Single-Domain Antibody-Nanoparticles: Promising Architectures for

1966

Bioconjugate Chem. 2009, 20, 1966–1974

Single-Domain Antibody-Nanoparticles: Promising Architectures for Increased Staphylococcus aureus Detection Specificity and Sensitivity Shannon Ryan,†,‡ Arnold J. Kell,†,§ Henk van Faassen,‡ Li-Lin Tay,| Benoit Simard,§ Roger MacKenzie,‡,⊥ Michel Gilbert,‡ and Jamshid Tanha*,‡,⊥,# Institute for Biological Sciences, Steacie Institute for Molecular Sciences, and Institute for Microstructural Sciences, National Research Council Canada, Ottawa, Ontario, Canada K1A 0R6, Department of Environmental Biology, Ontario Agricultural College, University of Guelph, Guelph, Ontario, Canada N1G 2W1, Department of Biochemistry, Microbiology and Immunology, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5. Received July 27, 2009

Because antibodies are highly target-specific and nanoparticles possess diverse, material-dependent properties that can be exploited in order to label and potentially identify biomolecules, the development of antibody-nanoparticle conjugates (nanoconjugates) has huge potential in biodiagnostics. Here, we describe a novel superparamagnetic nanoconjugate, one whose recognition component is a single-domain antibody. It is highly active toward its target Staphylococcus aureus, displays long shelf life, lacks cross-reactivity inherent to traditional homologue whole antibodies, and captures a few dozen S. aureus cells in a mixed cell population with ∼100% efficiency and specificity. We ascribe the excellent performance of our nanoconjugate to its single-domain antibody component and recommend it as a general purpose recognition element.

INTRODUCTION An area where antibody-nanoparticle conjugates (nanoconjugates) would be of great benefit is in the field of bacteria detection such as for identifying the presence of potentially harmful bacterial pathogens in food and water and in hospitals and elderly care facilities (1-6). Effective and timely preventive measures against such bacteria require that they be detected in trace amounts, thus calling for a detection system which is highly specific, sensitive, and fast. Superparamagnetic nanoconjugates are an attractive option for bacterial detection. By virtue of their magnetic property, they can magnetically confine/preconcentrate bacterial samples, leading to an increase in detection sensitivity. Though whole antibodies (e.g., IgGs, Figure 1) have been previously employed in a number of biodiagnostic applications to interface superparamagnetic nanoparticles (SPNs) to bacteria, antibodies have an inherent cross-reactivity associated with them because the Fc region is common to different antibodies and can interact with multiple targets despite the fact that their variable region (FV) is unique. For example, Ho et al. recently demonstrated that IgG-based nanoparticles interact with protein A (a surface protein unique to Staphylococcus aureus) as well as other surface markers such as protein M present on the surface of other bacteria (3, 7, 8). To avoid the cross-reactivity problems associated with traditional antibodies, smaller, single-chain variable fragments (scFv) lacking the Fc region can be utilized (Figure 1), but these fragments are frequently unstable and prone to aggregation. For many scFvs, the two binding domains can dissociate and then reassociate intermolecularly; on the surface * To whom correspondence should be addressed. Phone: 613-9907206. Fax: 613-952-9092. E-mail: [email protected]. † S. Ryan and A. J. Kell contributed equally to this work. ‡ Institute for Biological Sciences, National Research Council Canada. § Steacie Institute for Molecular Sciences, National Research Council Canada. | Institute for Microstructural Sciences, National Research Council Canada. ⊥ University of Guelph. # University of Ottawa.

of nanoparticles, this would lead to nanoparticle agglutination. A more attractive approach is to employ single-domain antibodies (sdAbs) (9). sdAbs (e.g., VH, Figure 1) can be readily cloned, resulting in small, chemically/thermally stable, and highly selective targeting moieties (10, 11). Nanoparticle agglutination is not an issue with VHs because of their single domain nature. Additionally, VHs display a lower number of reactive sites, and as a result, unlike larger IgGs or scFvs, very few VHs become linked to more than one nanoparticle during conjugation reactions, greatly reducing conjugation-induced nanoparticle aggregation. Smaller size also translates to higher binding density on the surface of nanoparticles. In this article, we demonstrate that a sdAb-SPN nanoconjugate can interact strongly and selectively with protein A expressed on the surface of S. aureus, leading to its sensitive magnetic confinement even in competition with other bacterial species (12). The nanoconjugate captures a few dozen S. aureus cells in a mixed-cell population with ∼100% efficiency and specificity.

EXPERIMENTAL PROCEDURES Materials. Salmonella typhimurium (ATCC19585), S. aureus (ATCC12598), Streptococcus pyogenes (ATCC12385), and Staphylococcus saprophyticus (ATCC15305) were obtained from American type Culture Collection (Manassas, VA). Streptococcus pneumoniae 35C was from an in-house stock (NRCC 4758). Xylose lysine desoxycholate (XLD), BairdParker (BP) agar, and chocolate agar plates were purchased from Oxoid (Nepean, ON, Canada). (By titration studies, it was shown that the XLD plate only supported the growth of S. typhimurium and not S. aureus and the BP plate only the growth of S. aureus and not S. typhimurium. The titers obtained when S. typhimurium and S. aureus were grown on nutrient broth (NB) and brain heart infusion (BHI, EMD Chemicals Inc., Darmstadt, Germany) media plates, respectively, were the same as those on selective media.) Protein A R-PE conjugate (1 mg/mL) was purchased from Innova Biosciences Ltd. (Babraham, Cambridge, UK). Methods. Synthesis of the Carboxylic Acid-Modified Superparamagnetic Nanoparticle, SPN, and Its Conjugate DeriVatiVe, 1-SPN. The synthesis of the silica encapsulated iron oxide

10.1021/bc900332r CCC: $40.75  2009 American Chemical Society Published on Web 09/14/2009

Single-Domain Antibody-Nanoparticle Architecture

Bioconjugate Chem., Vol. 20, No. 10, 2009 1967

Figure 1. Cartoon representation of an antibody and antibody fragments. The antibody domains are shown as ovals. An antibody, e.g., an IgG, comprises two identical heavy chains (yellow and blue) and two identical light chains (red and green). The Fc region of the IgG can interact indiscriminantly with several different bacterial targets, decreasing specificity. Alternatives to whole antibodies include scFv (a light chain binding domain (VL) plus a heavy chain binding domain (VH) covalently joined by a linker (orange)) and, most simply, VH domain.

nanoparticles has been previously described by us (13). Specifically, a commercially available ferrofluid (0.4 mL, ferrotec EMG304) was diluted to 400 mL with 18 Ω Millipore water. The resulting stock solution was further diluted (12 mL diluted to 45 mL) in Millipore water, and the solution was sonicated for 30 min. Following sonication, the ferrofluid solution was transferred to a 500 mL flask containing 400 mL isopropanol and mechanically stirred. Tetraethoxyorthosilane (0.140 mL) was added to the flask and stirred for ∼1 min at which time 6 mL of concentrated ammonium hydroxide was added. The nanoparticles were then stirred overnight. The resulting nanoparticles were ∼49 ( 7 nm in diameter consisting of a 10 nm iron oxide core and a smooth silicon dioxide shell with a thickness of 20-25 nm. Surface Modification of the Silica Encapsulated Iron Oxide Nanoparticles. Six 45-mL samples of the crude nanoparticle solution described above were transferred to 50 mL Falcon tubes, and 1.0 mL of aminopropyldiethoxymethylsilane (APDEMS) was added to each tube in order to modify the surface with a functional amine group. The mixture was shaken for 24 h and centrifuged at 5300 g for 20 min in order to precipitate the amine-modified nanoparticles. The contents of all of the tubes were combined and washed with ethanol (15 mL) and dimethylformamide (DMF, 15 mL) through repeated centrifugations. Following purification, the amine-modified nanoparticles were dispersed in a 1% solution of succinic anhydride and shaken overnight. The resulting carboxylic acid modified nanoparticles (SPN) were washed with fresh DMF several times to ensure all succinic anhydride had been removed. The nanoparticles were then dispersed in MES buffer (30 mM MES, 70 mM NaCl, pH 6.0) and centrifuged and redispersed in fresh MES buffer. The conversion efficiency for the reaction between the amine and the succinic anhydride was determined by reacting the resulting nanoparticle with fluorescamine, which did not result in a fluorescence response, suggesting that all of the accessible amine moieties had reacted with the succinic anhydride. The resulting SPN were surface modified with HVHP428 VH (1) (12) by first activating SPN (10 mL in MES buffer, 1 × 1014 nanoparticles) with N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC, 5 mg) and N-hydroxysuccinimiude (NHS, 5 mg) to yield the NHS-ester modified SPN. Following purification to remove any unreacted EDC and NHS, the nanoparticles were mixed with 1 (100 µL of a 1.8 mg/mL solution in phosphate-buffered saline (PBS)) and gently vortexed for at least 5 h. It is believed that 1 is covalently bound to the nanoparticle surface through primary amines, of which there are 7 on each antibody. Following modification with 1, the resulting 1-SPN particles were purified via repeated centrifugation cycles and redispersion in MES buffer to remove any

unbound 1. Finally, the nanoparticles were redispersed in 10 mL of MES buffer. In order to elucidate the number of HVHP428 (1) anchored to the surface of each SPN, 1-SPN was reacted with tetramethylrhodamine isothiocyanate (TRITC). The reaction was carried out in MES buffer with 1-SPN (6 × 1012 nanoparticles) and TRITC (50 µg, 1.2 × 10-7 mol) and the resulting TMRmodified 1-SPN was purified and analyzed by UV-visible spectroscopy in order to elucidate the number of sdAbs (1) on the surface of the nanoparticle. The UV-visible absorption spectra of freshly prepared, day-old 1-SPN and one-year-old 1-SPN (6 × 1012 nanoparticles/mL) and the original 1-SPN are presented in the Supporting Information (Figure S1). Analysis of these spectra suggests that TMR has reacted with the sdAb on 1-SPN, where there is an increase in absorption centered at 550 nm in comparison to the unmodified 1-SPN. When the number of TMR molecules was known, the number of sdAb (1) per nanoparticle could be elucidated by assuming that only one of the seven amine groups on 1 is responsible for anchoring it to the surface of the SPN. Construction of a calibration curve for TRITC in MES-buffered water and the comparison of the absorption intensity for the 1-SPN samples suggest that there are 12 and 4 sdAbs per 1-SPN for one-day-old nanoparticles and nanoparticles sitting in MES buffer at pH 6 for 1 year, respectively. This data are also presented in the Supporting Information (Figure S2). The slow loss of 1 over several months is consistent with the slow hydrolysis of silica surface ligands in aqueous solution. Construction and Binding Analysis of Pentamers. Standard cloning techniques were used to construct VH and VL (Figure 1) pentamers, 1P and 2P, respectively (14). 2P serves as the control pentamer. Pentamers were expressed, purified, and quantified as described (14, 15). The binding kinetics for the interaction of 1P with protein A and 2P with protein L were determined from surface plasmon resonance (SPR) data collected with BIACORE 3000 biosensor system (GE Healthcare, Baie d’Urfe´, QC, Canada). For the 1P analysis, 520 resonance units (RUs) of protein A or a Fab reference were immobilized on research grade CM5 sensor chips (GE Healthcare). For the 2P analysis, 680 RUs of protein L or 870 RUs of the Fab reference were immobilized. Immobilizations were carried out at protein concentrations of 25 µg/mL (protein A) or 50 µg/mL (protein L and Fab) in 10 mM sodium acetate buffer pH 4.5, using the amine coupling kit provided by the manufacturer. In all instances, analyses were carried out at 25 °C in 10 mM HEPES, pH 7.4, containing 150 mM NaCl, 3 mM EDTA, and 0.005% P20 at a flow rate of 20 µL/min, and surfaces were regenerated by washing with 50 mM HCl for 3 s. Data were evaluated using BIAeValuation 4.1 software (GE Healthcare).

1968 Bioconjugate Chem., Vol. 20, No. 10, 2009

Association rate constants, kas, were independently calculated from plots of kobs vs concentration. More than one dissociation rate constant, kd, could be calculated due to the heterogeneity in multivalent binding among the pentamer population. Therefore, more than one affinity (KA) could be obtained (see Figure 4 legend). Growth of Cells. A single S. aureus colony from a plate of BHI media was used to inoculate 15 mL of BHI media. The bacteria were grown overnight at 37 °C and 200 rpm. In the morning, the culture was spun down in a fixed rotor, Sorval RT6000B refrigerated centrifuge at 4000 rpm for 10 min, the supernatant was removed, and the cell pellet was resuspended in an appropriate buffer. The cells were respun, the supernatant was removed, and the cell pellet was resuspended followed by measuring cell density at OD600. Serial dilutions of the cells were spread on BHI plates at 37 °C for overnight growth. The cell titer was determined in the morning. An OD600 of 1.0 corresponded to 1 × 108 cells/mL. S. typhimurium was grown as described for S. aureus but with NB media (5 g peptone and 3 g meat extract in 1 L water, pH 7.0) (16). For S. typhimurium, an OD600 of 1 corresponded to 3 × 108 cells/mL. S. saprophyticus was prepared as outlined for S. aureus. An OD600 of 1 corresponded to 1 × 108 cells/mL. S. pyogenes was streaked onto chocolate agar plates from frozen stock and grown at 37 °C overnight. Plate contents were scraped into sterile PBS and adjusted to an OD600 of 1 which corresponded to 1 × 108 cells/ mL. S. pneumoniae was grown in media (Columbia broth 17.5 g, Todd Hewitt broth 15 g, 1% glucose in 1 L water, pH 7.0) at 37 °C, 5% CO2, without shaking. The cells were spun down as stated for S. aureus. An OD600 of 1 corresponded to 1 × 108 cells/mL. Cells were used in microagglutination assays, transmission electron microscopy (TEM), and scanning electron microscopy (SEM) experiments and capture studies. Binding Analysis by Flow Cytometry. Samples of 100 µL 1-SPN or SPN (1012 particles) were blocked in PBS/0.1% (v/v) Tween-20 overnight at 4 °C. To these, 1 µL aliquots of various concentrations of fluorescently active protein A (protein A R-PE conjugate) were added and the reaction volumes were increased by adding 200 µL of PBS/0.75% Tween-20. The samples were incubated for 1 h at 4 °C and were directly used for flow cytometry analysis with a FACSCanto Analyzer (BD Biosciences, Mississauga, ON, Canada). The R-PE fluorescence of 100 000 events per sample was analyzed, and data were generated using BD FACSDiVa software (BD Biosciences). Binding Analysis by Cell Microagglutination. Cell microagglutination assay was performed to find out if 1P and 1-SPN bind multivalently to S. aureus, in other words, find out if the apparent KA values of 1P and 1-SPN are increased compared to that of monomeric 1. If multivalentsand functionals1P and 1-SPN would be expected to cross-link and agglutinate cells. In the absence of agglutination, the cells settle out over time and form a circular dot at the center of the well. If there is agglutination, the cells settle out as a sheet (diffused pattern). Cell microagglutinations were performed essentially as described by Saito et al. (17). Twofold dilutions of pentamers or nanoparticles were performed in MES buffer from wells 1 to 11 of a microtiter plate. Well 12 had only buffer, and the total volume in each well was 50 µL. Subsequently, 1 OD600 unit of cells (see Growth of Cells) in 50 µL buffer was added to all wells, and the plate was incubated overnight at 4 °C. In the morning, photos of the plates were taken for further analysis. Minimum agglutination concentration (MAC), the minimum concentration of nanoparticles or pentamers that promoted bacterial agglutination, was used to express the agglutination activity of nanoparticles and pentamers. Binding Analysis by TEM Experiments. In general, we found that the most effective TEM images of the 1-SPN or SPN

Ryan et al.

interacting with the variety of bacteria could be obtained by extracting ∼1 µL of the solution in the microagglutination plates described above. That is, where the nanoparticles agglutinate effectively, 1 µL of the solution was extracted and expelled directly onto the TEM grid, dried, and analyzed via TEM. In the case where poor or no microagglutination is observed, the 1 µL of solution is extracted from the precipitate that has settled to the bottom of the microtiter plate. Here, because both 1-SPN and SPN are colloidally stable, there are significantly fewer nanoparticles interacting with the bacteria than in the case where agglutination is effective. The TEM images of the nanoparticlecell conjugates were acquired using a Philips CM20 FEG microscope operating at 200 keV. Capture Studies. 1-SPN and SPN were first blocked with 0.1% (v/v) Tween 20 in MES buffer overnight at 4 °C in an Eppendorf tube. One hundred microliters of the blocked nanoparticles (1 × 1012 1-SPN) were mixed with 1 µL of cells in MES buffer. The mixture was incubated at room temperature for 2 h with gentle rocking followed by a 1-h-long capture with a neodynium-based rare earth magnet (Nd0.23Fe0.74B0.03, Lee Valley, Ottawa, ON, Canada). The magnetic is rated at 5000 gauss. Fixing the sample on a homemade magnet stand, the supernatant was removed and plated on BHI media at various dilutions to determine the number of noncaptured cells. The nanoparticles were separated from the magnet, resuspended in 100 µL of MES buffer, and used to determine the number of captured cells. Capture efficiency was calculated as follows: capture efficiency ) (no. of captured cells/[no. of captured cells + no. of non-captured cells]) × 100 In specificity capture studies, 1 µL of each of S. aureus and S. typhimurium cells was added to 99 µL of overnight blocked conjugated or unconjugated nanoparticles (1 × 1012 1-SPN) in MES buffer/1% Tween 20. Incubation, nanoparticle capture, and fractionation of mixtures into captured and noncaptured cells were performed as above. Cell titration was performed as above except that the fractionated cells were plated on both XLD (selective for S. typhimurium) and BP (selective for S. aureus) media. Capture specificity was calculated on the basis relative number of S. aureus and S. typhimurium in the captured fraction, after normalizing for the total number of cells: capture specificity ) no. of S. aureus in the captured fraction/ no. of S. typhimurium in the captured fraction Alternatively, capture specificity was determined by direct comparison of the number of captured S. aureus vs the number of captured S. typhimurium in the mixed cell captured fraction. In some instances, a single wash step between the binding step and the fractionation step was included. The wash step comprised (i) resuspending the nanoparticles in 100 µL MES buffer/1% Tween 20, (ii) capturing the nanoparticles for 30 min, and (iii) removing the wash solution, which was subsequently titrated. Control experiments in which cells (S. aureus alone, S. typhimurium alone, or a mixture of both cells) were incubated with the MES buffer had the same total titer as the nanoparticletreated cells. Binding Analysis by Capture/SEM Experiments. Cell were grown as described above, centrifuged in a benchtop centrifuge (800 g, 5 min), resuspended in 1/2 MES buffer, and pelleted again. Cells were resuspended in 1012 1-SPN in MES buffer/ 0.1% Tween 20 (v/v) and incubated for binding for 30 min. The solutions were magnetically confined for 3 min and the supernatant removedsleaving the last 20 µL behind to ensure the confined materials are not disturbed and lost. The captured materials were resuspended in 200 µL MES buffer. This step was repeated, but the captured materials were resuspended in



Figure 3, there is a gradual shift in the population of fluorescently active nanoconjugate for 1-SPN as a function of protein A R-PE concentration, whereas the nanoparticle without the VH on its surface (SPN) does not show any binding activity (i.e., there is no significant shift in the fluorescence, suggesting the protein A does not interact with the nanoparticles). This demonstrates that 1-SPN effectively interacts with protein A through its sdAb moiety.

Figure 2. Cartoon representations of the pentameric sdAbs (1P), comprising five identical monomeric VH units (blue) assembled on a pentamerization domain, and the multivalent nanoparticles comprising a superparpamagnetic nanoparticle covalently modified with sdAbs (1SPN). The average valency of the year-old 1-SPN preparation used in this study was 4 sdAb.

200 µL sterile double-distilled H2O (ddH2O). The magnetic confinement was performed once more and the captured materials were resuspended in 50 µL ddH2O. One microliter suspensions were used for SEM imaging. SEM was performed on cells plated on Si substrate with a Hitachi S-4700 fieldemission scanning electron microscope. Samples were imaged with an acceleration voltage of 3 KV and at a working distance of 6 mm. Identical experiments were carried out with SNP.

RESULTS Synthesis of 1-SPN Nanoconjugate and Its Binding Activity Assessment by Flow Cytometry. In this investigation, we utilize a silica encapsulated iron oxide nanoparticle (SPN) that can be readily modified with a VH sdAb (HVHP428) (12) previously demonstrated to selectively interact with protein A (1), a surface marker unique to S. aureus bacteria, to generate 1-SPN (Figure 2). Details of the synthesis and modification of the SPN and the conjugation chemistry employed in this investigation are provided in the Methods section. To assess the functionality of 1-SPN, nanoparticles were incubated with fluorophore-conjugated protein A (protein A R-PE) and analyzed in a series of flow cytometry experiments. As highlighted in

Construction of 1P, the Pentameric Version of HVHP428 VH. An attractive feature of sdAbssa consequence of their stabilitysis the flexibility they allow for in terms of choosing optimal conjugation chemistry conditions and nanoparticle purification protocols (10, 11). Though the interactions between sdAbs and their ligand are quite selective, monomeric sdAbs may exhibit low affinities for target ligands. For example, the sdAb utilized in this investigation (1) has an affinity (KA) of 5.6 × 105 M-1 with protein A (12), the surface marker that will allow the modified SPN to interact selectively with S. aureus. Interestingly, the apparent KA, determined by SPR experiments, is increased by at least 1000-fold following the generation of a pentameric version of 1 (1P, Figures 2 and 4) (18). The increase of apparent KA is rationalized through the increased avidity brought on by the multivalent interactions between protein A and 1P. Because 1 is quite small (more than 10× smaller than IgGs), it can be conjugated to the surface of SPNs with a high binding density (>5-fold compared to IgGs and 2-fold compared to scFvs) (Figure 1) and will not promote nanoparticle aggregation as expected with scFvs and IgGs. The resulting 1-SPN is also expected to improve the apparent KA between the SPN-bound 1 and protein A simply because of increased avidity inherent to the now multimeric interactions (i.e., multiple 1 per SPN). To elucidate how the incorporation of the sdAb onto the nanoparticle surface affects the apparent KA between 1 and protein A, the binding of the 1-SPN was benchmarked against 1P by means of a cell microagglutination assay. (Attempts to obtain an apparent KA for the interaction of 1-SPN with protein A by SPR were not successful, presumably because the relatively large size of the nanoparticles prevented

Figure 3. Assessing the functionality of 1-SPN by flow cytometry. Nanoparticles were incubated with 0, 0.05, 0.1, 0.2, 0.5, and 1 µg of fluorescently active protein A (protein A R-PE) and assessed directly for binding by flow cytometry. (A) Binding, indicated as the rightward fluorescence shifts of nanoparticle populations, can be seen with the increase in protein A R-PE concentration. (B) Binding is shown as the mean fluorescence, extracted from data in (A), vs protein A R-PE concentration. Blue circles, 1-SPN; red triangles, SPN.


Ryan et al.

Figure 4. Assessing the binding activity of VH and VL pentamers by surface plasmon resonance. Sensorgram overlays showing the binding of 1P to immobilized protein A at 1, 2, 3, 4, 6, 8, and 10 nM concentrations (A) and of 2P to immobilized protein L at 1, 2, 2.5, 3, 3.5, 4, and 4.5 nM concentrations (B). Association and dissociation rate constants, ka and kd, respectively, and equilibrium constants, KA, for the bindings of 1P and 2P to proteins A and L, respectively, are shown on the graphs. 1P and 2P had minimum KAs of 5 × 108 M-1 and 5 × 109 M-1, respectively. With slower kds, 1P and 2P had KAs as high as 1.1 × 109 M-1 and 1.1 × 1010 M-1, respectively. RU, resonance units.

Figure 5. Representative microagglutination assays, where the concentrations of pentamers and SPNs decrease 2-fold from well 1 to well 11 with well 12 having only MES buffer.

their penetration into the dextran matrix on the sensor chip, making them inaccessible to essentially all of the immobilized protein A). Assessing the Binding Activity of 1-SPN Against 1P by Microagglutination Assays. Microagglutination assays are typically performed by incubating a constant number of target cells (in this case, S. aureus) with decreasing concentrations of an antibody (in this case, either 1P or 1-SPN) along a row of wells in a microtiter plate. In the absence of interactions between 1P (or 1-SPN) and S. aureus, the bacteria simply precipitate into a circular pellet in the bottom of the well (Figure 5, column 12 contains only bacteria cells). However, when 1P or 1-SPN are present and interact with the S. aureus there is a clumping or agglutination characterized by the absence of the circular pellet and the presence of a diffuse pattern in the well. As highlighted in Figure 5A, 1P is capable of agglutinating S. aureus cells up to well 6 (evidenced by the diffuse cell morphology), and beyond well 6, the concentration of 1P is too low for agglutination and the cells sediment as round dots. A control pentamer (2P), whose pentameric nature was confirmed by SPR (Figure 4), does not show any agglutination (Figure 5A), demonstrating that the activity of the 1P is due to its VH binding component and not the pentamerization domains (Figure 2). Interestingly 1-SPN is also capable of strongly interacting with

the S. aureus cells in the agglutination assay, where agglutination occurs up to well 5 (Figure 5B). Note that the starting concentrations of 1P and 1-SPN are 1.0 × 1015 pentamers/mL and 2.5 × 1012 nanoparticles/mL, respectively. There is a 2-fold dilution across the rows, translating to a MAC value of 3 × 1013 pentamers/mL and 1.5 × 1011 particles/mL for 1P and 1-SPN, respectively. That is, 1-SPN binds to protein A expressed on the surface of S. aureus 200 times more strongly than 1P. Importantly, in the absence of 1, SPN itself does not agglutinate with S. aureus under identical conditions. It is also noteworthy that 1-SPN is specific to S. aureus exclusively. As highlighted in Figure 5B, neither 1-SPN nor 1P agglutinate with S. typhimurium under identical conditions to that for the aforementioned S. aureus (only the circular pellet appears in these wells). Finally, the long-term stability of 1-SPN is impressive; the MAC values do not significantly decrease even following a year of storage in a pH 6 MES buffer at 4 °C, presumably due to the multivalency inherent to the four remaining sdAb on the surface of the 1-SPN. Assessing Capture Efficiency and Specificity of 1-SPN. In this series of experiments, 1-SPN is first incubated with S. aureus cells. If effectively and strongly labeled with 1-SPN, the resulting bacteria will be magnetically responsive, making it possible to isolate the cells from bulk solution with a simple


Figure 6. (A) Scheme representing the steps involved in the magnetic capture studies. Capture efficiency and specificity were calculated as described in the Methods section. (B) Capture efficiency of 1-SPN nanoconjugate toward S. aureus. The data are an average of five trials. (C) Capture specificity of 1-SPN in terms of the ratio of the number of S. aureus to S. typhimurium in the captured fraction. This is the average of two trials. S. aureus and S. typhimurium data were normalized for the total number of cells. The inset shows the number of cells (average of the two trials) captured with 1-SPN (white bars) and SPN (black bars). (D) Tabular presentation of results.

rare earth magnet (Figure 6A). In agreement with the microagglutination results, S. aureus cells are readily labeled with 1-SPN and can be magnetically confined. Nonspecific capture could be completely suppressed by coating nanoparticles overnight with 0.1% Tween 20 and performing the binding in the presence of 1% Tween 20. It is noteworthy that such conditions may denature other antibody fragments, but the chemical stability inherent to sdAbs allows them to maintain activity under these conditions. As can be seen in Figure 6 (B,D), 1-SPN is able to label and capture S. aureus with a capture efficiency of over 90% with as few as 180 cells (the number of cells subjected to capture studies ranged from 180-4000). As expected, the capture is mediated by 1, as evidenced by the poor capture efficiency exhibited by SPN (5%). Finally, a competition assay was performed involving several samples containing both S. aureus and S. typhimurium cells.


In these competition assays, mixtures of both bacteria were incubated with 1-SPN and subsequently magnetically confined. The capture selectivity was determined through the enumeration of the captured cells onto selective culture plates (BP plates for S. aureus, XLD plates for S. typhimurium), where the ratio of the S .aureus to S. typhimurium in the captured fraction was used to determine capture specificity. The total numbers of cells in the 1-SPN mixtures were 1170 S. aureus and 525 S. typhimurium cells in one trial and 2780 S. aureus and 2420 S. typhimurium in the second trial. As can be seen in Figure 6C,D, 1-SPN discriminately captures S. aureus over S. typhimurium with a ratio of 13:1. The number of S. typhimurium cells captured by 1-SPN is at the background level and is essentially the same as the numbers captured by SPN in the case of both bacteria (Figure 6C, inset). The capture selectivity could be further improved simply by including a wash step immediately following the binding step. As highlighted in Figure 7A,C, the capture efficiency was as high as 98% for 1-SPN vs only 0.8% for the SPN. (The total number of S. aureus cells was 338 in the 1-SPN mixture and 278 in the SPN mixture; the wash step removed 6 S. aureus from the 1-SPN mixture and 24 from the SPN mixture.) Without the wash step, 1-SPN capture efficiency would have remained at exactly the same level, but that of SPN would have been as high as 9%. This demonstrates that there is very little nonspecific absorption contributing to the interaction between 1-SPN and the S. aureus cells. In the case of specificity capture studies, we used a mixture of cells with as few as 80 S. aureus cells and 56 S. typhimurium cells (Figure 7B,C). The capture specificity was adequate (80 S. aureus/4 S. typhimurium) immediately following incubation and magnetic confinement and extremely high following the washing step after confinement (78 S. aureus and no S. typhimurium; the wash step removed 2 S. aureus and 4 S. typhimurium). Thus, without the wash step the capture specificity ratio would have been significantly lower. Assessing the Binding Specificity of 1-SPN by Capture Studies and Electron Microscopy. The binding (or lack of interaction) of both 1-SPN and SPN to S. aureus cells could also be qualitatively assessed by means of TEM analysis. Examples of these TEM binding experiments are provided in Figure 8. From these images, two factors are clear. First, 1-SPN interacts much more effectively over the entire surface of the S. aureus cells (Figure 8A) as opposed to SPN (Figure 8B), indicating that the sdAb mediates the specific interaction between the nanoparticles and the S. aureus cells. Second, in agreement with the microagglutination and capture specificity investigations highlighted above, the TEM analysis suggests that there is very little interaction between 1-SPN and the S. typhimurium cells (Figure 8C). Also as expected, there was essentially no interaction between SPN and S. typhimurium (data not shown). Previously an IgG-modified SPN conjugate was reported to interact with S. aureus, S. saprophyticus, and S. pyogenes (group A Streptococcus) through the Fc-region of the antibody (Figure 1) and surface ligands on the various bacteria (3). It is expected that such constructs may also cross-react with group C and G Streptococci through interaction with bacterial protein G (19). Because specificity/selectivity is an important characteristic that is required for most biodiagnostic applications, we also assessed the specificity of 1-SPN through capture/SEM experiments with S. saprophyticus, S. pyogenes, and S. pneumoniae (a protein G-containing Streptococcus species of bacteria). Mixtures of nanoparticles (1012) and cells (108) were incubated for binding and were subsequently subjected to magnetic confinement. While in the case of 1-SPN/S. aureus, the solution became clear in a couple of minutes following the magnetic confinement with cells clumped against the vessel wall, indicating a very strong


Ryan et al.

Figure 7. Cell capture efficiency and specificity of 1-SPN nanoconjugate when a single wash step is included after the binding step. Capture specificity is shown by comparing the number of S. aureus vs S. typhimurium in the mixed capture fraction. (C) Tabular presentation of results.

Figure 8. Representative TEM images of the 1-SPN-bacteria conjugates where there is a strong interaction between 1-SPN and S. aureus cells (A) in contrast to the inefficient labeling of S. typhimurium by the same nanoparticles (C). Control TEM images were also acquired confirming that SPN does not interact effectively with S. aureus cells in the absence of 1 (B).

Figure 9. Assessing the cross-reactivity of 1-SPN toward S. saprophyticus, S. pneumoniae, and S. pyogenes by SEM experiments. Cells and nanoparticles are shown by black and white arrows, respectively. The scale bar corresponds to 1 µm.

binding interaction between 1-SPN and S. aureus, in the case of SPN/S. aureus, the solution stayed cloudy over the course of the confinement and there was no visible captured cells. With S. pyogenes and S. pneumoniae, both 1-SPN and SPN hardly gave any captured cells, and by the end of the two washes (cycles of resuspension and confinement), there was no visible pack of confined cells. Only in the case of S. saprophyticus was some capture observed, albeit with a significant amount of cells still in the supernatant fraction. However, the capture was the same for both 1-SPN and SPN, indicating some nonspecific interactions between SPN and S. saprophyticus. An aliquot of the confined materials following the last wash was used for SEM imaging.

Figure 9 summarizes results of the capture studies in a series of SEM images. The top panels of Figure 9 depict the interaction of cells with 1-SPN, while bottom panels are the control experiments performed with SPN. In the S. aureus interaction with 1-SPN capture study, SEM images showed each of the captured cells (dark sphere, ∼1 µm) are flanked with 1-SPN with most cells partially covered by the NP aggregates, which indicates positive binding response between cells and 1-SPN. On the contrary, in the control experiment (S. aureus vs SPN), very large aggregates of SPN were observed with a few cells embedded in the SPN aggregates. Unlike the S. aureus vs 1-SPN interaction, the captured cells in the control experiment have no SPN on their surface, nor were they flanked by the


nanoparticles. This suggests the lack of interaction between the control nanoparticles and S. aureus. It is possible that the few cells trapped in the control experiment were from the remaining 20 µL of the solution left after the capture step rather than magnetically confined. SEM images of the interaction between 1-SPN with S. pyogenes and S. pneumoniae showed very similar patterns. Both images showed clusters of cells without any nanoparticles binding to their surfaces nor flanked in between the cells, which suggests the non-interactive nature of the 1-SPN to both S. pyogenes and S. pneumoniae. Control experiments of both species exposed to SPN also showed a similar noninteraction pattern. Despite the observation of very large aggregates of SPN in the images of the control study, there were no nanoparticles binding to the surface of the cells as observed in the case of S. aureus vs 1-SPN image. Since, in the case of both S. pyogenes and S. pneumoniae, there is no binding seen in the images as well as during the earlier confinement step, it is very likely that the cells seen in the images are carryover from the nonbinding fraction (see above and Experimental Procedures). We did, however, observe some interaction between both 1-SPN and SPN with S. saprophyticus as suggested by Figure 9. In both S. saprophyticus vs 1-SPN and S. saprophyticus vs SPN images, cells are flanked and partially covered by the nanoparticles, indicating a positive interaction between the nanoparticles and the cell. Because this is observed in both experiments, it is likely that the interactions are nonspecific, a conclusion that is also supported by the results of the capture step above. It is possible that such nonspecific interaction can be further reduced by passivating the nanoparticles surfaces with agents that suppress nonspecific interactions such as serum albumins. It is also very likely that the interaction between 1-SPN and S. saprophyticus could be distinguished from the interaction between 1-SPN and S. aureus in a real detection situation simply by looking at all of these interactions together.

DISCUSSION Protein A is (i) highly characteristic of S. aureus with some 99% of S. aureus strains having protein A on their surface (20, 21) and (ii) highly abundant on the surface of S. aureus (22). These characteristics make protein A an attractive marker for effective and specific detection of S. aureus (3, 21, 23-25). In fact, the commercially available latex agglutination test, which is a standard method for S. aureus detection, is based on protein A-antibody interaction (23, 25). Recently, a gold nanoparticlebased immunochromatographic test developed for S. aureus detection and based on IgG-protein A interaction gave an impressive sensitivity of 100% for 130 S. aureus strains and specificities of 94.7% and 100% for S. aureus when also tested against 19 and 17 Staphylococus spp. (non-S. aureus) and nonStaphylococci strains, respectively (25). The latex agglutination test, which operates on the same recognition principle as the immunochromatographic test, gave identical results (25). The one false positive seen in the case of S. xylosus in the Staphylococus spp. group was presumably due to the presence of protein G on the surface of S. xylosus, which can cross-react with antibodies. This protein G-based cross-reactivity should not exist in the case of our nanconjugate, a conclusion supported by the lack of interaction between 1-SPN and the protein G-displaying S. pneumoniae (Figure 9). Thus, on the basis of the above facts, the highly selective nature of the antibody-protein A pseudoimmune interaction (see Figure 1) and our own specificity/cross-reactivity studies, we expect 1-SPN architecture to be at least as specific as the previously described detection systems, which rely on whole antibody-protein A interactions. A potential advantage to the approach presented here involves the logical extension to dye or quantum dot-labeled superpara-


magnetic nanoparticles, which in addition to target preconcentration would provide detection signals. Alternatively, the detecting signals can be provided with SERS (surface enhanced Raman scattering)-active sdAb nanoconjugates (10) when used in combination with the present superparamagnetic nanoconjugates. Standard clinical detection methods, e.g., latex agglutination test for S. aureus, are time-consuming (takes a few days to obtain results) and laborious, as they involve steps such as plating and bacterial enrichment (21, 26). Detection platforms based on the present concept, besides offering better specificity (see above), would be fast and simple to perform. We have shown that it takes as little as 10 min to capture S. aureus with 1-SPN (unpublished results) and no more than 10 min to detect the captured S. aureus by SERS-active sdAb conjugates, with single cell sensitivity (28). Additional advantages inherent to nanoparticle-based targeting agents include superior labeling efficiencies owing to large surface area to volume ratios and the fact that the nanoparticles are colloidally stable and are small, so binding kinetics should be significantly improved in comparison to micrometer-sized analogues (27). On the basis of the performance of the present nanoconjugates and previously described SERS-active sdAb-silver nanoparticles (10) and the aforementioned properties of sdAbs, we strongly recommend sdAbs over other antibody formats as the recognition moiety of nanoparticles and nanowires.

CONCLUSION In summary, sdAbs are highly stable, active, and specific agents capable of mediating interactions between superparamagnetic nanoparticles and a few tens of pathogenic bacteria in mixed cell populations with exceptionally high specificity and efficiency. The use of sdAb eliminates much of the crossreactivity previously reported for nanoconjugates utilizing whole antibodies to interface the nanoparticles with the bacteria (3). These superparamagnetic nanoconjugates owe their excellent performance to a combination of multivalency and the unique properties of its sdAb component. A previously reported S. aureus-specific silver nanoconjugate which also relied on a sdAb for its recognition also performed very well (10), and it is expected that a combination of such nanoparticles could lead to very effective magnetic confinement and optical detection scaffolds for sensitive pathogen detection.

ACKNOWLEDGMENT This is National Research Council of Canada Publication 50005. We acknowledge Rebecca To for constructing pentameric single-domain antibodies, Tomoko Hirama for performing surface plasmon resonance experiments, and Tom Devecseri for preparing publication quality figures. We are indebted to Catherine Bibby for acquiring the TEM images. We also thank NRC-NSC-ITRI for funding. Supporting Information Available: Description of experimental procedures for quantifying the number of sdAbs on the nanoparticles and representative TEM images of the 1-SPN nanoparticles. This material is available free of charge via the Internet at http://pubs.acs.org.

LITERATURE CITED (1) Gu, H., Xu, K., Xu, C., and Xu, B. (2006) Biofunctional magnetic nanoparticles for protein separation and pathogen detection. Chem. Commun. (Camb.) 941–949. (2) Kalele, S. A., Kundu, A. A., Gosavi, S. W., Deobagkar, D. N., Deobagkar, D. D., and Kulkarni, S. K. (2006) Rapid detection of Escherichia coli by using antibody-conjugated silver nanoshells. Small 2, 335–338.

1974 Bioconjugate Chem., Vol. 20, No. 10, 2009 (3) Ho, K. C., Tsai, P. J., Lin, Y. S., and Chen, Y. C. (2004) Using biofunctionalized nanoparticles to probe pathogenic bacteria. Anal. Chem. 76, 7162–7168. (4) Su, X. L., and Li, Y. (2004) Quantum dot biolabeling coupled with immunomagnetic separation for detection of Escherichia coli O157:H7. Anal. Chem. 76, 4806–4810. (5) Tu, S.-I., Uknalis, J., Gore, M., Irwin, P., and Feder, I. (2003) Factors affecting the bacterial capture efficiency of immuno beads: A comparison between beads with different size and density. J. Rapid Methods Autom. Microbiol. 11, 35–46. (6) Tu, S.-I., Uknalis, J., Gore, M., and Irwin, P. (2002) The capture of Escherichia coli O157:H7 for light addressable potentiometric sensor (LAPS) using two different types of magnetic beads. J. Rapid Methods Autom. Microbiol. 10, 185–195. (7) Sjoquist, J., Movitz, J., Johansson, I. B., and Hjelm, H. (1972) Localization of protein A in the bacteria. Eur. J. Biochem. 30, 190–194. (8) Burova, L. A., Nagornev, V. A., Pigarevskii, P. V., Gladilina, M. M., Molchanova, I. V., Selivestrova, V. G., Tern, A., Lindahl, G., and Totolian, A. A. (1999) Streptococcal IgG Fc-binding proteins - initiation factors of experimental glomerulonephritis. Biull. Eksp. Biol. Med. 128, 548–552. (9) Holliger, P., and Hudson, P. J. (2005) Engineered antibody fragments and the rise of single domains. Nat. Biotechnol. 23, 1126–1136. (10) Tay, L., Tanha, J., Ryan, S., Veres, T. (2007) Detection of Staphylococci aureus cells with single domain antibody functionalized Raman nanoparobes. Photonics North (Armitage, J., Ed.) Proceedings of the SPIE, 6796: 67960C. (11) Arbabi-Ghahroudi, M., Tanha, J., and MacKenzie, R. (2005) Prokaryotic expression of antibodies. Cancer Metastasis ReV. 24, 501–519. (12) To, R., Hirama, T., Arbabi-Ghahroudi, M., MacKenzie, R., Wang, P., Xu, P., Ni, F., and Tanha, J. (2005) Isolation of monomeric human VHs by a phage selection. J. Biol. Chem. 280, 41395–41403. (13) Kell, A. J., and Simard, B. (2007) Vancomycin architecture dependence on the capture efficiency of antibody-modified microbeads by magnetic nanoparticles. Chem. Commun. 1227– 1229. (14) Zhang, J., Li, Q., Nguyen, T.-D., Tremblay, T. T. L., Stone, E., To, R., Kelly, J., and MacKenzie, C. R. (2004) A pentavalent single-domain antibody approach to tumor antigen discovery and the development of novel proteomics reagents. J. Mol. Biol. 341, 161–169. (15) Pace, C. N., Vajdos, F., Fee, L., Grimsley, G., and Gray, T. (1995) How to measure and predict the molar absorption coefficient of a protein. Protein Sci. 4, 2411–2423.

Ryan et al. (16) Sambrook, J., Fritsch, E. F., Maniatis, T. Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory: Cold Spring Harbor, NY, 1989. (17) Saito, T., Kawabata, S., Hirata, M., and Iwanaga, S. (1995) A novel type of limulus lectin-L6. Purification, primary structure, and antibacterial activity. J. Biol. Chem. 270, 14493–14499. (18) Zhang, J., Tanha, J., Hirama, T., Khieu, N. H., To, R., TongSevinc, H., Stone, E., Brisson, J. R., and MacKenzie, C. R. (2004) Pentamerization of single-domain antibodies from phage libraries: A novel strategy for the rapid generation of high-avidity antibody reagents. J. Mol. Biol. 335, 49–56. (19) Tashiro, M., and Montelione, G. T. (1995) Structures of bacterial immunoglobulin-binding domains and their complexes with immunoglobulins. Curr. Opin. Struct. Biol. 5, 471–481. (20) Forsgren, A. (1970) Significance of protein A production by Staphylococci. Infect. Immun. 2, 672–673. (21) Chang, Y. H., Chang, T. C., Kao, E. F., and Chou, C. (1996) Detection of protein A produced by Staphylococcus aureus with a fiber-optic-based biosensor. Biosci. Biotechnol. Biochem. 60, 1571–1574. (22) Kronvall, G., Quie, P. G., and Williams, R. C., Jr. (1970) Quantitation of Staphylococcal protein A: Determination of equilibrium constant and number of protein A residues on bacteria. J. Immunol. 104, 273–278. (23) Fukuda, S., Tatsumi, H., Igarashi, H., and Igimi, S. (2000) Rapid detection of Staphylococcus aureus using bioluminescent enzyme immunoassay. Lett. Appl. Microbiol. 31, 134–138. (24) Mongodin, E., Bajolet, O., Hinnrasky, J., Puchelle, E., and de Bentzmann, S. (2000) Cell wall-associated protein A as a tool for immunolocalization of Staphylococcus aureus in infected human airway epithelium. J. Histochem. Cytochem. 48, 523–534. (25) Huang, S. H. (2006) Gold nanoparticle-based immunochromatographic test for identification of Staphylococcus aureus from clinical specimens. Clin. Chim. Acta 373, 139–143. (26) Zhao, X., Hilliard, L. R., Mechery, S. J., Wang, Y., Bagwe, R. P., Jin, S., and Tan, W. (2004) A rapid bioassay for single bacterial cell quantitation using bioconjugated nanoparticles. Proc. Natl. Acad. Sci. U.S.A. 101, 15027–15032. (27) Gu, H. W., Xu, K. M., Xu, C. J., and Xu, B. (2006) Biofunctional magnetic nanoparticles for protein separation and pathogen detection. Chem. Commun. 941–949. (28) Huang, P.-J., Tay, L.-L., Tanha, J., Ryan, S., Chau, L.-K. (2009) Single-domain antibody-conjugated nanoaggregateembedded beads for targeted detection of pathogenic bacteria. Chemistry DOI: 10.1002/chem.200901397. BC900332R

Single-Domain Antibody-Nanoparticles: Promising Architectures for

Recommend Documents