Framework-Directed Amino-Acid Insertions Generated over 55-Fold Affinity-Matured Antibody Fragments That Enabled Sensitive Luminescent Immunoassays of Cortisol

Abstract

We generated three single-chain Fv fragments (scFvs) specific to cortisol according to our original affinity-maturation strategy and verified their utility in developing immunoassays. These scFv mutants (m-scFvs) had insertion of one, four, or six amino acid(s) in the framework region 1 of the V_H-domain and showed >55-fold higher affinity (K_a, 2.0 − 2.2 × 10¹⁰ M⁻¹) than the unmodified scFv (wt-scFv). Each m-scFv was fused with NanoLuc luciferase (NLuc) for the use in enzyme-linked immunosorbent assays (ELISAs). In these ELISA, the m-scFv–NLuc fusions were competitively reacted with immobilized cortisol residues and cortisol standards, and then the bound NLuc activity was monitored luminometrically. The luminescent ELISAs generated dose–response curves with extremely low midpoints (approx. 3 pg/assay) and were >150-fold more sensitive than the colorimetric ELISAs using wt-scFv and >8000-fold more sensitive than the ELISA using the parental native antibody. The luminescent ELISAs showed acceptable cross-reactivity patterns with related steroids, and the determination of control sera afforded cortisol levels in the reference range with satisfactory parallelism.

INTRODUCTION

Immunoassays inherently have excellent sensitivity and specificity, and thus have been essential for monitoring steroid hormones in clinical specimens.^1,2) To obtain even higher sensitivity, various approaches have been tried so far, e.g., employing new signal-generating molecules³⁾ or heterologous combinations of haptenic derivatives used for antibody production and derivatives for labeling of signal-generating molecules.⁴⁾ Although the most critical requirement for achieving high sensitivity is using antibodies showing high affinity against a target steroid,^5,6) a reliable way to produce was hardly available in conventional immunization-based formats. However, “antibody engineering,” which was developed in late 1980s, enabled the generation of “affinity-matured” antibodies with artificial amino-acid sequences via in vitro experiments.^7,8) In a typical strategy, suitable native antibodies (parental antibodies) were converted to single-chain Fv fragments (scFvs) and then randomly mutagenized to generate a diverse set (i.e., library) of scFv mutants (m-scFvs), each member of which displayed on bacteriophage particles. The scFv-displaying phage (scFv-Ph) species showing improved affinities were isolated, and their scFv moieties were converted into “soluble” (non-phage-linked) proteins to be available in immunoassay systems.

This strategy allowed us to produce m-scFvs specific to several haptenic antigens,^6,9–13) including estradiol-17β (E₂)^9,10) and cortisol,¹³⁾ which showed 10–150-fold higher affinity constants (K_a) than the “wild-type” scFvs (wt-scFvs) that are scFvs constructed by combining unmodified V_H and V_L sequences of the parental antibodies. As expected, these mutants enabled 3–100-fold more sensitive enzyme-linked immunosorbent assays (ELISAs).^6,9–13) Moreover, in the studies for anti-cortisol scFvs, we discovered that the insertion of amino acids between the positions 6 and 7 of the V_H domain, located in the framework region (FR) 1 (composed of amino acids at the position 1–30)¹⁴⁾ (Fig. 1A), dramatically enhanced the affinity.^15–17) Among 28 affinity-matured m-scFvs obtained in the study,^16,17) n4#335, having insertion of four amino acids [alanine (Ala)-phenylalanine (Phe)-Ala-aspartic acid (Asp)], showed the highest affinity (K_a 2.0 × 10¹⁰ M⁻¹), which was 56-fold higher than the wt-scFv (Fig. 1). This mutant was isolated from a scFv library lib-4/5 that contained the members each having serial four or five randomized residues.¹⁷⁾ Our original method CAP/DIR, i.e., “clonal array profiling (CAP)¹⁵⁾” combined with “dissociation-independent recovery (DIR),¹⁷⁾” functioned effectively for this selection process. Inspired by this success, we here explored two different libraries (lib-1–3 and lib-6) using CAP/DIR, and discovered new m-scFvs, n1#260 and n6#52, with comparable or even higher affinity (Fig. 1B). These m-scFvs and n4#335 were fused with NanoLuc luciferase (NLuc)^18,19) and used for ELISA. The resulting luminescent ELISAs were extremely more sensitive over the colorimetric ELISAs using the wt-scFv or parental native antibody.

Fig. 1. Summary of the Affinity Maturation of Cortisol-Specific scFvs via Amino Acid Insertion into the V_H-FR1

(A) The prototype antibody Ab-CS#3, prepared using the hybridoma method,²¹⁾ was converted into the scFv form (wt-scFv) by combining the V_H and V_L domains via the linker sequence, Val-Ser-Ser-[glycine (Gly)-Gly-Gly-Gly-Ser]₃-Thr.¹³⁾ We previously generated the library lib-4/5 by inserting random amino acids in the V_H-FR1 (between the positions 6 and 7) of wt-scFv, from which we found m-scFv n4#335.¹⁷⁾ (B) In the present study, the lib-1–3 and lib-6 were newly generated and screened similarly. Consequently, several improved m-scFvs with K_a values exceeding 1 × 10¹⁰ M⁻¹ were isolated. Among them, scFvs n1#260 and n6#52 (K_a ≥ 2.0 × 10¹⁰ M⁻¹) were selected and, with n4#335, further engineered to combine NLuc (Supplementary Figs. S3, S4).

MATERIALS AND METHODS

Generation and Characterization of Affinity-Matured m-scFvs

The anti-cortisol wt-scFv gene¹³⁾ was PCR-amplified using a reverse primer (a mixture of NNS-1, -2, -3, or NNS-6) (Supplementary Table S1) with a forward primer CS#3V_L-For.¹³⁾ NNS-1, -2, -3, and -6 comprised serial 1, 2, 3, or 6 NNS-degenerated codon(s) [(A/C/G/T)(A/C/G/T)(C/G)] that generate any of 20 standard amino acids. The resulting m-scFv gene libraries (lib-1–3 and lib-6) were each subcloned into the pEXmide 7 vector¹³⁾ and transformed into Escherichia coli TG1 cells. The transformant suspensions, containing 5.1 × 10⁶ (lib-1–3) and 1.9 × 10⁶ (lib-6) colony-forming units, were spread on agar plates, and incubated at 37 °C overnight. From each library, 4700 colonies were randomly selected and subjected to the CAP/DIR analyses (Supplementary Fig. S1). Clones individually generated the corresponding scFv-Phs in a single microwell. The resulting scFv-Phs were then ranked for binding activity against cortisol, and promising scFv-Phs were converted into soluble scFv proteins each with a FLAG tag at the C-terminus.^9–13) The K_a values of these scFvs were determined using the Scatchard analysis^15–17,20) (Supplementary Fig. S2).

Preparation of m-scFv–NLuc Fusion Proteins

The gene fragments encoding m-scFvs were amplified with pEX7-V_H-CP5¹⁷⁾ and CS#10V_L-For-2²¹⁾ primers, whereas the NLuc gene fragment in the plasmid containing the anti-thyroxine scFv-NLuc gene¹⁹⁾ was separately amplified (Supplementary Fig. S3). These gene fragments were spliced as usual^9–13) and introduced into E. coli XL1-Blue cells. A cloned transformant was grown, and periplasmic cell extracts containing m-scFv–NLuc protein were prepared^9–13) and used in the following ELISAs.

ELISAs

(a) Colorimetric ELISAs

In the 96-well colorless microplates (Costar#3590; Corning, Corning, NY, U.S.A.) coated with a conjugate of cortisol and bovine serum albumin (CS–BSA),^13,21) cortisol standard (50.0 µL) and the soluble wt-scFv or m-scFv protein (100 µL), both diluted in a phosphate-buffered saline (PBS) containing 1.0% gelatin (G-PBS), were incubated at 4 °C for 120 min. Bound scFvs were probed with a peroxidase-labeled anti-FLAG M2 antibody. Peroxidase activity was determined using o-phenylenediamine as a chromogen (490 nm).^9–13)

(b) Luminescent ELISAs

In the white microplates (Costar#3922; Corning) coated similarly, cortisol standard (50.0 µL) and the m-scFv–NLuc protein (100 µL) were incubated as described above. The microplates were washed, and a Nano-Glo reagent containing furimazine (Promega, Madison, WI, U.S.A.) was added. After 20 s, the luminescence was scanned using a Synergy HTX multimode reader (BioTek Instruments, Winooski, VT, U.S.A.).

RESULTS

Generation and Characterization of Affinity-Matured m-scFvs

We prepared the m-scFv libraries, lib-1–3 and lib-6, where the scFv members had an additional one to three and six fully-randomized amino acids (denoted as a–f), respectively, between the positions 6 and 7 (Fig. 1B). A portion of the bacterial libraries transformed with these genes were explored using CAP/DIR (Supplementary Fig. S1), and m-scFvs n1#260 and n6#52 were isolated from the lib-1–3 and lib-6, respectively. These m-scFvs, having the insertion of 1 [leucine (Leu)] and 6 [glutamine (Gln)-valine (Val)-glutamic acid (Glu)-threonine (Thr)-Thr-serine (Ser)] residues, showed K_a of 2.2, and 2.0 × 10¹⁰ M⁻¹, respectively (Supplementary Fig. S2), which was equal to or higher than that of n4#335 (Fig. 1A). Thus, the amino acid insertions enhanced the affinity by >55-fold over the wt-scFv and >400-fold over the prototype antibody Ab-CS#3 (K_a was 4.7 × 10⁷ M⁻¹; Fig. 1A).²¹⁾ In relation to the current capacity for manual operation of the CAP/DIR, we screened approximately 0.1 and 0.25% of cloned transformants for lib-1–3 and lib-6, respectively. Future automatization of the CAP/DIR might allow us to examine 100% of the libraries and serve novel scFv mutants with even higher affinity.

In a colorimetric ELISA monitoring the FLAG-tag attached at the C-terminus of scFvs (Fig. 2A), n1#260, n4#335, and n6#52 afforded dose–response curves with the midpoint (the cortisol amount that caused 50% inhibition) of 7.51, 11.7, and 8.04 pg/assay, respectively (Fig. 2B). Thus, the present mutagenesis increased the ELISA sensitivity by >40-fold over the ELISA using wt-scFv (the midpoint was 512 pg/assay) and by >2000-fold over the ELISA using the prototype antibody (the dose–response curve has been reported previously¹³⁾; the midpoint was 28 ng/assay).

Fig. 2. Summary of ELISAs Using scFvs

(A) Principles of colorimetric (left) and luminescent (right) ELISAs. (B) Dose–response curves of colorimetric ELISAs using m-scFvs (blue) and luminescent ELISAs using m-scFv–NLuc (magenta) [n1#260 (upper), n4#335 (middle), and n6#52 (lower)] were compared with those obtained by colorimetric ELISA using wt-scFv (green). Dose–response curves were constructed using GraphPad Prism (version 3.0; GraphPad Software). The unit “X g per assay” was used in the abscissa and refers to the total mass of X (g) of cortisol that was added to each microwell. The midpoints of the dose–response curves (pg/assay) are shown on the abscissa. The LODs of luminescent ELISAs using NLuc-fused n1#260, n4#335, and n6#52 were 0.389, 0.593, and 0.263 pg/assay, respectively. Vertical bars indicate the standard deviation (n = 4).

Conversion of the m-scFvs into NLuc-Fusions and Their Performance in ELISA

Genetic fusions of scFvs with signal-generating proteins function as “cloned” labeled-antibodies that ensure reproducibility of immunoassays.^19,21) Previously, we compared several luminometrically detectable enzymes and concluded that NLuc, an engineered enzyme (171 residues) derived from the luciferase of the deep-sea shrimp Oplophorus gracilirostris,¹⁸⁾ is the most suitable for developing sensitive immunoassays operatable under standard batch-by-batch-based formats.¹⁹⁾

Thus, fusion proteins in which NLuc was linked to the C-terminus of the aforementioned m-scFvs were prepared by expressing the m-scFv–NLuc genes in E. coli (Supplementary Fig. S3). Western blotting monitored using a FLAG-tag attached downstream of NLuc showed that the main product had the expected M_r (approx. 48000) (Supplementary Fig. S4). However, we found byproducts losing the scFv moiety, which were not observed in our previous preparations of similar NLuc fusions.¹⁹⁾ Degradation due to contaminating proteases was unlikely. The generation mechanism is currently under investigation.

We optimized the ELISA conditions such that the intensity of the B₀ luminescence (observed without a cortisol standard) should exceed 100-times the intensity of the non-specific luminescence (observed without m-scFv–NLuc). Luminescent ELISAs using m-scFv–NLuc fusions derived from n1#260, n4#335, and n6#52 were 2.3–3.9-fold more sensitive than colorimetric ELISAs using the corresponding m-scFvs (Fig. 2B). The midpoints and limit of detection (LOD)^9–13) were 3.01–3.31 and 0.263–0.593 pg/assay; thus, the luminescent ELISAs were >150-fold more sensitive than the colorimetric ELISAs using wt-scFv and >8000-fold more than the ELISA using the parental antibody. To the best of our knowledge, no previous study has described more sensitive cortisol ELISAs than those in our study.

These ELISAs showed acceptable cross-reactivity patterns with related corticosteroids (Fig. 3). Although >100% reactivity was recorded with 11-deoxycortisol, its normal serum levels are usually much lower (<10 ng/mL)²¹⁾ than those of cortisol (10–250 ng/mL)²¹⁾; thus, overestimation due to this steroid is unlikely. Notably, the reactivity with cortisone was significantly decreased compared to that in the ELISA using wt-scFv (from 45 to <5%). This fact suggested the potential of antibody engineering aimed at “specificity-maturation.” Since the present assay has excellent sensitivity, we could use over 100-fold diluted serum specimens and omit the pretreatment steps. In the determination of control sera, we obtained acceptable values with satisfactory parallelism between dilution rates (Fig. 4). Typical values (ng/mL) using NLuc-fusion of n1#260, n4#335, and n6#52 were 18.6, 20.4, and 19.8 for “level 1” serum, 158, 217, and 208 for “level 2” serum, and 306, 298, and 435 for “level 3” serum, respectively, which are within the range of the proposed “acceptable values” (level 1, 12.9–23.7; level 2, 144–267; and level 3, 285−>500, respectively). Determination of cortisol in biological fluids is still important for diagnosing the hypothalamic–pituitary–adrenal axis-associated diseases and testing the stress levels. The utility of this assay system for clinical diagnosis will be discussed in future studies.

Fig. 3. Cross-Reactivity (%) of Colorimetric ELISA Using wt-scFv and Luminescent ELISAs Using m-scFv–NLucs with Eight Endogenous and Three Synthetic Corticosteroids, as Determined Using the 50% Displacement Method^10–13)

Fig. 4. Serial Dilution Study for Measuring Serum Specimens

Commercially available control sera (Bio-Rad Lyphochek Immunoassay Plus Control, Level 1, 2, and 3) were serially diluted with G-PBS and subjected to luminescent ELISAs using NLuc-fused n1#260 (left), n4#335 (middle), and n6#52 (right). Vertical bars indicate the standard deviation (n = 4).

DISCUSSION

Currently, most antibodies used for clinical diagnosis are produced using the hybridoma methods. However, obtaining anti-steroid antibodies with K_a values in the 10¹⁰ (M⁻¹) range is difficult. To overcome this problem, we investigated antibody engineering over two decades^5,6) and developed a CAP/DIR system that efficiently discovers affinity-matured species from diverse m-scFv libraries.^15,17) In this study, the CAP/DIR system provided us with >55-fold affinity-matured m-scFvs from a library prepared via FR1-directed amino-acid insertion. This novel mutagenesis strategy targeting FR and not complementarity-determining regions (CDRs) that might directly interact with antigens²²⁾ was also developed in our laboratory.¹⁶⁾ The FR1 contributes mainly to the construction of the β-sheet sandwich that supports the CDR loops, and therefore, no researcher has focused this region as a “hot spot” of mutagenesis aimed at the affinity maturation. As demonstrated above, various insertion patterns (length and amino acid species) improved the affinity; however, other patterns deteriorated the affinity (data not shown). We are interested in the critical features required for the insertions to function positively, as well as the mechanism, i.e., how the inserted amino acids enhanced the affinity.

Antibody engineering also enabled the direct linkage of scFvs and luminescent enzymes, including NLuc, to provide “cloned” labeled antibodies.^19,21) Luminescence-based systems have advantages for achieving high assay sensitivity and shortening the overall assay duration by avoiding time-consuming incubation steps that are usually required in colorimetric and fluorometric endpoint readings.

In conclusion, by blushing up a wt-scFv via FR-1-directed mutagenesis and fusion with NLuc, we developed cortisol immunoassays exhibiting a sensitivity >150-fold higher than the assay using wt-scFv, which meant that a >8000-fold improvement in sensitivity was achieved by engineering the prototype native antibody (Fig. 1). In the near future, various high-performance “neo-antibodies” will be generated via the antibody engineering and used commonly in the diagnostic field.

Acknowledgments

This work was supported by JSPS KAKENHI (Grant Numbers: JP19K07021 and JP22K15268). We would like to thank Dr. Eskil Söderlind (Avena Partners AB, Sweden) and Professor Carl A. K. Borrebaeck (Lund University, Sweden) for permitting us to use the pEXmide 7 and 7' vectors.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1) Hampl R, Stárka L. Recent advances in immunoassay of steroid hormones. Trends Analyt. Chem., 8, 72–75 (1989).
• 2) Wudy SA, Schuler G, Sánchez-Guijo A, Hartmann MF. The art of measuring steroids: principles and practice of current hormonal steroid analysis. J. Steroid Biochem. Mol. Biol., 179, 88–103 (2018).
• 3) Immunoassay and other bioanalytical techniques. (Van Emon JM ed.) CRC Press, Boca Raton, FL (2007).
• 4) Hosoda H, Kobayashi N, Ishii N, Nambara T. Bridging phenomena in steroid immunoassays. The effect of bridge length on sensitivity in enzyme immunoassay. Chem. Pharm. Bull., 34, 2105–2111 (1986).
• 5) Kobayashi N, Oyama H. Antibody engineering toward high-sensitivity high-throughput immunosensing of small molecules. Analyst, 136, 642–651 (2011).
• 6) Oyama H, Kiguchi Y, Morita I, Yamamoto C, Higashi Y, Taguchi M, Tagawa T, Enami Y, Takamine Y, Hasegawa H, Takeuchi A, Kobayashi N. Seeking high-priority mutations enabling successful antibody-breeding: systematic analysis of a mutant that gained over 100-fold enhanced affinity. Sci. Rep., 10, 4807 (2020).
• 7) Handbook of therapeutic antibodies. (Dübel S ed.) Wiley-Blackwell, Hoboken, NJ (2010).
• 8) Introduction to Antibody Engineering. (Rüker F, Wozniac-Knopp G eds.) Springer, Switzerland (2021).
• 9) Kobayashi N, Oyama H, Kato Y, Goto J, Söderlind E, Borrebaeck CAK. Two-step in vitro antibody affinity maturation enables estradiol-17β assays with more than 10-fold higher sensitivity. Anal. Chem., 82, 1027–1038 (2010).
• 10) Oyama H, Yamaguchi S, Nakata S, Niwa T, Kobayashi N. “Breeding” diagnostic antibodies for higher assay performance: a 250-fold affinity-matured antibody mutant targeting a small biomarker. Anal. Chem., 85, 4930–4937 (2013).
• 11) Oyama H, Morita I, Kiguchi Y, Banzono E, Ishii K, Kubo S, Watanabe Y, Hirai A, Kaede C, Ohta M, Kobayashi N. One-shot in vitro evolution generated an antibody fragment for testing urinary cotinine with more than 40-fold enhanced affinity. Anal. Chem., 89, 988–995 (2017).
• 12) Morita I, Oyama H, Yasuo M, Matsuda K, Katagi K, Ito A, Tatsuda H, Tanaka H, Morimoto S, Kobayashi N. Antibody fragments for on-site testing of cannabinoids generated via in vitro affinity maturation. Biol. Pharm. Bull., 40, 174–181 (2017).
• 13) Oyama H, Morita I, Kiguchi Y, Morishita T, Fukushima S, Nishimori Y, Niwa T, Kobayashi N. A single-step “breeding” generated a diagnostic anti-cortisol antibody fragment with over 30-fold enhanced affinity. Biol. Pharm. Bull., 40, 2191–2198 (2017).
• 14) Kabat EA, Wu TT, Perry HM, Gottesman KS, Foeller C. Sequences of Proteins of Immunological Interest. U. S. Department of Health and Human Services, National Institutes of Health, U. S. Government Printing Office, Washington, DC (1991).
• 15) Kiguchi Y, Oyama H, Morita I, Morikawa M, Nakano A, Fujihara W, Inoue Y, Sasaki M, Saijo Y, Kanemoto Y, Murayama K, Baba Y, Takeuchi A, Kobayashi N. Clonal array profiling of scFv-displaying phages for high-throughput discovery of affinity-matured antibody mutants. Sci. Rep., 10, 14103 (2020).
• 16) Kiguchi Y, Oyama H, Morita I, Nagata Y, Umezawa N, Kobayashi N. The V_H framework region 1 as a target of efficient mutagenesis for generating a variety of affinity-matured scFv mutants. Sci. Rep., 11, 8201 (2021).
• 17) Kiguchi Y, Morita I, Tsuruno A, Kobayashi N. Retrieving dissociation-resistant antibody mutants: an efficient strategy for developing immunoassays with improved sensitivities. Biol. Pharm. Bull., 45, 1432–1437 (2022).
• 18) Hall MP, Unch J, Binkowski BF, Valley MP, Butler BL, Wood MG, Otto P, Zimmerman K, Vidugiris G, Machleidt T, Robers MB, Benink HA, Eggers CT, Slater MR, Meisenheimer PL, Klaubert DH, Fan F, Encell LP, Wood KV. Engineered luciferase reporter from a deep sea shrimp utilizing a novel imidazopyrazinone substrate. ACS Chem. Biol., 7, 1848–1857 (2012).
• 19) Oyama H, Kiguchi Y, Morita I, Miyashita T, Ichimura A, Miyaoka H, Izumi A, Terasawa S, Osumi N, Tanaka H, Niwa T, Kobayashi N. NanoLuc luciferase as a suitable fusion partner of recombinant antibody fragments for developing sensitive luminescent immunoassays. Anal. Chim. Acta, 1161, 238180 (2021).
• 20) Scatchard G. The attractions of proteins for small molecules and ions. Ann. N. Y. Acad. Sci., 51, 660–672 (1949).
• 21) Oyama H, Morita I, Kiguchi Y, Miyake S, Moriuchi A, Akisada T, Niwa T, Kobayashi N. Gaussia luciferase as a genetic fusion partner with antibody fragments for sensitive immunoassay monitoring of clinical biomarker. Anal. Chem., 87, 12387–12395 (2015).
• 22) Barbas CF III, Bain JD, Hoekstra DM, Lerner RA. Semisynthetic combinatorial antibody libraries: a chemical solution to the diversity problem. Proc. Natl. Acad. Sci. U.S.A., 89, 4457–4461 (1992).