IV) Computer Assisted Drug Design
B) Investigation of Target Compounds in Connection with Drug Design
1. Investigation of the Infrared Spectra of Labile Compounds Trapped in the Matrix at Low Temperatures
(Professor Elfi Kraka, Professor Dieter Cremer, and co-workers)
There are sophisticated matrix isolation techniques, which make it possible to trap labile reaction intermediates at low temperatures and to investigate them by spectroscopic means, especially by infrared spectroscopy. In this way, the spectra of many elusive species have been recorded that were not amenable to any other experimental technique. However, in many cases species trapped in the matrix cannot be identified because the recorded spectra are complicated superpositions of the spectra of several known and unknown compounds. In this situation, an ab initio calculation of the spectrum of the target compound is the best way of identifying trapped compounds.
In cooperation with several other research groups, we investigated labile species such as carbenes, strained rings, biradicals, and intermediates of oxidation processes. We calculated the infrared spectra of labile compounds and analyzed the calculated spectra so that they could be used for the identification of matrix-isolated compounds. The spectra (frequencies and intensities) were calculated with Coupled Cluster ab initio or DFT methods. Additional information was obtained by determining isotope shifts and polarization angles for spectra, which have been determined using linearly polarized light. Analysis of the infrared spectra is performed according to a new method developed members of the PCTC. This method leads to an accurate description of each vibrational mode and the determination of intrinsic frequencies and force constants that can be used for a description of electronic structure and chemical bonding.
Highlights of this work were the first identification of
a) m-dehydrobenzene:
- 1511,3-Didehydrobenzene (meta-benzyne), R. Marquardt, W. Sander, and E. Kraka, Angew. Chem. 108, 825 (1996).
- 178A CCSD(T) and DFT Investigation of m-Benzyne and 4-Hydroxy-m-benzyne, E. Kraka, D. Cremer, G. Bucher, H. Wandel, and W. Sander, Chem. Phys. Lett., 268, 313 (1997).
- 182Photochemistry of p-Benzoquinone Diazide Carboxylic Acids: Formation of 2,4-Didehydrophenols, W. Sander, G. Bucher, H. Wandel, E. Kraka, D. Cremer, and W.S. Sheldrick, J. Am. Chem. Soc., 119, 10660-10672 (1997).
b) p-dehydrobenzene
- 192p-Benzyne: R. Marquardt, A. Balster, W. Sander, E. Kraka, D. Cremer, J.G. Radziszewski, Angew. Chem., 110, 1001 (1998).
c) a carbonyl oxide that can be investigated in solution by NMR spectroscopy:
- 168Dimesitylketone O-Oxide: Verification of an Unusually Stable Carbonyl Oxide by NMR Chemical Shift Calculations, E. Kraka, C. P. Sosa, and D. Cremer, Chem. Phys. Lett. 260, 43 (1996).
d) an anti-Bredt compound containing a 1,3-bridged cyclopropene:
- 1604,4-Dimethylbicyclo[3.1.0]hexa-1(6),2-diene - A Highly Strained 1,3-Bridged Cyclopropene, R. Albers, W. Sander, C.-H. Ottosson and D. Cremer, Chem. Eur. J., 2, 967 (1996).
2. First Identification of m-Benzyne - A Joint Venture between Theory and Experiment
(Project leaders: Elfi Kraka and Dieter Cremer)
There is a class of anticancer drugs with unusual structures and extraordinary biological activity. This is the class of the enediyne-cytostatica/antibiotica, which have the capability of cleaving DNA. When properly triggered at body temperature, the enediynes can undergo cyclization to benzenoid biradicals, which then are capable of abstracting H atoms from the sugar phosphate backbone of DNA. This leads to double-stranded DNA cleavage and to death of the tumor cell. A key intermediate in this process is the benzenoid biradical 1, para-benzyne. It is generated by the naturally occurring enediyne-cytostatica/antibiotica and represents a "pair of scissors" that cuts through the DNA. Clearly, the naturally occurring enediyne-cytostatica/antibiotica are all optimized to fulfill a certain biological task, which not necessarily implies their usefulness as antitumor drugs. Therefore, much experimental research work has focused on the isolation of appropriate biradicals that can function as a synthetic "pair of scissors" that is more suitable in antitumor drugs. Possible candidates in this respect are the isomeric benzyne biradicals 2 and 3, namely meta- and ortho-benzyne.
In a fruitful interplay between theory and experiment, m-benzyne 2 was synthesized and investigated for the first time. Professor Sander and his research group (University of Bochum, Germany) managed to generate 2 photochemically by decarboxylation of a carbene derivative at low temperature in a matrix. However, the infrared spectrum of the isolated compound contained just one signal at 547 cm-1, which was difficult to relate to the vibrational modes of a benzene ring. Also, there was considerable doubt that 2 would lead to just one infrared-active signal in the vibrational spectrum.
In this situation we were asked by the experimentalists to calculate a reliable infrared spectrum of molecule 2 and to help with its identification. Due to the biradical character of 2, a reasonable description of the molecule could only be provided by high-level ab initio calculations. In the particular case of 2, we used state-of-the-art Coupled Cluster (CC) theory with all single (S) and double (D) excitations and a perturbative treatment of the triple (T) excitations, namely CCSD(T). With this method, we were able to calculate the infrared spectra of 2 using analytical energy gradients.
The calculated CCSD(T) infrared spectrum of 2 showed two very important things: 1) Biradical 2 indeed possesses a very strong infrared band at 547 cm-1 in agreement with experiment. 2) However, there are seven additional infrared bands of lower intensity, which had been overlooked by the experimentalists due to the low concentration of 2 in the matrix. Repetition of experiments under improved conditions led to the discovery of all bands predicted by theory and, thereby, to the confirmation that biradical 2 had been trapped in the matrix at low temperatures. In the following work, it was possible to completely characterize 2 and to determine its energetic properties. According to our results, m-benzynes such as 2 represents a synthetic "pair of scissors" that does not so rapidly cut through DNA and, therefore, can be handled much better than the more reactive 1.
It is interesting to note that DFT poorly fails to describe the infrared spectrum of m-benzyne. As can be seen in Figure 4, B3LYP totally fails because it predicts a bicyclic structure rather than the correct monocyclic m-benzyne structure. Accordingly, the calculated infrared spectrum has no similarity with the experimental one. Restricted BLYP DFT seems to provide a better description of the infrared spectrum of m-benzyne, however the calculations predict also a closed-shell bicyclus with a very long C1C3 bond rather than a monocyclic biradical. Hence the agreement between experimental and BLYP infrared spectrum is just accidental (for the wrong reason).
Figure 4. Comparison of the experimental infrared spectrum of m-benzyne (d) with various calculated infrared spectra. The best agreement is obtained for the coupled cluster spectrum (c) calculated at the CCSD(T)/6-311G(2d,2p) level of theory.