Photo from Adam Morse.
Summary written by Czeslaw Radziejewski, Ph.D.
Antibody Engineering & Therapeutics, held in December 2022, offered many opportunities to hear exciting and informative presentations by experts in the field, including Vaughn Smider, Ph.D., President, Applied Biomedical Science Institute, who discussed ultralong CDR H3-based knobs as the smallest antibody fragment and Jeff Allen, Ph.D. Vice President, Protein Sciences, Pelican Expression Technology, who discussed Large-scale production of knob peptides.
In 1997 Osvaldo Lopez and his colleagues  at the University of Nebraska analyzed transcripts encoding the variable regions of immunoglobulin heavy chains from adult and fetal bovine splenocytes. They were the first to notice the presence of long heavy chain CDR3s. The bovine CDR3s ranged in length from 13 to 28 amino acids, with the average length of CDR H3 being 21 residues in both adults and fetuses. This was longer than had been previously reported for other mammals. In a subset of bovine antibodies, CDR H3s are ultralong (50-70 AA). The structure of ultralong CDR H3s was solved by Wang et al.  and the results demonstrated that ultralong CDR H3s all adopt similar architectures, with each composed of a long protruding beta-ribbon “stalk” and diverse disulfide-bonded “knob” (PDB: 4K3D). Up to six cysteine residues can be found in sequences of bovine CDR H3, all involved in disulfide bridges. The loops within the knob domain are thought to be involved in antigen binding. This contrasts with human antibodies where the antigen binding surface is formed from six CDR loops. Bovine CDR H3s are enormously diverse, and the diversity is generated by somatic hypermutation.  There is little diversity in CDR H1 and CDR H2, and cows use one light chain. Cows are not unique in having antibodies characterized by long CDR H3s. Other animals with these antibodies include zebu, yak, American and European bison.
It was previously observed that some broadly neutralizing antibodies against HIV also have longer CDR3s. Sok et al.  showed that immunization in cows could elicit rapid generation of neutralizing anti-HIV antibodies. Using x-ray crystallography, cryo-electron microscopy, and site-directed mutagenesis, Stanfield et al.  elucidated the structure of one monoclonal antibody elicited in cows by immunization with the HIV envelope trimer and showed molecular details of the knob mini-domain binding to a cryptic site on the gp120 CD4 receptor.
Knob domains are reasonably similar in size and shape to cyclotides/knottins, such as prototypic Cyclotide Kalata B1 and other disulfide-bonded peptides. Clinical applications for T cell immunotherapies are now emerging for analogs of cyclotides, for example, inhibition of the Kv1.3 channel. The Kv1.3 potassium channel is expressed abundantly on activated T cells and mediates the cellular immune responses. Sea anemone ShK cyclotide peptide was grafted into the β-ribbon ‘stalk’ of the ultralong CDR H3 scaffold of a humanized bovine IgG and showed the ability to block the Kv1.3.  The analog of the ShK peptide called ShK-186 or dalazatide blocks this channel, suppresses T-cell activation and is in human trials as a therapeutic for autoimmune disease.
When the Covid pandemic started, Applied Biomedical Science Institute, headed by Dr. Smider, generated antibodies by injecting cows with the spike protein of SARS-CoV-2. The aim was to produce anti-SARS-CoV-2 antibodies and small, disulfide-bounded knob peptides. Serum antibodies from immunized cows neutralized live virus and bound to the neutralizing epitopes at the apex of the spike protein. Antibodies that were generated neutralized wild-type, beta variant, delta, and two antibodies also neutralized omicron variant in an ELISA type assay and pseudovirus assay. Interestingly, two antibodies, 5C1 and 4C1, by cryo-EM, were shown to have a unique mode of action by splitting apart spike protein trimer upon binding. Another antibody, SKD, formed a co-crystal with RBD, and, as in the case of the anti-HIV case, it bound to the spike protein through two loops of the knob mini-domain.
If the knob domain of CDR H3 could be made as an independent, stable, and active entity, it would be the smallest antibody fragment available to date. Dr. Smider reported that knobs peptides against spike protein could indeed be expressed as a fusion protein with a thioredoxin in the Origami strain of E. coli and the knob peptides could be liberated with an enterokinase. When examined by NMR spectroscopy, the knob peptides exhibited unique disulfide bond patterns. Knob peptides bound to spike protein with affinities similar to full antibodies, but neutralizing activity against the coronavirus was 2 to 3 times lower than full antibodies. The Institute team hypothesized that the beta-ribbon stalk clamps together the ends of the knob, and this might be essential for good affinity and viral neutralization. The team then added two amino acids on both ends of one of their knobs and, using phage display screening technology, found sequences IS-TV and HW-SF that indeed demonstrated improvement. Further investigation is needed to determine how these sequences contribute to the improvement.
Multiple potential applications exist for knob mini-domains, including generating bispecific and multispecific formats. Several companies (UCB, Merck, Twist, Minotaur Therapeutics) are now interested in using Ultralong CDR H3 to develop novel biopharmaceuticals. Academic groups are pursuing studies of knob domain peptides. 
In the Q&A part of the presentation, Dr. Smider addressed the potential immunogenicity of knob domain-based therapeutics by speculating that small disulfide-bonded peptides are not very well presented through the MHC pathway; therefore, immunogenicity might not be a significant obstacle in clinical settings.
Large-scale production of knob peptides (termed picobodies) was addressed by Jeff Allen, Ph.D. Vice President, Protein Sciences, Pelican Expression Technology. Pelican expression technology is based on Pseudomonas fluorescens, which is a gram-negative, obligate aerobe, non-pathogenic bacteria. Pseudomonas can be grown at a very large scale in animal-origin-free and antibiotic-free media to very large cell densities. The technology is commercially and clinically validated in the USA and used to produce several internationally approved therapeutics and vaccines. 
The company has multiple expression plasmids and a large number of engineered host strains. This enables a combinatorial approach to assess thousands of expression strains in parallel to find the optimal one for a particular target. High-throughput analytics allows the testing of protein quality in crude extracts. The company also uses mass spectrometry (MS) and HPLC as second-tier analytics for in-depth quality analysis. This technology enables selection of the best expression system. Pelican technology was used to produce “knob-alone” picobodies. On the lab scale, picobodies were recovered from the clarified harvest of homogenized cells by cation exchange and size exclusion chromatography (SEC).
Dr. Allen discussed two picobodies against SARS-CoV-2 spike protein, R2G3 and R2F12. After lab-scale purification, both picobodies showed low nanomolar binding to the target. The company performed additional biophysical characterization of the picobodies, which included circular dichroism analysis and hydrogen–deuterium exchange MS. Since SEC is not practical on the industrial scale, large-scale processing was devised to take advantage of the small size of picobodies and, instead of size exclusion chromatography, employs ultrafiltration before chromatography steps. Membrane-based throughput was scalable up to 1000 l and resulted in high recovery. In addition, picobodies were found inside the Pseudomonas cells and in the fermentation broth. Both sources could be used for the recovery of the target picobodies.
- Berens SJ, Wylie DE, Lopez OJ. Use of a single VH family and long CDR3s in the variable region of cattle Ig heavy chains. Int Immunol. 1997 Jan;9(1):189-99. doi: 10.1093/intimm/9.1.189.
- Wang F, Ekiert DC, Ahmad I, Yu W, Zhang Y, Bazirgan O, Torkamani A, Raudsepp T, Mwangi W, Criscitiello MF, Wilson IA, Schultz PG, Smider VV. Reshaping antibody diversity. Cell. 2013 Jun 6;153(6):1379-93. doi: 10.1016/j.cell.2013.04.049.
- Stanfield RL, Wilson IA, Smider VV. Conservation and diversity in the ultralong third heavy-chain complementarity-determining region of bovine antibodies. Sci Immunol. 2016 Jul;1(1):aaf7962. doi: 10.1126/sciimmunol.aaf7962.
- Sok D, Le KM, Vadnais M, Saye-Francisco KL, Jardine JG, Torres JL, Berndsen ZT, Kong L, Stanfield R, Ruiz J, Ramos A, Liang CH, Chen PL, Criscitiello MF, Mwangi W, Wilson IA, Ward AB, Smider VV, Burton DR. Rapid elicitation of broadly neutralizing antibodies to HIV by immunization in cows. Nature. 2017 Aug 3;548(7665):108-111. doi: 10.1038/nature23301.
- Stanfield RL, Berndsen ZT, Huang R, Sok D, Warner G, Torres JL, Burton DR, Ward AB, Wilson IA, Smider VV. Structural basis of broad HIV neutralization by a vaccine-induced cow antibody. Sci Adv. 2020 May 27;6(22):eaba0468. doi: 10.1126/sciadv.aba0468.
- Selvakumar P, Fernández-Mariño AI, Khanra N, He C, Paquette AJ, Wang B, Huang R, Smider VV, Rice WJ, Swartz KJ, Meyerson JR. Structures of the T cell potassium channel Kv1.3 with immunoglobulin modulators. Nat Commun. 2022 Jul 4;13(1):3854. doi: 10.1038/s41467-022-31285-5.
- Macpherson A, Laabei M, Ahdash Z, Graewert MA, Birtley JR, et al. (2021) The allosteric modulation of complement C5 by knob domain peptides. eLife 10:e63586.
- Coleman R, Orchard E, Lee Y. Pseudomonas fluorescens. Cell-line development of a commercially proven platform for biopharmaceutical manufacturing. BioProcess International. May 24, 2022.