Leveraging mRNAs sequences to express SARS-CoV-2 antigens in vivo

Chunxi Zeng; Xucheng Hou; Jingyue Yan; Chengxiang Zhang; Wenqing Li; Weiyu Zhao; Shi Du; Yizhou Dong

The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.04.01.019877 doi: bioRxiv preprint Next, we further studied the effects of functional motifs integrated with the lead S27a 3' UTR. Previous studies reported that abundant cis-regulatory elements existed in mRNA 3' UTR to modulate protein expression, such as functional motifs for RNA binding proteins (RBPs) 11 . Specific RBPs are able to bind to their cognate motifs within mRNA 3' UTR and enhance protein expression through various mechanisms 30 . One of the RBPs, QKI-7, was shown to enable cytoplasmic polyadenylation of mRNA and upregulation of protein expression after binding to its target motif called QKI Response Element (QRE) in 3' UTR 31 . Two QRE sequences were obtained from previous reports, namely QRE1 32 and QRE2 33 . Another RBP, HuR, was believed to stabilize target transcripts by binding to its cognate motifs called AU-Rich Element (ARE) in eukaryotic cells 34 . One such ARE sequence was obtained from human eIF4E mRNA 3' UTR 34 . Besides eukaryotic mRNA, the sindbis virus (SinV) evolved to utilize part of its viral 3' UTR to recruit HuR and induce potent stabilization of the viral RNA genome 35 . The SinV RNA regions responsible for HuR binding, including repeated sequence element 3 and U-rich element (together named R3U) was obtained. Additionally, the ribosome binding fragments from two internal ribosome entry sites (IRESs) of encephalomyocarditis virus (EMCV) 36 and foot-andmouth disease virus (FMDV) 37 were selected. These two fragments possess high binding affinities to ribosomal subunits and have the potential to retain and recycle the subunits for more efficient translation re-initiation. Based on these literature findings, we prepared mRNA transcripts with the integration of the following elements after S27a 3' UTR: QRE1, QRE2, R3U, ARE, EMCV, and FMDV (Supplementary Table 3 ). R3U was identified as a preferred functional motif from the results in both Hep3B and 293T cells (Fig. 4a) . Through the comprehensive UTR engineering, the optimal UTRs was identified as NCA-7d as the 5' UTR and S27a plus a functional motif R3U as the 3' UTR (named as NASAR). After several rounds of design and validation of UTRs, we then compared NASAR with S27a-45, the starting point of endogenous gene together with two additional control UTRs (MOD1 and MOD2) in the literature 16, 17 . NASAR was over 10-fold more potent than S27a-45, 4 to 7-fold better than CYBA or AG+G, and up to 2-fold superior to MOD1 and MOD2 in cell lines and primary cells ( Fig. 4b and Supplementary Fig. 3) . These results demonstrate that our concept, integration of endogenous UTRs with further de novo design, is an efficient strategy for UTRs design. Relative luciferase activity was normalized to that of S27a-45. All data are presented as the mean ± s.d. (n=4). Statistical significance in a and b was analyzed by the two-tailed Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001; n.s., not significant.