the detection of trypsin using a timeresolved fluorescence assay has been described before and resulted in an LOD of 1 ng trypsin (ca. 40 fmol).18 Our method with a detection limit of 2 pmol/L (100 amol) after a 15 min assay time using peptide 7 in E. coli lysate represents a 5 10-fold improvement compared to those assays and retains a very fast assay time (Figure 4). Other commercially available fluorescence-based protease assays are less sensitive (3 pmol) and require very long incubation times (up to 24 h).13 However, the applicability of our approach in serum samples needs to be further evaluated, and the direct comparison with the corresponding literature is therefore not possible. The determination of the protease activity of enteropeptidase involved the application of an enzyme cascade. This cascade system has been known for decades43 and has found biotechnological use in an ELISA-based detection system for α-fetoprotein (AFP).22 However, the reported LOD values have been rather high because of the substrates used. Thus, we hoped to optimize the signal amplification by the introduction of novel, 7361

dx.doi.org/10.1021/ac201274f |Anal. Chem. 2011, 83, 7356–7363

Analytical Chemistry FRET-based substrates. For the detection of enteropeptidase, specialized methods like the release of radiolabeled activation peptides or fluorophore-conjugated polymer-based superquenching exist. 25,44 These methods resulted in LODs of 27 pmol/L (380 amol) in buffer and 100 pmol/L (1 fmol) in serum, respectively, with an assay time of 15 90 min.25,44 Interestingly, with the use of the pro-FRET-assay and peptide 7, our novel technique offers a more than 10-fold improvement of the LOD in buffer (200 fmol/L = 10 amol, Figure 3) and more than 100-fold improvement in the complex biological mixture (200 fmol/L = 10 amol, Figure 5). Additionally, our technique is highly specific as the control proteases did not show any significant signal upon incubation with the substrates (Figures 2 and 3). However, as for the trypsin assay, the applicability of our approach in serum samples needs to be further evaluated, and the direct comparison with the corresponding literature is therefore not suitable. Furthermore, we want to use the cascade system to the detection of other proteases using appropriately modified peptide sequences. One target of a potential protease cascade would be the kallikrein-related renin angiotensin system, which is involved in the modulation of vascular contractility and injury response.45 Here, kallikrein activates prorenin into renin, which then converts angiotensinogen into angiotensin.45,46 Another possible example for an important cascade system occurs within the coagulation process, namely, the thrombin-catalyzed conversion of fibrinogen into insoluble fibrin.47,48 Thrombin was previously generated from prothrombin by the prothrombinase complex.49 In summary, we have established a highly sensitive FRETbased assay to determine protease activities. The potential of this method was demonstrated for trypsin and enteropeptidase. We obtained a 40 100-fold improvement in the LOD for enteropeptidase in E. coli lysate, which shows its high potential in complex biological mixtures. Additionally, this approach can be easily adapted for the detection of other proteases by slight changes of the peptide substrate cleavage site.

’ ASSOCIATED CONTENT

bS

Supporting Information. RP-HPLC chromatograms and MALDI mass spectra of peptides 7, 8, 11, and 12; peptide sequences, QRmax values, and the quenching efficiency E for peptides 1 12; intraday and interday data for the assay in buffer. This material is available free of charge via the Internet at http:// pubs.acs.org.

’ ACKNOWLEDGMENT We gratefully acknowledge the expert technical assistance from Susen Mairif. We thank Agneta Prasse and Mareen Pagel for tests with E. coli lysate and help with peptide synthesis as well as Dr. Andrew Hagan for proofreading. This work was supported by BMBF Innoprofile, project no. 03IP604, and the European Fund for Regional Development (EFRE). ’ REFERENCES (1) Turk, B. Nat. Rev. Drug Discovery 2006, 5, 785–799. (2) Mann, K.; Mann, M. Proteomics 2008, 8, 178–191. (3) Zhang, X.; Fang, A.; Riley, C. P.; Wang, M.; Regnier, F. E.; Buck, C. Anal. Chim. Acta 2010, 664, 101–113. (4) Drag, M.; Salvesen, G. S. Nat. Rev. Drug Discovery 2010, 9, 690–701.

ARTICLE

(5) Rawlings, N. D.; Tolle, D. P.; Barrett, A. J. Nucleic Acids Res. 2004, 32, D160–D164. (6) Ionescu, R. E.; Cosnier, S.; Marks, R. S. Anal. Chem. 2006, 78, 6327–6331. (7) Lefkowitz, R. B.; Schmid-Schonbein, G. W.; Heller, M. J. Anal. Chem. 2010, 82, 8251–8258. (8) Dorn, J.; Harbeck, N.; Kates, R.; Gkazepis, A.; Scorilas, A.; Soosaipillai, A.; Diamandis, E.; Kiechle, M.; Schmalfeldt, B.; Schmitt, M. Ann. Oncol. 2011, 22, 877–883. (9) Cunnane, G.; FitzGerald, O.; Hummel, K. M.; Gay, R. E.; Gay, S.; Bresnihan, B. Rheumatology (Oxford) 1999, 38, 34–42. (10) Puente, X. S.; Sanchez, L. M.; Overall, C. M.; Lopez-Otin, C. Nat. Rev. Genet. 2003, 4, 544–558. (11) Zhao, M.; Josephson, L.; Tang, Y.; Weissleder, R. Angew. Chem., Int. Ed. Engl. 2003, 42, 1375–1378. (12) Ekici, O. D.; Zhu, J.; Wah Chung, I. Y.; Paetzel, M.; Dalbey, R. E.; Pei, D. Biochemistry 2009, 48, 5753–5759. (13) Twining, S. S. Anal. Biochem. 1984, 143, 30–34. (14) Gasparian, M. E.; Ostapchenko, V. G.; Dolgikh, D. A.; Kirpichnikov, M. P. Biochemistry (Moscow) 2006, 71, 113–119. (15) Kukor, Z.; Toth, M.; Sahin-Toth, M. Eur. J. Biochem. 2003, 270, 2047–2058. (16) Mizukami, S.; Takikawa, R.; Sugihara, F.; Shirakawa, M.; Kikuchi, K. Angew. Chem., Int. Ed. Engl. 2009, 48, 3641–3643. (17) Simon, P.; Weiss, F. U.; Sahin-Toth, M.; Parry, M.; Nayler, O.; Lenfers, B.; Schnekenburger, J.; Mayerle, J.; Domschke, W.; Lerch, M. M. J. Biol. Chem. 2002, 277, 5404–5410. (18) Karp, M. T.; Suominen, A. I.; Hemmila, I.; Mantsala, P. I. J. Appl. Biochem. 1983, 5, 399–403. (19) Karvinen, J.; Laitala, V.; Makinen, M. L.; Mulari, O.; Tamminen, J.; Hermonen, J.; Hurskainen, P.; Hemmila, I. Anal. Chem. 2004, 76, 1429–1436. (20) Lee, J. S.; Kim, Y. K.; Vendrell, M.; Chang, Y. T. Mol. Biosyst. 2009, 5, 411–421. (21) Lamore, S. D.; Qiao, S.; Horn, D.; Wondrak, G. T. Photochem. Photobiol. 2010, 86, 1307–1317. (22) Lee, Y. M.; Jeong, Y.; Kang, H. J.; Chung, S. J.; Chung, B. H. Biosens. Bioelectron. 2009, 25, 332–337. (23) Tothill, I. E. Semin. Cell Dev. Biol. 2009, 20, 55–62. (24) Deng, B.; Wang, Y.; Zhu, P.; Xu, X.; Ning, X. Anal. Chim. Acta 2010, 683, 58–62. (25) Kumaraswamy, S.; Bergstedt, T.; Shi, X.; Rininsland, F.; Kushon, S.; Xia, W.; Ley, K.; Achyuthan, K.; McBranch, D.; Whitten, D. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 7511–7515. (26) Qiao, X.; Sun, L.; Wang, L.; Liang, Y.; Zhang, L.; Shan, Y.; Peng, X.; Liang, Z.; Zhang, Y. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2011, 879, 1439–1443. (27) Rosenthal, M. A.; Katz, H. B. Clin. Chim. Acta 2011, 412, 665–667. (28) Simeonov, P.; Berger-Hoffmann, R.; Hoffmann, R.; Strater, N.; Zuchner, T. Protein Eng. Des. Sel. 2011, 24, 261–268. (29) Avrameas, S. J. Immunol. Methods 1992, 150, 23–32. (30) Bobrow, M. N.; Harris, T. D.; Shaughnessy, K. J.; Litt, G. J. J. Immunol. Methods 1989, 125, 279–285. (31) Knappe, D.; Zahn, M.; Sauer, U.; Schiffer, G.; Strater, N.; Hoffmann, R. ChemBioChem 2011, 12, 874–876. (32) Grubbs, F. E. Ann. Math. Stat. 1950, 21, 27–58. (33) Grubbs, F. E. Technometrics 1969, 11, 1–21. (34) Mermet, J. M. Spectrochim. Acta, Part B 2008, 63, 166–182. (35) Posokhov, Y. O.; Merzlyakov, M.; Hristova, K.; Ladokhin, A. S. Anal. Biochem. 2008, 380, 134–136. (36) Roda, A.; Guardigli, M.; Michelini, E.; Mirasoli, M. Anal. Bioanal. Chem. 2009, 393, 109–123. (37) Appel, W. Clin. Biochem. 1986, 19, 317–322. (38) Beirao, L. H.; Mackie, I. M.; Teixeira, E.; Damian, C. Braz. Arch. Biol. Technol. 2001, 44, 33–40. (39) Howard, N.; Abell, C.; Blakemore, W.; Chessari, G.; Congreve, M.; Howard, S.; Jhoti, H.; Murray, C. W.; Seavers, L. C. A.; van Montfort, R. L. M. J. Med. Chem. 2006, 49, 1346–1355. 7362

dx.doi.org/10.1021/ac201274f |Anal. Chem. 2011, 83, 7356–7363

Analytical Chemistry

ARTICLE

(40) Geokas, M. C.; Largman, C.; Brodrick, J. W.; Johnson, J. H. Am. J. Physiol. 1979, 236, E77–E83. (41) Preiser, H.; Schmitz, J.; Maestracci, D.; Crane, R. K. Clin. Chim. Acta 1975, 59, 169–175. (42) Lefkowitz, R. B.; Schmid-Schonbein, G. W.; Heller, M. J. Electrophoresis 2010, 31, 2442–2451. (43) Kunitz, M. J. Gen. Physiol. 1939, 22, 429–446. (44) Magee, A. I.; Grant, D. A.; Hermon-Taylor, J.; Offord, R. E. Biochem. J. 1981, 197, 239–244. (45) Abdallah, R. T.; Keum, J. S.; Lee, M. H.; Wang, B.; Gooz, M.; Luttrell, D. K.; Luttrell, L. M.; Jaffa, A. A. J. Biol. Chem. 2010, 285, 35206–35215. (46) Schmaier, A. H. Am. J. Physiol.: Regul. Integr. Comp. Physiol. 2003, 285, R1–R13. (47) Centi, S.; Tombelli, S.; Minunni, M.; Mascini, M. Anal. Chem. 2007, 79, 1466–1473. (48) Salvagno, G. L.; Berntorp, E. Semin. Thromb. Hemostasis 2010, 36, 780–790. (49) Chen, Z.; Pelc, L. A.; Di Cera, E. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 19278–19283.

7363

dx.doi.org/10.1021/ac201274f |Anal. Chem. 2011, 83, 7356–7363