Novel Steroidal[17,16-d]pyrimidines Derived from Epiandrosterone and Androsterone: Synthesis, Characterization and Configuration-Activity Relationships

Two series of novel steroidal[17,16-d]pyrimidines derived from natural epiandrosterone and androsterone were designed and synthesized, and these compounds were screened for their potential anticancer activities. The preliminary bioassay indicated that some of these prepared compounds exhibited significantly good cytotoxic activities against human gastric cancer (SGC-7901), lung cancer (A549), and hepatocellular liver carcinoma (HepG2) cell lines compared with 5-fluorouracil (5-FU), epiandrosterone, and androsterone. Especially the respective pairs from epiandrosterone and androsterone showed significantly different inhibitory activities, and the possible configuration-activity relationships have also been summarized and discussed based on kinase assay and molecular docking, which indicated that the inhibition activities of these steroidal[17,16-d]pyrimidines might obviously be affected by the configuration of the hydroxyl group in the part of the steroidal scaffold.


Introduction
Cancer is a disorder that rigorously affects the human population worldwide despite significant improvements in new treatment options [1,2], which is the second leading cause of death in the world. The increasing multidrug resistance and side effects have become an important cause of clinical death in the twenty-first century. Although there are many effective therapies for cancer control, chemotherapy remains the main option to treat cancer disease.
It is well known that pyrimidine is a class of aromatic heterocyclic compounds that contain two nitrogen atoms at positions 1 and 3 of the six-membered ring. Heterocyclic compounds bearing the pyrimidine core are of tremendous interest because the pyrimidine structure, which is an important part of many endogenous substances, can easily interact with enzymes, genetic materials, and bio-components within the cell [3,4]. Up to now, many natural and synthetic pyrimidines derivatives have been of enormous importance and demonstrate a variety of pharmacological activities, including anticancer [5][6][7][8][9][10][11], antiviral [12][13][14][15], antifungal [16][17][18], antioxidant [19][20][21][22][23], antibacterial [24][25][26], antituberculosis [27,28], anticonvulsant [29][30][31][32][33][34], antimalarial [35][36][37][38][39][40], antihypertensive [41][42][43][44][45], and anti-inflammatory [46][47][48][49][50][51]. Pyrimidine scaffolds are privileged heterocycles in drug discovery because they have considerable pharmacological and chemical significance and are also easily soluble in water [41]. Moreover, many disubstituted pyrimidines are also used as successful moieties to construct novel functional molecules [4]. Many scientists have focused on the discovery and structural optimization of pyrimidine derivatives, and so many In addition, natural steroids from animal and plant metabolites are types of important secondary metabolites that are widely distributed in marine environments and are extremely important biomarkers in the field of marine chemistry. In the marine sedimentary environment, the growth and reproduction of in-situ plankton and the input of terrestrial higher plant debris are the main sources of steroids. In the past few years, many steroids have been used as prototype scaffolds for constructing diverse molecules with extensive pharmaceutical activities, especially for dehydroepiandrosterone (DHEA), epiandrosterone (EPIA), and androsterone (AND) steroids [60][61][62]. Some of these novel steroidal derivatives showed significant inhibitory activities on human tumor cells in In addition, natural steroids from animal and plant metabolites are types of important secondary metabolites that are widely distributed in marine environments and are extremely important biomarkers in the field of marine chemistry. In the marine sedimentary environment, the growth and reproduction of in-situ plankton and the input of terrestrial higher plant debris are the main sources of steroids. In the past few years, many steroids have been used as prototype scaffolds for constructing diverse molecules with extensive pharmaceutical activities, especially for dehydroepiandrosterone (DHEA), epiandrosterone (EPIA), and androsterone (AND) steroids [60][61][62]. Some of these novel steroidal derivatives showed significant inhibitory activities on human tumor cells in culture, suggesting that the natural four fused rings scaffold of steroids might be closely bound up with the cytotoxic activity [63][64][65][66][67]. Some of the EPIA and AND derivatives [67][68][69][70][71][72][73][74][75] that exhibited significant inhibition on cancer cell lines are shown in Figure 1A and Figure 1C, respectively.
Based on these observations, two series of steroidal [17,16-d]pyrimidines derived from epiandrosterone and androsterone were also designed and synthesized, which integrate the structural features of pyrimidines and EPIA/AND unit to a core molecule as indicated in Figure 2, and their cytotoxic effects on tumor cell lines (HepG2, A549, SGC-7901) were fully investigated by MTT colorimetric method, and the possible configuration-activity relationships have also been summarized and discussed based on the activity data, kinase assay, and molecular docking. These results may provide much useful information for the discovery of novel cytotoxic agents from natural steroids.
Molecules 2023, 28, x FOR PEER REVIEW 3 of culture, suggesting that the natural four fused rings scaffold of steroids might be close bound up with the cytotoxic activity [63][64][65][66][67]. Some of the EPIA and AND derivatives [6 75] that exhibited significant inhibition on cancer cell lines are shown in Figure 1A an Figure 1C, respectively. Based on these observations, two series of steroidal [17,16-d]pyrimidines derive from epiandrosterone and androsterone were also designed and synthesized, which int grate the structural features of pyrimidines and EPIA/AND unit to a core molecule indicated in Figure 2, and their cytotoxic effects on tumor cell lines (HepG2, A549, SG 7901) were fully investigated by MTT colorimetric method, and the possible configur tion-activity relationships have also been summarized and discussed based on the activi data, kinase assay, and molecular docking. These results may provide much useful info mation for the discovery of novel cytotoxic agents from natural steroids.

Chemistry
In this work, two series of heterocyclic steroidal [17,16-d]pyrimidines derived fro epiandrosterone and androsterone were conveniently prepared, and the general metho for the preparation of these steroidal [17,16-d]pyrimidines derivatives 3a-l and 6a-l is d scribed in Scheme 1. Scheme 1. Synthetic route for steroidal [17,16-d]pyrimidine derivatives. Reagents and conditions: RCHO, NaOH, MeOH, rt, yield 83-94%; b. Guanidine nitrate, t BuOK, t BuOH, reflux, yield  According to Scheme 1, these steroidal [17,16-d]pyrimidines derivatives 3a-l and 6 l have been conveniently synthesized via a two-step transformation from epiandrosteron or androsterone, respectively. First, the various substituted aromatic aldehydes we treated with epiandrosterone or androsterone via aldol condensation to obtain the inte mediates 2a-l and 5a-l. Subsequently, the reaction of intermediates 2a-l and 5a-l wi guanidine nitrate in the presence of potassium tert-butoxide have been conveniently pr cessed to obtain the target molecules.
All the newly synthesized steroidal [17,16-d]pyrimidines 3a-l and 6a-l gave satisfa tory chemical analyses, including 1 H NMR, 13 C NMR, and ESI-MS spectra analyses, an the chemical structures and physiochemical properties of these compounds were summ rized in the experimental (Some of them have been collected in the Supplementa

Chemistry
In this work, two series of heterocyclic steroidal [17,16-d]pyrimidines derived from epiandrosterone and androsterone were conveniently prepared, and the general method for the preparation of these steroidal [17,16-d]pyrimidines derivatives 3a-l and 6a-l is described in Scheme 1.
Molecules 2023, 28, x FOR PEER REVIEW 3 of 24 culture, suggesting that the natural four fused rings scaffold of steroids might be closely bound up with the cytotoxic activity [63][64][65][66][67]. Some of the EPIA and AND derivatives [67][68][69][70][71][72][73][74][75] that exhibited significant inhibition on cancer cell lines are shown in Figure 1A and Figure 1C, respectively. Based on these observations, two series of steroidal [17,16-d]pyrimidines derived from epiandrosterone and androsterone were also designed and synthesized, which integrate the structural features of pyrimidines and EPIA/AND unit to a core molecule as indicated in Figure 2, and their cytotoxic effects on tumor cell lines (HepG2, A549, SGC-7901) were fully investigated by MTT colorimetric method, and the possible configuration-activity relationships have also been summarized and discussed based on the activity data, kinase assay, and molecular docking. These results may provide much useful information for the discovery of novel cytotoxic agents from natural steroids.
According to Scheme 1, these steroidal [17,16-d]pyrimidines derivatives 3a-l and 6al have been conveniently synthesized via a two-step transformation from epiandrosterone or androsterone, respectively. First, the various substituted aromatic aldehydes were treated with epiandrosterone or androsterone via aldol condensation to obtain the intermediates 2a-l and 5a-l. Subsequently, the reaction of intermediates 2a-l and 5a-l with guanidine nitrate in the presence of potassium tert-butoxide have been conveniently processed to obtain the target molecules.
According to Scheme 1, these steroidal [17,16-d]pyrimidines derivatives 3a-l and 6a-l have been conveniently synthesized via a two-step transformation from epiandrosterone or androsterone, respectively. First, the various substituted aromatic aldehydes were treated with epiandrosterone or androsterone via aldol condensation to obtain the intermediates 2a-l and 5a-l. Subsequently, the reaction of intermediates 2a-l and 5a-l with guanidine nitrate in the presence of potassium tert-butoxide have been conveniently processed to obtain the target molecules.
All the newly synthesized steroidal [17,16-d]pyrimidines 3a-l and 6a-l gave satisfactory chemical analyses, including 1 H NMR, 13 C NMR, and ESI-MS spectra analyses, and the chemical structures and physiochemical properties of these compounds were summarized in the experimental (Some of them have been collected in the Supplementary Materials). For NMR analyses, the assignments of different signals are based on the chemical shifts and intensity patterns. All 1 H NMR spectra of molecules 3a-l and 6a-l presented distinctive signals of methine proton attached to the hydroxyl group, which always indicated a multiplet or broad singlet at about 3.29-3.55 ppm and 3.77-3.82 ppm, respectively. The signal for protons of the hydroxyl group showed doublet peaks at about 4.43-4.48 ppm in compounds 3a-l, but at 4.16-4.19 ppm in compounds 6a-l. The signal for protons of the 2-amino group attached to the pyrimidine ring resonated almost the same as a singlet between 5.36 ppm and 6.58 ppm. The other set of signals that appeared in their 1 H NMR spectra in the ranges 2.79-0.69 ppm belonged to the protons of epiandrosterone or androsterone scaffold, and the signals at lower fields were assigned to the signals of aromatic protons as described in general structures in Scheme 1. The 13 C NMR analysis of molecules 3a-l and 6a-l display obvious peaks in the alkyl region, indicating the presence of the epiandrosterone or androsterone scaffold, respectively. Other peaks appearing at lower fields were assigned to the carbon signals of the aromatic and heterocyclic moieties. The electron spray impact mass spectra (ESI-MS) for compounds 2a-l and 6a-l were measured on a WATERS ACQUITY UPLC ® H-CLASS PDA (Waters ® ) instrument, and the ESI-MS of target molecules exhibited obvious molecular peak [M + H] + in the positive ion mode. All the characteristic peaks observed within the 1 H NMR and 13 C NMR spectra for title compounds are given in the experimental section.

Inhibitory Effects of the Target Compounds
All newly synthesized steroidal [17,16-d]pyrimidines derivatives 3a-l, 6a-l and the intermediates 2a-l, 5a-l were evaluated for their potential in vitro cytotoxic effects on human gastric cancer (SGC-7901), lung cancer (A549), and hepatocellular liver carcinoma (HepG2) cell lines by the standard MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assay [76] using 5-FU (5-Fluorouracil) as a positive control. The preliminary screening results are summarized in the following Figure 3. Generally, as shown in Figure 3, we can find most of these steroidal derivatives displayed well in vitro cytotoxic activities against three human cancer cell lines except compounds 2l, 5g, 5h, 5i, 5j, 5l, 6g, and 6l. Notably, the compounds 3a-l exhibited obviously inhibitory activities against all tested cell lines with 75.0-84.1% growth inhibition at the concentration of 40 µg/mL. From Figure 3, we also can observe that the steroidal [17,16-d]pyrimidines (3a-l) derived from epiandrosterone presented significantly better inhibitory activities than that of steroidal [17,16-d]pyrimidines (6a-l) derived from androsterone. Materials). For NMR analyses, the assignments of different signals are based on the chemical shifts and intensity patterns. All 1 H NMR spectra of molecules 3a-l and 6a-l presented distinctive signals of methine proton attached to the hydroxyl group, which always indicated a multiplet or broad singlet at about 3.29-3.55 ppm and 3.77-3.82 ppm, respectively. The signal for protons of the hydroxyl group showed doublet peaks at about 4.43-4.48 ppm in compounds 3a-l, but at 4. 16-4.19 ppm in compounds 6a-l. The signal for protons of the 2-amino group attached to the pyrimidine ring resonated almost the same as a singlet between 5.36 ppm and 6.58 ppm. The other set of signals that appeared in their 1 H NMR spectra in the ranges 2.79-0.69 ppm belonged to the protons of epiandrosterone or androsterone scaffold, and the signals at lower fields were assigned to the signals of aromatic protons as described in general structures in Scheme 1. The 13 C NMR analysis of molecules 3a-l and 6a-l display obvious peaks in the alkyl region, indicating the presence of the epiandrosterone or androsterone scaffold, respectively. Other peaks appearing at lower fields were assigned to the carbon signals of the aromatic and heterocyclic moieties. The electron spray impact mass spectra (ESI-MS) for compounds 2a-l and 6a-l were measured on a WATERS ACQUITY UPLC ® H-CLASS PDA (Waters ® ) instrument, and the ESI-MS of target molecules exhibited obvious molecular peak [M + H] + in the positive ion mode. All the characteristic peaks observed within the 1 H NMR and 13 C NMR spectra for title compounds are given in the experimental section.

Inhibitory Effects of the Target Compounds
All newly synthesized steroidal [17,16-d]pyrimidines derivatives 3a-l, 6a-l and the intermediates 2a-l, 5a-l were evaluated for their potential in vitro cytotoxic effects on human gastric cancer (SGC-7901), lung cancer (A549), and hepatocellular liver carcinoma (HepG2) cell lines by the standard MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assay [76] using 5-FU (5-Fluorouracil) as a positive control. The preliminary screening results are summarized in the following Figure 3. Generally, as shown in Figure 3, we can find most of these steroidal derivatives displayed well in vitro cytotoxic activities against three human cancer cell lines except compounds 2l, 5g, 5h, 5i, 5j, 5l, 6g, and 6l. Notably, the compounds 3a-l exhibited obviously inhibitory activities against all tested cell lines with 75.0-84.1% growth inhibition at the concentration of 40 µg/mL. From Figure 3, we also can observe that the steroidal [17,16-d]pyrimidines (3a-l) derived from epiandrosterone presented significantly better inhibitory activities than that of steroidal [17,16-d]pyrimidines (6a-l) derived from androsterone.  The preliminary assay demonstrated that many of these novel steroidal derivatives displayed good inhibitory activities (Figure 3), so in order to further clarify the potential activities, the IC 50 values for all molecules were fully evaluated. The cytotoxic activities expressed as IC 50 values for all molecules are described in Table 1, which further confirmed that all the steroidal [17,16-d]pyrimidine derivatives 3a-l exhibited higher inhibition activity than that of compounds 6a-l and the commercial 5-FU under the same conditions, respectively. As indicated in Table 1, compounds 2b, 2d, 2h, 2i, 2k, 3a-l, 5c-e, and 6b have higher cytotoxicity activities (Entries 2, 4, 8, 9, 13-24, 27-29, and 38) against all tested cell lines. In particular, the compounds 3a, 3b, 3d, and 3k derived from epiandrosterone exhibited Molecules 2023, 28, 2691 6 of 23 significant inhibition (Entries 13, 14, 16, and 23) on all cancer cell lines compared to the positive control 5-FU, the activities of which were also more successful than those of the corresponding 6a, 6b, 6d, and 6k derived from androsterone. On the whole, the activities of molecules 3a-l are superior to that of molecules 6a-l, which confirms that the configuration of the molecule might have a significant effect on the inhibition and that the compounds with epiandrosterone scaffold are favorable for inhibition activities.
Moreover, the dose-response analysis of cell growth inhibition activities for high potential compounds 3a, 3b, 3d, 3k, and 5-FU have been displayed in Figure 4, which indicated that the target compounds significantly inhibited SGC-7901, A549, and HepG2 cell proliferation in a concentration-dependent manner. In particular, compound 3a containing a phenyl unit exhibited the highest potential inhibitory activities against all tested cell lines with the IC 50 values of 1.07 ± 0.22, 0.61 ± 0.19, and 0.51 ± 0.13 µg·mL −1 (Entry 13 in Table 1), respectively, which was significantly better than that of the control 5-FU and epiandrosterone.
As indicated in Table 1, compounds 2b, 2d, 2h, 2i, 2k, 3a-l, 5c-e, and 6b have higher cytotoxicity activities (Entries 2, 4, 8, 9, 13-24, 27-29, and 38) against all tested cell lines. In particular, the compounds 3a, 3b, 3d, and 3k derived from epiandrosterone exhibited significant inhibition (Entries 13, 14, 16, and 23) on all cancer cell lines compared to the positive control 5-FU, the activities of which were also more successful than those of the corresponding 6a, 6b, 6d, and 6k derived from androsterone. On the whole, the activities of molecules 3a-l are superior to that of molecules 6a-l, which confirms that the configuration of the molecule might have a significant effect on the inhibition and that the compounds with epiandrosterone scaffold are favorable for inhibition activities.
Moreover, the dose-response analysis of cell growth inhibition activities for high potential compounds 3a, 3b, 3d, 3k, and 5-FU have been displayed in Figure 4, which indicated that the target compounds significantly inhibited SGC-7901, A549, and HepG2 cell proliferation in a concentration-dependent manner. In particular, compound 3a containing a phenyl unit exhibited the highest potential inhibitory activities against all tested cell lines with the IC50 values of 1.07 ± 0.22, 0.61 ± 0.19, and 0.51 ± 0.13 µg·mL −1 (Entry 13 in Table 1), respectively, which was significantly better than that of the control 5-FU and epiandrosterone.

Selectivity Profiling of Compound 3a
Compound 3a was selected as a representative to further investigate the kinase selectivity profile against a panel of tyrosine kinases ( Table 2). The kinase inhibitory activities of compound 3a (at the concentration of 10 µM) were determined by ADP-Glo Protocol or LANCE Protocol. As shown in Table 2, compounds 3a displayed some inhibitory effect on the kinases of CDK1/CyclinA2, ALK, FGFR1, and FAK with inhibition rates of 22.51%, 17.36%, 11.82%, and 10.52%, respectively.

Selectivity Profiling of Compound 3a
Compound 3a was selected as a representative to further investigate the kinase selectivity profile against a panel of tyrosine kinases ( Table 2). The kinase inhibitory activities of compound 3a (at the concentration of 10 µM) were determined by ADP-Glo Protocol or LANCE Protocol. As shown in Table 2, compounds 3a displayed some inhibitory effect on the kinases of CDK1/CyclinA2, ALK, FGFR1, and FAK with inhibition rates of 22.51%, 17.36%, 11.82%, and 10.52%, respectively.

Structure and Activity Relationships (SARs) and Molecular Docking
The structure evolution here was to modify epiandrosterone or androsterone scaffold with a pyrimidine ring system (3a-1, 6a-l) and aromatic enones (2a-l, 5a-l), respectively. According to the aforementioned results indicated in Table 1, we can obtain the general structure-activity profile for these novel steroidal [17,16-d]pyrimidine derivatives ( Figure 5).
As indicated in Table 2, compound 3a displayed some inhibitory effect on the kinases of CDK1/CyclinA2, ALK, FGFR1, and FAK. In addition, it is widely known that the cyclindependent protein kinase (CDK) has diverse cellular roles, including regulation of the cell cycle and transcription and differentiation. It also plays a crucial role in regulation of the growth, proliferation, and differentiation of cancer cells. Blocking the CDK-driven pathway by inhibiting the intracellular tyrosine kinase domain of CDK has resulted in considerable improvements in tumor therapy. In particular, a variety of pyrimidine derivatives were synthesized and evaluated for their abilities to target CDK tyrosine kinases, such as AZD5438 (shown in Figure 1B) and dinaciciclib. In order to find a deeper explanation of the structure and inhibitory activities of these steroidal [17,16-d]pyrimidine derivatives, the possible docking modes of compounds 3a, 3b, 3l, 6a, 6b, and 6l with CDK1 (PDB code: 6GU6) were modeled (Figures 6-9). way by inhibiting the intracellular tyrosine kinase domain of CDK has resulted in considerable improvements in tumor therapy. In particular, a variety of pyrimidine derivatives were synthesized and evaluated for their abilities to target CDK tyrosine kinases, such as AZD5438 (shown in Figure 1B) and dinaciciclib. In order to find a deeper explanation of the structure and inhibitory activities of these steroidal [17,16-d]pyrimidine derivatives, the possible docking modes of compounds 3a, 3b, 3l, 6a, 6b, and 6l with CDK1 (PDB code: 6GU6) were modeled (Figures 6-9).     As shown in Figures 6-8B1,C1, the six compounds 3a, 3b, 3l, 6a, 6b, and 6l almost bound to the outer edge of the ATP-binding pocket sandwiched between the P-loop (Gly 11-Glu 12-Gly 13-Thr 14-Phe 15-Gly 16) and active loop with the protein backbone from Gln 132 to Gly 145, but were far from the active segment spans with the protein backbone from Asp 146 to Glu 173 that can form a platform recognizing the CDK1 substrate residues to either side of the site phosphotransfer. In addition, the Lys 33-Glu12-Asp 146 triad in the bottom of the binding pocket of CDK1 also impacts the binding of the ATP adenine ring and phosphate moieties [77]. The hydrogen atom from the amidogen of the pyrimidine ring of compounds 3a, 3b, 3l, 6a, 6b, and 6l was far from Lys 33 with a different distance from 6.8 Å to 7.0 Å. These may be the reasons that compounds displayed lower inhibitory effects on the kinases of CDK1/CyclinA2. It is clear that hydrogen bonds play an important role in the inhibition activities of compounds. The hydrogen atom from the amidogen of the pyrimidine ring of compounds 3a, 3b, 3l, 6a, 6b, and 6l can form one hydrogen bond with the oxygen atom of Gln 132 with the same distance of 2.1 Å shown in Figures 6-8B3,C3. As indicated in Figures 6-8, the hydroxyl group of compounds 3a, 3b, and 3l was oriented toward the Glu8 of CDK1, but the hydroxyl group of compounds 6a, 6b, and 6l was oriented toward the Lys 9 of CDK1. As a result, the hydrogen atom of the hydroxyl group of compounds 3a, 3b, and 3l from epiandrosterone and compounds 6a, 6b, and 6l from androsterone can form one different hydrogen bond with the oxygen atom of Glu 8 and Lys 9 with different distances, respectively. The hydrogen bond of the hydroxyl group from compound 3a is stronger than 3b and 3l with a shorter length of 2.1 Å, 2.2 Å, and 2.3 Å, respectively. Meanwhile, the significant difference in efficacy of compound 6l was mainly due to the longest distance (2.1 Å) of the hydrogen bond in the hydroxyl group, while the length of the hydrogen bond of the hydroxyl group of compounds 6a and 6b was 2.0 Å.
Additionally, hydrophobic forces also influence the interaction of inhibitors with kinases. A hydrophobic interaction between the non-polar residues of CDK1 and the atoms of steroidal [17,16-d]pyrimidine derivatives is shown in Figure 9. As shown in Figures 6-8B1,C1, the six compounds 3a, 3b, 3l, 6a, 6b, and 6l almost bound to the outer edge of the ATP-binding pocket sandwiched between the P-loop (Gly 11-Glu 12-Gly 13-Thr 14-Phe 15-Gly 16) and active loop with the protein backbone from Gln 132 to Gly 145, but were far from the active segment spans with the protein backbone from Asp 146 to Glu 173 that can form a platform recognizing the CDK1 substrate residues to either side of the site phosphotransfer. In addition, the Lys 33-Glu12-Asp 146 triad in the bottom of the binding pocket of CDK1 also impacts the binding of the ATP adenine ring and phosphate moieties [77]. The hydrogen atom from the amidogen of the pyrimidine ring of compounds 3a, 3b, 3l, 6a, 6b, and 6l was far from Lys 33 with a different distance from 6.8 Å to 7.0 Å. These may be the reasons that compounds displayed lower inhibitory effects on the kinases of CDK1/CyclinA2.
It is clear that hydrogen bonds play an important role in the inhibition activities of compounds. The hydrogen atom from the amidogen of the pyrimidine ring of compounds  3a, 3b, 3l, 6a, 6b, and 6l can form one hydrogen bond with the oxygen atom of Gln 132 with the same distance of 2.1 Å shown in Figures 6-8B3, C3. As indicated in Figures 6-8,  the hydroxyl group of compounds 3a, 3b, and 3l was oriented toward the Glu8 of CDK1, but the hydroxyl group of compounds 6a, 6b, and 6l was oriented toward the Lys 9 of The analysis from above helps to explain that all the steroidal [17,16-d]pyrimidine derivatives 3a-l exhibited higher inhibition activity than that of compounds 6a-l, respectively. On the one hand, it occupied more space in the binding pocket when the hydroxyl group was oriented towards Glu 8 (shown in Figures 6-8B1,C1). On the other hand, it had stronger hydrophobic force due to the skeleton of steroidal [17,16-d]pyrimidines near the hinge region of CDK1, especially the methyl of C 18 (shown in Figures 6-8B3,C3 and 9). The above analysis also helped to explain why compound 3a exhibited the highest potent inhibitory activities on all tested cell lines (Entry 13 in Table 1).
The analysis indicated that the configuration of hydroxy in the C3 position might be a key influence on the inhibitory activities of compounds. The molecular docking analyses were helpful in identifying the target steroidal [17,16-d]pyrimidines derived from epiandrosterone that could serve as potential lead compounds for the discovery of anticancer agents.

Conclusions
Twenty-four steroidal [17,16-d]pyrimidines derived from epiandrosterone(3a-l) and androsterone(6a-l) were designed and synthesized, and their in vitro inhibition activities on three cell lines were investigated. All the steroidal [17,16-d]pyrimidine derivatives 3a-l exhibited higher inhibition activities against SGC-7901, A549, and HepG2 cell lines than that of compounds 6a-l with the same substituents, respectively. Compound 3a, containing a phenyl unit, exhibited the highest potential inhibitory activities against all tested cell lines with the IC 50 values of 1.07 ± 0.22, 0.61 ± 0.19, and 0.51 ± 0.13 µg·mL −1 . In addition, the detailed SARs analysis based on the inhibition activities, kinase assay, and molecular docking model demonstrated that the configuration of the hydroxyl group in the C 3 position of A ring of steroidal scaffold might obviously affect the potential activities of these steroidal [17,16-d]pyrimidines. The β-configuration of the hydroxyl group of steroidal [17,16-d]pyrimidines performed better than an α-configuration of the hydroxyl group with the reason that it occupied more space in the binding pocket when the hydroxyl group was oriented towards Glu 8, and it also had stronger hydrophobic force due to the proximity of the skeleton of steroidal [17,16-d]pyrimidines to the hinge region of CDK1, and especially to the methyl of C 18 , which will provide key evidence for further structural optimization for the discovery of novel anticancer agents.

Instrumentation and Chemicals
All starting materials and reagents are commercially available and were used without further purification unless otherwise specified. 1 H NMR and 13 C NMR spectra were recorded on a Bruker Avance III 600 MHz FT-NMR spectrometer using DMSO-d 6 or CD 3 OD as the solvent and tetramethylsilane (TMS) as the internal standard. Chemical shifts were reported in δ (parts per million) values, and coupling constants n J were reported in Hz. Mass spectra were performed on a WATERS ACQUITY UPLC ® H-CLASS PDA (Waters ® ) instrument. Analytical thin-layer chromatography was carried out on precoated silica gel plates GF254 (Qindao Haiyang Chemical, China), and spots were visualized with ultraviolet light. The calculated logP values (logP), which are the logarithms of the partition coefficients for octan-1-ol/water, were determined using the CS ChemOffice Ultra program (version 12.0, Cambridge-Soft, Cambridge, MA, USA).

General Synthetic Procedure for Intermediates 2a-l and 5a-l
The intermediates 2a-l and 5a-l were synthesized via the classical aldol reaction. Generally, a solution of epiandrosterone or androsterone (1 mmol) in methanol (15 mL) was added to appropriate aldehyde (1.1 mmol) and sodium hydroxide (0.4 g, 10 mmol), respectively. The mixture was stirred at room temperature and detected by thin-layer chromatography. After the completion of this reaction, the mixture was poured into 40 mL of ice water with stirring. Then the precipitate was filtered and dried to obtain the corresponding powder 2a-l and 5a-l. Their basic physico-chemical properties and spectra data are as follows:

Molecular Docking Study
All the molecular docking simulations were carried out by the AutoDock 4.2 software [79]. The docking tutorial we used and the detailed AutoDock basic operation methods can be found at: https://autodock.scripps.edu/ (accessed on 6 January 2023). The protein preparation process of flexible docking mainly includes fixing the exact residues, removing irrelevant water molecules, adding hydrogen atoms and adding charges, etc. The crystal structure (PDB: 6GU6, https://www.rcsb.org/structure/6GU6, accessed on 6 January 2023) of the CDK1 bond to 3a, 3b, 3l, 6a, 6b, and 6l were used in the docking studies. We first removed inhibitor dinaciclib from the crystal structure, then put the target molecules 3a, 3b, 3l, 6a, 6b, and 6l in the binding site, and the energy was optimized using a genetic algorithm. Only the best-scoring ligand-protein complex was used for the binding site analysis. All the docking results were processed and modified in PyMOL 1.7.4.5 software (https://pymol.org, accessed on 6 January 2023) and LigPlot + v.2.2.7 software (https://www.ebi.ac.uk/thornton-srv/software/LigPlus/, accessed on 6 January 2023).