Absolute structure of the chiral pyrrolidine derivative (2S)-methyl (Z)-5-(2-tert-butoxy-1-cyano-2-oxoethylidene)pyrrolidine-2-carboxylate, a compound with low resonant scattering

The light-atom compound (2S)-methyl (Z)-5-(2-tert-butoxy-1-cyano-2-oxoethylidene)pyrrolidine-2-carboxylate is an enantiopure coordination partner for cations. Despite its only minor resonant scattering, the absolute structure was determined by a combination of diffraction, CD spectroscopy and theoretical calculations.


Introduction
Pyrrolidine derivatives have found applications as potential ligands, as organic intermediates and in medicinal chemistry. They can inhibit the activity of over-expressed protein tyrosine phosphatases (PTPs) of cancer cells and may be employed as anticancer drugs (IC 50 value is 3.65 AE 0.08 mM) (Chen et al., 2017). By forming imine or enamine intermediates with aldehydes and ketones, chiral monopyrrolidine derivatives have been widely used in asymmetric catalysis, and alkylation and acylation reactions of aldehydes and ketones have been achieved (Jensen et al., 2012). We report here the absolute configuration of the chiral pyrrolidine derivative (2S)-methyl (Z)-5-(2-tert-butoxy-1-cyano-2-oxoethylidene)pyrrolidine-2-carboxylate, (1) (Scheme 1). The compound has been synthesized and spectroscopically characterized by Pfaltz and co-workers (Pfaltz et al., 1977;Fritschi et al., 1988;Pfaltz, 1993); retention of the configuration at C1 may be assumed. No studies in medicinal chemistry have been conducted on (1), but a closely related compound was investigated, i.e. methyl 5-[1-cyano-2-oxo-2-(2,3,4-trimethoxyphenyl)ethylidene]prolinate was screened by the National Cancer Institution, USA, against ISSN 2053ISSN -2296 60 human tumour cell lines and showed moderate cell-growth inhibition at 10 mM concentration for renal cancer and leukemia (Ghinet et al., 2012). To the best of our knowledge, the structure of (1) has never been investigated and its absolute configuration has not been confirmed. Our assignment relies on a combination of diffraction experiments, experimental circular dichroism (CD) spectroscopy and theoretical calculations of these spectra. We will show that diffraction results, albeit with only a modest contribution of resonant scattering, and CD spectroscopy agree in their assignment of the absolute structure, whereas a diffraction experiment without relevant anomalous dispersion remains inconclusive.

Synthesis and crystallization
All reagents were commercially available and were used without further purification. The powder diffraction experiment was recorded at the Institute of Inorganic Chemistry, RWTH Aachen University, using a Stoe imaging-plate detector (IP-PSD). The diffractogram was recorded on a flat sample at ambient temperature in transmission mode using Cu K 1 radiation. The title compound was synthesized following the procedure of Pfaltz (Pfaltz et al., 1977;Fritschi et al., 1988;Pfaltz, 1993). The reaction combines S-configured pyroglutamic acid methyl ester and tert-butyl 2-cyanoacetate; retention of the configuration at the chiral centre (*) was expected [see Scheme 1 for a summary of the synthesis for (1) according to Pfaltz et al. (1977)] and is confirmed by the results reported in this work.
Crystals were grown by slow partial evaporation of a methanol solvent at ambient temperature over a period of one week. CHN  Wang and Englert Absolute structure of the chiral pyrrolidine derivative 1449 Table 1 Experimental details.
For both determinations: C 13 H 18 N 2 O 4 , M r = 266.29, orthorhombic, P2 1 2 1 2 1 , Z = 4. Experiments were carried out at 100 K using a D8 goniometer with an APEX CCD area detector. Absorption was corrected for by multi-scan methods (SADABS; Bruker, 2008). H atoms were treated by a mixture of independent and constrained refinement.
Crystal data a, b, c (Å ) 7.347 (4) Computer programs: SMART (Bruker, 2001), SAINT-Plus (Bruker, 2009), SHELXS2013 (Sheldrick, 2008), SHELXL2017 (Sheldrick, 2015) and PLATON (Spek, 2009). (1) is obtained as an essentially phasepure product; the shift of the calculated lines of two larger angles can be attributed to the different data-collection temperatures for the single-crystal and powder analyses. The IR spectrum shows an absorption associated with the triple bond in the nitrile group at (C N) = 2205 cm À1 , in good agreement with the reported frequency of 2207 cm À1 , and the 1 H NMR spectrum matches that available in the literature (Fritschi et al., 1988).

Refinement
Crystal data, data collection parameters and refinement results for both single-crystal X-ray diffraction experiments with Mo K (1mo) and Cu K (1cu) radiation are summarized in Table 1. H atoms attached to C atoms were introduced in calculated positions and treated as riding, with U iso (H) = 1.5U eq (C) for CH 3 groups and 1.2U eq (C) otherwise. For the H atom attached to an N atom, the positional coordinates and an isotropic displacement parameter were refined freely. For the diffraction experiment (1mo), resonant scattering is insignificant; no information can de deduced from the refined enantiopol parameter and its very high standard uncertainty.
For a better comparison with the results of (1cu) on the same single crystal, the same absolute structure model was chosen in both cases.

CD spectroscopy
The experimental electronic circular dichroism (CD) spectrum of (1) was recorded in methanol on a Chirascan circular dichroism chiroptical spectrometer at the Institutes of Biomedical Sciences of Shanxi University; it shows a positive Cotton effect at 278.40 nm and a negative Cotton effect at 245.60 nm.

Molecular structure
The chiral compound (1) was obtained as an essentially monophasic crystalline product. In view of its elemental composition, the determination of the absolute structure was expected to be challenging. With respect to resonant scattering, we calculated values of 6 and 33 for Friedif (Flack & Shmeuli, 2007) for diffraction experiments with Mo and Cu K radiation, respectively. Even the higher second value is dangerously low if the diffraction experiments are hampered by additional complications, such as disorder or twinning. An initial data collection with our standard set-up (1mo) was performed to ensure sufficient quality for the selected crystal and to confirm the chiral space group: even for high enantiomeric excesses, a small amount of racemic solid might precipitate (Bö hme & Fels, 2013).
The second data set collected with Cu K radiation resulted in slightly smaller standard uncertainties; all numerical values reported below therefore refer to (1cu) (see x3.3). As expected, the enantiopure compound (1) crystallized in a chiral space group. The asymmetric unit consists of a single molecule in the space group P2 1 2 1 2 1 ; Fig. 2 shows a displacement ellipsoid plot.
Atoms N1, N2, O3, O4, C3, C4 and C7-C10 define an almost planar core of the molecule shown in Fig. 3. The maximum deviation from that least-squares plane is 0.045 (2) Å for atom C3. Within the core plane, the pyrrolidine N-H group acts as a hydrogen-bond donor towards carbonyl atom O3; the hydrogen-bond geometry is summarized in Table 2.
The pyrrolidine ring is nonplanar and its C2 atom is significantly displaced from the above-defined plane by 0.364 (2) Å .
In contrast to the carboxylate group (C9/O3/O4), the C5/ O1/O2 group is not coplanar with the core of the molecule but The asymmetric unit of (1) based on data set (1cu), with displacement ellipsoids enclosing 50% of electron density.

Intermolecular contacts
The H atom of the pyrrolidine N-H group represents the only potential donor for classical hydrogen bonds. In addition to the intramolecular N-HÁ Á ÁO contact described above, it is involved in an intermolecular N-HÁ Á ÁN hydrogen bond to the nitrile group of a neighbouring molecule. The resulting chain runs along [100] (Fig. 4). The closest contacts perpendicular to this chain are due to nonclassical C-HÁ Á ÁO interactions. Numerical values and symmetry operators for the short contacts have been compiled in Table 2. 3.3. Absolute structure 3.3.1. Resonant scattering. Our first intensity data collection, i.e. the (1mo) data, had provided a consistent structure model without disorder and confirmed the quality of the chosen sample. As expected, however, the commonly applied methods for assigning the absolute structure gave inconclusive results for (1mo) with its negligible resonant scattering. The Flack (1983Flack ( , 2003, Parsons (Parsons et al., 2013) and Hooft (Hooft et al., 2010) parameters refined to values of ca 1, with standard uncertainties equally large; no conclusions could be drawn from these numbers. Therefore, a second diffraction experiment with Cu K radiation, i.e. the (1cu) data, was performed on the same single crystal. Fractional coordinates and derived geometry parameters agreed with the results of (1mo) within error, but resonant scattering was more pronounced and led to information about the absolute structure, i.e. the Flack (1983) parameter refined to À0.04 (12) 3.3.2. CD spectra. An independent assessment of the absolute structure of (1) relies on a comparison of the experimentally observed and theoretically calculated electronic circular dichroism (ECD) spectra; they are shown in Fig. 5.
The calculations were based initially on the molecular geometry obtained from (1cu). Ground-state geometry optimization and subsequent frequency calculations were performed via the density functional theory (DFT) method as implemented in GAUSSIAN09 (Frisch et al., 2009) using the B3LYP hybrid functional (Becke, 1993) and the 6-311++G(2d,p) basis set. The excitation energies, oscillator and rotational strengths of the excited singlet states for the optimized geometry were calculated according to the timedependent DFT (TDDFT) method with the same functional and basis set. The effects of the solvent (methanol) were included using the polarizable continuum model ( The planar core of (1). Table 3 Selected torsion angles ( ) for (1cu).

Figure 4
Intra-and intermolecular hydrogen bonds in the crystal of (1). H atoms not involved in hydrogen bonds have been omitted for clarity. (Tomasi et al., 2005) in the integral equation formalism (IEF).
With the PCM, a ground-state energy of À916.95 a.u. for (1) was obtained. 3.3.3. DFT energy levels and Kohn-Sham orbitals. The DFT energy levels show a HOMU-LUMO gap of 5.25 eV. A detailed analysis of the Kohn-Sham orbitals has been graphically summarized in Fig. 6. The two lowest unoccupied orbitals are dominated by a * region in the planar core and in the carboxylate group of the methyl ester (C5/O1/O2). The absolute value of the energy difference between these LUMO and LUMO+1 orbitals is 0.72 eV. Both of them may well act as electron-acceptor orbitals when electrons from the HOMO and HOMO-1 orbitals are excited. The HOMO is dominated by the region of the planar core of (1). The HOMO-1 essentially corresponds to a combination of + n N + n O orbitals; the energy difference of the HOMO and HOMO-1 amounts to 1.28 eV.
3.3.4. Rotational strengths and transition assignments. The contribution of different transition probabilities to the chiroptical properties of (1) were analyzed. The calculated excitation energies and oscillator and rotational strengths (in velocity form), as well as the transition assignments, have been compiled in Table 4. Results for the three excitations of the lowest energy conformer are given; they cover the spectral range 180 < < 350 nm.
Using the excitation energies and rotational strengths calculated by TDDFT, theoretical CD spectra for both stereoisomers of (1) were generated as the sum of Gaussians, centred at the calculated wavelengths calc with integral intensities proportional to the rotational strengths R of the corresponding transitions. The half bandwidths À at the Á" max /e of Gaussians were assumed as À = k calc 3/2 (Brown et al., 1971) with k = 0.00385 to best reproduce the experiment. The experimental spectrum and calculated spectra for both enantiomers have been compiled in Fig. 5. Ideally, experimental CD spectra of opposite enantiomers are mirror images of each other (Flack & Bernardinelli, 2003).
It is obvious that the CD curve calculated for S-configured (1) is in excellent agreement with the observed curve, with only a small blue shift in the calculated maximum. The agreement confirms that our spectroscopic interpretation of the DFT results is correct.
The observed CD curve consists of two absorption bands, i.e. a positive band around 278 nm arising from the first -* transition in which electrons are transferred from the HOMO to the LUMO (77%) and from the HOMO to the LUMO+1 (19%), and a negative band around 243 nm, which can also be ascribed to the second -* transition and a minor contribution of a -* transition. The main contribution to this significant negative -* transition, however, is associated with the transition from HOMO to LUMO+1 (77%) and from HOMO to LUMO (20%). The -* transition can be assigned to electronic excitation from HOMO-1 to LUMO. Thus, the optical properties of chiral compound (1)   Experimental (left) and calculated CD spectra for (1) in methanol. The spectrum in the centre corresponds to the (correct) S enantiomer and that on the right to the alternative R enantiomer.

Conclusion and outlook
The absolute structure of (1) could reliably be assigned as S, despite the limited contribution of resonant scattering; a lowtemperature diffraction experiment with Cu K radiation resulted in consistent values for the commonly applied enantiopol parameters. Their final standard uncertainties are still rather high, but our assignment is in agreement with the expected retention at the stereocentre of the starting material and could be further corroborated by the match between experimentally observed and theoretically calculated CD spectra. The associated Cotton effect was well reproduced by our TDDFT calculations, thus confirming that our methodology was suitable. We hope to use enantiopure (1) in future experiments as a ditopic ligand with the additional possibility to transfer central chirality from the ligand to its coordination complexes (Wang et al., 2015). For both structures, data collection: SMART (Bruker, 2001); cell refinement: SMART (Bruker, 2001); data reduction: (Bruker, 2009); program(s) used to solve structure: SHELXS2013 (Sheldrick, 2008); program(s) used to refine structure: SHELXL2017 (Sheldrick, 2015); software used to prepare material for publication: SHELXL2017 (Sheldrick, 2015) and PLATON (Spek, 2009).

Special details
Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å 2 )
x y z U iso */U eq O1 0.4701 (4) 0.3508 ( (12) Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.  (7) 0.0024 (7)  O2 0.0232 (7) 0.0229 (7) 0.0250 (7