Structurally Redesigned Bioorthogonal Reagents for Mitochondria-Specific Prodrug Activation.

The development of abiotic chemical reactions that can be performed in an organelle-specific manner can provide new opportunities in drug delivery and cell and chemical biology. However, due to the complexity of the cellular environment, this remains a significant challenge. Here, we introduce structurally redesigned bioorthogonal tetrazine reagents that spontaneously accumulate in mitochondria of live mammalian cells. The attributes leading to their efficient accumulation in the organelle were optimized to include the right combination of lipophilicity and positive delocalized charge. The best performing mitochondriotropic tetrazines enable subcellular chemical release of TCO-caged compounds as we show using fluorogenic substrates and mitochondrial uncoupler niclosamide. Our work demonstrates that a shrewd redesign of common bioorthogonal reagents can lead to their transformation into organelle-specific probes, opening the possibility to activate prodrugs and manipulate biological processes at the subcellular level by using purely chemical tools.

Introduction

Our ability to perform abiotic chemical reactions under stringent biological conditions has advanced significantly within the last decades. The ever growing repertoire of bioconjugation reactions allows for precise modification of biomolecules to enable detailed studies of their structure and function.[1−7] We have entered the era in which organic chemists can perform chemical reactions on complex biological systems and even inside living organisms.[8−10] However, the majority of biochemical processes in cells actually take place within distinct subcellular compartments. Our capability to perform chemical reactions in living cells with such precision is still in its infancy. Only a handful of recent examples indicate that this can be achieved.[11−16]

Within a cellular context, the reagents must surmount several barriers to reach the desired location and ideally accumulate within the specific compartment. This can be especially problematic because many factors influence the intracellular distribution of small molecules inside the cell.[17,18] The development of chemical reactions allowing for manipulation of molecules within a specific cellular organelle thus represents a significant challenge. However, the ability to chemically trigger, for example, the release of therapeutic molecules with organellar precision could provide us with new valuable strategies to achieve better therapeutic outcomes.[19]

Mitochondria are vital cellular organelles providing the energy for cells in the form of ATP. Beyond energy production, mitochondria are involved in numerous other biological processes. These include the generation of reactive oxygen species (ROS), the modulation of oxidation–reduction (redox) status, cell signaling, metabolism, and the initiation of apoptosis.[20] Due to the multifunctional role of mitochondria, the imbalance in any of these functions can result in severe disorders. Indeed, mitochondrial dysfunction has been linked to neurodegenerative diseases, obesity, diabetes, and cancer.[21−25] New approaches to combat these mitochondriopathies are thus highly desirable.

In this work, we show that by embedding specific features into the structure of 1,2,4,5-tetrazines, it is possible to achieve their spontaneous accumulation in mitochondria of live cells where they undergo selective reaction with trans-cyclooctene-modified molecules (Figure ). After optimization using fluorogenic substrate, we demonstrate the utility of the approach by organelle-specific reactivation of the protonophoric activity of a mitochondrial uncoupler, which ultimately leads to cell death. This work thus demonstrates that it is possible to achieve localization of bioorthogonal reagents to a specific intracellular site, and in this way, to activate prodrugs within a subcellular compartment using abiotic chemical transformations.

Figure 1

Overview of the workflow leading to the development of mitochondria-specific bioorthogonal release reaction. Mitochondriotropic tetrazines are selected from a pool of tetrazines based on the right combination of lipophilicity and positive delocalized charge. After addition to cells, the optimized tetrazine reagents react with TCO-caged compounds, enabling their reactivation in a mitochondria-specific manner.

Results and Discussion

Design and Selection of Mitochondriotropic Tetrazines

It is known that certain molecules are able to accumulate in mitochondria. It is the large negative membrane potential (Δψ approximately −180 mV) which is believed to drive the enrichment of lipophilic molecules containing a delocalized positive charge in this organelle.[26,27] In contrast to previous attempts to deliver various cargoes to mitochondria by, for example, conjugation to targeting moieties,[28] we sought to find 1,2,4,5-tetrazine derivatives having the inherent ability to accumulate in this organelle. With this goal in mind, we first synthesized a series of tetrazines bearing various heterocyclic substituents, into which we introduced the desired positive charge by alkylation (1a–1l) (Figure B, Supporting Information).

Figure 2

(A) Schematic presentation of the strategy used to select tetrazines having the ability to accumulate in mitochondria. (B) Chemical structures of 1,2,4,5-tetrazines used in this study. (C) Flow cytometry analysis of U2OS cells incubated with tetrazines 1a–1l (1 μM) for 10 min, labeled with fluorescent TCO-bodipy (1 μM) for 15 min and washed. (D) Confocal microscope images showing cellular localization of tetrazines 1b, 1e, and 1i in live U2OS cells. Mitotracker deep red dye (10 nM) was used to visualize cellular mitochondria. Cells in control experiments were loaded with TCO-bodipy alone (1 μM). Scale bar 10 μm.

We first examined the tetrazines 1a–l for their ability to localize in mitochondria. To visualize the tetrazines inside the cells, we used TCO-containing bodipy conjugate (TCO-bodipy), based on the previously described background free tame bodipy scaffold (Figure A).[29] The principle is that the free dye could be removed from the cells by several washes while the one that reacted with the tetrazines remains intracellular. The cells were incubated with 1 μM tetrazines for 10 min, then reacted with TCO-bodipy and after several washes, the total cellular fluorescence was measured using a flow cytometer. Under the selective pressure of low concentrations and short incubation times, only derivatives 1b, 1e, and 1i containing an isoquinolinium moiety showed sufficient cellular retention (Figures C and S14). Mitochondrial localization of the bodipy fluorescence for these derivatives was further confirmed using confocal microscopy (Figures D, S12, and S13).

To explain and rationalize the obtained data, we first calculated the partition coefficient logP values for the 12 tetrazines. Compound 1e has the highest logP value followed by 1i > 1d > 1c > 1b and then the rest of the tetrazines. Experimentally determined logP values for these tetrazines follow the same order (Table S2). Among these derivatives, compounds 1c and 1d are structurally different containing the anilinium moiety. Compounds 1b, 1e, and 1i, which best accumulate in mitochondria, all contain a delocalized positive charge embedded within the hydrophobic isoquinolinium moiety. However, the presence of quinolinium or isoquinolinium moiety alone is not sufficient for mitochondrial accumulation (see compounds 1a, 1f, 1g, and 1h).

To investigate if the accumulation of 1i is membrane potential driven, we performed the targeting experiments after pretreating the cells with carbonyl cyanide m-chlorophenyl hydrazone (CCCP), a known uncoupler of mitochondrial membrane potential (MMP). Depolarization of MMP by CCCP has led to a significant reduction of the fluorescent signal in the cells. Similarly, we did not observe any fluorescent signal in cells with intact MMP incubated with the noncharged tetrazine analogue 1i* (Figures A, S15, and S16). Collectively, these results imply that the overall lipophilicity and a delocalized positive charge is the correct combination for efficient accumulation of the compounds in the organelle, which is driven by the MMP.

Figure 3

(A) Accumulation of the tetrazines in mitochondria of live cells depends on the presence of the positive charge and is driven by the MMP. Shown are confocal microscope images using tetrazine 1i in U2OS cells as an example. Depolarization of the MMP by CCCP abolishes accumulation of 1i in mitochondria. The noncharged tetrazine 1i* does not accumulate in the organelle. Control cells were treated only with TCO-bodipy. Tetramethylrhodamine ethyl ester (TMRE) is sequestered only by active, polarized mitochondria. (B) Confocal microscope images showing that order of the reagent addition influences the organelle-specificity of the reaction. Cells loaded first with 1 μM 1i for 30 min, washed, and incubated with TCO-Reso (5 μM) for 4 h show a mitochondria specific reaction, while an additional unspecific intracellular signal develops in cells treated with the reagents in the reversed order. (C) Chemical structures of TCO-TPP-Reso and Coum-Tz. (D) Intracellular localization of TCO-Reso and TCO-TPP-Reso after labeling with Coum-Tz (1 μM for 5 min). Shown are confocal microscope images and overlays of line profiles showing the pixel intensities along the indicated thin white lines. (E) Fluorescence signal of released resorufin in cellular mitochondria after the addition of TCO-TPP-Reso (2.5 μM) for 15 min, washed and incubated with 1i (1 μM) for 16 h. Scale bars (right bottom corner) 10 μm.

Mitochondria-Specific Release Reaction in Live Cells

In addition to the increasing popularity of tetrazines as bioconjugation reagents,[30−39] we were especially intrigued by their use as trigger molecules in the so-called click-to-release reaction.[40−43] This type of chemistry is based on the reaction of specific dienophiles with tetrazines, which leads to the release of the dienophile-caged cargo in a sequence of tautomerization/elimination steps. This method has been successfully applied for uncaging nucleic acids and biologically active compounds or therapeutics from antibody–drug conjugates.[44−51] Therefore, we thought that our mitochondriotropic tetrazines would provide us with the unique possibility to release TCO-caged molecules specifically within mitochondria.

The releasing ability of 1,2,4,5-tetrazines depends on several factors such as pH and substitution pattern.[52] Because our cationic tetrazines comprise a structurally unprecedented motif, we first evaluated their releasing efficacy. Toward this goal, we employed our recently developed TCO-caged resorufin dye conjugate (TCO-Reso).[53] Our initial experiments performed in a well plate format revealed that tetrazines bearing a single methyl substituent release almost 50% of the resorufin dye from TCO-Reso within the first hour. In contrast, tetrazines having an aryl substituent at this position displayed a rather slow gradual release of the dye over 12 h (Figure S21B). Correlation between reaction rates and the release efficiency of the tetrazines can be found in the Supporting Information (Figure S11).

Next, we used the TCO-Reso to probe the release chemistry in live cells. U2OS cells were incubated with 1 μM tetrazines 1a–l, washed, and loaded with 5 μM TCO-Reso. After 16 h, the cells were analyzed by flow cytometry. Tetrazines capable of quick mitochondrial accumulation (1b, 1e, and 1i) led to markedly higher release of the resorufin in cells when compared to other derivatives (Figure S21C). To further optimize the release reaction, we experimented with the order at which the compounds were added to the cells. When the cells were loaded first with TCO-Reso followed by incubation with the tetrazines, we observed several differences. First, we detected a stronger fluorescent signal in the cells, suggesting that more resorufin had been released. This could be partially explained by the higher stability of TCO in the cellular environment when compared to the tetrazines (Figures S17–S20). Second, in addition to 1b, 1e, and 1i, we also observed a signal produced by tetrazines 1c and 1d. This could be rationalized by the sufficient lipophilicity (Table S2) and enough time for efficient cell penetration even at the low 1 μM concentration. In contrast, a markedly reduced fluorescent signal was observed in cells treated with the rest of the tetrazines (Figure S21C).

Despite the higher release efficiency observed in cells loaded first with the TCO-Reso construct, closer inspection by confocal microscopy revealed that a significant portion of the resorufin signal was not localized inside the mitochondrial compartment (Figure B). The lack of mitochondrial enrichment signal in the TCO-Reso construct is most likely the reason for the observed inferior specificity.

With the aim to further improve the organelle-specificity of the reaction, we decided to add a mitochondria-localizing triphenylphosphonium (TPP) moiety to the TCO-Reso construct (TCO-TPP-Reso, Supporting Information).[54] As expected, attachment of the TPP moiety to the TCO-caged resorufin has led to significant improvement in mitochondrial localization, as we confirmed after labeling the compound inside cells with coumarin tetrazine probe (Coum-Tz)[55] (Figure C and 3D). This fluorogenic tetrazine derivative reacts with the TCO-tagged compound but does not release the cargo (Figure S7). With TCO-TPP-Reso introduced into cells first, followed by the addition of tetrazine 1i, we observed a clear mitochondrial signal of the released resorufin (Figure E). Time-lapse experiment revealed that in the case of 1b, 1e, and 1i, the dye is released in the cells already after only 2 h of incubation. In contrast, the other tetrazines released much less of the compound and in some cases even did not release the dye at all (Figure S22). This highlights the importance of targeting also the second reagent to the organelle. Moreover, by targeting both reagents to mitochondria, we further lowered the tetrazine concentration to 100 nM and still observed efficient and specific resorufin release inside the cells (Figure A).

Figure 4

(A) Efficient and mitochondria-specific release reaction can be achieved even at low tetrazine concentrations. Scale bar 10 μm. (B) Resorufin release in one-cell mouse embryos. Scale bar 10 μm. (C) Pluripotency assessment using anti-Oct3/4 antibody. Scale bar 50 μm. (D) Time-lapse of embryo development after incubation with 1i or DMSO as a control.

To challenge the generality of the mitochondria-specific chemistry in a more complex biological system, we performed the release reaction in one-cell mouse embryos. The mouse embryos were first incubated with 5 μM TPP-TCO-Reso, washed, and incubated with 1 μM 1i for 2 h. Inspection of the cells under microscope confirmed a successful mitochondria-specific reaction (Figure B). Importantly, 1 μM concentration of 1i did not have any detrimental effects on the developmental potential of the embryos (Figures C and D and S23, Table S3; for toxicity of 1b, 1e, and 1i on different cell lines, see Table S4).

Intramitochondrial Delivery and Activation of Niclosamide

Delivery of certain type of drugs to mitochondria can have a beneficial therapeutic effect, especially in the treatment of cancer.[56,57] To demonstrate the utility of our system for mitochondria-specific activation of a biologically active compound, we prepared TCO-caged, TPP-tagged niclosamide (TCO-TPP-niclosamide, Supporting Information). Niclosamide (trade name Niclotide) is an FDA approved drug used for the treatment of intestinal parasite infections.[58] The compound acts as a mitochondrial uncoupler, and its mode of action involves the inhibition of oxidative phosphorylation and the modulation of ATP synthase activity.[59,60] Niclosamide was the subject of several recent studies and was found to have potent anticancer[61−64] and antiviral activity,[65] including promising activity against the recent SARS-CoV-2.[66,67] The protonophoric activity of niclosamide is linked to the weakly acidic phenolic hydroxyl group.[61] Based on this information, we designed and synthesized TCO-TPP-niclosamide containing the TCO and TPP substituents attached to this hydroxyl group via a self-immolative carbamate linker (Figure A). TCO-TPP-niclosamide showed excellent mitochondrial accumulation, as we confirmed by labeling the compound inside the cells using the Coum-Tz probe (Figure B and S24). HeLa cells pulsed with 5 μM solution of TCO-TPP-niclosamide for 1 h did not show any alteration in the membrane potential (TMRE positive staining), demonstrating that the protonophoric activity of this compound is switched off (Figure C). This activity could be efficiently restored back in the presence of 0.7 μM tetrazine 1i (Figure D). Importantly, cells treated under identical conditions with tetrazine 1i in the absence of TCO-TPP-niclosamide (Figure F) did not show membrane depolarization (positive TMRE signal).

Figure 5

Tetrazine-triggered reactivation of niclosamide uncoupler in mitochondria of live HeLa cells. (A) Chemical structure of TCO-TPP-niclosamide. (B) Cellular localization of TCO-TPP-niclosamide revealed after labeling with Coum-Tz. (C–F) Confocal microscope images showing the status of mitochondrial membrane potential in cells treated with the indicated compounds or their combination. The cells in panels C and D were pulsed with TCO-TPP-niclosamide (5 μM) for 1 h, washed to remove extracellular reagent, and further incubated with cultivation medium as a control (C) or medium containing 1i (0.7 μM) (D). Further controls included untreated cells (E) or cells treated with 1i (0.7 μM) (F). After 48 h, the cells were labeled with TMRE (100 nM, red). The nucleus was stained with Hoechst 33342 (blue). Scale bar is 10 μm. For a similar experiment with the nontargeted TCO-niclosamide, see the Supporting Information.

The determined IC50 for TCO-TPP-niclosamide was 40 ± 6 μM. After its mitochondrial chemical activation, the IC50 dropped to 0.8 ± 0.3 μM, which represents a 50-fold improvement in potency. For comparison, the IC50 of free niclosamide under the same conditions is >100 μM. For comparison, we also performed the release experiment with the nontargeted TCO-niclosamide and nontargeted tetrazines 1i* and 1l. The determined IC50 was in both cases >100 μM (Supporting Information). These data show that delivery of the prodrug to the organelle followed by its chemical activation leads to mitochondrial membrane depolarization and subsequent cell death at much lower efficient concentration.

Conclusion

In this work, we show that structural fine-tuning of 1,2,4,5-tetrazines enables their redesign into mitochondriotropic reagents. The structural features leading to their spontaneous enrichment in the organelle constitute the combination of a positive, delocalized charge and sufficient lipophilicity. These attributes are manifested in the new isoquinolinium mitochondrial localization signal. The optimized mitochondriotropic tetrazines are not toxic at working concentrations as determined on various cell lines and one-cell mouse embryos. We demonstrate the utility of this approach by tetrazine-triggered fluorophore release and by reactivation of mitochondrial uncoupler niclosamide from the respective precursors. By delivery and subsequent mitochondrial release, the potency of the latter compound was improved about 50-fold when compared to the TCO-caged prodrug and over 100-fold with respect to the parent unmodified compound. We believe that the ability to perform abiotic chemical reactions within particular organelles will open new avenues to manipulate and study biological processes at the subcellular level and that this strategy could be useful for the development of advanced organelle-targeted drug delivery systems.[68] Finally, the general concept presented here could be applied to the redesign of other popular bioorthogonal reagents and possibly be extended to other important cellular organelles.