Dual-Bioorthogonal Catalysis by a Palladium Peptide Complex

Artificial metalloenzymes (ArMs) enrich bioorthogonal chemistry with new-to-nature reactions while limiting metal deactivation and toxicity. This enables biomedical applications such as activating therapeutics in situ. However, while combination therapies are becoming widespread anticancer treatments, dual catalysis by ArMs has not yet been shown. We present a heptapeptidic ArM with a novel peptide ligand carrying a methyl salicylate palladium complex. We observed that the peptide scaffold reduces metal toxicity while protecting the metal from deactivation by cellular components. Importantly, the peptide also improves catalysis, suggesting involvement in the catalytic reaction mechanism. Our work shows how a palladium-peptide homogeneous catalyst can simultaneously mediate two types of chemistry to synthesize anticancer drugs in human cells. Methyl salicylate palladium LLEYLKR peptide (2-Pd) succeeded to simultaneously produce paclitaxel by depropargylation, and linifanib by Suzuki–Miyaura cross-coupling in cell culture, thereby achieving combination therapy on non-small-cell lung cancer (NSCLC) A549 cells.


■ INTRODUCTION
Nature has evolved to use relatively few metals to conduct biological reactions in living systems (Fe, Zn, Ni, Cu, Mg, and Mn). However, the range of possible chemical reactions could be greatly increased if abiotic transition metals could be used. 1 Since Bertozzi demonstrated that artificial chemical reactions can safely take place in living systems, bioorthogonal chemistry has been used to activate sensors and drugs, 2−5 repair tissues, 6 label biomolecules, 7−11 and modulate biological functions. 12−14 The use of transition-metal catalysts (TMC) to mediate bioorthogonal local activation of drugs has emerged as a potential type of anticancer treatment, increasing tolerability, and therefore effectiveness, of chemotherapeutics. 15−17 Bioorthogonal catalysis still has key challenges remaining, including metal toxicity and catalytic yield. 18 Furthermore, most bioorthogonal TMC are heterogeneous, relying on polymeric supports (polystyrene resins, 4,15,16 micelles, 19 hydrogels, 20 nanoreactors, 21,22 nanozymes, 23,24 and metal−organic frameworks (MOFs) 25,26 ), which can limit their application for treating solid tumors. Ways to overcome these limitations have employed biological delivery systems, including exosomes 17 and macrophages, 27 to encapsulate metals and avoid toxicity. Additionally, heterogeneous catalysts tend to have lower catalytic yield than homogeneous catalysts. 28 Organometallic complexes of Ru, Au, Pd, or Pt have been explored as homogeneous TMC to mediate bioorthogonal chemistry in cells, including alkyl deprotections, 29 hydroarylations, 30 cross-coupling ligations, 31 isomerisations, 32 and metathesis. 33 However, the standard ligands for complexation (e.g., phosphines 34 or N-heterocyclic carbenes 35 for palladium) are small-sized molecules and the metal can be leached easily by cellular proteins. 36 Such protein−metal interactions not only are a major cause of toxicity 37 but also lead to rapid deactivation of the metals catalytic properties. 34 To solve this problem, polymer-based homogeneous TMC have been applied in vitro; 38−43 however their questionable biocompatibility means that low catalyst concentrations have to be used.
Proteins have been used as biocompatible supports for metals, mimicking the active site of metalloenzymes and forming artificial metalloenzymes (ArMs). 44 However, there are few examples of ArMs as bioorthogonal, homogeneous, TMC in living systems, due to their prohibitive macromolecular size. 45−47 Reducing the protein structure to small peptides could overcome large structure limitations, which include delivery issues, in situ metal−protein assembly, and immunogenicity. 47,48 Metallopeptides are an exciting and highly appealing new type of bioorthogonal TMC, achieving homogeneous catalysis with minimal metal toxicity. 49,50 Here, we have synthesized a novel, bioorthogonal, homogeneous palladium peptide catalyst, consisting of a methyl salicylate tagged hydrophilic peptide (LLEYLKR) complexed to palladium. We then explored its catalytic properties in the context of cultured human non-small-cell lung cancer (NSCLC) A549 cells to demonstrate that simultaneous dual catalysis is possible.
■ RESULTS AND DISCUSSION Pd-Peptide Design and Synthesis. Salicylic acid and catechol are well-known chelating agents for palladium. 51−55 For example, catechol not only coordinates Pd(II) but also reduces and stabilizes Pd(0) species by forming o-quinone complexes. 56 We tested the metal chelating/reducing properties of catechol (L1) and methyl salicylate (L2), as these could be coupled in a peptide scaffold as metal binding sites. After incubation with Pd(OAc) 2 in deuterated DMSO for 1 h at 37°C , the capability of L1 and L2 to coordinate Pd was confirmed by 1 H NMR. L1-Pd showed a clear shift of the aromatic protons from 6.7 and 6.6 ppm to 6.3 and 6.2 ppm. Additionally, phenolic protons (8.8 ppm) disappeared, corroborating the palladium-catecholate complexation ( Figure  1a). 1 H NMR spectrum of L2-Pd also showed a shift of the aromatic protons from 7.7, 7.5, and 7.0 ppm to 7.6, 7.4, and 6.6 ppm, corresponding to the enol form of methyl salicylate after complexation with Pd 57 (Figure 1a; see Supporting Information Figure S1). The poor coordination yield of L2-Pd (10%) can be explained by having only one phenolic group donating electrons for complexation, while in the case of the catechol two phenolic electrons complete the coordination. Complete complexation was observed for the peptide ligands with Pd(II) (1 mol of Pd per mol of peptide; see ICP-OES, Figure 1f), possibly because of electron-donor groups from amino acid residues (see Supporting Information, Figures S7 and S8). To confirm the redox reaction, the UV−vis spectrum of L1 and L2 complexed with Pd was measured, displaying an absorbance increment at 300 nm compared to the source Pd(OAc) 2 ( Figure 1b). Based on these results, we decided to explore both ligands as a binding site for palladium on a peptide.
The peptide LLEYLKR was rationally selected from a pool of ribosomal peptides for being soluble in water (cLogP = −4.37) and for presenting hydrophobic (Leu, Tyr), acid (Glu), and basic (Lys, Arg) amino acid residues which have been reported as participating in catalytic mechanisms. 58 The LLEYLKR peptide was synthesized on a Wang resin using standard Fmoc SPPS with HBTU/DIPEA as the coupling combination. 4-{(4-Hydroxy-3-methoxycarbonyl)phenyl]-amino}-4-oxobutanoic acid (4) and 3,4-dihydroxyhydrocinnamic acid were coupled to the amino terminus of the peptide using HBTU/DIPEA.
The peptides were cleaved from the resin by treatment with TFA (5% DCM) and incubated in the presence of Pd(OAc) 2 to form the metallopeptides 1-Pd and 2-Pd ( Figure 1c). Both metallopeptides were purified by C18 SPE cartridges and characterized by mass spectrometry and ICP-OES (Figure 1d− f). Pd-peptides 1-Pd and 2-Pd showed relatively similar palladium mol % content per metallopeptide (52% and 51%, respectively); therefore full complexation was achieved (see full characterization in Supporting Information, Figures S6−S9 and Table S1). Importantly, mass spectra showed the remaining peak of the peptides 1 and 2, which dissociate under electrospray ionization analysis conditions (see Figure  1d (Figure 2a,b). These results agree with previously reported catalytic studies on peptides, suggesting that amino acid residues must play a role in the mechanism of catalysis. 58,60 But also, it is possible that higher concentration of hydrophobic substrate (e.g., ProRes) in hydrophobic pockets that might be formed by the peptide could accelerate the reaction rate. 40 Importantly, Suzuki−Miyaura crosscoupling reaction showed lower catalytic efficiency than the O-depropargylation catalysis. This observation can be explained by an in situ, yet incomplete, reduction of Pd(II) to Pd(0), which is the active catalyst for Suzuki−Miyaura crosscoupling reaction. 61 To further study the catalysis kinetics, Pd-peptides were then incubated with the nonfluorescent compound ProRes at different concentrations (40, 20, 10 μM), and fluorescence signal was monitored every 15 min over an 18 h period. As shown in the Figure 2c, the rate of product formation follows an exponential curve. Kinetic parameters were determined by plotting the Napierian logarithm of substrate ProRes concentrations versus time (see Supporting Information, Figure S10). Both metallopeptides 1-Pd and 2-Pd displayed pseudo-first-order kinetics (1-Pd, K = 0.1147 ± 0.05 and 2-Pd, K = 0.2175 ± 0.03 h −1 ) and a half-life of 6.04 and 3.19 h, respectively. Therefore, we decided to do further biological studies with the metallopeptide 2-Pd, having a depropargylation rate similar to previously reported bioorthogonal metal catalysts. 4,16,17,20 Pd(II) complexes are rapidly deactivated by proteins; 38 therefore the catalysis of ProRes by Pd-peptide (2-Pd) was tested in the presence of serum (see Supporting Information, Figure S11a). 2-Pd remained functional, while the free Pd salt lost its catalytic activity in serum, confirming the protective role of the peptide scaffold.
Cell Assays: Biocompatibility and in Situ Drug Synthesis. Paclitaxel (PTX) is an extremely potent microtubule inhibitor recommended for the treatment of the most common cancers, including breast, lung, and ovarian cancer. 62 Despite all the severe side effects, myelosuppression, peripheral neuropathy, and cardiac toxicity, paclitaxel is currently enrolled in more than 1000 clinical trials. There is a huge unmet clinical need for a paclitaxel prodrug that could be applied globally, but only activated locally, to avoid these terrible off-target effects. A propargylated paclitaxel prodrug (ProPTX) stable in cell culture and uncaged by heterogeneous palladium catalysts has been reported. 20 To further challenge the metallopeptide 2-Pd, it was essential to determine its capability to catalyze the activation of paclitaxel by ProPTX depropargylation. Under physiological conditions (37°C, PBS, pH 7.4) and in the presence of the metallopeptide 2-Pd, ProPTX was converted into the cytotoxic agent PTX and detected by LCMS (see Supporting Information, Figure S12).
Motivated by these results, we sought to prove the anticancer effect of both drugs (PTX and LNF). Cells were treated with PTX and LNF at a range of concentrations (up to 300 μM), and cell viability measurements were carried out after 5 days of treatment. As expected, PTX induced a very potent cytotoxic effect (EC 50 = 5.9 nM, Figure 3a) and LNF showed much lower activity (EC 50 = 3.7 μM, Figure 3b). We then aimed to confirm the innocuous effect of the individual components involved in the catalysis. First, the dose−response curves for the ProPTX prodrug and the two building blocks (A and B) were represented and their EC 50 values were calculated (EC 50 of ProPTX, 1.2 μM; A, 62.1 μM; B, 129.2 μM). As expected, ProPTX displayed >200-fold lower activity than PTX; 20 however the two building blocks (A and B) showed only >15-fold shift in the LNF apoptotic activity. The small gap between LNF and the two building blocks is mainly due to the low potency of LNF as a cytotoxic agent.
Next, a cell-based assay was performed to determine whether PTX and LNF show a synergic effect in non-small-cell lung cancer (NSCLC). A clinical study of paclitaxel in combination therapy with linifanib showed a reduced risk of progression or death in patients with NSCLC. 64 Cancer cells were treated with LNF at different concentrations (1 nM to 30 μM) in combination with a range of nontoxic concentrations of PTX (0−3 nM). These results demonstrated the synergic effect between LNF and PTX, flattening the dose−response curves of LNF and showing 80% cell death at only 0.3 nM PTX (Figure 4a). In order to confirm that dual-synthesis of drugs can be performed and building blocks A + B do not help catalytic depropargylation, e.g., by forming hydrophobic interiors, the deprotection of the sensor (ProRes) was tested with 2-Pd (Pd concentration of 6 μM) in the presence of A + B (2.5 or 25 μM, PBS, pH 7.4). The catalytic efficiency decreases as more A + B was added (see Supporting Information, Figure S11b), corroborating that any improvement in the therapeutic effect by dual-synthesis of drugs must be synergistic.
In parallel, to confirm the lack of toxicity of the catalyst, we performed a cell viability assay to evaluate the toxicity of the  Figure S14). When combining the catalysts and the precursors, cell viability decreased to 67% with prodrug ProPTX only and to 61% when incubated with building blocks A and B. In contrast, simultaneous treatment (incubation with ProPTX, A, and B) caused cell viability to decrease to 28% after 5 days ( Figure  5b). These results confirm the synergic effect when combining the deprotection and cross-coupling chemistry. Pd(II) deactivation by serum is a well reported issue and we can see this phenomenon occurring readily with Pd(OAc) 2 (see Supporting Information Figure S14). Importantly, the catalyst 2-Pd remains active in the presence of serum.
Finally, to validate that the combined treatment of Pdpeptide and drug precursors results in the same antiproliferative mode of action than the parent drugs PTX and LNF, we studied microtubule and nucleus stabilization by immunofluorescence. 20 Cells were fixed 24 h after treatment, incubated with cell nuclei DAPI stain and anti-α-tubulin IgG, and imaged by confocal microscopy. As shown in Figure 5c,d, negative controls did not induce changes in cell morphology (see controls with individual components in Supporting Information, Figures S15 and S16). In contrast, treatment of A549 cells with PTX + LNF led to round-shape morphology with nuclear fragmentation and microtubule condensation (Figure 5e). Importantly, equivalent morphological changes were observed in cells treated with the Pd-peptide and drug precursors combination, evidence that the anticancer effect mediated by the combination treatment is the result of in situ drug generation (Figure 5f).

■ CONCLUSIONS
A methyl salicylate peptide LLEYLKR (metallopeptide 2-Pd) efficiently forms a palladium complex and displays bioorthogonal catalytic properties in the presence of cultured lung cancer cells. Metallopeptide 2-Pd did not show any cytotoxic effect on its own, confirming the crucial role of the peptide to limit metal toxicity. Additionally, the peptide improved the catalytic efficiency of palladium, demonstrating its contribution to the mechanism of catalysis. The versatility of the catalyst in biological environments was exemplified by mediating two types of chemistry in the presence of cultured human cells: a propargyl deprotection and a Suzuki−Miyaura cross-coupling reaction. These can be executed in parallel, leading to a synergic effect of the two anticancer drugs by the simultaneous catalytic synthesis of paclitaxel and linifanib in human lung cancer cells. This is the first demonstration of multiple reactions being catalyzed in parallel by a homogeneous catalyst for drug synthesis. This opens the possibility of more advanced drug combination therapies with increased efficacy and reduced side effects and improved cancer targeting by the catalyst. Novel bioorthogonal homogeneous catalysts, as presented here, further facilitate the possibility of targeted catalysis by direct coupling to delivery vehicles such as antibodies, to overcome the current challenge of delivering the catalyst inside the human body to the desired location of action.
■ EXPERIMENTAL SECTION General. Chemicals and solvents were purchased from Sigma-Aldrich, abcr Germany, Axon Medchem, ChemPUR. FmocArg(Pbf)-OH, FmocLys(Boc)OH, FmocLeuOH, FmocTyr(tBu)OH, FmocGlu(tBu)OH were purchased from GL Biochem. All commercial amino acids are optically pure L-enantiomers. NMR spectra were recorded at ambient temperature on a 500 MHz Bruker Avance III spectrometer. Chemical shifts are reported in parts per million (ppm) relative to the solvent peak. Rf values were determined on Merck TLC silica gel 60 F254 plates under a 254 nm UV source. Purifications were carried out by Biotage Selekt flash column chromatography system or via semipreparative TLC chromatography on Merck TLC silica gel 60 F254 plates. All compounds are >95% pure as measured by either HPLC and NMR or HPLC. HPLC was performed on a Shimadzu LC-20AD system with a ReproSil-XR 120 C18, length 150 mm, i.d  DMF was added to the resin mixture, which was agitated for 2 h. The resin was then filtered and washed with DMF (×3), DCM (×3), MeOH (×3), Et 2 O (×2) and dried under vacuum for 30 min and the level of attachment estimated using the quantitative Fmoc test. 67 Fmoc removal was performed using 20% piperidine in DMF for 20 min (×2). The resin was then filtered and washed with DMF (×3), DCM (×3), MeOH (×3). Resin was swollen in 5 mL of DCM. N-Fmoc-amino acid (5 equiv) and HBTU (4.9 equiv) were dissolved in DMF (0.1 M). DIPEA (10 equiv) was added, and the resulting mixture was added to the resin. The resin was agitated for 40 min. The resin was washed with DMF (×3), DCM (×3), MeOH (×3), Et 2 O (×2). The completion of each coupling was verified using the qualitative ninhydrin test. 68 Cleavage
Biocompatibility Assays. Biocompatibility of metallopeptides was compared by performing dose−response studies in A549 cells. Cells were seeded in a 96-well plate format (at 1500 cells/well) and incubated for 48 h before treatment. Each well was then replaced with fresh medium (supplemented with 10% FBS) containing metallopeptides or Pd(OAc) 2 (100, 200, 400 μM) and incubated for 5 d. Untreated cells were incubated with DMSO (0.1% v/v). Experiments were performed in triplicate. PrestoBlue cell viability reagent (10% v/ v) was added to each well and the plate incubated for 60 min. Fluorescence emission was detected using a FLUOstar Omega multimode reader (excitation filter at 540 nm and emissions filter at 590 nm). All conditions were normalized to the untreated cells (100%).
Dose−Response Curves of Active and Inactive Agents. The antiproliferative activities of PTX/ProPTX and LNF/A/B were compared by performing dose−response studies against the A549 cells. Cells were seeded in a 96-well plate format (at 1500 cells/well) and incubated for 48 h before treatment. Each well was then replaced with fresh medium (supplemented with 10% FBS) containing PTX/ ProPTX (0.03 nM to 10 μM) or LNF/A/B (3 nM to 300 μM). Untreated cells were incubated with DMSO (0.1% v/v). After 5 d of incubation, cell viability was determined as described above. All conditions were normalized to the untreated cells (100%) and curves fitted using GraphPad Prism using a sigmoidal variable slope curve. Experiments were performed in triplicate.
Combination Therapy Assay. The antiproliferative activities of PTX and LNF combination treatment was done by performing dose− response studies against the A549 cells. Cells were seeded in a 96-well plate format (at 1500 cells/well) and incubated for 48 h before treatment. Each well was then replaced with fresh medium (supplemented with 10% FBS) containing PTX (0.03−3 nM) or/ and LNF (0.001−30 μM). Untreated cells were incubated with DMSO (0.1% v/v). After 5 d of incubation, cell viability was determined as described above. All conditions were normalized to the untreated cells (100%) and curves fitted using GraphPad Prism using a sigmoidal variable slope curve. Experiments were performed in triplicate.
Immunofluorescence Assay. A549 cells were seeded on 18 mm poly(L-lysine)-precoated coverslips in 12-well plates (50 000 cells/ well). Cells were incubated 24 h before treatment, and each well was replaced with fresh medium (supplemented with 10% FBS) containing: control, ProPTX Cells were permeabilized for 15 min in PBS, Tween (0.3% v/v) and washed 3 times with PBS. Coverslips were then covered with a blocking solution (PBS, 5% FBS, 0.3% Triton X-100) for 60 min. Cells were washed with PBS 3 times and incubated in an antibody dilution buffer (PBS, 1% BSA, 0.3% Triton X-100) containing anti-αtubulin mAb Alexa Fluor 488 (Santa Cruz) at a dilution of 1:200, overnight at 4°C. Coverslips were washed 3 times with PBS and mounted on Superfrost microscope slides (Thermo Fisher) with ProLong gold mounting medium with DAPI (Thermo Fisher). Cells were imaged using a scanning confocal inverted microscope Nikon scanning confocal A1Rsi+ with a 60× oil immersion objective. The images were acquired using the NIS-Elements program in a sequential mode and analyzed with ImageJ software to obtain maximal projections.

■ ACKNOWLEDGMENTS
We thank the Advanced Medical BioImaging Core Facility of the Charité-Universitätsmedizin Berlin (AMBIO) for support in acquisition of the imaging data.
■ ABBREVIATIONS USED