Construction and next-generation sequencing analysis of a large phage-displayed V(NAR) single-domain antibody library from six naïve nurse sharks

Feng, Mingqian; Bian, Hejiao; Wu, Xiaolin; Fu, Tianyun; Fu, Ying; Hong, Jessica; Fleming, Bryan D; Flajnik, Martin F; Ho, Mitchell

To analyze the diversity of the library, we performed deep sequencing of the whole library using NGS. Each sequencing was done twice in both forward and reverse direction for paired-end reads. The merging of paired-end reads was then performed with high stringency by combining the forward and reverse sequencing results to ensure accuracy for each unique sequence. Nearly 1.2 million full-length V NAR sequences with in-frame translation were used for further analysis with a focus on cysteine numbers, CDR3 length, amino acid variability, and V NAR type counts. This is the largest scale shark V NAR sequence analysis reported thus far. As shown in Figure 2A , the presence of two canonical cysteines located at both amino acid 21 and 82 are used as a key criterion to characterize Type I-IV V NAR s. The sequences that do not contain one or both of these cysteines are considered as other types (n = 56 508; ∼5% of the total V NAR s) because they do not fit in with the four known V NAR type families. The sequences that have both 21C and 82C (n = 1 138 843) are further categorized based on their placement of additional cysteines. Type IV V NAR s contain only two canonical cysteines found at position 21 and 82 (n = 19 494). Type I V NAR s contain an extra cysteine at position 34 (n = 281 361; ∼24% of the total V NAR s). This group can be further divided into subtypes based on how many additional cysteines they contain. The classical definition of a Type I V NAR describes an antibody that has a total of six cysteines. About 11% of the total V NAR s are classical Type I. Type II and III V NAR s have at least one extra cysteine at amino acid 28 (n = 837 988; 70% of the total V NAR s). Among them, ∼57% of the total V NAR s are classical Type II. Unique full-length V NAR s are defined as having one differing amino acid in sequence. As shown in Figure 2B , 85% (1 022 715) of the 1.2 million sequences only appeared once in NGS results, and 8.5% of the 1.2 million sequences appeared twice. Only 1% of the 1.2 million sequences appeared more than 10 times. The most frequently repeated clone appeared 2832 times in NGS results. The percentage of the different types of V NAR is plotted in the pie chart in Figure 2C . Based on the number of extra cysteines in these sequences, the Type I (six cysteines in total) and Type II (four cysteines in total) V NAR s are considered classical to others (other numbers of cysteines) as shown in Figure 2D and E, respectively. Representative sequences of Type I-IV shark V NAR s with different numbers of cysteines were randomly picked from NGS data and shown here as examples (Fig. 3) . The FRs and CDR1, CDR3, HV2, and HV4 are marked based on Stanfield et al. [11] and Fennell et al. [13] in Figure 3 . These sequences were aligned to sequences on IMGT database for CDR determination. As shown in Figure 3 , part of the HV2 sequence was identified as "CDR2" in the IMGT database.