Isolating Anti-amyloid Antibodies from Yeast-Displayed Libraries

Abstract

Conformational antibodies specific for amyloid-forming peptides and proteins are important for a range of biomedical applications, including detecting, inhibiting, and potentially treating protein aggregation disorders ranging from Alzheimer's to Parkinson's diseases. Generation of anti-amyloid antibodies is greatly complicated by the complex, heterogeneous and insoluble nature of amyloid antigens. Here we describe systematic methods for isolating and affinity maturing anti-amyloid antibodies using yeast surface display. Magnetic-activated cell sorting is used to sort single-chain antibody libraries positively for binding to amyloid antigens and negatively against the corresponding disaggregated antigens to remove antibodies that bind in a conformation-independent manner. Isolated lead antibody clones with conformational specificity are affinity matured via targeted CDR mutagenesis and magnetic-activated cell sorting.

Keywords: Aggregate; Aggregation; Alzheimer’s; Conformation; Conformational; Directed evolution; Fiber; Fibril; Oligomer; Parkinson’s; Yeast surface display.

Figure 1.. Overview of conformational anti-amyloid antibody isolation from yeast-displayed libraries.

Single-chain antibody libraries displayed on yeast are sorted negatively to remove clones which bind to disaggregated antigen and positively to enrich clones which bind to amyloid aggregates. Negative selections can be performed using either magnetic-activated cell sorting (MACS) or fluorescence-activated cell sorting (FACS), and positive selections are performed with amyloid aggregates immobilized on beads using MACS. After several rounds of enrichment, the selected antibodies are cloned directly into Fc-fusion plasmids and expressed in mammalian cells. Individual antibody clones are analyzed using flow cytometry for affinity to amyloid fibrils and conformational binding to fibrils in the presence of disaggregated antigen. Clones identified with affinity and conformational specificity for the target amyloid fibrils are then affinity-matured through the preparation of yeast surface displayed sub-libraries. Sub-libraries are sorted stringently by reducing antigen loading or the number of antigen-coated beads to select for clones with improved affinity and conformational specificity.

Figure 2.. Flow cytometry-based analysis of antibody affinity and conformational specificity.

Amyloid fibrils are immobilized on micron-sized magnetic beads and binding of soluble antibodies is evaluated using flow cytometry. (A) Flow cytogram displaying forward scatter (488 nm) versus side scatter (488 nm) results, with the singlet population of fibril-coated beads highlighted in the R6 gate. (B) Histogram of the counts of binding events in the R6 gate as a function of the antibody binding signal detected via fluorescence measurements (647 nm). (C) Relative binding of antibodies to Aβ fibrils as a function of the concentration of wild-type (WT) and affinity-matured (C1, C2 and C3) antibodies. The results are mean signals from the histogram shown in (B). (D) Antibody binding to fibrils in the presence of disaggregated Aβ. Each antibody (30 nM) was pre-incubated with different concentrations of disaggregated Aβ, and then the antibodies were evaluated for their ability to recognize Aβ fibrils. A control (non-conformational antibody) displays decreased binding to immobilized fibrils as the concentration of disaggregated Aβ is increased.

Figure 3.. Overview of the design of antibody sub-libraries for affinity maturation.

The design of sub-libraries for affinity maturation involves three major steps. First, the CDR positions to mutate are chosen based on the variability of each CDR position in natural antibody repertoires. The sites are ranked from most variable (most attractive for mutagenesis) to least variable (least attractive for mutagenesis), and highly conserved positions (>50% WT on average in natural antibody repertoires) are eliminated from consideration. Moreover, a subset of CDR sites are also excluded from consideration if their WT residues are Arg, Lys, His or Cys. Second, for the selected ~6–10 CDR sites with the highest variability, degenerate codons are chosen that encode the WT residue as well as 3–5 other amino acids based on maximizing the sum of the average site-specific frequencies of each encoded residue in natural antibody repertoires (referred to as natural diversity coverage). Degenerate codons with Arg, Lys and His are excluded. The libraries are designed to typically encode 10⁶-10⁸ variants. If there are multiple possible degenerate codons that encode the same set of amino acid mutations, codon selection is based on species-specific codon usage (e.g., S. cerevisiae codon usage). Third, mutagenic primers with degenerate codons and amplification primers without mutations are designed. One mutagenic primer is designed for each mutated CDR that contains the site-specific degenerate codons and 18–22 base pairs of framework DNA on both ends of the mutated CDR. Three additional primers are required for generating sub-libraries with a single mutated CDR. Typically two or three CDRs are mutated when generating sub-libraries for affinity maturation.