mRNArchitect: sequence design of mRNA medicines

mRNArchitect: sequence design of mRNA medicines | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results mRNArchitect : sequence design of mRNA medicines Magdalena A. Budzinska , Mohammad Zardbani , P. Sidharta Michelle , Jonathon Waterhouse , Theodore E. Leonard , Aayushi Ghodasara , Sneha Lakshman Das , Rachel Y. Chang , Natasha Chaudhary , View ORCID Profile Seth W. Cheetham , Timothy R. Mercer , View ORCID Profile Helen M. Gunter doi: https://doi.org/10.1101/2024.12.03.626696 Magdalena A. Budzinska 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Mohammad Zardbani 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site P. Sidharta Michelle 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Jonathon Waterhouse 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Theodore E. Leonard 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Aayushi Ghodasara 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Sneha Lakshman Das 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Rachel Y. Chang 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Natasha Chaudhary 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Seth W. Cheetham 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Seth W. Cheetham Timothy R. Mercer 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Helen M. Gunter 1 Australian Institute for Bioengineering and Nanotechnology, University of Queensland , Brisbane, QLD, Australia 2 BASE Facility, University of Queensland , Brisbane, QLD, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Helen M. Gunter For correspondence: h.gunter{at}uq.edu.au Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract Sequence design is the critical first step in developing a new mRNA medicine. The mRNA primary sequence is assembled from multiple elements, including the coding sequence of the target protein or gene, flanking untranslated regions that impact expression, and the poly(A) tail, which improves stability. The mRNA primary sequence can be optimized to improve the performance of an mRNA medicine, including increasing translation and stability, and reducing reactogenicity. Here, we introduce mRNArchitect , a software toolkit to assist in the design of mRNA sequences. mRNArchitect can rapidly assemble and optimize mRNA sequences based on user requirements, including GC content, secondary structure, codon optimization, and uridine depletion. The sequences generated by mRNArchitect can also be readily synthesized and have been experimentally validated across a variety of applications. We present mRNArchitect as a software toolkit to help scientists design new mRNA medicines. Availability and implementation mRNArchitect is available via a web portal at www.basefacility.org.au/software/ and Github at https://github.com/BaseUQ/mRNArchitect . mRNArchitect is open-source and freely available under MIT License. 1. INTRODUCTION Careful sequence design can markedly improve the performance of an mRNA by enhancing its translation, increasing stability and reducing reactogenicity 1 , 2 . Therefore, the primary sequence design is a key first step when developing a new mRNA vaccine or therapy. mRNA medicines can encode any coding sequence (CDS), from small peptides to large structural proteins. The CDS of interest is reverse translated into an mRNA sequence using codons optimised for high expression in the target organism or tissue 3 . The CDS can also be depleted of uridines that are recognised by receptors that initiate the innate immune response 4 . An mRNA medicine also encodes untranslated regions (UTR) that flank the CDS and regulate translation and stability. The ribosome binds the 5’ UTR and scans for the methionine start codon, while the 3’ UTR contains binding sites for proteins and non-coding RNAs that impact cell-specific expression. The mRNA ends with a poly(A) tail that protects against 3’ exonuclease degradation and improves stability. Sequence design also considers the formation of RNA secondary structures that stabilise the mRNA and interact with proteins. RNA structures, such as hairpins, can slow ribosome binding and translation, but protect against nuclease digestion. Double-stranded RNA (dsRNA) structures can also be detected by RIG-I and toll-like receptors that activate the innate immune response 5 . In addition to improving performance, the mRNA sequence design should also be compliant with manufacturing requirements. Ideally, the output mRNA sequences should be easily synthesized into DNA templates, avoid error-prone, repetitive sequences, and produce high yields of mRNA during in vitro transcription. Different mRNA design strategies may be preferred depending on the final clinical application. For example, mRNA vaccine sequences with extensive secondary structure may be preferred as they improve mRNA stability and confer an adjuvant effect 6 . Alternatively, therapeutic applications that require frequent dosing may avoid uridine and secondary structures to minimize potential inflammatory side effects 7 . To assist in the design of mRNA sequences for different applications, we have developed mRNArchitect , a software toolkit that can assemble and optimize mRNA sequences according to user specifications. Output sequences can be readily synthesised and have been experimentally validated in a wide range of cell-types and applications. We provide mRNArchitect , open-source software that assists in the design of new mRNA medicines. 2. FEATURES 2.1. Sequence input mRNArchitect enables the design of mRNAs based on user-supplied sequences. First, an open-reading frame encoding the user’s protein of interest is supplied ( see Figure 1A ). The coding sequence can be provided in either amino acid, RNA, or DNA format, and should initiate with a methionine codon (M, AUG) and terminate with one or more stop codons (e.g., *, UAA, UAG, UGA). Download figure Open in new tab Figure 1. Design of mRNA medicines using mRNArchitect . (A) Sequence inputs . mRNArchitect accepts 5’ and 3’ UTR sequences, a coding sequence (CDS) encoding the protein of interest (supplied as DNA, RNA or amino acid) and a poly(A) tail. (B) Optimisation parameters . CDS sequences are optimized to maximize CAI values within a given GC% range. Different sequence motifs can be avoided, including Restriction Enzyme binding sites or repeats which can both impede manufacture. Additionally, secondary structure is optimized according to Minimum Free Energy (MFE) calculations. (C) Outputs . mRNA candidate sequences are provided with an accompanying output table summarising the properties of the mRNA designs, which can be plotted to assist in sequence selection. (D) Sequence alignment . Multiple Sequence Alignment and sequence logo plot, showing the similarity between Luciferase mRNA designs generated by mRNArchitect , mRNAid and LinearDesign. (E) Codon optimality and mRNA secondary structure . Scatter plot comparing CAI and MFE of Luciferase mRNA designs generated by mRNArchitect , mRNAid and LinearDesign. (F) mRNA expression . Relative expression of different Luciferase mRNA designs was assayed in HEK293-T cells at 24h and 48h post transfection. Results were analysed using One-way ANOVA (****=p<0.0001). (G) Template manufacture . mRNArchitect designs include mRNAs with a range of CAI and MFE values, which impact manufacturability and expression. Users can also supply custom untranslated regions. By default, we suggest 5’ and 3’ UTRs based on human α-globin ( HBA1 ; Gene ID 3039) that have been validated in a wide range of cell types and applications. Additional 5’ UTR sequences are supplied, that are derived from further housekeeping genes 8 and minimal 5’ UTR sequences 9 . Finally, users can include a poly(A) tail in mRNA designs, either specifying its length, or supplying more complex designs, such as segmented poly(A) tails 10 . 2.2. Optimisation parameters mRNArchitect can optimize mRNA sequences according to a range of parameters. mRNArchitect is built upon the DNA Chisel framework 11 , which determines the optimal sequence based on a set of criteria, including GC content, codon translation speed, and mRNA secondary structures (e.g. hairpins; Figure 1B ). Default design parameters are derived from expression vector optimisation ( https://basebuddy.lbl.gov/ ), however custom parameters can be specified, including both hard- and soft constraints. Hard constraints, such as avoidance of microRNA seed sites, which can destabilise mRNAs when bound by complementary microRNAs, are strictly enforced. For soft constraints, such as codon optimisation, the constraint score is maximized. These parameters can be modified according to the user’s requirements and often depend on the target organism, tissue and intended application. There are several key parameters that impact mRNA expression and reactogenicity. To enhance translation, the CDS is codon-optimized to use preferred codons of the specified target organism (e.g. human or mouse). mRNArchitect also allows uridine depletion (UD), which excludes uridine at wobble positions in codons. Difficult sequence motifs, such as motifs associated with ribosomal frameshifting 12 , can also be avoided. mRNArchitect also allows users to impose a minimum and maximum GC fraction across a defined window, which is a key variable in the formation of stable RNA secondary structures. In addition to these performance considerations, it is necessary for mRNA sequences to be manufactured to a consistently high quality. mRNArchitect allows users to avoid restriction enzyme binding sites used during DNA template assembly and avoid low-complexity, repetitive sequences that cause errors during DNA synthesis and in vitro transcription. mRNArchitect optimisation parameters can be applied to either the full nucleotide sequence (“simple” mode), or across specified sub-regions of the sequence (“advanced” mode). This enables users to impose distinct optimisation parameters across different parts of the sequence – for example locally reducing secondary structures in the proximal part of the coding region 13 , or preserving exact nucleotide sequence where required, such as small interfering RNA (siRNA) binding sites. 2.3. Output mRNA sequences The mRNArchitect output provides fully assembled and optimized mRNA sequence candidates that are ready for synthesis ( Figure 1C ). Sequence outputs are accompanied by statistics describing their nucleotide- and structural composition. Reported statistics include the Codon Adaptation Index (CAI) score, which measures how closely the codon usage of the sequence matches the codon preference of a particular organism 3 , 14 . The MFE, which is calculated using the ViennaRNA package, represents the lowest Gibbs free energy change associated with the formation of secondary structures in RNA molecules due to intramolecular base pairing 15 . 3. RESULTS 3.1. Running mRNArchitect To demonstrate the use of mRNArchitect , we designed mRNAs encoding the Luciferase protein 16 . Sequence inputs were based on a modified version of the P. pyralis Luciferase amino acid sequence (GenBank: AB644228.1 ; Supplementary Table 1 ). We used 5’ and 3’ UTRs based on human- and mouse α-globin sequences respectively ( HBA1 ; Gene IDs 3039; 15122) and a 120 nt poly(A) tail. First, mRNArchitect was used to design Luciferase mRNAs using the human codon model 3 and UD, with all other parameters set to default values. This strategy yielded an mRNA design with a CAI of 0.859 a Minimum Free Energy (MFE) of −534.2, which we used as a point of comparison for further optimisation strategies ( Supplementary Figure 1A-B ). Different nucleotide inputs (encoding identical amino acid sequences) had a limited impact on the CAI and MFE of output sequences. Altering the mRNArchitect optimisation parameters impacted our designs more significantly. Luciferase mRNA designs optimised without UD showed a higher CAI and a lower MFE than designs with UD, despite only a modest increase in overall Uridine content (~4%) ( Supplementary Table 1 ). Moreover, increasing the maximum GC content to 70% or more, progressively increased the CAI and decreased the MFE of our designs. However, these designs were more difficult to manufacture with DNA synthesis ( Supplementary Figure 1C ). To investigate the limits of the mRNA sequences generated by mRNArchitect , we prepared designs that maximised and minimised CAI ( Figure 1D , Supplementary Table 1 ). The mRNArchitect maximum CAI design (termed “ mRNArchitect ” here on) was generated by relaxing the optimisation constraints of mRNArchitect (i.e. no UD, Min GC = 0, Max GC = 1.0; Avoid repeat length = 1000) so that only CAI value is considered for sequence optimisation. As a negative control, we designed a “minimum CAI” codon model, that assigns a high score to rare codons, thereby enriching them in the resulting design ( Supplementary Table 2 ). This codon model was used to prepare a Minimum CAI Luciferase mRNA sequence (termed “ MinCAI ” here on). We benchmarked the mRNArchitect and MinCAI designs against different publicly-available mRNA design tools, including mRNAid and LinearDesign 1 , 2 ( Supplementary Figure 1D-E ). mRNAid was used to design a Luciferase mRNA, with parameters that include CAI optimization with UD, MFE estimation using ViennaRNA, avoiding BsaI, Xba and XhoI cleavage sites, and a target GC content of 0.4 – 0.6 across a 60nt GC content window with all other parameters kept as default. LinearDesign, was also used to design a Luciferase mRNA, with the default parameters, and with lambda set to zero 2 . The mRNArchitect Luciferase design showed the highest similarity to mRNAid (94%), then LinearDesign (81%), then the Wild-type Luciferase (73%) and MinCAI designs (66%) ( Figure 1D , Supplementary Figure 1G ). Except for the mRNAid design, all sequences required specialized DNA synthesis due to difficult and repetitive sequences ( Supplementary Figure 1F ). We next compared the CAI and MFE of the Luciferase mRNA designs ( Figure 1E ; Supplementary Figure 1D-E ). The mRNArchitect and MinCAI Luciferase designs showed the highest (1.0) and lowest (0.62) CAI values, while mRNAid and LinearDesign had intermediate (0.965) and lower (0.734) values respectively. The MFE values were also variable, where the LinearDesign sequence had the lowest MFE (−1111.5), MinCAI had the highest (−351.9) and mRNArchitect and mRNAid designs showed intermediate MFE values (−685.3 and −593.0 respectively). 3.2. Luciferase mRNA We hypothesized that the different Luciferase mRNA designs would exhibit distinct expression levels and stability. To test this hypothesis, we synthesized the mRNA constructs at the BASE mRNA facility using standard methods 17 and transfected them into HEK293-T cells using Lipofectamine™ MessengerMAX™ (Invitrogen™). Luciferase activity was measured at 24 and 48 hours using the Luciferase Assay System (Promega) ( Figure 1F-G ). At 24 hours post-transfection, Luciferase expression peaked and varied significantly across the mRNA designs. The highest expression was observed for the mRNArchitect mRNA (p<0.0001 compared to all other treatments, based on 1-way ANOVA), while the lowest was recorded for the MinCAI design. The LinearDesign and mRNAid sequences exhibited intermediate expression levels. Specifically, mRNArchitect expression was 21-fold higher than MinCAI , sixfold higher than LinearDesign, and fivefold higher than mRNAid. Similar trends were observed at 48 hours, however expression of the mRNArchitect mRNA was now four-fold higher than the LinearDesign mRNA and 34-fold higher than the MinCAI mRNA. 3.3. eGFP mRNA Next, we prepared eGFP mRNA designs using mRNArchitect to validate its performance on a different amino acid sequence. First, an mRNA was designed using the default mRNArchitect parameters and UD, yielding an mRNA design with a CAI of 0.9 and an MFE of −217.4 ( Supplementary Table 3 ). Like the Luciferase optimisations (above), relaxing the eGFP design constraints (e.g. increasing GC content and avoiding UD), produced mRNA designs with progressively higher CAI and lower MFE values ( Supplementary Figure 2A-B ). The highest CAI was observed for an mRNArchitect Maximum CAI eGFP design (termed “ mRNArchitect ” here on), that was prepared using the same method described for the Luciferase Maximum CAI design (above). Further eGFP mRNA designs were also prepared using mRNAid and LinearDesign, using the optimisation parameters described for Luciferase (above). All designs can be easily manufactured using DNA synthesis, except for the LinearDesign mRNA ( Supplementary Figure 2C ). We compared the expression of eGFP for mRNA designs generated by mRNArchitect , mRNAid and LinearDesign ( Supplementary Figure 2D-F ). The mRNAs were synthesised at the BASE mRNA facility, and transfection followed the previously described methods (above). eGFP expression was assessed using flow cytometry (CytoFlex). At 24 hours post-transfection, there was no significant difference in the expression of the different mRNA designs, based on a 1-way ANOVA. However, by 48 hours, significant differences emerged: the mRNArchitect mRNA exhibited the highest expression, the mRNAid mRNA the lowest (p<0.0001), and the LinearDesign mRNA, an intermediate level (p=0.0028). The mRNArchitect mRNA expression was fivefold higher than mRNAid and twice as high as LinearDesign. Together, our results show that by modulating the mRNArchitect optimisation parameters, users can generate candidate mRNA sequences with differing sequence and expression properties. We observe a trade-off between the manufacturability and performance of mRNA designs, where mRNAs with higher CAI and lower MFE tend to be more highly expressed and more difficult to manufacture ( Figure 1G , Supplementary Table 1, 3 ). Users can specify different sequence optimisation parameters to suit the required performance and manufacturability of mRNA sequences according to their intended application. 3.5. 5’ Untranslated regions (UTRs) In addition to coding sequence (CDS) optimisation, mRNA expression is strongly influenced by the 5′ untranslated region (UTR) sequence. Therapeutic mRNAs often incorporate short 5′ UTRs derived from housekeeping genes such as α-globin , which are associated with stable, constitutive expression. We tested the impact of several different 5′ UTR sequences on expression of the mRNArchitect FLuc sequence. These included 5’ UTRs from housekeeping genes β-globin (ENSG00000244734), β-actin (ENSG00000075624), and albumin (ENSG00000163631) 8 , as well as a minimal 5’ UTR 9 ( Supplementary Figure 3A ). We also assessed 5′ UTRs with predicted low secondary structure that have been reported to drive strong reporter expression, including NELL2 (ENSG00000184613), ANKRD49 (ENSG00000168876), and ZFAND6 (ENSG00000086666) 18 . The mRNAs were synthesised and transfected into HEK293-T cells as described in Section 3.2 . All constructs with housekeeping 5’ UTRs had a luminescence equal to or greater than the control which included the α-globin 5’ UTR. Both β-actin and albumin UTRs showed significant increases in expression, based on a 1-way ANOVA (p=0.0003, p=0001 respectively; Supplementary Figure 3B ). Of the 5’ UTRs with predicted low secondary structure, NELL2 resulted in significant in expression (p=0.0036), while ANKRD49 and ZFAND6 resulted in significantly lower expression (p<0.0001 for both). These results highlight that the rational selection of 5′ UTRs can substantially enhance protein output, offering an additional layer of design flexibility for tailoring mRNA medicines to different therapeutic applications. 4. CONCLUSION The optimisation of mRNA sequences can markedly improve their translation and stability and reduce their reactogenicity. mRNArchitect can be used to design optimized mRNAs that encode a wide range of proteins, including vaccine antigens, peptides, antibodies, gene- and cell therapies. We have used mRNArchitect to design hundreds of mRNAs that have been experimentally validated across a wide range of applications. We provide mRNArchitect as a freely-available software toolkit for the assembly and optimization of mRNA sequences. Our results show that mRNAs composed entirely of preferred human codons, optimised for maximum CAI, show high and sustained expression in cell culture. This strategy is particularly effective for the design of mRNAs encoding short- to medium-length proteins. mRNAs composed of preferred codons are translated efficiently due to the high availability of complementary tRNAs, and tend to form stable secondary structures due to their high GC content ( Supplementary Figure 1B , E; 2B ) 19 . However, while a Maximum CAI optimisation supports robust translation of short- to medium-length proteins, it poses challenges for the manufacturability of mRNAs encoding longer proteins. It is therefore unlikely that a single mRNA sequence optimisation strategy will be universally appropriate. Instead, mRNA designs should consider the length and amino-acid composition of encoded proteins, and the intended application of the mRNA medicine. The field of mRNA design is advancing, with the identification of new features and sequences that regulate translation rates, subcellular localization, and stability. We anticipate that ongoing improvements in mRNA design, combined with an increased understanding of mRNA biology within the cell, will further enhance these designs. In concert, we will continue to improve mRNArchitect by incorporating new elements and features, supporting scientists in designing innovative mRNA medicines. Supplementary data Supplementary data are available online. Conflict of interest None. Funding We acknowledge the following sources of funding and support: National Health and Medical Research Council (GNT2014002 and GNT1161832) to T.R.M., Australian Research Council (DE230100036) to S.W.C., Medical Research Future Fund (MRFCRI000063 and MRFCRI000089) to S.W.C. and T.R.M., National Collaborative Research Infrastructure Strategy (NCRIS) to T.R.M., Therapeutic Innovation Australia (TIA) to S.W.C and T.R.M., Tour de Cure to S.W.C. and T.R.M., and The University of Queensland to S.W.C. and T.R.M. Data availability mRNArchitect is available through www.basefacility.org.au/software/ and can be downloaded for local installation from Github at https://github.com/BaseUQ/mRNArchitect . Supplemental Figure legends Supplementary Figure 1. Firefly Luciferase mRNA sequences generated using different mRNA design tools and parameters. (A-C) Different performance and manufacturability characteristics of Luciferase mRNA designs generated by mRNArchitect , with different CDS input sequences and optimisation parameters. These include mRNAs designed to meet stringent DNA template manufacturing requirements (Different inputs), without Uridine Depletion (No U depletion), with a maximum GC content of 70% (Max 70% GC), or 80% (Max 80% GC), or mRNAs designed without any further design constraints, to maximise CAI ( mRNArchitect ), or with a rare codon model ( MinCAI ). ( D-F ) Different performance and manufacturability characteristics of Luciferase mRNA designs generated by different software including mRNArchitect , mRNAid and LinearDesign, and the Luciferase Wild Type sequence ( P. pyralis ). (G) Percent identity matrix from Multiple Sequence Alignment of Luciferase mRNA sequences. Supplementary Figure 2. eGFP mRNA sequences, generated using different mRNA design tools and parameters. (A-C) Different performance and manufacturability characteristics of eGFP mRNA designs generated by mRNArchitect , using different optimisation parameters. These include mRNAs designed to meet stringent DNA template manufacturing requirements (low complexity), without Uridine Depletion (No U Depletion), with a maximum GC content of 70% (Max 70% GC), or 80% (Max 80% GC), or mRNAs designed to without any further design constraints, to maximise CAI ( mRNArchitect ). Further eGFP mRNA designs were prepared using mRNAid and LinearDesign. (D) Expression of different eGFP mRNA designs in HEK293 cells, determined by Flow cytometry, at 24h and 48h post transfection. Results were analysed using One-way ANOVA. (E) Fluorescence imaging of HEK293 cells transfected with mRNArchitect eGFP mRNA. (F) eGFP mRNA designs showed similar transfection rates in HEK293-T cells, determined by Flow cytometry. Results were analysed using One-way ANOVA (ns=p>0.05; *=p<0.05; **=p<0.01; ****=p<0.0001). Supplementary Figure 3. Impact of 5’ UTR sequence on Firefly Luciferase expression. (A) Four Firefly Luciferase mRNAs were designed, incorporating different 5’ UTR sequences derived from housekeeping genes, proceeded by an AGG co-transcriptional capping signal and followed by a Kozak sequence, an optimised Firefly Luciferase CDS, α-globin 3’ UTR and a 120-nucleotide poly(A) tail. Three further FLuc mRNAs were designed, with different A/T rich 5’ UTRs, that included an AGG co-transcriptional capping signal, but no Kozak sequence. (B) Firefly Luciferase expression was measured in HEK293-T cells at 24 hours post transfection and normalised against Nanoluciferase expression. (ns=p>0.05; **p<0.01; ***p<0.001; ****p<0.0001). Supplementary Tables Supplementary Table 1. Luciferase mRNA designs . Supplementary Table 2. Codon model used for MinCAI Luciferase mRNA design . Supplementary Table 3. eGFP mRNA designs . Acknowledgements The authors acknowledge the facilities and the scientific and technical assistance of the BASE mRNA Facility at The University of Queensland. BASE is supported by Therapeutic Innovation Australia (TIA). TIA is supported by the Australian Government through the National Collaborative Research Infrastructure Strategy (NCRIS) program. Footnotes This version of the manuscript has been revised to add new features of our software, and a new section that describes the impact of 5' UTR sequence on expression. References ↵ Vostrosablin , N. et al. mRNAid, an open-source platform for therapeutic mRNA design and optimization strategies . NAR Genomics and Bioinformatics 6 , qae028 ( 2024 ). ↵ Zhang , H. et al. Algorithm for optimized mRNA design improves stability and immunogenicity . Nature 621 , 396 – 403 ( 2023 ). OpenUrl CrossRef PubMed ↵ Holcomb , D. D. , Alexaki , A. , Katneni , U. & Kimchi-Sarfaty , C. The Kazusa codon usage database, CoCoPUTs, and the value of up-to-date codon usage statistics . Infection, genetics and evolution: journal of molecular epidemiology and evolutionary genetics in infectious diseases 73 , 266 – 268 ( 2019 ). OpenUrl CrossRef ↵ Karikó , K. , Buckstein , M. , Ni , H. & Weissman , D. Suppression of RNA recognition by Toll-like receptors: the impact of nucleoside modification and the evolutionary origin of RNA . Immunity 23 , 165 – 175 ( 2005 ). OpenUrl CrossRef PubMed Web of Science ↵ Pichlmair , A. et al. Activation of MDA5 requires higher-order RNA structures generated during virus infection . Journal of virology 83 , 10761 – 10769 ( 2009 ). OpenUrl Abstract / FREE Full Text ↵ Lee , J. , Woodruff , M. C. , Kim , E. H. & Nam , J.-H. Knife’s edge: Balancing immunogenicity and reactogenicity in mRNA vaccines . Experimental & Molecular Medicine 55 , 1305 – 1313 ( 2023 ). OpenUrl CrossRef PubMed ↵ Kiaie , S. H. et al. Recent advances in mRNA-LNP therapeutics: immunological and pharmacological aspects . Journal of nanobiotechnology 20 , 276 ( 2022 ). OpenUrl PubMed ↵ Ma , Q. et al. Optimization of the 5′ untranslated region of mRNA vaccines . Scientific Reports 14 , 19845 ( 2024 ). OpenUrl PubMed ↵ Trepotec , Z. et al. Maximizing the translational yield of mRNA therapeutics by minimizing 5′-UTRs . Tissue Engineering Part A 25 , 69 – 79 ( 2019 ). OpenUrl PubMed ↵ Trepotec , Z. , Geiger , J. , Plank , C. , Aneja , M. K. & Rudolph , C. Segmented poly (A) tails significantly reduce recombination of plasmid DNA without affecting mRNA translation efficiency or half-life . Rna 25 , 507 – 518 ( 2019 ). OpenUrl Abstract / FREE Full Text ↵ Zulkower , V. & Rosser , S. DNA Chisel, a versatile sequence optimizer . Bioinformatics 36 , 4508 – 4509 ( 2020 ). OpenUrl CrossRef PubMed ↵ Mulroney , T. E. et al. N 1-methylpseudouridylation of mRNA causes+ 1 ribosomal frameshifting . Nature 625 , 189 – 194 ( 2024 ). OpenUrl CrossRef PubMed ↵ Nieuwkoop , T. et al. Revealing determinants of translation efficiency via whole-gene codon randomization and machine learning . Nucleic acids research 51 , 2363 – 2376 ( 2023 ). OpenUrl CrossRef PubMed ↵ Kames , J. et al. TissueCoCoPUTs: novel human tissue-specific codon and codon-pair usage tables based on differential tissue gene expression . Journal of molecular biology 432 , 3369 – 3378 ( 2020 ). OpenUrl CrossRef PubMed ↵ Lorenz , R. et al. ViennaRNA Package 2.0 . Algorithms for molecular biology 6 , 26 ( 2011 ). OpenUrl CrossRef PubMed ↵ De Wet , J. R. , Wood , K. V. , DeLuca , M. , Helinski , D. R. & Subramani , S. Firefly luciferase gene: structure and expression in mammalian cells . Molecular and cellular biology 7 , 725 – 737 ( 1987 ). OpenUrl Abstract / FREE Full Text ↵ Leighton , L. J. et al. The design, manufacture and LNP formulation of mRNA for research use . Nature Protocols , doi: 10.1038/s41596-025-01174-4 ( 2025 ). OpenUrl CrossRef ↵ Lewis , C. J. et al. Quantitative profiling of human translation initiation reveals regulatory elements that potently affect endogenous and therapeutically modified mRNAs . bioRxiv , 2024.2002. 2028.582532 ( 2024 ). ↵ Trotta , E. On the normalization of the minimum free energy of RNAs by sequence length . PloS one 9 , e113380 ( 2014 ). OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted August 26, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following mRNArchitect: sequence design of mRNA medicines Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share mRNArchitect : sequence design of mRNA medicines Magdalena A. Budzinska , Mohammad Zardbani , P. Sidharta Michelle , Jonathon Waterhouse , Theodore E. Leonard , Aayushi Ghodasara , Sneha Lakshman Das , Rachel Y. Chang , Natasha Chaudhary , Seth W. Cheetham , Timothy R. Mercer , Helen M. Gunter bioRxiv 2024.12.03.626696; doi: https://doi.org/10.1101/2024.12.03.626696 Share This Article: Copy Citation Tools mRNArchitect : sequence design of mRNA medicines Magdalena A. Budzinska , Mohammad Zardbani , P. Sidharta Michelle , Jonathon Waterhouse , Theodore E. Leonard , Aayushi Ghodasara , Sneha Lakshman Das , Rachel Y. Chang , Natasha Chaudhary , Seth W. Cheetham , Timothy R. Mercer , Helen M. Gunter bioRxiv 2024.12.03.626696; doi: https://doi.org/10.1101/2024.12.03.626696 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Bioinformatics Subject Areas All Articles Animal Behavior and Cognition (7646) Biochemistry (17728) Bioengineering (13917) Bioinformatics (42038) Biophysics (21489) Cancer Biology (18637) Cell Biology (25554) Clinical Trials (138) Developmental Biology (13403) Ecology (19941) Epidemiology (2067) Evolutionary Biology (24368) Genetics (15624) Genomics (22547) Immunology (17764) Microbiology (40475) Molecular Biology (17208) Neuroscience (88756) Paleontology (667) Pathology (2842) Pharmacology and Toxicology (4834) Physiology (7659) Plant Biology (15175) Scientific Communication and Education (2047) Synthetic Biology (4304) Systems Biology (9835) Zoology (2272)