The guanidine metabolites of Ptilocaulis spiculifer and related compounds; isolation and synthesis

Abstract

Marine natural products possessing guanidine functionalities display a considerable array of biological activity and not surprisingly have attracted considerable synthetic interest. This review discusses the isolation of several guanidine containing metabolites, primarily from the sponge Ptilocaulis spiculifer, but also from other marine organisms. It also explores the synthetic methodologies adopted for their preparation and speculates on the structural similarity of the metabolite ptilomycalin A to abiotic guanidine based anionic receptor molecules.

---

Laura Heys

Laura Heys graduated (MChem) from Bangor University in 1999, she is currently pursuing a career in the chemical industry.

Christopher G. Moore

Christopher G. Moore graduated (BSc Hons) from Bangor in 1995 and is currently finishing his postgraduate studies which have focused on the total synthesis of guanidine containing marine metabolites.

Patrick J. Murphy

Patrick J. Murphy graduated (BSc Hons) from the University of Manchester Institute of Science and Technology in 1983 and went on to complete a PhD entitled ‘Synthetic routes towards naturally occurring thiotetronic acids’ in 1986 under the supervision of Dr John Brennan at the same institute. After a period of postdoctoral research at the University of Salford under the supervision of Professor Garry Procter (organosilicon chemistry) and Dr Mike Casey (reactions of chiral sulfoxides), he was appointed to the post of Lecturer in Organic Chemistry at the University of Wales, Bangor, gaining promotion to Senior Lecturer in 1999. Since gaining this post he has concentrated his research efforts on the development of new methodology and its applications in synthesis, with two key themes being the synthesis of naturally occurring guanidine containing marine alkaloids and new applications of Wittig chemistry.

---

1 Isolation of ptilomycalin A and related metabolites

Naturally occurring and synthetic compounds containing guanidine are of considerable interest due both to the hydrogen-bond mediated interaction of guanidinium ions and because of the wide range of biological activities these substances display.¹ Recent interest in marine natural products has seen a steady increase in the number of metabolites isolated from these sources and a metabolite of particular interest isolated from the sponge Ptilocaulis spiculifer² in 1989 is ptilomycalin A 1. This alkaloid has a unique and fascinating structure consisting of a pentacyclic guanidinium core linked to a spermidine unit via a ω-hydroxy acid spacer group. The same compound was also isolated from a Red Sea sponge of the genus Hemimycale sp.³ and subsequently from the sponge Batzella sp.⁴ and the starfishes Fromia monilis and Celerina heffernani⁵ (Scheme 1).

Scheme 1

Ptilomycalin A displays a remarkable range of biological activities including cytotoxicity against the following cell lines; P388 (IC₅₀ 0.1 μg mL⁻¹), L1210 (IC₅₀ 0.4 μg mL⁻¹), and KB (IC₅₀ 1.3 μg mL⁻¹) in addition to antifungal activity against Candida albicans (MIC 0.8 μg mL⁻¹) as well as very good antiviral activity (HSV) at a concentration of 0.2 μg mL⁻¹.²

Previous to the isolation of ptilomycalin A the related guanidinium alkaloids ptilocaulin 2 and isoptilocaulin 3 had been isolated from the same sponge in 1981⁶ (Scheme 2).

Scheme 2

Several metabolites related to ptilomycalin A have subsequently been isolated including the crambescidins 800, 816, 830 and 844 (4–7) from the Mediterranean sponge Crambe crambe⁷ and in the case of 4 also from the sponge Monanchora arbuscula.⁸ The structures of these closely related guanidines were elucidated in 1991 and differ from ptilomycalin A by the presence of a hydroxyspermidine instead of a spermidine residue, the presence of an extra hydroxy-substituent on C-13 (in the pyrrolidine ring) in the case of crambescidins 816, 830 and 844 and in the variation in length of the ω-hydroxy fatty acid. These polycyclic guanidines were reported to exhibit antiviral activity against Herpes simplex virus type 1 (HSV-1) and they were cytotoxic to L1210 murine leukaemia cells⁷ and crambescidin 816 has been shown to be a potent calcium channel blocker.⁹ In addition two related guanidinium alkaloids were isolated from the New Caledonian starfishes Celerina heffernani and Fromia monilis, and named celeromycalin 8, which bears an extra hydroxy function on the ω-hydroxy fatty acid and fromiamycalin 9 which contained a modified spermidine unit (Scheme 3).

Scheme 3

Interestingly the polar extracts of the starfish Fromia monilis also contained ptilomycalin A and crambescidin 800, which allows us to speculate that these metabolites may have arisen from a symbiotic organism common to both species. Both celeromycalin and fromiamycalin showed antiviral activities against the Herpes simplex virus (HSV-1) and were also found to be highly cytotoxic in anti-HIV assays on CEM 4 cells infected with human immunodeficiency virus (HIV-1) with CC-50 (cytotoxic concentration) of 0.11 μg mL⁻¹ (ptilomycalin A, crambescidin 800 and fromiamycalin) and 0.32 μg mL⁻¹ (celeromycalin).⁵

Possibly one of the most interesting of these metabolites to be isolated is 13,14,15-isocrambescidin 800 10, the structure of which was shown to contain a trans-ring junction in the guanidine pentacycle.^3,10 Surprisingly this molecule is substantially less cytotoxic to L1210 cells than the other crambescidins, and has no observed antiviral activity against HSV-1. This observation might suggest that the enclosed ionic pocket found in ptilomycalin A and the crambescidins and lacking in 10 might be conferring much of the biological activity found in these compounds (Scheme 4).

Scheme 4

In addition, a series of related guanidine alkaloids, the crambescins A, B, C1 and C2 11–14, were isolated from the same sponge Crambe crambe and displayed cytotoxicity against L1210 cells (IC₅₀ of less than 1 μg mL⁻¹) and were found to be ichthyotoxic (toxic against fish)¹¹ (Scheme 5).

Scheme 5

The batzelladines A–E 15–19, a series of guanidine alkaloids containing a tricyclic guanidine unit, were isolated from the Caribbean sponge Batzella sp. which was also found to contain the previously isolated metabolites ptilomycalin A, ptilocaulin, crambescin A, crambescidin 800 and 816.^4,12 Batzelladines A and B were reported to inhibit the binding of HIV-gp120 to the CD4 cell-surface receptor protein on T cells and to be of interest in the treatment of AIDS.⁴ The structures of the batzelladines A–E were originally elucidated by interpretation of the spectral data of the natural material and of the methanolysis products. The original stereochemical assignment of the alkaloids had been presumed to be one in which the methine hydrogens in the pyrrolidine ring, and those adjacent to the guanidine were all cis-, which was based on a previous literature precedent.² Subsequent work by Snider¹³ has shown that in the cases of batzelladine A and D the original assignment was incorrect and the relative stereochemistry of these metabolites was in fact trans. Both batzelladine A and B contain two guanidinium ring systems connected by a linear hydrocarbon chain, with the left-hand bicyclic guanidine being similar in nature to crambescidin A 11. It is interesting to note that the carbon skeleton present in the tricyclic guanidine nucleus of batzelladine E, together with the alkene geometry is identical to that found in ptilomycalin A 1 (Scheme 6).

Scheme 6

Following the isolation of these alkaloids, a further four batzelladine metabolites, termed batzelladine F–I 20–23, were obtained from the same source.¹⁴ These were interesting in so much as they all contained two tricyclic guanidine units one of which lacks an ester function, together with varying degrees of unsaturation and the presence of N-hydroxylation in all but 20. The relative stereochemical assignment of these metabolites was made by comparison with batzelladines A–E,^4,13 however synthetic work¹⁵ has shown that the original assignment of the left-hand tricycle of batzelladine F is possibly incorrect and that the stereochemistry is in fact an all cis-arrangement of the methine hydrogens (Scheme 7).

Scheme 7

As can be seen, these polycyclic guanidine alkaloids isolated from marine sponges possess a wide range of biological activities as well as intriguing molecular structures. It is thus not surprising that they have attracted considerable interest from several synthetic research groups and have also been the topic of some speculation as to the exact biological role and mechanism of action of these metabolites. The remainder of this review will focus on the synthetic routes applied to these metabolites and will comment on work relating to the biological aspects previously mentioned.

2 Synthetic contributions from the Snider group

The Snider research group has made several significant contributions to the synthesis of polycyclic guanidine alkaloids beginning with a synthesis, in racemic form, of the metabolite ptilocaulin 2 and the confirmation of its absolute stereo-chemistry by the synthesis of the unnatural enantiomer 25via addition of guanidine to the enone 24¹⁶ (Scheme 8).

Scheme 8 Reagents and conditions: (a) Guanidine, PhH, reflux 25 h, then HNO₃ (1% aq); 40%.

A conceptually similar strategy has been adopted by Snider in his synthetic approach to the crambescins, ptilomycalin A and the batzelladine alkaloids. The first report in this area detailed the addition of O-methylisourea to the enones 26 leading to the formation of the dihydropyrimidines 27 after desilylation, which were then converted into the guanidines 28 by treatment with methanolic ammonia. These key intermediates were then converted into crambescins C1 13 and C2 14 by hydrogenolysis, crambescin A 11 by mesylation, cyclisation and hydrogenolysis and to crambescin B 12 by cyclisation under basic conditions and hydrogenolysis¹⁷ (Scheme 9).

Scheme 9 Reagents and conditions: (a) 2 equiv. O-methylisoureido sulfate, NaHCO₃, DMF, 55 °C, 3 h, 94%; (b) HF, CH₃CN, rt, 1.5 h, 88%; (c) NH₃, NH₄OH, t-BuOH, MeOH, 60 °C, 3 days, (94% and 81% loss of one of the two Z-groups is observed); (d) H₂/Pd/C, HCl, CHCl₃ (88% and 91%); (e) MsCl, Et₃N, CH₂Cl₂, then Et₃N, CHCl₃, Δ, 24 h, (78%) then H₂/Pd/C, HCl, CHCl₃, 3 h (93%); (f) Et₃N, Δ, CHCl₃, 12 h, then H₂/Pd/C, HCl, CHCl₃, rt, 3 h, (89% and 91%).

This work set the scene for a similar synthetic approach to ptilomycalin A and preliminary model studies¹⁸ on the bis-enone 31 prepared by Knoevenagel condensation of β-ketoester 29 and aldehyde 30, illustrated that the addition of O-methylisourea was a feasible process leading to the bicyclic compound 32 as a 3∶1 mixture of trans- and cis-isomers. This reaction appears to lead to the incorrect stereochemistry at the pyrrolidine ring junction, however on reaction with methanolic ammonia, which generates the guanidine functionality, trans-32 undergoes isomerisation, presumably by an elimination and readdition process, with concomitant cyclisation to form the tricyclic model 33 of ptilomycalin A (Scheme 10).

Scheme 10 Reagents and conditions: (a) Piperidine, CH₂Cl₂, −20 °C, 2 days, 61%; (b) 2 equiv. O-methylisoureido sulfate, NaHCO₃, DMF, 50 °C, 2 h, 56%; (c) NH₃, NH₄OAc, MeOH, 60 °C, 4 days, 60%.

Following on from these preliminary studies Snider reported a synthesis of the methyl ester of the pentacyclic nucleus of ptilomycalin A 38via a convergent and biomimetic 14-step route.¹⁹ The key steps of this route involved the conversion of the bis-enone 34 to 38 in four steps. The addition of O-methylisourea to 34 gave a 1∶4 mixture of the two cis diastereomers 35a and the two trans diastereomers 35b in 52% yield. Ammonolysis converted this mixture of isoureas into a 1:1 mixture of the two cis diastereomeric guanidines 36a (cis, β) and 36b (cis, α) in 72% yield. Deprotection of the silyl ethers was accomplished by reaction with a 3:7 mixture of 50% aqueous HF and acetonitrile. Treatment of the resulting crude mixture (containing intermediate 37) with triethylamine in methanol gave a 60% yield of an approximately 65% pure 1.3∶1 mixture of 38 and the diastereomer 39 with an equatorial methyl ester. The remaining material was thought to consist of tri- and tetracyclic compounds from the undesired diastereomer 36b. Flash chromatography succeeded in separating 38 and 39 but gave only 80–85% pure material. Purification of this mixture, therefore, was best accomplished by treating the 1.3∶1 mixture of 38 and 39 with triethylamine in 1∶1 MeOH–H₂O to give tetracyclic alcohol 37. This was purified by flash chromatography and recycled with triethylamine in methanol to again give a 1.3∶1 mixture of 38 and 39, which were separated to give pure 38 (34% from 36a) and 39 (26% from 36a) (Scheme 11).

Scheme 11 Reagents and conditions: (a) O-methylisoureido sulfate, i-Pr₂EtN, DMSO, 80 °C, 1.5 h, 52%; (b) NH₃, NH₄OAc, t-BuOH, 60 °C, 40 h, 72%; (c) 3∶7 HF–CH₃CN, −30 °C, 3 days; (d) Et₃N, MeOH, 60 °C, 20 h (78%); (e) Et₃N, MeOH–H₂O (1∶1), 60 °C, 16 h, then (d), 38; 34% and 39; 26%.

Snider has also reported the first unequivocal synthesis (in racemic form) of one of the batzelladine alkaloids, batzelladine E, and in doing so was able to correct the structural misassignment of this metabolite which was initially given an E-stereochemistry in the unsaturated side chain.^13b The synthesis begins with the aldehyde 40 which was condensed with β-ketoester 41 under Knoevenagel conditions to give the enone 42. Reaction of this enone with O-methylisourea followed by ammonolysis gave the product 43 which on careful reduction led to the tricyclic guanidine 44. Deprotection of the amine function was followed by conversion to the terminal guanidine group, thus completing a synthesis of batzelladine E in 9 steps and in 3% overall yield from a commercially available starting material (Scheme 12).

Scheme 12 Reagents and conditions: (a) 0.33 equiv. piperidine, 0.30 equiv. AcOH, CH₂Cl₂, −20 °C, 2 days; (b) O-methylisourea, i-Pr₂EtN, DMSO, 55 °C, 4 h; (c) NH₃, NH₄OAc, t-BuOH, 60 °C, 1 day (14% from 40); (d) NaCNBH₃, NaH₂PO₄, MeOH, 25 °C, 88%; (e) 1:4 TFA–CH₂Cl₂, 25 °C, 5 min, 93%; (f) (BocNH)₂C[double bond, length half m-dash]S, 2-chloro-N-methylpyridinium chloride, CH₂Cl₂, NEt₃, 25 °C, 1 h, 64%; (g) 1∶1 TFA–CH₂Cl₂, 25 °C, 2 h, 90%.

As mentioned previously, Snider had also reported^13a the reassignment of the relative stereochemistry of batzelladines A 15 and D 18. This was accomplished by reaction of enone 45 with O-methylisourea under standard conditions leading to a 35% yield of a 6∶1 mixture of the dihydropyrimidine 46 and its cis-isomer. At this point in the synthesis the ketone function was ‘protected’ by reduction to the alcohol 47, which also serves to lock the pyrrolidine in a cis-configuration, which was then converted to the guanidine 48. Regeneration of the ketone function led to the formation of the tricyclic guanidine 49 which was reduced and saponified to give the carboxylic acid 50, a known degradation product of batzelladines A and D, thus leading to the conclusion that the stereochemistry in these metabolites was indeed trans- across the pyrrolidine ring system (Scheme 13).

Scheme 13 Reagents and conditions: (a) O-methylisourea hydrogen sulfate, i-Pr₂EtN, DMSO, 75 °C, 5 h; (b) NaBH₄, i-PrOH, 25 °C; (c) NH₃, NH₄OAc, MeOH, 60 °C, 2 days; (d) Dess–Martin, CH₂Cl₂, 25 °C, then MeOH, 25 °C, 12 h; (e) NaCNBH₃, NaH₂PO₄, MeOH, 25 °C, 16 h then 65 °C, 5 h; (f) NaOH, MeOH 25 °C, 18 h.

3 Synthetic contributions from the Overman group

The Overman research group has also been actively pursuing the synthesis of these polycyclic guanidine alkaloids, indeed, they have reported the first total synthesis of ptilomycalin A 1 using a tethered intramolecular Biginelli reaction.^20,21 Preliminary studies on this reaction,²⁰ which classically involves the condensation of an aldehyde, a β-ketoester and a urea to give a 3,4-dihydro-1H-pyrimidin-2-one, had indicated its potential as a synthetic methodology. A subsequent report²¹ illustrated that reaction of 51, which represents the urea and masked aldehyde portions of the reaction, with the β-ketoester 52 in the presence of morpholinium acetate (typical Knoevenagel conditions), led to the formation of the bicyclic pyrimidin-2-one 53. This product was cyclised to give the lower half of the pentacyclic guanidine core 54 by treatment under acidic conditions and converted to the aldehyde 55via Swern oxidation and O-methylation with MeOTf. These two intermediates differ from the corresponding portion in ptilomycalin A only by the stereochemistry at the ester position. Introduction of the upper half of the pentacycle was effected by reaction with the Grignard reagent 58 followed by Swern oxidation to the ketone 56; this was followed by deprotection of the silyl protecting group and treatment with a mixture of ammonia and ammonium acetate to give 57. Routine modification of the ester function to a protected spermidine unit was followed by epimerisation of the ester stereocentre by treatment with triethylamine and finally deprotection, leading to the first total synthesis of ptilomycalin A 1 (Scheme 14).

Scheme 14 Reagents and conditions: (a) Morpholine, AcOH, EtOH, Na₂SO₄, 70 °C, 61%; (b) PPTS, MeOH, 50 °C, then p-TosOH, CHCl₃, 23 °C, 96%; (c) Swern oxidation; (d) MeOTf, 2,6-di-tert-butylpyridine, CH₂Cl₂, 23 °C, 67%; (e) (S)-Z-EtCH(OTIPS)CH[double bond, length half m-dash]CHCH₂CH₂ MgBr (58), THF, −78 °C; (f) Swern oxidation, 58% (2 steps); (g) TBAF; (h) NH₃, NH₄OAc, t-BuOH, 60 °C, 51% (2 steps); (i) Pd(Ph₃P)₄, pyrrolidine, MeCN, 23 °C, 75%; (j) BocHN(CH₂)₃NH(CH₂)₄NHBoc, EDCl, DMAP, CH₂Cl₂, 23 °C, 60%; (k) NEt₃, MeOH, 65 °C, 50%; (l) HCO₂H, 23 °C, 100%. E = CO₂(CH₂)₁₅CO₂Allyl.

Overman has also applied the tethered intramolecular Biginelli reaction to the synthesis of tricyclic systems similar to those observed in the batzelladines and has reported that reaction of bicyclic guanidine 59, prepared in 32% yield from nonan-2-one, with methyl acetoacetate led to the formation of the tricycle 60 as a 10∶1 mixture of cis- and trans-isomers in 94% yield.^13c This structure is identical to a reported⁴ degradation product of batzelladine B in all respects, including the sign of rotation, thus confirming the absolute stereochemistry of the batzelladines to be that illustrated (Scheme 15).

Scheme 15 Reagents and conditions: (a) Methyl acetoacetate, morpholinium acetate (1 equiv.), CF₃CH₂OH, Na₂SO₄, 90 °C, 36 h, 94%.

Interestingly Overman has also reported that by judicious choice of reaction conditions it is possible to tune the selectivity of this reaction to obtain either the cis- or trans-pyrrolidine geometry.²² Thus treatment of either the urea derivative 61 (X = O, R = H or Bn) with benzyl acetoacetate under Knoevenagel conditions leads to the formation of cis-62 and trans-62 (X = O, R = H or Bn) in a 4∶1 ratio (80 and 81% yield respectively). However on reaction of 61 (X = O, R = Bn) and benzyl acetoacetate with the mild dehydrating agent polyphosphate ester (PPE) the stereoselectivity was reversed leading to a 4∶1 ratio of products with the trans-isomer predominating. Interestingly, treatment of the guanidine 61 (X = NH·HCl, R = OH) under identical Knoevenagel conditions led to the formation of trans- 62 (X = NH·HCl, R = OH) exclusively in 42% yield, which is in complete contrast to the stereoselection observed with ureas. Further work with the N-sulfonylguanidine 61 (X = MtrN, R = H, Bn) led to the formation of cis-62 and trans-62 in a 6–7∶1 ratio (X = MtrN, R = H, Bn; 61 and 84% yield respectively) under Knoevenagel conditions, but again reaction of 61 (X = MtrN, R = Bn) with PPE led to a reversal of selectivity with the trans-62 (X = MtrN, R = Bn) predominating in a 20∶1 ratio (61% yield) (Scheme 16).

Scheme 16 Reagents and conditions: (a) Benzyl acetoacetate (1.5 equiv.), morpholinium acetate (1.5 equiv), CF₃CH₂OH, Na₂SO₄, 60 °C, 48 h; (b) PPE, CH₂Cl₂, 23 °C, 48 h.

4 Synthetic contributions from the Murphy group

Within our own research group we have been pursuing the synthesis of both ptilomycalin A and the batzelladine alkaloids utilising a strategy based upon the double 1,4-addition of guanidine to bis-enones. Preliminary studies²³ illustrated that the formation of the pentacyclic guanidines 66 and 69 was possible using this strategy. Synthesis of the substrate for these model studies was straightforward in that reaction of lactones 63 with two equivalents of methylenetriphenylphosphorane followed by silyl protection of the intermediate phosphonium alkoxide gave the phosphoranes 64. Wittig reaction of 64a with 0.4 equivalents of succinaldehyde gave the symmetrical bis-enone 65 in 54% yield. Further reaction of this with one equivalent of guanidine, followed by removal of solvent, deprotection/cyclisation with methanolic HCl and counter ion exchange, afforded two products identified as the cis-product 66 and the corresponding trans-pentacycle in an approximate 4∶1 ratio, from which the major product could be isolated in 25% yield by recrystallisation.

Similarly reaction of the phosphorane 64a with an excess of succinaldehyde led to the formation of the aldehyde 67 in 43% yield from the lactone starting material; reaction of this with phosphorane 64b gave the unsymmetrical bis-enone 68. Reaction of 68 with guanidine under identical conditions to those previously employed, led to the formation of two pentacyclic guanidines in an approximate 4∶1 ratio and from which the major isomer 69 could be obtained by crystallisation in 20% yield (Scheme 17).

Scheme 17 Reagents and conditions: (a) 2 equiv. CH₂[double bond, length half m-dash]PPh₃, THF, −78 °C; (b) TBDMSCl, imidazole, DMF; (c) 0.4 equiv. succinaldehyde, THF, 48 h; 54% overall; (d) (i) guanidine, DMF, 3 h, (ii) MeOH, HCl, 0 °C–rt, 24 h; (iii) aq. NaBF₄ (sat.), (iv) trituration and crystallisation; 25% overall; (e) steps (a), (b) then 10 equiv. succinaldehyde, THF, 43%; (f) 64b, THF, 48 h; 37%; (g) as (d) 20%.

Following on from this work we also reported that the addition of guanidine to a series of bis-enones 70 followed by reduction with sodium borohydride led to the formation of the tricyclic guanidines 71, which, before the structural reassignments reported by Snider,¹³ were thought to be models of the saturated batzelladine alkaloids A and D²⁴ (Scheme 18).

Scheme 18 Reagents and conditions: (a) (i) Guanidine, DMF, 0 °C–rt, 5–8 h, (ii) 3∶1∶3 DMF–H₂O–MeOH, then NaBH₄, 16 h, (iii) HCl (aq), (iv) aq. NaBF₄ (sat.). R = Me, C₅H₁₁, C₉H₁₉, Ph, (22–33% yield).

With the isolation of batzelladine F, it became apparent that there was a strong correlation in structure between the left-hand portion of this molecule and the model compounds previously prepared by us;^15a we thus embarked upon a synthesis of the left-hand sub unit of this metabolite employing this methodology.^15b Reaction of the iodide 72 with the anion generated from deprotonation of phosphorane 73 led to the phosphorane 74, which on reaction with excess succinaldehyde gave the aldehyde 75. Subsequent reaction of this with a further equivalent of phosphorane 73 led to the formation of bis-enone 76, which on reaction with guanidine under standard conditions gave the tricycle 77. Comparison of the spectral data of 77 or the corresponding deprotected analogue 78 or acetate 79, illustrated a strong correlation between these materials and the naturally occurring batzelladine F thus leading to the conclusion that the reported stereochemistry is incorrect and is actually as illustrated²⁵ (Scheme 19).

Scheme 19 Reagents and conditions: (a) CH₃COCHPPh₃ (73), n-BuLi, THF, −78 °C–rt; (b) succinaldehyde, CH₂Cl₂, 24 h, 54% for 2 steps; (c) 73, CH₂Cl₂, 24 h, 91%; (d) (i) guanidine, DMF, 0 °C, 5 h; (ii) 3∶1∶3 DMF–H₂O–MeOH, then NaBH₄, 16 h; (iii) HCl (aq), (iv) aq. NaBF₄ (sat); 77, 29% overall; (e) (i) MeOH–HCl, (ii) aq. NaBF₄ (sat.), 78, 91%; (f) Ac₂O–Py, then HCl (2 M), 41%; (g) aq. NaBF₄ (sat), 79,100%.

5 Synthetic contributions from the Hiemstra group

A recent publication from the Hiemstra group describes the preparation of tricyclic guanidines from substituted pyrrolidin-2-ones utilising an N-acyliminium ion coupling reaction with silyl enol ethers and a direct guanylation with bis-Boc-thiourea and mercury(II) chloride.²⁶ They reported that reaction of the silyl enol ether 80 with lactam 81 led to the formation of the substituted lactam 82 in 63% yield. After conversion of 82 into the corresponding thiolactam, an Eschenmoser sulfide contraction procedure led to the formation of the vinylogous amide 83. Reduction and N-Boc protection of 83 gave the substrate 84 which was subjected to a three stage procedure which included protected guanidine 85 as an intermediate. This was cyclised under acidic conditions to give the tricyclic guanidine 86 in 33% overall yield, together with several other trans-substituted guanidines which could be recycled to give further 86 by treatment with ammonia and ammonium acetate in methanol at 60 °C²⁶ (Scheme 20).

Scheme 20 Reagents and conditions: (a) TMSOTf, −78 °C–rt, CH₂Cl₂, 18 h, 63%; (b) Lawesson’s reagent, PhCH₃, 80 °C, 10 min, 91%; (c) PhCOCH₂Br, Et₂O, rt, 18 h; (d) Et₃N, CH₂Cl₂, rt, 2 h, 83%; (e) PPh₃, CHCl₃, 60 °C, 18 h, 82%; (f) NaBH₃CN, 3∶1 AcOH–THF, 0 °C, 40 min, 99%; (g) Boc₂O, DIPEA, THF, rt, 18 h, 91%; (h) PCC, CH₂Cl₂, mol. sieves (4 Å), rt, 3 h, 91%; (i) CH(OMe)₃, H₂SO₄ (cat), MeOH, 50 °C, 5 h; (j) SC(NHBoc)₂, HgCl₂, Et₃N, DMF, 0 °C–rt, 18 h; X = O, (OMe)₂; (k) HCl, MeOH, rt, 3 h, 33% for three steps.

This synthesis is obviously applicable to the preparation of both ptilomycalin A 1 and the batzelladine alkaloids. Indeed Hiemstra has further reported²⁷ that bis-acetoxylactam 87 reacts with the silyl enol ether 88 in a stereocontrolled fashion to give 89, which is a potential precursor of 1 (Scheme 21).

Scheme 21 Reagents and conditions: (a) TMSOTf, −78 °C–rt, CH₂Cl₂, DIPEA, 1 h, 75%.

6 Synthetic contributions from the Rama Rao group

The Rama Rao research group has also reported²⁸ an enantiospecific synthesis of the tricyclic guanidine portion of batzelladine A, which was based upon the original stereochemical assignment made by Patil,⁴ and thus targeted an all cis-arrangement of hydrogens in the final product. In common with Hiemstra, the synthesis proceeds via a lactam, 91 prepared in eight steps from the azetidinone derivative 90, and introduction of the side chain was effected using an Eschenmoser sulfide contraction reaction, leading to the α,β-unsaturated ketone 92. After reduction of the alkene and ketone functions and N-Boc protection of the pyrrolidine nitrogen, the introduction of the remaining nitrogen containing groups was accomplished by sequential nucleophilic displacement of the secondary alcohols present on each side chain with azide (under Mitsunobu conditions), reduction and N-boc protection, ultimately leading to azide 93. This was converted to the cyclic urea 94 using 1,1′-carbonyldiimidazole, which in turn was cyclised by treatment with dimethylsulfate and hydrogenation to give the guanidine 95, after removal of the silyl protecting group (Scheme 22).

Scheme 22 Reagents and conditions: (a) C₉H₁₉COCH₂Br, CH₂Cl₂, rt, 30 min KHCO₃; (b) PPh₃, t-BuOK, t-BuOH, C₆H₆, heat, 65%; (c) TFA, CH₂Cl₂, 0 °C–rt, 30 min; (d) (imid)₂C[double bond, length half m-dash]O, THF, 0 °C–rt, 65%; (e) Me₂SO₄, C₆H₆, heat, 16 h; (f) H₂, Pd/BaSO₄, MeOH, 12 h, 65%; (g) HCl (1 M), MeOH, 50 °C, 2 h, 90%.

7 Ptilomycalin A as a potential host molecule

One important feature of the chemistry of guanidinium compounds is their ability to interact with anionic species, indeed, this is a key feature in the function of the amino acid arginine whose guanidine group serves as a key binding site for carboxylate and phosphate containing substrates in a wide range of biological applications.²⁹ It is apparent that ptilomycalin A 1 has an extraordinary molecular architecture in which the pentacyclic core (Fig. 1) can be considered to have a ‘cage’ like structure capable of binding strongly to an anion and possibly encapsulating it, indeed, Kashman referred to this core as the ‘vessel’ part of the molecule linked via the aliphatic ‘chain’ to a spermidine ‘anchor’.^2b In fact, one interesting feature of this metabolite is its structural similarity to abiotic guanidine based anionic receptor molecules.³⁰ Based on this hypothesis, Kashman^2b investigated the complexing ability of a trifluoroacetate derivative of 1 with various organic carboxylates and determined a scale of binding ability for N-acetylamino acids which was estimated to be as follows: L-N-acetylmethionate ≈ L-N-acetylvalinate > L-N-acetylalanate ≈ L-N-acetylisoleucinate >> L-N-acetylglycinate. In addition, there is evidence suggesting that the spermidine unit may also be involved in the binding of anionic species to ptilomycalin A.³¹ It is also interesting to note that ptilomycalin A, despite containing several polar functional groups, is a relatively non-polar molecule which is freely soluble in organic solvents such as chloroform.^2a These properties suggest an anionic binding capability possibly linked to strong lipophillic behaviour.²¹

Fig. 1

It has been speculated³² that the enclosed ionic ‘cage’ at the central guanidine sub-unit might indeed be acting as a specific anion recognition site thereby conferring much of the biological activity found in these compounds. In relation to this, it is interesting to note that the subsequently isolated 13,14,15-isocrambescidin 800 10 is substantially less cytotoxic to L1210 cells than other crambescidins and has no observed antiviral activity against HSV-1;¹⁰ this reduced activity may be due to the lack of this structural feature (Fig. 2).

Fig. 2

We have reported³² an interesting behaviour pattern in the interaction between the guanidinium and fluoroborate ions in the model compounds 96, 66 and 97. Fluoroborate can undergo a similar interaction with a guanidinium ion to that of the bidentate ligating interaction that is observed with a carboxylate or a phosphate. It was thus surprising to find that in compound 96 only one of the fluorine atoms of the fluoroborate anion was involved in a strong hydrogen bonding interaction with the guanidinium cation (to both N–H bonds). This observation led us to suppose that the guanidinium cavity of 96 was not of sufficient size to accommodate the fluoroborate anion. This behaviour was however not observed in the pentacyclic 6,6,5,6,6 model compound 66 which corresponds more closely to the structure of ptilomycalin A. In this case, the fluoroborate anion was involved in two separate non-symmetrical hydrogen bonding interactions, but was unable to achieve co-planarity with the guanidinium ion. However, near co-planarity was achieved in the case of the pentacyclic 7,6,5,6,7 model 97 which undergoes an identical hydrogen bonding pattern to 66 (Scheme 23). These observations offer some support to the theory of the ‘cage’ portion of ptilomycalin A being involved in a recognition process and that the structure of 1 represents an optimum host design for an as yet undetermined guest molecule.

Scheme 23

A further report from Hart and Grillot³³ describes the synthesis of an analogue 98 (Scheme 24), of ptilomycalin A 1, which models the guanidine core, chain and spermidine sub-units of the natural material and was prepared in a 12 step convergent sequence from acyclic precursors via a cyclic thiourea derivative. This molecule was prepared in an attempt to mimic the biological activity found in the natural system, however, it was found to be unstable and underwent an unidentified decomposition process over a period of a few weeks which precluded biological evaluation. It was suggested that this process involved cleavage of the ester linkage and the authors speculated that the role of the spiro-N,O-acetals present in ptilomycalin A 1 might be to protect the ester linkages from hydrolysis or aminolysis.

Scheme 24

8 Conclusion

Ptilomycalin A and related molecules are obviously of considerable interest from both a synthetic and biological perspective. The current state of the synthetic work offers several strategies for the synthesis of these metabolites and analogues thereof, with several total syntheses having been reported. The biological activities of these metabolites ranges over a broad spectrum, however, little is known about the exact mechanism of these activities despite much speculation. We believe that in the near future this area of investigation, together with continued synthetic efforts, will offer a valuable insight into these intriguing metabolites.

References

1. For leading references see: R. G. S. Berlinck, Nat. Prod. Rep., 1995, 377..
2. (a) Y. Kashman, S. Hirsh, O. J. McConnell, I. Ohtani, T. Kusumi and H. J. Kakisawa, J. Am. Chem. Soc., 1989, 111, 8925;; (b) I. Ohtani, T. Kusumi, H. Kakisawa, Y. Kashman and S. Hirsh, J. Am. Chem. Soc., 1992, 114, 8472.
3. R. G. S. Berlinck, J. C. Braekman, D. Daloze, K. Hallenga, R. Ottinger, I. Bruno and R. Riccio, Tetrahedron Lett., 1990, 31, 6531.
4. A. D. Patil, N. V. Kumar, W. C. Kokke, M. F. Bean, A. J. Freyer, C. De Brosse, S. Mai, A. Truneh, D. J. Faulkner, B. Carte, A. L. Breen, R. P. Hertzberg, R. K. Johnson, J. W. Westley and B. C. M. Potts, J. Org. Chem., 1995, 60, 1182.
5. E. Palagiano, S. De Marino, L. Minale, R. Riccio, F. Zollo, M. Iorizzi, J. B. Carré, C. Debitus, L. Lucarain and J. Provost, Tetrahedron, 1995, 51, 3675.
6. G. C. Harbour, A. A. Tymiak, K. L. Rinehart, P. D. Shaw, R. G. Hughes, S. A. Mizsak, J. H. Coats, G. E. Zurenko, L. H. Li and S. L. Kuentzel, J. Am. Chem. Soc., 1981, 103, 5604.
7. E. A. Jares-Erijman, R. Sakai and K. L. Rinehart, J. Org. Chem, 1991, 56, 5712.
8. R. Tavares, D. Daloze, J. C. Braekman, E. Hadju, G. Muricy and R. W. M. van Soest, Biochem. Syst. Ecol., 1994, 22, 645.
9. R. G. S. Berlinck, J. C. Braekman, D. Daloze, I. Bruno, R. Riccio, S. Ferri, S. Spampinato and E. Speroni, J. Nat. Prod., 1993, 56, 1007.
10. E. A. Jares-Erijman, A. L. Ingrum, J. R. Carney, K. L. Rinehart and R. Sakai, J. Org. Chem., 1993, 58, 4805.
11. (a) R. G. S. Berlinck, J. C. Braekman, D. Daloze, I. Bruno, R. Riccio, D. Rogeau and P. Amade, J. Nat. Prod., 1992, 55, 528;; (b) E. A. Jares-Erijman, A. A. Ingrum, F. Sun and K. L. Rinehart, J. Nat. Prod., 1993, 56, 2186.
12. Some confusion arose in the identification of the sponge Batzella sp.; a comparison of a specimen with a known specimen of Ptilocaulis sp. illustrated that the two were in fact the same.⁴.
13. (a) B. B. Snider, J. Chen, A. D. Patil and A. J. Freyer, Tetrahedron Lett., 1996, 37, 6977; (b) the alkene stereochemistry in batzelladine E has also been recently reassigned, see, B. B. Snider and J. Chen, Tetrahedron Lett., 1998, 39, 5697;; (c) the absolute stereochemistry of batzelladine B has been determined by synthesis of a degradation product and is as shown in the scheme, however this is the opposite of the arbitrary stereochemistry chosen by the authors of reference 4; A. S. Franklin, S. K. Ly, G. H. Mackin, L. E. Overman and A. J. Shaka, J. Org. Chem., 1999, 64, 1512..
14. A. D. Patil, A. J. Freyer, P. B. Taylor, B. Carté, G. Zuber, R. K. Johnson and D. J. Faulkner, J. Org. Chem, 1997, 62, 1814.
15. (a) G. P. Black, P. J. Murphy and N. D. A. Walshe, Tetrahedron, 1998, 54, 9481;; (b) G. P. Black, P. J. Murphy, A. Thornhill, N. D. A. Walshe and C. Zanetti, Tetrahedron, 1999, 55, 6547.
16. B. B. Snider and W. C. Faith, J. Am. Chem. Soc., 1984, 106, 1443.
17. (a) B. B. Snider and Z. Shi, J. Org. Chem., 1992, 57, 2526;; (b) B. B. Snider and Z. Shi, J. Org. Chem., 1993, 58, 3828.
18. B. B. Snider and Z. Shi, Tetrahedron Lett., 1993, 34, 2099.
19. B. B. Snider and Z. Shi, J. Am. Chem. Soc., 1994, 116, 549.
20. L. E. Overman and M. H. Rabinowitz, J. Org. Chem., 1993, 58, 3235.
21. L. E. Overman, M. H. Rabinowitz and P. A. Renhowe, J. Am. Chem. Soc., 1995, 117, 2657.
22. A. I. McDonald and L. E. Overman, J. Org. Chem., 1999, 64, 1520.
23. (a) P. J. Murphy, H. L. Williams, M. B. Hursthouse and K. M. A. Malik, J. Chem. Soc., Chem. Commun., 1994, 119;; (b) P. J. Murphy and H. L. Williams, J. Chem. Soc., Chem. Commun., 1994, 819;; (c) P. J. Murphy, H. L. Williams, D. E. Hibbs, M. B. Hursthouse and K. M. A. Malik, Tetrahedron, 1996, 52, 8315.
24. G. P. Black, P. J. Murphy, N. D. A. Walshe, D. E. Hibbs, M. B. Hursthouse and K. M. A. Malik, Tetrahedron Lett., 1996, 37, 6943.
25. Comments of this nature are also made in reference 8 of the recent paper by Overman.^13c.
26. S. Louwrier, M. Ostendorf, A. Tuynman and H. Hiemstra, Tetrahedron Lett., 1996, 37, 905.
27. (a) S. Louwrier, M. Ostendorf, A. Boom, H. Hiemstra and W. N. Speckamp, Tetrahedron, 1996, 52, 2603;; (b) S. Louwrier, A. Tuynman and H. Hiemstra, Tetrahedron, 1996, 52, 2629.
28. A. V. Rama Rao, M. K. Gurjar and J. Vasudevan, J. Chem. Soc., Chem. Commun., 1995, 1369.
29. C. L. Hannon and E. V. Anslyn, Bioorg. Chem. Front., 1993, 3, 193.
30. For leading references see: F. P. Schmidtchen and M. Berger, Chem. Rev., 1997, 97, 1609.
31. I. Ohtani, T. Kusumi and H. Kakisawa, Tetrahedron Let., 1992, 33, 2525.
32. P. J. Murphy, H. L. Williams, D. E. Hibbs, M. B. Hursthouse and K. M. A. Malik, Chem. Commun., 1996, 445.
33. (a) A.-L. Grillot and D. J. Hart, Tetrahedron, 1995, 51, 11377;; (b) A.-L. Grillot and D. J. Hart, Heterocycles, 1994, 39, 435.

---