Periplasmic Binding Protein-Based Detection of Maltose Using

Feb 14, 2013 - monitoring conformational shifts, we have instead utilized the maltose binding protein (MBP) in lieu of an antibody in an ELISA...
2 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Periplasmic Binding Protein-Based Detection of Maltose Using Liposomes: A New Class of Biorecognition Elements in Competitive Assays Katie A. Edwards and Antje J. Baeumner* Cornell University, Department of Biological and Environmental Engineering, 140 Riley-Robb Hall, Ithaca, New York 14853, United States S Supporting Information *

ABSTRACT: A periplasmic binding protein (PBP) was investigated as a novel binding species in a similar manner to an antibody in a competitive enzyme linked immunosorbent assay (ELISA), resulting in a highly sensitive and specific assay utilizing liposome-based signal amplification. PBPs are located at high concentrations (10−4 M) between the inner and outer membranes of gram negative bacteria and are involved in the uptake of solutes and chemotaxis of bacteria toward nutrient sources. Previous sensors relying on PBPs took advantage of the change in local environment or proximity of site-specific fluorophore labels resulting from the significant conformational shift of these proteins’ two globular domains upon target binding. Here, rather than monitoring conformational shifts, we have instead utilized the maltose binding protein (MBP) in lieu of an antibody in an ELISA. To our knowledge, this is the first PBP-based sensor without the requirement for engineering site-specific modifications within the protein. MBP conjugated fluorescent dye-encapsulating liposomes served to provide recognition and signal amplification in a competitive assay for maltose using amylose magnetic beads in a microtiter plate-based format. The development of appropriate binding buffers and competitive surfaces are described, with general observations expected to extend to PBPs for other analytes. The resulting assay was specific for D-(+)-maltose versus other sugar analogs including D-(+)-raffinose, sucrose, D-trehalose, D(+)-xylose, D-fructose, 1-thio-β-D-glucose sodium salt, D-(+)-galactose, sorbitol, glycerol, and dextrose. Cross-reactivity with Dlactose and D-(+)-glucose occurred only at concentrations >104-fold greater than D-(+)-maltose. The limit of detection was 78 nM with a dynamic range covering over 3 orders of magnitude. Accurate detection of maltose as an active ingredient in a pharmaceutical preparation was demonstrated. This method offers a significant improvement over existing enzymatic detection approaches that cannot discriminate between maltose and glucose and over existing fluorescence resonance energy transfer (FRET)-based detection methods that are sensitivity limited. In addition, it opens up a new strategy for the development of biosensors to difficult analytes refractory to immunological detection.

P

binding protein (MBP). MBP is a monomeric 42.5 kDa periplasmic binding protein with a moderate affinity for maltose (KD ∼1 μM) and high association rate constant (2.3 × 10−7 M−1s−1), leading to a rapid binding response.11 Often utilized as a labeling site for exogenous fluorophores in MBP is the side chain of residue 337, which undergoes a significant change in polarity upon the conformational shift. The local environment changes from hydrophilic to hydrophobic upon maltose binding, concomitant with, for example, a significant enhancement in fluorescence intensity (up to 180%) of an environmentally sensitive label in the presence of maltose.10 Other PBP-based assays have relied on fluorescence resonance energy transfer (FRET) to afford fluorescence quenching upon transfer of energy from donor to acceptor.12,13 FRET pairs have included two dyes,12,13 quantum dots and dyes,14 and green fluorescent protein (GFP) variants.15 Such assays have

eriplasmic binding proteins (PBPs) are located at high concentrations (10−4 M) between the inner and outer membranes of gram negative bacteria such as Escherichia coli and Salmonella typhimurium.1 They are involved in the uptake of solutes through coupling to transporters within the inner membrane serving to mediate their transport or chemotaxis toward nutrient sources.2 Examples of target compounds include inorganic ions, amino acids, peptides, and sugars.3−6 Characteristic of these protein structures are two globular domains connected by a hinge region, which undergoes a significant conformational change between its closed form when the ligand is bound deep within the cleft between the two domains and its open form when ligand is unbound.7 PBPs have been used as biorecognition elements in various biosensors, the majority of which take advantage of the significant conformational shift upon target binding.8 Such shifts place environmentally sensitive, site-specific fluorophore labels in different local environments, resulting in either an increase or decrease in fluorescence upon target binding.9,10 One often utilized periplasmic binding protein is the maltose © 2013 American Chemical Society

Received: November 11, 2012 Accepted: January 14, 2013 Published: February 14, 2013 2770

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

Figure 1. Competitive assay for maltose utilizing maltose binding protein (MBP)-tagged fluorescent dye-encapsulating liposomes in a similar manner to enzyme labeled antibodies in an ELISA. (a) The sample under investigation is mixed with amylose magnetic beads and dye-encapsulating liposomes (grey) tagged with MBP (green). The liposomes competitively bind to the magnetic beads (brown) or to maltose (blue) present in the sample during a 1 h incubation period. Depicted here is the presence (top) or absence (bottom) of maltose in the sample. (b) The microtiter plate is then placed over a 96-well magnet, drawing magnetic beads and any bound complexes to the base of the wells. (c) Unbound materials are washed away, and the liposomes remaining bound are lysed with surfactant to release their encapsulated dye (magenta). In the absence of maltose in the sample, more liposomes bind to the amylose magnetic beads resulting in greater signal intensity (bottom). In the presence of maltose in the sample (top), the MBP on the liposome surface is occupied by maltose, resulting in fewer interactions with the amylose magnetic beads and a signal inversely proportional to the maltose concentration.

linkage by α-D-glucosidase, generating two glucose molecules per input maltose molecule. The glucose molecules subsequently serve as substrates for glucose oxidase yielding a product suitable for colorimetric or fluorescence detection.19,20 Commercially available kits for this purpose reportedly detect between 10 pmol and 10 nmol of glucose per assay in a volume of 50 μL, yielding an effective limit of detection of 0.2 μM, but such enzyme-based approaches cannot inherently differentiate between maltose and glucose.19,20 Immunoassay approaches are not appropriate for many target molecules, including maltose, due to a lack of suitable antibodies.21,22 Interest in alternative biorecognition elements has increased as antibody development

demonstrated the unique capabilities of PBPs as sensing reagents, with limited cross-reactivity toward molecules other than their target analytes.16 These homogeneous assays have been reported to detect concentrations of maltose down to 50− 100 nM.13 However, drawbacks to FRET-based detection include the need for chemical modification or genetic engineering of the protein and reliance on ratiometric measurements, both of which can add complexity and expense in bioanalytical sensor development; high background; limited dynamic range; and often poor sensitivity.17,18 In commercial assays, maltose detection is frequently accomplished via enzymatic cleavage of its α(1→4)-glycosidic 2771

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

NJ). All other reagents were molecular biology grade and purchased from VWR Scientific (West Chester, PA) or SigmaAldrich Corp. (St Louis, MO). Fluorescence and UV/visible measurements were made using FLX800 and PowerWave XS microtiter plate readers, respectively (Bio-Tek Instruments, Winooski, VT). The procedures for liposome preparation,35 protein conjugation,35 and determination of the concentration of MBP on the liposome surface and liposome stability are provided in the Supporting Information. Evaluation of Competitive Surfaces. To generate sugarfunctionalized competitive surfaces, two covalent approaches targeting the sugars’ aldehyde groups were employed: sodium cyanoborohydride reduction and hydrazide formation of hydrazone bond. Subsequently, a direct-binding assay for the immobilized sugars by MBP was developed. For sodium cyanoborohydride reduction, bovine serum albumin (BSA) was diluted to 10 μg/mL in phosphate buffered saline (PBS: 137 mM NaCl, 2.7 mM KCl, 10 mM phosphate buffer, pH 7.4), and 100 μL was added to wells of black Maxisorp microtiter plates. The plates were then sealed with sealing film and stored overnight at 4 °C. Unbound protein was then removed, and the wells were washed with 200 μL of PBS prior to adding amylopectin, maltose, maltotriose, or maltoheptaose at 0−500 μg/mL in 200 mM sodium phosphate, pH 9.0, containing 1 mg/mL sodium cyanoborohydride. The plates were sealed at 37 °C with gentle shaking for 72 h, washed with 200 μL of PBS, then blocked with 0.1% (w/v) bovine serum albumin (BSA), 0.05% (w/v) Tween-20 in tris buffered saline (TBS: 20 mM Tris, 150 mM NaCl, 0.01% (w/v) NaN3), and sealed at 22 °C for 1 h (200 μL/well). The procedure for functionalizing carbohydrate binding microtiter plates with sugars was adapted from the manufacturer’s instructions for functionalization with antibodies.36 Amylose and amylopectin were dissolved at 20 mg/mL in dimethyl sulfoxide (DMSO) containing 50 μM sodium chloride and then further diluted to 500 μg/mL in 100 mM sodium acetate, pH 4.0, containing 75 μM sodium meta-periodate. The same procedure was carried out for maltose and maltotriose, though the initial dilution was made in water. The oxidation proceeded for 30 min at ambient temperature; then, the contents were further diluted to 1, 10, and 100 μg/mL with 100 mM sodium acetate, pH 5.5. A 100 μL aliquot was then added to the wells of a Carbobind microtiter plate, and the plate was sealed at 22 °C with gentle shaking for 1 h. The plate wells were washed with 200 μL of PBS containing 0.05% (w/v) Tween-20, then blocked with 0.1% (w/v) bovine serum albumin (BSA), 0.05% (w/v) Tween-20 in TBS, and sealed at 22 °C for 1 h (200 μL/well). For both types of microtiter plates, the solution was then removed and the wells were washed with 2 × 200 μL of PBS, then 2 × 200 μL of 50 mM 4-morpholineethanesulfonic acid sodium salt (MES), pH 6.5 containing 0.2 M sodium chloride and 0.01% (w/v) sodium azide (MESS). MBP conjugated liposomes diluted to 25 μM phospholipid or unconjugated MBP diluted to 1 μg/mL in MESS were added to all wells. The plates were sealed and incubated for 1 h at 22 °C on a vortex and then washed with 3 × 200 μL of MESS. For unconjugated MBP, biotinylated anti-MBP (1:10 000) was diluted in 50 mM MES, pH 6.5, 0.2 M sodium chloride, and 0.01% (w/v) sodium azide containing 0.1% (w/v) BSA (MESSB), and then, 100 μL was added to the washed wells. The plates were sealed and incubated for 1 h at 22 °C on a vortex, washed with 2 × 200 μL of MESS, and then 200 μL of TBS. The conditions for

toward certain molecules is often hindered by their lack of remarkable antigenic features, high toxicity to the host, or low affinity or unacceptable cross-reactivity of resulting antibodies toward closely related structures.23 A method using PBPs, which can be engineered against otherwise refractory targets of interest,24 thus would greatly expand the toolkit of assay developers. Dye-encapsulating liposomes have been used in lieu of enzymes, fluorophores, latex beads, or colloidal gold when tagged with biorecognition elements such as DNA probes, aptamers, gangliosides, antibodies, and streptavidin in microfluidic, lateral flow, microtiter plate, and microscopy applications.25−29 The advantages of liposomes for analytical purposes include the flexibility in lipid selection to allow for conjugation to or incorporation of both hydrophilic and hydrophobic biorecognition elements, the incorporation of hundreds of thousands of fluorescent dye molecules within their aqueous cores providing assay sensitivity, instantaneous signaling through surfactant induced release of encapsulant without need for time-dependent substrate conversion, and the protective nature of the interior toward encapsulants conferring increased long-term stability.28−32 Here, we sought to investigate PBPs as binding entities in a similar manner to antibodies used in immunosorbent assays, rather than relying on the detection of their conformational changes. To our knowledge, this is the first PBP-based sensor without the requirement for engineering site-specific modifications within the protein. We took advantage of the unique attributes of both species, combining the sensitivity and versatility of liposomes with the specificity of MBP to develop a sensitive competitive assay for maltose with fluorescence detection. This assay relied on amylose-functionalized magnetic beads to serve as a solid phase surface to compete with solution phase maltose for MBP-tagged liposomes encapsulating the fluorescent dye sulforhodamine B (SRB) (Figure 1). This method was employed to quantify maltose as an active ingredient in a pharmaceutical preparation used for the treatment of dry mouth.33,34

■

MATERIALS AND METHODS Maltose binding protein was purchased from VLI Research (Malvern, PA). Amylose magnetic beads were purchased from New England Biolabs, Inc. (Ipswich, MA). 1,2-Dipalmitoyl-snglycero-3-phosphocholine (DPPC), 1,2-dipalmitoyl-sn-glycero3-[phospho-rac-(1-glycerol)], sodium salt (DPPG), N-glutaryl 1,2-dipalmitoyl-sn-glycero-3-phosphatidylethanolamine (N-glutaryl-DPPE), and the extrusion membranes were purchased from Avanti Polar Lipids (Alabaster, AL). Sulforhodamine B (SRB) dye and streptavidin were purchased from Invitrogen (Carlsbad, CA). Medium binding polystyrene plates were manufactured by Axygen (Union City, CA). Fluorescein diphosphate (FDP) was purchased from Anaspec (San Jose, CA). Black Maxisorp high-binding polystyrene microtiter plates were purchased from Nunc (Roskilde, Denmark.) The Costar clear medium binding and the Carbobind carbohydrate binding 96-well microtiter plates were purchased from Corning, Inc. (Corning, NY). Biotinylated anti-MBP was purchased from Abcam (Cambridge, MA), and alkaline phosphatase-streptavidin (AP-StAv) was purchased from Rockland Immunochemicals, Inc. (Gilbertsville, PA). The potassium phosphate monobasic standard was purchased from Ricca Chemical Company (Arlington, TX). All buffers were prepared with HPLC grade water, manufactured by JT Baker (Phillipsburg, 2772

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

enzymatic amplification were as developed previously.28 Briefly, streptavidin alkaline phosphatase (StAv-AP) was diluted to 1:10 000 in 0.1% (w/v) BSA in TBS, and 100 μL was added per well. The plates were sealed and incubated for 1 h at 22 °C on a vortex and then washed with 3 × 200 μL of TBS. Fluorescein diphosphate (FDP) was diluted to 12.5 μM in 50 mM Tris containing 1 mM MgCl2 and 50 mM NaCl, pH 9.0; 100 μL was added to all wells, and then, the plates were read every 10 min for 30 min following 10 s of shaking at intensity level 1 (18 Hz) using a fluorescence plate reader with λex = 485/20 nm and λem = 528/20 nm. For MBP-tagged liposomes, the wells were washed with 3 × 200 μL of MESS; then, those liposomes remaining bound were lysed with 50 μL of 30 mM n-octyl-β-Dglucopyranoside (OG), and resulting fluorescence intensities were read at a gain of 65 with λex = 540/35 nm and λem = 590/ 20 nm. Effects of Buffer Composition on MBP Binding. To determine the effects of buffer composition on MBP binding, a competitive assay between solution phase maltose and amylose functionalized microtiter plates for free MBP was developed. Amylose coated microtiter plates were prepared at 100 μg/mL as described above, except the final plate washes were made in water rather than MESS. Maltose was diluted in buffers at various pH, salt content, and other additives (see Supporting Information) and added to every other row with remaining wells containing the respective buffers only. MBP was added such that its final concentration was 1 μg/mL and the final concentration of maltose was 1.333 μM. The plates were sealed and incubated for 1 h at 22 °C on a vortex and then washed with 3 × 200 μL of MESS. Signal development using biotinylated anti-MBP and StAv-AP with FDP and reading proceeded as described above. Competitive Assay for Maltose. Medium binding polystyrene microtiter plates were washed with PBS (200 μL/well), then blocked with 0.1% (w/v) BSA, 0.05% (w/v) Tween-20 in PBS, and sealed at 22 °C for 1 h (200 μL/well). The solution was then removed, and the wells were washed with 2 × 200 μL of PBS and then tapped dry onto several layers of Kimwipes. Three microliters of amylose magnetic beads were added to each well; then, dilutions of maltose (6.25 nM to 6.25 mM) in MESS were added (50 μL/well). MBP tagged SRBencapsulating liposomes diluted to 50 μM phospholipid in MESSB (50 μL/well) were then mixed with the well contents, plate sealed, and incubated for 1 h at 22 °C on a vortex. The plate was placed over a 96-well magnet, and unbound materials were carefully removed by aspiration. The remaining beads were washed with 3 × 200 μL of MESS in the same manner. Liposomes remaining bound were lysed with 100 μL of 30 mM OG; then, 50 μL of the well mixed supernatant was transferred to a black polystyrene microtiter plate, and the fluorescence of the lysed liposomes previously remaining bound to the beads was then read as described above. The fluorescence intensities were correlated to maltose concentrations using a 5-parameter logistic (5-PL, eq 1) using XLFit software (IDBS, Bridgewater, NJ): y=b+

a−b e ⎛ x d⎞ ⎜1 + ⎟ c ⎠ ⎝

()

slope factor, and e is an asymmetry factor.37,38 The limit of detection was defined as the concentration equivalent to the background signal minus 3× standard deviation39 and was determined on the basis of the 5-PL fit of the acquired data. The background signal was the signal resulting from the above steps using MESS without D-(+)-maltose and was used in all signal-to-noise calculations. To determine the above conditions, the following optimizations were carried out: bead volume varied from 1 to 5 μL; MBP labeled liposome incubation carried out for 15, 30, 45, or 60 min at 4, 22, or 37 °C; and MBP labeled liposome concentration varied from 5 to 50 μM phospholipid. To assess specificity, the following compounds were compared to D(+)-maltose monohydrate at the same concentrations (6.25 nM to 6.25 mM): sucrose, 1-thio-β-D-glucose sodium salt, D(+)-raffinose, D-lactose, D-(+)-glucose, trehalose, D-(+)-xylose, D-(+)-galactose, D-fructose, D-sorbitol, glycerol, maltotriose, maltoheptaose, and dextrose. To analyze Maxisal for maltose content, tablets labeled to contain 200 mg of anhydrous crystalline maltose were dissolved with shaking for 24 h in a simulated gastric fluid composed of 0.7% (v/v) hydrochloric acid and 0.2% (w/v) sodium chloride. The supernatant was filtered and diluted to a 10% (v/v) mixture with MESSB with anticipated maltose concentration of 1 μM prior to analysis in the competitive assay as described above. A control of the same dilution of simulated gastric fluid was compared to ensure this matrix did not adversely affect the analysis.

■

RESULTS AND DISCUSSION Competitive Surfaces. MBP recognizes maltose as well as higher order α(1→4) linked glucose polymers.40 Such maltodextrins, including maltotetraose, maltopentaose, maltoheptaose, cyclic heptaose, and amylose effectively compete with maltose (KD = 1.0 μM) for MBP binding41,42 and induce quenching of MBP fluorescence with similar affinity as maltotriose (KD = 0.16 μM).40 With these known potential binding entities, several options for generating a competitive surface were evaluated. These ranged from small sugar molecules to polymeric starch extracts covalently coupled either via sodium cyanoborohydride reduction to a protein coated surface or to a proprietary hydrazide-linked surface. Such conjugation approaches rely on covalent linkage to the aldehyde group in the open chain form of these sugar-based molecules. Other unsuccessful approaches are detailed in the Supporting Information. Overall, only large polymeric sugars yielded a binding response from MBP-tagged liposomes (Figure 2). For example, only amylopectin yielded a concentration-dependent increase in signal when conjugated to a BSA-coated surface but not maltose, maltotriose, or maltoheptaose (Figure 2a). Similarly, only amylopectin and amylose yielded a concentrationdependent increase in signal when conjugated to a proprietary hydrazide surface, not maltose or maltotriose (Figure 2b). A similar profile was observed with unconjugated MBP alone (Figure S-3, Supporting Information). It can easily be envisioned that a small molecule sugar linked only by its reducing group to the surface would be a challenging target for successful binding. Given the extensive network of well-defined hydrogen bonding and van der Waals interactions of MBP with its small target molecule,43 opening up of the cyclic sugar structure to its linear form and occupying the resulting aldehyde group via covalent interaction with the

(1)

where a is the response at zero concentration, b is the response at maximum concentration, x is the D-(+)-maltose concentration, c is the concentration yielding 50% response, d is a 2773

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

proved useful for screening of buffers and optimizing assay conditions. However, ultimately, amylose superparamagnetic beads were chosen as a competitive surface for the final assay. These 10 μm beads provided increased surface area and reduced mass transport limitations versus a flat microtiter platebased surface (Figure S-4, Supporting Information). We postulate that findings obtained here for MBP will be relevant also for other PBPs in the preparation of competitive binding surfaces for their respective target analytes. Assay Buffers. MBP was found to retain its competitive binding abilities between maltose and amylose in a variety of buffers with ranging pH, ionic strength, and additives. There was a concentration-dependent benefit toward adding up to 0.05% (w/v) Tween-20 (Figure S-6, Supporting Information); up to 0.1% (w/v) BSA (Figure S-8, Supporting Information); up to 200 mM sodium chloride (Figure 3) (higher

Figure 2. Competitive surface development. Various sugars (legend) were conjugated to either (a) BSA via sodium cyanoborohydride reduction or (b) a proprietary hydrazide surface. The fluorescence intensities resulting from binding of MBP-conjugated liposomes (yaxis) are plotted versus the sugar concentration present during conjugation (x-axis). Each data point represents the average of triplicate determinations with error bars showing one standard deviation of these results.

surface would limit the interactions that can take place. Specific interactions aside, steric factors likely also play a significant role when the target is immobilized. Crystallographic studies have shown that maltose, with an accessible surface area of 54.55 nm, has only minimal area that is solvent exposed (3.6%) when bound within the 1.8 nm deep cleft of MBP.43 Although larger linear multimers such as maltotriose and maltoheptaose may have some extension beyond the deep cleft,43 they likely would largely remain inaccessible when covalently linked as MBP binds to their reducing ends.43,44 The two lobes of this moderately sized protein must also undergo a large conformational shift of approximately 35° upon target binding43,45,46 which may impose additional steric constraints on surface recognition. On the contrary, using the same chemistries, polymeric supports such as amylose and amylopectin showed strong concentration-dependent responses (Figure 2). These large polysaccharides form starch, with amylose composed of linear α(1→4) linked glucose units and amylopectin composed of branched α(1→4) linked and α(1→6) linked glucosidic bonds. The known affinity of MBP for amylose is utilized in conjunction with elution via maltose in amylose chromatography columns to afford facile purification of recombinant proteins from E. coli where MBP is used as a fusion tag to enhance solubility.7,41,47,48 Both MBP alone and bulky MBPconjugated liposomes yielded a significant level of binding to these polysaccharide microtiter plate-based supports which

Figure 3. Effect of sodium chloride on MBP binding. The fluorescence intensity (a) or signal-to-noise ratio (b) versus the concentration of sodium chloride added to 50 mM MES, pH 6.5, in the presence or absence of 1 mM maltose and an amylose functionalized microtiter plate. Signal development was accomplished using biotinylated antiMBP, streptavidin alkaline phosphatase, and fluorescein diphosphate. Results plotted in terms signal-to-noise used the fluorescence intensity of the maltose containing wells divided by that from the buffer only wells. A decrease in the signal-to-noise ratio is favorable in this competitive format. Each column represents the average of triplicate determinations with error bars showing one standard deviation of these results.

concentrations of BSA and NaCl were not tested); and increasing buffer pH (Figure S-5, Supporting Information). Borate buffers at pH 6.0−8.0 were markedly less effective than acetate buffers or PBS over pH 4.0−9.0 (Figure S-5, Supporting Information). However, despite differences in effectiveness, MBP retained its binding ability to amylose and competitive binding with maltose under all conditions. This suggests that MBP, and presumably other PBPs, can be employed without significant restrictions on diluent. 2774

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

The choice of buffer for dilution of MBP-tagged dyeencapsulating liposomes yielded additional constraints. Liposomes are commonly stored and used in buffers that contain a high concentration of sucrose. For example, typical buffer compositions contain 200 mM sucrose and either 10 mM HEPES and 200 mM NaCl (HS)35 or PBS-based mixtures.25 The high concentration of sucrose was employed to yield a hyperosmotic external solution to help maintain the high concentration of SRB dye within the internal cavity. Liposomes stored in such solutions have been shown to maintain their encapsulant over a period of a year at 4 and 21 °C.28 However, preliminary experiments in this work indicated that the high concentration of sucrose was suboptimal for MBP binding, with less maltose binding in the presence of HS buffer containing 200 mM sucrose versus HS buffer alone (Figure S-7, Supporting Information). Hence, an alternate storage and dilution buffer was desirable. Over 160 buffers with ranging pH, ionic strength, and additives were screened for short-term dyeretention and size distribution properties. MES buffer yielded less leakage than HEPES or borate buffers (Figure S-9, Supporting Information); leakage decreased with increasing pH (Figure S-10, Supporting Information) and decreased with buffer concentration (Figure S-11, Supporting Information), and increasing sodium chloride concentration improved dye retention but had a negative impact on liposome suspension properties (Figure S-12, Supporting Information). Ultimately, the MBP-liposome buffer selected was a compromise between optimal MBP binding properties and those optimal for liposomes. It employed a reasonably high concentration of MES buffer (50 mM), a high pH within the buffering range of MES (pH 6.5), and an intermediate concentration of sodium chloride (200 mM). While beneficial for MBP binding, the surfactant Tween-20 would cause liposome lysis. However, a high concentration of BSA (0.1% (w/v)) also improved assay response of MBP-tagged liposomes (Figure S-13, Supporting Information); thus, it was added to the liposome diluent used in competitive maltose assays. The buffer composition employed for all further studies was 0.1% (w/v) BSA, 50 mM MES, 0.2 M sodium chloride, pH 6.5, containing 0.01% sodium azide. MBP Coverage. One of the challenges in the use of PBPs as biorecognition elements in biosensing is finding appropriate conditions for immobilization which maintain the conformational flexibility and activity of the protein.49,51 Coupling of MBP to liposomes in this work was accomplished using 1-ethyl3-(3-dimethylaminopropyl) carbodiimide (EDC) to link the carboxyl groups of the liposomes to amine groups on the protein. The conditions employed expectedly yielded increasing liposomal protein coverage with increasing protein concentration (Figure 4a). This approach yielded approximately 1.7− 27 molecules of MBP per liposome when coupled at input levels of 0.01−0.1 mol % MBP of total lipid. Such liposomes yielded specific binding in the maltose competitive assay. Those without MBP conjugation yielded no significant binding to the amylose magnetic beads and no maltose-dependent response, whereas those with MBP conjugated at increasing mol % yielded increasing interactions with the amylose magnetic beads and a significant response to maltose (Figure 4b). The useful target range of the competitive assay may be tailored through adjustments of the EDC concentrations or mol % coverage of carboxylic acid on the liposome surface to afford the necessary liposomal MBP coverage.35,50

Figure 4. (a) Coverage of MBP on liposomes expressed in terms of μmol MBP per mol phospholipid (y-axis) versus the protein concentration present during EDC-mediated conjugation. The MBP coverage was determined using a competitive ELISA for solution MBP versus surface immobilized MBP for a limited amount of biotinylated anti-MBP. Signal development was accomplished using streptavidin alkaline phosphatase and fluorescein diphosphate. (b) Competitive assay for maltose (x-axis) using amylose magnetic beads and liposomes conjugated to MBP at 0−0.10 mol % of total lipid (legend). The resulting fluorescence intensity is plotted on the y-axis (log scale). Each column represents the average of triplicate determinations with error bars showing one standard deviation of these results.

Specificity. Other researchers have shown that the intrinsic fluorescence of MBP is unaffected by glucose, α-methyl glucoside, lactose, and isomaltose, and these compounds do not affect the fluorescence quenching by maltose, when introduced at concentrations of 1 mM.40 MBP has not been observed to interact with species which contain fewer than two glucose molecules in an α-glycosidic linkage.44 Consistent with these results, MBP tagged liposomes exhibited excellent specificity toward maltose over other sugars. No binding was observed even at high concentrations (up to 6.25 mM) to dextrose, D-fructose, D-(+)-galactose, glycerol, D-(+)-raffinose, D-sorbitol, sucrose, 1-thio-β-D-glucose, trehalose, or D-(+)-xylose (Figures S14−S16, Supporting Information). This specificity was remarkable given the similarity of some of these structures. For example, maltose and trehalose both are composed of two glucose molecules but differ only in their connectivity (α(1→4) versus α(1→1)). Some competition with high concentrations (6.25 mM) of D-(+)-glucose and Dlactose was observed, which yielded a reduction in signal equivalent to that from maltose but only at a 10 000-fold higher concentration (Figure 5). The disaccharide lactose is similar to maltose, but they differ in that lactose is composed of galactose and glucose and lactose has a β-acetal versus an α-acetal linkage between their composite monosaccharides. While initially primarily the α-anomer, glucose gradually forms an equilibrium mixture of approximately one-third α and two-thirds β anomers 2775

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

plus three times its standard deviation, was 222 μM, yielding a response over more than 3 orders of magnitude. This compares favorably with literature reported detection limits for maltose of 1 mM using an enzyme-based amperometric biosensor,56 2 mg/L (equivalent to 5.84 μM) for a commercially available enzyme-based kit with UV detection, 57 and 0.01 g/L (equivalent to 29.2 μM) for HPLC with evaporative light scattering.58 The range of such methods spanned from 1 to 100 mM for the amperometric biosensor56 and 0.05−15.0 g/L (146 μM to 43.8 mM) for the HPLC method, respectively, indicating a broader assay range with lower detection limits for the PBP-based assay described within. The coefficients of variation of triplicate measurements ranged from 2% to 11%, with most less than 8% (Note: As the magnetic bead washing steps are done manually with a multichannel pipet, there exists the potential for aspiration of some of the bead complexes, resulting in loss of signal in affected wells and hence variability between replicate measurements. The reproducibility, and thus the calculated detection limits, could be further improved with a plate washer with magnetic bead handling capabilities.) The signal intensity in the absence of D -(+)-maltose was approximately 300 times that obtained from a saturating concentration of D-(+)-maltose, indicating significant signal enhancement by the SRB-encapsulating liposomes. The applicability of this method toward real samples was demonstrated by analysis of a pharmaceutical formulation of anhydrous crystalline maltose used for the treatment of dry mouth.33,34 Tablets of Maxisal, also formulated with stearic acid and magnesium stearate, were dissolved in a simulated gastric fluid and then diluted in assay buffer such that the anticipated maltose concentration was 1 μM maltose. The MBP-tagged SRB-encapsulating liposomes showed no release of their dye over storage in buffer alone during the assay period when diluted in a 10% (v/v) assay buffer mixture of the simulated gastric fluid or the same containing dissolved Maxisal tablets, indicating that liposomes were stable in the assay matrix (Figure S-20, Supporting Information). Additionally, the liposomes used in this analysis were prepared over 9 months previously and maintained strong functionality toward maltose. While it remains a subject for future work, this suggests that the liposomes retain their dye and MBP retains its function over long periods at 4 °C in the newly formulated MESS buffer. The competitive assay subsequently yielded accurate quantification of maltose (1.004 μM) in the control sample (1 μM maltose spiked into simulated gastric fluid), demonstrating that the simulated gastric fluid did not adversely affect assay performance. Quantification of maltose present in the dissolved Maxisal tablets led to determined maltose concentration within 5.7% of anticipated values (Table 1) and thus good correlation with the label claim.

Figure 5. Specificity of MBP toward D-(+)-maltose versus D(+)-glucose, D-lactose, and D-(+)-raffinose. Competitive assay for these sugars versus maltose using MBP-tagged SRB-encapsulating liposomes and amylose functionalized magnetic beads. The fluorescence intensity of bound liposomes versus the sugar concentration when sample and liposomes were diluted in 50 mM MES, pH 6.5, 0.2 M sodium chloride. Each point represents the average of triplicate determinations with error bars showing one standard deviation of these results.

when in aqueous solution.52,53 The glucose/galactose-binding protein has been shown to demonstrate a strong preference for binding to the β-anomer, not the α-anomer, of glucose.54,55 While not studied here, the slight cross-reactivity toward D(+)-glucose, but not its synonymous entity dextrose, may be a function of the time in solution prior to analysis and warrants further investigation. Overall, the exceptional specificity and limited cross-reactivity with few other sugars only at 104 greater concentrations lends this assay to analysis of pharmaceutical products and select foodstuffs. However, the known affinity toward higher order maltose polymers,40−42 also observed here toward maltotriose and maltoheptaose (Figure S-17, Supporting Information), suggests that this method would not be appropriate for specific analysis of samples containing maltodextrin, which is a common additive in many preparations. Assay Performance. The limit of detection, calculated from the 5PL fit of the data and the signal in the absence of D(+)-maltose minus three times its standard deviation, was 78 nM (Figure 6). The upper limit of the assay, determined using the signal resulting from D-(+)-maltose concentration at 2 mM

■

CONCLUSIONS Maltose binding protein was employed as a novel binding species in a similar manner to an antibody in a competitive enzyme linked immunosorbent assay (ELISA). This assay for maltose demonstrated that such periplasmic binding proteins can be used in analytical applications beyond those relying on monitoring conformational shifts. The assay for D-(+)-maltose utilizing MBP for target recognition was highly specific toward D-(+)-maltose versus other small molecule sugars. The binding of this protein was largely resilient to the buffer employed which suggests potential for use in many other assay formats that do not rely on liposome-based signal enhancement. This

Figure 6. Fluorescence intensity values resulting from the competitive assay for maltose. D-(+)-Maltose (●) concentrations ranging from 34 nM to 2 mM were assayed. The 5 PL fit of the experimental D(+)-maltose data is represented by the dashed line. Each point represents the average of triplicate determinations with error bars showing one standard deviation of these results. 2776

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

(5) Willis, R. C.; Morris, R. G.; Cirakoglu, C.; Schellenberg, G. D.; Gerber, N. H.; Furlong, C. E. Arch. Biochem. Biophys. 1974, 161, 64− 75. (6) Lu, D.; Boyd, B.; Lingwood, C. A. J. Biol. Chem. 1997, 272, 29033−29038. (7) Pattenden, L. K.; Thomas, W. G. Methods Mol. Biol. 2008, 421, 169−189. (8) Medintz, I. L.; Deschamps, J. R. Curr. Opin. Biotechnol. 2006, 17, 17−27. (9) Zhou, L.; Cass, A. Biosens. Bioelectron. 1991, 6, 445−450. (10) Gilardi, G.; Zhou, L. Q.; Hibbert, L.; Cass, A. E. Anal. Chem. 1994, 66, 3840−3847. (11) Miller, D. M., 3rd; Olson, J. S.; Pflugrath, J. W.; Quiocho, F. A. J. Biol. Chem. 1983, 258, 13665−13672. (12) Crochet, A. P.; Kabir, M. M.; Francis, M. B.; Paavola, C. D. Biosens. Bioelectron. 2010, 26, 55−61. (13) Medintz, I. L.; Goldman, E. R.; Lassman, M. E.; Mauro, J. M. Bioconjugate Chem. 2003, 14, 909−918. (14) Medintz, I. L.; Clapp, A. R.; Mattoussi, H.; Goldman, E. R.; Fisher, B.; Mauro, J. M. Nat. Mater. 2003, 2, 630−638. (15) Fehr, M.; Frommer, W. B.; Lalonde, S. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 9846−9851. (16) Park, K.; Lee, L. H.; Shin, Y. B.; Yi, S. Y.; Kang, Y. W.; Sok, D. E.; Chung, J. W.; Chung, B. H.; Kim, M. Biochem. Biophys. Res. Commun. 2009, 388, 560−564. (17) Szollosi, J.; Damjanovich, S.; Matyus, L. Cytometry 1998, 34, 159−179. (18) Ha, J. S.; Song, J. J.; Lee, Y. M.; Kim, S. J.; Sohn, J. H.; Shin, C. S.; Lee, S. G. Appl. Environ. Microbiol. 2007, 73, 7408−7414. (19) Maltose Assay Kit-ab83388, http://www.abcam.com/ps/ products/83/ab83388/documents/ ab83388%20Maltose%20Assay%20Kit%20(Website).pdf, Accessed December 10, 2012. (20) Maltose and Glucose Assay Kit, Catalog K618-100, http://www. biovision.com/manuals/K618.pdf?osCsid= 93rsk8j9m01m5180sgj2sosh05, Accessed December 10, 2012. (21) Shan, G.; Lipton, C.; Gee, S.; Hammock, B. In Handbook of Residue Analytical Methods for Agrochemicals; Lee, P. W., Ed.; John Wiley & Sons, Ltd: Chichester, 2002; pp 623−679. (22) Thorpe, R.; Thorpe, S. In Principles and Techniques of Biochemistry and Molecular Biology, 6th ed.; Wilson, K., Walker, J., Eds.; Cambridge University Press: New York, 2005; pp 292−348. (23) Tombelli, S.; Minunni, M.; Mascini, M. Biosens. Bioelectron. 2005, 20, 2424−2434. (24) Dwyer, M. A.; Hellinga, H. W. Curr. Opin. Struct. Biol. 2004, 14, 495−504. (25) Zaytseva, N. V.; Montagna, R. A.; Baeumner, A. J. Anal. Chem. 2005, 77, 7520−7527. (26) Edwards, K. A.; Baeumner, A. J. Anal. Bioanal. Chem. 2006, 386, 1613−1623. (27) Edwards, K. A.; Duan, F.; Baeumner, A. J.; March, J. C. Anal. Biochem. 2008, 380, 59−67. (28) Edwards, K. A.; Meyers, K. J.; Leonard, B. L.; Baeumner, A. J. Anal. Bioanal. Chem. 2012 , in press. (29) Campbell, R. L.; Wagner, D. B.; O'Connell, J. P. Solid phase assay with visual readout, U.S. Patent 4,703,017, Issued October 27, 1987. (30) Edwards, K. A.; Baeumner, A. J. Talanta 2006, 68, 1421−1431. (31) Rongen, H. A.; Bult, A.; van Bennekom, W. P. J. Immunol. Methods 1997, 204, 105−133. (32) Locascio-Brown, L.; Plant, A. L.; Horvath, V.; Durst, R. A. Anal. Chem. 1990, 62, 2587−2593. (33) Fox, P. C.; Cummins, M. J.; Cummins, J. M. J. Altern. Complementary Med. 2002, 8, 651−659. (34) Fox, P. C.; Cummins, M. J.; Cummins, J. M. J. Altern. Complementary Med. 2001, 7, 33−43. (35) Edwards, K. A.; Curtis, K. L.; Sailor, J. L.; Baeumner, A. J. Anal. Bioanal. Chem. 2008, 391, 1689−1702.

Table 1. Concentrations of Maltose Determined in Simulated Gastric Fluida sample

[maltose], μM

anticipated control (GI fluid) Maxisal in GI fluid sample 1 Maxisal in GI fluid sample 2

1.000 1.004 ± 0.017 1.049 ± 0.040 1.057 ± 0.046

Reported values are calculated from the 5PL fit of the calibration data. The concentration of each sample was determined in triplicate, with the values reported here equal to the average and standard deviation of these results. a

method may find utility in food, beverage, or pharmaceutical analyses, provided the samples do not contain higher order α(1→4) maltodextrins. This method was successfully employed to quantify maltose as an active ingredient in a pharmaceutical preparation for treatment of dry mouth, though it may also find utility toward detection and quantification of maltose used as excipients in biopharmaceutical formulations59 or formed as a metabolic product following patient administration. Maltose is an excipient of increasing use in pharmaceutical formulations, due to its minimal impact on blood sugar, its ability to prevent aggregation of biomolecules, and it being less hygroscopic than some other excipients.60,61 It is approved for use as a nutritional additive to intravenous fluids administered to postop patients or those with diabetes mellitus. However, between 1997 and 2009, deaths and other adverse events resulting from inappropriate insulin administration were reported as a consequence of the use of diabetic testing kits relying on glucose dehydrogenase pyrroloquinoline quinone (GDHPQQ).62 Such assays could not differentiate between glucose and other sugars either present as excipients or metabolites, resulting in falsely elevated glucose levels leading to these adverse events. The use of a PBP-based sensing system with its exquisite sugar specificity would avoid such unfortunate consequences.

■

ASSOCIATED CONTENT

* Supporting Information S

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: 607-255-5433. Fax: 607255-4449. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors are appreciative of partial support provided through a grant from the European Commission’s Seventh Framework Programme, Grant No. 244405.

■

REFERENCES

(1) Ames, G. F. Annu. Rev. Biochem. 1986, 55, 397−425. (2) Quiocho, F. A.; Ledvina, P. S. Mol. Microbiol. 1996, 20, 17−25. (3) Poole, K.; Hancock, R. E. Eur. J. Biochem. 1984, 144, 607−612. (4) Guyer, C. A.; Morgan, D. G.; Osheroff, N.; Staros, J. V. J. Biol. Chem. 1985, 260, 10812−10818. 2777

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778

Analytical Chemistry

Article

(36) Corning Carbo-BIND 96-well plates: Manufacturer’s instructions, Corning Incorporated: Corning, NY. (37) Gottschalk, P. G.; Dunn, J. R. Fitting Brendan’s Five-Parameter Logistic Curve, http://www.brendan.com/PDF%20files/BioRad%20Bulletin%203022%20-%20StatLIA5PL.pdf, Tech note 3022; accessed November 14, 2012. (38) Gottschalk, P. G.; Dunn, J. R. Anal. Biochem. 2005, 343, 54−65. (39) Thompson, V.; Schatzlein, D.; Mercuro, D. Spectroscopy 2003, 18, 112−114. (40) Szmelcman, S.; Schwartz, M.; Silhavy, T. J.; Boos, W. Eur. J. Biochem. 1976, 65, 13−19. (41) Ferenci, T.; Klotz, U. FEBS Lett. 1978, 94, 213−217. (42) Kellermann, O.; Szmelcman, S. Eur. J. Biochem. 1974, 47, 139− 149. (43) Spurlino, J. C.; Lu, G. Y.; Quiocho, F. A. J. Biol. Chem. 1991, 266, 5202−5219. (44) Thieme, R.; Lay, H.; Oser, A.; Lehmann, J.; Wrissenberg, S.; Boos, W. Eur. J. Biochem. 1986, 160, 83−91. (45) Gilardi, G.; Mei, G.; Rosato, N.; Agro, A. F.; Cass, A. E. Protein Eng. 1997, 10, 479−486. (46) Sharff, A. J.; Rodseth, L. E.; Spurlino, J. C.; Quiocho, F. A. Biochemistry 1992, 31, 10657−10663. (47) Fox, J. D.; Waugh, D. S. Methods Mol. Biol. 2003, 205, 99−117. (48) Kapust, R. B.; Waugh, D. S. Protein Sci. 1999, 8, 1668−1674. (49) Dattelbaum, A. M.; Baker, G. A.; Fox, J. M.; Iyer, S.; Dattelbaum, J. D. Bioconjugate Chem. 2009, 20, 2381−2384. (50) Edwards, K. A.; Meyers, K. J.; Leonard, B.; Connelly, J. T.; Wang, Y.; Holter, T.; Baeumner, A. J. Anal. Methods 2012, 4, 3948− 3955. (51) Li, Q. Z.; Cass, A. E. G. Biosens. Bioelectron. 1991, 6, 445−450. (52) O’Neil, M. J. Maltose, Entry #4472, The Merck Index, 13 ed.; Merck Research Laboratories: Whitehouse Station, NJ, 2001. (53) D-(+)-Glucose, G 8270 Product Information Sheet, SigmaAldrich, http://www.sigmaaldrich.com/etc/medialib/docs/Sigma/ Product_Information_Sheet/2/g8270pis.Par.0001.File.tmp/g8270pis. pdf, Accessed Nov. 1, 2012. (54) Vyas, M. N.; Vyas, N. K.; Quiocho, F. A. Biochemistry 1994, 33, 4762−4768. (55) Borrok, M. J.; Kiessling, L. L.; Forest, K. T. Protein Sci. 2007, 16, 1032−1041. (56) Mahosenaho, M.; Caprio, F.; Micheli, L.; Sesay, A. M.; Palleschi, G.; Virtanen, V. Microchim. Acta 2010, 170, 243−249. (57) Maltose/Sucrose/D-Glucose: UV-method for the determination of maltose, sucrose and D-glucose in foodstuffs and other materials, http://www.r-biopharm.com/ product_site.php?language= english&product_id=471, Accessed January 10, 2013. (58) Nogueira, L. C.; Silva, F.; Ferreira, I. M. P. L. V. O.; Trugo, L. C. J. Chromatogr., A 2005, 1065, 207−210. (59) Hekman, C.; Park, S.; Teng, W. Y.; Guzman, N. A.; Rossi, T. J. Pharm. Biomed. Anal. 1995, 13, 1249−1261. (60) Ochs, H.; Siegel, J.; Young, H. Stabilizers used in intravenous immunoglogulin products: A comparative review. In Pharmacy Practice News, August 2010; pp 1−8. (61) Joshi, V. Excipient Choice in Solid Dosage Forms. In Drug Development and Delivery, 2002; Vol. 2, No. 6. (62) FDA Public Health Notification: Potentially Fatal Errors with GDH-PQQ* Glucose Monitoring Technology, U.S. Food and Drug Administration, http://www.fda.gov/medicaldevices/safety/ alertsandnotices/publichealthnotifications/ucm176992.htm, Accessed October 18, 2012.

2778

dx.doi.org/10.1021/ac303258n | Anal. Chem. 2013, 85, 2770−2778