New insight on the structural features of the cytotoxic auristatins MMAE and MMAF revealed by combined NMR spectroscopy and quantum chemical modelling

Antibody-drug conjugates (ADCs) are emerging as a promising class of selective drug delivery systems in the battle against cancer and other diseases. The auristatins monomethyl auristatin E (MMAE) and monomethyl auristatin F (MMAF) appear as the cytotoxic drug in almost half of the state-of-the-art ADCs on the market or in late stage clinical trials. Here, we present the first complete NMR spectroscopic characterisation of these challenging molecules, and investigate their structural properties by a combined NMR and quantum chemical modelling approach. We find that in solution, half of the drug molecules are locked in an inactive conformation, severely decreasing their efficiency, and potentially increasing the risk of side-effects. Furthermore, we identify sites susceptible to future modification, in order to potentially improve the performance of these drugs.


NMR spectroscopic characterisation of MMAE
The sole aromatic residues in 1A (cis-conformer) and 1B (trans-conformer) were selected as suitable starting points for the NMR spectroscopic characterisation of MMAE. In the carbon spectrum, the C-1 (1) and C-1 (1) signals are well-resolved and appear at 144.1 (1A) and 143.9 (1B) ppm. In the HMBC spectrum ( Figure 5), the cross peaks from C-1 (1) to H-3 (1) and H-5 (1) at 7.34 ppm, H-7 (1) at 4.52 ppm and H-8 (1) at 4.24 ppm were visible. There were no HMBC correlations between C-1 (1) and H-2 (1) or H-6 (1). The conventional use of edHSQC ( Figure 3) and COSY (not shown) resulted in the identification of all signals in residue (1). The signals of residue (1) were identified and assigned in a similar fashion. It should be noted that all of the 1 H-and 13 C chemical shifts in residues (1) and (1) are similar. In fact, they differ by less than 0.1 ppm in the 1 H-NMR spectrum and 1.0 ppm in the 13 C-NMR spectrum.
As expected the remaining signals in the 1 H-and 13 C-NMR spectra could be assigned to residues (5) and (5). For example, H-2 (5) at 3.70 ppm and H-2 (4) at 4.80 ppm had HMBC correlations to C-1 (5) at 167.4 ppm thus assuring that all residues had been assigned correctly (the same patterns were identified in (5)). The chemical shifts for the H-2 (5) (3.70 ppm) and H-2 (5) (3.68 ppm) protons were significantly different than those reported for the corresponding protons in dolastatin 10 (2.65 ppm and 2.39 ppm).
While not mentioned previously, the commercial MMAE utilized in this study was supplied as a TFAsalt. The signals from TFA are not listed in the tables but appeared in the carbon spectrum as a quartet at 162.3 ppm (JC,F = 35.8 Hz) and a quartet at 117.9 ppm (JC,F = 291.8 Hz). In addition to these peaks, there was an unidentified signal at 101.4 ppm without edHSQC and HMBC correlations.

NMR spectroscopic characterisation of MMAF
The NMR spectroscopic characterisation of MMAF will not be discussed in detail since the guidelines provided above are directly applicable to MMAF. Instead, the chemical shift values of 2A (cisconformer) and 2B (trans-conformer) will be compared to each other and those observed for MMAE in order to uncover common trends for this class of compounds. For this comparison to be possible, the spectra of MMAF were also measured in deuterated methanol. It should be noted that MMAF was supplied as a TFA-salt. The chemical shifts of TFA are not listed in tables 3 and 4. In the 13 Cspectrum of MMAF, there is two different sets of signals for TFA which might reflect the existence of the two well-known isomers, two deviating salt forms or alternatively free TFA in addition to the salt. These signals appear at 167.7 ppm (q, JC,F = 35 Hz), 158.9 ppm (q, JC,F = 42 Hz), 118.0 ppm (q, JC,F = 292.4 Hz) and 116.0 ppm (q, JC,F = 284.4 Hz). In the spectra of MMAF, there is also an unknown signal at 101.4 ppm without HMBC or HSQC cross peaks.

Molecular coordinates
Below we tabulate the molecular coordinates for the isomers and transition states of the studied species, in standard XYZ format (Ångström units). All geometries computed at the TPSSh-D3(BJ) density functional theory level, with the COSMO solvation model simulating a methanol environment, using the dielectric constant ε=32.6. The stable points on the potential energy surface were optimised with the def2-TZVPP basis set, and the transition states using the def2-SVP basis set.