© Copyright 2005 American Chemical Society
JULY 5, 2005 VOLUME 21, NUMBER 14
Letters Vesicular Polydiacetylene Sensor for Colorimetric Signaling of Bacterial Pore-Forming Toxin Guangyu Ma and Quan Cheng* Department of Chemistry, University of California, Riverside, California 92521 Received February 10, 2005. In Final Form: May 10, 2005 A vesicle-based polydiacetylene biosensor for colorimetric detection of bacterial pore-forming toxin streptolysin O (SLO) is reported. The sensor was constructed with three lipid constituents: glycineterminated diacetylene lipid Gly-PCDA, cell membrane-mimicking component PC-DIYNE, and cholesterol (CHO), which serves as the bait molecule. UV irradiation led to photopolymerization of the diacetylene lipids that gave the material a blue appearance. Incubation of the vesicles with SLO from Streptococcus pyrogenes turned the vesicle solution red, and the color change was found to be correlated to SLO concentration. The optimal sensing performance was found with vesicles consisting of 71% Gly-PCDA, 25% CHO, and 4% PC-DIYNE, and a correlation relationship was obtained for 20 HU to 500 HU/mL, or 100 pM to 6.3 nM of SLO toxin. Transmission electron microscopy and dynamic light scattering was used for further characterization of the vesicular assemblies. Transmembrane pores (holes) with diameter around 30 nm were observed on the vesicle membranes, in particular on the peripheral of the membrane structures, suggesting pore formation by SLO toxin provides the driving force for the color change of the functional vesicles.
Conjugated polymers such as polydiacetylenes (PDA) have been widely investigated as “smart” materials because of their unique electrical and optical properties.1 The spatially aligned monomeric diacetylene lipids, under UV irradiation, undergo a photopolymerization process via a 1,4-addition mechanism and form π-conjugated polymer chains that give the material a colored appearance.2,3 PDA exhibits a distinctive chromatic transition upon stimulation by temperature increase, mechanical stress, or chemical solvents.4 The bio-interaction induced chromism of PDA was first demonstrated by Charych et al. using influenza virus.5 Since then, PDA colorimetric * Corresponding author. E-mail:
[email protected]. Tel: (951) 827-2702. (1) Leclerc, M. Adv. Mater. 1999, 11, 1491. (2) Ringsdorf, H.; Schlarb, B.; Venzmer, J. Angew. Chem., Int. Ed. Engl. 1988, 27, 113. (3) O’Brien, D.; Whitesides, T.; Klingbiel, R. J. Polym. Sci., Polym. Lett. Ed. 1981, 19, 95. (4) Okada, S.; Peng, S.; Spevak, W.; Charych, D. Acc. Chem. Res. 1998, 31, 229.
sensors have been developed for cholera toxin,6 E. coli,7 c-myc epitope,8 and, most recently, lipopolysaccharides.9 We report here a new PDA sensor for detection of a cytolytic pore-forming toxin using functional vesicular assemblies. Cytolytic pore-forming toxins (PFTs) comprise about 25% of all known bacterial protein toxins and are considered important for the virulence of many diseasecausing bacteria.10 These toxins can insert into target cell membranes to form transmembrane channels, promoting bacterial infection by destroying tissue cells and immune (5) Charych, D. H.; Nagy, J. O.; Spevak, W.; Bednarski, M. D. Science 1993, 261, 585. (6) Charych, D.; Cheng, Q.; Reichert, A.; Kuziemko, G.; Stroh, M.; Nagy, J. O.; Spevak, W.; Stevens, R. C. Chem. Biol. 1996, 3, 113. (7) Ma, Z.; Li, J.; Liu, M.; Cao, J.; Zou, Z.; Tu, J.; Jiang, L. J. Am. Chem. Soc. 1998, 120, 12678. (8) Kolusheva, S.; Kafri, R.; Katz, M.; Jelinek, R. J. Am. Chem. Soc. 2001, 123, 417-422. (9) Rangin, M.; Basu, A. J. Am. Chem. Soc. 2004, 126, 5038-5039. (10) Alouf, J. E.; Palmer, M. W. In The Comprehensive Sourcebook of Bacterial Protein Toxins, 2nd ed.; Alouf, J. E., Freer, J. H., Eds.; Academic Press: London, 1999.
10.1021/la050376w CCC: $30.25 © 2005 American Chemical Society Published on Web 06/11/2005
6124
Langmuir, Vol. 21, No. 14, 2005
Letters
Figure 2. Color change of the vesicle sensors in response to SLO toxin. Samples shown here are (from left) original vesicle, BSA control, incubated with 200 HU/mL and 500 HU/mL SLO. The incubation time is 30 min and the incubation temperature is 32 °C.
Figure 1. Structure of lipid constituents used in the fabrication of PDA vesicle sensor for pore-forming toxin SLO.
cells involved in the first-line of defense.11 One of the significant PFTs is streptolysin O (SLO) from Streptococcus pyrogenes, a cholesterol-dependent cytolysin that has served as a major virulence factor in streptococcus infections of soft tissue such as strep throat, impetigo, and necrotising fascitis.12 Upon binding to the cell membrane, SLO can generate transmembrane pores with sizes ranging from a few nanometers to 35 nm.10 Engineered PFTs have been extensively explored as the “trigger and switch” channels for stochastic sensing of analytes.13 Recently, the vesicle encapsulation method has been used for developing fluorescence14 and amperometric15 sensors for PFTs that take advantage of their pore-forming function. However, direct signaling of PFTs by using response-generating functional membranes has not yet been demonstrated. The SLO vesicle sensors reported here were constructed with three lipid constituents: glycine-terminated diacetylene monomer Gly-PCDA, 1,2-bis(10,12-tricosadiynoyl)sn-glycero-3-phosphocholine (PC-DIYNE), and cholesterol (CHO) (structure shown in Figure 1). Cholesterol serves as the bait molecule for SLO toxin since the first step in pore formation is believed to be its binding to CHO.11 Attaching the glycine molecule to the diacetylene lipid increases the headgroup size of the lipid, facilitating vesicle formation under mild conditions.16 PC-DIYNE is a membrane-mimicking component and was used to adjust the stability and phase transition property of the vesicles. To prepare vesicles, a chloroform solution containing all of the components in a proper ratio was placed in an amber (11) Bhakdi, S.; Bayley, H.; Valeva, A.; Walev, I.; Walker, B.; Weller, U.; Kehoe, M.; Palmer, M. Arch Microbiol. 1996, 165, 73-79. (12) Palmer, M. Toxicon. 2001, 39, 1681-1689. (13) (a) Bayley, H.; Cremer, P. S. Nature 2001, 413, 226. (b) Braha, O.; Gu, L. Q.; Zhou, L.; Lu, X.; Cheley, S.; Bayley, H. Nat. Biotechnol. 2000, 18, 1005. (14) Rausch, J. M.; Wimley, W. C. Anal. Biochem. 2001, 293, 258263. (15) Xu, D.; Cheng, Q. J. Am. Chem. Soc. 2002, 124, 14314-14315. (16) Cheng, Q.; Stevens, R. C. Langmuir 1998, 14, 1974.
vial, and the organic solvent was removed with a nitrogen stream. An appropriate amount of Tris buffer (pH 7.5, containing 0.1 M NaCl) was added to make a final GlyPCDA concentration in the range of 0.08-0.25 mg/mL. The solution was sonicated for 20 min followed by 1-h incubation at 4 °C. Photopolymerization of vesicles was achieved using an UV cross-linker (254 nm). Samples were pipetted into a 96-well microplate and irradiated at the intensity of 0.3J/cm2 to yield a blue color. Figure 2 shows the colorimetric response of the vesicle sensor against SLO toxin. The vesicle sensor consists of 71% Gly-PCDA, 25% CHO, and 4% PC-DIYNE (molar ratio), polymerized with 30 s UV irradiation. Initially, the vesicle solution showed a blue color. Addition of 200 HU/mL SLO toxin, followed by incubation at 32 °C for 30 min, turned the solution red (third cuvette from left). HU is the hemolytic unit for SLO toxin, defined as the amount of protein that causes 50% lysis of a 2% red blood cell suspension in phosphate buffered saline at pH 7.4. A solution of 200 HU/mL SLO corresponds roughly to 76 ng of protein or 2.5 nM in a 500 µL testing solution. Addition of more SLO toxin (500 HU/mL, fourth from left) results in deeper color change, verifying the sensing response while also indicating the high end of the response range. The control experiment was carried out with excess BSA (10 µg/mL) dissolved in the same Tris buffer used for SLO solution. Only a minor chromatic transition to blue/purple color was observed (Figure 1, second from left). We noted that the color change caused by the BSA control and the Tris buffer was virtually the same, which strongly suggests this is simply a thermal effect. An additional control experiment was carried out with base-denatured SLO toxin. No noticeable color change was observed for 200 HU/mL denatured SLO (data not shown). The working composition of the vesicle sensors was obtained by carrying out a series of screening experiments that optimize both the vesicle composition and experimental conditions. The key to inducing a color change by a membrane-targeting toxin appears to be the balance between the degree of polymer conjugation and the membrane fluidity. Different percentages of cholesterol, ranging from 5% to 29%, were first tested. We found out that when the content of cholesterol exceeds 29% only a faint blue color was observed after UV irradiation, indicating poor polymerization efficiency. To maintain the percentage of nonpolymerizabale components below 29%, a range of 25-29% cholesterol was employed, depending
Letters
Langmuir, Vol. 21, No. 14, 2005 6125
Figure 3. Absorption spectra for the PDA vesicle sensors in response to SLO toxin. (1) original vesicles, (2) addition of 200HU/mL SLO, and (3) addition of 500 HU/mL SLO. Insert is the calibration plot of colorimetric response (CR) vs SLO concentrations.
on whether a codopant (PC or PC-DIYNE) was used. Irradiation time is another important factor. Long irradiation time results in a deeper blue color of the vesicles, indicating a high degree of polymerization among lipids. However, this leads to formation of stable vesicles, which resist the pore-forming attack by the toxin due to a lack of fluidity in the membrane. We studied the polymerization of vesicles with 0.2, 0.5, 1, and 2 min UV irradiation. Irradiation between 0.5 and 1 min appeared to be sufficient to generate the blue color for detection, whereas the fluid nature of the membrane still can be largely preserved. The incubation temperature was also optimized. It is well documented that PDA vesicles are sensitive to temperature increase.17 SLO toxin, on the other hand, demonstrates its maximal activity at 37 °C. We observed that incubation at 37 °C caused some degree of chromatic transition, generating a large background signal that subsequently reduces the detection sensitivity. Room temperature incubation was attempted but more than 3 h incubation was generally required. The optimal temperature was determined to be 32 °C. Quantitative analysis of SLO toxin with vesicle sensors was carried out by spectroscopic measurements, and the absorbance spectra are shown in Figure 3. The vesicle solution shows a strong adsorption peak at 649 nm before toxin addition, indicative of blue color. After incubation with 200 HU/mL SLO toxin, the absorbance peak shifts to 545 nm with concurrent decrease at 649 nm, similar to other PDA sensors reported in the literature.4 At this time, the solution shows a red appearance. The color change of the vesicles increases with the concentration of SLO toxin. Colorimetric response (CR),5 the percent change before and after toxin incubation in the maximum absorption at 649 nm normalized with the total absorption at both 649 and 545 nm, is thus calculated for toxin quantification. The insert in Figure 3 shows the calibration plot of CR vs SLO concentrations. A correlation relationship was obtained in the range of 20HU to 500 HU/mL or 100 pM to 6.3 nM of SLO toxin. A response plateau was observed at high concentration, likely due to saturation of pore formation in the membrane. The SLO vesicle sensors were further characterized by transmission electron microscopy (TEM) and dynamic light scattering (DLS). Figure 4 shows the TEM micrographs. Vesicular structures are clearly evidenced for the lipid composition used in this study (Figure 4, top image).
Figure 4. TEM micrographs of the vesicular structures obtained with 71% Gly-PCDA, 25% CHO, and 4% PC-DIYNE (top). The images in the bottom show the formation of pores (indicated by the arrows) after incubation with 200 HU/mL SLO for 30 min. The bar is 200 nm.
Incubation of the vesicles with 200 HU/mL SLO significantly altered the vesicle structure. A considerable number of pores (holes) were observed on the membranes, in particular on the peripheral of the membrane structures (Figure 4, bottom images, indicated by arrows). The pore size was estimated to be around 30 nm, which agrees well with the literature.18 Additional characterization was performed with DLS, which monitored the size change of vesicles with SLO incubation. Table 1 summarizes the results. The three-component vesicles have a relatively large average size (384 ( 10 nm). Upon UV irradiation, the vesicle size is shrunken to 355 ( 12 nm, likely due to formation of conjugation that reduces the distance between lipids. Incubation of the vesicles with 500 HU/mL SLO toxin gave a size measurement of 380 nm, yielding a 25 nm increase. Previously, we observed that pore formation by SLO results in approximately 16-30 nm size increase for phosphatidylcholine (PC) vesicles.15 The value observed here falls into the same range, strongly suggesting a similar pore-forming process occurring on the functional polymeric membrane. Long incubation time (1 h) results in a larger increase in size and also an increase in CR. BSA control and Tris buffer, however, did not trigger any size increase but showed a slight decrease in vesicle size. The reason for size decrease in the controls is unclear. It is worth noting that incorporation of PC-DIYNE lipid into the vesicles appears to be critical to acquiring the sensing ability against the SLO toxin. The original vesicles that did not include PC-DIYNE lipid were too stable to allow pore formation to take place. For instance, the system with 71% Gly-PCDA and 29% cholesterol shows only 6% CR for 500 HU/mL SLO incubated at 37 °C. Admittedly, the role of PC-DIYNE in the membrane is not totally clear. We speculate that it destabilizes the membrane system slightly to lower the energy threshold for chromatic (17) (a) Tanaka, H.; Gomez, M. A.; Tonelli, A. E.; Thakur, M. Macromolecules 1989, 22, 1208-1215. (b) Song, J.; Cisar, J. S.; Bertozzi, C. R. J. Am. Chem. Soc. 2004, 126, 8459-8465.
6126
Langmuir, Vol. 21, No. 14, 2005
Letters
Table 1. DLS Results for PDA-based Sensing Vesicles Interacting with SLO and Controls
size (nm)
before UV
after UV
addition of 500 HU/mL SLO
addition of BSA (10 µg/mL)
treated with Tris buffer
384 ( 10
355 ( 12
380 ( 12
337 ( 28
337 ( 11
transition, similarly to the function of the “promotion lipid” used in a PDA-based cholera toxin sensor.6 Interestingly, incorporation of the same amount (4%) of regular phosphatidylcholine (PC) with saturated chains does not show any response against SLO, though PC is the most common matrix used in pore-forming study of SLO toxin.19 It is interesting to compare the performance of the PDA sensor for SLO detection to that by other techniques. The PDA method yields a detection limit of 20 HU/mL, which corresponds to about 100 pM. For comparison, an electrochemical SLO sensor that uses redox probes and encapsulating vesicles gives a detection limit of 5 HU/ mL,15 which is slightly better. Nevertheless, the colorimetric PDA sensor provides visual signaling that an electrochemical sensor cannot offer, and it does not involve complex procedures of vesicle encapsulation and surface immobilization. The specificity of the sensor is achieved by the well-defined cholesterol/SLO interaction on the membrane, which is a subject of extensive investigation.20 The incubation of SLO/vesicles at an elevated temperature has a slightly negative impact on the sensitivity as compared to those PDA sensors that operate at room temperature. Solution pH is another important factor. Charged PDA shows chromatic transition when exposed to solutions of high pH, and this has been observed in almost all PDA sensors with carboxylic headgroup. High salt concentration can cause precipitation of the vesicles, which adds another limitation for the PDA vesicle biosensing. In conclusion, we have developed a new PDA sensor for the detection of bacterial toxin streptolysin O. The colorimetric response appears to arise from the poreforming action of SLO on the biomimetic membrane, which leads to perturbation of the backbone conformation and thus the optical signal. This argument is experimentally supported by the results from TEM and DLS studies. The litmus paper fashion of toxin detection is advantageous for on-site testing applications. Future work will focus on
further optimization toward better sensitivity and expending the detection to include more PFTs. Experimental Section Gly-PCDA was synthesized from 10,12-pentacosadiynoic acid (PCDA) based on a published procedure16 and verified by NMR and mass spectrometry. PC-DIYNE was purchased from Avanti Polar Lipids (Alabaster, AL). A typical procedure of vesicle preparation is the following: A chloroform solution containing 0.25 mg of Gly-PCDA lipid (5.79 × 10-4 mmol), 0.078 mg of cholesterol, and 0.027 mg of PC-DIYNE was transferred into a 4-mL amber vial. The mixture was dried with a nitrogen stream. 3 mL of Tris buffer was then added into the vial, followed by probe sonication at room temperature for 20 min. The resulting clear solution was incubated at 4 °C for 1 h before use. Photopolymerization was subsequently carried out with a UV cross-linker. An aliquot of 450 µL of polymerized vesicle solution was transferred to a microcuvette and 50 µL of SLO solution of different concentrations was added to make the final concentration of the toxin ranging from 10 HU to 500 HU/mL. The vesicle solution was then incubated at 32 °C for 30 min along with the controls. Absorbance was measured before and after toxin incubation, from which CR values were calculated. TEM characterization was carried out with a Phillips CM300 transmission electron microscope operating at 200 kV. The samples were negatively stained with 2% uranyl acetate. DLS measurements were performed with a particle size analyzer from Brookhaven Instruments Corporation.
Acknowledgment. The authors thank Drs. Yushan Yan and Eric Hoek of UCR Department of Chemical Engineering for allowing us to use the DLS instrument, and Scott Phillips for critical reading of the manuscript. This work was supported by startup funds from UC Riverside and Eli Lilly analytical chemistry grant. LA050376W (18) Sekiya, K.; Danbara, H.; Yase, K.; Futaesaku, Y. J. Bacteriol. 1996, 178, 6998. (19) Zitzer, A.; Westover, E. J.; Covey, D. F.; Palmer, M. FEBS Lett. 2003, 553, 229-231. (20) Palmer, M. FEMS Microbiol Lett. 2004, 238, 281-9.
