Synthesis and Conformational Analysis of Naphthoxazine-Fused Phenanthrene Derivatives

The synthesis of new phenanthr[9,10-e][1,3]oxazines was achieved by the direct coupling of 9-phenanthrol with cyclic imines in the modified aza-Friedel–Crafts reaction followed by the ring closure of the resulting bifunctional aminophenanthrols with formaldehyde. Aminophenanthrol-type Mannich bases were synthesised and transformed to phenanthr[9,10-e][1,3]oxazines via [4 + 2] cycloaddition. Detailed NMR structural analyses of the new polyheterocycles as well as conformational studies including Density Functional Theory (DFT) modelling were performed. The relative stability of ortho-quinone methides (o-QMs) was calculated, the geometries obtained were compared with the experimentally determined NMR structures, and thereby, the regioselectivity of the reactions has been assigned.


Introduction
It is known that 9-phenanthrol is one of the most attractive structural units present in a large number of biologically active compounds. It is the identified inhibitor of the transient receptor potential melastatin (TRPM) 4 channels, a Ca 2+ activated non-selective cation channel whose mechanism of action remains to be determined [1][2][3][4]. Subsequent studies have proven that it modulates smooth muscle contraction in bladder and cerebral arteries, affects spontaneous activity in neurons and in the heart, and reduces lipopolysaccharide-induced cell death [5][6][7].
The Mannich reaction is one of the most important reactions in organic synthesis to form C-C bonds [8][9][10]. This synthetic pathway is widely used in the formation of secondary and tertiary amine derivatives and it is a key step in the synthesis of numerous bioactive molecules and complex natural products [11,12]. In one of the special variations of the modified Mannich reaction (mMR), 1-and 2-naphthol were applied as electron-rich aromatic compounds [13,14]. Mechanistically, the modified aza-Friedel-Crafts reaction can be interpreted as a special mMR, where electron-rich aromatic compounds such as 1-and 2-naphthol and their N-containing analogues are reacted with a wide range of cyclic imines to furnish aminonaphthols [15,16], aminoquinolinols, or aminoisoquinolinols [17-aminoquinolinols, or aminoisoquinolinols [17][18][19]. Accordingly, our first aim was to examine the reactivity of 9-phenanthrol with different cyclic imines in the modified aza-Friedel-Crafts reaction.
Recent publications pointed out that ortho-quinone methides (o-QMs) can also be generated from Mannich bases after thermal elimination of the amino group [20]. The reactive moiety thus formed can be stabilized by reactions with different dienophiles [14,21,22] or it can participate in [4 + 2] cycloadditions with cyclic imines to form new heterocycles [18,23,24]. Very recently the transformation of functionalised 1-aminobenzyl-2-naphthols via [4 + 2] cycloaddition has been examined [25]. It was found that the regio-and diastereoselectivity of the cycloaddition depend on the functional group of the phenyl substituent. Based on these findings, our aim was to synthesise aminobenzylphenanthrols or functionalised aminobenzylphenanthrols and to study their reactivity with cyclic imines as a novel precursor Mannich bases in [4 + 2] cycloaddition. Finally, we wanted to explore both the structure and conformational behaviour of the novel polyheterocycles by NMR spectroscopy and accompanying theoretical quantum chemical (QC) calculations.

Synthesis
To examine the possibility of the extension of 9-phenanthrol (1) in the modified aza-Friedel-Crafts reaction, 1 was reacted with 1.5 equiv. of 3,4-dihydroisoquinoline (2) [26] (Scheme 1) under neat conditions at 80 °C. After a reaction time of 120 min, the desired product (3) was isolated only with a yield of 10%. By increasing the temperature to 100 °C the yield of 3 increased slightly to 19%. Yields found under these conditions were not satisfactory and the appearance of side products was also observed. By changing the solvent to acetonitrile, treatment at reflux temperature (90 °C) afforded 3 in an isolated yield of 49%. The product was easily separated out from the reaction mixture that is acetonitrile proved to be an optimal solvent. To further improve the yield, the reaction was repeated under microwave (MW) conditions. In this case the reaction was driven at 100 °C and after a short reaction time (10 min) 3 was isolated in a yield of 67%, that could be improved to 82% by increasing the reaction time to 20 min (Table 1). It should be noted that the optimal workup procedure remained the filtration of the formed product from the cold reaction mixture. Subsequently, with the satisfactory optimal reaction conditions in hand, the extension possibility of the reaction was tested by using various cyclic imines, including 6,7-dihydrothieno[3 ,2-c]pyridine (4) [27] or 3,4-dihydro-β-carboline (6) [28]. These reactions afforded new 1-(9-phenanthrol-10-yl)-thieonopyridine 5 and 1-(9-phenanthrol-10-yl)-β-carboline 7 in yields of 92% and 76%, respectively (Table 1).  In the next step, the ring closures of aminophenanthrols 3, 5, and 7 were performed with 35% aqueous formaldehyde in CHCl 3 at room temperature. Reactions completed in relatively short time and phenanthrooxazines 8-10 were isolated in excellent yields by simple crystallization from n-hexane.
To test aminophenanthrols in cycloaddition reaction, first the synthesis of precursor 13 was achieved by reacting 9-phenanthrol (1) with morpholine in the presence of benzaldehyde. The reaction was carried out under solvent-free conditions at 80 • C. After a 4-h reaction, the desired 10-morpholinobenzyl-9-phenanthrol (13) was isolated by crystallization with n-hexane (Scheme 2).
Molecules 2020, 25, x 3 of 15 In the next step, the ring closures of aminophenanthrols 3, 5, and 7 were performed with 35% aqueous formaldehyde in CHCl3 at room temperature. Reactions completed in relatively short time and phenanthrooxazines 8-10 were isolated in excellent yields by simple crystallization from n-hexane. To test aminophenanthrols in cycloaddition reaction, first the synthesis of precursor 13 was achieved by reacting 9-phenanthrol (1) with morpholine in the presence of benzaldehyde. The reaction was carried out under solvent-free conditions at 80 °C. After a 4-h reaction, the desired 10-morpholinobenzyl-9-phenanthrol (13) was isolated by crystallization with n-hexane (Scheme 2). Scheme 2. Synthesis of 10-morpholinobenzyl-9-phenanthrol (13).
First, aminophenanthrol 13 was reacted with 3,4-dihydroisoquinoline 2 as dienophile. The reaction was performed in 1,4-dioxane under microwave irradiation at three different temperatures (60, 80, and 100 °C). In our first experiment, the reaction was performed at 60 °C and after a relatively short reaction time the desired product (14) was isolated only in a low yield (47%). Since the yield was not satisfactory, the reaction was repeated at 80 and 100 °C. As Table 2 shows, 80 °C and 15 min reaction time were found to be the optimal reaction conditions. The series of dienophiles was extended by using 6,7-dihydrothieno[3,2-c]pyridine 4 and 3,4-dihydro-β-carboline 6 (Scheme 3). The optimal conditions and related yields are listed in Table 2. Since two new stereogenic centres are generated during the reaction, two epimeric structures (a and b) can be obtained. The reaction was monitored by TLC and the compositions of the crude reaction mixtures were verified by 1 H-NMR analysis. All reactions were found to be diastereoselective and the relative configuration of H-9a:H-17 (14), H-9a:H-16 (15) and H-9a:H-18 (16) proved to be trans, based on the detailed NMR analysis (vide infra). First, aminophenanthrol 13 was reacted with 3,4-dihydroisoquinoline 2 as dienophile. The reaction was performed in 1,4-dioxane under microwave irradiation at three different temperatures (60, 80, and 100 • C). In our first experiment, the reaction was performed at 60 • C and after a relatively short reaction time the desired product (14) was isolated only in a low yield (47%). Since the yield was not satisfactory, the reaction was repeated at 80 and 100 • C. As Table 2 shows, 80 • C and 15 min reaction time were found to be the optimal reaction conditions. The series of dienophiles was extended by using 6,7-dihydrothieno[3,2-c]pyridine 4 and 3,4-dihydro-β-carboline 6 (Scheme 3). The optimal conditions and related yields are listed in Table 2. Since two new stereogenic centres are generated during the reaction, two epimeric structures (a and b) can be obtained. The reaction was monitored by TLC and the compositions of the crude reaction mixtures were verified by 1 H-NMR analysis. All reactions were found to be diastereoselective and the relative configuration of H-9a:H-17 (14), H-9a:H-16 (15) and H-9a:H-18 (16) proved to be trans, based on the detailed NMR analysis (vide infra).  Next, we wanted to investigate how o-QMs generated from functionalised aminophenanthrol derivatives can influence the [4 + 2] cycloaddition reaction. Accordingly, 9-phenanthrol and salicylic aldehyde were reacted in the presence of morpholine. The reaction was carried out under neat conditions at 80 °C. The characteristic spot formed according to the TLC was isolated and the NMR spectra of the formed compound confirmed the structure of 19. Formation of the dibenzo[a,c]xanthene (19) side-product can be explained by the elimination of water from diol 18. In this latter modified Mannich reaction, the availability of the nucleophile (morpholine) was postulated. Therefore, it was replaced by pyrrolidine applying again the above conditions (80 °C, neat). After a reaction time of 4 h, the expected phenanthrol derivative 21 was isolated in a yield of 76% (Scheme 4). Next, we wanted to investigate how o-QMs generated from functionalised aminophenanthrol derivatives can influence the [4 + 2] cycloaddition reaction. Accordingly, 9-phenanthrol and salicylic aldehyde were reacted in the presence of morpholine. The reaction was carried out under neat conditions at 80 • C. The characteristic spot formed according to the TLC was isolated and the NMR spectra of the formed compound confirmed the structure of 19. Formation of the dibenzo[a,c]xanthene (19) side-product can be explained by the elimination of water from diol 18. In this latter modified Mannich reaction, the availability of the nucleophile (morpholine) was postulated. Therefore, it was replaced by pyrrolidine applying again the above conditions (80 • C, neat). After a reaction time of 4 h, the expected phenanthrol derivative 21 was isolated in a yield of 76% (Scheme 4).  Next, we wanted to investigate how o-QMs generated from functionalised aminophenanthrol derivatives can influence the [4 + 2] cycloaddition reaction. Accordingly, 9-phenanthrol and salicylic aldehyde were reacted in the presence of morpholine. The reaction was carried out under neat conditions at 80 °C. The characteristic spot formed according to the TLC was isolated and the NMR spectra of the formed compound confirmed the structure of 19. Formation of the dibenzo[a,c]xanthene (19) side-product can be explained by the elimination of water from diol 18. In this latter modified Mannich reaction, the availability of the nucleophile (morpholine) was postulated. Therefore, it was replaced by pyrrolidine applying again the above conditions (80 °C, neat). After a reaction time of 4 h, the expected phenanthrol derivative 21 was isolated in a yield of 76% (Scheme 4).
To test the scope and limitations of the [4 + 2] cycloaddition, precursor 21 was first reacted with 3,4-dihydro-β-carboline as dienophile. The reaction was performed in 1,4-dioxane at three different temperatures (60, 80, and 100 • C) under microwave irradiation (Table 3). After the thermal decomposition of starting material 21, two types of o-QMs (22a and 22b) can be formed that can lead to the formation of two regioisomers and two diastereomers (Scheme 5) verified by 1 H-NMR analysis of the crude reaction mixture. Then, various reaction conditions were applied as depicted in Table 3. In all cases, reactions were found to be regio-and diastereoselective. The detailed NMR analysis (vide infra) confirmed that the formed/isolated product 24a is a phenanthroxazine and the relative configuration of H-9a:H-18 is trans. To test the scope and limitations of the [4 + 2] cycloaddition, precursor 21 was first reacted with 3,4-dihydro-β-carboline as dienophile. The reaction was performed in 1,4-dioxane at three different temperatures (60, 80, and 100 °C) under microwave irradiation (Table 3). After the thermal decomposition of starting material 21, two types of o-QMs (22a and 22b) can be formed that can lead to the formation of two regioisomers and two diastereomers (Scheme 5) verified by 1 H-NMR analysis of the crude reaction mixture. Then, various reaction conditions were applied as depicted in Table 3. In all cases, reactions were found to be regio-and diastereoselective. The detailed NMR analysis (vide infra) confirmed that the formed/isolated product 24a is a phenanthroxazine and the relative configuration of H-9a:H-18 is trans. Scheme 5. [4 + 2] Cycloaddition between 21 and 3,4-dihydro-β-carboline. The reaction was then extended by using 3,4-dihydroisoquinoline (2) and 6,7-dihydrothieno[3,2-c]pyridine (4) as cyclic imines (Scheme 6). In these cases, similar regio-and diastereoselectivity of the reactions were proved too. Detailed NMR analysis (vide infra) adequately supported that the isolated products are trans-phenanthroxazines (26a, 28a, Scheme 6). The optimal Scheme 5. [4 + 2] Cycloaddition between 21 and 3,4-dihydro-β-carboline.

Structural and Conformational Analysis
The complete 1 H and 13 C-NMR study of both diastereomeric phenanthr[9,10-e]oxazine derivatives 14-16 and regioisomeric/diastereotopic phenanthroxazine derivatives 24a, 26a, and 28a showed identical stereoisomerism (only one set of NMR spectra obtained). The NMR analyses of β-carboline derivative 16 and thiophene compound 28a are exemplarily utilised. The same stereoisomeric results were obtained for 14 and 24a, as well as 15 and 26a. For the corresponding structure elucidation, in addition to the NMR spectra (chemical shift, coupling constants, NOEs), quantum chemical calculations were also performed.

Detailed NMR Analysis of New Phenanthr[9,10-e]Oxazine 16
Determination of the relative configuration of 14-16 is based mainly on NOE interactions and their comparison with the corresponding calculated QC structures. As an example, β-carboline derivative 16 is examined (see Figure 1). Only one of the two possible diastereomers fits the experimental results: the position of methine hydrogens H-9a and H-18 was always found to be trans. The complete spatial information, found with the example of 16, is collected in Table 4, comparing NOE results with the calculated distances from QM calculations.

Structural and Conformational Analysis
The complete 1 H and 13 C-NMR study of both diastereomeric phenanthr[9,10-e]oxazine derivatives 14-16 and regioisomeric/diastereotopic phenanthroxazine derivatives 24a, 26a, and 28a showed identical stereoisomerism (only one set of NMR spectra obtained). The NMR analyses of β-carboline derivative 16 and thiophene compound 28a are exemplarily utilised. The same stereoisomeric results were obtained for 14 and 24a, as well as 15 and 26a. For the corresponding structure elucidation, in addition to the NMR spectra (chemical shift, coupling constants, NOEs), quantum chemical calculations were also performed.

Detailed NMR Analysis of New Phenanthr[9,10-e]Oxazine 16
Determination of the relative configuration of 14-16 is based mainly on NOE interactions and their comparison with the corresponding calculated QC structures. As an example, β-carboline derivative 16 is examined (see Figure 1). Only one of the two possible diastereomers fits the experimental results: the position of methine hydrogens H-9a and H-18 was always found to be trans. The complete spatial information, found with the example of 16, is collected in Table 4, comparing NOE results with the calculated distances from QM calculations.

Structural and Conformational Analysis
The complete 1 H and 13 C-NMR study of both diastereomeric phenanthr[9,10-e]oxazine derivatives 14-16 and regioisomeric/diastereotopic phenanthroxazine derivatives 24a, 26a, and 28a showed identical stereoisomerism (only one set of NMR spectra obtained). The NMR analyses of β-carboline derivative 16 and thiophene compound 28a are exemplarily utilised. The same stereoisomeric results were obtained for 14 and 24a, as well as 15 and 26a. For the corresponding structure elucidation, in addition to the NMR spectra (chemical shift, coupling constants, NOEs), quantum chemical calculations were also performed.

Detailed NMR Analysis of New Phenanthr[9,10-e]Oxazine 16
Determination of the relative configuration of 14-16 is based mainly on NOE interactions and their comparison with the corresponding calculated QC structures. As an example, β-carboline derivative 16 is examined (see Figure 1). Only one of the two possible diastereomers fits the experimental results: the position of methine hydrogens H-9a and H-18 was always found to be trans. The complete spatial information, found with the example of 16, is collected in Table 4, comparing NOE results with the calculated distances from QM calculations.    ψ-eq (smaller signal at higher field); ψ-ax (the corresponding broadened signal at lower field).

Detailed NMR Analysis of New Phenanthr[9,10-e]Oxazines 27a,b−28a,b
In analogy to previous investigations [25] and to the results of the structural analyses of 14-16 (vide supra), reactions starting from 21, in all cases, yielded trans isomers 24a, 26a, and 28a. Their NMR spectra are akin to the corresponding spectra of the compounds without an OH group at position 2 (14)(15)(16) and to compounds having a naphthalene moiety instead of phenanthrene [25].
To get the preferred stereoisomers of 28a,b, the trans/cis diastereomers of the regioisomers were calculated by the DFT method. Both the most stable structures and the corresponding energy differences are given in Figure 2. On the basis of these data, 28a could be confirmed to be the most stable structure. Energy differences around 1 kcal/mol to the next coming structure are rather unequivocal. Consequently, this structure of trans isomer 28a will be adjusted with the available experimental NMR [δ( 1 H)/ppm, δ( 13 C)/ppm, n J H,H /Hz)] and spatial NMR information (qualitative NOE s). ψ-eq (smaller signal at higher field); ψ-ax (the corresponding broadened signal at lower field).

Detailed NMR Analysis of New Phenanthr[9,10-e]Oxazines 27a,b−28a,b
In analogy to previous investigations [25] and to the results of the structural analyses of 14-16 (vide supra), reactions starting from 21, in all cases, yielded trans isomers 24a, 26a, and 28a. Their NMR spectra are akin to the corresponding spectra of the compounds without an OH group at position 2′ (14)(15)(16) and to compounds having a naphthalene moiety instead of phenanthrene [25].
To get the preferred stereoisomers of 28a,b, the trans/cis diastereomers of the regioisomers were calculated by the DFT method. Both the most stable structures and the corresponding energy differences are given in Figure 2. On the basis of these data, 28a could be confirmed to be the most stable structure. Energy differences around 1 kcal/mol to the next coming structure are rather unequivocal. Consequently, this structure of trans isomer 28a will be adjusted with the available experimental NMR [δ( 1 H)/ppm, δ( 13 C)/ppm, n JH,H/Hz)] and spatial NMR information (qualitative NOE′s).  First, all 1 H-NMR and 13 C-NMR chemical shifts for 28a could be unequivocally assigned. Hereby, the two singlets of protons H-16 and H-9a and the AX spin system of thiophene protons H-10 and H-11 can serve as useful starting points. The latter is the only spin system of aromatic protons with two doublets and with characteristic ortho-coupling constant of about 5.2 Hz. The high-field doublet (at 7.09 ppm) was assigned to the H-10 proton due to the substantial NOE to tertiary proton H-9a. On the other hand, no NOE between H-10 and the second tertiary proton (H-16) could be measured. The trans position of these two protons and, consequently, the trans stereochemistry of 28a can be proved. Furthermore, the HMBC connectivity of the same proton singlet (H-16) established the access to both the 1 H and 13 C assignments of the present aromatic moieties ((i) to C-6 , C-1 , and C-2 of the phenolic, and (ii) to C-16a and C-8b of the phenanthryl parts of the molecule), and to the piperidyl moiety via C-14. The trans stereochemistry of 28a could be proved additionally by the strong NOEs of H-16 to one of the H-14 protons, to the aromatic proton H-6' and to one of the H-14 protons. That is H-12a did not show any NOE to protons of the piperidyl ring.
The excellent agreement between the protons, proton distances, as calculated and the existence of strong NOEs in 28a can serve as an excellent further proof of the present regio-and diastereoselectivity. Similar 1 H/ 13 C-NMR studies of the isoquinoline (25a,b and 26a,b) and β-carboline analogues (23a,b and 24a,b) allowed to arrive at the same conclusion. These compounds also prefer the same regioisomerism in trans configuration of 24a and 26a, respectively.

General Methods
Melting points were determined on a Hinotek X-4 (Ningbo, China) melting point apparatus. Elemental analyses were performed with a Perkin-Elmer 2400 CHNS elemental analyzer (Waltham, MA, USA) in the Institute of Pharmaceutical Chemistry, University of Szeged. Merck Kieselgel 60F254 plates (Budapest, Hungary)were used for TLC. Microwave reactions were performed with a CEM Discover SP microwave reactor (Matthwes, NC, USA).
Quantum chemical calculations were performed using the Gaussian 09 program package [29] and carried out on LINUX clusters. The various different conformations and configurations of the studied compounds were optimised [30]. The B3LYP density functional method was selected for all calculations. The method is based on Becke s three-parameter hybrid functionals [31] and the correlation functional of Lee et al. [32]. All optimizations were carried out without any restriction at this B3LYP/6-311G(d,p) level of theory [33][34][35].