An Indirect Homologous Competitive Enzyme-linked Immunosorbent Assay for the Detection of a Class of Glycosylated Dihydrochalcones

Abstract

Hesperetin dihydrochalcone 4′-glucoside, 1 and phloretin 4′-glucoside, 2 belong to a family of dihydrochalcone glycosides that exhibit flavorant properties. We have developed a competitive, indirect homologous ELISA for the detection of targets 1 and 2 in fermentation media. Immunogen and coating antigen were prepared by conjugating hapten, 4-(3-oxo-3-(2,6-dihydroxy-4-glucoside phenyl)propyl) benzoic acid to thyroglobulin and bovine serum albumin, respectively. Antibodies raised in rabbits M6122, M6123 and M6124 and the coating antigen were screened and characterized to determine their optimum concentrations. The optimized ELISA, developed with antibody M6122, gave IC₅₀ values of 27.8 and 21.8 ng/mL for 1 and 2, respectively. Selectivity of the assay was assessed by measuring cross-reactivity of antibody M6122 to related congeners such as aglycones and the 2′-glycosides of hesperetin dihydrochalcone, 5 and phloretin, 6. Antibody M6122 showed very low recognition of 5 and virtually no recognition of the aglycones and 6.

INTRODUCTION

Flavonoids^1,2,3,4,5 are a large and diverse group of polyphenol compounds. They are present in many plants. Flavonoid patterns are normally characteristic of individual plants. They often are ubiquitous in leaves, stems, roots and fruits, and thus form an important part of the human diet. Structurally, they contain a fifteen carbon atom core, with a chroman ring (rings A and C) coupled to a second phenyl ring (ring B) through its C2, C3 or C4 position. Figure 1 depicts the structure of a citrus flavanone, neohesperidin as an example. Some of these compounds exhibit taste properties. One of the most active areas of flavonoid research is on their beneficial health effects, especially their antioxidant activities. The high antioxidant activity of some flavonoids is attributed to the presence of specific structural synthons that drive the inhibition of lipid peroxidation, as evidenced by in vitro analyses.^6,7 Epidemiological studies point to an inverse relationship between dietary flavonoid intake and incidence of coronary heart disease⁸, explained by their inhibitory effects on free radical oxidation of low-density lipids.⁹ They act as natural ultraviolet radiation filters, by scavenging oxygen free radicals generated by UV irradiation.¹⁰ Some phenolics can serve as good topical photo-protective agents. Hydroxyflavones are well-known for their broad-spectrum antimicrobial function¹¹ that aids in not only protecting plant life, but is also useful for the treatment of human diseases. One notable example is 5,6,7-trihydroxyflavone 7-glucuronide (baicalin), that has an inhibitory effect on the human immunodeficiency virus (HIV).¹²

Figure 1.

Chemical structures of neohesperedin, a flavanone glycoside and the hapten 4-(3-oxo-3-(2,6-dihydroxy-4-glucoside phenyl) propyl)benzoic acid.

Dihydrochalcones¹³ are a sub-class of flavonoids. The unique feature that distinguishes the dihydrochalcones from other flavonoids is the open-chain three-carbon structure linking the A- and B-rings in place of a heterocyclic C-ring (Table 1). This close structural correlation accounts for the co-occurrence of dihydrochalcones and flavanones as natural products. Dihydrochalcones are reported to exhibit a wide spectrum of bioactivities.^14,15,16,17 Simple dihydrochalcones such as phloretin, phloridzin and the recently discovered sieboldin¹⁸ possess anti-inflammatory and anti-hypertensive properties, and have important ramifications on cardiovascular disease and diabetes.¹⁹ Some dihydrochalcones and their glycosylated derivatives are found to be sweet. They are usually derived from bitter ingredients found in plants such as apple, citrus or tea leaves. Studies suggest a strong structure-“sweet” taste correlation for the dihydrochalcones.²⁰

Table 1.

Cross-reactivity (CR)^a of Ab M6122 to structurally related compounds.

Compound	CR (%)	Structure
hesperetin dihydrochalcone-4′- glucoside	100
phloretin-4′-glucoside (or trilobatin)	127
hesperetin dihydrochalcone	NIb
phloretin	NIb
hesperetin dihydrochalcone-2′- glucoside	14
phloretin 2′-glucoside	< 0.01

^aCR (%) was calculated as (IC₅₀ of the target analyte/IC₅₀ of the tested compound) × 100.

^bNot inhibited at 50,000 ng/mL. 3 experiments each with 4-well replicates.

Hesperetin dihydrochlcone 4′-glucoside, 1 and phloretin 4′-glucoside, 2, also commonly referred to as trilobatin (Table 1) are dihydrochalcone glycosides that are capable of imparting sweetness to food products, when used at high levels, but to our knowledge are not currently used commercially for this purpose. The 4′-glucoside targets elicit interesting taste properties when used at levels below their detection threshold.²¹ There is an interest in the large-scale synthesis of the target compounds 1 and 2 by microbial fermentation. This synthetic procedure may also yield other structurally similar, but undesirable dihydrochalcones, such as the corresponding dihydrochalcone 2′-glycosides (5, 6) and the aglycones (3, 4). Therefore, there is a need for developing a rapid and accurate method for the exclusive detection and measurement of the target 4′-glucosides in the desired fermentation broth. Currently, the method used for the measurement of compounds 1 and 2 is high-performance liquid chromatography (HPLC). Despite the HPLC method being sensitive and selective, limitations still exist; it is expensive, time-consuming and requires large quantities of solvent. In contrast, the enzyme-linked immunosorbent assay (ELISA) technique is a rapid, sensitive and cost- and time-effective tool, amenable to high-throughput, on-site screening tests for process development.²² The assay will allow a “real-time”, on-site detection and measurement of the targets, which will speed process optimization. The assay will be useful in two ways; one will be to test the fermentation media in order to measure the levels of targets 1 and 2 being produced during pathway optimization and process development. The other use could be to screen large enzyme libraries for specificity of the glycosylation step during enzyme discovery or optimization to narrow the search for an appropriate enzyme to develop a process. To the best of our knowledge, a polyclonal antibody-based ELISA for detection of dihydrochalcone glycosides such as 1 and 2 has not been previously reported, except for a single report on the production of monoclonal antibodies for quercetin flavonoid glycoside and its corresponding glucuronide.²³ Herein, we report the design and development of an indirect, homologous competitive ELISA for measurement of the desired target compounds in fermentation media.

MATERIALS AND METHODS

Chemicals and Instruments

The hapten coupling reagents, bovine serum albumin (BSA), thyroglobulin (Thy), Tween 20, 3,3′,5,5′-tetramethylbenzidine (TMB) were purchased from Sigma-Aldrich. Co. (St. Louis, MO). Goat anti-rabbit IgG peroxidase conjugate (GAR-HRP) was purchased from Abcam (Cambridge, MA). ELISA was performed on 96-well polystyrene microtiter plates (Nunc MaxiSorp, Roskilde, Denmark) and read spectrophotometrically with a microplate reader (Molecular Devices, Sunnyvale, CA) in dual wavelength mode (450-650 nm).

Buffers

All buffers were prepared with ultrapure deionized water. Phosphate buffered saline (1× PBS, pH 7.5), wash buffer (PBST): 1× PBS containing 0.05% Tween 20. Coating buffer (pH 9.6): 15 mM Na₂CO₃, 34.88 mM NaHCO₃, 3.08 mM NaN₃. Blocking agent: 1% BSA-PBS. Substrate buffer (pH 5.5): 0.1 M sodium acetate/citrate buffer. For the substrate solution, 0.4 mL of 0.6% TMB (in DMSO w/v) and 0.1 mL of 1% H₂O₂ were added to 25 mL sodium acetate-citrate buffer. Stop solution was 2N H₂SO₄.

Hapten Synthesis

Figure 1 depicts the structure of hapten 4-(3-oxo-3-(2,6-dihydroxy-4-glucoside phenyl) propyl)benzoic acid. The hapten was prepared by an enzyme-catalyzed (naringinase) reaction of 4-(3-oxo-3-(2,6-dihydroxy-4-neohesperidoside phenyl)propyl)benzoic acid. A jacketed flask equipped with a magnetic stirbar was equilibrated to 65 °C using a temperature controlled water circulator. To the flask was added a buffer solution (50 mL, 0.1 M KH₂PO₄, 0.2 M Na₂HPO₄, pH = 6.6) and naringinase (1.5 g). The reaction was stirred for 2 h at 65 °C. 4-(3-oxo-3-(2,6-dihydroxy-4-neohesperidoside phenyl)propyl)benzoic acid (1.2 g, 0.002 mol) was dissolved in 10 mL of distilled water and added to the stirred reaction. The reaction was cooled to 45 °C with stirring for 24 h or until complete by TLC analysis (BuOH/PrOH/H₂O, 10/5/4 (v/v)). The reaction was heated to 90 °C for 15 min to deactivate the enzyme. The enzyme was removed from the reaction solution by vacuum filtration. Purification was performed with HP-20 resin (0-60% MeOH in water) followed by HPLC (5-50% MeOH with 0.01% formic acid in water). Fractions were identified by NMR analysis and lyophilized to provide a light yellow powder (0.092 g, 9.9% yield). NMR analysis provided an estimated 90% pure product which was deemed acceptable for use in this assay.

¹H NMR (300 MHz, CD₃OD): δ 7.86-7.84 (2H, d, J = 8.0 Hz), 7.26-7.23 (2H, d, J = 8.0 Hz), 5.96 (2H, s), 4.94-4.91 (1H, d, J = 7.0 Hz), 3.91 (1H, d, J = 11.0 Hz), 3.76-3.73 (1H, m), 3.44-3.40 (4H, m), 3.32-3.30 (2H, q, J = 7.0 Hz, J = 1.5 Hz), 3.00-2.96 (2H, t, J = 15.0 Hz, J = 7.5 Hz); ¹³C NMR (75 MHz, CD₃OD): δ 206.22, 165.71, 164.98, 148.39, 148.23, 130.81, 129.42, 106.82, 105.24, 101.12, 96.44, 78.23, 77.86, 74.60, 71.14, 64.36, 62.34, 46.49, 31.76.

Preparation of Immunogen and Coating Antigen

The carbodiimide method²⁴ was employed for activation of the carboxylic acid group in the hapten. The hapten was coupled to porcine thyroglobulin (as the immunizing antigen) and bovine serum albumin (as the coating antigen) using the N-hydroxysulfosuccinimide ester method. Hapten was dissolved (3 mg, 0.0065 mmol) in 0.9 mL of 0.1 M PIPES (piperazine-N,N’-bis(2-ethanesulfonic acid)) buffer (pH 6.1). EDC, 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride, (1.4 mg, 0.0072 mmole) and N-hydroxy sulfosuccinimide sodium salt (1.6 mg, 0.0072 mmole), were added for preparation of the activated ester intermediate. After one hour, 2-mercaptoethanol (0.5 μL, 0.0072 mmole) was added with stirring for one minute. This latter step scavenges the unreacted EDC, which, if present, could cross-link with the protein resulting in unwanted conjugates in solution. Thyroglobulin (5 mg) was dissolved in 1 mL of a 0.25 M borate buffer (pH 8.8). The activated ester solution (0.6 mL) was added to the thyroglobulin (Thy) solution with rapid stirring. A similar procedure was followed for preparation of the coating antigen by conjugating hapten to BSA (dissolved 2.5 mg of BSA in 0.5 mL of borate buffer (pH 8.8)). Then, the activated ester solution (0.3 mL) was added to the BSA solution. The solutions were stirred at room temperature for 1.5 h, after which they were transferred to 4 °C for overnight incubation. Unreacted small molecules were removed by dialysis. The reaction mixtures containing the immunogen and coating antigen were dialyzed (10,000 MW cut-off) under stirring against PBS (0.01 M, pH 7.5), for 3 d, with frequent changing of the PBS solution to remove the unconjugated free hapten. Polyacrylamide gel electrophoresis and subsequent visualization with a SYPRO Ruby gel stain for glycoproteins showed that the haptens had coupled to the proteins. The resulting conjugates were stored at −20 °C for future use.

Immunization and Antiserum Preparation

The immunization procedure followed a previously published protocol.^24,25 Three female New Zealand white rabbits M6122, M6123, M6124, were each immunized with 500 μg of immunogen hapten-Thy (concentration of 5 mg/mL) in the initial immunization (pre-immune bleeds were collected from each rabbit prior to immunization). Following this, each rabbit received 300 μg of immunogen in the four subsequent boosts. A single test bleed was collected a month after the first immunization and two boosts separated by a one-week interval. A week after the test bleed collection, the third and fourth boosts were given, separated by a two-week interval. Production bleeds (20 mL from each rabbit) were collected in 3 months following the first immunization and all of the four boosts. Antiserum (from the production bleed) was obtained by centrifugation, stored at −20 °C, and used without further purification.

Development of Indirect Competitive ELISA

Although there are several different configurations for competitive ELISAs, we chose to develop the ELISA using the antigen immobilization format, following a previously reported procedure. In this assay format, the coating antigen was directly immobilized on the ELISA plate. A fixed amount of unlabeled primary antibody and labeled secondary antibody (tracer) were added. This implies an indirect detection method, which is contrary to direct detection wherein an enzyme conjugated primary antibody is employed. Use of non-purified and a single antigen-specific primary antibody and commercial availability of a wide variety of labeled secondary antibodies are major advantages of the indirect ELISA approach. In the indirect method, the labeled secondary antibody binds to different epitopes (antigenic determinants) of the primary antibody, providing greater signal amplification than that can be achieved with the direct method. The indirect ELISA is competitive when it is based on competition between target analyte in the sample and the immobilized or the coating antigen-hapten for a fixed amount of the primary antibody. As the concentration of the target analyte increases, the amount of coated antigen captured by the primary antibody decreases.

Details of the ELISA procedure followed for the reported assay are given in the subsequent sections.

Standard Curve Preparation

To evaluate the sensitivity and range of linearity of the ELISA, standard curves of both target compounds 1 and 2 were prepared in a 10% DMSO (dimethyl sulfoxide)-PBS dilution buffer (final dilution in the plate wells being 5% DMSO-PBS). In addition to two negative controls (no antibody blank wells in PBS and no coating antigen wells in 5% DMSO-PBS), a positive control in 5% DMSO-PBS with zero standard analyte were chosen. Nine standard concentrations were prepared from a five-fold serial dilution of a 50,000 ng/mL preparation. The limit of detection, set as IC₁₀, the assay sensitivity, represented by the IC₅₀ value and the linear range of detection, set as IC₂₀ – IC₈₀, were obtained from the calibration curve data. The data were fitted to a four-parameter logistic equation, using Igor Pro 6.22A, Wavemetrics, Inc.

Cross-Reactivity Determination

ELISA plates were coated with coating antigen at 0.1 μg/mL in coating buffer (100 μL/well) by overnight incubation at 4 °C. The plates were washed five times with 1× PBS containing 0.05% Tween 20. Plates were blocked for 0.5 h at room temperature with 1% BSA-PBS (350 μL/well) and washed five times. For the competitive inhibition, 50 μL/well of competing analytes (targets 1 and 2 and cross-reactants: hesperetin dihydrochalcone, 3; hesperetin dihydrochalcone 2′-glucoside, 5; phloretin, 4 and phloretin 2′-glucoside, 6, glucose and p-coumaric acid) were added starting at 50,000 ng/mL, followed by a five-fold serial dilution, ending at 0.128 ng/mL. Each ELISA plate had two competitors: target 1 and one cross-reactant, in two sets, with quadruplicate wells for each concentration. The competitor was co-incubated with 50 μL/well of primary antibody (at a dilution of 1 in 5000) for 1 h. After washing, secondary antibody, GAR-HRP, was added at a dilution of 1 in 25000, followed by an hour of incubation. This was followed by a 5× wash and addition of 100 μL of a solution containing TMB and 1% H₂O₂ in citrate buffer in each well, yielding a blue color when oxidized. The plate was incubated for 15 minutes before the reaction was stopped by the addition of 50 μL of 1N H₂SO₄ in each well. The absorbances were recorded at 450 nm, with background subtraction at 650 nm. The cross reactivity was calculated as [(IC₅₀ of target)/(IC₅₀ of cross-reactant)] × 100.

Matrix Effect Determination

Two commonly used culture media, LB Broth and SOC, were used in these studies. Bacteria (E. coli K12 strain) were grown in each broth to an OD₆₀₀ ≥ 4.0. The media was centrifuged and the supernatant was used for the matrix effect study. To determine interference effects from the matrix, neat fermentation media supernatant samples were spiked with the standard solution in 10% DMSO-PBS. Competitive curves with final concentrations of the target compounds 1 and 2 (five-fold serial dilution starting from 50,000 ng/mL to 0.128 ng/mL) were run in 5% DMSO-PBS (final dilution in plate wells) and in various dilutions of fermentation media (1:5, 1:20 and 1:100). IC₅₀ values obtained from each diluted curve were compared with those generated from the assay buffer (5% DMSO-PBS).

RESULTS AND DISCUSSION

Rationale for Design of Hapten

Hapten 4-(3-oxo-3-(2,6-dihydroxy-4-glucoside phenyl) propyl)benzoic acid (Figure 1) was chosen for the competitive assay development. The hapten with a –COOH functional group conjugated to Thy, was designed for immunizing to raise antibodies. Immunizing haptens typically are a close structural mimic of the target analyte having a functional group that can be used to couple the hapten to a protein. The carboxylic acid substituted hapten was chosen because it would likely generate antibodies that will detect both compounds 1 and 2, but not 3, 4, 5 and 6. The hapten was coupled to the protein at a point most distal to the sugar (providing discrimination between 2′- and 4′-glucosides) and close to the B phenyl ring farthest from the glycoside moiety, thus minimizing discrimination between compounds 1 and 2 and at the same time, precluding antibody detection of the aglycones 3 and 4. In a competitive ELISA, the target analyte competes with the immobilized, coating antigen-hapten for a fixed amount of antibody. The carboxylic acid group substituted hapten was used as both the immunogen and coating 3 antigen hapten in a homologous competitive ELISA format. In the following discussion, results are reported for the homologous ELISA developed with this hapten for target 1, with target 2 tested as a cross-reactant in the assay.

Antibody Characterization

Antisera collected from three rabbits M6122, M6123 and M6124 after the first immunization and two boosts separated by a one-week interval were subjected to titration by the homologous competitive ELISA. All three antisera showed significantly high titers (data not shown) indicative of the rabbits’ response to the immunogen. Checkerboard titrations (CBTs) were performed to determine the optimum concentrations of the antibody raised against hapten-Thy and coating antigen (hapten-BSA) with a 2-fold serial dilution of the antibody, starting from a 1000× dilution up to a 128,000× dilution, screened against a five-fold serial dilution of coating antigen (2 × 10⁴ ng/mL to 1.3 ng/mL). The determined optimum concentrations of antibody and coating antigen in the assay had minimum variations among the three rabbits tested. The IC₅₀ values obtained with antisera from rabbits M6123 and M6124, for target 1 as the inhibitor, were 243.8 and 114.5 ng/mL, respectively (Figure S5 in Supplementary Information). These values were at least an order of magnitude higher than that obtained with Ab M6122, demonstrating much lower assay sensitivities. Following this, the antisera obtained from rabbit M6122 was used for further immunoassay development. A CBT was performed with concentrations (250 ng/mL, 25 ng/mL and 0 ng/mL) of the target analyte 1 with four different antisera dilutions 5000×, 10,000×, 15,000× and 20,000× screened against a five-fold serial dilution of coating antigen concentrations (from 4000 ng/mL to 32 ng/mL). The combination of Ab M6122 and coating antigen, hapten-BSA that had over 80% inhibition at 250 ng/mL analyte concentration were screened using nine concentrations of target 1 ranging from 0.128 ng/mL to 5×10⁴ ng/mL. This dilution series was also used to assess the linear range of the assay. Wells with zero coating antigen and with zero primary antibody served as the negative controls. Wells with zero analyte concentration served as the positive controls. For targets 1 and 2 as inhibitors in the assay, IC₅₀ values of 27.8 and 21.8 ng/mL, respectively, were obtained (Figure 2).

Figure 2.

ELISA inhibition curves for target compounds A. hesperetin dihydrochalcone 4′-glucoside, 1 and B. phloretin 4′-glucoside, 2. Each data point represents the mean of quadruplicates.

Assay Optimization

The optimum concentrations of coating antigen hapten-BSA and Ab M6122 were 100 ng/mL and 1:10,000 dilution in the well, respectively. Since both targets 1 and 2 were only partially soluble in water, a series of co-solvents were screened and DMSO was selected to improve solubility. PBS buffer containing DMSO at different concentrations was tested to assess changes in assay sensitivity. A reduction in the optical densities and a decrease in assay sensitivity were observed with increasing DMSO concentrations (Figure S6 in Supplementary Information). Thus, a 10% DMSO-PBS in 0.15 M PBS, pH 7.4 buffer was selected to prepare the standard solutions and spiked fermentation media samples for assay optimization. The linear range of detection was 2 –300 ng/mL. The fermentation media supernatant samples were diluted as necessary to stay within the linear range of detection. Intra- and inter-plate variations (assays run on the same day as well as on different days) were determined by calculating the percent coefficient of variation (% CV) in the assay for both 1 and 2 The % CV was < 10%, implying minimum variations between wells and limited day-to-day variability.

Cross-reactivity (CR)

In order to determine antibody selectivity, compounds 3-6 (Table 1) with structural similarity to targets 1 and 2 and which are probable metabolites or precursors to the target compounds in the microbial system of production were investigated for relative cross-reactivity. With hapten-BSA as the competing coating antigen, the Ab M6122 showed >100% binding to target 2, phloretin 4′-glucoside (CR 127%, IC₅₀ 21.8 ng/mL). As expected, the antibody exhibited virtually no recognition of the aglycones hesperetin dihydrochalcone, 3 and phloretin, 4 because they do not inhibit even at 50,000 ng/mL. With the 2′-glucosides, the antibody exhibited a very low recognition of hesperetin dihydrochalcone 2′-glucoside, 5 (CR 14%, IC₅₀ 202 ng/mL, an order of magnitude greater than that observed for target 1 with an IC₅₀ of 27.8 ng/mL, indicating > 10× lower sensitivity). There was limited reactivity with phloridzin, 6 (CR <0.01%). Antisera obtained from rabbits M6123 and M6124 were also tested for cross-reactivity to hesperetin dihydrochalcone 2′-glucoside, 5 (see supplementary information Figure S2). Both antibodies displayed a very small level of cross-reactivity to 5, CR 15% for Ab M6123 and CR 21% for Ab M6124. Also, no cross-reactivity was observed with glucose and p-coumaric acid, other probable by-products in the microbial fermentation.

Matrix Effects

Variations in analyte reactivity in the presence of the matrix were tested by measuring the standard curves in fermentation media spiked with the target compounds 1 and 2. The standard curves in spiked fermentation media supernatant obtained at different dilutions of the media were compared with that obtained in the assay buffer (5% DMSO-PBS final dilution in wells). A difference in the values of percentages of control for the standard curve and corresponding dilution of the media (LB broth supernatant), observed at the IC₅₀ value of the standard curve in 5% DMSO-PBS was taken as a measure to estimate the matrix effect. For target 1, at 1 in 5 and 1 in 20 dilutions of the media, a difference in the percents of control of 4.66 ± 1.68 and 5.74 ± 1.68, respectively, between the standard curves for assay buffer and spiked media were obtained. At a 1 in 100 dilution of the media, the standard curve of 1 (Figure 3) was parallel to the one in 5% DMSO-PBS buffer, with a much lower difference of 0.26 ± 1.68. Correspondingly, for target 2 (Figure 4), 6.47 ± 1.66 at a 1 in 5 dilution and a difference of 0.5 ± 1.66 at a 1 in 20 dilution were observed, whereas at a 1 in 100 dilution, the difference was 1.5 ± 1.66. These small differences essentially signify absence of any matrix interference. Also, no matrix effect was observed with the SOC media supernatant at the tested dilutions (1 in 5, 1 in 25 and 1 in 125). For the recovery study, hesperetin dihydrochalcone-4′-glucoside, 1 dissolved in DMSO was spiked at 0.128-5×10⁴ ng/mL in 10 μL of clear fermentation media supernatant, and analyzed in triplicate. Table 2 shows the mean percent recovery data for target 1, given for four different analyte concentrations. A recovery of > 80% points to minimal interference from the fermentation media.

Figure 3.

Matrix effects evaluation for fermentation media supernatant (LB broth) spiked with target 1, hesperetin dihydrochalcone 4′-glucoside. Each data point represents the mean of quadruplicates. The standard curve in 5% DMSO-PBS was generated by taking the average of values from 8 ELISA plates.

Figure 4.

Matrix effects evaluation for fermentation media supernatant (LB broth) spiked with target phloretin 4′-glucoside 2. Each data point represents the mean of quadruplicates. The standard curve in 5% DMSO-PBS was generated by taking the average of values from 3 ELISA plates.

Table 2.

Recovery of hesperetin dihydrochalcone-4′-glucoside, 1 from spiked fermentation media.

Spiked Concentrationa (ng/mL)	Detected (ng/mL)	Mean recovery (%) (SD)
2000	2747.3	137.4 (39.5)
400	384.9	96.2 (13.7)
80	67.5	84.3 (17.1)
16	16.2	100.8 (27.1)

^aDifferent amounts of hesperetin dihydrochalcone-4′-glucoside, 1 (in DMSO) were added to the fermentation media supernatant. The spiked media samples were diluted 100-fold with PBS (5% DMSO) (n = 3).

In summary, the new competitive indirect, polyclonal antibody-based ELISA presented here was designed to serve as a class-selective immunoassay for the preferential detection of the target 4′-glucosides - hesperetin dihydrochalcone-4′-glucoside, 1 and phloretin-4′-glucoside, 2. The homologous assay has IC₅₀ values of 27.8 and 21.8 ng/mL for target compounds 1 and 2, respectively. There are many criteria for successful production of a target compound by fermentation or other approaches, one of which is reasonable yield. For this work, we selected a concentration of ~ 1 mg/mL of the target compounds. This is considered as a reasonable minimum yield for the high throughput screening process using whole cells or purified enzyme. Thus, even with the necessary dilutions of the fermentation media to avoid matrix effects, the sensitivity of the assay is more than adequate for use in monitoring during optimization of fermentation conditions. The immunoassay can be utilized on-site, further reducing the time and cost of the optimization process. It can also be used for continued monitoring during production.

ABBREVIATIONS USED

Ab

antibody

BSA

bovine serum albumin

CBT

checkerboard titration

CR

cross-reactivity

EDC

1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride

ELISA

enzyme-linked immunosorbent assay

GAR-HRP

goat-anti-rabbit IgG peroxidase conjugate

HRP

horseradish peroxidase

PBS

phosphate buffered saline

PBST

phosphate buffered saline-Tween-20

TMB

3,3′,5,5′-tetramethyl benzidine

Thy

thyroglobulin