Introduction

Messenger RNA (mRNA) vaccines represent a novel vaccination technology that employs mRNA as an antigen vector capable of eliciting robust immune responses against various pathogens or tumor cells. These vaccines introduce exogenous mRNA encoding antigens into host cells through expression systems, where they are translated into antigens to induce immune responses1. There have been significant advancements in mRNA vaccine development in recent years, particularly during the COVID-19 pandemic in 2020, which highlighted their potential to contribute to rapid and effective responses to emerging infectious diseases2. Moderna’s mRNA-1273 and BioNTech’s BNT162b2, the first COVID-19 vaccines to receive approval in numerous countries were mRNA vaccines. Both of the vaccines demonstrated high immunogenicity3,4.

However, mRNA vaccines face several challenges, including the instability of mRNA molecules, inefficient intracellular delivery and translation, and potential immunogenicity and toxicity5. To address these issues, it is necessary to optimize mRNA molecules through chemical modifications, delivery systems, sequence design, and other strategies6. The structure of mRNA vaccine nucleic acid molecules resembles that of mature mRNA in eukaryotic cells, comprising a 5ʹ cap, 5ʹ untranslated region (5ʹUTR), open reading frame, 3ʹ untranslated region (3ʹUTR) and poly(A) tail, arranged from 5ʹ to 3ʹ. These components can be modified to enhance the in vivo half-life of the mRNA and reduce undesirable immune reactions7. The 5ʹUTR and 3ʹUTR of mRNA play vital roles in regulating mRNA export from the nucleus, translation efficiency, subcellular localization, and mRNA stability8. UTRs significantly affect protein translation and mRNA stability through their interactions with RNA-binding proteins (RBPs)9,10; the 5ʹUTR, specifically, is a crucial factor determining mRNA translation efficiency and stability. Located between the cap structure and start codon, the 5′UTR binds to various transcription factors and ribosomes, thereby regulating initiation, scanning, and selection in mRNA translation11. The length, sequence, and secondary structure of the 5ʹUTR all play critical roles in mRNA translation12. Generally, lengthy or highly structured 5′UTRs can impede ribosome binding to the cap structure or the scanning process, thus reducing mRNA translation efficiency13. Therefore, when designing a 5ʹUTR for optimal translation, it is important to avoid GC-rich secondary structures in the UTR and to exclude the AUG codon to prevent premature translation initiation14. This study aimed to investigate the impact of different 5′UTRs on the translation efficiency of mRNA vaccines, providing an experimental foundation for the development of mRNA vaccines targeting various pathogens.

Results

Plasmid design and in vitro transcription of mRNA

As shown in Fig. 1a, five different kinds of 5ʹUTRs were synthesized and introduced upstream of the dual-reporter genes by PCR. The PCR products and vector plasmid were digested with restriction endonucleases and ligated to form five different recombinant plasmids as shown in Fig. 1b. The five recombinant plasmids were used as templates to be transcribed and capped to produce mRNA in vitro, as shown in Fig. 1c. The Gp145 expression plasmids and mRNA with different 5ʹUTRs were constructed in a similar way as the reporter genes (Fig. 1d–f).

Fig. 1
figure 1

Schematic representation of dual-reporter gene and Gp145 expression plasmids and mRNA containing different 5′UTR. (a) Five different synthetic UTRs were ligated to the reporter gene through PCR. (b) PCR products of (a) and vector plasmid pKMV-α-globin 3′UTR-PolyA were digested with XbaI/XhoI and ligated to generate different recombinant plasmids. (c) Five different recombinant plasmids were transcribed and capped in vitro to form reporter genes expression mRNA with different 5′UTR. (d) Three different synthetic UTRs were ligated to the Gp145 gene through PCR. (e) PCR products of (d) and vector plasmid pKMV-α-globin 3′UTR-PolyA were digested with XbaI/XhoI and ligated to generate different recombinant plasmids. (f) Three different recombinant plasmids were transcribed and capped in vitro to form Gp145 expression mRNA with different 5′UTR.

Full size image

Prediction of the secondary structure of 5ʹUTR mRNAs

Using the bioinformatics platform Paddlehelix, the secondary structures of the 5ʹUTR regions and first 30 amino acids of the coding sequences in the different mRNA constructs were predicted (Fig. 2)15. The thermodynamic stability of the first hairpin structure and position of the cap site were also predicted (Table 1).

Fig. 2
figure 2

Predicted secondary structure of different 5′UTR mRNA sequence. (a) α-globin 5′UTR mRNA. (b) Minimal 5′UTR mRNA. (c) β-globin 5′UTR mRNA. (d) CYBA 5′UTR mRNA. (e) albumin 5′UTR mRNA.

Full size image
Table 1 UTRs selected for initial assay validation.
Full size table

Reporter gene expression in cells transfected with recombinant plasmids containing different 5ʹUTRs

Recombinant plasmids with the various 5ʹUTRs were transfected into BHK-T7 cells. The expression of green fluorescent protein was assessed at 24, 48, and 72 h post-transfection using a CTL-ImmunoSpot S6 FluoroSpot, and fluorescence-positive cells were quantified. Furthermore, the expression of luciferase was measured using a bioluminescence detector for 96-well plates as shown in Fig. 3. Cells transfected with plasmids 1–5 exhibited conspicuous green fluorescence, in contrast with those transfected with the control plasmid pKMV-α-globin 3ʹUTR-PolyA, which showed no green fluorescence. At 24 h post-transfection, cells transfected with plasmid 4 (containing the cytochrome B-245α chain [CYBA] 5ʹUTR) displayed the highest number of green fluorescence-positive cells and most intense luciferase fluorescence. In comparison, cells transfected with plasmid 5 (containing the albumin 5ʹUTR) demonstrated lower numbers of green fluorescent cells and intensity of luciferase expression. The expression levels in cells transfected with plasmid 1 (containing the α-globin 5ʹUTR), plasmid 2 (containing the minimal 5ʹUTR), and plasmid 3 (containing the β-globin 5ʹUTR) were comparatively lower for both reporter genes, with no notable differences observed among these three plasmids. At 48 h post-transfection, an increase in both the number of green fluorescence-positive cells and intensity of luciferase fluorescence was observed in cells transfected with plasmids 1–5. The expression level was highest in cells with the CYBA 5ʹUTR, followed by the albumin 5ʹUTR and α-globin 5ʹUTR. The lowest expression level was found in cells with the minimal 5ʹUTR and β-globin 5ʹUTR. At 72 h post-transfection, a decrease in both the number of green fluorescence-positive cells and the intensity of luciferase fluorescence was noted in cells transfected with all five plasmids. Nonetheless, the expression level with the CYBA 5ʹUTR remained the highest.

Fig. 3
figure 3

BHK-T7 cells transfected with recombinant plasmids containing different 5′UTR. (a) Luciferase activity in BHK-T7 cells transfected with recombinant plasmids containing different 5′UTR. (b) Green fluorescence positive cell counts following transfection with plasmids containing different 5′UTR. (c) Fluorescence images of BHK-T7 cells transfected with plasmids containing different 5′UTR “*” P < 0.05; “**” P < 0.01; “***” P < 0.001; “****” P < 0.0001.

Full size image

Reporter gene expression in cells transfected with mRNA containing the different 5′UTRs

The expression of green fluorescence in HEK293T cells was detected by CTL-ImmunoSpot S6 FluoroSpot at 24, 48, and 72 h after transfection with mRNA 1–5 with different 5ʹUTRs. After counting cells, luciferase expression was evaluated using a 96-well bioluminescence detector. The results are shown in Fig. 4. Twenty-four hours after transfection, cells transfected with mRNAs 1–5 all showed significant green fluorescence, while the negative control did not show any fluorescence. The expression pattern differed from that of the plasmid template-transfected cells. Cells transfected with mRNA 3 (β-globin 5′UTR) showed the highest counts of green fluorescence-positive cells and highest intensity of luciferase fluorescence. Cells transfected with mRNA 5 (albumin 5′UTR) exhibited lower levels of green fluorescence and luciferase expression compared to mRNA 3. The levels of expression from cells transfected with mRNA 2 (minimal 5′UTR) were slightly lower than those of the first two, while cells transfected with mRNA 1 (α-globin 5′UTR) and mRNA 4 (CYBA 5′UTR) showed the lowest expression levels from both reporter genes. At 48 and 72 h after transfection, the expression levels of the two fluorescence reporter genes in mRNA 1–5-transfected cells were enhanced. However, the number of green fluorescence positive cells and intensity of luciferase fluorescence from mRNA 3 (β-globin 5ʹUTR) were still the highest, followed by mRNA 5 (albumin 5ʹUTR), mRNA 2 (minimal 5ʹUTR), mRNA 1 (α-globin 5ʹUTR), and mRNA 4 (CYBA 5ʹUTR).

Fig. 4
figure 4

HEK293T cells transfected with mRNA containing different 5′UTR. (a) Luciferase activity in HEK293T cells transfected with mRNA containing different 5′UTR B. Green fluorescence positive cell counts following transfection with mRNA containing different 5′UTR. Fluorescence images of HEK293T cells transfected with mRNA containing different 5′UTR. “*” P < 0.05; “**” P < 0.01; “***” P < 0.001; “****” P < 0.0001.

Full size image

Reporter gene expression in HEK293T cells transfected with HIV AEgp145 mRNA containing different 5ʹUTRs

Expression of the HIV-1 envelope protein in HEK293T cells transfected with AEgp145 mRNA containing different 5ʹUTRs is shown in Fig. 5. The results showed that both control cells and cells transfected with the mRNAs expressed the internal reference protein GAPDH. The theoretical size of envelope protein gp145 is 145 kDa. A specific band can be seen in this position, with expression of mRNA 7 (β-globin 5ʹUTR) the highest, followed by mRNA 8 (albumin 5ʹUTR), and mRNA 6 (α-globin 5ʹUTR).

Fig. 5
figure 5

Expression of different 5′UTR AEgp145-mRNA in HEK293T cells. (a) Expression of different 5′UTR AEgp145-mRNA in HEK293T cells. (b) Reaction strip gray scale analysis 1 mRNA6 (α-globin 5′UTR) 2 mRNA 8(albumin 5′UTR) 3 mRNA 7(β-globin 5′UTR) 4 HEK293T cells negative control 5 Marker “*” P < 0.05; “**” P < 0.01; “***” P < 0.001; “****” P < 0.0001.

Full size image

Discussion

In this study, to explore the effects of different 5ʹUTRs on protein translation of mRNA vaccines, NLUC-T2A-EGFP dual reporter genes expression plasmids containing five different 5ʹUTR sequences (α-globin, minimal, β-globin, CYBA, and albumin) were designed and constructed. The 5′ capped mRNA were generated by in vitro transcription and capping reactions. The 5ʹUTR mRNA constructs resulting in high reporter protein amounts were identified by transfection in HEK293T cells. Then, HIVgp145 mRNA vaccines containing different 5ʹUTR were constructed, and Gp145 expression levels were verified.

The sequence and structural characteristics of mRNA UTRs are critical for stability and translation of mRNA. Imburgio et al. reported that substitution of bases at positions + 1 to + 6 with any other bases in a wild-type sequence (GGGAGA) initiated by the T7 promoter had a negative effect on promoter strength16. Therefore, the starting sequence of the 5ʹUTR designed in this study was based on the wild-type T7 promoter sequence. The six nucleotides upstream of the ATG start codon, known as the Kozak initiation site, greatly impact the efficiency of protein translation and are located at the junction of the 5ʹUTR and the open reading frame17. N1-methyl-pseudouridine is known to enhance mRNA translation efficiency and overall protein synthesis by reducing eIF2α phosphorylation and increasing ribosome density on mRNA18. Consequently, we used N1-methyl-pseudouridine during in vitro transcription. To minimize the influence of Kozak starting sites on the efficiency of protein translation, sequences upstream of the ATG start codon were converted to the optimal Kozak consensus sequence (gccaccatg), wherein the purines, usually A at the − 3 position and G at the + 4 position, are the most crucial19. Sultana et al. demonstrated that an optimized 5ʹ untranslated region significantly enhances the translation efficiency of modified mRNA in cardiac and hepatic ischemic injuries, offering promising gene therapy applications for ischemic diseases20. Based on recent clinically applied research on mRNA and 5ʹUTR sequences, five 5′UTRs, the α-globin 5′UTR, minimal 5′UTR, β-globin 5′UTR, albumin 5′UTR, and CYBA 5′UTR, were included in this study. Natural UTRs from highly expressed genes (such as the α-globin and β-globin genes) are the preferred choice for synthesizing mRNAs21. The BNT162b2 vaccine from Pfizer/BioNTech uses the 5ʹUTR of human α-globin22. In a study of Trepotec Z, the expression level of an mRNA containing a mini 5ʹUTR was comparable to or even stronger than that of the human α-globin 5ʹUTR23. A study by Ferizi et al.24 showed that the human CYBA 5ʹUTR sequence can increase the translation efficiency of mRNA without affecting the half-life of mRNA transcripts. In this study, we evaluated the rank order of expression from the above five 5ʹUTR variants in plasmids and in vitro transcribed mRNA. Using plasmids to test the mRNA 5ʹUTRs offered a faster screening approach than using in vitro transcribed mRNA; however, our results showed that there was no correlation between plasmid-driven protein expression and in vitro-generated mRNA expression. Higher reporter gene expression levels were detected in the plasmids containing CYBA 5ʹUTR and albumin 5ʹUTR, while lower reporter gene expression was driven by plasmids containing the α-globin 5ʹUTR, minimal 5ʹUTR, and β-globin 5ʹUTR. When HEK293T cells were transfected with the mRNA UTR constructs, the reporter gene expression level of β-globin 5′UTR was the highest, and that of CYBA 5′UTR was the lowest. This discrepancy aligns with findings by Asrani et al., who noted differences in protein expression levels between plasmid-driven and in vitro transcribed mRNA-driven protein expression25. A possible reason for this difference between expression from plasmids and mRNAs may be that transcription occurs within nuclei of eukaryotic cells, while translation occurs within the cytoplasm after plasmid transfection. Protein translation is a complex process involving multiple protein–protein and protein-mRNA interactions. This process includes a number of RNA binding proteins that bind to mRNA transcripts prior to export from the nucleus26. Soucek et al. found that the poly(A) RNA binding protein Nab2 plays a key regulatory role in post-transcriptional processing, mRNA maturation, and gene expression regulation27. mRNA synthesized in the nucleus may undergo RNA editing or chemical modification, which can affect the stability and transcriptional or translation efficiency of RNA molecules. A study by Akhtar et al., for example, showed that M6A modification of mRNA may affect mRNA stability, recruitment of RNA-binding proteins, and translation and splicing28.

mRNA transcribed in vitro is directly transfected into the cytoplasm and thus bypasses these traditional mRNA modifications that occur in the nucleus. Therefore, the 5ʹUTR structure has a more significant effect on translation in this method, because it is the binding site for the pre-initiation complex in protein translation29,30. For example, binding of eukaryotic initiation factor-4A (eIF4A) to the 5ʹUTR is important for the unwinding initiation of protein translation31, and the secondary structure of the 5ʹUTR plays a key role in the binding of eIF4A and mRNA32. For translation initiation that depends on the 5ʹ cap structure, the ribosome must recognize and bind to this structure and then scan along the 5ʹUTR to the start codon, which may be affected by the length and secondary structure of the 5ʹUTR. Moreover, the position of secondary structures in the 5ʹUTR will have a unique effect on protein expression, and the stability of the hairpin structure and distance from the Cap will have a significant impact on translation efficiency. Based on the biological computing platform Paddlehelix, we predicted the secondary structure and hairpin structure stability of the different 5ʹUTR sequences. The predicted thermodynamic stabilities of the five 5ʹUTR hairpin structures were all less than − 30 kcal/mol. Among them, the stem-loop structures of the α-globin 5ʹUTR and CYBA 5ʹUTR were at the starting positions of the caps, the distance between the stem-loop structures of the albumin 5ʹUTR and cap was + 30, that of the minimal 5ʹUTR and cap was + 1, and that of the β-globin 5ʹUTR and cap was + 9. According to research by Babendure et al., when the stability of the hairpin structure is less than − 35 kcal/mol, translation efficiency is significantly correlated with its distance from the cap, translation efficiency increases linearly when the hairpin structure is from 0 to + 10 from the cap, and, when the hairpin structure is more than + 10 from the cap, translation efficiency begins to decline33. Unstructured 5ʹUTRs generally favor higher translation initiation rates34, whereas stem-loop structures immediately adjacent to the cap substantially suppress initiation33. Our results from the in vitro reporter gene expression experiment with mRNA constructs also showed that the distance between the hairpin structure and cap position affected mRNA translation efficiency.

To further investigate the impact of the different 5′UTRs on translation of mRNA vaccines expressing foreign genes, three of above five 5′UTRs (α-globin, β-globin, and albumin UTRs) were selected to construct mRNA vaccines expressing the HIV AEgp145 gene. HEK293T cells were transfected with the mRNA constructs, and, consistent with the results with dual-reporter gene mRNAs, the Gp145 protein expression level with the β-globin 5′UTR was the highest, followed by the albumin 5′UTR and α-globin 5′UTR.

In this study, AEgp145 gene expression with the different 5ʹUTR regions in mRNA vaccines were compared, with results highlighting the importance of the 5′UTR in vaccine expression efficiency. Through our analysis of the different 5′UTRs, we found that the β-globin 5′UTR resulted in the highest expression, as verified by both reporter gene and Gp145 expression. In future studies, immunogenicity of the Gp145 mRNA vaccine will be explored in an animal model and mRNA vaccines targeting other pathogens will be developed.

Methods

Design and selection of UTRs

We designed and evaluated five distinct 5′UTR sequences (Table 1), including a minimal 5′UTR comprising of only the T7 promoter, transcription start site, and Kozak sequence. The remaining four 5′UTR variants were derived from human genes: α-globin, β-globin, cytochrome B-245α chain (CYBA), and albumin. As for the 3′UTR sequences, all were sourced from the α-globin gene. A 110-nucleotide (nt) poly(A) tail was appended to the 3′UTR end. Both the 3ʹUTR and poly(A) sequence were synthesized by Tianyi Huiyuan (Beijing, China) and subsequently cloned into the pKMV vector and named pKMV-α-globin 3′UTR-PolyA. The secondary hairpin structures of the 5ʹUTR sequences were predicted using the Paddlehelix biological computing platform (Fig. 2).

Construction of dual-reporter gene expression plasmids containing different 5ʹUTRs

The coding sequences for nano-luciferase (NLUC) and enhanced green fluorescent protein (EGFP) were amplified from PNL-2 and PX458 plasmids, respectively, both of which were previously constructed in our laboratory. These sequences served as reporter genes and were linked via a T2A peptide, which mediates ribosome jumping and protein translation cutting, to form an NLUC-T2A-EGFP fusion construct. The five 5ʹUTR sequences were introduced into this NLUC-T2A-EGFP construct using PCR. For each 5ʹUTR, a specific set of primers was designed (Table S1). The PCR products and vector plasmid pKMV-α-globin 3′UTR-PolyA were digested with XbaI/XhoI and ligated to generate different recombinant plasmids. The successful ligation products were confirmed by sequencing, and the recombined plasmids were named plasmids 1–5.

Construction of HIV AEgp145 gene expression plasmids containing different 5ʹUTRs

The HIV AEgp145 coding sequence was amplified from a pVR-gp145 plasmid that was previously constructed in our laboratory. This sequence was then combined with three distinct 5ʹUTR sequences using PCR. Specific sets of primers were designed for each 5ʹUTR variant (Table 1). HIV AEgp145 gene expression plasmids were constructed in a similar way as the reporter gene plasmids. The recombinant plasmids were verified through sequencing and named as plasmids 6–8.

Generation of mRNAs

To generate in vitro transcribed mRNAs, plasmids were linearized downstream of the poly(A) tail by BspQI enzyme digestion and subsequently purified with DNA magnetic beads (Vazyme, China). The purified linear plasmids served as templates for in vitro transcription with T7 High Yield RNA Transcription Kit N1-Me-Pseudo UTP (Vazyme, China). Post-transcriptional mRNA products were purified using VAHTS RNA Magnetic Beads (Vazyme, China). RNA quality and concentration were assessed using a NanoDrop 2000C spectrophotometer (Thermo Fisher Scientific). A Vaccinia Capping System kit was utilized for mRNA capping following the operational guidelines provided. Finally, the capped mRNA product was purified using VAHTS RNA Magnetic Beads (Vazyme, China).

Cell culture

HEK293T and BHK-T7 cells (BHK cells stably expressing T7 RNA polymerase) were maintained by Zeng Yi at the Institute of Virus Disease Prevention and Control, Chinese Center for Disease Control and Prevention. These cell lines are cultured in low-glucose Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. All cell lines are incubated in a humidified environment at 37 °C with a 5% CO2 level.

Plasmid and mRNA transfection

BHK-T7 and HEK293T cells were seeded in 96-well plates at a density of 1 × 104 cells per well. Transfections with plasmid/mRNA were performed when cells reached 80% confluency. BHK-T7 cells were transfected with the different 5ʹUTR recombinant plasmids (plasmids 1–5) using Lipo8000 transfection reagent. The vector plasmid pKMV-α-globin 3ʹUTR-PolyA was used as a negative control. In parallel, the 5ʹUTR mRNAs were transfected into HEK293T cells using jetMESSENGER transfection reagent, with the pKMV-α-globin 3ʹUTR-PolyA mRNA serving as a negative control. All transfections were conducted according to the respective kit instructions.

Nano luciferase assay and fluorescence experiment

At 24, 48, and 72 h post-transfection, the expression of green fluorescence from the reporter gene (both plasmid and mRNA) was detected using a CTL-ImmunoSpot S6 FluoroSpot, and positive cells were quantified. Subsequently, the cells were lysed using Nano-Glo Luciferase Assay, and the intensity of luciferase expression was measured using a 96-well plate bioluminescence detector.

Protein detection. At 48 h post-transfection, HEK293T cells were collected in RIPA lysis buffer. Total protein concentrations were determined by BCA assay (ThermoFisher Scientific, Waltham, MA) following the manufacturer’s protocol. The cell extracts were fractionated by 10% SDS-PAGE and subjected to western blotting with 2G12 as the primary antibody (1:1000 dilution) and goat anti-human-HRP (1:2000 dilution) as the secondary antibody to evaluate expression of the Gp145 protein. The immunocomplexes were detected using an enhanced chemiluminescence system. The whole uncropped images of the original western blots in (Fig. S1).