Letter pubs.acs.org/NanoLett
Fast, Ratiometric FRET from Quantum Dot Conjugated Stabilized Single Chain Variable Fragments for Quantitative Botulinum Neurotoxin Sensing Joonseok Lee,† Melissa B. Brennan,‡ Rosemarie Wilton,*,‡ Clare E. Rowland,†,§ Elena A. Rozhkova,† Sara Forrester,∥ Daniel C. Hannah,§ Julia Carlson,⊥ Elena V. Shevchenko,† Daniel S. Schabacker,∥ and Richard D. Schaller*,†,§ †
Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States Biosciences, Argonne National Laboratory, Argonne, Illinois 60439, United States § Department of Chemistry, Northwestern University, Evanston, Illinois 60208, United States ∥ Global Security Sciences, Argonne National Laboratory, Argonne, Illinois 60439, United States ⊥ College of Agricultural & Life Sciences, University of Wisconsin, Madison, Wisconsin 53706, United States ‡
S Supporting Information *
ABSTRACT: Botulinum neurotoxin (BoNT) presents a significant hazard under numerous realistic scenarios. The standard detection scheme for this fast-acting toxin is a labbased mouse lethality assay that is sensitive and specific, but slow (∼2 days) and requires expert administration. As such, numerous efforts have aimed to decrease analysis time and reduce complexity. Here, we describe a sensitive ratiometric fluorescence resonance energy transfer scheme that utilizes highly photostable semiconductor quantum dot (QD) energy donors and chromophore conjugation to compact, single chain variable antibody fragments (scFvs) to yield a fast, fieldable sensor for BoNT with a 20−40 pM detection limit, toxin quantification, adjustable dynamic range, sensitivity in the presence of interferents, and sensing times as fast as 5 min. Through a combination of mutations, we achieve stabilized scFv denaturation temperatures of more than 60 °C, which bolsters fieldability. We also describe adaptation of the assay into a microarray format that offers persistent monitoring, reuse, and multiplexing. KEYWORDS: botulinum neurotoxin, FRET, protein sensor, scFv, quantum dot, microarray
B
availability of antitoxins that only offer benefits prior to the onset of symptoms, means to quickly identify and quantify BoNT by nonspecialized personnel such as first responders, preferably at the site of an event, would help to mitigate the threat. At present, the only universally accepted method of characterizing BoNT is the lab-based mouse lethality assay (MLA) that quantifies minimum lethal dose and examines botulism-related symptoms.2 While the method is sensitive down to 10 pg/mL, it requires 24−48 h to perform and necessitates multiple animals, which effectively reduces the stated sensitivity. More rapid variants of in vivo bioassays require 4 to 24 h.2 Recently, a wide array of in vitro detection assays has yielded significant progress with regard to sensitivity and rapidity of analysis,1−3 although no method has yet supplanted the MLA. Some of these approaches include a labbased enzyme-linked immunosorbent assays (ELISA)6,7 that
otulinum neurotoxin (BoNT), being the most toxic substance known with a median lethal dose of ∼1 ng/ kg, is a cause for significant concern with regard to public health, battlefield, and terrorism scenarios.1−3 Several serologically distinct neurotoxins are known to be produced by Clostridium botulinum and related clostridia. BoNT/A, B, E, and F cause human disease, and in aerosolized form, serotypes C and D are also considered likely to affect humans.3,4 The ∼150 kDa toxin is synthesized as a single polypeptide chain in complex with several nontoxic protective proteins. This “armored” complex facilitates transit of the toxin through the gastrointestinal tract, and rapidly dissociates at neutral to basic pH values.5 The toxin is activated by proteolytic nicking yielding two disulfide-linked segments, a 100 kDa heavy chain (HC) that is organized into domains for receptor binding (HCR) and translocation (HCT), and a light chain (LC) zincendopeptidase (see Figure 1) that cleaves proteins responsible for acetylcholine release within nerve synapses. Such cleavage can result in loss of neuronal activity, paralysis, and respiratory failure within 12 to 48 h.3 Given facile routes to the toxin partnered with its lethality, rapid onset of symptoms, and © XXXX American Chemical Society
Received: August 26, 2015
A
DOI: 10.1021/acs.nanolett.5b03442 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 1. Ratiometric FRET-based BoNT sensor components. (a) Ribbon/space-fill diagram of BoNT/A1 (Protein Data Bank entry no. 3BTA).14 Dashed lines show approximate HCR antigen binding locations for labeled scFvs sBOT1 and BOT2. (b) Optical properties of the QD donor and Dylight 650 acceptor.
present a challenge to fieldability and typically only find implementation in lab settings owing to temperature and environmental sensitivity that can cause these proteins to denature. However, as we explore here, approaches exist to increase the robustness of compact scFvs through targeted mutations that render increased stability of the desirable folded form yet maintain high target binding affinity. Here, we describe a rapid-readout, fieldable approach to sensing and quantization of BoNT, serotype A1, using an immunological sandwich assay and ratiometric FRET. We first investigate and demonstrate noncompetitive binding for proximal locations on the BoNT HCR such that optically labeled scFvs, herein termed BOT1 and BOT2, can respectively position donor and acceptor FRET chromophores in proximity upon binding. We describe and examine mutated versions of these scFvs that exhibit increased thermal stability while maintaining high binding affinity for BoNT. Red-emitting CdSe/ZnS core/shell QDs were conjugated to BOT1 and a deep-red organic dye (Dylite 650) was conjugated to BOT2. Upon the introduction of unlabeled BoNT HCR, purified toxin, or toxin complex to a solution containing the two scFvs, we rapidly observe FRET that causes reduction in QD PL with a commensurate increase in acceptor emission. We confirm the FRET process using time-resolved PL and utilize this ratiometric FRET signal to both sense and quantify the presence of BoNT in a homogeneous assay format. Using this approach, we determine a BoNT limit of detection near 20−40 pM, an adjustable dynamic range, and achieve very rapid readout times as fast as ∼5 min. We further exhibit adaptation of the assay into a microarray format, which demonstrates the potential for multiplexing of the method and expansion to multiple BoNT serotypes and threat agents. This broadly generalizable approach of scFv functionalization offers a path forward to fieldable pathogen sensing using stabilized materials and nonexpert implementation.
can exhibit sensitivity to BoNT enzymatic activity down to 350 pM,8,9 and compact lab-on-a-chip approaches.10−12 Photoluminescence (PL)-based detection transduction methods have been developed that exploit dipolar fluorescence resonance energy transfer (FRET) between a photoexcited donor chromophore and an adjacent acceptor, where the acceptor exhibits absorption at the PL energy of the donor.13,14 Although numerous FRET-based sensors rely on organic dyes as the donor chromophore (typically with a PL quencher acceptor), semiconductor nanocrystal quantum dots (QDs) have been explored as FRET donors because their properties mitigate several inherent difficulties in dye-to-dye FRET systems.8,15−17 Useful QD properties include broad absorption spectra, size-tunable absorption onset, and high molar absorption coefficients, which simplify means to optically excite QDs efficiently using wavelengths that avoid acceptor chromophore excitation. Also, size-controlled tuning of QD PL spectra facilitates large FRET overlap integrals. Moreover, their chemical and photochemical stability, including resistance to photobleaching, renders QDs more robust than many organic dyes.15−18 Immunological-based sensing of unlabeled toxins and pathogens offers high sensitivity and selectivity owing to the site-specific binding of antibodies to unique biomarkers. However, QD-antibody conjugates can become unsuitable for FRET applications. Specifically, the large size of full antibody, which can be 15 nm wide and 10 nm in length, can result in donor−acceptor distances that exceed the range over which FRET is generally efficient (usually
