Ultrasensitive Detection of Ricin Toxin in Multiple Sample Matrixes

May 22, 2015 - In this paper, we present the development and application of a single-molecule array (Simoa) for the detection of ricin toxin in human ...
0 downloads 9 Views 430KB Size
Subscriber access provided by UNIV OSNABRUECK

Article

Ultra-Sensitive Detection of Ricin Toxin in Multiple Sample Matrices Using Single Domain Antibodies Shonda T. Gaylord, Trinh L Dinh, Ellen R. Goldman, George P Anderson, Kevin C Ngan, and David R. Walt Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b00322 • Publication Date (Web): 22 May 2015 Downloaded from http://pubs.acs.org on May 26, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Ultra-Sensitive Detection of Ricin Toxin in Multiple Sample Matrices Using Single Domain Antibodies Shonda T. Gaylord, §†‡ Trinh L. Dinh,§‡ Ellen R. Goldman,‡ George P. Anderson,‡ Kevin C. Ngan§ and David R. Walt§* §

Department of Chemistry, Tufts University, 62 Talbot Ave, Medford, MA 02155 Center for Bio/Molecular Science and Engineering, Naval Research Laboratory, 4555 Overlook Ave SW, Washington DC 20375 ‡

ABSTRACT: Ricin is an extremely potent ribosomal inactivating protein listed as a Category B select agent. Although ricin intoxication is not transmittable from person to person, even a single ricin molecule can lead to cell necrosis because it inactivates 1,500 ribosomes per minute. Since there is currently no vaccine or therapeutic treatment for ricin intoxication ultra-sensitive analytical assays capable of detecting ricin in a variety of matrices are urgently needed to limit exposure to individuals as well as communities. In this paper, we present the development and application of a single molecule array (Simoa) for the detection of ricin toxin in human urine and serum. Single domain antibodies (sdAbs), among the smallest engineered binding fragments, were chemically coupled to the surface of paramagnetic beads for the sensitive detection of ricin toxin. The Simoa was able to detect ricin at levels of 10 fg/mL, 100 fg/mL and 1 pg/mL in buffer, urine and serum, respectively, in a fraction of the assay time need using ImmunoPCR. Using a fully automated state-of-the-art platform, the Simoa HD-1 Analyzer, the assay time was reduced to 64 minutes.

INTRODUCTION Since the 1990s, developers of biosensors have pursued the ultimate limit of analytical sensitivity: reliable single molecule detection. Recently, the principles of enzyme-linked immunosorbant assay (ELISA) have been utilized to design versatile and specific single molecule array (Simoa) assays, also called Digital ELISA1. The Simoa platform is a robust method that allows single molecule detection of enzymes isolated in discrete microwells. Instead of a fixed, planar, unstirred capture surface, Simoa uses paramagnetic beads to provide a high local concentration and distribution of the capture antibody within the sample 2. Proteins captured on the beads are sandwiched with biotinylated detector antibodies and labeled with streptavidin-β-galactosidase (SβG). Using high density microwell arrays, enzyme-labeled immunocomplexes formed on beads are isolated into 46-fL-sized microwells and sealed in the presence of a fluorogenic substrate, resorufin-β-Dgalactopyranoside (RDG). Single immunocomplexes are detected via the generation of a high local concentration of fluorescent product within each microwell 3,4. One area where ultra-high-sensitivity methods are urgently needed is the detection of biowarfare agents, including potent toxins such as ricin, of which just one molecule is enough to cause a cell to become necrotic. Fast and sensitive detection of ricin, in environmental, food and clinical samples, is critical for public health and national security. Simoa is ideally suited to address this challenge. Traditionally, Simoa technology uses monoclonal or polyclonal antibodies for the capture and detection of the target antigen. The production of antibodies, generated in response to immunization of an animal host with an immunogen, is costly, time consuming, and often results in antibodies unsuitable for therapeutics or field-deployable diagnostics. At tem-

peratures >60-70ᵒC, the heavy and light chains of conventional antibodies unfold and irreversibly aggregate rendering them unsuitable for use in hostile environments and necessitating low temperature storage5. In contrast, reverse transcriptasePCR cloning of mRNAs from immunized or naïve animals are used to create an immune library with specificity towards an antigen of interest for the generation of engineered variants: diabodies, triabodies, minibodies and single-domain antibodies (sdAbs) 6. The use of sdAbs, the smallest engineered binding fragments (~12 to 16 kDa), in immunoassays has gained momentum since the early 1990’s. sdAbs, are robust thermostable molecules with the ability to refold after denaturation to bind target antigen and can be generated for many different specificities7. Their extended shelf-life makes them an ideal alternative for diagnostic use. As an added benefit, a nonimmunized library can be generated from which antigen binders can be selected for against toxic or potentially lethal antigens by cloning sdAb genes from lymphocytes in llama blood and using phase display technology to identify antigen binders8. The application of such engineered antibodies for ricin diagnostics has been previously demonstrated using Luminex bead based assays, ELISAs and one-step silicon photonic microring resonator arrays with limits of detection of 64 pg/mL, 1 ng/mL, and 18 ng/mL, respectively 9,10. Ricin is a serious biological warfare threat because it is extremely potent, readily assessable, easily disseminated, with a wide pH tolerance, solubility and thermostability 11. Its use as a biological weapon is not presumptive but historical 12,13. As a type II ribosomal inactivating protein (RIP-II), ricin catalytically halts protein synthesis by preventing elongation factor-2 from associating with the ribosome during translocation. Ricin contains the majority of carbohydrates that bind to galactosecontaining glycoproteins and glycolipids on the surface of cells for transport into the cytosol 11. The two-chain structure

ACS Paragon Plus Environment

Analytical Chemistry

Ricin Assay (Bt-C8)

Ricin Assay (Bt-H1)

25000

6000 C8 anti-Ricin

20000

H1 anti-Ricin

D1 anti-Ricin Neg Control

MFI

MFI

C8 anti-Ricin

5000

H1 anti-Ricin

15000 10000

4000

D1 anti-Ricin

3000

Neg control

2000 5000

1000

0

0 0

A

0.064 0.32

1.6

8

40

200

1000

0

B

Ricin Conc. (ng/mL)

0.064 0.32

1.6

8

40

200

1000

Ricin Conc. (ng/mL)

Ricin Assay (Bt-D1)

Ricin Assay (Bt-B4)

12000

10000 C8 anti-Ricin

C8 anti-Ricin

10000

8000

H1 anti-Ricin

H1 anti-Ricin 8000

D1 anti-Ricin

6000

D1 anti-Ricin

Neg control

Neg Control

MFI

MFI

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 9

6000

4000

4000

2000

2000

0

0

-2000

0

C

0

0.064 0.32

1.6

8

40

200

1000

Ricin Conc. (ng/mL)

0.064 0.32

D

1.6

8

40

200

1000

Ricin Conc. (ng/mL)

Figure 1: Screening capture and detector sdAb pairs for ricin. Each sdAb pair was analyzed for its ability to detect ricin toxin in buffer using the MAGPIX assay. The B4 sdAb was not suitable as a capture Ab, while the C8 capture Ab performed best with all bt-sdAbs. The best observed pair was C8-D1 with a LOD of 64 pg/mL. A) C8, H1, D1 capture Abs with bt-C8 detector. B) C8, H1, D1 capture Abs with bt-H1 detector. C) C8, H1, D1 capture Abs with bt-D1detector. D) C8, H1, D1 capture Abs with bt-B4 detector.

of ricin is the contributing factor to its toxicity, as one ricin holotoxin can inactivate 1,500 ribosomes per minute 11. Individuals experiencing ricin poisoning through inhalation, ingestion or parenteral administration may present with symptoms clinically indistinguishable from ordinary infections such as gastrointestinal dysfunction, oropharyngeal pain, clinical characteristics similar to sepsis, and respiratory failure within 4 to 12 hours after exposure making symptomatic diagnosis extremely difficult 11. Although poisoning is not transmittable from person to person, there are currently no vaccines or therapeutics available resulting in an associated mortality rate of 1.9% after ingestion13; the least toxic route of poisoning. A variety of analytical methods for environmental and clinical monitoring of ricin exist and include: mouse bioassays 14, mass spectroscopy 15, immunosorbent assays (including colorimetric, chemiluminescent and electrochemiluminescent) 16, lateral flow assays 17 and in-vitro cell based assays 18. Currently, there are no clinically-validated methods for ricin detection in biological fluids 11 but the Centers for Disease Control and Prevention (CDC) has developed a test for the detection of ricinine (an alkaloid co-extracted with ricin) in urine as an indicator for ricin intoxication, making detection possible for up to 48 hours after exposure using high performance liquid chromatography-electrospray ionization-mass spectroscopy (HPLC-EIS-MS) 11. Ricinine may only serve as a surrogate marker or indicator of ricin intoxication, as the ricin toxin is

not actually being detected. Additionally, ricinine is expected to be present in the general population, with a recovery rate of 1.2% in persons who were not suspected of ricin exposure 19. Currently, there are no reports of human ricinine poisonings. Since ricin is rapidly absorbed into tissue, detection of the toxin in biological fluids has posed many problems. One of the most sensitive methods for ricin detection is Immuno-PCR (IPCR), which exploits the specificity of antibodies with the sensitivity of PCR. IPCR has a limit of detection (LOD) in buffer of 10 fg/mL 20,21, and was clinically tested as a diagnostic for the detection of ricin in serum and feces with a LOD of 1 pg/mL 22. The use of IPCR typically enhances the LOD by 100- to 10,000-fold in comparison to ELISA 21. As promising as this method initially appears to be, the sensitivity is still highly dependent on the enzymatic amplification of PCR reactions and requires 9 to 11 hours. With a reported “window of intervention” for ricin poisoning of 30 minutes to 1 hour 23, a much more rapid assay that does not sacrifice sensitivity is required to provide supportive care and address public health emergencies. Simoa has demonstrated significant superiority for bioanalysis by providing digital measurements of enzyme-labeled singulated beads with a 68,000 fold increase in sensitivity compared to ELISA as demonstrated with the enzymatic reporter β-galactosidase 4. Since its inception, Simoa has gained momentum as an alternative to conventional immunoassays

ACS Paragon Plus Environment

Page 3 of 9

100 Serum Urine Buffer

10

1

AEB

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

0.1

0.01

0.001

0.0001 0

0.01

0.1

1

10

100

1000

10000

Ricin, pg/mL

Figure 2: Combined digital and analog calibration curves. The log-log standard curve in buffer provided a 6-log concentration linear dynamic range with a measured LOD of 0.01 pg/mL, a 5-log linear dynamic range in urine with a measured LOD of 0.1 pg/mL and a 4-log linear dynamic range in serum with a measured LOD of 1 pg/mL. The dashed line indicates the boundary between the digital and analog read-out.

providing ultra-sensitive detection of prostate specific antigen, genomic DNA, HIV p24 protein, Dengue-specific antibodies24 and a variety of clinically relevant cytokines including: TNFα, interleukin-1α, interleukin-1β and interleukin-6 3,4,25,26. Here we detail the development of a ricin Simoa using sdAbs and present with a fully automated state-of-the-art platform, the Simoa HD-1 Analyzer, for the detection of ricin toxin. This method provides one of the shortest reported assay times with a clinically relevant limit of detection for ricin toxin. EXPERIMENTAL SECTION Materials. Carboxyl-functionalized paramagnetic beads (2.7-µm in diameter) were purchased from Agilent Technologies (Lexington, MA). Resorufin-β-D-galactopyranoside (RDG) was purchased from Life Technologies (Grand Island, NY). Phosphate buffered saline (PBS) packs, 2-(Nmorpholino)ethanesulfonic acid (MES) buffered saline packs, bovine serum albumin (BSA) in PBS (10%), 1-Ethyl-3-(3dimethylaminopropyl)carbodiimide hydrochloride (EDC) and NHS-long chain-long chain-biotin (BT-LC-LC-NHS) were purchased from Thermo Scientific (Waltham, MA). Tween 20, 1 M magnesium chloride and bovine serum albumin (BSA) were purchased from Sigma (St. Louis, MO). Streptavidin-β-galactosidase (SβG) was purchased from Quanterix (Lexington, MA). Optical fiber bundles were purchased from SCHOTT North America (Elmsford, NY). The fiber polisher and polishing consumables were purchased from Allied High Tech Products (Compton, CA). Single domain antibodies (sdAb) were provided by the Naval Research Laboratory (Washington, DC). Ricin toxin (RCA 60), ricin communis agglutinin (RCA120), ricin A chain, and ricin B were purchased from Vector Laboratories (Burlingame, CA). Rosetta DE3 and pet22b expression vector were purchased from Merck Millipore. Phycoerythrin conjugate with streptavidin was purchased from Columbia Bioscience (Columbia, MD). The Simoa HD-1 Analyzer and Homebrew Assay Kits were purchased from Quanterix Inc., (Lexington, MA). Homebrew kits include carboxyl-functionalized paramagnetic beads (2.7µm in diameter), 1-Ethyl-3-(3dimethylaminopropyl)carbodiimide hydrochloride (EDC), 157

nM Streptavidin-β-galactosidase (SβG) concentrate, 100 µM Resorufin-β-D-galactopyranoside (RDG), diluents (bead diluent, detector diluent, and SBG diluent), wash buffer 1, and wash buffer 2. Recombinant ricin A chain protein (Rivax) and recombinant vaccinia L1R fragment (rL1R) were provided by Edgewood Chemical Biological Center (Aberdeen, MD). Abrin toxoid was provided by Naval Medical Research Center (Silver Spring, MD). Staphylococcal enterotoxins type A-E (SEA-SEE) and Shigatoxin-1 and -2 were purchased from Toxin Technologies, Inc. (Sarasota, FL). Botulinum neurotoxins toxoids (Bot toxoid A-F) were purchased from Metabiologics, Inc. (Madison, WI). Pertussis toxin and Cholera toxin were purchased from List Biological Laboratories, Inc. (Campbell, CA). Recombinant bacillus collagen-like surface protein of anthracis (rBclA) fragment was purchased from Bei Resources (Manassas, VA). Preparation of Single Domain Antibodies. Ricin specific sdAbs C8, D1, H1, and B4, reported previously 27,28, were mobilized into the pet22b expression vector, without the upper hinge sequence, for improved protein production. D12f, reported previously 29, was also mobilized into the pet22b expression vector. Protein was produced in Rosetta DE3 and purified from the periplasm by a combination of immobilized metal affinity chromatography and size exclusion chromatography as previously described. The sdAb were biotinylated (bt-sdAb) using a 10-fold molar excess of BT-LC-LC-NHS 28,30 . Preparation of MAGPIX Assay. Sandwich assays were performed essentially as described previously using the MAGPIX platform (Luminex, Austin, TX) 30. Briefly, sdAb C8, D1, H1, along with control sdAb that recognized a nonricin target were immobilized to MAGPIX carboxylated microspheres using the two-step carbodiimide coupling protocol provided by the manufacturer. The sdAb coated microspheres were incubated with serial dilutions of ricin. The reactions were incubated at room temperature for 30 minutes, washed with PBS with 0.05% Tween 20 (PBST) and then incubated with 1 µg/mL bt-sdAb. The microspheres were washed again and then incubated with 2.5 µg/mL of a streptavidin– phycoerythrin conjugate (Columbia Biosciences) to generate the fluorescent complex. After a final wash and resuspending in 75 µL PBST, the assay results were evaluated on the MAGPIX platform. The obtained values were the median of at least 50 beads. The median fluorescence intensity (MFI) versus ricin concentration is plotted for each sdAb capturedetector pair; error bars are the 95% confidence value of the mean. Preparation of Capture Beads with Single Domain Antibodies. MES buffer (0.1M, 0.9% NaCl, pH 6.2), 1-ethyl-3-(3dimethylaminopropyl) carbodiimide hydrochloride (EDC), and PBS buffer (0.1M, 0.1M NaCl, pH 7.2) were equilibrated to room temperature before use. Carboxyl-functionalized (2.7-µm diameter) paramagnetic beads (100 µL) were washed three times with 500 µL of 1% T20-PBS and twice with 500 µL of MES buffer. Single domain antibodies (100 µg) were incubated with the beads while shaking for 15 minutes (total volume 100 µL). EDC (10 mg/mL) in MES buffer was prepared and 100 µL was added, mixed well, followed by the addition of 1 mL MES buffer. This mixture was incubated while shaking for 30 minutes. Beads were washed once with 1200 µL 1% T20-PBS and then incubated with 1200 µL 1% BSA-PBS for 40 minutes while shaking. Beads were then

ACS Paragon Plus Environment

Analytical Chemistry

Page 4 of 9

Protein Determination, Urine

Protein Determination, Serum

80000

50000

40000

Observed Actual

Ricin, pg/mL

60000

Ricin, pg/mL

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

40000

Observed Actual

30000

20000

20000 10000

0

A

0 1

2

3

4

5

6

B

1

2

Unknown Samples

3

4

5

6

Unknown Samples

Figure 3: Protein determination of unknown high, medium and low ricin concentrations spiked in urine and serum. A). Unknown ricin concentrations spiked in urine were measured to determine which samples were spiked with high, medium and low ricin amounts. Samples 3 and 4 were identified as the samples containing the lowest amounts of ricin while samples 2 and 6 were identified as having the highest amounts of ricin. B). Unknown ricin concentrations spiked in serum were measured to determine which samples were spiked with high, medium and low ricin amounts. Samples 1 and 5 were identified as the samples containing the lowest amounts of ricin while samples 2 and 6 were identified as having the highest amounts of ricin.

washed three times with 1% T20-PBS and stored in 200 µL bead storage buffer (50 mM Tris-HCl,150 mM NaCl, 10 mM EDTA, 0.1%, Tween-20, and 1% BSA). A two-step EDCcoupling protocol provided by Quanterix Inc. was used for preparation of sdAb beads for the HD-1 Analyzer. Briefly, 100 µL of pelleted beads were incubated with 190 µL of bead conjugation buffer and 10 µL of 10 mg/mL EDC solution. The bead-EDC mixture was vortexed and incubated while shaking for 30 minutes. The beads were then pelleted, washed, and incubated with 200 µL of 0.5 mg/mL sdAb solution. The solution was vortexed and incubated while shaking for 120 minutes. The beads were washed and blocked using 200 µL of bead blocking buffer for 30 minutes. After final washing, the sdAb capture beads were stored in bead diluent buffer at 4ᵒC. Capture of Ricin and Formation of Streptavidin-βGalactosidase Labeled Immunocomplexes. sdAb coated paramagnetic beads were incubated with ricin toxin at concentrations ranging from 10 ng/mL to 0.01 pg/mL in 1x PBS containing 0.3% Tween 20 and 0.1% BSA for 1 hour with shaking at 1,000 rpm. Undiluted urine and serum samples spiked with ricin toxin at concentrations ranging from 100 µg/mL to 1 pg/mL were diluted 1:100 (resulting in 1µg/mL to 10 fg/mL samples) in 1x PBS containing Tween 20 and BSA prior to incubation. After 1 hour, 1 µL of 1.5 µg/mL bt-sdAb were added and incubated while shaking for an additional 1 hour. Beads were washed 6 times with 5x PBS containing 0.1% Tween 20 wash buffer and incubated with 20 pM SβG in PBS buffer containing 0.5% FBS and 1 mM MgCl2 for 30 minutes while shaking. Beads were washed 12 times with wash buffer and once with sucrose buffer. Following the sucrose wash, beads were reconstituted in 10 µL of sucrose buffer and loaded onto fiber strips containing ~ 50,000 46-fL reaction wells. The strips were centrifuged at 10,000 x g for 5 minutes to trap individual beads into wells and imaged using an 8-fiber imager.

Limits of Detection in Biological Fluids. Human urine and serum samples spiked with ricin were prepared by the Naval Medical Research Center. Concentrations ranging from 100 µg/mL to 1 pg/mL were prepared in undiluted human urine and human serum at 50 µL volumes. Control samples were prepared with no ricin. Samples were shipped overnight and immediately diluted 100 times to achieve concentrations between 1 µg/mL and 10 fg/mL. Diluted samples were stored at 4oC for future analysis. Protein Determination in Biological Samples. Unknown samples containing ricin and blank samples containing no ricin were provided by the Naval Medical Research Center. Samples were diluted 100 times in assay buffer and the Simoa was performed as described above. Calibration curves were used to determine the ricin concentration of the unknown and blank samples. The linear-linear plots generated were used to obtain an equation for fitting AEB values. Percent error was used to determine the accuracy of each measurement. Imaging and Analysis. Imaging and analysis have been described previously 3,4. Two imaging platforms were used, an 8-fiber imager for the four-hour manual assay and a fully automated Simoa HD-1 Analyzer. Two quantification methods, digital and analog, provide Simoa with a wide dynamic range. The digital method of Simoa relies on the system following a Poisson distribution and derives from the ability to singulate and therefore count enzymatically "on" (active) wells. The analog method permits quantification at higher concentrations by measuring the entire fluorescence intensity. The digital read-out is determined by the ratio of active wells to total wells containing beads and is expressed as percent (%) active. Active wells are the number of wells within the array that contain a bead and exhibit at least a 20% increase in fluorescence in the four images from the five frames collected (frame 1 is t=0). A switch from the digital read-out to the analog read-out occurs under conditions that support >70% active beads because beads will be counted as "on" regardless of whether they contain one or multiple enzyme-labeled proteins. Since the

ACS Paragon Plus Environment

Page 5 of 9

Analytical Chemistry Inclusivity, Buffer 10

10

Serum Urine Buffer 1

RCA 60 RCA 120 A chain B chain Rivax

1

0.1

AEB

AEB

0.1

0.01

0.01 0.001 0

0.01

0.1

1

10

100

1000

Ricin, pg/mL

0.001 1

10

100

1000

Ricin, pg/mL

Inclusivity, Urine 10

RCA 60 RCA 120 A chain B chain Rivax

1

AEB

counting method cannot distinguish between a bead with one enzyme molecule vs. a bead with many enzymes, the analog method is used to convert the average intensity of the bead into a number that corresponds to the average enzyme per bead (AEB) 31. Combining these two methods permits the quantification of samples in the low fg/mL to high pg/mL range 31. All images were acquired at an excitation and emission wavelength for resorufin 31 using a CCD camera and standard imaging optics 4. Automated Simoa HD-1 Analyzer. 100 µg of sdAb were coupled onto paramagnetic beads according to the protocol stated above. 5x106/mL sdAb coated capture beads, 2 nM btsdAb, and 150 pM SβG solution were separately loaded into 14-ml plastic bottles. 100 µM RDG solution was loaded into a 4-mL plastic bottle. Each bottle was used for automatic dispensing in the HD-1 Analyzer. Ten-fold dilutions of ricin toxin, 0.01-1000 pg/mL, were prepared in triplicate in 1x PBS containing 0.3% Tween 20 and 0.1% BSA on a 96-well plate. The assay was performed using the 3-step assay mode. sdAb capture beads were incubated with 100 µL of each ricin toxin concentration for 15 minutes. Beads were then washed three times with wash buffer 1, and were then incubated with 100 µL of bt-sdAb solution for 5 minutes. Beads were washed again with wash buffer 1 and were incubated with 100 µL of SβG solution for 5 minutes. The beads were washed six times with wash buffer 1, one time with wash buffer 2, and resuspended in 25 µL RDG solution. All incubations were performed while shaking. 15 µL of capture paramagnetic beads in RDG solution were then loaded by gravity onto an array with 215,000 femtoliter reaction wells on a 24-array Simoa Disc32. The wells were sealed with fluorocarbon sealing oil. All results were represented as an AEB value. RESULTS Single and ensemble measurements of proteins have been previously described 31. Here we developed a Simoa for the detection of ricin toxin in biological fluids using two platforms, a manual 8-fiber imager and a fully automated Simoa HD-1 Analyzer. Ricin was captured by paramagnetic beads coated with sdAbs. Each captured protein molecule was labeled by the addition of bt-sdAbs, followed by SβG.

0

A

Figure 4: Combined digital and analog calibration curves using the Simoa HD-1 Analyzer. The log-log standard curve in each matrix provided a 5-log concentration linear dynamic range with a measured LOD of 1 pg/mL. The dashed line indicates the boundary between the digital and analog read-out.

0.1

0.01

0.001 0

1

B

10

100

1000

Ricin, pg/mL

Inclusivity, Serum 10

RCA 60 RCA 120 A chain B chain Rivax

1

AEB

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

0.1

0.01

0.001

C

0

1

10

100

1000

Ricin, pg/mL

Figure 5: Inclusivity tests in buffer using the Simoa HD-1 Analyzer. Each of the standards were performed in triplicate at 10fold dilutions ranging from 1 to 1000 pg/mL. In each matrix used, (A) buffer, (B) urine and (C) serum, RCA60 was detectable at all concentrations tested while RCA120 was only detectable at the 1000 pg/mL concentration. All other isoforms were undetectable at all concentrations.

ACS Paragon Plus Environment

Analytical Chemistry

Exclusivity 8

Buffer

6

AEB

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 9

Urine Serum

4

2

0 rL1

R

l lA -1 -2 s is oid S EA SEB EC 1 EC 2 S ED SEE id A id B id C id D id E id F nt ro lank era rBc t ox in t ox in rt u s hol Tox o S S xo xo xo xo B xo xo C rin a a e To t To t To t To t To t To v e C g g P i i t Sh Sh Ab Bo Bo Bo Bo os it i Bo Bo P

Figure 6: Exclusivity tests in buffer, urine, and serum using Simoa HD-1 Analyzer. Each of the standards was performed in triplicate at 1000 pg/mL. Only ricin (RCA60) was detectable at 1000 pg/mL in all matrices. All other toxins or toxoids or protein fragments were undetectable.

Individual beads were then isolated in femtoliter reaction wells in the presence of fluorogenic substrate, RDG, and sealed using a silicon gasket. Measuring the high local concentration of fluorescence in the reaction wells allowed the interrogation of individual protein molecules. Bulk Assays. The sdAbs C8, D1, H1, and B4 have been shown to bind distinct epitopes on ricin and can be used together as both capture and reporter reagents in sandwich immunoassays 27,28. Combinations of these sdAbs were used in MAGPIX assays. Previously, we determined that while B4 works well as a capture antibody in an ELISA format (Supplementary Figure 1), it did not function well when covalently immobilized onto microspheres; hence, D1, C8, and H1 were tested as capture sdAbs along with bt-C8, bt-D1, bt-H1 and btB4 sdAbs as detectors. Each sdAb was specific for a different epitope on the ricin target. C8 proved to be the best capture antibody in MAGPIX assays; when paired with bt-D1 it was possible to achieve detection down to 64 pg/ml with a signal to background ratio of ~9 (Figure 1). The D12f sdAb was developed as an improved sdAb that binds to the same epitope as C829. The improvement was mostly realized as an even higher thermostability with an equivalent binding affinity (Supplementary Figure 2). Digital and Analog Calibration Curves for Protein Determination using the Manual 8-Fiber Imager. The ricin Simoa was tested with two different capture sdAbs, C8 and D12f, and optimized for incubation times and Tween 20 concentration. All Simoa assays were performed using a D12fD1 sdAb pair, which was shown to perform similarly to the C8-D1 sdAb pair in bead based bulk ELISA (Supplementary Figure 3) and outperform the C8-D1 pair during Simoa optimizations due to a decrease in background signal

(Supplementary Figure 4). C8 and D12f have similar complementarity determining regions and bind the same epitope on ricin. We chose to assess the performance and application of the Simoa in buffer and two biological fluids in which ricin (or low molecular weight metabolites) has been detected— urine and serum. The LOD threshold was determined by the standard analytical criterion of three times the standard deviation of the background AEB. The Simoa was 1,000 times more sensitive than the sdAb-MAGPIX assay with a measured limit of detection (LOD) of 10 fg/mL (166 aM) in PBS buffer with a six-log concentration linear dynamic range (Figure 2). To establish the dynamic range of the Simoa in serum and urine, serial dilutions ranging between 100 µg/mL to 1 pg/mL were measured after a 1:100 dilution (1µg/mL to 10 fg/mL) in 1x PBS containing Tween 20 and BSA. The two highest dilutions tested, 1 µg/mL and 100 ng/mL, were not reported due to a loss of fluorescence intensity from frame 1 to frame 5 as a result of signal saturation at frame 1. As little as 100 fg/mL and 1 pg/mL of ricin was detectable in 1% urine and 1% serum, respectively (Figure 2). Each linear-linear plot had a R2 > 0.999. These curves demonstrate the ultra-sensitive capabilities of the Simoa for the detection of ricin in multiple sample matrices and exhibit broad linear dynamic ranges in each matrix. Given the inherent differences in sample composition and the sensitivity of the assay to the different sample matrices, calibration curves generated for 1% urine and 1% serum were used to determine the ricin concentration in unknown samples. These calibration curves enabled extrapolation of unknown sample concentrations based on AEB values. Protein Determination from Unknown Samples. Unknown samples provided by the Naval Medical Research Cen-

ACS Paragon Plus Environment

Page 7 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

ter were tested using the manual 8-fiber imager. Six unknown samples were provided in both urine and serum. Two high, two medium and two low samples were obtained to span the working range of the Simoa. Unknown samples 2 and 6 in both urine (Figure 3A) and serum (Figure 3B) generated the highest AEB values. Samples 3 and 4 in urine and samples 1 and 5 in serum generated the lowest AEB values. Samples 3 and 4 also provided the highest CV values in urine. All other CV values were