Determination of Aspergillus pathogens in agricultural products by a specific nanobody-polyclonal antibody sandwich ELISA

Aspergillus and its poisonous mycotoxins are distributed worldwide throughout the environment and are of particular interest in agriculture and food safety. In order to develop a specific method for rapid detection of Aspergillus flavus to forecast diseases and control aflatoxins, a nanobody, PO8-VHH, highly reactive to A. flavus was isolated from an immunized alpaca nanobody library by phage display. The nanobody was verified to bind to the components of extracellular and intracellular antigen from both A. flavus and A. parasiticus. To construct a sandwich format immunoassay, polyclonal antibodies against Aspergillus were raised with rabbits. Finally, a highly selective nanobody-polyclonal antibody sandwich enzyme-linked immunosorbent assay was optimized and developed. The results revealed that the detection limits of the two fungi were as low as 1 μg mL−1, and that it is able to detect fungal concentrations below to 2 μg mg−1 of peanut and maize grains in both artificially and naturally contaminated samples. Therefore, we here provided a rapid and simple method for monitoring Aspergillus spp. contamination in agricultural products.

In 1993, a group of Belgian scientists found a type of antibody in the blood of camelids (camels, llamas, and alpacas) produce a unique subclass of antibodies that naturally lack light chains, referred to as heavy chain antibody 15 . The variable domain (VH) of such heavy-chain antibodies is formed by only one variable domain (VHH), which contains the antigen binding site 16 . Recombinant expression of these VHHs yields a single domain heavy-chain antibody, termed "nanobody" 17 .
Unlike polyclonal and monoclonal antibodies, nanobody can be isolated together with their coding sequence by phage display, expressed with a high yield with a bacterial expression system and readily extracted from the periplasm space while still retaining their monoclonal properties [17][18][19] . In addition, the low expression yield and poor stability of scFv limit their development 20,21 , while nanobodies have the advantages of strong stability, good solubility, antigen combined with good performance and low immunogenicity than scFv 22 . With these benefits, recent success in generating camelid nanobodies prompted our interest in developed and applied for diagnostic and therapeutic purposes 23,24 , and nanobodies are promising reagents in the next generation of immunoassays.
An increasing number of nanobodies, especially in our laboratory, have been isolated against aflatoxin and applied in immunoassay 25,26 . However, as yet not nanobody which have been described against the antigens of A. flavus or A. parasiticus have been developed into a specific and sensitive sandwich ELISA, which is able to detect the presence of these fungal species. Based on the facts, we prepared two antigens, extracellular antigens and intracellular antigens (mycelia lysate) of A. flavus, as immunogens to generate nanobodies and rabbit polyclonal antibody, which can recognize certain components of A. flavus. After panning, high-affinity Aspergillus-specific nanobodies, EA3-VHH, PO6-VHH and PO8-VHH were successfully isolated from an immunized alpaca nanobody library by phage display. By comparison, PO8-VHH, highly selective to A. flavus was finally chosen to develop a sensitive direct sandwich immunoassay for aflatoxigenic Aspergillus pathogens. Immunoblot analyses demonstrated the binding of the PO8-VHH to the components of extracellular and intracellular antigen from both A. flavus and A. parasiticus. Direct sandwich ELISAs based on the PO8-VHH (capture antibody) and a polyclonal antibody (detection antibody) were developed. The assay is sensitive, accurate, and cost-effective, with limits of detection as low as 1 μg mL −1 in PBS and 2 μg mg −1 mycelia in peanut or maize, making it suitable for rapid detection of Aspergillus spp. contamination in agricultural products. To the best of our knowledge, this is the first report of the development of a nanobody for direct and species-specific detection of A. flavus.

Results
Antigen preparation and nanobody library. To generate antibodies against aflatoxigenic A. flavus, extracellular antigens and intracellular antigens (mycelia lysate) from a representative strain A. flavus 3.4408 producing high level of aflatoxin were prepared as antigen for immunization of alpaca and rabbit. This strain was identified morphologically and chemically to be an aflatoxin producer as determined by HPLC (data not shown). After the sixth immunization, polyclonal antisera from the immunized alpaca and rabbit antisera were analyzed by indirect ELISA. The results showed a clear robust humoral response up to 1:128000. Total RNA was extracted from immunized alpaca's blood and used to transcribe into cDNA for construction of nanobody library. A phagemid library with a size of 7 × 10 8 clones was constructed and 20 clones randomly selected were sequenced to evaluate library diversity, which displayed a good diversity and all contained the expected inserts (data not shown) and was subsequently used for panning.
Biopanning and phage ELISA analysis. The constructed phage library containing approximately 10 12 phage particles was screened against extracellular antigens and mycelia lysate of A. flavus through three rounds of panning. In order to select phage-displayed VHHs that specifically recognize antigens, washing with PBST was increased each round by five times (10,15, and 20 times in the first, second and third rounds, respectively), while the amount of antigens coated in plate were fixed. Under such stringent conditions, titer of the output phage was increased steadily after each round of panning, which means an efficient enrichment of specific phage (Table 1). Subsequently, the finally panned library served as a valuable pool for further selection of highly reactive antibodies.
In order to isolate soluble nanobodies with high affinity, 48 clones were randomly selected and affinity of phage-VHH toward antigens was tested by a phage ELISA. Among them, 4 clones that exhibited a positive reaction toward extracellular antigens and 5 clones toward mycelia lysate of A. flavus with varied signal intensity were selected for further analyses (Fig. 1). Sequencing analyses of the selected 4 clones against extracellular antigens from A. flavus showed one group of antibodies: clone 3, 14, 41 and 48 carried identical nucleotide sequences (marked in Fig. 1A). These clones were then named EA3-VHH, and its deduced amino acid sequences were illustrated in Fig. 2A. And sequencing analyses of the selected 5 clones toward mycelia lysate of A. flavus were divided   into two groups of antibodies: clone 8, 17, 19 and 32 contained identical sequences; clone 6 carried different nucleotide sequence (marked in Fig. 1B). These clones were then named PO6-VHH and PO8-VHH, respectively, and their deduced amino acid sequences were illustrated in Fig. 2B. However, further sequence analysis of these three nanobodies showed that EA3-VHH and PO8-VHH carried identical nucleotide sequences. Therefore, the phage-displayed PO6-VHH and PO8-VHH were confirmed to be originated from alpaca and were used for further characterization.

Soluble nanobody ELISA and immunoblot analysis. To further reveal their binding activity toward
A. flavus and A. parasiticus that also produced aflatoxins (data not shown), soluble ELISAs with two nanobodies purified from large-scale expression were carried out. The results confirmed that both still had the highest reactivity to A. flavus 3.4408, whereas the PO6-VHH had lowest reactive to A. flavus73# and A. parasiticus (Fig. 3). Meanwhile, PO8-VHH has a better binding ability to A. flavus73# and A. parasiticus than PO6-VHH. These results suggested that CDRs in the PO8-VHH, with variation in both the sequence and length, may play an important role in its interaction with antigens resulting in the highest binding activity. For further analyses of antigen-binding properties of the PO6-VHH and PO8-VHH, the mycelia from other A. flavus and A. parasiticus isolated from diseased peanut in China were detected by immunoblots, and the results showed that PO8-VHH could bind to the same antigen displayed a single band (45-kDa) in A. flavus and A. parasiticus, whereas the PO6-VHH could only bind to the antigen in A. flavus 3.4408 but not bind to other A. flavus and A. parasiticus (Fig. 4). These results further indicated that the PO8-VHH apparently contributes a proper structural conformation and is most specific to the antigen that is commonly and constitutively present on fungal mycelia of both A. flavus and A. parasiticus, although the precise component recognized by the nanobody is unknown. Therefore, the PO8-VHH may be a proper candidate for further genetic manipulation for wide use in immunoassays.
To further evaluate the specificity of the PO8-VHH, ten isolates of Aspergillus with different levels aflatoxin and other fungal or bacteria species (Table 2) were used for indirect ELISA analysis. The results indicated that the PO8-VHH had a high affinity for Aspergillus, especially for A. flavus and A. parasiticus but did not cross-react  with other fungal or bacteria species (Table 2). These results congruously suggested that the PO8-VHH is specific to Aspergillus genus and can be able to specifically detect Aspergillus in the condition of complex pollution of different fungal pathogens, as Tsai pointed out that antibodies against moulds are specific at the genus level and only react to closely related genera 3 .
To evaluate the affinity of the PO8-VHH, the detection limit of the PO8-VHH (200 nM≈4 μg mL −1 ) was measured by indirect ELISA with different fungal concentrations. As shown in Fig. 5, the detection limit of PO8-VHH  Table 2. Affinity and specificity of PO8-VHH determined by ELISA analyses. a Plate wells were coated with mycelia of different fungal or bacteria species, followed by adding PO8-VHH. For wells containing the PO8-VHH, mouse anti-HA antibody and HRP-labeled goat anti-mouse antibody were additionally added. Colorimetric reactions were performed by adding TMB substrate and measured the absorbance at 450 nm.
Each sample was carried out in triplicate and the scale presents an arbitrary set of the mean of OD 450nm readings. b Aflatoxins B 1 were determined by HPLC. Aspergillus and other fungal or bacteria species were cultured in Czapek medium at 28 °C for 5 days. Values represent mean ± SD of triplicate assays.
towards A. flavus and A. parasiticus was about 10 μg mL −1 . Therefore, the PO8-VHH expressed in E. coli can be better used for detecting aflatoxigenic Aspergillus pathogens directly and simply.
Standard curve of the sandwich ELISA. To optimize sandwich ELISA detection through the PO8-VHH and polyclonal antibody, the capture antibody and detection antibody were compared for their applicability and sensitivity. The A. flavus and A. parasiticus mycelia with different concentrations in PBS (0, 10 −2 , 10 −1 , 1, 10, 10 2 , 10 3 , 10 4 μg mL −1 ) were used to construct a standard curve. Additionally, a standard curve for the polyclonal antibody (capture antibody) and PO8-VHH (detection antibody) was also constructed, but the value of blank was so high no matter how optimized the concentration (data not shown). Finally, a standard curve for the PO8-VHH (capture antibody) at 0.5 μg mL −1 combined with polyclonal antibody (detection antibody) at 1 μg mL −1 was constructed though optimization (Fig. 6). The results showed that the detection limits were about 1 μg mL −1 of A. flavus and A. parasiticus for sandwich ELISA.
Sandwich ELISA for rapid immunological detection of Aspergillus contamination in agricutural products. To determine the limit of quantification, different concentrations of mycelia from A. flavus 3.4408 and A. parasiticus mixed with grinded peanut or maize (1 mg mL −1 ) were prepared, and the homogenates were assayed by the optimized sandwich ELISA. The results showed that the detection limit for the PO8-VHH and polyclonal antibody was approximately 2 μg mg −1 , which was identical to the detection limit in PBS but with a lower and slower color development (Fig. 7). Under the optimized sandwich ELISA condition, the fungal concentration displayed a good correlation between logarithmic concentration of mycelia and OD 450nm values (R 2 > 0.99 for both fungi).
To study the feasibility of PO8-VHH combined with polyclonal antibody for rapid immunological detection, sandwich ELISA was used to detect naturally Aspergillus contaminated peanut and maize samples collected   Table 3). The results indicated that both samples can be detected effectively without any false-positive signals in the healthy peanut and maize. Therefore, sandwich ELISA based on the prokaryote-expressed PO8-VHH may be preferable for monitoring naturally Aspergillus-contaminated samples in agricultural products.

Discussion
In this study, two kinds of antigens, extracellular antigens and intracellular antigens (mycelia lysate) from A. flavus were designed to include the surface proteins and crude extract proteins of A. flavus. In order to obtain high affinity nanobodies for A. flavus, we first explored the dose of antigens immunized rabbits to get high titer alpaca antiserum. Results indicate that about 1 mg mL −1 extracellular antigens and 100 mg mL −1 mycelia lysate (fungal concentration) were confirmed as good immunogen for eliciting antibodies.
In order to select phage-displayed VHHs that specifically recognize antigens, a optimized solid phase protocol was performed to select the nanobodies through 3 rounds of panning. Washing with PBST was increased each round by five times, while the amount of antigens coated in plate were fixed so as to efficient enrichment of selective phage. Finally, EA3-VHH, PO6-VHH and PO8-VHH were successfully isolated and cloned the gene of a high-affinity Aspergillus-specific nanobody from an immunized alpaca antibody library by phage display. By comparison, EA3-VHH and PO8-VHH carried identical nucleotide sequences. It is speculated that PO8-VHH recognized protein of A. flavus could be easily secreted and therefore easily detected. Therefore, it is possible to detect A. flavus in the samples, and this nanobody could be used to establish an immunosensor assay for detection of the samples. And results of both soluble nanobody ELISA and sandwich ELISA showed that PO8-VHH could recognizes an antigen that both exist in extracellular and intracellular of A. flavus and A. parasiticus. Specificity and reactivity assays indicated that the PO8-VHH had binding for all Aspergillus species and had no cross-reactivity with non-Aspergillus species.
Moreover, a nanobody-polyclonal antibody sandwich ELISA based on the PO8-VHH was optimized and developed. This assay was able to detect the fungus at a concentration as low as 1 μg mL −1 , the same magnitude as reported by others 10 . It is a pity that the detection limit in peanut and maize was a fungal concentration of 2 μg mg −1 , with less sensitivity than that reported (1 μg g −1 ) thus far 14 . Although the sensitivity is not as good as that of scFv 14 , but as others have reported 20,21 , the nanobody is more stable than scFv and has so small molecular weight , and the most important advantage of nanobody is that it has strong and fast tissue penetration, which is able to enter the dense tissue such as solid tumors to play a role 27,28 . As we have shown in Fig. 4, it is found that PO8-VHH could bind to the same antigen displayed a single band (45-kDa) in A. flavus and A.  parasiticus, which is of great value for exploring target molecules in Aspergillus pathogens by using the advantage of penetrate the cell membrane. We will use the immunoaffinity column to enrich the single component and determine the target molecule by mass spectrometry in next studies. Of course, we will also be committed to developing a more sensitive method based on PO8-VHH, such as the construction of AP-labeled functional nanobodies, to improve the detection limit of A. flavus in future work. When sandwich ELISA and the HPLC method were used to evaluate samples that were naturally contaminated with aflatoxins, A. flavus antigens were detected in all samples containing aflatoxins (Table 3). Therefore, the sandwich ELISA would have potential to detect aflatoxigenic Aspergillus in naturally contaminated samples. We found that there is still a certain regularity here: for example, samples GX73 and GX54 showed the lowest ELISA signal, and their aflatoxins were lowest; the ELISA signal of samples AF-inoculated, AP-inoculated and GX51 were top three, and their aflatoxins also ranked the top three. However, we also found that ELISA and aflatoxin levels determined by HPLC did not correspond totally in this study because there were instances when high levels of aflatoxin samples showed low ELISA readings in Table 3. This result may be due to the differences in substrate composition and growth conditions (temperature, pH, water activity) since they were collected from different sources that affect aflatoxin levels 29 .
In summary, we have developed a sandwich ELISA based on a nanobody and polyclonal antibody which is selective for effectively detecting aflatoxin producing strains of Aspergillus spp. in agricultural products. This method is likely to be useful for detecting the fungi before they are mature enough to produce high levels of aflatoxin. Thus this is an excellent early detection system. To our knowledge, this is the first example of the development of a nanobody for detection of Aspergillus pathogens.
Unless otherwise stated, all other chemicals and organic solvents were of analytical reagent grade or better. Water was obtained from a MilliQ purification system (Millipore).
Safety. This study involving the care and use of animals was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Laboratory Animal Monitoring Committee of Hubei Province and performed accordingly.
Pure aflatoxin standards were handled in a hood with extreme caution. All items coming in contact with aflatoxins, phage and bacterial cultures (glassware, vials, tubes, ELISA plates, etc.) were immersed in a 10% bleach solution for 1-2 h before they were discarded or autoclaved.
Antigen preparation and immunization. A representative isolate of A. flavus 3.4408, a potential aflatoxin producer was used for antigen preparation. The mycelia were cultured in Czapek medium on a shaker (200 rpm) at 28 °C for 5 days. The mycelia were harvested by filtration through four layers of cheesecloth. Extracellular antigens in the medium were prepared according to Tsai and Cousin 3 . In brief, the cultural fluid was separated by filtration and then through a 0.45-μm membrane. The filtrate was extracted with 80% saturated ammonium sulfate by stirring for 2 h at 4 °C. After the precipitate was suspended in PBS and then dialyzed overnight, the filtrate was concentrated and the concentration was determined. Intracellular antigens (mycelia lysate) prepared according to the method of Yong and Cousin 6 . For the preparation of mycelia lysate, mycelia were suspended in PBS with a final concentration of 3.7% formalin solution. After inactivation overnight at 4 °C, the mycelia were then frozen immediately in liquid nitrogen and grinded into fine powder. The powdered mycelia were weighed and suspended in PBS buffer. The suspended mycelia were homogenized for 3 times in a high pressure homogenizer ATS1500 at 100 bar and further up to 1000 bar for 4 times so as to fully disrupt mycelia. Disruptions of mycelia were confirmed by electron microscope.
A 2-year-old male alpaca was immunized subcutaneously with 1 mg mL −1 extracellular antigens and 100 mg mL −1 mycelia lysate (fungal concentration) mixed with the same volume of Freund's complete Adjuvant. Five additional injections were given at 2-week intervals. After the last booster, 2 × 10 milliliters blood was used for mRNA extraction.
The New Zealand White female rabbit was injected intramuscularly with similar method of alpaca. Boosters were given every other week for a total of six times. Seven days after the final injection, the rabbit was bled to collect whole blood and antiserum. Immunoglobulin G (IgG) was purified from the antiserum using octanoic acid-saturated ammonium sulfate precipitation 30 . Then, the purified polyclonal antibody was freeze-dried and stored at −20 °C.
Preimmune serum was collected as a negative control. The antibody titers were assayed by indirect ELISA.
Phage-displayed Library Construction. According to the titer results, the sixth blood was selected to extract RNA and construct the library. The library was constructed as previously described 24 . In brief, total RNA was extracted from alpaca's blood and transcribed into cDNA. Nanobody gene fragments encoding the VHH variable domains were amplified by PCR using a previously published procedure 31 . Two pairs of primers containing two different SfiI sites (underlined) are as followed, forward primer VHH-F (CAT GCC ATG ACT GTG GCC CAG GCG GCC CAG KTG CAG CTC GTG GAG TC) targeting the framework 1 region, reverse primers VHH-R1 (CAT GCC ATG ACT CGC GGC CGG CCT GGC CGT CTT GTG GTT TTG GTG TCT TGG G) and VHH-R2 (CAT GCC ATG ACT CGC GGC CGG CCT GGC CAT GGG GGT CTT CGC TGT GGT GCG) corresponding to the IgG2 and IgG3 hinge region, respectively. The PCR products and pComb3X phagemid vector were separately digested with SfiI and subsequently ligated to generate pComb3X/nanobody constructs. The resulting recombinant phagemids were transformation into the competent cells of E. coli ER 2738. The cells were then plated on LB ampicillin agar plates to estimate the library size. Twenty clones were selected from the LB-amp plate and sequenced to evaluate library diversity.
Panning. The constructed phage library containing approximately 10 8 phage particles was screened against A. flavus through three rounds of panning. A solid phase protocol was performed to select the nanobodies with high affinity. Briefly, the wells were washed with PBST increased five times in each round (10,15 and 20 times in the first, second and third rounds, respectively), while the amount of antigens were fixed. Under these stringent conditions, the ratios of output and input phages increased steadily (Table 1), with about 10-fold increased phage recovery after the third round of panning compared with the first one, demonstrating an efficient enrichment of specific antibodies. After three rounds of panning, 48 clones were randomly selected from the final output titer plate and infected with M13KO7 helper phage for further characterized by phage enzyme-linked immunosorbent assay (ELISA). Clones binding to antigens but not to any other protein like BSA or OVA were deemed positive and selected for further analyses.
Sequencing analysis of nanobodies. The rescued phages were applied to phage ELISA with 1 mg mL −1 antigens as previously described 32 . Plasmid DNAs from the positive clones were extracted and sequenced using the primer gback (GCCCCCTTATTAGCGTTTGCCATC).
Soluble nanobody ELISA. Soluble nanobodies were expressed in bacteria as previously described 26 .
The antigen-binding characteristics of nanobodies were identified by indirect ELISA. The protocol for indirect ELISA was similar to that described by Hu et al. 32 with slight modification. 96-well microtiter plates were respectively coated with 100 μL of 100 μg mL −1 extracellular antigens and mycelia lysate from A. flavus and A. parasiticus in PBS per well overnight at 4 °C and blocking with 5% skimmed milk in PBS at 37 °C for 2 h. After three times washing with PBST, 200 nM purified nanobodies was added to each well and incubated for 1 h at 37 °C. After three washing cycles, 100 μL of anti-HA mouse monoclonal antibody was added to each well following by 1 h incubation at 37 °C. Then the plate wells were incubated with 100 μL of HRP-labeled goat anti-mouse IgG antibody for 1 h at 37 °C. After washing, 100 μL peroxidase substrate solution system were incubated for 15 min at 37 °C. Enzyme reactivity was stopped by adding 50 μL of 2 M H 2 SO 4 and the absorbance was detected at 450 nm by a microplate reader. Some other protein like BSA or OVA served as the control.
To further evaluate the specificity of the nanobody, ten isolates of Aspergillus spp. with different level aflatoxin and other genera of fungi or bacteria were used for indirect ELISA analysis. The remaining steps were performed the same as soluble nanobody-ELISA.
Immunoblot analysis of nanobody. Immunoblots of extracellular antigens and mycelia from A. flavus and A. parasiticus was detected with nanobodies. In immunoblot assays, extracellular antigens and mycelia of A. flavus and A. parasiticus boiled for 5 min were applied to 12% (w/v) SDS-PAGE and then transferred onto nitrocellulose membrane (Millipore). The membrane was washed with TBST three times for 15 min and then blocked with 5% non-fat milk for 1 h at room temperature, followed by incubation with nanobodies (100 nM) for 2 h. After incubating with the nanobody solution and washed with TBST, the membranes were followed by addition of anti-HA mouse antibody (1:1000 dilution) and HRP-labeled goat anti-mouse IgG antibody (1:2000 dilution) for 1 h at room temperature, respectively.
The membrane was then washed and incubated with ECL (USA) solution in a dark room for 10 min.
Scientific RepoRts | 7: 4348 | DOI:10.1038/s41598-017-04195-6 Detection of A. flavus by sandwich ELISA. A sandwich ELISA procedure requires testing match pair antibodies, capture and detection antibodies that detect different epitopes on the target antigen so that they do not interfere with the other antibody binding. To optimize sandwich ELISA match pair antibodies through the nanobody and polyclonal antibody, the capture antibody and detection antibody were compared for their applicability and sensitivity. Different concentrations of mycelia from A. flavus and A. parasiticus ranging from approximately 10 -2 to 10 4 μg mL −1 were determined by a sandwich ELISA according to the following protocol: 96-well microtiter plates were coated with 100 μL of capture antibody (nanobody) at 0.5 μg mL −1 at 4 °C overnight; three times washing with PBST were performed and each well was blocked with 300 μL of 5% non-fat milk at 37 °C for 2 h; after three times washing with PBST, 100 μL of ten-fold serial diluted mycelia (A. flavus and A. parasiticus) were added at 37 °C for 1 h; washing cycles were repeated and 100 μL of the polyclonal antibody at 1 μg mL −1 was added and incubated at 37 °C for 1 h; after washing, 100 μL HRP-labeled goat anti-rabbit IgG antibody (1:10000 dilution) was incubated at 37 °C for 1 h; after six times washing, 100 μL of TMB solution were incubated at 37 °C for 15 min; the reaction was then terminated by adding 50 μL of 2 M H 2 SO 4 and the absorbance values was detected at 450 nm.
Preparation of contaminated agricultural products samples for sandwich ELISA analysis. To determine the limit of quantification, 0.1 g of grinded peanut or maize suspended in 10 mL PBST was then 10-fold diluted and mixed with different concentrations of mycelia ranging from 10 −2 to 10 4 μg mg −1 , a 100 μL homogenates was added into plate wells for sandwich ELISA analysis.
For analysis of the feasibility of nanobodies combined with polyclonal antibody for detecting Aspergillus-infected sample, sandwich ELISA was used to detect naturally infected peanut and maize samples collected from different areas and 4 samples inoculated with A. flavus and A. parasiticus, respectively (Table 3).