Preparation of circular RNA

Abstract

A linear precursor RNA molecule is provided, having a 5' end and a 3' end, and comprising in the 5'-to-3' direction i) a first self-splicing sequence, ii) a sequence of interest, and iii) a second self-splicing sequence, wherein the linear precursor RNA molecule is capable of forming a circular RNA upon splicing of said first and second self-splicing sequences; wherein the first self-splicing sequence comprises a first homopolymeric moiety located between said 5' end and a splice site first self-splicing sequence, and / or wherein the second self-splicing sequence comprises a second homopolymeric moiety located between said 3' end and a splice site of the second self-splicing sequence. The invention may improve the production and / or purification of circular RNA.

Description

[0001] PREPARATION OF CIRCULAR RNA

[0002] TECHNICAL FIELD

[0003] The present invention relates to a method of producing circular RNA, to a linear precursor RNA intended for production of circular RNA, to a composition comprising circular RNA, linear precursor RNA and other RNA impurities, and to methods of separating circular RNA from such a composition.

[0004] BACKGROUND

[0005] The use of RNA, in particular mRNA, for pharmaceutical applications is a fast-emerging field and the rapid development of effective vaccines against COVID-19 has clearly demonstrated the great potential of mRNA technology. This has paved the way for a range of future novel vaccines and therapies. Synthetic biology enables rapid tailoring of the mRNA sequence, which makes mRNA- based vaccines an attractive therapeutic strategy for future pandemics, as well as endemics and other infectious diseases such as rabies, Zika, and cytomegalovirus infection. In addition to this, there are several clinical studies ongoing which are based on mRNA treatment of diseases, such as HIV, various types of cancers, cystic fibrosis and diabetes. mRNA technology has also shown to be a good candidate for use within personalized medicine.

[0006] Production of mRNA is typically performed by in vitro transcription (IVT), followed one or more steps of purification to remove IVT components and by-products. This can include precipitation and filtering steps including tangential flow filtration (TFF) and / or different modes of chromatography. For example, mRNA can be purified by affinity chromatography based on the interaction between the mRNA 3' polyA tail and a complementary oligo(dT) ligand, e.g. as described in W02014152031A1.

[0007] Beyond conventional mRNA molecules for therapeutic application, the use of circular RNA is gaining attention. Reasons for the growing interest in circular RNA include its longevity and reduced immunogenicity mediated by the closed-loop structure without a 5' or 3' end. Circular RNA is considered to offer several advantages over traditional linear mRNA molecules.

[0008] A major problem when working with mRNA is the presence of degrading ribonucleases (RNases). As a circular RNA molecule does not have a free 5' and 3' ends, it is less prone to degradation by exonucleases. In addition, the lack of 5' phosphate groups in circular RNA significantly reduces recognition by the innate immune response. Prolonged protein expression (around 1 week) in cells has been reported for circular RNA as compared to normal mRNA (around 2-3 days). Furthermore, uncapped linear mRNAs may activate innate immune receptors and consequently cause an immune response in patients and / or limit clinical efficacy. As an example, innate immune receptors such as Toll-like receptor 8 may be activated by uridine, and this effect can be suppressed by the use of modified nucleotides, like pseudouridines. However, modified nucleotides are expensive and may alter the structural properties of mRNA. By contrast, circular RNAs have been found to be less immunogenic (Wesselhbft et al 2019), even when using unmodified nucleotides. It is also foreseen that the increased in vivo stability of circular RNA will allow reducing the dosage, which could also be also expected to raise less immune response in patients. In addition, the development process is shorter, as there is no need for optimization of combinations of modified nucleotides and this group of RNA molecules likely require less effort in removing uridine nucleotides from the coding region.

[0009] Furthermore, the absence of 5' cap and 3' polyA tail in circular RNA simplifies the production process which can therefore be shorter, with minimal losses. From a regulatory point of view the manufacturing is simplified as no analysis methods are needed for these two features.

[0010] Circular RNAs exist at low abundance in nature, and different strategies have been developed for synthetic circularization of RNA fragments. Different techniques can be used to generate circular RNA molecules, all of which depend, as a first step, on the provision of a linear precursor RNA, e.g. by IVT. To date, the state-of-the-art technique for generating circular RNA of some length is based on transcribed functional sequence elements that facilitate the formation of a circular RNA molecule. More specifically, one method of RNA circularization utilizes self-splicing ribozyme subunits that can be incorporated in the linear precursor RNA. For example, the permuted intron-exon (PIE) splicing strategy uses an exon pair flanked by permuted group I catalytic intron fragments. Under circularization conditions, the linear precursor RNA containing the 3' and 5' PIE fragments undergo a catalytic reaction resulting in splicing of the introns and fusing of the respective 5' and 3' exon ends to form a circular RNA molecule.

[0011] Currently available methods to produce circular RNA struggle with low circularization efficiency, especially for larger RNAs. The circularization efficiency can to some extent be improved e.g. by addition of homology arms, but efficiencies are typically quite low (20-30 % of total RNA) for RNAs of around 2000 nucleotides (nt) and significantly lower for larger mRNAs.

[0012] Another major challenge in the preparation of circular RNA for therapeutic applications is the removal of RNA impurities. As a result of the circularization reaction, the reaction sample contains, in addition to the desired circular RNA, (a) unreacted linear precursor RNA, (b) intermediates, (c) nicked RNA, i.e. RNA that has re-opened after circularization and (d) shorter linear fragments that have been cleaved from the linear precursor. Intermediates may include linear RNA molecules where only one intron has been cleaved off and the other intron (including any homology arms, etc) remains. These RNA species (a)-(d) represent impurities, and it is desirable to separate these from the circular target RNA. However, these RNA impurities are highly charged and have large hydrodynamic radii, very similar to the circular RNA. Chromatographic techniques such as ion exchange and hydrophobic interaction chromatography currently do not achieve satisfactory resolution when separating these species at large scale, or may require harsh and non-aqueous buffer conditions. Thus, the presence of these RNA species, which are very similar to the target circular RNA, makes purification of circular RNA more difficult as compared to purification of conventional mRNA.

[0013] The currently most widely used method for removing linear RNA species is based on enzymatic degradation of linear RNA specifically by the exonuclease RNase R. After IVT and circularization, the RNase R degrades the linear species while the target circular RNA largely remains intact and can then be further purified to remove protein components of the reaction. However, this procedure adds process steps, and suffer from low specificity and efficiency which results in partial degradation of the circular RNA.

[0014] Another option is separation of circular and linear RNA species by high-performance liquid chromatography (HPLC). However, this approach has limited scalability and requires the use of toxic solvents.

[0015] A new approach for purification of circular RNA is outlined in WO2023242425A1. This document proposes a circular RNA comprising a protein encoding region and an RNA aptamer, which is intended to bind to an affinity ligand, to allow binding of the circular RNA to an affinity chromatography resin, followed by elution of the circular RNA.

[0016] Taken together, efficient purification of circular RNA is continuing to be a challenge.

[0017] SUMMARY OF THE INVENTION

[0018] It is an object of the invention to overcome or at least partly alleviate drawbacks of the prior art.

[0019] Accordingly, it is an object of the invention to provide a linear precursor RNA that facilitates the production and / or purification of circular RNA. It is also an object of the invention to provide improved methods for production and / or purification of circular RNA.

[0020] These and other objects are achieved by a linear precursor RNA molecule having a 5' end and a 3' end, comprising in the 5'-to-3' direction: i) a first self-splicing sequence, ii) a sequence of interest, and iii) a second self-splicing sequence, wherein the linear precursor RNA molecule is capable of forming a circular RNA upon splicing of said first and second self-splicing sequences, wherein the first self-splicing sequence comprises a first homopolymeric moiety located between said 5' end and a splice site first self-splicing sequence, and / or wherein the second self-splicing sequence comprises a second homopolymeric moiety located between said 3' end and a splice site of the second self-splicing sequence. Thus, at least one of said first self-splicing sequence and said second self-splicing sequence comprises a homopolymeric moiety.

[0021] Advantageously, the homopolymeric moiety or moieties can be utilized as a purification handle in a method of separating circularized RNA from linear impurities, such as precursor RNA, intermediates and cleaved self-splicing sequences. Furthermore, it was surprisingly found that the proposed linear precursor RNA having at least one homopolymeric moiety has the potential to increase circularization efficiency.

[0022] In embodiments, the first and / or the second homopolymeric moiety may comprise at least one homopolymeric sequence of at least 6, such as at least 10, consecutive identical units. In embodiments, the first and / or the second homopolymeric moiety comprises up to 75 units.

[0023] In embodiments, the first and / or the second homopolymeric moiety may comprise or consist of at least one homopolymeric sequence which is an interrupted sequence comprising at least two repeating segments separated by a single intervening unit or shorter sequence. Each repeating segment may consist of at least 3 repeating units, such as nucleotides.

[0024] Where present, the first homopolymeric moiety preferably comprises or is a nucleic acid sequence. Where present, the second homopolymeric moiety preferably comprises or is a nucleic acid sequence. Hence, the homopolymeric moiety or moieties may comprise repeating units which are nucleotides.

[0025] For example, the first and / or the second homopolymeric moiety may comprise at least one sequence selected from polyA, poly U, polyC and poly(G), preferably a polyA sequence, wherein said sequence optionally includes one or more modified nucleotides. Modified nucleotides may be of the same type. Alternatively, two modified nucleotides may represent different types of modification.

[0026] The first homopolymeric moiety may be located adjacent to the 5' end of the linear RNA molecule. Optionally said first homopolymeric moiety may be located downstream of a 5' start sequence, such as directly following said 5' start sequence. In embodiments, the first self-splicing sequence may comprise said first homopolymeric moiety located between said 5' end and a splice site first self-splicing sequence. The second homopolymeric moiety may be located adjacent to the 3' end of the linear RNA molecule. Optionally said second homopolymeric moiety may be located upstream of a 3' end sequence, such as immediately preceding said 3' end sequence. In embodiments, the second selfsplicing sequence comprises said second homopolymeric moiety located between said 3' end and a splice site of the second self-splicing sequence.

[0027] Where the linear precursor comprises both of said first and second homopolymeric moieties, the first and second homopolymeric moieties preferably do not bind to each other. For example, the first homopolymeric moiety and the second homopolymeric moiety may be nucleic acid sequences which are non-complementary or uncapable of hybridizing to each other to each other. Optionally, the first and second homopolymeric moieties may be identical to each other.

[0028] The linear precursor RNA molecule may comprise a first joining sequence (El) located between the sequence of interest and the splice site first self-splicing sequence, and a second joining sequence located (E2) located between the sequence of interest and the splice site of the second self-splicing sequence, wherein said first and second joining sequences are capable of joining to form a circular RNA molecule containing the sequence of interest. For example, the first self-splicing sequence may comprise a 3' self-splicing intron fragment and the second self-splicing sequence may comprise a 5' self-splicing intron fragment. For example, the first self-splicing sequence may comprise a sequence that forms part of a permuted intron-exon (PIE), and the second self-splicing sequence may comprise a sequence that forms part of a permuted intron-exon. Optionally the linear precursor RNA may comprise a circularizing ribozyme.

[0029] In embodiments, the first self-splicing sequence comprises a first pairing sequence, and the second self-splicing sequence comprises a second pairing sequence, wherein said first and second pairing sequences are capable of hybridizing with each other. Where the first self-splicing sequence comprises the first homopolymeric moiety, the first pairing sequence may be located between the first homopolymeric moiety and the splice site in the 5'-to-3' direction. Where the second selfsplicing sequence comprises the second homopolymeric moiety, the second pairing sequence may be located between the splice site and the second homopolymeric moiety in the 5'-to-3' direction.

[0030] The sequence of interest may comprise at least one coding sequence encoding a protein of interest and optionally at least one translation regulatory sequence, such as an IRES sequence.

[0031] Alternatively or additionally, the sequence of interest may comprise at least one of a regulatory RNA sequence, protein binding RNA sequence, or a guide RNA sequence.

[0032] In another aspect, the invention provides a template DNA molecule encoding a precursor

[0033] RNA molecule as described herein. In another aspect, the invention provides a method of producing a circular RNA molecule comprising a sequence of interest, the method comprising incubating a precursor linear RNA molecule as described herein under conditions allowing splicing of the splicing sequences and circularization to form a circular RNA molecule containing the sequence of interest.

[0034] It was surprisingly found that the method disclosed herein, using the linear precursor of the first aspect, increased circularization efficiency. Thus, the yield of circular RNA may be increased, and the content of linear precursor and other impurities remaining after the circularization reaction may be reduced. Furthermore, further separation of impurities is facilitated by the presence of the homopolymeric moiety or moieties of the linear precursor.

[0035] The method may comprise a step of performing an in vitro transcription (IVT) reaction of the template DNA molecule of the second aspect, to provide the precursor linear RNA molecule. Performing an in vitro transcription (IVT) reaction typically comprises providing an IVT reaction mixture comprising template DNA, ribonucleoside triphosphates, DNA dependent RNA polymerase and a buffer system, wherein the step of incubating the linear precursor RNA molecule comprises incubating said IVT reaction mixture.

[0036] In the method of producing circular RNA, the linear precursor RNA molecule may be incubated at a temperature in the range of from 20°C to 56°C, such as from about 37°C to about 56 °C. The incubation time may be from 10 minutes to 24 hours, such as from 10 minutes to 5 hours, such as from 30 minutes to 5 hours. The pH during the incubation may be in the range of 6.5-8. Optionally GTP and / or ATP may be added to the composition comprising the precursor RNA prior to or during the incubation.

[0037] The method of producing circular RNA may optionally further comprise a step of separating the circular RNA from the linear precursor RNA molecule and / or other RNA species formed during incubation.

[0038] In another aspect, the invention provides a composition comprising: a linear precursor RNA molecule, a circular RNA molecule comprising the sequence of interest of the precursor RNA molecule, and optionally, at least one impurity selected from: i) an RNA molecule comprising at least part of the first self-splicing sequence but lacking the second self-splicing sequence, ii) an RNA molecule comprising at least part of the second self-splicing sequence but lacking the first selfsplicing sequence, iii) a cleaved first self-splicing sequence, and iv) a cleaved second self-splicing sequence. The composition may represent a crude reaction mixture resulting from the IVT reaction (incubation), in which case the composition comprises circular RNA, RNA impurities as well as IVT components. Alternatively, the composition may be an IVT reaction mixture that has been processed or purified by one or more steps, e.g. by size exclusion chromatography, which step(s) may serve to remove IVT components such as digested template DNA, but which may not remove the linear precursor RNA and the aforementioned linear RNA impurities i)-iv).

[0039] Where the composition is an IVT reaction mixture it may further comprise template DNA encoding the precursor RNA, optionally digested by deoxyribonuclease (DNase) in which case the composition typically also contains DNase ribonucleoside triphosphates; a DNA dependent RNA polymerase, and any cofactors required for nucleic acid polymerization catalyzed by said DNA dependent polymerase; and a buffer.

[0040] The composition may additionally comprise at least one component selected from a pyrophosphatase, an RNase inhibitor, and a detergent. In embodiments, the composition may contain an RNase inhibitor. In embodiments, the composition may be RNase free.

[0041] Where the composition has been subjected to one or more conditioning or purification steps, some or all of the non-RNA components residual from an IVT reaction may have been removed. Hence, the composition may be essentially free of non-RNA IVT reaction components such as NTPs, digested DNA template, RNA polymerase and other proteins.

[0042] In a further aspect, the invention provides a method of separating a desired circular RNA from a linear RNA species comprising at least one homopolymeric moiety, the method comprising a) providing a composition as described above or a composition obtainable or obtained by the method of producing circular RNA as described herein, b) contacting the composition with a chromatography material comprising a ligand coupled to a support material, said ligand having a binding affinity for the homopolymeric moiety of the linear precursor RNA molecule, under conditions allowing the ligand to bind to the homopolymeric moiety, c) obtaining the circular RNA in a flowthrough fraction, d) optionally, eluting bound non-circular RNA species from the chromatography material and optionally regenerating the chromatography material.

[0043] The ligand may be a nucleic acid molecule that is capable of binding to, such as hybridizing with, the homopolymeric moiety. For example, the ligand may be an oligo(dT) ligand and the homopolymeric moiety may comprise a polyA sequence.

[0044] In embodiments, the support material may be a convection-based matrix, optionally selected from the group consisting of a porous polymer membrane, a filter, a fibrous matrix and a porous monolith. In other embodiments, the support material may be a chromatography resin, comprising beads selected from porous beads, non-porous beads, and magnetic beads. Porous beads may be formed of a polysaccharide or derivative thereof, such as agarose or a derivative thereof.

[0045] Optionally, the method may comprise, prior to step b, a step of subjecting the composition to a size-based separation step, such as filtration or size exclusion chromatography.

[0046] Preferred aspects of the present disclosure are described below in the detailed description and in the dependent claims. It is noted that the invention relates to all possible combinations of features recited in the claims.

[0047] BRIEF DESCRIPTION OF THE DRAWINGS

[0048] These and other aspects of the present invention will now be described in more detail, with reference to the appended drawings showing embodiments of the invention, in which:

[0049] Fig. 1 is a schematic illustration of the formation of circular RNA from a linear precursor RNA.

[0050] Figs. 2a-d are schematic illustrations of linear RNA precursors (Figs. 2a, 2c) according to embodiments of the present invention and exemplary RNA species (circular RNA and impurities) resulting from the circularization reaction (Figs. 2b, 2d).

[0051] Figs. 3a-b are schematic illustrations of linear RNA precursors (Fig. 3a) according to embodiments of the present invention and exemplary RNA species (circular RNA and impurities) resulting from the circularization reaction (Fig. 3b).

[0052] Fig 4a is a schematic illustration of a linear RNA precursor according to embodiments of the invention, and Fig. 4b illustrates shows a hypothetical formation thereof during the circularization.

[0053] Fig. 5 is a fragment analyzer image showing the results of circularization of three different RNA constructs according to embodiments of the invention, and their respective controls, as described in Experiment 1.

[0054] Fig 6a is a chart showing the % circular RNA, of total RNA, resulting from the circularization of a precursor according to embodiments of the invention at different time points, as described in Experiment 2. Fig. 6b shows the results in terms of % intermediate RNA for the same experiment.

[0055] Fig. 7a is a chromatogram demonstrating separation of RNA species (peak 1) and IVT components (nucleotides and proteins; peak 2) by size exclusion chromatography (SEC) described in Experiment 3. Fig. 7b is a fragment analyzer image showing the content of RNA fragments in the IVT input sample and SEC peaks 1 and 2. Fig. 7c is a fragment analyzer graph showing the fragment size of input IVT reaction sample (dotted line), the SEC elution peak 1 (dashed line) and the SEC elution peak 2 (solid line).

[0056] Fig. 8 is a chromatogram from the oligo(dT) affinity chromatography run of Experiment 3. Fig. 9a-c shows the results from RNA fragment analysis of samples from Experiment 3. Fig. 9a is an image generated by the Fragment Analyzer System showing the respective composition of the input sample, the oligo(dT) flowthrough fractions (FT) and the oligo(dT) eluate (E) fraction from the affinity chromatography run of Experiment 3 (500 mM KCI). Fig. 9b-c are graphs generated by the Fragment Analyzer System showing the fragment composition of total RNA (dotted line), oligo(dT) flowthrough fractions (solid line) and oligo(dT) eluate (dashed line).

[0057] Fig. 10a is an image of the fragment analyzer capillary electrophoresis gel showing the content of RNA fragments resulting from circularization of RNA precursors constructs having two homopolymeric moieties of 10, 25, 50 or 75 recurring units, respectively (Experiment 4). "CircGIuc" denotes the control, i.e. a precursor lacking homopolymeric moieties. Fig. 10b is a chart visualizing the relative content (%) of different RNA species based on total RNA. For each construct, the bars represent, in the order from left to right: precursor RNA, circular RNA, intermediate RNA.

[0058] Fig. 11 is a bar chart visualizing the content (%) of different RNA species based on total RNA resulting from Experiment 5 which investigated the effect of heating on circularization efficiency. For each construct, the bars represent, in the order from left to right: precursor RNA, circular RNA, intermediate RNA.

[0059] Fig. 12 is a bar chart showing the content (%) of different RNA species based on total RNA resulting from Experiment 6 which investigated the circularization efficiency of constructs having homopolymeric moieties of the same length but using different nucleotides (C, G A or U) as the recurring unit.

[0060] Fig. 13 is a bar chart showing the content (%) of different RNA species based on total RNA resulting from Experiment 7 which compared the circularization efficiency of RNA precursors containing one or two homopolymeric moieties, respectively.

[0061] As illustrated in the figures, some features may be exaggerated for illustrative purposes and, thus, are provided to illustrate the general structures of embodiments of the present invention. Like reference numerals refer to like elements throughout.

[0062] DETAILED DESCRIPTION

[0063] The present inventors unexpectedly found that the addition of a homopolymeric moiety to a non-circularizing part of a linear precursor RNA molecule increased the circularization efficiency and / or reduced the amount of impurities that require removal during subsequent purification. Furthermore, the inventors propose that circular RNA produced from such a precursor RNA molecule can be readily separated from RNA impurities in a flow-through chromatography method using a chromatography material capable of binding to the homopolymeric moiety.

[0064] The terms "circular RNA" (alternatively, "circRNA", "cRNA"), "circular RNA molecule" and "circular RNA sequence" refer to an RNA polynucleotide that has a closed-loop structure without a 5' or 3' end. Hence, a circular RNA does not comprise a 5' end or a 3' end.

[0065] The term "linear" in the context of an RNA molecule or fragment refers to an RNA molecule that has a 5' and a 3' end. Thus, "linear RNA" is an RNA molecule that contains a finite sequence of ribonucleotides. A linear RNA may adopt any spatial conformation or structure, as long as the 5' end and the 3' end are not joined.

[0066] The terms "linear precursor RNA", "linear precursor RNA molecule", "linear RNA precursor", "precursor RNA", "precursor" and the like herein refer to an RNA polynucleotide that is not circular, but has a 5' end and a 3' end, and additionally contains at least one sequence element that enables or facilitates a circularization reaction to create a circular RNA. The sequence element(s) that enable(s) circularization may be a self-splicing sequence, such as a self-splicing ribozyme element, and may optionally include auxiliary sequences such as homology arms or other splicing enhancers, e.g. exonic splicing enhancers. The sequence element(s) may comprise at least two catalytic subunits.

[0067] For example, self-splicing ribozyme subunits may be derived or engineered (e.g. comprising consensus sequences) from group I or group II catalytic introns or fragments thereof. One method of RNA circularization by self-splicing ribozyme subunits is the permuted intron-exon (PIE) splicing strategy which uses an exon pair flanked by permuted group I catalytic intron fragments. As schematically illustrated in Fig. 1 (prior art), the 5' end of a linear RNA sequence to be circularized contains a 3' self-splicing intron 11 and corresponding exon El forming a 3' splicing PIE fragment, and the 3' end of the RNA sequence to be circularized contains a 5' self-splicing intron 12 and corresponding exon E2 (forming a 5' splicing PIE fragment). Under circularization conditions, the linear precursor RNA containing the 3' and 5' PIE fragments undergo a catalytic reaction resulting in splicing of the introns and fusing of the respective 5' and 3' exon ends to form a circular RNA molecule C. A sequence of interest, designated to form part of the circular RNA molecule, is positioned between the exons El and E2 of the linear precursor, and forms part of the circular RNA molecule. The PIE elements may comprise a sequence derived from Cyanobacterium anabaena, T4 phage or Tetrahymena. In addition to the circular RNA molecules and the cleaved-off introns, intermediates and by-products such as nicked RNA are also formed (not shown).

[0068] The general structure of a linear precursor RNA molecule P according to the present invention is illustrated in Fig. 2a. The linear precursor RNA molecule has a 5' end and a 3' end. A sequence 30 intended to form the circularized RNA molecule, also referred to herein as a circularizing sequence or circularizing fragment, is flanked by a first self-splicing sequence 10 and a second self-splicing sequence 20. The first self-splicing sequence 10 is provided upstream of the circularizing sequence as seen in the 5'-to-3' direction and comprises a homopolymeric moiety 101 which may optionally include interrupting sequences as described below. The second self-splicing sequence 20 is provided downstream of the circularizing sequence as seen in the 5'-to-3' direction. In the embodiment represented in this figure, the homopolymeric moiety 101 is arranged at or close to the 5' end of the self-splicing sequence 10. However, it is envisaged that the homopolymeric moiety may instead be located in the second self-splicing sequence 20, in which case the first self-splicing sequence 10 may lack a homopolymeric moiety.

[0069] Each of the self-splicing sequences 10, 20 comprises a splice site adjacent to the circularizing fragment, indicated by arrows in the figure. The self-splicing sequence 10 and / or the self-splicing sequence 20 may comprise additional sequence elements, such as pairing sequences (also referred to as homology arms). Such additional sequence elements may be arranged between the homopolymeric moiety 101 and the circularizing fragment. Homology arms are commonly used to enhance circularization efficiency. Other examples of additional sequence elements include spacer sequences. Another example of an additional sequence element is a 5' start sequence, i.e. a short sequence (e.g. 3-10 or 3-8 nucleotides) marking the 5' end of the transcribed RNA sequence. Another example of an additional sequence element is a 3' end sequence of e.g. 3-10 or 3-8 nucleotides at the 3' end of the precursor.

[0070] Under circularization conditions the self-splicing sequences 10, 20 are cleaved from the linear precursor molecule, and the ends of the circularizing fragment 30 are joined so as to generate i) a closed circular RNA molecule C containing the circularizing sequence 30, and ii) cleaved fragments 10' and 20', as schematically shown in Fig. 2b. The cleaved fragment 10' comprises the homopolymeric moiety 101. In addition to the circular RNA molecule C and the cleaved fragments 10', 20', the circularization reaction typically also results in the formation of iii) intermediate forms of RNA, exemplified in this figure by a fragment I containing the first self-splicing sequence 10 and the circularization sequence 30, but lacking the second self-splicing sequence, which has been cleaved off. Other intermediate RNA species that may be formed during a circularization reaction may include a fragment having the circularization sequence and the second self-splicing sequence, i.e. where only the first self-splicing sequence has been cleaved off. Lastly, another impurity that may be formed is iv) nicked RNA 40. Advantageously, the present invention has been found to provide improved circularization efficiency and to result in lower levels of impurities such as precursor RNA and intermediate forms of RNA, compared to a control lacking homopolymeric moieties. Fig 2c illustrates another embodiment of a linear precursor RNA molecule P according to the present invention, which differs from the embodiment of Fig 2a in that the first self-splicing sequence 10 comprises a first homopolymeric moiety 101 and the second self-splicing sequence 20 comprises a second homopolymeric moiety 201. In the embodiment represented in this figure, the first homopolymeric moiety 101 is arranged at the 5' end of the self-splicing sequence 10, and the second homopolymeric moiety 201 is arranged at the 3' end of the self-splicing sequence 20. The homopolymeric moieties 10, 20 may be identical or non-identical nucleic acid sequences. The homopolymeric moieties 10, 20 may be non-complementary nucleic acid sequences, or otherwise uncapable of hybridizing to each other, e.g. due to the presence of modifications which prevent hybridization. Any additional sequence elements as described above with reference to Fig 2a may be arranged between the first homopolymeric moiety 101 and the circularizing fragment, and / or between the circularizing fragment and the second homopolymeric moiety 201, and / or arranged at the 5' end and / or 3' end or the precursor P.

[0071] Circularization of the precursor of Fig. 2c generates at least i) a closed circular RNA molecule C containing the circularizing sequence 30, and ii) cleaved fragments 10' and 20", as schematically shown in Fig. 2d. The cleaved fragments 10', 20" comprise the first homopolymeric moiety 101 and the second homopolymeric moiety 201, respectively. Intermediate and nicked RNA species may also be formed (not shown).

[0072] Figure 3a illustrates an embodiment of a linear precursor RNA comprising permuted intronexon pairs. The linear precursor RNA comprises first and second homopolymeric moieties 101, 201 as described above with reference to Fig. 2c-d, although it is envisaged that the precursor may alternatively contain only one homopolymeric moiety as described with reference to Fig. 2a. The first self-splicing sequence 10 comprises a 3' self-splicing intron 11, and the second self-splicing sequence 20 comprises a 5' self-splicing intron 12. The circularizing fragment 30 comprises a sequence of interest 301 flanked by an exon pair El, E2. In the linear precursor RNA the intron 11 of the first selfsplicing sequence is positioned adjacent exon El of the circularizing fragment, and the intron 12 is positioned adjacent exon E2 of the circularizing fragment. Under circularization conditions, splicing of the introns 11, 12 results in the formation of a circularized fragment where the free 5' end of the first exon El is joined to the 3' end of the second exon E2, as shown in Fig. 3b. The cleaved-off fragments contain the first homopolymeric moiety 101 and the self-splicing intron 11, and the second homopolymeric moiety 201 and the self-splicing intron 12, respectively. Any nicked RNA (not illustrated) would also contain the exons El, E2. Other non-circular intermediates (not illustrated) may contain exons El, E2 and additionally one of the introns 11 or 12. Fig. 4a illustrates a variant of this embodiment which additionally contains a first homology arm 102 positioned between the first homopolymeric moiety 101 and the intron 11, and a second homology arm 202 position between the intron 12 and the second homopolymeric moiety 201. In a variant (not illustrated) of this embodiment, the precursor RNA still contains homology arms 102, 202 as shown in Figs. 4a-b but lacks either one of the homopolymeric moieties 101, 201, i.e., the precursor in such embodiments comprises homology arms 102, 202 but only one homopolymeric moiety, which may be located either on the 5' of homology arm 102, or on the 3' side of homology arm 202.

[0073] The homology arms 102, 202 are RNA sequences that are capable of complementary binding to each other. Hybridization between the homology arms may assist in correctly positioning other sequence elements, such as PIE fragments, to facilitate the splicing and circularization event, which may increase circularization efficiency for a linear precursor RNA containing homology arms as compared to a precursor lacking homology arms.

[0074] It is believed that the present invention may be particularly useful in further improving the circularization efficiency of linear precursor RNA which contains external homology arms, i.e., homology arms which do not form part of the circularizing fragment, but which are contained in the parts of the molecule that are spliced off. Not wishing to be bound by any particular theory, it is hypothesized that the presence of two homopolymeric moieties, which do not bind to each other, could provide a leverage effect that promotes the splicing and subsequent circularization. Fig 4b shows a possible configuration of the RNA molecule during the circularization process. Alternatively, the homopolymeric moiety or moieties could serve to improve or facilitate the hybridization between the homology arms, e.g. by stabilizing the double helix formed, and thereby improve splicing. As demonstrated herein, also the presence of a single homopolymeric moiety provides improved circularization of precursor RNA (see Experiment 7).

[0075] The first homopolymeric moiety and the second homopolymeric moiety may be any homopolymeric moiety that can be attached to the linear precursor RNA without compromising the circularization. However, for the purpose of process economy, the homopolymeric moieties are preferably polynucleotides that can be produced together with the rest of the RNA precursor molecule, such as by in vitro transcription. Alternative homopolymeric moieties may for example be amino acid sequences, for instance where the RNA molecule is produced by chemical synthesis instead of enzymatically by IVT.

[0076] As used herein, "homopolymeric" refers to a polymeric moiety formed of recurring units (sometimed referred to as repeating units), where the recurring units are the same to an extent that is notable enough to distinguish it from the random variation of a natural polymer, such as a nucleic acid. For example, a homopolymeric moiety may have at least 80 % identical units. The number of identical units may be calculated based on the sequence of the homopolymeric moiety, and need not be all in a consecutive sequence. For instance, it is envisaged that a homopolymeric sequence may be an interrupted sequence, in which two or more segments of identical recurring units are separated by or one or more shorter, interrupting segments of one or more recurring units. The recurring units of such interrupting sequences need not be identical to each other. As an example, in the context of nucleic acid homopolymeric moieties, an interrupting sequence may be a restriction site. As another example, a first sequence of a repeating first unit Xmimay be followed by an interrupting unit Y or sequence [S], followed by a second sequence of said first unit Xm2, and these three segments may optionally be repeated. In the case of an interrupted sequence, the segments of identical recurring units may contain at least 3 identical units, such as at least 4 or at least 5 identical units. Thus, in such embodiments the indices ml and m2 may, independently and for each segment, represent an integer of at least 3, and up to e.g. 30 or 50. The interrupting unit or sequence may be at least one unit, such as a nucleotide, and would typically consist of fewer units than an individual segment of identical units.

[0077] According to one embodiment, the interrupted sequence may be according to an exemplary formula AXI-Y-AX2-Y-AX3. A denotes adenine, wherein said adenine may be a naturally occurring or a modified adenine and / or analogues or derivatives of adenine, as described below. Y denotes a sequence of nucleotides which comprises other nucleotides than adenine such as guanine, uracil, and cytosine which may be naturally occurring or comprise a modification, as described below. It is also possible that the Y segment comprises adenine as defined below, as long as also other nucleotides are comprised in the Y segment. Furthermore, Y may denote any other modified nucleotide as disclosed below, or a modified adenine as defined below. X represents an integer of at least 10, such as 10, or 15, or 20, or 25, or 30, or 35, or 40, or 45, or 50. The number of nucleotides comprised in Y is at least 1 or 2, and up to about half the amount of adenines in one adjacent Axsegment or both adjacent Axsegments, or more than half the amount of adenines in one adjacent Axsegment or both adjacent Axsegments, as long as the Y segment comprises less nucleotides than one adjacent Axsegment or both adjacent Axsegments. Each Axsegment may comprise different numbers of adenines. Each Y segment may comprise a different number of nucleotides. The sequence of Ax-Y may also be repeated multiple times. In one embodiment, an Axis the first and the final segment of the interrupted sequence. In one embodiment, a Y segment is the final segment of the interrupted sequence.

[0078] The advantage of including an interrupting sequence in the homopolymeric moiety is that it improves the transcription accuracy of the RNA during the IVT. A long, continuous homogenic polynucleotide tails is disadvantageous as the RNA polymerase will start to 'slip', a phenomenon called transcriptional slippage, which will result in the incorporation of an incorrect number of As. (see Beverly, et al., Analytical and Bioanalytical Chemistry (2018) 410:1667-1677)

[0079] Additionally, DNA encoding a long continuous homogenic polynucleotide can be genomically unstable and more recombination events are observed in bacterial culture during plasmid production. It was shown that this recombinatory effect can be reduced by using segmented PolyA tails. (see Trepotec, et al., 2019, RNA 25:507-518)

[0080] Notably, in the case of recurring nucleotides, "identical" units does not preclude the presence of various modifications on said nucleotides. For instance, a polyA sequence may include natural adenosine nucleotides as well as one or more modified versions or analogs of adenosine.

[0081] In embodiments, including also embodiments where a homopolymeric moiety does not comprise an interrupted sequence, a homopolymeric moiety may comprise or consist of at least 6 consecutive, identical recurring units. Hence, the first and / or the second homopolymeric moiety may comprise at least 6 consecutive identical units, such as at least 10, at least at least 12, at least 15, at least 20, at least 25, or at least 30 consecutive, identical recurring units. For a nucleic acid homopolymeric moiety, said recurring unit may be A, G, C or U, optionally including natural or synthetic modified variants thereof.

[0082] Furthermore, it is envisaged that a homopolymeric moiety may contain multiple homopolymeric segments, which do not need to be identical. To illustrate, a nucleic acid homopolymeric moiety may comprise a first segment, e.g. a polyA sequence, followed by a second segment, e.g. a polyG sequence, wherein each of said segments comprises at least 6, such as at least 10, at least 12, at least 15, at least 20, at least 25 recurring units (in this example A and G, respectively). However, where the first and second homopolymeric moieties are nucleic acids that are not entirely formed of the same nucleotide, the respective sequences should be chosen such that the first homopolymeric moiety and the second homopolymeric moiety do not hybridize under circularization conditions in a manner detrimental to splicing and circularization of the circularizing fragment.

[0083] Each of the first and the second homopolymeric moiety may have any suitable length, for instance up to 100 recurring units, such as up to 75 recurring units, optionally including several different homopolymeric segments. As an example, each of the first and second homopolymeric moieties may comprise or consist of at least one sequence selected from Ak, Uj, Cnand Gpand combinations thereof, wherein each of k, j, n and p independently represents an integer from 6 to 75, preferably 10-50, wherein said sequence optionally includes modified nucleotide variants. In embodiments where a homopolymeric moiety comprises a polyG segment, or at least a polyG segment of a certain length, such as 6 guanine nucleotides or more, the polyG segment is not located at the 5' end of the precursor molecule as such a polyG sequence may hinder initiation of transcription by T7 RNA polymerase. However, a polyG segment is not expected to hinder the T7 polymer during the elongation phase of transcription, and thus a polyG segment may be located elsewhere in the homopolymeric moiety or moieties, such as anywhere in a 3' positioned homopolymeric moiety, or positioned as a further segment, downstream of a non-polyG segment (e.g. a non-polyG 5' start sequence), of a 5' located homopolymeric moiety. For example, the linear precursor may have a 5' start sequence that contains no more than 5, such as no more than 3, consecutive G nucleotides. In embodiments, a polyG sequence may be an interrupted sequence. In embodiments, a 5' homopolymeric moiety does not comprise a polyG sequence of more than 3 consecutive guanine nucleotides.

[0084] In embodiments, the first homopolymeric moiety may comprise or consist of a polyA sequence of from 20 to 50, such as from 25 to 50, adenosine units.

[0085] The linear precursor RNA may contain pairing sequences, or homology arms, provided in the parts of the molecule that are cleaved off during splicing. The homology arms are RNA sequences that are capable of complementary base-pairing with each other. Hybridization between the homology arms may facilitate the splicing and circularization event to increase circularization efficiency for a linear precursor RNA. However, the homology arms should preferably not be complementary, or at least have less than 50% (e.g., less than 45%, less than 40%, less than 35%, less than 30%, less than 25%) base pairing with unintended sequences in the RNA (e.g., non-homology arm sequences). Homology arms may be synthetic sequences.

[0086] Homology arms may be provided anywhere in the self-splicing sequences 10, 20. In embodiments of the invention comprising a first homopolymeric moiety, a first homology arm may be arranged downstream of the first homopolymeric moiety in the 5'-to-3' direction. In embodiments of the invention comprising a second homopolymeric moiety, a second homology arm may be arranged upstream of the second homopolymeric moiety in the 5'-to-3' direction.

[0087] The RNA precursor may contain one or more spacer sequences. A spacer sequence, or a plurality of spacer sequences may be contained in the circularizing fragment and / or the self-splicing sequences. For example, the first self-splicing sequence 10 and / or the second self-splicing sequence 20 may contain one or more spacer or linker sequences. For example, a PIE fragment may contain a spacer sequence. A homology arm may be joined to an adjacent homopolymeric moiety via a spacer sequence. For example, in the 5'-to-3' direction, a first homopolymeric moiety may be followed by a spacer sequence of e.g. 10-20 nt, followed by the first homology arm and intron / exon region. At the 3' end of the circularizing sequence and the other intron / exon region, a second homology arm may be followed by a spacer sequence of e.g. 10-20 nt, followed by a second homopolymeric moiety.

[0088] The circularizing fragment contains the target sequence of interest to provide in the form of a circular RNA molecule. The target sequence of interest may comprise a functional sequence selected from a protein encoding sequence, a regulatory RNA sequence, protein binding RNA sequence, a guide RNA sequence, and combinations thereof.

[0089] The DNA sequence serving as template to generate the precursor RNA may thus contain one or more sequences encoding proteins or sequences encoding other functional RNAs. The DNA template may contain codon-optimized sequences or sequences optimized for reduced immunogenicity such as uridine depletion.

[0090] In embodiments, the sequence of interest may comprise a protein coding sequence and typically also at least one regulatory sequence, such as an internal ribosome entry site (IRES). In some embodiments the sequence of interest may contain multiple protein coding sequences, wherein each protein coding sequence may be preceded by a regulatory sequence upstream of the coding sequence. Where the sequence of interest comprises multiple protein encoding regions, the protein encoding regions may encode the same or different polypeptides.

[0091] Optionally, the sequence of interest may additionally contain untranslated regions (UTRs), such as a 3' UTR and / or a 5' UTR.

[0092] A protein encoding sequence may encode a polypeptide. The polypeptide may be a therapeutic polypeptide. In embodiments, the polypeptide may comprise an antibody heavy chain, an antibody light chain, a single-chain antibody, an antibody fragment, an antibody mimetic protein, an enzyme or a cytokine. In embodiments, the polypeptide may be an antigen, such as an antigen derived from a virus, such as an influenza virus, coronavirus (e.g., SARS-CoV-1 , SARS-CoV-2, or MERS- related virus), Ebola virus, Dengue virus, human immunodeficiency virus (HIV), hepatitis A virus (HAV), hepatitis B virus (HBV), hepatitis C virus (HCV), herpes simplex virus (HSV), respiratory syncytial virus (RSV), rhinovirus, cytomegalovirus (CMV), zika virus, human papillomavirus (HPV), human metapneumovirus (hMPV), human parainfluenza virus type 3 (PIV3), Epstein-Barr virus (EBV), or chikungunya virus. Alternatively, the antigen may be derived from a bacterium, such as Staphylococcus aureus, Moraxella, Chlamydia trachomatis, Borrelia, Bacillus anthracis, Salmonella typhi, Mycobacterium tuberculosis, Propionibacterium acnes, or Haemophilus influenzae. The circular RNA molecule produced as described herein may be used as a vaccine.

[0093] A translation regulatory sequence may be provided upstream of a protein encoding region as seen in the 5'-to-3' direction of the linear RNA precursor. Alternatively, a translation regulatory sequence may be provided downstream of a protein encoding region as seen in the 5'-to-3' direction of the linear precursor. In the latter case, the translation regulatory sequence may still be adequately positioned in relation to the protein encoding region in the circularized RNA. Examples of translation regulatory sequences include internal ribosome entry site (IRES) sequences and riboregulatory sequences, such as riboswitch sequences.

[0094] IRES sequences useful in the present context are known in the art. For example, an IRES sequence may be derived from Coxsackievirus, Encephalomyocarditis virus (EMCV), Dicistroviruses, hepatitis C virus (HCV), poliovirus (PV), enterovirus 71 (EV71 ), human rhinovirus (HRV), foot-and- mouth disease virus (FMDV), or may be a consensus sequence or a synthetic sequence engineered e.g. to provide improved stability and translatability.

[0095] Optionally, a spacer sequence may be present within the circularizing fragment at any suitable position between the exons El, E2 of the circularizing fragment, such as between the exon El and the sequence of interest, and / or between the sequence of interest and the exon E2.

[0096] Optionally, any part of the linear precursor molecule, such as the circularizing fragment or part thereof (e.g. the sequence of interest, a protein encoding sequence, a regulatory sequence, an IRES sequence, a spacer sequence or a PIE element) or the self-splicing sequences or any part thereof (e.g. a PIE element, a pairing sequence, a spacer sequence or a homopolymeric moiety) may independently comprise one or more modified nucleotides, analogues or derivatives of purines and pyrimidines. Such modified nucleotides, analogues or derivatives may be present at any suitable content, such as from about 1 mol% to 30 mol%.

[0097] Modified nucleotides may have natural modifications or may be synthetic, non-naturally occurring modifications. Modified nucleotides, analogues or derivatives of purines and pyrimidines may include e.g. 1-methyl- adenine, 2-methyl-adenine, 2-methylthio-N-6-isopentenyl-adenine, N6- methyl-adenine, N6- isopentenyl-adenine, 2-thio-cytosine, 3-methyl-cytosine, 4-acetyl-cytosine, 5- methyl-cytosine, 2,6-diaminopurine, 1-methyl-guanine, 2-methyl-guanine, 2,2-dimethyl-guanine, 7- methyl- guanine, inosine, 1-methyl-inosine, pseudouracil (5-uracil), dihydro-uracil, 2-thio-uracil, 4- thio-uracil, 5- carboxymethylaminomethyl-2-thio-uracil, 5-(carboxyhydroxymethyl)-uracil, 5-fluoro- uracil, 5-bromo-uracil, 5-carboxymethylaminomethyl-uracil, 5-methyl-2-thio-uracil, 5-methyl- uracil, N-uracil-5-oxy acetic acid methyl ester, 5-methylaminomethyl-uracil, 5- methoxyaminomethyl-2-thio- uracil, 5'-methoxycarbonylmethyl-uracil, 5-methoxy-uracil, uracil-5-oxyacetic acid methyl ester, uracil-5-oxyacetic acid (v), 1-methyl-pseudouracil, queosine, p-D-mannosyl-queosine, phosphoramidates, phosphorothioates, peptide nucleotides, methylphosphonates, 7- deazaguanosine, 5-methylcytosine, N6-methyladenosine, and inosine.

[0098] Examples of modified nucleotides include pseudouridine, Nl-methylpseudouridine, 2- thiouridine, 4' -thiouridine, 5- methylcytosine, 2-thio-l- methyl-l-deaza-pseudouridine, 2-thio-l- methyl-pseudouridine, 2-thio-5-aza-uridine, 2-thio- dihydropseudouridine, 2-thio-dihydrouridine, 2- thio-pseudouridine, 4-methoxy-2-thio-pseudouridine, 4-methoxy-pseudouridine, 4-thio-l-methyl- pseudouridine, 4-thio-pseudouridine, 5-aza-uridine, dihydropseudouridine, 5-methyluridine, 5- methyluridine, 5-methoxyuridine, and 2'-O-methyl uridine.

[0099] The circularizing sequence may have a size in the range of 20-10000 nucleotides (nt), such as 20-1000 nt, such as 50-1000 nt, such as 100-1000 nt, such as 200-10000 nt, such as 200-6000 nt, such as 200-4500 nt, such as 1000-10000 nt, or 1000-4500 nt. The linear precursor RNA molecule is longer than the circularizing fragment, and may comprise tens or hundreds of nucleotides in addition to those forming the circularizing sequence. In the experimental examples described herein, the precursor contains an additional 323 nucleotides including homology arms but excluding the length of any homopolymeric moieties. Where present, each homology arms may comprise e.g. 10-50 nucleotides, such as from 12 to 50 nucleotides. Furthermore, each of the homopolymeric moieties may comprise from 6 and up to at least 75 nt, such as 10-75 nt or 10-50 nt.

[0100] The linear precursor RNA molecule may produced by any suitable method known in the art, but may advantageously be produced in a cell-free system by in vitro transcription (IVT). The production of RNA by IVT is a well-established enzyme-based reaction, where a DNA dependent RNA polymerase (generally T7, SP6 or T3 RNA polymerase) catalyses the synthesis of the target mRNA from nucleotide triphosphates (NTPs) substrates using a DNA template encoding the RNA precursor. In addition, the IVT reaction typically requires other components such as polymerase cofactor Mg2+(e.g., provided in the form of MgCk or MgOAc), and suitable buffer conditions and pH. Additional components and reagents may also be used, depending on the process design, for example pyrophosphatase, RNase inhibitor and / or a zwitterionic or non-ionic detergent (e.g., TritonX, Tween20 or Tween80). The DNA template is produced in advance, usually by enzymatic linearization of purified plasmid DNA, but it can also be produced by amplification of the region of interest using PCR, or produced by rolling circle amplification, or may be chemically synthesized. The DNA template typically includes a promoter sequence.

[0101] To produce circular RNA from the linear precursor RNA molecule described herein, the linear precursor RNA is incubated at circularization conditions, which allow the self-splicing sequences to cleave off and the circularizing fragment to form a closed circular RNA molecule. In detail, a method of generating circular RNA may thus comprise providing a composition comprising a linear precursor molecule as disclosed herein in a buffer environment and incubating said composition to allow the circularization to occur.

[0102] The composition may be the IVT reaction mixture. Hence, the linear precursor may be produced by IVT and undergo splicing and circularization continuously and in parallel, in the same reaction vessel. In such embodiments, incubation may be considered to start once the IVT reactants (at least DNA template, ribonucleoside triphosphates, RNA polymerase and necessary cofactors) have been mixed in a suitable aqueous environment under conditions allowing the splicing and circularization event.

[0103] As used herein, the term "circularization conditions" refers to the minimum conditions required for circularization to occur. Generally, circularization may be facilitated by an aqueous, buffered environment having a pH in the range of from 7 to 10, such as from 7 to 9.5, such as from 7.5 or from 7.9 to 9.5. An example of a suitable buffer is Tris pH 7.5. The incubation temperature may be in the range of from 20 °C to 56 °C. Circularization may also be also facilitated by the presence of GTP and / or ATP, such as at least 2mM GTP. Circularization may also be facilitated by the presence of Mg2+, such as at least 10 mM Mg2+. During incubation under such conditions, a linear precursor RNA comprising group I catalytic introns can undergo the double transesterification event that is characteristic for group I catalytic introns in the presence of co-factors Mg2+and GTP.

[0104] The incubation time is not particularly limited, as the circular RNA molecule starts to form within minutes and is typically stable under circularization conditions and may be kept under such conditions for an extended period of time, e.g. a week, without degrading. However, for example, in the present method, the incubation time may be from 5 minutes to 24 hours, such as from 10 minutes to 24 hours, from 20 minutes to 24 hours, from 15 minutes to 24 hours, from 10 minutes to 12 hours, from 15 minutes to 12 hours, from 20 minutes to 12 hours, , from 10 minutes to 6 hours, from 20 minutes to 6 hours, from 30 minutes to 6 hours, such as 30-180 minutes, such as 30-120 minutes.

[0105] The incubation temperature may be from 20°C to 56°C, and may preferably be from 28°C to 56 °C, such as about 37°C. Incubation at temperatures higher than 37 °C, at least for a short period of time, may increase the circularization efficiency.

[0106] In embodiments, the incubation may be performed at a temperature of in the range of from 20°C to 37°C for a time period of 2-5 hours.

[0107] As the circularization progresses, the amount of circular RNA in the composition increases. However, since the circularization efficiency is not 100 %, the incubated composition typically still contains linear precursor RNA, in addition to the circular RNA and the cleaved self-splicing sequences 10' and 20' (cf. Fig. 2b). Furthermore, the incubated mixture often contains other impurities, such as intermediate and nicked RNA fragment. Surprisingly, the present inventors found that the circularization efficiency was significantly increased by providing a linear precursor RNA molecule having homopolymeric moieties as disclosed herein. For example, under favourable (but not necessarily optimized) conditions, the invention was seen to provide a substantial increase of circular RNA content, and reduced formation of intermediate RNA, during the production of circular RNA from a linear precursor over an incubation time period of 30-120 minutes (see Fig. 6a).

[0108] After the desired incubation time, the circular RNA is to be separated from other reactants and reaction products. In embodiments where the incubated composition is an IVT mixture, the circular RNA must be separated from the IVT reactants, such as DNA template, RNA polymerase and free NTPs, in addition to the linear precursor RNA, and other impurities. Generally, plasmid DNA is subjected to enzymatic degradation into smaller fragments by addition of DNase, in order to assist size-based separation. Hence, a reaction mixture comprising circular RNA and other components as described herein may be subjected to size-based separation e.g. by tangential flow filtration (TFF) or size exclusion chromatography (SEC), e.g. using a size exclusion chromatography resin such as Sepharose 6 Fast Flow (Cytiva) with a sample load of up to 0.2 column volumes. The SEC must not be HPLC, as HPLC has limited scalability and requires use of toxic solvents. By such a size-based separation step, the RNA species may be separated from smaller components such as NTPs, digested DNA template, and optionally RNA polymerase.

[0109] However, size-based methods are typically not able to separate the target circular RNA from similar-sized RNA impurities. To this end, the invention provides the possibility to use affinity chromatography using a chromatography material functionalized with a ligand which has a binding affinity for the homopolymeric moiety. By contacting a composition comprising the circular RNA and least one impurity (e.g. linear precursor RNA, cleaved self-splicing sequences, and / or intermediate RNA) with such a chromatography material, the circular RNA, which does not have the homopolymeric moieties and is unable to bind to the ligand, can be obtained in a flow-through fraction while the linear precursor and other impurities which contain the homopolymeric moieties are bound by the chromatography material. After collecting the circular RNA, the bound impurities can be eluted by conventional means and the chromatography material can be regenerated and used for a new purification cycle, if desired.

[0110] The chromatography material comprises a support material and at least one type of ligand coupled to the support material, the at least one ligand having a binding affinity for the first and second homopolymeric moieties. Preferably the material is functionalized with at least one ligand capable of binding the homopolymeric moiety, and if needed a second ligand capable of binding to the second homopolymeric moiety.

[0111] As used herein, the term "ligand" is a molecule that has a known or unknown affinity for a given entity. "Affinity ligand" refers to a moiety or molecule that binds reversibly and selectively or preferentially with high affinity to a target entity through a specific interaction with a binding site of the component. An affinity ligand may be immobilized to a support material. The support may optionally be provided in other forms such as a fiber, a membrane, a fibrous matrix, a filter, a porous monolith, a particle or a bead, such as a gel bead as used in chromatography resins. Particles or beads can be porous or non-porous. Particles or beads may include magnetic beads. Supports in the form of beads or particles can be used as a packed bed or in a suspended form. Suspended forms include those known as expanded beds and pure suspensions, in which the particles or beads are free to move.

[0112] With regard to specific materials, the support may comprise a polymeric material. Polymeric materials include materials of natural or synthetic polymers, and combinations thereof. For example, the support may comprise a polyhydroxy polymer, such as a polysaccharide. Examples of polysaccharides include e.g. dextran, starch, cellulose, pullulan, agar, agarose etc, including derivatives thereof. Polysaccharides are inherently hydrophilic with low degrees of nonspecific interactions, they provide a high content of reactive (activatable) hydroxyl groups and they are generally stable towards alkaline cleaning solutions used in bioprocessing. The support may comprise agar or agarose, or derivatives thereof. Optionally the support may comprise crosslinked agarose. Suitable supports can easily be prepared according to standard methods, such as inverse suspension gelation (S Hjerten: Biochim Biophys Acta 79(2), 393-398 (1964)). In embodiments, especially advantageous for large-scale separations, polysaccharide-based beaded support can be adapted to increase its rigidity using the methods described in US6602990 or US7396467 which renders the matrix more suitable for high flow rates.

[0113] Synthetic polymers useful as material for the support include polyvinyl alcohol, polyhydroxyalkyl acrylates, polyhydroxyalkyl methacrylates, polyacrylamides, polymethacrylamides etc. In case of hydrophobic polymers, such as matrices based on divinyl and monovinyl-substituted benzenes, the surface of the matrix can be hydrophilized to expose hydrophilic groups as defined above to a surrounding aqueous liquid. Such polymers are easily produced according to standard methods. As an alternative, a commercially available product, such as SOURCE™ (Cytiva) may be used as the support material.

[0114] Alternatively, the solid support according to the invention comprises a support of inorganic nature, e.g. silica, zirconium oxide etc.

[0115] In embodiments, the ligand may be coupled to a support which is a convection-based chromatography matrix, meaning a matrix that allows convective flow of liquids. Such convectionbased chromatography matrix may be a porous polymer membrane, a filter, a fibrous matrix, or a porous monolith. Examples of a porous polymer membrane include Mustang™ membranes (Cytiva) and Sartobind™ membranes (Sartorius). A fibrous support may be based on electrospun polymeric fibers or cellulose fibers. A fibrous support may optionally comprise non-woven fibers. A fibrous matrix may thus be a non-woven fibrous matrix. Fibers may have a cross-sectional diameter of 10- 1000 nm, such as 200-800 nm, 200-400 nm or 300-400 nm. Such a fibrous support can be found in a HiTrap Fibro™ unit (Cytiva). Alternative fibrous supports are disclosed in e.g. WO2019137869A1 and W02018011600A1.

[0116] The ligand may be attached to the support via known coupling techniques utilizing e.g. thiol, amino and / or carboxy groups present in ligand. Bisepoxides, epichlorohydrin, CNBr, N- hydroxysuccinimide (NHS) etc. are well-known coupling reagents. Between the support and the ligand, a spacer molecule can be introduced, which improves the availability of the ligand and / or facilitates the chemical coupling thereof to the support.

[0117] Where the homopolymeric moieties comprise homopolynucleotides, the affinity ligand typically comprises an oligonucleotide that is complementary to at least a portion of the respective homopolymeric moiety. For example, where the homopolymeric moiety or moieties comprise(s) a polyA sequence, the ligand may be an oligo(dT) ligand. Chromatography materials having oligo(dT) ligands are known in the art, and are commercially available e.g. as CIMmultus® Oligo dT (Sartorius BIA Separations). As another example, Fibro™ membrane chromatography materials functionalized with oligo(dT) ligands, and methods of performing separation methods using such materials, are described in WO2022162018A1 and WO2024022894A1 which are incorporated herein in their entireties. An oligo(dT) ligand may be a (dT)io-so ligand, preferably a (dT)i2-3o ligand, such as a (dT)2o ligand.

[0118] Alternatively, the ligand may be another type of molecule that is capable of binding to the homopolymeric moiety. For example, the ligand may be an immobilized metal affinity chromatography (IMAC) ligand, which is capable of binding to e.g. a polyA-containing moiety. An IMAC ligand is typically charged with a metal ion such as a nickel or zinc ion.

[0119] The circular RNA may be separated from at least one non-circular or linear RNA species comprising at least one homopolymeric moiety, by a method comprising a) Providing a composition comprising the circular RNA and the linear precursor RNA, and optionally other impurities. The composition may e.g. be a crude IVT mixture or a partially purified IVT mixture. b) Contacting the composition with a chromatography material comprising a ligand coupled to a support material, said ligand having a binding affinity for the homopolymeric moiety, under conditions allowing the ligand to bind to the homopolymeric moiety. Binding conditions may comprise a pH in the range of 6.5-8, such as 7-8, and may comprise the presence of a salt, e.g. KCI or NaCI, at a concentration of 50-1000 mM, such as 100-600 mM or 200-500 mM. For example a binding buffer used in step a) may be 10 mM Tris buffer pH 7.5, and EDTA ImM and KCI in the range of 200-500 mM. Other suitable buffers may include HEPES buffer and phosphate buffer pH 7.5. c) obtaining the circular RNA in a flowthrough fraction. This step may comprise washing the chromatography material with an aqueous phase, such as a binding or wash buffer, which does not disrupt the interaction between the ligand and the homopolymeric moiety of the linear RNA species, that is, does not result in elution of bound entities. The circular RNA may be collected in one or more flow-through fractions forming a sample enriched in circular RNA. This sample may be subjected to further purification, conditioning (e.g., buffer exchange), concentration and / or processing steps, as needed depending on the requirements on the final circular RNA product. For example, the sample may be subjected to one or more steps of TFF, encapsulation with a pharmaceutical carrier, such as a lipid nanoparticle, and / or sterile filtration. d) Optionally, after obtaining the circular RNA, bound non-circular RNA species may be eluted from the chromatography material using an elution buffer which provides conditions interrupting the interaction between the ligand and the homopolymeric moiety. For example, the elution buffer may have a lower conductivity, or salt concentration, than the binding / wash buffer. An exemplary elution buffer contains 10 mM Tris, 1 mM EDTA, pH 7.5. Where recovery of the bound impurities is of no interest, the elution step may be omitted. e) Optionally, after obtaining the circular RNA, the chromatography material may be subjected to cleaning-in-place (CIP) as known in the art, using for instance an alkaline solution such as 0.1 M NaOH, or a solution of a chaotropic salt such as 3 M guanidine-HCI or 5 M urea. The CIP may be performed directly after obtaining the circular RNA, thus omitting a step of eluting bound RNA species. Alternatively, if it is desirable to collect the bound species, CIP may be performed after eluting and collecting the eluate.

[0120] Lastly, after eluting the bound impurities and / or cleaning the chromatography material by CIP, the chromatography material can be regenerated and optionally used for a new separation cycle by repeating steps b-c, steps b-c-d, or steps b-c-d-e as desired.

[0121] In an embodiment, the process for separation of a desired circular RNA from a non-circular RNA comprises contacting the composition with a chromatography material comprising a ligand coupled to a support material, said ligand having a binding affinity for the homopolymeric moiety, under conditions allowing the ligand to bind to the homopolymeric moiety, followed by a step of TFF or SEC. Alternatively, the TFF or SEC may be performed before the contacting of the composition with a chromatography material comprising a ligand coupled to a support material, said ligand having a binding affinity for the homopolymeric moiety, under conditions allowing the ligand to bind to the homopolymeric moiety. Thus, any of the order Affinity chromatography - SEC / TFF or SEC / TFF - Affinity chromatography may be performed. Regarding all other details or process steps, the process above will apply.

[0122] While the invention is described herein with respect to exemplary embodiments, the skilled person will appreciate that the invention is not limited to these. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim.

[0123] Use of the verb "to comprise" and its conjugations does not exclude the presence of elements or steps other than those stated. The article "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0124] EXAMPLES

[0125] The invention is exemplified by the following experimental examples using RNA produced by IVT and containing three different coding regions and combinations of homopolymeric moieties.

[0126] DNA templates and RNA construct nomenclature

[0127] The DNA template sequence to be transcribed into RNA was based on the intron / exon sequence named "Ana2.0 (full)" described in the Supplementary information of Wesselhoeft et al., Nature Communications (2018)9:2629. Each DNA template construct contained a T7 promotor sequence followed by a AGG codon and the 5' part of the Ana2.0 sequences, which included the complementary 19 nt homology arm. The IRES sequence and the protein coding region were framed by the 'permissive spacer' 1 and 2 from the same publication. The 3' intron / exon region, also including the homology arm, was followed by a restriction enzyme site for template linearization. Different homopolymeric moieties were added after the AGG codon (5') and before the restriction enzyme site (3') on either both or single ends by conventional restriction enzyme cloning. Table 1 presents selected sequence elements of the of the self-splicing sequences of the precursor RNA molecules used in the present examples. Table 1. Selected transcribed DNA template sequence elements

[0128] As coding regions used in the present examples, sequences encoding Green Fluorescent Protein (GFP), Gaussia Luciferase (GLuc) or Firefly Luciferase (FLuc) were used. In the Examples below the RNA constructs are named with reference to the respective coding region and 5' and 3' homopolymeric moieties, as follows:

[0129] Circfcoding region)-(first homopolymeric moiety) / (second homopolymeric moiety)

[0130] For example, a construct with a circularizing sequence comprising a sequence encoding GLuc and having a homopolymeric moiety of 25 adenosines located 5' of the circularizing sequence, and a homopolymeric moiety of 25 adenosines located 3' of the circularizing sequence is referred to as "circGLuc-A25 / A25".

[0131] Table 2 indicates the sizes of circular RNA molecules and the respective precursor RNA molecules features in the examples below. Table 2. 1

[0132] Plasmid propagation and purification

[0133] The template plasmid was propagated in TOPIO E.coli cells in LB (for Maxiprep) or TB (for Gigaprep) medium supplemented with neomycin antibiotics (final concentration 25 pg / mL) until stationary phase was reached at 37 °C. After cell harvest by centrifugation the plasmid was isolated by either by QIAGEN Plasmid Giga kit or HiSpeed Plasmid Maxi kit according to the manufacturer's recommendations.

[0134] Plasmid linearization and purification

[0135] Prior to IVT, DNA templates were linearized by EcoRI digestion and purified by isopropanol precipitation and resuspended in in RNase-free water.

[0136] In vitro transcription

[0137] In vitro transcription (IVT) was performed using the following IVT components: 7.5 mM of each nucleotide (A, C, G, U, all unmodified), 2 U / pl RNase inhibitor, 0.002 U / pl Pyrophosphatase, 4 U / pl T7 polymerase and 150 pg / mL EcoRI-linearized DNA template in RNase-free water supplemented with IVT buffer (final concentration was lx). The lOx IVT buffer contained 400 mM Tris-HCI, pH 8.0, 500 mM magnesium acetate, 100 mM DTT and 0.2% TritonX-100. After mixing the components, the IVT reaction mixture was incubated at 37 °C in a heat block shaking with 1000 rpm for 2.5 h.

[0138] After the incubation, DNase I was added (final cone. 1.5 U / pg DNA) followed by incubation for 30 min at 37 °C in a heat block shaking with 1000 rpm. After DNase I treatment, EDTA was added to a final concentration of 60 mM. Analysis of RNA fragments

[0139] Capillary electrophoresis of IVT samples was performed with an Agilent 5200 Fragment Analyzer System with associated ProSize Data Analysis software (Agilent Technologies) using the instrument manufacturer's RNA Kit (15 NT). The IVT samples were diluted 1:25 and heated for 2 min to 70 °C. 2 pl of the heated mixture were transferred into 22 pl diluent marker before analysis using the DNF-471E33-SSTotalRNA 15 nt method provided by the instrument manufacturer.

[0140] Total RNA concentration was determined by size exclusion chromatography as follows: The IVT sample (subjected to DNase I treatment and EDTA addition, as described above) was diluted 50x with 10 mM Tris pH 7.5 and 1 mM EDTA, prior to loading of 2 mL sample onto a Prepacked Prep 26 / 10 Sepharose 6 Fast Flow SEC column (Cytiva). Absorbance was recorded at 260 nm during the runs. The RNA concentration in the injected samples was determined from the measured absorbance values at 260 nm in the chromatogram by the peak area. The corresponding RNA mass (in mg) present in the injected volume was calculated using the mass extinction coefficient of singlestranded RNA (mg1mL cm1) with a 0.2 cm path length UV cell.

[0141] Experiment 1: Addition of polyA stretches improves circularization efficiency of RNA of various sizes

[0142] Homopolymeric stretches of 50 adenosine units (i.e., polyAso) were added to three different mRNAs by cloning as described above and compared to their respective controls with respect to circularization efficiency.

[0143] In detail, three constructs circGLuc-A50 / A50, circGFP-A50 / A50 and circFLuc-A50 / A50 were compared to their respective controls (circGIuc-AO / AO, circGFP-AO / AO and circFLuc-AO / AO). After IVT the circularization efficiency was analyzed by capillary electrophoresis as described above. Fig. 5 shows the results obtained from the fragment analyzer system. The results indicate that based on total RNA, all three polyA-tailed constructs exhibited an increase in % circular RNA for all of the investigated constructs and reduced level of linear precursor, as presented in Table 3. Hence, the circularization efficiency was not influenced by difference in construct size. Concentration analysis indicated that IVT yields for the polyA-containing constructs were slightly below the control values.

[0144] Table 3 Experiment 2: Construct with homopolymeric stretches has similar IVT kinetics and less impurities compared to the control

[0145] This example compared the IVT kinetics of circular RNA production for constructs with and without homopolymeric stretches.

[0146] In detail, circGLuc and circGLuc-A50 / A50 constructs were produced by IVT as described above with incubation time extending up to 2h. The circularization efficiency was analyzed over time at 30 min intervals up to 2 h of incubation during the IVT. Capillary electrophoresis indicated that at all time points tested, the percentage of circular RNA was higher for the circGLuc-A50 / A50 construct than for the control. Fig. 6a is a bar chart showing the % circular RNA achieved at 30 minutes, 60 minutes, 90 minutes and 120 minutes incubation time, respectively. As can be seen in this figure, at all tested time points the percentage of circular RNA was remarkably higher for the circGLuc- A50 / A50 construct compared to the circGLuc control, and the constructs seemed to follow a similar kinetics over time.

[0147] Table 4 shows the content of undesired intermediate RNA at each tested time point for the circGLuc and the circGLuc-A50 / A50 constructs, respectively. The circGLuc-A50 / A50 construct resulted in much lower levels of intermediate RNA compared to the control at all tested time points. The data is graphically represented in Fig 6b.

[0148] Table 4. % intermediate RNA

[0149] Experiment 3: Purification of circular RNA by chromatography

[0150] This example demonstrates purification of circular RNA by a two-step process: size exclusion chromatography followed by oligo(dT) affinity chromatography.

[0151] The circGLuc-A50 / A50 material produced in Experiment 2 (incubation time 2 h) was purified by a two-step protocol. The IVT material, which contained at least four different species of RNA (circular RNA, precursor RNA, intermediate RNA, and cleaved intron fragments) was applied on a Sepharose 6FF chromatography column (Cytiva) to remove the free nucleotides and IVT proteins. Briefly, after EDTA addition to the IVT reaction mixture as described above, the IVT material was loaded onto a Prepacked Prep 26 / 10 Sepharose 6 Fast Flow column (Cytiva). The sample loading volume was 2 mL, the buffer was 10 mM Tris pH 7.5 with 1 mM EDTA, flow velocity was 510 mL / min, and run time approximately 10 minutes. The results are shown in Figs. 7a-c. As can be seen in the chromatogram of Fig. 7a, two main elution peaks were obtained, of which peak 1 (at around 20 mL) contained the RNA species, and peak 2 (at around 50 mL) contained the smaller components, such as protein, and nucleotides, including digested DNA template. In this SEC step, as expected, no separation of the different RNA species was achieved, as shown by capillary electrophoresis, Fig 7b-c. In Fig. 7b, the strong band slightly below 1500 nt represents the circular RNA. A faint band between the precursor and the circular species represents the intermediate RNA species. Fig. 7c plots the fragment size of input IVT sample (dotted line), elution peak 1 (dashed line) and elution peak 2 (solid line). In conclusion, the SEC purification yields a mixture of RNA species (containing precursor, circular RNA, introns and other aberrant species) free of smaller contaminants such as nucleotides and IVT proteins.

[0152] Next, the RNA-containing peak 1 fraction from the SEC column was subjected to affinity chromatography using a Fibro chromatography membrane (Cytiva) functionalized with an oligo(dT)zo affinity ligand as described in WO2022162018A1 and WO2024 / 022894. A 6 ml sample with a total RNA concentration of 0.25 mg / ml, preconditioned with KCI to a final concentration of either 500 or 200 mM were prepared for oligo(dT) purification. The samples were loaded in binding buffer with 10 mM Tris, pH 7.5, 1 mM EDTA and either 500 or 200 mM KCI with 0.8 ml / min (corresponds to 30 s residence time). After washing with for 4 column volumes (CV) with binding buffer and 6 CV in 60 % binding buffer (corresponds to 300 mM or 120 mM KCI respectively) the bound fraction was eluted in 10 mM Tris, pH 7.5, 1 mM EDTA. The total run time was approximately 10 min.

[0153] The oligo(dT) affinity chromatography was performed twice with either 200 or 500 mM KCI in the binding buffer with similar results, as shown in the chromatogram of Fig 8, in which FT denotes the flowthrough and E denotes the eluate. Four fractions of the flow-through and the eluate fraction (indicated by arrows is Fig. 8) were analyzed by capillary electrophoresis as described above and compared to the crude IVT input sample (lane 2). As shown in Fig. 9a the flowthrough fractions (lanes 3-6) contained pure circular RNA, while the other species were found in the eluate (lane 7), which contained very little circular RNA. Lane 2 shows the crude IVT material and lane 1 is the ladder. Fig. 9b (500 mM KCI) and Fig. 9c (200 mM KCI) show a graphical representation of the same results for the input IVT sample, the first flow-through fraction, and the elution fraction (as indicated by arrows in Fig.9a). The peak at 1406 nt contains circular RNA whereas the peak at 1813 nt represents the precursor. The small shoulder between the circular RNA peak and the precursor peak represents non-circularized intermediate RNA. As seen from these figures, the flow-through fraction contains no precursor RNA or intermediate RNA. 4: Circularization efficiency of homopolymeric moieties having different lengths

[0154] This example compared the circularization efficiency of RNA constructs having polyA homopolymeric moieties of different lengths ranging from 10 to 75 adenosine units. Five different constructs, circGLuc-AlO / AlO ("A10 / A10" in Fig. 10b), circGLuc-A25 / A25 ("A25 / A25" in Fig. 10b), circGLuc-A50 / A50 ("A50 / A50" in Fig. 10b), circGLuc-A75 / A75 ("A75 / A75" in Fig. 10b) and control circGLuc ("A0 / A0" in Fig. 10b), produced by IVT as described above were analyzed by capillary electrophoresis. RNA concentration was measured after 2.5h IVT incubation time. Fig. 10a shows the result of capillary electrophoresis, and Fig. 10b shows the concentration of different RNA species in each sample (for each sample, the bars represent in the direction from left to right: precursor, circular, intermediate). It was seen that all constructs containing polyA homopolymeric moieties gave increased circularization efficiency, and reduced levels of intermediate RNA species, in relation to the control. Moreover, the circularization efficiency increased with increasing length of the polyA homopolymeric moieties.

[0155] 5: Effect on circularization after IVT

[0156] This experiment investigated whether heating after IVT produced similar effects on circularization of RNA constructs with and without homopolymeric moieties. circGLuc-A50 / A50 and circGLuc-AO / AO ("A0 / A0") constructs were heated to 55°C for 15 min after 2.5 h of IVT incubation. Control circGLuc-A50 / A50 and circGLuc-AO / AO constructs were not heated but stored on ice after IVT. Analysis by capillary electrophoresis indicated that heating increased circularization efficiency for both constructs to a similar extent. Fig. 11 plots the RNA content for all samples, and for each sample (group) the bars represent in the direction from left to right: circular, precursor, intermediates. As evident from this figure, heating after IVT improved % circular RNA and reduced % precursor and intermediate RNA species both for circGLuc-AO / AO and circGLuc-A50 / A50. Hence, this heat-induced improvement seems neither increased nor decreased by the presence of homopolymeric moieties. eric moieties of different nucleotides

[0157] In this experiment investigated the circularization efficiency of constructs having homopolymeric moieties of the same length but using different nucleotides (C, G A or U) as the recurring unit.

[0158] The test constructs circGLuc-C25 / C25, circGLuc-G25 / G25, circGLuc-A25 / A25 and circGLuc-

[0159] U25 / U25 were produced by IVT together with the control circGLuc-AO / AO and analyzed as described above. Both analysis by capillary electrophoresis and the concentration measurement indicated that for the circGLuc-G25 / G25 and circGLuc-U25 / U25 constructs the yield was below that obtained for circGLuc-A25 / A25. For circGLuc-G25 / G25, the low IVT yield was expected as a result of the 5' polyG sequence (AGG start sequence followed by the polyG) impeding the polymerase performance. However it is envisaged that a polyG sequence could be designed or positioned differently within the precursor to avoid this limitation.

[0160] Fig. 12 shows the % precursor (left bar) and circular (right bar) RNA, respectively. Due to the low IVT yield of circGluc-G25 / G25, no analysis of circularization efficiency could be performed for this construct. Circularization efficiency was greatly improved for circGLuc-A25 / A25 and for circGLuc- U25 / U25. However, as the IVT yields for circGLuc-U25 / U25 was much below the yield for circGLuc- A25 / A25, the total amount of circular RNA for circGLuc-U25 / U25 was not increased in relation to the control circGLuc-AO / AO. The construct circGLuc-C25 / C25, which gave only a slight decrease of IVT yields, exhibited low circularization efficiency, with level even below the circGLuc-AO / AO control. eric moiety can increase circularization

[0161] In this experiment constructs with single homopolymeric moieties of 50A were tested to determine if whether a single homopolymeric moiety is enough to increase circularization of a linear precursor.

[0162] In detail, two new constructs, each containing a 50 nucleotide polyA sequence on either the 5' (circGluc-A50 / A0) or 3' (circGLuc-A0 / 50A) end of the precursor RNA were compared to circGLuc- A50 / A50 and circGLuc-AO / AO, respectively. Capillary electrophoresis and concentration measurement indicated that both constructs with single polyA homopolymeric moieties improved circularization compared to the circGLuc-AO / AO control. Fig. 13 shows the % precursor (left bar, black) and % circular (right bar, white) RNA, respectively, for the respective constructs. As can be seen, a 3' polyA homopolymeric moiety increased % circular RNA from below 40 % (control) to nearly 50 %. The 5' polyA moiety (circGluc-A50 / A0) improved circularization efficiency even more. The greatest improvement was observed for the circGLuc-A50 / A50 construct which had homopolymeric moieties at both ends of the precursor.

[0163] Interestingly, the IVT yields of the circGLuc-A0 / A50 construct were in the same range as the circGLuc-AO / AO control, whereas the IVT yields of the circGLuc-A50 / A0 were lower than for the control and in the range of the circGluc-A50 / A50 construct. This indicates that the drop in IVT yields observed for the circGLuc-A50 / A50 in Example 1 may be due to the presence of a 5' homopolymeric moiety. REFERENCES

[0164] W02014152031A1

[0165] WO2023242425A1 Wesselhoeft et al., 2019, Molecular Cell 74, 508-520

[0166] Wesselhoeft et al., Nature Communications (2018)9:2629

[0167] S Hjerten: Biochim Biophys Acta 79(2), 393-398 (1964)

[0168] US6602990

[0169] US7396467 WO2019137869A1

[0170] W02018011600A1

[0171] WO2022162018A1

[0172] WO2024022894A1

[0173] Trepotec, et al., 2019, RNA 25:507-518 Beverly, et al., Analytical and Bioanalytical Chemistry (2018) 410:1667-1677

Claims

CLAIMS1. A linear precursor ribonucleic acid (RNA) molecule having a 5' end and a 3' end, comprising in the 5'-to-3' direction i) a first self-splicing sequence, ii) a sequence of interest, and iii) a second self-splicing sequence, wherein the linear precursor RNA molecule is capable of forming a circular RNA upon splicing of said first and second self-splicing sequences; wherein the first self-splicing sequence comprises a first homopolymeric moiety located between said 5' end and a splice site first self-splicing sequence, and / or wherein the second self-splicing sequence comprises a second homopolymeric moiety located between said 3' end and a splice site of the second self-splicing sequence.

2. The linear RNA molecule of claim 1, wherein the first and / or the second homopolymeric moiety comprises at least one homopolymeric sequence of at least 6, such as at least 10, consecutive repeating units, and optionally comprises up to 75 units.

3. The linear RNA molecule of claim 1, wherein the first and / or the second homopolymeric moiety comprises at least one homopolymeric sequence which is an interrupted sequence comprising at least two repeating segments separated by a single intervening unit or shorter sequence, wherein each repeating segment consists of at least 3 repeating units.

4. The linear RNA molecule of any one of the preceding claims, wherein the first homopolymeric moiety comprises or is a nucleic acid sequence, and / or the second homopolymeric moiety comprises or is a nucleic acid sequence.

5. The linear RNA molecule of any one of the preceding claims, wherein said homopolymeric moiety comprises at least one sequence selected from polyA, poly U, polyC and polyG, preferably a polyA sequence, wherein said sequence optionally includes one or more modified nucleotides.

6. The linear RNA molecule according to claim 5, wherein the homopolymeric sequence is an interrupted sequence comprising multiple repeated polyA segments, separated by intervening sequences that are not adenine or adenine variants or homologues.

7. The linear RNA molecule according to claim 6, wherein the repeating segments comprise at least 25 units of adenines, adenine variants or adenine homologues, and the intervening sequence comprise half the amount of units compared to the polyA segment.

8. The linear RNA molecule of any one of the preceding claims, wherein said first and / or second homopolymeric moiety comprises at least one sequence selected from (A)k, ( U)j, (C)nand (G)p, preferably Ak, and combinations thereof, wherein each of k, j, n and p independently represents aninteger from 6 to 75, preferably 10-50, wherein said sequence optionally includes one or more modified nucleotides.

9. The linear RNA molecule of any one of the preceding claims, wherein the first homopolymeric moiety is located adjacent to the 5' end of the linear RNA molecule, optionally downstream of a 5' start sequence.10 The linear RNA molecule of any one of the preceding claims, wherein the second homopolymeric moiety is located adjacent to the 3' end of the linear RNA molecule, optionally upstream of a 3' end sequence.

11. The linear RNA molecule of any one of the preceding claims, wherein the first self-splicing sequence comprises said first homopolymeric moiety located between said 5' end and a splice site first self-splicing sequence, and wherein the second self-splicing sequence comprises said second homopolymeric moiety located between said 3' end and a splice site of the second self-splicing sequence, wherein the first and second homopolymeric moieties do not bind to each other.

12. The linear RNA molecule of claim 11, wherein the first homopolymeric moiety and the second homopolymeric moiety are nucleic acid sequences which are non-complementary or uncapable of hybridizing to each other to each other.

13. The linear RNA molecule of claims 11 or 12, wherein the first and second homopolymeric moieties are identical to each other.

14. The linear precursor RNA molecule of claim 1, further comprising a first joining sequence (El) located between the sequence of interest and the splice site first self-splicing sequence, and a second joining sequence located (E2) located between the sequence of interest and the splice site of the second self-splicing sequence, wherein said first and second joining sequences are capable of joining to form a circular RNA molecule containing the sequence of interest.

15. The linear RNA molecule of any one of the preceding claims, wherein the first self-splicing sequence comprises a first pairing sequence, and the second self-splicing sequence comprises a second pairing sequence, wherein said first and second pairing sequences are capable of hybridizing with each other.

16. The linear RNA molecule of any one of the preceding claims, wherein the sequence of interest comprises a coding sequence encoding a protein of interest and optionally a translation regulatory sequence, such as an IRES sequence.

17. The linear RNA molecule of any one of the preceding claims, wherein the sequence of interest has a length of 20-10000 nt, such as 20-1000 nt, such as 50-1000 nt, such as 100-1000 nt, such as 200-10000 nt, such as 200-6000 nt, such as 200-4500 nt, such as 1000-10000 nt, such as 1000-4500 nt.

18. A template DNA molecule encoding the RNA molecule of any one of claims 1-17.

19. A method of producing a circular RNA molecule comprising a sequence of interest, the method comprising incubating a precursor linear RNA molecule according to any one of the claims 1- 17 under conditions allowing splicing of the self-splicing sequences and circularization to form a circular RNA molecule containing the sequence of interest.

20. The method of claim 19, comprising a step of performing an in vitro transcription (IVT) reaction of the template DNA molecule of claim 18 to provide said precursor linear RNA molecule.

21. The method of claim 20, wherein performing an in vitro transcription (IVT) reaction comprises providing an IVT reaction mixture comprising template DNA, ribonucleoside triphosphates, DNA dependent RNA polymerase and a buffer system, and said step of incubating the linear precursor RNA molecule comprises incubating said IVT reaction mixture.

22. The method of any one of claims 19-21, wherein the linear precursor RNA molecule is incubated at a pH of 6.5-8.

23. The method of any one of claims 19-22, wherein the linear precursor RNA molecule is incubated at a temperature in the range of from 20°C to 56°C, such as from about 37°C to about 56 °C.

24. The method of any one of claims 19-23, wherein the linear precursor RNA molecule is incubated for an incubation time of from 10 minutes to 24 hours, such as from 10 minutes to 5 hours, such as from 30 minutes to 5 hours.

25. The method of any one of claims 19-24, further comprising separating the circular RNA from the linear precursor RNA molecule and / or other RNA species formed during incubation.

26. A composition comprising: a linear precursor RNA molecule according to any one of the claims 1-17, a circular RNA molecule comprising the sequence of interest of the precursor RNA molecule, and optionally, at least one impurity selected from: i) an RNA molecule comprising the first self-splicing sequence but lacking the second self-splicing sequence, ii) an RNA molecule comprising the second self-splicing sequence but lacking the first self-splicing sequence, iii) a cleaved first self-splicing sequence, and iv) a cleaved second self-splicing sequence.

27. The composition of claim 26, wherein the composition is an IVT reaction mixture and further comprisestemplate DNA according to claim 18, optionally digested by deoxyribonuclease (DNase); optionally, DNase; ribonucleoside triphosphates; a DNA dependent RNA polymerase, and any cofactors required for nucleic acid polymerization catalyzed by said DNA dependent polymerase; and a buffer optionally, at least one component selected from a pyrophosphatase, an RNase inhibitor and a detergent28. A method of separating a desired circular RNA from a non-circular RNA species comprising at least one homopolymeric moiety, the method comprising a) providing a composition according to claim 26 or 27 or a composition obtainable by the method of any one of claims 19-25, b) contacting the composition with a chromatography material comprising a ligand coupled to a support material, said ligand having a binding affinity for the homopolymeric moiety, under conditions allowing the ligand to bind to the homopolymeric moiety, c) obtaining the circular RNA in a flowthrough fraction, d) optionally, eluting bound non-circular RNA species from the chromatography material and optionally regenerating the chromatography material.

29. The method of claim 28, wherein the ligand is a nucleic acid molecule that is capable of binding to the homopolymeric moiety.

30. The method of claim 29, wherein the ligand is an oligo(dT) ligand and the homopolymeric moiety comprises a polyA sequence.

31. The method of any one of claims 28-30, wherein the support material is a convectionbased matrix, optionally selected from the group consisting of a porous polymer membrane, a filter, a fibrous matrix and a porous monolith.

32. The method of any one of claims 28-30, wherein the support material comprises beads selected from porous beads, non-porous beads, and magnetic beads.

33. The method of claim 32, wherein the beads are porous beads formed of a polysaccharide or derivative thereof, such as agarose or a derivative thereof.

34. The method of any one of claims 28-33, wherein the method comprises, prior to step b, or after step b, a step of subjecting the composition to a size-based separation step, such as filtration or size exclusion chromatography.

Images