A Cleavable C2-Symmetric trans-Cyclooctene Enables Fast and Complete Bioorthogonal Disassembly of Molecular Probes

Bioorthogonal chemistry is bridging the divide between static chemical connectivity and the dynamic physiologic regulation of molecular state, enabling in situ transformations that drive multiple technologies. In spite of maturing mechanistic understanding and new bioorthogonal bond-cleavage reactions, the broader goal of molecular ON/OFF control has been limited by the inability of existing systems to achieve both fast (i.e., seconds to minutes, not hours) and complete (i.e., >99%) cleavage. To attain the stringent performance characteristics needed for high fidelity molecular inactivation, we have designed and synthesized a new C2-symmetric trans-cyclooctene linker (C2TCO) that exhibits excellent biological stability and can be rapidly and completely cleaved with functionalized alkyl-, aryl-, and H-tetrazines, irrespective of click orientation. By incorporation of C2TCO into fluorescent molecular probes, we demonstrate highly efficient extracellular and intracellular bioorthogonal disassembly via omnidirectional tetrazine-triggered cleavage.

HRMS analysis was carried out using methanol solutions (concentration: 10 ppm) on an Agilent 6230 LC TOFMS mass spectrometer equipped with an Agilent Dual AJS ESI-Source. The mass spectrometer was connected to a liquid chromatography system of the 1100/1200 series from Agilent Technologies (Palo Alto, CA, USA). The system consisted of a 1200SL binary gradient pump, a degasser, column thermostat, and an HTC PAL autosampler (CTC Analytics AG, Zwingen, Switzerland).

4-Hydroxycyclooct-2-ene-1-one (9)
In a PESCHL UV-Photoreactor (Borosilicate glass; THQ 150 Z1 immersion lamp, 150 W) 1,3-cyclooctadiene 8 (4.3 g, 5 mL, 40.2 mmol) was dissolved in dry DCM (400 mL). Meso-tetraphenylporphyrin 1 (50 mg, 0.08 mmol) was added and the mixture was cooled to -5 °C. Molecular oxygen was bubbled through the solution and the UV-lamp was switched on. The solution was irradiated under constant oxygen flow, while the temperature was maintained between -10 °C and 0 °C. The reaction was monitored via 1 H NMR, drawing samples every hour. After 5 h further meso-tetraphenylporphyrin (50 mg, 0.08 mmol) was added to the brownish solution, as the photosensitizer had been degraded. After irradiating for additional 9 h the oxygen-bubbling was stopped and the solution was transferred to a 500 mL round bottom flask. DBU 2 (630 µL, 4.2 mmol) was added and the mixture was stirred at room temperature for 13 h. The solvent was evaporated and the residual solid was purified by column chromatography (33-50% EtOAc in hexanes, gradient elution) to give 9 (2.01 g, 63%) as a white solid. Analytical data matched that reported in the literature;

Sar-C 2 TCO-Sar (13)
Bis-axial-11 (100 mg, 0.70 mmol) and CDI (342.1 mg, 2.10 mmol) were dissolved in dry THF (1.4 mL) and the solution was stirred for 3 h under an atmosphere of argon. The reaction mixture was diluted with Et 2 O (50 mL) and washed with H 2 O (3 x 50 mL) and brine (1 x 20 mL), dried over Na 2 SO 4 and concentrated. The residual white solid was dissolved in dry MeCN (2.8 mL), methyl iodide (435 µL, 7.00 mmol) was added dropwise, and the mixture was stirred at room temperature for 19 h. The solvent and volatiles were evaporated to obtain the methylated imidazolium carbamate 12 as a yellow solid (343 mg). 12 was immediately dissolved in dry DMF (2.8 mL) and DIPEA (980 µL, 5.6 mmol) followed by sarcosine methylester hydrochloride (390. 8

General procedure for Tz-acids
A well-blended mixture of nitrile #1 (1 eq.) and nitrile #2 or formamidine acetate salt (5 eq.) and Zn(OTf) 2 (5 mol%) was treated dropwise with hydrazine monohydrate (25 eq.) while cooling in an ice bath. The mixture was allowed to warm up to room temperature and stirred for 10 min, after which it was stirred at 60 °C for the specified time. The crude reaction mixture was poured onto ice-water (50 mL). After addition of NaNO 2 (4 eq.) the solution was acidified with aqueous 2N HCl solution. The mixture was extracted with EtOAc, dried over Na 2 SO 4 , filtered and concentrated. The crude product was purified by column chromatography.

2)
Click kinetics Sample preparation rTCO-dPEG 4 -OH (6) 12 and Sar-C 2 TCO-Sar (13) were dissolved in PBS (pH 7.4, 10 mM) to reach an approximate concentration above 2 mM. The exact concentration was determined by absorbance titration with DMT (1) (extinction coefficient 510 M -1 cm -1 at 520 nm), quantifying the decrease in tetrazine absorbance upon reaction with TCO. The initial PBS stock was diluted to prepare solutions for stopped-flow analysis at a final TCO concentration of 2 mM. Stock solutions of tetrazines (1, 3, 4a, 4b, 5, 17, 18, 19, 20) were prepared in DMSO at a concentration of 10 mM. Serial dilution into PBS (pH 7.4, 10 mM) was used to prepare solutions for stopped-flow analysis at a tetrazine concentration of 100 µM.

Stopped-flow spectrophotometry
Stopped-flow measurements were performed using an SX20-LED stopped-flow spectrophotometer (Applied Photophysics) equipped with a 535nm LED (optical pathlength 10mm, full width half-maximum 34nm) to monitor the characteristic tetrazine visible light absorbance (520-540 nm). The reagent syringes were loaded with tetrazine and TCO compound (6 or 13) solutions and the instrument was primed. Subsequent data were collected in triplicate to sextuplicate for each tetrazine. Reactions were conducted at 37 °C and recorded automatically at the time of acquisition.

Data analysis
Data sets were analyzed by fitting an exponential decay using Prism 6 (Graphpad) to calculate the observed pseudo-first order rate constants that were converted into second order rate constants by dividing through the concentration of excess TCO compound (6, 13).

Instrument & solvents
All release experiments were performed on a Nexera X2® UHPLC system (Shimadzu®) with a temperature-controlled autosampler at 25 °C or 37 °C. For buffered LCMS conditions, the aqueous solvent was prepared by addition of 625 µL of 10 M ammonium formate (BioUltra, Sigma-Aldrich) to 2.5 L of HPLC-grade water followed by adjusting the pH to 8.5 by addition of 100 µL of 25% aqueous ammonia (for HPLC, LiChropur, Merck). The pH of this volatile buffer declines over time and was thus freshly prepared each day. HPLC-grade acetonitrile was used without any additives.

Release kinetics with rTCO-AF350 (7)
A 20 mM stock solution of rTCO-Alexa Fluor 350 (rTCO-AF350, 7) in DMSO was diluted with citrate-phosphate buffer (10 mM, pH 7) to a final concentration of 100 µM directly in an LCMS sample vial and mixed. 20 mM tetrazine stock solutions were diluted with citrate-phosphate buffer (10 mM, pH 7) to give 200 µM concentrations. These solutions were added to the TCO sample and mixed (1:1) to initiate the cleavage/release reaction. Reactions were conducted at 25 °C and monitoring was performed by HPLC analysis (fluorescence and MS detection; ammonium formate 2.5 mM, pH 8.4 / acetonitrile gradient).

Release measurements with AF594-C 2 TCO-AF594 (15) and AF350-cTCO-Sar-OEt
Analytical stock solutions and buffers. 10 mM stock solutions of the release probes AF594-C 2 TCO-AF594 (15) and AF350-cTCO-Sar-OEt 12 in DMSO were prepared and further diluted with citrate-phosphate buffer (as indicated below), so that the final concentration of DMSO in buffer was never higher than 2%. Tetrazine stock solutions were prepared in DMSO at concentrations ranging from 10-100 mM and diluted with citrate-phosphate buffer (as indicated below), so that the final concentration of DMSO in buffer was never higher than 5%. Citrate-phosphate buffer (10 mM) was prepared by dilution from standard stock solutions of 0.1 M citric acid and 0.2 M Na 2 HPO 4 with water and the pH verified and adjusted as needed to within ±0.05 pH units by digital pH metering.

Selected chromatograms
All chromatograms were screened for m/z of starting materials, potential intermediates and products. No intermediates (= clicked but not released yet) could be detected, indicating instantaneous release (click = rate-determining step of overall cleavage).

Stability of C 2 TCO
Analytical stock solutions and solvents. The stability of C 2 TCO was assessed in PBS buffer (10 mM), PBS buffer (10 mM) + 1 mM L-glutathione and full cell growth media (DMEM (fluorobrite) + 10% FBS). Therefore, a 200 mM stock solution of Sar-C 2 TCO-Sar (13) in DMSO was prepared and further diluted to a concentration of 0.2 mM with the respective solvent (final concentration of DMSO = 0.1%).
Stability measurements. All solutions of Sar-C 2 TCO-Sar (13) were incubated at 37 °C and aliquots were taken after 6, 12, 24, and 48 h. Triplicate aliquots of each solvent were spiked with excess 88.3 mM stock solution of DMT (1) in DMSO and mixed. Upon reaction with TCO, the resulting tetrazine absorbance was measured on a Thermo Fisher Scientific NanoDrop One C in cuvette mode at 25 °C at 520 nm. This procedure with adding excess tetrazine and measuring the absorbance of the solution upon reaction with TCO was repeated two more times (standard addition). The absorbance was further plotted against the concentration of tetrazine at each addition and the concentration of reacted tetrazine equated the corresponding concentration of TCO present in the solution at each time point. The stability was determined to be >97% for all analyzed samples, except for 48 h incubation in PBS + 1 mM L-glutathione (89±2%).

Stability of HPA (19) and HK (20)
We have not experienced any issues related to the stability of the H-tetrazines 19 and 20 when stored at -20 °C (either as solids or as DMSO stock solutions). When working with PBS solutions of both tetrazines at room temperature, we noticed that HPA (19) was significantly more stable than HK (20). To measure the stability of 19 and 20, both tetrazines were incubated in PBS (5% DMSO) at a concentration of 500 µM. The characteristic absorbance at 520 nm was then monitored by UV/Vis (Shimadzu UV1800 Spectrophotometer), showing that HPA (19) was stable in PBS (>97%) for at least 9 days. HK (20) was confirmed to be less stable, but still showed sufficient stability over several hours (e.g. 87% after 18 h in PBS). HK (20) S16

5) Aryl-Tz problem: Isolation and structure elucidation of non-releasing dead ends
PyrMe (4a) + rTCO-glycine rTCO-glycine 12 was dissolved in DMSO to prepare a 50 mM stock solution. 528 µL (26.4 µmol, 6 mg) of that solution were diluted into 50 mL of CitPhos buffer (10 mM, pH 7), for a final concentration of 0.528 mM rTCO-glycine. The solution was then sparged with argon for 15 minutes to reduce dissolved oxygen and minimize potential aromatization, whereupon 295 µL of a 98.3 mM stock solution of PyrMe tetrazine 4a (1.1 equiv., 5.03 mg) were then added. At these reagent concentrations the click reaction is expected to be done in <10 minutes. The release reaction was allowed to proceed under argon overnight, a timeline after which no further release was observed by serial HPLC analyses. The buffer solution was evaporated to near-dryness and then redissolved in 1 mL of DMSO. The DMSO solution was loaded directly onto a 30 g Biotage Snap C18 Ultra column and the non-releasing isomer isolated with an ammonium formate (2.5 mM, pH 8.4) / acetonitrile gradient. HPLC analysis indicated ≥ 99% purity ( Figure S1). A single fraction containing the purified material was immediately concentrated, stored at -80 °C under argon, and 1 day later dissolved in CD 3 OD for NMR analysis

Antibody Labeling
Cetuximab (2.0 mg/mL, as supplied by the manufacturer) was concentrated by centrifugal filtration (Amicon 100K filter) to a concentration of 4 mg/mL. The concentrated stock solution was then bufferexchanged by 40K Zeba spin column into PBS-Bicarb buffer (phosphate buffered saline plus 100 mM NaHCO 3 ) that had been freshly adjusted to pH 8.4, as the pH of PBS-Bicarb stock solutions rises over time in storage. The concentration of the buffer-exchanged antibody was re-checked by Nanodrop (A280) and found to be 4.15 mg/mL (28.5 µM). In parallel, 6-azidohexanoic acid sulfoNHS ester (Click Chemistry Tools, 1251-5) was dissolved in MQ water to prepare a 500 µM stock solution.

Optimized Labeling Procedure
Labeling reactions were prepared by adding 1.2 -1.6 -2.0 equivalents of the sulfoNHS ester to 100 µL of the antibody solution in PBS-Bicarb. After 20 minutes, the reactions were purified by 40K Zeba spin column (pre-conditioned with PBS). The concentration of the antibody solutions was rechecked following the spin column to verify recovery and account for microscale dilution effects. Aliquots of each N 3 -labeled antibody (12 µL) were then combined with 2 µL of a stock solution of DBCO-C 2 TCO-AF594 (21) (1.7 mM in DMSO) for a final reaction concentration of 242 µM dye (~10X antibody concentration). The reactions were allowed to proceed at 4 °C overnight, then purified by two successive 40K Zeba spin columns (PBS).
Nanodrop analysis of the fluorescently labeled antibodies was used to determine the final dye/label concentrations, per routine, with the following parameters for extinction coefficients and correction factors; cetuximab: ε 280 = 215,000 M -1 cm -1 , AF594: ε 594 = 93,000 M -1 cm -1 , correction factor: CF 280 = 0.705. The experimentally determined DOL was consistent for the three labeling reactions, with ~0.7 dye molecules per equivalent of N 3 -sulfoNHS, resulting in antibodies with a DOL of 0.87, 1.12, and 1.45 dyes/Ab respectively. [Note: the CF280 for DBCO-C 2 TCO-AF594 in PBS was experimentally measured and gives consistent results for all AF594-labeled antibodies (irrespective of conjugation chemistry) in our hands; this differs from the CF280 value of 0.56 reported by ThermoFisher for AlexaFluor 594].
To verify intact click-reactivity of the C 2 TCO probes and Tz/TCO cleavage in situ, these antibodies were combined on a microscale with HPA (19) and allowed to react for ten minutes, followed by a 40K Zeba column to remove free/cleaved dye. Nanodrop analysis revealed excellent cleavage efficiency for all three Ab-AF594 conjugates, with an inverse correlation between DOL and cleavage efficiency, ranging from ≥97% cleavage at a DOL of 0.87 to 92% cleavage at DOL 1.45. In earlier pilot experiments with a higher degree of N 3 labeling, we noted a consistent trend toward lower C 2 TCO reactivity as the number of azides per antibody increased, suggestive of an intramolecular reaction of N 3 with C 2 TCO.
Pilot Experiments: concern for C 2 TCO -N 3 cross-reactivity Initial Labeling reactions were prepared by adding 2 -20 equiv. of the 6-azidohexanoic acid sulfoNHS ester to 100 µL of the antibody solution in PBS-Bicarb. After 20 minutes, the reactions were purified by 40K Zeba spin column (pre-conditioned with PBS), labeled with DBCO-C 2 TCO-AF594 (21, 90 µM in the click reaction, 8-10 molar equiv. relative to the mAb) and then characterized as above. In these initial test experiments the efficiency of cleavage varied strikingly as a function of the number of equivalents of azide used in the initial labeling reaction. This was true irrespective of labeling time/temperature, suggesting that it was unlikely to be due to issues of general stability for C 2 TCO (which is excellent in even more demanding biofluids/temperatures, see SI Section 4), and more likely to be related to an interaction between the azide tags and the C 2 TCO. In order to test that hypothesis, we conducted additional follow-experiments systematically varying both the antibody DOL (by controlling the number of equivalents of 6-azidohexanoic acid sulfoNHS ester in the initial N 3 -labeling reaction) and the concentration of DBCO-C 2 TCO-AF594 (21) in the subsequent click reaction. Plotting the data by observed AF594 DOL (rather than equivalents azide) illustrates the trend: Higher cleavage efficiency was observed: i) at lower DOL; ii) at higher concentrations of 21, consistent with more rapid reaction of the Ab-N 3 moieties with the DBCO in solution. At a concentration of 240 µM DBCO in excess and a rate constant of 0.3 M -1 s -1 , 13 complete reaction (>99%) is expected to require ~18 hours. In both cases, cleavage efficiency is increased by conditions that reduce the likelihood (time/stoichiometry) of an intramolecular reaction between an antibody-linked C 2 TCO and a neighboring azide, i.e. on a nearby lysine, within the ~3 nm molecular reach of the DBCO-C 2 TCO linker.

C 2 TCO -Azide reactivity and kinetics
To confirm cycloaddition of C 2 TCO with azide-functionalities we reacted DBCO-C 2 TCO-AF594 (21, 50 µM in PBS) with 50 mM 2-azidoethanol. While we observed an immediate reaction of DBCO and N 3 (in agreement with the reported second order rate constant of 0.3 M -1 s -1 , 13 which predicts 99% conversion in ~5 min at these concentrations), extended LCMS monitoring demonstrated cycloaddition of N 3 to the C 2 TCO moiety, with a second order rate constant of 0.0011 ± 0.0002 M -1 s -1 : Taken together the results support slow reaction between Ab-linked C 2 TCO and neighboring N 3 -lysine residues due to increased local concentrations. We therefore recommend a switch to other bioconjugation methods when designing probes with a higher DOL.

Cell Culture
A-431 cells were purchased from the American Tissue Culture Collection (ATCC) and passaged in DMEM (10% FBS, 1% penicillin/streptomycin) according to the specifications from ATCC. Cells were first grown in a 150 mm cell culture dish and then seeded on Millicell 8-well EZ slides (Millipore) for imaging. After 24-48 hours confluency was assessed and cells were fixed with 4% paraformaldehyde in PBS (10 min) prior to EGFR imaging.

Cellular Imaging and in-situ cleavage
A-431 cells were stained with 70 nM cetuximab-C 2 TCO-AF594 at 4 °C for three hours, then rinsed with PBS three times. An Olympus BX-63 upright automated epifluorescence microscope was used to acquire fluorescent images. The stained cells were imaged to establish baseline brightness and register coordinates for serial imaging after Tz addition. After collecting the initial image, the slide was removed from the microscope and a solution of 500 µM HPA (19) in PBS was added to the well. The automatic stage was returned to the coordinates of the initial image collection and serial images collected (400 msec exposure) at two-minute intervals. FIJI was used to extract line-intensity profiles for each image for quantitative comparison of signal intensity vs. time ( Figure S1).