Synthesis and Cytotoxicity Evaluation of a Series of 3-Alkenyl-2-Hydroxy-1 , 4-Naphthoquinones Obtained by an Efficient Knoevenagel Condensation

A modified and efficient Knoevenagel condensation procedure was developed to synthesize the title compounds using β-alanine and acetic acid as catalysts, showing good to excellent yields. We used lawsone with suitable aliphatic aldehydes including isobutyraldehyde, 3-methylbutyraldehyde, 2-ethylbutyraldehyde, phenylacetaldehyde, 2-phenyl-propionaldehyde, among others. These compounds were submitted to cytotoxic screening against three tumor cell lines presenting good to excellent cytotoxic profiles.


Introduction
The synthesis of 3-alkenyl-2-hydroxy-1,4-naphthoquinones has been a topic of permanent interest in Brazil and abroad. 1,2These compounds share structural features with natural lapachol (1), the most abundant quinone found in the core wood of various Bignoniaceae.Lapachol has a long tradition of use as a precursor of modified quinones, particularly for derivatives aimed at trials and screening of their multipurpose biological profiles. 3ooker, in 1936, 4 published a systematic synthetic procedure for a series of 3-alkenyl-2-hydroxy-1,4-naphtho quinones using lawsone (2) with suitable aliphatic aldehydes with hydrochloric acid as catalyst.Hooker used propionaldehyde, n-butyraldehyde, valeraldehyde, heptanaldehyde, phenylacetaldehyde and hydrocynamaldehyde (3a-f), obtaining the corresponding products in low to modest yields (23-48%, Table 1).In this work, Hooker also pointed that he was unable to obtain the condensation product from isobutyraldehyde, and from acetaldehyde he notes that a rapid consumption of the formed product was observed. 4Fifty years later, in 1986, Bock et al. 5 published an update of this methodology (see Table 1).
As pointed above, the use of modified Knoevenagel catalysts and water scavenger could reduce undesired reactions like lawsone dimerization and side chain acid cyclization, common side reactions in this type of condensation. 14Finally, the synthesized compounds were subjected to an evaluation screening for their potentials as antitumor lead compounds against different cancer cell lines.In this study, the biological screening also included a comparison with some other available furan and pyran derived hydroxymethyl-1,4-naphthoquinones obtained previously in our group by an acid catalyzed Prins cyclization of lapachol and norlapachol with paraformaldehyde and the newly corresponding acetyl derivatives.

General information
Air-and moisture-sensitive reactions were carried out under argon atmosphere.Reagents were purchased from Sigma-Aldrich, Dinâmica or Vetec and distilled or used without further purification.Reactions were monitored by thin layer chromarography (TLC) analysis on precoated silica gel plates (Merck, Kieselgel 60 GF 254 ) and compounds were visualized with UV light.Column chromatography was performed on silica gel 60 (70-230 mesh, Merck).Melting points were measured in open capillary tubes in a Quimis apparatus and are uncorrected.The infrared (IR) spectra were recorded on an IFS66 Bruker spectrophotometer using KBr discs or Varian Mercury 640IR with attenuated total reflectance (ATR).High resolution mass spectrometry (HRMS) analyses were performed on a matrix assisted laser desorption ionization-Scheme 2. Two-step synthesis of 3,4-dehydropyrano-1,4-naphthoquinones using Knoevenagel conditions in the obtention of 2-hydroxy-3-alkenyl-1,4-naphthoquinone. 7
Method B: in a flask, it was added lawsone (2, 0.2 g, 1.2 mmol), β-alanine (0.06 g, 0.67 mmol) and toluene (40 mL), followed by glacial acetic acid (0.034 g, 0.56 mmol) and the corresponding aldehyde (2.3 mmol).The reaction system was kept under reflux using Dean-Stark apparatus.Inspection of the progress of the reaction was made by TLC for about 1 h.The solvent was removed and the compound was purified by column chromatography with silica gel and hexanes:ethyl acetate 90:10 as eluent.

Synthesis of 10b and 11b
In a preheated flask filled with formic acid (25 mL), it was added paraformaldehyde (1.32 g, 44.3 mmol) followed by corresponding lapachol or norlapachol (22 mmol).The reaction was heated for 2 h at 90 o C, and water (50 mL) was added and the mixture was kept under reflux for 12 h.The reaction mixture was cooled, neutralized with Na 2 CO 3 saturated solution and the resulting precipitate was submitted to vacuum filtration.The compounds were purified by column chromatography with silica gel using mixtures of dichlorometane:hexanes and dichloromethane:ethyl acetate with increasing polarity.

Synthesis of 10c and 11c
In a round bottom flask with 10a or 11a (200 mg), it was added acetic anhydride (1 mL, 10.6 mmol) followed by pyridine (1 mL, 12.4 mmol).After 2.5 h, after TLC inspection, the product was precipitated in ice-water and submitted to vacuum filtration and column chromatography with silica gel using mixtures of dichlorometane:hexanes and dichloromethane:ethyl acetate with increasing polarity.

In vitro citotoxicity assay
The human lung carcinoma cell line (NCI-H292), human larynx carcinoma cell line (HEP-2) and breast adenocarcinoma (MCF-7) were purchased from the Adolfo Lutz Institute, São Paulo, Brazil.A Dulbecco's Modified Eagle's medium (DMEM), enriched with 10% of fetal bovine serum, 1% of L-glutamine and 1% of antibiotics (penicillin and streptomycin), was used for cell cultivation and to perform the tests.The cytotoxic activity was investigated using the 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. 19ell suspensions were diluted to 10 5 cells mL -1 , suitably prepared and distributed in plates of culture with 96 wells (225 μL in each well), then incubated at 37 o C in a humid atmosphere with 5% of CO 2 .After 24 h, 25 μL of either the synthesized compounds or the reference drugs (doxorubicine and cisplatin) were added to each well.The plates were incubated again at 37 o C for 72 h.Then, 25 μL of MTT solution (5 mg mL -1 ) were added to each well, and the mixture was incubated at 37 o C for 2 h.At the end of this period, the culture medium with the MTT excess was aspirated and after that, 100 μL of DMSO were added to each well to dissolve the formazan crystals. 20The optical density (OD) of the wells was measured at 540 nm and compared to the control (cells with medium only).The data represent the mean of two experiments in triplicate and were expressed as means ± standard deviation (SD). 21The concentration at which 50% survival of cells was observed (IC 50 ), the effective concentration causing 50% decrease in cell viability (EC 50 ) and their 95% confidence intervals were determined from nonlinear regression using the program SigmaPlot version 11. 22

Results and Discussion
The preliminary data found by the present method in the synthesis of norlapachol 4g showed a 95% yield, and resulted in patent filing. 14In this method, benzene was substituted by toluene, a greener and safer solvent with a higher boiling point.As far as we know, this remarkable result is better than any previously reported in the literature.For example, the classic Hooker's oxidative degradation of natural lapachol 1 to obtain norlapachol 4g results in yields around 50%. 16 Initially, we planned to use the inexpensive n-butyraldehyde as a model to observe the importance of some reaction conditions employed in original Kopanski methodology (Table 3), in particular the proportion of acetic acid/β-alanine catalyst, the amount of aldehyde, the substitution of acetic acid by p-TsOH, and the substitution of β-alanine by triethylamine or L-proline (Table 3).
From the data of Table 3, we conclude that the acetic acid or β-alanine alone were not effective as catalysts (entries 1 and 2), but the use of the mixture (entry 3) results in a good yield of 64%.Increase of acid alone or β-alanine, however (entries 4 and 5, respectively), did not contribute to further increase the observed yields.An increase of twofold in the aldehyde drops the yield (entry 6), probably caused by competition for self-condensation; however, the increase of catalyst in a 4:1:1 proportion of aldehyde:acetic acid:β-alanine results in the best yield of 81% (entry 7).Increasing two-fold the proportion of either catalyst or both (entries 8, 9 and 10) drops the observed yield.Changing the acetic acid by triethylamine or p-toluenesulfonic acid (entries 11 and 12) also drops the yield.The use of a known condensation catalyst L-proline (entry 13) was also unsuccesful.The reaction seems to be sensitive to the amount of acid and β-alanine, however, all the reactions with excess acid catalysis results in the formation of the bis-adduct dimer of lawsone, a polar product visible by TLC inspection that was studied in a previous article by our group. 23We chose to compare this modification of the original Kopanski procedure (named method A, entry 3), with this new method B (entry 9, Table 3) in a series of experiments conducted to verify the scope of the method with a range of suitable aldehydes.These results are shown in Table 4, and we planned to submit initially the available propionaldehyde (3a), n-butyraldehyde (3b), n-valeraldehyde (3c), phenylacetaldehyde (3e), n-hexanaldehyde (3p), isobutyraldehyde (3g), isovaleraldehyde (3i) and 2-ethylbutyraldehyde (3j) to condense with lawsone (2) using both conditions A and B that differ only by B using twice the amount of catalyst that A uses (Table 4).The best yields were obtained in either method for the more substituted alkyl, specially those with an extra alkyl at alpha position (see entries 6-8).This should reflect an easier elimination step at the formation of a more stable substituted double bond.This effect could be extended to the stabilization of the double bond by conjugation with a phenyl group (entry 4).The use of twice the amount of catalyst in method B should give better and more consistent yields without exceptions.At a glance we could see that method B was generally better than the previous one, varying from an excelent yield of 95 and 96% for phenylaceltadehyde (3e) and 2-ethylbutyraldehyde (3j) (entries 4 and 8), respectively, to good yields (85, 86 and 85%) for isobutyraldehyde (3g), isovaleraldehyde (3i) and 2-phenyl-propionaldehyde (3q) (entries 6, 7 and 10), respectively.Linear alkyl aldehydes are also otained in good yields of 70 and 76%, respectively, for n-valeraldehyde (3c) and n-hexanaldehyde (3p) (entries 3 and 9), representing better yields than those afforded using method A. The 3-phenyl-propionaldehyde (3f) and even the cyclopentanone (3r) furnish modest yields of 47 and 36% (entries 5 and 11, Table 4), respectively, using method B. These last two precursors and 2-phenyl-propionaldehyde (3q) did not react (entries 5, 10 and 11) at the conditions used in method A by TLC inspection.However, the method did not improve the yield of propionaldehyde (3a), entry 1.With the aid of some minor condition modifications, as the use of β-alanine alone as catalyst with 2 h reaction time, we are able to improve almost twice the yield of 3a to 31%.We believe that the extremely low boiling point of propionaldehyde could explain the observed low yield.The compound 2-phenyl-propionaldehyde (3q) was only recently described and synthesized by the aid of a Pd-catalyzed hydroxyinvolved enolate-type reaction with lawsone and suitable alkenes.Previously, this type of functionalization of a 2-hydroxy-naphthoquinone used labor-intensive alkylation conditions 24 or phenyliodonium ylides BF 3 catalyzed cross coupling. 25Although Knoevenagel condensations with acyclic or cyclic ketones as enolates are common, 26 it is noteworthy that in our report the cyclopentanone yielding compound 4r was used almost always as an electrophile, 27 and as far as we know, this compound is new.The similar reaction with cyclohexanone gives no results with the sort of different modifications described before in this work.
The synthesis of the furan (10b) and pyran (11b) 3-hydroxymethyl-1,4-naphthoquinones were performed by the reaction of lapachol 1 or norlapachol 4g with paraformaldehyde in formic acid (Scheme 3), resulting in the condensation of the alkene quinone side chain and the formaldehyde by a Prins reaction, 28,29 followed by in situ cyclization, in 76 and 51% yield, respectively.The corresponding acetates 10c and 11c were obtained by straightforward acetylation procedure in 86 and 54% yield, respectively.
Concerning the biological activity, the title compounds 4e, 4i, 4j, and 4p were previously tested together for antifungal activity, 30 4g and 4i for molluscicidal activity, 31,32 4g for antimycobacterial activity, 33 4b, 4g and 4i for trypanocidal activity, 34 4g and 4i for leishmanicial activity, 35,36 and 4e was evaluated for cytotoxicity, 37 antifungal 30 and anticancer activities. 24Although the majority of the title compounds were known and tested elsewhere, they were all tested together in the present study for cytotoxic screening against tumor cell lines HEP-2, MCF-7 and NCI-H292 (Table 5).
Concerning the inhibition of growth of HEP-2 tumor cells, the compounds alkenyl-1,4-naphthoquinone series 4b, 4c, 4i, 4j, 4p and 4q showed the highest percentuals of inhibition, superior to those from lapachol 1 and reference compound doxorubicin (Table 5).Compound 4c showed a higher percentual of inhibition (95.6%) against (HEP-2) cell lines when compared to 4j with 94.1%, 4p with 94%, 4b with 93.5%, 4i with 93% and 4q with 91.2%.In this series the norlapachol 4g showed the lowest inhibition percentual (27.3%) (Table 5).Concerning the MCF-7 cell line tested the only compound with high activity was 4j with 77.9% inhibition, showing a better profile than the isomer with a linear alkyl side chain 4p with 53.6% inibition (Table 5).Notably, in these experimental trials the compound norlapachol 4g did not inhibit the growth of MCF-7 cells or of NCI-H292.In NCI-H292, the highest inhibition for the alkenyl-1,4-naphthoquinones 4 were 85.9, 85.7 and 85.3% for compounds 4i, 4j and 4r, respectively (Table 5).The presence of an additional phenyl ring in compounds 4e, 4f and 4q did not improve the biological profile.From the data on Table 5 we could see that the series of alkenyl-1,4-naphthoquinones 4 showed a remarkable selectivity to the larynx tumor cell lines, when compared to other tumor cell lines.
for 10c in HEP-2 cell lines to 92.7% for 11c in NCI-H292 cell lines.The acetylation of the hydroxymethyl groups enhances the observed activities in all tested cell lines, either in pyran 10b or furan 11b series (Table 5).
In order to determine the half maximal inhibitory concentration (IC 50 ) values (μM) displayed in Table 6, we decided to adopt a threshold for using only compounds with at least 75% inhibition in all cell lines tested (at 25 mg mL -1 concentration).Thus, compounds 4j, 10a-c and 11a-c are those satisfying this criterion as can be observed from the data in Table 5.The 3-alkenyl-2-hydroxy-1,4-naphthoquinone compound 4j was shown to be the least cytotoxic for all tested cell lines (Table 6).For the MCF-7 tumor cell line, compounds 11a-c were more active than 10a-c, β-nor-lapachone (11a) being more cytotoxic than 11b, 11c, and even more cytotoxic than the positive control.All tested cyclized compounds showed a remarkable citotoxic profile against the breast MCF-7 tumor cell line, ranging from 0.10 μM for 11a to 5.51 μM for 10b.Also in this tumor cell line, the hydroxymethyl 10b/11b and acetate derivatives 10c/11c resulted in decreased cytotoxicity when compared to the parent compounds β-lapachone (10)  and β-nor-lapachone (11), although 10c did not result in changes to the observed activity (Table 6).The tumor cell lines NCI-H292 and HEP-2 showed more sensitivity to the acetylated derivative 10c, with IC 50 of 2.26 and 6.81 μM, respectively, and also for the less active 11c, with IC 50 values of 11.13 and 13.6 μM, respectively.In general, the tested compounds showed remarkable and consistent cytotoxicity in tested cancer cell lines, although the compounds presented high inhibitory effect in the growth of Buffalo green monkey kidney (BGMK) cells used to access the selectivity, which means a non-selective effect and high Data are presented as half maximal inhibitory concentration (IC 50 ) values and 95% confidence intervals obtained by nonlinear regression for all cell lines.Doxorubicin was used as positive control.Only compounds with an IC 50 value lower than 5 μg mL -1 for at least one cell line were considered active.MCF-7: breast adenocarcinoma cell line; NCI-H292: human lung carcinoma cell line; HEP-2: human larynx carcinoma cell line; BGMK: Buffalo green monkey kidney cell line; ND: not determined.
A very different scenario is obtained when the cyclized derivatives 10a-c and 11a-c were compared with the open chain alkenyl 1,4-naphthoquinones 4, with all compounds showing excellent profiles of inhibition, ranging from 99% cytoxicity for the series tested, except for the series of cyclic furan-1,4-naphthoquinone and 3-hydroxymethyl derivatives 11a-c, which showed a good selectivity index for MCF-7.

Conclusions
In this work a very efficient Knoevenagel procedure was presented for the obtention of 2-hydroxy-3-alkenyl-1,4-naphthoquinones (4) from lawsone (2) and suitable aldehydes, using β-alanine and acetic acid as catalysts.The method runs with overall good yields in a systematic fashion.Some of the parameters briefly studied here could be easily altered resulting in improved reaction yields, especially when using non-linear chain aldehydes.We also presented the synthesis of a series of furan and pyranderived hydroxymethyl-1,4-naphthoquinones obtained by an acid catalyzed Prins cyclization of lapachol and norlapachol and corresponding acetyl derivatives.The 2-hydroxy-3-alkenyl-1,4-naphthoquinone (4) series was in general less cytotoxic than the cyclized furan-and pyrannaphthoquinones 10/11.The β-nor-lapachone (11a) was the most active compound in the MCF-7 tumor cell line.In general, the furan-(β-nor-lapachone) (11) derivatives were more active than pyran-β-lapachone (10), and the presence of the hydroxyl and acetate groups decreased the cytotoxicity of β-lapachone and β-nor-lapachone.Also the 2-(2-ethyl-but-1-enyl)-3-hydroxy- [1,4]naphthoquinone (4j) was shown to be the less cytotoxic for all tumor cell lines tested.

Table 3 .
Variation in conditions for the condensation of lawsone 2 with

Table 4 .
Knoevenagel a Yields after column chromatography; b see Experimental section.

Table 6 .
Cytotoxic activity expressed as IC 50 of compounds 4j