Conversion of Natural Narciclasine to Its C-1 and C-6 Derivatives and Their Antitumor Activity Evaluation: Some Unusual Chemistry of Narciclasine

During the search for a general, efficient route toward the synthesis of C-1 analogues of narciclasine, natural narciclasine was protected and converted to its C-1 enol derivative using a novel semi-synthetic route. Attempted conversion of this material to its triflate in order to conduct cross-coupling at C-1 resulted in a triflate at C-6 that was successfully coupled with several functionalities. Four novel compounds were fully deprotected after seven steps and subjected to evaluation for cytotoxic activity against three cancer cell lines. Only one derivative showed moderate activity compared to that of narciclasine. Spectral and physical data are provided for all new compounds.


Introduction
Narciclasine (1) and pancratistatin (2) (Figure 1), have been among the most studied constituents of the Amaryllidaceae family of alkaloids since their isolation from Narcissus bulbs in 1967 [1], and Pancratium littorale in 1984 [2,3], respectively. Their unique antitumor activity has led not only to many synthetic approaches, but also to investigations of unnatural derivatives with a focus on providing more bioavailable compounds. The syntheses of these natural products and their unnatural derivatives, and the study of their biological activities have been reviewed on many occasions, with the last major reports published in 2016 and 2017 [4][5][6][7][8][9][10][11][12][13][14][15]. Synthetic activity continued, and further creative approaches to both 1 and 2 and their related congeners have been reported since [16][17][18][19].
The minimum pharmacophore for the Amaryllidaceae alkaloids has been suggested [20], and changes in structure are allowed in the "northwest bay region," the space outside C-1 and C-10 indicated in Figure 1, without detriment to biological activity. Many derivatives of pancratistatin have been made and tested [21][22][23][24][25][26]. Some of the unnatural derivatives showed enhanced biological activities, namely the pancratistatin C-1 benzoate reported by Pettit [27] and the C-1 benzoyloxymethyl compounds reported by us [28,29] (both with nanomolar activity). These findings provided further impetus to continue this research with pancratistatin as well as narciclasine. There are not as many unnatural derivatives reported for the latter alkaloid compared to the research volume with pancratistatin. We have published several fully synthetic approaches to C-7 and C-10 analogues, namely compounds 3, 4, and 5, shown in Figure 1 [30][31][32][33]. Only one of these, 10-azanarciclasine (4), displayed activity comparable to the natural product. The minimum pharmacophore for the Amaryllidaceae alkaloids has been suggested [20], and changes in structure are allowed in the "northwest bay region," the space outside C-1 and C-10 indicated in Figure 1, without detriment to biological activity. Many derivatives of pancratistatin have been made and tested [21][22][23][24][25][26]. Some of the unnatural derivatives showed enhanced biological activities, namely the pancratistatin C-1 benzoate reported by Pettit [27] and the C-1 benzoyloxymethyl compounds reported by us [28,29] (both with nanomolar activity). These findings provided further impetus to continue this research with pancratistatin as well as narciclasine. There are not as many unnatural derivatives reported for the latter alkaloid compared to the research volume with pancratistatin. We have published several fully synthetic approaches to C-7 and C-10 analogues, namely compounds 3, 4, and 5, shown in Figure  1 [30][31][32][33]. Only one of these, 10-azanarciclasine (4), displayed activity comparable to the natural product.
The preparation of unnatural derivatives of pancratistatin or narciclasine is arduous and generally requires a unique approach for each derivative. For this reason, we have recently embarked on a program seeking to convert natural narciclasine, available by extraction from daffodil bulbs, to its C-1 derivatives. In this paper, we report the synthesis of several derivatives prepared by cross-coupling, along with the interesting and unique reactivity of narciclasine. The results of initial biological evaluations are provided for the new derivatives.

Initial Approach to the Synthesis of C-1 Analogues
We initiated our approach to C-1 analogues by pursuing established chemistry to functionalize the C-1 position (Scheme 1). The first attempt involved oxidation of protected narciclasine to the corresponding enone 7. We expected to be able to prepare the C-1 vinyl bromide for cross-coupling by bromination, followed by a dehydrobromination process. This approach was met with failure because of the instability of 7 and its isomerization to enamide 8. The preparation of unnatural derivatives of pancratistatin or narciclasine is arduous and generally requires a unique approach for each derivative. For this reason, we have recently embarked on a program seeking to convert natural narciclasine, available by extraction from daffodil bulbs, to its C-1 derivatives. In this paper, we report the synthesis of several derivatives prepared by cross-coupling, along with the interesting and unique reactivity of narciclasine. The results of initial biological evaluations are provided for the new derivatives.

Initial Approach to the Synthesis of C-1 Analogues
We initiated our approach to C-1 analogues by pursuing established chemistry to functionalize the C-1 position (Scheme 1). The first attempt involved oxidation of protected narciclasine to the corresponding enone 7. We expected to be able to prepare the C-1 vinyl bromide for cross-coupling by bromination, followed by a dehydrobromination process. This approach was met with failure because of the instability of 7 and its isomerization to enamide 8.
Epoxidation of narciclasine, as previously reported by Pettit [27], seemed to be another viable option for functionalizing the C-1 position. Initial attempts to open epoxide 10 with various nucleophiles proved to be a more difficult task than originally accounted for. Epoxide 10 was inert to several conditions, and upon treatment with KCN, the reaction yielded a compound that was later identified as a C-1 enol C-2 nitrile (11).
When trying to form the syn-epoxide with N-bromoacetamide (NBA), the reaction afforded the C-1 enol 12a instead of the desired bromohydrin intermediate [34]. Hydrogenation of this intermediate was attempted as it would provide a short conversion of narciclasine to pancratistatin. The enol, or, more accurately, the vinylogous phenol, did not undergo hydrogenation even under high pressure (500 psi). The formation of the stable enol was both unusual and surprising, but it provided the means for potential cross-coupling through pseudohalide groups (either tosylate or triflate) at C-1, and we have focused on this particular strategy.
Narciclasine (1) was protected as an acetonide at C-3/C-4, and the C-2/C-7 hydroxyl groups were acylated to furnish diacetate 9 along with monoacetate 14 in a ratio of 2.2:1, respectively, and trace amounts of triacetate 15 (Scheme 2). In addition, narciclasine was also converted to its peracetate 16 (Scheme 3) in order to compare the steric influence of the acetonide group versus acetates on the chemical events taking place at C-1. Epoxidation of narciclasine, as previously reported by Pettit [27], seemed to be another viable option for functionalizing the C-1 position. Initial attempts to open epoxide 10 with various nucleophiles proved to be a more difficult task than originally accounted for. Epoxide 10 was inert to several conditions, and upon treatment with KCN, the reaction yielded a compound that was later identified as a C-1 enol C-2 nitrile (11).
When trying to form the syn-epoxide with N-bromoacetamide (NBA), the reaction afforded the C-1 enol 12a instead of the desired bromohydrin intermediate [34]. Hydrogenation of this intermediate was attempted as it would provide a short conversion of narciclasine to pancratistatin. The enol, or, more accurately, the vinylogous phenol, did not undergo hydrogenation even under high pressure (500 psi). The formation of the stable enol was both unusual and surprising, but it provided the means for potential cross-coupling through pseudohalide groups (either tosylate or triflate) at C-1, and we have focused on this particular strategy.
Narciclasine (1) was protected as an acetonide at C-3/C-4, and the C-2/C-7 hydroxyl groups were acylated to furnish diacetate 9 along with monoacetate 14 in a ratio of 2.2:1, respectively, and trace amounts of triacetate 15 (Scheme 2). In addition, narciclasine was also converted to its peracetate 16 (Scheme 3) in order to compare the steric influence of the acetonide group versus acetates on the chemical events taking place at C-1. Treatment of either narciclasine peracetate (16) or narciclasine diacetate (9) with NBA (or NBS) provided the C-1 enols 17a and 12a, respectively, along with the C-2 epimers 17b and 12b, as shown in Scheme 3. It is believed that this reaction proceeds by the formation of a bromonium species and its opening by water at the C-1 position. Elimination of the benzylic bromine follows, yielding the C-1 enol ( Figure 2). This is accompanied by the partial epimerization at C-2 through keto-enol tautomerization, facilitated by the generation of HBr and the resulting acidic medium. The suggested mechanism for this transformation is shown in Figure  2. The ratio of the corresponding C-2 epimers appears to depend on the reaction time and the protecting groups on the C-ring. The ratio of enols 17a/17b was 4:1 after five minutes, while the ratio of 12a/12b was 1:2 after 30 min. The reaction proceeds almost instantly and can be quenched within minutes. The faster it is quenched, the lesser the amount of C-2 epimer observed. Treatment of either narciclasine peracetate (16) or narciclasine diacetate (9) with NBA (or NBS) provided the C-1 enols 17a and 12a, respectively, along with the C-2 epimers 17b and 12b, as shown in Scheme 3. It is believed that this reaction proceeds by the formation of a bromonium species and its opening by water at the C-1 position. Elimination of the benzylic bromine follows, yielding the C-1 enol ( Figure 2). This is accompanied by the partial epimerization at C-2 through keto-enol tautomerization, facilitated by the generation of HBr and the resulting acidic medium. The suggested mechanism for this transformation is shown in Figure 2. The ratio of the corresponding C-2 epimers appears to depend on the reaction time and the protecting groups on the C-ring. The ratio of enols 17a/17b was 4:1 after five minutes, while the ratio of 12a/12b was 1:2 after 30 min. The reaction proceeds almost instantly and can be quenched within minutes. The faster it is quenched, the lesser the amount of C-2 epimer observed. ization at C-2 through keto-enol tautomerization, facilitated by the generation of HBr and the resulting acidic medium. The suggested mechanism for this transformation is shown in Figure  2. The ratio of the corresponding C-2 epimers appears to depend on the reaction time and the protecting groups on the C-ring. The ratio of enols 17a/17b was 4:1 after five minutes, while the ratio of 12a/12b was 1:2 after 30 min. The reaction proceeds almost instantly and can be quenched within minutes. The faster it is quenched, the lesser the amount of C-2 epimer observed.

Synthesis of Cross-Coupling Substrates
With the enols 12ab, 17ab in hand, we set out to prepare the C-1 triflate and subjected the resulting compounds to cross-coupling. However, early trials resulted in sluggish reactions with low yields of the desired vinyl triflate. It was believed that such results were observed because of potential competition between the C-1 enol and the ring-B lactam. The mixture of C-1 enols 12a and 12b was treated with acetic anhydride in order to obtain imide 22 that would then be converted to the C-1 triflate (Scheme 4). Surprisingly, upon treating together the C-2 epimers 12a and 12b with acetic anhydride, the mixture equilibrated under these conditions and produced the enol acetate 24 with the correct C-2 stereochemistry rather than the expected diastereomeric mixture of imide 22. The enol acetate 24 was erroneously further functionalized under triflation conditions, as shown on Scheme 4.

Synthesis of Cross-Coupling Substrates
With the enols 12ab, 17ab in hand, we set out to prepare the C-1 triflate and subjected the resulting compounds to cross-coupling. However, early trials resulted in sluggish reactions with low yields of the desired vinyl triflate. It was believed that such results were observed because of potential competition between the C-1 enol and the ring-B lactam. The mixture of C-1 enols 12a and 12b was treated with acetic anhydride in order to obtain imide 22 that would then be converted to the C-1 triflate (Scheme 4). Surprisingly, upon treating together the C-2 epimers 12a and 12b with acetic anhydride, the mixture equilibrated under these conditions and produced the enol acetate 24 with the correct C-2 stereochemistry rather than the expected diastereomeric mixture of imide 22. The enol acetate 24 was erroneously further functionalized under triflation conditions, as shown on Scheme 4. served because of potential competition between the C-1 enol and the ring-B lactam. The mixture of C-1 enols 12a and 12b was treated with acetic anhydride in order to obtain imide 22 that would then be converted to the C-1 triflate (Scheme 4). Surprisingly, upon treating together the C-2 epimers 12a and 12b with acetic anhydride, the mixture equilibrated under these conditions and produced the enol acetate 24 with the correct C-2 stereochemistry rather than the expected diastereomeric mixture of imide 22. The enol acetate 24 was erroneously further functionalized under triflation conditions, as shown on Scheme 4. Three initial cross-coupling partners were selected (phenylboronic acid, phenyl acetylene and potassium vinyl trifluoroborate), and cross-coupling reactions were further performed on what was believed to be "triflate" 23, later determined to be 25. "Triflate" 23 reacted with moderate yields with all three reagents.
During further investigation of the reactivity of the C-1 triflate 23 and in the cross-coupling reactions, we noticed that the experimental data of the expected imide 22 and the product of a Prévost reaction of diacetate 9 were identical. This fact cast serious doubt on the identity of the C-1 cross-coupled products. From 1 H-15 N-NMR HSQC and HMBC experiments, it was determined that the imide 22 had been misassigned, and it is in fact acetate 24. The structure of enol acetate 24 is consistent with the experimental data and would explain how the stereochemistry at C-2 has been "fixed" and "locked", since the presence of an enol acetate Three initial cross-coupling partners were selected (phenylboronic acid, phenyl acetylene and potassium vinyl trifluoroborate), and cross-coupling reactions were further performed on what was believed to be "triflate" 23, later determined to be 25. "Triflate" 23 reacted with moderate yields with all three reagents.
During further investigation of the reactivity of the C-1 triflate 23 and in the crosscoupling reactions, we noticed that the experimental data of the expected imide 22 and the product of a Prévost reaction of diacetate 9 were identical. This fact cast serious doubt on the identity of the C-1 cross-coupled products. From 1 H-15 N-NMR HSQC and HMBC experiments, it was determined that the imide 22 had been misassigned, and it is in fact acetate 24. The structure of enol acetate 24 is consistent with the experimental data and would explain how the stereochemistry at C-2 has been "fixed" and "locked", since the presence of an enol acetate would prevent any further keto-enol tautomerization and the epimerization at C-2, which is only possible while the C-1 functionality is a free enol. We therefore had to re-evaluate the structures of the expected cross-coupling substrate 23.
Additional 1 H-15 N-NMR HSQC and HMBC experiments were performed on 23 and showed the absence of -NH and a large change in δ N chemical shift from 150 ppm (amide) to 315 ppm (pyridine-like nitrogen atom). Upon inspection of IR, the characteristic band for lactams was absent (~1670 cm −1 ). The cross-coupling conditions (Sonogashira, Suzuki) were still working, although in low to moderate yield. We noted that C-N cross-couplings of that nature have not been reported in the literature and that such a scenario would not fit in Buchwald-Hartwig cross-coupling conditions [35,36]. Based on these additional data, we have proposed that the cross-couplings occurred at the C-6 position of triflate 25 (Scheme 4).
The synthesis of the vinyl cross-coupling product (26) provided more data useful to elucidate the cross-coupling process at the C-6 position. Upon analysis with 1 H-15 N-NMR HMBC, two correlation signals were observed: N-H (4a) and N-H (vinyl) (Figure 3).
Molecules 2022, 27, x FOR PEER REVIEW would prevent any further keto-enol tautomerization and the epimerization at C only possible while the C-1 functionality is a free enol. We therefore had to re-e structures of the expected cross-coupling substrate 23.
Additional 1 H-15 N-NMR HSQC and HMBC experiments were performed showed the absence of -NH and a large change in δN chemical shift from 150 ppm 315 ppm (pyridine-like nitrogen atom). Upon inspection of IR, the characteristic b tams was absent (~1670 cm −1 ). The cross-coupling conditions (Sonogashira, Suzuk working, although in low to moderate yield. We noted that C-N cross-couplings o have not been reported in the literature and that such a scenario would not fit in Hartwig cross-coupling conditions [35,36]. Based on these additional data, we hav that the cross-couplings occurred at the C-6 position of triflate 25 (Scheme 4).
The synthesis of the vinyl cross-coupling product (26) provided more data u cidate the cross-coupling process at the C-6 position. Upon analysis with 1 H-15 N-N two correlation signals were observed: N-H (4a) and N-H (vinyl) (Figure 3). A correlation signal between the vinylic proton and nitrogen could only log in two situations: if the vinyl group was directly bound to nitrogen, or if it was latter scenario is the most likely to be the correct one as the coupling constants ag structure. After such findings, further investigations of the structure of the fully products obtained from the cross-couplings performed on the C-6 triflate was th It is important to note the lack of a carbonyl peak from 13 C-NMR in the region ppm on the fully deprotected products. This region is where the lactam carbonyl ciclasine is expected.
Deprotection of the C-1 enol 17a, as well as the C-6 derivatives, was relativ A correlation signal between the vinylic proton and nitrogen could only logically occur in two situations: if the vinyl group was directly bound to nitrogen, or if it was at C-6. The latter scenario is the most likely to be the correct one as the coupling constants agree with the structure. After such findings, further investigations of the structure of the fully deprotected products obtained from the cross-couplings performed on the C-6 triflate was thus required. It is important to note the lack of a carbonyl peak from 13  region around 169 ppm on the fully deprotected products. This region is where the lactam carbonyl peak of narciclasine is expected.
Deprotection of the C-1 enol 17a, as well as the C-6 derivatives, was relatively straightforward. In the case of 17a, treatment with NaOMe in methanol did not promote the formation of product 31, so it was decided to use concentrated HCl in THF. After quenching the reaction with solid NaHCO 3, it yielded the desired deprotected product 31. In the case of the C-6 derivatives, treatment with potassium carbonate in methanol followed immediately by treatment with 3 M HCl gave the desired deprotected products, as shown in Scheme 5. However, in the case of compound 26, the standard conditions shown in Scheme 5 promoted cyclization of the vinyl group (Scheme 6). It is believed that the deprotection of 26 produces 32, an α,β-unsaturated imine that could undergo intramolecular Michael addition with the phenol to yield dihydropyran 33.   However, in the case of compound 26, the standard conditions shown in Scheme 5 pr moted cyclization of the vinyl group (Scheme 6). It is believed that the deprotection of 2 produces 32, an α,β-unsaturated imine that could undergo intramolecular Michael additio with the phenol to yield dihydropyran 33.  The four deprotected compounds 29, 30, 31 and 33 (Figure 4), together with narciclasine (1) and pancratistatin (2) controls, were subjected to biological evaluation. The results are shown in Table 1.    We used three in vitro cancer models-neuroblastoma BE(2)-C, lung squamous cell carcinoma H157 and lung adenocarcinoma cells A549 (Table 1, Supplementary Figure S1). As expected [33], the narciclasine and pancratistatin controls showed good nanomolar activity against all cancer cell lines. Out of the synthesized compounds, only phenyl derivative 29 showed moderate single digit micromolar activity against BE(2)-C and H157 cells. Phenylacetylene derivative 30 also displayed activity, albeit of significantly diminished potency, and most likely through a different mode of action due to marked structural differences. Surprisingly, enol 31 failed to register any pronounced activity against any of the cell lines. Although structurally, 31 resembles both 1 and 2, the stability of such enol functionality in a biological medium is an important consideration. Overall, the B-ring lactam appears to be an important part of the cytotoxic pharmacophore and the removal of the lactamic carbonyl has a deleterious effect on activity, although it does not abolish it altogether.

Materials and Methods
All solvents were distilled and kept dry before usage. Unless otherwise stated, all reactions were done in an inert atmosphere (Ar or N 2 ). All reagents were obtained from commercial sources. Nuclear magnetic resonance (NMR) analyses were performed on Bruker Avance AV 300, Bruker Avance III HD 400 and Bruker Avance AV 600 digital NMR spectrometers, running Topspin 2.1 and 3.5 software. The probes are furnished with VT (variable temperature) and gradient equipment. Chemical shifts are given in δ, relative integral, multiplicity (singlet (s), doublet (d), triplet (t), quartet (q), multiplet (m)) and coupling constants (J) in Hz. Melting points (m.p.) were measured in a capillary apparatus. Mass spectra (HRMS) measurements were determined on an LTQ Orbitrap XL. The molecular mass-associated ion was measured by electron ionization, electrospray ionization or fast atom bombardment. Infrared (IR) spectra were recorded on an FT-IR spectrophotometer as neat and are reported in wave numbers (cm −1 ) and intensity (broad (br), strong (s), medium (m), weak (w)). Column chromatography was performed on flash grade 60 silica gel. Thin-layer chromatography (TLC) was performed on silica gel 60 F254-coated aluminum sheets. TLC plates were visualized using UV and stained with iodine, cerium ammonium molybdate (CAM), KMnO 4 solutions, FeCl 3 solutions, ninhydrin solutions or 2,4-dinitrophenylhydrazine (2,4-DNP) solutions.
Diacetate 9 (250 mg, 0.58 mmol) was solubilized in DCM (15 mL) and an equal volume of phosphate buffer (pH = 8.0) was added to the solution. The biphasic system was cooled to 0 • C, recrystallized m-CPBA (350 mg, 2.02 mmol) was added, and the reaction was left stirring overnight. The reaction was quenched with a saturated solution of Na 2 S 2 O 3 (15 mL), transferred to a separatory funnel and further washed with NaHCO 3 and extracted with DCM (3 × 15 mL). The combined organic layers were dried over MgSO 4 , concentrated, and adsorbed on 10% deactivated silica gel. The product was purified through flash column chromatography (n-hex:EA 1:1). Title compound was obtained (130 mg, 0.29 mmol, 50% yield) as a white solid.  13 [27].
Diacetate-acetonide narciclasine 9 (287 mg, 0.66 mmol) was solubilized in a 2.3:1 mixture of THF:H2O (7 mL/3 mL) and cooled to 0 • C in an ice-salt bath, and the flask was covered with aluminum foil to protect it from light. Recrystallized NBS (142 mg, 0.80 mmol) was added in the solution and the colorless solution turned yellow. The reaction was left stirring in the dark and at 0 • C for 30 min and quenched with 1 mL of a saturated solution of Na 2 S 2 O 3 . The yellow color should fade, and the reaction mixture is concentrated on a rotary evaporator to remove THF, while the remaining solid product with the aqueous solution is diluted with DCM (10 mL) and extracted with DCM (3 × 10 mL). The combined organic layers are dried over MgSO 4 , filtered, and concentrated to yield 200 mg (0.44 mmol, 66% yield) of a mixture of C-1 enol (C-2 epimer) 12a and C-1 enol 12b as in a 1:2 ratio. After purification through flash column chromatography (DCM:MeOH:Acetone 200:1:1 to 50:1:1), 12a was obtained as a white foam and 12b as a white film. Note: these compounds epimerize in solution.
Narciclasine tetraacetate 16 (50 mg, 0.1 mmol) was dissolved in a mixture of 3:1 THF:H 2 O (3 mL THF, 1 mL water) in an RBF under argon atmosphere. The reaction mixture was then cooled down to 0 • C using an ice bath. When at 0 • C, NBA (35 mg, 0.25 mmol) was added in one portion. The reaction mixture then turned yellow and was left to stir for 5 min.
The reaction mixture was worked up by the addition of H 2 O, followed by liquid-liquid extraction. The aqueous layer was extracted with 3 × 10 mL DCM. The combined organics were then extracted with 10 mL of brine and dried over Na 2 SO 4 . The crude mixture was purified using flash column chromatography (1:1 Hexanes:EA). The resulting product was a beige powder with a 4:1 ratio (17a:17b) (35 mg, 0.07 mmol, 70%).
Method A: Enol acetate 24 (30 mg, 0.06 mmol) was charged into a flame-dried Schlenk flask, solubilized in dry 1,2-DME (1 mL) and cooled to −5 • C in an ice-salt bath. Recrystallized 18-crown-6 ether (catalytic amount) and KH (30% in oil, one drop) were added at −5 • C under argon, and the reaction mixture turned yellow. A solution of TfCl (3.63 M in DCM, 50 µL, 0.18 mmol) was added to the reaction mixture, and the color changed to a bright yellow. The mixture was left stirring for 2 h in the ice-salt bath. The reaction progress was observed through TLC (DCM:MeOH 30:1), and new spots were observed but starting material was still present. The reaction mixture was diluted with DCM and filtered through a plug of celite; the organic phase was collected and concentrated in the rotavap. The crude product was purified by flash column chromatography with 10% deactivated silica gel (DCM:MeOH 200:1). Finally, 6 mg of title compound (0.03 mmol, 16% yield) was obtained as a film, and 18 mg of starting material was recovered.
Method B: NBS (124 mg, 0.70 mmol) was added to a stirred solution of NAR acetonide diacetate 9 (150 mg, 0.35 mmol) in THF/HOAc (2 mL/2 mL). The reaction mixture was stirred for 20 min, and then it was carefully diluted with EtOAc/satd. NaHCO 3 (15 mL/15 mL). When the gas evolution ceased, the layers were separated. The aqueous layer was further extracted with EtOAc (2 × 10 mL). The combined organic layers were washed with 5% aq. Na 2 S 2 O 3 (2 × 10 mL), then brine (10 mL), dried over anhydrous MgSO 4 , filtered and concentrated to afford a crude solid residue of enol acetate that was used as is in the next step. To the crude product from the previous step dissolved in DCM (3 mL) at 0 • C was added pyridine (226 mL, 2.80 mmol) and Tf 2 O (235 mL, 1.40 mmol). The reaction mixture was stirred at this temperature for 3 h, after which it was quenched by the addition of 10% aq. citric acid solution (5 mL). The reaction mixture was diluted with DCM (20 mL) and water (5 mL). The layers were separated, and the organic layer was further washed with 10% aq. citric acid solution (2 × 5 mL) and brine (5 mL). The organic layer was dried over MgSO 4 , filtered, and concentrated to afford a residue that was chromatographed on silica gel using hexanes/EtOAc as an eluent (2:1 to 1:1) to afford the enol triflate product 25 as a yellow film (135 mg, 60.3% yield).
To a stirred solution of triflate 25 (60 mg, 0.096 mmol) in dioxane (1.0 mL), phenylboronic acid (17 mg, 0.139 mmol), Pd(PPh 3 ) 4 (8 mg, 0.006 mmol), and Net 3 (20 µL, 0.143 mmol) were added. The reaction mixture was heated to 85 • C for 2 h. It was then cooled to RT and filtered through a plug of celite and washed with DCM. The organic filtrate was concentrated to afford a residue that was adsorbed on 10% deactivated silica gel and purified by flash column chromatography using DCM:MeOH as eluent (220:1 to 200:1) to afford the phenyl coupling as a white solid (32 mg, 60% yield).  13

Conclusions
Throughout our attempts to synthesize C-1 derivatives of narciclasine, whose biological activity could be compared with those of the corresponding C-1 unnatural derivatives of pancratistatin, four novel C-1 and C-6 analogues were obtained. It was observed that installation of the C-1 enol moiety enhanced the reactivity of the B-ring lactam, so that triflation reactions yielded an imino triflate/triflyl imidate and subsequent cross-couplings occurred at the C-6 position. The four new analogues were subjected to biological testing and the SAR database was expanded with results that indicate that the removal of the lactamic carbonyl has a deleterious effect on activity, although it does not abolish it altogether. Our efforts will continue in order to design a new route taking such activity into consideration and to hopefully synthesize C-1 analogues of narciclasine that could be directly compared to similar pancratistatin derivatives developed and evaluated previously [8,9]. We will report any new results in due course.