Nat. Prod. Rep., 2003, 20(5), 476 - 493   DOI: 10.1039/b208140g	Review Article
Quinoline, quinazoline and acridone alkaloids

Joseph P. Michael
Molecular Sciences Institute, School of Chemistry, University of the Witwatersrand, Wits 2050, South Africa. E-mail: jmichael@aurum.chem.wits.ac.za

Covering: July 2001 to June 2002. Previous review: Nat. Prod. Rep., 2002, 19, 742

This review covers the isolation, structure determination, synthesis and biological activity of quinoline, quinazoline and acridone alkaloids from plant, microbial and animal sources. The literature from July 2001 to June 2002 is reviewed, and 125 references are cited.

The past year has brought to light an astonishing number of new quinoline alkaloids, some from unusual sources. By contrast, the number of known quinoline alkaloids isolated from previously unexplored sources was relatively small. The compounds of interest and their sources are listed in Table 1.^1–32

Species	Alkaloida	Ref.
Antidesma membranaceum	(+)-(17RS)-17-(-D-Glucopyranosyloxy)antidesmoneb66	1
	(+)-(17RS)-8-Deoxo-17-(-D-glucopyranosyloxy)antidesmoneb67
Aplidium tabascum (tunicate)	(+)-Lepadin Fbrel-96	2
	(+)-Lepadin Gbrel-98
	(+)-Lepadin Hbrel-99
Aquilegia ecalcarata	7-Hydroxy-4-[5-(hydroxymethyl)-2-furyl]quinolin-2(1H)-oneb68	3
Casimiroa edulis	5,6-Dimethoxy-2-(3-methoxyphenyl)quinolin-4(1H)-oneb1	4
	5,6-Dimethoxy-2-(3,4-dimethoxyphenyl)quinolin-4(1H)-oneb2
	5,6-Dimethoxy-2-(2,5,6-trimethoxyphenyl)quinolin-4(1H)-oneb3
Didemnum sp. (tunicate)	(+)-Lepadin Dbrel-94	5
	(–)-Lepadin D, quaternary derivativeb
	(–)-Lepadin Ebrel-95
	(–)-Lepadin Fbrel-96
Esenbeckia conspecta	Flindersiamine 58	6
	Maculosidine
	8-Methoxy-N-methylflindersine (zanthobungeanine)
Galipea officinalis	2-[2-(3,4-Dimethoxyphenyl)ethyl]quinoline	7
	2-[2-(3,4-Dimethoxyphenyl)ethyl]-1,2,3,4-tetrahydroquinolineb4
	2-[2-(3-Hydroxy-4-methoxyphenyl)ethyl]-4-methoxyquinoline (or 3-methoxy-4-hydroxyphenyl isomer)b5
	Maculosidine
	2-[2-(3,4-Methylenedioxyphenyl)ethyl]-1,2,3,4-tetrahydroquinolineb6
	1-Methyl-2-propyl-1,2,3,4-tetrahydroquinolineb7
	2-Pentyl-1,2,3,4-tetrahydroquinolineb8
Geocrinia laevis (frog)	Decahydroquinoline 195A 115	8
Glycosmis parviflora (= G. citrifolia)	Flindersine 34	9
	N-Methylflindersine 45
Halfordia kendack	(+)-trans-Deacetoxyerioaustralasineb29	10
	(+)-trans-Deacetoxyerioaustralasine hydrateb30
	(+)-trans-1-Epideacetoxyerioaustralasine hydrateb32
	(+)-trans-Erioaustralasine 28
	(+)-trans-Erioaustralasine hydrateb31
Haplophyllum acutifolium	Flindersine 34	11
	Haplophytin-Ab33
	(+)-Haplophytin-B (= evoxine) 35
Haplophyllum patavinum	(–)-Edulinine	12
	Haplopine 55
	(+)-Isoplatydesmine
	(–)-Ribalinine
	Skimmianine 57
Hyrtios erecta (sponge)	6-Bromo-4-hydroxyquinolin-2(1H)-oneb102	13
	6,7-Dibromo-4-hydroxyquinolin-2(1H)-oneb103
Ladenbergia oblongifolia	(–)-Cinchonidicinol 71	14
	(–)-Epicinchonicinol 72
	Dihydrocinchonicinol/dihydrocinchonidicinolb73 (mixture)
Limonia crenulata	Integriquinolone	15
Melicope semecarpifolia (= M. confusa, Evodia merrillii)	2-Acetylevolitrineb50	16
	2-Acetylpteleineb51
	Dictamnine 53
	(3R)-(–)-8,9-Dimethoxygeibalansine 36
	cis-(+)-7,8-Dimethoxymyrtopsine 37
	(2S)-(–)-7,8-Dimethoxyplatydesmine 38
	Evolitrine 54
	Semecarpifolineb14
Micromonospora sp. IM 2670	7-(1-Methyl-2-oxopropyl)streptonigrinb78	17
	Streptonigrin 77
Nitraria sibirica	Dihydroschoberineb75	18
Oryza sativa cv. Heugjinmi	Methyl 6-hydroxy-2-oxo-1,2-dihydroquinoline-4-carboxylateb69	19
Oryza sativa cv. Mihyangbyo	Methyl 6-methoxy-2-oxo-1,2-dihydroquinoline-4-carboxylateb70	20
Penicillium citrinum 90648	Quinolactacin A1b79	21
	Quinolactacin A2b80
Pseudomonas fluorescens G308	N-Mercapto-4-formylcarbostyrilb83	22
Raulinoa echinata	Flindersiamine 58	23
	Kokusaginine 56
	Maculine 59
	N-Methyl-2-nonylquinolin-4-one 15
	N-Methyl-2-phenylquinolin-4-one 16
	2-Nonylquinolin-4(1H)-one 17
	Skimmianine 57
Sarcomelicope megistophylla	Cyclomegistineb18	24
	Megistoquinone Ib60	25
	Megistoquinone IIb62
Scolopendra subspinipes mutilans (centipede)	Scolopendrineb104	26
Severinia buxifolia	N-Methylswietinidine-B	27,28
Streptomyces SNA15896	(–)-SW-163Cb84	29,30
	(–)-SW-163Eb85
Zanthoxylum dimorphophyllum	-Fagarine 64	31
Zanthoxylum hyemale	(R)-(–)-Geibalansineb39	32
	Hyemalineb63
aOnly new alkaloids and new records for a given species are listed in the Table. Structures of known alkaloids, if not specifically numbered, may be found in previous reviews in this series.  bNew alkaloids.

While 2-arylquinoline alkaloids are not uncommon metabolites of plants belonging to the Rutaceae, it is unusual for them to bear substituents on the carbocyclic ring of the quinoline nucleus. However, no fewer than three 2-aryl-5,6-dimethoxyquinolin-4-one alkaloids, 1–3, were isolated from the leaves of the medicinally useful central American plant Casimiroa edulis.⁴ The alkaloids were also detected in the seed kernels by TLC. The structures were determined by a combination of spectroscopic methods, and the substitution patterns were inferred in part by comparing chemical shifts with those reported for related flavonoids.

A GC-MS method for detecting the constituents in the essential oils and extracts from the trunk bark of Galipea officinalis, another medicinally valuable plant from South America, has permitted the identification of a further five minor alkaloids, 4–8.⁷ The structures of the new alkaloids and a further ten known alkaloids were deduced on the basis of their mass spectral fragmentation patterns only. G. officinalis is the only species in this genus to contain tetrahydroquinoline alkaloids, and it is noteworthy that four of the five new alkaloids belong to this group. Some of the better-known G. officinalis tetrahydroquinolines, including cuspareine 9, galipeine 10, galipinine 11 and angustureine 12, have been shown to possess antimalarial activity against both chloroquine-sensitive and chloroquine-resistant strains of the malaria parasite Plasmodium falciparum, with galipinine being the most active compound (IC₅₀ 0.09–0.9 µg cm^–3 for the resistant strain).³³ The same four alkaloids were cytotoxic towards human fibroblast cells, but in this case cuspareine was the most effective (IC₅₀ 5.8–8.5 µg cm^–3). 2-Propylquinoline 13, from the species G. longiflora, has proved to be clinically effective as a trypanocidal agent in mice chronically infected with Trypanosoma cruzi, an encouraging result for the treatment of conditions such as Chagas disease.³⁴

Semecarpifoline 14 is an atypical metabolite from the root bark of the Taiwanese tree Melicope semecarpifolia, which is better known as a source of furoquinoline alkaloids.¹⁶ The positions of the substituents in this new quinolin-2-one were ascertained by HMBC and nOe experiments. The isolation of 14 marks the first occurrence of a methoxymethyl substituent in a rutaceous quinoline alkaloid, although the authors seem not to have recognised that this is a precedent-setting feature.

¹H and ¹³C NMR spectra have been reported in full for N-methyl-2-nonylquinolin-4-one 15, N-methyl-2-phenylquinolin-4-one 16 and 2-nonylquinolin-4(1H)-one 17, three well-known alkaloids recently isolated from the previously unexplored Brazilian shrub Raulinoa echinata.²³ Compounds 15 and 17 showed moderate antitrypanosomal activity against trypomastigote forms of T. cruzi (IC₅₀ 134.9 and 100.9 µg cm^–3, respectively), and the latter was also weakly fungicidal towards Leucoagaricus gongylophorus, a symbiotic fungus of leaf-cutting ants.

The New Caledonian tree Sarcomelicope megistophylla, noteworthy for producing unusual quinoline alkaloids, has yielded perhaps its most astounding natural product to date.²⁴ The optically inactive cyclomegistine 18, another in the series of putative oxidation products of an oxygenated precursor such as melicopicine 19 (cf. Ref. 35a), contains a cyclobuta[b]quinoline nucleus that is unprecedented both in nature and as a synthetic product. While spectroscopic methods pointed to the gross structure, the four-membered ring and the stereochemistry of its substituents could only be confirmed unambiguously by X-ray diffraction analysis. A plausible biosynthesis from melicopicine involves oxidative cleavage of ring A to give a butadiene intermediate 20, photochemical electrocyclic ring closure of which leads to the observed product. Interestingly, when melicopicine was treated with hydrogen peroxide in acetic acid while being irradiated, cyclomegistine was isolated in low yield. The new alkaloid was moderately cytotoxic towards L1210 leukaemia cells (IC₅₀ 80 µM).

A number of short synthetic approaches to simple naturally occurring 2-substituted quinolines merit brief mention (Scheme 1). Imino Diels–Alder reaction between the benzaldimine 21 and ethyl vinyl ether could be induced by samarium iodide to give the alkaloid 2-phenylquinoline 22 in modest yield.³⁶ Various ruthenium catalysts, and in particular the Grubbs catalyst RuCl₂(CHPh)(PCy₃)₂, induced oxidative cyclisation of 2-aminobenzyl alcohol 23 with a range of enolisable ketones to give quinoline products, among them the alkaloids 22, 24 and 25, by a modified Friedländer reaction.³⁷ Irradiation of the trans-2-aminocinnamoyl derivatives 26 in acetonitrile induced trans–cis isomerisation and subsequent cyclisation to give the quinolines 22 and 24 in good yield.³⁸ Cross-coupling of 2-chloroquinoline 27 with phenylmagnesium halides could be accomplished in the presence of a palladium catalyst to give 2-phenylquinoline 22 in 80% yield,³⁹ or somewhat less efficiently (65%) with iron(III) tris(acetylacetonate).⁴⁰

	Scheme 1 Reagents: i, EtOCHCH2, SmI2(THF)2(0.1 equiv.), CH2Cl2, rt; ii, RuCl2(CHPh)(PCy3)2(0.01 equiv.), KOH (1 equiv.), ketone (2 equiv.), dioxane, 80 °C; iii, h(high pressure Hg lamp, Pyrex filter), MeCN, 15–25 min; iv, PhMgCl, Pd(dba)2(5 mol%), dppf (5 mol%), THF, –5 °C; v, PhMgBr, Fe(acac)3(10 mol%), THF, –30 °C.

A group of structurally unprecedented polycyclic quinolinone alkaloids formally incorporating a 3-geranylquinoline motif was isolated from the Australian rutaceous plant Eriostemon australasius ssp. banksii (= E. banksii) in 1993⁴¹ (cf. Ref. 35b). One of these compounds, trans-erioaustralasine 28, has recently turned up in an apparently taxonomically remote source, Halfordia kendack, together with another four novel alkaloids of similar structure.¹⁰trans-Deacetoxyerioaustralasine 29 gave essentially the same NMR spectra as 28, but the unusual N-acetoxymethyl substituent was replaced by N-methyl. The remaining three minor alkaloids, trans-deacetoxyerioaustralasine hydrate 30, trans-erioaustralasine hydrate 31 and trans-1-epideacetoxyerioaustralasine hydrate 32, are 1,2-diols formally derived by hydrolysis of the epoxide rings of 28 and 29. The relative stereochemistry of 30 was deduced by analysis of coupling constants and NOESY experiments, while minor differences in the spectra of 30 and the other two diols led to the assignment of structures 31 and 32.

The novel tricyclic hemiterpenoid alkaloid haplophytin-A 33, isolated from a methanol extract of Haplophyllum acutifolium, is a simple methoxy derivative of the common alkaloid flindersine 34, which was also found in the extract.¹¹ However, the structure reported for haplophytin-B, another apparently new metabolite from the same source, is actually identical to that of the well-known alkaloid evoxine 35—confusingly, also known as haploperine. The name haplophytin-B should thus be abandoned in favour of the long-established name evoxine.⁴² Neither 33 nor 35 showed activity when tested against various bacteria and fungi.

Neither optical rotations nor absolute configurations were reported for the three linearly fused tricyclic quinoline alkaloids 8,9-dimethoxygeibalansine 36, 7,8-dimethoxymyrtopsine 37 and 7,8-dimethoxyplatydesmine 38 when they were first described in 1992 as metabolites of the New Caledonian plant Dutaillyea baudouinii⁴³ (cf. Ref. 35c), or in a later isolation of 37 and 38 from Dictamnus dasycarpus⁴⁴ (cf. Ref. 35d). The recent isolation of the same three alkaloids from a methanol extract of the root bark of Melicope semecarpifolia has now permitted some stereochemical inferences to be drawn.¹⁶ Analogy with the known (3S)-(+)-9-methoxygeibalansine, for instance, suggested that the laevorotatory pyrano[2,3-b]quinoline 36 ([]_D –18.6, c 0.065, CHCl₃) should have the (3R) configuration as illustrated. A similar comparison drawn between (2R)-(+)-8-methoxyplatydesmine and 38 ([]_D –10.3, c 0.16, CHCl₃) led to assignment of the (2S) configuration for the latter. Although the absolute configuration of (+)-7,8-dimethoxymyrtopsine 37 ([]_D +16.2, c 0.165, CHCl₃) was not deduced, the NOESY spectrum indicated that the hydrogen substituents at C-2 and C-3 were cis to each other. Interestingly enough, the parent (3R)-(–)-geibalansine 39 was recently isolated as a new natural product from the stem bark of Zanthoxylum hyemale, and its absolute configuration was established by the Horeau method.³² The absolute configuration of the (3S)-(+)-enantiomer, a metabolite of Geijera balansae, was established as recently as 2000 by total synthesis⁴⁵ (cf. Ref. 35e).

Several short syntheses of terpenoid quinoline alkaloids and related model systems were published during the review period. The first reported synthesis of the simple hemiterpenoid alkaloid orixiarine 40 has as its central feature the formation of the 3-substituted quinolin-2-one 41 by condensation of the malonate derivative 42 with N-methylaniline at 200 °C, after which sequential treatment with phosphorus oxychloride and sodium methoxide yielded the target alkaloid.⁴⁶ A new one-pot synthesis of pyrano[3,2-c]quinolin-5-ones by reaction of 4-hydroxyquinolin-2-ones 43 or 44 with ,-unsaturated aldehydes in the presence of ytterbium(III) trifluoromethanesulfonate in acetonitrile at reflux yielded, among other products, flindersine 34 (50%) from 43 and prenal, N-methylflindersine 45 (57%) from 44 and prenal, and zanthosimuline 46 (55%) from 44 and citral.⁴⁷ Another potentially useful route to pyrano[3,2-c]quinolin-5-one alkaloids, applied so far only to model systems, involved the in situ formation of the 3-methylenequinoline-2,4-dione 47 from 44 and paraformaldehyde in boiling dioxane, followed by Diels–Alder trapping with various alkenes to give products of general structure 48.⁴⁸ Also worth noting is the finding that the manganese(III) acetate-mediated radical-reaction of 44 with alkenes to give furoquinolinones, e.g., 49, could be improved when carried out in the ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate.⁴⁹

While hemiterpenoid 2,3-dihydrofuro[2,3-b]quinoline alkaloids (cf. Section 1.3 above) invariably betray their 3-prenylquinoline origins by bearing a three-carbon substituent at C-2, the substituent is lost in their biosynthetically more advanced furo[2,3-b]quinoline cousins. It is thus intriguing that two unprecedented 2-acetylfuro[2,3-b]quinoline alkaloids have been isolated from the root bark of Melicope semecarpifolia.¹⁶ The structures of the two compounds, 2-acetylevolitrine 50 and 2-acetylpteleine 51, were elucidated spectroscopically, with one- and two-dimensional NMR spectroscopic methods providing unambiguous support for the structures and the location of the substituents. Once again, however, the authors seem not to have recognised the novelty of their discovery. It is worth noting that the antiplatelet aggregation activity of the plant extract was traced to a suite of more conventional furoquinoline alkaloids, among them confusameline 52, dictamnine 53, evolitrine 54, haplopine 55, kokusaginine 56 and skimmianine 57.

The X-ray crystal structure of the known alkaloid flindersiamine 58, isolated for the first time from Raulinoa echinata, has been reported.²³ This alkaloid and its congeners kokusaginine, skimmianine and maculine 59 showed antifungal activity against Leucoagaricus gongylophorus, but were ineffective towards Trypanosoma cruzi (cf. Section 1.2).

A further hitherto unprecedented variation in the oxidation pattern found in the astonishing suite of quinoline alkaloids from Sarcomelicope megistophylla (cf. Section 1.2) involves the oxidation of the aromatic ring up to the quinone level.²⁵ Megistoquinone I 60, fully characterised by spectroscopic techniques, is probably derived biogenetically by oxidation of a precursor such as acronycidine 61, a major alkaloid from the same source. Indeed, treatment of 61 with nitric acid yielded 60, an experiment first reported in 1950 during the structural elucidation of 61 by degradation experiments.⁵⁰ A second metabolite, megistoquinone II 62, is apparently derived from 60 by hydrolysis of the furan ring followed by methylation of the resulting enol. While a number of quinoline-5,8-quinones have been found in sponges, this is the first report of their occurrence in terrestrial plants. Both new alkaloids showed antibacterial properties, with minimum inhibitory concentrations ranging from 2.35 to 5.25 mg cm^–3 for 60 and 0.73 to 1.23 mg cm^–3 for 62. Acronycidine itself was inactive in the same battery of tests.

The first example of a naturally occurring furo[3,2-g]quinoline alkaloid has been reported as a constituent of the stem bark of Zanthoxylum hyemale, a South American plant with a history of use in folk medicine.³² The structure of this unique compound, hyemaline 63, was established by spectroscopic methods, and the positioning of substituents was based on HMBC experiments.

Haplopine 55, skimmianine 57, -fagarine 64 and other furoquinoline alkaloids that occur in the genus Haplophyllum have been found to show more pronounced estrogenic activity than dihydrofuroquinoline, quinolin-2-one and quinolin-4-one alkaloids from the same genus.⁵¹ The related alkaloid dictamnine 53 has proved to be a feeding deterrent against two insect pests (Sitophilus zeamays and Trilobium castaneum) responsible for spoilage of stored products.⁵²

Two new glucoside derivatives of antidesmone 65, a recently discovered alkaloid from the African tree Antidesma membranaceum (Euphorbiaceae) (cf. Ref. 35f), have been detected in leaf extracts of the same plant.¹ The salient structural features of the novel metabolites, (+)-(17RS)-17-(-D-glucopyranosyloxy)antidesmone 66 and (+)-(17RS)-8-deoxo-17-(-D-glucopyranosyloxy)antidesmone 67, were revealed by thorough NMR and mass spectrometric analysis. Although both compounds were mixtures of diastereomers at C-17 in the side chain, comparison with reported ¹H NMR spectroscopic data for alk-2-yl -D-glucopyranosides allowed unambiguous differentiation of the (17R) and (17S) forms, and integration of the methyl signals indicated the (R)/(S) ratio in 66 and 67 to be 3 : 1 and 1.3 : 1, respectively. Analysis of coupling constants showed that the proton at C-5 was in an equatorial or pseudoequatorial position in both alkaloids, while the maxima in the CD spectrum of compound 66 were the same as in antidesmone itself, from which the (S) absolute configuration at C-5 was inferred. The presence of the new alkaloids is thought to imply a biological function for antidesmone as a vehicle for the storage or transport of glucosides. In the meantime, antidesmone has been found to display potent and selective antitrypanosomal activity against Trypanosoma cruzi, the pathogenic agent of Chagas disease (IC₅₀ 0.02 µg ml^–1), but little or no activity against T. brucei rhodesiense, an extracellular protozoan parasite responsible for African sleeping sickness, or against Leishmania donovani, the cause of visceral leishmaniasis.⁵³

Three new carbostyril derivatives have been reported from atypical plant sources and characterised by spectroscopic methods. Bioactivity-guided fractionation of an ethyl acetate extract of the Chinese medicinal plant Aquilegia ecalcarata (Ranunculaceae) led to the isolation of 7-hydroxy-4-[5-(hydroxymethyl)-2-furyl]quinolin-2(1H)-one 68, the furyl substituent of which is unprecedented in a quinoline alkaloid.³ The purified alkaloid was moderately cytotoxic towards two human cancer cell lines (IC₅₀ 8.8–10.1 µM) in in vitro tests. The less exotic alkaloid methyl 6-hydroxy-2-oxo-1,2-dihydroquinoline-4-carboxylate 69 was isolated from the aleurone (protein-containing) layer of the dark purple anthocyanin-pigmented rice cultivar Oryza sativa cv. Heugjinmi (Gramineae).¹⁹ Structural characterisation of the new alkaloid and its readily prepared acetate derivative by spectroscopic methods was straightforward. This is the first report of the compound from a natural source, although it has previously been synthesised.⁵⁴ Compound 69 may well act as an anti-oxidant in the plant, since it was found to exhibit moderate anti-oxidative activity in a radical-scavenging assay (IC₅₀ 36.4 µg ml^–1). A different rice cultivar, Oryza sativa cv. Mihyangbyo, yielded the corresponding 6-methoxy compound 70, which exhibited moderate antineoplastic activity towards the human leukemia cell line U937 (IC₅₀ 118.1 µg ml^–1).²⁰

A brief note describing the constituents of the bark of Peruvian Ladenbergia oblongifolia, a previously unexplored plant of the Rubiaceae with no apparent history of medicinal use, reported the isolation and full spectroscopic characterisation of the known alkaloids cinchonidicinol 71 and epicinchonicinol 72, as well as a mixture of the epimeric alkaloids dihydrocinchinicinol and dihydrocinchonidicinol 73.¹⁴ The latter may well be new alkaloids, since no previous reference to them in Chemical Abstracts could be traced. None of these compounds showed antimalarial activity towards a chloroquine-sensitive strain of Plasmodium falciparum. In this regard, several synthetic analogues of known antimalarial alkaloids such as quinine and cinchonidine having the general structure 74 showed potent activity against a chloroquine resistant strain of P. falciparum, and may be promising alternatives for currently used drugs in antimalarial chemotherapy.⁵⁵

Dihydroschoberine 75 is a novel decahydroquinoline alkaloid isolated from a chloroform extract of the aerial parts of Nitraria sibirica (Nitrareaceae, Zygophyllaceae).¹⁸ The natural product, an optically inactive oil, proved to be identical to the product obtained by catalytic hydrogenation of schoberine 76, a well-known metabolite of the genus Nitraria, in acetic acid. The stereochemistry of the ring junction in the new alkaloid was not reported, but is presumably unchanged from that in schoberine itself.

Bioassay-guided HPLC fractionation of the fermentation broth of the actinomycete strain Micromonospora sp. IM 2670 yielded the well-known quinolinequinone antibiotic streptonigrin 77 and its novel N-(1-methyl-2-oxopropyl) derivative 78.¹⁷ The structure of 78 was deduced largely on the basis of NMR spectra and chemical shift comparisons with streptonigrin and a model N-(1-methyl-2-oxopropyl)aniline. Both compounds induced apoptosis in human neuroblastoma SH-SY5Y cells containing wild-type p53 (a tumour-suppressing protein functioning as a key component of the cellular emergency response mechanism), although streptonigrin was more active (IC₅₀ 0.05 µM, vs. 0.9 µM).

The isolation of quinolactacin A from the entomopathogenic fungus Penicillium sp. EPF-6 was reported in last year's review^56,57 (cf. Ref. 35g). Although a novel 3-(sec-butyl)-2,3-dihydro-1H-pyrrolo[3,4-b]quinoline-1,9(4H)-dione structure was proposed for the alkaloid, the stereochemistry at the two stereogenic centres was not elucidated. However, two diastereomeric metabolites having the same gross structure, isolated more recently from Penicillium citrinum 90648 and thoroughly characterised by a range of spectroscopic techniques, cast light on this stereochemical ambiguity.²¹ On the basis of significant nOe differences involving the NH, NCH₃ and 3-H protons on the one hand, and the protons on the conformationally restricted sec-butyl group on the other, the authors proposed relative (S*,R*) and (S*,S*) stereochemistries for quinolactacins A1 and A2, respectively, as shown in 79 and 80. H-3 and the methyl group of the sec-butyl chain appear to be anti in the former compound, while H-3 and the methylene group are anti in the latter. The chemical shifts of quinolactacin A2 were very similar to those previously reported for quinolactacin A, to which structure 80 should now also be assigned. Quinolactacin A2 was found to be a more pronounced inhibitor of acetylcholinesterase than quinolactacin A1 (IC₅₀ 19.8 vs. 280 µM); its selectivity for acetylcholinesterase rather than butyrylcholinesterase (IC₅₀ 650 µM) may offer opportunities for exploration in the treatment of Alzheimer-type dementia.

2-Alkyl-4-hydroxyquinolines (or their quinolin-4-one tautomers) have long been known as metabolites of Pseudomonas organisms, from which source their alternative name of pseudans is derived. While the recent isolation of the 2-heptyl and 2-nonyl homologues 81 and 82 (pseudans VII and IX) from Pseudomonas aeruginosa ATCC 15692 was not in itself novel, the point of interest is that these compounds turned out to be novel membrane-associated iron chelators (MAIC) in the cytoplasmic membrane of iron-rich bacterial cells.⁵⁸ The structure of 82 was confirmed by spectroscopic comparison with a sample prepared by alkylating the dianion of 2-methylquinolin-4-ol with octyl bromide. Although in vitro complexation of synthetic pseudan IX with iron(III) chloride could not be verified spectroscopically, pink micelles similar to those found in the ethanol extract of cell membranes were formed on mixing the precursors. Also, after growth of the cells in the presence of ⁵⁵FeCl₃, the radioactivity co-eluted with pseudan IX from an LH-20 Sephadex gel filtration column.

The simple antibiotic N-mercapto-4-formylcarbostyril 83 is a novel metabolite of Pseudomonas fluorescens (strain G308), a microorganism originally found growing on barley leaves.²² The structure was deduced from its relatively simple spectra and those of its easily prepared S-acetyl derivative. Although no mention was made of the fact, this is the first occurrence of the N-mercapto substituent—effectively, part of the rare N-mercaptoamide functional group—in a quinolinone alkaloid. The purified antibiotic exhibited good activity against a range of plant pathogenic fungi, including Fusarium oxysporum, F. culmorum, Cladosporium cucumerinum and Colletotrichum lagenarium.

Two new octadepsipeptides bearing pendent 3-hydroxyquinaldic acid chromophores have been isolated from the culture broth of the Streptomyces strain SNA15896, collected from a Japanese soil sample.^29,30 The symmetrical dimer (–)-SW-163C 84 incorporates a central disulfide bridge, while the related compound (–)-SW-163E 85 possesses the same periphery as 84 but differs only in its inclusion of an internal dithioacetal linkage. Compound SW-163E is, in fact, the S-ethyl analogue of three antibiotics reported in 1989, and coded as UK-63,052 86, UK-65,662 87 and UK-63,598 88.⁵⁹ The stereochemistry of all these natural products has not been elucidated, although it could no doubt be inferred from the nature of the amino acid constituents. Both SW-163C and SW-163E displayed good antibacterial activity, especially against Gram-positive bacteria. More importantly, potent antitumour activity was demonstrated in in vitro tests against various murine and human tumour cell lines, with 85 being about a hundredfold more active than 84 (IC₅₀ 0.2–1.6 vs. 17–140 nM, respectively). When in vivo activity was assessed in mice implanted with P388 leukaemia, SW-163E was remarkable in prolonging life span at a dose of 0.01 mg kg^–1, but proved to be acutely toxic at higher doses (LD₅₀ 0.6 mg kg^–1; cf. > 100 mg kg^–1 for 84).

Several model studies aimed at the synthesis of quinoline-containing natural products require brief mention. Evolving approaches to the synthesis of lavendamycin 89 by Holzapfel et al. have involved optimising the palladium-catalysed aminocarbonylation of 2-chloroquinolines with amines in the presence of carbon monoxide to give intermediates such as 90, Bischler–Napieralski cyclisation of which produces the antibiotic's pentacyclic core.⁶⁰ Nicolaou and co-workers have reported a stereocontrolled synthesis of the macrocycle 91, envisaged as a key intermediate in the synthesis of the complex thiopeptide antibiotic thiostrepton.^61,62 The dihydroquinoline building block 92 employed in this route was prepared in a multi-step procedure from quinaldic acid. Coincidentally, Hashimoto and co-workers almost simultaneously reported a similar multi-step synthesis of the dihydroquinoline 93 starting from 5,6,7,8-tetrahydroquinoline.⁶³ Compound 93 is proposed as an intermediate en route to siomycin D₁, another member of the thiostrepton group of peptide antibiotics.

Several new lepadins, a group of decahydroquinoline alkaloids isolated from ascidians (tunicates or sea squirts) collected on the Great Barrier Reef, have been described in two independent publications that appeared almost simultaneously. A new species of the animal belonging to the genus Didemnum yielded four novel lepadins, the structures of which were elucidated with the aid of a comprehensive suite of spectroscopic techniques.⁵ In the case of (+)-lepadin D rel-94, for example, analysis of coupling constants in combination with nOe experiments indicated cis-fusion of the decahydroquinoline nucleus, chair conformations of both rings, a trans-diaxial arrangement of the C-2 and C-3 methyl and hydroxy substituents, and a 1,3-diaxial relationship between 5-H and 8a-H. This analysis established the relative stereochemistry of the stereogenic centres on the ring system, but both the configuration of the hydroxy group on the side chain and the molecule's absolute configuration remain unknown despite attempts to form a p-bromobenzoyl derivative for crystallographic analysis. Similar considerations led to the elucidation of the relative structures 95 and 96 for (–)-lepadins E and F, respectively, the last-named differing from the others in having an equatorial methyl substituent at C-2. The fourth alkaloid was simply the protonated form of lepadin D, probably with chloride as counter-ion. All the new alkaloids displayed biological activity: lepadins D–F were weakly antifungal, while the mildly cytotoxic 95 and 96 were also moderate inhibitors of tyrosine kinase p56. Most importantly, all four natural products showed significant antitrypanosomal and antiplasmodial activity, the most potent compound being lepadin F. The isolates may thus serve as novel lead compounds for the development of new antimalarial drugs.

The three new lepadins described in the second publication were isolated from the ascidian Aplidium tabascum.² Apart from its optical rotation (+5.5 rather than –1.5), lepadin F proved to be identical to the compound described above; slight differences in the reported ¹H NMR spectra were a consequence of measurement in different solvents (C₆D₆vs. CDCl₃). It is unlikely that the two isolates are mirror images; it is more probable that minor impurities have distorted the small specific rotations recorded. The structure drawn for lepadin F in this publication was actually ent-96, but the relative stereochemistry in the two representations is the same. [In this regard, it is pertinent to note that a recent enantioselective total synthesis of three lepadins, e.g., (–)-lepadin A 97, by Kibayashi's group⁶⁴ (cf. Ref. 35h) has established a correlation between absolute configuration and optical rotation that may be valid for other members of the lepadin family.] (+)-Lepadin G rel-98 is a dienoate analogue of lepadin F, while (+)-lepadin H rel-99 is the octa-2,4-dienoate ester of lepadin D. In this study, molecular modelling was also used for ascertaining the minimum energy conformations of the alkaloids, and substantiated the conformational deductions based on NMR data. Interestingly, the preferred cis-fused twin-chair decahydroquinoline ring system for lepadins G and H places the ring nitrogen equatorial, as shown in 100, whereas it is known that the nitrogen is axial in lepadin A, as depicted in 101.

The Okinawan sponge Hyrtios erecta was the source of the two simple brominated carbostyrils 102 and 103.¹³ Both are novel natural products, although the former was previously reported as a synthetic compound.⁶⁵ The compounds were good but unselective inhibitors of the neural and inducible isozymes of nitric oxide synthase. Another new alkaloid, represented as the 2-hydroxyquinoline 104 but more probably the tautomeric carbostyril, was isolated from whole body extracts of the centipede Scolopendra subspinipes mutilans, which is used in traditional Chinese and Korean medicines.²⁶ Treatment of 104, which has been named scolopendrine, with diazomethane gave 105, HMBC correlations on which assisted in the location of the substituents. Scolopendrine's benzyl group at C-7 and sulfate group at C-8 are unprecedented substituents among the quinoline alkaloids.

Continuing uncertainty over the proposed structure 106 for the sponge metabolite halitulin has prompted Banwell and co-workers to undertake the short convergent synthesis shown in Scheme 2.⁶⁶ Palladium(0)-mediated coupling of the 3,4-diiodopyrrole 107 with pinacolborane followed by Suzuki–Miyaura cross-coupling of the resulting bis(boronic ester) 108 with the bromoquinoline 109 produced the desilylated pyrrole 110 in an overall yield of 29% based on 107. The structure of compound 110 was confirmed by single-crystal X-ray crystallography. The 3-methylazecane 111, previously prepared by Banwell's group⁶⁷ during a synthesis of the putative alkaloid haliclorensin (originally thought to be 112, but now known to be a diazacyclotetradecane⁶⁸), was converted in three steps into the unstable trifluoromethanesulfonate 113, which readily alkylated the potassium salt of pyrrole 110 to give halitulin tetramethyl ether 114 in 87% yield. Unfortunately, all attempts to cleave the methyl ethers failed. However, detailed comparison of the spectroscopic data reported for halitulin with those obtained on 110 and 114 showed disturbing discrepancies in both the cyclic amine and the 3,4-bis(quinolin-5-yl)pyrrole portions of structure 106. In the interim, Usuki et al. have reported an enantioselective synthesis of (R)-(+)-112⁶⁹ (renamed isohaliclorensin), the ¹³C NMR data for which were claimed to be comparable with those for the corresponding substructure of halitulin. Final resolution of the disputed structure of halitulin must thus await complete total synthesis or re-collection and further spectroscopic investigation.

	Scheme 2 Reagents: i, pinacolborane (6 equiv.), PdCl2(dppf)(14 mol%), Et3N (10 equiv.), dioxane, 85 °C, 24 h; ii, 109(1.4 equiv.), Pd(PPh3)4(10 mol%), Na2CO3(2 M, 10 equiv.), PhMe–MeOH (6 : 5), 65 °C, 24 h; iii, H2CCHCO2Me, AcOH, reflux; iv, LiAlH4, THF, 18 °C; v, Tf2O, 2,6-di-tert-butyl-4-methylpyridine, CH2Cl2, 0 °C; vi, KHMDS +110, THF, 0 °C, then add 113, 0 °C.

The fascinating aspect of the synthesis of decahydroquinoline 195A 115 (commonly known as pumiliotoxin C) by Mori and co-workers is the use of a nitrogen fixation process to introduce the nitrogen atom as an N-1 synthon (Scheme 3).⁷⁰ Nitrogen, bound in a Ti–N complex prepared from titanium(IV) tetrakis(isopropoxide), lithium metal and trimethylsilyl chloride in the presence of molecular nitrogen gas, was efficiently incorporated into the keto-alkyne precursor 116 to give the azabicyclic product 117 in 79% yield. Catalytic hydrogenation over palladium on carbon followed by protection of the basic nitrogen site yielded the cis-fused decahydroquinoline 118 as the major diastereomer (50%), probably because of the easier approach of the reagent from the precursor's convex face. The synthesis of the racemic target compound rac-115, characterised as the hydrochloride salt, was completed by standard methods as illustrated.

	Scheme 3 Reagents: i, Ti–N complex [from Li–Ti(OPri)4–Me3SiCl (10 : 1 : 15)+ N2], CsF, THF, –78 °C, then rt, 20 h; ii, H2(1 atm), 5% Pd/C, MeOH; iii, ClCO2Bn, K2CO3, CH2Cl2, 0 °C to rt; iv, DIBAL-H, PhCH3, –78 °C, 15 min, then MeOH, aq. NH4Cl, rt; v, Ph3PCH2, THF, rt.

The use of tetra-O-pivaloyl--D-galactopyranosylamine 119 as a chiral auxiliary in enantioselective and stereochemically complementary syntheses of several stereoisomers of pumiliotoxin C has been investigated by Kunz and co-workers.⁷¹ Imines 120 formed from this precursor underwent highly diastereoselective annulation with the diene 121 in the presence of zinc(II) chloride to give the dihydropyridones 122 and 123 in 68% yield (Scheme 4). Diastereoselective conjugate addition of appropriate cuprates then yielded the 2,6-cis-disubstituted piperidinones 124 and 125 which, with suitable manipulation of the functionalised side chains, afforded octahydroquinolinones 126 and 127 possessing opposite absolute configurations at the C-2 and C-8a sites. Treating isomer 127 with lithium dimethylcuprate in the presence of trimethylsilyl chloride proceeded with excellent diastereoselectivity to give the expected (5R)-methyl product, but resulted in the unanticipated formation of a trans-fused decahydroquinoline 128, apparently because the chiral auxiliary steered protonation of the intermediate enolate. To substantiate the involvement of the chiral auxiliary in 127, exchange with a phenoxycarbonyl protecting group prior to conjugate addition yielded the expected cis-fused decahydroquinoline 129, a compound previously prepared by Comins and Dehghani in their 1993 synthesis of (–)-pumiliotoxin C.⁷² Epimerisation of the N-phenoxycarbonyl analogue of 128 to give 129 could also be rapidly effected upon treatment with triethylamine. Alternatively, replacing the auxiliary in 128 with a benzyloxycarbonyl protecting group before defunctionalising the ketone via the dithiolane intermediate 130 eventually completed the total synthesis of trans-4a-epi-pumiliotoxin C 131, which was characterised by spectroscopy and X-ray crystallography as the laevorotatory hydrochloride salt. Finally, similar transformations performed on the octahydroquinolinone 126 as shown in the Scheme resulted in the synthesis of cis-4a-epi-perhydrodecahydroquinoline 219A hydrochloride 132, which gave NMR spectroscopic data that agreed with those reported in the literature.

	Scheme 4 Reagents: i, R(CH2)3CHO (R = H, vinyl), 4 molecular sieves, pentane; ii, ZnCl2·Et2O, THF, –20 °C; iii, EtOCH(Me)O(CH2)4MgCl, CuBr·SMe2, Me3SiCl; iv, n-PrMgCl, CuCl, BF3·Et2O, THF, –78 °C; v, PPTS, CH2Cl2; vi, TPAP, N-methylmorpholine-N-oxide, 4 molecular sieves, CH2Cl2, rt; vii, NaOH, dibenzo-18-crown-6, C6H6, reflux; viii, K2OsO4·2H2O (cat.), NaIO4, dioxane–H2O; ix, Me2CuLi, THF, –78 °C, then Me3SiCl; x, Bu4NF, THF, –20 °C; xi, aq. HCl, MeOH, rt; xii, ClCO2Ph, aq. NaHCO3; xiii, Me2CuLi, BF3·Et2O, THF, –78 °C; xiv, Et3N, THF, 2 min; xv, ClCO2Bn, aq. NaHCO3; xvi, HSCH2CH2SH, BF3·Et2O, CH2Cl2, 0 °C to rt; xvii, H2, Raney Ni, PriOH, rt; xviii, aq. HCl; xix, n-PrMgCl, CuBr·SMe2, THF, –78 °C, then 126, Me3SiCl; xx, NaHMDS, THF, –78 °C, then 2-N(SO2CF3)2-5-Cl-pyridine, –78 °C to –20 °C; xxi, H2, 5% Pd/C, Li2CO3, MeOH, rt.

Few new quinazoline alkaloids were reported during the period under review. 3-(2-Hydroxyphenyl)quinazolin-4(3H)-one 133 has been claimed as a new alkaloid from the roots of Isatis indigotica,⁷³ but in fact it was reported from the same source in 1997—by the same research group!⁷⁴ The 3a-methoxy-substituted vasicinone analogue 134 (incorrectly named as 3-methoxyvasicinone) was isolated from leaves of Eranthemum nervosum, and its structure was established by electrospray mass spectrometry in combination with IR and NMR spectroscopy.⁷⁵ An unusual feature in the structural assignment of 3-hydroxyrutaecarpine 135, a new natural product isolated from leaves and twigs of Leptothyrsa sprucei (Rutaceae), was the use of gradient ¹H–¹⁵N HMBC NMR spectroscopy at natural nitrogen abundance for establishing some of the skeletal connectivities.⁷⁶ The more complex peptide-derived alkaloid (–)-serantrypinone 136, obtained from mycelial extracts of the fungus Penicillium thymicola IBT 5891, is a hydroxylated congener of the known spiroquinazoline alantrypinone 137, and effectively arises by replacing an alanine residue with serine.⁷⁷ The similarity in the CD spectra of the two metabolites suggested that serantrypinone and alantrypinone share the same (3R,14R,17S) absolute configuration.

The alkaloids of Adhatoda vasica, an important plant in Ayurvedic medicine, are principally derivatives of vasicine 138 and vasicinone 139. Recent tests on the unseparated alkaloid fraction showed anti-inflammatory activity in the HET-CAM assay approximately equal to that of hydrocortisone.⁷⁸ Several synthetic tricyclic vasicinone analogues of general structure 140 proved to be effective bronchodilators in in vitro tests with guinea pig tracheal chain pre-contracted with acetylcholine, the most active compound being the 8,9-dimethoxy derivative.⁷⁹ Isaindigotone 141, a naturally occurring vasicinone analogue isolated from Isatis tinctoria, and seven synthetic derivatives displayed a range of biological effects related to anti-inflammatory activity, including inhibition of superoxide, nitric oxide and prostaglandin-E₂ production.⁸⁰ Isaindigotone itself was a superior scavenger of superoxide generated in the hypoxanthine/xanthine oxidase system (IC₅₀ 42.2 nM), no doubt because of its free phenolic group, whereas derivatives 142 and 143 were, respectively, the most effective inhibitors of PGE₂ (IC₅₀ 80 nM) and nitrite (1.8 µM) accumulation in mouse macrophages.

The efficient extraction and quantitative analysis of tryptanthrin 144, the anti-inflammatory principle of Isatis tinctoria (woad), have been reproducibly achieved by accelerated solvent extraction (ASE) in a commercial instrument followed by HPLC with electrospray-MS detection.⁸¹ Analysis of over 70 woad samples from different sources indicated a highly variable tryptanthrin content of 0.56–16.74 × 10^–3 % by mass. In further pharmacological developments, tryptanthrin has been found to ameliorate artificially induced colitis in mice, at the same time suppressing weight loss, tissue damage and subsequent mortality;⁸² this relatively non-toxic alkaloid thus has potential for treating inflammatory bowel disease. It also effected cell differentiation and apoptosis of both monocytic (U-937) and promyelocytic (HL-60) leukaemia cells,⁸³ and suppressed the growth of azoxymethane-induced intestinal tumours in F344 rats.⁸⁴ Tryptanthrin has shown activity as an antitrypanosomal agent against Trypanosoma brucei; furthermore 11 synthetic analogues were also active, the most potent compounds, 145 (EC₅₀ 0.82 µM) and 146 (EC₅₀ 0.4 µM), being substantially more effective than the parent alkaloid (EC₅₀ 23.0 µM).⁸⁵ Another protozoan parasite, Leishmania donovani, was sensitive to no fewer than 27 synthetic analogues of tryptanthrin, 13 of the compounds displaying IC₅₀ values of less than 100 ng ml^–1 as well as being more toxic to the parasite than to mammalian cell lines; the most active compounds were 147 and 148 (IC₅₀ 16 ng ml^–1).⁸⁶ (For comparison, the clinically prescribed antileishmanial agent amphotericin has an IC₅₀ value of 416 ng ml^–1). Structure–activity relationships revealed a reasonable correlation between activity and some calculated and measured molecular properties such as molecular density, LUMO energies and redox potentials (determined by cyclic voltammetry), the implication of the latter two being that electron transfer involving the carbonyl groups might be a crucial factor in the mechanism of action of the compounds.

The revival of interest in the antimalarial alkaloid febrifugine 149 seems set to continue. A brief review by Takeuchi and Harayama includes useful information on the history, structure determination, synthesis, biological activity and structure–activity relationships of the alkaloid and its isomer isofebrifugine 150.⁸⁷ However, the unacceptable emetic effects of these two compounds have prevented their clinical use as antimalarial agents. In attempts to find analogues with improved properties, Oshima and co-workers investigated the biological activity of 151 and 152, products of condensation of the parent alkaloids with acetone, as well as eleven other synthetic derivatives of 149 and 150 and a further fifteen analogues of 151 and 152.⁸⁸ All compounds exhibited good in vitro antimalarial activity against P. falciparum, though none were as potent as febrifugine itself (EC₅₀ 7.0 × 10^–10 M). Additionally, all compounds were cytotoxic to FM3A mouse mammary cells, but only four of them, viz.153–156, were significantly more selective (> 500 : 1) towards the parasite than towards the mammary cells, 156 being a staggering 3100 times as selective. These findings prompted in vivo studies of the antimalarial efficacy of the more accessible racemates of 153–155 in mice infected with P. berghei. The first two proved to be more potent (ED₅₀ 1.3 and 2.9 µmol kg^–1) than chloroquine or artemisinin (ED₅₀ 4.7 and 17.7 µmol kg^–1), and only slightly less effective than febrifugine (ED₅₀ 1 µmol kg^–1). When injected intraperitoneally, they also prolonged the survival rate of infected mice. Thus this study appears to have revealed exciting new lead compounds for the development of novel antiplasmodial drugs.

Three simple syntheses of quinazoline alkaloids are illustrated in Scheme 5. A straightforward condensation between anthranilic acid 157 and various imidates of general formula RC(NH)OMe in boiling methanol produced a range of 2-substituted quinazolin-4(3H)-ones, among them the alkaloids 2-methylquinazolin-4(3H)-one 158 (42%) and glycosminine 159 (65%).⁸⁹ A short synthesis of deoxyvasicinone 160 showcasing the reductive cyclisation of N-(2-nitrobenzoyl)pyrrolidin-2-one 161 with carbon monoxide in DMF in the presence of triethylamine and a catalytic amount of selenium was also applied to the synthesis of substituted and six- to eight-membered ring C analogues 162 in yields that were generally above 75%.⁹⁰ The shortest synthesis of all, entailing reaction of isatoic anhydride 163 with lactams 164 (n = 1–3) under solvent-free conditions in a microwave oven, gave deoxyvasicinone in 92% yield within 6 minutes, and was also used for preparing the six- and seven-membered ring C congeners 162 (R = H; n = 2, 3) in 90% and 89% yields, respectively.⁹¹

	Scheme 5 Reagents: i, MeOH, 25 °C, 30 min, then 80 °C, 6 h; ii, CO (5 atm), Se (5 mol%), NEt3(1 equiv.), DMF, 100 °C, 10–20 h; iii, microwave irradiation (450 W), 6 min.

Treatment of anthranilamide 165 with the succinic anhydrides 166 and esterification of the intermediate succinanilic acids produced esters 167, chemoselective reduction of which with lithium aluminium hydride followed by aqueous work-up smoothly yielded the quinazolinones 168 and 169 (Scheme 6).⁹² The latter is a natural product, pegamine, and this is the first report of its synthesis (in 89% overall yield from 165). Both 168 and 169 underwent further ring closure under Mitsunobu conditions to give (–)-vasicinone 139 and deoxyvasicinone 160. The (S) absolute configuration of (–)-vasicinone was established by NMR spectroscopic analysis of its Mosher esters, a technique that also indicated its enantiomeric excess to be 97–98%. In an extension of this work, pegamine 169 was oxidised with pyridinium chlorochromate to give the masked aldehyde isovasicinone 170 (64%), Friedländer condensation of which with 2-aminobenzaldehyde followed by benzylic oxidation of the resulting quinoline completed a synthesis of luotonin F 171, an anti-tumour alkaloid originally obtained from Peganum nigellastrum.⁹³

	Scheme 6 Reagents: i, Et2O–C6H6–dioxane (2 : 2 : 1), rt, 2 h; ii, CH2N2, Et2O, 0 °C to rt; iii, LiAlH4, THF, 0 °C to rt, then aq. NH4Cl, evaporation in vacuo; iv, PPh3, EtO2CNNCO2Et, THF, rt; v, PCC, 4 molecular sieves, CH2Cl2, rt; vi, 2-aminobenzaldehyde, KOH, EtOH, reflux; vii, CrO3, H5IO6, DMF, rt.

The current popularity of luotonin A 172 as a synthetic target is a consequence of its cytotoxicity towards murine leukemia P-388 cells. A further three syntheses of this alkaloid were published during the review period. The key steps are shown in Scheme 7. Adaptation of the microwave synthesis described above entailed heating isatoic anhydride with the known pyrroloquinoline 173 under solvent-free conditions in a microwave oven to give luotonin A in 85% yield within seven minutes.⁹¹ Alternatively, the same pyrroloquinoline 173, prepared in a novel manner by Friedländer synthesis from 1-allylpyrrolidine-2,3-dione 174 and 2-aminobenzaldehyde followed by removal of the N-allyl protecting group, was acylated with 2-nitrobenzoyl chloride, after which reductive cyclisation completed the synthesis of 172.⁹⁴ In the third route, in situ conversion of the amide 175 into the O-silylimidate 176 in a sealed tube at 150 °C preceded spontaneous intramolecular hetero Diels–Alder reaction to give luotonin A in 46% yield.⁹⁵

	Scheme 7 Reagents: i, isatoic anhydride 163, microwave irradiation (450 W), 7 min; ii, 2-aminobenzaldehyde, p-TsOH (cat.), PhMe, reflux (Dean–Stark apparatus); iii, PdCl2(5 mol%), PPh3(20 mol%), DMF–H2O (4 : 1), reflux; iv, HCl (6 M), reflux; v, NaH, THF, 60 °C, then 2-nitrobenzoyl chloride, 0 °C to 50 °C; vi, Fe, AcOH–EtOH (1 : 1), reflux; vii, o-anisic acid, bis(2-oxo-3-oxazolidinyl)phosphinic chloride, Et3N, CH2Cl2; viii, Pd2(dba)3, dppf, CuCN, Et4N+ CN–, dioxane, reflux; ix, TMSCl, ZnCl2, Et3N, PhMe, 150 °C (sealed tube).

Much of the recent synthetic activity aimed at the synthesis of febrifugine analogues has centred on the stereoselective construction of the piperidinol moiety, as, for instance, in the asymmetric synthesis of model 2-substituted piperidin-3-ols such as 177 by Enders et al.⁹⁶ The ongoing researches of Kobayashi have concentrated on the ring opening of racemic⁹⁷ and enantiomerically pure⁹⁸ semicyclic N,O-acetals with nucleophiles in the presence of Lewis acid catalysts. The systematic model studies performed by these workers culminated in the diastereoselective transformation of (S)-178 into 179, which proceeded in 81% yield and a diastereomeric ratio of 86 : 14 (Scheme 8). This compound was converted via the 1 : 1 diastereomeric mixture of epoxides 180 into the 3-substituted quinazolin-4-one 181, oxidation and intramolecular reductive amination of which yielded the protected isofebrifugine analogue 182. The same workers had previously converted racemic 182 into (±)-isofebrifugine⁹⁹ (cf. Ref. 35i); similar hydrolysis of the enantiomer (+)-182 with boiling hydrochloric acid completed the synthesis of (+)-isofebrifugine 150. The less efficient alternative reaction of (3S)-178 with the silyl enol ether 183 was mediated by trimethylsilyl triflate to yield 184 (51%, syn : anti ratio 93 : 7), which was also converted into the isofebrifugine precursor 182 as illustrated.

	Scheme 8 Reagents: i, H2CCHCH2SiMe3(2 equiv.), TMSOTf (0.2 equiv.), MeCN, –20 °C; ii, MCPBA, CH2Cl2, rt; iii, quinazolin-4-ol, KOH (0.2 equiv.), MeOH, reflux; iv, Dess–Martin reagent, CH2Cl2, rt; v, Et3SiH, TMSOTf, MeCN, 0 °C; vi, HCl (6 M), reflux, 2 h; vii, (3S)-178, 183(2 equiv.), TMSOTf (2.5 equiv.), MeCN, rt.

The short synthesis of rutaecarpine 185 and related alkaloids shown in Scheme 9 commenced with reaction of anthranilate esters 186 with 4,5-dichloro-1,2,3-dithiazolium chloride (Appel's salt) 187 to produce the imines 188, which reacted with substituted tryptamines to yield the 2-cyanoquinazolin-4-ones 189.¹⁰⁰ Simply heating these intermediates with trifluoroacetic anhydride and hydrogen chloride gas completed syntheses of rutaecarpine 185, hortiacine 190, euxylophoricine A 191 and euxylophoricine D 192 in 90–95% yields. Also shown in the Scheme is another short synthesis of rutaecarpine, the key step of which was a Fischer indole synthesis using the readily prepared quinazolinedione 193.¹⁰¹

	Scheme 9 Reagents: i, pyridine, CH2Cl2, rt, 2–3 h; ii, CH2Cl2, rt, 24–31 h; iii, TFAA, HCl(g), 120–130 °C; iv, PhCHO, Ac2O, reflux; v, O3, CH2Cl2, –78 °C, then Me2S; vi, PhNHNH2, EtOH; vii, PPA, 180 °C.

The reaction of benzoxazine 194 with methyl anthranilate followed by thermal recyclisation, although rapidly giving the quinazolin-4-one 195 in a poor overall yield of 4.7%, was featured in a short new synthesis of circumdatin F 196 (Scheme 10).¹⁰² Electrophilic bromination of 195 yielded 197 as a mixture of two diastereomers, probably arising from restricted rotation around the N–aryl bond. Conversion of 197 into the amine 198 followed by thermally-induced lactam formation completed the synthesis of the target alkaloid 196. Similarly, acid-induced conversion of the bromoacetamide 199 into the corresponding quinazolinone followed by treatment with ammonia yielded sclerotigenin 200 (16% overall yield), while heating 199 and the methyl homologue 201 in DMF at reflux produced the oxygen analogues of sclerotigenin and circumdatin, 202 and 203, respectively, in 20% yield.

	Scheme 10 Reagents: i, methyl anthranilate, AcOH, reflux; ii, DMF, reflux; iii, Br2, NaOAc, HOAc, 60 °C; iv, NaN3, H2O–PriOH (1 : 5), reflux; v, H2(220 psi), 5% Pd/C, EtOH, rt; vi, MeCN, reflux.

Avendaño and co-workers have investigated the acylation of a range of diketopiperazines 204, prepared by standard methods from the respective N-Boc dipeptides, with 2-azidobenzoyl chloride via the silyl imidates 205 (Scheme 11).¹⁰³ With the glycine derivative of 204 (i.e., R = H), selective monoacylation on N(4) to give 206 was ascribed to a boat-like conformation of the silylated intermediate, with the indolyl substituent folding in such a way that N(1) was blocked. Selectivity was also good with the (S)-alanine derivative of 204 [R = (S)-Me], but less impressive with the (R)-alanine and (S)-valine analogues [R = (R)-Me and (S)-Prⁱ], which gave almost equal amounts of the N(1)-acylated products. All of the acylated products 206 could be cyclised by an intramolecular Staudinger reaction upon treatment with tributylphosphine to complete syntheses of (–)-glyantrypine 207, (–)-fumiquinazoline F 208, fumiquinazoline G 209 and fiscalin B 210, respectively. In an alternative synthesis of fiscalin B, treatment of the quinazolinedione 211 with lithium hexamethyldisilazide and N-Boc-3-indolylmethyl bromide gave a mixture of the 1,4-syn- and anti-disubstituted products 212 (46%) and 213 (31%). Deprotection of the minor isomer was accomplished with boron tribromide to give fiscalin B in 88% yield.¹⁰⁴ The syn preference in this example was somewhat surprising, since alkylation of the methyl analogue of 211 gave mainly 1,4-anti-disubstituted products,^104,105 a feature that was also borne out by SCF-MO calculations.¹⁰⁶ This research group has also reported a variety of model studies on the formation of the pyrazino[2,1-b]quinazoline-3,6-dione ring system viaN-acyliminium ion intermediates.^107,108

	Scheme 11 Reagents: i, TMSCl (2 equiv.), Et3N (2 equiv.), CH2Cl2, rt; ii, 2-N3C6H4COCl (1.1 equiv.), CH2Cl2, rt; iii, Bu3P, PhMe, rt; iv, LiHMDS (10 equiv.), THF, –78 °C, 10 min; v, N-Boc-3-indolylmethyl bromide (2 equiv.), THF, –78 °C; vi, BBr3, CH2Cl2, –78 °C.

The first reported syntheses of three related pyrazino[2,1-b]quinazoline-3,6-dione alkaloids, verrucine A 214, verrucine B 215 and anacine 216, have been accomplished by exploiting peptide assembly on Sasrin resin (Scheme 12).¹⁰⁹ For example, the resin-bound L-glutamine derivative 217 was sequentially condensed with anthranilic acid and Fmoc-protected L-phenylalanine chloride to give the resin-bound tripeptide 218. Intramolecular dehydration followed by treatment with piperidine, a general procedure developed by Wang and Ganesan¹¹⁰ and optimised by He and Snider¹¹¹ (cf. Ref. 35j), afforded amidine 219. Cyclisation with concomitant detachment from the resin was effected by overnight heating in a mixture of acetonitrile and 1,2-dichloroethane to give N-tritylverrucine A 220 in 17% overall yield from 217; only 0.8% of the corresponding 1,4-anti-disubstituted isomer was isolated. The removal of the trityl group was achieved reductively with triethylsilane in trifluoroacetic acid to give (+)-verrucine A 214, the specific rotation of which was much greater than for the natural product isolated from its fungal source. Similar reaction sequences employing D-phenylalanine and L-leucine afforded (+)-verrucine B 215 and (+)-anacine 216, respectively, in overall yields of 14.5% and 9.3% based on 217. The absolute configuration of the former, not assigned when it was originally isolated, has thus been established unambiguously. The originally proposed benzodiazepine structure 221 for anacine, revised to 216 after the isolation of the verrucines, has also been unequivocally refuted. It is worth noting that verrucine B and anacine underwent aerial oxidation in solution, the former yielding the hydroxy product 222. This type of reaction may explain the origin of the hydroxy group in related pyrazinoquinazoline alkaloids.

	Scheme 12 Reagents: i, 20% piperidine in DMF, 15 min, then repeat; ii, EDC (13.4 equiv.), anthranilic acid (12.1 equiv.), DMF or NMP, rt, 19 h; iii, Fmoc-L-Phe-Cl (5.1 equiv.), pyridine (29 equiv.), CH2Cl2, rt, 13 h, then workup, then repeat in DMF; iv, PPh3(12 equiv.), I2(11.1 equiv.), EtNPri2(25 equiv.), CH2Cl2, rt, 15 h; v, 20% piperidine in CH2Cl2, rt, 30 min; vi, MeCN–(CH2Cl)2(1 : 1), reflux overnight; vii, TFA–Et3SiH–CH2Cl2(2 : 2 : 1), rt, 15 min.

The advanced intermediate 223, previously used by Snider and Zeng in a synthesis of the Aspergillus metabolite fumiquinazoline A¹¹² (cf. Ref. 35k), has been transformed into another two complex fumiquinazolines by the same workers (Scheme 13).¹¹³ Condensation of 223 with a selenocysteine derivative, (R)-FmocNHCH(CH₂SePh)CO₂H, yielded the quinazoline precursor 224, subjection of which to the Ganesan cyclisation procedure (cf. Scheme 12) sequentially afforded the benzoxazine and amidine intermediates 225 and 226. Simply heating crude 226 in acetonitrile–acetic acid (25 : 1) at reflux set off a cascade of reactions that culminated in the formation of a mixture of 227 and its oxygen-bridged isomer 228 in yields of 56% and 14%, respectively, based on 225. Compound 227 could be partially converted into 228 by further heating, and recovered 227 could be recycled. Finally, standard transformations on both products completed the first reported total syntheses of (–)-fumiquinazolines E 229 and C 230, respectively. A similar set of reactions on the appropriate analogue of 224, designed to produce (–)-fumiquinazoline H 231, was less easily accomplished, and required replacement of the Cbz protecting group by Fmoc in the benzoxazine equivalent of 225 before satisfactory cyclisation could be effected.

	Scheme 13 Reagents: i, (R)-FmocNHCH(CH2SePh)CO2H, EDAC, MeCN; ii, PPh3, Br2, EtNPri2; iii, piperidine (10 equiv.), EtOAc, rt, 10 min; iv, MeCN–HOAc (25 : 1), reflux, 2 h; v, MeCN–HOAc (100 : 1), reflux, 2 h; vi, HCl (0.2 M), MeOH, 25 °C; vii, H2(1 atm), Pd/C, 30 min; viii, H2(4 atm), Pd/C, 30 h.

Known acridone alkaloids isolated from new sources during the period covered by this review include 1-hydroxy-3-methoxy-N-methylacridone 232 from Feronia limonia (= F. elephantum),¹¹⁴ 5-hydroxynoracronycine 233 from Swinglea citrumelo,¹¹⁵ and citrusinine-II 234, glycocitrine-IV 235 and 5-hydroxynoracronycine from S. glutinosa.¹¹⁶ The last-named species was also the source of a novel alkaloid, 1,3,5-trihydroxy-2,8-diprenyl-N-methylacridone 236, which is one of the very few acridone alkaloids to bear a substituent at C-8, and the only one to have a prenyl substituent at this position.

The biological activities of acridone alkaloids continue to attract interest. Citrusinine-II 234, citrusinine-I 237, severifoline 238 and buxifoliadine-H 239, isolated from the root bark of Severinia buxifolia, were cytotoxic towards nasopharyngeal carcinoma KB cells (ED₅₀ 0.09–0.82 µg ml^–1); 237 and 239 were also marginally active against hepatoma 3B cells (ED₅₀ 6.6 and 5.2 µg ml^–1), whereas 237, buxifoliadine-B 240 and buxifoliadine-D 241 were cytotoxic towards colon carcinoma colo-205 cells (ED₅₀ 0.58–6.3 µg ml^–1).²⁸ The four alkaloids from S. glutinosa mentioned in the previous paragraph, and glycocitrine-IV in particular, were cytotoxic towards human fibroblast (HeLa) cells, and also showed antiplasmodial activity against both chloroquine-sensitive and -resistant strains of Plasmodium falciparum.¹¹⁶ Three acridones from the African medicinal plant Fagara macrophylla, viz. 1-hydroxy-3-methoxy-N-methylacridone 232, arborinine 242 and xanthoxoline 243, showed potent antifeedant activity against final stage larvae of the lepidopteran pest Spodoptera frugiperda, but only xanthoxoline was effective against S. littoralis.¹¹⁷

Acronycine 244 is a promising lead compound for the development of novel antitumour agents. Recent progress in the synthesis and evaluation of novel acronycine analogues has been summarised in two timely reviews, both of which highlight the unique pharmacological profile of the derivative S23906-1 245 when tested on aggressive human lung, ovarian and colon cancer xenografts in nude mice.^118,119 Details of the experimental testing of this potent new anticancer agent and its probable mode of action were described in two full publications.^120,121

1-Oxo-2-hydroxy-1,2-dihydroacronycine 246, easily prepared by permanganate oxidation of acronycine, has proved to be a useful precursor for preparing biologically active nitrogen-containing derivatives both by replacement of the methoxy group and by further manipulation of the pyran ring; it has also been transformed into the cytotoxic furanone derivative 247 by treatment with sodium hydroxide in methanol.¹²² Isoacronycine 248, synthesised from 1,3-dihydroxy-N-methylacridone and prenal, has itself been transformed into various derivatives by functionalisation of the double bond in the pyran ring, but only 249 and 250 were appreciably cytotoxic towards L-1210 leukaemia cells.¹²³ Further synthetic analogues of acronycine to exhibit promising cytotoxicity included benzophenanthroline derivatives such as 251 and 252,¹²⁴ while some more exotic pyrazole-fused systems of general structure 253 were substantially more active towards L-1210 cells than the parent alkaloid, and even inhibited the proliferation of several human solid tumour cells.¹²⁵