Magnetic Nanoparticle-Based Platform for ... - ACS Publications

Oct 8, 2015 - O157:H7 can cause bloody diarrhea, hemolytic uremic syndrome, or even death. The pathogenicity of E. coli O157:H7 is mainly caused by th...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Magnetic Nanoparticle-Based Platform for Characterization of Shigalike Toxin 1 from Complex Samples Fang-Yin Kuo, Bo-Yao Chang, Ching-Yi Wu, Kwok-Kong Tony Mong,* and Yu-Chie Chen* Department of Applied Chemistry, National Chiao Tung University, Hsinchu 300, Taiwan S Supporting Information *

ABSTRACT: Foodborne illness outbreaks resulting from contamination of Escherichia coli O157:H7 remain a serious concern in food safety. E. coli O157:H7 can cause bloody diarrhea, hemolytic uremic syndrome, or even death. The pathogenicity of E. coli O157:H7 is mainly caused by the expression of Shiga-like toxins (SLTs), i.e., SLT-1 and SLT-2. SLTs are pentamers composed of a single A and five B subunits. In this study, we propose a magnetic nanoparticle (MNP)-based platform to rapidly identify SLT-1 from the complex cell lysate of E. coli O157:H7. The core of the MNPs is made of iron oxide, whereas the surface of the core is coated with a thin layer of alumina (Fe3O4@Al2O3 MNPs). The Fe3O4@ Al2O3 MNPs are functionalized with pigeon ovalbumin (POA), which contains Gal-α(1→4)-Gal-β(1→4)-GlcNAc termini that can bind SLT1B selectively. Furthermore, POA is a phosphate protein. Thus, it can be easily immobilized on the surface of the Fe3O4@Al2O3 MNPs through aluminum phosphate chelation under microwave heating within 1.5 min. The generated POA−Fe3O4@Al2O3 MNPs are capable of effectively enriching SLT-1B from complex cell lysates simply by pipetting 20 μL of the sample in and out of the tip in a vial for ∼1 min. To release SLT-1 from the MNPs, Gal-α(1→ 4)-Gal disaccharides were used for displacement. The released target species are sufficient to be identified by matrix-assisted laser desorption/ionization mass spectrometry. Although the sample volume used in this approach is small (20 μL) and the enrichment time is short (1 min), the selectivity of this approach toward SLT-1B is quite good. We have demonstrated the effectiveness of this approach for rapid determination of the presence of SLT-1 from complex cell lysates and ham/juice samples based on the detection of SLT-1B.

T

Pigeon (Columba livia) egg white constitutes a rich pool of natural glycoproteins,14 among which the pigeon ovalbumin (POA) (∼60%) is the most abundant.14 POA contains Galα(1→4)-Gal-β(1→4)-GlcNAc termini.14,22−24 In addition to the carbohydrate ligand, POA also contains phosphate (OPO32−) groups.25 POA-immobilized gels have been used to selectively purify SLT-1.9 In addition, POA immobilized on the surface of alumina oxide-coated Fe3O4 (POA−Fe3O4@ Al2O3) MNPs through phosphate−aluminum(III) chelation has been used as effective probes to selectively target bacteria, such as uropathogenic E. coli25 and Pseudomonas aeruginosa,26 which contain binding sites for Gal-α(1→4)-Gal. However, POA glycoproteins need purification from chicken egg white before use for preparation of POA−Fe3O4@Al2O3 MNPs in those previous studies. The purification of POA from pigeon egg white is time-consuming. Thus, in this study, we immobilized POA on the surface of the Fe3O4@Al2O3 MNPs directly from complex pigeon egg white proteins. The Fe3O4@ Al2O3 MNPs were used as affinity probes to bind with POA, which was obtained from complex pigeon egg white proteins,

he bacterium Escherichia coli O157:H7 is a foodborne pathogen that has frequently caused foodborne illness outbreaks.1−5 The pathogenicity of this bacterial strain is mainly attributed to the expression of Shiga-like toxins (SLTs), which can cause bloody diarrhea and hemolytic uremic syndrome.6−8 Treatment of HeLa cell lines with as little as 1.5 ng mL−1 SLT-1 can inhibit cell growth up to 50%.9 SLTs are composed of a single A and five B subunits.10−12 The A subunit can inactivate ribosomes, leading to the inhibition of the biosynthesis of eukaryotic proteins. The B subunits can specifically bind to glycolipid ligands,13−16 i.e., globotriaosylceramide (Gal-α(1→4)-Gal-β(1→4)-Glc-β-O-ceramide) (Gb3) on the cell membrane of target cells. Taking advantage of such specific carbohydrate−protein recognition, Gb3 analogues have been frequently used as probes to interact with SLTs.9,17−22 For example, Gb3 analogues immobilized on magnetic nanoparticles (MNPs)17 and gold nanoparticles (NPs)20,21 have been successfully used to isolate B subunits derived from SLT-1 in the complex cell lysate of E. coli. The toxin trapping capacity was estimated to be ∼30 μg of B subunits trapped on per minigram of Gb3-functionalized MNPs.17 In those studies, Gb3 analogues were obtained through organic synthesis. If Gb3containing molecules are already available from natural products, the availability will not be a problem. © XXXX American Chemical Society

Received: July 20, 2015 Accepted: September 28, 2015

A

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry Scheme 1. Cartoon Illustration of Our Approach Used in This Study

obtained from a local pigeon farm. Amicon Ultra 4 centrifugal filters were purchased from Millipore (Billerica, MA, U.S.A.). E. coli O157:H7 (BCRC 13085) was purchased from Bioresource Collection and Research Center (Hsinchu, Taiwan), whereas E. coli DH5α containing pUCVT1 (SLT100) that can express SLT-19 was kindly provided by Professor Hiroshi Morie (Department of Microbiology, Fujita Health University, School of Medicine, Japan). Pall acrodisc syringe filters with Supor membrane (PN-4612; pore size, 0.2 μm) were obtained from Voigt Global Distribution (Lawrence, KS, U.S.A.). 10Mecaptodecyl 4-O-(α-D-galactopyranosyl)-β-D-galactopyranoside (Gal-α(1→4)-Gal) was synthesized in laboratory (see the Supporting Information). Apple juice and ham were obtained from a local supermarket. Preparation of SLTs from Cell Lysates. Genetically modified E. coli BL21/pETSTX1 that can express SLT-1 was cultured in LB broth (25 mg mL−1, 10 mL) containing kanamycin (30 μg mL−1) at 37 °C for 4−6 h. When the optical density at the wavelength of 600 nm (OD600) reached higher than 0.3, IPTG (2.38 mg) was added to the bacterial suspension to initiate protein expression. The final concentration of IPTG was 1 mM. After incubation at 37 °C for 24 h, the bacterial suspension was centrifuged at 6000 rpm for 10 min, followed by elimination of the supernatant. The remaining bacterial cells were rinsed twice with Tris buffer (pH 7.4, 15 mM, 1 mL) under centrifugation at 6000 rpm for 10 min. The rinsed bacterial cells were then resuspended in urea (8 M, 1 mL) and incubated at 37 °C for 2 h to lyse the cells. The resultant cell lysates were centrifuged at 6000 rpm for 10 min, and the precipitates were removed. The supernatant containing SLT-1 was filtered through a Pall acrodisc syringe filter (pore size, 0.2 μm), whereas the filtrate was collected. The filtrate was treated with a ZipTip to remove most salts. The resultant filtrate (100 μL) containing SLT-1 was mixed with Tris buffer (pH 7.4, 15 mM, 100 μL), which was used as the model sample for the experiments. E. coli O157:H7 is a Risk Group 2 pathogen. Thus, for safety concern, the experiments involving the use of E. coli O157:H7

through aluminum(III)−phosphate chelation. The resultant conjugates were mainly composed of the POA−Fe3O4@Al2O3 MNPs because POA is the dominant phosphoprotein in egg white. To accelerate the functionalization process of the Fe3O4@Al2O3 MNPs, the binding experiment was conducted under microwave heating.27,28 Within a very short duration (∼1.5 min), POA was bound to the surface of the Fe3O4@ Al2O3 MNPs. Such an approach can reduce the time of POA purification and functionalization of MNPs. The resulting POA−Fe3O4@Al2O3 MNPs were employed as affinity probes to target SLT-1 from complex samples (Scheme 1). The cell lysates were derived from genetically modified E. coli and wildtype E. coli O157:H7, both of which can express SLTs and were used as the model samples. Matrix-assisted laser desorption/ ionization mass spectrometry (MALDI-MS) was applied to characterize the species trapped by the POA−Fe3O4@Al2O3 MNPs.

■

EXPERIMENTAL SECTION Reagents and Materials. Iron(III) chloride, aluminum isopropoxide, α-cyano-4-hydroxycinnamic acid (CHCA), isopropyl β-D-1-thiogalactopyranoside (IPTG), kanamycin, human serum albumin (HSA), 2,5-dihydroxybenzoic acid (2,5-DHB), brilliant blue G, ammonium hydroxide solution, trypsin (from bovine pancreas), peanut agglutinin, cytochrome c, iodoacetic acid (IAA), and dithiothreitol (DTT) were purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.). Iron(II) chloride, ammonium bicarbonate, urea, and tris(hydroxymethyl)-aminomethane (Tris) were purchased from J.T. Baker (Phillipsburg, NJ, U.S.A.). Acetonitrile (99%) was purchased from Merck (Darmstadt, Germany). Luria−Bertani (LB) broth, tryptic soy broth (TSB), and yeast extract (Y) were purchased from Becton Dickinson (Franklin Lakes, NJ, U.S.A.). Tetraethylorthosilicate (TEOS) was purchased from Fluka (Buchs, St. Gallen, Switzerland). Trifluoroacetic acid (TFA) (99%) and phosphoric acid (85%) were purchased from Riedel-de Haen (Buchs, St. Gallen, Switzerland). ZipTips were obtained from Millipore (Bedford, MA, U.S.A.). Pigeon (C. livia) eggs were B

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Instrumentation. All the mass spectra were obtained at positive ion mode using an AutoFlex III MALDI mass spectrometer (Bruker Daltonics, Breman, Germany). When analyzing the analytes larger than 3500 Da, linear mode was utilized and the laser size was set to large. The voltages were set as follows: ion source 1, 20.05 kV; ion source 2, 18.65 kV; lens, 6.82 kV; detector gain, 1169 V. The ion suppression was set at m/z 3500. The mass spectra were collected within the range of m/z 4000−10 000. When analyzing peptides, reflectron mode was used and the laser size was set to medium. The voltages were set as follows: ion source 1, 19.05 kV; ion source 2, 16.59 kV; lens, 8.78 kV; reflector 1, 21.06 kV; reflector 2, 9.72 kV; detector gain, 1645 V. The ion suppression was set at m/z 900. The mass spectra were collected at the range of m/z 1000− 2500. Infrared (IR) spectra were obtained using a FTS3100 FTIR spectrometer (Digilab, U.S.A.). Hysteresis curves were obtained using a superconducting quantum interference device (SQUID) magnetometer (MPMS5, Quantum Design, U.S.A.). ζ-Potentials were measured using a Delsa Nano light scattering from Beckman Coulter (Brea, CA, U.S.A.). Transmission electron microscopy (TEM) images were obtained using a JEM2000 FXII TEM from JEOL (Tokyo, Japan). UV−vis absorption spectra were obtained using NanoVue plus (Buckinghamshire, England).

were operated in our Biosafety Level 2 laboratory. E. coli O157:H7 was cultured in TSBY broth (25 mg mL−1, 12 mL) at 37 °C for 12 h. TSBY was prepared by mixing TSB (12 g) and yeast extract (Y) (2 g) in deionized water (400 mL). The resultant bacterial suspension was rinsed twice with Tris buffer (pH 7.4, 15 mM, 1 mL) under centrifugation at 6000 rpm for 10 min. The cells were lysed by resuspension in urea (8 M, 1 mL) and incubation at 37 °C for 2 h. The supernatant presumably containing SLTs was filtered through a Pall acrodisc syringe filter (pore size, 0.2 μm), and the filtrate was collected. The filtrate was treated with a ZipTip to remove most salts. The resultant filtrate (100 μL) containing SLT-1 was mixed with the Tris buffer (pH 7.4, 15 mM, 100 μL) and used as the model sample for the experiments. Preparation of SLT-Contaminated Food Samples. The toxin-contaminated food samples were prepared by spiking cell lysates of E. coli O157:H7 to the food sample. A 50-fold diluted apple juice sample (50 μL) prepared in Tris buffer spiked with cell lysates (50 μL) was first prepared followed by 2-fold dilution with Tris buffer (pH 7.4, 15 mM) prior to enrichment experiments. When preparing SLT-1-contaminated ham, one colony of E. coli O157:H7 was cultured with a piece of ham (∼0.9 g) soaked in TSBY broth (25 mg mL−1, 3 mL) at 37 °C for 12 h. After centrifugation, the resultant suspension was removed and the remaining ham was vortex-mixed in Tris buffer (pH 7.4, 15 mM, 3 mL) for 2 h. The supernatant was mixed with urea (8 M, 0.1 mL) and incubated at 37 °C for 2 h followed by desalting treatment. The resultant filtrate (20 μL) presumably containing SLT-1 was mixed with Tris buffer (pH 7.4, 15 mM, 20 μL) prior to enrichment experiments. Enrichment of SLTs Using the POA−Fe3O4@Al2O3 MNPs as Affinity Probes. The details of preparing of the POA−Fe3O4@Al2O3 MNPs are described in the Supporting Information. When performing the toxin enrichment experiments, the POA−Fe3O4@Al2O3 MNPs (50 μg) were mixed with the aforementioned samples (20 μL) by pipetting for 1 min followed by magnetic separation. After discarding the supernatant, the remaining POA−Fe3O4 MNP−target species conjugates were rinsed twice with Tris buffer (pH 7.4, 15 mM, 20 μL) by pipetting. Galα(1→4)Gal glycoside (80 nM, 2 μL) prepared in Tris buffer (pH 7.4, 15 mM, 20 μL) was added to the rinsed conjugates to release the target species trapped by the MNPs. After magnetic separation, the supernatant (2 μL) was mixed with CHCA (15 mg mL−1, 2 μL). The mixture (2 μL) was deposited on a sample plate. After solvent evaporation, the sample was readily analyzed by MALDI-MS. Tryptic Digestion of SLT-1. To confirm the identity of the species trapped by the POA−Fe3O4@Al2O3 MNPs, tryptic digestion was conducted. To speed up the analysis process, tryptic digestion was conducted under microwave heating. The trapped species on the POA−Fe3O4@Al2O3 MNPs were released by adding Galα(1→4)Gal (80 nM, 10 μL), followed by mixing with bare Fe3O4 MNPs (5 μg). The function of the bare Fe3O4 MNPs is to aid protein attachment on the MNPs for ease of enzymatic digestion29 and to accelerate the heating process.27,28,30,31 The mixture was incubated in a microwave oven (power, 900 W) for 60 s, followed by the addition of trypsin (23.3 ng mL−1, 2 μL, pH 8). The mixture was then heated in the microwave oven for 90 s (power, 450 W). After magnetic separation, the supernatant (2 μL) was mixed with CHCA (15 mg mL−1, 2 μL). After solvent evaporation, the sample was ready for characterization by MALDI-MS.

■

RESULTS AND DISCUSSION The POA−Fe3O4@Al2O3 MNPs were used as nanoprobes for targeting SLT-1B in this work. Initially, Fe3O4 MNPs were prepared followed by coating with a SiO2 layer for the attachment of the alumina. Alumina prepared from the sol−gel reaction was then coated on the surface of the Fe3O4@SiO2 MNPs to generate the Fe3O4@Al2O3 MNPs. The prepared Fe3O4, Fe3O4@SiO2, and Fe3O4@Al2O3 MNPs were characterized by IR spectroscopy (Figure S1). The band that appeared at ∼580 cm−1 corresponded to the stretching vibration mode of Fe−O,32 whereas the band that appeared at ∼3400 cm−1 corresponded to the OH stretching (Figure S1A). The IR spectrum of the Fe3O4@SiO2 MNPs contained two absorption bands at 797 and 1095 cm−1 corresponding to the asymmetrical and symmetrical vibration modes of the Si− O−Si bond,33 respectively (Figure S1B). Additionally, the band at 948 cm−1 representing Si−OH34 implicated the presence of silanol groups on the surface of the Fe3O4 MNPs (Figure S1B). The new band shown at 1401 cm−1 representing Al−OH35 appeared in the IR spectrum of the Fe3O4@Al2O3 MNPs (Figure S1C). The results confirmed that an alumina layer was successfully coated on the surface of the MNPs. Figure S2A shows the TEM image of the Fe3O4@Al2O3 MNPs. The MNPs were not uniform. We estimated the average size of the MNPs to be ∼30 nm. In addition, the magnetic property of the MNPs was examined by SQUID (Figure S2B). The magnetization value decreased from ∼55 emu g−1 (black curve) to ∼40 emu g−1 (red curve) after the Fe3O4 MNPs were coated with a thin layer of SiO2 followed by alumina. This reduction is understandable because the SiO2 and alumina coating do not contribute to magnetization, and the weight of the coating was taken into account of the total weight. We then used the generated Fe3O4@Al2O3 MNPs as substrate to bind with POA directly from denatured pigeon egg white proteins; as such, the POA−Fe3O4@Al2O3 MNPs were produced. To confirm the binding of POA on the surface of the Fe3O4@Al2O3 MNPs, 0.1% phosphoric acid containing MALDI matrix was mixed with the generated MNPs, which C

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 1. MALDI mass spectra of (A) intact cells, (B) the cell lysate of the E. coli DH5α containing pUCVT1 (SLT100), (C) the eluted species obtained after using the POA−Fe3O4@Al2O3 MNPs to selectively enrich target species from the sample used in panel B followed by elution with Gal-α(1→4)-Gal (80 nM, 2 μL), and (D) the tryptic digest of the same sample used in panel C. SLT-1B+ and SLT-1B2+ indicate the peaks derived from the singly and doubly charged ions of SLT-1B, respectively. Red dots indicate the peaks derived from SLT-1B, while the blue dots indicate the peaks derived from trypsin self-digest.

would release the POA (MW = ∼53 000 Da).36 This result was confirmed by MALDI-MS. Figure S3 shows the resultant MALDI mass spectrum. Several peaks appeared at m/z ∼13 200, ∼17 600, ∼26 400, and ∼53 000, corresponding to the multiple charged ions of POA, i.e., POA4+, POA3+, POA2+, and POA+, respectively. The results showed that POA was successfully immobilized on the MNPs. The amount of POA bound to the Fe3O4@Al2O3 MNPs was estimated to be ∼0.49 nmol mg−1. Furthermore, the ζ-potential changed from approximately −10.9 to −21.7 mV at pH 7 (Tris buffer, 10 mM) after POA was bound to the surface of the Fe3O4@Al2O3 MNPs. The isoelectric point of POA was 4.68. Thus, the ζpotential shifted to a higher value. This result provided further evidence that the Fe3O4@Al2O3 MNPs were functionalized by POA. At outset, the binding behavior of the generated POA− Fe3O4@Al2O3 MNPs toward nontarget proteins was examined. Cytochrome c was selected as the test sample. Parts A and B of Figure S4 show the direct MALDI mass spectra of the sample containing cytochrome c (pH 7.4) obtained before and after using the POA−Fe3O4@Al2O3 MNPs as the affinity probes to enrich target species from the sample, respectively. Although the multiply charged ion peaks derived from cytochrome c appeared in the mass spectrum, there was no peak observed in the mass spectrum after enrichment. The results demonstrated that our approach cannot be used to detect cytochrome c. One might wonder why cytochrome c cannot be trapped by the POA−Fe3O4@Al2O3 MNPs. Indeed, cytochrome c could be trapped by the POA−Fe3O4@Al2O3 MNPs at pH 7.4 through electrostatic interactions. However, it could not be released from the MNPs with addition of Galα(1→4)Gal glycoside. As we described in the Experimental Section, the two-step enrichment was conducted by using the POA−Fe3O4@Al2O3

MNPs to trap target species (first step) from samples followed by releasing target species with addition of Galα(1→4)Gal glycoside (second step). Thus, only the target species have binding sites with Galα(1→4)Gal can be released from the MNPs and detected by MALDI-MS. Additionally, peanut agglutinin is a lectin that can bind with the ligand containing Galβ1-4GalNAc.37 Thus, it was also selected as the model protein to examine if our nanoprobes have binding affinity toward this lectin. Figure S4C shows the direct MALDI mass spectrum of peanut agglutinin. It is clear that the multiply charged ions derived from peanut agglutinin appeared in the mass spectrum. However, there was no peak found in the mass spectrum (Figure S4D) after using the POA−Fe3O4@Al2O3 MNPs as the affinity probes to selectively trap target species from the sample containing peanut agglutinin. The results demonstrated that our approach has no binding capacity toward this lectin. Next, the cell lysate of the genetically engineered E. coli that can express SLT-1 was used as the model sample. Figure 1A shows the MALDI mass spectrum of the intact E. coli cells. The peaks at m/z 4367, 5099, and 6206 were dominant in the mass spectrum. Figure 1B shows the MALDI mass spectrum of the cell lysate of E. coli treated by a high concentration of urea. The peak at m/z 9744 dominated the mass spectrum. Presumably, this peak was derived from the B subunit of SLT-1 (SLT-1B) (MW = 9743 Da). We then used the POA−Fe3O4@Al2O3 MNPs as affinity probes (50 μg) to trap target species from the same cell lysate sample (20 μL), followed by elution with free Gal-α(1→4)-Gal disaccharide ligands. Enrichment and elution were simply conducted by mixing the sample solution with the MNPs by pipetting for 1 min. The eluted species from the MNPs were then characterized by MALDI-MS. Figure 1C shows the MALDI mass spectrum of the eluted species. Only D

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry Table 1. List of the Sequences That Correspond to Ions Observed in Figures 1D and 2D sequence no.

theoret value

exptl value

protein

VGDKELFTNR

sequence

44−53

1178.6

SLT-1B

KELFTNRWNLQSLLLSA

47−63

2033.3

YNDDDTFTVKVGDKELFTNR

34−53

2378.1

144−161

1962.1

50−69

2163.3

1178.5a 1178.1b 2033.0a 2032.8b 2378.4a 2378.1b 1962.4a 1962.2b 2163.5a 2163.9b

SDSSCKSAYPGQITSNMF LGEDNINVVEGNEQFISASK a

SLT-1B SLT-1B trypsin trypsin

Figure 1D. bFigure 2D.

Figure 2. MALDI mass spectra of (A) intact cells, (B) the cell lysate of E. coli O157:H7, (C) the cell lysate obtained after using the POA−Fe3O4@ Al2O3 MNPs (50 μg) as affinity probes to selectively enrich target species from the cell lysate (20 μL) used in panel B, and (D) the tryptic digest of the eluted species used in panel C. SLT-1B+ and SLT-1B2+ indicate the peaks derived from the singly and doubly charged ions of SLT-1B, respectively. Red dots indicate the peaks derived from SLT-1B, while the blue dots indicate the peaks derived from trypsin self-digest.

two peaks at m/z 4872 and 9744, presumably derived from the doubly charged and singly charged SLT-1B, respectively, appeared in the mass spectrum. The eluted species were further digested by trypsin and characterized by MALDI-MS. Figure 1D shows the resultant MALDI mass spectrum. The peaks at m/z 1178.5 (nos. 44−53), 2033.0 (nos. 47−63), and 2378.4 (nos. 34−53) represent the sequences derived from SLT-1B (Table 1). In addition, the peaks at m/z 1962.4 and 2163.5 are the peptides derived from trypsin self-digestion (Table 1). These results demonstrated that the POA−Fe3O4 MNPs could be used to selectively trap SLT-1B from a complex cell lysate sample. Furthermore, only a small volume (20 μL) of sample was needed for the entire analysis, and the enrichment time was as short as 1 min. We further used wild-type E. coli O157:H7 as the model sample to demonstrate the feasibility of using this approach for analyzing real samples. Figure 2A shows the MALDI mass spectrum of the intact E. coli O157:H7. Figure 2B shows the MALDI mass spectrum of the cell lysate from E. coli O157:H7.

The peaks at m/z 4872 and 9744, derived from the doubly charged and singly charged ions of SLT-1B, respectively, appeared in the mass spectrum. In additional, a peak at m/z 5457 was observed in the same mass spectrum. Figure 2C shows the MALDI mass spectrum obtained after using the POA−Fe3O4@Al2O3 MNPs as the affinity probes to enrich the cell lysate of E. coli O157:H7 (20 μL). Only the ion peaks at m/ z 4872 and 9744 were observed in the mass spectrum. To confirm the species enriched by the MNPs, trypsin digestion was further conducted. Figure 2D shows the MALDI mass spectrum of the resultant trypsin digest of the eluted species from the POA−Fe3O4@Al2O3 MNPs. The peaks at m/z 1178.1 (nos. 44−53), 2032.8 (nos. 47−63), and 2378.1 (nos. 34−53) corresponding to the sequences derived from SLT-1B appeared in the mass spectrum. These results indicated that the POA− Fe3O4@Al2O3 MNPs could selectively enrich SLT-1B from the complex mixture. No ion peaks derived from the SLT-1A were observed in the MALDI mass spectra. Presumably, SLT-1A is less abundant than SLT-1B since SLT-1 is composed of one A E

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 3. (A) MALDI mass spectrum of the 100-fold diluted cell lysate sample that was used to obtain Figure 2B. (B) MALDI mass spectrum obtained after using the POA−Fe3O4@Al2O3 MNPs (50 μg) as affinity probes to selectively enrich target species from the cell lysate (20 μL) used in panel A.

Figure 4. MALDI mass spectra of the apple juice sample (20 μL) containing the cell lysate of E. coli O157:H7 obtained (A) before and (B) after enrichment by the POA−Fe3O4@Al2O3 MNPs (50 μg). MALDI mass spectra of the ham sample cultured with E. coli O157:H7 obtained (C) before and (D) after enrichment by the POA−Fe3O4@Al2O3 MNPs (50 μg). The details for preparing the SLT-contaminated ham sample are described in the Experimental Section.

and five B subunits. Additionally, the molecular size of A subunits (∼32 kDa) is much larger than that of B subunits (∼9.7 kDa). MALDI-MS is more sensitive to the analytes with the molecular size of B subunit than those with the molecular weight larger than 10 kDa such as A subunit. Therefore, no ions derived from SLT-1A were observed in the resultant MALDI mass spectra obtained from present cell lysate samples. Additionally, no ions derived from SLT-2 were observed in the MALDI mass spectra of the cell lysate samples. Presumably, it is because that the binding affinity of POA toward SLT-2 is much lower than toward SLT-1.9 The poor binding interaction between POA and SLT-2 has been studied in the previous report.9 To examine the sensitivity of this approach, we diluted the cell lysate of wild-type E. coli O157:H7 by Tris buffer with a dilution factor of 100 and conducted the same enrichment experiment similar to that used to obtain Figure 2B. Figure 3A shows the direct MALDI mass spectrum of the diluted cell

lysate sample. No apparent ions peaks were observed in the mass spectrum. Figure 3B shows the MALDI mass spectrum of the diluted cell lysate sample obtained after enriched by the POA−Fe3O4@Al2O3 MNPs. The peaks at m/z 4872 and 9744 representing SLT-1B appeared in the mass spectrum. These results indicated that the POA−Fe3O4@Al2O3 MNPs could be used to selectively enrich a low concentration of SLT-1B from the complex cell lysate sample. In addition, we also used Bradford dye-binding assay38,39 to estimate the protein concentration in the diluted cell lysate samples. HSA was used as the standard protein in Bradford assay. Figure S5A shows the calibration curve of HSA incubated with the Bradford dye by plotting the absorbance at the wavelength of 609 nm versus the HSA concentration. The inset shows the corresponding absorption spectra. Figure S5B shows the absorption spectra of the sample containing SLT-1B after being incubated with the Bradford dye. The sample was prepared using the same steps to obtain Figure 2B. The F

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry concentration of SLT-1B was estimated to be ∼44.3 nM. Since the enrichment factor was 10 when preparing the sample for Bradford assay, the original concentration of SLT-1B in the cell lysate was ∼4.4 nM. Therefore, the concentration of SLT-1B used to obtain the result in Figure 3 was ∼44 pM. That is, our approach could be used to enrich a very low concentration of SLT-1B from a small volume (20 μL) of complex cell lysate sample. Compared with a previous report,40 which only achieved a detection limit of ∼300 ng mL−1 (∼40 nM), our approach was much more sensitive. We also used apple juice spiked with the cell lysate of E. coli O157:H7 and a piece of ham contaminated by E. coli O157:H7 as the model samples. Parts A and B of Figure 4 show the MALDI mass spectra of the apple juice spiked with the cell lysate before and after enrichment by the POA−Fe3O4@Al2O3 MNPs, respectively. No apparent ion peaks were observed in the mass spectrum before enrichment. After enrichment, the doubly and singly charged ions at m/z 4872 and 9744 derived from SLT-1B appeared in the MALDI mass spectrum. These results demonstrated that our approach could be used in determining the presence of the SLT-1B in the complex juice sample. Additionally, a piece of ham contaminated by E. coli O157:H7 was also used as the model sample. Parts C and D of Figure 4 show the MALDI mass spectra of the ham sample contaminated by E. coli O157:H7 obtained before and after being treated by the POA−Fe3O4@Al2O3 MNPs, respectively. The results show that there was no apparent peak found in the MALDI mass spectrum before enrichment (Figure 4C). It may be due to the complexity and the presence of high salts in the ham sample, resulting in the difficulty in revealing any ions in the MALDI mass spectrum. However, the ion peaks at m/z 4872 and 9744 representing the doubly charged and singly charged ions of SLT-1B appeared in the MALDI mass spectrum after enrichment. These results indicate that our nanoprobes have the capacity to selectively enrich SLT-1 even from very complex samples.

operation. Thus, we believe that this approach is potentially suitable to be used in the rapid identification of SLT-1 based on the detection of SLT-1B from suspicious samples.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b02712. Additional experimental details, synthetic scheme, and figures of IR spectra, TEM image and magnetic hysteresis curves, MALDI mass spectra, and Bradford assay (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] Phone: +886-3-51-31204 Fax: +886-3-5723764. *E-mail: [email protected]. Phone: +886-3-5131527. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the Ministry of Science and Technology of Taiwan (MOST102-2627-M-009-002) for financial support of this work. We are very grateful to Professor Yu-Chen Chen (Tajen University, Taiwan), Mr. Chao-Ming Su, Ms. Ling-Tun Kung, and Mr. Lung-Hsing Liu for their kind help in searching fresh pigeon eggs.

■

REFERENCES

(1) Rodrigue, D. C.; Mast, E. E.; Greene, K. D.; Davis, J. P.; Hutchinson, M. A.; Wells, J. G.; Barrett, T. J.; Griffin, P. M. J. Infect. Dis. 1995, 172, 1122−1125. (2) Wells, J. G.; Shipman, L. D.; Greene, K. D.; Sowers, E. G.; Green, J. H.; Cameron, D. N.; Downes, F. P.; Martin, M. L.; Griffin, P. M.; Ostroff, S. M. J. Clin. Microbiol. 1991, 29, 985−989. (3) Lahti, E.; Eklund, M.; Ruutu, P.; Siitonen, A.; Rantala, L.; Nuorti, P.; Honkanen-Buzalski, T. Eur. J. Clin. Microbiol. Infect. Dis. 2002, 21, 189−195. (4) http://www.cdc.gov/ecoli/2014/O157H7-05-14/index.html, accessed on Oct. 8, 2015. (5) Orden, J. A.; Cid, D.; Ruiz-Santa-Quiteria, J. A.; García, S.; Martínez, S.; de la Fuente, R. J. Appl. Microbiol. 2002, 93, 29−35. (6) Keir, L. S.; Marks, S. D.; Kim, J. J. Drug Des., Dev. Ther. 2012, 6, 195−208. (7) Hussein, H. S. J. Anim. Sci. 2007, 85, E63−E72. (8) Jay, M. T.; Garrett, V.; Mohle-Boetani, J. C.; Barros, M.; Farrar, J. A.; Rios, R.; Abbott, S.; Sowadsky, R.; Komatsu, K.; Mandrell, R.; Sobel, J.; Werner, B. Clin. Infect. Dis. 2004, 39, 1−7. (9) Tomoda, H.; Arai, M.; Koyama, N.; Matsui, H.; O̅ mura, S.; Obata, R.; Lee, Y. C. Anal. Biochem. 2002, 311, 50−56. (10) Donohue-Rolfe, A.; Acheson, D. W.; Kane, A. V.; Keusch, G. T. Infect. Immun. 1989, 57, 3888−3893. (11) Fraser, M. E.; Chernaia, M. M.; Kozlov, Y. V.; James, M. N. Nat. Struct. Biol. 1994, 1, 59−64. (12) Fraser, M. E.; Fujinaga, M.; Cherney, M. M.; Melton-Celsa, A. R.; Twiddy, E. M.; O'Brien, A. D.; James, M. N. J. Biol. Chem. 2004, 279, 27511−27517. (13) Tarrago-Trani, M. T.; Storrie, B. Protein Expression Purif. 2004, 38, 170−176. (14) Suzuki, N.; Khoo, K. H.; Chen, H. C.; Johnson, J. R.; Lee, Y. C. J. Biol. Chem. 2001, 276, 23221−23229. (15) Suzuki, N.; Laskowski, M., Jr.; Lee, Y. C. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 9023−9028.

■

CONCLUSIONS We have demonstrated an effective approach for detection and identification of SLT-1B from complex samples using POA− Fe3O4@Al2O3 MNPs as affinity probes in conjunction with MALDI-MS analysis. POA can be simply enriched and modified directly from pigeon egg white proteins using the Fe3O4@Al2O3 MNPs as nanoprobes. The cost and time spent in obtaining purified proteins as the probe molecules are greatly reduced because the traditional time-consuming protein purification steps are skipped. Immobilization of POA on the Fe3O4@Al2O3 MNPs was accelerated by microwave irradiation. Furthermore, this approach is highly sensitive and only requires a microsized volume of sample (20 μL). The selective enrichment of SLT-1 used in this approach involves two steps: the first step is selective trapping of SLT-1 by the POA− Fe3O4@Al2O3 MNPs, and the second step is the displacement of the bound target species by competing Galα(1−4)Gal disaccharide ligands. It only takes ∼10 min to complete the entire analysis for one sample including sample pretreatment (enrichment, rinse, and release of target species, ∼3 min) and MS analysis (sample preparation and sample introduction to the mass spectrometer, ∼7 min). Additionally, the POA− Fe3O4@Al2O3 MNPs can be used to selectively enrich SLT-1 from very complex cell lysates or complex food samples. Thus, the main advantages of this approach include a short analysis time, high sensitivity, high selectivity, low cost, and simple G

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry (16) Lingwood, C. A.; Law, H.; Richardson, S.; Petric, M.; Brunton, J. L.; De Grandis, S.; Karmali, M. J. Biol. Chem. 1987, 262, 8834−8839. (17) Lin, P. C.; Yu, C. C.; Wu, H. T.; Lu, Y. W.; Han, C. L.; Su, A. K.; Chen, Y. J.; Lin, C. C. Biomacromolecules 2013, 14, 160−168. (18) Yosief, H. O.; Iyer, S. S.; Weiss, A. A. Infect. Immun. 2013, 81, 2753−2760. (19) Kale, R. R.; McGannon, C. M.; Fuller-Schaefer, C.; Hatch, D. M.; Flagler, M. J.; Gamage, S. D.; Weiss, A. A.; Iyer, S. S. Angew. Chem., Int. Ed. 2008, 47, 1265−1268. (20) Chien, Y. Y.; Jan, M. D.; Adak, A. K.; Tzeng, H. C.; Lin, Y. P.; Chen, Y. J.; Wang, K. T.; Chen, C. T.; Chen, C. C.; Lin, C. C. ChemBioChem 2008, 9, 1100−1109. (21) Nagatsuka, T.; Uzawa, H.; Sato, K.; Kondo, S.; Izumi, M.; Yokoyama, K.; Ohsawa, I.; Seto, Y.; Neri, P.; Mori, H.; Nishida, Y.; Saito, M.; Tamiya, E. ACS Appl. Mater. Interfaces 2013, 5, 4173−4180. (22) Kulkarni, A. A.; Fuller, C.; Korman, H.; Weiss, A. A.; Iyer, S. S. Bioconjugate Chem. 2010, 21, 1486−1493. (23) Suzuki, N.; Khoo, K. H.; Chen, C. M.; Chen, H. C.; Lee, Y. C. J. Biol. Chem. 2003, 278, 46293−46306. (24) Stromberq, N.; Marklund, B. I.; Lund, B.; Ilyer, D.; Hamers, A.; Gaastra, W.; Karlsson, K. A.; Normark, S. EMBO J. 1990, 9, 2001− 2010. (25) Liu, J.-C.; Tsai, P.-J.; Lee, Y.-C.; Chen, Y.-C. Anal. Chem. 2008, 80, 5425−5432. (26) Liu, J.-C.; Chen, W.-J.; Li, C.-W.; Mong, K.-K.; Tsai, P.-J.; Tsai, T.-L.; Lee, Y.-C.; Chen, Y.-C. Analyst 2009, 134, 2087−2094. (27) Lin, H.-Y.; Chen, W.-Y.; Chen, Y.-C. Anal. Bioanal. Chem. 2009, 394, 2129−2136. (28) Lin, J.-Y.; Chen, Y.-C. Talanta 2011, 86, 200−207. (29) Huang, S.-Y.; Chen, Y.-C. Anal. Chem. 2013, 85, 3347−3354. (30) Chen, W.-Y.; Chen, Y.-C. Anal. Chem. 2007, 79, 2394−2401. (31) Chen, W.-Y.; Chen, Y.-C. Anal. Chem. 2007, 79, 8061−8066. (32) Lee, Y. J.; Jun, K. W.; Park, J. Y.; Potdar, H. S.; Chikate, R. C. J. Ind. Eng. Chem. 2008, 14, 38−44. (33) Sun, L.; Zhang, X.; Zheng, C.; Wu, Z.; Xia, X.; Li, C. RSC Adv. 2012, 2, 9904−9913. (34) Mei, X.; Chen, D.; Li, N.; Xu, Q.; Ge, J.; Li, H.; Yang, B.; Xu, Y.; Lu, J. Soft Matter 2012, 8, 5309−5316. (35) Parida, K. M.; Pradhan, A. C. Ind. Eng. Chem. Res. 2010, 49, 8310−8318. (36) Campbell, K. P.; MacLennan, D. H. J. Biol. Chem. 1981, 256, 4626−4632. (37) Lotan, R.; Skutelsky, E.; Danon, D.; Sharon, N. J. Biol. Chem. 1975, 250, 8518−8523. (38) Bradford, M. M. Anal. Biochem. 1976, 72, 248−254. (39) Georgiou, C. D.; Grintzalis, K.; Zervoudakis, G.; Papapostolou, I. Anal. Bioanal. Chem. 2008, 391, 391−403. (40) Goldman, E. R.; Clapp, A. R.; Anderson, G. P.; Uyeda, H. T.; Mauro, J. M.; Medintz, I. L.; Mattoussi, H. Anal. Chem. 2004, 76, 684− 688.

H

DOI: 10.1021/acs.analchem.5b02712 Anal. Chem. XXXX, XXX, XXX−XXX