Coupling Single Domain Antibodies to Catalytic Hairpin Assemblies for Homogeneous Assays

Immunoassays are widely used in various fields, including biomedical research, clinical diagnostics, and environmental monitoring. Single domain antibodies (sdAbs) provide small, tailorable, recognition elements that have been integrated into immunoassays for detecting a myriad of targets. Deoxyribonucleic acid (DNA) circuits are synthetic molecular devices composed of DNA strands that can perform logical or computational operations. They have a range of applications, including biosensing, diagnostics, and drug delivery. Recently, an immunoassay method was reported that used catalytic hairpin assemblies (CHA) with antibodies for homogeneous protein detection. The CHA process uses DNA hairpins that interact in the presence of a catalytic DNA sequence. This paper presents a new strategy to couple the recognition of sdAbs with CHA circuits using genetic fusions of sdAbs with rhizavidin (rz), a dimeric biotin binding protein. A pair of sdAb‐rz constructs is each functionalized with a biotinylated DNA sequence that represents half of the catalyst. When both sdAbs bind to the target protein, a signal is generated through the CHA circuit. The split catalyst approach amplifies signals through a DNA circuit without washing steps. Furthermore, this method distinctively utilizes programmable DNA circuits, which are highly modular and can accommodate new targets without disrupting the assay.


Introduction
Immunoassays are analytical techniques in which antibodies serve as recognition elements, and the interaction between the DOI: 10.1002/adsr.202300111antibody and antigen can be used for quantification. [1]They play an important part in a number of applications including biomedical research, clinical diagnostics, environmental monitoring, and food testing.While other innovative approaches such as DNA microarrays, lab-on-a-chip devices, microfluidics, and nanoparticle-based assays are being explored elsewhere, this paper focuses on the immunoassay-based approach.There are many configurations of immunoassays, with the enzyme linked immunosorbent assay (ELISA) being one of the most widely utilized. [2]The ELISA is a heterogeneous assay requiring several incubation and washing steps.In contrast, homogeneous assays require no washing steps and consist of simply mixing the sample with the detection reagents and measuring signal in one-pot.
Single domain antibodies (sSdAbs), also known as nanobodies, are the recombinantly expressed variable heavy chains derived from heavy chain only antibodies found in sharks and camelids. [3]An advantage of recombinantly expressed antibody binding domains, such as sdAbs is that they can be engineered for specific assay formats.In general, sdAbs and genetic fusions of sdAbs with other protein domains are soluble and well expressed.SdAbs have been integrated into immunoassay formats for a variety of target antigens ranging from ricin toxin, [4] Ebola, [5] Covid 19, [6] and caffeine. [7]NA circuits, also known as nucleic acid circuits, are artificial programmable molecular devices composed of DNA strands that can carry out logical or computational operations.[8,9] They are typically constructed by carefully designing DNA sequences with specific base-pairing properties that allow them to self-assemble into desired structures or circuits.[10][11][12] Earlier research results suggest that DNA circuits can be used for a variety of applications, such as biosensing, diagnostics, and drug delivery.[10,13] One of the most notable features of DNA circuits is their ability to function without the need for enzymes or other external factors, making them suitable for use in a wide range of environments.In addition, DNA circuits can be designed to perform a range of functions, including amplification of nucleic acid sequences, detection of specific DNA or RNA sequences, and even computation.[8,[14][15][16] They can also be used in combination with other molecular components, such as proteins or nanoparticles, to create hybrid devices that can carry out more complex tasks.[8,10,17] A recent study introduced a novel method that enables the use of catalytic hairpin assemblies (CHA), [18] an amplification type of DNA circuit, [19][20][21] for protein detection in a homogeneous assay format.[22] Unlike the enzyme-linked immunosorbent assay (ELISA) which requires washing steps and other sample manipulation, this format has a simplified method, and signal amplification is achieved through a specialized DNA circuit. In paticular, the CHA system involves the use of DNA hairpins that interact in the presence of a catalytic DNA sequence, which can be integrated with several schemes for generating output, including fluorescence reporters.[18] The key to integrating CHA amplification with antibody recognition was coupling the protein recognition event to the generation of the DNA catalyst.This was achieved using a "split catalyst" strategy where antibodies are coupled to DNA which forms the catalyst only when the antibodies bind their target antigen.In this paper both the antibody and DNA were biotinylated, and the antibody-DNA conjugate was formed using streptavidin to bridge the antibody and DNA.[10] This method combines the strengths of immunoassays and DNA circuits. By elimnating the need for washing and utilizing DNA amplification, this method offers a faster and simpler assay format compared to a traditional ELISA.It's worth to highlight that sensitive one-step ELISA protocols have been developed that cut down the number of incubation steps and, can achieve detection limits down to the pg mL −1 range.[23] However, the onestep ELISA still involves at least one washing step, and requires individual optimization for each assay, involving meticulous adjustments to the molar ratio between the primary and secondary antibodies, or calibration of the color development duration.
Here, we implemented a new strategy to couple the recognition of sdAbs with CHA circuits.Our method employs genetic fusions of sdAbs with rhizavidin (rz), a dimeric biotin binding protein [24,25] that can bind to any biotinylated DNA.This is a homogeneous assay requiring no washing steps.Only the use of a different sdAb-rz pair is required to modify the assay for a new target.Using a split catalyst approach, we demonstrated the detection of multiple protein targets: the toxin ricin, the nucleocapsid protein from SARS-CoV-2, and chikungunya virus like particles.As with the method using biotinylated traditional antibodies, [22] this method required two different sdAb-DNA conjugates to bind the target simultaneously in order to form the split catalyst and initiate the DNA amplification circuit.An advantage of our approach using sdAb-rz is that it provides a more defined assay that is not dependent on the extent and location of antibody biotinylation.The design scheme of our strategy is illustrated in Figure 1 and details of the split catalyst are shown in Figure S1, Supporting Information .For simplicity, the complex of sdAb-rz with biotinylated DNA is abbreviated as sdAb-DNA.At the onset of the process, both sdAb1-DNA and sdAb2-DNA simultaneously bind to the target molecule.As a result of this binding, the two DNA strands come into close proximity, enabling them to form a hybrid DNA strand.This hybrid DNA strand serves as the catalyst that triggers the CHA amplification circuit.Initially, the catalyst hybridizes with the ACH1 hairpin via domains 1, 2, and 3, which induces conformational changes in the ACH1 hairpin, causing the remaining domains to open up.This allows the ACH1 hair-pin to hybridize with the ACH2 hairpin.The hybridization process continues until the ACH2 hairpin has opened up completely, releasing the catalyst, allowing it to participate in another reaction cycle resulting in further signal amplification.
A reporter DNA complex, called ACRP, that consists of two strands, one labeled with a fluorophore and the other with a quencher that engage in FRET (Foster resonance energy transfer) is utilized to detect the completion of each cycle.The ACRP complex binds to the exposed domains on the ACH1 hairpin, and undergoes a strand displacement mechanism. [9]This mechanism effectively separates the quencher from the fluorophore, disrupting the FRET pair and resulting in detectable fluorescence emission.

Ricin Assay and Optimization
We chose the toxin ricin as the first target for demonstration of this homogeneous assay format.It has been implicated in both accidental and intentional poisoning; rapid and reliable detection of ricin is important for timely administration of supportive care. [26]Ricin is a ≈60 kDa heterodimeric toxin consisting of an A chain and a B chain, and sdAbs recognizing both the A and B chain of ricin have been isolated, including sdAbs that have been engineered for increased stability. [4,27,28]Additionally, genetic fusions of anti-ricin sdAbs with proteins such as rz, a dimeric biotin binding protein, have been produced. [25,29]e chose four anti-ricin sdAbs, two with specificity for the A chain (D12f and F6H2Y), and two for the B chain (B4 and B7), as candidates to express as fusions with rz.Prior to production of the fusions we confirmed that all four sdAbs were compatible in a sandwich format (Figure S2, Supporting Information).These experiments confirmed previous work that had shown that the D12f and F6H2Y sdAbs bind to distinct epitopes and performed well when used in a sandwich immunoassay format. [30]The four rz fusions produced well, with typical yields of D12f-rz, B4-rz, and F6H2Y-rz between 20 and 36 mg mL −1 ; B7-rz production was lower, averaging 5 mg mL −1 .
We exclusively focused on the D12f-rz and F6H2Y-rz sdAb pair in the first sets of experiments to couple sdAb-rz with biotinylated DNA and integrate them into a homogeneous assay with signal generation by CHA.The biotinylated strands that are conjugated to the sdAb-rz and can come together to form the split catalyst are termed TB and B*C.Sequences for TB and B*C were based off the prior work. [22]Each of these DNA fragments contained a terminal biotin at either the 5′ end for TB or the 3′ for B*C; a poly T linker, a six base complementary region, and a portion of the catalyst sequence (Figure S1, Supporting Information).Typically D12f-rz was conjugated to the TB strand while F6H2Y-rz was conjugated to B*C.
Our initial experiments utilized either TNaK (a tris-based buffer with both sodium and potassium chloride) or phosphate buffered saline (PBS) as the reaction buffer and resulted in poor detection, with very low signal.By adding MgCl 2 to the PBS buffer, we realized greatly improved signal.At maximum we used a top MgCl 2 concentration of 12.5 mm, a value used in previous work with structural DNA [31] ; PBS with 12.5 mm MgCl 2 is referred to as PBSMg.As shown in the left panel of Figure 2,   even reducing the concentration of MgCl 2 by 25% resulted in decreased signal.The stability of DNA circuits is influenced by the concentration of MgCl 2 . [31]Existing research has shown that elevated concentrations of MgCl 2 (> 12.5 mm) minimally alter the structural integrity of DNA components. [32]Consequently, further increasing the MgCl 2 concentration is not expected to substantially raise the detection.
Another important variable was the addition of free biotin to the conjugation mix after the sdAb-rz and biotinylated oligonucleotide had been mixed for 30 min.A gradient of biotin concentrations was selected to effectively investigate the impact of introducing additional biotin and, if needed, fine-tune the optimal levels of free biotin to effectively quench the conjugation reaction.The effect of added biotin is shown in the right panel of Figure 2. In the absence of ricin, there is a large signal when no biotin is added, but adding biotin substantially reduces the background.However, even with the addition of biotin, we have not been able to achieve background signal as low as seen with samples that include only the CHA mix (ACH1, ACH2, and ACRP).A series of control experiments ensured that there was no signal when only one of the sdAb-DNA constructs was added in the presence of ricin, or when ricin was added by itself to CHA mix with no sdAb-DNA (Figure S3, Supporting Information).A series of negative control experiments in which no ricin was present were also performed as shown in Figure S3 (Supporting Information).
Using PBS with 12.5 mm MgCl 2 and biotin added to at least 40 μm after the sdAb-rz conjugation to biotinylated DNA, we examined different combinations of anti-ricin sdAbs as shown in Figure 3. Similar to the sandwich immunoassay (Figure S2,Supporting Information), the best pair was D12f and F6H2Y, both of which bind to the ricin A chain.However, the F6H2Y and B7 pair also performed well and could be utilized in situations where it is important to verify the presence of both ricin chains.Although the A chain of ricin is responsible for ribosome inactivation, purified A chain has minimal toxicity as the B chain is required for cellular entry.Therefore, the intact heterodimer of ricin is necessary for its potent toxicity. [26] performed a dose-response curve to investigate the detection of ricin using the D12f and F6H2Y pair, as shown in Figure 4. Reliable detection of ricin was achieved down to 3.1 nm, as indicated by the green stars.Although we were able to detect 1.6 nm ricin above background in some instances, detection at this level was not consistently reproducible.

Detection of Additional Targets
Our second target was the nucleocapsid protein from SARS-CoV-2 (N).N is one of the most abundant proteins in cells infected by SARS-CoV-2, and provides a good target for diagnostic assays. [33]e had previously developed three sdAbs (E2, C2, and B6) that can be used in different combinations in sandwich assays for the detection of N. [6] Unlike with ricin, where we only had a general  idea of the epitope recognized by the sdAbs, the crystal structure of each of the three sdAbs in complex with domains of N have been solved. [34]Both C2 and B6 bind adjacent epitopes on the N terminal RNA binding domain, and E2 recognizes the C terminal dimerization domain.Each of the combinations of sd-Abs was examined (Figure S4, Supporting Information).All of the pairings were successful, while the B6f and E2 combination showed the highest signal.We performed a dose-response curve using this pair and were able to detect down to 3 nm as illustrated in Figure 5.Our assay performed at least equivalent to the BIO-CREDIT COVID-19 Ag commercial rapid test which was shown to detect to 250 ng mL −1 (≈6 nm) N as assessed in a 2020 study, however the other six antigen tests evaluated were able to detect at least 25 ng mL −1 (≈0.6 nm), with two detecting reliably to 2.5 ng mL −1 (0.06 nm). [35]Ultimately more sensitive detection is desired to produce a test that provides a reliable measure of diagnosing COVID-19 infections, however our study demonstrates the potential utility of this assay.
The final target we examined was chikungunya virus (CHIKV) in the format of virus like particles (VLPs).CHIKV is a reemergent mosquito borne virus that is found in many parts of the world including Africa, Asia, and the Americas.Although usually not fatal, CHIKV can cause debilitating symptoms with roughly 30-40% of infected people experiencing long-term symptoms such as arthralgia. [36]VLPs are virus surrogates that contain the envelope proteins displayed on the surface of the virus, but lack the virus's genetic material so are not infectious.We had previously isolated sdAbs that bind CHIKV VLPs and clone CC3 was demonstrated to work in sandwich assays for detection. [37]or this target we conjugated the same sdAb-rz to each of the DNA components of the catalyst.There are multiple copies of the CHIKV envelope proteins on the surface of the VLP, which affords many places that CC3 can bind on the same VLP.We first performed the assay using biotinylated DNA that contained different lengths of poly T linker between the biotin and complementary segment (Figure S5, Supporting Information).Each worked essentially the same.We chose to use a medium poly T linker to perform a dose response experiment with CHIKV VLPs, data is shown in Figure 6.We achieved reproducible detection at 33 pm, and often could detect 16 pm (Figure S6, Supporting Information), where the concentration is in terms of VLPs that have multiple copies of envelope protein on their surface.Further work may improve the reproducible limit of detection.These experiments show the feasibility of using the assay to detect samples containing live virus, which displays the same envelope proteins that are assembled on the VLP.

Conclusions
We have successfully demonstrated the ability to combine sdAb recognition with a DNA amplification circuit.While this represents a promising proof of concept, in order to make the assays more useful, we need to improve the limits of detection by at least five-fold.One potential avenue for improvement is to investigate a variety of concentration conditions for the CHA components and sdAb-DNA to determine how the current system can be further optimized.Additionally, another strategy is to couple the DNA circuit with a second stage of CHA to achieve further amplification. [38]This approach can result in several orders of magnitude of additional amplification, but it is crucial to carefully design and prepare the assay to minimize background in the absence of target.By exploring these options, we can work toward improving the performance of our system and increasing the usefulness of the assays.The use of sdAbs with DNA circuits in immunoassays holds great promise for the development of more efficient, sensitive, and specific assays, and could potentially augment the field of detection technology.

Experimental Section
Reagents and Instrumentation: Unless otherwise specified, chemicals were purchased from VWR (Radnor, PA), Thermo Fisher Scientific (Waltham, MA), or Sigma-Aldrich (St. Louis, MO).Molecular biology reagents were sourced from New England Biolabs (Ipswich, MA).Ricin was purchased from Vector Laboratories (Newark, CA), N protein from SARS-CoV-2 was from ACRO Biosystems (Newark, DE), and Chikungunya VLPs were purchased from The Native Antigen Company (Kidlington, UK).
Oligo DNA sequences were synthesized by Integrated DNA Technologies (Coralville, IA); all non-dye labeled oligonucleotides were purchased with polyacrylamide gel electrophoresis (PAGE) purification, and all dye-labeled oligonucleotides were HPLC purified.The hairpin, reporter, and catalyst sequences were designed based on those previously reported [22] and are provided in Table S1 (Supporting Information).
With the exception of B6f, all sdAb sequences have been reported previously (ricin binding sdAbs: D12f, [27] F6H2Y, [30] B4, [4] B7 [28] ; N binding sdAbs: C2 and E2, [6] and CHIKV binding sdAb: CC3 [37] ).B6f is a derivative of the N binding sdAb, B6, [6] which has point mutations in the framework that enabled improved protein production.Genetic fusions of sd-Abs with rz were produced as described previously. [29]Sequences of all new constructs were verified using sequencing services through Eurofins Genomics (Louisville, KY); the protein sequence of all the sdAb-rz sequences used in this work are listed in Table S2 (Supporting Information).The sdAb-rz fusion proteins were produced and purified according to our previously described protocol for the production of sdAb constructs. [6,37]oncentration of the sdAb-rz was based on the absorbance at 280 nm as measured in a Nanodrop and the extinction coefficient calculated from the amino acid sequence using the expasy ProtParam tool. [39]The sdAbrz were aliquoted and stored frozen at −80 °C until use.Tecan Spark plate readers were employed for this work.All offered temperature control above room temperature, and one of the plate readers was able to both heat and cool.
Preparation of CHA Reagents: Each of the oligonucleotide was dissolved at 100 μm in TE buffer.Oligonucleotides were either stored refrigerated or aliquoted with some portions kept at −20 °C for longer term storage.Prior to annealing, the hairpins, ACH1 and ACH2 were diluted to 5 μm in TNaK buffer (20 mm Tris pH 7.5, 140 mm NaCl, 5 mm KCl).To prepare the reporter (ACRP), ACRPF (the fluorophore containing strand) at 5 μm was mixed with ACRPQ (the quencher containing strand) at 10 μm in TNaK.For annealing, the oligonucleotide solutions were heated at 90 °C for 10 min and cooled to 4 °C at a rate of 1 °C per minute.Annealed CHA reagents were produced in 200 μL batches and kept refrigerated or on ice until use.The batches of annealed oligos were stable for at least two weeks.
Coupling of sdAb-rz to DNA: The biotinylated strands that are conjugated to the sdAb-rz and can come together to form the split catalyst are termed TB and B*C (see Table S1 and Figure S1, Supporting Information).Versions of these oligonucleotides with three poly T linker lengths (short, medium, and long with lengths of 7, 16, and 25 bases respectively) were procured.Figure S1 (Supporting Information) shows the parts of the two DNA strands (TB and B*C) that comprise the split catalyst which include the biotin, a poly T linker, complementary domain, spacer, and the split catalyst sequence.Unless specified otherwise, ricin detection was with the short linker, while the medium poly T linker was used for N protein and Chikungunya VLP assays.
Detection requires both an sdAb-rz construct conjugated to TB and a second one conjugated to B*C.Depending on the nature of the target antigen the two sdAbs should either recognize different regions of the target, or in cases like a Chikungunya VLP which contain repeating protein motifs, the same sdAb can be coupled separately to TB and B*C.
Freshly thawed aliquots of sdAb-rz were used for coupling to biotinylated DNA (TB or B*C).Coupling was performed in PBSMg (PBS with 12.5 mm MgCl 2 ).First 1 μL of the sdAb-rz (37 μm stock) was added to PB-SMg for a final concentration of ≈1.5 μm.Next 0.5 μL of the biotinylated DNA (100 μm stock) was added for a final concentration of ≈2 μm.After a 30 min incubation at room temperature, biotin was added and the samples were put on ice.Typically, the addition of biotin consisted of 10 μL of either a 1 mm or a 0.5 mm biotin solution in PBS, however other conditions such as 10 μL of 0.1 mm were also explored.The sdAb-DNA was either used the same day or stored at −20 °C.
Performing Assays: Assays were performed in black 384 well low bind plates.Conditions were always run in duplicate, and the same conditions were tested on at least two separate days.Typically per well we added 50 μL buffer (PBSMg unless otherwise specified), 1.5 μL each of the sdAb-DNA in the pair (each sdAb-DNA made up as described above), 0.8 μL target in positive wells, and 3 μL CHA mix.The CHA mix consisted of equal volumes of the annealed ACH1, ACH2, and ACRP.CHA mix was always used within a few minutes of preparation.Immediately after the CHA mix was added the plate inserted into a Tecan plate reader and data recorded using the settings for FAM (exciting at 450 nm and recording emission at 520 nm).Read times were typically at least 30 min and not longer than 90 min.Experiments were performed at 37 °C unless noted otherwise.
Statistical Analysis: Background noise was corrected in the preprocessing data by subtracting the initial fluorescence signal from the kinetics fluorescence data, prior to the addition of the target.All kinetics runs were performed in duplicate and the data are presented as mean ± SEM.SigmaPlot was used for graphing all data, unless otherwise specified.

Figure 1 .
Figure 1.Scheme for coupling target recognition by sdAbs with CHA-based amplification.This represents a homogeneous assay where the DNAfunctionalized sdAbs, DNA hairpins (ACH1 and ACH2), and reporter complex (ACRP) are all present in solution.Two sdAb constructs that recognize different epitopes on a target are functionalized with DNA that comprises a split catalyst (sdAb1 with DNA sequence 1 and sdAb2 with DNA sequence 2, 3).Only when both sdAb constructs bind the target simultaneously can the ACH1 hairpin (blue) hybridize to the catalytic sequence (1, 2, 3), opening its hairpin.This enables the second hairpin (ACH2, pink) to hybridize to ACH1, displacing the catalytic segment.The reporter can hybridize with the unpaired region of ACH1 (2, 5, 6) causing an increase in fluorescence.The split catalyst is released and available for further reactions.Sequence segments are indicated by number.An asterisk represents a complementary strand, and gray shading indicates base pairing.The reporter (ACRP) consists of a fluorophore (F) whose fluorescence is quenched by a quencher dye in proximity (Q).When the F-containing strand of ACRP hybridizes with ACH1, the fluorophore is no longer in proximity to the quencher and a fluorescence increase can be monitored.

Figure 2 .
Figure 2. Effect of added MgCl 2 and biotin on target detection.The left panel shows decreasing amounts of MgCl 2 in the buffer used for both making the sdAb-DNA and detecting the ricin target.Percentages refer to the mix of PBSMg with PBS, where the concentration of MgCl 2 in PBSMg is 12.5 mm (100%) and no MgCl 2 is added to PBS (0%).The right panel shows the addition of increasing amounts of biotin when assembling the conjugates of sdAb-rz with biotinylated-DNA.Free biotin is added after the sdAb-rz has incubated for a half an hour with the biotinylated-DNA.In both these plots the filled symbols represent the signal in the presence of ricin (12 nm) and the open symbols are the same condition in the absence of ricin.

Figure 3 .
Figure 3. Combinations of different anti-ricin sdAb-DNA conjugate pairs.The four ricin binding sdAb-rz were utilized in combination with each other for the detection of ricin (25 nm).Samples with ricin have filled symbols and are denoted as "+" in the legend.Samples without ricin have symbols that are not filled and are denoted as "-".The names of the sdAbs have been abbreviated in the legend: D = D12f-rz, F = F6H2Y-rz, 7 = B7-rz, 4 = B4-rz.The first sdAb of the pair was conjugated to the TB DNA strand, while the second was conjugated to the B*C DNA strand.

Figure 4 .
Figure 4. Fluorescence kinetics of dose-response curve for ricin detection.The optimal conditions for detecting ricin were used.The D12f-rz was conjugated to TB and F6H2Y was conjugated to B*C.The minimum concentration of ricin that could be detected reproducibly was 3.1 nm.

Figure 5 .
Figure 5. Fluorescence kinetics of dose-response curve for N detection.The B6f-rz and E2-rz combination was used with B6f-rz conjugated to TB and E2-rz conjugated to B*C.The minimum concentration of N that could be detected reproducibly was 3 nm.

Figure 6 .
Figure 6.Fluorescence kinetics of dose-response curve for CHIKV VLPs detection.The CC3-rz was conjugated separately to TB and B*C.The minimum concentration of CHIKV VLPs that could be detected reproducibly was 33 pm.