Immunoaffinity chromatography occupies a special place in modern bioprocessing. By exploiting the exquisite specificity of antibodies for their antigens, it allows a desired protein to be pulled out of a complex mixture based on epitopes displayed on its surface, a capability that applies to nearly every antigen imaginable. Yet the technique is expensive and technically demanding at industrial scale, and its success hinges on the properties of the immobilized ligand. Conventional immunoglobulins are large, fragile molecules that can denature under the harsh acidic or alkaline conditions used to strip bound product from a column, shortening resin lifespan and raising costs. The variable domain of camelid heavy-chain antibodies, known as VHH or nanobody, has emerged as a compelling alternative. These roughly 15-kilodalton fragments are small, remarkably stable, and resistant to conditions that destroy ordinary antibodies, and they have already proven their worth commercially in resins such as the CaptureSelect family used across the biopharmaceutical industry.

Recombinant factor VII presents a particularly demanding purification challenge. The protein is expressed in baby hamster kidney cells and is both thermally sensitive and structurally fragile, requiring a rapid series of downstream steps to preserve its integrity. Because the drug is administered in its active form, activation at the end of the bioprocess is essential. At least one affinity chromatography step is indispensable, and the newest commercial solution, VIISelect from GE Healthcare, relies on a camelid VHH ligand discovered from an immune llama antibody library. Immune libraries, however, require animal immunization, a drawback that limits their use against toxic, poorly immunogenic, or otherwise unsuitable antigens. Synthetic libraries sidestep this limitation entirely, allowing binders to be discovered against an essentially unrestricted range of targets through rational design of molecular diversity.

The research team, led by Aliasghar Rahimian, Ali Nabati, Hooman Askari, and Mahdi Aminian, drew on two anti-factor VII nanobodies, designated VHH19 and VHH24, which they had previously isolated from a phage-displayed synthetic VHH library built on the cAbBCII-10 scaffold. This framework is prized for its exceptional stability, its plasticity for grafting diverse antigen-binding loops, and its acceptable expression yield in bacteria. The two binders carried affinity constants of 10 to the power of minus eight and 5.8 times 10 to the power of minus eight, respectively, indicating tight binding to the coagulation factor. The team expressed the nanobodies in the periplasm of TG1 bacterial cells on a five-liter scale, recovered them through osmotic shock, and purified them sequentially on nickel-affinity resin and Sephacryl S-100 gel filtration to remove aggregates and polymers. Yields reached approximately 0.8 milligrams per liter for VHH19 and 1 milligram per liter for VHH24, with reducing SDS-PAGE revealing the expected doublet of native and disulfide-reduced forms around 15 kilodaltons.

Before committing the binders to a column format, the researchers verified their specificity by enzyme-linked immunosorbent assay. Both purified VHHs produced strong, concentration-dependent signals on wells coated with recombinant factor VII, while signals on wells coated with diluted human blood plasma, used as a source of non-specific antigens, remained at baseline across all coating concentrations. This clean discrimination against the thousands of proteins present in plasma is a critical prerequisite for any ligand intended to purify a therapeutic protein from biological feedstocks. Encouraged by these results, the team coupled each nanobody to cyanogen bromide-activated cross-linked agarose beads, a classical chemistry that converts agarose hydroxyl groups into cyanate esters reactive with protein amino groups. By measuring protein remaining in the coupling supernatant, they determined that coupling efficiency exceeded 95 percent for both binders, yielding resins containing approximately 2 milligrams of immobilized VHH per milliliter of gel.

The decisive test came in column chromatography, where the chemistry of calcium binding to the Gla domain of factor VII shaped the optimization strategy. Calcium association is essential to the conformation and enzymatic activity of the protein, so the team screened four buffering systems that varied the presence of calcium chloride and EDTA during binding and elution. The outcome was unambiguous: only the VHH19-based resin, designated Seph4B-VHH19, captured factor VII effectively, and only when calcium chloride was present in the binding and washing buffers. Under those conditions the factor VII concentration in the flow-through was depleted almost completely, whereas procedures lacking calcium produced partial or negligible adsorption. EDTA proved unable to elute the bound protein, leaving acidic glycine-hydrochloric acid buffer at pH 2.2 as the effective desorption agent. The VHH24 column and an uncoupled control resin retained the protein poorly or not at all under every condition tested, and both were excluded from further study.

With the optimal conditions established, the researchers benchmarked their resin against the commercial VIISelect medium using breakthrough curve analysis on a partially purified, inactive factor VII feed. Seph4B-VHH19 processed 28 plus or minus 1 milliliter of feed before breakthrough, yielding 2.3 plus or minus 0.15 milligrams of eluted protein per milliliter of resin, compared with 42 plus or minus 1 milliliter and 4.08 plus or minus 0.06 milligrams per milliliter for VIISelect. The authors attribute roughly half of this capacity gap to ligand density: the commercial resin reportedly carries about 6 milligrams of ligand per milliliter, double the loading achieved with their research-scale periplasmic expression system. They argue the comparison should be read as a feasibility demonstration rather than a head-to-head performance contest, noting that industrial expression improvements and more sophisticated coupling chemistries could readily narrow the difference.

Two further results stand out for bioprocess engineers. First, the resin proved operationally robust: it performed consistently across four consecutive chromatography cycles, and although cleaning with 0.1 molar sodium hydroxide before a fifth cycle cost 15 percent of the dynamic binding capacity, the column retained that reduced capacity after a full month of storage at 4 degrees Celsius. The limited alkaline tolerance marks a clear target for future sequence optimization and cleaning-procedure refinement. Second, and perhaps more striking, the column accomplished on-column activation of the drug substance. When a mixture of partially purified inactive factor VII and human plasma was passed through Seph4B-VHH19, the resin excluded a substantial portion of the plasma contaminants, and the eluted factor VII appeared as two chains of 20 to 30 kilodaltons on reducing SDS-PAGE, the signature of autocleavage. Activity assays confirmed the transformation: eluate fractions reached 26.8 kilo-international units per milliliter, roughly two hundred times the 0.12 KIU per milliliter measured in the pre-column sample, while flow-through and wash fractions carried almost none of the activity.

This activation step matters because commercial factor VII production is not complete without it; in existing downstream processes, activation occurs as a byproduct of concentration during ion-exchange chromatography. The nanobody column reproduced that behavior naturally, concentrating the protein enough to trigger autocleavage as it bound. The authors place their work in a broader context of nanobody-based affinity media, which have already been engineered for demanding targets such as adeno-associated virus serotype 8 and for separating closely related product forms of low-abundance plasma proteins. Nanobodies resist aggregation because they lack the variable light-chain domain, their frameworks can be fully humanized, and the cAbBCII-10 scaffold offers exceptional plasticity for CDR grafting. The synthetic library approach eliminates animal immunization and permits rational targeting of difficult antigens, though the absence of in vivo affinity maturation means isolated clones may need additional in vitro optimization to match the affinity profiles of immune-derived counterparts.

The study also exposes a knowledge gap that the authors hope the field will fill. Few publications assess how different conjugation methods and linkers affect the performance of VHH affinity columns, and the cyanogen bromide chemistry used here, while simple and traditional, immobilizes ligands in random orientations. Site-specific immobilization strategies, including engineered tag systems, have been shown to improve both capacity and resin stability compared with random amine coupling, and the team has published the sequence of VHH19 to encourage such exploration. For an industry that already generates billions of dollars in value from VHH-based CaptureSelect resins, the demonstration that a fully synthetic, animal-free discovery pipeline can produce a functional purification ligand for a fragile, medically critical coagulation protein signals a practical new route to affinity media, one that could extend to toxic, poorly immunogenic, or otherwise inaccessible targets where immunized llamas simply cannot help.

Subject of Research: Synthetic camelid nanobody binders for immunoaffinity purification of recombinant coagulation factor VII

Article Title: Application of camelid nanobody binders derived from a novel full-synthetic VHH library in immunoaffinity purification of recombinant coagulation factor VII

Article References: Rahimian, A., Nabati, A., Askari, H., & Aminian, M. (2026). Application of camelid nanobody binders derived from a novel full-synthetic VHH library in immunoaffinity purification of recombinant coagulation factor VII. Heliyon, 12(15), Article e45524. https://doi.org/10.1016/j.heliyon.2026.e45524

Image Credits: AI Generated

DOI: Not provided

Keywords: nanobodies, VHH, synthetic antibody library, coagulation factor VII, immunoaffinity chromatography, bioprocessing, hemophilia, affinity resin, phage display, recombinant protein purification, cyanogen bromide coupling, downstream processing

Nanobodies, the single-domain antibody fragments originally derived from camelids, have become indispensable instruments across modern biology. Because they are compact, remarkably stable, and easy to engineer, they serve as conformational stabilizers, crystallization chaperones, biosensors, intracellular targeting agents, and imaging probes. Clinically, the class has already produced approved drugs such as caplacizumab, and nanobody-based therapeutics are advancing against cancer, autoimmune disorders, and infectious disease. In structural biology, their most celebrated role has been locking fragile membrane proteins into defined functional states, a strategy that made high-resolution structures of G-protein-coupled receptors possible more than a decade ago.

Traditionally, these binders are obtained by immunizing llamas or other animals with the protein of interest, an approach that fails when antigens are toxic, unstable, or require trapping in artificial conditions such as extreme pH or fully liganded states. Synthetic libraries sidestep these constraints entirely. Instead of an immune system, researchers use combinatorial gene libraries encoding billions of randomized antibody fragments, then select binders in vitro. Among the display technologies available, ribosome display is uniquely powerful because it couples each protein to its encoding messenger RNA without any cells, allowing theoretical diversities of 10^12 to 10^13 to be explored in a test tube, far beyond what phage or yeast display can practically achieve.

The existing benchmark, the Seeger concave sybody library, has enabled countless selections, including binders that trap membrane proteins in specific conformations. However, it carries a hidden incompatibility. Its nanobody constructs lack the C-terminal histidine tag that the Legobody system requires to recognize its target. Legobody, an approximately 120-kilodalton scaffold assembled from a sybody, an engineered maltose-binding protein, and a nanobody-binding antibody fragment, dramatically enlarges small protein complexes so that cryo-EM image alignment becomes feasible. Without the tag, every sybody recovered from the established libraries must be subcloned and re-engineered before it can be used, adding days of work and eroding throughput precisely where speed matters most.

The new S1.0 library eliminates that detour by embedding the required C-terminal histidine tag and a short Legobody-compatible linker directly into the library scaffold. But the redesign went beyond a simple tagging exercise. Drawing on structural analyses of published concave sybody-antigen complexes, the team noticed that these synthetic binders engage their targets mainly through the CDR1 and CDR2 loops, unlike natural nanobodies, which typically lean heavily on CDR3. In the concave library, the short seven-residue CDR3 and the largely fixed CDR1 had constrained where binding could occur. The researchers therefore extended CDR3 by one residue, raised the number of randomized positions in CDR3 and CDR1 to seven each, and deliberately reduced variability at one CDR2 position, restricting it to serine, tyrosine, arginine, and glutamate to tune surface chemistry while preserving the concave paratope architecture.

The mathematics of library design dictated careful restraint. Nineteen fully randomized positions would imply a theoretical diversity of roughly 5.2 × 10^24, unreachable by any physical display system. Instead, the team employed semi-randomization, enriching bulky aromatic and charged residues at positions likely to contact antigen while weighting choices according to amino-acid frequencies observed in natural nanobody interfaces, yielding a practical theoretical diversity of 10^18. The gene library itself was assembled through overlapping polymerase chain reactions, Type IIS restriction digestion, and T4 ligation, then transcribed into messenger RNA and stored in aliquots at minus 80 degrees Celsius.

Quality control came from deep sequencing. Approximately 9 × 10^7 reads of the naive library revealed that 98 percent of sequences were unique, and most designed amino-acid compositions at randomized positions matched expectations closely, particularly in CDR1, where 23 of 32 compositions showed discrepancies below 10 percent. Some positions, especially in CDR3, deviated more substantially, a caveat the authors acknowledge, though the overall fidelity confirmed that the construction strategy had faithfully captured the intended chemical space.

Functional validation proceeded against two test proteins. In head-to-head selections against calmodulin, performed in parallel with the benchmark concave library, S1.0 produced 45 positive clones from 96 screened, compared with 61 for the established library, but the S1.0 binders showed statistically significantly stronger calcium-dependent ELISA signals. Selections against thermostable green fluorescent protein, enzymatically biotinylated for magnetic bead capture, were even more striking: after one round of ribosome display and two rounds of phage display with off-rate selection, an enrichment factor of 427 was recorded, and ten unique sybody sequences emerged from twelve colonies analyzed. Fluorescence-detection size-exclusion chromatography showed that all ten shifted the elution profile of the fluorescent target, and biolayer interferometry confirmed nanomolar binding, with the best binder, named SH2, achieving a dissociation constant of 1.9 nanomolar.

The decisive experiment came next. When sybodies SH1 and SH2 were mixed with TGP and Legobody components, both assembled into stable complexes that could be pulled down through the maltose-binding protein’s affinity for amylose resin, detected directly by gel electrophoresis with no re-engineering of any kind. That intrinsic compatibility transforms the workflow for structural studies: a selection campaign that once ended with weeks of subcloning can now feed binders straight into Legobody assembly and cryo-EM grid preparation.

The authors are candid about remaining limitations. Validation covered only two soluble targets, and performance against the small membrane proteins that motivate the entire endeavor remains to be demonstrated. Sequencing also revealed instructive surprises, such as the strong underrepresentation of isoleucine at position 29 among functional binders despite its 20 percent abundance in the design, evidence that effective functional diversity, not raw theoretical diversity, ultimately governs library success. The team argues that accumulating sequencing data from naive and enriched libraries, mined with machine-learning approaches and shared through public repositories, should guide the next generation of synthetic nanobody libraries, turning empirical selection outcomes into ever-smarter designs.

Subject of Research: Design and validation of a ribosome display library for synthetic nanobody selection

Article Title: Design and validation of a ribosome display library for synthetic nanobody selection

Article References: Design and validation of a ribosome display library for synthetic nanobody selection. (n.d.). https://doi.org/10.1007/s44307-026-00112-z

Image Credits: AI Generated

DOI: 10.1007/s44307-026-00112-z

Keywords: nanobodies, sybodies, ribosome display, synthetic antibody library, Legobody, cryo-EM, structural biology, membrane proteins, CDR randomization, biolayer interferometry, phage display, protein engineering