Figure 1 displays the non-formylated methionine process for protein synthesis in eukaryotic cells using E. coli Met-tRNAi. First, we engineered and heterologously expressed fMet-tRNAi, wherein two types of E. coli Met-tRNAi were introduced: E. coli Met-tRNAi/EcMRS and E. coli Met-tRNAi/human methionyl tRNA synthetase (HMRS). We compared the activity of human Met-tRNAi and E. coli Met-tRNAi by monitoring the expression of PII such as green fluorescent protein (GFP) or fluorescent N-term-labeled HIV transcription activator protein (Tat) as PII model proteins. Human Met-tRNAi was used as a positive control for the fluorescence-labeled E. coli Met-tRNAi-transfected cells (Fig. 1). Codons and amino acids were then assigned to individual tRNAs through selective aminoacylation of the tRNAs using aminoacyl tRNA synthetases [43]. Additional file 1: Fig. S1 [42, 43] shows that tRNAs have a cloverleaf structure and contain all invariant and semi-invariant sequences in both prokaryotic and eukaryotic initiators [44,45,46,47].
The preservation of the sequence in eukaryotic tRNAis is a unique property of tRNAi (Additional file 1: Fig. S1A). The unique characteristics of prokaryotic tRNAi may explain the strong conservation of the Watson–Crick base pair at the end of the receptor stem and the presence of the 11-purine:24-pyrimidine base pair (instead of the 11-pyrimidine:24-purine base pair) (Additional file 1: Fig. S1). Three consecutive GC pairs, including sequences of three guanines and three cysteines, were formed in the anticodon stem in both prokaryotic and eukaryotic tRNAis (Table 1). The schematic in Fig. 1 summarizes the experiment and compares the activity of human and E. coli tRNAis during the initiation of PII synthesis in mammalian cells.
We assessed E. coli fluorescent Met-tRNAi, using the size-exclusion chromatography (SEC) method to separate unlabeled methionine-conjugated E. coli tRNAi and 5-FAM or methionine (Fig. 2A). The peaks shown on the left of Fig. 2A represent free methionine (210 nm, 13 min), the fluorescent label alone (210 nm, 12 min), and the fluorescence detector (FLD). We previously reported the N-terminal labeling of target proteins with the strong chromophore Cy5-Met [48,49,50,51]. Purified Cy5-Met was chemically synthesized and then purified to validate the occurrence of protein synthesis mediated by single E. coli fluorescent Met-tRNAi.
We also chemically synthesized 5-FAM-Met, which was then coupled to synthetic E. coli tRNAi, as demonstrated by magnetic resonance spectroscopy. E. coli tRNAi was generated via in vitro transcription and separated according to mass from 5-FAM (473.39 Da), E. coli tRNAi (26.07 kDa), and methionine alone (149 Da) (Fig. 2A). Here, 5-FAM-labeled methionine-charged E. coli tRNAi clearly overlaid the fluorescence detection spectra and could be distinguished by its absorbance peak of 210 nm at 5.3 min (Fig. 2B) and a peak at 260 nm at 5.3 min.
The peak corresponding to 5.3 min was further analyzed via matrix-assisted laser desorption ionization time-of-flight mass spectrometry (Fig. 2C and Additional file 1: Fig. S2). As an additional control, we compared the methionylation of fluorescent methionine-conjugated human tRNAi with fluorescent methionine-conjugated E. coli tRNAi using 2% agarose gel electrophoresis, followed by imaging with 260-nm UV irradiation and an FLD (Additional file 1: Fig. S3). To study the effect of the non-formylated methionine-conjugated E. coli tRNAi on translation recognition in HeLa cells, we purified EcMRS based on its affinity for E. coli fluorescent Met-tRNAi. This enzyme is activated by the amine group of unpurified fluorescent methionine, indicating that non-formylated, rather than formylated, methionine is charged by EcMRS. The effect on a single fluorescent methionine-binding E. coli tRNAi was measured for fluorescently labeled synthetic proteins using E. coli Met-tRNAi/EcMRS.
To evaluate whether E. coli Met-tRNAi conjugated with the N-terminal fluorophore 5-FAM (Ex: 492 nm; Em: 512 nm), HeLa cells were extracted from the C-terminal interaction partner fused to the AVI-Tagged Tat protein to confirm fluorescent Tat expression (Fig. 2D). To verify the translation efficiency of the fluorescence-labeled protein using E. coli Met-tRNAi/EcMRS pairs, we prepared a single fluorescent methionine-conjugated E. coli tRNAi with an optimized N-terminal recognition motif, and the C-terminal interaction partner was fused to AviTag. Fluorescent protein imaging was performed using biotin double-labeled Tat on polyethylene streptavidin-coated quartz slides.
Additional file 1: Fig. S4 presents gel images of the fluorescence-labeled E. coli Met-tRNAi-mediated N-terminal labeling of Tat proteins. To assess E. coli tRNAi, we investigated whether the fluorescent E. coli Met-tRNAi was involved in the translation of the reading frame using the Tat protein, a factor essential for the transcription of the HIV-1 genome.
Fluorescent-labeled human tRNAi charged endogenous human MRS (lane 2) and E. coli tRNAi charged EcMRS (lanes 3, 4, and 5), identified via Tat expression as a positive control (17 kDa), were analyzed through fluorescence scanning and SDS-PAGE (Additional file 1: Fig. S4). The results show that E. coli Met-tRNAi/EcMRS pairs migrated in a fluorescent band pattern with fluorescent-labeled Tat protein expression as the positive control. One of the two suggested hypotheses was that the E. coli Met-tRNAi/HMRS pair with the fluorescent band would result in Tat identification as a 17-kDa band when expressed Tat was dyed with Coomassie brilliant blue staining solution (lane 1 and 2 in Additional file 1: Fig. S4). HMRS and EcMRS utilization led to the identification of mRNAs involved in Tat expression (Figs. 3 and 4). In contrast, the efficiency of protein initiation synthesis by fluorescent-Met-tRNAi/EcMRS was not well detected because of the disruption of endogenous protein expression in the Coomassie-stained gel.
Our results show that E. coli Met-tRNAi enables target protein synthesis in live HeLa cells and allows visualization of the proteins in the fluorescent imaging molecules in response to protein translation mediated by tRNAi/MRS pairs. In addition, fluorescence-tagged Tat protein expression occurred simultaneously with Tat protein expression based on tRNAi translation. Various results for the fluorescent E. coli Met-tRNAi strategy mediated PII synthesis are presented in Table 2.
As shown in Fig. 2, fluorescent methionine was chemically synthesized and then purified to confirm the occurrence of protein synthesis [40, 52, 53]. We introduced fluorescence-labeled Tat protein with non-purified fluorescence-labeled Met-conjugated human tRNAi to HeLa lysates. To demonstrate further the evaluation of this approach, the non-purified fluorescent Tat protein obtained through sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE), we examined the fluorescent Tat, which was double-tagged with biotin, on streptavidin-coated slides. The initiation of protein synthesis suggested that tRNAis can become novel regulators that maintain the translational reading frame.
Therefore, the results of this study provide new insights into the optimization of initiation tRNAi motifs to allow control over the initiation of protein synthesis. Moreover, our observations of the fluorophore-E. coli Met-tRNAi bases further advance the understanding of protein expression in eukaryotic cells by revealing the various initiating tRNA sequence motifs involved in protein synthesis and fidelity. Therefore, the results of this study provide new insights into the roles of various tRNAis in protein fidelity and could be used to develop a strategy for the optimization of initiation tRNAi motifs to facilitate precise protein synthesis.
We validated Tat expression using a fluorescence-labeled, methionine-conjugated human tRNAi probe (Fig. 1). As previously reported, EGFP expression was measured in an L-Met–deficient medium, demonstrating the translational recognition of human Met-tRNAi-mediated selective protein expression in HeLa cells [38, 40]. In the Met-sufficient growth medium, cells were cultured with the other 19 natural amino acids, while fluorescence-labeled methionine was added to the cells during the stationary growth phase. To initiate E. coli Met-tRNAi-mediated selective protein synthesis, we co-transfected an EGFP-expressing plasmid with human Met-tRNAi into HeLa cells. As shown in Fig. 3A, EGFP-expressing cells exhibited lower levels of EGFP mRNA, and EGFP expression was reduced in both methionine-free and -sufficient cells.
EGFP mRNA levels were determined using quantitative real-time polymerase chain reaction (qRT-PCR) and were correlated with green fluorescence (Fig. 3A, left). Despite the presence of low levels of endogenous tRNAi in the L-Met–free medium, which indicates low levels of EGFP expression, this strategy was found to be applicable to the use of E. coli tRNAi as the positive control (Fig. 3A). The flow cytometry results indicated that EGFP expression gradually increased in the methionine-supplemented medium, as did the level of EGFP. As shown in Fig. 3 (right and left panels), the analysis of EGFP via flow cytometry following incubation in Met-deficient media indicated that EGFP was highly expressed in Met-supplemented media (GFP population 1.64, dark violet), while there was a clear quantitative separation of EGFP mRNA expressed in Met-deficient medium (GFP population 0.67, orange).
Flow cytometry dot plots of EGFP expression in methionine-free cell lines expressing EGFP were also analyzed to confirm consistency with the EGFP mRNA levels. The level of green emission observed from EGFP-expressing cells was approximately two-fold higher (GFP population 1.64, dark violet) than that from methionine-free HeLa cells (GFP population 0.67, orange; Fig. 3A). Based on these results, E. coli Met-tRNAi can be used for protein synthesis in methionine-free cells. The expression of GFP mRNA increased after the introduction of EGFP-transfected HeLa cells, but L-Met deficiency and L-Met sufficiency acted as positive controls in correlation with HeLa cell viability (Fig. 3B). Thus, the effect of L-Met on HeLa cell viability occurs during initial translation. The decrease in the viability of HeLa cells without methionine was not significant compared with that of control HeLa cells with sufficient methionine (Fig. 3B). This result provides experimental proof of the Hoffman effect [54], whereby malignant cells can endogenously synthesize high levels of methionine, which is insufficient to sustain cancer growth. Human tumor cells have been reported to depend on exogenous methionine, whereas normal tissue may depend on homocysteine to meet methionine needs [55, 56]. This means that malignant cells can adapt to glucose-poor conditions by adopting alternative metabolic pathways that support growth [57,58,59]. Therefore, by introducing our methionine-deficient medium condition, the viability characteristics of methionine-dependent HeLa cells were efficiently applied to the experimental system to illustrate the significance of E. coli tRNAi-mediated initiation protein synthesis and previous reports [55, 56].
To evaluate the activity of E. coli tRNAi, EGFP was expressed by transfecting HeLa cells in a L-Met–free medium. The E. coli Met-tRNAi/EcMRS-charged E. coli tRNAi, which was used as a bio-orthogonal reaction, was paired with proteins produced in eukaryotic hosts (Fig. 4A). The EGFP mRNA levels determined through qRT-PCR demonstrated a correlation with the E. coli Met-tRNAi/EcMRS pairs. In the experiment involving E. coli Met-tRNAi/EcMRS pairs alone, EGFP mRNA expression (blue bars), EGFP expression dot plots (Additional file 1: Fig. S4, GFP group 0.58, blue), and EGFP mRNA expression induced with L-Met were observed in the methionine-deficient medium (Fig. 4A; GFP population 0.35, orange). In addition, EGFP mRNA expression was observed after introducing E. coli Met-tRNAi/EcMRS pairs alone with L-Met–sufficient HeLa cells used as a control (Fig. S4: GFP population 0.01, red). Figure 4B shows that HeLa cell viability was induced by GFP mRNA expression. In HeLa cells without methionine, cell viability, which was introduced by adding 1, 5, and 10 nM of the E. coli Met-tRNAi/EcMRS pair (blue bar), was dependent on the concentration change of the HeLa cell control with sufficient methionine. The E. coli Met-tRNAi/EcMRS pair showed significant toxicity at a concentration of 10 nM, as shown in Fig. 4B.
The results suggest the possibility of an association with the GFP mRNA expression level introduced by the E. coli Met-tRNAi/EcMRS pair (blue bar) shown in Fig. 4A. As shown in Fig. 4, the expression of GFP mRNA decreased after the introduction of the E. coli Met-tRNAi/HMRS pair (green bar), unlike that in the E. coli Met-tRNAi/EcMRS (blue bar) pair comparison experiment. The expression of GFP mRNA increased after introducing the E. coli Met-tRNAi/HMRS pairs (green bars), as shown in Fig. 4C. The expression of E. coli Met-tRNAi/HMRS pairs was higher under methionine-deficient HeLa cell conditions than that observed when using E. coli Met-tRNAi/EcMRS pairs (Fig. 4C, GFP population 14.8, green). Furthermore, the expression of GFP mRNA decreased after introducing the E. coli MRS pair. The results for the HMRS group were compared with that of the introduction of an E. coli initiator charged with EcMRS (Fig. 4C; GFP population 4.81, blue) as a replacement for HMRS (Fig. 4C, GFP population 14.8, green). We found that the expression of GFP mRNA increased after the introduction of EGFP expression.
By comparing the results obtained for MRS substitution, we found that MRS from humans and E. coli synthesized PIIs in the cytoplasm. These results indicate that various tRNAis potentially act on different tRNA sequence motif interactions.
These results suggest that introducing E. coli Met-tRNAi/EcMRS pairs in the initial translation frame for GFP mRNA expression inhibits protein initiation due to cytotoxicity.
Next, to determine the viability of E. coli tRNAi in non-formylated methionine used in the initial translational reading frame, samples with non-formylated and formylated methionine were compared by monitoring the viability of HeLa cells. In methionine-free HeLa cells, non-formylated methionine-conjugated E. coli Met-tRNAi/EcMRS pairs (1, 5, and 10 nM) exhibited a significantly higher level of toxicity compared with that seen in formylated methionine-conjugated E. coli Met-tRNAi/HMRS pairs (Fig. 5 and Additional file 1: Figs. S5 and S6). Non-formylated methionine bound to E. coli tRNAi was associated with lower cell viability than formylated methionine bound to E. coli tRNAi.
In the absence of the introduction of protein initiation, the HeLa cell viability results suggest that methionine is a complex metabolic pathway that is associated with cancer at multiple levels. However, the lower HeLa cell viability observed when non-formylated methionine was introduced during the initiation of protein synthesis suggests that the cytotoxicity induced in the early translation frame of GFP mRNA expression is significantly associated with EcMRS. Generally, most cells readily synthesize methionine from homocysteine but cannot proliferate in a medium where methionine has been replaced with homocysteine. Therefore, the results shown in Figs. 5A and 5B suggest that the increased viability of HeLa cells in the methionine-deficient state (orange bars) could mean that HeLa cells replaced methionine with homocysteine [55] are associated with increased HeLa cell viability.
Although the accurate determination of the efficiency of synthesized proteins remains controversial, catalyzing the linking of amino acids to cognate transfer RNAs (tRNAs) could be controlled by inducing the PII expression. We found that the selective linking of a fluorescent methionine to E. coli Met-tRNAi was highly precise, as shown in Fig. 5. Moreover, every EcMRS paired with methionine could be promoted as a newly synthesized protein in HeLa cells. Many types of aaRS can be tracked via their pairing with cognate tRNAs to investigate the fidelity of protein translation.
