Enzymatic Modification of the 5’ Cap with Photocleavable ONB‐Derivatives Using GlaTgs V34A

Abstract

The 5’ cap of mRNA plays a critical role in mRNA processing, quality control and turnover. Enzymatic availability of the 5’ cap governs translation and could be a tool to investigate cell fate decisions and protein functions or develop protein replacement therapies. We have previously reported on the chemical synthesis of 5’ cap analogues with photocleavable groups for this purpose. However, the synthesis is complex and post‐synthetic enzymatic installation may make the technique more applicable to biological researchers. Common 5’ cap analogues, like the cap 0, are commercially available and routinely used for in vitro transcription. Here, we report a facile enzymatic approach to attach photocleavable groups site‐specifically to the N² position of m⁷G of the 5’ cap. By expanding the substrate scope of the methyltransferase variant GlaTgs V34A and using synthetic co‐substrate analogues, we could enzymatically photocage the 5’ cap and recover it after irradiation.

A facile enzymatic approach to attach photolabile groups to the 5’ cap is reported. It starts from commercially available m⁷Gp₃G and avoids complicated synthesis. By expanding the substrate scope of the methyltransferase variant GlaTgs V34A and using AdoMet analogues we successfully photocaged the 5’ cap and could incorporate it into mRNA by in vitro transcription. The cap 0 was recovered after irradiation.

Introduction

The 5’ cap is a hallmark of eukaryotic transcripts produced by RNA polymerase II, such as mRNAs and long non‐coding RNAs (lncRNAs), but also short RNAs, like snRNAs and miRNAs. The 5’‐5’‐triphosphate linkage between a methylated guanosine and the first transcribed nucleotide is formed co‐transcriptionally. ^[1] The 5’ cap plays a critical role in processing of pre‐mRNA, including splicing and polyadenylation, in nuclear export and also in protection from exoribonucleases. ^[2] One fundamental and highly conserved cap‐dependent process is translation in eukaryotes, which requires direct interaction with the eukaryotic initiation factor 4E (eIF4E). ^[3] Modifications of the 5’ cap that impact the binding affinity to eIF4E also affect translation efficiency. ^[4] These findings promoted the idea of introducing modifications to the 5’ cap to modulate translation.

The regulation of translation plays a major role in cell differentiation and proliferation. ^[5] For external control, photocleavable protecting groups are of interest as they can be removed by light. Light is a bioorthogonal trigger and can be controlled to act only in a defined location at a specific time. Therefore, the ability to manipulate bioactive molecules by light provides access to spatiotemporal control of their activity. ^[6] The “photocaging” of nucleosides, nucleotides and oligonucleotides is an established method. ^[7] These photocaged bioactive molecules show the potential to improve drug delivery, bioavailability or reduction of side effects in biomedical applications. ^[8]

Recent advances have enabled the use of mRNA as a new modality in medicine, e. g. as vaccine against infectious diseases, for personalized cancer treatment, and for treatment of autoimmune diseases. ^[9] However, methods to control gene expression at the mRNA level are scarce.

Recent advancements in the direction of post‐transcriptional control of gene expression have led to the development of optochemical tools. Among those, many rely on short regulatory RNAs, like siRNAs, miRNAs, morpholinos or aptamers. ^[10] Other strategies act directly on the mRNA itself by introducing multiple photocleavable groups into the backbone, ^[11] or structural motifs into the 5’ UTR. ^[12] While this approach is versatile and straightforward compared to the addition of other oligonucleotides, it still requires multiple photolabile groups or the introduction of structural motifs into the 5’ UTR of mRNA.[ ¹¹ , ¹² ]

In a recent study, we introduced FlashCaps, i. e. 5’ cap analogues that carry a photolabile group connected via a carbamate linker to the exocyclic N² position of the m⁷G. This approach yielded promising results, showing efficient optical control of mRNA translation by introducing a single modification and not altering the sequence. However, an elaborate setup is required to synthesize, purify and analyze these FlashCaps. ^[13]

Here we report an enzymatic route to synthesize 5’ cap analogues bearing photolabile groups at the N² position of m⁷G. Using a one‐step synthesis procedure to produce S‐adenosyl‐l‐methionine (AdoMet) analogues, we were able to transfer ortho‐nitrobenzyl (ONB)‐derivatives with a methyltransferase to the N² position of the 5’ cap. The Giardia lamblia trimethylguanosine synthase variant V34A (GlaTgs V34A) efficiently transferred the ONB‐derivatives and irradiation led to partial recovery of the native cap analogue in vitro. ^[14] The 5’ cap analogues were successfully incorporated into mRNA during in vitro transcription. After transfection of mammalian cells with modified reporter‐mRNA, a luciferase assay validated higher protein expression after irradiation.

Results and Discussion

We aimed to develop a fast and efficient method to enzymatically photocage the 5’ cap as an alternative to the elaborate chemical synthesis and purification of light‐activatable FlashCaps. To achieve this, the natural 5’ cap analogue, m⁷Gp₃G (1), was modified with a methyltransferase variant (GlaTgs V34A) previously engineered for improved promiscuity.[ ¹⁴ , ¹⁵ ] We prepared a set of synthetic analogues of the co‐substrate AdoMet bearing derivatives of the photocleavable ONB group at the sulfonium center. Specifically, AdoMet analogues with the ortho‐nitrobenzyl (AdoONB 2 a), ^[16] 4‐bromo‐2‐nitrobenzyl (AdoPBr 2 b) or 4,5‐dimethoxy‐2‐nitrobenzyl (AdoDMNB 2 c) (Supporting Information section 4) were synthesized and tested in enzymatic reactions with 1 (Scheme 1a). The steric demand of the photocleavable group on the resulting products 3 a–c was anticipated to interfere with the availability of these modified 5’ caps for proteins responsible for further processing or binding of mRNA. The photocaged caps 3 a–c should then become activated by irradiation with UV‐light (365 nm) to reconstitute the native 5’ cap (1) available for natural binding partners (Scheme 1b).

Scheme 1.

General concept of the enzymatic approach for photocaging the 5’ cap. a) Enzymatic transfer of a photocleavable group from different AdoMet analogues (2 a–c) to the exocyclic N² position of the m⁷G of cap 0 (1) using GlaTgs V34A as promiscuous methyltransferase. b) Scheme illustrating photo‐uncaging of the N² modified cap analogues by irradiation with 365 nm for 30 s in order to retrieve the unmodified 5’ cap analogue (1) again. PG: photocleavable group.

In a previous approach to photocage the 5’ cap, we had focused on the N7 position of the guanosine. ^[4b] However, at this position, the photolysis of ONB‐derivatives was not successful in the sense that the native 5’ cap structure was not reconstituted. In contrast, the enzymatic modification at the N⁶ position of adenosine with ONB and nitropiperonly groups at internal sites of RNA was successful and could also be removed by light. ^[16a] In the current study, we chose the exocyclic N² position of the 5’ cap m⁷G as modification site. We reasoned photolysis should be possible without extensive side‐product formation and lead to the reconstitution of the native 5’ cap structure.

First, we tested whether GlaTgs V34A is able to transfer the sterically demanding groups, as it was only reported to transfer small aliphatic or para‐substituted benzylic residues before. ^[4a] The enzymatic reactions were incubated and analyzed via HPLC in order to evaluate the formation of modified 5’ cap analogues 3 a–c (Figure 1C). The HPLC analysis showed that transfer of aromatic rings is possible, even with substituents in the ortho position. Importantly, the ONB group from 2 a and even the red‐shifted DMNB group from 2 c were transferred, demonstrating the applicability of GlaTgs V34A to transfer photocleavable groups to 5’ cap analogues and generate 3 a–c as well as 4. However, we observed that each additional substituent on the aromatic ring lowered the conversion (Table 1). Purification of 3 a–c and 4 was performed via HPLC to remove residual reactants and degradation products as the AdoMet analogues are instable in neutral pH. ^[17] After purification of the photocaged 5’ cap analogues via HPLC, their absorbance spectra were measured, confirming the presence of the photocleavable group at the m⁷Gp₃G (Figure 1B).

Figure 1.

Enzymatic 5’ cap modification. a) Enzymatically modified 5’ cap analogues made in this study. b) Absorbance spectra of 1 and 5’ cap analogues made in this study. c) HPLC chromatograms of the enzymatic modification reaction with GlaTgs V34A with indicated AdoMet analogues. The reaction was stopped either immediately (0 h) or after 8 h at 37 °C (8 h) by incubation at 65 °C for 10 min. Colored peaks correspond to indicated compounds (as validated by LC/MS) and areas were used to calculate conversions. The HPLC chromatograms show a representative of n=3 independent replicates.

Table 1.

Conversion of m⁷Gp₃G (1) with different photocleavable groups by enzymatic modification. The values result from an integration of peaks in HPLC analysis and represent the average and standard deviation. of n=3 independent replicates.

AdoMet analogue	Conversion [%]
AdoBenzyl	69±7
2 a	49±3
2 b	38±6
2 c	40±9

To test whether irradiation of 3 a–c with UV‐light leads to an efficient photolysis reaction and re‐formation of 1, we irradiated the purified 5’ cap analogues and analyzed the resulting mixture via HPLC (Figure 2). After irradiation for 30 s, the photolysis reaction was completed, and no starting material remained detectable. At the same time, a new peak corresponding to the native 5’ cap was observed (Figure 2, S5). As a control, compound 4 with a benzyl ring instead of a photocleavable group was tested but did not lead to formation of 1 or another new product upon irradiation (Figure 2). These data show that enzymatically produced 5’ cap analogues 3 a–c with photolabile ONB‐derivatives at the N² position of the m⁷G are in principle suitable for releasing the native 5’ cap (1) by irradiation.

Figure 2.

Photolysis of the photocaged 5’ cap analogues (3 a–c) and control (4). HPLC traces of purified 3 a–c and 4 before and after irradiation (365 nm, 30 s). Guanosine was added as an internal standard. The HPLC chromatograms show a representative of n=3 independent measurements.

However, upon irradiation of 3 a–c, in addition to the desired peak for the native 5’ cap 0 (1), we observed several smaller peaks, indicating the formation of side‐products during the photolysis reaction (Figure 2). Although the irradiation of the photocaged 5’ cap analogues 3 a–c does not yield quantitative conversion to 1, a significant fraction was successfully photo‐deprotected (Table 2). This supports our hypothesis that photocaging of an exocyclic position is favorable for native 5’ cap reconstitution.

Table 2.

Conversion to the natural 5’ cap (1) after irradiation. The values result from an integration of peaks in HPLC analysis and represent the average and standard deviation of n=3 independent experiments.

Cap analogue	Conversion to 1 [%]
4	0±0
3 a	43±3
3 b	36±5
3 c	33±2

As the ONB‐caged 5’ cap 3 a showed the most efficient recovery of 1 among the tested photolabile groups, we wanted to find out whether this modified 5’ cap could be incorporated into a reporter mRNA (Figure 3a). In vitro transcription using T7 polymerase was performed in the presence of purified 3 a for transcriptional priming. The ONB‐cap 3 a was as well incorporated as cap 0 during in vitro transcription and similar amounts of capped mRNA were obtained after digestion of uncapped mRNA (Figure 3a). Irradiation of the mRNA at 365 nm for 30 s did not cause degradation of mRNA capped with 1 or 3 a (Figure 3a). To evaluate the effect of modifications at the N² position of m⁷G on translation, we subsequently transfected HeLa cells with cap‐caged‐mRNA and cap 0‐mRNA as control and measured the resulting luminescence (Figure 3b).

Figure 3.

mRNA production and translation studies. a) Polyacrylamide gel electrophoresis of Gaussia luciferase (GLuc) mRNA (=821 nt) made by in vitro transcription (Supporting Information Table S1) capped with 1 or 3 a, respectively, with or without irradiation before loading (365 nm, 30 s). b) In‐cell translation studies based on luminescence measurements performed with non‐irradiated and irradiated (365 nm, 30 s) GLuc‐mRNA transfected into cells. The bars represent the average of n=6 independent experiments and the error bars represent the standard deviation. Statistical analysis: two‐tailed t‐test. p<0.1:*, p<0.05:**,p<0.01:***

Compared to cap 0‐mRNA, luciferase activity was reduced to 49 %, when the 5’ cap was modified with the ONB group (Figure 3a), confirming that interaction with eIF4E and translation are impaired. The irradiation of the mRNA prior transfection led to a 1.3‐fold increase in luciferase activity compared to the sample without irradiation. These data show that the photocleavable group can also be removed from cap‐caged mRNA and that the released cap 0‐mRNA can be translated. The relatively low increase suggests that the small photocleavable group does not result in complete inhibition of translation. In addition, we observed already in vitro that the native cap structure is only partially restored (43 %).

These results show that enzymatic transfer of photocleavable groups to the 5’ cap and subsequent photo‐deprotection to the native 5’ cap are possible. This work provides a first proof‐of‐concept that enzymatic 5’ cap modification is an alternative route to obtain mRNAs whose translation can be activated by light.

Conclusion

We expanded the substrate scope of a promiscuous methyltransferase to photocleavable groups. We showed that ONB‐based photocleavable groups can be efficiently transferred to the 5’ cap and yield photoactivatable bioactive compounds. This enzymatic approach offers advantages compared to the synthetic route, as it is fast and cost‐efficient, starting from commercially available 5’ cap. Several photocleavable groups have been tested and analyzed in terms of absorbance, enzymatic conversion and recovery of the native 5’ cap after irradiation. The irradiation experiments revealed that 33–43 % of the native 5’ cap are restored after irradiation, which is a reasonable improvement compared to the enzymatic N7 modification of G with ONB‐groups, which did not yield the native compound at all. ^[4b] This confirms that the electronic situation of the imidazolium ion upon N7G modification of the 5’ cap plays a critical role for side‐product formation with ONB‐derived groups. However, compared to our recent publication, in which the photocleavable group is connected via a carbamate linker to the N² position of the m⁷G of the 5’ cap, the recovery is reduced. ^[13] This is likely due to the self‐immolative property of the carbamate linker, releasing CO₂, as well as the radical cleavage mechanism of ONB‐derivatives. ^[18] Possibly, the close proximity of the photocleavable group to the purine ring system promotes side‐product formation, which would explain the observations made in this and other studies.

Additional cell experiments to analyze protein production support the partial recovery of the native 5’ cap after irradiation. The amount of protein produced from the photocaged mRNA after irradiation results in 61 % relative to cap 0‐mRNA. This is a ∼1.3‐fold increase in translation when irradiated mRNA was transfected.

Taken together, the enzymatic transfer of ONB‐derivatives to the N² position of the 5’ cap offers an alternative to the synthesis of 5’ cap analogues. In fact, the preparation is faster and potentially more widely applicable, if the synthetic facilities for full cap synthesis are not available. This work also showed that the N² position of m⁷G of the 5’ cap is more applicable to gain optical control over 5’ cap interactions than the N7 position, as it was possible to partially recover the native cap after irradiation. Luciferase assays revealed enhanced protein production after irradiation of the N² modified 5’ cap, supporting the hypothesis that the native cap is recovered after irradiation. However, the amount of recovered 5’ cap and the relatively low translation inhibition elicited by ONB‐derivatives result in a rather minor turn‐on effect after irradiation. As such, this study provides a proof‐of‐concept that the exocyclic N² position is a suitable target for the modification with photolabile groups. The reported results form the basis for future approaches to gain optical control over mRNA via enzymatic modification. One limitation is the promiscuity of GlaTgs V34A. Other methyltransferases or further protein engineering of GlaTgs would enable the transfer of modern photolabile groups like coumarins or BODIPYs yielding efficient inhibition due to steric demand and a more effective turn‐on as other photolysis mechanisms produce less side‐products.

Experimental Section

Enzymatic modification of the m⁷GpppG cap analogue using GlaTgs V34A: For enzymatic modification, m⁷Gp₃G (0.4 mM), AdoMet analogue (1.2 mM), MTAN (4 μM), and LuxS (4 μM) were incubated for 8 h at 23 °C in a final volume of 25 μL with GlaTgs V34A (50–70 μM). The reaction was stopped by heating to 65 °C for 10 min. The denatured enzymes were removed by centrifugation for 10 min at 21,130 g and 4 °C and the reaction mixture was analyzed by analytical HPLC. The modified cap analogues were isolated by analytical HPLC. The collected product fractions were lyophilized. The product was taken up in 10 μL double distilled water (ddH₂O). The resulting solution was used for in vitro transcription.

Irradiation of samples: The respective cap analogues were dissolved in ddH₂O to give a solution with a final concentration of 500 μM. The solution (10 μL) was transferred into a PCR‐tube and irradiated. LEDs (LED Engin) were used to irradiate mRNA and cap analogues. The UV−A‐LED (λ_max=365 nm) was operated with 5 V and 600 mA in a custom‐made LED setup at 23 °C. The samples were irradiated in a PCR‐tube. The mRNA and cap analogues were irradiated at 365 nm (142 mW/cm²) for 30 s, unless otherwise noted. Cap analogues were analyzed by HPLC. Irradiated mRNA was transfected into HeLa cells (Merck). Luminescence was measured by Tecan (Tecan Infinite^© M1000 PRO plate reader).

Absorbance spectra analysis: The analysis of the absorbance properties of the photocaged 5’ cap analogues was performed using a quartz cuvette (Hellma) together with the FP‐8500 fluorescence spectrometer (Jasco). The respective cap analogues were dissolved in water at a final concentration of 100 μM. For the absorbance measurement, 20 μL of the solution were further diluted in water to yield a final volume of 100 μL (20 μM), which was transferred into the cuvette and measured. The values were normalized to the highest measured value of each measurement.

In vitro transcription of mRNA: The DNA template required for the in vitro transcription was synthesized by PCR, in which the DNA sequence coding for Gaussia luciferase (GLuc), were amplified from plasmids containing the respective sequence. After purification (NucleoSpin Gel and PCR Clean up, Macherey‐Nagel), the resulting linear dsDNA was used as template (200 ng). The concentration was measured at 260 nm with the Tecan Infinite^© M1000 PRO. The in vitro transcription was performed with T7 polymerase (Thermo Scientific) in transcription buffer (40 mM Tris/HCl, 25 mM NaCl, 8 mM MgCl₂, 2 mM spermidine (HCl)₃) by adding a A/C/UTP (0.5 mM) mix, GTP (0.25 mM), the respective cap analogue (1 mM), T7 RNA polymerase (50 U) (Thermo Scientific) and pyrophosphatase (0.1 U) (Thermo Scientific) for 4 h at 37 °C. After the reaction, the DNA template was digested in presence of 2 U DNase I for 1 h at 37 °C and then mRNAs were purified using the RNA Clean & Concentrator™‐5 Kit (Zymo Research). To digest non‐capped RNAs, 10 U of the RNA 5’‐polyphosphatase (Epicentre) as well as the supplied reaction buffer were added to purified mRNAs. After an incubation period of 30 min at 37 °C, 0.5 U of the 5’‐3’ exoribonuclease XRN‐1 (NEB) and MgCl₂ (5 mM) were added. The reaction mixture was incubated for 60 min at 37 °C. Subsequently, capped mRNAs were purified using the RNA Clean & Concentrator^TM −5 Kit (Zymo Research).

In‐cell luminescence assay: One day prior to transfection, HeLa cells were seeded in a 96‐well plate (30,000/well) and cultured in minimal essential medium (MEM) with antibiotics. The cells were transfected with mRNA (100 ng) in Opti‐MEM (10 μL) using Lipofectamine™ MessengerMAX™ Transfection Reagent (0.3 μL) in Opti‐MEM (9.6 μL). The cells were incubated with the mRNA/Lipofectamine™ MessengerMAX™ mixture for 6 h at 37 °C in a total volume of 100 μL. Subsequently, the cells were incubated overnight at 37 °C in media. At 24 h post transfection the supernatant was collected. To perform the luminescence measurement the Gaussia‐Juice Luciferase Assay‐Kit (PJK GmbH) was used. The supernatant of the previously prepared samples was transferred to a 96‐well plate (5 μL supernatant per well). Afterwards, 50 μL of a reaction mixture (Reconstruction buffer and Coelenterazine) were added to the wells and the luminescence activity was measured using a Tecan Infinite^© M1000 PRO plate reader. The activity in relative light units (RLU) was determined with an integration time of 3 s. Differently capped mRNAs were used. Ap₃G‐capped mRNA represents cap‐independent translation and was subtracted as background from the other samples. All values were normalized to m⁷Gp₃G capped mRNA.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Klöcker N., Anhäuser L., Rentmeister A., ChemBioChem 2023, 24, e202200522.

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.