Fig. 5. Screening of gene shuffled IFN-a libraries. (A) Summary of the shuffled IFN-a libraries created and screened during this study. The BC9 library (library 1) was produced by shuffling an IFN-a gene family PCR product derived from amplification of human genomic DNA with 12 sets of degenerate PCR primers (1). Library 2 was derived from shuffling four of the top hits from the tertiary screen of the BC9 library (sequences BC9.1, BC9.3, BC9.6, BC9.7). Library 3 is a synthetically shuffled library that incorporated into the consensus IFN backbone an ensemble of 20 structurally conservative changes extracted from the natural IFN-a sequence space. The primary screen for all libraries was a high throughput Th1 differentiation assay in which crude IFN-a library supernatants from transiently transfected COS cells were examined (two sample dilutions per clone) for their ability to induce Th1 differentiation on primary human CD4⁺ T cells in comparison to IFN-a8. Those clones with Th1 inducing activity similar to or better than an on-plate IFN-a8 control were subjected to the secondary screen. Variation among supernatants was generally <2-fold, as determined by Western blot analysis. For the secondary screen, the DNA corresponding to the selected clones was retransfected into COS cells, the amount of IFN protein in the cell supernatant was quantified by Western blot and the normalized cell supernatants were assayed in both the Th1 differentiation and Daudi tritiated thymidine uptake assays, using an eight-point dose-response curve. In the tertiary screen, purified and BCA-quantified protein preparations were examined for their Th1-differentiation, antiproliferation, and antiviral potency in the Th1, Daudi, and Huh7 EMCV assays, respectively. (B) Historical overview of the discovery of the B9X14 and B9X25 shuffled IFN-as.

1. Chang CC, Chen TT, Cox BW, Dawes GN, Stemmer WP, Punnonen J, Patten PA (1999) Nat Biotechnol 17:793-797.

Fig. 6. The protein sequence of B9.1.1 is aligned with respect to the wild-type human IFN-as, including LeIFN-E, a pseudogene that contributed sequences to B9.1.1 by crossbreeding.

Fig. 7. The protein sequence of B9.1.1 and B9.1.2 is aligned with respect to the B9X series IFN molecules.

Fig. 8. Suppression of HIV-1 by B9X14 and B9X25. HIV-1 infected primary CD4+T cells were incubated in the presence or absence of various concentrations of B9X14, B9X25, IFN-a2b and IFN-con1. The amount of p24 antigen in the culture supernatant was determined by ELISA (mean ± SD, n = 3).

Fig. 9. The Th1 inducing activity of B9X16 and B9X22 is blocked with a polyclonal antibody against human IFN-a. Primary human naïve T cells were exposed to (A) 20 ng/ml of B9X16 or (B) 20 ng/ml of B9X22 in the presence of a 2-fold serial dilution of rabbit polyclonal anti-human IFN-a antibody (PBL Biomedical Labortories, Piscataway, NJ; catalog no. 31130-1). As a control for background IFN-g expression, naïve T cells were incubated with the anti-IFN-a antibody dilution series alone and no IFN. Treatment with B9X16 (A) or B9X22 (B) in the absence of antibody produced »1,600 pg/ml of IFN- g in the culture supernatant.

Fig. 10. Binding specificity of biotinylated B9X14 to Daudi cells. Single cell suspensions of Daudi cells (2 ´ 10⁵ cells/well) were preincubated with binding buffer alone (■), 5 mg/ml unlabelled B9X14 (□) or 5 mg/ml IFNAR-2 neutralising antibody (x) before assaying binding of the indicated concentrations of carbohydrate biotinylated B9X14 (A) or amine biotinylated B9X14 (B). Binding of biotinylated B9X14 to Daudi cells is fully blocked by unlabelled B9X14 and anti-IFNAR2 antibodies.