Radical reactions with 3H-quinazolin-4-ones: synthesis of deoxyvasicinone, mackinazolinone, luotonin A, rutaecarpine and tryptanthrin

Abstract

Alkyl, aryl, heteroaryl and acyl radicals have been cyclised onto the 2-position of 3H-quinazolin-4-one. The side chains containing the radical precursors were attached to the nitrogen atom in the 3-position. The cyclisations take place by aromatic homolytic substitution hence retain the aromaticity of the 3H-quinazolin-4-one ring. The highest yields were obtained using hexamethylditin to facilitate cyclisation rather than reduction without cyclisation. The alkaloids deoxyvasicinone 2, mackinazolinone 3, tryptanthrin 4, luotonin A 5 and rutaecarpine 8 were synthesised by radical cyclisation onto 3H-quinazolin-4-one.

---

The 3H-quinazolin-4-one ring system is important to the biological activity of both naturally occurring alkaloids, biosynthesised from anthranilic acid, and pharmaceuticals. The alkaloids include vasicinone 1 and deoxyvasicinone 2,¹ mackinazolinone 3,² tryptanthrin 4,³ luotonins A 5, B 6 and E 7⁴ and rutaecarpine 8.⁵ 3H-Quinazolin-4-one alkaloids have been recently reviewed.⁶ All the 3H-quinazolin-4-one natural products have interesting biological activity and have therefore been extensively investigated for useful pharmaceutical activity. The 3H-quinazolin-4-one ring is regarded as a ‘privileged structure’ in combinatorial synthesis.⁷ These are structures which represent molecules that are capable of binding at multiple sites with high affinity and facilitate more rapid discovery of useful medicinally active compounds.⁷

Our study involved the development of protocols involving radical cyclisation for the synthesis of polycyclic 3H-quinazolin-3-ones (Scheme 1). The protocols have also been used for the synthesis of novel polycyclic quinazolinones including the natural products deoxyvasicinone 2, mackinazolinone 3, tryptanthrin 4, luotonin A 5 and rutaecarpine 8.

Scheme 1 General synthetic protocol.

Radical cyclisation onto heteroarenes has been developed in recent years to considerable advantage for the synthesis of novel polycyclic heteroarenes. Examples of these cyclisations include: a. alkyl radicals onto pyrroles,^8,9 imidazoles,⁸ pyrazoles,¹⁰ indoles,^9,11,12 1,2,3-triazoles,¹³ pyridinium salts,¹⁴ and quinolones;¹⁵ b. acyl radicals onto pyrroles,¹⁶ quinolines,¹⁷ pyridines¹⁸ and arenes;¹⁹ c. aryl radicals onto indoles,^20,21 pyrroles,⁹ pyridones,²² and 5-amino- and 5-hydroxyuracils,²³ 2-quinolones,²⁴ quinolines²⁵ and pyridines.²⁶

All of the above cyclisations are ‘oxidative’ i.e. the intermediate π-radicals are not reduced by triorganometal hydrides [e.g. tributyltin hydyride (Bu₃SnH)] as normally observed for these reagents. The cyclisations proceed by aromatic homolytic substitution with abstraction of hydrogen in a rearomatisation process. Aromatic homolytic substitution has been recently reviewed²⁷ and the mechanism of Bu₃SnH mediated ‘oxidative’ cyclisation elaborated.²⁸ The pyrimidin-4-one ring of the quinazolin-4-ones has some aromaticity and therefore aromatic homolytic substitution could be predicted, and was observed in our studies, as shown in Scheme 1 (9 to 11via the π-radical 10). However, the lower aromaticity could favour reductive cyclisation in which the intermediate π-radical 10 is intercepted by reagents such as Bu₃SnH. Our prediction that radical cyclisation onto the quinazolin-4-one ring would be ‘oxidative’ was supported by the ‘oxidative’ radical cyclisation onto related ring systems, e.g. pyrimidine-2,4-diones,^23,24 quinolinones¹⁵ and pyridinones.^22,29,30 However, the lower aromaticity in the pyrrole ring of indole, facilitates both reductive and oxidative cyclisation depending on the conditions.^11,12,31 The radical intermediate 10, whether a π-radical or not, is still a strongly stabilised anilyl radical and therefore the rate of reduction by Bu₃SnH to yield 12 is probably too low to be competitive with loss of hydrogen to yield the ‘oxidised’ product. The reactions could also be regarded as exo-cyclisations onto imines which are well known.

We used two general methodologies to synthesise the radical precursors. Firstly, heteroarenes containing an NH group facilitate N-alkylation and provide a suitable synthetic route to radical precursors for cyclisation. The radical leaving group is introduced as part of the N-alkyl substituent. Secondly, there is a wide variety of protocols for the synthesis of 3-substituted quinazolinones using ring synthesis from anthranilic acid derivatives, which were employed as required.

Aryl radical cyclisation

Aryl radicals have been successfully cyclised onto pyrroles, indoles and pyrazoles.³² Aryl radicals are very reactive and the most likely to cyclise onto quinazolinones. Therefore, suitable radical precursors for 5- and 6-exo cyclisation were prepared (Scheme 2). The 5-membered ring cyclisation with the precursor 16 using Bu₃SnH gave only the reduced uncyclised product 19 (Scheme 3). Even syringe pump addition of Bu₃SnH, with either Et₃B (r.t.) or AIBN (110 °C) as initiators, gave only 19. We considered that the lack of cyclisation could be due to 1,5-hydrogen abstraction from the 2-position of the quinazolinone as shown in Scheme 3 (route b: 18 to 21). This possibility was eliminated by repeating the reaction with Bu₃SnD which gave only deuteriation on the aryl radical position, and none on the 2-position of the quinazolinone, and these results indicated that the rate of cyclisation was not favourable compared to reduction with Bu₃SnH. Therefore, hexamethylditin [(Me₃Sn)₂] was used so that the intermediate radical (cf.9 in Scheme 1) was not reduced. The reaction was repeated with (Me₃Sn)₂ which resulted in a small amount of the cyclised product 20 (18%) as well as uncyclised 19 (65%) as the major compound.

Scheme 2 Reagents and conditions: i, tert-BuOK, DMF, 63% (16 from 14), 38% (17 from 15).

Scheme 3 Reagents and conditions: i, Et₃B, PhMe, r.t., Bu₃SnH (fast addition) 96% (19), slow addition, 30% (19); AIBN, Bu₃SnH (slow addition), PhMe, reflux, 45% (19); Et₃B, TTMSS, 35% (19); (Me₃Sn)₂, tert-BuPh, reflux, 18% (20), 65% (19).

The analogous six-membered ring aryl radical cyclisation using precursor 17 gave an excellent yield of the cyclised product 22 (92%) with no reduction using (Me₃Sn)₂ (Scheme 4). The reaction with Bu₃SnH again gave largely the uncyclised reduced product 23 (55%) but did also yield a small amount of the cyclised product 22 (8%). We have observed before that 5-ring cyclisation of radicals onto heteroarenes is difficult due to strain whereas 6-ring cyclisation is more favourable.^8,10,16,32 Therefore, the higher yields of cyclisation from 17 relative to 16 are expected. The results indicate that the oxidative route is dominant (i.e. loss of hydrogen from the π-radical intermediate 10) for cyclisation onto quinazolinones as observed for other heteroarenes.

Scheme 4 Reagents and conditions: i, (Me₃Sn)₂, tert-BuPh, reflux, 92% (22), 0% (23); Et₃B, PhMe, r.t., Bu₃SnH (fast addition), 8% (22), 55% (23).

3H-4-Oxoquinazolin-2-yl radical cyclisation

As shown in Scheme 5 we envisaged an alternative route to polycyclic quinazolinones by cyclisation of 3H-4-oxoquinazolin-2-yl radicals onto pendant side chains attached to the 3-position instead of cyclisation of side chain radicals onto the quinazolinone moiety. 3H-4-Oxoquinazolin-2-yl radicals,³³ generated photolytically from 3-(but-1-en-4-yl)-3H-quinazolin-4-one, have been reported to undergo 5-exo cyclisation onto the pendant alkene.^33a The procedure however was not suitable for synthetic application. During our studies, a synthesis of luotonin A analogues using 3H-4-oxoquinazolin-2-yl radicals was reported.^33b

Scheme 5 Reagents and conditions: i, Br₂, EtOH, 50% (26); ii, NaH, DMF, BnBr, 46% (27); NaI, PhCH₂CH₂Br, 6% (28); iii, 27: Bu₃SnH, Et₃B, slow addition, 62% (19); Bu₃GeH, Et₃B, slow addition, 41% (19); (Me₃Sn)₂, tert-BuPh, reflux, 41% (19), 27% (20); 28: (Me₃Sn)₂, tert-BuPh, reflux, 0% (23), 97% (22).

The plan of our protocol was to alkylate 2-bromo-3H-quinazolin-4-one to provide suitable radical precursors (Scheme 5). 2-Bromo-3H-quinazolin-4-one 26 was prepared by an adapted literature procedure from 2-mercapto-3H-quinazolin-4-one 24.³⁴ Alkylations with activated halides, e.g. benzyl bromide, allyl bromide and methyl iodide were successful but alkylations with unactivated halides were very poor or failed. Alkylation with activated propargyl halides has also recently been reported.^33b A maximum yield (6%) of alkylation was obtained with 2-phenylethyl bromide and alkylation with 3-phenylpropyl bromide failed. Alkylation with the corresponding triflates failed to give improved results. We suggest that the anion of 2-bromo-3H-4-oxoquinazoline is too stabilised and hence not nucleophilic enough to react with unactivated halides.

Various protocols were attempted to circumvent the alkylation problem. For instance, the successful bromination suggested that other 2-thioxo-2,3-dihydro-1H-quinazolin-4-ones with the 3-alkyl side chain in place could also be converted to the corresponding 2-bromo compounds. In order to test this hypothesis, 2-(phenylethyl)-2-thioxo-2,3-dihydro-1H-quinazolin-4-one 30 was prepared by ring synthesis using the reaction between 2-(methoxycarbonyl)phenyl isothiocyanate and amines.³⁵ The structure of 30 was confirmed by X-ray crystallography (Fig. 1). However, the bromination procedure yielded only the corresponding disulfide instead of the 2-bromoquinazolinone 28.

Fig. 1 X-Ray structure of 3-(2-phenylethyl)-2-thioxo-2,3-dihydro-1H-quinazolin-4-one 30 with atom labelling.

The mechanism of these brominations is unknown. The most likely intermediate is the sulfenyl bromide 25 which can either lose a proton and undergo substitution with bromide, or undergo direct substitution with bromide. The intermediate sulfenyl bromide from 2-(phenylethyl)-2-thioxo-2,3-dihydro-1H-quinazolin-4-one cannot lose a proton which is possibly significant in directing the reaction towards disulfide formation. Examples from the literature show bromoimine formation for thioureas with N,N′-substitution (i.e. can lose a proton)³⁶ and disulfides and sulfides from thioamides.³⁷

We carried out radical cyclisations with the precursors 27 and 28 (Scheme 5). As observed for the cyclisation of precursor 16, the 5-membered ring cyclisation from 27 was unfavourable with both Bu₃SnH and Bu₃GeH and yielded only the reduced uncyclised product 19. Radical abstraction of hydrogen from Bu₃GeH is 20 times slower than from Bu₃SnH which would favour cyclisation over reduction,³⁸ but still only reduction was observed from the intermediate 29a. When (Me₃Sn)₂ was used, a moderate yield of cyclisation product (27%) was obtained but reduction was still the major route (41%). The 6-membered ring cyclisation from precursor 28via29b gave a quantitative yield of cyclisation to 22 with none of the reduced compound 23 formed. This cyclisation could either proceed by 5-exo cyclisation followed by a neophyl rearrangement or directly by 6-endo cyclisation. The results show that radical cyclisation can be used onto, or from, the quinazolinone moiety, but that a useful method of synthesis of the 2-bromoquinazolinones is still required. Addition of 3H-4-oxoquinazolin-2-yl radicals onto isonitriles has also recently been reported.^33b

Alkyl radical cyclisation

Several methods have been used for the synthesis of both deoxyvasicinone 2 and mackinazolinone 3 but none involving radicals.³⁹ Our protocol using cyclisation of an alkyl radical onto 3H-quinazolin-4-one rings was aimed at the synthesis of both natural products (Scheme 6). Routine alkylation yielded the required radical precursors 31a,b. An initial study with Bu₃SnH using precursor 31a yielded only reduced uncyclised product 33a (78%). The rate of cyclisation of the alkyl radical is obviously slow, and it is intercepted by Bu₃SnH to yield the respective reduced uncyclised product 33a. Therefore, (Me₃Sn)₂ was used to facilitate cyclisation which gave moderate yields of deoxyvasicinone 2 and mackinazolinone 3.

Scheme 6 Reagents and conditions: i, NaH, DMF, a. 1-chloro-3-iodopropane, 51%; b. 1-chloro-4-iodobutane, 65%; ii, NaI, acetone, reflux, 53% (31a), 67% (31b); iii, (Me₃Sn)₂, tert-BuPh, reflux, hν: a. 31a gave 20% (2), 13% (32a) and 6% (33a); b. 31b gave 30% (3), 23% (32b) and 0% (33b); Et₃B (20 equiv.), tert-BuPh, air (yields by ¹H NMR analysis): a. 31a gave 40% (2), 10% (32a) and 30% (33a); b. 32b gave 61% (3), 23% (32b) and 0% (33b).

The use of Et₃B mediated reactions gave better yields of both 2 and 3 (Scheme 6). We suggest that the ethyl radical generated from the reaction between Et₃B and oxygen is able to abstract iodine from the radical precursors 31a,b to yield alkyl radicals and ethyl iodide, which is lost from the reaction. The hydrogen in the π-radical intermediates (cf.10) is also most likely abstracted by ethyl radicals to facilitate rearomatisation.⁴⁰ Unusually, some cyclised reduced material 32a,b was obtained suggesting disproportionation was also taking place. GCMS analysis of the crude product from the Et₃B facilitated reaction with 31a indicated traces (ca. 5%) of 2-ethyl-3-propyl-3H-quinazolin-4-one and 3-pentyl-3H-quinazolin-4-one indicating addition of ethyl radicals to intermediates. No uncyclised reduced material 33b was observed for the 6-ring cyclisation of 31b. Again, this provides evidence that 5-ring cyclisation onto heteroarenes is strained and that 6-ring cyclisation (84% of cyclised products) is more favourable.

Heteroaryl radical cyclisation

During our studies, Pd(0) catalysed cyclisations were reported for 2-bromoindole⁴¹ and 2-bromoquinoline^41,42 moieties onto the 2-position of quinazolin-4-one to yield rutaecarpine in poor yield (24% maximum) and luotonin A in good yield (86%) respectively.

Luotonin A 5. The luotonins make up a group of pyrroloquinazolino-quinoline alkaloids of which the pentacyclic luotonins A 5, B 6 and E 7 are of most interest. The luotonins are used in traditional Chinese medicine and are reported to exhibit activity against a range of ailments including rheumatism, inflammation, influenza, hepatitis and leukemia.⁴³ Luotonin A has been reported to show activity as an antitumour compound and is an inhibitor for human DNA topoisomerase I.⁴³ The compounds were isolated some ten years ago⁴ and luotonin A has been a popular synthetic target.^43,44 Two radical syntheses of luotonin A have been reported; a radical domino cyclisation reaction using vinyl radical cyclisation onto nitriles⁴⁵ and a bimolecular reaction involving radical addition onto isonitriles.^33b

The synthesis of luotonin A 5 using 2-quinolinyl radicals was carried out as a further example of cyclisation onto the quinazolinone moiety (Scheme 7). The use of 2-quinolinyl radicals in cyclisation has two literature precedents; synthesis of camptothecin using cyclisation of 2-quinolinyl radicals onto a pyridone moiety²⁹ and synthesis of 10,11-methylene-14-azaxamptothecin and 14-azacamptothecin using cyclisation of 2-quinolinyl radicals onto 3H-pyrimidin-4-one moieties.³⁰ These cyclisations also proceeded by ‘oxidative cyclisation’ as observed for the 3H-quinazolin-4-ones.

Scheme 7 Reagents and conditions: i, (Me₃Sn)₂, tert-BuPh, reflux, hν: 51% (5), 15% (36); Et₃B (20 equiv.), PhMe, r.t., air: Bu₃SnH (fast addition), 0% (5), 53% (36) Bu₃SnH (slow addition), 14% (5), 32% (36); Bu₃GeH (fast addition), 18% (5), 11% (36).

The required starting material 34 was prepared by a literature procedure from 2-chloroquinoline-2-carbaldehyde⁴⁶ and reacted under several radical conditions. Unusually, the yield from 5-exo cyclisation was significant. The use of (Me₃Sn)₂ gave a reasonable yield (53%) of luotonin A 5 but also some of the reduced product 36. The yield was high in comparison to the 5-exo cyclisation of the phenyl analogue 16. Luotonin A was even obtained with Bu₃SnH and Bu₃GeH, clearly indicating that the cyclisation is more favourable for 2-quinolinyl radicals as compared to phenyl radicals. The yield was not as high as the Pd(0)-catalysed cyclisation⁴² but further optimisation could improve the yield.

Rutaecarpine 8. Rutaecarpine 8 has been synthesised by a wide variety of protocols but none involving radicals.⁴⁷ Our synthesis of rutaecarpine uses indol-2-yl radical cyclisation onto the quinazolinone motif as a further example of natural product synthesis using the general protocol (Scheme 8). Indol-2-yl radicals have been previously used in radical cyclisation onto pendant 1-ω-alkenes⁴⁸ and -arenes⁴⁹ and in H-translocation reactions towards the synthesis of mitomycin.⁵⁰

Scheme 8 Reagents and conditions: i, PBr₃, DCM, 56% (38); ii, tert-BuOK, DMF, 3H-quinazolin-4-one, 39% (39); iii, NBS, DCM, 0 °C, 30 min, 38% (40); iv, (Me₃Sn)₂, tert-BuPh, reflux, hν: 55% (8), 0% (39); Et₃B (20 equiv.), PhMe. r.t., air: Bu₃SnH (slow addition), 15% (8), 57% (39).

The preparation of the 2-bromoindole part of the precursor proved troublesome even with protection of the side chain and the NH. The use of NBS has been reported in the literature but the reaction is very sensitive, and can be adversely affected by the NH and side chain groups.^48,49,51 We finally used an adapted literature procedure⁴¹ whereby the 3H-quinazolin-4-one moiety was used as the ‘protective group’ for the indole side chain (i.e.39). The bromination was rapid and short reaction times and low temperature gave the best yields of the 2-bromoindole precursor 40. As reported in the literature for 2-bromoindoles,^48,49,51,52 decomposition of the product was a problem until it was purified. The free indole-NH did not interfere and was therefore not protected. A blank reaction between NBS and 3-methylquinazolin-4-one gave no reaction after two days indicating that in the NBS reaction with 39, bromination of the quinazolinone ring is unlikely to be the cause of the decomposition.

3-[2-(1H-Indol-3-yl)ethyl]-4(3H)-quinazolinone 39 was prepared by alkylation of 4(3H)-quinazolinone with 3-(2-bromoethyl)-1H-indole 38 which was prepared by bromination of tryptophol 37.

Radical cyclisation of the precursor 40 gave the predicted 6-exo cyclisation of the intermediate radical 41 to yield rutaecarpine 8. Only cyclisation was obtained when (Me₃Sn)₂ was used and even reductive conditions with Bu₃SnH yielded a small amount of cyclisation product. The 6-exo cyclisation yields were similar to the equivalent cyclisation with aryl radicals (see Scheme 4). Longer reactions times led to decomposition.

Acyl radicals—synthesis of tryptanthrin

Tryptanthrin has been synthesised by a range of protocols.^3,6,53

We prepared an authentic sample of tryptanthrin 4 by a literature procedure in order to obtain full spectroscopic data for comparison.⁵³

Aromatic acyl radical cyclisation has recently been shown to be a useful synthetic technique.^17–19,54 Although 5-membered ring cyclisation was reported to be unsuccessful,¹⁹ we have shown that 5-ring cyclisation onto the 3H-quinazolin-4-one moeity was possible (e.g. the luotonin synthesis). Therefore, we carried out the syntheses as shown in Scheme 9. The aryl-CO bond in the intermediate acyl radical 44 is strong enough to avoid decarbonylation which is a rapid reaction for alkyl-CO radicals.⁵⁵ However, it is possible that we failed to isolate products resulting from CO loss.

Scheme 9 Reagents and conditions: i, Bu₃P, PhSeSePh, DCM, 73% (42); ii, LiOH, EtOH, H₂O; iii, Bu₃P, PhSeSePh, DCM, 68% (47); iv, (Me₃Sn)₂, tert-BuPh, reflux, hν, 3% (4); PhH, r.t., hν, 12 h, 15% (4); PhH, reflux, hν, 10 h, 11% (4); AIBN, PhH, reflux, hν, 0% (4); v, (Me₃Sn)₂, PhH, reflux, hν, 39% (50).

The starting material 42 was prepared in one step by a literature procedure⁵⁶ and converted to the acyl selenide 43 by standard procedures.¹⁶ Several conditions were used based on literature reports.^17–19,54 The highest yield (15%) was obtained by photolysis at r.t. Although the yield is poor, we believe this is the first example of a 5-exo acyl radical cyclisation onto a heterocycle. When AIBN was added to the reaction, an intractable mixture was obtained. Heating under reflux was not required and UV photolysis alone was enough to facilitate the reaction, presumably by homolysis of the carbonyl-SePh bond. In a blank reaction, heating under reflux in benzene yielded only unaltered starting material after 24 h. The mechanism is unclear, other than 5-exo acyl radical cyclisation followed by hydrogen abstraction from the resulting π-radical intermediate. Large amounts of diphenyl diselenide were isolated indicating CO–Se bond homolysis.

We also investigated the 6-membered ring cyclisation because these had proved to be more successful than the 5-ring cyclisations (Scheme 9). The acyl radical precursor 47 was prepared by standard procedures. The ester 45 was synthesised by alkylation of 3H-quinazolin-4-one with methyl 2-(bromomethyl)-benzoate. Cyclisation using Et₃B only, AIBN only, UV photolysis only and slow addition of TTMSS with Et₃B as initiator all gave intractable mixtures. However, reaction with (Me₃Sn)₂ gave a moderate yield of the unusual product 50 along with an intractable mixture of other products. The structure of the hydroxylated product 50 was confirmed by X-ray crystallography (Fig. 2). The product indicated 6-ring cyclisation as expected but that the product 48 was unstable to the reaction conditions and was readily oxidised introducing an OH group onto the newly formed ring.

Fig. 2 X-Ray structure of 11-hydroxy-11H-isoquinolino[3,2-b]quinazolin-6,13-dione 50 with atom labelling.

Analogous reactions with 2-indolyl acyl radicals have yielded a range of unexpected products which included a ‘quinone’ product caused by oxidation of the benzylic methylene during the reaction.^17,18 2-Cyanoprop-2-yl radicals from the breakdown of AIBN were proposed as the abstracting radicals. In our reaction, no AIBN is present so we suggest that any of the intermediate radicals (e.g. intermediate acyl or π-radical, PhSe˙) can abstract the benzylic hydrogen, which is a favourable process, to form a stable radical intermediate 49. Although air was excluded, traces of oxygen are likely to be present and to react rapidly with trimethylstannyl radicals to form a peroxyl radical (Me₃SnOO˙).^45,57 Combination of the peroxyl radical with the benzylic radical 49 would yield a peroxide which could easily break down to yield 50. Alternatively, a standard auto-oxidation mechanism with the traces of oxygen present could explain this unusual product.

Conclusions

Our results show that radical cyclisation on the quinazolinone moiety can be used for synthesis. The results show that radical cyclisation is also favourable for radicals centred on the 2-position of the quinazolinone moiety but that a better method for the synthesis of 2-bromoquinazolinones is still required.

Experimental

General

Commercial dry solvents were used in all reactions except for light petroleum and ethyl acetate which were distilled from CaCl₂ and dichloromethane which was distilled from CaH₂. Light petroleum refers to the bp 40–60 °C fraction. Sodium hydride was obtained as 60% dispersion in oil. A 2.5 M solution of n-butyl lithium in hexane was used. A solution of Et₃B in hexane (1.0 M) was used in all cases. Melting points were determined on a Stuart Scientific SMP3 melting point apparatus and are uncorrected. Elemental analyses were determined on a Perkin Elmer 2400 CHN Elemental Analyser. Infrared spectra were recorded on a Perkin-Elmer Paragon 1000 FT-IR spectrophotometer on NaCl plates. ¹H (400 MHz) and ¹³C (100 MHz) NMR spectra were recorded on a Bruker DPX-400 instrument, or ¹H (250 MHz) spectra were obtained using a Bruker AC-250 spectrometer. Spectra were obtained from solutions in CDCl₃ with TMS as the internal standard for ¹H NMR spectra and deuteriochloroform as the standard for ¹³C NMR spectra unless otherwise specified. Chemical shifts are given in parts per million (ppm) and J values in hertz (Hz). Assignments were made using a combination of COSY and HMQC analysis. Mass spectra were recorded on a Jeol SX102 high resolution mass spectrometer or were run by the EPSRC MS Service at the University of Wales, Swansea. GCMS was carried out on a Fisons 8000 series mass spectrometer using a 15 m × 0.25 mm DB-5 GC column. TLC using silica as absorbent was carried out with aluminium backed plates coated with silica gel (Merck Kieselgel 60 F₂₅₄), and TLC using alumina as absorbent was carried out with aluminium backed plates coated with neutral aluminium oxide (Merck 150 F₂₅₄, TypeT). Silica gel (Merck Kieselgel 60 H silica) was used for column chromatography unless otherwise specified. Column chromatography using alumina was carried out with Aldrich aluminium oxide, activated neutral, Brockmann 1, STD Grade, 150 mesh size. Prep-TLC was carried out using aluminium oxide (Merck 60 PF₂₅₄, Type E). Bu₃GeH was prepared from GeCl₄ by a known procedure.³⁸

General procedure for alkylation of 3H-quinazolin-4-ones. 3-(2-Iodobenzyl)-3H-quinazolin-4-one 16. Potassium tert-butoxide (1.35 g, 12 mmol) was added to 3H-quinazolin-4-one 13 (1.17 g, 8 mmol) in dry DMF (50 cm³) and the mixture stirred for 1 h under an atmosphere of nitrogen. 2-Iodobenzyl bromide 14 (2.84 g, 9.6 mmol) was added and the reaction mixture stirred for a further 16 h. The mixture was diluted with DCM and washed with H₂O and brine. The organic layer was dried and evaporated under reduced pressure. The crude product was purified by column chromatography using silica gel as absorbent and light petroleum–EtOAc (2 : 1) as eluent yielding 3-(2-iodobenzyl)-3H-quinazolin-4-one 16 as pale yellow crystals (1.82 g, 5.0 mmol, 63%), mp 94–95 °C; Found: M⁺, 362.9987. C₁₅H₁₁IN₂O requires 362.9989; ν_max(thin film)/cm⁻¹ 3055, 1689, 1469, 1230, 962 and 734; δ_H 5.71 (2 H, s, CH₂), 6.93 (1 H, ddd, J 7.9, 7.4, 1.7, BnH-4), 7.09 (1 H, dd, J 7.9, 1.8, BnH-3), 7.22 (1 H, ddd, J 8.0, 7.4, 1.2, BnH-5), 7.45 (1 H, ddd, J 8.0, 5.5, 1.7, 6-H), 7.75–7.64 (2 H, m, 7,8-H), 7.81 (1 H, dd, J 8.0, 1.2, BnH-6), 8.11 (1 H, s, 2-H) and 8.26 (1 H, dd, J 8.0, 1.6, 5-H); δ_C 54.90 (CH₂), 98.6 (C), 122.1 (4a-C), 126.9 (5-C), 127.5 (6-C), 127.6 (8-C), 128.9 (CH), 129.1 (CH), 129.9 (CH), 134.5 (7-C), 137.7 (8a-C), 139.9 (CH), 146.4 (2-C), 148.0 (C) and 161.1 (4-C); m/z (EI) 252 (M⁺, 98%), 129 (43), 107 (41), 90 (100), 89 (77), 76 (51), 63 (58), 50 (42) and 40 (85).

3-[2-(Bromophenyl)ethyl]-3H-quinazolin-4-one 17. The general procedure for alkylation with 2-(2-bromophenyl)ethyl methanesulfonate 15 (4.90 g, 18.6 mmol) for 48 h was used to yield 3-[2-(bromophenyl)ethyl]-3H-quinazolin-4-one 17 as colourless crystals (1.93 g, 5.9 mmol, 38%), mp 119–120 °C; Found: M⁺, 328.0206. C₁₆H₁₃⁷⁹BrN₂O requires 328.0206; ν_max(thin film)/cm⁻¹ 3423, 1972, 1609, 1470, 1069, 1027 and 773; δ_H 3.25 (2 H, t, J 7.2, CH₂), 4.25 (2 H, t, J 6.9, NCH₂), 7.22–7.07 (3 H, m, ArH), 7.58–7.48, (2 H, m, ArH), 7.78–7.64 (3 H, m, ArH) and 8.34 (1 H, ddd, J 8.1 1.5 0.5, 5-H); δ_C 35.3 (CH₂), 46.7 (NCH₂), 122.0 (C), 124.4 (C), 126.6 (CH), 127.2 (CH), 127.5 (CH), 127.9 (CH), 128.9 (CH), 131.4 (CH), 133.1 (CH), 134.2 (CH), 136.7 (C), 146.4 (CH), 148.1 (C) and 161.1 (C); m/z (EI) 330/328 (M⁺, 52/51%), 251 (49), 249 (56) and 52 (14).

Cyclisation reactions of 3-(2-iodobenzyl)-3H-quinazolin-4-one 16.
General procedure for Bu₃SnH reactions using Et₃B as initiator. A solution of Bu₃SnH (0.76 g, 2.6 mmol) and 3-(2-iodobenzyl)-3H-quinazolin-4-one (0.44 g, 1.2 mmol) in dry toluene (40 cm³) was deoxygenated under an atmosphere of nitrogen and stirred for 1 h. Triethylborane (3.6 cm³, 3.6 mmol) was added via a needle through a septum; the needle was then left open to allow air (oxygen) to enter the reaction and the mixture stirred for 1 h. More triethylborane (3.6 cm³, 3.6 mmol) was added and the reaction mixture stirred for a further 10 h. Dilute hydrochloric acid was added to extract the protonated quinazolone products into the aqueous layer. The aqueous layer was washed with light petroleum to remove tributyltin residues. The aqueous layer was basified with sodium hydroxide to pH 14 and extracted with DCM. The combined organic layers were dried and evaporated under reduced pressure yielding the reduced product 3-benzyl-3H-quinazolin-4-one 19 as colourless crystals (0.27 g, 1.2 mmol, 96%), mp 116–117 °C (lit.,⁵⁸ 117–118 °C); ν_max(thin film)/cm⁻¹ 3419, 1675, 1609, 1559, 1473, 1369, 1332, 1161, 1026, 773, 709, 694 and 645; δ_H 5.08 (2 H, s, CH₂), 7.24–7.23 (5 H, m), 7.48 (1 H, ddd, J 8.2 6.3 1.8, 6-H), 7.74–7.68 (2 H, m, 7,8-H), 8.11 (1 H, s, 2-H) and 8.32 (1 H, dd, J 8.2 2.0, 5-H); δ_C 49.6 (CH₂), 122.2 (4a-C), 126.9 (5-H), 127.4 (6-H), 127.5 (8-H), 128.0 (CH), 128.3 (CH), 129.0 (CH), 134.3 (7-H), 135.8 (C), 146.4 (2-H), 148.0 (8a-C) and 161.1 (4-C). The data were identical to that in the literature.⁵⁸ The procedure was repeated using TTMSS in place of Bu₃SnH to yield 3-benzyl-3H-quinazolin-4-one 19 (35%).

General procedure for slow addition of Bu₃SnH reactions using Et₃B as initiator. The general procedure for Bu₃SnH reactions using Et₃B as initiator was repeated except that Bu₃SnH was added by a syringe pump over 6 h to give 3-benzyl-3H-quinazolin-4-one 19 (30%).

General procedure for slow addition of Bu₃SnH reactions using AIBN as initiator. The general procedure for Bu₃SnH reactions using Et₃B as initiator was repeated except that AIBN (0.25 molar equiv.) was used as initiator and the reaction was heated under reflux for 6 h to yield 3-benzyl-3H-quinazolin-4-one 19 (45%).

General procedure for reactions using photolysis and hexamethylditin. A solution of 3-(2-iodobenzyl)-3H-quinazolin-4-one 16 (0.36 g, 1.0 mmol) and hexamethylditin (0.99 g, 3.0 mmol) in tert-butylbenzene (20 cm³) in a two-necked pyrex flask (5 × 1 cm and 25 cm high, wall thickness = 1 mm), was purged with nitrogen for 30 min. The mixture was irradiated with a combined 300 W sunlamp at 150 °C for 10 h. The reaction mixture was cooled to room temperature, diluted with MeOH and evaporated under reduced pressure to a small volume. The residue was purified using column chromatography with light petroleum as eluent to remove the tert-butylbenzene. A precipitate of polymeric dimethyltin was produced. The product was eluted with ethyl acetate and further worked-up as for the general procedure for Bu₃SnH reactions using Et₃B as initiator. 3-Benzyl-3H-quinazolin-4-one 19 and cyclised 12H-isoindolo[1,2-b]quinazolin-10-one 20 were obtained as an inseparable white solid. Analysis using ¹H NMR spectroscopy showed 19 (65%) and 20 (18%). 12H-Isoindolo[1,2-b]quinazolin-10-one 20 was identified in the mixture and therefore not fully characterised. δ_H 5.19 (2 H, s, CH₂), 7.28–7.35 (3 H, m), 7.51 (1 H, ddd, J 8.0, 8.0, 1.6, 7-H), 7.81 (1 H, ddd, J 8.0, 8.0, 1.6, 8-H), 7.85 (1 H, dd, 8.0, 1.6, 6-H), 8.21 (1 H, d, J 7.2, 4-H) and 8.41 (1 H, dd, J, 8.0, 1.6, 9-H); δ_C 49.6 (CH₂), 120.6 (C), 123.4 (CH), 123.5 (CH), 126.4 (CH), 126.4 (CH), 127.3 (CH), 128.8 (CH), 132.3 (CH), 132.6 (C), 134.2 (CH), 139.6 (C), 149.4 (C), 154.9 (C) and 160.5 (C). The data were the same as that in the literature.⁵⁹ GC-MS analysis showed the ratio of the two products to be in the ratio 4 : 1 of the reduced produced and cyclised product: 12H-isoindolo-[1,2-b]-quinazolin-10-one; R_T 22.8 min, m/z 234 (M⁺, 95%), 205 (24), 179 (15), 151 (16), 130 (25), 102 (28), 91 (100) and 77 (20) and 3-benzyl-3H-quinazolin-4-one; R_T 24.5 min, m/z 236 (M⁺, 51%), 130, (37), 91 (100) and 65 (27).

3-(2-Deuterio)benzyl-3H-quinazolin-4-one. The general procedure for reductive reactions with Et₃B with Bu₃SnD as reductant were used with 3-(2-iodobenzyl)-3H-quinazolin-4-one 16 to yield 3-(2-deuterio)benzyl-3H-quinazolin-4-one as colourless crystals (41%); mp 100–101 °C (Found: M⁺, 238.1084. C₁₅DH₁₁N₂O requires 238.1085); ν_max(thin film)/cm⁻¹ 3421, 2359, 1670, 1610, 1471, 1396, 1153 and 773; δ_H 5.21 (2 H, s, CH₂), 7.38–7.29 (4 H, m, PhH), 7.52 (1 H, dd, J 8.1 1.4, 6-H), 7.79–7.69 (2 H, m, 7,8-H), 8.12 (1 H, s, 2-H) and 8.34 (1 H, ddd, J 8.1 2.1 0.6, 5-H); δ_C 49.7 (CH₂), 104.1 (CD), 122.2 (4a-C), 126.9 (5-C), 127.4 (6-C), 127.6 (CH), 128.0 (CH), 128.3 (CH), 128.9 (CH), 129.1 (CH), 134.3 (CH), 135.7 (C), 146.3 (5-C), 148.1 (8a-C) and 161.1 (4a-C); m/z (EI) 237 (M⁺, 48%), 235 (21), 130 (33), 92 (100), 91 (75), 65 (21) and 51 (23).

Cyclisation reactions of 3-[2-(bromophenyl)ethyl]-3H-quinazolin-4-one 17.
Hexamethylditin. The general procedure for reactions using photolysis and hexamethylditin yielded 5,6-dihydroisoquino[1,2-b]quinazolin-8-one 22 as colourless crystals (92%), mp 195–197 °C (lit.,⁶⁰ 196 °C); Found: M⁺, 249.1025. C₁₆H₁₂N₂O requires 249.1022; ν_max(thin film)/cm⁻¹ 3421, 1682, 1590, 1556, 1471, 1150, 762 and 692; δ_H 3.06 (2 H, t, J 6.4, CH₂), 4.35 (2 H, t, J 6.4, NCH₂), 7.22–7.13 (1 H, m, ArH), 7.43–7.37 (3 H, m, ArH), 7.71–7.69 (2 H, m, ArH), 8.25 (1 H, d, J 7.6, ArH) and 8.43–8.40 (1 H, m, ArH); δ_C 27.4 (CH₂), 39.5 (NCH₂), 120.7 (C), 126.5 (CH), 126.8 (CH), 127.5 (CH), 127.6 (CH), 127.9 (CH), 129.5 (C), 131.7 (CH), 134.2 (CH), 137.0 (C), 147.7 (C), 149.3 (C) and 161.6 (CO); m/z (EI) 249 (M⁺, 17%), 247 (100), 128 (20), 116 (46), 104 (30), 90 (34), 89 (62), 77 (85), 76 (97), 63 (53) and 39 (54). One CH signal was coincident or could not be detected.
Bu₃SnH and Et₃B. The general procedure for Bu₃SnH reactions using Et₃B as initiator yielded an inseparable mixture of 3-[2-(phenylethyl)]-3H-quinazolin-4-one 23 (55%) and 5,6-dihydroisoquino[1,2-b]quinazolin-8-one 22 (8%). The yields were estimated using ¹H NMR spectroscopic analysis.

3-Benzyl-2-bromo-3H-quinazolin-4-one 27. The general procedure for alkylation of 3H-quinazolin-4-ones (3 h) yielded 3-benzyl-2-bromo-3H-quinazolin-4-one 27 as colourless crystals (46%), mp 109–111 °C; Found: M⁺, 315.0128. C₁₅H₁₁BrN₂O requires 315.0128; ν_max(thin film)/cm⁻¹ 3457, 3064, 1653, 1558, 1494, 1430, 1339, 1234, 1134, 1074, 1023, 901 and 819; δ_H 5.56 (2 H, s CH₂), 7.29–7.37 (5 H, m, PhH), 7.52 (1 H, ddd, J 8.2, 8.1, 2.0, 7-H), 7.66 (1 H, dd, J 8.2, 1.2, 8-H), 7.79 (1 H, ddd, J 8.1, 8.0, 1.2, 6-H) and 8.27 (1 H, dd, J 8.0, 2.0, 5-ArH); δ_C 51.8 (CH₂), 120.5 (4a-C), 126.9 (6/8-CH), 127.5 (CH), 127.6 (6/8-CH), 127.7 (5-CH), 127.8 (CH), 127.9 (7-CH), 135.1 (CH), 135.4 (C), 136.2 (8a-C), 147.2 (2-C) and 161.6 (4-C); m/z (EI) 318 (M⁺, 15%), 317 (100), 315 (100), 237 (75), 235 (39) and 52 (79).

2-Bromo-3-(phenylethyl)-3H-quinazolin-4-one 28. The general procedure for alkylation of 3H-quinazolin-4-ones (12 h) yielded 2-bromo-3-(phenylethyl)-3H-quinazolin-4-one as colourless crystals (6%), mp 109–110 °C; Found: M⁺, 328.0206. C₁₆H₁₃⁷⁹BrN₂O requires 328.0206; mp 195–197 °C; ν_max(thin film)/cm⁻¹ 3060, 2930, 2359, 1866, 1714, 1655, 1621, 1604, 1493, 1452, 1405, 1072, 1028, 938 and 908; δ_H 2.99 (2 H, t, J 7.6, CH₂), 4.37 (2 H, t, J 7.6, NCH₂), 7.26–7.15 (5 H, m, PhH), 7.34–7.30 (1 H, m, 6-H), 7.49 (1 H, d, J 7.6, 8-H), 7.67–7.63 (1 H, m, 7-H) and 8.20 (1 H, d, J 7.6, 5-H); δ_C 34.0 (CH₂), 42.2 (NCH₂), 114.5 (4a-C), 114.9 (CH), 123.1 (8-C), 126.4 (6-C), 128.3 (5-C), 128.4 (CH), 128.9 (CH), 134.8 (7-C), 138.5 (C), 138.7 (C), 151.2 (2-C) and 162.3 (4-C); m/z (EI) 266 (M⁺, 16%), 264 (15), 249 (20), 185 (100), 129 (52), 102 (23), 90 (35) and 40 (53).

Cyclisation reactions of 3-benzyl-2-bromo-3H-quinazolin-4-one 27.
Bu₃SnH and Et₃B. The general procedure for Bu₃SnH reactions using Et₃B as initiator were used with Bu₃SnH added by syringe pump over 6 h and the reaction stirred for a further 5 h. Work-up yielded 3-benzyl-3H-quinazolin-4-one 19 as colourless crystals (0.15 g, 0.62 mmol, 62%). The data were identical to authentic material. A repeat experiment using Bu₃GeH in place of Bu₃SnH gave 3-benzyl-3H-quinazolin-4-one 19 in reduced yield (47%).
Hexamethylditin. The general procedure for reactions using photolysis and hexamethylditin yielded an inseparable mixture of 12H-isoindolo-[1,2-b]quinazolin-10-one 20 (27%) and 3-benzyl-3H-quinazolin-4-one 19 (41%). The yields were estimated using ¹H NMR spectroscopic analysis.

Cyclisation reactions of 2-bromo-3-(phenylethyl)-3H-quinazolin-4-one 28.
Hexamethylditin. The general procedure for reactions using photolysis and hexamethylditin (24 h) yielded 5,6-dihydroisoquino[1,2-b]quinazolin-8-one 22 as colourless crystals (97%). The data were identical to authentic material.

Cyclisation reactions of 3-(3-iodopropyl)-3H-quinazolin-4-one 31a.
Hexamethylditin. The general procedure for reactions using photolysis and hexamethylditin with 3-(3-iodopropyl)-3H-quinazolin-4-one 31a (24 h) yielded: 2,3-dihydro-1H-pyrrolo[2,1-b]quinazolin-9-one 2 as colourless crystals (20%), mp 190–192 °C (lit.,⁶¹ 196–198 °C); ν_max(thin film)/cm⁻¹ 3418, 2355, 1651, 1621, 1470, 1385, 1285, 775, 695 and 667; δ_H 2.35–2.23 (2 H, m, CH₂), 3.18 (2 H, t, J 12.6, CH₂), 4.21 (2 H, t, J 11.5, NCH₂), 7.47–7.62 (1 H, m, ArH), 7.76–7.62 (2 H, m, ArH) and 8.28 (1 H, dd, J 12.6, 2.4, ArH); δ_C 19.5 (CH₂), 32.5 (CH₂), 46.5 (NCH₂), 120.4 (C), 126.2 (CH), 126.4 (CH), 126.8 (CH), 134.2 (CH), 149.1 (C), 159.5 (C) and 161.0 (C). 2,3,3a,4-Tetrahydro-1H-pyrrolo[2,1-b]quinazolin-9-one 32a as colourless crystals (12%), mp 187–188 °C (lit.,⁶² 184–185 °C); ν_max(thin film)/cm⁻¹ 3421, 1633, 1503, 1470, 1434, 1337, 1151, 752 and 695; δ_H 2.32–2.29 (2 H, m, 2-CH₂), 3.19 (1 H, t, J 8.0, 1-H), 3.75–3.67 (2 H, m, CHCH₂), 4.21 (1 H, dd, J 8.0, 7.2, 1-H), 4.5 (1 H, bs, NH), 5.03 (1 H, dd, J 7.2, 5.6, 3a-H), 6.69 (1 H, d, J 12.9, ArH), 6.92–6.88 (1 H, m, ArH), 7.27–7.24 (1 H, m, ArH) and 7.91–7.87 (1 H, dd, J 12.3, 9.9, ArH); δ_C 21.7 (CH₂), 33.2 (CHCH₂), 44.3 (NCH₂), 69.9 (CH), 114.9 (CH), 118.4 (CH), 119.9 (C), 126.8 (CH), 132.9 (CH), 149.1 (C) and 162.2 (C). 3-Propyl-3H-quinazolin-4-one 33a as colourless crystals (6%). The data were identical with those from an independently prepared authentic sample.
Bu₃SnH and Et₃B. 3-Propyl-3H-quinazolin-4-one 33a (78%) was the only product.
Et₃B only. The general procedure for Bu₃SnH and Et₃B reactions were used except that the Bu₃SnH was not added. ¹H NMR spectral analysis showed 2,3-dihydro-1H-pyrrolo[2,1-b]quinazolin-9-one 2 (40%), 2,3,3a,4-tetrahydro-1H-pyrrolo[2,1-b]quinazolin-9-one 32a (10%) and 3-propyl-3H-quinazolin-4-one 33a (30%).

Cyclisation reactions of 3-(3-iodobutyl)-3H-quinazolin-4-one 31b.
(Me₃Sn)₂ and photolysis. The general procedure for reactions using photolysis and hexamethylditin (24 h) yielded: 6,7,8,9-tetrahydropyrido[2,1-b]quinzolin-11-one 3 as colourless crystals (30%), mp 94–95 °C (lit.,⁶³ 96–97 °C); Found: MH⁺, 201.1020. C₁₂H₁₃N₂O requires 201.1022; ν_max(thin film)/cm⁻¹ 3423, 2948, 2110, 1655, 1614, 1565, 1477, 1173, 1102, 990, 870, 771 and 583; δ_H 2.18–1.91 (4 H, m, CH₂CH₂), 3.00 (2 H, t, J 6.6, NCCH₂), 4.09 (2 H, t, J 6.3, NCH₂), 7.44–7.40 (1 H, m, ArH), 7.59 (1 H, dd, J 8.3, 0.5, ArH), 7.73–7.69 (1 H, m, ArH) and 8.26 (1 H, dd, J 8.0 1.5, ArH); δ_C 19.3 (CH₂), 22.1 (CH₂), 31.9 (CH₂), 42.3 (CH₂), 120.4 (C), 126.1 (CH), 126.4 (CH), 126.6 (CH), 134.1 (CH), 147.4 (C), 154.9 (C) and 162.2 (C); m/z (EI) 200 (M⁺, 50%), 199 (40), 90 (33), 76 (34), 63 (33), 50 (34) and 41 (90). 5,5a,6,7,8,9-Hexahydropyrido[2,1-b]quinazolin-11-one 32b as colourless crystals (23%), mp 128–130 °C (lit.,² 132–133 °C); Found: MH⁺, 203.1180. C₁₂H₁₅N₂O requires 203.1179; ν_max(thin film)/cm⁻¹ 3398, 3058, 2938, 2224, 1632, 1518, 1416, 1218, 1153, 1026, 754 and 692; δ_H 1.52–1.49 (2 H, m), 1.91–1.75 (4 H, m), 2.66–2.59 (1 H, m, 9-H), 4.25 (1 H, m, NH), 4.64–4.59 (1 H, m, 9-H), 4.78 (1 H, dd, J 13.5, 3.1, 5a-H), 6.57 (1 H, d, J 7.1, ArH), 6.79 (1 H, ddd, J 7.8, 2.9, 1.0, ArH), 7.27–7.22 (1 H, m, ArH) and 7.90 (1 H, dd, J 7.8 1.6, ArH); δ_C 22.3 (CH₂), 24.2 (CH₂), 33.4 (CH₂), 42.0 (CH₂), 68.5 (CH), 113.7 (CH), 115.2 (C), 118.8 (CH), 128.7 (CH), 133.4 (CH), 145.9 (C) and 164.3 (C); m/z (EI) 202 (M⁺, 30%), 146 (42), 119 (48), 92 (42), 55 (79) and 41 (100).
Et₃B only. The general procedure for Bu₃SnH and Et₃B reactions were used except that the Bu₃SnH was not added. ¹H NMR spectral analysis showed 6,7,8,9-tetrahydro-pyrido[2,1-b]quinzolin-11-one 3 (61%) and 5,5a,6,7,8,9-hexahydropyrido[2,1-b]quinazolin-11-one 32b (23%).

Cyclisation of 3-[(2-bromoquinolin-3-yl)methyl]-4(3H)-quinazolinone 34.
(Me₃Sn)₂ and photolysis. The general procedure for reactions using photolysis and hexamethylditin (24 h) yielded: luotonin A 5 (51%). All data were the same as reported.⁴⁵ 3-[(Quinolin-3-yl)methyl]-3H-quinazolin-4-one 36, colourless crystals, mp 252–253 °C; Found: MH⁺, 288.1130. C₁₈H₁₃N₃O requires 288.1131; ν_max(thin film)/cm⁻¹ 3420, 2357, 2107, 1646, 1558, 1450, 1361, 1332, 510 and 454; δ_H 5.39 (2 H, s, CH₂), 7.58–7.53 (2 H, m, ArH), 7.81–7.70 (4 H, m, ArH), 8.10 (1 H, dd, J 8.9, 0.4, ArH), 8.17 (1 H, d, J 1.8, ArH), 8.35–8.32 (1 H, m, ArH) and 8.99 (1 H, d, J 1.8, ArH); δ_C 47.7 (CH₂), 122.1 (CH), 126.9 (CH), 127.3 (CH), 127.7 (2 × CH), 127.9 (CH), 128.6 (2 × C), 129.3 (CH), 130.0 (CH), 134.6 (CH), 135.4 (CH), 145.9 (CH), 147.9 (C), 148.0 (C), 150.1 (CH) and 161.1 (C); m/z (EI) 287 (M⁺, 45%), 270 (42), 132 (100), 115 (70), 89 (20), and 63 (21).
Bu₃SnH and Et₃B. The general procedure for Bu₃SnH reactions using Et₃B as initiator were used with Bu₃SnH added by syringe pump over 5 h and the reaction stirred for a further 5 h to yield luotonin A 5 (14%) and 3-[(quinolin-3-yl)methyl]-3H-quinazolin-4-one 36 (32%). When the Bu₃SnH was added at the beginning of the reaction only 3-[(quinolin-3-yl)methyl]-3H-quinazolin-4-one 36 was obtained (53%).
Bu₃GeH and Et₃B. Luotonin A (18%) and [(3-quinolin-3-yl)methyl]-3H-quinazolin-4-one 36 (11%) were obtained.

Cyclisation of 3-[2-(2-bromoindol-3-yl)ethyl]-4(3H)-quinazolinone 40.
Bu₃SnH addition. The general procedure for Bu₃SnH reactions using Et₃B as initiator were used with Bu₃SnH added by syringe pump over 12 h to yield rutaecarpine 8 (15%), mp 254–255 °C (lit.,⁴¹ 256–257 °C) and 3-[2-(1H-indol-3-yl)ethyl]-4(3H)-quinazolinone 39 (57%), both as colourless crystals. The data for rutaecarpine were the same as those in the literature.⁴¹ The data for 39 were the same as obtained in the previous synthesis.
(Me₃Sn)₂ and photolysis. The general procedure for reactions using photolysis and hexamethylditin (6 h) yielded rutaecarpine (55%) as the only product.

2-(4-Oxo-4H-quinazolin-3-yl)selenobenzoic acid Se-phenyl ester 43. Tributylphosphine (2.5 cm³, 10 mmol) was slowly added to diphenyl diselenide (2.47 g, 7.9 mmol) in anhydrous DCM and the mixture left to stir for 15 min. 2-(4-Oxo-4H-quinazolin-3-yl)benzoic acid 42 (1.40 g, 5.2 mmol) was then added and the reaction stirred for 12 h till no precipitate was visible. The reaction mixture was diluted with DCM and washed with H₂O (3 × 50 cm³) and brine (50 cm³). The organic layer was dried over MgSO₄ and evaporated to dryness under reduced pressure. The crude product was purified by column chromatography using silica gel as absorbent and light petroleum/EtOAc (4:1). The resulting yellow solid was recrystallised from hot EtOH yielding 2-(4-oxo-4H-quinazolin-3-yl)selenobenzoic acid Se-phenyl ester 43 as long yellow needles (1.55 g, 3.8 mmol, 73%); (Found: M⁺, 407.0289. C₂₁H₁₄N₂O₂Se requires 407.0293); mp 234–236 °C; ν_max(thin film)/cm⁻¹ 3165, 2831, 2063, 1798, 1519, 1367, 1118, 960 and 915; δ_H 7.36 − 7.30 (3 H, m, ArH), 7.42 (1 H, dd, J 9.2 1.6, ArH), 7.55 − 7.45 (3 H, m, ArH), 7.70–7.66 (1 H, m, ArH), 7.81–7.72 (3 H, m, ArH), 8.02 (1 H, s, 2-H), 8.14 (1 H, dd, J 9.2 1.6, ArH) and 8.34–8.32 (1 H, m, ArH); δ_C 122.3 (C), 125.9 (C), 127.1 (5-C), 127.6 (6-C), 127.7 (CH), 129.2 (CH), 129.4 (2 × CH), 129.4 (CH), 129.8 (CH), 130.1 (CH), 133.6 (C), 133.7 (CH), 134.6 (CH), 136.0 (2 × CH), 136.9 (C), 145.6 (2-C), 148.0 (C), 160.7 (4-C) and 192.7 (CO); m/z (EI) 407 (M⁺, 100%), 251 (39), 235 (55) and 223 (46).

Radical reactions of 2-(4-oxo-4H-quinazolin-3-yl)selenobenzoic acid Se-methyl ester 43.
UV and reflux. 2-(4-Oxo-4H-quinazolin-3-yl)selenobenzoic acid Se-phenyl ester 43 (0.24 g, 0.59 mmol) was heated under reflux in benzene (20 cm³) for 12 h under UV irradiation. The solvent was removed under vacuum and the crude product was purified by column chromatography using silica gel as absorbent and light petroleum–EtOAc (2 : 1) yielding tryptanthrin 4 as yellow crystals (0.02 g, 0.07 mmol, 13%), mp 213–214 °C (lit.⁵³ 215–217 °C); Found: MH⁺, 249.0659. C₁₅H₉N₂O₂ requires 249.0660; δ_H 7.43 (1 H, t, J 8.0, ArH), 7.68 (1 H, t, J 8.0, ArH), 7.79 (1 H, t, J 8.0, ArH), 7.86 (1 H, t, J 8.0, ArH), 7.92 (1 H, d, J 8.0, ArH), 8.04 (1 H, d, J 8.0, ArH), 8.44 (1 H, d, J 8.0, ArH) and 8.64 (1 H, d, J 8.0, ArH); δ_C 118.0 (CH), 122.0 (C), 123.8 (C), 125.4 (CH), 127.2 (CH), 127.6 (CH), 130.2 (CH), 130.7 (CH), 135.1 (CH), 138.3 (CH), 146.4 (C), 146.7 (C), 158.1 (C) and 171.1 (2 × CH); m/z (EI) 248 (M⁺, 51%), 220 (22), 130 (28), 102 (54), 90 (100), 76 (81), 63 (55) and 50 (74). The data were identical to the literature data⁵³ and those of authentic material. Repeating the reaction for 10 h gave a yield of 11% of 4. The procedure was repeated using the same conditions but with the addition of AIBN (2.0 equiv.) for 8 h. An intractable mixture was obtained. A blank reaction with no sunlamp photolysis or added AIBN, in benzene heated under reflux for 24 h gave a quantitative recovery of unaltered starting material 43.
(Me₃Sn)₂, sunlamp. The reaction between 2-(4-oxo-4H-quinazolin-3-yl)selenobenzoic acid Se-phenyl ester 47 (0.29 g, 0.72 mmol) and hexamethylditin (0.70 g, 2.15 mmol) was carried out in tert-butylbenzene (20 cm³). The mixture was heated under reflux and irradiated with a sunlamp for 4 h. Dilute hydrochloric acid was added to the cooled reaction mixture to extract the protonated quinazolone products into the aqueous layer. The aqueous layer was washed with light petroleum to remove tin residues. The aqueous layer was basified with sodium hydroxide to pH 14 and extracted with DCM. The combined organic layers were dried and evaporated under reduced pressure. The residue was purified by column chromatography using silica gel as absorbent and light petroleum–EtOAc (2 : 1) yielding tryptanthrin 4 (5 mg, 0.02 mmol, 3%).
Sunlamp. When the sunlamp irradiation alone was used for 30 min, the highest yield of tryptanthrin (15%) was obtained.

2-[(4-Oxo-4H-quinazolin-3-yl)methyl]benzoic methyl ester 45. N-Bromosuccinimide (19.0 g, 107.1 mmol, 1.5 equiv.), methyl 2-methylbenzoate (10.0 cm³, 71.4 mmol) and AIBN (1.16 g, 7.1 mmol) were refluxed in cyclohexane in the dark for 5 h. The reaction mixture was cooled to 0 °C and filtered. The filtrate was diluted with DCM, washed with H₂O (3 × 50 cm³) and brine (50 cm³) and then dried and evaporated to dryness under reduced pressure. The unpurified methyl 2-bromomethylbenzoate was added to a solution of 3H-quinazolin-4-one (7.3 g, 50 mmol) and potassium tert-butoxide (07.3 g, 65 mmol) in DMF (50 cm³) that had been stirred for 40 min. The resulting mixture was stirred for 12 h then diluted with DCM, and the solution washed with H₂O and brine. The organic layer was dried and evaporated under reduced pressure. The crude product was purified by column chromatography using silica gel as absorbent and light petroleum–EtOAc (4 : 1) as eluent yielding 2-[(4-oxo-4H-quinazolin-3-yl)methyl]benzoic methyl ester as a pale oil 45 (4.98 g, 16.6 mmol, 34%); (Found: M⁺, 295.1077. C₁₇H₁₄N₂O₃ requires 295.1077); ν_max(thin film)/cm⁻¹ 3541, 3417, 1715, 1681, 1563, 1367, 1322, 1079, 968 and 774; δ_H 3.88 (3 H, s, OMe), 5.62 (2 H, s, CH₂), 7.31–7.22 (2 H, m, ArH), 7.70–7.69 (2 H, m, ArH), 7.97 (1 H, dd, J 7.8 1.4, ArH) and 8.28–8.23 (2 H, m, ArH); δ_C 47.9 (CH₂), 52.4 (Me), 122.2 (4a-C), 126.9 (CH), 127.3 (CH), 127.5 (CH), 128.0 (CH), 128.6 (C), 129.1 (CH), 131.3 (CH), 133.0 (CH), 134.7 (CH), 147.1 (CH), 148.1 (8a-C), 161.4 (4-C) and 167.4 (CO₂Me); m/z (EI) 262 (M⁺, 53%), 132 (80), 118 (30), 102 (60), 91 (100), 77 (58), 63 (33) and 50 (23).

2-[(4-Oxo-4H-quinazolin-3-yl)methyl]benzoic methyl acid 46. A solution of 2-[(4-oxo-4H-quinazolin-3-yl)methyl]benzoic methyl ester 45 (1.93 g, 6.14 mmol) in ethanolic LiOH (1 M, 200 cm³) and water (20 cm³) was stirred for 18 h. The mixture was acidified with HCl and extracted with EtOAc. The organic layers were combined and washed with H₂O and brine. The organic extract was dried over MgSO₄ and evaporated to dryness under reduced pressure yielding 2-[(4-oxo-4H-quinazolin-3-yl)methyl]benzoic methyl acid 46 as a colourless oil (0.63 g, 2.1 mmol, 34%); Found: M⁺, 281.0918. C₁₆H₁₂N₂O₃ requires 281.0921; m/z (EI) 262 (M⁺, 31%), 146 (35), 132 (86), 118 (38), 104 (47), 90 (58), 77 (100), 63 (28), 51 (32) and 43 (53). The carboxylic acid 46 was converted to the acyl selenide without further characterisation.

2-[(4-Oxo-4H-quinazolin-3-yl)methyl]seleno benzoic acid Se-phenyl ester 47. The same procedure as for the seleno-ester 43 was used to convert 2-[(4-oxo-4H-quinazolin-3-yl)methyl]benzoic methyl acid 46 to 2-[(4-oxo-4H-quinazolin-3-yl)methyl]seleno benzoic acid Se-phenyl ester 47 as colourless crystals (0.31 g, 0.71 mmol, 68%), mp 314–316 °C; Found: M⁺H⁺, 421.0450. C₂₂H₁₇N₂O₂⁸⁰Se requires 421.0451; ν_max(thin film)/cm⁻¹ 3361, 1911, 1659, 1589, 1542, 1392, 1336, 1188, 801, 760 and 722; δ_H 5.43 (2 H, s, CH₂), 7.37–7.35 (1 H, m, ArH), 7.53–7.45 (6 H, m, ArH), 7.61–7.59 (2 H, ArH), 7.79–7.73 (2 H, m, ArH), 8.04 (1 H, dd, J 7.6 1.3, ArH), 8.13 (1 H, s, 2-H) and 8.33–8.30 (1 H, m, ArH); δ_C 47.2 (CH₂), 122.1 (4a-C), 126.3 (C), 126.9 (CH), 127.4 (CH), 127.5 (CH), 128.5 (CH), 129.4 (CH), 129.6 (CH), 129.6 (CH), 129.7 (CH), 133.2 (CH), 133.6 (C), 134.4 (CH), 136.2 (CH), 137.6 (C), 147.0 (CH), 148.0 (8a-C), 161.3 (4-C) and 196.7 (CO); m/z (Electrospray) 421 (M⁺, 7%), 263 (60), 156 (45), 129 (100), 102 (43), 89 (46), 77 (60), 63 (23) and 51 (33).

11-Hydroxy-11H-isoquinolino[3,2-b]quinazolin-6,13-dione 50. The reaction between 2-[(4-oxo-4H-quinazolin-3-yl)methyl]seleno benzoic acid Se-phenyl ester 47 (0.044 g, 0.1 mmol) and hexamethylditin (0.099 g, 0.3 mmol) was carried out in benzene (20 cm³) and was refluxed and irradiated with a UV lamp for 3 h. The benzene was removed under vacuum and the crude product was purified by column chromatography using silica gel as absorbent and light petroleum–EtOAc (4 : 1) as eluent yielding 11-hydroxy-11H-isoquinolino[3,2-b]quinazolin-6,13-dione 50 as colourless needles (0.011 g, 0.04 mmol, 39%), mp 176–177 °C (CHCl₃); Found: (M + H)⁺, 279.0768. C₁₆H₁₁N₂O₃ requires 279.0764; ν_max(thin film)/cm⁻¹ 3421, 2919, 2360, 1699, 1683, 1653, 1559, 1465, 1260, 1100, 963, 910, 728 and 667; δ_H 5.29 (1 H, d, J 4.6, OH), 7.21 (1 H, d, J 4.6, CH), 7.69–7.65 (2 H, m, ArH), 7.80 (1 H, d, J 8.0, ArH), 7.93–7.84 (2 H, m, ArH), 8.10 (1 H, d, J 8.0, ArH), 8.32 (1 H, dd, J 7.6 1.2, ArH) and 8.39 (1 H, dd, J 8.0 1.2, ArH); δ_C 75.5 (CH), 121.9 (C), 126.7 (CH), 128.0 (CH), 128.7 (CH), 129.2 (C), 129.6 (CH), 130.0 (CH), 130.3 (CH), 135.6 (CH), 135.7 (CH), 137.0 (C), 142.6 (ArC), 146.7 (C), 162.7 (C) and 176.6 (C); m/z (Electrospray) 278 (M⁺, 50%), 262 (43), 249 (40), 221 (26), 135 (38), 119 (36), 105 (100), 91 (45), 84 (60), 77 (68) and 51 (69). The structure was confirmed by X-ray crystallography.

X-Ray crystallography

Data were collected at 150(2) K on a Bruker SMART 1000 diffractometer with sealed tube source⁶⁴ for 30 and Bruker-Nonius CCD diffractometer with rotating anode source⁶⁵ for 50. The structures were solved by direct methods and refined by full-matrix least-squares on F² using the SHELXTL suite of programs.⁶⁶ All the non-hydrogen atoms were refined with anisotropic atomic displacement parameters and hydrogen atoms were inserted at calculated positions using a riding model except for H(2) for which coordinates were freely refined in both structures. In 30, molecules form into chains via intermolecular H-bonds N(2)–H(2)⋯O(1′). The structure of 50 additionally contains one molecule of CHCl₃ in the asymmetric unit and molecules are linked into chains via strong intermolecular H-bonds O(2)–H(2)⋯N(2′). Off-set π–π stacking is also dominant with separations in the range 3.33–3.5 Å.

Crystal data for 30: C₁₆H₁₄N₂OS, M = 282.35, monoclinic, Cc, a = 4.6405(7), b = 25.329(4), c = 12.1910(17) Å, β = 97.047(2)°, U = 1422.1(4) ų, T = 150(2) K, Z = 4, µ(Mo–Kα) = 0.224 mm⁻¹, 5350 reflections measured, 3005 unique (R_int = 0.0368) which were used in all calculations, wR = 0.1590 for all data, R1 = 0.0602 for 2351 data with F² ≥ 2σ(F²). Absolute structure parameter = 0.13(13)—thus reliably determined.

Crystal data for 50: C₁₇H₁₁Cl₃N₂O₃, M = 397.63, triclinic, P1̄, a = 6.6511(3), b = 8.6274(3), c = 14.5395(7) Å, α = 90.123, β = 94.525(3), γ = 99.594(3)°, U = 819.97(6) ų, T = 120(2) K, Z = 2, µ(Mo–Kα) = 0.579 mm⁻¹, 11250 reflections measured, 3715 unique (R_int = 0.0370) which were used in all calculations, wR = 0.0939 for all data, R1 = 0.0353 for 2898 data with F² ≥ 2σ(F²).‡

Acknowledgements

We thank the EPSRC (DTA award) and Loughborough University for a Postgraduate Studentship (T. S.), and the EPSRC Mass Spectrometry Unit, Swansea University, Wales for mass spectra. We also thank the EPSRC National Crystallography Service at Southampton University, UK for collecting X-ray diffraction data for compound 50.

References

1. A. Al-Shamma, S. Drake, D. L. Flynn, L. A. Mitscher, Y. H. Park, G. S. R. Rao, A. Simpson, J. K. Swayze, T. Veysogly and S. T.-S. Wu, J. Nat. Prod., 1981, 745–747.
2. S. R. Johns and J. A. Lamberton, J. Chem. Soc., Chem. Commun., 1965, 267; J. S. Fitzgerald, S. R. Johns, J. A. Lamberton and A. H. Redcliffe, Aust. J. Chem., 1966, 19, 151–158.
3. M. Hamburger, Phytochem. Rev., 2002, 1, 333.
4. Z.-Z. Ma, Y. Hano, T. Nomura and Y.-J. Chen, Heterocycles, 1997, 46, 541–546.
5. Y. Asahina, Acta Phytochim., 1922, 1, 67; Y. Asahina and K. Kashiwaki, J. Pharm. Soc. Jpn, 1915, 35, 1293.
6. J. P. Michael, Nat. Prod. Rev., 2004, 21, 650–668.
7. D. A. Horton, G. T. Bourne and M. L. Smythe, Chem. Rev., 2003, 103, 893–930.
8. F. Aldabbagh, W. R. Bowman and E. Mann, Tetrahedron Lett., 1997, 38, 7937–7940; F. Aldabbagh, W. R. Bowman, E. Mann and A. M. Z. Slawin, Tetrahedron, 1999, 55, 8111–8128.
9. Y. Antonio, E. De La Cruz, E. Galeazzi, A. Guzman, B. L. Bray, R. Greenhouse, L. J. Kurz, D. A. Lustig, M. L. Maddox and J. M. Muchowski, Can. J. Chem., 1994, 72, 15–22.
10. S. M. Allin, W. R. S. Barton, W. R. Bowman and T. McInally, Tetrahedron Lett., 2002, 43, 4191–4193.
11. F. E. Ziegler and M. Belema, J. Org. Chem., 1997, 62, 1083–1094 and references therein.
12. C. J. Moody and C. L. Norton, J. Chem. Soc., Perkin Trans. 1, 1997, 2639–2643.
13. J. Marco-Contelles and M. Rodríquez-Fernández, Tetrahedron Lett., 2000, 41, 381–384; J. Marco-Contelles and M. Rodríquez-Fernández, J. Org. Chem., 2001, 66, 3717–3725.
14. J. A. Murphy and M. S. Sherburn, Tetrahedron, 1991, 47, 4077–4088.
15. Y. Z. Osornio, L. D. Miranda, R. Cruz-Almaza and J. D. Muchowski, Tetrahedron Lett., 2004, 45, 2855–2858.
16. S. M. Allin, W. R. S. Barton, W. R. Bowman and T. McInally, Tetrahedron Lett., 2001, 42, 7887–7890.
17. M.-L. Bennasar, T. Roca and F. Ferrando, Org. Lett., 2006, 8, 561–564.
18. M.-L. Bennasar, T. Roca and F. Ferrando, J. Org. Chem., 2005, 70, 9077–9080.
19. M.-L. Bennasar, T. Roca and F. Ferrando, Tetrahedron Lett., 2004, 45, 5605–5609.
20. S. R. Flannagan, D. C. Harrowven and M. Bradley, Tetrahedron Lett., 2003, 44, 1795–1798.
21. T. C. T. Ho and K. Jones, Tetrahedron, 1997, 53, 8287–8294.
22. A. Nadin and T. Harrison, Tetrahedron Lett., 1999, 40, 4073–4076.
23. K. C. Mujumbar and P. P. Mukhopadhyay, Synthesis, 2003, 920–924; K. C. Mujumbar, P. K. Basu, P. P. Mukhopadhyay, S. Sarkar, S. K. Ghosh and P. Biswas, Tetrahedron, 2003, 59, 2151–2157.
24. K. Orito, Y. Satoh, H. Nishizawa, R. Harada and M. Tokuda, Org. Lett., 2000, 2, 2535–2537.
25. D. C. Harrowven, B. J. Sutton and S. Coulton, Tetrahedron, 2002, 58, 3387–3400; D. C. Harrowven, B. J. Sutton and S. Coulton, Tetrahedron Lett., 2001, 42, 2907–2910.
26. D. C. Harrowven, B. J. Sutton and S. Coulton, Tetrahedron Lett., 2001, 42, 9061–9064; D. C. Harrowven, B. J. Sutton and S. Coulton, Org. Biomol. Chem., 2003, 4047–4057.
27. Review: A. Studer and M. Bossart, in Radicals in Organic Synthesis, ed. P. Renaud and M. P. Sibi, Wiley-VCH, Weinheim, 2001, vol. 2, pp. 62–80.
28. A. L. J. Beckwith, W. R. Bowman, V. W. Bowry, E. Mann, J. Parr and J. M. D. Storey, Angew. Chem., Int. Ed., 2004, 43, 95–98.
29. D. L. Comins, H. Hong and G. Jianhua, Tetrahedron Lett., 1994, 35, 5331–5334.
30. M. A. Elban, W. Sun, B. M. Eisenhauer, R. Gao and S. M. Hecht, Org. Lett., 2006, 8, 3513–3516; N. J. Rahier, K. Chen, R. Gao, B. M. Eisenhauer and S. M. Hecht, Org. Lett., 2003, 5, 835–837.
31. J. B. Bremner and W. Sengpracha, Tetrahedron, 2005, 61, 941–953.
32. S. M. Allin, W. R. Bowman, M. R. J. Elsegood, V. McKee, R. Karim and S. S. Rahman, Tetrahedron, 2005, 61, 2689–2696.
33. (a) C. Kaneko, K. Kasai, N. Katagiri and T. Chiba, Chem. Pharm. Bull., 1986, 34, 3672–3681; (b) R. Tangirala, S. Anthony, K. Agami, Y. Pommier and D. P. Curran, Synlett, 2005, 2843–2846.
34. R. Hull and M. L. Swain, J. Chem. Soc., Perkin Trans. 1, 1976, 653–659.
35. J. Azizian, A. A. Mohammadi and A. R. Karimi, Synth. Commun., 2003, 33, 415–420.
36. A. Andrzejewska, M. A. Pagano, F. Meggio, A. M. Brunati and Z. Kazimierczuk, Bioorg. Med. Chem., 2003, 11, 3997–4002; W. M. David, D. Kumar and S. M. Kerwin, Bioorg. Med. Chem. Lett., 2000, 10, 2509–2512; J. W. Ellingboe, W. Spinelli, M. W. Winkley, T. T. Nguyen, R. A. Parsons, I. F. Moubarak, J. M. Kitzen, D. Von Engen and J. F. Bagli, J. Med. Chem., 1992, 35, 705–716; C. R. Petrie, H. B. Cottam, P. A. McKernan, R. K. Robins and G. R. Revanker, J. Med. Chem., 1985, 28, 1010–1016; M. C. Huang, T. L. Avery, R. L. Blakley, J. A. Secrist and J. A. Montgomery, J. Med. Chem., 1984, 27, 800–802; R. E. Holmes and R. K. Robins, J. Am. Chem. Soc., 1964, 86, 1242–1245; A. G. Beaman, J. F. Gerster and R. K. Robins, J. Org. Chem., 1962, 27, 986–990.
37. M. K. Ali and M. McDermott, Tetrahedron Lett., 2002, 43, 6271–6273; M. Joshaghani, A. R. Khosropour, H. Jafary and I. Mohammadpoor-Baltork, Phosphorus, Sulfur Silicon Relat. Elem., 2005, 180, 117–123.
38. W. R. Bowman, S. L. Krintel and M. B. Schilling, Org. Biomol. Chem., 2004, 585–592; W. R. Bowman, S. L. Krintel and M. B. Schilling, Synlett, 2004, 1215–1218.
39. J.-F. Liu, P. Ye, K. Sprague, K. Sargent, D. Yohannes, C. M. Baldino, C. J. Wilson and S.-C. Ng, Org. Lett., 2005, 7, 3363–3365; E. S. Lee, J.-G. Park and Y. Jahng, Tetrahedron Lett., 2003, 44, 1883–1886; A. Kamal, K. V. Ramana, H. B. Antaki and A. V. Ramana, Tetrahedron Lett., 2002, 43, 6861–6863; K. G. Nazerenko, T. I. Shyrokaya and A. A. Tolemkev, Synth. Commun., 2003, 33, 303.
40. W. R. Bowman, M. R. J. Elsegood, A. J. Fletcher, P. J. Lovell and J. M. Pedersen, J. Org. Chem., 2005, 70, 10615–10618; W. R. Bowman, A. J. Fletcher, P. J. Lovell and J. M. Pedersen, Synlett, 2004, 1905–1908.
41. T. Harayama, A. Hori, G. Serban, Y. Morikami, T. Matsumoto, H. Abe and Y. Takeuchi, Tetrahedron, 2004, 60, 10645–10649.
42. T. Harayama, Y. Morikami, Y. Shigeta, H. Abe and Y. Takeuchi, Synlett, 2003, 847–848.
43. A. Cagir, S. H. Jones, R. Gao, B. M. Eisenhauer and S. M. Hecht, J. Am. Chem. Soc., 2003, 125, 13628–13629.
44. H. Twin and R. A. Batey, Org. Lett., 2004, 6, 4913–4916; S. B. Mhaske and N. P. Argade, J. Org. Chem., 2004, 69, 4563–4566; S. P. Chavan and R. Sivappa, Tetrahedron, 2004, 60, 9931–9935; D. Osborne and P. Stevenson, Tetrahedron Lett., 2002, 43, 5469–5470; M. Toyota, C. Komori and M. Ihara, Heterocycles, 2002, 56, 101–103; S. Dallavalle and L. Merlini, Tetrahedron Lett., 2002, 43, 1835–1837; P. Molina, A. Tárraga and A. González-Tejero, Synthesis, 2000, 1523–1525; Z.-Z. Ma, Y. Hano, T. Nomura and Y.-J. Chen, Heterocycles, 1999, 51, 1593–1596; H. Wang and A. Ganesan, Tetrahedron Lett., 1998, 39, 9097–9098.
45. W. R. Bowman, M. O. Cloonan, A. J. Fletcher and T. Stein, Org. Biomol. Chem., 2005, 3, 1460–1467.
46. (a) M. A. Ciufolini and F. Roschangar, Tetrahedron, 1997, 53, 11049–11060; (b) H. Abe, T. Harayama, A. Hori, T. Matsumoto, Y. Morikami, G. Serban and Y. Takeuchi, Tetrahedron, 2004, 60, 10645–10649.
47. Selected syntheses: (a) T. Harayama, A. Hori, G. Serban, Y. Morikami, T. Matsumoto, H. Abe and Y. Takeuchi, Tetrahedron, 2004, 60, 10645–10649; (b) S. P. Chavan and R. Sivappa, Tetrahedron Lett., 2004, 45, 997–999; P. K. Mohanta and K. Kim, Tetrahedron Lett., 2002, 43, 3993–3996; S. H. Lee, S. I. Kim, J. G. Park, E. S. Lee and Y. Jahng, Heterocycles, 2001, 55, 1555–1559; A. Ikuta, H. Urabe and H. Nashimura, J. Nat. Prod., 1998, 61, 1012–1014; J. Bergman and S. Bergman, J. Org. Chem., 1985, 50, 1246–1255; J. Kokosi, I. Hermecz, G. Szaz and Z. Meszaros, Tetrahedron Lett., 1981, 22, 4861–4862; Y. Asahina, R. H. F. Manske and R. Robinson, J. Chem. Soc., 1927, 1708–1710.
48. A. P. Dobbs, K. Jones and K. T. Veal, Tetrahedron Lett., 1995, 36, 4857–4860.
49. A. Fiumana and K. Jones, Tetrahedron Lett., 2000, 41, 4209–4211.
50. G. W. Gribble, H. L. Fraser and J. C. Badenock, Chem. Commun., 2001, 805–806.
51. R. Liu, P. Zhang, T. Gan and J. M. Cook, J. Org. Chem., 1997, 62, 7447–7456; U. Schmidt and S. Weinberger, Synthesis, 1996, 28–30.
52. A. R. Katritzky and K. Akutagava, Tetrahedron Lett., 1985, 5935–5938.
53. R. Ajaykumar, P. A. Bapu, D. S. Deevi, G. S. Kumar, K. B. S. Kumar, N. V. S. R. Mamidi, C. P. Narasimhulu, P. Prasanna, R. C. Puranik, C. S. Rao, C. V. L. Rao, S. Rajaopal, R. Rajagopalan, B. Renuka, K. V. A. Seshu, V. M. Sharma, D. Subramanyam and A. Venkateswarlu, Bioorg. Med. Chem. Lett., 2002, 12, 2303–2307.
54. M.-L. Bennasar, T. Roca and F. Ferrando, J. Org. Chem., 2006, 71, 1746–1749.
55. C. Chatgilialoglu, D. Crich, M. Komatsu and I. Ryu, Chem. Rev., 1999, 99, 1991–2069.
56. S. Johne and M. Süsse, J. Prakt. Chem., 1984, 326, 1027–1033.
57. J. A. Howard and J. C. Tait, Can. J. Chem., 1979, 57, 2761; S. B. Tong, I. P. Beletskaya, A. S. Sigeev, V. A. Kuzmin and A. S. Tatikolov, J. Chem. Soc., Perkin Trans. 2, 2000, 107.
58. M. Baghbanzadeh, M. Dabiri, A. Mohammadi and P. Salehi, Synth. Commun., 2005, 279–287.
59. S. Petersen and E. Tietze, Justus Liebigs Ann. Chem., 1959, 623, 166–176; K. Nesmerak, H. Pelouchova, V. Vsetecka, I. Nemec and J. Gabriel, Folia Microbiol. (Prague), 1998, 43, 39–41.
60. L. Jung, I. Kranmacher and J.-F. Stambach, Heterocycles, 1990, 31, 2131–2138.
61. T. Kametani, C. Van Loc, T. Higa, M. Koizumi, M. Ihara and K. Fukumoto, J. Am. Chem. Soc., 1977, 2306–2308.
62. V. Gatta and R. L. Vittory, Gazz. Chim. Ital., 1969, 59.
63. J. S. Yadav and B. V. Reddy, Tetrahedron Lett., 2002, 43, 1905–1908.
64. SMART and SAINT software for CCD diffractometers, Bruker AXS Inc., Madison, WI, 2001.
65. R. W. W. Hooft, COLLECT data collection software, Nonius B.V., 1998.
66. G. M. Sheldrick, SHELXTL user manual, version 6.10, Bruker AXS Inc., Madison, WI, 2001.

---

Footnotes

† Electronic supplementary information (ESI) available: Preparation and analytical details of 3H-quinazolin-4-one radical precursors and products. See DOI: 10.1039/b614075k

‡ CCDC reference numbers 622451 and 622452. For crystallographic data in CIF or other electronic format see DOI: 10.1039/b614075k

---