Verubulin (Azixa) Analogues with Increased Saturation: Synthesis, SAR and Encapsulation in Biocompatible Nanocontainers Based on Ca2+ or Mg2+ Cross-Linked Alginate

Tubulin-targeting agents attract undiminished attention as promising compounds for the design of anti-cancer drugs. Verubulin is a potent tubulin polymerization inhibitor, binding to colchicine-binding sites. In the present work, a series of verubulin analogues containing a cyclohexane or cycloheptane ring 1,2-annulated with pyrimidine moiety and various substituents in positions 2 and 4 of pyrimidine were obtained and their cytotoxicity towards cancer and non-cancerous cell lines was estimated. The investigated compounds revealed activity against various cancer cell lines with IC50 down to 1–4 nM. According to fluorescent microscopy data, compounds that showed cytotoxicity in the MTT test disrupt the normal cytoskeleton of the cell in a pattern similar to that for combretastatin A-4. The hit compound (N-(4-methoxyphenyl)-N,2-dimethyl-5,6,7,8-tetrahydroquinazolin-4-amine) was encapsulated in biocompatible nanocontainers based on Ca2+ or Mg2+ cross-linked alginate and it was demonstrated that its cytotoxic activity was preserved after encapsulation.


Introduction
Tubulin is a globular protein essential for the functioning of all eukaryotic cells.Heterodimers composed of αand β-tubulin molecules polymerize to form microtubules, which play a key role in the formation of cytoskeleton, cell division and intracellular transport [1][2][3].Tubulin-and microtubule-targeting agents have been the subject of intensive research and some of them are now widely used both in chemotherapy and as research tools to study the details of microtubule functioning.Despite widespread clinical use of antimicrotubule drugs, research in the field is constantly ongoing and associated mainly with the search for new alternatives with lower toxicities, better bioavailability and effectivity against resistant tumors [4][5][6][7][8][9][10].
The binding of tubulin-targeting agents to different regions of α,β-tubulin dimer causes its conformation switch and regulates microtubule behavior [3,4].The compounds interacting with the taxol-binding site and stabilizing microtubules [11], as well as microtubuledestabilizing agents that bind to the vinca domain of tubulin [12], are used in chemotherapy.Several representatives of a large group of colchicine-binding site ligands have also been introduced into clinical trials, the most successful of which was fosbretabulin (combretastatin A-4 prodrug) [13].However, so far none of these compounds has received full and final approval for clinical use.Considering the fact that many ligands of the colchicine-binding site are both efficient and easily accessible synthetically, they attract significant interest of researchers, and extensive efforts are being made to develop new molecules with this type of action [13][14][15].Note that the common limitations for the use of these compounds as drug candidates are their low selectivity towards non-tumor cells (resulting in high general toxicity) and, for certain molecules, poor bioavailability.
The binding of tubulin-targeting agents to different regions of α,β-tubulin dimer causes its conformation switch and regulates microtubule behavior [3,4].The compounds interacting with the taxol-binding site and stabilizing microtubules [11], as well as microtubule-destabilizing agents that bind to the vinca domain of tubulin [12], are used in chemotherapy.Several representatives of a large group of colchicine-binding site ligands have also been introduced into clinical trials, the most successful of which was fosbretabulin (combretastatin A-4 prodrug) [13].However, so far none of these compounds has received full and final approval for clinical use.Considering the fact that many ligands of the colchicine-binding site are both efficient and easily accessible synthetically, they attract significant interest of researchers, and extensive efforts are being made to develop new molecules with this type of action [13][14][15].Note that the common limitations for the use of these compounds as drug candidates are their low selectivity towards non-tumor cells (resulting in high general toxicity) and, for certain molecules, poor bioavailability.
In 2021, Banerjee et al. [33] showed that replacement of the fused benzene ring in quinazoline for a saturated 5-membered ring can lead to compounds with high cytotoxicity (see example in Figure 1).Recently, J. Tan et al. [34] described a large series of verubulin analogues, in one of which the quinazoline core was replaced with tetrahydroquinazoline (Figure 1).This compound not only retained high activity but also demonstrated the best selectivity (by 4-5 times) for tumor cells, despite the fact that most of the active substances in the series were not selective at all [34].The result obtained seemed  [36], some of its previously described analogues [25,33,34] and compounds studied in the present work.
In 2021, Banerjee et al. [33] showed that replacement of the fused benzene ring in quinazoline for a saturated 5-membered ring can lead to compounds with high cytotoxicity (see example in Figure 1).Recently, J. Tan et al. [34] described a large series of verubulin analogues, in one of which the quinazoline core was replaced with tetrahydroquinazoline (Figure 1).This compound not only retained high activity but also demonstrated the best selectivity (by 4-5 times) for tumor cells, despite the fact that most of the active substances in the series were not selective at all [34].The result obtained seemed interesting in the context of the strategy of designing pharmaceutically relevant molecules, including less toxic ones, by increasing their three-dimensionality ("escaping the flatland"), which has Pharmaceuticals 2023, 16, 1499 3 of 23 been widely discussed in recent decades [37][38][39][40].Therefore, it seemed interesting to conduct a systematic study of verubulin analogues with a reduced degree of flatness due to the replacement of the benzene ring fused with pyrimidine for a six-or seven-membered saturated ring B (Figure 1).In the present work, we synthesized a series of such analogues, also varying their conformational rigidity and the substituents in positions 2 and 4 of heterocyclic core A (Figure 1), and studied them in cell viability assays on cell lines of cancer and non-cancerous etiology and other biotests.
Additionally, we explored the possibility to improve the toxicological profile of a selected compound via its encapsulation into nanoparticles formed from Ca 2+ or Mg 2+ cross-linked alginate biopolymers.Alginate is a biodegradable and biocompatible copolymer of guluronic and mannuronic acids, which has already received US Food and Drug Administration (FDA) approval for medical applications [41].This natural anionic polymer (biomaterial) has been widely investigated and used in many areas of biomedical science and technology due to its biocompatibility, low toxicity, relatively low cost, softness and easy gelation upon addition of divalent cations (Ca 2+ and Mg 2+ ) [42].The role of Ca 2+ and Mg 2+ ions is to form a framework structure of polysaccharide macromolecules, accompanied by a significant compactization of the initial macromolecules of alginate.Alginate has promising biopharmaceutical properties such as pH sensitivity, biocompatibility, biodegradability, mucoadhesiveness, lack of toxicity and immunogenicity, which make it attractive for the development of carriers for drug immobilization [43].In our previous work, we have demonstrated that encapsulation of an isoxasole derivative with anti-cancer activity in alginate biopolymer can significantly increase its selectivity towards cancer cells [44].

Chemistry
4-Aminopyrimidines 2a-r were obtained via nucleophilic substitution of chlorine in the corresponding 4-cloropyrimidines upon the treatment with secondary amines in the presence of a catalytic amount of HCl (Schemes 1 and 2).In most cases, the reactions of 4-chloropyrimidines 3a-c with amines proceeded smoothly, affording target heterocycles in moderate to good yields.interesting in the context of the strategy of designing pharmaceutically relevant molecules, including less toxic ones, by increasing their three-dimensionality ("escaping the flatland"), which has been widely discussed in recent decades [37][38][39][40].Therefore, it seemed interesting to conduct a systematic study of verubulin analogues with a reduced degree of flatness due to the replacement of the benzene ring fused with pyrimidine for a six-or seven-membered saturated ring B (Figure 1).In the present work, we synthesized a series of such analogues, also varying their conformational rigidity and the substituents in positions 2 and 4 of heterocyclic core A (Figure 1), and studied them in cell viability assays on cell lines of cancer and non-cancerous etiology and other biotests.Additionally, we explored the possibility to improve the toxicological profile of a selected compound via its encapsulation into nanoparticles formed from Ca 2+ or Mg 2+ crosslinked alginate biopolymers.Alginate is a biodegradable and biocompatible copolymer of guluronic and mannuronic acids, which has already received US Food and Drug Administration (FDA) approval for medical applications [41].This natural anionic polymer (biomaterial) has been widely investigated and used in many areas of biomedical science and technology due to its biocompatibility, low toxicity, relatively low cost, softness and easy gelation upon addition of divalent cations (Ca 2+ and Mg 2+ ) [42].The role of Ca 2+ and Mg 2+ ions is to form a framework structure of polysaccharide macromolecules, accompanied by a significant compactization of the initial macromolecules of alginate.Alginate has promising biopharmaceutical properties such as pH sensitivity, biocompatibility, biodegradability, mucoadhesiveness, lack of toxicity and immunogenicity, which make it attractive for the development of carriers for drug immobilization [43].In our previous work, we have demonstrated that encapsulation of an isoxasole derivative with anti-cancer activity in alginate biopolymer can significantly increase its selectivity towards cancer cells [44].

Chemistry
4-Aminopyrimidines 2a-r were obtained via nucleophilic substitution of chlorine in the corresponding 4-cloropyrimidines upon the treatment with secondary amines in the presence of a catalytic amount of HCl (Schemes 1 and 2).In most cases, the reactions of 4chloropyrimidines 3a-c with amines proceeded smoothly, affording target heterocycles in moderate to good yields.The reaction of 2,4-dichlorotetrahydroquinazoline (3d) with 4-methoxy-N-methylaniline led to a multicomponent mixture of product, from which only the prevailing product 2n resulted from two-fold substitution and compound 2m could be isolated in low yields (Scheme 2).By reducing the time and using the excess of dichloride 3d, we managed to obtain tetrahydroquinazoline 2m as the main product, though during the purification the yield of this compound decreased.The interaction of dichloride 3d and sterically hindered 6-methoxytetrahydroquinoline was even more complicated and required a long reflux, after which the compounds 2o-r could be isolated from the reaction mixture (Scheme 2).The structure of all the obtained compounds was unambiguously determined via NMR spectroscopy, using 2D techniques, when necessary (see Supplementary Materials).

Cytotoxicity of Verubulin Analogues and SAR Analysis
The obtained series of verubulin analogues 2a-r (which also included three by-products isolated in the reactions according to Scheme 2) were initially evaluated for cytotoxicity against commonly used human breast cancer cell line MCF7' (fast-growth subclone), human lung epithelial carcinoma cell line A549 and etiologically non-cancer immortalized lung fibroblast cell line VA13 (WI38 subline 2RA).Cytotoxicity against immortalized human embryonic kidney cell line HEK293T, characterized by a high growth rate, was also studied regarding cytotoxicity against non-cancerous fast-growth cells.A standard MTT assay [45] was performed; verubulin (1) served as a positive reference.The data obtained are presented in Table 1 (for dose-response curves, see Supplementary Materials).
Saturated verubulin derivative 2c retained high cytotoxicity of the parent molecule (1) similarly to its analogue with a chlorine substituent in position 2 of ring A (2m), described by Tan et al. [34].Compounds 2c and 2m are more cytotoxic than their counterpart with a more polar amino group at C2 (2i), which is consistent with molecular dynamics modeling data (see below).Increasing the size of the alicycle B in the verubulin analogue 2c to a seven-membered ring (2k) reduces cytotoxicity by more than 10 times.A noticeable drop in activity during the expansion of ring B is most likely due to the steric hindrances it creates.Thus, this result denotes a limitation imposed on the volume of alicyclic ring B.
Saturated verubulin analogues (2c, 2k), regardless of the size of ring B, demonstrated no selectivity towards tumor cells in the pair A549/VA13.It should be noted that verubulin The reaction of 2,4-dichlorotetrahydroquinazoline (3d) with 4-methoxy-N-methylaniline led to a multicomponent mixture of product, from which only the prevailing product 2n resulted from two-fold substitution and compound 2m could be isolated in low yields (Scheme 2).By reducing the time and using the excess of dichloride 3d, we managed to obtain tetrahydroquinazoline 2m as the main product, though during the purification the yield of this compound decreased.The interaction of dichloride 3d and sterically hindered 6-methoxytetrahydroquinoline was even more complicated and required a long reflux, after which the compounds 2o-r could be isolated from the reaction mixture (Scheme 2).The structure of all the obtained compounds was unambiguously determined via NMR spectroscopy, using 2D techniques, when necessary (see Supplementary Materials).

Bioactivity Testing 2.2.1. Cytotoxicity of Verubulin Analogues and SAR Analysis
The obtained series of verubulin analogues 2a-r (which also included three byproducts isolated in the reactions according to Scheme 2) were initially evaluated for cytotoxicity against commonly used human breast cancer cell line MCF7 (fast-growth subclone), human lung epithelial carcinoma cell line A549 and etiologically non-cancer immortalized lung fibroblast cell line VA13 (WI38 subline 2RA).Cytotoxicity against immortalized human embryonic kidney cell line HEK293T, characterized by a high growth rate, was also studied regarding cytotoxicity against non-cancerous fast-growth cells.A standard MTT assay [45] was performed; verubulin (1) served as a positive reference.The data obtained are presented in Table 1 (for dose-response curves, see Supplementary Materials).
Saturated verubulin derivative 2c retained high cytotoxicity of the parent molecule (1) similarly to its analogue with a chlorine substituent in position 2 of ring A (2m), described by Tan et al. [34].Compounds 2c and 2m are more cytotoxic than their counterpart with a more polar amino group at C2 (2i), which is consistent with molecular dynamics modeling data (see below).Increasing the size of the alicycle B in the verubulin analogue 2c to a seven-membered ring (2k) reduces cytotoxicity by more than 10 times.A noticeable drop in activity during the expansion of ring B is most likely due to the steric hindrances it creates.Thus, this result denotes a limitation imposed on the volume of alicyclic ring B.
Saturated verubulin analogues (2c, 2k), regardless of the size of ring B, demonstrated no selectivity towards tumor cells in the pair A549/VA13.It should be noted that verubulin and its saturated analogue 2m which has been reported to be selective in the pair MCF7/MCF10A [34] demonstrated no selectivity in the pair A549/VA13 either.Conformational restriction of 4-methoxy-N-methylaniline moiety in saturated verubulin derivative 2c led to compound 2e with slightly less activity, but with an IC 50 still in the nanomolar concentration range.As seen in Figure 1, a similar change in activity was observed for the rigid analogue of verubulin obtained by Wang et al. [25].The same modification for verubulin analogue 2k with a seven-membered ring B resulted in compound 2l, which was seven to twelve times less active.Thus, the cytotoxicity in the series of both "open" and "rigid" compounds, depending on ring B, changed as follows: benzene > cyclohexane > cycloheptane.
In a series of conformationally restricted tetrahydroquinazolines, the important role of a non-polar substituent at position 2 of ring A was also demonstrated: the replacement of the methyl group by an amino group slightly reduced cytotoxicity against cancer cells (2j vs. 2e).Similar replacement with a hydroxyl group led to a complete loss of cytotoxicity (2o vs. 2e), which was, however, largely restored if the hydroxyl was converted to non-polar ether (O-iPr) moiety (2q vs. 2o).
Finally, analysis of SAR in the series of obtained verubulin analogues shows the critical role of the para-methoxy substituent in ring C. Its removal reduced cytotoxicity by 2-3 orders of magnitude (2a vs. 2c, 2g vs. 2i), while its replacement with the methyl group was less detrimental, although it still reduced the activity by an order of magnitude (2b vs. 2c, 2j vs. 2h).Compound 2f containing a meta-nitro group in ring C was inactive.Miscellaneous compounds 2n, 2p and 2r obtained as by-products in the synthetic schemes were moderately cytotoxic.

Effect of Verubulin Analogues on Microtubule Dynamics
Since verubulin is known to stimulate microtubule disassociation, we performed an immunofluorescent microscopy study to determine the response of microtubule dynamics in carcinoma A549 cells to the treatment with representative compounds 2b,c,e.As shown in Figure 2, exposure to 5-10 µM of the compounds for 24 h leads to a pronounced or complete disruption of microtubules (Figure 2A-C), and the pattern of a diffuse distribution of the stained microtubule subunits corresponds to that caused by combretastatin CA-4 (Figure 2E).Aberrant cell morphology-contraction and rounding (Figure 2A,B) or contraction and elongation (Figure 2C)-was also observed.As expected, immunofluorescent staining of microtubules in the cells treated with compound 2a, characterized by low cytotoxicity, revealed no effect on the microtubule cytoskeleton even at very high concentrations of this compound (Figure 2D).
An additional study of the effect of compounds 2c and 2e on microtubule assembly in vitro was monitored by recording the change in optical density at 355 nm during the GTP-dependent polymerization on crude preparation of tubulin (Tb) and microtubuleassociated proteins (MAPs) from mouse brains.The obtained graphs of the change in optical density versus time are shown in Figure 3.It was shown that both 2c and 2e inhibited the initial rate of tubulin polymerization in a dose-dependent manner similarly to verubulin.The concentrations of the compounds were the same as in Figure 2. On the basis of initial rates (from 6 to 16 min for control and from 12 to 32 min for others) of the linear change in optical density at 355 nm vs. compound concentrations, where each point is the average of two different experiments with three repeats in each, the IC 50 were calculated for 2c, 2e and verubulin: 7.2 ± 1.9 µM; 26 ± 6 µM and 3.1 ± 1.1 µM, respectively (Figure S2).It should be mentioned that unlike control probe (DMSO) in the presence of 2c, 2e and verubulin there are no features of depolymerization which causes a plateau in the control curve (Figure 3).

Molecular Dynamics Simulations for Compound 2c
The obtained data (Figures 2 and 3) confirm that new compounds significantly disturb the assembly of microtubules and exhibit characteristics similar to tubulin polymerization inhibitors, particularly such as verubulin.As was mentioned above, the latter binds to the colchicine site of tubulin, which is located in the β-subunit at the interface with the α-subunit of the α,β-heterodimer.The data of X-ray diffraction analysis of tubulin complexes with verubulin (PDB ID: 5XKF) [46] and several of its analogues [29,30,33,34] have been collected recently.To get an idea of how the saturated analogue of verubulin 2c (the most active in the series) binds to the protein, we performed molecular dynamics simulations (Figure 4).The plot of the root mean square deviations (RMSDs, Figure 4A) for the protein, GDP, GTP and ligand non-hydrogen atoms and the visual analysis of the trajectory confirm that system stability is retained from about the fiftieth nanosecond until the end of the course of the production simulation (100 ns), although the position of the ligand is slightly adjusted compared to the docking pose (by a shift of tetrahydroquinazoline residue).
vs. 2c, 2j vs. 2h).Compound 2f containing a meta-nitro group in ring C was inactive.Miscellaneous compounds 2n, 2p and 2r obtained as by-products in the synthetic schemes were moderately cytotoxic.

Effect of Verubulin Analogues on Microtubule Dynamics
Since verubulin is known to stimulate microtubule disassociation, we performed an immunofluorescent microscopy study to determine the response of microtubule dynamics in carcinoma A549 cells to the treatment with representative compounds 2b,c,e.As shown in Figure 2, exposure to 5-10 µM of the compounds for 24 h leads to a pronounced or complete disruption of microtubules (Figure 2A-C), and the pattern of a diffuse distribution of the stained microtubule subunits corresponds to that caused by combretastatin CA-4 (Figure 2E).Aberrant cell morphology-contraction and rounding (Figure 2A,B) or contraction and elongation (Figure 2C)-was also observed.As expected, immunofluorescent staining of microtubules in the cells treated with compound 2a, characterized by low cytotoxicity, revealed no effect on the microtubule cytoskeleton even at very high concentrations of this compound (Figure 2D).As shown in Figure 4B, the saturated analogue of verubulin, compound 2c, is oriented in the colchicine-binding site similar to the parent molecule.Ring B in 2c is located in the hydrophobic pocket within the side chains of Ala252β, Ala314β, Ala315β, Ile316β.Both the methyl group attached to tetrahydroquinazoline and the methoxy group of 4methoxyphenyl residue also form important hydrophobic interactions with the β-tubulin subunit (the former is exposed in the hydrophobic subpocket formed by the side chain of Leu250β, Leu253β and the aliphatic chain of Gln245, while the latter is located close to the aliphatic chains of Met257β and Lys350β).The N 1 atom of ring A of verubulin analogues, as was shown by Banerjee et al. [29], can form a key water-mediated hydrogen bond with the main chain of Cys239β, and since such an interaction is water-mediated it is theoretically possible for compound 2c, despite some differences in the location of ring A compared to verubulin.change in optical density at 355 nm vs. compound concentrations, where each point is the average of two different experiments with three repeats in each, the IC50 were calculated for 2c, 2e and verubulin: 7.2 ± 1.9 µM; 26 ± 6 µM and 3.1 ± 1.1 µM, respectively (Figure S2).It should be mentioned that unlike control probe (DMSO) in the presence of 2c, 2e and verubulin there are no features of depolymerization which causes a plateau in the control curve (Figure 3).

Molecular Dynamics Simulations for Compound 2c
The obtained data (Figures 2 and 3) confirm that new compounds significantly disturb the assembly of microtubules and exhibit characteristics similar to tubulin polymerization inhibitors, particularly such as verubulin.As was mentioned above, the latter binds to the colchicine site of tubulin, which is located in the β-subunit at the interface with the α-subunit of the α,β-heterodimer.The data of X-ray diffraction analysis of tubulin complexes with verubulin (PDB ID: 5XKF) [46] and several of its analogues [29,30,33,34] have been collected recently.To get an idea of how the saturated analogue of verubulin 2c (the most active in the series) binds to the protein, we performed molecular dynamics simulations (Figure 4).The plot of the root mean square deviations (RMSDs, Figure 4A) for the protein, GDP, GTP and ligand non-hydrogen atoms and the visual analysis of the trajectory confirm that system stability is retained from about the fiftieth nanosecond until the end of the course of the production simulation (100 ns), although the As shown in Figure 4B, the saturated analogue of verubulin, compound 2c, is oriented in the colchicine-binding site similar to the parent molecule.Ring B in 2c is located in the hydrophobic pocket within the side chains of Ala252β, Ala314β, Ala315β, Ile316β.Both the methyl group attached to tetrahydroquinazoline and the methoxy group of 4methoxyphenyl residue also form important hydrophobic interactions with the β-tubulin subunit (the former is exposed in the hydrophobic subpocket formed by the side chain of Leu250β, Leu253β and the aliphatic chain of Gln245, while the latter is located close to the aliphatic chains of Met257β and Lys350β).The N 1 atom of ring A of verubulin analogues, as was shown by Banerjee et al. [29], can form a key water-mediated hydrogen bond with the main chain of Cys239β, and since such an interaction is water-mediated it is theoretically possible for compound 2c, despite some differences in the location of ring A compared to verubulin.The binding mode of compound 2c corresponds to the observed SAR patterns, in particular, to a decrease in activity when the non-polar group at position 2 of ring A was replaced by a polar one, as well as a loss of activity upon removal of the methoxy group in ring C (see above, Table 1).The dramatic difference in cytotoxicity of compounds 2o and 2j (both with polar substituents at position 2 of ring A) is most likely due to the disruption of the above-mentioned water-mediated hydrogen bonds of the N 1 with the key amino acid residue Cys239β caused by the hydroxyl group of 2o, capable of forming stronger hydrogen bonds than the amino group of its counterpart 2j.Overall, molecular modeling and SAR data further confirm that saturated verubulin analogue 2c retains the mode of action of the parent molecule.The binding mode of compound 2c corresponds to the observed SAR patterns, in particular, to a decrease in activity when the non-polar group at position 2 of ring A was replaced by a polar one, as well as a loss of activity upon removal of the methoxy group in ring C (see above, Table 1).The dramatic difference in cytotoxicity of compounds 2o and 2j (both with polar substituents at position 2 of ring A) is most likely due to the disruption of the above-mentioned water-mediated hydrogen bonds of the N 1 with the key amino acid residue Cys239β caused by the hydroxyl group of 2o, capable of forming stronger hydrogen bonds than the amino group of its counterpart 2j.Overall, molecular modeling and SAR data further confirm that saturated verubulin analogue 2c retains the mode of action of the parent molecule.

Apoptosis Induction by Verubulin Analogues
Apoptosis induction by verubulin and hit compounds 2c,e,i,j,m was studied by flow cytometry to reveal the mechanism of HCT116 tumor cell death.Apoptosis is an important and active regulatory pathway of cell growth and proliferation.Annexin V is a calciumdependent phospholipid-binding protein with a high affinity for phosphatidylserine (PS), a membrane component normally localized to the internal face of the cell membrane.Early in the apoptotic pathway, molecules of PS are translocated to the outer surface of the cell membrane where Annexin V can readily bind them [47,48].
All tested compounds were found to significantly induce apoptosis of HCT116 cells after 48 h of exposure (Figure 5).Nearly 77.5% of untreated HCT116 cells were viable (Annexin V (−) and 7-AAD (−)).For verubulin (1), the fraction of viable cells was reduced compared to the control, and almost half of the studied cells (49.3%) were found at the early (Annexin V (+) and 7-AAD (−) 16.3%) or late (Annexin V (+) and 7-AAD (+) 33%) apoptosis stages.The number of viable cells was about 50% of the entire studied population of HCT116 cells.Saturation of ring B (2c) was shown to affect the activation of apoptosis: the population of apoptotic cells decreased down to 30% when treated with 2c (Figure 5D).Further conformational restriction (2e) led to a decrease in the number of viable cells down to 30% and a significant increase in the number of cells at the late apoptosis stage (Figure 5E).The replacement of the methyl group in position 2 of the pyrimidine ring in compound 2c by chlorine (2m) had no effect on the apoptotic pattern compared to verubulin, while its replacement by a NH 2 group (2i) changed it: the percentage of early apoptotic cells increased significantly (46.3%) and the population of cells in late apoptosis decreased.Thus, the introduction of an amino group into the pyrimidine ring affects the rate of cell entry into apoptosis.For compound 2j, the population of cells in early apoptosis increased up to 56.4%.dependent phospholipid-binding protein with a high affinity for phosphatidylserine (PS), a membrane component normally localized to the internal face of the cell membrane.Early in the apoptotic pathway, molecules of PS are translocated to the outer surface of the cell membrane where Annexin V can readily bind them [47,48].
All tested compounds were found to significantly induce apoptosis of HCT116 cells after 48 h of exposure (Figure 5).Nearly 77.5% of untreated HCT116 cells were viable (Annexin V (−) and 7-AAD (−)).For verubulin (1), the fraction of viable cells was reduced compared to the control, and almost half of the studied cells (49.3%) were found at the early (Annexin V (+) and 7-AAD (−) 16.3%) or late (Annexin V (+) and 7-AAD (+) 33%) apoptosis stages.The number of viable cells was about 50% of the entire studied population of HCT116 cells.Saturation of ring B (2c) was shown to affect the activation of apoptosis: the population of apoptotic cells decreased down to 30% when treated with 2c (Figure 5D).Further conformational restriction (2e) led to a decrease in the number of viable cells down to 30% and a significant increase in the number of cells at the late apoptosis stage (Figure 5E).The replacement of the methyl group in position 2 of the pyrimidine ring in compound 2c by chlorine (2m) had no effect on the apoptotic pattern compared to verubulin, while its replacement by a NH2 group (2i) changed it: the percentage of early apoptotic cells increased significantly (46.3%) and the population of cells in late apoptosis decreased.Thus, the introduction of an amino group into the pyrimidine ring affects the rate of cell entry into apoptosis.For compound 2j, the population of cells in early apoptosis increased up to 56.4%.The obtained data demonstrate that saturation of ring B in the verubulin molecule, though preserving both cytotoxicity and the effect on microtubules, does not lead to an increase in the selectivity.Therefore, we studied the opportunity to improve the toxicological profile for compound 2c via its encapsulation in Ca 2+ /Mg 2+ cross-linked alginate The obtained data demonstrate that saturation of ring B in the verubulin molecule, though preserving both cytotoxicity and the effect on microtubules, does not lead to an increase in the selectivity.Therefore, we studied the opportunity to improve the toxicological profile for compound 2c via its encapsulation in Ca 2+ /Mg 2+ cross-linked alginate nanoparticles as a drug delivery system.Compound 2c was encapsulated in alginate (Alg)-based nanocontainers cross-linked with Ca 2+ and Mg 2+ ions synthesized following the procedure described in [44,49].The molar ratio both for [Alginate monomer units]/[Ca 2+ ] and [Alginate monomer units]/[Mg 2+ ] was 10 ÷ 1.After purification and lyophilization, white complexes were obtained which were dissolved in water that resulted in (Alg)-Ca 2+ and (Alg)-Mg 2+ nanocontainer solutions, respectively.
Aminopyrimidine 2c was transformed into a protonated form 2c•HCl for better interaction with the carboxyl groups of anionic nanocontainers (see Supplementary Materials).
The formation of polysaccharide nanocontainers as well as their interaction with aminopyrimidine 2c•HCl was controlled by means of dynamic light scattering (DLS), laser microelectrophoresis and UV spectrophotometry (Figure S4).
The synthesized nanocontainers as well as nanocontainers filled with 2c were easily dispersed in bi-distilled water.In all cases, dynamic light scattering found one type of composite microgel with narrow particle size distribution (see Figure S5 for typical distribution functions).It was established that the effective hydrodynamic diameter (D h ) of the initial alginate macromolecules in an aqueous solution is 680 nm.The interaction of linear alginate both with Ca 2+ and Mg 2+ ions leads to a significant contraction in particle size: for samples (Alg)-Ca 2+ and (Alg)-Mg 2+ , the particle diameters are 150 nm and 205 nm, respectively (Table 2).The effective diameters of 2c•HCl-(Alg)-Ca 2+ and 2c•HCl-(Alg)-Mg 2+ decrease down to 130 nm and 106 nm, respectively, that confirms the formation of the formulations.The polydispersity of alginate-based nanocarriers was evaluated using PDI [50]  The incorporation of aminopyrimidine 2c into microgels (entries 1 and 3) was also confirmed using UV spectroscopy (Figure S4).The absorption spectrum of protonated aminopyrimidine 2c shows two pronounced electronic transitions at wavelengths of 225 nm and 300 nm.In the absorption spectra of the interaction products of both microgels with 2c, there is a sharp decrease in the peak intensity at 225 nm.The peak intensity at 300 nm does not change significantly.The obtained result indicates the electrostatic interaction of the protonated form of 2c with the carboxyl groups of the microgels [51,52].In turn, the preservation of the peak at 300 nm can be used for the quantitative determination of the 2c content in microgels.A calibration curve (Figure S4) was plotted to quantify the 2c content in the nanocontainers.It was established that both (Alg)-Ca 2+ and (Alg)-Mg 2+ bound 16 wt.%aminopyrimidine 2c that corresponds to the quantitative binding of 2c.In addition, the nature of the cross-linking ions does not affect the binding capacity of nanocontainers towards aminopyrimidine 2c.
To evaluate the ability of the nanocontainers to retain the loaded aminopyrimidine 2c, external dialysis solutions after purification of the microgels were analyzed by measuring the optical density at two wavelengths of 225 nm (D 225 ) and 300 nm (D 300 ), corresponding to the two maxima in the spectrum of the control aminopyrimidine solution (Figure S4).The D 225 and D 300 in external dialysis solutions after purification of systems 2c•HCl-(Alg)-Ca 2+  and 2c•HCl-(Alg)-Mg 2+ were found to be 0.18 and 0.012, respectively, which is comparable to the corresponding value for the solvent (bi-distilled water).That confirms that the loaded compound is retained in the nanocontainers.
The MTT assay was carried out on the A549 cell line and non-tumor hTERT-immortalized human fibroblasts for 48 h (Table 2; for cell survival curves, see Supplementary Materials).While (Alg)-Ca 2+ and (Alg)-Mg 2+ nanocontainers demonstrated no or low toxicity against cell lines of different etiology (entries 1 and 3, respectively), their formulations with compound 2c were found to preserve cytotoxic activity at the level of free amine 2c.
The differences in cell survival were demonstrated on non-tumor human hTERTimmortalized fibroblasts.While for the Mg 2+ formulation (entry 3) IC 50 values obtained for A549 and hTERT-immortalized fibroblast cell lines were the same, the Ca 2+ formulation (entry 1) demonstrated two-fold selectivity towards tumor cells.
The data obtained allow us to conclude that the cytotoxic activity against the A549 cell line is preserved under the action of the compound 2c in alginate formulations compared to its free form.The encapsulation of the amine in (Alg)-Ca 2+ nanoparticles is more promising from the point of view of selectivity to cancer cells.

S N Ar Reactions of 4-Chloropyrimidines 2a-d with Amines (General Method)
To the solution of chloropyrimidine 2a-d (0.55 mmol) and corresponding amine (0.66 mmol) in isopropanol (5 mL), two drops of HCl were added.The reaction mixture was refluxed for 5 h.When chloropyrimidine 2d was used as starting compound, reagent ratios and reaction time were modified, see below.Isopropanol was evaporated under reduced pressure and, to the residue, water (15 mL)  MCF7 , VA13, A549 and HEK293T cell lines were maintained in DMEM/F-12 (Thermo Fisher Scientific, Waltham, MA, USA) culture medium containing 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA, USA) and 50 µg/mL penicillin and 0.05 mg/mL streptomycin at 37 • C (Thermo Fisher Scientific, Waltham, MA, USA) in 5% CO 2 .Cells were maintained at 37 • C in a humidified MCO-18AC incubator (Sanyo, Tokyo, Japan) supplied with 5% CO 2 .Cell cultures were tested for the absence of mycoplasma and validated by STR.

Cell Viability Assay (MTT Assay)
The cytotoxicity of the substances was tested using the 3-(4,5-dimethylthiazol-2-yl)2,5diphenyl tetrazolium bromide (MTT) assay [41] with some modifications.A total of 2500 cells per well for the MCF7 , HEK293T, A549 cell lines or 4000 cells per well for the VA-13 cell line were plated out in 135 µL of DMEM-F12 media (Gibco, Thermo Fisher Scien-tific, Waltham, MA, USA) in a 96-well plate and incubated in the 5% CO 2 incubator for the first 20 h without treatment.Then, 15.8 µL of media-DMSO solutions of tested substances (final DMSO concentrations in the media were 0.5% or less) was added to the cells (triplicate each) and cells were treated for 64 h with 0.023 mg/L-50 mg/L (eight dilutions) of compounds 2f,n,o,r; 1.14 µg/L-2.5 mg/L (eight dilutions) of compounds 2a,b,d,g,h,k,l,p,q; 57.2 ng/L-125 µg/L (eight dilutions) of compounds 2e,i,j; 22.9 ng/L-50 µg/L (eight dilutions) of compounds 1,2c,m.Doxorubicin was used as a control substance.Then, the MTT reagent (Paneco LLC, Moscow, Russia) was added to cells up to a final concentration of 0.5 g/L (10 × stock solution in PBS was used) and incubated for 2 h at 37 • C in the incubator, under an atmosphere of 5% CO 2 .Then, the MTT solution was discarded, and 140 µL of DMSO (PharmaMed LLC, Moscow, Russia) was added.The plates were swayed on a shaker (120 rpm) to dissolve the formazan.The absorbance was measured using a microplate reader (VICTOR X5 Light Plate Reader, PerkinElmer, Waltham, MA, USA) at a wavelength of 555 nm (in order to measure formazan concentration).The results were used to construct dose-response graphs and to estimate IC 50 values (IC 50 is the concentration resulting in half of the maximal cytotoxic effect) with GraphPad Prism V, GraphPadSoftware, Inc., San Diego, CA, USA.

Immunofluorescent Microscopy
A549 cells were seeded on coverslips (treated with poly-l-lysine at a concentration of 33 mg/L for an hour).Compounds were added up to the corresponding concentration after 24 h attachment and growth of the cells.Then, cells were incubated 24 h.
The cells were fixed with methanol (precooled in liquid nitrogen for 20 min), no additional permeabilization was performed.To prevent non-specific binding, cells were incubated with 4% BSA in PBS for one hour.Then, a solution of mouse monoclonal antibodies against alpha-tubulin cross-linked with Alexa 488 fluoroform (Thermo Fisher Scientific, Waltham, MA, USA, 32-2588) at a concentration of 5 µg/mL in 1% BSA solution in PBS was added to the cells, and the slides were incubated at +5 • C in a water chamber for 24 h.After the end of the incubation time, the cells were washed several times with PBS buffer solution and fixed on a glass slide with Moviol.Next day, cells were imaged with a ZEISS LSM 900 (ZEISS Microscopy, GmbH, Germany) and images processed with ImageJ.

Tubulin + MAP Polymerization
A tubulin + MAP polymerization assay was conducted as previously described in [57].
Dynamic light scattering (DLS) measurements were carried out in 0.15 M NaCl aqueous solution using a Photocor Complex photometer (Photocor Instruments, Moscow, Russia) equipped with a He-Ne 10 mW laser (λ = 633 nm) as the light source.The data were processed using DynaLS (DynaLS Version 2.7.1) software.Mean hydrodynamic sizes of particles were determined by DLS at the fixed scattering angle (90 • ).Software provided by the manufacturer was employed to calculate diameter values.Electrophoretic mobility (EPM) of particles was measured by laser microelectrophoresis in a thermostatic cell using a Brookhaven Zeta Plus (Brookhaven Instruments, Holtsville, NY, USA) device with the corresponding software.UV/vis spectroscopy measurements were performed with a UV-PC PE-5400UV spectrophotometer (Ecros, Moscow, Russia).

Preparation of Nanocontainers
All nanocontainers were synthesized at room temperature.A dilute solution of CaCl 2 (1 mL) containing 0.56 mg of the compound was added dropwise to 50 mL of a 0.1% Na-Alg solution.A dilute solution of MgSO 4 (1 mL) containing 0.60 mg of the compound was added dropwise to 50 mL of a 0.1% Na-Alg solution.
The resulting mixtures was intensely stirred for 24 h and placed in dialysis bags (MWCO~12 kDa, «Sigma», Merck, Rahway, NJ, USA) and dialyzed for 24 h against deionized water.After dialysis, the solutions were lyophilized.

Preparation of Nanocontainers Filled with 2c
A solution containing 1 mg of aminopyrimidine in 50 µL HCl was added dropwise to 5 mL of an aqueous solution of the Alg-Ca 2+ (and Alg-Mg 2+ ) nanocontainer containing 5 mg of the compound.The reaction mixture was intensely stirred for 24 h and neutralized by NaOH solution to pH~7.High-molecular products were purified by flow dialysis.Then, lyophilization was used.

Cytotoxicity Studies of Nanocontainers Filled with 2c
Lung adenocarcinoma A549, breast adenocarcinoma MDA-MB-231 and non-tumor fibroblasts immortalized with h-TERT cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA).Cells were routinely propagated in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (HyClone, Logan, UT, USA), 2 mM L-glutamine, 100 U/mL penicillin and 100 µg/mL streptomycin (PanEco, Moscow, Russia) at 37 • C, 5% CO 2 in a humidified atmosphere.Cells in the logarithmic phase of growth were used in the experiments.
Cytotoxicity was assessed in a formazan conversion assay (MTT test) after a 48 h exposure.After the completion of drug exposure, 0.5 mg/mL 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide (MTT reagent) was added to cells for 2 h, the culture medium was removed, cells were resuspended in 100 µL DMSO and the optical densities were measured on a Multiscan FC plate spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) at a wavelength of 571 nm.The percentage of surviving cells for each dose was calculated as the quotient of the average optical density in the wells after incubation with this dose to the average optical density of the control wells (the values of the latter are taken as 100%).Five independent experiments were performed for each concentration.Standard deviations did not exceed 10%.

Molecular Dynamics Simulation
Molecular dynamics simulation was performed according to a previously published detailed protocol [57] using the model obtained from the tetramer structure of α,β-tubulin dimer PDB ID: 6GJ4 [58] and CHARMM36/CGenFF 4.4 force field [59,60] in GROMACS 2021.2 software [61] (the starting structure of the protein-ligand complex was obtained by

Figure 3 .
Figure 3. Influence of compounds 2c and 2e in comparison with verubulin (positive control) on Tb + MAP polymerization as a time-dependent increase in absorbance at 355 nm.Background-subtracted data were normalized to control probe.This figure shows the results of the representative experiment.Each curve is the average of three repeats; each point is mean ± SEM.

Figure 3 .
Figure 3. Influence of compounds 2c and 2e in comparison with verubulin (positive control) on Tb + MAP polymerization as a time-dependent increase in absorbance at 355 nm.Backgroundsubtracted data were normalized to control probe.This figure shows the results of the representative experiment.Each curve is the average of three repeats; each point is mean ± SEM.

Figure 4 .
Figure 4. (A) Root mean square deviation of the protein, GDP, GTP and ligand (2c) non-hydrogen atoms during molecular dynamics simulation of the complex with the α,β-tubulin heterodimer.(B) Binding mode of compound 2c refined using molecular dynamics simulation (detailed view of the binding site in the β-subunit).Ligand 2c is represented by a green-colored stick model; the amino acid residues located within 3 Å of the ligand are represented by beige stick models.The position of verubulin (PDB ID: 5XKF, matched to the β-subunit) is shown by a mauve-colored stick model.The guanosine triphosphate molecule bound in the α-subunits is shown by stick models on the right in the background.The hydrogen atoms are omitted for clarity.

Figure 4 .
Figure 4. (A) Root mean square deviation of the protein, GDP, GTP and ligand (2c) non-hydrogen atoms during molecular dynamics simulation of the complex with the α,β-tubulin heterodimer.(B) Binding mode of compound 2c refined using molecular dynamics simulation (detailed view of the binding site in the β-subunit).Ligand 2c is represented by a green-colored stick model; the amino acid residues located within 3 Å of the ligand are represented by beige stick models.The position of verubulin (PDB ID: 5XKF, matched to the β-subunit) is shown by a mauve-colored stick model.The guanosine triphosphate molecule bound in the α-subunits is shown by stick models on the right in the background.The hydrogen atoms are omitted for clarity.

Figure 5 .
Figure 5. Apoptotic profile of HCT116 cells treated with selected compounds after 48 h.Concentration of compounds 2 × IC50 (see Supplementary Materials).

Figure 5 .
Figure 5. Apoptotic profile of HCT116 cells treated with selected compounds after 48 h.Concentration of compounds 2 × IC 50 (see Supplementary Materials).

2. 3 .
Encapsulation of Compound 2c in Biocompatible Nanocontainers Based on Ca 2+ /Mg 2+ Cross-Linked Alginate and Investigation of Cytotoxicity of Formulations via MTT Assay . It was shown that the synthesized nanocontainers are characterized by PDI values not exceeding 0.05.Negative EPM values indicate that both 2c•HCl-(Alg)-Ca 2+ and 2c•HCl-(Alg)-Mg 2+ are characterized by high aggregative stability.

Table 1 .
Cytotoxicity effects calculated as IC 50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 1 .
Cytotoxicity effects calculated as IC50 for 64 h treatment of the cell lines of different etiology.

Table 2 .
Characteristics of aqueous dispersions of Ca 2+ and Mg 2+ cross-linked nanocontainers filled with immobilized compound 2c and their components.