Optimized Construction\nof a Yeast SICLOPPS Library\nfor Unbiased <i>In Vivo</i> Selection of Cyclic Peptides

DNA-encoded libraries hold great potential for discovering\nsmall,\ncyclized peptides with drug potential. Split-intein circular ligation\nof peptides and proteins (SICLOPPS) is a well-established method for <i>in vivo</i> selection of cyclic peptides targeting specific\nintracellular components. However, the method has mainly been used\nin prokaryotic cells. In contrast, selection studies performed directly\nin eukaryotic cells allow for the identification of cyclic peptides\npromoting a functional outcome, without the need to define a specific\ncellular target. Here, we report the construction of a Saccharomyces cerevisiae-specific SICLOPPS library\nof 80 million members, via careful optimization of several steps to\nincrease the size of the library. Individual library members were\nshown to be correctly expressed and processed in yeast. High-throughput\nsequencing was conducted on the randomized primer used for library\nconstruction and the pure yeast SICLOPPS library isolated from Escherichia coli. A distinct guanine insertion bias\nwas observed in the peptide-encoding, randomized sequence, which was\nprimarily attributed to the degenerate primer used to introduce the\nrandomized sequence. Moreover, high-throughput sequencing was performed\non the library before and after the induction of cyclic peptide expression\nin yeast. Importantly, expression of the SICLOPPS library in S. cerevisiae caused only a marginal further sequence\nbias. Our work paves the way for selection studies using a large and\ndiverse library to identify cyclic peptides of therapeutic interest\nthat promote a specific phenotypic outcome in eukaryotic organisms,\nwith yeast representing a beneficial model system due to its high\ntransformation efficiency.