Syntheses and crystal structures of a nitro–anthracene–isoxazole and its oxidation product

The title compounds arose as unexpected by-products of an iodination reaction: in each case the fused-ring and isoxazole planes are almost perpendicular to each other.


Chemical context
In the course of our study of aryl-isoxazole amide (AIM) antitumor agents, we have a standard operating procedure to identify by-products of the synthesis , and have used the mechanistic insights gained in order to optimize and improve subsequent syntheses.
During recent structure-activity relationship studies, we encountered complications in constructing sterically hindered examples, which we desired for their calculated pharmacokinetic properties. After obtaining mediocre results with bromine as a leaving group in Suzuki couplings, we pursued a fairly routine alternative of moving to the next halogen down in the periodic table. We have encountered more complications in this study than in the previous twenty papers we have published in this area (e.g. Weaver, Stump et al., 2020 andWeaver et al., 2015), and herein report the crystal structures of two compounds observed.
Using conditions usually reported for iodination, the main product observed for reaction of (II) was the nitro ester (I) rather than the expected iodo product (III), which was obtained in small amounts (Fig. 1). The nitro product so obtained exhibits most of the stereoelectronic properties of previously studied analogues that we have considered to be essential for their biological activity (Han et al., 2009). The nitro group is disordered and found in two distinct conformations in the unit cell. We attribute this to an extreme perieffect, which substantially raises the energy of the co-planar conformer.
In order to improve on the accuracy of the crystal structure of (I) we attempted numerous recrystallizations; however, what was observed was the addition of oxygen to compound (I), which we attribute to cycloaddition of dioxygen to an endo-peroxide (IV) (Klaper et al., 2016), and ring opening with loss of a leaving group to the oxidation product anthraquinone (V). Usually, anthracenes are oxidized in vivo predominantly by cytochrome P450, leading to a potentially toxic arene oxide (Silverman et al., 2014). The rationale for the isoxazole series is that the C-5 isoxazole methyl group represents an opportunity for safer metabolism (Natale et al., 2010). The observation in this manuscript suggests that intramolecular dioxygenation, which would likely be mediated in vivo by mono amine oxidase (MAO), is another plausible route (Silverman, 2002). The observation of a possible endoperoxide pathway in this study suggests that the metabolism of these 10-substituted anthracenyl isoxazole analogues could go through dioxygenation catalysed by COX (cyclooxygenase) and other prostaglandin synthases in vivo (Silverman, 2002).

Structural commentary
The first title compound (I), C 21 H 16 N 2 O 5 , crystallizes in the monoclinic Cc space group with two independent molecules in the asymmetric unit (Fig. 2). The dihedral angle between the anthracene ring mean plane and the isoxazole ring mean plane indicate near orthogonality: 88.67 (16) and 85.64 (16) for molecules A (containing C1) and B (containing C22), respectively. Each independent anthryl ring contains a 10nitro group with the O atoms disordered over two orientations. The isoxazole group and its attached ethyl ester moiety are virtually co-planar, with the twist angles found to be 3.1 (2) between the C15-C17/O1/N1 and O2/C19/O3/C20 planes in molecule A, and 4.2 (2) between the C36-C38/O6/ N3 and O7/C40/O8/C41 planes in molecule B. The ester ethyl group is exo-with respect to the anthryl ring in the solid state but this conformation is not completely retained in solution as the proton NMR indicates significant anisotropy at the methyl group of the ethyl ester ( = 0.41), which indicates at the very least a significant population of the endo-orientation. In addition, many of our other reported anthracenyl isoxazole esters have shown the ester ethyl group in an endo-orienta- The asymmetric unit of compound (I) showing displacement ellipsoids drawn at the 50% probability level. The structure on the left is molecule A and that on the right is molecule B.

Figure 3
The asymmetric unit of compound (V) with displacement ellipsoids drawn at the 50% probability level.

Figure 1
Preparation and molecular structures of the title compounds.
The second title compound (V), C 21 H 17 NO 5 , crystallizes in the monoclinic P2 1 /c space group with one independent molecule in the asymmetric unit (Fig. 3). The anthrone ring system is virtually planar with an r.m.s. deviation of 0.029 Å . Like the other anthracenyl isoxazole structures we have reported (vide supra), the isoxazole ring is orthogonal to the anthracene ring, with a dihedral angle of 89.65 (5) . The ester ethyl group is in endo-orientation and the C19-O3-C20-C21 grouping is twisted [torsion angle = 86.7 (2) ].

Supramolecular features
In compound (I), weak C-HÁ Á ÁO hydrogen bonds between adjacent A molecules (C7-H7Á Á ÁO4 and C1-H1Á Á ÁO5) form a column running perpendicular to the [101] direction. Molecule B lies between the columns and its O7 atom accepts a hydrogen bond from H3 of molecule A (Table 1, Fig. 4). There is an aromaticstacking interaction with a centroidcentroid separation of 3.537 (5) Å between the planes of the C22-C25/C32/C33 and C1-C4/C11/C12 rings. Ainteraction is observed at a distance of 3.774 Å from atom C42 to the plane centroid.
In the crystal of compound (V), inversion dimers linked by pairwise O2-H2Á Á ÁO1 hydrogen bonds occur (Table 2, Fig. 5). A short contact distance between the isoxazole ring of one molecule (ring mean plane C15-C17/O5N1) and the carbonyl oxygen (O4) of another molecule [3.1486 (16) Å ] may contribute to the head-to-head, tail-to-tail arrangement in the crystal structure, also shown in Fig. 8b.
The Hirshfeld surface of compound V is mapped over d norm in a fixed color scale of À0.58 (red) to 1.31 (blue) arbitrary units (Fig. 8a), showing two short contacts from OÁ Á ÁH hydrogen bonds in red spots. The delineated two-dimensional fingerprint plots ( Fig. 9) Table 2 Hydrogen-bond geometry (Å , ) for (V).

Figure 4
The partial packing of compound (I). For clarity, only hydrogen bonds C1-H1Á Á ÁO5 i and C3-H3Á Á ÁO7 ii are shown as dashed lines, and H atoms not involved in these hydrogen bonds are removed.

Figure 6
(a) The Hirshfeld surface of (I) mapped over d norm . Short and long contacts are indicated as red and blue spots, respectively. Contacts with distances approximately equal to the sum of the van der Waals radii are colored white. (b) Weakinteractions are shown as green dashed lines on a surface mapped over curvedness. Thestacking is indicated by the green flat regions surrounded by dark blue edges.
also identifiable from the Hirshfeld surface mapped over the shape-index property (Fig. 8b).

Database survey
A search for the 9-nitroanthracenyl moiety in the Cambridge Structural Database (CSD version 5.43, November 2021 update; Groom et al., 2016) resulted in 14 hits, of which two crystal structures of 9-nitroanthracene itself were reported, namely refcodes NTRANT (Trotter, 1959) and NTRANT01 (Glagovich et al., 2004). The reported angles between the NO 2 plane and the anthracene plane are 84.78 and 69.40 , respec-tively, which agree with our observation of the disordered NO 2 group in (I).

Synthesis and crystallization
Iodination of aromatic hydrocarbons with molecular iodine has been accomplished by several methods, typically using an oxidizing agent to generate the iodonium cation electrophile. Among the conditions we surveyed, fuming nitric acid in particular (Bansal et al., 1987) with the anthracene isoxazole (II), appears to consistently produce the nitrated anthryl (I) rather than the desired iodo product (III). The anthryl isoxazole ester (II) was prepared as previously described (Mosher et al., 1996), and recrystallized before use. The ester The two-dimensional fingerprint plots for (I) delineated into (a) HÁ Á ÁH contacts, (b) OÁ Á ÁH/HÁ Á ÁO contacts, (c) CÁ Á ÁH/HÁ Á ÁC contacts, and (d) NÁ Á ÁH/HÁ Á ÁN contacts. Other contact contributions less than 5% are omitted.

Figure 8
(a) The Hirshfeld surface of (V) mapped over d norm . Short and long contacts are indicated as red and blue spots, respectively. Contacts with distances approximately equal to the sum of the van der Waals radii are colored white. Hydroxyl and carbonyl groups on the anthrone ring contributed major short contacts. (b)interactions (anthrone to anthrone and carbonyl to isoxazole ring) andinteraction (C-H bond to carbonyl) are shown as orange-red spots with green dashed lines in the shape-index map.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 3. In compound (I), the nitro group is disordered in each of the two independent molecules in the asymmetric unit. The occupancies of each disordered part were refined, converging to 0.572 (13) and 0.428 (13) for molecule A, and 0.64 (3) and 0.36 (3) for molecule B. EADP constraints were applied (Sheldrick, 2015) to each nitro group. The C-bound hydrogen atoms on both compounds were fixed geometrically and treated as riding with C-H = 0.95-0.98 Å and refined with U iso (H) = 1.2U eq (CH, CH 2 ) or 1.5U eq (CH 3 ). The O-bound H atom in (V) was found in a difference-Fourier map and refined freely. Four reflections (110, 110, 111 and 111) in compound (I) and four reflections (100, 10 4 5, 110 and 011) in compound (V) affected by the beam stop were omitted from the final cycles of refinement because of poor agreement between the observed and calculated intensities. The absolute structure of (I) was indeterminate in the present refinement.  For both structures, data collection: APEX2 (Bruker, 2012); cell refinement: SAINT (Bruker, 2018); data reduction: SAINT (Bruker, 2018); program(s) used to solve structure: SHELXS (Sheldrick, 2008); program(s) used to refine structure: SHELXL2018/1 (Sheldrick, 2015); molecular graphics: OLEX2 (Dolomanov et al., 2009); software used to prepare material for publication: OLEX2 (Dolomanov et al., 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.