Gene-product expression constructs and systems

Abstract

Constructs for expression of a gene product can comprise a circular RNA construct comprising sequence for encoding the gene product operably linked to an I RES-like sequence. The IRES-like sequence can be derived from a Hoxa gene or other developmental gene. Circular RNA constructs can be utilized to produce a gene product, treat an individual with a therapeutic gene product, or vaccinate an individual with an immunogenic gene product.

Description

GENE-PRODUCT EXPRESSION CONSTRUCTS AND SYSTEMSCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Appl. No. 63 / 695,706, entitled “Gene-Product Expression Constructs and Systems,” filed September 17, 2024, the disclosure of which is hereby incorporated by reference in its entirety.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0002] This invention was made with Government support under contract HD086634 awarded by the National Institutes of Health. The Government has certain rights in the invention.TECHNOLOGICAL FIELD

[0003] The present disclosure relates to gene-product expression systems, including circular RNA systems that utilize untranslated regions of eukaryotic homeobox (Hox) mRNA and other developmental gene mRNA.BACKGROUND

[0004] Homeobox (Hox) genes encode a family of transcription factors that act as master regulators of tissue development of multicellular species. Their spatiotemporal expression during development is tightly controlled such that local concentrations of Hox genes influence anatomical development such as limb morphogenesis. In vertebrates, there are four paralogous clusters of Hox genes: A, B, C, and D. These clusters yield e.g., in mice, the following clusters of genes: Hoxa, Hoxab, Hoxac, and Hoxad. It has been reported that mRNA molecules of several Hoxa genes include a sequence within the 5’-untranslated region (5’-UTR) that folds into an RNA secondary structural formation similar to internal ribosome entry site (IRES) sequences commonly found in viral genomes, allowing for cap-independent translation of these transcripts (see., e.g., S. Xue, et al., Nature. 2015 Jan 1 ;517(7532):33-8, the disclosure of which is hereby incorporated by reference). Herein, eukaryotic RNA elements with IRES activity are referred to as to “IRES- like” RNA elements to distinguish from viral IRESs. Further reports, however, claimed that IRES- like sequences were inaccurately annotated in Hoxa genes and that these RNA molecules lacked ability to perform cap-independent translation (see, e.g., C. Akirtava, et al., Proc Natl Acad Sci U S A. 2022 Sep 6;119(36):e2122170119; and I. P. Ivanov, et al., Proc Natl Acad Sci U S A. 2022 Mar 1;119(9):e2117226119; the disclosures of which are provided for context). Due to theseconflicting reports, it had remained unclear whether Hoxa genes include IRES-like sequences in their 5’-UTRs.SEQUENCE LISTING

[0005] This application hereby incorporates by reference the material of the electronic Sequence Listing filed concurrently herewith. The material in the electronic Sequence Listing is submitted as an XML file entitled “S31-09125PCT.xml” created on September 5, 2025, which has a file size of 49,277 bytes, and is hereby incorporated by reference in its entirety.SUMMARY

[0006] This summary is meant to provide some examples and is not intended to be limiting of the scope of the technologies to be claimed. For example, any feature included in an example of this summary is not required by the claims, unless the claims explicitly recite the features. Various features and steps as described elsewhere in this disclosure may be included in the examples summarized here, and the features and steps described here and elsewhere can be combined in a variety of ways.

[0007] In some aspects, the techniques described herein relate to a composition, including: a circular RNA construct including a sequence for encoding a gene product that is operably linked to an IRES-like sequence derived from a 5 -UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

[0008] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

[0009] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence includes one of: SEQ ID NOs: 1-5, 37, or 38.

[0010] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA 11, CHRDL1, or DLX1.

[0011] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence includes one of: SEQ ID NOs: 6-9, 39 or 40.

[0012] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a non-self gene product.

[0013] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a heterologous gene product.

[0014] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

[0015] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a therapeutic protein or peptide including: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

[0016] In some aspects, the techniques described herein relate to a composition, wherein the gene product is an immunogen including: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

[0017] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a cytotoxin including: TNF, a granzyme, a perforin, a caspase, a ricin, or a cellpenetrating peptide.

[0018] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a gene editing tool including: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

[0019] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a molecular marker including: a fluorescent protein, a luciferase, or a peroxidase.

[0020] In some aspects, the techniques described herein relate to a composition, including: a DNA construct including a sequence for transcribing a linear mRNA molecule that is operably linked to a promoter and poly-A signal; wherein the linear mRNA molecule includes a sequence including two introns, a gene sequence split into a 3'-gene sequence and a 5'-gene sequence, and an IRES-like sequence derived from a 5'-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene; wherein the IRES- like sequence is downstream of the 3'-gene sequence and upstream the 5'-gene sequence; whereby processing of the linear mRNA molecule by a spliceosome yields a circular RNA construct including the gene sequence operably linked to the IRES-like sequence; whereby the processing of the linear mRNA molecule by the spliceosome also conjoins the 5'-gene sequence and the 3'-gene sequence to yield the gene sequence in a configuration that encodes a product of the gene sequence.

[0021] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

[0022] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence includes one of: SEQ ID NOs: 10-14, 41 or 42.

[0023] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA 11, CHRDL1, or DLX1.

[0024] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence includes one of: SEQ ID NOs: 15-18, 43 or 44.

[0025] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a non-self gene product.

[0026] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a heterologous gene product.

[0027] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

[0028] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a therapeutic protein or peptide including: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

[0029] In some aspects, the techniques described herein relate to a composition, wherein the gene product is an immunogen including: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

[0030] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a cytotoxin including: TNF, a granzyme, a perforin, a caspase, a ricin, or a cellpenetrating peptide.

[0031] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a gene editing tool including: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

[0032] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a molecular marker including: a fluorescent protein, a luciferase, or a peroxidase.

[0033] In some aspects, the techniques described herein relate to a method of producing a gene product, including: contacting a eukaryotic cell with a circular RNA construct, wherein the circular RNA construct includes a sequence for encoding the gene product that is operably linked to an IRES-like sequence derived from a 5'-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

[0034] In some aspects, the techniques described herein relate to a method, wherein the I RES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, Chrd / 1, or Dlx1.

[0035] In some aspects, the techniques described herein relate to a method, wherein the I RES-like sequence includes one of: SEQ ID NOs: 1-5, 37, or 38.

[0036] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA 11, CHRDL1, or DLX1.

[0037] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence includes one of: SEQ ID NOs: 6-9, 39, or 40.

[0038] In some aspects, the techniques described herein relate to a method, wherein the gene product is a non-self gene product.

[0039] In some aspects, the techniques described herein relate to a method, wherein the gene product is a heterologous gene product.

[0040] In some aspects, the techniques described herein relate to a method, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

[0041] In some aspects, the techniques described herein relate to a method, wherein the gene product is a therapeutic protein or peptide including: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

[0042] In some aspects, the techniques described herein relate to a method, wherein the gene product is an immunogen including: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

[0043] In some aspects, the techniques described herein relate to a method, wherein the gene product is a cytotoxin including: TNF, a granzyme, a perforin, a caspase, a ricin, or a cellpenetrating peptide.

[0044] In some aspects, the techniques described herein relate to a method, wherein the gene product is a gene editing tool including: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

[0045] In some aspects, the techniques described herein relate to a method, wherein the gene product is a molecular marker including: a fluorescent protein, a luciferase, or a peroxidase.

[0046] In some aspects, the techniques described herein relate to a method of treating an individual, including: administering to the individual a circular RNA construct, wherein the circularRNA construct includes a sequence for encoding a therapeutic gene product that is operably linked to an IRES-like sequence derived from a 5'-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

[0047] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

[0048] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence includes one of: SEQ ID NOs: 1-5, 37, or 38.

[0049] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA 11, CHRD , or DLX1.

[0050] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence includes one of: SEQ ID NOs: 6-9, 39 or 40.

[0051] In some aspects, the techniques described herein relate to a method, wherein the gene product is a non-self gene product.

[0052] In some aspects, the techniques described herein relate to a method, wherein the gene product is a heterologous gene product.

[0053] In some aspects, the techniques described herein relate to a method, wherein the therapeutic gene product includes: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

[0054] In some aspects, the techniques described herein relate to a method of vaccinating an individual, including: administering to the individual a circular RNA construct, wherein the circular RNA construct includes a sequence for encoding an immunogenic gene product that is operably linked to an IRES-like sequence derived from a 5'-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

[0055] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

[0056] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence includes one of: SEQ ID NOs: 1-5, 37, or 38.

[0057] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA 11, CHRDU, or DLX1.

[0058] In some aspects, the techniques described herein relate to a method, wherein the IRES-like sequence includes one of: SEQ ID NOs: 6-9, 39, or 40.

[0059] In some aspects, the techniques described herein relate to a method, wherein the gene product is a non-self gene product

[0060] In some aspects, the techniques described herein relate to a method, wherein the gene product is a heterologous gene product.

[0061] In some aspects, the techniques described herein relate to a method, wherein the immunogenic gene product includes: a protein or a peptides derived from a pathogen, or a cancer- related neoantigen.

[0062] In some aspects, the techniques described herein relate to a composition, including: a DNA construct including a sequence for transcribing a linear mRNA molecule that is operably linked to a promoter capable of driving in vitro transcription in a cell-free buffer; wherein the linear mRNA molecule includes a sequence including two self-splicing ribozyme introns, a gene sequence encoding a gene product, and an IRES-like sequence derived from a 5'-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene; wherein the IRES-like sequence is downstream of the gene sequence; whereby the linear mRNA molecule is configured to self-splice via the two self-splicing ribozyme introns to yield a circular RNA construct including the gene sequence operably linked to the IRES- like sequence; whereby the self-splicing of the linear mRNA molecule by the self-splicing ribozyme introns also conjoins the gene sequence with the IRES-like sequence using complementary exon junction sequences to operably link the gene sequence with the IRES-like sequence.

[0063] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdH, or Dlx1.

[0064] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence includes one of: SEQ ID NOs: 10-14, 41 or 42.

[0065] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA 11, CHRDL1, or DLX1.

[0066] In some aspects, the techniques described herein relate to a composition, wherein the IRES-like sequence includes one of: SEQ ID NOs: 15-18, 43 or 44.

[0067] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a non-self gene product.

[0068] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a heterologous gene product.

[0069] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

[0070] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a therapeutic protein or peptide including: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

[0071] In some aspects, the techniques described herein relate to a composition, wherein the gene product is an immunogen including: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

[0072] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a cytotoxin including: TNF, a granzyme, a perforin, a caspase, a ricin, or a cellpenetrating peptide.

[0073] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a gene editing tool including: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

[0074] In some aspects, the techniques described herein relate to a composition, wherein the gene product is a molecular marker including: a fluorescent protein, a luciferase, or a peroxidase.BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The description and claims will be more fully understood with reference to the following figures and data graphs, which are presented as exemplary embodiments of the invention and should not be construed as a complete recitation of the scope of the invention.

[0076] Figure 1A provides a schematic of a circular RNA construct.

[0077] Figures 1 B and 1C each provide a schematic of a DNA construct to yield a circularRNA construct.

[0078] Figures 2A-2D provide schematics and data results of an assessment of Hoxa9 mRNA isoform expression in the mouse embryo by single-molecule fluorescent in situ hybridization (smFISH). Figure 2A: Model depicting the expression pattern of HoxA cluster genes in mouse embryos. The Hoxa9 gene is expressed in posterior somites and the neural tube (NT), and canbe expressed in mRNA isoforms with either a long 5’ UTR and full CDS (isoform A; ENSMUST00000048680.8) or short 5’ UTR and truncated CDS (isoform B; ENSMUST00000114425.3), as annotated in mouse GENCODE release M34 (GRCm39). A third mRNA isoform C is annotated in NCBI to contain the short 5’ UTR and the full CDS (NM_010456.4). TIE, translation inhibitory element; IRES, internal ribosome entry site. Regions targeted by smFISH probes are indicated as “IRES” and “CDS” with the respective coordinates of the ENSEMBL-annotated isoforms. Figure 2B: Representative images of E11 mouse embryo sections immunostained with DAPI, the Hoxa9 CDS probe, the Hoxa9 I RES-like probe and the 3- plex Negative Control probes. Boxed regions #1 and #2 indicate the regions for which quantifications were performed with results plotted in Fig. 2D. Arrows point to individual somites. Figure 2C: Representative zoomed-in views of the numbered boxes within Fig. 2B. #1 represents somites 20-28, while #2 represents the neural tube. Schematics of the potential mRNA isoforms labeling is indicated in Fig. 3A. Figure 2D: Quantification of the proportion of the long and short 5’ UTR isoforms in the respective regions from Fig. 2C. Co-localized CDS and I RES-like probe signals represent the long 5’ UTR isoform, while the non-colocalized CDS-only probe signals represent the short 5’ UTR isoform. IRES-only signal is not detected at a biologically relevant level. The long 5’ UTR isoform is overall the major isoform in the posterior somites and the neural tube, n = 3 embryos each.

[0079] Figures 3A-3F provide schematics and data results of an assessment of Hoxa cluster mRNA 5’ UTR isoform expression by PacBio sequencing and Hoxa9 mRNA abundance and translation in embryonic tissues. Figure 3A: Illustration of a genome browser snapshot showing the transcript models predicted by IsoQuant around Hoxa9 and HoxalO loci (chr6:52,200,050- 52,217,850). There was no novel isoform identified that indicates any Hoxa9 / 10 fusion transcripts. Only two novel transcripts were detected for the HoxalO gene. The Hoxa9-201 IRES-containing reference transcript together with Hoxa9-202 and Hoxa9-204 were detected. Figure 3B: Schematic of Hoxa9 mRNA amplicons used for RT-qPCR analysis. Figure 3C: RT-qPCR results from micro-dissected E11.5 mouse embryonic tissue. Values shown are the difference in Ct values between tissue samples (3 biological replicates each) and a respective no-RT control. Error bars are standard deviation (SD), n=3. Figure 3D: RT-qPCR results using 0.1 fg of a plasmid containing the Hoxa9 I RES-like element and CDS as template. Values shown are the difference in Ct values between DNA samples (n=3) and a no-RT control as in Fig. 3B. Error bars are SD, n=3. Figure 3D: Representative sucrose gradient fractionation trace from a E11 .5 neural tube and somite sample and quantification of the fraction of total mRNA found in each of the five portions of the gradient, as demarcated in the schematic. Figure 3F: Three biological replicates wereperformed and the values for each Hoxa9 mRNA (ENSMUST00000048680) amplicon are shown as mean + / - standard error of the mean (SEM), n=3. qPCR amplicons correspond to as illustrated in Fig. 3A. For comparison, a highly translated housekeeping mRNA, Nup / 1 (ENSMUST00000225805), is also shown (dotted black line), as well as the respective mRNA ORF lengths.

[0080] Figures 4A-4J provide schematics and data results of an improved reporter system and of Hoxa IRES-like elements are independent of transcription in promoterless reporters. Figure 4A: Schematic of SV40 promoter-driven monocistronic Nanoluc (NIuc) with an N-terminal 3xHA tag, the optimized N\uc-$-globin (NLB) reporter mRNA that leads to a fusion protein, and the control Firefly (Flue) reporter mRNA with the calculated molecular weight of their encoded protein in kiloDalton (kDa). mRNA reporters contain hHBB as a control 5’ UTR. IRES-like elements were introduced into the NIuc 5’ UTR instead of hHBB. Figure 4B: Reporter genes were transiently transfected into mouse C3H / 10T1 / 2 cells and expressed from plasmids. Relative luciferase activity is expressed as a Nluc / Fluc ratio. Average luciferase activity ± standard error of the mean (SEM), n = 5; NLB was normalized to 1; A.U., arbitrary units. Figure 4C: EMSFold yields a high confidence prediction of the structure of the designed NLB reporter (average per-residue model confidence score plddt of 0.78). Predicted structure is colored by the model confidence. For comparison, the crystal structure of shrimp Nanoluc (PDB ID 5IBO) and chain B of rabbit hemoglobin (PDB ID 2RAO) are shown, in dark and light gray, respectively. Crystal structures aligned to the respective domains in the NLB reporter and the root-mean-square deviation (RMSD) for each of them are provided. Figure 4D: Substitution mutations were mapped onto the linear WT P4 sequence (labeled according to conservation) that were tested in context of the full- length Hoxa9 IRES-like element (323 nt). Nts critical for IRES activity are highlighted, also in Fig. 4E. Numbers refer to nucleotide positions within the Hoxa9 5’ UTR. The suspected 6-nt putative and common E-box is aligned accordingly. Included in Fig. 4D are SEQ ID NOs: 5, 20, 23, 30, & 31 (Table 1). Figure 4E: Schematic representation of the mouse Hoxa9 secondary structure model of the 180-nt long Hoxa9 IRES-like element RNA with P4 highlighted. The secondary structure model of the P4 stem-loop and disruptive substitution mutations mapped onto the P4 stem. P4 mutants moderately active are labeledand inactive mutants are labeled M12 and M13 mutants are new. Included in Fig. 4E are SEQ ID NOs: 5, 20, 23, & 30-32 (Table 1). Figure 4F: Substitution mutations are mapped onto the linear P4 sequence, together with the IRES-like activity (+, + / -, -) of the corresponding reporter mRNAs in context of the native spacer or the acf / n(inv) sequence. Included in Fig. 4F are SEQ ID NOs: 5 & 19-29 (Table 1). Figure 4G: Schematic representation of the mouse Hoxa9 secondary structure model of the 180 nucleotides(nt) long Hoxa9 I RES-like element RNA (termed a9 IRESiso) containing four pairing elements P1- P4 and a putative pseudoknot (PK), and P4 highlighted. Numbers refer to nucleotide positions within the Hoxa9 5’ UTR. Secondary structure model of the P4 stem-loop and substitution mutations mapped onto the P4 structure. Numbers refer to nucleotide positions within the Hoxa9 5’ UTR. P4 stem mutations were introduced either in one strand to disrupt P4, or as compensatory mutations in both strands to restore P4. P4 mutants active in mediating I RES-like activity in the context of the fusion to the native spacer (P4-native) or the acf / n(inv) 5’ UTR sequence (P4- acfzn(inv)) (normalized Fluc / Rluc< 0.5 A. U.) are labeled "+", moderately active mutants (Fluc / Rluc < 0.5, > 1.0 A.U.) are labeledand inactive mutants (Fluc / Rluc > 0.5 A.U.) are labeled Sequence critical for IRES activity is highlighted in yellow. Included in Fig. 4G are SEQ ID NOs: 5, 19-29, & 33 (Table 1). Figure 4H: Schematic of the topology of regulatory elements in the mouse Hoxa95’ UTR and promoterless reporter assay design. The 323 nt-long Hoxa9 full-length (FL) I RES-like RNA element (a9 1 RES-like FL) harbors the P4 stem-loop. Cryptic promoter activity from Hoxa I RES-like elements (a3, a5, a9) is tested by inserting full-length IRES-like elements upstream in a NLB reporter mRNA plasmid lacking the SV40 promoter (ASV40). Viral IRES controls (EMCV, HCV), an empty vector control (empty; no insert in the 5’ UTR region) and control 5’ UTR (hHBB) were included. A co-transfected control reporter ( HBB-Fluc) under an active SV40 promoter served as reference. Figure 4I: Normalized NIuc mRNA levels from promoterless NLB constructs were measured in transiently plasmid-transfected C3H / 10T1 / 2 cells as in (B). Cells from the same transfection were split in half for mRNA and protein analysis. Average NIuc mRNA levels are expressed as respective globin / NupL1 mRNA levels ± SEM, n = 2-7. Promotercontaining plasmids (+SV40) of empty and hHBB inserts served as positive controls for expression, empty and hHBB 5’ UTR serve as negative controls for ASV40 constructs; +SV40- hHBB was normalized to 1. Figure 4J: Normalized Nluc / Fluc luciferase activity from promoterless constructs. Average luciferase activity ± SEM, n = 4-8; +SV40- / ? / 7BB was normalized to 1 ; ns, not significant.

[0081] Figures 5A-5C provide schematics and data results showing IRES activity of a circRNA reporter system. Figure 5A: Experimental outline of the circRNA reporter assay based on the mRuby-ZKSCAN-splitEGFP plasmid for the screening of IRES activity of the different tested insert sequences, including inverse sequences as controls for circRNA translation activity dependent on insert length and GC-content. Following plasmid transfection, the expression of the reporter system under CMV promoter control leads to the linear pre-mRNA which is circularized through spliceosome-mediated backsplicing (grey box). Cells were harvested after 24 hrs and 72 hrs, and their mRuby signal (transfection control) and EGFP signal (readout for IRES activity) wasdetected. Figure 5B: Calculated median fluorescence intensities (MFIs) of EGFP of the mRuby+ subtractions are shown at 24 hrs and 72 hrs. Bar graphs are indicating mean values ± SEM, n = 3-6. Empty vector control was normalized to 1 ; ns, not significant. Figure 5C: Inverse sequences of active cellular IRESes \ were tested as controls for circRNA translation activity dependent on insert length and GC-content, which are identical in the forward and inverse sequences. We did pairwise comparison of the respective inserts. MFIs of EGFP of the mRuby+ subtractions are shown at 72 hrs. Bar graphs are indicating mean values ± SEM, n = 3-6. Empty vector control was normalized to 1. Source data are available online for this figure. In all figures, data was presented as mean, SD or SEM as stated, and *p < 0.05 was considered significant (ns: p > 0.05; *p < 0.05; **p < 0.01 ; ***p < 0.001 ; ****p < 0.0001). Tests, two-tailed unpaired Student’s t-test if not stated otherwise.

[0082] Figures 6A-6F provide schematics and data results showing Hoxa I RES-like activity in circRNA reporters. Figure 6A: Experimental outline of the circRNA reporter assay based on the mRuby-ZKSCAN-splitEGFP plasmid for the screening of IRES-like activity of different insert sequences of Hoxa IRES-like elements. Figure 6B: Calculated MFIs of EGFP of the mRuby+ subfractions are shown at 24 hrs, 48 hrs, and 72 hrs. Bar graphs are indicating mean values ± SEM, n = 3-6. Empty vector control was normalized to 1 ; ns, not significant. Figure 6C: Schematic overview of the procedure to validate the linear and circRNA content of the HEK293T cells post transfection using RNase R digestion followed by RT-qPCR quantification (primers indicated as red and green arrows). Outward directed circRNA-specific qPCR primers exclusively detect circRNAs. The EGFP primer set only leads to an amplification product after successful backsplicing. Figure 6D: The linear and circRNA content of the cells is shown after RNase R digestion relative to undigested samples after normalization to NupH mRNA for each construct. The enrichment of circRNA content over RNase R digestion indicates the successful backsplicing in cells (left). Relative circRNA / linear RNA levels increase upon RNase R digestion after normalization to NupH (right). Bar graphs are indicating mean values ± SEM, n = 2-3. In the case of n = 2 we do not show error bars. Figure 6E: We tested the inverse sequences of hHBB, EMCV, HCV, Hoxa9 and Hoxa5 as controls and calculated MFIs of EGFP of the mRuby+ subtractions 72 hrs. Bar graphs indicate mean values ± SEM, n = 3-6. Empty vector control was normalized to 1 . Figure 6F: Calculated MFIs of EGFP of the mRuby+ subtractions are shown at 24 hrs and 72 hrs. Mutants (M2, M5, M12, M13) are tested against the FL IRES-like Hoxa9 WT sequence. Bar graphs are indicating mean values ± SEM, n = 4-7. Empty vector control was normalized to 1. Source data are available online for this figure.

[0083] Figure 7 provides a schematic of an example of a method for detecting circRNA after plasmid transfection via RNase R digestion and RT-qPCR. Calculated mean fluorescence intensities (MFIs) of the mRuby+ subtractions are shown after normalization to the empty vector control in dependency of the tested insert sequences 120h post transfection. Bar graphs are indicating mean values ± SEM, n = 4. Experimental outline to proof the circular RNA content of the HEK293T cells, generated by spliceosome mediated backsplicing, after plasmid DNA transfection in order to validate the origin of the observed EGFP signal. Therefore, transfected cells were harvested 5 days post-transfection and FACS sorted according to their mRuby / EGFP signal. Afterwards, total RNA extraction was performed on the double positive cell fraction and subsequently digested with RNase R (1 U / pg RNA) for 30 min at 37°C (negative control was incubated with RNase R reaction buffer). Quantitative real time PCR was used for final circular RNA quantification (EGFP primer will only lead to a product of 94 nt length after successful back splicing). The 115 nt long mRuby product was used for linear pre-RNA quantification. The enrichment of circular RNA over the RNase R digestion is shown in Fig. 5D. Cell numbers of the EGFP+ cell fractions of the indicated samples shown for one representative experiment used for RNase R-qPCR.

[0084] Figures 8A-8E provide schematics and data results showing methods of synthesizing circRNA reporter constructs using a synthetic in vitro protocol. Figure 8A: Schematic showing design and synthesis of circRNA reporter constructs using T7 RNA polymerase in vitro transcription (IVT) and self-splicing ribozyme introns. Figure 8B: Schematic detailing circular RNA formation with group I catalytic introns from the phage T4 thymidylate synthase (td) gene. Figure 8C: Experimental outline of the circRNA reporter assay for assessing design and synthesis of circRNA reporter constructs generated by in vitro transcription (IVT) and self-splicing ribozyme introns. Figure 8D: Gel analysis of NIuc RNA before and after RNAase R treatment. Figure 8D: TapeStation chip analysis of NIuc RNA before and after RNAase R treatment.

[0085] Figure 9 provides schematics and data results showing expression from IVT synthesized circRNA reporter constructs when transfected into cells. On left is a schematic of experimental outline to assess IVT synthesized circRNA reporter constructs via luciferase expresssion. On right are data resutls of luciferase expression from IVT synthesized circRNA reporter constructs.DETAILED DESCRIPTION

[0086] Several embodiments of the disclosure are directed towards genetic expression systems for expressing a gene product. In many embodiments, a genetic expression systemcomprises a circular RNA (circRNA) construct. In some embodiments, a DNA-based construct that is transcribed and processed via cellular components (e.g., within cell or cell lysate or equivalent) to yield a circRNA construct. In some embodiments, a DNA-based construct that is transcribed and processed via in vitro transcription (IVT) and self-splicing ribozymes to yield a circRNA construct. In several embodiments, a circRNA construct comprises a sequence for encoding translation of gene product, which is in operable connection with an I RES-I ike sequence. When within a cell or a cell-free in vitro protein expression system, the IRES-like sequence can recruit ribosomes to translate the gene product, which can be achieved by cap-independent translation. In some implementations, a circRNA construct consists of only an IRES-like system and gene product sequence (inclusive of a stop codon). In some implementations, a circRNA construct comprises an IRES-like system and gene product sequence and further comprises other sequences that may be of benefit, such as RNA stability sequences, intron sequences, splice site sequences, self-spicing ribozyme intron sequences, and poly-A sequences.

[0087] In many embodiments, a circRNA construct comprises a eukaryotic IRES-like sequence. In several embodiments, the IRES-like sequence is derived from a 5’-UTR sequence of a Hoxa gene. IRES-like sequences that can be utilized can be derived from the following genes of mouse: Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, and Dlx1. IRES-like sequences that can be utilized can be derived from the following genes of human: H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDL1, and DLX1.

[0088] In several embodiments, systems and methods are for producing a gene product from a circRNA construct. In many embodiments, an expression system is used to express a gene product from a circRNA construct. The expression system can be a biological cell or a cell-free protein expression system. A cell-free protein expression system can be a cellular lysate or a buffered system and is to comprise at least the minimal components for translating RNA into polypeptides. In many embodiments, the expression system comprises the circRNA construct having a gene product sequence operably linked to an IRES-like sequence. In some embodiments, a circRNA construct is added to an expression system, which can be achieved by introducing the construct into a cell or a cell-free protein expression system. In some embodimenets, a circRNA construct transcribed from a DNA construct to yield the circRNA construct. In these embodiments, a DNA construct is added to an expression system, which can be achieved by introducing the construct into a cell or a cell-free protein expression system. Alternatively, a DNA construct can be integrated within a host’s DNA or maintained as an exosome or plasmid within a host such that the DNA construct is heritable to progeny during cell proliferation. Transcription of the circRNA construct can be performed within the cell or within acell-free RNA transcription system. A cell-free RNA transcription system can be a cellular lysate or a buffered system and is to comprise at least the minimal components for trancribing DNA into RNA. In some implementations, a cell-free RNA transcription system further comprises minimal components for processing expressed linear RNA into the circRNA construct. In some implementations, a cell-free translation system and a cell-free transcription system are combined or otherwise provided as a single system. The combined or single system is configured to assemble circRNA constructs from a DNA construct and to further express gene products from the circRNA construct. In some embodiments, a cell-free RNA transcription and protein expression system comprises minimal components for: transcribing DNA into RNA, processing transcribed linear RNA to yield a circRNA construct, and expressing a gene product from the circRNA construct, within a cell to produce a gene product encoded by the circRNA. In some embodiments, a circRNA construct is introduced into a cell-free protein translation system to produce a gene product encoded by the circRNA.

[0089] Many embodiments are directed towards applications and uses of circRNA constructs. Many new and emerging therapeutics and vaccines are based on mRNA constructs that express a gene product upon delivery into a recipient (see, e.g., H. Parhiz, et al., Lancet. 2024 Mar 23;403(10432): 1192-1204, the disclosure of which is hereby incorporated by reference). Accordingly, circRNA constructs with a Hoxa gene IRES-like sequences can be utilized to produce a gene within a recipient in order to provide a therapeutic and / or immunogenic effect. circRNA Constructs

[0090] Several embodiments of the disclosure are directed towards systems for gene product expression using circRNA constructs. Provided in Fig. 1A is an example of a circRNA construct comprising a gene sequence operably linked to an IRES-like sequence. The gene sequence can encode for a gene product and can be translated via the recruitment of ribosomes by the IRES- like sequence. Translation can be performed in a cap-independent manner. The sequence of the gene product includes a start codon (ATG) and stop codon (TAA, TAG, or TGA). Generally, the IRES sequence is upstream (i.e., 5’-) of the gene sequence. In some implementations, a Kozak sequence is provided at the start codon, which may enhance translation. Synthesis and / or manufacturing of the circRNA construct can be accomplished through any known or relevant method, including in vivo, in vitro, and chemical synthesis methods.

[0091] The circRNA can be single stranded and can comprise naturally occurring, modified, and / or synthetic nucleosides. Accordingly, the sugar moiety, nucleobase, or both can be naturally occurring, modified, and / or synthetic. Internucleoside linkages of the circRNA can be natural (e.g.,phosphodiester bond) or can be synthetic and / or modified. Examples of synthetic and / or modified internucloside linkages include (but are not limited to): phosphorothioate bonds, dimethylene sulfone bonds, A / -(2-aminoethyl) glycine bonds. Phosphorothioate bonds can reduce endonuclease activity.

[0092] The sugar moiety of a nucleoside generally comprises a furanosyl ring, which can include a hydroxyl or a substituent at the 2’, 4’, and / or 5’ positions. Examples of substituents include (but are not limited to): 2'-O-methyl, 2’-O-methoxy-ethyl, and 2’-fluoro. The sugar moiety can also be bicyclic (or bridged), such as configurations of sugar moieties of locked nucleic acids (LNA). LNAs typically comprise a bridge between 2’ oxygen and 4’ carbon, resulting in overall improved stability and thermodynamics.

[0093] The nucleobase of a nucleoside can be a common naturally occurring nucleobase utilized in RNA or DNA, such as adenine (A), cytosine (C), guanine (G), thymine (T), and uracil (U). The nucleobase can be a modified and / or synthetic. Examples of modified and / or synthetic nucleobases include (but are not limited to): pseudouridine, A / 1-methy-pseudouridine, 2- aminopropyladenine, 5-hydroxymethyl cytosine, xanthine, hypoxanthine, 2-aminoadenine, 6-N- methylguanine, 6-N-methyladenine, 2-propyladenine , 2-thiouracil, 2-thiothymine and 2- thiocytosine.

[0094] The I RES-like sequence of the circRNA can be sequence derived from a Hoxa gene or other gene having IRES-like functionality. As described herein, it has been found that untranslated portions of Hoxa mRNA and other developmental genes can recruit ribosomes via an IRES-like RNA sequence and structure element to perform cap-independent translation from circRNAs. Prior to this work, it was disputed whether Hoxa genes include IRES-like sequences in its 5’-UTR and could perform cap-independent translation. Similar disputes also arose for the ChrdH and Dlx1 genes. Results of experiments using circRNA constructs, however, show that these constructs having IRES-like sequences operably linked with a gene product sequence induce translation of the gene product. Notably, these IRES-like sequences have high conservation among mammalian species, and as such, Hoxa, ChrdH, and Dlx1 IRES-like sequences derived from a variety of a mammalian species can be utilized, including mouse and human. In various embodiments, a circRNA construct comprises an IRES-like sequence derived from mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdH, or Dlx1. In various embodiments, a circRNA construct comprises an IRES-like sequence derived from human H0XA3, H0XA4, H0XA5, H0XA9, H0XA 11, CHRDL1, or DLX1. It has been further found that Hoxa9 IRES-like sequence includes a minimal active element termed P4-native that is sufficient to provide the IRES-like function. The P4-native element includes a 35-nucleotide stem-loop that comprisesnucleotides critical for cap-independent translation. In some implementations, a circRNA construct comprises a mouse Hoxa I RES-like sequence is selected from SEQ ID NOs: 1-5. In some implementations, a circRNA construct comprises a human HOXA IRES-like sequence is selected from SEQ ID NOs: 6-9. In some implementations, a circRNA construct comprises a mouse ChrdH IRES-like sequence is selected from SEQ ID NO: 37. In some implementations, a circRNA construct comprises a mouse Dlx1 IRES-like sequence is selected from SEQ ID NO: 38. In some implementations, a circRNA construct comprises a human CHRDL1 IRES-like sequence is selected from SEQ ID NO: 39. In some implementations, a circRNA construct comprises a human DLX1 IRES-like sequence is selected from SEQ ID NO: 40. In some implementations, a circRNA construct comprises an IRES-like sequence having at least 50% homology to a sequence selected from SEQ ID Nos: 1-9 and 37-40. In some implementations, a circRNA construct comprises an IRES-like sequence having at least 60% homology to a sequence selected from SEQ ID Nos: 1-9 and 37-40. In some implementations, a circRNA construct comprises an IRES-like sequence having at least 70% homology to a sequence selected from SEQ ID Nos: 1-9 and 37-40. In some implementations, a circRNA construct comprises an IRES- like sequence having at least 80% homology to a sequence selected from SEQ ID Nos: 1-9 and 37-40. In some implementations, a circRNA construct comprises an IRES-like sequence having at least 90% homology to a sequence selected from SEQ ID Nos: 1-9 and 37-40. In some implementations, a circRNA construct comprises an IRES-like sequence having at least 95% homology to a sequence selected from SEQ ID Nos: 1-9 and 37-40. In some implementations, a circRNA construct comprises an IRES-like sequence having at least 99% homology to a sequence selected from SEQ ID Nos: 1-9 and 37-40.

[0095] The Hoxa and developmental gene IRES-like sequences were assessed for their ability to induce transgene expression from circRNA utilizing a backsplicing split-EGFP assay that only allows translation of the EGFP reporter up maturation of a pre-mRNA molecule into a circRNA construct. Similar to viral IRES sequences, the Hoxa, ChrdH, and Dlx1 IRES-like sequences induced EGFP expression from these circRNA constructs in human cells, validating their ability to initiate cap-independent translation. Further, circRNA constructs utilizing Hoxa, ChrdH, or Dlx1 IRES-like elements to induce transgene expression have several advantages over constructs that utilize viral IRES sequences. Many of the Hoxa IRES-like elements are shorter in length than viral IRES sequences, allowing for more compact sequences. Hoxa, ChrdH, and Dlx1 IRES-like elements lack stretches of poly-uridines that are common in viral IRES sequences, which have been demonstrated to induce immunogenic responses. Hoxa, ChrdH, and Dlx1 IRES- like elements further have fewer potential stop codons in their sequences as compared to viralIRES sequences, reducing the likelihood of premature translation termination. In fact, it has been documented that viral IRES sequences can have significant hurdles when utilized for therapeutic purposes (see, e.g., M. J. Unti and S. R. Jaffrey, Cell Chem Biol. 2024 Jan 18;31 (1):163-176.e5, the disclosure of which is incorporated herein by reference).

[0096] The gene sequences of the circRNA can encode any gene product to be translated, as appropriate to the application. Examples of gene products that can be expressed include (but are not limited to): therapeutic proteins or peptides, immunogens, cytotoxins, genetic editing tools, and molecular markers. The gene sequence can be codon optimized, which may yield higher translation. In some implementations, the gene product is a non-self gene product, meaning the gene product is different from the gene product naturally associated with IRES-like sequence. For example, when using a sequence from the 5’-UTR of mouse Hoxa9, the gene product is not mouse Hoxa9. In some implementations, the gene product is a heterologous gene product, meaning the gene product is from a different species than from the species from which the IRES- like sequences was derived.

[0097] Therapeutic proteins or peptides can be any gene product that provides therapeutic benefit. Examples of therapeutic proteins or peptides include (but are not limited to): hormones, growth factors, cytokines, chemokines, antigen-binding products, and blood factors. Examples of hormones include (but are not limited to): amylin, angiotensin, calcitonin, cholecystokinin, gastrin, ghrelin, glucagon, glucagon-like peptide, growth hormone, insulin, leptin, oxytocin, prolactin, somatostatin, thyroid-stimulating hormone, and vasopressin. Examples of growth factors include (but are not limited to): BDNF, BMP, EGF, EPO, FGF, IGF, neuregulin, neurotrophin, PDGF, TPO, and VEGF. Examples of cytokines include (but are not limited to): CSFs, ILs, TGFs, and TNFs. Examples of chemokines include (but are not limited to): CCLs and CXCLs. Examples of antigenbinding factors include antibodies, T-cell receptors, MHC molecules, and any gene product comprising an antigen binding region derived from an antibody, T-cell receptor, or MHC molecule. Examples of blood factors include (but not limited to): Factors I through XIII. Therapeutic proteins or peptides can also be utilized for replacement therapy, such as replacement of an enzyme or transcription factor in a congenital disorder.

[0098] Immunogens can be any gene product to stimulate an immune response, as would be beneficial in vaccination. Examples of immunogens include (but are not limited to) proteins or peptides derived from pathogens and cancer-related neoantigens. Examples of pathogens that can be targeted include (but are not limited to) coronavirus, influenza virus, dengue virus, Zika virus, EBV, CMV, HIV, HAV, HBV, HCV, HPV, HSV, RSV, VZV, norovirus, rhinovirus, rabies virus, tuberculosis, meningococcal tetanus, cholera, pneumococcal, and botulinum.

[0099] Cytotoxin can be any gene product to induce toxicity in a cell, including inducing programmed cell death, necrosis, and / or lysis of the cell. Examples of cytotoxins include (but are not limited to) TNF, granzymes, perforin, caspases, ricin, and cell-penetrating peptides. Cytotoxins can be utilized to induce toxicity in unwanted cells, such as neoplastic cells. Cytotoxins can be targeted to particular cell types via conjugation to a ligand such as an antigen-binding domain or other molecules having specific interaction with its target. Alternatively (or in addition), cytotoxins can be spatiotemporally controlled, which can be achieved utilizing (for example) photocleavable cages.

[0100] Genetic editing tools can be any gene product that provides a means for conferring editing of nucleic acids, especially chromosomal DNA. Examples of genetic editing tools include (but are not limited to): Cas nucleases of CRISPR / Cas systems, TALENs, and zinc-finger nucleases. Genetic editing tools may be combined with other components to confer the desired editing result. For example, Cas nucleases can be utilized with a guide RNA to such that gene editing is performed at a desired location on a DNA molecule.

[0101] Molecular markers can be any gene product that provides an indication of gene expression. Examples of molecular markers include (but are not limited to): fluorescent proteins, luciferases, and peroxidases. Molecular markers can be utilized in laboratory and preclinical techniques to study the ability to deliver and express gene products from circRNAs.

[0102] A circRNA construct can be generated from a DNA construct, which can be achieved by expressing the DNA construct within a cell, a cell lysate, or a buffered system comprising at least minimal components for RNA transcription and / or RNA processing. In some implementations, a DNA construct comprises sequences for gene product expression from a circRNA construct, such as (for example) a sequence encoding a gene product that is configured to be operatively linked to an IRES-like sequence. In some implementations, a DNA construct comprises a sequence for one or more elements involved in processing an expressed RNA transcript to yield a circRNA construct, such as (for example) introns, splice donor sequences, splice acceptor sequences, self-splicing ribozyme introns, and exon junction sequences. In some implementations, a DNA construct comprises one or more sequences for RNA transcription, such as (for example) an RNA polymerase promoter or poly-A signals. In some implementations, a DNA construct is configured to prevent functional gene product expression of the gene product that is to be operatively linked to the IRES-like sequence when constructed into a circRNA construct. Various configurations can be used to prevent functional gene product expression directly from the DNA construct. For example, the DNA construct can be configured to have the sequence encoding the gene product split into two portions. In some implementations, thesequence encoding the gene product is split into a 5’-gene sequence portion and a 3’-gene sequence portion. In some implementations, within the DNA construct, the 3’-gene sequence portion is upstream of the 5’-gene sequence portion. In another example, the DNA construct can be configured to have the sequence encoding the gene product upstream of the IRES-like sequence. In some implementations, when the DNA construct is configured to have the sequence encoding the gene product upstream of the IRES-like sequence, the DNA comprises an exon junction sequence with an ATG-start codon operatively linked to the IRES sequence (e.g., the exon junction sequence with ATG is positioned proximally downstream of the IRES-like sequence).

[0103] Provided in Fig. 1 B is an example of a DNA construct to yield a circRNA for expressing a gene product, which is processed via cellular splicing complexes. The construct is configured to express a linear mRNA transcript, can be expressed within a cell, a cellular lysate, or a buffered solution comprising at least the minimal components for RNA polymerase transcription. Upon transcription, the expressed linear construct is configured to be processed by cellular splicing complexes to yield the circRNA construct, which can be processed within a cell, a cellular lysate, or a buffered solution comprising minimal components for intron splicing.

[0104] The DNA construct of Fig. 1 B comprises sequences to express an mRNA transcript. The mRNA transcript can include a sequence for a gene product that is split into two portions, a 5’-gene portion (5’ Gene) and a 3’-gene portion (3’ Gene). The mRNA transcript further includes an IRES-like sequence (IRES) and stop codon (stop). The IRES-like sequence is upstream of the 5’-gene portion, and the stop codon is downstream the 3’-gene portion. The 3’-gene portion and stop codon, however, are upstream from the IRES and 5’-gene portion, which prevents translation of the gene product from RNA transcripts that are transcribed from the DNA construct. Flanking the IRES-like sequence and portioned gene sequences are two introns, a first intron that is upstream (5’ intron) and a second intron that is downstream (3’ intron). The 3’ intron comprises a splice donor (spD) and the 5’ intron comprises a splice acceptor (spA). To transcribe the linear mRNA transcript, the DNA construct includes an RNA polymerase promotor (Promoter) and poly- A signal (pA Sig) that are operably linked with the sequences of the mRNA transcript.

[0105] Upon transcription, cellular splicing complexes (e.g., spliceosomes) process the linear mRNA to yield the circRNA construct (Fig. 1 B). The cellular splicing complexes to remove introns. Utilizing the splice donor (spD) of the 3’ intron and the splice acceptor (spA) of the 5’ intron, the sequences of the 5’-gene product and the 3’-gene product are spliced together to yield the circRNA construct, which is functional. Upon ligation via the cellular splicing complexes, the circRNA construct comprises a sequence encoding a gene product that is operably linked withthe IRES-like sequence. As shown, the sequence encoding a gene product is proximate to and downstream of the IRES-like sequence.

[0106] Provided in Fig. 1 C is an example of a DNA construct to yield a circRNA for expressing a gene product, which is capable of being transcribed and processed in vitro (e.g., via IVT). Although the example is capable of yielding circRNA constructs in an in vitro setting, the construct can be used within any biological system for RNA transcription, such as within a cell, a cellular lysate, or or a buffered solution comprising at least the minimal components for RNA polymerase transcription. The DNA construct is configured to express a linear mRNA transcript, which can be expressed within an in vitro system, such as a buffered solution comprising minimal components for RNA polymerase transcription. Any IVT-capable RNA polymerase can be utilized, such as (for example) T7 RNA polymerase, T3 RNA polymerase, and SP6 RNA polymerase. The DNA construct comprises sequences encoding self-splicing ribozyme introns and exon junctions provided in a configuration to yield a circRNA construct. As transcription occurs, the self-spicing ribozyme introns are processed and removed and the junction sequences provide a means for ligation, generating the circRNA constructs. The self-spicing ribozyme introns can be Group I introns or Group II introns. If using Group I introns, the buffered solution (e.g., the IVT buffer) can comprise a free guanosine cofactor, which can be a guanosine nucleosides or nucleotide. Group II introns can self-splice without any cofactor. Group I and Group II introns can be found within genomes of species throughout the biological kingdoms. Further description and examples of Group I and Group II introns are provided within the following publications: R. M. O. D. S. Gomes, K. J. G. D Silva, and R. C. Theodore, Genet Mol Biol. 2024 Mar 25;47Suppl 1 (Suppl 1):e20230228; and A. M. Lambowitz and S. Zimmerly, Cold Spring Harb Perspect Biol. 2011 Aug 1 ;3(8):a003616; the disclosures of which are incorporated herein by reference.

[0107] The DNA construct of Fig. 1C comprises sequences to express an mRNA transcript. The mRNA transcript can include a sequence for a gene product (Gene). The mRNA transcript further includes an IRES-like sequence (IRES) and stop codon (stop), configured with the IRES- like sequence downstream the sequence for the gene product and stop codon. Downstream of and proximate to the IRES-like sequence is a first exon junction sequence (E1). Upstream of and proximate to the sequence encoding the gene product is a second exon junction sequence (E2). Flanking the E2-Gene-stop-IRES-E1 sequence IRES-like are two self-splicing ribozyme introns, a first self-splicing intron that is upstream (3’ S-S Ribozyme Intron) and a second self-splicing intron that is downstream (5’ S-S Ribozyme Intron). To transcribe the linear mRNA transcript, the DNA construct includes an RNA polymerase promoter (Prom) that is operably linked with the sequences of the mRNA transcript. The RNA polymerase promoter can be (for example) apromoter for T7 RNA polymerase, T3 RNA polymerase, or SP6 RNA polymerase. The DNA construct can further comprise overhang sequences that are complementary to one another, which can help promote the processing of the expressed RNA transcript (e.g., increasing the interactions the ribozymes and exon junction sequences).

[0108] As the RNA transcript is transcribed, the self-splicing ribozyme introns self-process, yielding the circRNA construct (Fig. 1C). First, the self-splicing intron that is downstream of the E1 exon junction sequences is processed by nucleophilic attack via a guanosine moiety (G), resulting in removal of the 5’ self-splicing ribozyme intron. The first self-splicing reaction results in a hydroxyl group at the 3’-end of the RNA transcript, which is then utilized as a nucleophile within the second self-splicing reaction, resulting in the removal the 3’ self-splicing ribozyme intron. The self-splicing reactions result in the joining of the 3’-end and 5’-end of the RNA transcript, which is assisted by the E1 and E2 exon junction sequences that provide complementary sequences to promote the ligation and circulation of the RNA transcript. Upon ligation, the circRNA construct comprises a sequence encoding a gene product that is operably linked with the IRES-like sequence. As shown, the sequence encoding a gene product is proximate to and downstream of the IRES-like sequence.

[0109] Various IRES-like sequences can be used within a DNA construct configured to yield a circRNA construct (e.g., the examples of DNA constructs protrayed within Figs. 1 B and 10). In various embodiments, a DNA construct configured to yield a circRNA construct comprises an IRES-like sequence derived from mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, Chrd / 1, or Dlx1. In various embodiments, a DNA construct configured to yield a circRNA construct comprises an IRES-like sequence derived from human H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDL1, or DLX1. In some implementations, a DNA construct comprises a mouse Hoxa IRES-like sequence is selected from SEQ ID NOs: 10-14. In some implementations, a DNA construct comprises a human HOXA IRES-like sequence is selected from SEQ ID NOs: 15-18. In some implementations, a DNA construct comprises a mouse Chrd / 1 IRES-like sequence is selected from SEQ ID NO: 41. In some implementations, a DNA construct comprises a mouse Dlx1 IRES- like sequence is selected from SEQ ID NO: 42. In some implementations, a DNA construct comprises a human CHRDL1 IRES-like sequence is selected from SEQ ID NO: 43. In some implementations, a DNA construct comprises a human DLX1 IRES-like sequence is selected from SEQ ID NO: 40. In some implementations, a DNA construct comprises an IRES-like sequence having at least 50% homology to a sequence selected from SEQ ID Nos: 10-18 and 41-44. In some implementations, a DNA construct comprises an IRES-like sequence having at least 60% homology to a sequence selected from SEQ ID Nos: 10-18 and 41-44. In some implementations,a DNA construct comprises an I RES-like sequence having at least 70% homology to a sequence selected from SEQ ID Nos: 10-18 and 41-44. In some implementations, a DNA construct comprises an IRES-like sequence having at least 80% homology to a sequence selected from SEQ ID Nos: 10-18 and 41-44. In some implementations, a DNA construct comprises an IRES- like sequence having at least 90% homology to a sequence selected from SEQ ID Nos: 10-18 and 41-44. In some implementations, a DNA construct comprises an IRES-like sequence having at least 95% homology to a sequence selected from SEQ ID Nos: 10-18 and 41-44. In some implementations, a DNA construct comprises an IRES-like sequence having at least 99% homology to a sequence selected from SEQ ID Nos: 10-18 and 41-44.

[0110] A DNA construct can comprise several other features to enhance and / or control expression of the linear mRNA transcript. The DNA construct can include transcription start sites, enhancers, conditional activators, and / or any other relevant genetic feature useful in the provision of gene expression, each of which can be operationally linked with the linear mRNA transcript. In some embodiments, the linear mRNA transcript provides a sequence for encoding an expression marker for detecting expression. In certain embodiments, a DNA construct may include additional features, such as an origin of replication, one or more splice sites, restriction enzyme cut sites, and other features for propagation, cloning, inserting, or molecularly tagging the linear mRNA transcript. Certain DNA constructs are circular (e.g., plasmid, BAC, etc.), while other nucleic acids are linear. Synthesis and / or manufacturing of the DNA construct can be accomplished through any known or relevant method, including in vivo, in vitro, and chemical synthesis methods.Gene-Product Expression Systems

[0111] Several embodiments of the disclosure are directed towards expression systems for producing gene products from a circRNA. In many embodiments, a circRNA construct is utilized for expression of a gene product. Any of the circRNA constructs or DNA constructs for generating a circRNA construct described herein can be utilized. Generally, a circRNA construct comprises a sequence encoding a gene product operably linked to an IRES-like sequence derived from a mouse Hoxa gene, ChrdH, or Dlx1, or human HOXA gene, CHRDL1, or DLX1.

[0112] In many embodiments, an expression system comprises a circRNA construct and the minimal components necessary for gene-product translation. Accordingly, in many embodiments, a gene expression system comprises a circRNA construct, ribosomes, tRNA, amino acids, and an energy source (e.g., ATP and / or GTP). Other components that can be included within a gene expression system can include initiation factors, release factors, cellular components for properprotein folding, cellular components for post-translational modification, and co-factors utilized by the gene-product.

[0113] In several embodiments, an expression system comprises a circRNA construct within a eukaryotic cell, which can naturally provide all components necessary for translation, protein folding, and post-translational modification. In some embodiments, an expression system is a cellular lysate (with or without outer membranes removed). Additional components can be added to the cell or cellular lysate, such as (for example) amino acids and energy sources, or any other component that can enhance translation. When utilizing a cell or cell-lysate, any type of eukaryotic cell can be utilized that is capable of translating a gene product utilizing an I RES-like sequence derived from a mouse Hoxa gene or human HOXA gene or similar. IRES-mediated translation has been observed in yeast, fungi, plant cells, and animal cells (including insect, worm, and mammalian cells). Hoxa genes are expressed in all bilaterian animals, and thus in some embodiments, a bilaterian animal cell (or lysate thereof) is utilized along with a circRNA construct as an expression system. In some embodiments, a mammalian cell is utilized. In some embodiments, a mouse cell is utilized. In some embodiments, a primate cell is utilized. In some embodiments, a human cell is utilized. A bilaterian animal cell can be an in vitro cell (e.g., grown in culture), an ex vivo cell (e.g., extracted from a live animal), or an in vivo cell (e.g., within a live animal).

[0114] In some embodiments, an expression system comprises a circRNA construct and minimal components necessary for gene-product translation within a cell-free buffered solution. The components for gene-product translation can be derived from any appropriate source, such as cellular lysates or extracted components. In some embodiments, the gene product is transgene product, meaning the gene product is not naturally expressed in the host cell.

[0115] Many embodiments are also directed to the use of expression systems that utilize DNA constructs to generate circRNA constructs. When a DNA construct is utilized, the expression system also comprises minimal components necessary for mRNA transcription and processing. Components for mRNA transcription can include an RNA polymerases, nucleotide-triphosphates, cellular components for splicing, and energy sources. In many embodiments, the eukaryotic cell, cell lysates, or cell-free buffered solution to be utilized for translation of the circRNA construct is also utilized for mRNA transcription such that transcription and translation are occurring simultaneously.

[0116] Gene products can be extracted, purified, or otherwise enriched after expression within an expression system. RNA constructs, DNA constructs, and / or gene products can be stored in solution, lyophilized and stored as a salt, or by any other appropriate means.Therapeutics and Vaccine Applications

[0117] Many embodiments of the disclosure are directed toward the use of circRNA constructs (or gene products thereof) as a therapeutic or vaccine. Generally, a circRNA construct can be delivered to a recipient such that a gene product can be produced in order to provide a therapeutic benefit and / or immunogenic response. A circRNA can be provided in a sterile solution and administered to the recipient via various routes, as appropriate for the treatment and / or vaccination. Recipients can be various animals that would be treated and / or vaccinated, including (but not limited to): humans, farm animals, zoo animals, domestic animals, and research animals.

[0118] In several embodiments, a pharmaceutical composition is for the administration of a therapeutic or a vaccine, the composition comprising a circRNA construct within a suitable pharmaceutically acceptable diluent or carrier. The diluent can be water, saline or a buffered solution. Examples of buffers that can be utilized include (but are not limited to) phosphate- buffered solution (PBS), tromethamine / tromethamine-HCI, and acetic acid / sodium acetate. The solution provided can be sterile or otherwise clinical grade for administration.

[0119] In many embodiments, pharmaceutical compositions further comprise one or more excipients. Examples of excipients that can be utilized with a circRNA construct include squalene, polyethylene glycol, lipids, SM-102, ALC-0315, triglycerides, diacyl glycerol, cholesterol, gelatin, phosphocholine, glucose, sucrose, lactose, cellulose, galactose, mannose, amylase, albumin, glycine, monosodium glutamate, magnesium stearate, sodium salt, potassium salt, aluminum salts, talc, silicic acid, viscous paraffin, hydroxymethylcellulose, and polyvinylpyrrolidone. Pharmaceutical compositions for vaccination may also include adjuvants to improve immune response. Examples of adjuvants include psuedouridine-modified RNA, lipid nanoparticles, ammonium lipids, mannan, and protamine peptides Compositions and methods for the formulation of pharmaceutical compositions will depend on a number of criteria, such as route of administration, systemic / local distribution, dosage, and storage / shelf-life.

[0120] Various routes of administration can be utilized, including (but not limited to) oral administration, buccal administration, sublingual administration, aerosol or nebulizer administration, nasal administration, transdermal administration, and administration by injection. Various mean of injection can be utilized, including (but not limited to) intravenous, intra-arterial, intrathecal, intralymphatic, subcutaneous, intramuscular, and intratumor.

[0121] In some embodiments, a pharmaceutical composition is administered in a therapeutically or immunogenically effective amount as part of a course of treatment. As used in this context, to "treat" means to ameliorate or prevent at least one symptom of the disorder to betreated or to provide a beneficial physiological effect (e.g., induce immunogenic sera). A therapeutically or immunogenically effective amount can be an amount sufficient to prevent, reduce, ameliorate or eliminate the symptoms of diseases or pathological conditions susceptible to such treatment.

[0122] Dosage, toxicity and therapeutic efficacy of a pharmaceutical composition can be determined, e.g., by standard pharmaceutical procedures in cell cultures or experimental animals, e.g., for determining the LD5o (the dose lethal to 50% of the population) and the ED5o (the dose therapeutically effective in 50% of the population). The dose ratio between toxic and therapeutic effects is the therapeutic index and it can be expressed as the ratio LD50 / ED50. Compounds that exhibit high therapeutic indices are preferred.

[0123] An "effective amount" is an amount sufficient to effect beneficial or desired results. For example, a therapeutic amount is one that achieves the desired therapeutic effect. This amount can be the same or different from a prophylactically effective amount, which is an amount necessary to prevent onset of disease or disease symptoms. An effective amount can be administered in one or more administrations, applications or dosages. A therapeutically effective amount of a composition depends on the composition selected. The compositions can be administered one from one or more times per day to one or more times per week; including once every other day. The skilled artisan will appreciate that certain factors may influence the dosage and timing required to effectively treat a subject, including but not limited to the severity of the disease or disorder, previous treatments, the general health and / or age of the subject, and other diseases present. Moreover, treatment of a subject with a therapeutically effective amount of a pharmaceutical composition described herein can include a single treatment or a series of treatments. For example, several divided doses may be administered daily, one dose, or cyclic administration of the compounds to achieve the desired therapeutic result. A single small molecule compound may be administered, or combinations of various small molecule compounds may also be administered.

[0124] Generally, dosage of a circRNA construct can be between about 1 mcg and about 1 mg, within a volume of about 0.05 ml_ to about 2 mL, which can be a therapeutically or immunogenically effective amount. In various embodiments, dosage of a circRNA construct is between about 1 mcg and about 50 mcg, between about 25 mcg and about 75 mcg, between about 50 mcg and about 100 mcg, between about 75 mcg and about 125 mcg, between about 100 mcg and about 150 mcg, between about 125 mcg and about 175 mcg, between about 150 mcg and about 200 mcg, between about 175 mcg and about 225 mcg, between about 200 mcg and about 250 mcg, between about 225 mcg and about 275 mcg, between about 250 mcg andabout 300 mcg, between about 275 mcg and about 325 mcg, between about 300 mcg and about350 mcg, between about 325 mcg and about 375 mcg, between about 350 mcg and about 400 mcg, between about 375 mcg and about 425 mcg, between about 400 mcg and about 450 mcg, between about 425 mcg and about 475 mcg, between about 450 mcg and about 500 mcg, between about 475 mcg and about 525 mcg, between about 500 mcg and about 550 mcg, between about 525 mcg and about 575 mcg, between about 550 mcg and about 600 mcg, between about 575 mcg and about 625 mcg, between about 600 mcg and about 650 mcg, between about 625 mcg and about 675 mcg, between about 650 mcg and about 700 mcg, between about 675 mcg and about 725 mcg, between about 700 mcg and about 750 mcg, between about 725 mcg and about 775 mcg, between about 750 mcg and about 800 mcg, between about 775 mcg and about 825 mcg, between about 800 mcg and about 850 mcg, between about 825 mcg and about 875 mcg, between about 850 mcg and about 900 mcg, between about 875 mcg and about 925 mcg, between about 900 mcg and about 950 mcg, between about 925 mcg and about 975 mcg, or between about 950 mcg and about 1 mg.

[0125] Although circRNA may be stable and produce gene products for up to several days after administration, dosing may be repeated as may be necessary to achieve the therapeutic or immunogenic result desired. In various embodiments, a dosage regimen comprises a single dose, two doses, three doses, four doses, five doses, six doses, seven doses, eight doses, nine doses, ten doses, or more than 10 doses. In various embodiments, dosage is repeated daily, every two days, every three days, every four days, every five days, every six days, every seven days, every ten days, every fourteen days, every 21 days, every 28 days, every two months, every three months, every four months, every five months, every six months, or annually. As can be understood, the precise dosing and dosing regimen can be determined as appropriate to the desired result an in accordance with preclinical and clinical data.

[0126] Preservatives and other additives, like antimicrobial, antioxidant, chelating agents, and inert gases, can also be present. (See generally, Remington's Pharmaceutical Sciences, 16th Edition, Mack, (1980), the disclosure of which is incorporated herein by reference.)EXAMPLES AND DATA

[0127] Experimental data generated and detailed herein provide support for the various embodiments of gene-product expression constructs and systems. In particular, the data establish that I RES-like sequences from Hoxa genes provide a means for cap-independent translation of genes utilizing circRNA constructs. Other experimental results are also described that further establish that the I RES-like sequences within Hoxa genes are sufficient for protein translation.This establishes the use of I RES-like sequences within circRNA constructs, which have value in various applications including therapeutics and vaccines.STUDY: Assessment of Hoxa IRES-like Function

[0128] Messenger RNA molecules have evolved to contain conserved functional regions that have distinct regulatory roles in post-transcriptional gene expression. The mRNA elements embedded in the untranslated regions (UTRs) attached to the protein coding sequence (CDS) harbor regulatory capacity to control mRNA translation and stability. One such category of regulatory RNA elements resides in the 5’ UTR of eukaryotic certain mRNAs are internal ribosome entry sites (IRESes) that facilitate alternate internal, cap-independent translation initiation, bypassing most required steps of conventional initiation. A strong body of work over decades has unraveled clear IRES activity of diverse structured RNA elements in viruses for host-dependent translation of the viral genome. Such viral IRESes are often employed to sustain translation of viral mRNAs upon host cell infection for which many require protein co-factors. These findings led to the fundamental question decades ago: Do such regulatory sequences and functional roles also exist in eukaryotes, and how can we find and characterize them?

[0129] Cellular IRES-like RNA elements represent eukaryotic mRNA elements that have IRES activity similar to viral IRES elements. They often mediate continued translation upon stimulation or stress (e.g., heat shock, nutrient deprivation, stress responses, cell proliferation, and apoptosis). For example, the first IRES in a cellular mRNA was discovered in the immunoglobulin heavy chain binding protein BiP, a stress-induced chaperone, by its continued activity in poliovirus-infected cells when bulk host translation was inhibited. Many IRES-containing viral transcripts even solely rely on the IRES element for translation. In contrast, mammalian mRNAs are capped and conventional cap-dependent translation initiation is very strong compared to most known cases of IRES-like elements in steady-state conditions which can be dynamic based on cell state. However, IRES-like elements have roles in a variety of biological functions in mammals even under conditions of ongoing cap-dependent translation, such as the translation of two different proteins from a naturally occurring single bicistronic transcript guiding organismal development. An unbiased screen across human genes suggests that IRES-like RNA elements may be present in ~10% of the genome. Therefore, it is reasonable that IRES-like regulation may contribute to widespread gene-specific translation in many mammalian mRNAs. In some cases, IRES-like elements may function as translation enhancers that modulate total protein levels. Of great interest is that IRES-like elements can make more direct contacts with the ribosome either by virtue of RNA-binding IRES trans-acting factors (ITAFs) or direct base pairing or tertiaryinteractions with ribosomal RNA. This may provide mechanisms by which the ribosome may have more regulatory roles in mRNA translation.

[0130] Such increased possibilities for interactions between IRES-like mRNA elements and the ribosome, given the complexity of translation regulation encoded in mammalian mRNAs renders important the discovery and determination of the contribution of mammalian IRES-like activity to overall protein levels. A key reason for our poor understanding of cellular IRES-like mRNA elements is the lack, limitations, and challenges of existing tools used so far for the validation and functional description of cellular IRES-like activity because many of these methods are prone to experimental artifacts and noise (See, e.g., R. J. Jackson, Cold Spring Harb Perspect Biol. 2013 Feb 1 ;5(2):a011569; and S. R. Thompson, Wiley Interdiscip Rev RNA. 2012 Sep- Oct;3(5):697-705; the disclosures of which are hereby incorporated by reference). Indeed, the caveats of the existing often reporter-based techniques led to false positive detection of IRES activity. This is particularly important with the use of bicistronic reporter constructs to detect IRES activity which are prone to artifacts. Particularly technically challenging is the distinction of the effect of an RNA element on transcription and / or translation as it may act on both processes. There is also no gold standard about which tests a cellular IRES needs to pass to determine genuine IRES-like activity.

[0131] Previous work has shown that the ribosome is recruited to what is potentially a structured 5’ UTR IRES-like mRNA element in a subset of Homeobox (Hox) mRNAs for ribosome- directed regulation of gene expression (S. Xue, et al., Nature. 2015 Jan 1 ;517(7532):33-8, the disclosure of which is hereby incorporated by reference in its entirety for all purposes). The transcription factors of the Hox cluster are body plan master regulators and are among the most spatiotemporally controlled transcripts. IRES-like elements have alleged to be involved with gene expression of several Hoxa mRNAs and anterior-posterior patterning of the mouse axial skeleton (S. Xue, 2015, cited supra; and N. Kondrashov, et al., Cell. 2011 Apr 29;145(3):383-397, the disclosure of which is hereby incorporated by reference in its entirety for all purposes). It has been disputed within the field whether these IRES-like elements do in fact provide IRES-dependent translational control in mammalian development (C. Akirtava, et al., Proc Natl Acad Sci U S A. 2022 Sep 6;119(36):e2122170119; and I. P. Ivanov, et al., Proc Natl Acad Sci U S A. 2022 Mar 1 ;119(9):e2117226119; the disclosures of which are provided for context). Because of the limitations to assess cap dependent gene translation, the field has not established an affirmative conclusion on whether Hoxa genes contain IRES-like elements.

[0132] A strong combinatorial approach allows for a comprehensive assessment of the contribution of IRES-like activity to mRNA translation. Herein, we present a diverse set ofmethodologies to investigate I RES-like activity, which provides assessment from several different angles and yield an affirmative conclusion. In this study, I RES-like elements are assessed via the following methodologies:1. Single-molecule fluorescent in situ hybridization (smFISH): Tissue-specific detection of I RES-like containing mRNAs in the embryo using I RES-like and CDS region-specific probes. This assay can assess and precisely quantify in which tissues a certain transcript isoform, and functional RNA elements therein, is present across physiological expression patterns in the mouse embryo.2. Pacbio long-read sequencing: Isoform detection of tissue-specifically expressed I RES- like containing mRNA variants.3. mRNA isoform quantification and translation level: qPCR-dependent absolute quantification of isoform abundance and isoform-sensitive polysome-qPCR detection for analysis of relative levels of translation across isoforms.4. Promoterless reporter assay: Sensitive detection of luciferase and mRNA levels from Nanoluciferase (Nluc)-£-g lobin fusion mRNA reporters to detect artificial cryptic promoter activity from DNA sequences encoding I RES-like elements.5. Circular RNA (circRNA) reporter system: Plasmid-derived RNA reporters encoding split-EGFP reporters generated by backsplicing that employ IRES-like elements for translation in cells and allow for mutagenesis analysis. Together, this versatile tool set will enable the field to confidently and universally assess IRES-like activity from eukaryotic mRNA elements.ResultsSingle-molecule mRNA imaging: Hoxa9 mRNA isoforms are locally expressed in specific embryo tissues

[0133] We here aim to visualize the quantitative tissue-specific expression pattern of the 5’ UTR that contains an IRES-like element compared to other isoforms in a tissue. We used the IRES-like elements of the mouse Hoxa cluster as examples of potentially having mammalian mRNA elements with IRES-like activity. In a prior study, a selective knock-out of a putative cellular IRES-like element within Hoxa9 was in mice led to a homeotic transformation and diminished HOXA9 protein expression in the somites and neural tube of developing embryos (S. Xue, 2015, cited supra). Importantly, due to a cap- repressive element at the 5’ end of several Hoxa 5’ UTRs, the putative IRES-like element was alleged to be capable of translation initiation for these mRNAs. Previously, transcript-level detection was limited to whole mount in situ hybridization (ISH)targeting almost the entire Hoxa9 CDS length in wildtype and homozygous Hoxa9 IRES-like knockout stage E11.5 embryos, which resulted in no change in mRNA expression. However, this approach was insensitive to different mRNA isoforms.

[0134] We use the RNAScope technique for smFISH assessment of Hoxa9 mRNA isoforms in sectioned embryos using mRNA small tiling probes that allows distinguishing mRNA isoforms with different 5’ UTR variants. smFISH allows for precise quantification of the expression and abundance of different Hoxa9 mRNA isoforms. The target-specific signal is amplified in intact cells and tissues of sectioned embryos. Single mRNA molecules can be imaged and precisely quantified as individual dots in the tissue. There are two full-length protein coding mRNA isoforms and a third isoform annotated for mouse Hoxa9 mRNA with different protein coding regions and 5’ UTRs. We here term them mRNA isoform A-C for convenience (A: long 5’ UTR, 1266 nt, complete CDS, ENSMUST00000048680; B: short 5’ UTR, 85 nt, short CDS, ENSMUST00000114425; C: short 5’ UTR, 85 nt, complete CDS, NM_010456.4) (Fig. 2A). Both A and C isoforms encode the full-length functional HOXA9 protein, while B isoform generates an early stop codon through alternative splicing, resulting in the lack of the C-terminal homeobox domain. The CDS probes (green) tile a shared region of the CDS between all isoforms and the IRES probes (red) tile the 323 nt-long IRES-like region in the long 5’ UTR isoform A. We can use co-localization of the probes to distinguish A and C isoforms at the single-molecule level: colocalized signal indicates the long 5’ UTR isoform, while the CDS-only signal indicates the other two shorter isoforms. In the sectioned E11 embryo, we focused on somites 20-28, as HOXA9 is strongly expressed caudally to somite level 21-22, which corresponds to the tenth thoracic vertebrae. Notably, a fully penetrant homeotic transformation of the thirteenth thoracic vertebra (T13) is evident in mice with a genetic deletion of the Hoxa9 IRES-like element (See S. Xue, 2015, cited supra). There is an overall clear tissue-specific distribution of the signal and mRNA staining with low background in the longitudinal section of wildtype E11 embryos for the long 5’ UTR mRNA isoform (Fig. 2B). We find that for Hoxa9 mRNA in longitudinal embryo sections, both the short and long 5’ UTR Hoxa9 mRNA isoforms are expressed. In both the neural tube and posterior somites (somites 20-28), the long 5’ UTR isoform A is predominantly expressed in nearly 80% of total Hoxa9 mRNA (Figs. 2C-2D). No IRES-only signal is expected to exist and is not observed (Fig. 2D). Before, with CDS-targeting probes used in whole-mount ISH, we were blind to these distinct isoform expression patterns and their overall abundances (See S. Xue, 2015, cited supra). These new data indicate that already at the transcription start site (TSS) level, a tissue-specific decision is made to predominantly employ the long mRNA isoform containing the IRES-like element in Hoxa9 mRNAs in somites and the neural tube that leads to translation of thehomeobox-containing HOXA9 transcription factor. smFISH with IRES-like-specific probes allows for distinct detection of localized IRES-like-containing mRNA isoform expression in their endogenous tissue environmentPacBio long-read sequencing: Hoxa9 mRNA isoforms are expressed in embryo tissues

[0135] We next sought to confirm the existence of the Hoxa9 mRNA isoforms described previously and the putative Hoxa9 / a10 fusion transcript (C. Akirtava, 2022, cited supra), that was purported to explain the existence of the IRES-containing 5’ UTR. We performed PacBio long- read sequencing from micro-dissected wildtype FVB / NJ E11.5 mouse embryo somite and neural tube tissues. PacBio HiFi long-read sequencing allows for accurate sequencing of full-length mRNAs and isoform identification. It can discover which possible mRNA isoform variants are expressed, but is limited in terms of accurate comparison of mRNA abundances across molecules with different lengths. Additionally, technical challenges that make data interpretation difficult, such as reverse transcription, library preparation, sequencing, alignment artifacts, and RNA degradation, can lead to false positive transcripts being identified. This is especially true for transcripts with identical intron chains or splice junctions, but different TSSs or transcription end sites (TES), which is the case for the TSS of isoform A and C. Given the complex problem of determining isoforms even from long-read sequenced samples at this time, transcript discovery and quantification were performed using the IsoQuant algorithm, a new gold standard for transcript annotation (Fig. 3A). We use the same PacBio kit as was previously used for generation of the public ENCODE long-read data set which was previously independently employed to claim no detection of the IRES-containing long Hoxa9 mRNA isoform and detection of a Hoxa9 / a10 fusion transcript by presenting raw reads. Importantly, in our data, we do not detect the Hoxa9 / a10 fusion transcript (Fig. 3A). Instead, we actually find unannotated transcript isoforms for HoxalO mRNA (Fig. 3A). With PacBio diffusion-based sequencing runs, shorter fragments will be sequenced much more than longer ones introducing strong technical bias towards shorter mRNA isoforms. Given our parameters using the GENCODE release M34 as gene annotation, we detect the reference 1266 nt-long 5’ UTR isoform A (ENSMUST00000048680) that encodes the IRES- like RNA element and the HOXA9 transcription factor, but we could not detect the short 5’ UTR Hoxa9 mRNA isoform C (NM_010456.4). We believe that this is due to the algorithm prioritizing isoforms A and B that are present in the gene annotation, while being unable to distinguish between isoforms A and C due to them having the same intron chain, only differing in their TSS. Nevertheless, we detect an annotated isoform B (ENSMUST00000114425) with a short 5’ UTR of 85 nts with a truncated CDS producing a protein lacking the homeodomain (Fig. 3A). Thus, weconfirm the existence of the annotated short and long Hoxa9 mRNA isoforms in micro-dissected E11.5 mouse embryo somite and neural tube tissues. We further elucidate that the long 5’ UTR Hoxa9 mRNA isoform A is not due to a Hoxa9 / a10 fusion transcript.RT-qPCR and polysome-qPCR: Hoxa9 mRNA isoforms are abundant and translatedin embryonic tissues

[0136] Given a 5’ UTR and mRNA isoform repertoire in which IRES-like elements may be included or not, it is important to determine if an IRES-like containing mRNA is translated or not. This can be achieved by polysome fractionation of tissue material and isoform-specific RT-qPCR analysis to determine to which extent an IRES-like-containing isoform is associated with translating ribosomes.

[0137] We next aimed at absolute quantification of mRNA isoforms of Hoxa9 in embryo tissues with primer sets that span different 5’ UTR and CDS regions (Fig. 3B). With this, we directly quantified the abundance of the Hoxa9 IRES-like-containing mRNA by reverse transcription (RT)-quantitative PCR (qPCR) on cDNA from micro-dissected E11.5 mouse embryo limb buds or neural tube / somites. Contrary to the previous E11.5 somite data (See I. P. Ivanov, 2022, cited supra) that reported an ~8 Ct cycle difference between IRES and CDS amplicons, we found that the CDS2 and IRES-like amplicons only differed by ~4 Ct cycles (Fig. 3C). This discrepancy could be due to differences in tissue harvesting techniques or RT, 3’ end-biases, or ability to process highly structured RNA elements. Furthermore, the CDS 5’ and 85 nt 5’ UTR amplicons, the latter representing the short 5’ UTR previously proposed to be the primary isoform (See C. Akirtava, 2022, cited supra), had similar Ct cycle counts to the IRES-like amplicons, suggesting that the CDS2 amplicon may overestimate Hoxa9 mRNA abundance or that the Hoxa9 mRNA isoform landscape may be more complex as can be resolved by tissue dissection and RT- qPCR. To absolutely quantify abundance and to account for primer efficiency differences, we cloned a Hoxa9 DNA template from E11.5 neural tube / somite cDNA using a primer set spanning all qPCR amplicons. Sequencing of this product showed correct splicing, further confirming that the IRES-like element is present in fully processed mRNAs. We used serial dilutions of this DNA template to create a standard curve in the qPCR (one dilution is shown in Fig. 3D). Accordingly, we determined that CDS2 is more than 2-fold as abundant as the CDS 5’ or 85 nt 5’ UTR amplicon, which are found to be at similar levels. From this we can conclude that the IRES-like element is at least 8-15% of the abundance of the CDS 5’ or 85 nt 5’ UTR amplicon.

[0138] Next, we performed sucrose gradient analysis of polysomes to examine the association of specific mRNA isoforms with ribosomes by RT-qPCR analysis on micro-dissectedE11.5 neural tube / somite samples (Fig. 3E), using the same primer sets as for isoform quantification (Fig. 3B). We find two clusters of mRNA isoforms in terms of their distribution across the sucrose gradient from E11.5 neural tube and somites (Fig. 3F). We find that IRES-like amplicons were present in polysomes, peaking in the light polysome (2-5 ribosomes) fraction, importantly indicating that IRES-like-containing mRNAs are not only expressed but also actively translated. The CDS and 85 nt 5’ UTR amplicons also peaked in the light polysomes, showing a more similar pattern to the IRES-like amplicons than to the highly translated housekeeping mRNA NupH with a longer CDS (1764 nt, NM_170591.1). The differences in 60S subunit / monosome abundance between the IRES-like amplicons and the CDS and 85 nt 5’ UTR amplicons may be indicative of different kinetics and modes of ribosome recruitment between Hoxa9 mRNA isoforms which will require further mechanistic experimentation to be unraveled. Overall, we find all detected Hoxa9 mRNA isoforms to be efficiently translated in embryonic tissues.Promoterless luciferase reporters: assessment of cryptic promoter activity from known IRES-like elements

[0139] Next, we sought to distinguish the activity of an IRES-like element in the 5’ UTR of a reporter gene on translation at the mRNA level from a possible simultaneous function on transcription at the DNA level. For this, we uncoupled the two processes by using promoterless luciferase-encoding reporter mRNAs into which IRES-like elements can be inserted upstream of the CDS and any artificial reporter expression due to promoter-like activity of the insert can be detected at the mRNA and luciferase activity level. To achieve robust detection of mRNA expression and luciferase activity from the same sample, we need sensitive reporters.

[0140] Previous studies have used the firefly luciferase (Flue) CDS fused to a rabbit -globin reporter gene generating a fusion protein, wherein the -globin gene contains an intron that can be used for reliable and quantitative distinction of mRNA expression from transfected plasmid in RT-qPCR analysis, especially for weak expression. We here aimed at improving the sensitivity of this reporter by using Nanoluc luciferase (NIuc) which is a much more sensitive and smaller enzyme than Flue, ideal to increase the detection limit of promoterless mRNA constructs. We fused the short NIuc ORF (621 nt) attached to an N-terminal 3xHA tag, to the intron-containing [3- globin gene (Fig. 4A). A cell will generate a Nluc-3-g / ob / n (NLB) fusion protein, instead of retaining the NIuc stop codon, in order to avoid creating a nonsense-mediated decay target triggered by a long 3’ UTR length. Another advantage of the NLB mRNA and fusion protein is a more sensitive, precise ratiometric quantitative luminescence measurement for comparing translation efficiencies on mRNAs through a co-transfected Flue control plasmid. The 5’ and 3’ UTRs from humanhemoglobin subunit beta (hHBB') mRNA serve as a reference for an efficiently expressed mammalian mRNA. When we compare simian virus 40 (SV40) promoter-driven expression of relative activity of the original NIuc and the NLB reporter enzymes in transiently transfected C3H10T 1 / 2 mouse embryonic mesenchymal cells, we detect a 70-fold decrease in overall relative NIuc luciferase activity with NLB (Fig. 4B). The structure prediction of the NLB fusion protein with the flexible 3xHA tag and interdomain linker using ESMFold reveals that NIuc and p-globin can adopt their native folds, with a less compact thus less stable fold of the NIuc p-barrel (Fig. 4C). Given that the luciferase activity of NIuc is tightly linked to substrate oxidation in its central cavity for luminescence, this effect on the native NIuc folding state may explain the reduced activity of NLB (Fig. 4B). Nonetheless, we aimed to leverage the advantage of easier and higher confidence mRNA normalization of the spliced reporter mRNA. We still retain sufficient sensitivity and dynamic range at lower activity levels of NLB compared to NIuc alone. Thus, we further employed NLB for our assays.

[0141] The distinction of a 5’ UTR element acting at the transcription and / or at the translation level is a difficult problem. In many instances it may be acting at both. A method to distinguish mRNA IRES-like activity in the cytoplasm from potential cryptic promoter-like activity of the corresponding DNA region in the nucleus can be assessed utilizing “promoterless” reporter plasmids. The rationale is that it cannot be excluded that at the DNA level, especially for conserved 5’ UTR regions, IRES-like-containing regions may harbor promoter-like activity which may lead to generation of capped mRNA isoforms with shorter 5’ UTRs. The IRES-like regions of the Hoxa9 gene are higher conserved than the rest of the 5’ UTR in the long isoform. Independently, genome-wide mouse RefSeq transcripts annotated with long, hyperconserved 5’ UTRs were experimentally confirmed to be expressed and IRES-like activity was allegedly detected for many of them, while cryptic promoter activity was indicated for a very small number. Bicistronic reporter mRNAs in which the second ORF is driven by the IRES is common to study IRES activity, but is plagued by the possibility of cryptic promoter activity and by the unusually long, uncommon bicistronic nature of the mRNA for mammalian cells.

[0142] In the Hoxa9 IRES-like RNA structure, we identified a short stem-loop termed P4 that directly recruits the 40S ribosomal subunit to the IRES-like element through mRNA-rRNA interactions with rRNA expansion segment 9S. P4 acts as a “translation enhancer” as it increases reporter mRNA translation when added to m7G-capped mRNA reporters. Additionally, simply removing any spacer sequence to the start codon abolishes P4 IRES-like activity in plasmid- derived experiments. To characterize its necessary features for translation enhancement, we introduced a series of 3-4 nucleotide (nt) mutations into P4. Overall, we found the 3’ arm of P4 tobe responsible for I RES-like activity and to bind to the 40S ribosome. We focused on the 4-nt mutation M5 in the P4 stem-loop 3’ arm which for example exhibited reduced binding of P4 RNA to human ES9S-edited 40S yeast ribosomes and caused a decrease in endogenous Hoxa9 mRNA translation efficiency compared to other Hoxa mRNAs in differentiated mouse ESCs, when introduced into the genomic Hoxa9 locus. The latter is a cytoplasmic effect on translation upon genome editing and transcriptional induction of Hoxa9 mRNA in respective M5-edited or WT cell clones. This allowed us to conclude that the M5 mutation decreases the translation efficiency of Hoxa9 mRNA in the cytoplasm independent from transcription. Nevertheless, it appears that there is the possibility of a 6-nt long putative E-box motif (consensus motif: 5’-CANNTG-3’), a rather degenerate sequence in the genome to which transcriptions factors can bind, that can map to the P4 3’ arm region at the DNA level (Figs. 4D&4F). It would overlap with previously characterized mutants M5 (inactive) and M2 (moderately active) (Fig. 4F-4G). We designed two new mutants M12 (2 nt) and M13 (3 nt) flanking each side of the putative E-box and tested these four P4 mutants in the context of the 323 nt-long full-length Hoxa9 IRES-like element (Figs. 4D-4E).

[0143] The rationale for promoterless plasmid constructs, however artificial this may be, is that in absence of a promoter that drives reporter mRNA expression from a defined transcription site, cryptic promoter-like activity can arise from DNA inserts upstream of the start codon. To create promoterless reporter mRNAs, we used the NLB fusion protein construct (Figs. 4A-4B), but deleted the upstream SV40 promoter in the plasmid (ASV40). All promoterless constructs contain the hHBB 3’ UTR. An empty or the hHBB 5’ UTR served as 5’ UTR controls and promotercontaining versions (SV40+) of the two served as positive controls for NLB expression. We included a series of Hoxa IRES-like elements of various lengths (101-323 nt), viral encephalomyocarditis virus (EMCV, 575 nt) and hepatitis C virus (HCV, 372 nt) IRES controls, as well as the P4 mutants in the full-length Hoxa9 IRES-like element (Fig. 4H). An / ? / 7BB-Fluc construct served a transfection control and we measured relative Nluc / Fluc activity and relative -globin reporter mRNA levels (Fig. 4I). Compared to the strongly transcribed empty and hHBB SV40+ controls we see diminished transcript production from ASV40 constructs for the empty, hHBB 5’ UTR, and EMCV IRES. However, we detect artificial promoter activity and corresponding mRNA abundance of up to 50% of the hHBB SV40+ control for the HCV IRES, Hoxa3, a5 and a9 IRES-like elements. This indicates the unspecific artificial promoter-like activity of DNA elements placed in otherwise promoterless plasmids. In absence of a promoter, longer stretches of DNA can be biased to be used as promoter sequences that otherwise would not be active in native chromatin. Indeed, it has been independently shown that the HCV IRES DNA sequence can have promoter activity in reporter mRNAs, which highlights the artificial nature of the assay. Incomparison to the WT Hoxa9 IRES-like element, M2, M5 and 13 mutations show decreased promoter activity while M12 is less affecting Hoxa9 IRES-like derived artificial promoter activity. M12 is only a 2 nt- mutation 3’ of the putative E-box mapped to P4 and less affects transcript production compared to the WT Hoxa9 IRES-like DNA. Next, we accounted for the relative luciferase activity corresponding to the relative luciferase transcript abundance (Fig. 4J). We find that overall, the luciferase activity measured from the corresponding samples agrees well with detected mRNA levels. We measure less luciferase expression in the case of Hoxa3 and Hoxa9 IRES-like elements. Random promoter activity and alternate transcription site choice in the 5’ UTR inserts may lead to transcripts that are initiated early or out of frame of the main NLB start codon. We conclude that in promoterless plasmids, it is more likely that any 5’ UTR region is favored to be used as a transcription start site. This seems plausible as the activity of transcriptional regulatory elements in 5’ UTR DNA regions is common. Overall, regulatory elements co-existing at the DNA and RNA-level in highly conserved genomic regions, and being repurposed in the nucleus and cytoplasm, may indicate the need for more intricate levels of gene expression of highly conserved genes. circRNA: Cellular IRES elements drive circRNA reporter translation

[0144] We first tested IRES activity of ten cellular 5’ UTR elements in a circRNA reporter. circRNAs have diverse cellular roles in health and disease, but only a few hard-to-identify functional circRNA-encoded mammalian proteins are known. There is an increasing interest in exploiting circRNAs for sustained translation of stable therapeutic biologies, for which important optimization steps have recently been achieved. circRNAs present a powerful tool to study capindependent translation. We adapted a plasmid-based circRNA system that uses cellular backsplicing of a reporter pre-mRNA into circRNA for the study of cellular IRES elements (Fig. 5A). In the split-EGFP reporter plasmid, a full-length (FL) EGFP is fused from an N- and C-terminal fragment by backsplicing mediated by ZKSCAN1 sites, with the IRES element upstream of the 5’ EGFP fragment. Such split-EGFP approaches have recently been invaluable for optimization of large gene delivery for muscular dystrophy therapy. In our case, FL EGFP is exclusively translated through ribosome-recruitment by the IRES. The linear reporter mRNA does not encode full EGFP, but two opposing fragments (Fig. 5A). Promoter activity from the insert would result in a linear capped mRNA that would only translate the 5’ EGFP fragment. The linear mRNA also encodes cap-initiated mRuby as a transfection control. We apply the circRNA reporter to test five recently identified ultraconserved 5’ UTR elements with IRES activity from a screen: ChrdH, Dlx1, Gdf5, Sema3a, Zfx. We also included five 5’ UTR elements with previously assigned cellular IRESelements: Bcl2, c-Myc, CACNA1A, Cofilin, and Fmr1. We included three viral IRES elements from encephalomyocarditis virus (EMCV), hepatitis C virus (HCV) and coxsackievirus B3 (CVB3) as references, as well as human hemoglobin subunit beta hHBB) 5’ UTR as a negative control (Fig. 5A). EMCV, HCV, CVB3 and hHBB were not tested as controls in previous studies, and we therefore have included them in our study. We transiently transfected human HEK293T cells and measured relative EGFP signal of the mRuby+ cells from 24 and 72 hours (hrs) post transfection (Fig. 5B). At 24 hrs, compared to the empty vector, we observe low background IRES activity for hHBB, no activity for EMCV and strong HCV and CVB3 IRES activity (Fig. 5B). The EMCV, HCV and CVB3 IRES mediate 0-1.4-fold, 4-5-fold and 12.5-13.6-fold higher IRES activity compared to the hHBB control at 24 and 72 hrs, respectively (Fig. 5B). We find that in this assay, EMCV follows slower kinetics than the other two viral IRESes and is only robustly active from 72 hrs on. We find that for the cellular IRES elements, at 24 and 72 hrs, ChrdU and Dlx1 are very active, and c-Myc and Bcl2 are moderately active, with 1.7-1.5-fold, 3.7-2.3-fold, 1.2-fold, and 1.4-1.5-fold higher IRES activity compared to hHBB, respectively. Cofilin and Fmr1 are active only at 72 hrs with a 1.4-fold and 1.2-fold increased IRES activity over hHBB', and Gdf5, Sema3a, Zfx, CACNA1A are inactive even at 72 hrs (Fig. 5B). Multiple upstream AUGs (uAUGs) in the insert sequence out-of- frame of EGFP, particularly for CACNA1A, may contribute to the observed absence of IRES activity. From the five cellular IRES elements from, we find that 2 / 5 are active in HEK cells. For all five IRES activity has been found to vary across different cell types (Byeon et al., 2021). It remains to be determined whether required cofactor ITAFs absent in HEK are needed for IRES activity regulation of the inactive constructs. When applied to various cell types, we think the circRNA assay is a great tool to also filter out false positive IRES activity possibly detected by orthogonal methods. To control for the specificity of circRNA translation dependent on insert length, RNA structure and GC-content, we tested the inverse sequences of all active IRES sequences from Fig. 5B (hHBB, EMCV, HCV, CVB3, Dlx1, ChrdU, Bcl2, c-Myc, Cofilin, Fmr1) compared to their forward sequence (Fig. 5C). We observe that all inverse inserts tested completely or nearly completely diminished IRES activity to the levels of the empty vector or hHBB (Fig. 5C). These data confirm that cellular or viral IRES elements can specifically recruit ribosomes to circRNAs for internal translation initiation. These findings reflect the usefulness of circRNAs to describe bona fide IRES elements from previous screen results and other studies, where this analysis has not yet been possible. circRNA: Hoxa IRES-like elements and mutagensesis in circRNA reporters

[0145] Previous work from the Barna lab and collaborators has shown that the ribosome is recruited to structured 5’ UTR IRES-like mRNA elements in a subset of Hoxa mRNAs for translation. The transcription factors of the Hox cluster are master regulators of body plan formation and are among the most spatiotemporally controlled transcripts. I ES-like elements are critical for gene expression of several Hoxa mRNAs and anterior-posterior patterning of the mouse axial skeleton as foundational examples of IRES-dependent translational control in mammalian development.

[0146] We tested a subset of Hoxa IRES-like elements, including the Hoxa9 FL IRES-like element and its shorter derivative P4-native in circRNAs (Fig. 6A). Upon transient transfection we measured the relative EGFP signal of mRuby+ cells after 24, 48, and 72 hrs (Fig. 6B). Compared to the controls, we see activity for all Hoxa-derived IRES-like elements to about half the levels of HCV. Hoxa9 IRES-like FL and P4-native display similar activity. Hoxa5 IRES-like FL drives the strongest activity of the Hoxa elements tested. To emphasize that the Hoxa9 IRES-like FL has half the activity of the strongest viral IRES tested (HCV), which reveals considerable activity, sustaining a comparable activity to HCV and EMCV at 120 hrs (Fig. 7). For Hoxa IRES-like elements, we also detect uAUGs in the insert sequences, but they do not prevent IRES-like activity. As EGFP accumulates over time, the EMCV IRES only shows activity in HEK293T cells with increasing incubation times over 72 hrs, but the relative activities between all the Hoxa IRES- like elements are similar at 48 and 72 hrs. To qualitatively show that the detected IRES-like activity is due to circRNA translation, we confirmed the existence of cytoplasmic circRNAs by RNase R- treatment of total RNA to degrade all RNAs except circRNAs. We FACS-sorted mRuby+ / EGFP+ cells at 5 days post transfection, and treated total RNA with RNase R for RT-qPCR analysis (Fig. 6C). Normalized mRNA levels after + / -RNase R indicate that linear mRNA is depleted in the hHBB, HCV, Hoxa9 and Hoxa5 samples, and circRNA species accumulate 8-38-fold for all four constructs (Fig. 6D, left, Fig. 7). The circ / linear RNA ratio indicates that the circRNA is enriched 4-14-fold for the four constructs (Fig. 6D, right). This RNase R-mediated enrichment of circRNA by degradation of linear species, including to an extent of the abundant reference mRNA NupH, shows the existence of circRNA irrespective of insert identity. Variability in the detected relative circRNA levels between inserts (Fig. 6D) may stem from the elaborate RNase R procedure (Fig. 7). circRNA detection confidently supports that the EGFP-signal is derived from IRES-like activity. Similar to other cellular IRES-like elements, the inverse insert controls tested for Hoxa5 and Hoxa9 specifically reduced IRES-like activity to that of the empty vector or hHBB (Fig. 6E and 7).

[0147] In the Hoxa9 IRES-like structure, we previously identified a 35 nt-long stem-loop termed P4 that recruits the 40S ribosomal subunit through mRNA-rRNA interactions with rRNAexpansion segment 9S as visualized by cryo-EM. P4 acts as a “translation enhancer” that can increase m7G-capped mRNA reporter translation. Through mutagenesis (Figs. 4D and 4E), we isolated a 4-nucleotide (nt) P4 mutation, termed M5, that selectively decreased ribosome binding and endogenous Hoxa9 mRNA translation. This allowed us to conclude that the M5 mutation decreases translation of Hoxa9 mRNA in the cytoplasm independent from transcription. Nevertheless, it appears that there is the possibility of a 6-nt-long putative E-box motif (consensus motif: 5’-CANNTG-3’). It is a rather degenerate sequence in the genome to which transcription factors may bind, that can be mapped to the P4 in the DNA (Fig. 4D). It would overlap with mutants M5 (inactive) and M2 (moderately active) (Fig. 4E). We designed two new mutants M12 (2 nt) and M13 (3 nt) flanking each side of the putative E-box. Particularly, we tested these four 2-4 nt P4 mutants in context of the 323 nt-long Hoxa9 I RES-like FL in circRNAs (Figs. 4D and 4E). At 24 and 72 hrs, we observe that M2, M5 and M13 are clearly inactive and display EGFP expression at the level of the empty vector and hHBB, while M12 is dampened, to half the activity of the wild type (WT) at 72 hrs (which is not significant) (Fig. 6F). As previous mutations changed 3-4 nt, mutating only 2 nt in M12 in the distal P4 may not be sufficient to affect overall Hoxa9 I RES-like activity. Indeed, a previous 3-nt mutation (M2) overlapping M12 also showed moderate activity (Figs. 4D and 4E).

[0148] Overall, our studies show that plasmid-derived split-reporter circRNAs are an elegant and sensitive system to investigate a range of I RES-like element activities and their mutagenesis, in any transfectable cell type. A major argument for circRNAs, and for their superiority over artifact-prone, bicistronic reporter mRNAs, is that promoter activity of IRES inserts as a source of false positive IRES activity can be excluded. We also provide appropriate controls and clearly confirm specific IRES activity from several cellular I RES-like elements, including Hoxa3, a5 and a9. Strong IRES activity found in circRNAs is also directly relevant for emerging synthetic mRNA medicine. circRNA: Ribozyme-mediated generation of synthetic circRNAs with cellular IRES sequences

[0149] We aimed to test the activity of viral and cellular IRESes in the UTR of in vitro synthesized circRNAs. This is different from a plasmid-based back-spliced split-EGFP circRNA reporter system that relies on plasmid transfection into cells where the cellular RNA processing and spliceosome machinery fuse two split fragments in a pre-mRNA by back-splicing mediated by ZKSCAN1 introns into a full-length (FL) EGFP ORF. The IRES resides upstream of the 5’ EGFP portion. This approach allowed for IRES-screening and -mutagenesis, and has been important for optimization of large gene delivery for the therapy of muscular dystrophy. As nuclear-generated circRNAs mimic physiological circRNA generation and require export to the cytoplasm, this approach is well suited to study endogenous functions and regulation of circRNA- and IRES- translation, but the reporter pre-mRNA is very inefficiently back-spliced.

[0150] Different from that, we here aimed to study the IRES activity of diverse IRESes in synthetic circRNAs eventually useful at scale for therapeutics (Fig. 8A). We evaluated specific IRES-performance in circRNAs in cells and in in vitro translation reactions. For this, we generated circRNAs in vitro that encode either the intact EGFP or NIuc sequence upstream of an IRES from an in vitro transcript (IVT) (Fig. 8A). This RNA with reporter-IRES topology is circularized by two permuted 5’- and 3’-flanking split group I catalytic introns from the phage T4 thymidylate synthase (td) gene in reverse order that together have self-splicing ribozyme activity (Fig. 8B). This ribozyme has previously been successfully used for circRNA synthesis (REF). Group I introns consist of two structural domains, P4-P6 and P3-P9, split into a 3’ and 5’ intron, that assemble at the RNA level through distally extended interactions to fold into an active ribozyme (Fig. 8B). Group I introns splice themselves out without assistance from the spliceosome or other proteins, and instead rely on magnesium and a free guanosine to initiate the splicing reaction. First, we generated a DNA template using an A-tail-T7 promoter-fused forward primer complementary to the 3’ intron and a T-tail-attached reverse primer annealing to the 3’ end of the 5’ intron (Fig. 8A). We realized that these A / T-extensions at the template ends improved the in vitro circularization efficiency during the IVT. We assume that the complementary distal ends help anneal the DNA ends in a circular manner which may bring the two td introns closer together for enhanced ribozyme cleavage, as observed before with RNA-encoded complementary ends. After selfsplicing, in the circRNA there remains a defined previously confirmed E1 / E2 exon-exon fusion site and scar sequence of here 58 nt, essential for group I intron activity, between the IRES and the ORF start codon (Fig. 8B). Despite relying on processing to a consistent IRES-ORF fusion site irrespective of IRES identity, any disadvantage of that is absorbed by placing the IRES downstream of the ORF. This architecture ensures that remaining uncapped linear RNA after circRNA purification cannot contribute to reporter expression. Additionally, to control for potential effects by IRES inserts of different length, structure, and GC-content on ribozyme efficiency, we pairwise test the forward and inverse IRES inserts in circRNAs (Fig. 8C).

[0151] We aimed to test diverse IRESes in circRNA reporters and carefully selected three viral and four cellular IRESes previously confirmed to have IRES activity in plasmid-derived circRNAs, as well as human hemoglobin subunit beta (hHBB) 5’ UTR as a negative control (Fig. 8C). We included the viral IRESes of hepatitis C virus (HCV), coxsackievirus B3 (CVB3). Among the four cellular IRESes with previously confirmed cellular IRES activity, ChrdU and Dlx1 wererecently identified as ultraconserved 5’ UTR elements in a screen for cell type-specific IRES- mediated initiation in bicistronic mRNA reporters, which were among the strongest cellular IRESes in plasmid-based circRNAs. These IRESes are important for mouse embryonic development and impact the translation efficiency of their endogenous mRNAs. Dlx1 acts as a homeobox transcription factor with critical roles in craniofacial patterning, and neuron differentiation and survival in the brain. We also include the structured Hoxa9 and Hoxa55’ UTR “I RES-like” mRNA elements from the Homeobox (Hox) gene cluster of transcription factors, with previously assigned cellular IRES activity in linear reporters and confirmed in plasmid-derived circRNAs. Hoxa I RES-like elements were initially found to mediate IRES-dependent, ribosome- directed translation regulation critical for anterior-posterior patterning of the mouse embryo axial skeleton. The Hoxa9 IRES-like element acts as a translation enhancer by directly recruiting 40S ribosomal subunits through mRNA-rRNA interactions with rRNA expansion segment 9S (ES9S) for translation initiation.

[0152] Before application of circRNAs in experiments, we enriched them after IVT. Quality control gel electrophoresis for these IRESes after IVT showed that the linear mRNA is the predominant RNA species in the IVT compared to the circRNA for NIuc (Fig. 8D). To degrade all RNAs except circRNAs, we digest the IVT with the 3’-to-5’ exonuclease RNase R which only cleaves linear RNA species with accessible ends (see Fig. 8B), leaving circular forms intact. RNAse R treatment and column purification of the remaining RNA removes the majority of linear RNA, visible as an enriched single circRNA band, independent of insert identity, on formaldehyde gels (Fig. 8D) or TapeStation chips (Fig. 8E) comparing RNA before and after RNase R digestion. For some constructs, we detect remaining RNase R-resistant spliced-out td introns as stable side products of low molecular weight and residual linear mRNA. For Dlx1, clear enrichment of circRNAs required 90 min RNase R digestion instead of 30 min (Fig. 8D). Reporter ORF translation is exclusively derived from bona fide IRES activity as remaining traces of linear mRNA species are uncapped and will be unstable in cells and are expected to be poorly initiated both in cells and in translation extracts. With an IVT-based circRNA system external to the cell, there is no concern about cryptic artificial promoter activity from the IRES insert in a plasmid, that may contribute to translation through generation of linear capped EGFP or NIuc mRNAs.

[0153] We next apply the purified circRNAs containing different IRESes and reporters to either in-cell transfection or to in vitro translation reactions. For in-cell assays (Fig. 9), we transfect purified circRNAs into HEK293T cells and harvest the cells 24 hrs post transfection (hpt). For EGFP circRNAs we FACS-sort EGFP+ cells and quantify the EGFP+ signal. For NIuc circRNAs,we co-transfect circRNAs with a linear capped / ? / 7BB-Fluc reference mRNA, lyse the cells, and quantify relative Nluc / Fluc luminescence.Conclusions

[0154] We present an extensive spread of different techniques to critically investigate cellular IRES-like elements applied to the Hoxa9 IRES-like element example. We find support for it to be expressed as part of a long 5’ UTR isoform mRNA A that is visualized by smFISH imaging, translated in mouse embryo tissues, and drives IRES-like initiation capacity in circRNA reporters. We envision that this roadmap is widely transferable to IRES-like elements from any species. Wth regard to the translation profile of Hoxa9 mRNA isoforms by sucrose gradient fractionation, from combined neural tube and somite tissues by RT-qPCR, all isoforms with the long and the short Hoxa9 5’ UTR are detected in contrast to what we see with smFISH isoform imaging (Figs. 2A-2C; 80% long compared to 20% short 5’ UTR in smFISH). However, smFISH is more tissuespecific than micro-dissection of the embryo for tissue isolation which can never be complete and thus cannot clearly distinguish between tissues to the level of precision an imaging approach can. An alternative explanation is that there may be a roadblock, for example a chemical modification or structure, in the Hoxa9 mRNA that prevents efficient cDNA synthesis but not hybridization of the RNAscope probes.

[0155] It has previously been suggested that several Hox mRNA 5’ UTRs can have promoter activity in plasmid-derived reporters in transient transfection-based cell culture assays. The importance of this transcriptional regulation remains to be determined in the in vivo developmental context of an organism. In the case of Hoxa9 mRNA expression, in vivo evidence exists only at the level of translation. Importantly, even if respective DNA sequences corresponding to 5’ UTR regions have functionally significant promoter activity in the nucleus, it does not preclude them from co-existing functionally at the DNA and RNA level and also serving as mRNA regulatory elements in translation. The question remains whether these sequences are so highly conserved due to the selective pressure of maintaining these putative functions or due to hyper-efficient repair of any mutations in these regions. Additionally, it is critical to point out that the ability to predict reporter assay activities from genomic features, while an important result, is not actually inconsistent with their function as translational regulatory elements. The presence of transcriptional regulatory elements in 5’ UTR regions and their activity at the DNA level are common and unsurprising. Neither are alternative TSS / 5’ UTRs that result from the presence of such regulation. It would be extraordinary to expect DNA- and RNA-level regulatory sequences to evolve in mutual exclusivity. On the contrary, transcriptional and translational effects may co-vary, potentially with observing greater levels of conservation. Thus, it is interesting but not surprising to discover that the conserved 5’ UTR regions are enriched in other regulatory elements such as transcriptional regulatory elements, especially in the most conserved ones.

[0156] I ES-like elements in the genome may provide new layers of regulatory specificity to gene expression. Similar to transcription where c / s-acting promoters or enhancers are guided in trans by DNA binding factors, I ES-like elements may provide the basis for more directed and regulatory translational control by the ribosome or additional RNA binding factors. This may diversify the expression of the genome in time and space. As the field is turning to more genomewide methods to identify I RES-like elements, this can provide an entry point for in-depth characterization of individual examples. By having a tool set of technologies to dive more in-depth into the characterization of select 5’ UTRs, this will broaden our understanding of the function of such elements. Of importance, circRNA reporter assays have the advantage that only strong IRES-like activity will be detected while cell-type specificity can be included. Particularly, strong IRES-like elements found through this assay are directly relevant for synthetic biology and recently emerging mRNA therapeutic applications. Meanwhile, at the organismal level, smFISH is capable of detecting selective mRNA isoforms with varying 5’ UTRs in embryo sections. Together, the roadmap presented in this study provides a systematic methodology for characterizing 5’ UTR mRNA isoforms and their ability to be entry points for ribosome-mediated control of gene expression.Experimental MethodsCell culture and transfection

[0157] Mouse C3H / 10T1 / 2 (ATCC: CCL-226) cells or human HEK393T (ATCC: CRL-3216) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco, 11965-118) containing 2 mM L- glutamine, supplemented with 10% fetal calf serum (Gibco, ES009-B), 100 U / ml penicillin and 0.1 mg / ml streptomycin (EMD Millipore, TMS-AB2-C or Gibco, 15140-122) at 37°C in 5% CO2- buffered incubators.

[0158] For luciferase assays, ~0.6 X 106 C3H / 10T1 / 2 cells were seeded per well in 12-well dishes and transfected the following day with total 2 pg of plasmid (1 pg NIuc and 1 pg Flue control plasmid) using 4 pL Lipofectamine 2000 (Invitrogen, 11668-019) and Opti-MEM (Gibco, 11058- 021) according to the manufacturer’s instructions in serum-free and antibiotic-free DMEM. For transfection with monocistronic NIuc constructs (pGL3-Nluc or pGL3-NLB), 1 pg of a HBB-Fluc control plasmid (pGL3-HBB-Fluc; pKL039) was co-transfected per well. The medium waschanged to regular DM EM 4-6 hours after transfection and cells were collected 24 hrs posttransfection and samples were split in half for protein and mRNA analysis.

[0159] For the mRuby-ZKSCAN-spEGFP plasmid reporter assay and transfection of circRNA plasmid into into HEK393T cells, -0.08 X 106 cells were seeded per well in 24-well dishes 16 hrs before transfection (total volume 500 pl / well). After overnight incubation at 37°C in 5% CO2 the medium was replaced by 450 pL antibiotic free DM EM and cells were transfected with 1 pg of plasmid DNA using 1 pL Lipofectamine 3000 combined with 1 pL P3000 (Invitrogen, L3000001) and Opti-MEM (Gibco, 11058-021) according to the manufacturer’s instructions in serum-free and antibiotic-free DM EM. The medium was changed to 1 ml_ regular DM EM 6 hrs after transfection and cells were collected 24, 48, or 72 hrs post transfection. In order to obtain further information about the reaction kinetics over time and to detect even weak I RES-like activity, the incubation times were varied between 24, 48 and 72 hrs. For testing these later time points, the number of seeded cells on day 0 were adapted to 16,000 cells and an additional media change was performed 48 hrs post transfection for the 72 hrs test series. For verification of circRNA content of the transfected HEK293T cells, 1.6 X 106 cells HEK293T cells were seeded 16 hrs before transfection in 15 cm cell culture dishes. Cells (covered with 8 mL antibiotic free DMEM) were transfected with 45 pg of plasmid DNA using 180 pL Lipofectamine 2000 (Invitrogen, 11668-019) according to the manufacturer’s instructions. 24 hrs post transfection, the media was replaced with 25 mL of full cell culture media (DMEM supplemented with 10% FBS, 1% penicillinstreptomycin and 1% glutamine). After 5 days of incubation, cells were harvested.Mice

[0160] Mice were housed under a 12 hrs light / dark cycle with free access to food and water. FVB / NJ mice (JAX, 001800) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA) and used as wildtype. Pregnant females were euthanized at E11.5, the uterus was dissected and embryos were taken out and placed into 10% fetal bovine serum (FBS) in DMEM / F12 without phenol red (Gibco, 21041025). Embryos were individually micro-dissected to isolate somites and neural tube or limb buds, respectively, as described in(23). For RNA extraction and sucrose gradient fractionation, the tissues were dissociated with 1% trypsin (Gibco 27250018) in Hanks’ Balanced Salt Solution (HBSS, Thermo Fisher, 14025-076) at 37°C prior to lysis. For RNAScope, E11.5 wildtype embryos were sectioned longitudinally as described separately. For PacBio long- read sequencing, E11 .5 wildtype embryos were micro-dissected and total RNA was extracted from tissues. All animal work was performed in accordance with protocols approved by Stanford University’s Administrative Panel on Laboratory Animal Care.Plasmid Construction

[0161] The following plasmids have been described previously: mRuby3-ZK-spEGFP and mRuby3-ZK- spEGFP-IRES2(EMCV) (C. K. Chen, et al., 2021 , cited supra). For NIuc and NLB luciferase reporter constructs, the following plasmids were used: pGL3- FLB-fusion-HBB (pKL082, this study), encoding for a monocistronic firefly luciferase (Flue) fused to human hemoglobin (hHBB) 5’ UTR and -globin (exons 1-3, including exon 2,3 intron) reporter gene fused at its 3’ end. To generate the pGL3-NLB-fusion-HBB5’ (pMH001) plasmid that encodes a NIuc reporter gene fused to a rabbit fi-globin reporter gene (intron included) resulting in a Nluc / p- globin fusion protein termined NLB, by deleting the stop codon of the NIuc ORF. For this, first pGL3-FLB-fusion-HBB (pKL082) was digested with EcoRV / Van91l (Thermo Scientific, FD0303, FD0714), creating a linearized fi-globin encoding plasmid, with non-overlapping sticky ends excising the Flue reporter gene. After digestion, the digested backbone plasmid was purified by agarose gel extraction according to manufacturer’s instructions (NEB, T1020S). Using primers encoding for a fi-globin 5’-overlap, two NIuc gene fragments were amplified from pcDNA3.1- 5'UTR-3xHA-NLuc (pKL401) template and fused by overlap PCR. The EcoRV / Van97 / -digested pKL082 was then ligated with the NIuc insert downstream of the 5’ hHBB and upstream of the [3- globin reporter gene via Gibson assembly (NEB, E2621S), generating pGL3-NLB-fusion-HBB5’ (pMH001). The second construct is the pGL3-Nluc-HBB5’ (pMH002) plasmid for which pKL082 was digested with EcoRV / Xbal (Thermo Scientific, FD0303, FD0684). The linearized plasmid (excising the Flue reporter gene) was purified by agarose gel extraction. A NIuc insert was amplified from pKL401 using oligonucleotides with corresponding Gibson-overlap sequences for the EcoRV / Xbal-digested pKL082 backbone. The NIuc insert was purified via agarose gel extraction. Gibson assembly of the insert with pKL082 generated plasmid pMH002, which contains no -globin reporter gene after NIuc. For generating pGL3-NLB-hHBB3'-HBB5' (pMH023), the hHBB 3’ UTR was amplified with primers M H 008 / M H 009 and Gibson-inserted into Agel / Xbal (Thermo Scientific, FD1464, FD0684) digested pGL3-NLB-fusion-HBB5’ (pMH001).

[0162] In order to generate the series of promoterless pGL3-NLB plasmids without the SV40 promoter containing candidate 5’ UTRs which were analyzed for their potential promoter-like activity, we generate a SV40A version by cutting out the SV40 promoter from pMH023 that contains the hHBB 5’ UTR by BgllllHindlll-digest and Gibson assembly using the NEBuilder HiFi DNA Assembly Master Mix (NEB, E2621S) of the annealed oligo 1 LL042 / LL043 to seal the ends and generate pLL015. Second, to generate an SV40A empty vector, pMH023 is Bglll / EcoRV- digested and followed by Gibson insertion of the annealed oligo 2 LL044 / LL045 to seal the endsand generate pLL016. The IRES-like 5’ UTR sequences to be tested (including hHBB, HCV, EMCV, Hoxa9, Hoxa3 and Hoxa5) were PCR-amplified using the AccuPrime Pfx DNA Polymerase (Thermo, Invitrogen, 1876525) and the primers LL046-LL057 and LL68 / LL69 cloned into the Bglll-site of pl_L016 to generate SV40A-versions of pGL3-NLB-hHBB3’ in the series of plasmids pLL015 and pLL017-pl_L022. For SV40+ positive control empty vector, pMH023 is digested with Hindlll / EcoRV to cut out the hHBB 5’ UTR and followed by Gibson insertion of the annealed oligo 3 LL064 / LL065 seal the ends and generate pLL026. To generate mutants in the context of the full-length Hoxa9 IRES-like element, the mutated versions of the sequences (Hoxa9-M2l M5 / M12 / M13) were generated by overlap stitch PCR using the primers LL056 / LL057 and LL070-077 and pl_L022 as template, and inserted using Gibson assembly into the Bglll / EcoRV-digested pMH023 plasmid to generate plasmids pLL036-pl_L039.

[0163] To generate ZKSCAN-split EGFP plasmid constructs for circRNA reporter assays, the sequences which were analyzed for their potential IRES-like activity (including hHBB, HCV, Hoxa9, P4-native, Hoxa3 and Hoxa5) were amplified by PCR using the AccuPrime Pfx DNA Polymerase (Thermo, Invitrogen, 1876525) with primers LL017-LL028 and cloned into the EcoRV cloning side of the mRuby-ZKSCAN-spEGFP-EcoRV (pKL477) plasmid (see also(52)) by Gibson Assembly using the Gibson HiFi DNA Assembly Kit (NEB, E2621S) according to manufacturer’s instructions. The mutated versions of the tested sample sequences (Hoxa9-M2 / M5 / M12 / M13) were generated by PCR amplification using primer LL027 / LL028 and LL070-LL077 and cloned into the EcoRV-cloning side of the mRuby-ZKSCAN-spEGFP-EcoRV (pKL477) plasmid as described before to generate plasmids pLL040-pl_L043.

[0164] After transformation of NEB 5-alpha competent E. coll (High Efficiency) bacteria (NEB, C2987I) and following Midiprep (QIAGEN Plasmid Plus Midi Kit, 12945), mutations, cloning boundaries and coding sequences in all plasmids were verified by DNA sequencing (Eurofins Genomics or Microsynth).Mouse Embryo smFISH and Image Analysis

[0165] The E11 C57 mouse embryo sagittal frozen sections (Zyagen, MF-104-11-C57) were used as standard fresh frozen sections. The RNAScope probes were custom designed and purchased from ACDBio. The probe targeting the Hoxa9 mRNA CDS region is termed Mm- Hoxa9-O6 and targets the 1267-1892 region of NM_010456.3 excluding the short isoform-specific intron region 1581-1753. The probe targeting the IRES-like region is termed Mm-Hoxa9-O5 and targets the 776-1142 region of NM_010456.3. Mm-Hoxa9-O5 and Mm-Hoxa9-06 probes were directly ordered from ACDBio. smFISH was conducted according to the manufacturer’sinstructions. Confocal images were acquired with a custom-built inverted spinning disk microscope. A Zeiss Axio Observer Z1 microscope was coupled to the Perkin Elmer UltraVIEW Vox spinning disk confocal microscopy system with an encoded ASI MS2000 motorized piezo stage equipped with Plan-Apochromat 10x / 0.45 and 63x / 1 .20 objectives (Carl Zeiss microscopy). The 405-nm, 488-nm and 561-nm solid-state laser lines were paired with the 452W25, 510W20, and 588W21 Semrock emission filters to minimize crosstalk between different fluorophores. Images were acquired with the Velocity Acquisition suite 6.3 (Perkin Elmer). All acquired high- resolution images of the posterior somites (somite number 20-28) and neural tube (the boxed region as depicted in the main figure) were analyzed using custom Python scripts. In brief, local fluorescence intensity thresholding was applied to call individual fluorescence puncta. Then for each punctum from the CDS-probe channel, if it partially overlaps with a punctum from the IRES- probe channel, it is classified as a long 5’ UTR Hoxa9 mRNA isoform; otherwise it is classified as a short 5’ UTR isoform. The puncta count from multiple images in the same embryo were then pooled together and used to quantify the respective isoform ratio.PacBio Long-read Sequencing, Library Generation and Analysis

[0166] Neural tube / somites were micro-dissected from E11.5 FVB / NJ mouse embryos. The uterus was washed in 1xPBS twice to remove blood. Each embryo was transferred into a plate containing filming media one by one, where micro-dissection was performed to take out the tissues. Filming media (10% FBS (Sigma-Aldrich, TMS-013-B) in DMEM / F-12 with HEPES without phenol red (Gibco, 11039021)) was used during micro-dissection. Excess media was removed with a pipette without centrifugation, and the sample was then flash frozen in liquid nitrogen. Tissue was then rapidly ground by hand for 30 seconds on ice with a disposable RNase free microcentrifuge pestle (Fisher Scientific, 12-141-364). Immediately after, 1 mL of TRIzol reagent was added (Thermo Fisher, 15596026). The sample was lysed and mixed by vigorously inverting the tube several times, and then had the total RNA extracted following the manufacturer’s protocol, using a cleanup column (Zymo Research, R1013) instead of ethanol precipitation. Subsequently, the RNA samples were treated with TurboDNase (Thermo Fisher, AM2238) at 37°C for 30 minutes according to the manufacturer protocol and cleaned up using a column.

[0167] The Iso-seq library was prepared using SMRTbell prep kit 3.0 (PacBio, 102-182-700) following the kit’s instructions with some added details below. For the section “Purification of amplified cDNA with SMRTbell cleanup beads”, 84 L of SMRTbell cleanup beads were added to the sample. Prepared library was then sequenced with PacBio Sequel II by GenomeSequencing Service Center at Stanford University, and also processed to obtain HiFi reads. Reference genome and comprehensive gene annotation on the reference chromosomes, scaffolds, assembly patches, and haplotypes were obtained from GENCODE release M32 (GRCm39). HiFi reads were aligned to the genome using minimap2 with default settings. Isoform discovery was performed using IsoQuant with default settings. Transcript model covering Hoxa9 and HoxalO (chr6:52,200,050-52,217,850) was visualized using IGV.Mouse Embryo RNA Isolation and RT-qPCR

[0168] All mouse work was reviewed and approved by the Stanford Administrative Panel on Laboratory Animal Care (APLAC). FVB / NJ mice (JAX, 001800) were housed in standard conditions with 12 hrs light / dark cycles, ambient temperatures between 20-26°C, and humidity between 30 and 70%. To generate embryos, a male was housed with one or two females and the females monitored for vaginal plug formation. On the day a plug was observed, the female was considered to be pregnant at E0.5. Embryos were harvested from the pregnant female at E11.5 and the limb buds and neural tube / somites dissected in filming media (10% fetal bovine serum (FBS) in DMEM / F12, no phenol red, 21041025). Each tissue was dissociated using 1% trypsin (Gibco 27250018) in Hanks’ Balanced Salt Solution (HBSS, Thermo Fisher, 14025-076) at 37 °C and the trypsin was neutralized with filming media. Cells were washed with 1xPBS and then RNA was extracted with TRIzol (Thermo Fisher, 15596026) following manufacturer’s instructions. The RNA was treated with TURBO DNase (Ambion, AM2696) at 37°C for 30 minutes to remove genomic DNA and purified using the Purelink RNA Mini Kit (Ambion, 12183018) following the manufacturer’s instructions. cDNA synthesis was performed using 1 pg RNA with the Script Reverse Transcription Supermix kit (Bio-Rad, 1708841) according to manufacturer’s instructions. The cDNA was diluted 10-fold and 4 uL used as a template for qPCR with SsoAdvanced SYBR Green supermix (Bio-Rad, 1725270) on a CFX384 machine (Bio-Rad).

[0169] In order to perform absolute quantification, a DNA standard template was made by cloning the Hoxa9 IRES-CDS from E11.5 neural tube and somite cDNA using primers KL596 / KL110. The PCR product was cloned into the pCR4 TOPO vector using the Zero Blunt TOPO Cloning Kit (Invitrogen, 450245) following kit instructions and sequenced to confirm that the plasmid insert was from the mature spliced Hoxa9 transcript sequence. A standard curve was made for each amplicon by performing qPCR on 10-fold serial dilutions (1 pg to 0.1 fg) and used to determine the starting quantity of each amplicon in the embryonic cDNA samples.Mouse Embryo Sucrose Gradient Fractionation Analysis and RT-qPCR

[0170] For sucrose gradient fractionation of lysates of micro-dissected mouse embryo tissues from mouse tissues, E11.5 mouse embryo neural tube and somites were dissected and dissociated with 1% trypsin (Gibco 27250018) and 100 pg / mL cycloheximide (CHX) (Sigma- Aldrich, C7698- 1G) in HBSS at 37°C. After neutralization with cold filming media with 100 pg / mL cycloheximide, the cells were washed with cold 1xPBS with 100 pg / mL cycloheximide and lysed in lysis buffer (20 mM Tris-HCI pH 7.5 (Ambion, AM9850G, and Ambion, AM9855G), 150 mM NaCI (Ambion, AM9759), 15 mM MgCI2 (Ambion, AM9530G), 1 mM DTT (Ambion, 10197777001), 8% glycerol (Sigma-Aldrich, G5516), 1 % Triton X-100 (Sigma-Aldrich, T8787), 100 pg / ml Cycloheximide (CHX) (Sigma-Aldrich, C7698-1G), 20 U / ml TURBO DNase (Ambion, AM2238), and Complete Protease Inhibitor EDTA-free (Sigma-Aldrich, 11836170001) in nuclease-free water (Thermo Fisher Scientific, 10977015)) at 4°C for 30 minutes on a rotator at 4°C with occasional vortexing, followed by sequential centrifugation at 1800 g for 5 min at 4°C and then 10,000*g for 5 min at 4°C. The clarified cytoplasmic extract was loaded onto 25-50% sucrose (Fisher Scientific, S5-12) (w / v) gradients (20 mM Tris pH 7.5, 100 mM NaCI, 15 mM MgCI2, 100 pg / ml cycloheximide), made by sequentially freezing 50%, 43.75%, 37.5%, 31.25%, and 25% sucrose. Gradients were centrifuged in a Beckman SW60 rotor at 35,000 rpm for 2.5 hrs at 4°C in a Beckman L8-80M ultracentrifuge and then fractionated on a Density Gradient Fraction System (Brandel, BR-188) using PeakChart software (v1.02) with continuous A260 measurements. To each fraction we added 100 pg of in vitro transcribed Renilla luciferase RNA as a spike-in control and detected it using Rluc-specific primers. RNA was extracted using Acid- PhenokChloroform, pH 4.5 (with IAA, 125:24:1) (Ambion, AM9722), incubating for 5 min at 65°C followed by centrifugation at 21 ,000xg for 10 min. The aqueous phase was mixed 1 :1 with ethanol and RNA isolated using the RNA Clean and Concentrator-5 Kit (Zymo Research, R1016). The RNA was treated with TURBO DNase (Ambion, AM2238) for 30 minutes at 37°C and purified with the RNA Clean and Concentrator-5 Kit (Zymo Research, R1016). cDNA synthesis was performed using the iScript Reverse Transcription Supermix kit (Bio-Rad, 1708841) with 200 ng of RNA, diluted 20-fold, and 4 pl used as a template for qPCR with SsoAdvanced SYBR Green supermix (Bio-Rad, 1725270) on a CFX384 machine (Bio-rad). The Ct value of each amplicon from each fraction was first normalized to the Ct value of the luciferase RNA spike- in, converted from Iog2 to linear values, and normalized to the total abundance of the mRNA across all fractions.ESMFold and Fusion Protein Structure Prediction

[0171] ESMFold, as implemented in the ESM Metagenomic Atlas was used to predict the structure of the NLB reporter, given the low depth of co-evolutionary information available for the Nanoluc domain. Molecular graphics created and analysis performed with UCSF ChimeraX.Luciferase Activity Assay after Plasmid Transfection

[0172] Mouse C3H / 10T1 / 2 cells in 12-well plates were transiently transfected (Lipofectamine 2000, Invitrogen, 11668-019) with 1 ug NLB and 1 ug hHBB-F\uc control plasmid. Cells were washed with 1x PBS (Gibco, 14190-250) and collected by trypsinization 24 hours posttransfection for Nluc / Fluc luciferase activity assays. 60% the cells were used for assaying luciferase activity using the Nano-Gio Dual-Luciferase Reporter Assay System (Promega, N1610) to measure Nanoluc (NIuc) and Firefly (Flue) luciferase activities, the rest 40% was collected in TRIzol (Invitrogen, 15596) for total RNA purification with the RNA Clean and Concentrator-5 Kit (Zymo Research, R1016), TURBO DNase (Ambion, AM2238) digest, repeat column purification, and detection of relative mRNA levels by RT-qPCR (see RT-qPCR section). For luciferase assays, cells were lysed in 50 pl of 1x passive lysis buffer (Promega, E1941) and directly assayed or frozen at -20°C. After thawing, cell debris and nuclei were removed by centrifugation for 1 min at 13,000 rpm. 25 pl of supernatant was assayed in a 96 well black-non-transparent plate (Falcon, 732-2194) for luciferase activity by mixing with 35 pl of Nano-Gio Dual-Luciferase Reporter Assay System substrates (Promega, N1610, concentrated substrates for promoterless assays and 1 :10 dilutions in water for all other assays) using manual multi-channel pipetting. Flue and NIuc activities were measured on a GloMax Explorer Microplate Reader (Promega). Luciferase reporter activity is expressed as a ratio between NIuc and Flue (Nluc / Fluc) which was compared to respective normalized mRNA levels. For monocistronic Nluc-b-g / ob / n-fusion constructs, Nluc / Fluc luciferase activity was normalized to b-globin / Nupl1 mRNA levels to quantify variation in mRNA expression. Each experiment was performed a minimum of three independent times.Quantitative RT-PCR (RT-qPCR) Analysis

[0173] Cells transfected with pGL3-NLB-fusion constructs were collected in 500 pL TRIzol (Invitrogen, 15596). Total RNA was isolated from the aqueous phase using Zymo Clean & Concentrator-5 column (Zymo Research, R1016) and treated with TURBO DNase (Ambion, AM2238) followed by a second Zymo column purification step. For reverse transcription- quantitative PCR (RT- qPCR) analysis, cDNA was synthesized from 100-200 ng of total RNA using iScript Supermix (Bio-Rad, 1708840) containing random hexamer primers, according to themanufacturer’s instructions. PCR reactions were assembled in 384-well plates using 2.5 pL of a 1 :4-1 :5 dilution of a cDNA reaction, 300 nM of target-specific primer mix and the SsoAdvanced SYBR Green supermix (Bio-Rad, 1725270) or5x EvaGreen qPCR-Mix (Bio-Budget, 80-5820000) in a final volume of 10 pl per well. SYBR green detection qPCR was performed on a CFX384 machine (Bio-Rad) or QuantStudio qPCR machine (Thermo, applied biosystems, 15721248). Data was analyzed and converted to relative RNA quantity manually or using CFX manager (BioRad). circRNA Plasmid-based EG FP Reporter Assay

[0174] Cells transiently transfected with mRuby-ZKSCAN-spEGFP plasmids were harvested 24, 48 or 72 hrs post transfection. After resuspension, cells were transferred into 1.5 ml Eppendorf tubes. Cells were washed with 500 pL IxDPBS and finally resuspended in 180 pL FACS buffer (1xPBS, 1% BSA, 0.1 % NaN3) for the 24 hrs and 48 hrs test series and in 230 pL for the 72 hrs test series. The mRuby and EGFP signal intensity was then measured by using the ABI Attune Acoustic Focusing Cytometer (Thermo Scientific). FACS data were exported as FCS files and further analyzed by FlowJo v10.8 (BD). Low quality data derived from cells measured after a clogging event of the FACS machine were initially excluded during quality control. Afterwards, viable cells and singlets were identified based on theFSC and SSC channel. Finally, transfected cells were identified based on their mRuby signal, and the EGFP signal of this positive subtraction was measured and used for calculation of the median fluorescence intensities (MFIs).Quantification and Analysis of circRNA Content by qPCR

[0175] After 5 days of incubation, cells were harvested and transferred into 15 mL falcon tubes. After a quick centrifugation (500 g, 5 min at RT), media was replaced with 6 mL of full cell culture media. In order to reduce the noise level of untransfected cells during the following assay, cells were FACS sorted by the Flow Cytometry Core Facility of the Medical Faculty at the University of Bonn based on their mRuby and EGFP signal intensity. TRIzol-based total RNA extraction (Invitrogen, 15596026) was then performed on the double positive sorted cell fraction and purified using the RNA Clean & Concentrator -5 kit (Zymo, R1016). Remaining DNA contamination was removed by TURBO DNase treatment for 30 min at 37°C (Thermo, AM2238) and second purification using the same Clean & Concentrator -5 column as beforehand. The obtained RNA was split into two and incubated with (positive sample) or without (negative control) RNase R (Biozym Scientific, B3539-172010) at a working concentration 1 U / pg RNA in presence of RNase R reaction buffer for 30 min at 37°C and 500 rpm shaking. After subsequent RNApurification using the PureLink RNA Mini Kit (Thermo, 12183018A), equal amounts of 200-500 ng RNA per construct were used for initial cDNA synthesis using the iScript RT supermix (Biorad, 1708841), according to manufacturer’s instructions, and diluted 1 :5 before final circRNA quantification by qPCR (my- Budget 5x EvaGreen qPCR-Mix II) with 300 nM of target-specific primer mix in a final volume of 10 pl.

[0176] Raw qPCR data were transformed (2A(-1*(dCt))), normalized to the housekeeping mRNA Nup / 1 and linear and circular RNA content was visualized after RNase R digestion relative to the undigested control samples. Additionally, the ratio of circRNA enrichment over RNase R digestion relative to the linear RNA enrichment over RNase R digestion was calculated for comparison between the different tested constructs.In vitro synthesis of circular RNAs and purification

[0177] In order to generate IVT templates which include the required T7 promoter side, which were used subsequently for the main circular RNA production, 200ng of Plasmid template DNA (constructs: pKL480; pll001 ; pll002; PKp001-PKp005; PKpO42-PKpO48) were used in a 50pl AccuPrime Pfx PCR reaction (Thermo Fisher Scientific, cat. no. 12344024) with the primer pair PLL078 & PLL079 following the manufactures recommendations (95°C 2min; 35x 95°C 20sec, 58°C 45sec, 68°C 3min; 68°C 10min; 4°C storage) supplemented with 1.5pl of DMSO and 0.5pl MgSC 50mM. Correct sizes of the generated templates were evaluated by a standard 1% agarose gel electrophoresis. Correct templates were cut out and purified using the NEB Monarch Gel extraction kit (cat. No. T1020L) following manufactures instructions with an adapted elution volume of 5pl ultra-pure DNase / RNase free water. 1 g of the purified IVT templates were used subsequently in a standard IVT T7 reaction following manufactures instructions (NEB HiScribe T7 High Yield RNA Synthesis Kit cat. no. E2040S) supplemented with 1 l RiboLock (Thermo Fisher Scientific cat. no. EO0381) and 2pl DTT. The reaction, set up in a DNA low bind tube (Eppendorf cat. No. 022431021), was incubated for 2h at 37°C at 800rpm. Afterwards, 68pl ultra- pure RNase / DNase free water was added to the reaction mixture without removing the tubes from the heatblock, after first adding 10pl 10x TURBO DNase reaction buffer and 2pl TURBO DNase (Thermo Fisher Scientific cat. no. AM2238). The reaction was incubated for 20min at 37°C at 800rpm and the RNA products were subsequently purified using the PureLink RNA Mini-Kit (Thermo Fisher Scientific cat. no. 12183018A) following manufactures instructions. The total RNA concentration was measured by using a Thermo Fisher Scientific NanoDrop One. In order to purify the generated circular RNAs, linear side products as well as unmature non-circularized RNAs were degraded over Ribonuclease R (RNaseR) digestion (Biozyme Scienfitic cat. no172010). 1pg of total RNA input was digested with 1 Unit of RNaseR following manufactures recommendations and incubated for 45min at 37°C and 500rpm (Dlx1 purification required a 90min RNaseR digestion). The final circular RNA was purified using the PureLink RNA Mini-Kit (Thermo Fisher Scientific cat. no. 12183018A). The successful purification of circular RNA was finally evaluated using the Agilent Technologies 4200 TapeStation system using the High Sensitive RNA Screen Analysis (Agilent Technologies cat. no. 5067-5579).Table 1. Sequences

Claims

WHAT IS CLAIMED IS:

1. A composition, comprising: a circular RNA construct comprising a sequence for encoding a gene product that is operably linked to an I RES-like sequence derived from a 5’-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

2. The composition of claim 1 , wherein the I RES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

3. The composition of claim 1 or 2, wherein the I RES-like sequence comprises one of: SEQ ID NOs: 1-5, 37, or 38.

4. The composition of claim 1 , wherein the I RES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDL1, or DLX1.

5. The composition of claim 1 or 4, wherein the I RES-like sequence comprises one of: SEQ ID NOs: 6-9, 39 or 40.

6. The composition of any one of claims 1-5, wherein the gene product is a non-self gene product.

7. The composition of any one of claims 1-6, wherein the gene product is a heterologous gene product.

8. The composition of any one of claims 1-7, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

9. The composition of claim 8, wherein the gene product is a therapeutic protein or peptide comprising: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

10. The composition of claim 8, wherein the gene product is an immunogen comprising: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

11. The composition of claim 8, wherein the gene product is a cytotoxin comprising: TNF, a granzyme, a perforin, a caspase, a ricin, or a cell-penetrating peptide.

12. The composition of claim 8, wherein the gene product is a gene editing tool comprising: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

13. The composition of claim 8, wherein the gene product is a molecular marker comprising: a fluorescent protein, a luciferase, or a peroxidase.

14. A composition, comprising: a DNA construct comprising a sequence for transcribing a linear mRNA molecule that is operably linked to a promoter and poly-A signal; wherein the linear mRNA molecule comprises a sequence comprising two introns, a gene sequence split into a 3’-gene sequence and a 5’-gene sequence, and an I RES-like sequence derived from a 5’-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene; wherein the I RES- like sequence is downstream of the 3’-gene sequence and upstream the 5’-gene sequence; whereby processing of the linear mRNA molecule by a spliceosome yields a circular RNA construct comprising the gene sequence operably linked to the I RES-like sequence; whereby the processing of the linear mRNA molecule by the spliceosome also conjoins the 5’-gene sequence and the 3’-gene sequence to yield the gene sequence in a configuration that encodes a product of the gene sequence.

15. The composition of claim 14, wherein the I RES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdH, or Dlx1.

16. The composition of claim 14 or 15, wherein the I RES-like sequence comprises one of: SEQ ID NOs: 10-14, 41 or 42.

17. The composition of claim 14, wherein the I RES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDL1, or DLX1.

18. The composition of claim 14 or 17, wherein the I RES-like sequence comprises one of: SEQ ID NOs: 15-18, 43 or 44.

19. The composition of any one of claims 14-18, wherein the gene product is a non-self gene product.

20. The composition of any one of claims 14-19, wherein the gene product is a heterologous gene product.

21. The composition of any one of claims 14-20, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

22. The composition of claim 21 , wherein the gene product is a therapeutic protein or peptide comprising: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

23. The composition of claim 21 , wherein the gene product is an immunogen comprising: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

24. The composition of claim 21 , wherein the gene product is a cytotoxin comprising: TNF, a granzyme, a perforin, a caspase, a ricin, or a cell-penetrating peptide.

25. The composition of claim 21 , wherein the gene product is a gene editing tool comprising: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

26. The composition of claim 21 , wherein the gene product is a molecular marker comprising: a fluorescent protein, a luciferase, or a peroxidase.

27. A method of producing a gene product, comprising: contacting a eukaryotic cell with a circular RNA construct, wherein the circular RNA construct comprises a sequence for encoding the gene product that is operably linked to an IRES- like sequence derived from a 5’-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

28. The method of claim 27, wherein the I RES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

29. The method of claim 27 or 28, wherein the I RES-like sequence comprises one of: SEQ ID NOs: 1-5, 37, or 38.

30. The method of claim 27, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDL1, or DLX1.

31. The method of claim 27 or 30, wherein the IRES-like sequence comprises one of: SEQ ID NOs: 6-9, 39, or 40.

32. The method of any one of claims 27-31 , wherein the gene product is a non-self gene product.

33. The method of any one of claims 27-32, wherein the gene product is a heterologous gene product.

34. The method of any one of claims 27-33, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

35. The method of claim 34, wherein the gene product is a therapeutic protein or peptide comprising: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

36. The method of claim 34, wherein the gene product is an immunogen comprising: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

37. The method of claim 34, wherein the gene product is a cytotoxin comprising: TNF, a granzyme, a perforin, a caspase, a ricin, or a cell-penetrating peptide.

38. The method of claim 34, wherein the gene product is a gene editing tool comprising: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

39. The method of claim 34, wherein the gene product is a molecular marker comprising: a fluorescent protein, a luciferase, or a peroxidase.

40. A method of treating an individual, comprising: administering to the individual a circular RNA construct, wherein the circular RNA construct comprises a sequence for encoding a therapeutic gene product that is operably linked to an IRES-like sequence derived from a 5’-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

41. The method of claim 40, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

42. The method of claim 40 or 41 , wherein the IRES-like sequence comprises one of: SEQ ID NOs: 1-5, 37, or 38.

43. The method of claim 40, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDU, or DLX1.

44. The method of claim 40 or 43, wherein the I RES-like sequence comprises one of: SEQ I D NOs: 6-9, 39 or 40.

45. The method of any one of claims 40-44, wherein the gene product is a non-self gene product.

46. The method of any one of claims 40-45, wherein the gene product is a heterologous gene product.

47. The method of any one of claims 40-46, wherein the therapeutic gene product comprises: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

48. A method of vaccinating an individual, comprising: administering to the individual a circular RNA construct, wherein the circular RNA construct comprises a sequence for encoding an immunogenic gene product that is operablylinked to an IRES-like sequence derived from a 5’-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene.

49. The method of claim 48, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

50. The method of claim 48 or 49, wherein the IRES-like sequence comprises one of: SEQ ID NOs: 1-5, 37, or 38.

51. The method of claim 48, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDU, Of DLXl.

52. The method of claim 48 or 51, wherein the IRES-like sequence comprises one of: SEQ ID NOs: 6-9, 39, or 40.

53. The method of any one of claims 48-52, wherein the gene product is a non-self gene product.

54. The method of any one of claims 48-53, wherein the gene product is a heterologous gene product.

55. The method of any one of claims 48-54, wherein the immunogenic gene product comprises: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

56. A composition, comprising: a DNA construct comprising a sequence for transcribing a linear mRNA molecule that is operably linked to a promoter capable of driving in vitro transcription in a cell-free buffer; wherein the linear mRNA molecule comprises a sequence comprising two self-splicing ribozyme introns, a gene sequence encoding a gene product, and an IRES-like sequence derived from a 5’-UTR of a mouse Hoxa gene, a mouse developmental gene, a human HOXA gene, or a human developmental gene; wherein the IRES-like sequence is downstream of the gene sequence; whereby the linear mRNA molecule is configured to self-splice via the two self-splicing ribozyme introns to yield a circular RNA construct comprising the gene sequence operably linked to the IRES-like sequence; whereby the self-splicing of the linear mRNA molecule by the self-splicingribozyme introns also conjoins the gene sequence with the IRES-like sequence using complementary exon junction sequences to operably link the gene sequence with the IRES-like sequence.

57. The composition of claim 56, wherein the IRES-like sequence is derived from one of mouse Hoxa3, Hoxa4, Hoxa5, Hoxa9, Hoxa11, ChrdU, or Dlx1.

58. The composition of claim 56 or 57, wherein the IRES-like sequence comprises one of: SEQ ID NOs: 10-14, 41 or 42.

59. The composition of claim 56, wherein the IRES-like sequence is derived from one of human H0XA3, H0XA4, H0XA5, H0XA9, H0XA11, CHRDU, or DLX1.

60. The composition of claim 56 or 59, wherein the IRES-like sequence comprises one of: SEQ ID NOs: 15-18, 43 or 44.

61. The composition of any one of claims 56-60, wherein the gene product is a non-self gene product.

62. The composition of any one of claims 56-61 , wherein the gene product is a heterologous gene product.

63. The composition of any one of claims 56-62, wherein the gene product is a therapeutic protein or peptide, an immunogen, a cytotoxin, a genetic editing tool, or a molecular marker.

64. The composition of claim 63, wherein the gene product is a therapeutic protein or peptide comprising: a hormone, a growth factor, a cytokine, a chemokine, an antigen-binding product, or a blood factor.

65. The composition of claim 63, wherein the gene product is an immunogen comprising: a protein or a peptides derived from a pathogen, or a cancer-related neoantigen.

66. The composition of claim 63, wherein the gene product is a cytotoxin comprising: TNF, a granzyme, a perforin, a caspase, a ricin, or a cell-penetrating peptide.

67. The composition of claim 63, wherein the gene product is a gene editing tool comprising: a Cas nuclease of a CRISPR / Cas system, a TALEN, or a zinc-finger nuclease.

68. The composition of claim 63, wherein the gene product is a molecular marker comprising: a fluorescent protein, a luciferase, or a peroxidase.

Images