Isolated Variable Domains of an Antibody Can Assemble on Blood Coagulation Factor VIII into a Functional Fv-like Complex

Abstract

Single-chain variable fragments (scFv) are antigen-recognizing variable fragments of antibodies (FV) where both subunits (V_L and V_H) are connected via an artificial linker. One particular scFv, iKM33, directed against blood coagulation factor VIII (FVIII) was shown to inhibit major FVIII functions and is useful in FVIII research. We aimed to investigate the properties of iKM33 enabled with protease-dependent disintegration. Three variants of iKM33 bearing thrombin cleavage sites within the linker were expressed using a baculovirus system and purified by two-step chromatography. All proteins retained strong binding to FVIII by surface plasmon resonance, and upon thrombin cleavage, dissociated into V_L and V_H as shown by size-exclusion chromatography. However, in FVIII activity and low-density lipoprotein receptor-related protein 1 binding assays, the thrombin-cleaved iKM33 variants were still inhibitory. In a pull-down assay using an FVIII-affinity sorbent, the isolated V_H, a mixture of V_L and V_H, and intact iKM33 were carried over via FVIII analyzed by electrophoresis. We concluded that the isolated V_L and V_H assembled into scFv-like heterodimer on FVIII, and the isolated V_H alone also bound FVIII. We discuss the potential use of both protease-cleavable scFvs and isolated Fv subunits retaining high affinity to the antigens in various practical applications such as therapeutics, diagnostics, and research.

Keywords: antibody engineering; coagulation factor VIII; low-density lipoprotein receptor-related protein 1; single-chain variable fragment; thrombin.

Figure 1

Design of iKM33 variants cleavable by thrombin. (A) Structural diagram of the constructs, which include a signal peptide gp67, V_H and V_L domains connected by a linker with thrombin cleavage site, and 10 × His tag for purification. The signal peptide is cleaved off during protein secretion upon expression in Sf9 cells. (B) Amino acid sequences of the linkers in three cleavable scFv variants and parental (uncleavable) iKM33 [34]. The cleavage positions are indicated (^▼).

Figure 2

Binding of FVIII to iKM33 variants by SPR. Each scFv variant was covalently immobilized on a sensor chip. FVIII was injected over the chip at concentrations of 3.125 nM, 12.5 nM, 50 nM, and 200 nM. Resulting K_D values were calculated using the 1:1 binding model.

Figure 3

Cleavage of iKM33 variants by thrombin. The scFv variants were incubated with thrombin for 10 min at 37 °C. The samples of iKM33-v1 (v1), iKM33-v2 (v2), and iKM33-v3 (v3) were treated with 0.05 µg of thrombin per 1 µg of protein substrate. The reactions were stopped by adding PPACK. The parental iKM33 (31 kDa) was found not to be cleavable in preliminary experiments. (A) The samples with (+) and without thrombin cleavage (−) were loaded on SDS-PAGE gels (3 µg protein per well), and the gel was stained by Coomassie blue. The bands of V_H (~13 kDa) and V_L (~17 kDa) were identified based on their theoretical molecular weight. (B) SE-FPLC analysis of the thrombin-cleaved samples. MW—elution profile of molecular weight markers: thyroglobulin bovine 670 kDa (1), γ-globulin bovine 158 kDa (2), ovalbumin chicken 44 kDa (3), myoglobin horse 17 kDa (4), and vitamin B₁₂ 1.35 kDa (5). The insert shows a standard curve based on the markers 2–5.

Figure 4

Thrombin cleavage of FVIII and iKM33-v2. (A) Samples with FVIII (0.5 µg per sample), iKM33-v2 (0.055 µg per sample), and a mix of FVIII (0.5 µg per sample) and iKM33-v2 (0.055 µg per sample) at a molar ratio of 1:1 were incubated for 10 min with increasing amounts of thrombin (0, 8.7, 69, and 555 ng per sample). The reactions were stopped by PPACK, and the samples were analyzed by SDS-PAGE followed by gel silver staining. The identities of the selected bands are indicated as follows: (i) FVIII fragments: A1, A2, A3-C1-C2, and (ii) iKM33-v2 fragments: scFv—single-chain form, V_H and V_L subunits. (B) Quantitative densitometry of selected FVIII and iKM33-v2 fragments (representative results from one of three experiments).

Figure 5

Effects of scFv iKM33 variants on FVIII activity in TG assay. Serially diluted iKM33 variants (60 nM to 0.008 nM) were incubated for 15 min at room temperature with 116 pM FVIII and tested in TG assay. The results are normalized to the control sample (no iKM33). Mean ± SD, n = 3. Statistical analysis: two-way ANOVA, Tukey’s multiple comparisons test for each dilution, no significant difference, p > 0.05.

Figure 6

Effect of thrombin-cleaved iKM33-v2 on FVIII activity. (A) Three samples were prepared as (i) scFv, (ii) scFv treated with thrombin followed by termination of reaction by PPACK (scFv/Thromb/PPACK), and (iii) buffer treated with thrombin followed by addition of PPACK (control) (Thrombin/PPACK). Serially diluted samples (iKM33 concentration from 1000 to 0.24 nM) were incubated with FVIII (145 pM) and tested by CS assay. The results were normalized to FVIII incubated with thrombin-treated buffer. Mean ± SD, n = 4. (B) The samples were prepared similarly to panel A, but thrombin was removed from the reactions using biotinylated PPACK followed by incubation with streptavidin beads. Serially diluted samples (iKM33 concentration from 600 to 0.3 nM) were incubated with FVIII (116 pM) for 15 min and tested by TG assay. The results were normalized to FVIII incubated with Tris-BSA buffer (pH 7.2). Mean ± SD, n = 3.

Figure 7

Effect of thrombin-cleaved iKM33-v2 on the binding of FVIII to LRP1. In SPR assay, LRP1 was immobilized on a chip via amine coupling and tested for binding with FVIII (100 nM) preincubated without or with iKM33-v2 at a ratio of 1, 2, or 5 (A). In a similar set-up, immobilized LRP1 was tested with thrombin-cleaved iKM33-v2 (reaction was terminated by PPACK and confirmed for cleavage by SDS-PAGE) (B). Both experiments were repeated three times with similar results.

Figure 8

Testing interactions between FVIII and isolated iKM33 subunits. The interaction was evaluated by a pull-down assay using FVIII (BDD), iKM33-v1, and its thrombin-cleaved fragments (as shown on top) and FVIII-affinity sorbent; the resulting samples were analyzed by SDS-PAGE followed by gel silver staining. (A) Lanes 1–10 contain the samples, and lanes 11–13 contain control proteins. Specifically, lanes 1–6 contain FVIII (~2 µg), lanes 1–5 and 7–9 contain the scFv and its fragments (used at 3-fold molar excess over FVIII in pull-down assay) as follows: lanes 1 and 7 contain the scFv; lane 2 contains the thrombin-cleaved scFv, lanes 3 and 8 contain the V_H, lanes 4 and 9 contain the V_L, and lanes 5 and 10 contain a mixture of the V_H and V_L. (B) Analysis of dose-dependence of the scFv fragments used in a similar experiment: lanes 1–12 contain FVIII (~2 µg), lanes 1–9 contain the scFv, thrombin-cleaved scFv, and a mixture of V_H and V_L. In each group shown on top, the molar ratio of FVIII to the scFv fragments increases as 1:1, 1:2, and 1:5. Lane 10 contains the V_H and lane 11 contains the V_L; both subunits were used at 5-fold molar excess over FVIII. In both panels, lane M contains a molecular weight marker. Each experiment was performed independently two times with similar results.