Can an antibody bind picomolar to a floppy loop?
Why a flexible epitope is a hard target
Antibodies tend to be raised against, and to bind best to, parts of an antigen that hold a defined shape. A loop that samples many conformations costs entropy on binding and gives the paratope nothing stable to grip. When a programme stalls on a “difficult target”, disorder in the epitope is one of the common underlying reasons — and it is usually invisible until after the campaign has failed.
This paper is a worked example of the opposite outcome. Molecular dynamics on the target region identified the Leu1181–Leu1189 loop as one of the two most flexible stretches in complement factor H domains CCP18–20. That flexible loop is exactly what the antibody binds, at picomolar affinity with a slow off-rate.
The measured result
| Parameter | Value | Method |
|---|---|---|
| Affinity (VHH4 × CFH CCP18-20) | 1.71 ± 0.66 pM | SPR, Biacore 8K |
| Dissociation | Notably slow | SPR, 1800 s dissociation |
| Binding enthalpy | Exothermic; KD below the accurate range of the instrument | ITC |
| Complex structure | 2.6 Å resolution | X-ray crystallography |
| Epitope | Leu1181–Leu1189 loop of CCP20 | Crystal structure |
Route to the clone: an alpaca was immunised with recombinant CFH CCP18-20, a VHH phagemid library was built from peripheral blood B cells, and three rounds of biopanning returned ten VHH clones. Six expressed and purified cleanly; SPR screening identified VHH4.
How it grips something that will not hold still
Two structural features do the work, and both are unusual.
An exceptionally long CDR3. VHH4’s CDR3 runs 24 residues (Ala99–Tyr122), long even among VHHs. Rather than the typical VHH mode of inserting that loop into a cavity on the antigen, VHH4 does the reverse: it captures the antigen’s Trp1183 side chain inside its own CDR3. The buried surface area of that contact is 242.2 Ų.
A framework cysteine that should not be there. An intramolecular disulfide forms between Cys106 in the CDR3 and Cys50 in the framework region. A cysteine at that framework position occurs in under 1% of sequences. It appears to have been selected because it pins the very long CDR3 into the one conformation that fits the target — the antibody solved the flexibility problem by making itself rigid first.
Alanine scanning across ten interface residues confirmed the architecture: mutating the antigen’s Trp1183 dropped affinity below the micromolar range, and on the antibody side Tyr118 and Gly119 each cost roughly 9.4 kcal/mol of binding enthalpy.
The biology this enabled
The loop VHH4 binds is the same region targeted by anti-CFH autoantibodies in atypical haemolytic uraemic syndrome, a disease where CFH dysfunction carries a high probability of end-stage renal disease. In a haemolysis assay VHH4 reproduced the pathogenic behaviour: it blocked CFH binding to sialic acid on sheep erythrocytes and induced lysis, while leaving CFH’s cofactor activity for C3b degradation intact. It is, in effect, a defined recombinant stand-in for a patient autoantibody — which is what makes it useful as a research tool.
This work was an academic collaboration and the paper declares no conflicts of interest. We describe it here as a capability example, not as a product.
Have a target other campaigns have failed on?
Flexible epitopes, membrane proteins, conserved self-antigens — tell us the target and what has already been tried. A PhD scientist reads it and replies within two business days. No NDA needed to start.
Talk to a scientist →Frequently asked questions
Can you raise antibodies against disordered or flexible epitopes?
Sometimes, and this paper is the existence proof — 1.71 pM against the most flexible loop in the target region. It is not guaranteed for any given epitope. What improves the odds is a large selected repertoire rather than a designed or narrowly screened panel, because the solutions that work tend to be structurally unusual.
Do you work with VHH as well as conventional antibodies?
Yes. VHH and single-domain discovery from immunised camelids is a standard campaign type, and single-domain engineering — solubility, stability, humanization — runs alongside it.
Why did you use an alpaca rather than a rabbit here?
The scientific question needed a single-domain binder small enough to interfere with one interaction surface while leaving another intact, which is exactly what happened: VHH4 blocked CFH binding to sialic acid without blocking C3b degradation. An IgG would have been more likely to occlude both.
How long does a VHH campaign take?
Immunisation through characterised recombinant clones runs on a similar timescale to a rabbit campaign, roughly four to six months for a typical target.
Do we own the resulting clones?
In commercial engagements, yes — sequences included, with no downstream royalties in standard scope. The work described here was an academic collaboration with separate arrangements.
Want this run on your target?
Reaching an epitope a conventional antibody cannot is the reason to use a single domain at all. These are the campaigns that produce them.
- VHH / single-domain discovery — selection for extra-long HCDR3 and HCDR2/HCDR3 cysteines
- Functional and neutralizing discovery — when selective blocking, not binding, is the endpoint
- VHH humanization — taking a camelid domain toward clinical grade
Every campaign is scoped individually and quoted. Email info@abwizbio.com with your target or lead sequence, or see all 25 publications and the full service menu.