Antibody-drug conjugates (ADCs) allow selective targeting of cytotoxic drugs to cancer cells presenting tumor-associated surface markers, thereby minimizing systemic toxicity. Traditionally, the drug is conjugated nonselectively to cysteine or lysine residues in the antibody. However, these strategies often lead to heterogeneous products, which make optimization of the biological, physical, and pharmacological properties of an ADC challenging. Here we demonstrate the use of genetically encoded unnatural amino acids with orthogonal chemical reactivity to synthesize homogeneous ADCs with precise control of conjugation site and stoichiometry. p -Acetylphenylalanine was site-specifically incorporated into an anti-Her2 antibody Fab fragment and full-length IgG in Escherichia coli and mammalian cells, respectively. The mutant protein was selectively and efficiently conjugated to an auristatin derivative through a stable oxime linkage. The resulting conjugates demonstrated excellent pharmacokinetics, potent in vitro cytotoxic activity against Her2 + cancer cells, and complete tumor regression in rodent xenograft treatment models. The synthesis and characterization of homogeneous ADCs with medicinal chemistry-like control over macromolecular structure should facilitate the optimization of ADCs for a host of therapeutic uses.

Summary Unlike humans or mice, some species have limited genome encoded combinatorial diversity potential, yet mount a robust antibody response. Cows are unusual in having exceptionally long CDR H3 loops and few V-regions, but the mechanism for creating diversity is not understood. Deep sequencing revealed that ultralong CDR H3s contain a remarkable complexity of cysteines, suggesting that disulfide-bonded mini-domains may arise during repertoire development. Indeed, crystal structures of two cow antibodies reveal that these CDR H3s form a very unusual architecture composed of a β-strand “stalk” that supports a structurally diverse, disulfide-bonded, “knob” domain. Sequence analysis suggests that diversity arises from somatic hypermutation of an ultralong DH with a severe codon bias towards mutation to cysteine. These unusual antibodies can be elicited to recognize defined antigens through the knob domain. Thus, the bovine immune system produces an antibody repertoire composed of CDR H3s of unprecedented length that fold into a diversity of mini-domains generated through combinations of somatically generated disulfides.

Three genetic complementation groups of rodent cells are defective for both repair of x-ray-induced double-strand breaks and V(D)J recombination. Cells from one group lack a DNA end-binding activity that is biochemically and antigenically similar to the Ku autoantigen. Transfection of complementary DNA (cDNA) that encoded the 86-kilodalton subunit of Ku rescued these mutant cells for DNA end-binding activity, x-ray resistance, and V(D)J recombination activity. These results establish a role for Ku in DNA repair and recombination. Furthermore, as a component of a DNA-dependent protein kinase, Ku may initiate a signaling pathway induced by DNA damage.

No immunogen to date has reliably elicited broadly neutralizing antibodies (bnAbs) to HIV in humans or animal models. Advances in the design of immunogens (BG505 SOSIP) that antigenically mimic the HIV envelope glycoprotein (Env)1 have improved the elicitation of potent isolate-specific Ab responses in rabbits2 and macaques3, but so far failed to induce bnAbs. One possible contributor to this failure is that the relevant antibody repertoires are poorly suited to target somewhat occluded conserved epitope regions on Env relative to exposed variable epitopes. To test this hypothesis, we immunized four cows with BG505 SOSIP. The antibody repertoire of cows contains long third heavy chain complementary determining regions (HCDR3) with an ultralong subset that can reach over 70 amino acids in length4–9. Remarkably, BG505 SOSIP immunization resulted in rapid elicitation of broad and potent serum antibody responses in all four cows. Longitudinal serum analysis for one cow showed the development of neutralization breadth (20%, n = 117 cross-clade isolates) in 42 days and 96% breadth (n = 117) at 381 days. A monoclonal antibody (mAb) isolated from this cow harbored an ultralong HCDR3 of 60 amino acids and neutralized 72% of cross-clade isolates (n = 117) with a potent median IC50 of 0.028 μg/ml. We note that breadth was elicited with a single trimer immunogen and did not require additional envelope diversity. Immunization of cows may provide an avenue to rapidly generate antibody prophylactics and therapeutics to address disease agents that have evolved to avoid human antibody responses.

Bispecific antibodies were constructed using genetically encoded unnatural amino acids with orthogonal chemical reactivity. A two-step process afforded homogeneous products in excellent yield. Using this approach, we synthesized an anti-HER2/anti-CD3 bispecific antibody, which efficiently crosslinked HER2+ cells and CD3+ cells. In vitro effector-cell mediated cytotoxicity was observed at picomolar concentrations.

Antibody conjugates are widely used as diagnostics and imaging reagents. However, many such conjugates suffer losses in sensitivity and specificity due to nonspecific labeling techniques. We have developed methodology to site-specifically conjugate oligonucleotides to antibodies containing a genetically encoded unnatural amino acid with orthogonal chemical reactivity. These oligobody molecules were used in immuno-PCR assays to detect Her2 + cells with greater sensitivity and specificity than nonspecifically coupled fragments, and can detect extremely rare Her2 + cells in a complex cellular environment. Such designed antibody-oligonucleotide conjugates should provide sensitive and specific reagents for diagnostics, as well as enable other unique applications based on oligobody building blocks.

We have devised a phage display system in which an expanded genetic code is available for directed evolution. This system allows selection to yield proteins containing unnatural amino acids should such sequences functionally outperform ones containing only the 20 canonical amino acids. We have optimized this system for use with several unnatural amino acids and provide a demonstration of its utility through the selection of anti-gp120 antibodies. One such phage-displayed antibody, selected from a naïve germline scFv antibody library in which six residues in V H CDR3 were randomized, contains sulfotyrosine and binds gp120 more effectively than a similarly displayed known sulfated antibody isolated from human serum. These experiments suggest that an expanded ''synthetic'' genetic code can confer a selective advantage in the directed evolution of proteins with specific properties. directed evolution ͉ phage display ͉ unnatural amino acids W ith few exceptions, the genetic codes of all known organisms specify only the 20 canonical amino acids for protein synthesis. Yet it is quite possible that the ability to encode additional amino acids and their corresponding chemical functionalities would be evolutionarily advantageous, especially since nature's choice of 20 could have been arbitrarily fixed at the point of transition between communal and Darwinian evolution paradigms and subsequently sustained by the code's inertia (1). Furthermore, in the limited scope of laboratory-directed evolution, which concerns only one or few specific functions over a short time rather than general organismal fitness over thousands of years, one can easily envision a selective advantage conferred by additional amino acids. Recent developments in our laboratory allow us to explore this possibility. Specifically, orthogonal tRNA/aminoacyl-tRNA synthetase (aaRS) pairs capable of incorporating various unnatural amino acids into proteins in response to unique nonsense and frameshift codons have been added to the translational machinery of Escherichia coli (2). These E. coli (X-E. coli) can now be used for evolution of protein function wherein 21 building blocks, rather than the common 20, are available.Several unnatural amino acids were initially chosen, on the basis of their unique chemistries, for use in our system. For example, X-E. coli genetically encoding the bidentate metalchelating amino acid bipyridyl-alanine (3) are well-suited for the evolution of redox and hydrolytic catalysts, as metal ion binding would not require preorganized primary and secondary ligand shells. Similarly, X-E. coli encoding the reactive 4-boronophenylalanine (4) are well-suited for evolution of receptors specific for glycoproteins or serine protease inhibitors, because the boronate group can form covalent complexes with diols or reactive serine residues. In addition, X-E. coli genetically encoding otherwise posttranslationally modified amino acids, such as sulfotyrosine (5), can be used for evolution of properties that exploit the unique chemical characterist...

X-ray-sensitive hamster cells in complementation groups 4, 5, 6, and 7 are impaired for both double-strand break repair and V(D)J recombination. Here we show that in two mutant cell lines (XR-V15B and XR-V9B) from group 5, the genetic defects are in the gene encoding the 86-kDa subunit of the Ku autoantigen, a nuclear protein that binds to double-stranded DNA ends. These mutants express Ku86 mRNA containing deletions of 138 and 252 bp, respectively, and the encoded proteins contain internal, in-frame deletions of 46 and 84 amino acids. Two X-ray-resistant revertants of XR-V15B expressed two Ku86 transcripts, one with and one without the deletion, suggesting that reversion occurred by activation of a silent wild-type allele. Transfection of fulllength cDNA encoding hamster Ku86 into XR-V15B cells resulted in a complete rescue of DNA-end-binding (DEB) activity and Ku70 levels, suggesting that Ku86 stabilizes the Ku70 polypeptide. In addition, cells expressing wild-type levels of DEB activity were fully rescued for X-ray resistance and V(D)J recombination, whereas cells expressing lower levels of DEB activity were only partially rescued. Thus, Ku is an essential component of the pathway(s) utilized for the resolution of DNA double-strand breaks induced by either X rays or V(D)J recombination, and mutations in the Ku86 gene are responsible for the phenotype of group 5 cells.All cells possess a mechanism for repairing DNA doublestrand breaks (DSBs) produced by ionizing radiation. Cells of the immune system must also resolve DNA DSBs produced by V(D)J recombination of the immunoglobulin and T-cell receptor genes during development of B and T cells (reviewed in reference 28). In fact, the biochemical pathways for DSB repair and V(D)J recombination have a number of common factors. Evidence of this was first described for the severe combined immune deficient (scid) mouse, which is hypersensitive to ionizing radiation because of defective DSB repair (1, 12, 18) and immune deficient because of defective V(D)J recombination (29). Subsequently, other X-ray-sensitive rodent cell lines defective in DSB repair were also found to have impaired V(D)J recombination (17,27,34,45,47). These cell lines fall into four complementation groups, 4, 5, 6, and 7, with group 7 corresponding to the scid defect (43, 52). The human genes capable of rescuing these mutants are designated XRCC4, XRCC5, XRCC6, and XRCC7, respectively (XRCC denotes X ray cross-complementing) (for a review, see reference 51).Recently, it has been shown that the Ku protein is involved in both DSB repair and V(D)J recombination. Ku is an abundant nuclear protein identified originally as an autoantigen recognized by sera from patients with autoimmune diseases, including scleroderma-polymyositis overlap syndrome and systemic lupus erythematosus (31